id	sid	tid	token	lemma	pos
ejpam-86	1	1	european	european	PROPN
ejpam-86	1	2	journal	journal	PROPN
ejpam-86	1	3	of	of	ADP
ejpam-86	1	4	pure	pure	ADJ
ejpam-86	1	5	and	and	CCONJ
ejpam-86	1	6	applied	apply	VERB
ejpam-86	1	7	mathematics	mathematic	NOUN
ejpam-86	1	8	vol	vol	NOUN
ejpam-86	1	9	.	.	PROPN
ejpam-86	2	1	1	1	NUM
ejpam-86	2	2	,	,	PUNCT
ejpam-86	2	3	no	no	INTJ
ejpam-86	2	4	.	.	NOUN
ejpam-86	2	5	1	1	NUM
ejpam-86	2	6	,	,	PUNCT
ejpam-86	2	7	2008	2008	NUM
ejpam-86	2	8	,	,	PUNCT
ejpam-86	2	9	(	(	PUNCT
ejpam-86	2	10	4	4	NUM
ejpam-86	2	11	-	-	SYM
ejpam-86	2	12	37	37	NUM
ejpam-86	2	13	)	)	PUNCT
ejpam-86	2	14	issn	issn	PROPN
ejpam-86	2	15	1307	1307	NUM
ejpam-86	2	16	-	-	SYM
ejpam-86	2	17	5543	5543	NUM
ejpam-86	3	1	–	–	PUNCT
ejpam-86	3	2	www.ejpam.com	www.ejpam.com	X
ejpam-86	3	3	honorary	honorary	PROPN
ejpam-86	3	4	invited	invite	VERB
ejpam-86	3	5	paper	paper	NOUN
ejpam-86	3	6	multivariate	multivariate	NOUN
ejpam-86	3	7	regression	regression	NOUN
ejpam-86	3	8	models	model	NOUN
ejpam-86	3	9	with	with	ADP
ejpam-86	3	10	power	power	NOUN
ejpam-86	3	11	exponential	exponential	ADJ
ejpam-86	3	12	random	random	ADJ
ejpam-86	3	13	errors	error	NOUN
ejpam-86	3	14	and	and	CCONJ
ejpam-86	3	15	subset	subset	VERB
ejpam-86	3	16	selection	selection	NOUN
ejpam-86	3	17	using	use	VERB
ejpam-86	3	18	genetic	genetic	ADJ
ejpam-86	3	19	algorithms	algorithm	NOUN
ejpam-86	3	20	with	with	ADP
ejpam-86	3	21	information	information	NOUN
ejpam-86	3	22	complexity	complexity	PROPN
ejpam-86	3	23	minhui	minhui	PROPN
ejpam-86	3	24	liu1	liu1	PROPN
ejpam-86	3	25	,	,	PUNCT
ejpam-86	3	26	hamparsum	hamparsum	VERB
ejpam-86	3	27	bozdogan2,∗	bozdogan2,∗	ADJ
ejpam-86	3	28	1	1	NUM
ejpam-86	3	29	department	department	NOUN
ejpam-86	3	30	of	of	ADP
ejpam-86	3	31	statistics	statistic	NOUN
ejpam-86	3	32	,	,	PUNCT
ejpam-86	3	33	operations	operation	NOUN
ejpam-86	3	34	and	and	CCONJ
ejpam-86	3	35	management	management	NOUN
ejpam-86	3	36	science	science	NOUN
ejpam-86	3	37	,	,	PUNCT
ejpam-86	3	38	university	university	PROPN
ejpam-86	3	39	of	of	ADP
ejpam-86	3	40	tennessee	tennessee	PROPN
ejpam-86	3	41	,	,	PUNCT
ejpam-86	3	42	2	2	NUM
ejpam-86	3	43	stokley	stokley	PROPN
ejpam-86	3	44	management	management	NOUN
ejpam-86	3	45	center	center	NOUN
ejpam-86	3	46	,	,	PUNCT
ejpam-86	3	47	knoxville	knoxville	PROPN
ejpam-86	3	48	tennessee	tennessee	PROPN
ejpam-86	3	49	,	,	PUNCT
ejpam-86	3	50	37996	37996	NUM
ejpam-86	3	51	-	-	SYM
ejpam-86	3	52	0562	0562	NUM
ejpam-86	3	53	u.s.a	u.s.a	NOUN
ejpam-86	3	54	.	.	PUNCT
ejpam-86	3	55	abstract	abstract	PROPN
ejpam-86	3	56	.	.	PUNCT
ejpam-86	4	1	in	in	ADP
ejpam-86	4	2	this	this	DET
ejpam-86	4	3	paper	paper	NOUN
ejpam-86	4	4	we	we	PRON
ejpam-86	4	5	introduce	introduce	VERB
ejpam-86	4	6	and	and	CCONJ
ejpam-86	4	7	develop	develop	VERB
ejpam-86	4	8	two	two	NUM
ejpam-86	4	9	different	different	ADJ
ejpam-86	4	10	novel	novel	NOUN
ejpam-86	4	11	multivariate	multivariate	NOUN
ejpam-86	4	12	regression	regression	NOUN
ejpam-86	4	13	models	model	NOUN
ejpam-86	4	14	with	with	ADP
ejpam-86	4	15	power	power	NOUN
ejpam-86	4	16	exponential	exponential	NOUN
ejpam-86	4	17	(	(	PUNCT
ejpam-86	4	18	pe	pe	ADJ
ejpam-86	4	19	)	)	PUNCT
ejpam-86	4	20	random	random	ADJ
ejpam-86	4	21	errors	error	NOUN
ejpam-86	4	22	for	for	ADP
ejpam-86	4	23	the	the	DET
ejpam-86	4	24	first	first	ADJ
ejpam-86	4	25	time	time	NOUN
ejpam-86	4	26	.	.	PUNCT
ejpam-86	5	1	our	our	PRON
ejpam-86	5	2	first	first	ADJ
ejpam-86	5	3	model	model	NOUN
ejpam-86	5	4	assumes	assume	VERB
ejpam-86	5	5	that	that	SCONJ
ejpam-86	5	6	the	the	DET
ejpam-86	5	7	observations	observation	NOUN
ejpam-86	5	8	are	be	AUX
ejpam-86	5	9	independent	independent	ADJ
ejpam-86	5	10	and	and	CCONJ
ejpam-86	5	11	the	the	DET
ejpam-86	5	12	second	second	ADJ
ejpam-86	5	13	model	model	NOUN
ejpam-86	5	14	assumes	assume	VERB
ejpam-86	5	15	that	that	SCONJ
ejpam-86	5	16	the	the	DET
ejpam-86	5	17	observations	observation	NOUN
ejpam-86	5	18	are	be	AUX
ejpam-86	5	19	dependent	dependent	ADJ
ejpam-86	5	20	.	.	PUNCT
ejpam-86	6	1	these	these	DET
ejpam-86	6	2	two	two	NUM
ejpam-86	6	3	models	model	NOUN
ejpam-86	6	4	coincide	coincide	VERB
ejpam-86	6	5	only	only	ADV
ejpam-86	6	6	when	when	SCONJ
ejpam-86	6	7	the	the	DET
ejpam-86	6	8	shape	shape	NOUN
ejpam-86	6	9	parameter	parameter	NOUN
ejpam-86	6	10	of	of	ADP
ejpam-86	6	11	the	the	DET
ejpam-86	6	12	multivariate	multivariate	NOUN
ejpam-86	6	13	power	power	NOUN
ejpam-86	6	14	exponential	exponential	NOUN
ejpam-86	6	15	(	(	PUNCT
ejpam-86	6	16	mpe	mpe	ADJ
ejpam-86	6	17	)	)	PUNCT
ejpam-86	6	18	distribution	distribution	NOUN
ejpam-86	6	19	is	be	AUX
ejpam-86	6	20	equal	equal	ADJ
ejpam-86	6	21	to	to	ADP
ejpam-86	6	22	one	one	NUM
ejpam-86	6	23	which	which	PRON
ejpam-86	6	24	corresponds	correspond	VERB
ejpam-86	6	25	to	to	ADP
ejpam-86	6	26	the	the	DET
ejpam-86	6	27	multivariate	multivariate	NOUN
ejpam-86	6	28	normal	normal	ADJ
ejpam-86	6	29	distribution	distribution	NOUN
ejpam-86	6	30	.	.	PUNCT
ejpam-86	7	1	we	we	PRON
ejpam-86	7	2	develop	develop	VERB
ejpam-86	7	3	method	method	NOUN
ejpam-86	7	4	of	of	ADP
ejpam-86	7	5	moments	moment	NOUN
ejpam-86	7	6	(	(	PUNCT
ejpam-86	7	7	mom	mom	NOUN
ejpam-86	7	8	)	)	PUNCT
ejpam-86	7	9	and	and	CCONJ
ejpam-86	7	10	the	the	DET
ejpam-86	7	11	maximum	maximum	ADJ
ejpam-86	7	12	likelihood	likelihood	NOUN
ejpam-86	7	13	(	(	PUNCT
ejpam-86	7	14	ml	ml	NOUN
ejpam-86	7	15	)	)	PUNCT
ejpam-86	7	16	methods	method	NOUN
ejpam-86	7	17	to	to	PART
ejpam-86	7	18	estimate	estimate	VERB
ejpam-86	7	19	the	the	DET
ejpam-86	7	20	model	model	NOUN
ejpam-86	7	21	parameters	parameter	NOUN
ejpam-86	7	22	.	.	PUNCT
ejpam-86	8	1	the	the	DET
ejpam-86	8	2	model	model	NOUN
ejpam-86	8	3	selection	selection	NOUN
ejpam-86	8	4	criteria	criterion	NOUN
ejpam-86	8	5	such	such	ADJ
ejpam-86	8	6	as	as	ADP
ejpam-86	8	7	aic	aic	PROPN
ejpam-86	8	8	and	and	CCONJ
ejpam-86	8	9	icomp(ifim	icomp(ifim	NOUN
ejpam-86	8	10	)	)	PUNCT
ejpam-86	8	11	for	for	ADP
ejpam-86	8	12	both	both	DET
ejpam-86	8	13	models	model	NOUN
ejpam-86	8	14	are	be	AUX
ejpam-86	8	15	derived	derive	VERB
ejpam-86	8	16	.	.	PUNCT
ejpam-86	9	1	two	two	NUM
ejpam-86	9	2	simulation	simulation	NOUN
ejpam-86	9	3	examples	example	NOUN
ejpam-86	9	4	and	and	CCONJ
ejpam-86	9	5	a	a	DET
ejpam-86	9	6	real	real	ADJ
ejpam-86	9	7	example	example	NOUN
ejpam-86	9	8	on	on	ADP
ejpam-86	9	9	a	a	DET
ejpam-86	9	10	benchmark	benchmark	ADJ
ejpam-86	9	11	data	datum	NOUN
ejpam-86	9	12	set	set	VERB
ejpam-86	9	13	are	be	AUX
ejpam-86	9	14	given	give	VERB
ejpam-86	9	15	to	to	PART
ejpam-86	9	16	show	show	VERB
ejpam-86	9	17	the	the	DET
ejpam-86	9	18	applications	application	NOUN
ejpam-86	9	19	of	of	ADP
ejpam-86	9	20	these	these	DET
ejpam-86	9	21	two	two	NUM
ejpam-86	9	22	models	model	NOUN
ejpam-86	9	23	in	in	ADP
ejpam-86	9	24	subset	subset	ADJ
ejpam-86	9	25	selection	selection	NOUN
ejpam-86	9	26	of	of	ADP
ejpam-86	9	27	the	the	DET
ejpam-86	9	28	best	good	ADJ
ejpam-86	9	29	predictors	predictor	NOUN
ejpam-86	9	30	.	.	PUNCT
ejpam-86	10	1	a	a	DET
ejpam-86	10	2	genetic	genetic	ADJ
ejpam-86	10	3	algorithm	algorithm	NOUN
ejpam-86	10	4	(	(	PUNCT
ejpam-86	10	5	ga	ga	NOUN
ejpam-86	10	6	)	)	PUNCT
ejpam-86	10	7	approach	approach	NOUN
ejpam-86	10	8	is	be	AUX
ejpam-86	10	9	used	use	VERB
ejpam-86	10	10	to	to	PART
ejpam-86	10	11	obtain	obtain	VERB
ejpam-86	10	12	the	the	DET
ejpam-86	10	13	estimates	estimate	NOUN
ejpam-86	10	14	of	of	ADP
ejpam-86	10	15	the	the	DET
ejpam-86	10	16	model	model	NOUN
ejpam-86	10	17	parameters	parameter	NOUN
ejpam-86	10	18	and	and	CCONJ
ejpam-86	10	19	to	to	PART
ejpam-86	10	20	carry	carry	VERB
ejpam-86	10	21	out	out	ADP
ejpam-86	10	22	the	the	DET
ejpam-86	10	23	subset	subset	ADJ
ejpam-86	10	24	selection	selection	NOUN
ejpam-86	10	25	of	of	ADP
ejpam-86	10	26	the	the	DET
ejpam-86	10	27	best	good	ADJ
ejpam-86	10	28	predictors	predictor	NOUN
ejpam-86	10	29	under	under	ADP
ejpam-86	10	30	these	these	DET
ejpam-86	10	31	two	two	NUM
ejpam-86	10	32	different	different	ADJ
ejpam-86	10	33	model	model	NOUN
ejpam-86	10	34	types	type	NOUN
ejpam-86	10	35	.	.	PUNCT
ejpam-86	11	1	key	key	ADJ
ejpam-86	11	2	words	word	NOUN
ejpam-86	11	3	:	:	PUNCT
ejpam-86	11	4	multivariate	multivariate	NOUN
ejpam-86	11	5	power	power	NOUN
ejpam-86	11	6	exponential	exponential	ADJ
ejpam-86	11	7	distribution	distribution	NOUN
ejpam-86	11	8	,	,	PUNCT
ejpam-86	11	9	multivariate	multivariate	NOUN
ejpam-86	11	10	regression	regression	NOUN
ejpam-86	11	11	,	,	PUNCT
ejpam-86	11	12	aic	aic	PROPN
ejpam-86	11	13	,	,	PUNCT
ejpam-86	11	14	icomp	icomp	PROPN
ejpam-86	11	15	,	,	PUNCT
ejpam-86	11	16	model	model	NOUN
ejpam-86	11	17	selection	selection	NOUN
ejpam-86	11	18	,	,	PUNCT
ejpam-86	11	19	genetic	genetic	ADJ
ejpam-86	11	20	algorithm	algorithm	NOUN
ejpam-86	11	21	.	.	PUNCT
ejpam-86	12	1	1	1	X
ejpam-86	12	2	.	.	X
ejpam-86	12	3	introduction	introduction	NOUN
ejpam-86	12	4	and	and	CCONJ
ejpam-86	12	5	objectives	objective	NOUN
ejpam-86	12	6	during	during	ADP
ejpam-86	12	7	the	the	DET
ejpam-86	12	8	past	past	ADJ
ejpam-86	12	9	fifty	fifty	NUM
ejpam-86	12	10	years	year	NOUN
ejpam-86	12	11	,	,	PUNCT
ejpam-86	12	12	multivariate	multivariate	VERB
ejpam-86	12	13	normal	normal	ADJ
ejpam-86	12	14	distribution	distribution	NOUN
ejpam-86	12	15	has	have	AUX
ejpam-86	12	16	enjoyed	enjoy	VERB
ejpam-86	12	17	a	a	DET
ejpam-86	12	18	significant	significant	ADJ
ejpam-86	12	19	role	role	NOUN
ejpam-86	12	20	in	in	ADP
ejpam-86	12	21	the	the	DET
ejpam-86	12	22	development	development	NOUN
ejpam-86	12	23	of	of	ADP
ejpam-86	12	24	many	many	ADJ
ejpam-86	12	25	important	important	ADJ
ejpam-86	12	26	multivariate	multivariate	NOUN
ejpam-86	12	27	modeling	model	VERB
ejpam-86	12	28	techniques	technique	NOUN
ejpam-86	12	29	including	include	VERB
ejpam-86	12	30	the	the	DET
ejpam-86	12	31	multivariate	multivariate	NOUN
ejpam-86	12	32	regression	regression	NOUN
ejpam-86	12	33	models	model	NOUN
ejpam-86	12	34	.	.	PUNCT
ejpam-86	13	1	in	in	ADP
ejpam-86	13	2	many	many	ADJ
ejpam-86	13	3	practical	practical	ADJ
ejpam-86	13	4	applications	application	NOUN
ejpam-86	13	5	such	such	ADJ
ejpam-86	13	6	as	as	ADP
ejpam-86	13	7	in	in	ADP
ejpam-86	13	8	behavioral	behavioral	ADJ
ejpam-86	13	9	and	and	CCONJ
ejpam-86	13	10	social	social	ADJ
ejpam-86	13	11	sciences	science	NOUN
ejpam-86	13	12	,	,	PUNCT
ejpam-86	13	13	biometrics	biometric	NOUN
ejpam-86	13	14	,	,	PUNCT
ejpam-86	13	15	chemometrics	chemometric	NOUN
ejpam-86	13	16	,	,	PUNCT
ejpam-86	13	17	econometrics	econometric	NOUN
ejpam-86	13	18	,	,	PUNCT
ejpam-86	13	19	environmental	environmental	ADJ
ejpam-86	13	20	sciences	science	NOUN
ejpam-86	13	21	,	,	PUNCT
ejpam-86	13	22	and	and	CCONJ
ejpam-86	13	23	financial	financial	ADJ
ejpam-86	13	24	modeling	modeling	NOUN
ejpam-86	13	25	to	to	PART
ejpam-86	13	26	name	name	VERB
ejpam-86	13	27	a	a	DET
ejpam-86	13	28	few	few	ADJ
ejpam-86	13	29	,	,	PUNCT
ejpam-86	13	30	we	we	PRON
ejpam-86	13	31	can	can	AUX
ejpam-86	13	32	not	not	PART
ejpam-86	13	33	any	any	PRON
ejpam-86	13	34	longer	long	ADV
ejpam-86	13	35	assume	assume	VERB
ejpam-86	13	36	the	the	DET
ejpam-86	13	37	multinormality	multinormality	NOUN
ejpam-86	13	38	on	on	ADP
ejpam-86	13	39	the	the	DET
ejpam-86	13	40	set	set	NOUN
ejpam-86	13	41	of	of	ADP
ejpam-86	13	42	dependent	dependent	ADJ
ejpam-86	13	43	variables	variable	NOUN
ejpam-86	13	44	or	or	CCONJ
ejpam-86	13	45	the	the	DET
ejpam-86	13	46	random	random	ADJ
ejpam-86	13	47	error	error	NOUN
ejpam-86	13	48	term	term	NOUN
ejpam-86	13	49	of	of	ADP
ejpam-86	13	50	the	the	DET
ejpam-86	13	51	model	model	NOUN
ejpam-86	13	52	.	.	PUNCT
ejpam-86	14	1	real	real	ADJ
ejpam-86	14	2	data	datum	NOUN
ejpam-86	14	3	often	often	ADV
ejpam-86	14	4	show	show	VERB
ejpam-86	14	5	significant	significant	ADJ
ejpam-86	14	6	departures	departure	NOUN
ejpam-86	14	7	from	from	ADP
ejpam-86	14	8	normality	normality	NOUN
ejpam-86	14	9	and	and	CCONJ
ejpam-86	14	10	normality	normality	NOUN
ejpam-86	14	11	may	may	AUX
ejpam-86	14	12	not	not	PART
ejpam-86	14	13	be	be	AUX
ejpam-86	14	14	tenable	tenable	ADJ
ejpam-86	14	15	,	,	PUNCT
ejpam-86	14	16	especially	especially	ADV
ejpam-86	14	17	when	when	SCONJ
ejpam-86	14	18	the	the	DET
ejpam-86	14	19	tails	tail	NOUN
ejpam-86	14	20	are	be	AUX
ejpam-86	14	21	thicker	thick	ADJ
ejpam-86	14	22	or	or	CCONJ
ejpam-86	14	23	thinner	thin	ADJ
ejpam-86	14	24	than	than	ADP
ejpam-86	14	25	those	those	PRON
ejpam-86	14	26	of	of	ADP
ejpam-86	14	27	normal	normal	ADJ
ejpam-86	14	28	distributions	distribution	NOUN
ejpam-86	14	29	.	.	PUNCT
ejpam-86	15	1	for	for	ADP
ejpam-86	15	2	this	this	DET
ejpam-86	15	3	reason	reason	NOUN
ejpam-86	15	4	,	,	PUNCT
ejpam-86	15	5	to	to	PART
ejpam-86	15	6	achieve	achieve	VERB
ejpam-86	15	7	more	more	ADJ
ejpam-86	15	8	flexibility	flexibility	NOUN
ejpam-86	15	9	in	in	ADP
ejpam-86	15	10	statistical	statistical	ADJ
ejpam-86	15	11	modeling	modeling	NOUN
ejpam-86	15	12	and	and	CCONJ
ejpam-86	15	13	model	model	NOUN
ejpam-86	15	14	selection	selection	NOUN
ejpam-86	15	15	,	,	PUNCT
ejpam-86	15	16	and	and	CCONJ
ejpam-86	15	17	to	to	PART
ejpam-86	15	18	robustify	robustify	VERB
ejpam-86	15	19	many	many	ADJ
ejpam-86	15	20	multivariate	multivariate	ADJ
ejpam-86	15	21	statistical	statistical	ADJ
ejpam-86	15	22	procedures	procedure	NOUN
ejpam-86	15	23	,	,	PUNCT
ejpam-86	15	24	the	the	DET
ejpam-86	15	25	purpose	purpose	NOUN
ejpam-86	15	26	of	of	ADP
ejpam-86	15	27	this	this	DET
ejpam-86	15	28	paper	paper	NOUN
ejpam-86	15	29	is	be	AUX
ejpam-86	15	30	to	to	PART
ejpam-86	15	31	develop	develop	VERB
ejpam-86	15	32	novel	novel	ADJ
ejpam-86	15	33	techniques	technique	NOUN
ejpam-86	15	34	in	in	ADP
ejpam-86	15	35	multivariate	multivariate	NOUN
ejpam-86	15	36	regression	regression	NOUN
ejpam-86	15	37	models	model	NOUN
ejpam-86	15	38	for	for	ADP
ejpam-86	15	39	nonnormal	nonnormal	ADJ
ejpam-86	15	40	data	datum	NOUN
ejpam-86	15	41	under	under	ADP
ejpam-86	15	42	the	the	DET
ejpam-86	15	43	general	general	ADJ
ejpam-86	15	44	class	class	NOUN
ejpam-86	15	45	of	of	ADP
ejpam-86	15	46	:	:	PUNCT
ejpam-86	15	47	multivariate	multivariate	NOUN
ejpam-86	15	48	power	power	NOUN
ejpam-86	15	49	exponential	exponential	NOUN
ejpam-86	15	50	(	(	PUNCT
ejpam-86	15	51	pe	pe	NOUN
ejpam-86	15	52	)	)	PUNCT
ejpam-86	15	53	distributions	distribution	NOUN
ejpam-86	15	54	by	by	ADP
ejpam-86	15	55	broadening	broaden	VERB
ejpam-86	15	56	the	the	DET
ejpam-86	15	57	usual	usual	ADJ
ejpam-86	15	58	multivariate	multivariate	NOUN
ejpam-86	15	59	normal	normal	ADJ
ejpam-86	15	60	assumption	assumption	NOUN
ejpam-86	15	61	on	on	ADP
ejpam-86	15	62	the	the	DET
ejpam-86	15	63	random	random	ADJ
ejpam-86	15	64	errors	error	NOUN
ejpam-86	15	65	under	under	ADP
ejpam-86	15	66	various	various	ADJ
ejpam-86	15	67	assumptions	assumption	NOUN
ejpam-86	15	68	.	.	PUNCT
ejpam-86	16	1	in	in	ADP
ejpam-86	16	2	regression	regression	NOUN
ejpam-86	16	3	models	model	NOUN
ejpam-86	16	4	,	,	PUNCT
ejpam-86	16	5	the	the	DET
ejpam-86	16	6	∗corresponding	∗corresponde	VERB
ejpam-86	16	7	author	author	NOUN
ejpam-86	16	8	.	.	PUNCT
ejpam-86	17	1	email	email	NOUN
ejpam-86	17	2	address	address	NOUN
ejpam-86	17	3	:	:	PUNCT
ejpam-86	17	4	bozdogan@utk.edu	bozdogan@utk.edu	PROPN
ejpam-86	17	5	(	(	PUNCT
ejpam-86	17	6	h.	h.	PROPN
ejpam-86	17	7	bozdogan	bozdogan	PROPN
ejpam-86	17	8	)	)	PUNCT
ejpam-86	17	9	http://www.ejpam.com	http://www.ejpam.com	X
ejpam-86	18	1	4	4	NUM
ejpam-86	19	1	c	c	X
ejpam-86	19	2	©	©	NOUN
ejpam-86	19	3	2007	2007	NUM
ejpam-86	19	4	ejpam	ejpam	NOUN
ejpam-86	19	5	all	all	DET
ejpam-86	19	6	rights	right	NOUN
ejpam-86	19	7	reserved	reserve	VERB
ejpam-86	19	8	.	.	PUNCT
ejpam-86	20	1	m.	m.	PROPN
ejpam-86	20	2	liu	liu	PROPN
ejpam-86	20	3	and	and	CCONJ
ejpam-86	20	4	h.	h.	PROPN
ejpam-86	20	5	bozdogan	bozdogan	PROPN
ejpam-86	20	6	/	/	SYM
ejpam-86	20	7	eur	eur	PROPN
ejpam-86	20	8	.	.	PUNCT
ejpam-86	21	1	j.	j.	PROPN
ejpam-86	21	2	pure	pure	PROPN
ejpam-86	21	3	appl	appl	PROPN
ejpam-86	21	4	.	.	PROPN
ejpam-86	21	5	math	math	PROPN
ejpam-86	21	6	,	,	PUNCT
ejpam-86	21	7	1	1	NUM
ejpam-86	21	8	(	(	PUNCT
ejpam-86	21	9	2008	2008	NUM
ejpam-86	21	10	)	)	PUNCT
ejpam-86	21	11	,	,	PUNCT
ejpam-86	21	12	(	(	PUNCT
ejpam-86	21	13	4	4	NUM
ejpam-86	21	14	-	-	SYM
ejpam-86	21	15	37	37	NUM
ejpam-86	21	16	)	)	PUNCT
ejpam-86	21	17	5	5	NUM
ejpam-86	21	18	random	random	ADJ
ejpam-86	21	19	error	error	NOUN
ejpam-86	21	20	terms	term	NOUN
ejpam-86	21	21	are	be	AUX
ejpam-86	21	22	generally	generally	ADV
ejpam-86	21	23	assumed	assume	VERB
ejpam-86	21	24	to	to	PART
ejpam-86	21	25	be	be	AUX
ejpam-86	21	26	normally	normally	ADV
ejpam-86	21	27	distributed	distribute	VERB
ejpam-86	21	28	.	.	PUNCT
ejpam-86	22	1	however	however	ADV
ejpam-86	22	2	,	,	PUNCT
ejpam-86	22	3	since	since	SCONJ
ejpam-86	22	4	data	datum	NOUN
ejpam-86	22	5	often	often	ADV
ejpam-86	22	6	are	be	AUX
ejpam-86	22	7	non	non	ADJ
ejpam-86	22	8	-	-	ADJ
ejpam-86	22	9	normal	normal	ADJ
ejpam-86	22	10	,	,	PUNCT
ejpam-86	22	11	the	the	DET
ejpam-86	22	12	normality	normality	NOUN
ejpam-86	22	13	assumption	assumption	NOUN
ejpam-86	22	14	is	be	AUX
ejpam-86	22	15	not	not	PART
ejpam-86	22	16	always	always	ADV
ejpam-86	22	17	tenable	tenable	ADJ
ejpam-86	22	18	especially	especially	ADV
ejpam-86	22	19	when	when	SCONJ
ejpam-86	22	20	the	the	DET
ejpam-86	22	21	tails	tail	NOUN
ejpam-86	22	22	are	be	AUX
ejpam-86	22	23	thicker	thick	ADJ
ejpam-86	22	24	or	or	CCONJ
ejpam-86	22	25	thinner	thin	ADJ
ejpam-86	22	26	than	than	ADP
ejpam-86	22	27	those	those	PRON
ejpam-86	22	28	of	of	ADP
ejpam-86	22	29	normal	normal	ADJ
ejpam-86	22	30	distributions	distribution	NOUN
ejpam-86	22	31	.	.	PUNCT
ejpam-86	23	1	the	the	DET
ejpam-86	23	2	power	power	NOUN
ejpam-86	23	3	exponential	exponential	NOUN
ejpam-86	23	4	(	(	PUNCT
ejpam-86	23	5	pe	pe	NOUN
ejpam-86	23	6	)	)	PUNCT
ejpam-86	23	7	distribution	distribution	NOUN
ejpam-86	23	8	family	family	NOUN
ejpam-86	23	9	which	which	PRON
ejpam-86	23	10	is	be	AUX
ejpam-86	23	11	introduced	introduce	VERB
ejpam-86	23	12	by	by	ADP
ejpam-86	23	13	subbotin	subbotin	NOUN
ejpam-86	23	14	(	(	PUNCT
ejpam-86	23	15	[	[	X
ejpam-86	23	16	32	32	NUM
ejpam-86	23	17	]	]	PUNCT
ejpam-86	23	18	)	)	PUNCT
ejpam-86	23	19	and	and	CCONJ
ejpam-86	23	20	popularized	popularize	VERB
ejpam-86	23	21	by	by	ADP
ejpam-86	23	22	box	box	NOUN
ejpam-86	23	23	and	and	CCONJ
ejpam-86	23	24	tiao	tiao	NOUN
ejpam-86	23	25	(	(	PUNCT
ejpam-86	23	26	[	[	X
ejpam-86	23	27	6	6	NUM
ejpam-86	23	28	]	]	NUM
ejpam-86	23	29	)	)	PUNCT
ejpam-86	23	30	,	,	PUNCT
ejpam-86	23	31	has	have	AUX
ejpam-86	23	32	been	be	AUX
ejpam-86	23	33	used	use	VERB
ejpam-86	23	34	in	in	ADP
ejpam-86	23	35	modeling	model	VERB
ejpam-86	23	36	economic	economic	ADJ
ejpam-86	23	37	and	and	CCONJ
ejpam-86	23	38	financial	financial	ADJ
ejpam-86	23	39	data	datum	NOUN
ejpam-86	23	40	as	as	ADP
ejpam-86	23	41	a	a	DET
ejpam-86	23	42	generalization	generalization	NOUN
ejpam-86	23	43	of	of	ADP
ejpam-86	23	44	normal	normal	ADJ
ejpam-86	23	45	distribution	distribution	NOUN
ejpam-86	23	46	in	in	ADP
ejpam-86	23	47	recent	recent	ADJ
ejpam-86	23	48	years	year	NOUN
ejpam-86	23	49	(	(	PUNCT
ejpam-86	23	50	e.g.	e.g.	ADV
ejpam-86	23	51	[	[	X
ejpam-86	23	52	28	28	NUM
ejpam-86	23	53	,	,	PUNCT
ejpam-86	23	54	33	33	NUM
ejpam-86	23	55	,	,	PUNCT
ejpam-86	23	56	34	34	NUM
ejpam-86	23	57	]	]	PUNCT
ejpam-86	23	58	)	)	PUNCT
ejpam-86	23	59	.	.	PUNCT
ejpam-86	24	1	gómez	gómez	NOUN
ejpam-86	24	2	-	-	PUNCT
ejpam-86	24	3	villegas	villegas	PROPN
ejpam-86	24	4	and	and	CCONJ
ejpam-86	24	5	sá	sá	PROPN
ejpam-86	24	6	nchez	nchez	PROPN
ejpam-86	24	7	-	-	PUNCT
ejpam-86	24	8	manzano	manzano	PROPN
ejpam-86	24	9	et	et	PROPN
ejpam-86	24	10	al	al	PROPN
ejpam-86	24	11	.	.	PUNCT
ejpam-86	25	1	(	(	PUNCT
ejpam-86	25	2	[	[	X
ejpam-86	25	3	20	20	NUM
ejpam-86	25	4	,	,	PUNCT
ejpam-86	25	5	31	31	NUM
ejpam-86	25	6	]	]	PUNCT
ejpam-86	25	7	)	)	PUNCT
ejpam-86	25	8	proposed	propose	VERB
ejpam-86	25	9	multivariate	multivariate	NOUN
ejpam-86	25	10	and	and	CCONJ
ejpam-86	25	11	matrix	matrix	NOUN
ejpam-86	25	12	generalizations	generalization	NOUN
ejpam-86	25	13	of	of	ADP
ejpam-86	25	14	the	the	DET
ejpam-86	25	15	pe	pe	PROPN
ejpam-86	25	16	family	family	NOUN
ejpam-86	25	17	of	of	ADP
ejpam-86	25	18	distributions	distribution	NOUN
ejpam-86	25	19	and	and	CCONJ
ejpam-86	25	20	studied	study	VERB
ejpam-86	25	21	their	their	PRON
ejpam-86	25	22	properties	property	NOUN
ejpam-86	25	23	in	in	ADP
ejpam-86	25	24	relation	relation	NOUN
ejpam-86	25	25	to	to	PART
ejpam-86	25	26	multivariate	multivariate	VERB
ejpam-86	25	27	elliptically	elliptically	ADV
ejpam-86	25	28	contoured	contour	VERB
ejpam-86	25	29	(	(	PUNCT
ejpam-86	25	30	ec	ec	NOUN
ejpam-86	25	31	)	)	PUNCT
ejpam-86	25	32	distributions	distribution	NOUN
ejpam-86	25	33	.	.	PUNCT
ejpam-86	26	1	zeckhauser	zeckhauser	NOUN
ejpam-86	26	2	and	and	CCONJ
ejpam-86	26	3	thompson	thompson	PROPN
ejpam-86	26	4	(	(	PUNCT
ejpam-86	26	5	[	[	X
ejpam-86	26	6	36	36	NUM
ejpam-86	26	7	]	]	PUNCT
ejpam-86	26	8	)	)	PUNCT
ejpam-86	26	9	were	be	AUX
ejpam-86	26	10	probably	probably	ADV
ejpam-86	26	11	the	the	DET
ejpam-86	26	12	first	first	ADJ
ejpam-86	26	13	who	who	PRON
ejpam-86	26	14	attempted	attempt	VERB
ejpam-86	26	15	to	to	PART
ejpam-86	26	16	study	study	VERB
ejpam-86	26	17	the	the	DET
ejpam-86	26	18	simple	simple	ADJ
ejpam-86	26	19	multiple	multiple	ADJ
ejpam-86	26	20	linear	linear	ADJ
ejpam-86	26	21	regression	regression	NOUN
ejpam-86	26	22	model	model	NOUN
ejpam-86	26	23	with	with	ADP
ejpam-86	26	24	pe	pe	NOUN
ejpam-86	26	25	error	error	NOUN
ejpam-86	26	26	terms	term	NOUN
ejpam-86	26	27	in	in	ADP
ejpam-86	26	28	a	a	DET
ejpam-86	26	29	short	short	ADJ
ejpam-86	26	30	but	but	CCONJ
ejpam-86	26	31	an	an	DET
ejpam-86	26	32	incomplete	incomplete	ADJ
ejpam-86	26	33	paper	paper	NOUN
ejpam-86	26	34	.	.	PUNCT
ejpam-86	27	1	liu	liu	PROPN
ejpam-86	27	2	and	and	CCONJ
ejpam-86	27	3	bozdogan	bozdogan	PROPN
ejpam-86	27	4	(	(	PUNCT
ejpam-86	27	5	[	[	X
ejpam-86	27	6	23	23	NUM
ejpam-86	27	7	]	]	PUNCT
ejpam-86	27	8	)	)	PUNCT
ejpam-86	27	9	developed	develop	VERB
ejpam-86	27	10	a	a	DET
ejpam-86	27	11	ga	ga	PROPN
ejpam-86	27	12	on	on	ADP
ejpam-86	27	13	ga	ga	PROPN
ejpam-86	27	14	(	(	PUNCT
ejpam-86	27	15	or	or	CCONJ
ejpam-86	27	16	ga	ga	PROPN
ejpam-86	27	17	engineering	engineering	NOUN
ejpam-86	27	18	)	)	PUNCT
ejpam-86	27	19	approach	approach	NOUN
ejpam-86	27	20	for	for	ADP
ejpam-86	27	21	pe	pe	PROPN
ejpam-86	27	22	multiple	multiple	ADJ
ejpam-86	27	23	regression	regression	NOUN
ejpam-86	27	24	and	and	CCONJ
ejpam-86	27	25	subset	subset	ADJ
ejpam-86	27	26	selection	selection	NOUN
ejpam-86	27	27	of	of	ADP
ejpam-86	27	28	variables	variable	NOUN
ejpam-86	27	29	with	with	ADP
ejpam-86	27	30	information	information	NOUN
ejpam-86	27	31	-	-	PUNCT
ejpam-86	27	32	theoretic	theoretic	NOUN
ejpam-86	27	33	complexity	complexity	NOUN
ejpam-86	27	34	(	(	PUNCT
ejpam-86	27	35	icomp	icomp	ADJ
ejpam-86	27	36	)	)	PUNCT
ejpam-86	27	37	criterion	criterion	NOUN
ejpam-86	27	38	.	.	PUNCT
ejpam-86	28	1	the	the	DET
ejpam-86	28	2	closed	closed	ADJ
ejpam-86	28	3	form	form	NOUN
ejpam-86	28	4	expressions	expression	NOUN
ejpam-86	28	5	of	of	ADP
ejpam-86	28	6	the	the	DET
ejpam-86	28	7	inverse	inverse	NOUN
ejpam-86	28	8	fisher	fisher	PROPN
ejpam-86	28	9	information	information	NOUN
ejpam-86	28	10	matrix	matrix	NOUN
ejpam-86	28	11	(	(	PUNCT
ejpam-86	28	12	ifim	ifim	NOUN
ejpam-86	28	13	)	)	PUNCT
ejpam-86	28	14	and	and	CCONJ
ejpam-86	28	15	icomp(ifim	icomp(ifim	NOUN
ejpam-86	28	16	)	)	PUNCT
ejpam-86	28	17	for	for	ADP
ejpam-86	28	18	pe	pe	PROPN
ejpam-86	28	19	multiple	multiple	ADJ
ejpam-86	28	20	linear	linear	ADJ
ejpam-86	28	21	regression	regression	NOUN
ejpam-86	28	22	models	model	NOUN
ejpam-86	28	23	were	be	AUX
ejpam-86	28	24	also	also	ADV
ejpam-86	28	25	given	give	VERB
ejpam-86	28	26	.	.	PUNCT
ejpam-86	28	27	finding	find	VERB
ejpam-86	28	28	proper	proper	ADJ
ejpam-86	28	29	criterion	criterion	NOUN
ejpam-86	28	30	or	or	CCONJ
ejpam-86	28	31	measure	measure	NOUN
ejpam-86	28	32	for	for	ADP
ejpam-86	28	33	the	the	DET
ejpam-86	28	34	comparison	comparison	NOUN
ejpam-86	28	35	of	of	ADP
ejpam-86	28	36	competing	compete	VERB
ejpam-86	28	37	models	model	NOUN
ejpam-86	28	38	is	be	AUX
ejpam-86	28	39	important	important	ADJ
ejpam-86	28	40	to	to	PART
ejpam-86	28	41	select	select	VERB
ejpam-86	28	42	correct	correct	ADJ
ejpam-86	28	43	regression	regression	NOUN
ejpam-86	28	44	models	model	NOUN
ejpam-86	28	45	.	.	PUNCT
ejpam-86	29	1	aic	aic	PROPN
ejpam-86	29	2	-	-	PUNCT
ejpam-86	29	3	type	type	NOUN
ejpam-86	29	4	criteria	criterion	NOUN
ejpam-86	29	5	(	(	PUNCT
ejpam-86	29	6	[	[	X
ejpam-86	29	7	2–5	2–5	NOUN
ejpam-86	29	8	]	]	PUNCT
ejpam-86	29	9	)	)	PUNCT
ejpam-86	29	10	are	be	AUX
ejpam-86	29	11	the	the	DET
ejpam-86	29	12	most	most	ADV
ejpam-86	29	13	widely	widely	ADV
ejpam-86	29	14	used	use	VERB
ejpam-86	29	15	information	information	NOUN
ejpam-86	29	16	-	-	PUNCT
ejpam-86	29	17	based	base	VERB
ejpam-86	29	18	criteria	criterion	NOUN
ejpam-86	29	19	for	for	ADP
ejpam-86	29	20	model	model	NOUN
ejpam-86	29	21	selection	selection	NOUN
ejpam-86	29	22	in	in	ADP
ejpam-86	29	23	recent	recent	ADJ
ejpam-86	29	24	years	year	NOUN
ejpam-86	29	25	.	.	PUNCT
ejpam-86	30	1	however	however	ADV
ejpam-86	30	2	,	,	PUNCT
ejpam-86	30	3	the	the	DET
ejpam-86	30	4	penalty	penalty	NOUN
ejpam-86	30	5	term	term	NOUN
ejpam-86	30	6	used	use	VERB
ejpam-86	30	7	in	in	ADP
ejpam-86	30	8	aic	aic	PROPN
ejpam-86	30	9	-	-	PUNCT
ejpam-86	30	10	type	type	NOUN
ejpam-86	30	11	criteria	criterion	NOUN
ejpam-86	30	12	is	be	AUX
ejpam-86	30	13	insufficient	insufficient	ADJ
ejpam-86	30	14	to	to	PART
ejpam-86	30	15	measure	measure	VERB
ejpam-86	30	16	the	the	DET
ejpam-86	30	17	model	model	NOUN
ejpam-86	30	18	complexity	complexity	NOUN
ejpam-86	30	19	which	which	PRON
ejpam-86	30	20	has	have	AUX
ejpam-86	30	21	been	be	AUX
ejpam-86	30	22	cited	cite	VERB
ejpam-86	30	23	by	by	ADP
ejpam-86	30	24	many	many	ADJ
ejpam-86	30	25	authors	author	NOUN
ejpam-86	30	26	(	(	PUNCT
ejpam-86	30	27	see	see	VERB
ejpam-86	30	28	,	,	PUNCT
ejpam-86	30	29	e.g.	e.g.	ADV
ejpam-86	30	30	[	[	X
ejpam-86	30	31	29	29	NUM
ejpam-86	30	32	]	]	PUNCT
ejpam-86	30	33	)	)	PUNCT
ejpam-86	30	34	.	.	PUNCT
ejpam-86	31	1	bozdogan	bozdogan	PROPN
ejpam-86	31	2	’s	’s	PART
ejpam-86	31	3	icomp	icomp	ADJ
ejpam-86	31	4	criteria	criterion	NOUN
ejpam-86	31	5	(	(	PUNCT
ejpam-86	31	6	[	[	X
ejpam-86	31	7	7,9,10,12–14	7,9,10,12–14	X
ejpam-86	31	8	]	]	X
ejpam-86	31	9	)	)	PUNCT
ejpam-86	31	10	improve	improve	VERB
ejpam-86	31	11	aic	aic	PROPN
ejpam-86	31	12	-	-	PUNCT
ejpam-86	31	13	type	type	NOUN
ejpam-86	31	14	criteria	criterion	NOUN
ejpam-86	31	15	by	by	ADP
ejpam-86	31	16	using	use	VERB
ejpam-86	31	17	an	an	DET
ejpam-86	31	18	information	information	NOUN
ejpam-86	31	19	-	-	PUNCT
ejpam-86	31	20	theoretic	theoretic	NOUN
ejpam-86	31	21	measure	measure	NOUN
ejpam-86	31	22	of	of	ADP
ejpam-86	31	23	“	"	PUNCT
ejpam-86	31	24	overall	overall	ADJ
ejpam-86	31	25	”	"	PUNCT
ejpam-86	31	26	model	model	NOUN
ejpam-86	31	27	complexity	complexity	NOUN
ejpam-86	31	28	based	base	VERB
ejpam-86	31	29	on	on	ADP
ejpam-86	31	30	the	the	DET
ejpam-86	31	31	generalized	generalized	ADJ
ejpam-86	31	32	covariance	covariance	NOUN
ejpam-86	31	33	complexity	complexity	NOUN
ejpam-86	31	34	index	index	NOUN
ejpam-86	31	35	of	of	ADP
ejpam-86	31	36	van	van	PROPN
ejpam-86	31	37	emden	emden	PROPN
ejpam-86	31	38	(	(	PUNCT
ejpam-86	31	39	[	[	X
ejpam-86	31	40	16	16	NUM
ejpam-86	31	41	]	]	SYM
ejpam-86	31	42	)	)	PUNCT
ejpam-86	31	43	.	.	PUNCT
ejpam-86	32	1	icomp(ifim	icomp(ifim	NOUN
ejpam-86	32	2	)	)	PUNCT
ejpam-86	32	3	is	be	AUX
ejpam-86	32	4	the	the	DET
ejpam-86	32	5	most	most	ADV
ejpam-86	32	6	general	general	ADJ
ejpam-86	32	7	form	form	NOUN
ejpam-86	32	8	of	of	ADP
ejpam-86	32	9	icomp	icomp	NOUN
ejpam-86	32	10	.	.	PUNCT
ejpam-86	33	1	in	in	ADP
ejpam-86	33	2	this	this	DET
ejpam-86	33	3	paper	paper	NOUN
ejpam-86	33	4	,	,	PUNCT
ejpam-86	33	5	we	we	PRON
ejpam-86	33	6	extend	extend	VERB
ejpam-86	33	7	the	the	DET
ejpam-86	33	8	work	work	NOUN
ejpam-86	33	9	of	of	ADP
ejpam-86	33	10	liu	liu	PROPN
ejpam-86	33	11	and	and	CCONJ
ejpam-86	33	12	bozdogan	bozdogan	PROPN
ejpam-86	33	13	(	(	PUNCT
ejpam-86	33	14	[	[	X
ejpam-86	33	15	23	23	NUM
ejpam-86	33	16	]	]	PUNCT
ejpam-86	33	17	)	)	PUNCT
ejpam-86	33	18	and	and	CCONJ
ejpam-86	33	19	study	study	VERB
ejpam-86	33	20	the	the	DET
ejpam-86	33	21	multivariate	multivariate	NOUN
ejpam-86	33	22	regression	regression	NOUN
ejpam-86	33	23	models	model	NOUN
ejpam-86	33	24	with	with	ADP
ejpam-86	33	25	pe	pe	INTJ
ejpam-86	33	26	random	random	ADJ
ejpam-86	33	27	errors	error	NOUN
ejpam-86	33	28	under	under	ADP
ejpam-86	33	29	various	various	ADJ
ejpam-86	33	30	assumptions	assumption	NOUN
ejpam-86	33	31	.	.	PUNCT
ejpam-86	34	1	as	as	ADP
ejpam-86	34	2	a	a	DET
ejpam-86	34	3	short	short	ADJ
ejpam-86	34	4	hand	hand	NOUN
ejpam-86	34	5	notation	notation	NOUN
ejpam-86	34	6	,	,	PUNCT
ejpam-86	34	7	we	we	PRON
ejpam-86	34	8	abbreviate	abbreviate	VERB
ejpam-86	34	9	these	these	DET
ejpam-86	34	10	models	model	NOUN
ejpam-86	34	11	as	as	ADP
ejpam-86	34	12	mvper	mvper	NOUN
ejpam-86	34	13	models	model	NOUN
ejpam-86	34	14	.	.	PUNCT
ejpam-86	35	1	we	we	PRON
ejpam-86	35	2	develop	develop	VERB
ejpam-86	35	3	and	and	CCONJ
ejpam-86	35	4	use	use	VERB
ejpam-86	35	5	the	the	DET
ejpam-86	35	6	genetic	genetic	ADJ
ejpam-86	35	7	algorithms	algorithm	NOUN
ejpam-86	35	8	(	(	PUNCT
ejpam-86	35	9	gas	gas	NOUN
ejpam-86	35	10	)	)	PUNCT
ejpam-86	35	11	for	for	ADP
ejpam-86	35	12	both	both	CCONJ
ejpam-86	35	13	the	the	DET
ejpam-86	35	14	parameter	parameter	NOUN
ejpam-86	35	15	estimation	estimation	NOUN
ejpam-86	35	16	of	of	ADP
ejpam-86	35	17	the	the	DET
ejpam-86	35	18	models	model	NOUN
ejpam-86	35	19	and	and	CCONJ
ejpam-86	35	20	for	for	ADP
ejpam-86	35	21	subset	subset	ADJ
ejpam-86	35	22	selection	selection	NOUN
ejpam-86	35	23	of	of	ADP
ejpam-86	35	24	best	good	ADJ
ejpam-86	35	25	predictors	predictor	NOUN
ejpam-86	35	26	with	with	ADP
ejpam-86	35	27	information	information	NOUN
ejpam-86	35	28	criteria	criterion	NOUN
ejpam-86	35	29	as	as	ADP
ejpam-86	35	30	our	our	PRON
ejpam-86	35	31	fitness	fitness	NOUN
ejpam-86	35	32	function	function	NOUN
ejpam-86	35	33	.	.	PUNCT
ejpam-86	36	1	we	we	PRON
ejpam-86	36	2	note	note	VERB
ejpam-86	36	3	that	that	SCONJ
ejpam-86	36	4	,	,	PUNCT
ejpam-86	36	5	in	in	ADP
ejpam-86	36	6	the	the	DET
ejpam-86	36	7	literature	literature	NOUN
ejpam-86	36	8	,	,	PUNCT
ejpam-86	36	9	this	this	DET
ejpam-86	36	10	work	work	NOUN
ejpam-86	36	11	seems	seem	VERB
ejpam-86	36	12	to	to	PART
ejpam-86	36	13	be	be	AUX
ejpam-86	36	14	the	the	DET
ejpam-86	36	15	first	first	ADJ
ejpam-86	36	16	to	to	PART
ejpam-86	36	17	attempt	attempt	VERB
ejpam-86	36	18	to	to	PART
ejpam-86	36	19	study	study	VERB
ejpam-86	36	20	mvper	mvper	NOUN
ejpam-86	36	21	models	model	NOUN
ejpam-86	36	22	utilizing	utilize	VERB
ejpam-86	36	23	modern	modern	ADJ
ejpam-86	36	24	optimization	optimization	NOUN
ejpam-86	36	25	techniques	technique	NOUN
ejpam-86	36	26	such	such	DET
ejpam-86	36	27	the	the	DET
ejpam-86	36	28	gas	gas	NOUN
ejpam-86	36	29	and	and	CCONJ
ejpam-86	36	30	information	information	NOUN
ejpam-86	36	31	criteria	criterion	NOUN
ejpam-86	36	32	as	as	ADP
ejpam-86	36	33	our	our	PRON
ejpam-86	36	34	fitness	fitness	NOUN
ejpam-86	36	35	function	function	NOUN
ejpam-86	36	36	.	.	PUNCT
ejpam-86	37	1	more	more	ADV
ejpam-86	37	2	specifically	specifically	ADV
ejpam-86	37	3	,	,	PUNCT
ejpam-86	37	4	we	we	PRON
ejpam-86	37	5	consider	consider	VERB
ejpam-86	37	6	the	the	DET
ejpam-86	37	7	multivariate	multivariate	NOUN
ejpam-86	37	8	regression	regression	NOUN
ejpam-86	37	9	model	model	NOUN
ejpam-86	37	10	:	:	PUNCT
ejpam-86	37	11	yn×p	yn×p	PROPN
ejpam-86	37	12	=	=	SYM
ejpam-86	37	13	xn×qbq×p	xn×qbq×p	PROPN
ejpam-86	38	1	+	+	PUNCT
ejpam-86	38	2	en×p	en×p	PROPN
ejpam-86	38	3	,	,	PUNCT
ejpam-86	38	4	p+	p+	VERB
ejpam-86	38	5	q	q	PROPN
ejpam-86	38	6	≤	≤	PROPN
ejpam-86	38	7	n	n	CCONJ
ejpam-86	38	8	,	,	PUNCT
ejpam-86	38	9	rank(x	rank(x	PROPN
ejpam-86	38	10	)	)	PUNCT
ejpam-86	38	11	=	=	SYM
ejpam-86	39	1	q	q	X
ejpam-86	39	2	(	(	PUNCT
ejpam-86	39	3	1.1	1.1	NUM
ejpam-86	39	4	)	)	PUNCT
ejpam-86	39	5	where	where	SCONJ
ejpam-86	39	6	y	y	NOUN
ejpam-86	39	7	=	=	PUNCT
ejpam-86	40	1	[	[	X
ejpam-86	40	2	y1,y2	y1,y2	PROPN
ejpam-86	40	3	,	,	PUNCT
ejpam-86	40	4	·	·	PUNCT
ejpam-86	40	5	·	·	PUNCT
ejpam-86	40	6	·	·	PUNCT
ejpam-86	40	7	,	,	PUNCT
ejpam-86	40	8	yp	yp	X
ejpam-86	40	9	]	]	X
ejpam-86	40	10	=	=	X
ejpam-86	41	1	[	[	X
ejpam-86	41	2	y(1),y(2	y(1),y(2	NOUN
ejpam-86	41	3	)	)	PUNCT
ejpam-86	41	4	,	,	PUNCT
ejpam-86	41	5	·	·	PUNCT
ejpam-86	41	6	·	·	PUNCT
ejpam-86	41	7	·	·	PUNCT
ejpam-86	41	8	,	,	PUNCT
ejpam-86	41	9	y(n)]′	y(n)]′	PROPN
ejpam-86	41	10	is	be	AUX
ejpam-86	41	11	the	the	DET
ejpam-86	41	12	matrix	matrix	NOUN
ejpam-86	41	13	of	of	ADP
ejpam-86	41	14	n	n	DET
ejpam-86	41	15	observations	observation	NOUN
ejpam-86	41	16	on	on	ADP
ejpam-86	41	17	p	p	NOUN
ejpam-86	41	18	response	response	NOUN
ejpam-86	41	19	variables	variable	NOUN
ejpam-86	41	20	.	.	PUNCT
ejpam-86	42	1	x	x	X
ejpam-86	43	1	=	=	PUNCT
ejpam-86	44	1	[	[	X
ejpam-86	44	2	x1,x2	x1,x2	PROPN
ejpam-86	44	3	,	,	PUNCT
ejpam-86	44	4	·	·	PUNCT
ejpam-86	44	5	·	·	PUNCT
ejpam-86	44	6	·	·	PUNCT
ejpam-86	44	7	,	,	PUNCT
ejpam-86	44	8	xq	xq	X
ejpam-86	44	9	]	]	X
ejpam-86	44	10	=	=	PUNCT
ejpam-86	44	11	[	[	X
ejpam-86	44	12	x(1),x(2	x(1),x(2	NOUN
ejpam-86	44	13	)	)	PUNCT
ejpam-86	44	14	,	,	PUNCT
ejpam-86	44	15	·	·	PUNCT
ejpam-86	44	16	·	·	PUNCT
ejpam-86	44	17	·	·	PUNCT
ejpam-86	44	18	,	,	PUNCT
ejpam-86	44	19	x(n)]′	x(n)]′	PROPN
ejpam-86	44	20	is	be	AUX
ejpam-86	44	21	the	the	DET
ejpam-86	44	22	matrix	matrix	NOUN
ejpam-86	44	23	of	of	ADP
ejpam-86	44	24	(	(	PUNCT
ejpam-86	44	25	n×	n×	NOUN
ejpam-86	44	26	q	q	NOUN
ejpam-86	44	27	)	)	PUNCT
ejpam-86	44	28	constant	constant	ADJ
ejpam-86	44	29	terms	term	NOUN
ejpam-86	44	30	on	on	ADP
ejpam-86	44	31	non	non	ADJ
ejpam-86	44	32	-	-	ADJ
ejpam-86	44	33	stochastic	stochastic	ADJ
ejpam-86	44	34	predictor	predictor	NOUN
ejpam-86	44	35	variables	variable	NOUN
ejpam-86	44	36	,	,	PUNCT
ejpam-86	44	37	b	b	X
ejpam-86	44	38	=	=	SYM
ejpam-86	45	1	[	[	X
ejpam-86	45	2	b1,b2	b1,b2	PROPN
ejpam-86	45	3	,	,	PUNCT
ejpam-86	45	4	·	·	PUNCT
ejpam-86	45	5	·	·	PUNCT
ejpam-86	45	6	·	·	PUNCT
ejpam-86	45	7	,	,	PUNCT
ejpam-86	45	8	bp	bp	PROPN
ejpam-86	45	9	]	]	PUNCT
ejpam-86	45	10	=	=	PUNCT
ejpam-86	46	1	[	[	X
ejpam-86	46	2	b(1),b(2	b(1),b(2	PROPN
ejpam-86	46	3	)	)	PUNCT
ejpam-86	46	4	,	,	PUNCT
ejpam-86	46	5	·	·	PUNCT
ejpam-86	46	6	·	·	PUNCT
ejpam-86	46	7	·	·	PUNCT
ejpam-86	46	8	,	,	PUNCT
ejpam-86	46	9	b(q)]′	b(q)]′	NOUN
ejpam-86	46	10	is	be	AUX
ejpam-86	46	11	the	the	DET
ejpam-86	46	12	coefficient	coefficient	NOUN
ejpam-86	46	13	matrix	matrix	NOUN
ejpam-86	46	14	and	and	CCONJ
ejpam-86	46	15	e	e	NOUN
ejpam-86	46	16	=	=	PUNCT
ejpam-86	47	1	[	[	X
ejpam-86	47	2	ε1	ε1	NOUN
ejpam-86	47	3	,	,	PUNCT
ejpam-86	47	4	ε2	ε2	PROPN
ejpam-86	47	5	,	,	PUNCT
ejpam-86	47	6	·	·	PUNCT
ejpam-86	47	7	·	·	PUNCT
ejpam-86	47	8	·	·	PUNCT
ejpam-86	47	9	,	,	PUNCT
ejpam-86	47	10	εp	εp	ADP
ejpam-86	47	11	]	]	X
ejpam-86	47	12	=	=	PUNCT
ejpam-86	48	1	[	[	X
ejpam-86	48	2	ε(1	ε(1	NOUN
ejpam-86	48	3	)	)	PUNCT
ejpam-86	48	4	,	,	PUNCT
ejpam-86	48	5	ε(2	ε(2	NOUN
ejpam-86	48	6	)	)	PUNCT
ejpam-86	48	7	,	,	PUNCT
ejpam-86	48	8	·	·	PUNCT
ejpam-86	48	9	·	·	PUNCT
ejpam-86	48	10	·	·	PUNCT
ejpam-86	48	11	,	,	PUNCT
ejpam-86	48	12	ε(n)]′	ε(n)]′	PROPN
ejpam-86	48	13	is	be	AUX
ejpam-86	48	14	the	the	DET
ejpam-86	48	15	unobservable	unobservable	ADJ
ejpam-86	48	16	random	random	ADJ
ejpam-86	48	17	error	error	NOUN
ejpam-86	48	18	term	term	NOUN
ejpam-86	48	19	matrix	matrix	NOUN
ejpam-86	48	20	.	.	PUNCT
ejpam-86	49	1	the	the	DET
ejpam-86	49	2	error	error	NOUN
ejpam-86	49	3	terms	term	NOUN
ejpam-86	49	4	are	be	AUX
ejpam-86	49	5	assumed	assume	VERB
ejpam-86	49	6	to	to	PART
ejpam-86	49	7	be	be	AUX
ejpam-86	49	8	multivariate	multivariate	NOUN
ejpam-86	49	9	pe	pe	INTJ
ejpam-86	49	10	rather	rather	ADV
ejpam-86	49	11	than	than	ADP
ejpam-86	49	12	normal	normal	ADJ
ejpam-86	49	13	.	.	PUNCT
ejpam-86	50	1	applying	apply	VERB
ejpam-86	50	2	the	the	DET
ejpam-86	50	3	vector	vector	NOUN
ejpam-86	50	4	operator	operator	NOUN
ejpam-86	50	5	v	v	ADP
ejpam-86	50	6	ec	ec	PROPN
ejpam-86	50	7	(	(	PUNCT
ejpam-86	50	8	·	·	PUNCT
ejpam-86	50	9	)	)	PUNCT
ejpam-86	50	10	(	(	PUNCT
ejpam-86	50	11	[	[	X
ejpam-86	50	12	30	30	NUM
ejpam-86	50	13	,	,	PUNCT
ejpam-86	50	14	p.21	p.21	NOUN
ejpam-86	50	15	]	]	X
ejpam-86	50	16	)	)	PUNCT
ejpam-86	50	17	to	to	ADP
ejpam-86	50	18	(	(	PUNCT
ejpam-86	50	19	1.1	1.1	NUM
ejpam-86	50	20	)	)	PUNCT
ejpam-86	50	21	,	,	PUNCT
ejpam-86	50	22	the	the	DET
ejpam-86	50	23	multivariate	multivariate	NOUN
ejpam-86	50	24	regression	regression	NOUN
ejpam-86	50	25	model	model	NOUN
ejpam-86	50	26	can	can	AUX
ejpam-86	50	27	be	be	AUX
ejpam-86	50	28	transformed	transform	VERB
ejpam-86	50	29	to	to	ADP
ejpam-86	50	30	a	a	DET
ejpam-86	50	31	univariate	univariate	ADJ
ejpam-86	50	32	regression	regression	NOUN
ejpam-86	50	33	problem	problem	NOUN
ejpam-86	50	34	given	give	VERB
ejpam-86	50	35	by	by	ADP
ejpam-86	50	36	v	v	NUM
ejpam-86	50	37	ec(y	ec(y	NOUN
ejpam-86	50	38	′	′	NOUN
ejpam-86	50	39	)	)	PUNCT
ejpam-86	51	1	=	=	PUNCT
ejpam-86	51	2	v	v	PART
ejpam-86	51	3	ec(b′x	ec(b′x	NOUN
ejpam-86	51	4	′	′	NUM
ejpam-86	51	5	)	)	PUNCT
ejpam-86	52	1	+	+	CCONJ
ejpam-86	52	2	v	v	NUM
ejpam-86	52	3	ec(e′	ec(e′	NOUN
ejpam-86	52	4	)	)	PUNCT
ejpam-86	52	5	=	=	PUNCT
ejpam-86	53	1	(	(	PUNCT
ejpam-86	53	2	x	x	PROPN
ejpam-86	53	3	⊗	⊗	PROPN
ejpam-86	53	4	ip)v	ip)v	PROPN
ejpam-86	53	5	ec(b′	ec(b′	PROPN
ejpam-86	53	6	)	)	PUNCT
ejpam-86	53	7	+	+	CCONJ
ejpam-86	53	8	v	v	NUM
ejpam-86	53	9	ec(e′	ec(e′	NOUN
ejpam-86	53	10	)	)	PUNCT
ejpam-86	53	11	(	(	PUNCT
ejpam-86	53	12	1.2	1.2	NUM
ejpam-86	53	13	)	)	PUNCT
ejpam-86	53	14	or	or	CCONJ
ejpam-86	53	15	ynp×1	ynp×1	NUM
ejpam-86	53	16	=	=	PUNCT
ejpam-86	54	1	(	(	PUNCT
ejpam-86	54	2	x	x	PROPN
ejpam-86	54	3	⊗	⊗	PROPN
ejpam-86	54	4	ip)bpq×1	ip)bpq×1	PROPN
ejpam-86	54	5	+	+	CCONJ
ejpam-86	54	6	εnp×1	εnp×1	PROPN
ejpam-86	54	7	(	(	PUNCT
ejpam-86	54	8	1.3	1.3	NUM
ejpam-86	54	9	)	)	PUNCT
ejpam-86	54	10	m.	m.	NOUN
ejpam-86	54	11	liu	liu	PROPN
ejpam-86	54	12	and	and	CCONJ
ejpam-86	54	13	h.	h.	PROPN
ejpam-86	54	14	bozdogan	bozdogan	PROPN
ejpam-86	54	15	/	/	SYM
ejpam-86	54	16	eur	eur	PROPN
ejpam-86	54	17	.	.	PUNCT
ejpam-86	55	1	j.	j.	PROPN
ejpam-86	55	2	pure	pure	PROPN
ejpam-86	55	3	appl	appl	PROPN
ejpam-86	55	4	.	.	PROPN
ejpam-86	55	5	math	math	PROPN
ejpam-86	55	6	,	,	PUNCT
ejpam-86	55	7	1	1	NUM
ejpam-86	55	8	(	(	PUNCT
ejpam-86	55	9	2008	2008	NUM
ejpam-86	55	10	)	)	PUNCT
ejpam-86	55	11	,	,	PUNCT
ejpam-86	55	12	(	(	PUNCT
ejpam-86	55	13	4	4	NUM
ejpam-86	55	14	-	-	SYM
ejpam-86	55	15	37	37	NUM
ejpam-86	55	16	)	)	PUNCT
ejpam-86	55	17	6	6	NUM
ejpam-86	55	18	where	where	SCONJ
ejpam-86	55	19	ynp×1	ynp×1	NOUN
ejpam-86	55	20	=	=	PUNCT
ejpam-86	56	1	[	[	X
ejpam-86	56	2	y′(1),y	y′(1),y	NUM
ejpam-86	56	3	′	′	NOUN
ejpam-86	56	4	(	(	PUNCT
ejpam-86	56	5	2	2	NUM
ejpam-86	56	6	)	)	PUNCT
ejpam-86	56	7	,	,	PUNCT
ejpam-86	56	8	·	·	PUNCT
ejpam-86	56	9	·	·	PUNCT
ejpam-86	56	10	·	·	PUNCT
ejpam-86	56	11	,	,	PUNCT
ejpam-86	56	12	y′(n	y′(n	PROPN
ejpam-86	56	13	)	)	PUNCT
ejpam-86	56	14	]	]	PUNCT
ejpam-86	57	1	′	′	NUM
ejpam-86	57	2	,	,	PUNCT
ejpam-86	57	3	bpq×1	bpq×1	PROPN
ejpam-86	57	4	=	=	PUNCT
ejpam-86	58	1	[	[	X
ejpam-86	58	2	b′(1),b	b′(1),b	ADJ
ejpam-86	58	3	′	′	NUM
ejpam-86	58	4	(	(	PUNCT
ejpam-86	58	5	2	2	NUM
ejpam-86	58	6	)	)	PUNCT
ejpam-86	58	7	,	,	PUNCT
ejpam-86	58	8	·	·	PUNCT
ejpam-86	58	9	·	·	PUNCT
ejpam-86	58	10	·	·	PUNCT
ejpam-86	58	11	,	,	PUNCT
ejpam-86	58	12	b′(q)]′	b′(q)]′	PROPN
ejpam-86	58	13	and	and	CCONJ
ejpam-86	58	14	εnp×1	εnp×1	PROPN
ejpam-86	58	15	=	=	PUNCT
ejpam-86	59	1	[	[	X
ejpam-86	59	2	ε′(1	ε′(1	NOUN
ejpam-86	59	3	)	)	PUNCT
ejpam-86	59	4	,	,	PUNCT
ejpam-86	59	5	ε	ε	PROPN
ejpam-86	59	6	′	′	NUM
ejpam-86	59	7	(	(	PUNCT
ejpam-86	59	8	2	2	NUM
ejpam-86	59	9	)	)	PUNCT
ejpam-86	59	10	,	,	PUNCT
ejpam-86	59	11	·	·	PUNCT
ejpam-86	59	12	·	·	PUNCT
ejpam-86	59	13	·	·	PUNCT
ejpam-86	59	14	,	,	PUNCT
ejpam-86	59	15	ε′(n	ε′(n	PROPN
ejpam-86	59	16	)	)	PUNCT
ejpam-86	59	17	]	]	PUNCT
ejpam-86	59	18	′.	′.	NOUN
ejpam-86	59	19	ip	ip	VERB
ejpam-86	59	20	is	be	AUX
ejpam-86	59	21	the	the	DET
ejpam-86	59	22	(	(	PUNCT
ejpam-86	59	23	p	p	X
ejpam-86	59	24	×	×	PROPN
ejpam-86	59	25	p	p	X
ejpam-86	59	26	)	)	PUNCT
ejpam-86	59	27	dimensional	dimensional	ADJ
ejpam-86	59	28	identity	identity	NOUN
ejpam-86	59	29	matrix	matrix	NOUN
ejpam-86	59	30	.	.	PUNCT
ejpam-86	60	1	⊗	⊗	PROPN
ejpam-86	60	2	is	be	AUX
ejpam-86	60	3	kronecker	kronecker	NOUN
ejpam-86	60	4	product	product	NOUN
ejpam-86	60	5	operator	operator	NOUN
ejpam-86	60	6	(	(	PUNCT
ejpam-86	60	7	[	[	X
ejpam-86	60	8	30	30	NUM
ejpam-86	60	9	,	,	PUNCT
ejpam-86	60	10	p.12	p.12	NOUN
ejpam-86	60	11	]	]	X
ejpam-86	60	12	)	)	PUNCT
ejpam-86	60	13	.	.	PUNCT
ejpam-86	61	1	in	in	ADP
ejpam-86	61	2	this	this	DET
ejpam-86	61	3	paper	paper	NOUN
ejpam-86	61	4	,	,	PUNCT
ejpam-86	61	5	we	we	PRON
ejpam-86	61	6	let	let	VERB
ejpam-86	61	7	v	v	NOUN
ejpam-86	61	8	ec′	ec′	VERB
ejpam-86	61	9	(	(	PUNCT
ejpam-86	61	10	·	·	PUNCT
ejpam-86	61	11	)	)	PUNCT
ejpam-86	61	12	denote	denote	NOUN
ejpam-86	61	13	(	(	PUNCT
ejpam-86	61	14	v	v	X
ejpam-86	61	15	ec(·))′.	ec(·))′.	NOUN
ejpam-86	61	16	the	the	DET
ejpam-86	61	17	rest	rest	NOUN
ejpam-86	61	18	of	of	ADP
ejpam-86	61	19	the	the	DET
ejpam-86	61	20	paper	paper	NOUN
ejpam-86	61	21	is	be	AUX
ejpam-86	61	22	organized	organize	VERB
ejpam-86	61	23	as	as	SCONJ
ejpam-86	61	24	follows	follow	VERB
ejpam-86	61	25	.	.	PUNCT
ejpam-86	62	1	in	in	ADP
ejpam-86	62	2	section	section	NOUN
ejpam-86	62	3	2	2	NUM
ejpam-86	62	4	,	,	PUNCT
ejpam-86	62	5	we	we	PRON
ejpam-86	62	6	introduce	introduce	VERB
ejpam-86	62	7	the	the	DET
ejpam-86	62	8	multivariate	multivariate	NOUN
ejpam-86	62	9	pe	pe	ADP
ejpam-86	62	10	distribution	distribution	NOUN
ejpam-86	62	11	.	.	PUNCT
ejpam-86	63	1	in	in	ADP
ejpam-86	63	2	section	section	NOUN
ejpam-86	63	3	3	3	NUM
ejpam-86	63	4	,	,	PUNCT
ejpam-86	63	5	we	we	PRON
ejpam-86	63	6	study	study	VERB
ejpam-86	63	7	type	type	NOUN
ejpam-86	63	8	i	i	PROPN
ejpam-86	63	9	mvper	mvper	NOUN
ejpam-86	63	10	model	model	NOUN
ejpam-86	63	11	and	and	CCONJ
ejpam-86	63	12	in	in	ADP
ejpam-86	63	13	section	section	NOUN
ejpam-86	63	14	4	4	NUM
ejpam-86	63	15	we	we	PRON
ejpam-86	63	16	study	study	VERB
ejpam-86	63	17	type	type	NOUN
ejpam-86	63	18	ii	ii	PROPN
ejpam-86	63	19	mvper	mvper	PROPN
ejpam-86	63	20	model	model	NOUN
ejpam-86	63	21	which	which	PRON
ejpam-86	63	22	takes	take	VERB
ejpam-86	63	23	the	the	DET
ejpam-86	63	24	dependency	dependency	NOUN
ejpam-86	63	25	structure	structure	NOUN
ejpam-86	63	26	of	of	ADP
ejpam-86	63	27	the	the	DET
ejpam-86	63	28	data	datum	NOUN
ejpam-86	63	29	into	into	ADP
ejpam-86	63	30	account	account	NOUN
ejpam-86	63	31	.	.	PUNCT
ejpam-86	64	1	section	section	NOUN
ejpam-86	64	2	5	5	NUM
ejpam-86	64	3	gives	give	VERB
ejpam-86	64	4	the	the	DET
ejpam-86	64	5	derivation	derivation	NOUN
ejpam-86	64	6	of	of	ADP
ejpam-86	64	7	aic	aic	PROPN
ejpam-86	64	8	and	and	CCONJ
ejpam-86	64	9	icomp(ifim	icomp(ifim	NOUN
ejpam-86	64	10	)	)	PUNCT
ejpam-86	64	11	for	for	ADP
ejpam-86	64	12	these	these	DET
ejpam-86	64	13	two	two	NUM
ejpam-86	64	14	types	type	NOUN
ejpam-86	64	15	of	of	ADP
ejpam-86	64	16	mvper	mvper	NOUN
ejpam-86	64	17	models	model	NOUN
ejpam-86	64	18	.	.	PUNCT
ejpam-86	65	1	section	section	NOUN
ejpam-86	65	2	6	6	NUM
ejpam-86	65	3	outlines	outline	NOUN
ejpam-86	65	4	and	and	CCONJ
ejpam-86	65	5	presents	present	VERB
ejpam-86	65	6	the	the	DET
ejpam-86	65	7	general	general	ADJ
ejpam-86	65	8	background	background	NOUN
ejpam-86	65	9	of	of	ADP
ejpam-86	65	10	the	the	DET
ejpam-86	65	11	genetic	genetic	ADJ
ejpam-86	65	12	algorithms	algorithm	NOUN
ejpam-86	65	13	(	(	PUNCT
ejpam-86	65	14	gas	gas	NOUN
ejpam-86	65	15	)	)	PUNCT
ejpam-86	65	16	.	.	PUNCT
ejpam-86	66	1	in	in	ADP
ejpam-86	66	2	section	section	NOUN
ejpam-86	66	3	7	7	NUM
ejpam-86	66	4	,	,	PUNCT
ejpam-86	66	5	we	we	PRON
ejpam-86	66	6	give	give	VERB
ejpam-86	66	7	two	two	NUM
ejpam-86	66	8	simulation	simulation	NOUN
ejpam-86	66	9	examples	example	NOUN
ejpam-86	66	10	and	and	CCONJ
ejpam-86	66	11	a	a	DET
ejpam-86	66	12	real	real	ADJ
ejpam-86	66	13	model	model	NOUN
ejpam-86	66	14	selection	selection	NOUN
ejpam-86	66	15	example	example	NOUN
ejpam-86	66	16	on	on	ADP
ejpam-86	66	17	a	a	DET
ejpam-86	66	18	benchmark	benchmark	ADJ
ejpam-86	66	19	macro	macro	ADJ
ejpam-86	66	20	-	-	ADJ
ejpam-86	66	21	economic	economic	ADJ
ejpam-86	66	22	data	datum	NOUN
ejpam-86	66	23	in	in	ADP
ejpam-86	66	24	subset	subset	NOUN
ejpam-86	66	25	selection	selection	NOUN
ejpam-86	66	26	of	of	ADP
ejpam-86	66	27	best	good	ADJ
ejpam-86	66	28	predictors	predictor	NOUN
ejpam-86	66	29	under	under	ADP
ejpam-86	66	30	both	both	CCONJ
ejpam-86	66	31	the	the	DET
ejpam-86	66	32	type	type	NOUN
ejpam-86	66	33	i	i	PRON
ejpam-86	66	34	and	and	CCONJ
ejpam-86	66	35	type	type	PROPN
ejpam-86	66	36	ii	ii	PROPN
ejpam-86	66	37	mvper	mvper	NOUN
ejpam-86	66	38	models	model	NOUN
ejpam-86	66	39	to	to	PART
ejpam-86	66	40	illustrate	illustrate	VERB
ejpam-86	66	41	the	the	DET
ejpam-86	66	42	versatility	versatility	NOUN
ejpam-86	66	43	of	of	ADP
ejpam-86	66	44	our	our	PRON
ejpam-86	66	45	new	new	ADJ
ejpam-86	66	46	approach	approach	NOUN
ejpam-86	66	47	.	.	PUNCT
ejpam-86	67	1	section	section	NOUN
ejpam-86	67	2	8	8	NUM
ejpam-86	67	3	concludes	conclude	VERB
ejpam-86	67	4	the	the	DET
ejpam-86	67	5	paper	paper	NOUN
ejpam-86	67	6	.	.	PUNCT
ejpam-86	68	1	2	2	X
ejpam-86	68	2	.	.	X
ejpam-86	68	3	multivariate	multivariate	NOUN
ejpam-86	68	4	pe	pe	PROPN
ejpam-86	68	5	distributions	distribution	NOUN
ejpam-86	68	6	a	a	DET
ejpam-86	68	7	random	random	ADJ
ejpam-86	68	8	variable	variable	NOUN
ejpam-86	68	9	z	z	NOUN
ejpam-86	68	10	is	be	AUX
ejpam-86	68	11	pe	pe	PRON
ejpam-86	68	12	distributed	distribute	VERB
ejpam-86	68	13	if	if	SCONJ
ejpam-86	68	14	the	the	DET
ejpam-86	68	15	density	density	NOUN
ejpam-86	68	16	function	function	NOUN
ejpam-86	68	17	of	of	ADP
ejpam-86	68	18	z	z	PROPN
ejpam-86	68	19	is	be	AUX
ejpam-86	68	20	f(z;µ	f(z;µ	PROPN
ejpam-86	68	21	,	,	PUNCT
ejpam-86	68	22	σ	σ	PROPN
ejpam-86	68	23	,	,	PUNCT
ejpam-86	68	24	β	β	X
ejpam-86	68	25	)	)	PUNCT
ejpam-86	68	26	=	=	SYM
ejpam-86	68	27	1	1	NUM
ejpam-86	68	28	σ	σ	NUM
ejpam-86	68	29	γ	γ	X
ejpam-86	68	30	(	(	PUNCT
ejpam-86	68	31	1	1	NUM
ejpam-86	68	32	+	+	NUM
ejpam-86	68	33	1	1	NUM
ejpam-86	68	34	2β	2β	NUM
ejpam-86	68	35	)	)	PUNCT
ejpam-86	68	36	21	21	NUM
ejpam-86	69	1	+	+	SYM
ejpam-86	69	2	1	1	NUM
ejpam-86	69	3	2β	2β	NOUN
ejpam-86	69	4	exp	exp	NOUN
ejpam-86	69	5	(	(	PUNCT
ejpam-86	69	6	−1	−1	NOUN
ejpam-86	69	7	2	2	NUM
ejpam-86	69	8	∣∣∣∣	∣∣∣∣	PROPN
ejpam-86	69	9	z	z	NOUN
ejpam-86	69	10	−	−	PROPN
ejpam-86	69	11	µ	µ	PROPN
ejpam-86	69	12	σ	σ	X
ejpam-86	69	13	∣∣∣∣	∣∣∣∣	NOUN
ejpam-86	69	14	2β	2β	NOUN
ejpam-86	69	15	)	)	PUNCT
ejpam-86	69	16	,	,	PUNCT
ejpam-86	69	17	(	(	PUNCT
ejpam-86	69	18	2.1	2.1	NUM
ejpam-86	69	19	)	)	PUNCT
ejpam-86	69	20	where	where	SCONJ
ejpam-86	69	21	the	the	DET
ejpam-86	69	22	parameters	parameter	NOUN
ejpam-86	69	23	−∞	−∞	ADP
ejpam-86	69	24	<	<	X
ejpam-86	69	25	µ	µ	X
ejpam-86	69	26	<	<	X
ejpam-86	69	27	∞	∞	PROPN
ejpam-86	69	28	and	and	CCONJ
ejpam-86	69	29	σ	σ	PROPN
ejpam-86	69	30	>	>	X
ejpam-86	69	31	0	0	NUM
ejpam-86	69	32	are	be	AUX
ejpam-86	69	33	location	location	NOUN
ejpam-86	69	34	and	and	CCONJ
ejpam-86	69	35	scale	scale	NOUN
ejpam-86	69	36	parameters	parameter	NOUN
ejpam-86	69	37	,	,	PUNCT
ejpam-86	69	38	respectively	respectively	ADV
ejpam-86	69	39	,	,	PUNCT
ejpam-86	69	40	and	and	CCONJ
ejpam-86	69	41	β	β	X
ejpam-86	69	42	>	>	X
ejpam-86	69	43	0	0	PUNCT
ejpam-86	69	44	is	be	AUX
ejpam-86	69	45	the	the	DET
ejpam-86	69	46	shape	shape	NOUN
ejpam-86	69	47	parameter	parameter	NOUN
ejpam-86	69	48	,	,	PUNCT
ejpam-86	69	49	which	which	PRON
ejpam-86	69	50	is	be	AUX
ejpam-86	69	51	related	relate	VERB
ejpam-86	69	52	to	to	ADP
ejpam-86	69	53	the	the	DET
ejpam-86	69	54	kurtosis	kurtosis	NOUN
ejpam-86	69	55	parameter	parameter	NOUN
ejpam-86	69	56	.	.	PUNCT
ejpam-86	70	1	a	a	DET
ejpam-86	70	2	family	family	NOUN
ejpam-86	70	3	of	of	ADP
ejpam-86	70	4	unimodel	unimodel	PROPN
ejpam-86	70	5	symmetric	symmetric	ADJ
ejpam-86	70	6	curves	curve	NOUN
ejpam-86	70	7	with	with	ADP
ejpam-86	70	8	different	different	ADJ
ejpam-86	70	9	shapes	shape	NOUN
ejpam-86	70	10	for	for	ADP
ejpam-86	70	11	different	different	ADJ
ejpam-86	70	12	values	value	NOUN
ejpam-86	70	13	of	of	ADP
ejpam-86	70	14	β	β	X
ejpam-86	70	15	can	can	AUX
ejpam-86	70	16	be	be	AUX
ejpam-86	70	17	represented	represent	VERB
ejpam-86	70	18	by	by	ADP
ejpam-86	70	19	the	the	DET
ejpam-86	70	20	above	above	ADJ
ejpam-86	70	21	density	density	NOUN
ejpam-86	70	22	.	.	PUNCT
ejpam-86	71	1	when	when	SCONJ
ejpam-86	71	2	β	β	X
ejpam-86	71	3	=	=	NOUN
ejpam-86	71	4	0.5	0.5	NUM
ejpam-86	71	5	,	,	PUNCT
ejpam-86	71	6	the	the	DET
ejpam-86	71	7	laplace	laplace	NOUN
ejpam-86	71	8	distribution	distribution	NOUN
ejpam-86	71	9	,	,	PUNCT
ejpam-86	71	10	and	and	CCONJ
ejpam-86	71	11	when	when	SCONJ
ejpam-86	71	12	β	β	X
ejpam-86	71	13	→	→	SYM
ejpam-86	71	14	∞	∞	NUM
ejpam-86	71	15	the	the	DET
ejpam-86	71	16	uniform	uniform	ADJ
ejpam-86	71	17	distribution	distribution	NOUN
ejpam-86	71	18	arises	arise	VERB
ejpam-86	71	19	.	.	PUNCT
ejpam-86	72	1	particularly	particularly	ADV
ejpam-86	72	2	when	when	SCONJ
ejpam-86	72	3	β	β	X
ejpam-86	72	4	=	=	SYM
ejpam-86	72	5	1	1	NUM
ejpam-86	72	6	,	,	PUNCT
ejpam-86	72	7	the	the	DET
ejpam-86	72	8	density	density	NOUN
ejpam-86	72	9	becomes	become	VERB
ejpam-86	72	10	normal	normal	ADJ
ejpam-86	72	11	distribution	distribution	NOUN
ejpam-86	72	12	.	.	PUNCT
ejpam-86	73	1	so	so	ADV
ejpam-86	73	2	pe	pe	VERB
ejpam-86	73	3	distribution	distribution	NOUN
ejpam-86	73	4	can	can	AUX
ejpam-86	73	5	be	be	AUX
ejpam-86	73	6	seen	see	VERB
ejpam-86	73	7	as	as	ADP
ejpam-86	73	8	a	a	DET
ejpam-86	73	9	generalized	generalized	ADJ
ejpam-86	73	10	normal	normal	ADJ
ejpam-86	73	11	distribution	distribution	NOUN
ejpam-86	73	12	.	.	PUNCT
ejpam-86	74	1	standard	standard	ADJ
ejpam-86	74	2	pe	pe	NOUN
ejpam-86	74	3	distribution	distribution	NOUN
ejpam-86	74	4	is	be	AUX
ejpam-86	74	5	the	the	DET
ejpam-86	74	6	pe	pe	PROPN
ejpam-86	74	7	distribution	distribution	NOUN
ejpam-86	74	8	with	with	ADP
ejpam-86	74	9	mean	mean	NOUN
ejpam-86	74	10	zero	zero	NUM
ejpam-86	74	11	and	and	CCONJ
ejpam-86	74	12	σ	σ	NUM
ejpam-86	74	13	=	=	SYM
ejpam-86	74	14	1	1	X
ejpam-86	74	15	.	.	PUNCT
ejpam-86	74	16	according	accord	VERB
ejpam-86	74	17	to	to	ADP
ejpam-86	74	18	gómez	gómez	PROPN
ejpam-86	74	19	et	et	PROPN
ejpam-86	74	20	al	al	PROPN
ejpam-86	74	21	.	.	PUNCT
ejpam-86	75	1	(	(	PUNCT
ejpam-86	75	2	[	[	X
ejpam-86	75	3	20	20	NUM
ejpam-86	75	4	]	]	NUM
ejpam-86	75	5	)	)	PUNCT
ejpam-86	75	6	,	,	PUNCT
ejpam-86	75	7	a	a	DET
ejpam-86	75	8	multivariate	multivariate	NOUN
ejpam-86	75	9	generalization	generalization	NOUN
ejpam-86	75	10	of	of	ADP
ejpam-86	75	11	the	the	DET
ejpam-86	75	12	pe	pe	PROPN
ejpam-86	75	13	family	family	NOUN
ejpam-86	75	14	of	of	ADP
ejpam-86	75	15	distributions	distribution	NOUN
ejpam-86	75	16	,	,	PUNCT
ejpam-86	75	17	denoted	denote	VERB
ejpam-86	75	18	by	by	ADP
ejpam-86	75	19	pep(µ,σ	pep(µ,σ	PROPN
ejpam-86	75	20	,	,	PUNCT
ejpam-86	75	21	β	β	NOUN
ejpam-86	75	22	)	)	PUNCT
ejpam-86	75	23	,	,	PUNCT
ejpam-86	75	24	is	be	AUX
ejpam-86	75	25	defined	define	VERB
ejpam-86	75	26	as	as	ADP
ejpam-86	75	27	f(z;µ,σ	f(z;µ,σ	PROPN
ejpam-86	75	28	,	,	PUNCT
ejpam-86	75	29	β	β	NOUN
ejpam-86	75	30	)	)	PUNCT
ejpam-86	75	31	=	=	SYM
ejpam-86	75	32	pγ(p/2	pγ(p/2	X
ejpam-86	75	33	)	)	PUNCT
ejpam-86	76	1	πp/2γ(1	πp/2γ(1	NOUN
ejpam-86	77	1	+	+	CCONJ
ejpam-86	77	2	p	p	X
ejpam-86	77	3	2β	2β	NOUN
ejpam-86	77	4	)	)	PUNCT
ejpam-86	77	5	21+p/2β	21+p/2β	NUM
ejpam-86	78	1	|σ|−1/2	|σ|−1/2	NOUN
ejpam-86	78	2	exp(−1	exp(−1	X
ejpam-86	78	3	2	2	NUM
ejpam-86	78	4	(	(	PUNCT
ejpam-86	78	5	(	(	PUNCT
ejpam-86	78	6	z−	z−	X
ejpam-86	78	7	µ)′−1(z−	µ)′−1(z−	NOUN
ejpam-86	78	8	µ))β	µ))β	NOUN
ejpam-86	78	9	)	)	PUNCT
ejpam-86	78	10	(	(	PUNCT
ejpam-86	78	11	2.2	2.2	NUM
ejpam-86	78	12	)	)	PUNCT
ejpam-86	79	1	where	where	SCONJ
ejpam-86	79	2	z	z	NOUN
ejpam-86	79	3	=	=	PUNCT
ejpam-86	80	1	[	[	X
ejpam-86	80	2	z1	z1	PROPN
ejpam-86	80	3	,	,	PUNCT
ejpam-86	80	4	z2	z2	PROPN
ejpam-86	80	5	,	,	PUNCT
ejpam-86	80	6	·	·	PUNCT
ejpam-86	80	7	·	·	PUNCT
ejpam-86	80	8	·	·	PUNCT
ejpam-86	80	9	,	,	PUNCT
ejpam-86	80	10	zp]′	zp]′	PROPN
ejpam-86	80	11	is	be	AUX
ejpam-86	80	12	a	a	DET
ejpam-86	80	13	p	p	NOUN
ejpam-86	80	14	dimension	dimension	NOUN
ejpam-86	80	15	random	random	ADJ
ejpam-86	80	16	vector	vector	NOUN
ejpam-86	80	17	,	,	PUNCT
ejpam-86	80	18	µ∈	µ∈	PRON
ejpam-86	80	19	rp	rp	NOUN
ejpam-86	80	20	,	,	PUNCT
ejpam-86	80	21	σ	σ	PROPN
ejpam-86	80	22	is	be	AUX
ejpam-86	80	23	a	a	DET
ejpam-86	80	24	(	(	PUNCT
ejpam-86	80	25	p	p	X
ejpam-86	80	26	×	×	PROPN
ejpam-86	80	27	p	p	X
ejpam-86	80	28	)	)	PUNCT
ejpam-86	80	29	positive	positive	ADJ
ejpam-86	80	30	definite	definite	ADJ
ejpam-86	80	31	symmetric	symmetric	ADJ
ejpam-86	80	32	matrix	matrix	NOUN
ejpam-86	80	33	,	,	PUNCT
ejpam-86	80	34	and	and	CCONJ
ejpam-86	80	35	β	β	X
ejpam-86	80	36	∈	∈	PROPN
ejpam-86	80	37	(	(	PUNCT
ejpam-86	80	38	0,∞	0,∞	NOUN
ejpam-86	80	39	)	)	PUNCT
ejpam-86	80	40	is	be	AUX
ejpam-86	80	41	the	the	DET
ejpam-86	80	42	shape	shape	NOUN
ejpam-86	80	43	parameter	parameter	NOUN
ejpam-86	80	44	.	.	PUNCT
ejpam-86	81	1	if	if	SCONJ
ejpam-86	81	2	β	β	PROPN
ejpam-86	81	3	is	be	AUX
ejpam-86	81	4	given	give	VERB
ejpam-86	81	5	,	,	PUNCT
ejpam-86	81	6	this	this	DET
ejpam-86	81	7	multivariate	multivariate	NOUN
ejpam-86	81	8	generalized	generalize	VERB
ejpam-86	81	9	pe	pe	NOUN
ejpam-86	81	10	distribution	distribution	NOUN
ejpam-86	81	11	is	be	AUX
ejpam-86	81	12	actually	actually	ADV
ejpam-86	81	13	a	a	DET
ejpam-86	81	14	multivariate	multivariate	NOUN
ejpam-86	81	15	elliptically	elliptically	ADV
ejpam-86	81	16	contoured	contour	VERB
ejpam-86	81	17	(	(	PUNCT
ejpam-86	81	18	ec	ec	NOUN
ejpam-86	81	19	)	)	PUNCT
ejpam-86	81	20	distribution	distribution	NOUN
ejpam-86	81	21	ecp(µ,σ	ecp(µ,σ	PROPN
ejpam-86	81	22	,	,	PUNCT
ejpam-86	81	23	g	g	NOUN
ejpam-86	81	24	)	)	PUNCT
ejpam-86	81	25	with	with	ADP
ejpam-86	81	26	the	the	DET
ejpam-86	81	27	probability	probability	NOUN
ejpam-86	81	28	density	density	NOUN
ejpam-86	81	29	generator	generator	NOUN
ejpam-86	81	30	g(t	g(t	PROPN
ejpam-86	81	31	)	)	PUNCT
ejpam-86	82	1	=	=	SYM
ejpam-86	82	2	exp(−1	exp(−1	PUNCT
ejpam-86	82	3	2	2	NUM
ejpam-86	82	4	t	t	NOUN
ejpam-86	82	5	β	β	NOUN
ejpam-86	82	6	)	)	PUNCT
ejpam-86	82	7	(	(	PUNCT
ejpam-86	82	8	[	[	X
ejpam-86	82	9	18	18	NUM
ejpam-86	82	10	,	,	PUNCT
ejpam-86	82	11	p.46	p.46	NOUN
ejpam-86	82	12	]	]	NOUN
ejpam-86	82	13	)	)	PUNCT
ejpam-86	82	14	.	.	PUNCT
ejpam-86	83	1	further	far	ADV
ejpam-86	83	2	,	,	PUNCT
ejpam-86	83	3	it	it	PRON
ejpam-86	83	4	is	be	AUX
ejpam-86	83	5	a	a	DET
ejpam-86	83	6	special	special	ADJ
ejpam-86	83	7	case	case	NOUN
ejpam-86	83	8	of	of	ADP
ejpam-86	83	9	symmetric	symmetric	ADJ
ejpam-86	83	10	kotz	kotz	PROPN
ejpam-86	83	11	type	type	NOUN
ejpam-86	83	12	distribution	distribution	NOUN
ejpam-86	83	13	(	(	PUNCT
ejpam-86	83	14	[	[	X
ejpam-86	83	15	18	18	NUM
ejpam-86	83	16	,	,	PUNCT
ejpam-86	83	17	p.76	p.76	NOUN
ejpam-86	83	18	]	]	X
ejpam-86	83	19	)	)	PUNCT
ejpam-86	83	20	with	with	ADP
ejpam-86	83	21	n	n	NOUN
ejpam-86	83	22	=	=	SYM
ejpam-86	83	23	1	1	NUM
ejpam-86	83	24	.	.	PUNCT
ejpam-86	83	25	when	when	SCONJ
ejpam-86	83	26	p	p	NOUN
ejpam-86	83	27	=	=	NOUN
ejpam-86	83	28	1	1	NUM
ejpam-86	83	29	,	,	PUNCT
ejpam-86	83	30	(	(	PUNCT
ejpam-86	83	31	2.2	2.2	NUM
ejpam-86	83	32	)	)	PUNCT
ejpam-86	83	33	reduces	reduce	VERB
ejpam-86	83	34	to	to	ADP
ejpam-86	83	35	(	(	PUNCT
ejpam-86	83	36	2.1	2.1	NUM
ejpam-86	83	37	)	)	PUNCT
ejpam-86	83	38	.	.	PUNCT
ejpam-86	84	1	sánchez	sánchez	PROPN
ejpam-86	84	2	-	-	PUNCT
ejpam-86	84	3	manzano	manzano	PROPN
ejpam-86	84	4	et	et	PROPN
ejpam-86	84	5	al	al	PROPN
ejpam-86	84	6	.	.	PUNCT
ejpam-86	85	1	(	(	PUNCT
ejpam-86	85	2	[	[	X
ejpam-86	85	3	31	31	NUM
ejpam-86	85	4	]	]	PUNCT
ejpam-86	85	5	)	)	PUNCT
ejpam-86	85	6	gave	give	VERB
ejpam-86	85	7	the	the	DET
ejpam-86	85	8	definition	definition	NOUN
ejpam-86	85	9	of	of	ADP
ejpam-86	85	10	matrix	matrix	NOUN
ejpam-86	85	11	variate	variate	NOUN
ejpam-86	85	12	pe	pe	NOUN
ejpam-86	85	13	distribution	distribution	NOUN
ejpam-86	85	14	.	.	PUNCT
ejpam-86	86	1	a	a	DET
ejpam-86	86	2	random	random	ADJ
ejpam-86	86	3	(	(	PUNCT
ejpam-86	86	4	n	n	NUM
ejpam-86	86	5	×	×	NOUN
ejpam-86	86	6	p	p	NOUN
ejpam-86	86	7	)	)	PUNCT
ejpam-86	86	8	matrix	matrix	NOUN
ejpam-86	86	9	z	z	NOUN
ejpam-86	86	10	has	have	VERB
ejpam-86	86	11	a	a	DET
ejpam-86	86	12	(	(	PUNCT
ejpam-86	86	13	n	n	NUM
ejpam-86	86	14	×	×	NOUN
ejpam-86	86	15	p)-variate	p)-variate	ADV
ejpam-86	86	16	pe	pe	X
ejpam-86	86	17	distribution	distribution	NOUN
ejpam-86	86	18	,	,	PUNCT
ejpam-86	86	19	denoted	denote	VERB
ejpam-86	86	20	as	as	ADP
ejpam-86	86	21	z	z	NOUN
ejpam-86	86	22	∼	∼	NOUN
ejpam-86	86	23	mpep×n(m	mpep×n(m	PROPN
ejpam-86	86	24	,	,	PUNCT
ejpam-86	86	25	φ	φ	PROPN
ejpam-86	86	26	,	,	PUNCT
ejpam-86	86	27	σ	σ	PROPN
ejpam-86	86	28	,	,	PUNCT
ejpam-86	86	29	β	β	NOUN
ejpam-86	86	30	)	)	PUNCT
ejpam-86	86	31	with	with	ADP
ejpam-86	86	32	parameters	parameter	NOUN
ejpam-86	86	33	m	m	VERB
ejpam-86	86	34	,	,	PUNCT
ejpam-86	86	35	a	a	DET
ejpam-86	86	36	(	(	PUNCT
ejpam-86	86	37	n×	n×	NOUN
ejpam-86	86	38	p	p	NOUN
ejpam-86	86	39	)	)	PUNCT
ejpam-86	86	40	matrix	matrix	NOUN
ejpam-86	86	41	;	;	PUNCT
ejpam-86	86	42	φ	φ	PROPN
ejpam-86	86	43	,	,	PUNCT
ejpam-86	86	44	a	a	PRON
ejpam-86	86	45	(	(	PUNCT
ejpam-86	86	46	n×	n×	NOUN
ejpam-86	86	47	n	n	CCONJ
ejpam-86	86	48	)	)	PUNCT
ejpam-86	86	49	positive	positive	ADJ
ejpam-86	86	50	definite	definite	ADJ
ejpam-86	86	51	matrix	matrix	NOUN
ejpam-86	86	52	;	;	PUNCT
ejpam-86	86	53	σ	σ	PROPN
ejpam-86	86	54	,	,	PUNCT
ejpam-86	86	55	a	a	DET
ejpam-86	86	56	(	(	PUNCT
ejpam-86	86	57	p×	p×	NOUN
ejpam-86	86	58	p	p	NOUN
ejpam-86	86	59	)	)	PUNCT
ejpam-86	86	60	positive	positive	ADJ
ejpam-86	86	61	definite	definite	ADJ
ejpam-86	86	62	matrix	matrix	NOUN
ejpam-86	86	63	and	and	CCONJ
ejpam-86	86	64	β	β	X
ejpam-86	86	65	∈	∈	PROPN
ejpam-86	86	66	(	(	PUNCT
ejpam-86	86	67	0,∞	0,∞	NOUN
ejpam-86	86	68	)	)	PUNCT
ejpam-86	86	69	if	if	SCONJ
ejpam-86	86	70	v	v	X
ejpam-86	86	71	ec(z	ec(z	NUM
ejpam-86	86	72	′	′	NOUN
ejpam-86	86	73	)	)	PUNCT
ejpam-86	86	74	∼	∼	NOUN
ejpam-86	86	75	penp(v	penp(v	NOUN
ejpam-86	86	76	ec(m	ec(m	X
ejpam-86	86	77	′),φ⊗	′),φ⊗	PROPN
ejpam-86	86	78	σ	σ	PROPN
ejpam-86	86	79	,	,	PUNCT
ejpam-86	86	80	β	β	NOUN
ejpam-86	86	81	)	)	PUNCT
ejpam-86	86	82	.	.	PUNCT
ejpam-86	87	1	(	(	PUNCT
ejpam-86	87	2	2.3	2.3	NUM
ejpam-86	87	3	)	)	PUNCT
ejpam-86	87	4	m.	m.	NOUN
ejpam-86	87	5	liu	liu	PROPN
ejpam-86	87	6	and	and	CCONJ
ejpam-86	87	7	h.	h.	PROPN
ejpam-86	87	8	bozdogan	bozdogan	PROPN
ejpam-86	87	9	/	/	SYM
ejpam-86	87	10	eur	eur	PROPN
ejpam-86	87	11	.	.	PUNCT
ejpam-86	88	1	j.	j.	PROPN
ejpam-86	88	2	pure	pure	PROPN
ejpam-86	88	3	appl	appl	PROPN
ejpam-86	88	4	.	.	PROPN
ejpam-86	88	5	math	math	PROPN
ejpam-86	88	6	,	,	PUNCT
ejpam-86	88	7	1	1	NUM
ejpam-86	88	8	(	(	PUNCT
ejpam-86	88	9	2008	2008	NUM
ejpam-86	88	10	)	)	PUNCT
ejpam-86	88	11	,	,	PUNCT
ejpam-86	88	12	(	(	PUNCT
ejpam-86	88	13	4	4	NUM
ejpam-86	88	14	-	-	SYM
ejpam-86	88	15	37	37	NUM
ejpam-86	88	16	)	)	PUNCT
ejpam-86	88	17	7	7	NUM
ejpam-86	88	18	the	the	DET
ejpam-86	88	19	matrix	matrix	NOUN
ejpam-86	88	20	form	form	NOUN
ejpam-86	88	21	of	of	ADP
ejpam-86	88	22	the	the	DET
ejpam-86	88	23	density	density	NOUN
ejpam-86	88	24	function	function	NOUN
ejpam-86	88	25	of	of	ADP
ejpam-86	88	26	z	z	PROPN
ejpam-86	88	27	is	be	AUX
ejpam-86	88	28	then	then	ADV
ejpam-86	88	29	given	give	VERB
ejpam-86	88	30	by	by	ADP
ejpam-86	88	31	f(z;m	f(z;m	PROPN
ejpam-86	88	32	,	,	PUNCT
ejpam-86	88	33	φ	φ	PROPN
ejpam-86	88	34	,	,	PUNCT
ejpam-86	88	35	σ	σ	PROPN
ejpam-86	88	36	,	,	PUNCT
ejpam-86	88	37	β	β	X
ejpam-86	88	38	)	)	PUNCT
ejpam-86	88	39	=	=	SYM
ejpam-86	89	1	k|φ|−p/2|σ|−n/2	k|φ|−p/2|σ|−n/2	PROPN
ejpam-86	89	2	exp(−1	exp(−1	X
ejpam-86	89	3	2	2	NUM
ejpam-86	89	4	(	(	PUNCT
ejpam-86	89	5	tr((z	tr((z	NOUN
ejpam-86	89	6	−m)′−1(z	−m)′−1(z	PROPN
ejpam-86	89	7	−m)σ−1))β	−m)σ−1))β	NOUN
ejpam-86	89	8	)	)	PUNCT
ejpam-86	89	9	(	(	PUNCT
ejpam-86	89	10	2.4	2.4	NUM
ejpam-86	89	11	)	)	PUNCT
ejpam-86	89	12	where	where	SCONJ
ejpam-86	89	13	k	k	NOUN
ejpam-86	89	14	=	=	PUNCT
ejpam-86	89	15	npγ(np/2	npγ(np/2	ADJ
ejpam-86	89	16	)	)	PUNCT
ejpam-86	89	17	πnp/2γ(1	πnp/2γ(1	NOUN
ejpam-86	89	18	+	+	CCONJ
ejpam-86	89	19	np/2β)21+np/2β	np/2β)21+np/2β	NOUN
ejpam-86	89	20	.	.	PUNCT
ejpam-86	90	1	if	if	SCONJ
ejpam-86	90	2	z	z	PROPN
ejpam-86	90	3	∼mpep×n(m	∼mpep×n(m	PROPN
ejpam-86	90	4	,	,	PUNCT
ejpam-86	90	5	φ	φ	PROPN
ejpam-86	90	6	,	,	PUNCT
ejpam-86	90	7	σ	σ	PROPN
ejpam-86	90	8	,	,	PUNCT
ejpam-86	90	9	β	β	NOUN
ejpam-86	90	10	)	)	PUNCT
ejpam-86	90	11	,	,	PUNCT
ejpam-86	90	12	then	then	ADV
ejpam-86	90	13	z	z	NOUN
ejpam-86	90	14	′	′	NUM
ejpam-86	90	15	∼mpen×p(m	∼mpen×p(m	PROPN
ejpam-86	90	16	′,σ	′,σ	NOUN
ejpam-86	90	17	,	,	PUNCT
ejpam-86	90	18	φ	φ	NUM
ejpam-86	90	19	,	,	PUNCT
ejpam-86	90	20	β	β	NOUN
ejpam-86	90	21	)	)	PUNCT
ejpam-86	90	22	.	.	PUNCT
ejpam-86	91	1	some	some	DET
ejpam-86	91	2	probabilistic	probabilistic	ADJ
ejpam-86	91	3	characteristics	characteristic	NOUN
ejpam-86	91	4	of	of	ADP
ejpam-86	91	5	z	z	NOUN
ejpam-86	91	6	are	be	AUX
ejpam-86	91	7	given	give	VERB
ejpam-86	91	8	by	by	ADP
ejpam-86	91	9	:	:	PUNCT
ejpam-86	91	10	e[z	e[z	X
ejpam-86	91	11	]	]	X
ejpam-86	91	12	=	=	PUNCT
ejpam-86	91	13	m	m	PROPN
ejpam-86	91	14	,	,	PUNCT
ejpam-86	91	15	v	v	X
ejpam-86	91	16	ar[v	ar[v	ADJ
ejpam-86	91	17	ec(z	ec(z	NUM
ejpam-86	91	18	′	′	NUM
ejpam-86	91	19	)	)	PUNCT
ejpam-86	91	20	]	]	PUNCT
ejpam-86	92	1	=	=	SYM
ejpam-86	92	2	21	21	NUM
ejpam-86	92	3	/	/	SYM
ejpam-86	92	4	βγ(np+2	βγ(np+2	PROPN
ejpam-86	92	5	2β	2β	NUM
ejpam-86	92	6	)	)	PUNCT
ejpam-86	92	7	npγ(np	npγ(np	NOUN
ejpam-86	92	8	2β	2β	NUM
ejpam-86	92	9	)	)	PUNCT
ejpam-86	92	10	(	(	PUNCT
ejpam-86	92	11	φ⊗	φ⊗	NOUN
ejpam-86	92	12	σ	σ	PROPN
ejpam-86	92	13	)	)	PUNCT
ejpam-86	92	14	,	,	PUNCT
ejpam-86	92	15	γ1[z	γ1[z	PROPN
ejpam-86	92	16	]	]	X
ejpam-86	93	1	=	=	SYM
ejpam-86	93	2	0	0	NUM
ejpam-86	93	3	,	,	PUNCT
ejpam-86	93	4	γ2[z	γ2[z	NOUN
ejpam-86	93	5	]	]	X
ejpam-86	93	6	=	=	SYM
ejpam-86	93	7	(	(	PUNCT
ejpam-86	93	8	np)2γ(np	np)2γ(np	ADV
ejpam-86	93	9	2β	2β	NOUN
ejpam-86	93	10	)	)	PUNCT
ejpam-86	93	11	γ(np+4	γ(np+4	PROPN
ejpam-86	93	12	2β	2β	NOUN
ejpam-86	93	13	)	)	PUNCT
ejpam-86	93	14	γ2(np+2	γ2(np+2	ADJ
ejpam-86	93	15	2β	2β	NOUN
ejpam-86	93	16	)	)	PUNCT
ejpam-86	94	1	−	−	PROPN
ejpam-86	94	2	np(np+	np(np+	PROPN
ejpam-86	94	3	2	2	NUM
ejpam-86	94	4	)	)	PUNCT
ejpam-86	94	5	,	,	PUNCT
ejpam-86	94	6	e[(tr((z	e[(tr((z	PROPN
ejpam-86	94	7	−m)′−1(z	−m)′−1(z	PROPN
ejpam-86	94	8	−m)σ−1))s	−m)σ−1))s	X
ejpam-86	94	9	]	]	PUNCT
ejpam-86	94	10	=	=	SYM
ejpam-86	94	11	2s	2s	NUM
ejpam-86	94	12	/	/	SYM
ejpam-86	94	13	βγ(np+2s	βγ(np+2	NOUN
ejpam-86	94	14	2β	2β	NOUN
ejpam-86	94	15	)	)	PUNCT
ejpam-86	94	16	γ(np	γ(np	ADP
ejpam-86	94	17	2β	2β	NUM
ejpam-86	94	18	)	)	PUNCT
ejpam-86	94	19	,	,	PUNCT
ejpam-86	94	20	where	where	SCONJ
ejpam-86	94	21	s	s	NOUN
ejpam-86	94	22	is	be	AUX
ejpam-86	94	23	a	a	DET
ejpam-86	94	24	positive	positive	ADJ
ejpam-86	94	25	integer	integer	NOUN
ejpam-86	94	26	,	,	PUNCT
ejpam-86	94	27	γ1	γ1	NOUN
ejpam-86	94	28	and	and	CCONJ
ejpam-86	94	29	γ2	γ2	PROPN
ejpam-86	94	30	are	be	AUX
ejpam-86	94	31	multidimensional	multidimensional	ADJ
ejpam-86	94	32	asymmetry	asymmetry	NOUN
ejpam-86	94	33	(	(	PUNCT
ejpam-86	94	34	skewness	skewness	NOUN
ejpam-86	94	35	)	)	PUNCT
ejpam-86	94	36	and	and	CCONJ
ejpam-86	94	37	kurtosis	kurtosis	VERB
ejpam-86	94	38	coefficients	coefficient	NOUN
ejpam-86	94	39	(	(	PUNCT
ejpam-86	94	40	[	[	X
ejpam-86	94	41	26	26	NUM
ejpam-86	94	42	]	]	PUNCT
ejpam-86	94	43	)	)	PUNCT
ejpam-86	94	44	defined	define	VERB
ejpam-86	94	45	as	as	ADP
ejpam-86	94	46	γ1[z	γ1[z	NOUN
ejpam-86	94	47	]	]	X
ejpam-86	95	1	=	=	SYM
ejpam-86	95	2	e[((v	e[((v	NUM
ejpam-86	95	3	ec(z	ec(z	NUM
ejpam-86	96	1	′)−	′)−	PROPN
ejpam-86	96	2	v	v	ADP
ejpam-86	96	3	ec(m	ec(m	NOUN
ejpam-86	96	4	′))′(v	′))′(v	VERB
ejpam-86	96	5	ar[v	ar[v	ADP
ejpam-86	96	6	ec(z	ec(z	NUM
ejpam-86	96	7	′−1(v	′−1(v	PROPN
ejpam-86	96	8	ec(z	ec(z	NUM
ejpam-86	96	9	′)−	′)−	PROPN
ejpam-86	96	10	v	v	ADP
ejpam-86	96	11	ec(m	ec(m	X
ejpam-86	96	12	′3	′3	PROPN
ejpam-86	96	13	]	]	PUNCT
ejpam-86	96	14	,	,	PUNCT
ejpam-86	96	15	γ2[z	γ2[z	PROPN
ejpam-86	96	16	]	]	X
ejpam-86	96	17	=	=	SYM
ejpam-86	96	18	e[((v	e[((v	NUM
ejpam-86	96	19	ec(z	ec(z	NUM
ejpam-86	96	20	′)−	′)−	PROPN
ejpam-86	96	21	v	v	ADP
ejpam-86	96	22	ec(m	ec(m	NOUN
ejpam-86	96	23	′))′(v	′))′(v	VERB
ejpam-86	96	24	ar[v	ar[v	ADP
ejpam-86	96	25	ec(z	ec(z	NUM
ejpam-86	96	26	′−1(v	′−1(v	PROPN
ejpam-86	96	27	ec(z	ec(z	NUM
ejpam-86	96	28	′)−	′)−	PROPN
ejpam-86	96	29	v	v	ADP
ejpam-86	96	30	ec(m	ec(m	X
ejpam-86	96	31	′2	′2	X
ejpam-86	96	32	]	]	X
ejpam-86	96	33	−np(np+	−np(np+	NOUN
ejpam-86	96	34	2	2	NUM
ejpam-86	96	35	)	)	PUNCT
ejpam-86	96	36	.	.	PUNCT
ejpam-86	97	1	from	from	ADP
ejpam-86	97	2	the	the	DET
ejpam-86	97	3	above	above	NOUN
ejpam-86	97	4	,	,	PUNCT
ejpam-86	97	5	it	it	PRON
ejpam-86	97	6	is	be	AUX
ejpam-86	97	7	easy	easy	ADJ
ejpam-86	97	8	to	to	PART
ejpam-86	97	9	show	show	VERB
ejpam-86	97	10	that	that	SCONJ
ejpam-86	97	11	e[(v	e[(v	NOUN
ejpam-86	97	12	ec(z	ec(z	PUNCT
ejpam-86	98	1	′)−	′)−	PROPN
ejpam-86	98	2	v	v	ADP
ejpam-86	98	3	ec(m	ec(m	NOUN
ejpam-86	98	4	′))′(v	′))′(v	VERB
ejpam-86	98	5	ar[v	ar[v	ADP
ejpam-86	98	6	ec(z	ec(z	NUM
ejpam-86	98	7	′−1(v	′−1(v	PROPN
ejpam-86	98	8	ec(z	ec(z	NUM
ejpam-86	98	9	′)−	′)−	PROPN
ejpam-86	98	10	v	v	ADP
ejpam-86	98	11	ec(m	ec(m	NOUN
ejpam-86	98	12	′	′	NUM
ejpam-86	98	13	)	)	PUNCT
ejpam-86	98	14	)	)	PUNCT
ejpam-86	98	15	]	]	PUNCT
ejpam-86	99	1	=	=	PUNCT
ejpam-86	99	2	np	np	INTJ
ejpam-86	99	3	.	.	PUNCT
ejpam-86	99	4	for	for	ADP
ejpam-86	99	5	a	a	DET
ejpam-86	99	6	given	give	VERB
ejpam-86	99	7	β	β	PRON
ejpam-86	99	8	>	>	X
ejpam-86	99	9	0	0	NUM
ejpam-86	99	10	,	,	PUNCT
ejpam-86	99	11	z	z	PROPN
ejpam-86	99	12	actually	actually	ADV
ejpam-86	99	13	is	be	AUX
ejpam-86	99	14	a	a	DET
ejpam-86	99	15	matrix	matrix	NOUN
ejpam-86	99	16	variate	variate	NOUN
ejpam-86	99	17	ec	ec	NOUN
ejpam-86	99	18	distribution	distribution	NOUN
ejpam-86	99	19	denoted	denote	VERB
ejpam-86	99	20	as	as	ADP
ejpam-86	99	21	z	z	NOUN
ejpam-86	99	22	∼	∼	NOUN
ejpam-86	99	23	en	en	ADP
ejpam-86	99	24	,	,	PUNCT
ejpam-86	99	25	p(m	p(m	NOUN
ejpam-86	99	26	,	,	PUNCT
ejpam-86	99	27	φ⊗	φ⊗	NOUN
ejpam-86	99	28	σ	σ	PROPN
ejpam-86	99	29	,	,	PUNCT
ejpam-86	99	30	ψ	ψ	NOUN
ejpam-86	99	31	)	)	PUNCT
ejpam-86	99	32	by	by	ADP
ejpam-86	99	33	gupta	gupta	PROPN
ejpam-86	99	34	and	and	CCONJ
ejpam-86	99	35	varga	varga	PROPN
ejpam-86	99	36	(	(	PUNCT
ejpam-86	99	37	[	[	X
ejpam-86	99	38	21	21	NUM
ejpam-86	99	39	,	,	PUNCT
ejpam-86	99	40	p.20	p.20	X
ejpam-86	99	41	,	,	PUNCT
ejpam-86	99	42	p.26	p.26	ADJ
ejpam-86	99	43	]	]	PUNCT
ejpam-86	99	44	)	)	PUNCT
ejpam-86	99	45	with	with	ADP
ejpam-86	99	46	h(t	h(t	PROPN
ejpam-86	99	47	)	)	PUNCT
ejpam-86	99	48	=	=	SYM
ejpam-86	100	1	k	k	PROPN
ejpam-86	100	2	exp(−tβ/2	exp(−tβ/2	PROPN
ejpam-86	100	3	)	)	PUNCT
ejpam-86	100	4	.	.	PUNCT
ejpam-86	101	1	for	for	ADP
ejpam-86	101	2	n	n	NOUN
ejpam-86	101	3	=	=	SYM
ejpam-86	101	4	1	1	NUM
ejpam-86	101	5	and	and	CCONJ
ejpam-86	101	6	φ	φ	NUM
ejpam-86	101	7	=	=	PUNCT
ejpam-86	101	8	in	in	ADP
ejpam-86	101	9	,	,	PUNCT
ejpam-86	101	10	z	z	NOUN
ejpam-86	101	11	′	′	NOUN
ejpam-86	102	1	and	and	CCONJ
ejpam-86	102	2	z	z	AUX
ejpam-86	102	3	have	have	VERB
ejpam-86	102	4	the	the	DET
ejpam-86	102	5	same	same	ADJ
ejpam-86	102	6	distribution	distribution	NOUN
ejpam-86	102	7	.	.	PUNCT
ejpam-86	103	1	the	the	DET
ejpam-86	103	2	parameters	parameter	NOUN
ejpam-86	103	3	in	in	ADP
ejpam-86	103	4	the	the	DET
ejpam-86	103	5	definition	definition	NOUN
ejpam-86	103	6	of	of	ADP
ejpam-86	103	7	a	a	DET
ejpam-86	103	8	matrix	matrix	NOUN
ejpam-86	103	9	multivariate	multivariate	NOUN
ejpam-86	103	10	pe	pe	ADP
ejpam-86	103	11	distribution	distribution	NOUN
ejpam-86	103	12	are	be	AUX
ejpam-86	103	13	not	not	PART
ejpam-86	103	14	uniquely	uniquely	ADV
ejpam-86	103	15	defined	define	VERB
ejpam-86	103	16	as	as	SCONJ
ejpam-86	103	17	shown	show	VERB
ejpam-86	103	18	in	in	ADP
ejpam-86	103	19	the	the	DET
ejpam-86	103	20	following	follow	VERB
ejpam-86	103	21	theorem	theorem	NOUN
ejpam-86	103	22	.	.	PUNCT
ejpam-86	103	23	theorem	theorem	VERB
ejpam-86	103	24	2.1	2.1	NUM
ejpam-86	103	25	.	.	PUNCT
ejpam-86	104	1	letz	letz	PROPN
ejpam-86	104	2	∼mpep×n(m1,φ1,σ1	∼mpep×n(m1,φ1,σ1	PROPN
ejpam-86	104	3	,	,	PUNCT
ejpam-86	104	4	β1	β1	PROPN
ejpam-86	104	5	)	)	PUNCT
ejpam-86	104	6	and	and	CCONJ
ejpam-86	104	7	at	at	ADP
ejpam-86	104	8	the	the	DET
ejpam-86	104	9	same	same	ADJ
ejpam-86	104	10	timez	timez	NOUN
ejpam-86	104	11	∼mpep×n(m2,φ2,σ2	∼mpep×n(m2,φ2,σ2	NOUN
ejpam-86	104	12	,	,	PUNCT
ejpam-86	104	13	β2	β2	NOUN
ejpam-86	104	14	)	)	PUNCT
ejpam-86	104	15	.	.	PUNCT
ejpam-86	105	1	if	if	SCONJ
ejpam-86	105	2	z	z	NOUN
ejpam-86	105	3	is	be	AUX
ejpam-86	105	4	non	non	ADJ
ejpam-86	105	5	-	-	ADJ
ejpam-86	105	6	degenerate	degenerate	ADJ
ejpam-86	105	7	,	,	PUNCT
ejpam-86	105	8	then	then	ADV
ejpam-86	105	9	there	there	PRON
ejpam-86	105	10	exist	exist	VERB
ejpam-86	105	11	positive	positive	ADJ
ejpam-86	105	12	constant	constant	ADJ
ejpam-86	105	13	c	c	NOUN
ejpam-86	105	14	such	such	ADJ
ejpam-86	105	15	that	that	SCONJ
ejpam-86	105	16	m2	m2	PROPN
ejpam-86	105	17	=	=	PROPN
ejpam-86	105	18	m1	m1	PROPN
ejpam-86	105	19	,	,	PUNCT
ejpam-86	105	20	σ2	σ2	PROPN
ejpam-86	105	21	=	=	SYM
ejpam-86	105	22	cς1	cς1	PROPN
ejpam-86	105	23	,	,	PUNCT
ejpam-86	105	24	φ2	φ2	NOUN
ejpam-86	105	25	=	=	PUNCT
ejpam-86	105	26	φ1	φ1	PROPN
ejpam-86	105	27	/	/	SYM
ejpam-86	105	28	c	c	PROPN
ejpam-86	105	29	and	and	CCONJ
ejpam-86	105	30	β2	β2	NOUN
ejpam-86	105	31	=	=	SYM
ejpam-86	105	32	β1	β1	PROPN
ejpam-86	105	33	.	.	PUNCT
ejpam-86	106	1	proof	proof	NOUN
ejpam-86	106	2	:	:	PUNCT
ejpam-86	106	3	the	the	DET
ejpam-86	106	4	proof	proof	NOUN
ejpam-86	106	5	follows	follow	VERB
ejpam-86	106	6	along	along	ADP
ejpam-86	106	7	the	the	DET
ejpam-86	106	8	lines	line	NOUN
ejpam-86	106	9	given	give	VERB
ejpam-86	106	10	in	in	ADP
ejpam-86	106	11	gupta	gupta	PROPN
ejpam-86	106	12	and	and	CCONJ
ejpam-86	106	13	varga	varga	PROPN
ejpam-86	106	14	(	(	PUNCT
ejpam-86	106	15	[	[	X
ejpam-86	106	16	21	21	NUM
ejpam-86	106	17	,	,	PUNCT
ejpam-86	106	18	p.23	p.23	X
ejpam-86	106	19	]	]	NOUN
ejpam-86	106	20	)	)	PUNCT
ejpam-86	106	21	.	.	PUNCT
ejpam-86	107	1	since	since	SCONJ
ejpam-86	107	2	z	z	PROPN
ejpam-86	107	3	is	be	AUX
ejpam-86	107	4	symmetric	symmetric	ADJ
ejpam-86	107	5	about	about	ADP
ejpam-86	107	6	m1	m1	PROPN
ejpam-86	107	7	as	as	ADV
ejpam-86	107	8	well	well	ADV
ejpam-86	107	9	as	as	ADP
ejpam-86	107	10	about	about	ADP
ejpam-86	107	11	m2	m2	PROPN
ejpam-86	107	12	,	,	PUNCT
ejpam-86	107	13	then	then	ADV
ejpam-86	107	14	m2	m2	PROPN
ejpam-86	107	15	=	=	PROPN
ejpam-86	107	16	m1	m1	PROPN
ejpam-86	107	17	.	.	PUNCT
ejpam-86	108	1	let	let	VERB
ejpam-86	108	2	m	m	NOUN
ejpam-86	108	3	=	=	VERB
ejpam-86	108	4	m1	m1	PROPN
ejpam-86	108	5	and	and	CCONJ
ejpam-86	108	6	σl	σl	NOUN
ejpam-86	108	7	=	=	PUNCT
ejpam-86	109	1	[	[	X
ejpam-86	109	2	lσij	lσij	X
ejpam-86	109	3	]	]	X
ejpam-86	109	4	,	,	PUNCT
ejpam-86	109	5	i	i	PRON
ejpam-86	109	6	,	,	PUNCT
ejpam-86	109	7	j	j	PROPN
ejpam-86	109	8	=	=	SYM
ejpam-86	109	9	1	1	NUM
ejpam-86	109	10	,	,	PUNCT
ejpam-86	109	11	·	·	PUNCT
ejpam-86	109	12	·	·	PUNCT
ejpam-86	109	13	·	·	PUNCT
ejpam-86	109	14	,	,	PUNCT
ejpam-86	109	15	p	p	X
ejpam-86	109	16	;	;	PUNCT
ejpam-86	109	17	l	l	NOUN
ejpam-86	109	18	=	=	SYM
ejpam-86	109	19	1	1	NUM
ejpam-86	109	20	,	,	PUNCT
ejpam-86	109	21	2	2	NUM
ejpam-86	109	22	,	,	PUNCT
ejpam-86	109	23	φl	φl	ADJ
ejpam-86	109	24	=	=	PUNCT
ejpam-86	110	1	[	[	X
ejpam-86	110	2	lφij	lφij	X
ejpam-86	110	3	]	]	X
ejpam-86	110	4	,	,	PUNCT
ejpam-86	110	5	i	i	PRON
ejpam-86	110	6	,	,	PUNCT
ejpam-86	110	7	j	j	PROPN
ejpam-86	110	8	=	=	SYM
ejpam-86	110	9	1	1	NUM
ejpam-86	110	10	,	,	PUNCT
ejpam-86	110	11	·	·	PUNCT
ejpam-86	110	12	·	·	PUNCT
ejpam-86	110	13	·	·	PUNCT
ejpam-86	110	14	,	,	PUNCT
ejpam-86	110	15	n	n	CCONJ
ejpam-86	110	16	;	;	PUNCT
ejpam-86	110	17	l	l	NOUN
ejpam-86	110	18	=	=	SYM
ejpam-86	110	19	1	1	NUM
ejpam-86	110	20	,	,	PUNCT
ejpam-86	110	21	2	2	NUM
ejpam-86	110	22	.	.	PUNCT
ejpam-86	110	23	m.	m.	NOUN
ejpam-86	110	24	liu	liu	PROPN
ejpam-86	110	25	and	and	CCONJ
ejpam-86	110	26	h.	h.	PROPN
ejpam-86	110	27	bozdogan	bozdogan	PROPN
ejpam-86	110	28	/	/	SYM
ejpam-86	110	29	eur	eur	PROPN
ejpam-86	110	30	.	.	PUNCT
ejpam-86	111	1	j.	j.	PROPN
ejpam-86	111	2	pure	pure	PROPN
ejpam-86	111	3	appl	appl	PROPN
ejpam-86	111	4	.	.	PROPN
ejpam-86	111	5	math	math	PROPN
ejpam-86	111	6	,	,	PUNCT
ejpam-86	111	7	1	1	NUM
ejpam-86	111	8	(	(	PUNCT
ejpam-86	111	9	2008	2008	NUM
ejpam-86	111	10	)	)	PUNCT
ejpam-86	111	11	,	,	PUNCT
ejpam-86	111	12	(	(	PUNCT
ejpam-86	111	13	4	4	NUM
ejpam-86	111	14	-	-	SYM
ejpam-86	111	15	37	37	NUM
ejpam-86	111	16	)	)	PUNCT
ejpam-86	111	17	8	8	NUM
ejpam-86	111	18	−2	−2	NOUN
ejpam-86	111	19	0	0	NUM
ejpam-86	111	20	2	2	NUM
ejpam-86	111	21	−2	−2	NOUN
ejpam-86	111	22	0	0	NUM
ejpam-86	111	23	2	2	NUM
ejpam-86	111	24	0	0	NUM
ejpam-86	111	25	0.5	0.5	NUM
ejpam-86	111	26	1	1	NUM
ejpam-86	111	27	x	x	SYM
ejpam-86	111	28	10	10	NUM
ejpam-86	111	29	−3	−3	NOUN
ejpam-86	111	30	−2	−2	NOUN
ejpam-86	111	31	0	0	NUM
ejpam-86	111	32	2	2	NUM
ejpam-86	111	33	−2	−2	NOUN
ejpam-86	111	34	0	0	NUM
ejpam-86	111	35	2	2	NUM
ejpam-86	111	36	0	0	NUM
ejpam-86	111	37	0.1	0.1	NUM
ejpam-86	111	38	0.2	0.2	NUM
ejpam-86	111	39	−2	−2	NOUN
ejpam-86	111	40	0	0	NUM
ejpam-86	111	41	2	2	NUM
ejpam-86	111	42	−2	−2	NOUN
ejpam-86	111	43	0	0	NUM
ejpam-86	111	44	2	2	NUM
ejpam-86	111	45	0	0	NUM
ejpam-86	111	46	0.2	0.2	NUM
ejpam-86	111	47	0.4	0.4	NUM
ejpam-86	111	48	−2	−2	NOUN
ejpam-86	111	49	0	0	NUM
ejpam-86	111	50	2	2	NUM
ejpam-86	111	51	−2	−2	NOUN
ejpam-86	111	52	0	0	NUM
ejpam-86	111	53	2	2	NUM
ejpam-86	111	54	0	0	NUM
ejpam-86	111	55	0.2	0.2	NUM
ejpam-86	111	56	0.4	0.4	NUM
ejpam-86	111	57	beta=0.25	beta=0.25	NOUN
ejpam-86	111	58	beta=1	beta=1	PROPN
ejpam-86	112	1	(	(	PUNCT
ejpam-86	112	2	multivariate	multivariate	VERB
ejpam-86	112	3	normal	normal	ADJ
ejpam-86	112	4	distribution	distribution	NOUN
ejpam-86	112	5	)	)	PUNCT
ejpam-86	112	6	beta=2	beta=2	PROPN
ejpam-86	112	7	beta=10	beta=10	NOUN
ejpam-86	112	8	figure	figure	VERB
ejpam-86	112	9	1	1	NUM
ejpam-86	112	10	:	:	PUNCT
ejpam-86	112	11	matrix	matrix	VERB
ejpam-86	112	12	pe	pe	NOUN
ejpam-86	112	13	density	density	NOUN
ejpam-86	112	14	function	function	NOUN
ejpam-86	112	15	with	with	ADP
ejpam-86	112	16	n	n	NOUN
ejpam-86	112	17	=	=	SYM
ejpam-86	112	18	1	1	NUM
ejpam-86	112	19	,	,	PUNCT
ejpam-86	112	20	p	p	NOUN
ejpam-86	112	21	=	=	SYM
ejpam-86	112	22	2	2	NUM
ejpam-86	112	23	,	,	PUNCT
ejpam-86	112	24	φ	φ	PROPN
ejpam-86	112	25	=	=	SYM
ejpam-86	112	26	i1	i1	PROPN
ejpam-86	112	27	,	,	PUNCT
ejpam-86	112	28	σ	σ	PROPN
ejpam-86	112	29	=	=	PROPN
ejpam-86	112	30	i2	i2	PROPN
ejpam-86	112	31	and	and	CCONJ
ejpam-86	112	32	variate	variate	NOUN
ejpam-86	112	33	β	β	PROPN
ejpam-86	112	34	.	.	PUNCT
ejpam-86	113	1	let	let	AUX
ejpam-86	113	2	k(i	k(i	PROPN
ejpam-86	113	3	)	)	PUNCT
ejpam-86	113	4	denote	denote	VERB
ejpam-86	113	5	the	the	DET
ejpam-86	113	6	p	p	ADJ
ejpam-86	113	7	-	-	PUNCT
ejpam-86	113	8	dimensional	dimensional	ADJ
ejpam-86	113	9	vector	vector	NOUN
ejpam-86	113	10	whose	whose	DET
ejpam-86	113	11	ith	ith	NOUN
ejpam-86	113	12	entry	entry	NOUN
ejpam-86	113	13	is	be	AUX
ejpam-86	113	14	1	1	NUM
ejpam-86	113	15	and	and	CCONJ
ejpam-86	113	16	all	all	DET
ejpam-86	113	17	the	the	DET
ejpam-86	113	18	others	other	NOUN
ejpam-86	113	19	are	be	AUX
ejpam-86	113	20	0	0	NUM
ejpam-86	113	21	.	.	PUNCT
ejpam-86	114	1	let	let	VERB
ejpam-86	114	2	l(i	l(i	NOUN
ejpam-86	114	3	)	)	PUNCT
ejpam-86	115	1	denote	denote	VERB
ejpam-86	115	2	the	the	DET
ejpam-86	115	3	n	n	ADV
ejpam-86	115	4	-	-	PUNCT
ejpam-86	115	5	dimensional	dimensional	ADJ
ejpam-86	115	6	vector	vector	NOUN
ejpam-86	115	7	whose	whose	DET
ejpam-86	115	8	ith	ith	NOUN
ejpam-86	115	9	entry	entry	NOUN
ejpam-86	115	10	is	be	AUX
ejpam-86	115	11	1	1	NUM
ejpam-86	115	12	and	and	CCONJ
ejpam-86	115	13	all	all	DET
ejpam-86	115	14	the	the	DET
ejpam-86	115	15	others	other	NOUN
ejpam-86	115	16	are	be	AUX
ejpam-86	115	17	0	0	NUM
ejpam-86	115	18	.	.	PUNCT
ejpam-86	116	1	since	since	SCONJ
ejpam-86	116	2	z	z	PROPN
ejpam-86	116	3	is	be	AUX
ejpam-86	116	4	nondegenerate	nondegenerate	ADJ
ejpam-86	116	5	,	,	PUNCT
ejpam-86	116	6	it	it	PRON
ejpam-86	116	7	must	must	AUX
ejpam-86	116	8	have	have	VERB
ejpam-86	116	9	an	an	DET
ejpam-86	116	10	element	element	NOUN
ejpam-86	116	11	zi0j0	zi0j0	PROPN
ejpam-86	116	12	which	which	PRON
ejpam-86	116	13	is	be	AUX
ejpam-86	116	14	non	non	ADJ
ejpam-86	116	15	-	-	ADJ
ejpam-86	116	16	degenerate	degenerate	ADJ
ejpam-86	116	17	.	.	PUNCT
ejpam-86	117	1	since	since	SCONJ
ejpam-86	117	2	zi0j0	zi0j0	PROPN
ejpam-86	117	3	=	=	SYM
ejpam-86	117	4	l′(i0)zk(j0	l′(i0)zk(j0	PROPN
ejpam-86	117	5	)	)	PUNCT
ejpam-86	117	6	,	,	PUNCT
ejpam-86	117	7	from	from	ADP
ejpam-86	117	8	theorem	theorem	ADJ
ejpam-86	117	9	4	4	NUM
ejpam-86	117	10	of	of	ADP
ejpam-86	117	11	sánchez	sánchez	PROPN
ejpam-86	117	12	-	-	PUNCT
ejpam-86	117	13	manzano	manzano	PROPN
ejpam-86	117	14	et	et	PROPN
ejpam-86	117	15	al	al	PROPN
ejpam-86	117	16	.	.	PUNCT
ejpam-86	118	1	(	(	PUNCT
ejpam-86	118	2	[	[	X
ejpam-86	118	3	31	31	NUM
ejpam-86	118	4	]	]	PUNCT
ejpam-86	118	5	)	)	PUNCT
ejpam-86	118	6	,	,	PUNCT
ejpam-86	118	7	there	there	PRON
ejpam-86	118	8	are	be	VERB
ejpam-86	118	9	zi0j0	zi0j0	PROPN
ejpam-86	118	10	∼	∼	NOUN
ejpam-86	118	11	pe(mi0j0	pe(mi0j0	PUNCT
ejpam-86	118	12	,	,	PUNCT
ejpam-86	118	13	1φi0i01σj0j0	1φi0i01σj0j0	NUM
ejpam-86	118	14	,	,	PUNCT
ejpam-86	118	15	β1	β1	PROPN
ejpam-86	118	16	)	)	PUNCT
ejpam-86	118	17	and	and	CCONJ
ejpam-86	118	18	zi0j0	zi0j0	PROPN
ejpam-86	118	19	∼	∼	NOUN
ejpam-86	118	20	pe(mi0j0	pe(mi0j0	PUNCT
ejpam-86	118	21	,	,	PUNCT
ejpam-86	118	22	2φi0i02σj0j0	2φi0i02σj0j0	NUM
ejpam-86	118	23	,	,	PUNCT
ejpam-86	118	24	β2	β2	PROPN
ejpam-86	118	25	)	)	PUNCT
ejpam-86	118	26	.	.	PUNCT
ejpam-86	119	1	so	so	ADV
ejpam-86	119	2	we	we	PRON
ejpam-86	119	3	have	have	AUX
ejpam-86	119	4	β2	β2	VERB
ejpam-86	119	5	=	=	SYM
ejpam-86	119	6	β1	β1	PROPN
ejpam-86	119	7	and	and	CCONJ
ejpam-86	119	8	2φi0i02σj0j0	2φi0i02σj0j0	NUM
ejpam-86	119	9	=	=	SYM
ejpam-86	119	10	1φi0i01σj0j0	1φi0i01σj0j0	NUM
ejpam-86	119	11	.	.	PUNCT
ejpam-86	120	1	from	from	ADP
ejpam-86	120	2	gupta	gupta	PROPN
ejpam-86	120	3	and	and	CCONJ
ejpam-86	120	4	varga	varga	PROPN
ejpam-86	120	5	(	(	PUNCT
ejpam-86	120	6	[	[	X
ejpam-86	120	7	21	21	NUM
ejpam-86	120	8	,	,	PUNCT
ejpam-86	120	9	p.24	p.24	PROPN
ejpam-86	120	10	]	]	X
ejpam-86	120	11	)	)	PUNCT
ejpam-86	120	12	,	,	PUNCT
ejpam-86	120	13	there	there	PRON
ejpam-86	120	14	must	must	AUX
ejpam-86	120	15	be	be	AUX
ejpam-86	120	16	φ2	φ2	PROPN
ejpam-86	120	17	⊗	⊗	PROPN
ejpam-86	120	18	σ2	σ2	PROPN
ejpam-86	120	19	=	=	PROPN
ejpam-86	120	20	φ1	φ1	PROPN
ejpam-86	120	21	⊗	⊗	PROPN
ejpam-86	120	22	σ1	σ1	PROPN
ejpam-86	120	23	.	.	PUNCT
ejpam-86	121	1	by	by	ADP
ejpam-86	121	2	the	the	DET
ejpam-86	121	3	theorem	theorem	ADJ
ejpam-86	121	4	1.3.16	1.3.16	NUM
ejpam-86	121	5	of	of	ADP
ejpam-86	121	6	gupta	gupta	PROPN
ejpam-86	121	7	and	and	CCONJ
ejpam-86	121	8	varga	varga	PROPN
ejpam-86	121	9	(	(	PUNCT
ejpam-86	121	10	[	[	X
ejpam-86	121	11	21	21	NUM
ejpam-86	121	12	,	,	PUNCT
ejpam-86	121	13	p.13	p.13	X
ejpam-86	121	14	]	]	PUNCT
ejpam-86	121	15	)	)	PUNCT
ejpam-86	121	16	,	,	PUNCT
ejpam-86	121	17	there	there	PRON
ejpam-86	121	18	exists	exist	VERB
ejpam-86	121	19	a	a	DET
ejpam-86	121	20	nonzero	nonzero	ADJ
ejpam-86	121	21	real	real	ADJ
ejpam-86	121	22	number	number	NOUN
ejpam-86	121	23	c	c	NOUN
ejpam-86	121	24	such	such	ADJ
ejpam-86	121	25	that	that	DET
ejpam-86	121	26	σ2	σ2	NOUN
ejpam-86	121	27	=	=	SYM
ejpam-86	121	28	cς1	cς1	VERB
ejpam-86	121	29	and	and	CCONJ
ejpam-86	121	30	φ2	φ2	PROPN
ejpam-86	121	31	=	=	SYM
ejpam-86	121	32	φ1	φ1	PROPN
ejpam-86	121	33	/	/	SYM
ejpam-86	121	34	c.	c.	PROPN
ejpam-86	121	35	if	if	SCONJ
ejpam-86	121	36	φ2	φ2	PROPN
ejpam-86	121	37	=	=	SYM
ejpam-86	121	38	φ1	φ1	PROPN
ejpam-86	121	39	=	=	SYM
ejpam-86	121	40	φ	φ	PROPN
ejpam-86	121	41	,	,	PUNCT
ejpam-86	121	42	then	then	ADV
ejpam-86	121	43	c	c	NOUN
ejpam-86	121	44	=	=	SYM
ejpam-86	121	45	1	1	NUM
ejpam-86	121	46	and	and	CCONJ
ejpam-86	121	47	σ2	σ2	PROPN
ejpam-86	121	48	=	=	SYM
ejpam-86	121	49	σ1	σ1	PROPN
ejpam-86	121	50	.	.	PUNCT
ejpam-86	122	1	¤	¤	VERB
ejpam-86	122	2	from	from	ADP
ejpam-86	122	3	the	the	DET
ejpam-86	122	4	proof	proof	NOUN
ejpam-86	122	5	of	of	ADP
ejpam-86	122	6	theorem	theorem	ADJ
ejpam-86	122	7	2.1	2.1	NUM
ejpam-86	122	8	,	,	PUNCT
ejpam-86	122	9	the	the	DET
ejpam-86	122	10	only	only	ADJ
ejpam-86	122	11	case	case	NOUN
ejpam-86	122	12	that	that	SCONJ
ejpam-86	122	13	a	a	DET
ejpam-86	122	14	matrix	matrix	NOUN
ejpam-86	122	15	multivariate	multivariate	NOUN
ejpam-86	122	16	pe	pe	ADP
ejpam-86	122	17	distribution	distribution	NOUN
ejpam-86	122	18	is	be	AUX
ejpam-86	122	19	not	not	PART
ejpam-86	122	20	uniquely	uniquely	ADV
ejpam-86	122	21	defined	define	VERB
ejpam-86	122	22	is	be	AUX
ejpam-86	122	23	that	that	SCONJ
ejpam-86	122	24	there	there	PRON
ejpam-86	122	25	exist	exist	VERB
ejpam-86	122	26	positive	positive	ADJ
ejpam-86	122	27	constant	constant	ADJ
ejpam-86	122	28	c	c	NOUN
ejpam-86	122	29	such	such	ADJ
ejpam-86	122	30	that	that	DET
ejpam-86	122	31	σ2	σ2	NOUN
ejpam-86	122	32	=	=	SYM
ejpam-86	122	33	cς1	cς1	VERB
ejpam-86	122	34	and	and	CCONJ
ejpam-86	122	35	φ2	φ2	PROPN
ejpam-86	122	36	=	=	SYM
ejpam-86	122	37	φ1	φ1	PROPN
ejpam-86	122	38	/	/	SYM
ejpam-86	122	39	c.	c.	PROPN
ejpam-86	123	1	so	so	ADV
ejpam-86	123	2	if	if	SCONJ
ejpam-86	123	3	the	the	DET
ejpam-86	123	4	parameter	parameter	NOUN
ejpam-86	123	5	σ	σ	PROPN
ejpam-86	123	6	or	or	CCONJ
ejpam-86	123	7	φ	φ	PROPN
ejpam-86	123	8	is	be	AUX
ejpam-86	123	9	known	know	VERB
ejpam-86	123	10	,	,	PUNCT
ejpam-86	123	11	then	then	ADV
ejpam-86	123	12	the	the	DET
ejpam-86	123	13	distribution	distribution	NOUN
ejpam-86	123	14	is	be	AUX
ejpam-86	123	15	uniquely	uniquely	ADV
ejpam-86	123	16	defined	define	VERB
ejpam-86	123	17	by	by	ADP
ejpam-86	123	18	the	the	DET
ejpam-86	123	19	other	other	ADJ
ejpam-86	123	20	three	three	NUM
ejpam-86	123	21	parameters	parameter	NOUN
ejpam-86	123	22	.	.	PUNCT
ejpam-86	124	1	therefore	therefore	ADV
ejpam-86	124	2	,	,	PUNCT
ejpam-86	124	3	in	in	ADP
ejpam-86	124	4	the	the	DET
ejpam-86	124	5	rest	rest	NOUN
ejpam-86	124	6	of	of	ADP
ejpam-86	124	7	this	this	DET
ejpam-86	124	8	paper	paper	NOUN
ejpam-86	124	9	,	,	PUNCT
ejpam-86	124	10	we	we	PRON
ejpam-86	124	11	only	only	ADV
ejpam-86	124	12	consider	consider	VERB
ejpam-86	124	13	the	the	DET
ejpam-86	124	14	case	case	NOUN
ejpam-86	124	15	that	that	SCONJ
ejpam-86	124	16	parameter	parameter	PROPN
ejpam-86	124	17	φ	φ	PROPN
ejpam-86	124	18	is	be	AUX
ejpam-86	124	19	known	know	VERB
ejpam-86	124	20	to	to	PART
ejpam-86	124	21	handle	handle	VERB
ejpam-86	124	22	the	the	DET
ejpam-86	124	23	uniqueness	uniqueness	NOUN
ejpam-86	124	24	issue	issue	NOUN
ejpam-86	124	25	.	.	PUNCT
ejpam-86	125	1	an	an	DET
ejpam-86	125	2	advantage	advantage	NOUN
ejpam-86	125	3	of	of	ADP
ejpam-86	125	4	pe	pe	ADJ
ejpam-86	125	5	distribution	distribution	NOUN
ejpam-86	125	6	is	be	AUX
ejpam-86	125	7	that	that	SCONJ
ejpam-86	125	8	it	it	PRON
ejpam-86	125	9	is	be	AUX
ejpam-86	125	10	adaptive	adaptive	ADJ
ejpam-86	125	11	to	to	ADP
ejpam-86	125	12	both	both	CCONJ
ejpam-86	125	13	peakedness	peakedness	NOUN
ejpam-86	125	14	and	and	CCONJ
ejpam-86	125	15	flatness	flatness	NOUN
ejpam-86	125	16	in	in	ADP
ejpam-86	125	17	the	the	DET
ejpam-86	125	18	data	datum	NOUN
ejpam-86	125	19	by	by	ADP
ejpam-86	125	20	varying	vary	VERB
ejpam-86	125	21	the	the	DET
ejpam-86	125	22	values	value	NOUN
ejpam-86	125	23	of	of	ADP
ejpam-86	125	24	β	β	PROPN
ejpam-86	125	25	.	.	PUNCT
ejpam-86	126	1	when	when	SCONJ
ejpam-86	126	2	β	β	NOUN
ejpam-86	126	3	increases	increase	NOUN
ejpam-86	126	4	,	,	PUNCT
ejpam-86	126	5	the	the	DET
ejpam-86	126	6	sharpness	sharpness	NOUN
ejpam-86	126	7	diminishes	diminish	VERB
ejpam-86	126	8	.	.	PUNCT
ejpam-86	127	1	figure	figure	NOUN
ejpam-86	127	2	1	1	NUM
ejpam-86	127	3	represents	represent	VERB
ejpam-86	127	4	the	the	DET
ejpam-86	127	5	plot	plot	NOUN
ejpam-86	127	6	of	of	ADP
ejpam-86	127	7	(	(	PUNCT
ejpam-86	127	8	2.2	2.2	NUM
ejpam-86	127	9	)	)	PUNCT
ejpam-86	127	10	with	with	ADP
ejpam-86	127	11	n	n	NOUN
ejpam-86	127	12	=	=	SYM
ejpam-86	127	13	1	1	NUM
ejpam-86	127	14	,	,	PUNCT
ejpam-86	127	15	p	p	NOUN
ejpam-86	127	16	=	=	SYM
ejpam-86	127	17	2	2	NUM
ejpam-86	127	18	,	,	PUNCT
ejpam-86	127	19	φ	φ	PROPN
ejpam-86	127	20	=	=	SYM
ejpam-86	127	21	i1	i1	PROPN
ejpam-86	127	22	,	,	PUNCT
ejpam-86	127	23	σ	σ	PROPN
ejpam-86	127	24	=	=	PROPN
ejpam-86	127	25	i2	i2	PROPN
ejpam-86	127	26	and	and	CCONJ
ejpam-86	127	27	the	the	DET
ejpam-86	127	28	shape	shape	NOUN
ejpam-86	127	29	parameter	parameter	NOUN
ejpam-86	127	30	β	β	NOUN
ejpam-86	127	31	.	.	PUNCT
ejpam-86	128	1	the	the	DET
ejpam-86	128	2	relationship	relationship	NOUN
ejpam-86	128	3	between	between	ADP
ejpam-86	128	4	β	β	NOUN
ejpam-86	128	5	and	and	CCONJ
ejpam-86	128	6	γ2	γ2	PROPN
ejpam-86	128	7	for	for	ADP
ejpam-86	128	8	the	the	DET
ejpam-86	128	9	multivariate	multivariate	NOUN
ejpam-86	128	10	pe	pe	ADP
ejpam-86	128	11	distribution	distribution	NOUN
ejpam-86	128	12	is	be	AUX
ejpam-86	128	13	shown	show	VERB
ejpam-86	128	14	in	in	ADP
ejpam-86	128	15	figure	figure	NOUN
ejpam-86	128	16	2	2	NUM
ejpam-86	128	17	.	.	PUNCT
ejpam-86	128	18	for	for	ADP
ejpam-86	128	19	β	β	X
ejpam-86	128	20	=	=	SYM
ejpam-86	128	21	1	1	NUM
ejpam-86	128	22	,	,	PUNCT
ejpam-86	128	23	(	(	PUNCT
ejpam-86	128	24	2.2	2.2	NUM
ejpam-86	128	25	)	)	PUNCT
ejpam-86	128	26	is	be	AUX
ejpam-86	128	27	a	a	DET
ejpam-86	128	28	multivariate	multivariate	NOUN
ejpam-86	128	29	normal	normal	ADJ
ejpam-86	128	30	distribution	distribution	NOUN
ejpam-86	128	31	and	and	CCONJ
ejpam-86	128	32	for	for	ADP
ejpam-86	128	33	β	β	X
ejpam-86	128	34	→	→	SYM
ejpam-86	128	35	∞	∞	PROPN
ejpam-86	128	36	,	,	PUNCT
ejpam-86	128	37	(	(	PUNCT
ejpam-86	128	38	2.2	2.2	NUM
ejpam-86	128	39	)	)	PUNCT
ejpam-86	128	40	is	be	AUX
ejpam-86	128	41	a	a	DET
ejpam-86	128	42	multivariate	multivariate	NOUN
ejpam-86	128	43	uniform	uniform	ADJ
ejpam-86	128	44	distribution	distribution	NOUN
ejpam-86	128	45	.	.	PUNCT
ejpam-86	129	1	we	we	PRON
ejpam-86	129	2	note	note	VERB
ejpam-86	129	3	that	that	SCONJ
ejpam-86	129	4	as	as	SCONJ
ejpam-86	129	5	the	the	DET
ejpam-86	129	6	dimension	dimension	NOUN
ejpam-86	129	7	of	of	ADP
ejpam-86	129	8	the	the	DET
ejpam-86	129	9	distribution	distribution	NOUN
ejpam-86	129	10	gets	get	VERB
ejpam-86	129	11	larger	large	ADJ
ejpam-86	129	12	,	,	PUNCT
ejpam-86	129	13	the	the	DET
ejpam-86	129	14	kurtosis	kurtosis	NOUN
ejpam-86	129	15	gets	get	VERB
ejpam-86	129	16	larger	large	ADJ
ejpam-86	129	17	for	for	ADP
ejpam-86	129	18	both	both	PRON
ejpam-86	129	19	β	β	X
ejpam-86	129	20	<	<	X
ejpam-86	129	21	1	1	NUM
ejpam-86	129	22	and	and	CCONJ
ejpam-86	129	23	β	β	X
ejpam-86	129	24	>	>	X
ejpam-86	130	1	1	1	X
ejpam-86	130	2	.	.	PUNCT
ejpam-86	131	1	what	what	PRON
ejpam-86	131	2	this	this	PRON
ejpam-86	131	3	means	mean	VERB
ejpam-86	131	4	is	be	AUX
ejpam-86	131	5	that	that	SCONJ
ejpam-86	131	6	for	for	ADP
ejpam-86	131	7	large	large	ADJ
ejpam-86	131	8	dimensional	dimensional	ADJ
ejpam-86	131	9	data	datum	NOUN
ejpam-86	131	10	sets	set	NOUN
ejpam-86	131	11	,	,	PUNCT
ejpam-86	131	12	we	we	PRON
ejpam-86	131	13	expect	expect	VERB
ejpam-86	131	14	heavy	heavy	ADJ
ejpam-86	131	15	fat	fat	ADJ
ejpam-86	131	16	tails	tail	NOUN
ejpam-86	131	17	.	.	PUNCT
ejpam-86	132	1	therefore	therefore	ADV
ejpam-86	132	2	,	,	PUNCT
ejpam-86	132	3	we	we	PRON
ejpam-86	132	4	should	should	AUX
ejpam-86	132	5	use	use	VERB
ejpam-86	132	6	flexible	flexible	ADJ
ejpam-86	132	7	distributional	distributional	ADJ
ejpam-86	132	8	models	model	NOUN
ejpam-86	132	9	to	to	PART
ejpam-86	132	10	capture	capture	VERB
ejpam-86	132	11	the	the	DET
ejpam-86	132	12	heavy	heavy	ADJ
ejpam-86	132	13	fat	fat	ADJ
ejpam-86	132	14	tail	tail	NOUN
ejpam-86	132	15	behavior	behavior	NOUN
ejpam-86	132	16	of	of	ADP
ejpam-86	132	17	large	large	ADJ
ejpam-86	132	18	dimensional	dimensional	ADJ
ejpam-86	132	19	data	data	NOUN
ejpam-86	132	20	sets	set	NOUN
ejpam-86	132	21	.	.	PUNCT
ejpam-86	133	1	m.	m.	NOUN
ejpam-86	133	2	liu	liu	PROPN
ejpam-86	133	3	and	and	CCONJ
ejpam-86	133	4	h.	h.	PROPN
ejpam-86	133	5	bozdogan	bozdogan	PROPN
ejpam-86	133	6	/	/	SYM
ejpam-86	133	7	eur	eur	PROPN
ejpam-86	133	8	.	.	PUNCT
ejpam-86	134	1	j.	j.	PROPN
ejpam-86	134	2	pure	pure	PROPN
ejpam-86	134	3	appl	appl	PROPN
ejpam-86	134	4	.	.	PROPN
ejpam-86	134	5	math	math	PROPN
ejpam-86	134	6	,	,	PUNCT
ejpam-86	134	7	1	1	NUM
ejpam-86	134	8	(	(	PUNCT
ejpam-86	134	9	2008	2008	NUM
ejpam-86	134	10	)	)	PUNCT
ejpam-86	134	11	,	,	PUNCT
ejpam-86	134	12	(	(	PUNCT
ejpam-86	134	13	4	4	NUM
ejpam-86	134	14	-	-	SYM
ejpam-86	134	15	37	37	NUM
ejpam-86	134	16	)	)	PUNCT
ejpam-86	134	17	9	9	NUM
ejpam-86	134	18	0.5	0.5	NUM
ejpam-86	134	19	1	1	NUM
ejpam-86	134	20	1.5	1.5	NUM
ejpam-86	134	21	2	2	NUM
ejpam-86	134	22	2.5	2.5	NUM
ejpam-86	134	23	3	3	NUM
ejpam-86	134	24	3.5	3.5	NUM
ejpam-86	134	25	4	4	NUM
ejpam-86	134	26	4.5	4.5	NUM
ejpam-86	134	27	5	5	NUM
ejpam-86	134	28	5.5	5.5	NUM
ejpam-86	134	29	6	6	NUM
ejpam-86	134	30	−10	−10	SYM
ejpam-86	134	31	−5	−5	ADV
ejpam-86	134	32	0	0	NUM
ejpam-86	134	33	5	5	NUM
ejpam-86	134	34	10	10	NUM
ejpam-86	134	35	15	15	NUM
ejpam-86	134	36	20	20	NUM
ejpam-86	134	37	25	25	NUM
ejpam-86	134	38	30	30	NUM
ejpam-86	134	39	35	35	NUM
ejpam-86	134	40	40	40	NUM
ejpam-86	134	41	relationship	relationship	NOUN
ejpam-86	134	42	between	between	ADP
ejpam-86	134	43	beta	beta	NOUN
ejpam-86	134	44	and	and	CCONJ
ejpam-86	134	45	kurtosis	kurtosis	VERB
ejpam-86	134	46	k	k	PROPN
ejpam-86	134	47	ur	ur	INTJ
ejpam-86	134	48	to	to	ADP
ejpam-86	134	49	si	si	PROPN
ejpam-86	134	50	s	s	PART
ejpam-86	134	51	beta	beta	PROPN
ejpam-86	134	52	p=2	p=2	PROPN
ejpam-86	134	53	p=3	p=3	PROPN
ejpam-86	134	54	p=4	p=4	PROPN
ejpam-86	134	55	p=5	p=5	ADJ
ejpam-86	134	56	p=6	p=6	NOUN
ejpam-86	134	57	figure	figure	NOUN
ejpam-86	134	58	2	2	NUM
ejpam-86	134	59	:	:	PUNCT
ejpam-86	134	60	the	the	DET
ejpam-86	134	61	relationship	relationship	NOUN
ejpam-86	134	62	between	between	ADP
ejpam-86	134	63	β	β	NOUN
ejpam-86	134	64	and	and	CCONJ
ejpam-86	134	65	kurtosis	kurtosis	NOUN
ejpam-86	134	66	.	.	PUNCT
ejpam-86	135	1	3	3	X
ejpam-86	135	2	.	.	X
ejpam-86	135	3	type	type	NOUN
ejpam-86	135	4	i	i	PRON
ejpam-86	135	5	multivariate	multivariate	VERB
ejpam-86	135	6	pe	pe	INTJ
ejpam-86	135	7	regression	regression	NOUN
ejpam-86	135	8	model	model	NOUN
ejpam-86	135	9	in	in	ADP
ejpam-86	135	10	model	model	NOUN
ejpam-86	135	11	(	(	PUNCT
ejpam-86	135	12	1.1	1.1	NUM
ejpam-86	135	13	)	)	PUNCT
ejpam-86	135	14	,	,	PUNCT
ejpam-86	135	15	if	if	SCONJ
ejpam-86	135	16	we	we	PRON
ejpam-86	135	17	assume	assume	VERB
ejpam-86	135	18	that	that	SCONJ
ejpam-86	135	19	the	the	DET
ejpam-86	135	20	rows	row	NOUN
ejpam-86	135	21	of	of	ADP
ejpam-86	135	22	the	the	DET
ejpam-86	135	23	random	random	ADJ
ejpam-86	135	24	error	error	NOUN
ejpam-86	135	25	matrix	matrix	NOUN
ejpam-86	135	26	e	e	NOUN
ejpam-86	135	27	are	be	AUX
ejpam-86	135	28	drawn	draw	VERB
ejpam-86	135	29	independently	independently	ADV
ejpam-86	135	30	from	from	ADP
ejpam-86	135	31	pep(0,σ	pep(0,σ	PROPN
ejpam-86	135	32	,	,	PUNCT
ejpam-86	135	33	β	β	NOUN
ejpam-86	135	34	)	)	PUNCT
ejpam-86	135	35	distribution	distribution	NOUN
ejpam-86	135	36	,	,	PUNCT
ejpam-86	135	37	i.e.	i.e.	X
ejpam-86	135	38	,	,	PUNCT
ejpam-86	135	39	ε(1	ε(1	NOUN
ejpam-86	135	40	)	)	PUNCT
ejpam-86	135	41	,	,	PUNCT
ejpam-86	135	42	ε(2	ε(2	NOUN
ejpam-86	135	43	)	)	PUNCT
ejpam-86	135	44	,	,	PUNCT
ejpam-86	135	45	·	·	PUNCT
ejpam-86	135	46	·	·	PUNCT
ejpam-86	135	47	·	·	PUNCT
ejpam-86	135	48	,	,	PUNCT
ejpam-86	135	49	ε(n	ε(n	NOUN
ejpam-86	135	50	)	)	PUNCT
ejpam-86	135	51	are	be	AUX
ejpam-86	135	52	i.i.d	i.i.d	ADP
ejpam-86	135	53	.	.	PUNCT
ejpam-86	136	1	and	and	CCONJ
ejpam-86	136	2	ε(1	ε(1	NOUN
ejpam-86	136	3	)	)	PUNCT
ejpam-86	136	4	∼	∼	NOUN
ejpam-86	136	5	pep(0,σ	pep(0,σ	NOUN
ejpam-86	136	6	,	,	PUNCT
ejpam-86	136	7	β	β	NOUN
ejpam-86	136	8	)	)	PUNCT
ejpam-86	136	9	,	,	PUNCT
ejpam-86	136	10	then	then	ADV
ejpam-86	136	11	y(i	y(i	NOUN
ejpam-86	136	12	)	)	PUNCT
ejpam-86	136	13	∼	∼	NOUN
ejpam-86	136	14	pep(b′x(i),σ	pep(b′x(i),σ	NOUN
ejpam-86	136	15	,	,	PUNCT
ejpam-86	136	16	β	β	NOUN
ejpam-86	136	17	)	)	PUNCT
ejpam-86	136	18	.	.	PUNCT
ejpam-86	137	1	we	we	PRON
ejpam-86	137	2	denote	denote	VERB
ejpam-86	137	3	the	the	DET
ejpam-86	137	4	multivariate	multivariate	NOUN
ejpam-86	137	5	linear	linear	PROPN
ejpam-86	137	6	regression	regression	NOUN
ejpam-86	137	7	model	model	NOUN
ejpam-86	137	8	under	under	ADP
ejpam-86	137	9	this	this	DET
ejpam-86	137	10	assumption	assumption	NOUN
ejpam-86	137	11	as	as	ADP
ejpam-86	137	12	type	type	NOUN
ejpam-86	137	13	i	i	PROPN
ejpam-86	137	14	mvper	mvper	NOUN
ejpam-86	137	15	model	model	NOUN
ejpam-86	137	16	.	.	PUNCT
ejpam-86	138	1	in	in	ADP
ejpam-86	138	2	this	this	DET
ejpam-86	138	3	case	case	NOUN
ejpam-86	138	4	,	,	PUNCT
ejpam-86	138	5	the	the	DET
ejpam-86	138	6	likelihood	likelihood	NOUN
ejpam-86	138	7	function	function	NOUN
ejpam-86	138	8	is	be	AUX
ejpam-86	138	9	l(b	l(b	PROPN
ejpam-86	138	10	,	,	PUNCT
ejpam-86	138	11	σ	σ	PROPN
ejpam-86	138	12	,	,	PUNCT
ejpam-86	138	13	β|y	β|y	NUM
ejpam-86	138	14	,	,	PUNCT
ejpam-86	138	15	x	x	NOUN
ejpam-86	138	16	)	)	PUNCT
ejpam-86	138	17	=	=	SYM
ejpam-86	138	18	k1	k1	NOUN
ejpam-86	138	19	exp(−1	exp(−1	NUM
ejpam-86	138	20	2	2	NUM
ejpam-86	138	21	n∑	n∑	NOUN
ejpam-86	138	22	i=1	i=1	PROPN
ejpam-86	139	1	(	(	PUNCT
ejpam-86	139	2	(	(	PUNCT
ejpam-86	139	3	y(i)−b′x(i	y(i)−b′x(i	NOUN
ejpam-86	139	4	)	)	PUNCT
ejpam-86	139	5	)	)	PUNCT
ejpam-86	139	6	′σ−1(y(i)−b′x(i	′σ−1(y(i)−b′x(i	PROPN
ejpam-86	139	7	)	)	PUNCT
ejpam-86	139	8	)	)	PUNCT
ejpam-86	139	9	)	)	PUNCT
ejpam-86	140	1	β	β	X
ejpam-86	140	2	)	)	PUNCT
ejpam-86	140	3	,	,	PUNCT
ejpam-86	140	4	(	(	PUNCT
ejpam-86	140	5	3.1	3.1	NUM
ejpam-86	140	6	)	)	PUNCT
ejpam-86	140	7	where	where	SCONJ
ejpam-86	140	8	k1	k1	NOUN
ejpam-86	140	9	=	=	SYM
ejpam-86	140	10	pnγn(p	pnγn(p	PROPN
ejpam-86	140	11	2	2	NUM
ejpam-86	140	12	)	)	PUNCT
ejpam-86	140	13	πnp/2γn(1	πnp/2γn(1	NOUN
ejpam-86	140	14	+	+	CCONJ
ejpam-86	140	15	p	p	NOUN
ejpam-86	140	16	2β	2β	NOUN
ejpam-86	140	17	)	)	PUNCT
ejpam-86	140	18	2n+np	2n+np	NUM
ejpam-86	140	19	2β	2β	NUM
ejpam-86	140	20	|σ|−n/2	|σ|−n/2	PROPN
ejpam-86	140	21	.	.	PUNCT
ejpam-86	141	1	the	the	DET
ejpam-86	141	2	log	log	NOUN
ejpam-86	141	3	likelihood	likelihood	NOUN
ejpam-86	141	4	function	function	NOUN
ejpam-86	141	5	is	be	AUX
ejpam-86	141	6	l(b	l(b	PROPN
ejpam-86	141	7	,	,	PUNCT
ejpam-86	141	8	σ	σ	PROPN
ejpam-86	141	9	,	,	PUNCT
ejpam-86	141	10	β|y	β|y	NUM
ejpam-86	141	11	,	,	PUNCT
ejpam-86	141	12	x	x	NOUN
ejpam-86	141	13	)	)	PUNCT
ejpam-86	141	14	≡	≡	PROPN
ejpam-86	141	15	logl(b	logl(b	PROPN
ejpam-86	141	16	,	,	PUNCT
ejpam-86	141	17	σ	σ	PROPN
ejpam-86	141	18	,	,	PUNCT
ejpam-86	141	19	β|y	β|y	NUM
ejpam-86	141	20	,	,	PUNCT
ejpam-86	141	21	x	x	X
ejpam-86	141	22	)	)	PUNCT
ejpam-86	141	23	=	=	SYM
ejpam-86	141	24	n	n	CCONJ
ejpam-86	141	25	log(pγ(p	log(pγ(p	PROPN
ejpam-86	141	26	2))−	2))−	NUM
ejpam-86	141	27	np	np	NOUN
ejpam-86	141	28	2	2	NUM
ejpam-86	141	29	log(π)−	log(π)−	NOUN
ejpam-86	141	30	n	n	CCONJ
ejpam-86	141	31	log	log	NOUN
ejpam-86	141	32	γ(1	γ(1	PROPN
ejpam-86	142	1	+	+	CCONJ
ejpam-86	142	2	p	p	NOUN
ejpam-86	142	3	2β	2β	NOUN
ejpam-86	142	4	)	)	PUNCT
ejpam-86	143	1	−	−	ADP
ejpam-86	143	2	n(1	n(1	NOUN
ejpam-86	143	3	+	+	CCONJ
ejpam-86	143	4	p/2β	p/2β	ADJ
ejpam-86	143	5	)	)	PUNCT
ejpam-86	143	6	log	log	VERB
ejpam-86	143	7	2	2	NUM
ejpam-86	143	8	−n	−n	SYM
ejpam-86	143	9	2	2	NUM
ejpam-86	143	10	log	log	NOUN
ejpam-86	143	11	|σ|	|σ|	PROPN
ejpam-86	143	12	−	−	PROPN
ejpam-86	143	13	1	1	NUM
ejpam-86	143	14	2	2	NUM
ejpam-86	143	15	∑n	∑n	PROPN
ejpam-86	143	16	i=1((y(i)−b′x(i	i=1((y(i)−b′x(i	PROPN
ejpam-86	143	17	)	)	PUNCT
ejpam-86	143	18	)	)	PUNCT
ejpam-86	143	19	′σ−1(y(i)−b′x(i)))β	′σ−1(y(i)−b′x(i)))β	NOUN
ejpam-86	143	20	.	.	PUNCT
ejpam-86	144	1	(	(	PUNCT
ejpam-86	144	2	3.2	3.2	NUM
ejpam-86	144	3	)	)	PUNCT
ejpam-86	144	4	let	let	VERB
ejpam-86	144	5	θ=	θ=	NOUN
ejpam-86	144	6	(	(	PUNCT
ejpam-86	144	7	b′	b′	NUM
ejpam-86	144	8	,	,	PUNCT
ejpam-86	144	9	v	v	ADP
ejpam-86	144	10	ec′(σ	ec′(σ	PROPN
ejpam-86	144	11	)	)	PUNCT
ejpam-86	144	12	,	,	PUNCT
ejpam-86	144	13	β)′	β)′	NOUN
ejpam-86	144	14	,	,	PUNCT
ejpam-86	144	15	ε(i	ε(i	PROPN
ejpam-86	144	16	)	)	PUNCT
ejpam-86	144	17	=	=	SYM
ejpam-86	144	18	y(i)−b′x(i	y(i)−b′x(i	PROPN
ejpam-86	144	19	)	)	PUNCT
ejpam-86	144	20	and	and	CCONJ
ejpam-86	144	21	ti	ti	NOUN
ejpam-86	144	22	=	=	SYM
ejpam-86	144	23	ε′−1	ε′−1	X
ejpam-86	144	24	(	(	PUNCT
ejpam-86	144	25	i	i	NOUN
ejpam-86	144	26	)	)	PUNCT
ejpam-86	144	27	ε(i	ε(i	NOUN
ejpam-86	144	28	)	)	PUNCT
ejpam-86	144	29	.	.	PUNCT
ejpam-86	145	1	differentiating	differentiate	VERB
ejpam-86	145	2	(	(	PUNCT
ejpam-86	145	3	3.2	3.2	NUM
ejpam-86	145	4	)	)	PUNCT
ejpam-86	145	5	with	with	ADP
ejpam-86	145	6	respect	respect	NOUN
ejpam-86	145	7	to	to	ADP
ejpam-86	145	8	b	b	NUM
ejpam-86	145	9	,	,	PUNCT
ejpam-86	145	10	v	v	NOUN
ejpam-86	145	11	ec(σ	ec(σ	NOUN
ejpam-86	145	12	)	)	PUNCT
ejpam-86	145	13	and	and	CCONJ
ejpam-86	145	14	β	β	X
ejpam-86	145	15	respectively	respectively	ADV
ejpam-86	145	16	,	,	PUNCT
ejpam-86	145	17	we	we	PRON
ejpam-86	145	18	have	have	VERB
ejpam-86	145	19	∂l(θ	∂l(θ	PROPN
ejpam-86	145	20	)	)	PUNCT
ejpam-86	145	21	∂b	∂b	PROPN
ejpam-86	145	22	=	=	PUNCT
ejpam-86	146	1	β	β	X
ejpam-86	146	2	n∑	n∑	X
ejpam-86	146	3	i=1	i=1	PROPN
ejpam-86	147	1	tβ−1	tβ−1	INTJ
ejpam-86	147	2	i	i	PRON
ejpam-86	147	3	v	v	VERB
ejpam-86	147	4	ec(σ−1ε(i)x	ec(σ−1ε(i)x	PROPN
ejpam-86	148	1	′	′	NUM
ejpam-86	148	2	(	(	PUNCT
ejpam-86	148	3	i	i	NOUN
ejpam-86	148	4	)	)	PUNCT
ejpam-86	148	5	)	)	PUNCT
ejpam-86	148	6	,	,	PUNCT
ejpam-86	148	7	(	(	PUNCT
ejpam-86	148	8	3.3	3.3	NUM
ejpam-86	148	9	)	)	PUNCT
ejpam-86	148	10	∂l(θ	∂l(θ	PROPN
ejpam-86	148	11	)	)	PUNCT
ejpam-86	148	12	∂v	∂v	PROPN
ejpam-86	148	13	ec(σ	ec(σ	PRON
ejpam-86	148	14	)	)	PUNCT
ejpam-86	148	15	=	=	PUNCT
ejpam-86	149	1	−n	−n	ADJ
ejpam-86	149	2	2	2	NUM
ejpam-86	149	3	v	v	NOUN
ejpam-86	149	4	ec(σ−1	ec(σ−1	X
ejpam-86	149	5	)	)	PUNCT
ejpam-86	150	1	+	+	CCONJ
ejpam-86	150	2	β	β	X
ejpam-86	150	3	2	2	NUM
ejpam-86	150	4	n∑	n∑	NOUN
ejpam-86	150	5	i=1	i=1	PRON
ejpam-86	151	1	tβ−1	tβ−1	NOUN
ejpam-86	151	2	i	i	PRON
ejpam-86	151	3	v	v	VERB
ejpam-86	151	4	ec(σ−1ε(i)ε	ec(σ−1ε(i)ε	NOUN
ejpam-86	151	5	′−1	′−1	X
ejpam-86	151	6	(	(	PUNCT
ejpam-86	151	7	i	i	NOUN
ejpam-86	151	8	)	)	PUNCT
ejpam-86	151	9	)	)	PUNCT
ejpam-86	151	10	,	,	PUNCT
ejpam-86	151	11	(	(	PUNCT
ejpam-86	151	12	3.4	3.4	NUM
ejpam-86	151	13	)	)	PUNCT
ejpam-86	151	14	m.	m.	NOUN
ejpam-86	151	15	liu	liu	PROPN
ejpam-86	151	16	and	and	CCONJ
ejpam-86	151	17	h.	h.	PROPN
ejpam-86	151	18	bozdogan	bozdogan	PROPN
ejpam-86	151	19	/	/	SYM
ejpam-86	151	20	eur	eur	PROPN
ejpam-86	151	21	.	.	PUNCT
ejpam-86	152	1	j.	j.	PROPN
ejpam-86	152	2	pure	pure	PROPN
ejpam-86	152	3	appl	appl	PROPN
ejpam-86	152	4	.	.	PROPN
ejpam-86	152	5	math	math	PROPN
ejpam-86	152	6	,	,	PUNCT
ejpam-86	152	7	1	1	NUM
ejpam-86	152	8	(	(	PUNCT
ejpam-86	152	9	2008	2008	NUM
ejpam-86	152	10	)	)	PUNCT
ejpam-86	152	11	,	,	PUNCT
ejpam-86	152	12	(	(	PUNCT
ejpam-86	152	13	4	4	NUM
ejpam-86	152	14	-	-	SYM
ejpam-86	152	15	37	37	NUM
ejpam-86	152	16	)	)	PUNCT
ejpam-86	152	17	10	10	NUM
ejpam-86	152	18	and	and	CCONJ
ejpam-86	152	19	∂l(θ	∂l(θ	PROPN
ejpam-86	152	20	)	)	PUNCT
ejpam-86	153	1	∂β	∂β	PROPN
ejpam-86	153	2	=	=	PUNCT
ejpam-86	154	1	np	np	PROPN
ejpam-86	154	2	2β2	2β2	NUM
ejpam-86	154	3	ψ(1	ψ(1	VERB
ejpam-86	154	4	+	+	PUNCT
ejpam-86	154	5	p	p	NOUN
ejpam-86	154	6	2β	2β	NOUN
ejpam-86	154	7	)	)	PUNCT
ejpam-86	155	1	+	+	CCONJ
ejpam-86	155	2	np	np	PRON
ejpam-86	155	3	2β2	2β2	NUM
ejpam-86	155	4	log	log	NOUN
ejpam-86	155	5	2−	2−	NUM
ejpam-86	155	6	1	1	NUM
ejpam-86	155	7	2	2	NUM
ejpam-86	155	8	n∑	n∑	NOUN
ejpam-86	155	9	i=1	i=1	PROPN
ejpam-86	155	10	tβi	tβi	NOUN
ejpam-86	155	11	log	log	NOUN
ejpam-86	155	12	ti	ti	NOUN
ejpam-86	155	13	,	,	PUNCT
ejpam-86	155	14	(	(	PUNCT
ejpam-86	155	15	3.5	3.5	NUM
ejpam-86	155	16	)	)	PUNCT
ejpam-86	155	17	where	where	SCONJ
ejpam-86	155	18	ψ	ψ	X
ejpam-86	155	19	(	(	PUNCT
ejpam-86	155	20	·	·	PUNCT
ejpam-86	155	21	)	)	PUNCT
ejpam-86	155	22	=	=	PUNCT
ejpam-86	156	1	d	d	X
ejpam-86	156	2	log	log	NOUN
ejpam-86	156	3	γ(x)/dx	γ(x)/dx	INTJ
ejpam-86	156	4	is	be	AUX
ejpam-86	156	5	called	call	VERB
ejpam-86	156	6	digamma	digamma	PROPN
ejpam-86	156	7	function	function	NOUN
ejpam-86	156	8	(	(	PUNCT
ejpam-86	156	9	[	[	X
ejpam-86	156	10	1	1	NUM
ejpam-86	156	11	]	]	NUM
ejpam-86	156	12	)	)	PUNCT
ejpam-86	156	13	,	,	PUNCT
ejpam-86	156	14	or	or	CCONJ
ejpam-86	156	15	psi	psi	NOUN
ejpam-86	156	16	function	function	NOUN
ejpam-86	156	17	.	.	PUNCT
ejpam-86	157	1	to	to	PART
ejpam-86	157	2	derive	derive	VERB
ejpam-86	157	3	and	and	CCONJ
ejpam-86	157	4	construct	construct	VERB
ejpam-86	157	5	the	the	DET
ejpam-86	157	6	fisher	fisher	PROPN
ejpam-86	157	7	information	information	NOUN
ejpam-86	157	8	matrix	matrix	NOUN
ejpam-86	157	9	(	(	PUNCT
ejpam-86	157	10	fim	fim	NOUN
ejpam-86	157	11	)	)	PUNCT
ejpam-86	157	12	and	and	CCONJ
ejpam-86	157	13	its	its	PRON
ejpam-86	157	14	inverse	inverse	NOUN
ejpam-86	157	15	ifim	ifim	NOUN
ejpam-86	157	16	,	,	PUNCT
ejpam-86	157	17	we	we	PRON
ejpam-86	157	18	have	have	VERB
ejpam-86	157	19	∂2(θ	∂2(θ	ADJ
ejpam-86	157	20	)	)	PUNCT
ejpam-86	157	21	∂b′∂b	∂b′∂b	PUNCT
ejpam-86	158	1	=	=	PUNCT
ejpam-86	158	2	β	β	X
ejpam-86	158	3	n∑	n∑	X
ejpam-86	158	4	i=1	i=1	PROPN
ejpam-86	158	5	(	(	PUNCT
ejpam-86	158	6	ipq	ipq	NOUN
ejpam-86	158	7	⊗	⊗	PROPN
ejpam-86	158	8	v	v	NOUN
ejpam-86	158	9	ec′(ip))(iq	ec′(ip))(iq	NOUN
ejpam-86	158	10	⊗mi	⊗mi	NUM
ejpam-86	158	11	⊗	⊗	PROPN
ejpam-86	158	12	ip)(v	ip)(v	ADJ
ejpam-86	158	13	ec(iq)⊗	ec(iq)⊗	NOUN
ejpam-86	158	14	ipq	ipq	NOUN
ejpam-86	158	15	)	)	PUNCT
ejpam-86	158	16	,	,	PUNCT
ejpam-86	158	17	(	(	PUNCT
ejpam-86	158	18	3.6	3.6	NUM
ejpam-86	158	19	)	)	PUNCT
ejpam-86	158	20	where	where	SCONJ
ejpam-86	158	21	mi	mi	NOUN
ejpam-86	159	1	=	=	PROPN
ejpam-86	159	2	−tβ−1	−tβ−1	PUNCT
ejpam-86	159	3	i	i	PRON
ejpam-86	159	4	v	v	X
ejpam-86	159	5	ec(σ−1)v	ec(σ−1)v	NOUN
ejpam-86	159	6	ec′(x(i)x	ec′(x(i)x	PUNCT
ejpam-86	159	7	′	′	NUM
ejpam-86	159	8	(	(	PUNCT
ejpam-86	159	9	i))−	i))−	NOUN
ejpam-86	159	10	2(β	2(β	NUM
ejpam-86	159	11	−	−	NOUN
ejpam-86	160	1	1)tβ−2	1)tβ−2	NUM
ejpam-86	161	1	i	i	INTJ
ejpam-86	161	2	(	(	PUNCT
ejpam-86	161	3	σ−1ε(i)x	σ−1ε(i)x	PROPN
ejpam-86	161	4	′	′	NUM
ejpam-86	161	5	(	(	PUNCT
ejpam-86	161	6	i))⊗	i))⊗	NOUN
ejpam-86	161	7	(	(	PUNCT
ejpam-86	161	8	σ−1ε(i)x	σ−1ε(i)x	NUM
ejpam-86	161	9	′	′	NUM
ejpam-86	161	10	(	(	PUNCT
ejpam-86	161	11	i	i	NOUN
ejpam-86	161	12	)	)	PUNCT
ejpam-86	161	13	)	)	PUNCT
ejpam-86	161	14	.	.	PUNCT
ejpam-86	162	1	∂2l(θ	∂2l(θ	NOUN
ejpam-86	162	2	)	)	PUNCT
ejpam-86	162	3	∂v	∂v	PROPN
ejpam-86	162	4	ec′(σ)∂v	ec′(σ)∂v	NOUN
ejpam-86	162	5	ec(σ	ec(σ	NUM
ejpam-86	162	6	)	)	PUNCT
ejpam-86	162	7	=	=	SYM
ejpam-86	163	1	(	(	PUNCT
ejpam-86	163	2	ip2	ip2	PROPN
ejpam-86	163	3	⊗	⊗	PROPN
ejpam-86	163	4	v	v	ADP
ejpam-86	163	5	ec′(ip))(ip	ec′(ip))(ip	PROPN
ejpam-86	163	6	⊗	⊗	PROPN
ejpam-86	163	7	∂2l(θ	∂2l(θ	NOUN
ejpam-86	163	8	)	)	PUNCT
ejpam-86	163	9	∂σ∂σ	∂σ∂σ	NOUN
ejpam-86	163	10	⊗	⊗	PROPN
ejpam-86	163	11	ip)(v	ip)(v	PROPN
ejpam-86	163	12	ec(ip)⊗	ec(ip)⊗	PROPN
ejpam-86	163	13	ip2	ip2	PROPN
ejpam-86	163	14	)	)	PUNCT
ejpam-86	163	15	)	)	PUNCT
ejpam-86	163	16	,	,	PUNCT
ejpam-86	163	17	(	(	PUNCT
ejpam-86	163	18	3.7	3.7	NUM
ejpam-86	163	19	)	)	PUNCT
ejpam-86	163	20	where	where	SCONJ
ejpam-86	163	21	∂2l(θ	∂2l(θ	NOUN
ejpam-86	163	22	)	)	PUNCT
ejpam-86	163	23	∂σ∂σ	∂σ∂σ	NOUN
ejpam-86	163	24	=	=	SYM
ejpam-86	163	25	n	n	PROPN
ejpam-86	163	26	2v	2v	PROPN
ejpam-86	163	27	ec(σ	ec(σ	CCONJ
ejpam-86	163	28	−1)v	−1)v	X
ejpam-86	163	29	ec′−1	ec′−1	PROPN
ejpam-86	163	30	)	)	PUNCT
ejpam-86	163	31	−β	−β	NOUN
ejpam-86	163	32	2	2	NUM
ejpam-86	163	33	n∑	n∑	NOUN
ejpam-86	163	34	i=1	i=1	PRON
ejpam-86	164	1	tβ−1	tβ−1	ADV
ejpam-86	164	2	i	i	NOUN
ejpam-86	164	3	(	(	PUNCT
ejpam-86	164	4	v	v	NOUN
ejpam-86	164	5	ec(σ−1)v	ec(σ−1)v	VERB
ejpam-86	164	6	ec′−1ε(i)ε	ec′−1ε(i)ε	NOUN
ejpam-86	164	7	′−1	′−1	SYM
ejpam-86	164	8	(	(	PUNCT
ejpam-86	164	9	i	i	NOUN
ejpam-86	164	10	)	)	PUNCT
ejpam-86	164	11	)	)	PUNCT
ejpam-86	165	1	+	+	ADP
ejpam-86	165	2	v	v	NUM
ejpam-86	165	3	ec(σ−1ε(i)ε	ec(σ−1ε(i)ε	NOUN
ejpam-86	165	4	′−1	′−1	X
ejpam-86	165	5	(	(	PUNCT
ejpam-86	165	6	i	i	NOUN
ejpam-86	165	7	)	)	PUNCT
ejpam-86	165	8	)	)	PUNCT
ejpam-86	165	9	v	v	NOUN
ejpam-86	165	10	ec′−1	ec′−1	NOUN
ejpam-86	165	11	)	)	PUNCT
ejpam-86	165	12	)	)	PUNCT
ejpam-86	165	13	−β(β−1	−β(β−1	NUM
ejpam-86	165	14	)	)	PUNCT
ejpam-86	165	15	2	2	NUM
ejpam-86	165	16	n∑	n∑	NOUN
ejpam-86	165	17	i=1	i=1	PROPN
ejpam-86	165	18	tβ−2	tβ−2	PROPN
ejpam-86	166	1	i	i	PRON
ejpam-86	166	2	(	(	PUNCT
ejpam-86	166	3	σ−1ε(i)ε	σ−1ε(i)ε	NUM
ejpam-86	166	4	′−1	′−1	SYM
ejpam-86	166	5	(	(	PUNCT
ejpam-86	166	6	i	i	NOUN
ejpam-86	166	7	)	)	PUNCT
ejpam-86	166	8	)	)	PUNCT
ejpam-86	167	1	⊗	⊗	PROPN
ejpam-86	167	2	(	(	PUNCT
ejpam-86	167	3	σ−1ε(i)ε	σ−1ε(i)ε	NUM
ejpam-86	167	4	′−1	′−1	SYM
ejpam-86	167	5	(	(	PUNCT
ejpam-86	167	6	i	i	NOUN
ejpam-86	167	7	)	)	PUNCT
ejpam-86	167	8	)	)	PUNCT
ejpam-86	167	9	.	.	PUNCT
ejpam-86	168	1	∂2l(θ	∂2l(θ	NOUN
ejpam-86	168	2	)	)	PUNCT
ejpam-86	168	3	∂β2	∂β2	ADJ
ejpam-86	168	4	=	=	PUNCT
ejpam-86	168	5	−np	−np	X
ejpam-86	168	6	β3ψ(1	β3ψ(1	PUNCT
ejpam-86	168	7	+	+	PUNCT
ejpam-86	168	8	p	p	PRON
ejpam-86	168	9	2β	2β	NOUN
ejpam-86	168	10	)	)	PUNCT
ejpam-86	168	11	−	−	PROPN
ejpam-86	168	12	np2	np2	NOUN
ejpam-86	168	13	4β4ψ	4β4ψ	NOUN
ejpam-86	168	14	′(1	′(1	PROPN
ejpam-86	168	15	+	+	CCONJ
ejpam-86	168	16	p	p	NOUN
ejpam-86	168	17	2β	2β	NOUN
ejpam-86	168	18	)	)	PUNCT
ejpam-86	168	19	−	−	PROPN
ejpam-86	169	1	np	np	PRON
ejpam-86	169	2	β3	β3	INTJ
ejpam-86	169	3	log	log	VERB
ejpam-86	169	4	2−	2−	NUM
ejpam-86	169	5	1	1	NUM
ejpam-86	169	6	2	2	NUM
ejpam-86	169	7	n∑	n∑	NOUN
ejpam-86	169	8	i=1	i=1	PROPN
ejpam-86	169	9	tβi	tβi	VERB
ejpam-86	169	10	log2	log2	PROPN
ejpam-86	169	11	ti	ti	PROPN
ejpam-86	169	12	,	,	PUNCT
ejpam-86	169	13	(	(	PUNCT
ejpam-86	169	14	3.8	3.8	NUM
ejpam-86	169	15	)	)	PUNCT
ejpam-86	169	16	where	where	SCONJ
ejpam-86	169	17	ψ′	ψ′	PROPN
ejpam-86	169	18	(	(	PUNCT
ejpam-86	169	19	·	·	PUNCT
ejpam-86	169	20	)	)	PUNCT
ejpam-86	169	21	=	=	SYM
ejpam-86	169	22	d2	d2	PROPN
ejpam-86	169	23	log	log	PROPN
ejpam-86	169	24	γ(x)/dx2	γ(x)/dx2	PROPN
ejpam-86	169	25	is	be	AUX
ejpam-86	169	26	called	call	VERB
ejpam-86	169	27	trigamma	trigamma	PROPN
ejpam-86	169	28	function	function	PROPN
ejpam-86	169	29	.	.	PUNCT
ejpam-86	170	1	∂2l(θ	∂2l(θ	NOUN
ejpam-86	170	2	)	)	PUNCT
ejpam-86	171	1	∂v	∂v	PROPN
ejpam-86	171	2	ec′(σ)∂b	ec′(σ)∂b	X
ejpam-86	172	1	=	=	NOUN
ejpam-86	172	2	β	β	X
ejpam-86	172	3	n∑	n∑	X
ejpam-86	172	4	i=1	i=1	PROPN
ejpam-86	172	5	(	(	PUNCT
ejpam-86	172	6	ipq	ipq	NOUN
ejpam-86	172	7	⊗	⊗	PROPN
ejpam-86	172	8	v	v	PROPN
ejpam-86	172	9	ec′(ip))(iq	ec′(ip))(iq	PROPN
ejpam-86	172	10	⊗ni	⊗ni	NUM
ejpam-86	172	11	⊗	⊗	PROPN
ejpam-86	172	12	ip)(v	ip)(v	PROPN
ejpam-86	172	13	ec(iq)⊗	ec(iq)⊗	PROPN
ejpam-86	172	14	ip2	ip2	PROPN
ejpam-86	172	15	)	)	PUNCT
ejpam-86	172	16	,	,	PUNCT
ejpam-86	172	17	(	(	PUNCT
ejpam-86	172	18	3.9	3.9	NUM
ejpam-86	172	19	)	)	PUNCT
ejpam-86	173	1	where	where	SCONJ
ejpam-86	173	2	ni	ni	PROPN
ejpam-86	173	3	=	=	PROPN
ejpam-86	173	4	−tβ−1	−tβ−1	PUNCT
ejpam-86	173	5	i	i	PRON
ejpam-86	173	6	v	v	VERB
ejpam-86	173	7	ec(σ−1)v	ec(σ−1)v	ADJ
ejpam-86	173	8	ec′−1ε(i)x′(i	ec′−1ε(i)x′(i	NOUN
ejpam-86	173	9	)	)	PUNCT
ejpam-86	173	10	)	)	PUNCT
ejpam-86	173	11	−(β	−(β	VERB
ejpam-86	173	12	−	−	PROPN
ejpam-86	174	1	1)tβ−2	1)tβ−2	NUM
ejpam-86	174	2	i	i	INTJ
ejpam-86	174	3	(	(	PUNCT
ejpam-86	174	4	σ−1ε(i)x	σ−1ε(i)x	PROPN
ejpam-86	174	5	′−1	′−1	X
ejpam-86	174	6	(	(	PUNCT
ejpam-86	174	7	i	i	NOUN
ejpam-86	174	8	)	)	PUNCT
ejpam-86	174	9	ε(i)ε	ε(i)ε	PROPN
ejpam-86	174	10	′−1	′−1	PUNCT
ejpam-86	174	11	(	(	PUNCT
ejpam-86	174	12	i	i	NOUN
ejpam-86	174	13	)	)	PUNCT
ejpam-86	174	14	)	)	PUNCT
ejpam-86	174	15	.	.	PUNCT
ejpam-86	175	1	∂2l(θ	∂2l(θ	NOUN
ejpam-86	175	2	)	)	PUNCT
ejpam-86	175	3	∂β∂b	∂β∂b	PROPN
ejpam-86	175	4	=	=	PUNCT
ejpam-86	175	5	n∑	n∑	PROPN
ejpam-86	175	6	i=1	i=1	PROPN
ejpam-86	176	1	tβ−1	tβ−1	NOUN
ejpam-86	176	2	i	i	PRON
ejpam-86	176	3	v	v	ADP
ejpam-86	176	4	ec(σ−1ε(i)x′(i	ec(σ−1ε(i)x′(i	NUM
ejpam-86	176	5	)	)	PUNCT
ejpam-86	176	6	)	)	PUNCT
ejpam-86	177	1	+	+	CCONJ
ejpam-86	177	2	β	β	AUX
ejpam-86	177	3	n∑	n∑	X
ejpam-86	177	4	i=1	i=1	X
ejpam-86	178	1	tβ−1	tβ−1	VERB
ejpam-86	178	2	i	i	PRON
ejpam-86	178	3	log(ti)v	log(ti)v	VERB
ejpam-86	178	4	ec(σ−1ε(i)x′(i	ec(σ−1ε(i)x′(i	NOUN
ejpam-86	178	5	)	)	PUNCT
ejpam-86	178	6	)	)	PUNCT
ejpam-86	178	7	.	.	PUNCT
ejpam-86	179	1	(	(	PUNCT
ejpam-86	179	2	3.10	3.10	NUM
ejpam-86	179	3	)	)	PUNCT
ejpam-86	179	4	∂2l(θ	∂2l(θ	NOUN
ejpam-86	179	5	)	)	PUNCT
ejpam-86	179	6	∂β∂v	∂β∂v	PROPN
ejpam-86	179	7	ec(σ	ec(σ	PUNCT
ejpam-86	179	8	)	)	PUNCT
ejpam-86	179	9	=	=	SYM
ejpam-86	180	1	1	1	NUM
ejpam-86	180	2	2	2	NUM
ejpam-86	180	3	n∑	n∑	NOUN
ejpam-86	180	4	i=1	i=1	PRON
ejpam-86	181	1	tβ−1	tβ−1	NOUN
ejpam-86	181	2	i	i	PRON
ejpam-86	181	3	v	v	VERB
ejpam-86	181	4	ec(σ−1ε(i)ε	ec(σ−1ε(i)ε	NOUN
ejpam-86	181	5	′−1	′−1	X
ejpam-86	181	6	(	(	PUNCT
ejpam-86	181	7	i	i	NOUN
ejpam-86	181	8	)	)	PUNCT
ejpam-86	181	9	)	)	PUNCT
ejpam-86	182	1	+	+	VERB
ejpam-86	182	2	β	β	X
ejpam-86	182	3	2	2	NUM
ejpam-86	182	4	n∑	n∑	NOUN
ejpam-86	182	5	i=1	i=1	PRON
ejpam-86	182	6	tβ−1	tβ−1	VERB
ejpam-86	182	7	i	i	PRON
ejpam-86	182	8	log(ti)v	log(ti)v	VERB
ejpam-86	182	9	ec(σ−1ε(i)ε	ec(σ−1ε(i)ε	NOUN
ejpam-86	182	10	′−1	′−1	X
ejpam-86	182	11	(	(	PUNCT
ejpam-86	182	12	i	i	NOUN
ejpam-86	182	13	)	)	PUNCT
ejpam-86	182	14	)	)	PUNCT
ejpam-86	182	15	.	.	PUNCT
ejpam-86	183	1	(	(	PUNCT
ejpam-86	183	2	3.11	3.11	NUM
ejpam-86	183	3	)	)	PUNCT
ejpam-86	183	4	some	some	PRON
ejpam-86	183	5	of	of	ADP
ejpam-86	183	6	the	the	DET
ejpam-86	183	7	details	detail	NOUN
ejpam-86	183	8	of	of	ADP
ejpam-86	183	9	the	the	DET
ejpam-86	183	10	above	above	ADJ
ejpam-86	183	11	derivations	derivation	NOUN
ejpam-86	183	12	are	be	AUX
ejpam-86	183	13	given	give	VERB
ejpam-86	183	14	in	in	ADP
ejpam-86	183	15	the	the	DET
ejpam-86	183	16	appendix	appendix	NOUN
ejpam-86	183	17	.	.	PUNCT
ejpam-86	184	1	since	since	SCONJ
ejpam-86	184	2	there	there	PRON
ejpam-86	184	3	are	be	VERB
ejpam-86	184	4	no	no	DET
ejpam-86	184	5	closed	closed	ADJ
ejpam-86	184	6	form	form	NOUN
ejpam-86	184	7	solutions	solution	NOUN
ejpam-86	184	8	to	to	ADP
ejpam-86	184	9	the	the	DET
ejpam-86	184	10	likelihood	likelihood	NOUN
ejpam-86	184	11	equations	equation	NOUN
ejpam-86	184	12	,	,	PUNCT
ejpam-86	184	13	numerical	numerical	ADJ
ejpam-86	184	14	methods	method	NOUN
ejpam-86	184	15	such	such	ADJ
ejpam-86	184	16	as	as	ADP
ejpam-86	184	17	genetic	genetic	ADJ
ejpam-86	184	18	algorithms	algorithm	NOUN
ejpam-86	184	19	or	or	CCONJ
ejpam-86	184	20	newton	newton	PROPN
ejpam-86	184	21	-	-	PUNCT
ejpam-86	184	22	raphson	raphson	PROPN
ejpam-86	184	23	iterative	iterative	NOUN
ejpam-86	184	24	method	method	NOUN
ejpam-86	184	25	can	can	AUX
ejpam-86	184	26	be	be	AUX
ejpam-86	184	27	used	use	VERB
ejpam-86	184	28	to	to	PART
ejpam-86	184	29	obtain	obtain	VERB
ejpam-86	184	30	the	the	DET
ejpam-86	184	31	maximum	maximum	ADJ
ejpam-86	184	32	likelihood	likelihood	NOUN
ejpam-86	184	33	estimators	estimator	NOUN
ejpam-86	184	34	(	(	PUNCT
ejpam-86	184	35	mles	mle	NOUN
ejpam-86	184	36	)	)	PUNCT
ejpam-86	184	37	.	.	PUNCT
ejpam-86	185	1	here	here	ADV
ejpam-86	185	2	we	we	PRON
ejpam-86	185	3	give	give	VERB
ejpam-86	185	4	a	a	DET
ejpam-86	185	5	method	method	NOUN
ejpam-86	185	6	to	to	PART
ejpam-86	185	7	compute	compute	VERB
ejpam-86	185	8	the	the	DET
ejpam-86	185	9	method	method	NOUN
ejpam-86	185	10	of	of	ADP
ejpam-86	185	11	moments	moment	NOUN
ejpam-86	185	12	(	(	PUNCT
ejpam-86	185	13	mom	mom	NOUN
ejpam-86	185	14	)	)	PUNCT
ejpam-86	185	15	estimates	estimate	NOUN
ejpam-86	185	16	which	which	PRON
ejpam-86	185	17	can	can	AUX
ejpam-86	185	18	be	be	AUX
ejpam-86	185	19	used	use	VERB
ejpam-86	185	20	as	as	ADP
ejpam-86	185	21	the	the	DET
ejpam-86	185	22	starting	start	VERB
ejpam-86	185	23	values	value	NOUN
ejpam-86	185	24	to	to	PART
ejpam-86	185	25	calculate	calculate	VERB
ejpam-86	185	26	the	the	DET
ejpam-86	185	27	mles	mle	NOUN
ejpam-86	185	28	.	.	PUNCT
ejpam-86	186	1	the	the	DET
ejpam-86	186	2	steps	step	NOUN
ejpam-86	186	3	of	of	ADP
ejpam-86	186	4	mom	mom	NOUN
ejpam-86	186	5	are	be	AUX
ejpam-86	186	6	:	:	PUNCT
ejpam-86	186	7	m.	m.	PROPN
ejpam-86	186	8	liu	liu	PROPN
ejpam-86	186	9	and	and	CCONJ
ejpam-86	186	10	h.	h.	PROPN
ejpam-86	186	11	bozdogan	bozdogan	PROPN
ejpam-86	186	12	/	/	SYM
ejpam-86	186	13	eur	eur	PROPN
ejpam-86	186	14	.	.	PUNCT
ejpam-86	187	1	j.	j.	PROPN
ejpam-86	187	2	pure	pure	PROPN
ejpam-86	187	3	appl	appl	PROPN
ejpam-86	187	4	.	.	PROPN
ejpam-86	187	5	math	math	PROPN
ejpam-86	187	6	,	,	PUNCT
ejpam-86	187	7	1	1	NUM
ejpam-86	187	8	(	(	PUNCT
ejpam-86	187	9	2008	2008	NUM
ejpam-86	187	10	)	)	PUNCT
ejpam-86	187	11	,	,	PUNCT
ejpam-86	187	12	(	(	PUNCT
ejpam-86	187	13	4	4	NUM
ejpam-86	187	14	-	-	SYM
ejpam-86	187	15	37	37	NUM
ejpam-86	187	16	)	)	PUNCT
ejpam-86	187	17	11	11	NUM
ejpam-86	187	18	step	step	NOUN
ejpam-86	187	19	1	1	NUM
ejpam-86	187	20	:	:	PUNCT
ejpam-86	187	21	compute	compute	VERB
ejpam-86	187	22	b̂	b̂	NOUN
ejpam-86	187	23	=	=	SYM
ejpam-86	187	24	(	(	PUNCT
ejpam-86	187	25	x	x	X
ejpam-86	187	26	′x)−1x	′x)−1x	PROPN
ejpam-86	187	27	′y	′y	PROPN
ejpam-86	187	28	.	.	PUNCT
ejpam-86	188	1	step	step	NOUN
ejpam-86	188	2	2	2	NUM
ejpam-86	188	3	:	:	PUNCT
ejpam-86	188	4	let	let	VERB
ejpam-86	188	5	di	di	X
ejpam-86	188	6	=	=	PUNCT
ejpam-86	188	7	ε′(i)v	ε′(i)v	NOUN
ejpam-86	188	8	ar(y(i))−1ε(i	ar(y(i))−1ε(i	PROPN
ejpam-86	188	9	)	)	PUNCT
ejpam-86	188	10	.	.	PUNCT
ejpam-86	189	1	di	di	PROPN
ejpam-86	189	2	actually	actually	ADV
ejpam-86	189	3	is	be	AUX
ejpam-86	189	4	the	the	DET
ejpam-86	189	5	squared	square	VERB
ejpam-86	189	6	mahalanobis	mahalanobis	ADJ
ejpam-86	189	7	distances	distance	NOUN
ejpam-86	189	8	.	.	PUNCT
ejpam-86	190	1	by	by	ADP
ejpam-86	190	2	the	the	DET
ejpam-86	190	3	probability	probability	NOUN
ejpam-86	190	4	characteristics	characteristic	NOUN
ejpam-86	190	5	of	of	ADP
ejpam-86	190	6	multivariate	multivariate	NOUN
ejpam-86	190	7	pe	pe	ADP
ejpam-86	190	8	distribution	distribution	NOUN
ejpam-86	190	9	,	,	PUNCT
ejpam-86	190	10	we	we	PRON
ejpam-86	190	11	have	have	VERB
ejpam-86	190	12	e[t2i	e[t2i	NOUN
ejpam-86	190	13	]	]	PUNCT
ejpam-86	190	14	=	=	PUNCT
ejpam-86	191	1	22	22	NUM
ejpam-86	191	2	/	/	SYM
ejpam-86	191	3	βγ(p+4	βγ(p+4	NOUN
ejpam-86	191	4	2β	2β	NOUN
ejpam-86	191	5	)	)	PUNCT
ejpam-86	191	6	γ	γ	X
ejpam-86	191	7	(	(	PUNCT
ejpam-86	191	8	p	p	ADJ
ejpam-86	191	9	2β	2β	NOUN
ejpam-86	191	10	)	)	PUNCT
ejpam-86	191	11	.	.	PUNCT
ejpam-86	192	1	then	then	ADV
ejpam-86	192	2	e[d2	e[d2	VERB
ejpam-86	192	3	i	i	PRON
ejpam-86	192	4	]	]	PUNCT
ejpam-86	192	5	=	=	PUNCT
ejpam-86	193	1	e	e	X
ejpam-86	193	2	[	[	PUNCT
ejpam-86	193	3	p2γ2	p2γ2	ADP
ejpam-86	193	4	(	(	PUNCT
ejpam-86	193	5	p	p	ADJ
ejpam-86	193	6	2β	2β	NOUN
ejpam-86	193	7	)	)	PUNCT
ejpam-86	193	8	22	22	NUM
ejpam-86	193	9	/	/	SYM
ejpam-86	193	10	βγ2(p+2	βγ2(p+2	PROPN
ejpam-86	193	11	2β	2β	NOUN
ejpam-86	193	12	)	)	PUNCT
ejpam-86	193	13	t2i	t2i	CCONJ
ejpam-86	193	14	]	]	PUNCT
ejpam-86	193	15	=	=	PUNCT
ejpam-86	193	16	p2γ	p2γ	PROPN
ejpam-86	193	17	(	(	PUNCT
ejpam-86	193	18	p	p	NOUN
ejpam-86	193	19	2β	2β	NOUN
ejpam-86	193	20	)	)	PUNCT
ejpam-86	193	21	γ(p+4	γ(p+4	NOUN
ejpam-86	193	22	2β	2β	NOUN
ejpam-86	193	23	)	)	PUNCT
ejpam-86	193	24	γ2(p+2	γ2(p+2	ADV
ejpam-86	193	25	2β	2β	NOUN
ejpam-86	193	26	)	)	PUNCT
ejpam-86	193	27	.	.	PUNCT
ejpam-86	194	1	so	so	ADV
ejpam-86	194	2	we	we	PRON
ejpam-86	194	3	can	can	AUX
ejpam-86	194	4	compute	compute	VERB
ejpam-86	194	5	β̂	β̂	ADP
ejpam-86	194	6	as	as	ADP
ejpam-86	194	7	the	the	DET
ejpam-86	194	8	solution	solution	NOUN
ejpam-86	194	9	of	of	ADP
ejpam-86	194	10	p2γ	p2γ	PROPN
ejpam-86	194	11	(	(	PUNCT
ejpam-86	194	12	p	p	NOUN
ejpam-86	194	13	2β̂	2β̂	NUM
ejpam-86	194	14	)	)	PUNCT
ejpam-86	194	15	γ(p+4	γ(p+4	PROPN
ejpam-86	194	16	2β̂	2β̂	NUM
ejpam-86	194	17	)	)	PUNCT
ejpam-86	194	18	γ2(p+2	γ2(p+2	ADV
ejpam-86	194	19	2β̂	2β̂	NUM
ejpam-86	194	20	)	)	PUNCT
ejpam-86	195	1	=	=	SYM
ejpam-86	196	1	1	1	NUM
ejpam-86	196	2	n	n	NUM
ejpam-86	196	3	n∑	n∑	NOUN
ejpam-86	196	4	i=1	i=1	PROPN
ejpam-86	197	1	d̂2	d̂2	NOUN
ejpam-86	198	1	i	i	PRON
ejpam-86	198	2	where	where	SCONJ
ejpam-86	198	3	d̂i	d̂i	VERB
ejpam-86	198	4	=	=	SYM
ejpam-86	198	5	(	(	PUNCT
ejpam-86	198	6	y(i	y(i	PROPN
ejpam-86	198	7	)	)	PUNCT
ejpam-86	198	8	−	−	PROPN
ejpam-86	198	9	b̂′x(i))′−1(y(i	b̂′x(i))′−1(y(i	NOUN
ejpam-86	198	10	)	)	PUNCT
ejpam-86	198	11	−	−	PROPN
ejpam-86	198	12	b̂′x(i	b̂′x(i	NOUN
ejpam-86	198	13	)	)	PUNCT
ejpam-86	198	14	)	)	PUNCT
ejpam-86	199	1	and	and	CCONJ
ejpam-86	199	2	s	s	X
ejpam-86	199	3	=	=	PUNCT
ejpam-86	199	4	(	(	PUNCT
ejpam-86	199	5	y	y	PROPN
ejpam-86	199	6	−	−	PROPN
ejpam-86	199	7	xb̂)′(y	xb̂)′(y	PROPN
ejpam-86	199	8	−	−	PROPN
ejpam-86	199	9	xb̂)/n	xb̂)/n	PROPN
ejpam-86	199	10	is	be	AUX
ejpam-86	199	11	the	the	DET
ejpam-86	199	12	sample	sample	NOUN
ejpam-86	199	13	covariance	covariance	NOUN
ejpam-86	199	14	matrix	matrix	NOUN
ejpam-86	199	15	.	.	PUNCT
ejpam-86	200	1	step	step	NOUN
ejpam-86	200	2	3	3	NUM
ejpam-86	200	3	:	:	PUNCT
ejpam-86	200	4	compute	compute	VERB
ejpam-86	200	5	σ̂	σ̂	NOUN
ejpam-86	200	6	=	=	PUNCT
ejpam-86	200	7	pγ	pγ	PROPN
ejpam-86	200	8	(	(	PUNCT
ejpam-86	200	9	p	p	NOUN
ejpam-86	200	10	2β̂	2β̂	NUM
ejpam-86	200	11	)	)	PUNCT
ejpam-86	200	12	21	21	NUM
ejpam-86	200	13	/	/	SYM
ejpam-86	200	14	β̂γ(p+2	β̂γ(p+2	PROPN
ejpam-86	200	15	2β̂	2β̂	NUM
ejpam-86	200	16	)	)	PUNCT
ejpam-86	201	1	s.	s.	PROPN
ejpam-86	201	2	4	4	NUM
ejpam-86	201	3	.	.	PUNCT
ejpam-86	201	4	type	type	NOUN
ejpam-86	201	5	ii	ii	PROPN
ejpam-86	201	6	multivariate	multivariate	NOUN
ejpam-86	201	7	pe	pe	INTJ
ejpam-86	201	8	regression	regression	NOUN
ejpam-86	201	9	model	model	NOUN
ejpam-86	201	10	if	if	SCONJ
ejpam-86	201	11	we	we	PRON
ejpam-86	201	12	assume	assume	VERB
ejpam-86	201	13	that	that	SCONJ
ejpam-86	201	14	the	the	DET
ejpam-86	201	15	random	random	ADJ
ejpam-86	201	16	error	error	NOUN
ejpam-86	201	17	terms	term	NOUN
ejpam-86	201	18	v	v	ADP
ejpam-86	201	19	ec(e′	ec(e′	NOUN
ejpam-86	201	20	)	)	PUNCT
ejpam-86	201	21	in	in	ADP
ejpam-86	201	22	model	model	NOUN
ejpam-86	201	23	(	(	PUNCT
ejpam-86	201	24	1.2	1.2	NUM
ejpam-86	201	25	)	)	PUNCT
ejpam-86	201	26	has	have	VERB
ejpam-86	201	27	a	a	DET
ejpam-86	201	28	multivariate	multivariate	NOUN
ejpam-86	201	29	pe	pe	ADP
ejpam-86	201	30	distribution	distribution	NOUN
ejpam-86	201	31	penp(0,φ⊗	penp(0,φ⊗	NOUN
ejpam-86	201	32	σ	σ	PROPN
ejpam-86	201	33	,	,	PUNCT
ejpam-86	201	34	β	β	NOUN
ejpam-86	201	35	)	)	PUNCT
ejpam-86	201	36	,	,	PUNCT
ejpam-86	201	37	i.e.	i.e.	X
ejpam-86	201	38	,	,	PUNCT
ejpam-86	201	39	v	v	NUM
ejpam-86	201	40	ec(y	ec(y	NOUN
ejpam-86	201	41	′	′	NOUN
ejpam-86	201	42	)	)	PUNCT
ejpam-86	201	43	∼	∼	NOUN
ejpam-86	201	44	penp((x	penp((x	NOUN
ejpam-86	201	45	⊗	⊗	PROPN
ejpam-86	201	46	ip)bpq×1,φ⊗	ip)bpq×1,φ⊗	ADV
ejpam-86	201	47	σ	σ	PROPN
ejpam-86	201	48	,	,	PUNCT
ejpam-86	201	49	β	β	NOUN
ejpam-86	201	50	)	)	PUNCT
ejpam-86	201	51	or	or	CCONJ
ejpam-86	201	52	with	with	ADP
ejpam-86	201	53	the	the	DET
ejpam-86	201	54	matrix	matrix	NOUN
ejpam-86	201	55	notation	notation	NOUN
ejpam-86	201	56	y	y	PROPN
ejpam-86	201	57	∼	∼	NOUN
ejpam-86	201	58	mpen×p(xb	mpen×p(xb	PROPN
ejpam-86	201	59	,	,	PUNCT
ejpam-86	201	60	φ	φ	PROPN
ejpam-86	201	61	,	,	PUNCT
ejpam-86	201	62	σ	σ	PROPN
ejpam-86	201	63	,	,	PUNCT
ejpam-86	201	64	β	β	NOUN
ejpam-86	201	65	)	)	PUNCT
ejpam-86	201	66	,	,	PUNCT
ejpam-86	201	67	where	where	SCONJ
ejpam-86	201	68	φ	φ	PROPN
ejpam-86	201	69	is	be	AUX
ejpam-86	201	70	a	a	DET
ejpam-86	201	71	(	(	PUNCT
ejpam-86	201	72	n	n	NUM
ejpam-86	201	73	×	×	NOUN
ejpam-86	201	74	n	n	CCONJ
ejpam-86	201	75	)	)	PUNCT
ejpam-86	201	76	positive	positive	ADJ
ejpam-86	201	77	definite	definite	ADJ
ejpam-86	201	78	symmetric	symmetric	ADJ
ejpam-86	201	79	matrix	matrix	NOUN
ejpam-86	201	80	and	and	CCONJ
ejpam-86	201	81	known	know	VERB
ejpam-86	201	82	,	,	PUNCT
ejpam-86	201	83	then	then	ADV
ejpam-86	201	84	the	the	DET
ejpam-86	201	85	density	density	NOUN
ejpam-86	201	86	function	function	NOUN
ejpam-86	201	87	of	of	ADP
ejpam-86	201	88	e	e	PROPN
ejpam-86	201	89	is	be	AUX
ejpam-86	201	90	:	:	PUNCT
ejpam-86	201	91	f(e	f(e	NOUN
ejpam-86	201	92	;	;	PUNCT
ejpam-86	201	93	φ	φ	NUM
ejpam-86	201	94	,	,	PUNCT
ejpam-86	201	95	σ	σ	PROPN
ejpam-86	201	96	,	,	PUNCT
ejpam-86	201	97	β	β	X
ejpam-86	201	98	)	)	PUNCT
ejpam-86	201	99	=	=	SYM
ejpam-86	201	100	k|φ|−p/2|σ|−n/2	k|φ|−p/2|σ|−n/2	PROPN
ejpam-86	201	101	exp(−1	exp(−1	X
ejpam-86	201	102	2	2	NUM
ejpam-86	201	103	(	(	PUNCT
ejpam-86	201	104	tr(σ−1e′−1e))β	tr(σ−1e′−1e))β	NUM
ejpam-86	201	105	)	)	PUNCT
ejpam-86	201	106	(	(	PUNCT
ejpam-86	201	107	4.1	4.1	NUM
ejpam-86	201	108	)	)	PUNCT
ejpam-86	201	109	or	or	CCONJ
ejpam-86	201	110	,	,	PUNCT
ejpam-86	201	111	the	the	DET
ejpam-86	201	112	density	density	NOUN
ejpam-86	201	113	function	function	NOUN
ejpam-86	201	114	of	of	ADP
ejpam-86	201	115	y	y	PROPN
ejpam-86	201	116	is	be	AUX
ejpam-86	201	117	:	:	PUNCT
ejpam-86	201	118	f(y	f(y	NOUN
ejpam-86	201	119	;	;	PUNCT
ejpam-86	201	120	b	b	PROPN
ejpam-86	201	121	,	,	PUNCT
ejpam-86	201	122	φ	φ	PROPN
ejpam-86	201	123	,	,	PUNCT
ejpam-86	201	124	σ	σ	PROPN
ejpam-86	201	125	,	,	PUNCT
ejpam-86	201	126	β	β	X
ejpam-86	201	127	)	)	PUNCT
ejpam-86	202	1	=	=	SYM
ejpam-86	202	2	k|φ|−p/2|σ|−n/2	k|φ|−p/2|σ|−n/2	PROPN
ejpam-86	202	3	exp(−1	exp(−1	X
ejpam-86	202	4	2	2	NUM
ejpam-86	202	5	(	(	PUNCT
ejpam-86	202	6	tr(σ−1(y	tr(σ−1(y	NOUN
ejpam-86	202	7	−xb)′−1(y	−xb)′−1(y	ADP
ejpam-86	202	8	−xb)))β	−xb)))β	ADV
ejpam-86	202	9	)	)	PUNCT
ejpam-86	202	10	.	.	PUNCT
ejpam-86	203	1	(	(	PUNCT
ejpam-86	203	2	4.2	4.2	NUM
ejpam-86	203	3	)	)	PUNCT
ejpam-86	203	4	we	we	PRON
ejpam-86	203	5	denote	denote	VERB
ejpam-86	203	6	the	the	DET
ejpam-86	203	7	multivariate	multivariate	NOUN
ejpam-86	203	8	regression	regression	NOUN
ejpam-86	203	9	model	model	NOUN
ejpam-86	203	10	under	under	ADP
ejpam-86	203	11	this	this	DET
ejpam-86	203	12	assumption	assumption	NOUN
ejpam-86	203	13	as	as	ADP
ejpam-86	203	14	type	type	NOUN
ejpam-86	203	15	ii	ii	PROPN
ejpam-86	203	16	mvper	mvper	PROPN
ejpam-86	203	17	model	model	NOUN
ejpam-86	203	18	.	.	PUNCT
ejpam-86	204	1	it	it	PRON
ejpam-86	204	2	is	be	AUX
ejpam-86	204	3	noted	note	VERB
ejpam-86	204	4	that	that	SCONJ
ejpam-86	204	5	type	type	NOUN
ejpam-86	204	6	i	i	PRON
ejpam-86	204	7	and	and	CCONJ
ejpam-86	204	8	type	type	PROPN
ejpam-86	204	9	ii	ii	PROPN
ejpam-86	204	10	mvper	mvper	NOUN
ejpam-86	204	11	models	model	NOUN
ejpam-86	204	12	coincide	coincide	VERB
ejpam-86	204	13	only	only	ADV
ejpam-86	204	14	if	if	SCONJ
ejpam-86	204	15	β	β	X
ejpam-86	204	16	=	=	SYM
ejpam-86	204	17	1	1	NUM
ejpam-86	204	18	and	and	CCONJ
ejpam-86	204	19	φ	φ	NUM
ejpam-86	204	20	=	=	PUNCT
ejpam-86	204	21	in	in	ADP
ejpam-86	204	22	,	,	PUNCT
ejpam-86	204	23	which	which	PRON
ejpam-86	204	24	corresponds	correspond	VERB
ejpam-86	204	25	to	to	ADP
ejpam-86	204	26	the	the	DET
ejpam-86	204	27	multivariate	multivariate	NOUN
ejpam-86	204	28	normal	normal	ADJ
ejpam-86	204	29	distribution	distribution	NOUN
ejpam-86	204	30	.	.	PUNCT
ejpam-86	205	1	when	when	SCONJ
ejpam-86	205	2	φ	φ	PROPN
ejpam-86	205	3	=	=	SYM
ejpam-86	205	4	in	in	ADP
ejpam-86	205	5	,	,	PUNCT
ejpam-86	205	6	each	each	DET
ejpam-86	205	7	row	row	NOUN
ejpam-86	205	8	of	of	ADP
ejpam-86	205	9	e	e	NOUN
ejpam-86	205	10	,	,	PUNCT
ejpam-86	205	11	or	or	CCONJ
ejpam-86	205	12	y	y	PROPN
ejpam-86	205	13	,	,	PUNCT
ejpam-86	205	14	are	be	AUX
ejpam-86	205	15	still	still	ADV
ejpam-86	205	16	considered	consider	VERB
ejpam-86	205	17	as	as	ADP
ejpam-86	205	18	an	an	DET
ejpam-86	205	19	observation	observation	NOUN
ejpam-86	205	20	of	of	ADP
ejpam-86	205	21	a	a	DET
ejpam-86	205	22	p	p	X
ejpam-86	205	23	dimensional	dimensional	ADJ
ejpam-86	205	24	random	random	ADJ
ejpam-86	205	25	vector	vector	NOUN
ejpam-86	205	26	and	and	CCONJ
ejpam-86	205	27	the	the	DET
ejpam-86	205	28	n	n	NUM
ejpam-86	205	29	rows	row	NOUN
ejpam-86	205	30	have	have	VERB
ejpam-86	205	31	the	the	DET
ejpam-86	205	32	same	same	ADJ
ejpam-86	205	33	distribution	distribution	NOUN
ejpam-86	205	34	.	.	PUNCT
ejpam-86	206	1	but	but	CCONJ
ejpam-86	206	2	the	the	DET
ejpam-86	206	3	n	n	NOUN
ejpam-86	206	4	observations	observation	NOUN
ejpam-86	206	5	are	be	AUX
ejpam-86	206	6	assumed	assume	VERB
ejpam-86	206	7	to	to	PART
ejpam-86	206	8	be	be	AUX
ejpam-86	206	9	uncorrelated	uncorrelate	VERB
ejpam-86	206	10	rather	rather	ADV
ejpam-86	206	11	than	than	ADP
ejpam-86	206	12	independent	independent	ADJ
ejpam-86	206	13	from	from	ADP
ejpam-86	206	14	each	each	DET
ejpam-86	206	15	other	other	ADJ
ejpam-86	206	16	.	.	PUNCT
ejpam-86	207	1	the	the	DET
ejpam-86	207	2	identity	identity	NOUN
ejpam-86	207	3	matrix	matrix	NOUN
ejpam-86	207	4	in	in	ADP
ejpam-86	207	5	reflects	reflect	VERB
ejpam-86	207	6	the	the	DET
ejpam-86	207	7	lack	lack	NOUN
ejpam-86	207	8	of	of	ADP
ejpam-86	207	9	correlation	correlation	NOUN
ejpam-86	207	10	among	among	ADP
ejpam-86	207	11	the	the	DET
ejpam-86	207	12	observations	observation	NOUN
ejpam-86	207	13	.	.	PUNCT
ejpam-86	208	1	m.	m.	PROPN
ejpam-86	208	2	liu	liu	PROPN
ejpam-86	208	3	and	and	CCONJ
ejpam-86	208	4	h.	h.	PROPN
ejpam-86	208	5	bozdogan	bozdogan	PROPN
ejpam-86	208	6	/	/	SYM
ejpam-86	208	7	eur	eur	PROPN
ejpam-86	208	8	.	.	PUNCT
ejpam-86	209	1	j.	j.	PROPN
ejpam-86	209	2	pure	pure	PROPN
ejpam-86	209	3	appl	appl	PROPN
ejpam-86	209	4	.	.	PROPN
ejpam-86	209	5	math	math	PROPN
ejpam-86	209	6	,	,	PUNCT
ejpam-86	209	7	1	1	NUM
ejpam-86	209	8	(	(	PUNCT
ejpam-86	209	9	2008	2008	NUM
ejpam-86	209	10	)	)	PUNCT
ejpam-86	209	11	,	,	PUNCT
ejpam-86	209	12	(	(	PUNCT
ejpam-86	209	13	4	4	NUM
ejpam-86	209	14	-	-	SYM
ejpam-86	209	15	37	37	NUM
ejpam-86	209	16	)	)	PUNCT
ejpam-86	209	17	12	12	NUM
ejpam-86	209	18	under	under	ADP
ejpam-86	209	19	the	the	DET
ejpam-86	209	20	assumption	assumption	NOUN
ejpam-86	209	21	of	of	ADP
ejpam-86	209	22	type	type	NOUN
ejpam-86	209	23	ii	ii	PROPN
ejpam-86	209	24	mvper	mvper	PROPN
ejpam-86	209	25	model	model	NOUN
ejpam-86	209	26	,	,	PUNCT
ejpam-86	209	27	the	the	DET
ejpam-86	209	28	likelihood	likelihood	NOUN
ejpam-86	209	29	function	function	NOUN
ejpam-86	209	30	is	be	AUX
ejpam-86	209	31	:	:	PUNCT
ejpam-86	209	32	l(b	l(b	PROPN
ejpam-86	209	33	,	,	PUNCT
ejpam-86	209	34	σ	σ	PROPN
ejpam-86	209	35	,	,	PUNCT
ejpam-86	209	36	β	β	X
ejpam-86	209	37	)	)	PUNCT
ejpam-86	209	38	=	=	SYM
ejpam-86	210	1	k|φ|−p/2|σ|−n/2	k|φ|−p/2|σ|−n/2	PROPN
ejpam-86	210	2	exp(−1	exp(−1	X
ejpam-86	210	3	2	2	NUM
ejpam-86	210	4	(	(	PUNCT
ejpam-86	210	5	tr(σ−1(y	tr(σ−1(y	NOUN
ejpam-86	210	6	−xb)′−1(y	−xb)′−1(y	ADP
ejpam-86	210	7	−xb)))β	−xb)))β	ADV
ejpam-86	210	8	)	)	PUNCT
ejpam-86	210	9	,	,	PUNCT
ejpam-86	210	10	(	(	PUNCT
ejpam-86	210	11	4.3	4.3	NUM
ejpam-86	210	12	)	)	PUNCT
ejpam-86	210	13	and	and	CCONJ
ejpam-86	210	14	the	the	DET
ejpam-86	210	15	log	log	NOUN
ejpam-86	210	16	likelihood	likelihood	NOUN
ejpam-86	210	17	function	function	NOUN
ejpam-86	210	18	is	be	AUX
ejpam-86	210	19	:	:	PUNCT
ejpam-86	210	20	l(b	l(b	PROPN
ejpam-86	210	21	,	,	PUNCT
ejpam-86	210	22	σ	σ	PROPN
ejpam-86	210	23	,	,	PUNCT
ejpam-86	210	24	β	β	NOUN
ejpam-86	210	25	)	)	PUNCT
ejpam-86	210	26	≡	≡	PROPN
ejpam-86	210	27	logl(b	logl(b	PROPN
ejpam-86	210	28	,	,	PUNCT
ejpam-86	210	29	σ	σ	PROPN
ejpam-86	210	30	,	,	PUNCT
ejpam-86	210	31	β	β	NOUN
ejpam-86	210	32	)	)	PUNCT
ejpam-86	210	33	=	=	SYM
ejpam-86	210	34	log(npγ(np	log(npγ(np	NOUN
ejpam-86	210	35	2	2	NUM
ejpam-86	210	36	)	)	PUNCT
ejpam-86	210	37	)	)	PUNCT
ejpam-86	211	1	−	−	PROPN
ejpam-86	211	2	np	np	INTJ
ejpam-86	211	3	2	2	NUM
ejpam-86	211	4	log(π)−	log(π)−	NOUN
ejpam-86	211	5	log	log	NOUN
ejpam-86	211	6	γ(1	γ(1	PROPN
ejpam-86	212	1	+	+	CCONJ
ejpam-86	212	2	np	np	INTJ
ejpam-86	212	3	2β	2β	NUM
ejpam-86	212	4	)	)	PUNCT
ejpam-86	212	5	−	−	PROPN
ejpam-86	213	1	(	(	PUNCT
ejpam-86	213	2	1	1	NUM
ejpam-86	213	3	+	+	NUM
ejpam-86	213	4	np/2β	np/2β	NOUN
ejpam-86	213	5	)	)	PUNCT
ejpam-86	213	6	log	log	VERB
ejpam-86	213	7	2	2	NUM
ejpam-86	213	8	−p	−p	NOUN
ejpam-86	213	9	2	2	NUM
ejpam-86	213	10	log	log	NOUN
ejpam-86	213	11	|φ|−n	|φ|−n	NUM
ejpam-86	213	12	2	2	NUM
ejpam-86	213	13	log	log	NOUN
ejpam-86	213	14	|σ|	|σ|	PROPN
ejpam-86	213	15	−	−	PROPN
ejpam-86	213	16	1	1	NUM
ejpam-86	213	17	2(tr(σ−1(y	2(tr(σ−1(y	PROPN
ejpam-86	213	18	−xb)′−1(y	−xb)′−1(y	ADP
ejpam-86	213	19	−xb)))β	−xb)))β	ADV
ejpam-86	213	20	.	.	PUNCT
ejpam-86	214	1	(	(	PUNCT
ejpam-86	214	2	4.4	4.4	NUM
ejpam-86	214	3	)	)	PUNCT
ejpam-86	214	4	let	let	VERB
ejpam-86	214	5	θ=(b′	θ=(b′	PROPN
ejpam-86	214	6	,	,	PUNCT
ejpam-86	214	7	v	v	ADP
ejpam-86	214	8	ec′(σ	ec′(σ	PROPN
ejpam-86	214	9	)	)	PUNCT
ejpam-86	214	10	,	,	PUNCT
ejpam-86	214	11	β)′.	β)′.	ADP
ejpam-86	214	12	differentiating	differentiating	NOUN
ejpam-86	214	13	(	(	PUNCT
ejpam-86	214	14	4.4	4.4	NUM
ejpam-86	214	15	)	)	PUNCT
ejpam-86	214	16	with	with	ADP
ejpam-86	214	17	respect	respect	NOUN
ejpam-86	214	18	to	to	ADP
ejpam-86	214	19	b	b	NUM
ejpam-86	214	20	,	,	PUNCT
ejpam-86	214	21	v	v	NOUN
ejpam-86	214	22	ec(σ	ec(σ	NOUN
ejpam-86	214	23	)	)	PUNCT
ejpam-86	214	24	and	and	CCONJ
ejpam-86	214	25	β	β	X
ejpam-86	214	26	respectively	respectively	ADV
ejpam-86	214	27	,	,	PUNCT
ejpam-86	214	28	we	we	PRON
ejpam-86	214	29	have	have	VERB
ejpam-86	214	30	∂l(θ	∂l(θ	PROPN
ejpam-86	214	31	)	)	PUNCT
ejpam-86	215	1	∂b	∂b	PROPN
ejpam-86	215	2	=	=	SYM
ejpam-86	215	3	v	v	PROPN
ejpam-86	215	4	ec	ec	PROPN
ejpam-86	215	5	(	(	PUNCT
ejpam-86	215	6	(	(	PUNCT
ejpam-86	215	7	∂l(θ	∂l(θ	PROPN
ejpam-86	215	8	)	)	PUNCT
ejpam-86	215	9	∂b	∂b	PROPN
ejpam-86	215	10	)	)	PUNCT
ejpam-86	215	11	′−1e′−1e))β−1v	′−1e′−1e))β−1v	NOUN
ejpam-86	215	12	ec(σ−1e′−1x	ec(σ−1e′−1x	NOUN
ejpam-86	215	13	)	)	PUNCT
ejpam-86	215	14	,	,	PUNCT
ejpam-86	215	15	(	(	PUNCT
ejpam-86	215	16	4.5	4.5	X
ejpam-86	215	17	)	)	PUNCT
ejpam-86	215	18	∂l(θ	∂l(θ	PROPN
ejpam-86	215	19	)	)	PUNCT
ejpam-86	215	20	∂v	∂v	PROPN
ejpam-86	215	21	ec(σ	ec(σ	PRON
ejpam-86	215	22	)	)	PUNCT
ejpam-86	215	23	=	=	PUNCT
ejpam-86	215	24	v	v	ADP
ejpam-86	215	25	ec(∂l(θ	ec(∂l(θ	PROPN
ejpam-86	215	26	)	)	PUNCT
ejpam-86	215	27	∂σ	∂σ	PROPN
ejpam-86	215	28	)	)	PUNCT
ejpam-86	216	1	=	=	PUNCT
ejpam-86	216	2	−n	−n	PROPN
ejpam-86	216	3	2v	2v	PROPN
ejpam-86	216	4	ec(σ	ec(σ	CCONJ
ejpam-86	216	5	−1	−1	NOUN
ejpam-86	216	6	)	)	PUNCT
ejpam-86	217	1	+	+	CCONJ
ejpam-86	217	2	β	β	X
ejpam-86	217	3	2	2	NUM
ejpam-86	217	4	(	(	PUNCT
ejpam-86	217	5	tr(σ−1e′−1e))β−1v	tr(σ−1e′−1e))β−1v	PROPN
ejpam-86	217	6	ec(σ−1e′−1e	ec(σ−1e′−1e	PROPN
ejpam-86	217	7	)	)	PUNCT
ejpam-86	217	8	(	(	PUNCT
ejpam-86	217	9	4.6	4.6	NUM
ejpam-86	217	10	)	)	PUNCT
ejpam-86	217	11	and	and	CCONJ
ejpam-86	217	12	∂l(θ	∂l(θ	PROPN
ejpam-86	217	13	)	)	PUNCT
ejpam-86	217	14	∂β	∂β	PROPN
ejpam-86	217	15	=	=	PUNCT
ejpam-86	217	16	np	np	PROPN
ejpam-86	217	17	2β2ψ(1	2β2ψ(1	NUM
ejpam-86	217	18	+	+	CCONJ
ejpam-86	217	19	np	np	PRON
ejpam-86	217	20	2β	2β	NOUN
ejpam-86	217	21	)	)	PUNCT
ejpam-86	218	1	+	+	CCONJ
ejpam-86	218	2	np	np	PRON
ejpam-86	218	3	2β2	2β2	NUM
ejpam-86	218	4	log	log	NOUN
ejpam-86	218	5	2	2	NUM
ejpam-86	218	6	−1	−1	NOUN
ejpam-86	218	7	2(tr(σ−1e′−1e))β	2(tr(σ−1e′−1e))β	NUM
ejpam-86	218	8	log(tr(σ−1e′−1e	log(tr(σ−1e′−1e	NOUN
ejpam-86	218	9	)	)	PUNCT
ejpam-86	218	10	)	)	PUNCT
ejpam-86	218	11	.	.	PUNCT
ejpam-86	219	1	(	(	PUNCT
ejpam-86	219	2	4.7	4.7	NUM
ejpam-86	219	3	)	)	PUNCT
ejpam-86	219	4	to	to	PART
ejpam-86	219	5	derive	derive	VERB
ejpam-86	219	6	and	and	CCONJ
ejpam-86	219	7	construct	construct	VERB
ejpam-86	219	8	the	the	DET
ejpam-86	219	9	fim	fim	NOUN
ejpam-86	219	10	and	and	CCONJ
ejpam-86	219	11	ifim	ifim	NOUN
ejpam-86	219	12	of	of	ADP
ejpam-86	219	13	the	the	DET
ejpam-86	219	14	model	model	NOUN
ejpam-86	219	15	parameters	parameter	NOUN
ejpam-86	219	16	,	,	PUNCT
ejpam-86	219	17	we	we	PRON
ejpam-86	219	18	have	have	VERB
ejpam-86	219	19	∂l2(θ	∂l2(θ	PROPN
ejpam-86	219	20	)	)	PUNCT
ejpam-86	219	21	∂b′∂b	∂b′∂b	PUNCT
ejpam-86	220	1	=	=	SYM
ejpam-86	220	2	∂	∂	NUM
ejpam-86	220	3	∂b′	∂b′	NOUN
ejpam-86	220	4	(	(	PUNCT
ejpam-86	220	5	∂l(θ	∂l(θ	PROPN
ejpam-86	220	6	)	)	PUNCT
ejpam-86	220	7	∂b	∂b	PROPN
ejpam-86	220	8	)	)	PUNCT
ejpam-86	221	1	=	=	PUNCT
ejpam-86	221	2	(	(	PUNCT
ejpam-86	221	3	ipq	ipq	NOUN
ejpam-86	221	4	⊗	⊗	PROPN
ejpam-86	221	5	v	v	NUM
ejpam-86	221	6	ec′(ip))(iq	ec′(ip))(iq	PROPN
ejpam-86	221	7	⊗	⊗	PROPN
ejpam-86	221	8	∂l2(θ	∂l2(θ	PROPN
ejpam-86	221	9	)	)	PUNCT
ejpam-86	221	10	∂b′∂b′	∂b′∂b′	ADV
ejpam-86	221	11	⊗	⊗	NUM
ejpam-86	221	12	ip)(v	ip)(v	VERB
ejpam-86	221	13	ec(iq)⊗	ec(iq)⊗	NOUN
ejpam-86	221	14	ipq	ipq	NOUN
ejpam-86	221	15	)	)	PUNCT
ejpam-86	221	16	(	(	PUNCT
ejpam-86	221	17	4.8	4.8	NUM
ejpam-86	221	18	)	)	PUNCT
ejpam-86	221	19	where	where	SCONJ
ejpam-86	221	20	∂l2(θ	∂l2(θ	NOUN
ejpam-86	221	21	)	)	PUNCT
ejpam-86	221	22	∂b′∂b′	∂b′∂b′	ADV
ejpam-86	221	23	=	=	SYM
ejpam-86	221	24	β(tr(σ−1e′−1e))β−1	β(tr(σ−1e′−1e))β−1	PROPN
ejpam-86	221	25	∂(σ−1e′−1x	∂(σ−1e′−1x	NOUN
ejpam-86	221	26	)	)	PUNCT
ejpam-86	221	27	∂b′	∂b′	NOUN
ejpam-86	222	1	+	+	ADV
ejpam-86	222	2	(	(	PUNCT
ejpam-86	222	3	σ−1e′−1x)⊗	σ−1e′−1x)⊗	X
ejpam-86	222	4	∂(β(tr(σ−1e′−1e))β−1	∂(β(tr(σ−1e′−1e))β−1	NUM
ejpam-86	222	5	)	)	PUNCT
ejpam-86	222	6	∂b′	∂b′	NOUN
ejpam-86	222	7	,	,	PUNCT
ejpam-86	222	8	(	(	PUNCT
ejpam-86	222	9	4.9	4.9	NUM
ejpam-86	222	10	)	)	PUNCT
ejpam-86	222	11	∂(σ−1e′−1x	∂(σ−1e′−1x	NOUN
ejpam-86	222	12	)	)	PUNCT
ejpam-86	222	13	∂b′	∂b′	NOUN
ejpam-86	222	14	=	=	SYM
ejpam-86	222	15	−v	−v	NOUN
ejpam-86	222	16	ec(σ−1)v	ec(σ−1)v	PROPN
ejpam-86	222	17	ec′(x	ec′(x	PROPN
ejpam-86	222	18	′−1x	′−1x	PROPN
ejpam-86	222	19	)	)	PUNCT
ejpam-86	222	20	(	(	PUNCT
ejpam-86	222	21	4.10	4.10	NUM
ejpam-86	222	22	)	)	PUNCT
ejpam-86	222	23	and	and	CCONJ
ejpam-86	222	24	∂(β(tr(σ−1e′−1e))β−1	∂(β(tr(σ−1e′−1e))β−1	NUM
ejpam-86	222	25	)	)	PUNCT
ejpam-86	222	26	∂b′	∂b′	NOUN
ejpam-86	222	27	=	=	SYM
ejpam-86	222	28	−2β(β	−2β(β	NOUN
ejpam-86	222	29	−	−	PROPN
ejpam-86	222	30	1)(tr(σ−1e′−1e))β−2σ−1e′−1x	1)(tr(σ−1e′−1e))β−2σ−1e′−1x	NUM
ejpam-86	222	31	.	.	PUNCT
ejpam-86	223	1	(	(	PUNCT
ejpam-86	223	2	4.11	4.11	NUM
ejpam-86	223	3	)	)	PUNCT
ejpam-86	223	4	∂l2(θ	∂l2(θ	PROPN
ejpam-86	223	5	)	)	PUNCT
ejpam-86	223	6	∂v	∂v	PROPN
ejpam-86	223	7	ec′(σ)∂v	ec′(σ)∂v	NOUN
ejpam-86	223	8	ec(σ	ec(σ	NUM
ejpam-86	223	9	)	)	PUNCT
ejpam-86	223	10	=	=	SYM
ejpam-86	223	11	(	(	PUNCT
ejpam-86	223	12	ip2	ip2	PROPN
ejpam-86	223	13	⊗	⊗	PROPN
ejpam-86	223	14	v	v	ADP
ejpam-86	223	15	ec′(ip))(ip	ec′(ip))(ip	PROPN
ejpam-86	223	16	⊗	⊗	PROPN
ejpam-86	223	17	∂l2(θ	∂l2(θ	PROPN
ejpam-86	223	18	)	)	PUNCT
ejpam-86	223	19	∂σ∂σ	∂σ∂σ	PROPN
ejpam-86	223	20	⊗	⊗	PROPN
ejpam-86	223	21	ip)(v	ip)(v	PROPN
ejpam-86	223	22	ec(ip)⊗	ec(ip)⊗	PROPN
ejpam-86	223	23	ip2	ip2	PROPN
ejpam-86	223	24	)	)	PUNCT
ejpam-86	223	25	(	(	PUNCT
ejpam-86	223	26	4.12	4.12	NUM
ejpam-86	223	27	)	)	PUNCT
ejpam-86	223	28	where	where	SCONJ
ejpam-86	223	29	∂l2(θ	∂l2(θ	NOUN
ejpam-86	223	30	)	)	PUNCT
ejpam-86	223	31	∂σ∂σ	∂σ∂σ	NOUN
ejpam-86	223	32	=	=	SYM
ejpam-86	223	33	n	n	PROPN
ejpam-86	223	34	2v	2v	PROPN
ejpam-86	223	35	ec(σ	ec(σ	CCONJ
ejpam-86	223	36	−1)v	−1)v	X
ejpam-86	223	37	ec′−1	ec′−1	PROPN
ejpam-86	223	38	)	)	PUNCT
ejpam-86	223	39	−β	−β	NOUN
ejpam-86	223	40	2	2	NUM
ejpam-86	223	41	(	(	PUNCT
ejpam-86	223	42	tr(σ−1e′−1e))β−1(v	tr(σ−1e′−1e))β−1(v	NOUN
ejpam-86	223	43	ec(σ−1)v	ec(σ−1)v	PROPN
ejpam-86	223	44	ec′−1e′−1eς−1	ec′−1e′−1eς−1	PROPN
ejpam-86	223	45	)	)	PUNCT
ejpam-86	224	1	+	+	ADP
ejpam-86	224	2	v	v	PRON
ejpam-86	224	3	ec(σ−1e′−1eς−1)v	ec(σ−1e′−1eς−1)v	ADJ
ejpam-86	224	4	ec′−1	ec′−1	NOUN
ejpam-86	224	5	)	)	PUNCT
ejpam-86	224	6	)	)	PUNCT
ejpam-86	225	1	−β(β−1	−β(β−1	NUM
ejpam-86	225	2	)	)	PUNCT
ejpam-86	225	3	2	2	NUM
ejpam-86	225	4	(	(	PUNCT
ejpam-86	225	5	tr(σ−1e′−1e))β−2	tr(σ−1e′−1e))β−2	NOUN
ejpam-86	225	6	(	(	PUNCT
ejpam-86	225	7	(	(	PUNCT
ejpam-86	225	8	σ−1e′−1eς−1)⊗	σ−1e′−1eς−1)⊗	X
ejpam-86	225	9	(	(	PUNCT
ejpam-86	225	10	σ−1e′−1eς−1	σ−1e′−1eς−1	PROPN
ejpam-86	225	11	)	)	PUNCT
ejpam-86	225	12	)	)	PUNCT
ejpam-86	225	13	.	.	PUNCT
ejpam-86	226	1	(	(	PUNCT
ejpam-86	226	2	4.13	4.13	X
ejpam-86	226	3	)	)	PUNCT
ejpam-86	226	4	m.	m.	NOUN
ejpam-86	226	5	liu	liu	PROPN
ejpam-86	226	6	and	and	CCONJ
ejpam-86	226	7	h.	h.	PROPN
ejpam-86	226	8	bozdogan	bozdogan	PROPN
ejpam-86	226	9	/	/	SYM
ejpam-86	226	10	eur	eur	PROPN
ejpam-86	226	11	.	.	PUNCT
ejpam-86	227	1	j.	j.	PROPN
ejpam-86	227	2	pure	pure	PROPN
ejpam-86	227	3	appl	appl	PROPN
ejpam-86	227	4	.	.	PROPN
ejpam-86	227	5	math	math	PROPN
ejpam-86	227	6	,	,	PUNCT
ejpam-86	227	7	1	1	NUM
ejpam-86	227	8	(	(	PUNCT
ejpam-86	227	9	2008	2008	NUM
ejpam-86	227	10	)	)	PUNCT
ejpam-86	227	11	,	,	PUNCT
ejpam-86	227	12	(	(	PUNCT
ejpam-86	227	13	4	4	NUM
ejpam-86	227	14	-	-	SYM
ejpam-86	227	15	37	37	NUM
ejpam-86	227	16	)	)	PUNCT
ejpam-86	227	17	13	13	NUM
ejpam-86	227	18	∂l2(θ	∂l2(θ	NOUN
ejpam-86	227	19	)	)	PUNCT
ejpam-86	227	20	∂β2	∂β2	ADJ
ejpam-86	227	21	=	=	PUNCT
ejpam-86	227	22	−np	−np	X
ejpam-86	227	23	β3ψ(1	β3ψ(1	PUNCT
ejpam-86	227	24	+	+	CCONJ
ejpam-86	227	25	np	np	INTJ
ejpam-86	227	26	2β	2β	NUM
ejpam-86	227	27	)	)	PUNCT
ejpam-86	227	28	−	−	PUNCT
ejpam-86	227	29	n2p2	n2p2	NOUN
ejpam-86	227	30	4β4	4β4	NUM
ejpam-86	227	31	ψ	ψ	SYM
ejpam-86	227	32	′(1	′(1	PROPN
ejpam-86	227	33	+	+	CCONJ
ejpam-86	227	34	np	np	INTJ
ejpam-86	227	35	2β	2β	NUM
ejpam-86	227	36	)	)	PUNCT
ejpam-86	227	37	−	−	PROPN
ejpam-86	228	1	np	np	PRON
ejpam-86	228	2	β3	β3	INTJ
ejpam-86	228	3	log	log	VERB
ejpam-86	228	4	2	2	NUM
ejpam-86	228	5	−	−	NOUN
ejpam-86	228	6	1	1	NUM
ejpam-86	228	7	2(tr(σ−1e′−1e))β	2(tr(σ−1e′−1e))β	NUM
ejpam-86	228	8	log2(tr(σ−1e′−1e	log2(tr(σ−1e′−1e	NOUN
ejpam-86	228	9	)	)	PUNCT
ejpam-86	228	10	)	)	PUNCT
ejpam-86	228	11	.	.	PUNCT
ejpam-86	229	1	(	(	PUNCT
ejpam-86	229	2	4.14	4.14	NUM
ejpam-86	229	3	)	)	PUNCT
ejpam-86	229	4	∂l2(θ	∂l2(θ	PROPN
ejpam-86	229	5	)	)	PUNCT
ejpam-86	230	1	∂β∂b	∂β∂b	PROPN
ejpam-86	230	2	=	=	PUNCT
ejpam-86	230	3	(	(	PUNCT
ejpam-86	230	4	tr(σ−1e′−1e))β−1v	tr(σ−1e′−1e))β−1v	PROPN
ejpam-86	230	5	ec(σ−1e′−1x)(1	ec(σ−1e′−1x)(1	PROPN
ejpam-86	230	6	+	+	CCONJ
ejpam-86	230	7	β	β	X
ejpam-86	230	8	log(tr(σ−1e′−1e	log(tr(σ−1e′−1e	NOUN
ejpam-86	230	9	)	)	PUNCT
ejpam-86	230	10	)	)	PUNCT
ejpam-86	230	11	)	)	PUNCT
ejpam-86	230	12	.	.	PUNCT
ejpam-86	231	1	(	(	PUNCT
ejpam-86	231	2	4.15	4.15	NUM
ejpam-86	231	3	)	)	PUNCT
ejpam-86	231	4	∂l2(θ	∂l2(θ	PROPN
ejpam-86	231	5	)	)	PUNCT
ejpam-86	231	6	∂β∂v	∂β∂v	PROPN
ejpam-86	231	7	ec(σ	ec(σ	PUNCT
ejpam-86	231	8	)	)	PUNCT
ejpam-86	231	9	=	=	SYM
ejpam-86	232	1	1	1	NUM
ejpam-86	232	2	2(tr(σ−1e′−1e))β−1v	2(tr(σ−1e′−1e))β−1v	NUM
ejpam-86	232	3	ec(σ−1e′−1eς−1	ec(σ−1e′−1eς−1	NUM
ejpam-86	232	4	)	)	PUNCT
ejpam-86	232	5	(	(	PUNCT
ejpam-86	232	6	1	1	NUM
ejpam-86	232	7	+	+	CCONJ
ejpam-86	232	8	β	β	X
ejpam-86	232	9	log(tr(σ−1e′−1e	log(tr(σ−1e′−1e	NOUN
ejpam-86	232	10	)	)	PUNCT
ejpam-86	232	11	)	)	PUNCT
ejpam-86	232	12	)	)	PUNCT
ejpam-86	232	13	.	.	PUNCT
ejpam-86	233	1	(	(	PUNCT
ejpam-86	233	2	4.16	4.16	NUM
ejpam-86	233	3	)	)	PUNCT
ejpam-86	233	4	∂l2(θ	∂l2(θ	PROPN
ejpam-86	233	5	)	)	PUNCT
ejpam-86	233	6	∂v	∂v	PROPN
ejpam-86	233	7	ec′(σ)∂b	ec′(σ)∂b	PROPN
ejpam-86	233	8	=	=	PUNCT
ejpam-86	233	9	(	(	PUNCT
ejpam-86	233	10	ipq	ipq	PROPN
ejpam-86	233	11	⊗	⊗	PROPN
ejpam-86	233	12	v	v	NUM
ejpam-86	233	13	ec′(ip))(iq	ec′(ip))(iq	PROPN
ejpam-86	233	14	⊗	⊗	PROPN
ejpam-86	233	15	∂l2(θ	∂l2(θ	PROPN
ejpam-86	233	16	)	)	PUNCT
ejpam-86	233	17	∂σ∂b′	∂σ∂b′	PROPN
ejpam-86	233	18	⊗	⊗	PROPN
ejpam-86	233	19	ip)(v	ip)(v	PROPN
ejpam-86	233	20	ec(iq)⊗	ec(iq)⊗	PROPN
ejpam-86	233	21	ip2	ip2	PROPN
ejpam-86	233	22	)	)	PUNCT
ejpam-86	233	23	(	(	PUNCT
ejpam-86	233	24	4.17	4.17	NUM
ejpam-86	233	25	)	)	PUNCT
ejpam-86	233	26	where	where	SCONJ
ejpam-86	233	27	∂l2(θ	∂l2(θ	NOUN
ejpam-86	233	28	)	)	PUNCT
ejpam-86	233	29	∂σ∂b′	∂σ∂b′	NOUN
ejpam-86	234	1	=	=	PUNCT
ejpam-86	234	2	−β(tr(σ−1e′−1e))β−1v	−β(tr(σ−1e′−1e))β−1v	PROPN
ejpam-86	234	3	ec(σ−1)v	ec(σ−1)v	PROPN
ejpam-86	234	4	ec′−1e′−1x	ec′−1e′−1x	PROPN
ejpam-86	234	5	)	)	PUNCT
ejpam-86	234	6	−β(β	−β(β	PROPN
ejpam-86	235	1	−	−	PROPN
ejpam-86	235	2	1)(tr(σ−1e′−1e))β−2(σ−1e′−1x)⊗	1)(tr(σ−1e′−1e))β−2(σ−1e′−1x)⊗	NOUN
ejpam-86	235	3	(	(	PUNCT
ejpam-86	235	4	σ−1e′−1eς−1	σ−1e′−1eς−1	PROPN
ejpam-86	235	5	)	)	PUNCT
ejpam-86	235	6	.	.	PUNCT
ejpam-86	236	1	(	(	PUNCT
ejpam-86	236	2	4.18	4.18	NUM
ejpam-86	236	3	)	)	PUNCT
ejpam-86	236	4	further	further	ADJ
ejpam-86	236	5	details	detail	NOUN
ejpam-86	236	6	of	of	ADP
ejpam-86	236	7	the	the	DET
ejpam-86	236	8	above	above	ADJ
ejpam-86	236	9	derivations	derivation	NOUN
ejpam-86	236	10	are	be	AUX
ejpam-86	236	11	given	give	VERB
ejpam-86	236	12	in	in	ADP
ejpam-86	236	13	the	the	DET
ejpam-86	236	14	appendix	appendix	NOUN
ejpam-86	236	15	of	of	ADP
ejpam-86	236	16	the	the	DET
ejpam-86	236	17	paper	paper	NOUN
ejpam-86	236	18	.	.	PUNCT
ejpam-86	237	1	by	by	ADP
ejpam-86	237	2	the	the	DET
ejpam-86	237	3	probability	probability	NOUN
ejpam-86	237	4	characteristics	characteristic	NOUN
ejpam-86	237	5	of	of	ADP
ejpam-86	237	6	multivariate	multivariate	NOUN
ejpam-86	237	7	pe	pe	ADP
ejpam-86	237	8	distribution	distribution	NOUN
ejpam-86	237	9	,	,	PUNCT
ejpam-86	237	10	we	we	PRON
ejpam-86	237	11	have	have	VERB
ejpam-86	237	12	e[((v	e[((v	NUM
ejpam-86	237	13	ec(y	ec(y	NOUN
ejpam-86	238	1	′)−	′)−	PROPN
ejpam-86	238	2	v	v	ADP
ejpam-86	238	3	ec(b′x	ec(b′x	NOUN
ejpam-86	238	4	′))′v	′))′v	PROPN
ejpam-86	238	5	ar(v	ar(v	NUM
ejpam-86	238	6	ec(y	ec(y	VERB
ejpam-86	238	7	′−1(v	′−1(v	PROPN
ejpam-86	238	8	ec(y	ec(y	PUNCT
ejpam-86	239	1	′)−	′)−	PROPN
ejpam-86	239	2	v	v	ADP
ejpam-86	239	3	ec(b′x	ec(b′x	NOUN
ejpam-86	239	4	′2	′2	NOUN
ejpam-86	239	5	]	]	X
ejpam-86	239	6	=	=	SYM
ejpam-86	239	7	(	(	PUNCT
ejpam-86	239	8	np)2γ(np	np)2γ(np	ADV
ejpam-86	239	9	2β	2β	NOUN
ejpam-86	239	10	)	)	PUNCT
ejpam-86	239	11	γ(np+4	γ(np+4	PROPN
ejpam-86	239	12	2β	2β	NOUN
ejpam-86	239	13	)	)	PUNCT
ejpam-86	239	14	γ2(np+2	γ2(np+2	ADJ
ejpam-86	239	15	2β	2β	NOUN
ejpam-86	239	16	)	)	PUNCT
ejpam-86	239	17	.v	.v	ADV
ejpam-86	240	1	then	then	ADV
ejpam-86	240	2	,	,	PUNCT
ejpam-86	240	3	the	the	DET
ejpam-86	240	4	method	method	NOUN
ejpam-86	240	5	of	of	ADP
ejpam-86	240	6	moment	moment	NOUN
ejpam-86	240	7	estimate	estimate	NOUN
ejpam-86	240	8	of	of	ADP
ejpam-86	240	9	β	β	X
ejpam-86	240	10	,	,	PUNCT
ejpam-86	240	11	denoted	denote	VERB
ejpam-86	240	12	as	as	ADP
ejpam-86	240	13	β̂	β̂	ADP
ejpam-86	240	14	,	,	PUNCT
ejpam-86	240	15	can	can	AUX
ejpam-86	240	16	be	be	AUX
ejpam-86	240	17	obtained	obtain	VERB
ejpam-86	240	18	by	by	ADP
ejpam-86	240	19	solving	solve	VERB
ejpam-86	240	20	the	the	DET
ejpam-86	240	21	equation	equation	NOUN
ejpam-86	240	22	:	:	PUNCT
ejpam-86	240	23	(	(	PUNCT
ejpam-86	240	24	(	(	PUNCT
ejpam-86	240	25	v	v	NOUN
ejpam-86	240	26	ec(y	ec(y	NOUN
ejpam-86	240	27	′)−	′)−	PROPN
ejpam-86	240	28	v	v	ADP
ejpam-86	240	29	ec(b̂′x	ec(b̂′x	ADJ
ejpam-86	240	30	′))′−1(v	′))′−1(v	NOUN
ejpam-86	240	31	ec(y	ec(y	PUNCT
ejpam-86	241	1	′)−	′)−	PROPN
ejpam-86	241	2	v	v	ADP
ejpam-86	241	3	ec(b̂′x	ec(b̂′x	ADJ
ejpam-86	241	4	′2	′2	NOUN
ejpam-86	241	5	=	=	SYM
ejpam-86	241	6	(	(	PUNCT
ejpam-86	241	7	np)2γ(np	np)2γ(np	NOUN
ejpam-86	241	8	2β̂	2β̂	NUM
ejpam-86	241	9	)	)	PUNCT
ejpam-86	241	10	γ(np+4	γ(np+4	PROPN
ejpam-86	241	11	2β̂	2β̂	NUM
ejpam-86	241	12	)	)	PUNCT
ejpam-86	242	1	γ2(np+2	γ2(np+2	ADJ
ejpam-86	242	2	2β̂	2β̂	NUM
ejpam-86	242	3	)	)	PUNCT
ejpam-86	242	4	where	where	SCONJ
ejpam-86	242	5	s	s	VERB
ejpam-86	242	6	=	=	SYM
ejpam-86	242	7	(	(	PUNCT
ejpam-86	242	8	v	v	NUM
ejpam-86	242	9	ec(y	ec(y	NOUN
ejpam-86	242	10	′)−	′)−	PROPN
ejpam-86	242	11	v	v	ADP
ejpam-86	242	12	ec(b̂′x	ec(b̂′x	ADJ
ejpam-86	242	13	′))(v	′))(v	NOUN
ejpam-86	242	14	ec(y	ec(y	VERB
ejpam-86	243	1	′)−	′)−	PROPN
ejpam-86	243	2	v	v	ADP
ejpam-86	243	3	ec(b̂′x	ec(b̂′x	ADJ
ejpam-86	243	4	′))′	′))′	PROPN
ejpam-86	243	5	and	and	CCONJ
ejpam-86	243	6	b̂	b̂	NOUN
ejpam-86	243	7	=	=	SYM
ejpam-86	243	8	(	(	PUNCT
ejpam-86	243	9	x	x	X
ejpam-86	243	10	′x)−1x	′x)−1x	PROPN
ejpam-86	243	11	′y	′y	PROPN
ejpam-86	243	12	is	be	AUX
ejpam-86	243	13	the	the	DET
ejpam-86	243	14	method	method	NOUN
ejpam-86	243	15	of	of	ADP
ejpam-86	243	16	moment	moment	NOUN
ejpam-86	243	17	estimate	estimate	NOUN
ejpam-86	243	18	of	of	ADP
ejpam-86	243	19	b.	b.	PROPN
ejpam-86	243	20	by	by	ADP
ejpam-86	243	21	anderson	anderson	PROPN
ejpam-86	243	22	and	and	CCONJ
ejpam-86	243	23	fang	fang	X
ejpam-86	243	24	(	(	PUNCT
ejpam-86	243	25	[	[	X
ejpam-86	243	26	17	17	NUM
ejpam-86	243	27	,	,	PUNCT
ejpam-86	243	28	p.215	p.215	NOUN
ejpam-86	243	29	]	]	PUNCT
ejpam-86	243	30	)	)	PUNCT
ejpam-86	243	31	,	,	PUNCT
ejpam-86	243	32	the	the	DET
ejpam-86	243	33	unbiased	unbiased	ADJ
ejpam-86	243	34	estimator	estimator	NOUN
ejpam-86	243	35	of	of	ADP
ejpam-86	243	36	σ	σ	PROPN
ejpam-86	243	37	can	can	AUX
ejpam-86	243	38	be	be	AUX
ejpam-86	243	39	obtained	obtain	VERB
ejpam-86	243	40	as	as	ADP
ejpam-86	243	41	σ̂	σ̂	NOUN
ejpam-86	243	42	=	=	SYM
ejpam-86	243	43	npγ(np	npγ(np	X
ejpam-86	243	44	2β̂	2β̂	NUM
ejpam-86	243	45	)	)	PUNCT
ejpam-86	244	1	(	(	PUNCT
ejpam-86	244	2	n−	n−	NOUN
ejpam-86	244	3	q)21	q)21	NOUN
ejpam-86	244	4	/	/	SYM
ejpam-86	244	5	β̂γ(np+2	β̂γ(np+2	NOUN
ejpam-86	244	6	2β̂	2β̂	NUM
ejpam-86	244	7	)	)	PUNCT
ejpam-86	245	1	(	(	PUNCT
ejpam-86	245	2	y	y	PROPN
ejpam-86	245	3	−xb̂)′−1(y	−xb̂)′−1(y	PROPN
ejpam-86	245	4	−xb̂	−xb̂	NUM
ejpam-86	245	5	)	)	PUNCT
ejpam-86	245	6	.	.	PUNCT
ejpam-86	246	1	under	under	ADP
ejpam-86	246	2	the	the	DET
ejpam-86	246	3	assumption	assumption	NOUN
ejpam-86	246	4	of	of	ADP
ejpam-86	246	5	type	type	NOUN
ejpam-86	246	6	ii	ii	PROPN
ejpam-86	246	7	mvper	mvper	PROPN
ejpam-86	246	8	model	model	NOUN
ejpam-86	246	9	,	,	PUNCT
ejpam-86	246	10	indeed	indeed	ADV
ejpam-86	246	11	the	the	DET
ejpam-86	246	12	sample	sample	NOUN
ejpam-86	246	13	size	size	NOUN
ejpam-86	246	14	is	be	AUX
ejpam-86	246	15	only	only	ADV
ejpam-86	246	16	1	1	NUM
ejpam-86	246	17	.	.	PUNCT
ejpam-86	247	1	according	accord	VERB
ejpam-86	247	2	to	to	ADP
ejpam-86	247	3	gupta	gupta	PROPN
ejpam-86	247	4	and	and	CCONJ
ejpam-86	247	5	varga	varga	PROPN
ejpam-86	247	6	(	(	PUNCT
ejpam-86	247	7	[	[	X
ejpam-86	247	8	21	21	NUM
ejpam-86	247	9	,	,	PUNCT
ejpam-86	247	10	p.224	p.224	NOUN
ejpam-86	247	11	]	]	PUNCT
ejpam-86	247	12	)	)	PUNCT
ejpam-86	247	13	,	,	PUNCT
ejpam-86	247	14	the	the	DET
ejpam-86	247	15	mles	mle	NOUN
ejpam-86	247	16	of	of	ADP
ejpam-86	247	17	the	the	DET
ejpam-86	247	18	model	model	NOUN
ejpam-86	247	19	parameters	parameter	NOUN
ejpam-86	247	20	do	do	AUX
ejpam-86	247	21	not	not	PART
ejpam-86	247	22	exist	exist	VERB
ejpam-86	247	23	without	without	ADP
ejpam-86	247	24	imposing	impose	VERB
ejpam-86	247	25	some	some	DET
ejpam-86	247	26	restrictions	restriction	NOUN
ejpam-86	247	27	on	on	ADP
ejpam-86	247	28	σ	σ	PROPN
ejpam-86	247	29	and	and	CCONJ
ejpam-86	247	30	φ	φ	NUM
ejpam-86	247	31	even	even	ADV
ejpam-86	247	32	if	if	SCONJ
ejpam-86	247	33	φ	φ	PROPN
ejpam-86	247	34	is	be	AUX
ejpam-86	247	35	known	know	VERB
ejpam-86	247	36	.	.	PUNCT
ejpam-86	248	1	for	for	ADP
ejpam-86	248	2	n	n	PRON
ejpam-86	248	3	≥	≥	NOUN
ejpam-86	248	4	p	p	X
ejpam-86	248	5	,	,	PUNCT
ejpam-86	248	6	since	since	SCONJ
ejpam-86	248	7	h(t	h(t	NUM
ejpam-86	248	8	)	)	PUNCT
ejpam-86	248	9	=	=	SYM
ejpam-86	249	1	k	k	PROPN
ejpam-86	249	2	exp(−tβ/2	exp(−tβ/2	PROPN
ejpam-86	249	3	)	)	PUNCT
ejpam-86	249	4	is	be	AUX
ejpam-86	249	5	monotone	monotone	ADJ
ejpam-86	249	6	decreasing	decrease	VERB
ejpam-86	249	7	on	on	ADP
ejpam-86	249	8	(	(	PUNCT
ejpam-86	249	9	0,+∞	0,+∞	NUM
ejpam-86	249	10	)	)	PUNCT
ejpam-86	249	11	,	,	PUNCT
ejpam-86	249	12	there	there	PRON
ejpam-86	249	13	is	be	VERB
ejpam-86	249	14	b̂mle	b̂mle	NOUN
ejpam-86	249	15	=	=	SYM
ejpam-86	249	16	b̂	b̂	NOUN
ejpam-86	249	17	m.	m.	NOUN
ejpam-86	249	18	liu	liu	PROPN
ejpam-86	249	19	and	and	CCONJ
ejpam-86	249	20	h.	h.	PROPN
ejpam-86	249	21	bozdogan	bozdogan	PROPN
ejpam-86	249	22	/	/	SYM
ejpam-86	249	23	eur	eur	PROPN
ejpam-86	249	24	.	.	PUNCT
ejpam-86	250	1	j.	j.	PROPN
ejpam-86	250	2	pure	pure	PROPN
ejpam-86	250	3	appl	appl	PROPN
ejpam-86	250	4	.	.	PROPN
ejpam-86	250	5	math	math	PROPN
ejpam-86	250	6	,	,	PUNCT
ejpam-86	250	7	1	1	NUM
ejpam-86	250	8	(	(	PUNCT
ejpam-86	250	9	2008	2008	NUM
ejpam-86	250	10	)	)	PUNCT
ejpam-86	250	11	,	,	PUNCT
ejpam-86	250	12	(	(	PUNCT
ejpam-86	250	13	4	4	NUM
ejpam-86	250	14	-	-	SYM
ejpam-86	250	15	37	37	NUM
ejpam-86	250	16	)	)	PUNCT
ejpam-86	250	17	14	14	NUM
ejpam-86	250	18	by	by	ADP
ejpam-86	250	19	theorem	theorem	ADJ
ejpam-86	250	20	7.1.1	7.1.1	NUM
ejpam-86	250	21	of	of	ADP
ejpam-86	250	22	(	(	PUNCT
ejpam-86	250	23	[	[	X
ejpam-86	250	24	21	21	NUM
ejpam-86	250	25	,	,	PUNCT
ejpam-86	250	26	p.226	p.226	NOUN
ejpam-86	250	27	]	]	PUNCT
ejpam-86	250	28	)	)	PUNCT
ejpam-86	250	29	and	and	CCONJ
ejpam-86	250	30	the	the	DET
ejpam-86	250	31	mle	mle	PROPN
ejpam-86	250	32	of	of	ADP
ejpam-86	250	33	σ	σ	PROPN
ejpam-86	250	34	is	be	AUX
ejpam-86	250	35	σ̂mle	σ̂mle	NOUN
ejpam-86	250	36	=	=	SYM
ejpam-86	250	37	p	p	X
ejpam-86	250	38	λmax	λmax	NOUN
ejpam-86	250	39	(	(	PUNCT
ejpam-86	250	40	y	y	PROPN
ejpam-86	250	41	−xb̂)′−1(y	−xb̂)′−1(y	PROPN
ejpam-86	250	42	−xb̂	−xb̂	NOUN
ejpam-86	250	43	)	)	PUNCT
ejpam-86	250	44	where	where	SCONJ
ejpam-86	250	45	λmax	λmax	NOUN
ejpam-86	250	46	is	be	AUX
ejpam-86	250	47	the	the	DET
ejpam-86	250	48	maximum	maximum	ADJ
ejpam-86	250	49	point	point	NOUN
ejpam-86	250	50	of	of	ADP
ejpam-86	250	51	the	the	DET
ejpam-86	250	52	function	function	NOUN
ejpam-86	250	53	f(λ	f(λ	NOUN
ejpam-86	250	54	)	)	PUNCT
ejpam-86	250	55	=	=	SYM
ejpam-86	250	56	λnp/2h(λ	λnp/2h(λ	NOUN
ejpam-86	250	57	)	)	PUNCT
ejpam-86	250	58	by	by	ADP
ejpam-86	250	59	theorem	theorem	NOUN
ejpam-86	250	60	7.1.3	7.1.3	PROPN
ejpam-86	250	61	of	of	ADP
ejpam-86	250	62	(	(	PUNCT
ejpam-86	250	63	[	[	X
ejpam-86	250	64	21	21	NUM
ejpam-86	250	65	,	,	PUNCT
ejpam-86	250	66	p.235	p.235	NOUN
ejpam-86	250	67	]	]	PUNCT
ejpam-86	250	68	)	)	PUNCT
ejpam-86	250	69	.	.	PUNCT
ejpam-86	251	1	it	it	PRON
ejpam-86	251	2	is	be	AUX
ejpam-86	251	3	easy	easy	ADJ
ejpam-86	251	4	now	now	ADV
ejpam-86	251	5	to	to	PART
ejpam-86	251	6	show	show	VERB
ejpam-86	251	7	that	that	DET
ejpam-86	251	8	λmax	λmax	NOUN
ejpam-86	251	9	=	=	SYM
ejpam-86	251	10	(	(	PUNCT
ejpam-86	251	11	np	np	INTJ
ejpam-86	251	12	β̂mle	β̂mle	PUNCT
ejpam-86	251	13	)	)	PUNCT
ejpam-86	251	14	1	1	X
ejpam-86	251	15	/	/	SYM
ejpam-86	251	16	β̂mle	β̂mle	NOUN
ejpam-86	251	17	by	by	ADP
ejpam-86	251	18	the	the	DET
ejpam-86	251	19	lemma	lemma	PROPN
ejpam-86	251	20	2	2	NUM
ejpam-86	251	21	of	of	ADP
ejpam-86	251	22	(	(	PUNCT
ejpam-86	251	23	[	[	X
ejpam-86	251	24	17	17	NUM
ejpam-86	251	25	,	,	PUNCT
ejpam-86	251	26	p.204	p.204	NUM
ejpam-86	251	27	]	]	PUNCT
ejpam-86	251	28	)	)	PUNCT
ejpam-86	251	29	.	.	PUNCT
ejpam-86	252	1	however	however	ADV
ejpam-86	252	2	,	,	PUNCT
ejpam-86	252	3	if	if	SCONJ
ejpam-86	252	4	we	we	PRON
ejpam-86	252	5	substitute	substitute	VERB
ejpam-86	252	6	σ̂mle	σ̂mle	NOUN
ejpam-86	252	7	and	and	CCONJ
ejpam-86	252	8	b̂mle	b̂mle	NOUN
ejpam-86	252	9	into	into	ADP
ejpam-86	252	10	the	the	DET
ejpam-86	252	11	log	log	NOUN
ejpam-86	252	12	likelihood	likelihood	NOUN
ejpam-86	252	13	function	function	NOUN
ejpam-86	252	14	(	(	PUNCT
ejpam-86	252	15	4.4	4.4	NUM
ejpam-86	252	16	)	)	PUNCT
ejpam-86	252	17	,	,	PUNCT
ejpam-86	252	18	the	the	DET
ejpam-86	252	19	log	log	NOUN
ejpam-86	252	20	likelihood	likelihood	NOUN
ejpam-86	252	21	function	function	NOUN
ejpam-86	252	22	of	of	ADP
ejpam-86	252	23	β	β	NOUN
ejpam-86	252	24	becomes	become	VERB
ejpam-86	252	25	:	:	PUNCT
ejpam-86	252	26	f(β	f(β	NUM
ejpam-86	252	27	)	)	PUNCT
ejpam-86	253	1	=	=	PUNCT
ejpam-86	253	2	log(npγ(np	log(npγ(np	NOUN
ejpam-86	253	3	2	2	NUM
ejpam-86	253	4	)	)	PUNCT
ejpam-86	253	5	)	)	PUNCT
ejpam-86	254	1	−	−	PROPN
ejpam-86	254	2	np	np	INTJ
ejpam-86	254	3	2	2	NUM
ejpam-86	254	4	log(π)−	log(π)−	NOUN
ejpam-86	254	5	log	log	NOUN
ejpam-86	254	6	γ(1	γ(1	PROPN
ejpam-86	255	1	+	+	CCONJ
ejpam-86	255	2	np	np	INTJ
ejpam-86	255	3	2β	2β	NUM
ejpam-86	255	4	)	)	PUNCT
ejpam-86	255	5	−	−	NOUN
ejpam-86	256	1	log	log	NOUN
ejpam-86	256	2	2−np	2−np	NUM
ejpam-86	256	3	2	2	NUM
ejpam-86	256	4	log	log	NOUN
ejpam-86	256	5	p−	p−	NOUN
ejpam-86	256	6	p	p	NOUN
ejpam-86	256	7	2	2	NUM
ejpam-86	256	8	log	log	NOUN
ejpam-86	256	9	|φ|	|φ|	PROPN
ejpam-86	256	10	−n	−n	ADV
ejpam-86	256	11	2	2	NUM
ejpam-86	256	12	log	log	NOUN
ejpam-86	256	13	|(y	|(y	NUM
ejpam-86	257	1	−xb̂)′−1(y	−xb̂)′−1(y	PROPN
ejpam-86	257	2	−xb̂)|+	−xb̂)|+	VERB
ejpam-86	257	3	np	np	ADP
ejpam-86	257	4	2β	2β	NUM
ejpam-86	257	5	log(np	log(np	NUM
ejpam-86	257	6	2β	2β	NOUN
ejpam-86	257	7	)	)	PUNCT
ejpam-86	258	1	−	−	PROPN
ejpam-86	259	1	np	np	INTJ
ejpam-86	259	2	2β	2β	NOUN
ejpam-86	259	3	which	which	PRON
ejpam-86	259	4	actually	actually	ADV
ejpam-86	259	5	has	have	VERB
ejpam-86	259	6	no	no	DET
ejpam-86	259	7	maximum	maximum	NOUN
ejpam-86	259	8	.	.	PUNCT
ejpam-86	260	1	we	we	PRON
ejpam-86	260	2	get	get	VERB
ejpam-86	260	3	around	around	ADP
ejpam-86	260	4	this	this	DET
ejpam-86	260	5	problem	problem	NOUN
ejpam-86	260	6	by	by	ADP
ejpam-86	260	7	providing	provide	VERB
ejpam-86	260	8	two	two	NUM
ejpam-86	260	9	methods	method	NOUN
ejpam-86	260	10	to	to	PART
ejpam-86	260	11	compute	compute	VERB
ejpam-86	260	12	the	the	DET
ejpam-86	260	13	mles	mle	NOUN
ejpam-86	260	14	of	of	ADP
ejpam-86	260	15	type	type	NOUN
ejpam-86	260	16	ii	ii	PROPN
ejpam-86	260	17	mvper	mvper	PROPN
ejpam-86	260	18	model	model	NOUN
ejpam-86	260	19	parameters	parameter	NOUN
ejpam-86	260	20	.	.	PUNCT
ejpam-86	261	1	one	one	NUM
ejpam-86	261	2	method	method	NOUN
ejpam-86	261	3	is	be	AUX
ejpam-86	261	4	to	to	PART
ejpam-86	261	5	maximize	maximize	VERB
ejpam-86	261	6	the	the	DET
ejpam-86	261	7	log	log	NOUN
ejpam-86	261	8	likelihood	likelihood	NOUN
ejpam-86	261	9	function	function	NOUN
ejpam-86	261	10	directly	directly	ADV
ejpam-86	261	11	with	with	ADP
ejpam-86	261	12	an	an	DET
ejpam-86	261	13	algorithm	algorithm	NOUN
ejpam-86	261	14	such	such	ADJ
ejpam-86	261	15	as	as	ADP
ejpam-86	261	16	the	the	DET
ejpam-86	261	17	ga	ga	NOUN
ejpam-86	261	18	given	give	VERB
ejpam-86	261	19	in	in	ADP
ejpam-86	261	20	bozdogan	bozdogan	NOUN
ejpam-86	261	21	and	and	CCONJ
ejpam-86	261	22	liu	liu	PROPN
ejpam-86	261	23	(	(	PUNCT
ejpam-86	261	24	[	[	X
ejpam-86	261	25	23	23	NUM
ejpam-86	261	26	]	]	PUNCT
ejpam-86	261	27	)	)	PUNCT
ejpam-86	261	28	.	.	PUNCT
ejpam-86	262	1	the	the	DET
ejpam-86	262	2	other	other	ADJ
ejpam-86	262	3	method	method	NOUN
ejpam-86	262	4	is	be	AUX
ejpam-86	262	5	to	to	PART
ejpam-86	262	6	employ	employ	VERB
ejpam-86	262	7	a	a	DET
ejpam-86	262	8	two	two	NUM
ejpam-86	262	9	step	step	NOUN
ejpam-86	262	10	procedure	procedure	NOUN
ejpam-86	262	11	given	give	VERB
ejpam-86	262	12	as	as	SCONJ
ejpam-86	262	13	follows	follow	VERB
ejpam-86	262	14	:	:	PUNCT
ejpam-86	262	15	step	step	NOUN
ejpam-86	262	16	1	1	NUM
ejpam-86	262	17	:	:	PUNCT
ejpam-86	262	18	use	use	VERB
ejpam-86	262	19	a	a	DET
ejpam-86	262	20	set	set	NOUN
ejpam-86	262	21	of	of	ADP
ejpam-86	262	22	m	m	PRON
ejpam-86	262	23	pairs	pair	NOUN
ejpam-86	262	24	of	of	ADP
ejpam-86	262	25	observations	observation	NOUN
ejpam-86	262	26	randomly	randomly	ADV
ejpam-86	262	27	selected	select	VERB
ejpam-86	262	28	from	from	ADP
ejpam-86	262	29	the	the	DET
ejpam-86	262	30	original	original	ADJ
ejpam-86	262	31	data	datum	NOUN
ejpam-86	262	32	to	to	PART
ejpam-86	262	33	compute	compute	VERB
ejpam-86	262	34	b̂mle	b̂mle	NOUN
ejpam-86	262	35	and	and	CCONJ
ejpam-86	262	36	σ̂mle	σ̂mle	NOUN
ejpam-86	262	37	by	by	ADP
ejpam-86	262	38	considering	consider	VERB
ejpam-86	262	39	the	the	DET
ejpam-86	262	40	shape	shape	NOUN
ejpam-86	262	41	parameter	parameter	NOUN
ejpam-86	262	42	β	β	PROPN
ejpam-86	262	43	fixed	fix	VERB
ejpam-86	262	44	.	.	PUNCT
ejpam-86	263	1	step	step	NOUN
ejpam-86	263	2	2	2	NUM
ejpam-86	263	3	:	:	PUNCT
ejpam-86	263	4	substitute	substitute	NOUN
ejpam-86	263	5	b̂mle	b̂mle	NOUN
ejpam-86	263	6	and	and	CCONJ
ejpam-86	263	7	σ̂mle	σ̂mle	NOUN
ejpam-86	263	8	obtained	obtain	VERB
ejpam-86	263	9	in	in	ADP
ejpam-86	263	10	step	step	NOUN
ejpam-86	263	11	1	1	NUM
ejpam-86	263	12	and	and	CCONJ
ejpam-86	263	13	the	the	DET
ejpam-86	263	14	original	original	ADJ
ejpam-86	263	15	sample	sample	NOUN
ejpam-86	263	16	data	datum	NOUN
ejpam-86	263	17	into	into	ADP
ejpam-86	263	18	the	the	DET
ejpam-86	263	19	log	log	NOUN
ejpam-86	263	20	likelihood	likelihood	NOUN
ejpam-86	263	21	function	function	NOUN
ejpam-86	263	22	(	(	PUNCT
ejpam-86	263	23	4.4	4.4	NUM
ejpam-86	263	24	)	)	PUNCT
ejpam-86	263	25	,	,	PUNCT
ejpam-86	263	26	then	then	ADV
ejpam-86	263	27	β̂mle	β̂mle	VERB
ejpam-86	263	28	is	be	AUX
ejpam-86	263	29	the	the	DET
ejpam-86	263	30	value	value	NOUN
ejpam-86	263	31	of	of	ADP
ejpam-86	263	32	β	β	PRON
ejpam-86	263	33	which	which	PRON
ejpam-86	263	34	maximizes	maximize	VERB
ejpam-86	263	35	the	the	DET
ejpam-86	263	36	log	log	NOUN
ejpam-86	263	37	likelihood	likelihood	NOUN
ejpam-86	263	38	function	function	NOUN
ejpam-86	263	39	of	of	ADP
ejpam-86	263	40	β	β	PROPN
ejpam-86	263	41	.	.	PROPN
ejpam-86	264	1	5	5	X
ejpam-86	264	2	.	.	X
ejpam-86	264	3	information	information	NOUN
ejpam-86	264	4	criteria	criterion	NOUN
ejpam-86	264	5	for	for	ADP
ejpam-86	264	6	multivariate	multivariate	NOUN
ejpam-86	264	7	pe	pe	PROPN
ejpam-86	264	8	regression	regression	NOUN
ejpam-86	264	9	models	model	NOUN
ejpam-86	264	10	in	in	ADP
ejpam-86	264	11	recent	recent	ADJ
ejpam-86	264	12	years	year	NOUN
ejpam-86	264	13	,	,	PUNCT
ejpam-86	264	14	information	information	NOUN
ejpam-86	264	15	-	-	PUNCT
ejpam-86	264	16	based	base	VERB
ejpam-86	264	17	criteria	criterion	NOUN
ejpam-86	264	18	such	such	ADJ
ejpam-86	264	19	as	as	ADP
ejpam-86	264	20	akaike	akaike	ADJ
ejpam-86	264	21	’s	’s	ADV
ejpam-86	264	22	aic	aic	PROPN
ejpam-86	264	23	(	(	PUNCT
ejpam-86	265	1	[	[	X
ejpam-86	265	2	2–5	2–5	NOUN
ejpam-86	265	3	]	]	X
ejpam-86	265	4	)	)	PUNCT
ejpam-86	265	5	,	,	PUNCT
ejpam-86	265	6	which	which	PRON
ejpam-86	265	7	compromises	compromise	VERB
ejpam-86	265	8	between	between	ADP
ejpam-86	265	9	the	the	DET
ejpam-86	265	10	goodness	goodness	NOUN
ejpam-86	265	11	-	-	PUNCT
ejpam-86	265	12	of	of	ADP
ejpam-86	265	13	-	-	PUNCT
ejpam-86	265	14	fit	fit	NOUN
ejpam-86	265	15	and	and	CCONJ
ejpam-86	265	16	the	the	DET
ejpam-86	265	17	model	model	NOUN
ejpam-86	265	18	complexity	complexity	NOUN
ejpam-86	265	19	,	,	PUNCT
ejpam-86	265	20	have	have	AUX
ejpam-86	265	21	been	be	AUX
ejpam-86	265	22	widely	widely	ADV
ejpam-86	265	23	used	use	VERB
ejpam-86	265	24	in	in	ADP
ejpam-86	265	25	statistical	statistical	ADJ
ejpam-86	265	26	modeling	modeling	NOUN
ejpam-86	265	27	and	and	CCONJ
ejpam-86	265	28	model	model	NOUN
ejpam-86	265	29	selection	selection	NOUN
ejpam-86	265	30	.	.	PUNCT
ejpam-86	266	1	however	however	ADV
ejpam-86	266	2	,	,	PUNCT
ejpam-86	266	3	the	the	DET
ejpam-86	266	4	penalty	penalty	NOUN
ejpam-86	266	5	term	term	NOUN
ejpam-86	266	6	used	use	VERB
ejpam-86	266	7	in	in	ADP
ejpam-86	266	8	aic	aic	PROPN
ejpam-86	266	9	-	-	PUNCT
ejpam-86	266	10	type	type	NOUN
ejpam-86	266	11	criteria	criterion	NOUN
ejpam-86	266	12	,	,	PUNCT
ejpam-86	266	13	that	that	ADV
ejpam-86	266	14	is	is	ADV
ejpam-86	266	15	,	,	PUNCT
ejpam-86	266	16	the	the	DET
ejpam-86	266	17	number	number	NOUN
ejpam-86	266	18	of	of	ADP
ejpam-86	266	19	free	free	ADJ
ejpam-86	266	20	parameters	parameter	NOUN
ejpam-86	266	21	,	,	PUNCT
ejpam-86	266	22	is	be	AUX
ejpam-86	266	23	insufficient	insufficient	ADJ
ejpam-86	266	24	to	to	PART
ejpam-86	266	25	measure	measure	VERB
ejpam-86	266	26	the	the	DET
ejpam-86	266	27	model	model	NOUN
ejpam-86	266	28	complexity	complexity	NOUN
ejpam-86	266	29	as	as	SCONJ
ejpam-86	266	30	noted	note	VERB
ejpam-86	266	31	by	by	ADP
ejpam-86	266	32	many	many	ADJ
ejpam-86	266	33	authors	author	NOUN
ejpam-86	266	34	(	(	PUNCT
ejpam-86	266	35	see	see	VERB
ejpam-86	266	36	,	,	PUNCT
ejpam-86	266	37	e.g.	e.g.	ADV
ejpam-86	266	38	[	[	X
ejpam-86	266	39	29	29	NUM
ejpam-86	266	40	]	]	SYM
ejpam-86	266	41	)	)	PUNCT
ejpam-86	266	42	.	.	PUNCT
ejpam-86	267	1	icomp	icomp	PROPN
ejpam-86	267	2	criteria	criterion	NOUN
ejpam-86	267	3	(	(	PUNCT
ejpam-86	267	4	[	[	X
ejpam-86	267	5	7	7	NUM
ejpam-86	267	6	,	,	PUNCT
ejpam-86	267	7	9	9	NUM
ejpam-86	267	8	,	,	PUNCT
ejpam-86	267	9	10	10	NUM
ejpam-86	267	10	,	,	PUNCT
ejpam-86	267	11	12–14	12–14	NUM
ejpam-86	267	12	]	]	PUNCT
ejpam-86	267	13	)	)	PUNCT
ejpam-86	267	14	improve	improve	VERB
ejpam-86	267	15	aic	aic	PROPN
ejpam-86	267	16	-	-	PUNCT
ejpam-86	267	17	type	type	NOUN
ejpam-86	267	18	criteria	criterion	NOUN
ejpam-86	267	19	by	by	ADP
ejpam-86	267	20	using	use	VERB
ejpam-86	267	21	an	an	DET
ejpam-86	267	22	informationtheoretic	informationtheoretic	ADJ
ejpam-86	267	23	measure	measure	NOUN
ejpam-86	267	24	of	of	ADP
ejpam-86	267	25	“	"	PUNCT
ejpam-86	267	26	overall	overall	ADJ
ejpam-86	267	27	”	"	PUNCT
ejpam-86	267	28	model	model	NOUN
ejpam-86	267	29	complexity	complexity	NOUN
ejpam-86	267	30	based	base	VERB
ejpam-86	267	31	on	on	ADP
ejpam-86	267	32	the	the	DET
ejpam-86	267	33	generalized	generalized	ADJ
ejpam-86	267	34	covariance	covariance	NOUN
ejpam-86	267	35	complexity	complexity	NOUN
ejpam-86	267	36	index	index	NOUN
ejpam-86	267	37	of	of	ADP
ejpam-86	267	38	van	van	PROPN
ejpam-86	267	39	emden	emden	PROPN
ejpam-86	267	40	(	(	PUNCT
ejpam-86	267	41	[	[	X
ejpam-86	267	42	16	16	NUM
ejpam-86	267	43	]	]	PUNCT
ejpam-86	267	44	)	)	PUNCT
ejpam-86	267	45	.	.	PUNCT
ejpam-86	268	1	icomp	icomp	PROPN
ejpam-86	268	2	criteria	criterion	NOUN
ejpam-86	268	3	can	can	AUX
ejpam-86	268	4	be	be	AUX
ejpam-86	268	5	defined	define	VERB
ejpam-86	268	6	in	in	ADP
ejpam-86	268	7	several	several	ADJ
ejpam-86	268	8	ways	way	NOUN
ejpam-86	268	9	.	.	PUNCT
ejpam-86	269	1	the	the	DET
ejpam-86	269	2	most	most	ADV
ejpam-86	269	3	general	general	ADJ
ejpam-86	269	4	form	form	NOUN
ejpam-86	269	5	of	of	ADP
ejpam-86	269	6	icomp	icomp	NOUN
ejpam-86	269	7	,	,	PUNCT
ejpam-86	269	8	referred	refer	VERB
ejpam-86	269	9	to	to	ADP
ejpam-86	269	10	as	as	ADP
ejpam-86	269	11	icomp(ifim	icomp(ifim	NOUN
ejpam-86	269	12	)	)	PUNCT
ejpam-86	269	13	,	,	PUNCT
ejpam-86	269	14	exploits	exploit	VERB
ejpam-86	269	15	the	the	DET
ejpam-86	269	16	well	well	ADV
ejpam-86	269	17	-	-	PUNCT
ejpam-86	269	18	known	know	VERB
ejpam-86	269	19	asymptotic	asymptotic	ADJ
ejpam-86	269	20	optimality	optimality	NOUN
ejpam-86	269	21	properties	property	NOUN
ejpam-86	269	22	of	of	ADP
ejpam-86	269	23	the	the	DET
ejpam-86	269	24	mle	mle	NOUN
ejpam-86	269	25	’s	’s	PART
ejpam-86	269	26	,	,	PUNCT
ejpam-86	269	27	and	and	CCONJ
ejpam-86	269	28	uses	use	VERB
ejpam-86	269	29	the	the	DET
ejpam-86	269	30	ifim	ifim	NOUN
ejpam-86	269	31	to	to	PART
ejpam-86	269	32	measure	measure	VERB
ejpam-86	269	33	the	the	DET
ejpam-86	269	34	complexity	complexity	NOUN
ejpam-86	269	35	of	of	ADP
ejpam-86	269	36	a	a	DET
ejpam-86	269	37	model	model	NOUN
ejpam-86	269	38	.	.	PUNCT
ejpam-86	270	1	for	for	ADP
ejpam-86	270	2	both	both	DET
ejpam-86	270	3	types	type	NOUN
ejpam-86	270	4	of	of	ADP
ejpam-86	270	5	criteria	criterion	NOUN
ejpam-86	270	6	,	,	PUNCT
ejpam-86	270	7	the	the	DET
ejpam-86	270	8	model	model	NOUN
ejpam-86	270	9	with	with	ADP
ejpam-86	270	10	the	the	DET
ejpam-86	270	11	smallest	small	ADJ
ejpam-86	270	12	score	score	NOUN
ejpam-86	270	13	is	be	AUX
ejpam-86	270	14	chosen	choose	VERB
ejpam-86	270	15	to	to	PART
ejpam-86	270	16	be	be	AUX
ejpam-86	270	17	the	the	DET
ejpam-86	270	18	best	good	ADJ
ejpam-86	270	19	model	model	NOUN
ejpam-86	270	20	.	.	PUNCT
ejpam-86	271	1	m.	m.	PROPN
ejpam-86	271	2	liu	liu	PROPN
ejpam-86	271	3	and	and	CCONJ
ejpam-86	271	4	h.	h.	PROPN
ejpam-86	271	5	bozdogan	bozdogan	PROPN
ejpam-86	271	6	/	/	SYM
ejpam-86	271	7	eur	eur	PROPN
ejpam-86	271	8	.	.	PUNCT
ejpam-86	272	1	j.	j.	PROPN
ejpam-86	272	2	pure	pure	PROPN
ejpam-86	272	3	appl	appl	PROPN
ejpam-86	272	4	.	.	PROPN
ejpam-86	272	5	math	math	PROPN
ejpam-86	272	6	,	,	PUNCT
ejpam-86	272	7	1	1	NUM
ejpam-86	272	8	(	(	PUNCT
ejpam-86	272	9	2008	2008	NUM
ejpam-86	272	10	)	)	PUNCT
ejpam-86	272	11	,	,	PUNCT
ejpam-86	272	12	(	(	PUNCT
ejpam-86	272	13	4	4	NUM
ejpam-86	272	14	-	-	SYM
ejpam-86	272	15	37	37	NUM
ejpam-86	272	16	)	)	PUNCT
ejpam-86	272	17	15	15	NUM
ejpam-86	272	18	for	for	ADP
ejpam-86	272	19	a	a	DET
ejpam-86	272	20	general	general	ADJ
ejpam-86	272	21	multivariate	multivariate	NOUN
ejpam-86	272	22	linear	linear	NOUN
ejpam-86	272	23	or	or	CCONJ
ejpam-86	272	24	nonlinear	nonlinear	ADJ
ejpam-86	272	25	model	model	NOUN
ejpam-86	272	26	,	,	PUNCT
ejpam-86	272	27	aic	aic	PROPN
ejpam-86	272	28	is	be	AUX
ejpam-86	272	29	defined	define	VERB
ejpam-86	272	30	as	as	ADP
ejpam-86	272	31	:	:	PUNCT
ejpam-86	272	32	aic	aic	PROPN
ejpam-86	272	33	=	=	PUNCT
ejpam-86	272	34	−2	−2	PROPN
ejpam-86	272	35	logl	logl	NOUN
ejpam-86	272	36	(	(	PUNCT
ejpam-86	272	37	θ̂	θ̂	NUM
ejpam-86	272	38	)	)	PUNCT
ejpam-86	273	1	+	+	CCONJ
ejpam-86	273	2	2k	2k	NUM
ejpam-86	273	3	(	(	PUNCT
ejpam-86	273	4	5.1	5.1	NUM
ejpam-86	273	5	)	)	PUNCT
ejpam-86	273	6	where	where	SCONJ
ejpam-86	273	7	k	k	PROPN
ejpam-86	273	8	is	be	AUX
ejpam-86	273	9	the	the	DET
ejpam-86	273	10	number	number	NOUN
ejpam-86	273	11	of	of	ADP
ejpam-86	273	12	free	free	ADJ
ejpam-86	273	13	parameters	parameter	NOUN
ejpam-86	273	14	estimated	estimate	VERB
ejpam-86	273	15	within	within	ADP
ejpam-86	273	16	the	the	DET
ejpam-86	273	17	model	model	NOUN
ejpam-86	273	18	.	.	PUNCT
ejpam-86	274	1	so	so	ADV
ejpam-86	274	2	for	for	ADP
ejpam-86	274	3	type	type	NOUN
ejpam-86	274	4	i	i	PROPN
ejpam-86	274	5	mvper	mvper	PROPN
ejpam-86	274	6	model	model	NOUN
ejpam-86	274	7	,	,	PUNCT
ejpam-86	274	8	we	we	PRON
ejpam-86	274	9	have	have	VERB
ejpam-86	274	10	aicmodel	aicmodel	NOUN
ejpam-86	274	11	i	i	PRON
ejpam-86	274	12	=	=	SYM
ejpam-86	274	13	−2n	−2n	PROPN
ejpam-86	274	14	log(pγ(p	log(pγ(p	PROPN
ejpam-86	274	15	2)γ(1	2)γ(1	NOUN
ejpam-86	274	16	+	+	CCONJ
ejpam-86	274	17	p	p	X
ejpam-86	274	18	2β̂	2β̂	NUM
ejpam-86	274	19	)	)	PUNCT
ejpam-86	274	20	)	)	PUNCT
ejpam-86	275	1	+	+	CCONJ
ejpam-86	275	2	np	np	PRON
ejpam-86	275	3	log(π	log(π	NOUN
ejpam-86	275	4	)	)	PUNCT
ejpam-86	275	5	+	+	NUM
ejpam-86	275	6	2n(1	2n(1	NUM
ejpam-86	275	7	+	+	NUM
ejpam-86	275	8	p/2β̂	p/2β̂	PROPN
ejpam-86	275	9	)	)	PUNCT
ejpam-86	275	10	log	log	VERB
ejpam-86	275	11	2	2	NUM
ejpam-86	276	1	+	+	NOUN
ejpam-86	276	2	n	n	PRON
ejpam-86	276	3	log	log	VERB
ejpam-86	276	4	|σ̂|+	|σ̂|+	PROPN
ejpam-86	276	5	∑n	∑n	PROPN
ejpam-86	276	6	i=1((y(i)−b̂	i=1((y(i)−b̂	VERB
ejpam-86	276	7	′	′	ADJ
ejpam-86	276	8	x(i	x(i	PROPN
ejpam-86	276	9	)	)	PUNCT
ejpam-86	276	10	)	)	PUNCT
ejpam-86	276	11	′σ̂−1(y(i)−b̂	′σ̂−1(y(i)−b̂	NOUN
ejpam-86	277	1	′	′	NUM
ejpam-86	277	2	x(i)))β̂	x(i)))β̂	NOUN
ejpam-86	277	3	+2pq	+2pq	PUNCT
ejpam-86	277	4	+	+	CCONJ
ejpam-86	277	5	p(p+	p(p+	PROPN
ejpam-86	277	6	1	1	NUM
ejpam-86	277	7	)	)	PUNCT
ejpam-86	277	8	+	+	CCONJ
ejpam-86	277	9	2	2	NUM
ejpam-86	277	10	(	(	PUNCT
ejpam-86	277	11	5.2	5.2	NUM
ejpam-86	277	12	)	)	PUNCT
ejpam-86	277	13	and	and	CCONJ
ejpam-86	277	14	for	for	ADP
ejpam-86	277	15	type	type	NOUN
ejpam-86	277	16	ii	ii	PROPN
ejpam-86	277	17	mvper	mvper	PROPN
ejpam-86	277	18	model	model	NOUN
ejpam-86	277	19	,	,	PUNCT
ejpam-86	277	20	we	we	PRON
ejpam-86	277	21	have	have	VERB
ejpam-86	277	22	aicmodel	aicmodel	NOUN
ejpam-86	277	23	ii	ii	PROPN
ejpam-86	278	1	=	=	SYM
ejpam-86	278	2	−2	−2	NOUN
ejpam-86	278	3	log(npγ(np	log(npγ(np	NOUN
ejpam-86	278	4	2	2	NUM
ejpam-86	278	5	)	)	PUNCT
ejpam-86	278	6	)	)	PUNCT
ejpam-86	279	1	+	+	CCONJ
ejpam-86	279	2	np	np	PRON
ejpam-86	279	3	log(π	log(π	NOUN
ejpam-86	279	4	)	)	PUNCT
ejpam-86	279	5	+	+	CCONJ
ejpam-86	279	6	2	2	NUM
ejpam-86	279	7	log	log	NOUN
ejpam-86	279	8	γ(1	γ(1	PROPN
ejpam-86	280	1	+	+	CCONJ
ejpam-86	280	2	np	np	INTJ
ejpam-86	280	3	2β̂	2β̂	NUM
ejpam-86	280	4	)	)	PUNCT
ejpam-86	281	1	+	+	CCONJ
ejpam-86	281	2	(	(	PUNCT
ejpam-86	281	3	2	2	NUM
ejpam-86	281	4	+	+	CCONJ
ejpam-86	281	5	np	np	PROPN
ejpam-86	281	6	/	/	SYM
ejpam-86	281	7	β̂	β̂	NUM
ejpam-86	281	8	)	)	PUNCT
ejpam-86	281	9	log	log	NOUN
ejpam-86	281	10	2	2	NUM
ejpam-86	282	1	+	+	NOUN
ejpam-86	282	2	p	p	NOUN
ejpam-86	282	3	log	log	NOUN
ejpam-86	282	4	|φ|+	|φ|+	PROPN
ejpam-86	282	5	n	n	ADV
ejpam-86	282	6	log	log	VERB
ejpam-86	282	7	|σ̂|+	|σ̂|+	PROPN
ejpam-86	282	8	(	(	PUNCT
ejpam-86	282	9	tr(σ̂−1(y	tr(σ̂−1(y	SCONJ
ejpam-86	282	10	−xb̂)′−1(y	−xb̂)′−1(y	PROPN
ejpam-86	282	11	−xb̂)))β̂	−xb̂)))β̂	VERB
ejpam-86	282	12	+2pq	+2pq	PUNCT
ejpam-86	283	1	+	+	CCONJ
ejpam-86	283	2	p(p+	p(p+	NOUN
ejpam-86	283	3	1	1	NUM
ejpam-86	283	4	)	)	PUNCT
ejpam-86	283	5	+	+	CCONJ
ejpam-86	283	6	2	2	X
ejpam-86	283	7	.	.	X
ejpam-86	283	8	(	(	PUNCT
ejpam-86	283	9	5.3	5.3	NUM
ejpam-86	283	10	)	)	PUNCT
ejpam-86	283	11	the	the	DET
ejpam-86	283	12	definition	definition	NOUN
ejpam-86	283	13	of	of	ADP
ejpam-86	283	14	icomp(ifim	icomp(ifim	NOUN
ejpam-86	283	15	)	)	PUNCT
ejpam-86	283	16	uses	use	VERB
ejpam-86	283	17	the	the	DET
ejpam-86	283	18	concept	concept	NOUN
ejpam-86	283	19	of	of	ADP
ejpam-86	283	20	maximal	maximal	ADJ
ejpam-86	283	21	covariance	covariance	NOUN
ejpam-86	283	22	complexity	complexity	NOUN
ejpam-86	283	23	which	which	PRON
ejpam-86	283	24	is	be	AUX
ejpam-86	283	25	defined	define	VERB
ejpam-86	283	26	as	as	ADP
ejpam-86	283	27	:	:	PUNCT
ejpam-86	283	28	definition	definition	NOUN
ejpam-86	283	29	5.1	5.1	NUM
ejpam-86	283	30	a	a	DET
ejpam-86	283	31	maximal	maximal	ADJ
ejpam-86	283	32	information	information	NOUN
ejpam-86	283	33	theoretic	theoretic	NOUN
ejpam-86	283	34	measure	measure	NOUN
ejpam-86	283	35	of	of	ADP
ejpam-86	283	36	complexity	complexity	NOUN
ejpam-86	283	37	of	of	ADP
ejpam-86	283	38	a	a	DET
ejpam-86	283	39	covariance	covariance	NOUN
ejpam-86	283	40	matrix	matrix	NOUN
ejpam-86	283	41	σ	σ	NOUN
ejpam-86	283	42	of	of	ADP
ejpam-86	283	43	a	a	DET
ejpam-86	283	44	multivariate	multivariate	NOUN
ejpam-86	283	45	normal	normal	ADJ
ejpam-86	283	46	distribution	distribution	NOUN
ejpam-86	283	47	is	be	AUX
ejpam-86	283	48	c1	c1	NOUN
ejpam-86	283	49	(	(	PUNCT
ejpam-86	283	50	σ	σ	PROPN
ejpam-86	283	51	)	)	PUNCT
ejpam-86	283	52	≡	≡	PROPN
ejpam-86	283	53	maxt	maxt	PROPN
ejpam-86	283	54	c0(σ	c0(σ	PROPN
ejpam-86	283	55	)	)	PUNCT
ejpam-86	284	1	=	=	PUNCT
ejpam-86	284	2	p	p	DET
ejpam-86	284	3	2	2	NUM
ejpam-86	284	4	log	log	NOUN
ejpam-86	284	5	(	(	PUNCT
ejpam-86	284	6	tr	tr	VERB
ejpam-86	284	7	(	(	PUNCT
ejpam-86	284	8	σ	σ	NOUN
ejpam-86	284	9	)	)	PUNCT
ejpam-86	284	10	p	p	NOUN
ejpam-86	284	11	)	)	PUNCT
ejpam-86	284	12	−	−	PROPN
ejpam-86	284	13	1	1	NUM
ejpam-86	284	14	2	2	NUM
ejpam-86	284	15	log	log	NOUN
ejpam-86	284	16	|	|	ADV
ejpam-86	284	17	σ|	σ|	NOUN
ejpam-86	284	18	,	,	PUNCT
ejpam-86	284	19	(	(	PUNCT
ejpam-86	284	20	5.4	5.4	NUM
ejpam-86	284	21	)	)	PUNCT
ejpam-86	284	22	where	where	SCONJ
ejpam-86	284	23	the	the	DET
ejpam-86	284	24	maximum	maximum	NOUN
ejpam-86	284	25	is	be	AUX
ejpam-86	284	26	taken	take	VERB
ejpam-86	284	27	over	over	ADP
ejpam-86	284	28	the	the	DET
ejpam-86	284	29	orthonormal	orthonormal	ADJ
ejpam-86	284	30	transformation	transformation	NOUN
ejpam-86	284	31	t	t	NOUN
ejpam-86	284	32	of	of	ADP
ejpam-86	284	33	the	the	DET
ejpam-86	284	34	overall	overall	ADJ
ejpam-86	284	35	coordinate	coordinate	NOUN
ejpam-86	284	36	system	system	NOUN
ejpam-86	284	37	x1	x1	PROPN
ejpam-86	284	38	,	,	PUNCT
ejpam-86	284	39	x2	x2	PROPN
ejpam-86	284	40	,	,	PUNCT
ejpam-86	284	41	·	·	PUNCT
ejpam-86	284	42	·	·	PUNCT
ejpam-86	284	43	·	·	PUNCT
ejpam-86	284	44	,	,	PUNCT
ejpam-86	284	45	xp	xp	INTJ
ejpam-86	284	46	.	.	PUNCT
ejpam-86	285	1	for	for	ADP
ejpam-86	285	2	more	more	ADJ
ejpam-86	285	3	details	detail	NOUN
ejpam-86	285	4	on	on	ADP
ejpam-86	285	5	(	(	PUNCT
ejpam-86	285	6	5.4	5.4	NUM
ejpam-86	285	7	)	)	PUNCT
ejpam-86	285	8	,	,	PUNCT
ejpam-86	285	9	we	we	PRON
ejpam-86	285	10	refer	refer	VERB
ejpam-86	285	11	the	the	DET
ejpam-86	285	12	readers	reader	NOUN
ejpam-86	285	13	to	to	ADP
ejpam-86	285	14	(	(	PUNCT
ejpam-86	285	15	[	[	X
ejpam-86	285	16	9,13,14	9,13,14	NUM
ejpam-86	285	17	]	]	PUNCT
ejpam-86	285	18	)	)	PUNCT
ejpam-86	285	19	.	.	PUNCT
ejpam-86	286	1	for	for	ADP
ejpam-86	286	2	a	a	DET
ejpam-86	286	3	multivariate	multivariate	NOUN
ejpam-86	286	4	normal	normal	ADJ
ejpam-86	286	5	linear	linear	NOUN
ejpam-86	286	6	or	or	CCONJ
ejpam-86	286	7	nonlinear	nonlinear	ADJ
ejpam-86	286	8	structural	structural	ADJ
ejpam-86	286	9	model	model	NOUN
ejpam-86	286	10	,	,	PUNCT
ejpam-86	286	11	we	we	PRON
ejpam-86	286	12	define	define	VERB
ejpam-86	286	13	the	the	DET
ejpam-86	286	14	general	general	ADJ
ejpam-86	286	15	form	form	NOUN
ejpam-86	286	16	of	of	ADP
ejpam-86	286	17	icomp(ifim	icomp(ifim	NOUN
ejpam-86	286	18	)	)	PUNCT
ejpam-86	286	19	as	as	ADP
ejpam-86	286	20	icomp	icomp	NOUN
ejpam-86	286	21	(	(	PUNCT
ejpam-86	286	22	ifim	ifim	NOUN
ejpam-86	286	23	)	)	PUNCT
ejpam-86	286	24	=	=	SYM
ejpam-86	286	25	−2	−2	NOUN
ejpam-86	286	26	logl(θ̂	logl(θ̂	VERB
ejpam-86	286	27	)	)	PUNCT
ejpam-86	287	1	+	+	CCONJ
ejpam-86	287	2	2c1(f̂−1(θ̂	2c1(f̂−1(θ̂	NUM
ejpam-86	287	3	)	)	PUNCT
ejpam-86	287	4	)	)	PUNCT
ejpam-86	287	5	,	,	PUNCT
ejpam-86	287	6	(	(	PUNCT
ejpam-86	287	7	5.5	5.5	NUM
ejpam-86	287	8	)	)	PUNCT
ejpam-86	287	9	where	where	SCONJ
ejpam-86	287	10	c1	c1	PROPN
ejpam-86	287	11	denotes	denote	VERB
ejpam-86	287	12	the	the	DET
ejpam-86	287	13	maximal	maximal	ADJ
ejpam-86	287	14	information	information	NOUN
ejpam-86	287	15	-	-	PUNCT
ejpam-86	287	16	theoretic	theoretic	NOUN
ejpam-86	287	17	complexity	complexity	NOUN
ejpam-86	287	18	of	of	ADP
ejpam-86	287	19	f̂−1	f̂−1	PROPN
ejpam-86	287	20	,	,	PUNCT
ejpam-86	287	21	the	the	DET
ejpam-86	287	22	estimated	estimate	VERB
ejpam-86	287	23	ifim	ifim	NOUN
ejpam-86	287	24	given	give	VERB
ejpam-86	287	25	in	in	ADP
ejpam-86	287	26	(	(	PUNCT
ejpam-86	287	27	5.4	5.4	NUM
ejpam-86	287	28	)	)	PUNCT
ejpam-86	287	29	,	,	PUNCT
ejpam-86	287	30	and	and	CCONJ
ejpam-86	287	31	θ̂	θ̂	NUM
ejpam-86	287	32	is	be	AUX
ejpam-86	287	33	the	the	DET
ejpam-86	287	34	mle	mle	PROPN
ejpam-86	287	35	vector	vector	NOUN
ejpam-86	287	36	.	.	PUNCT
ejpam-86	288	1	so	so	ADV
ejpam-86	288	2	,	,	PUNCT
ejpam-86	288	3	for	for	ADP
ejpam-86	288	4	type	type	NOUN
ejpam-86	288	5	i	i	PROPN
ejpam-86	288	6	mvper	mvper	PROPN
ejpam-86	288	7	model	model	NOUN
ejpam-86	288	8	,	,	PUNCT
ejpam-86	288	9	we	we	PRON
ejpam-86	288	10	have	have	VERB
ejpam-86	288	11	icom(ifim)model	icom(ifim)model	NOUN
ejpam-86	288	12	i	i	NOUN
ejpam-86	288	13	=	=	SYM
ejpam-86	288	14	−2n	−2n	PROPN
ejpam-86	288	15	log(pγ(p	log(pγ(p	PROPN
ejpam-86	288	16	2)γ(1	2)γ(1	NOUN
ejpam-86	289	1	+	+	CCONJ
ejpam-86	289	2	p	p	X
ejpam-86	289	3	2β̂	2β̂	NUM
ejpam-86	289	4	)	)	PUNCT
ejpam-86	289	5	)	)	PUNCT
ejpam-86	290	1	+	+	CCONJ
ejpam-86	290	2	np	np	PRON
ejpam-86	290	3	log(π	log(π	NOUN
ejpam-86	290	4	)	)	PUNCT
ejpam-86	290	5	+	+	NUM
ejpam-86	290	6	2n(1	2n(1	NUM
ejpam-86	290	7	+	+	NUM
ejpam-86	290	8	p/2β̂	p/2β̂	PROPN
ejpam-86	290	9	)	)	PUNCT
ejpam-86	290	10	log	log	VERB
ejpam-86	290	11	+	+	NOUN
ejpam-86	290	12	n	n	PRON
ejpam-86	290	13	log	log	VERB
ejpam-86	290	14	2|σ̂|	2|σ̂|	NUM
ejpam-86	290	15	+	+	CCONJ
ejpam-86	291	1	∑n	∑n	PROPN
ejpam-86	291	2	i=1((y(i)−b̂	i=1((y(i)−b̂	NOUN
ejpam-86	291	3	′	′	ADJ
ejpam-86	291	4	x(i	x(i	PROPN
ejpam-86	291	5	)	)	PUNCT
ejpam-86	291	6	)	)	PUNCT
ejpam-86	291	7	′σ̂−1(y(i)−b̂	′σ̂−1(y(i)−b̂	NOUN
ejpam-86	292	1	′	′	NUM
ejpam-86	292	2	x(i)))β̂	x(i)))β̂	NOUN
ejpam-86	292	3	+	+	CCONJ
ejpam-86	292	4	2c1(f̂−1	2c1(f̂−1	NUM
ejpam-86	292	5	(	(	PUNCT
ejpam-86	292	6	θ̂)model	θ̂)model	NOUN
ejpam-86	292	7	i	i	PRON
ejpam-86	292	8	)	)	PUNCT
ejpam-86	292	9	(	(	PUNCT
ejpam-86	292	10	5.6	5.6	NUM
ejpam-86	292	11	)	)	PUNCT
ejpam-86	292	12	and	and	CCONJ
ejpam-86	292	13	for	for	ADP
ejpam-86	292	14	type	type	NOUN
ejpam-86	292	15	ii	ii	PROPN
ejpam-86	292	16	mvper	mvper	PROPN
ejpam-86	292	17	model	model	NOUN
ejpam-86	292	18	,	,	PUNCT
ejpam-86	292	19	we	we	PRON
ejpam-86	292	20	have	have	VERB
ejpam-86	292	21	icom(ifim)model	icom(ifim)model	PROPN
ejpam-86	292	22	ii	ii	NOUN
ejpam-86	292	23	=	=	SYM
ejpam-86	292	24	−2	−2	NOUN
ejpam-86	292	25	log(npγ(np	log(npγ(np	NOUN
ejpam-86	292	26	2	2	NUM
ejpam-86	292	27	)	)	PUNCT
ejpam-86	292	28	)	)	PUNCT
ejpam-86	293	1	+	+	CCONJ
ejpam-86	293	2	np	np	PRON
ejpam-86	293	3	log(π	log(π	NOUN
ejpam-86	293	4	)	)	PUNCT
ejpam-86	293	5	+	+	CCONJ
ejpam-86	293	6	2	2	NUM
ejpam-86	293	7	log	log	NOUN
ejpam-86	293	8	γ(1	γ(1	PROPN
ejpam-86	294	1	+	+	CCONJ
ejpam-86	294	2	np	np	INTJ
ejpam-86	294	3	2β̂	2β̂	NUM
ejpam-86	294	4	)	)	PUNCT
ejpam-86	295	1	+	+	CCONJ
ejpam-86	295	2	(	(	PUNCT
ejpam-86	295	3	2	2	NUM
ejpam-86	295	4	+	+	CCONJ
ejpam-86	295	5	np	np	PROPN
ejpam-86	295	6	/	/	SYM
ejpam-86	295	7	β̂	β̂	NUM
ejpam-86	295	8	)	)	PUNCT
ejpam-86	295	9	log	log	NOUN
ejpam-86	295	10	2	2	NUM
ejpam-86	296	1	+	+	NOUN
ejpam-86	296	2	p	p	NOUN
ejpam-86	296	3	log	log	NOUN
ejpam-86	296	4	|φ|+	|φ|+	PROPN
ejpam-86	296	5	n	n	ADV
ejpam-86	296	6	log	log	VERB
ejpam-86	296	7	|σ̂|+	|σ̂|+	PROPN
ejpam-86	296	8	(	(	PUNCT
ejpam-86	296	9	tr(σ̂−1(y	tr(σ̂−1(y	ADP
ejpam-86	296	10	−xb̂)′−1(y	−xb̂)′−1(y	PROPN
ejpam-86	296	11	−xb̂)))β̂	−xb̂)))β̂	X
ejpam-86	296	12	+2c1(f̂−1	+2c1(f̂−1	NOUN
ejpam-86	296	13	(	(	PUNCT
ejpam-86	296	14	θ̂)model	θ̂)model	NOUN
ejpam-86	296	15	ii	ii	NOUN
ejpam-86	296	16	)	)	PUNCT
ejpam-86	296	17	.	.	PUNCT
ejpam-86	297	1	(	(	PUNCT
ejpam-86	297	2	5.7	5.7	NUM
ejpam-86	297	3	)	)	PUNCT
ejpam-86	297	4	m.	m.	NOUN
ejpam-86	297	5	liu	liu	PROPN
ejpam-86	297	6	and	and	CCONJ
ejpam-86	297	7	h.	h.	PROPN
ejpam-86	297	8	bozdogan	bozdogan	PROPN
ejpam-86	297	9	/	/	SYM
ejpam-86	297	10	eur	eur	PROPN
ejpam-86	297	11	.	.	PUNCT
ejpam-86	298	1	j.	j.	PROPN
ejpam-86	298	2	pure	pure	PROPN
ejpam-86	298	3	appl	appl	PROPN
ejpam-86	298	4	.	.	PROPN
ejpam-86	298	5	math	math	PROPN
ejpam-86	298	6	,	,	PUNCT
ejpam-86	298	7	1	1	NUM
ejpam-86	298	8	(	(	PUNCT
ejpam-86	298	9	2008	2008	NUM
ejpam-86	298	10	)	)	PUNCT
ejpam-86	298	11	,	,	PUNCT
ejpam-86	298	12	(	(	PUNCT
ejpam-86	298	13	4	4	NUM
ejpam-86	298	14	-	-	SYM
ejpam-86	298	15	37	37	NUM
ejpam-86	298	16	)	)	PUNCT
ejpam-86	298	17	16	16	NUM
ejpam-86	299	1	the	the	DET
ejpam-86	299	2	estimated	estimate	VERB
ejpam-86	299	3	observed	observe	VERB
ejpam-86	299	4	inverse	inverse	NOUN
ejpam-86	299	5	fisher	fisher	PROPN
ejpam-86	299	6	information	information	NOUN
ejpam-86	299	7	matrices	matrix	NOUN
ejpam-86	299	8	for	for	ADP
ejpam-86	299	9	type	type	NOUN
ejpam-86	299	10	i	i	PRON
ejpam-86	299	11	and	and	CCONJ
ejpam-86	299	12	type	type	NOUN
ejpam-86	299	13	ii	ii	PROPN
ejpam-86	299	14	mvper	mvper	NOUN
ejpam-86	299	15	models	model	NOUN
ejpam-86	299	16	,	,	PUNCT
ejpam-86	299	17	f̂−1(θ̂)model	f̂−1(θ̂)model	INTJ
ejpam-86	299	18	i	i	PROPN
ejpam-86	299	19	and	and	CCONJ
ejpam-86	299	20	f̂−1(θ̂)model	f̂−1(θ̂)model	PROPN
ejpam-86	299	21	ii	ii	PROPN
ejpam-86	299	22	,	,	PUNCT
ejpam-86	299	23	can	can	AUX
ejpam-86	299	24	be	be	AUX
ejpam-86	299	25	computed	compute	VERB
ejpam-86	299	26	from	from	ADP
ejpam-86	299	27	the	the	DET
ejpam-86	299	28	results	result	NOUN
ejpam-86	299	29	of	of	ADP
ejpam-86	299	30	sections	section	NOUN
ejpam-86	299	31	3	3	NUM
ejpam-86	299	32	and	and	CCONJ
ejpam-86	299	33	4	4	NUM
ejpam-86	299	34	accordingly	accordingly	ADV
ejpam-86	299	35	by	by	ADP
ejpam-86	299	36	f̂−1(θ̂	f̂−1(θ̂	NOUN
ejpam-86	299	37	)	)	PUNCT
ejpam-86	299	38	=	=	SYM
ejpam-86	300	1	−	−	PROPN
ejpam-86	300	2			PROPN
ejpam-86	300	3			PROPN
ejpam-86	300	4	∂l2(θ	∂l2(θ	PROPN
ejpam-86	300	5	)	)	PUNCT
ejpam-86	300	6	∂b′∂b	∂b′∂b	PROPN
ejpam-86	301	1	∂l2(θ	∂l2(θ	PROPN
ejpam-86	301	2	)	)	PUNCT
ejpam-86	301	3	∂v	∂v	PROPN
ejpam-86	301	4	ec′(σ)∂b	ec′(σ)∂b	VERB
ejpam-86	301	5	∂l2(θ	∂l2(θ	PROPN
ejpam-86	301	6	)	)	PUNCT
ejpam-86	301	7	∂β∂b	∂β∂b	PROPN
ejpam-86	301	8	∂l2(θ	∂l2(θ	PROPN
ejpam-86	301	9	)	)	PUNCT
ejpam-86	301	10	∂b′∂v	∂b′∂v	PROPN
ejpam-86	301	11	ec(σ	ec(σ	NOUN
ejpam-86	301	12	)	)	PUNCT
ejpam-86	301	13	∂l2(θ	∂l2(θ	PROPN
ejpam-86	301	14	)	)	PUNCT
ejpam-86	301	15	∂v	∂v	PROPN
ejpam-86	301	16	ec′(σ)∂v	ec′(σ)∂v	NOUN
ejpam-86	301	17	ec(σ	ec(σ	NUM
ejpam-86	301	18	)	)	PUNCT
ejpam-86	301	19	∂l2(θ	∂l2(θ	PROPN
ejpam-86	301	20	)	)	PUNCT
ejpam-86	301	21	∂β∂v	∂β∂v	PROPN
ejpam-86	301	22	ec(σ	ec(σ	NOUN
ejpam-86	301	23	)	)	PUNCT
ejpam-86	301	24	∂l2(θ	∂l2(θ	PROPN
ejpam-86	301	25	)	)	PUNCT
ejpam-86	301	26	∂b′∂β	∂b′∂β	PROPN
ejpam-86	302	1	∂l2(θ	∂l2(θ	PROPN
ejpam-86	302	2	)	)	PUNCT
ejpam-86	302	3	∂v	∂v	PROPN
ejpam-86	302	4	ec′(σ)∂β	ec′(σ)∂β	ADJ
ejpam-86	302	5	∂l2(θ	∂l2(θ	PROPN
ejpam-86	302	6	)	)	PUNCT
ejpam-86	302	7	∂β2	∂β2	PROPN
ejpam-86	302	8			PROPN
ejpam-86	302	9			PROPN
ejpam-86	302	10	θ̂	θ̂	PROPN
ejpam-86	302	11	.	.	PUNCT
ejpam-86	303	1	note	note	VERB
ejpam-86	303	2	that	that	SCONJ
ejpam-86	303	3	the	the	DET
ejpam-86	303	4	expected	expect	VERB
ejpam-86	303	5	fisher	fisher	PROPN
ejpam-86	303	6	information	information	NOUN
ejpam-86	303	7	matrix	matrix	NOUN
ejpam-86	303	8	and	and	CCONJ
ejpam-86	303	9	its	its	PRON
ejpam-86	303	10	inverse	inverse	NOUN
ejpam-86	303	11	for	for	ADP
ejpam-86	303	12	the	the	DET
ejpam-86	303	13	type	type	NOUN
ejpam-86	303	14	i	i	PRON
ejpam-86	303	15	and	and	CCONJ
ejpam-86	303	16	type	type	PROPN
ejpam-86	303	17	ii	ii	PROPN
ejpam-86	303	18	mvper	mvper	NOUN
ejpam-86	303	19	models	model	NOUN
ejpam-86	303	20	involve	involve	VERB
ejpam-86	303	21	complicated	complicated	ADJ
ejpam-86	303	22	forms	form	NOUN
ejpam-86	303	23	of	of	ADP
ejpam-86	303	24	expected	expect	VERB
ejpam-86	303	25	values	value	NOUN
ejpam-86	303	26	that	that	PRON
ejpam-86	303	27	is	be	AUX
ejpam-86	303	28	difficult	difficult	ADJ
ejpam-86	303	29	to	to	PART
ejpam-86	303	30	compute	compute	VERB
ejpam-86	303	31	.	.	PUNCT
ejpam-86	304	1	therefore	therefore	ADV
ejpam-86	304	2	,	,	PUNCT
ejpam-86	304	3	in	in	ADP
ejpam-86	304	4	what	what	PRON
ejpam-86	304	5	follows	follow	VERB
ejpam-86	304	6	,	,	PUNCT
ejpam-86	304	7	it	it	PRON
ejpam-86	304	8	suffices	suffice	VERB
ejpam-86	304	9	for	for	SCONJ
ejpam-86	304	10	us	we	PRON
ejpam-86	304	11	to	to	PART
ejpam-86	304	12	use	use	VERB
ejpam-86	304	13	the	the	DET
ejpam-86	304	14	complexity	complexity	NOUN
ejpam-86	304	15	of	of	ADP
ejpam-86	304	16	the	the	DET
ejpam-86	304	17	estimated	estimate	VERB
ejpam-86	304	18	observed	observe	VERB
ejpam-86	304	19	inversefisher	inversefisher	ADJ
ejpam-86	304	20	information	information	NOUN
ejpam-86	304	21	matrix	matrix	NOUN
ejpam-86	304	22	(	(	PUNCT
ejpam-86	304	23	ifim	ifim	NOUN
ejpam-86	304	24	)	)	PUNCT
ejpam-86	304	25	above	above	ADV
ejpam-86	304	26	in	in	ADP
ejpam-86	304	27	our	our	PRON
ejpam-86	304	28	numerical	numerical	ADJ
ejpam-86	304	29	examples	example	NOUN
ejpam-86	304	30	.	.	PUNCT
ejpam-86	305	1	6	6	X
ejpam-86	305	2	.	.	X
ejpam-86	305	3	genetic	genetic	ADJ
ejpam-86	305	4	algorithms	algorithm	NOUN
ejpam-86	305	5	(	(	PUNCT
ejpam-86	305	6	gas	gas	NOUN
ejpam-86	305	7	)	)	PUNCT
ejpam-86	305	8	in	in	ADP
ejpam-86	305	9	this	this	DET
ejpam-86	305	10	section	section	NOUN
ejpam-86	305	11	to	to	PART
ejpam-86	305	12	be	be	AUX
ejpam-86	305	13	complete	complete	ADJ
ejpam-86	305	14	and	and	CCONJ
ejpam-86	305	15	for	for	ADP
ejpam-86	305	16	the	the	DET
ejpam-86	305	17	benefit	benefit	NOUN
ejpam-86	305	18	of	of	ADP
ejpam-86	305	19	the	the	DET
ejpam-86	305	20	general	general	ADJ
ejpam-86	305	21	readership	readership	NOUN
ejpam-86	305	22	of	of	ADP
ejpam-86	305	23	the	the	DET
ejpam-86	305	24	paper	paper	NOUN
ejpam-86	305	25	,	,	PUNCT
ejpam-86	305	26	we	we	PRON
ejpam-86	305	27	give	give	VERB
ejpam-86	305	28	the	the	DET
ejpam-86	305	29	general	general	ADJ
ejpam-86	305	30	background	background	NOUN
ejpam-86	305	31	and	and	CCONJ
ejpam-86	305	32	the	the	DET
ejpam-86	305	33	working	working	NOUN
ejpam-86	305	34	of	of	ADP
ejpam-86	305	35	the	the	DET
ejpam-86	305	36	genetic	genetic	ADJ
ejpam-86	305	37	algorithms	algorithm	NOUN
ejpam-86	305	38	(	(	PUNCT
ejpam-86	305	39	gas	gas	NOUN
ejpam-86	305	40	)	)	PUNCT
ejpam-86	305	41	for	for	ADP
ejpam-86	305	42	estimating	estimate	VERB
ejpam-86	305	43	model	model	NOUN
ejpam-86	305	44	parameters	parameter	NOUN
ejpam-86	305	45	and	and	CCONJ
ejpam-86	305	46	model	model	NOUN
ejpam-86	305	47	selection	selection	NOUN
ejpam-86	305	48	contemporaneously	contemporaneously	ADV
ejpam-86	305	49	.	.	PUNCT
ejpam-86	306	1	genetic	genetic	ADJ
ejpam-86	306	2	algorithm	algorithm	NOUN
ejpam-86	306	3	(	(	PUNCT
ejpam-86	306	4	ga	ga	PROPN
ejpam-86	306	5	)	)	PUNCT
ejpam-86	306	6	(	(	PUNCT
ejpam-86	306	7	see	see	VERB
ejpam-86	306	8	,	,	PUNCT
ejpam-86	306	9	e.g.	e.g.	ADV
ejpam-86	306	10	,	,	PUNCT
ejpam-86	306	11	goldberg	goldberg	PROPN
ejpam-86	306	12	(	(	PUNCT
ejpam-86	306	13	[	[	X
ejpam-86	306	14	19	19	NUM
ejpam-86	306	15	]	]	NUM
ejpam-86	306	16	)	)	PUNCT
ejpam-86	306	17	,	,	PUNCT
ejpam-86	306	18	holland	holland	PROPN
ejpam-86	306	19	(	(	PUNCT
ejpam-86	306	20	[	[	X
ejpam-86	306	21	22	22	NUM
ejpam-86	306	22	]	]	SYM
ejpam-86	306	23	)	)	PUNCT
ejpam-86	306	24	,	,	PUNCT
ejpam-86	306	25	mitchell	mitchell	PROPN
ejpam-86	306	26	(	(	PUNCT
ejpam-86	306	27	[	[	X
ejpam-86	306	28	27	27	NUM
ejpam-86	306	29	]	]	NUM
ejpam-86	306	30	)	)	PUNCT
ejpam-86	306	31	)	)	PUNCT
ejpam-86	306	32	is	be	AUX
ejpam-86	306	33	a	a	DET
ejpam-86	306	34	randomized	randomized	ADJ
ejpam-86	306	35	,	,	PUNCT
ejpam-86	306	36	population	population	NOUN
ejpam-86	306	37	-	-	PUNCT
ejpam-86	306	38	based	base	VERB
ejpam-86	306	39	heuristic	heuristic	ADJ
ejpam-86	306	40	optimization	optimization	NOUN
ejpam-86	306	41	technique	technique	NOUN
ejpam-86	306	42	that	that	PRON
ejpam-86	306	43	belong	belong	VERB
ejpam-86	306	44	to	to	ADP
ejpam-86	306	45	the	the	DET
ejpam-86	306	46	general	general	ADJ
ejpam-86	306	47	class	class	NOUN
ejpam-86	306	48	of	of	ADP
ejpam-86	306	49	evolutionary	evolutionary	ADJ
ejpam-86	306	50	algorithms	algorithm	NOUN
ejpam-86	306	51	(	(	PUNCT
ejpam-86	306	52	eas	ea	NOUN
ejpam-86	306	53	)	)	PUNCT
ejpam-86	306	54	.	.	PUNCT
ejpam-86	307	1	ga	ga	PROPN
ejpam-86	307	2	has	have	VERB
ejpam-86	307	3	significant	significant	ADJ
ejpam-86	307	4	advantages	advantage	NOUN
ejpam-86	307	5	such	such	ADJ
ejpam-86	307	6	that	that	SCONJ
ejpam-86	307	7	it	it	PRON
ejpam-86	307	8	is	be	AUX
ejpam-86	307	9	independent	independent	ADJ
ejpam-86	307	10	from	from	ADP
ejpam-86	307	11	the	the	DET
ejpam-86	307	12	complexity	complexity	NOUN
ejpam-86	307	13	of	of	ADP
ejpam-86	307	14	the	the	DET
ejpam-86	307	15	problem	problem	NOUN
ejpam-86	307	16	at	at	ADP
ejpam-86	307	17	hand	hand	NOUN
ejpam-86	307	18	,	,	PUNCT
ejpam-86	307	19	and	and	CCONJ
ejpam-86	307	20	not	not	PART
ejpam-86	307	21	likely	likely	ADJ
ejpam-86	307	22	to	to	PART
ejpam-86	307	23	be	be	AUX
ejpam-86	307	24	restricted	restrict	VERB
ejpam-86	307	25	to	to	ADP
ejpam-86	307	26	a	a	DET
ejpam-86	307	27	local	local	ADJ
ejpam-86	307	28	optimal	optimal	ADJ
ejpam-86	307	29	solution	solution	NOUN
ejpam-86	307	30	,	,	PUNCT
ejpam-86	307	31	and	and	CCONJ
ejpam-86	307	32	it	it	PRON
ejpam-86	307	33	is	be	AUX
ejpam-86	307	34	easy	easy	ADJ
ejpam-86	307	35	to	to	PART
ejpam-86	307	36	use	use	VERB
ejpam-86	307	37	in	in	ADP
ejpam-86	307	38	many	many	ADJ
ejpam-86	307	39	difficult	difficult	ADJ
ejpam-86	307	40	optimization	optimization	NOUN
ejpam-86	307	41	problems	problem	NOUN
ejpam-86	307	42	.	.	PUNCT
ejpam-86	308	1	yang	yang	PROPN
ejpam-86	308	2	and	and	CCONJ
ejpam-86	308	3	honavar	honavar	NOUN
ejpam-86	308	4	(	(	PUNCT
ejpam-86	308	5	[	[	X
ejpam-86	308	6	35	35	NUM
ejpam-86	308	7	]	]	PUNCT
ejpam-86	308	8	)	)	PUNCT
ejpam-86	308	9	use	use	VERB
ejpam-86	308	10	ga	ga	PROPN
ejpam-86	308	11	for	for	ADP
ejpam-86	308	12	the	the	DET
ejpam-86	308	13	selection	selection	NOUN
ejpam-86	308	14	of	of	ADP
ejpam-86	308	15	a	a	DET
ejpam-86	308	16	subset	subset	NOUN
ejpam-86	308	17	of	of	ADP
ejpam-86	308	18	attributes	attribute	NOUN
ejpam-86	308	19	or	or	CCONJ
ejpam-86	308	20	features	feature	NOUN
ejpam-86	308	21	to	to	PART
ejpam-86	308	22	represent	represent	VERB
ejpam-86	308	23	the	the	DET
ejpam-86	308	24	patterns	pattern	NOUN
ejpam-86	308	25	to	to	PART
ejpam-86	308	26	be	be	AUX
ejpam-86	308	27	classified	classify	VERB
ejpam-86	308	28	with	with	ADP
ejpam-86	308	29	neural	neural	ADJ
ejpam-86	308	30	network	network	NOUN
ejpam-86	308	31	(	(	PUNCT
ejpam-86	308	32	nn	nn	NOUN
ejpam-86	308	33	)	)	PUNCT
ejpam-86	308	34	.	.	PUNCT
ejpam-86	309	1	bozdogan	bozdogan	PROPN
ejpam-86	309	2	(	(	PUNCT
ejpam-86	309	3	[	[	X
ejpam-86	309	4	8	8	NUM
ejpam-86	309	5	]	]	PUNCT
ejpam-86	309	6	)	)	PUNCT
ejpam-86	309	7	who	who	PRON
ejpam-86	309	8	introduced	introduce	VERB
ejpam-86	309	9	the	the	DET
ejpam-86	309	10	ga	ga	PROPN
ejpam-86	309	11	in	in	ADP
ejpam-86	309	12	statistical	statistical	ADJ
ejpam-86	309	13	model	model	NOUN
ejpam-86	309	14	selection	selection	NOUN
ejpam-86	309	15	,	,	PUNCT
ejpam-86	309	16	uses	use	VERB
ejpam-86	309	17	the	the	DET
ejpam-86	309	18	ga	ga	PROPN
ejpam-86	309	19	in	in	ADP
ejpam-86	309	20	the	the	DET
ejpam-86	309	21	multiple	multiple	ADJ
ejpam-86	309	22	regression	regression	NOUN
ejpam-86	309	23	model	model	NOUN
ejpam-86	309	24	for	for	ADP
ejpam-86	309	25	subset	subset	ADJ
ejpam-86	309	26	selection	selection	NOUN
ejpam-86	309	27	of	of	ADP
ejpam-86	309	28	the	the	DET
ejpam-86	309	29	best	good	ADJ
ejpam-86	309	30	predictors	predictor	NOUN
ejpam-86	309	31	for	for	ADP
ejpam-86	309	32	intelligent	intelligent	ADJ
ejpam-86	309	33	data	datum	NOUN
ejpam-86	309	34	mining	mining	NOUN
ejpam-86	309	35	under	under	ADP
ejpam-86	309	36	the	the	DET
ejpam-86	309	37	normality	normality	NOUN
ejpam-86	309	38	assumption	assumption	NOUN
ejpam-86	309	39	.	.	PUNCT
ejpam-86	310	1	in	in	ADP
ejpam-86	310	2	ga	ga	PROPN
ejpam-86	310	3	,	,	PUNCT
ejpam-86	310	4	the	the	DET
ejpam-86	310	5	criterion	criterion	NOUN
ejpam-86	310	6	to	to	PART
ejpam-86	310	7	rank	rank	NOUN
ejpam-86	310	8	solutions	solution	NOUN
ejpam-86	310	9	is	be	AUX
ejpam-86	310	10	often	often	ADV
ejpam-86	310	11	called	call	VERB
ejpam-86	310	12	a	a	DET
ejpam-86	310	13	fitness	fitness	NOUN
ejpam-86	310	14	function	function	NOUN
ejpam-86	310	15	.	.	PUNCT
ejpam-86	311	1	a	a	DET
ejpam-86	311	2	set	set	NOUN
ejpam-86	311	3	of	of	ADP
ejpam-86	311	4	solutions	solution	NOUN
ejpam-86	311	5	is	be	AUX
ejpam-86	311	6	called	call	VERB
ejpam-86	311	7	a	a	DET
ejpam-86	311	8	generation	generation	NOUN
ejpam-86	311	9	of	of	ADP
ejpam-86	311	10	population	population	NOUN
ejpam-86	311	11	.	.	PUNCT
ejpam-86	312	1	a	a	DET
ejpam-86	312	2	specific	specific	ADJ
ejpam-86	312	3	solution	solution	NOUN
ejpam-86	312	4	is	be	AUX
ejpam-86	312	5	called	call	VERB
ejpam-86	312	6	an	an	DET
ejpam-86	312	7	individual	individual	NOUN
ejpam-86	312	8	in	in	ADP
ejpam-86	312	9	the	the	DET
ejpam-86	312	10	population	population	NOUN
ejpam-86	312	11	.	.	PUNCT
ejpam-86	313	1	ga	ga	PROPN
ejpam-86	313	2	improves	improve	VERB
ejpam-86	313	3	solutions	solution	NOUN
ejpam-86	313	4	by	by	ADP
ejpam-86	313	5	generating	generate	VERB
ejpam-86	313	6	a	a	DET
ejpam-86	313	7	new	new	ADJ
ejpam-86	313	8	generation	generation	NOUN
ejpam-86	313	9	of	of	ADP
ejpam-86	313	10	population	population	NOUN
ejpam-86	313	11	on	on	ADP
ejpam-86	313	12	the	the	DET
ejpam-86	313	13	base	base	NOUN
ejpam-86	313	14	of	of	ADP
ejpam-86	313	15	current	current	ADJ
ejpam-86	313	16	population	population	NOUN
ejpam-86	313	17	through	through	ADP
ejpam-86	313	18	a	a	DET
ejpam-86	313	19	series	series	NOUN
ejpam-86	313	20	of	of	ADP
ejpam-86	313	21	ga	ga	PROPN
ejpam-86	313	22	operators	operator	NOUN
ejpam-86	313	23	,	,	PUNCT
ejpam-86	313	24	such	such	ADJ
ejpam-86	313	25	as	as	ADP
ejpam-86	313	26	crossover	crossover	NOUN
ejpam-86	313	27	and	and	CCONJ
ejpam-86	313	28	mutation	mutation	NOUN
ejpam-86	313	29	.	.	PUNCT
ejpam-86	314	1	the	the	DET
ejpam-86	314	2	first	first	ADJ
ejpam-86	314	3	generation	generation	NOUN
ejpam-86	314	4	of	of	ADP
ejpam-86	314	5	population	population	NOUN
ejpam-86	314	6	to	to	PART
ejpam-86	314	7	start	start	VERB
ejpam-86	314	8	the	the	DET
ejpam-86	314	9	ga	ga	NOUN
ejpam-86	314	10	process	process	NOUN
ejpam-86	314	11	is	be	AUX
ejpam-86	314	12	generated	generate	VERB
ejpam-86	314	13	as	as	ADP
ejpam-86	314	14	a	a	DET
ejpam-86	314	15	set	set	NOUN
ejpam-86	314	16	of	of	ADP
ejpam-86	314	17	“	"	PUNCT
ejpam-86	314	18	wildly	wildly	ADV
ejpam-86	314	19	’	'	PUNCT
ejpam-86	314	20	guessed	guess	VERB
ejpam-86	314	21	or	or	CCONJ
ejpam-86	314	22	randomly	randomly	ADV
ejpam-86	314	23	generated	generate	VERB
ejpam-86	314	24	solutions	solution	NOUN
ejpam-86	314	25	.	.	PUNCT
ejpam-86	315	1	for	for	SCONJ
ejpam-86	315	2	the	the	DET
ejpam-86	315	3	ga	ga	PROPN
ejpam-86	315	4	to	to	PART
ejpam-86	315	5	evolve	evolve	VERB
ejpam-86	315	6	,	,	PUNCT
ejpam-86	315	7	a	a	DET
ejpam-86	315	8	solution	solution	NOUN
ejpam-86	315	9	needs	need	VERB
ejpam-86	315	10	to	to	PART
ejpam-86	315	11	be	be	AUX
ejpam-86	315	12	represented	represent	VERB
ejpam-86	315	13	in	in	ADP
ejpam-86	315	14	binary	binary	ADJ
ejpam-86	315	15	string	string	NOUN
ejpam-86	315	16	format	format	NOUN
ejpam-86	315	17	.	.	PUNCT
ejpam-86	316	1	a	a	DET
ejpam-86	316	2	binary	binary	ADJ
ejpam-86	316	3	string	string	NOUN
ejpam-86	316	4	represents	represent	VERB
ejpam-86	316	5	a	a	DET
ejpam-86	316	6	solution	solution	NOUN
ejpam-86	316	7	and	and	CCONJ
ejpam-86	316	8	it	it	PRON
ejpam-86	316	9	is	be	AUX
ejpam-86	316	10	often	often	ADV
ejpam-86	316	11	called	call	VERB
ejpam-86	316	12	a	a	DET
ejpam-86	316	13	chromosome	chromosome	NOUN
ejpam-86	316	14	.	.	PUNCT
ejpam-86	317	1	the	the	DET
ejpam-86	317	2	individuals	individual	NOUN
ejpam-86	317	3	in	in	ADP
ejpam-86	317	4	the	the	DET
ejpam-86	317	5	current	current	ADJ
ejpam-86	317	6	population	population	NOUN
ejpam-86	317	7	are	be	AUX
ejpam-86	317	8	used	use	VERB
ejpam-86	317	9	to	to	PART
ejpam-86	317	10	generate	generate	VERB
ejpam-86	317	11	the	the	DET
ejpam-86	317	12	new	new	ADJ
ejpam-86	317	13	population	population	NOUN
ejpam-86	317	14	.	.	PUNCT
ejpam-86	318	1	there	there	PRON
ejpam-86	318	2	are	be	VERB
ejpam-86	318	3	different	different	ADJ
ejpam-86	318	4	strategies	strategy	NOUN
ejpam-86	318	5	to	to	PART
ejpam-86	318	6	generate	generate	VERB
ejpam-86	318	7	a	a	DET
ejpam-86	318	8	new	new	ADJ
ejpam-86	318	9	population	population	NOUN
ejpam-86	318	10	.	.	PUNCT
ejpam-86	319	1	one	one	NUM
ejpam-86	319	2	strategy	strategy	NOUN
ejpam-86	319	3	commonly	commonly	ADV
ejpam-86	319	4	used	use	VERB
ejpam-86	319	5	is	be	AUX
ejpam-86	319	6	the	the	DET
ejpam-86	319	7	so	so	ADV
ejpam-86	319	8	-	-	PUNCT
ejpam-86	319	9	called	call	VERB
ejpam-86	319	10	called	call	VERB
ejpam-86	319	11	“	"	PUNCT
ejpam-86	319	12	natural	natural	ADJ
ejpam-86	319	13	”	"	PUNCT
ejpam-86	319	14	selecting	select	VERB
ejpam-86	319	15	strategy	strategy	NOUN
ejpam-86	319	16	.	.	PUNCT
ejpam-86	320	1	with	with	ADP
ejpam-86	320	2	this	this	DET
ejpam-86	320	3	strategy	strategy	NOUN
ejpam-86	320	4	,	,	PUNCT
ejpam-86	320	5	the	the	DET
ejpam-86	320	6	chance	chance	NOUN
ejpam-86	320	7	of	of	ADP
ejpam-86	320	8	an	an	DET
ejpam-86	320	9	individual	individual	NOUN
ejpam-86	320	10	being	be	AUX
ejpam-86	320	11	selected	select	VERB
ejpam-86	320	12	is	be	AUX
ejpam-86	320	13	proportional	proportional	ADJ
ejpam-86	320	14	to	to	ADP
ejpam-86	320	15	the	the	DET
ejpam-86	320	16	ratio	ratio	NOUN
ejpam-86	320	17	:	:	PUNCT
ejpam-86	320	18	rj	rj	PROPN
ejpam-86	320	19	=	=	SYM
ejpam-86	320	20	∆fitnessj/∆fitness	∆fitnessj/∆fitness	X
ejpam-86	320	21	(	(	PUNCT
ejpam-86	320	22	6.1	6.1	NUM
ejpam-86	320	23	)	)	PUNCT
ejpam-86	320	24	where	where	SCONJ
ejpam-86	320	25	∆fitnessj	∆fitnessj	NOUN
ejpam-86	320	26	=	=	SYM
ejpam-86	320	27	fitnessmax	fitnessmax	NOUN
ejpam-86	320	28	−	−	PROPN
ejpam-86	320	29	fitnessj	fitnessj	NOUN
ejpam-86	320	30	and	and	CCONJ
ejpam-86	320	31	∆fitness	∆fitness	NOUN
ejpam-86	320	32	is	be	AUX
ejpam-86	320	33	the	the	DET
ejpam-86	320	34	mean	mean	NOUN
ejpam-86	320	35	of	of	ADP
ejpam-86	320	36	∆fitnessj	∆fitnessj	PROPN
ejpam-86	320	37	.	.	PUNCT
ejpam-86	321	1	the	the	DET
ejpam-86	321	2	chance	chance	NOUN
ejpam-86	321	3	of	of	ADP
ejpam-86	321	4	an	an	DET
ejpam-86	321	5	individual	individual	NOUN
ejpam-86	321	6	being	be	AUX
ejpam-86	321	7	selected	select	VERB
ejpam-86	321	8	is	be	AUX
ejpam-86	321	9	proportional	proportional	ADJ
ejpam-86	321	10	to	to	ADP
ejpam-86	321	11	this	this	DET
ejpam-86	321	12	ratio	ratio	NOUN
ejpam-86	321	13	.	.	PUNCT
ejpam-86	322	1	in	in	ADP
ejpam-86	322	2	other	other	ADJ
ejpam-86	322	3	words	word	NOUN
ejpam-86	322	4	,	,	PUNCT
ejpam-86	322	5	an	an	DET
ejpam-86	322	6	individual	individual	ADJ
ejpam-86	322	7	m.	m.	NOUN
ejpam-86	322	8	liu	liu	PROPN
ejpam-86	322	9	and	and	CCONJ
ejpam-86	322	10	h.	h.	PROPN
ejpam-86	322	11	bozdogan	bozdogan	PROPN
ejpam-86	322	12	/	/	SYM
ejpam-86	322	13	eur	eur	PROPN
ejpam-86	322	14	.	.	PUNCT
ejpam-86	323	1	j.	j.	PROPN
ejpam-86	323	2	pure	pure	PROPN
ejpam-86	323	3	appl	appl	PROPN
ejpam-86	323	4	.	.	PROPN
ejpam-86	323	5	math	math	PROPN
ejpam-86	323	6	,	,	PUNCT
ejpam-86	323	7	1	1	NUM
ejpam-86	323	8	(	(	PUNCT
ejpam-86	323	9	2008	2008	NUM
ejpam-86	323	10	)	)	PUNCT
ejpam-86	323	11	,	,	PUNCT
ejpam-86	323	12	(	(	PUNCT
ejpam-86	323	13	4	4	NUM
ejpam-86	323	14	-	-	SYM
ejpam-86	323	15	37	37	NUM
ejpam-86	323	16	)	)	PUNCT
ejpam-86	323	17	17	17	NUM
ejpam-86	323	18	with	with	ADP
ejpam-86	323	19	a	a	DET
ejpam-86	323	20	ratio	ratio	NOUN
ejpam-86	323	21	of	of	ADP
ejpam-86	323	22	two	two	NUM
ejpam-86	323	23	is	be	AUX
ejpam-86	323	24	twice	twice	ADV
ejpam-86	323	25	as	as	ADV
ejpam-86	323	26	likely	likely	ADJ
ejpam-86	323	27	to	to	PART
ejpam-86	323	28	be	be	AUX
ejpam-86	323	29	selected	select	VERB
ejpam-86	323	30	as	as	ADP
ejpam-86	323	31	an	an	DET
ejpam-86	323	32	individual	individual	NOUN
ejpam-86	323	33	with	with	ADP
ejpam-86	323	34	a	a	DET
ejpam-86	323	35	ratio	ratio	NOUN
ejpam-86	323	36	of	of	ADP
ejpam-86	323	37	one	one	NUM
ejpam-86	323	38	.	.	PUNCT
ejpam-86	324	1	this	this	DET
ejpam-86	324	2	approach	approach	NOUN
ejpam-86	324	3	is	be	AUX
ejpam-86	324	4	called	call	VERB
ejpam-86	324	5	the	the	DET
ejpam-86	324	6	proportional	proportional	ADJ
ejpam-86	324	7	selection	selection	NOUN
ejpam-86	324	8	.	.	PUNCT
ejpam-86	325	1	there	there	PRON
ejpam-86	325	2	are	be	VERB
ejpam-86	325	3	also	also	ADV
ejpam-86	325	4	other	other	ADJ
ejpam-86	325	5	selection	selection	NOUN
ejpam-86	325	6	strategies	strategy	NOUN
ejpam-86	325	7	such	such	ADJ
ejpam-86	325	8	as	as	ADP
ejpam-86	325	9	the	the	DET
ejpam-86	325	10	rank	rank	NOUN
ejpam-86	325	11	order	order	NOUN
ejpam-86	325	12	selection	selection	NOUN
ejpam-86	325	13	,	,	PUNCT
ejpam-86	325	14	and	and	CCONJ
ejpam-86	325	15	so	so	ADV
ejpam-86	325	16	forth	forth	ADV
ejpam-86	325	17	.	.	PUNCT
ejpam-86	326	1	a	a	DET
ejpam-86	326	2	pair	pair	NOUN
ejpam-86	326	3	of	of	ADP
ejpam-86	326	4	individuals	individual	NOUN
ejpam-86	326	5	selected	select	VERB
ejpam-86	326	6	from	from	ADP
ejpam-86	326	7	the	the	DET
ejpam-86	326	8	current	current	ADJ
ejpam-86	326	9	population	population	NOUN
ejpam-86	326	10	are	be	AUX
ejpam-86	326	11	used	use	VERB
ejpam-86	326	12	to	to	PART
ejpam-86	326	13	generate	generate	VERB
ejpam-86	326	14	a	a	DET
ejpam-86	326	15	pair	pair	NOUN
ejpam-86	326	16	of	of	ADP
ejpam-86	326	17	new	new	ADJ
ejpam-86	326	18	solutions	solution	NOUN
ejpam-86	326	19	,	,	PUNCT
ejpam-86	326	20	often	often	ADV
ejpam-86	326	21	called	call	VERB
ejpam-86	326	22	“	"	PUNCT
ejpam-86	326	23	offsprings	offspring	NOUN
ejpam-86	326	24	”	"	PUNCT
ejpam-86	326	25	,	,	PUNCT
ejpam-86	326	26	through	through	ADP
ejpam-86	326	27	the	the	DET
ejpam-86	326	28	ga	ga	PROPN
ejpam-86	326	29	operator	operator	NOUN
ejpam-86	326	30	crossover	crossover	NOUN
ejpam-86	326	31	.	.	PUNCT
ejpam-86	327	1	crossover	crossover	NOUN
ejpam-86	327	2	mimics	mimic	VERB
ejpam-86	327	3	the	the	DET
ejpam-86	327	4	process	process	NOUN
ejpam-86	327	5	of	of	ADP
ejpam-86	327	6	mating	mating	NOUN
ejpam-86	327	7	.	.	PUNCT
ejpam-86	328	1	the	the	DET
ejpam-86	328	2	pair	pair	NOUN
ejpam-86	328	3	of	of	ADP
ejpam-86	328	4	chromosomes	chromosome	NOUN
ejpam-86	328	5	chosen	choose	VERB
ejpam-86	328	6	for	for	ADP
ejpam-86	328	7	crossover	crossover	NOUN
ejpam-86	328	8	is	be	AUX
ejpam-86	328	9	controlled	control	VERB
ejpam-86	328	10	by	by	ADP
ejpam-86	328	11	the	the	DET
ejpam-86	328	12	crossover	crossover	NOUN
ejpam-86	328	13	probability	probability	NOUN
ejpam-86	328	14	(	(	PUNCT
ejpam-86	328	15	pc	pc	NOUN
ejpam-86	328	16	)	)	PUNCT
ejpam-86	328	17	which	which	PRON
ejpam-86	328	18	is	be	AUX
ejpam-86	328	19	an	an	DET
ejpam-86	328	20	input	input	NOUN
ejpam-86	328	21	parameter	parameter	NOUN
ejpam-86	328	22	of	of	ADP
ejpam-86	328	23	the	the	DET
ejpam-86	328	24	algorithm	algorithm	NOUN
ejpam-86	328	25	.	.	PUNCT
ejpam-86	329	1	crossover	crossover	NOUN
ejpam-86	329	2	point	point	NOUN
ejpam-86	329	3	,	,	PUNCT
ejpam-86	329	4	where	where	SCONJ
ejpam-86	329	5	the	the	DET
ejpam-86	329	6	binary	binary	PROPN
ejpam-86	329	7	string	string	NOUN
ejpam-86	329	8	is	be	AUX
ejpam-86	329	9	broken	break	VERB
ejpam-86	329	10	for	for	ADP
ejpam-86	329	11	crossover	crossover	NOUN
ejpam-86	329	12	,	,	PUNCT
ejpam-86	329	13	is	be	AUX
ejpam-86	329	14	picked	pick	VERB
ejpam-86	329	15	randomly	randomly	ADV
ejpam-86	329	16	along	along	ADP
ejpam-86	329	17	each	each	DET
ejpam-86	329	18	pair	pair	NOUN
ejpam-86	329	19	of	of	ADP
ejpam-86	329	20	parent	parent	NOUN
ejpam-86	329	21	chromosomes	chromosome	NOUN
ejpam-86	329	22	.	.	PUNCT
ejpam-86	330	1	in	in	ADP
ejpam-86	330	2	the	the	DET
ejpam-86	330	3	algorithm	algorithm	NOUN
ejpam-86	330	4	,	,	PUNCT
ejpam-86	330	5	we	we	PRON
ejpam-86	330	6	give	give	VERB
ejpam-86	330	7	three	three	NUM
ejpam-86	330	8	choices	choice	NOUN
ejpam-86	330	9	of	of	ADP
ejpam-86	330	10	crossover	crossover	NOUN
ejpam-86	330	11	types	type	NOUN
ejpam-86	330	12	corresponding	correspond	VERB
ejpam-86	330	13	to	to	ADP
ejpam-86	330	14	different	different	ADJ
ejpam-86	330	15	locations	location	NOUN
ejpam-86	330	16	of	of	ADP
ejpam-86	330	17	crossover	crossover	NOUN
ejpam-86	330	18	points	point	NOUN
ejpam-86	330	19	.	.	PUNCT
ejpam-86	331	1	in	in	ADP
ejpam-86	331	2	the	the	DET
ejpam-86	331	3	following	following	NOUN
ejpam-86	331	4	,	,	PUNCT
ejpam-86	331	5	“	"	PUNCT
ejpam-86	331	6	|	|	ADV
ejpam-86	331	7	”	"	PUNCT
ejpam-86	331	8	represents	represent	VERB
ejpam-86	331	9	a	a	DET
ejpam-86	331	10	crossover	crossover	NOUN
ejpam-86	331	11	point	point	NOUN
ejpam-86	331	12	.	.	PUNCT
ejpam-86	332	1	•	•	NUM
ejpam-86	332	2	single	single	ADJ
ejpam-86	332	3	point	point	NOUN
ejpam-86	332	4	crossover	crossover	VERB
ejpam-86	332	5	parent	parent	NOUN
ejpam-86	332	6	a	a	DET
ejpam-86	332	7	10|001001001	10|001001001	NUM
ejpam-86	332	8	↓↑	↓↑	NOUN
ejpam-86	332	9	parent	parent	NOUN
ejpam-86	332	10	b	b	NOUN
ejpam-86	332	11	00|010011000	00|010011000	NUM
ejpam-86	332	12	→	→	SYM
ejpam-86	332	13	offspring	offspre	VERB
ejpam-86	332	14	a	a	DET
ejpam-86	332	15	10|010011000	10|010011000	NUM
ejpam-86	332	16	offspring	offspring	NOUN
ejpam-86	332	17	b	b	NOUN
ejpam-86	332	18	00|001001001	00|001001001	NUM
ejpam-86	332	19	•	•	NUM
ejpam-86	332	20	two	two	NUM
ejpam-86	332	21	point	point	NOUN
ejpam-86	332	22	crossover	crossover	VERB
ejpam-86	332	23	parent	parent	NOUN
ejpam-86	332	24	a	a	DET
ejpam-86	332	25	10|001001|001	10|001001|001	NUM
ejpam-86	332	26	↓↑	↓↑	NOUN
ejpam-86	332	27	parent	parent	NOUN
ejpam-86	332	28	b	b	PROPN
ejpam-86	332	29	00|010011|000	00|010011|000	NUM
ejpam-86	332	30	→	→	PUNCT
ejpam-86	332	31	offspring	offspre	VERB
ejpam-86	332	32	a	a	DET
ejpam-86	332	33	10|010011|001	10|010011|001	NUM
ejpam-86	332	34	offspring	offspring	NOUN
ejpam-86	332	35	b	b	NOUN
ejpam-86	332	36	00|001001|000	00|001001|000	NUM
ejpam-86	332	37	•	•	NOUN
ejpam-86	332	38	uniform	uniform	NOUN
ejpam-86	332	39	crossover	crossover	NOUN
ejpam-86	332	40	bits	bit	NOUN
ejpam-86	332	41	are	be	AUX
ejpam-86	332	42	randomly	randomly	ADV
ejpam-86	332	43	switched	switch	VERB
ejpam-86	332	44	between	between	ADP
ejpam-86	332	45	parents	parent	NOUN
ejpam-86	332	46	:	:	PUNCT
ejpam-86	332	47	parent	parent	NOUN
ejpam-86	332	48	a	a	DET
ejpam-86	332	49	10001001001	10001001001	NUM
ejpam-86	332	50	↓↑	↓↑	NOUN
ejpam-86	332	51	parent	parent	NOUN
ejpam-86	332	52	b	b	PROPN
ejpam-86	332	53	00010011000	00010011000	NUM
ejpam-86	332	54	→	→	SYM
ejpam-86	332	55	possible	possible	ADJ
ejpam-86	332	56	offspring	offspring	NOUN
ejpam-86	332	57	a	a	DET
ejpam-86	332	58	00000011001	00000011001	NUM
ejpam-86	332	59	possible	possible	ADJ
ejpam-86	332	60	offspring	offspring	NOUN
ejpam-86	332	61	b	b	SYM
ejpam-86	332	62	10011001000	10011001000	NUM
ejpam-86	332	63	mutation	mutation	NOUN
ejpam-86	332	64	is	be	AUX
ejpam-86	332	65	another	another	DET
ejpam-86	332	66	parameter	parameter	NOUN
ejpam-86	332	67	or	or	CCONJ
ejpam-86	332	68	operator	operator	NOUN
ejpam-86	332	69	used	use	VERB
ejpam-86	332	70	in	in	ADP
ejpam-86	332	71	ga	ga	PROPN
ejpam-86	332	72	to	to	PART
ejpam-86	332	73	realize	realize	VERB
ejpam-86	332	74	a	a	DET
ejpam-86	332	75	global	global	ADJ
ejpam-86	332	76	search	search	NOUN
ejpam-86	332	77	.	.	PUNCT
ejpam-86	333	1	during	during	ADP
ejpam-86	333	2	mutation	mutation	NOUN
ejpam-86	333	3	,	,	PUNCT
ejpam-86	333	4	each	each	DET
ejpam-86	333	5	bit	bit	NOUN
ejpam-86	333	6	in	in	ADP
ejpam-86	333	7	a	a	DET
ejpam-86	333	8	binary	binary	ADJ
ejpam-86	333	9	string	string	NOUN
ejpam-86	333	10	can	can	AUX
ejpam-86	333	11	change	change	VERB
ejpam-86	333	12	from	from	ADP
ejpam-86	333	13	0	0	NUM
ejpam-86	333	14	to	to	ADP
ejpam-86	333	15	1	1	NUM
ejpam-86	333	16	,	,	PUNCT
ejpam-86	333	17	or	or	CCONJ
ejpam-86	333	18	from	from	ADP
ejpam-86	333	19	1	1	NUM
ejpam-86	333	20	to	to	ADP
ejpam-86	333	21	0	0	NUM
ejpam-86	333	22	,	,	PUNCT
ejpam-86	333	23	with	with	ADP
ejpam-86	333	24	a	a	DET
ejpam-86	333	25	user	user	NOUN
ejpam-86	333	26	input	input	NOUN
ejpam-86	333	27	probability	probability	NOUN
ejpam-86	333	28	called	call	VERB
ejpam-86	333	29	the	the	DET
ejpam-86	333	30	mutation	mutation	NOUN
ejpam-86	333	31	probability	probability	NOUN
ejpam-86	333	32	(	(	PUNCT
ejpam-86	333	33	pm	pm	NOUN
ejpam-86	333	34	)	)	PUNCT
ejpam-86	333	35	.	.	PUNCT
ejpam-86	334	1	hence	hence	ADV
ejpam-86	334	2	,	,	PUNCT
ejpam-86	334	3	the	the	DET
ejpam-86	334	4	searching	searching	NOUN
ejpam-86	334	5	process	process	NOUN
ejpam-86	334	6	can	can	AUX
ejpam-86	334	7	jump	jump	VERB
ejpam-86	334	8	to	to	ADP
ejpam-86	334	9	another	another	DET
ejpam-86	334	10	area	area	NOUN
ejpam-86	334	11	of	of	ADP
ejpam-86	334	12	the	the	DET
ejpam-86	334	13	fitness	fitness	NOUN
ejpam-86	334	14	landscape	landscape	NOUN
ejpam-86	334	15	,	,	PUNCT
ejpam-86	334	16	instead	instead	ADV
ejpam-86	334	17	of	of	ADP
ejpam-86	334	18	being	be	AUX
ejpam-86	334	19	limited	limit	VERB
ejpam-86	334	20	in	in	ADP
ejpam-86	334	21	a	a	DET
ejpam-86	334	22	local	local	ADJ
ejpam-86	334	23	optimum	optimum	ADJ
ejpam-86	334	24	area	area	NOUN
ejpam-86	334	25	.	.	PUNCT
ejpam-86	335	1	our	our	PRON
ejpam-86	335	2	ga	ga	PROPN
ejpam-86	335	3	also	also	ADV
ejpam-86	335	4	allows	allow	VERB
ejpam-86	335	5	the	the	DET
ejpam-86	335	6	elitism	elitism	NOUN
ejpam-86	335	7	rule	rule	NOUN
ejpam-86	335	8	(	(	PUNCT
ejpam-86	335	9	er	er	INTJ
ejpam-86	335	10	)	)	PUNCT
ejpam-86	335	11	.	.	PUNCT
ejpam-86	336	1	when	when	SCONJ
ejpam-86	336	2	the	the	DET
ejpam-86	336	3	elitism	elitism	NOUN
ejpam-86	336	4	rule	rule	NOUN
ejpam-86	336	5	is	be	AUX
ejpam-86	336	6	applied	apply	VERB
ejpam-86	336	7	,	,	PUNCT
ejpam-86	336	8	the	the	DET
ejpam-86	336	9	best	good	ADJ
ejpam-86	336	10	solution	solution	NOUN
ejpam-86	336	11	of	of	ADP
ejpam-86	336	12	a	a	DET
ejpam-86	336	13	generation	generation	NOUN
ejpam-86	336	14	will	will	AUX
ejpam-86	336	15	be	be	AUX
ejpam-86	336	16	copied	copy	VERB
ejpam-86	336	17	without	without	ADP
ejpam-86	336	18	changes	change	NOUN
ejpam-86	336	19	to	to	ADP
ejpam-86	336	20	the	the	DET
ejpam-86	336	21	next	next	ADJ
ejpam-86	336	22	generation	generation	NOUN
ejpam-86	336	23	.	.	PUNCT
ejpam-86	337	1	er	er	INTJ
ejpam-86	337	2	guarantees	guarantee	VERB
ejpam-86	337	3	the	the	DET
ejpam-86	337	4	individual	individual	NOUN
ejpam-86	337	5	with	with	ADP
ejpam-86	337	6	the	the	DET
ejpam-86	337	7	best	good	ADJ
ejpam-86	337	8	fitness	fitness	NOUN
ejpam-86	337	9	in	in	ADP
ejpam-86	337	10	the	the	DET
ejpam-86	337	11	current	current	ADJ
ejpam-86	337	12	generation	generation	NOUN
ejpam-86	337	13	to	to	PART
ejpam-86	337	14	survive	survive	VERB
ejpam-86	337	15	in	in	ADP
ejpam-86	337	16	the	the	DET
ejpam-86	337	17	next	next	ADJ
ejpam-86	337	18	generation	generation	NOUN
ejpam-86	337	19	.	.	PUNCT
ejpam-86	338	1	in	in	ADP
ejpam-86	338	2	other	other	ADJ
ejpam-86	338	3	words	word	NOUN
ejpam-86	338	4	,	,	PUNCT
ejpam-86	338	5	the	the	DET
ejpam-86	338	6	best	good	ADJ
ejpam-86	338	7	solution	solution	NOUN
ejpam-86	338	8	is	be	AUX
ejpam-86	338	9	passed	pass	VERB
ejpam-86	338	10	from	from	ADP
ejpam-86	338	11	one	one	NUM
ejpam-86	338	12	generation	generation	NOUN
ejpam-86	338	13	to	to	ADP
ejpam-86	338	14	the	the	DET
ejpam-86	338	15	next	next	ADJ
ejpam-86	338	16	and	and	CCONJ
ejpam-86	338	17	the	the	DET
ejpam-86	338	18	survival	survival	NOUN
ejpam-86	338	19	of	of	ADP
ejpam-86	338	20	the	the	DET
ejpam-86	338	21	fittest	fit	ADJ
ejpam-86	338	22	is	be	AUX
ejpam-86	338	23	achieved	achieve	VERB
ejpam-86	338	24	until	until	SCONJ
ejpam-86	338	25	the	the	DET
ejpam-86	338	26	ga	ga	PROPN
ejpam-86	338	27	converges	converge	VERB
ejpam-86	338	28	.	.	PUNCT
ejpam-86	339	1	the	the	DET
ejpam-86	339	2	outline	outline	NOUN
ejpam-86	339	3	of	of	ADP
ejpam-86	339	4	the	the	DET
ejpam-86	339	5	ga	ga	PROPN
ejpam-86	339	6	procedures	procedure	NOUN
ejpam-86	339	7	for	for	ADP
ejpam-86	339	8	model	model	NOUN
ejpam-86	339	9	parameter	parameter	PROPN
ejpam-86	339	10	estimation	estimation	NOUN
ejpam-86	339	11	and	and	CCONJ
ejpam-86	339	12	model	model	NOUN
ejpam-86	339	13	selection	selection	NOUN
ejpam-86	339	14	is	be	AUX
ejpam-86	339	15	summarized	summarize	VERB
ejpam-86	339	16	as	as	SCONJ
ejpam-86	339	17	follows	follow	VERB
ejpam-86	339	18	:	:	PUNCT
ejpam-86	339	19	step	step	NOUN
ejpam-86	339	20	1	1	NUM
ejpam-86	339	21	:	:	PUNCT
ejpam-86	339	22	create	create	VERB
ejpam-86	339	23	a	a	DET
ejpam-86	339	24	generation	generation	NOUN
ejpam-86	339	25	of	of	ADP
ejpam-86	339	26	population	population	NOUN
ejpam-86	339	27	with	with	ADP
ejpam-86	339	28	a	a	DET
ejpam-86	339	29	given	give	VERB
ejpam-86	339	30	population	population	NOUN
ejpam-86	339	31	size	size	NOUN
ejpam-86	339	32	.	.	PUNCT
ejpam-86	340	1	step	step	NOUN
ejpam-86	340	2	2	2	NUM
ejpam-86	340	3	:	:	PUNCT
ejpam-86	340	4	encode	encode	VERB
ejpam-86	340	5	the	the	DET
ejpam-86	340	6	individuals	individual	NOUN
ejpam-86	340	7	into	into	ADP
ejpam-86	340	8	binary	binary	ADJ
ejpam-86	340	9	strings	string	NOUN
ejpam-86	340	10	.	.	PUNCT
ejpam-86	341	1	m.	m.	PROPN
ejpam-86	341	2	liu	liu	PROPN
ejpam-86	341	3	and	and	CCONJ
ejpam-86	341	4	h.	h.	PROPN
ejpam-86	341	5	bozdogan	bozdogan	PROPN
ejpam-86	341	6	/	/	SYM
ejpam-86	341	7	eur	eur	PROPN
ejpam-86	341	8	.	.	PUNCT
ejpam-86	342	1	j.	j.	PROPN
ejpam-86	342	2	pure	pure	PROPN
ejpam-86	342	3	appl	appl	PROPN
ejpam-86	342	4	.	.	PROPN
ejpam-86	342	5	math	math	PROPN
ejpam-86	342	6	,	,	PUNCT
ejpam-86	342	7	1	1	NUM
ejpam-86	342	8	(	(	PUNCT
ejpam-86	342	9	2008	2008	NUM
ejpam-86	342	10	)	)	PUNCT
ejpam-86	342	11	,	,	PUNCT
ejpam-86	342	12	(	(	PUNCT
ejpam-86	342	13	4	4	NUM
ejpam-86	342	14	-	-	SYM
ejpam-86	342	15	37	37	NUM
ejpam-86	342	16	)	)	PUNCT
ejpam-86	342	17	18	18	NUM
ejpam-86	342	18	step	step	NOUN
ejpam-86	342	19	3	3	NUM
ejpam-86	342	20	:	:	PUNCT
ejpam-86	342	21	rank	rank	VERB
ejpam-86	342	22	each	each	DET
ejpam-86	342	23	individual	individual	NOUN
ejpam-86	342	24	in	in	ADP
ejpam-86	342	25	the	the	DET
ejpam-86	342	26	population	population	NOUN
ejpam-86	342	27	according	accord	VERB
ejpam-86	342	28	to	to	ADP
ejpam-86	342	29	the	the	DET
ejpam-86	342	30	given	give	VERB
ejpam-86	342	31	fitness	fitness	NOUN
ejpam-86	342	32	function	function	NOUN
ejpam-86	342	33	.	.	PUNCT
ejpam-86	343	1	step	step	NOUN
ejpam-86	343	2	4	4	NUM
ejpam-86	343	3	:	:	PUNCT
ejpam-86	343	4	select	select	VERB
ejpam-86	343	5	individuals	individual	NOUN
ejpam-86	343	6	to	to	PART
ejpam-86	343	7	be	be	AUX
ejpam-86	343	8	used	use	VERB
ejpam-86	343	9	to	to	PART
ejpam-86	343	10	generate	generate	VERB
ejpam-86	343	11	the	the	DET
ejpam-86	343	12	new	new	ADJ
ejpam-86	343	13	population	population	NOUN
ejpam-86	343	14	.	.	PUNCT
ejpam-86	344	1	step	step	NOUN
ejpam-86	344	2	5	5	NUM
ejpam-86	344	3	:	:	PUNCT
ejpam-86	344	4	do	do	AUX
ejpam-86	344	5	crossover	crossover	VERB
ejpam-86	344	6	on	on	ADP
ejpam-86	344	7	selected	select	VERB
ejpam-86	344	8	individuals	individual	NOUN
ejpam-86	344	9	with	with	ADP
ejpam-86	344	10	a	a	DET
ejpam-86	344	11	given	give	VERB
ejpam-86	344	12	crossover	crossover	NOUN
ejpam-86	344	13	type	type	NOUN
ejpam-86	344	14	and	and	CCONJ
ejpam-86	344	15	crossover	crossover	VERB
ejpam-86	344	16	probability	probability	NOUN
ejpam-86	344	17	and	and	CCONJ
ejpam-86	344	18	create	create	VERB
ejpam-86	344	19	a	a	DET
ejpam-86	344	20	new	new	ADJ
ejpam-86	344	21	population	population	NOUN
ejpam-86	344	22	.	.	PUNCT
ejpam-86	345	1	step	step	NOUN
ejpam-86	345	2	6	6	NUM
ejpam-86	345	3	:	:	PUNCT
ejpam-86	345	4	do	do	VERB
ejpam-86	345	5	mutation	mutation	NOUN
ejpam-86	345	6	on	on	ADP
ejpam-86	345	7	the	the	DET
ejpam-86	345	8	new	new	ADJ
ejpam-86	345	9	population	population	NOUN
ejpam-86	345	10	with	with	ADP
ejpam-86	345	11	a	a	DET
ejpam-86	345	12	given	give	VERB
ejpam-86	345	13	mutation	mutation	NOUN
ejpam-86	345	14	probability	probability	NOUN
ejpam-86	345	15	.	.	PUNCT
ejpam-86	346	1	step	step	NOUN
ejpam-86	346	2	7	7	NUM
ejpam-86	346	3	:	:	PUNCT
ejpam-86	346	4	do	do	AUX
ejpam-86	346	5	ga	ga	NOUN
ejpam-86	346	6	engineering	engineering	NOUN
ejpam-86	346	7	on	on	ADP
ejpam-86	346	8	the	the	DET
ejpam-86	346	9	new	new	ADJ
ejpam-86	346	10	population	population	NOUN
ejpam-86	346	11	with	with	ADP
ejpam-86	346	12	a	a	DET
ejpam-86	346	13	given	give	VERB
ejpam-86	346	14	engineering	engineering	NOUN
ejpam-86	346	15	probability	probability	NOUN
ejpam-86	346	16	.	.	PUNCT
ejpam-86	347	1	step	step	NOUN
ejpam-86	347	2	8	8	NUM
ejpam-86	347	3	:	:	PUNCT
ejpam-86	347	4	do	do	VERB
ejpam-86	347	5	elitism	elitism	NOUN
ejpam-86	347	6	if	if	SCONJ
ejpam-86	347	7	required	require	VERB
ejpam-86	347	8	.	.	PUNCT
ejpam-86	348	1	elitism	elitism	NOUN
ejpam-86	348	2	means	mean	VERB
ejpam-86	348	3	that	that	SCONJ
ejpam-86	348	4	the	the	DET
ejpam-86	348	5	best	good	ADJ
ejpam-86	348	6	individual	individual	NOUN
ejpam-86	348	7	in	in	ADP
ejpam-86	348	8	current	current	ADJ
ejpam-86	348	9	population	population	NOUN
ejpam-86	348	10	is	be	AUX
ejpam-86	348	11	guaranteed	guarantee	VERB
ejpam-86	348	12	to	to	PART
ejpam-86	348	13	be	be	AUX
ejpam-86	348	14	included	include	VERB
ejpam-86	348	15	in	in	ADP
ejpam-86	348	16	the	the	DET
ejpam-86	348	17	new	new	ADJ
ejpam-86	348	18	population	population	NOUN
ejpam-86	348	19	.	.	PUNCT
ejpam-86	349	1	step	step	NOUN
ejpam-86	349	2	9	9	NUM
ejpam-86	349	3	:	:	PUNCT
ejpam-86	349	4	replace	replace	VERB
ejpam-86	349	5	the	the	DET
ejpam-86	349	6	current	current	ADJ
ejpam-86	349	7	population	population	NOUN
ejpam-86	349	8	with	with	ADP
ejpam-86	349	9	the	the	DET
ejpam-86	349	10	new	new	ADJ
ejpam-86	349	11	population	population	NOUN
ejpam-86	349	12	.	.	PUNCT
ejpam-86	350	1	step	step	NOUN
ejpam-86	350	2	10	10	NUM
ejpam-86	350	3	:	:	PUNCT
ejpam-86	350	4	repeat	repeat	VERB
ejpam-86	350	5	steps	step	NOUN
ejpam-86	350	6	2	2	NUM
ejpam-86	350	7	-	-	SYM
ejpam-86	350	8	9	9	NUM
ejpam-86	350	9	until	until	SCONJ
ejpam-86	350	10	a	a	DET
ejpam-86	350	11	certain	certain	ADJ
ejpam-86	350	12	condition	condition	NOUN
ejpam-86	350	13	of	of	ADP
ejpam-86	350	14	the	the	DET
ejpam-86	350	15	result	result	NOUN
ejpam-86	350	16	is	be	AUX
ejpam-86	350	17	satisfied	satisfied	ADJ
ejpam-86	350	18	.	.	PUNCT
ejpam-86	351	1	the	the	DET
ejpam-86	351	2	ga	ga	PROPN
ejpam-86	351	3	for	for	ADP
ejpam-86	351	4	model	model	NOUN
ejpam-86	351	5	parameter	parameter	PROPN
ejpam-86	351	6	estimation	estimation	NOUN
ejpam-86	351	7	and	and	CCONJ
ejpam-86	351	8	the	the	DET
ejpam-86	351	9	ga	ga	PROPN
ejpam-86	351	10	for	for	ADP
ejpam-86	351	11	model	model	NOUN
ejpam-86	351	12	selection	selection	NOUN
ejpam-86	351	13	share	share	NOUN
ejpam-86	351	14	the	the	DET
ejpam-86	351	15	shame	shame	NOUN
ejpam-86	351	16	outline	outline	VERB
ejpam-86	351	17	above	above	ADV
ejpam-86	351	18	,	,	PUNCT
ejpam-86	351	19	but	but	CCONJ
ejpam-86	351	20	with	with	ADP
ejpam-86	351	21	different	different	ADJ
ejpam-86	351	22	objective	objective	ADJ
ejpam-86	351	23	functions	function	NOUN
ejpam-86	351	24	(	(	PUNCT
ejpam-86	351	25	fitness	fitness	NOUN
ejpam-86	351	26	functions	function	NOUN
ejpam-86	351	27	)	)	PUNCT
ejpam-86	351	28	and	and	CCONJ
ejpam-86	351	29	solution	solution	NOUN
ejpam-86	351	30	representations	representation	NOUN
ejpam-86	351	31	.	.	PUNCT
ejpam-86	352	1	this	this	PRON
ejpam-86	352	2	reflects	reflect	VERB
ejpam-86	352	3	one	one	NUM
ejpam-86	352	4	of	of	ADP
ejpam-86	352	5	the	the	DET
ejpam-86	352	6	significant	significant	ADJ
ejpam-86	352	7	advantages	advantage	NOUN
ejpam-86	352	8	of	of	ADP
ejpam-86	352	9	ga	ga	PROPN
ejpam-86	352	10	method	method	NOUN
ejpam-86	352	11	.	.	PUNCT
ejpam-86	353	1	once	once	ADV
ejpam-86	353	2	a	a	DET
ejpam-86	353	3	ga	ga	PROPN
ejpam-86	353	4	is	be	AUX
ejpam-86	353	5	set	set	VERB
ejpam-86	353	6	up	up	ADP
ejpam-86	353	7	,	,	PUNCT
ejpam-86	353	8	it	it	PRON
ejpam-86	353	9	can	can	AUX
ejpam-86	353	10	be	be	AUX
ejpam-86	353	11	expanded	expand	VERB
ejpam-86	353	12	to	to	PART
ejpam-86	353	13	solve	solve	VERB
ejpam-86	353	14	different	different	ADJ
ejpam-86	353	15	problems	problem	NOUN
ejpam-86	353	16	easily	easily	ADV
ejpam-86	353	17	by	by	ADP
ejpam-86	353	18	only	only	ADV
ejpam-86	353	19	changing	change	VERB
ejpam-86	353	20	the	the	DET
ejpam-86	353	21	fitness	fitness	NOUN
ejpam-86	353	22	function	function	NOUN
ejpam-86	353	23	and	and	CCONJ
ejpam-86	353	24	the	the	DET
ejpam-86	353	25	representation	representation	NOUN
ejpam-86	353	26	of	of	ADP
ejpam-86	353	27	solution	solution	NOUN
ejpam-86	353	28	space	space	NOUN
ejpam-86	353	29	.	.	PUNCT
ejpam-86	354	1	we	we	PRON
ejpam-86	354	2	will	will	AUX
ejpam-86	354	3	explain	explain	VERB
ejpam-86	354	4	the	the	DET
ejpam-86	354	5	fitness	fitness	NOUN
ejpam-86	354	6	function	function	NOUN
ejpam-86	354	7	and	and	CCONJ
ejpam-86	354	8	solution	solution	NOUN
ejpam-86	354	9	representation	representation	NOUN
ejpam-86	354	10	for	for	ADP
ejpam-86	354	11	model	model	NOUN
ejpam-86	354	12	parameter	parameter	PROPN
ejpam-86	354	13	estimation	estimation	NOUN
ejpam-86	354	14	and	and	CCONJ
ejpam-86	354	15	model	model	NOUN
ejpam-86	354	16	selection	selection	NOUN
ejpam-86	354	17	in	in	ADP
ejpam-86	354	18	following	follow	VERB
ejpam-86	354	19	sections	section	NOUN
ejpam-86	354	20	.	.	PUNCT
ejpam-86	355	1	the	the	DET
ejpam-86	355	2	ga	ga	PROPN
ejpam-86	355	3	engineering	engineering	NOUN
ejpam-86	355	4	(	(	PUNCT
ejpam-86	355	5	or	or	CCONJ
ejpam-86	355	6	ga	ga	NOUN
ejpam-86	355	7	cloning	cloning	NOUN
ejpam-86	355	8	)	)	PUNCT
ejpam-86	355	9	in	in	ADP
ejpam-86	355	10	step	step	NOUN
ejpam-86	355	11	7	7	NUM
ejpam-86	355	12	above	above	ADV
ejpam-86	355	13	is	be	AUX
ejpam-86	355	14	a	a	DET
ejpam-86	355	15	new	new	ADJ
ejpam-86	355	16	ga	ga	NOUN
ejpam-86	355	17	operator	operator	NOUN
ejpam-86	355	18	we	we	PRON
ejpam-86	355	19	developed	develop	VERB
ejpam-86	355	20	to	to	PART
ejpam-86	355	21	improve	improve	VERB
ejpam-86	355	22	the	the	DET
ejpam-86	355	23	evolution	evolution	NOUN
ejpam-86	355	24	of	of	ADP
ejpam-86	355	25	the	the	DET
ejpam-86	355	26	ga	ga	PROPN
ejpam-86	355	27	process	process	NOUN
ejpam-86	355	28	.	.	PUNCT
ejpam-86	356	1	this	this	PRON
ejpam-86	356	2	will	will	AUX
ejpam-86	356	3	be	be	AUX
ejpam-86	356	4	explained	explain	VERB
ejpam-86	356	5	in	in	ADP
ejpam-86	356	6	section	section	NOUN
ejpam-86	356	7	6.4	6.4	NUM
ejpam-86	356	8	.	.	PUNCT
ejpam-86	357	1	the	the	DET
ejpam-86	357	2	pseudo	pseudo	NOUN
ejpam-86	357	3	code	code	NOUN
ejpam-86	357	4	of	of	ADP
ejpam-86	357	5	the	the	DET
ejpam-86	357	6	steps	step	NOUN
ejpam-86	357	7	of	of	ADP
ejpam-86	357	8	the	the	DET
ejpam-86	357	9	ga	ga	PROPN
ejpam-86	357	10	outlined	outline	VERB
ejpam-86	357	11	above	above	ADV
ejpam-86	357	12	is	be	AUX
ejpam-86	357	13	given	give	VERB
ejpam-86	357	14	in	in	ADP
ejpam-86	357	15	the	the	DET
ejpam-86	357	16	appendix	appendix	NOUN
ejpam-86	357	17	.	.	PUNCT
ejpam-86	358	1	6.1	6.1	NUM
ejpam-86	358	2	.	.	PUNCT
ejpam-86	358	3	ga	ga	PROPN
ejpam-86	358	4	for	for	ADP
ejpam-86	358	5	model	model	NOUN
ejpam-86	358	6	parameter	parameter	NOUN
ejpam-86	358	7	estimation	estimation	NOUN
ejpam-86	358	8	we	we	PRON
ejpam-86	358	9	use	use	VERB
ejpam-86	358	10	ga	ga	PROPN
ejpam-86	358	11	to	to	PART
ejpam-86	358	12	estimate	estimate	VERB
ejpam-86	358	13	the	the	DET
ejpam-86	358	14	maximum	maximum	ADJ
ejpam-86	358	15	likelihood	likelihood	NOUN
ejpam-86	358	16	estimators	estimator	NOUN
ejpam-86	358	17	(	(	PUNCT
ejpam-86	358	18	mles	mle	NOUN
ejpam-86	358	19	)	)	PUNCT
ejpam-86	358	20	of	of	ADP
ejpam-86	358	21	the	the	DET
ejpam-86	358	22	multivariate	multivariate	NOUN
ejpam-86	358	23	regression	regression	NOUN
ejpam-86	358	24	model	model	NOUN
ejpam-86	358	25	parameters	parameter	NOUN
ejpam-86	358	26	;	;	PUNCT
ejpam-86	358	27	the	the	DET
ejpam-86	358	28	coefficients	coefficient	NOUN
ejpam-86	358	29	and	and	CCONJ
ejpam-86	358	30	the	the	DET
ejpam-86	358	31	estimated	estimate	VERB
ejpam-86	358	32	inverse	inverse	NOUN
ejpam-86	358	33	-	-	PUNCT
ejpam-86	358	34	fisher	fisher	PROPN
ejpam-86	358	35	information	information	NOUN
ejpam-86	358	36	matrix	matrix	NOUN
ejpam-86	358	37	(	(	PUNCT
ejpam-86	358	38	ifim	ifim	NOUN
ejpam-86	358	39	)	)	PUNCT
ejpam-86	358	40	of	of	ADP
ejpam-86	358	41	the	the	DET
ejpam-86	358	42	model	model	NOUN
ejpam-86	358	43	.	.	PUNCT
ejpam-86	359	1	we	we	PRON
ejpam-86	359	2	use	use	VERB
ejpam-86	359	3	the	the	DET
ejpam-86	359	4	negative	negative	ADJ
ejpam-86	359	5	log	log	NOUN
ejpam-86	359	6	likelihood	likelihood	NOUN
ejpam-86	359	7	function	function	NOUN
ejpam-86	359	8	as	as	ADP
ejpam-86	359	9	the	the	DET
ejpam-86	359	10	fitness	fitness	NOUN
ejpam-86	359	11	function	function	NOUN
ejpam-86	359	12	.	.	PUNCT
ejpam-86	360	1	with	with	ADP
ejpam-86	360	2	this	this	DET
ejpam-86	360	3	choice	choice	NOUN
ejpam-86	360	4	,	,	PUNCT
ejpam-86	360	5	the	the	DET
ejpam-86	360	6	best	good	ADJ
ejpam-86	360	7	solution	solution	NOUN
ejpam-86	360	8	is	be	AUX
ejpam-86	360	9	the	the	DET
ejpam-86	360	10	minimum	minimum	ADJ
ejpam-86	360	11	fitness	fitness	NOUN
ejpam-86	360	12	value	value	NOUN
ejpam-86	360	13	.	.	PUNCT
ejpam-86	361	1	we	we	PRON
ejpam-86	361	2	encode	encode	VERB
ejpam-86	361	3	each	each	DET
ejpam-86	361	4	model	model	NOUN
ejpam-86	361	5	parameter	parameter	NOUN
ejpam-86	361	6	to	to	PART
ejpam-86	361	7	be	be	AUX
ejpam-86	361	8	estimated	estimate	VERB
ejpam-86	361	9	in	in	ADP
ejpam-86	361	10	a	a	DET
ejpam-86	361	11	binary	binary	ADJ
ejpam-86	361	12	string	string	NOUN
ejpam-86	361	13	of	of	ADP
ejpam-86	361	14	fixed	fix	VERB
ejpam-86	361	15	length	length	NOUN
ejpam-86	361	16	,	,	PUNCT
ejpam-86	361	17	which	which	PRON
ejpam-86	361	18	is	be	AUX
ejpam-86	361	19	given	give	VERB
ejpam-86	361	20	by	by	ADP
ejpam-86	361	21	the	the	DET
ejpam-86	361	22	investigator	investigator	NOUN
ejpam-86	361	23	as	as	ADP
ejpam-86	361	24	an	an	DET
ejpam-86	361	25	input	input	NOUN
ejpam-86	361	26	.	.	PUNCT
ejpam-86	362	1	since	since	SCONJ
ejpam-86	362	2	the	the	DET
ejpam-86	362	3	parameters	parameter	NOUN
ejpam-86	362	4	are	be	AUX
ejpam-86	362	5	real	real	ADJ
ejpam-86	362	6	numbers	number	NOUN
ejpam-86	362	7	,	,	PUNCT
ejpam-86	362	8	we	we	PRON
ejpam-86	362	9	use	use	VERB
ejpam-86	362	10	the	the	DET
ejpam-86	362	11	following	follow	VERB
ejpam-86	362	12	scheme	scheme	NOUN
ejpam-86	362	13	to	to	PART
ejpam-86	362	14	encode	encode	VERB
ejpam-86	362	15	them	they	PRON
ejpam-86	362	16	.	.	PUNCT
ejpam-86	363	1	given	give	VERB
ejpam-86	363	2	a	a	DET
ejpam-86	363	3	real	real	ADJ
ejpam-86	363	4	interval	interval	NOUN
ejpam-86	363	5	[	[	X
ejpam-86	363	6	a	a	X
ejpam-86	363	7	,	,	PUNCT
ejpam-86	363	8	b	b	NOUN
ejpam-86	363	9	]	]	PUNCT
ejpam-86	363	10	and	and	CCONJ
ejpam-86	363	11	the	the	DET
ejpam-86	363	12	length	length	NOUN
ejpam-86	363	13	of	of	ADP
ejpam-86	363	14	binary	binary	PROPN
ejpam-86	363	15	string	string	PROPN
ejpam-86	363	16	l	l	PROPN
ejpam-86	363	17	,	,	PUNCT
ejpam-86	363	18	the	the	DET
ejpam-86	363	19	binary	binary	PROPN
ejpam-86	363	20	string	string	NOUN
ejpam-86	363	21	000	000	NUM
ejpam-86	363	22	·	·	PUNCT
ejpam-86	363	23	·	·	PUNCT
ejpam-86	363	24	·	·	PUNCT
ejpam-86	363	25	000	000	NUM
ejpam-86	363	26	represents	represent	VERB
ejpam-86	363	27	a	a	PRON
ejpam-86	363	28	and	and	CCONJ
ejpam-86	363	29	111	111	NUM
ejpam-86	363	30	·	·	PUNCT
ejpam-86	363	31	·	·	PUNCT
ejpam-86	363	32	·	·	PUNCT
ejpam-86	363	33	111	111	NUM
ejpam-86	363	34	represents	represent	VERB
ejpam-86	363	35	b.	b.	NOUN
ejpam-86	363	36	adding	add	VERB
ejpam-86	363	37	binary	binary	NOUN
ejpam-86	363	38	1	1	NUM
ejpam-86	363	39	to	to	ADP
ejpam-86	363	40	an	an	DET
ejpam-86	363	41	existing	exist	VERB
ejpam-86	363	42	binary	binary	ADJ
ejpam-86	363	43	number	number	NOUN
ejpam-86	363	44	increases	increase	VERB
ejpam-86	363	45	its	its	PRON
ejpam-86	363	46	real	real	ADJ
ejpam-86	363	47	value	value	NOUN
ejpam-86	363	48	by	by	ADP
ejpam-86	363	49	(	(	PUNCT
ejpam-86	363	50	b−a)/2l−1	b−a)/2l−1	NOUN
ejpam-86	363	51	.	.	PUNCT
ejpam-86	364	1	with	with	ADP
ejpam-86	364	2	this	this	DET
ejpam-86	364	3	approach	approach	NOUN
ejpam-86	364	4	,	,	PUNCT
ejpam-86	364	5	decoding	decode	VERB
ejpam-86	364	6	a	a	DET
ejpam-86	364	7	binary	binary	ADJ
ejpam-86	364	8	string	string	NOUN
ejpam-86	364	9	to	to	PART
ejpam-86	364	10	real	real	ADJ
ejpam-86	364	11	number	number	NOUN
ejpam-86	364	12	is	be	AUX
ejpam-86	364	13	easy	easy	ADJ
ejpam-86	364	14	.	.	PUNCT
ejpam-86	365	1	for	for	ADP
ejpam-86	365	2	example	example	NOUN
ejpam-86	365	3	,	,	PUNCT
ejpam-86	365	4	if	if	SCONJ
ejpam-86	365	5	l	l	NOUN
ejpam-86	365	6	=	=	SYM
ejpam-86	365	7	5	5	NUM
ejpam-86	365	8	,	,	PUNCT
ejpam-86	365	9	then	then	ADV
ejpam-86	365	10	10010	10010	NUM
ejpam-86	365	11	represents	represent	VERB
ejpam-86	365	12	the	the	DET
ejpam-86	365	13	real	real	ADJ
ejpam-86	365	14	value	value	NOUN
ejpam-86	365	15	a+(b−a)×	a+(b−a)×	NOUN
ejpam-86	365	16	18/31	18/31	NUM
ejpam-86	365	17	=	=	SYM
ejpam-86	365	18	(	(	PUNCT
ejpam-86	365	19	13a+	13a+	NUM
ejpam-86	365	20	18b)/31	18b)/31	NUM
ejpam-86	365	21	.	.	PUNCT
ejpam-86	366	1	with	with	ADP
ejpam-86	366	2	the	the	DET
ejpam-86	366	3	above	above	ADJ
ejpam-86	366	4	encoding	encoding	NOUN
ejpam-86	366	5	approach	approach	NOUN
ejpam-86	366	6	,	,	PUNCT
ejpam-86	366	7	we	we	PRON
ejpam-86	366	8	first	first	ADV
ejpam-86	366	9	obtain	obtain	VERB
ejpam-86	366	10	the	the	DET
ejpam-86	366	11	starting	starting	NOUN
ejpam-86	366	12	point	point	NOUN
ejpam-86	366	13	and	and	CCONJ
ejpam-86	366	14	search	search	VERB
ejpam-86	366	15	the	the	DET
ejpam-86	366	16	interval	interval	NOUN
ejpam-86	366	17	with	with	ADP
ejpam-86	366	18	other	other	ADJ
ejpam-86	366	19	methods	method	NOUN
ejpam-86	366	20	,	,	PUNCT
ejpam-86	366	21	such	such	ADJ
ejpam-86	366	22	as	as	ADP
ejpam-86	366	23	the	the	DET
ejpam-86	366	24	method	method	NOUN
ejpam-86	366	25	of	of	ADP
ejpam-86	366	26	moments	moment	NOUN
ejpam-86	366	27	(	(	PUNCT
ejpam-86	366	28	mom	mom	NOUN
ejpam-86	366	29	)	)	PUNCT
ejpam-86	366	30	.	.	PUNCT
ejpam-86	367	1	then	then	ADV
ejpam-86	367	2	,	,	PUNCT
ejpam-86	367	3	we	we	PRON
ejpam-86	367	4	use	use	VERB
ejpam-86	367	5	the	the	DET
ejpam-86	367	6	ga	ga	PROPN
ejpam-86	367	7	to	to	PART
ejpam-86	367	8	estimate	estimate	VERB
ejpam-86	367	9	the	the	DET
ejpam-86	367	10	mles	mle	NOUN
ejpam-86	367	11	.	.	PUNCT
ejpam-86	368	1	m.	m.	NOUN
ejpam-86	368	2	liu	liu	PROPN
ejpam-86	368	3	and	and	CCONJ
ejpam-86	368	4	h.	h.	PROPN
ejpam-86	368	5	bozdogan	bozdogan	PROPN
ejpam-86	368	6	/	/	SYM
ejpam-86	368	7	eur	eur	PROPN
ejpam-86	368	8	.	.	PUNCT
ejpam-86	369	1	j.	j.	PROPN
ejpam-86	369	2	pure	pure	PROPN
ejpam-86	369	3	appl	appl	PROPN
ejpam-86	369	4	.	.	PROPN
ejpam-86	369	5	math	math	PROPN
ejpam-86	369	6	,	,	PUNCT
ejpam-86	369	7	1	1	NUM
ejpam-86	369	8	(	(	PUNCT
ejpam-86	369	9	2008	2008	NUM
ejpam-86	369	10	)	)	PUNCT
ejpam-86	369	11	,	,	PUNCT
ejpam-86	369	12	(	(	PUNCT
ejpam-86	369	13	4	4	NUM
ejpam-86	369	14	-	-	SYM
ejpam-86	369	15	37	37	NUM
ejpam-86	369	16	)	)	PUNCT
ejpam-86	369	17	19	19	NUM
ejpam-86	369	18	6.2	6.2	NUM
ejpam-86	369	19	.	.	PUNCT
ejpam-86	370	1	ga	ga	PROPN
ejpam-86	370	2	for	for	ADP
ejpam-86	370	3	model	model	NOUN
ejpam-86	370	4	selection	selection	NOUN
ejpam-86	370	5	in	in	ADP
ejpam-86	370	6	the	the	DET
ejpam-86	370	7	ga	ga	PROPN
ejpam-86	370	8	for	for	ADP
ejpam-86	370	9	model	model	NOUN
ejpam-86	370	10	selection	selection	NOUN
ejpam-86	370	11	,	,	PUNCT
ejpam-86	370	12	we	we	PRON
ejpam-86	370	13	use	use	VERB
ejpam-86	370	14	the	the	DET
ejpam-86	370	15	icomp(ifim	icomp(ifim	NOUN
ejpam-86	370	16	)	)	PUNCT
ejpam-86	370	17	as	as	ADP
ejpam-86	370	18	the	the	DET
ejpam-86	370	19	fitness	fitness	NOUN
ejpam-86	370	20	function	function	NOUN
ejpam-86	370	21	.	.	PUNCT
ejpam-86	371	1	the	the	DET
ejpam-86	371	2	algorithm	algorithm	NOUN
ejpam-86	371	3	can	can	AUX
ejpam-86	371	4	also	also	ADV
ejpam-86	371	5	be	be	AUX
ejpam-86	371	6	easily	easily	ADV
ejpam-86	371	7	edited	edit	VERB
ejpam-86	371	8	to	to	PART
ejpam-86	371	9	use	use	VERB
ejpam-86	371	10	other	other	ADJ
ejpam-86	371	11	model	model	NOUN
ejpam-86	371	12	selection	selection	NOUN
ejpam-86	371	13	criteria	criterion	NOUN
ejpam-86	371	14	.	.	PUNCT
ejpam-86	372	1	we	we	PRON
ejpam-86	372	2	encode	encode	VERB
ejpam-86	372	3	each	each	DET
ejpam-86	372	4	model	model	NOUN
ejpam-86	372	5	using	use	VERB
ejpam-86	372	6	the	the	DET
ejpam-86	372	7	following	follow	VERB
ejpam-86	372	8	scheme	scheme	NOUN
ejpam-86	372	9	.	.	PUNCT
ejpam-86	373	1	each	each	DET
ejpam-86	373	2	model	model	NOUN
ejpam-86	373	3	,	,	PUNCT
ejpam-86	373	4	or	or	CCONJ
ejpam-86	373	5	subset	subset	NOUN
ejpam-86	373	6	,	,	PUNCT
ejpam-86	373	7	is	be	AUX
ejpam-86	373	8	encoded	encode	VERB
ejpam-86	373	9	as	as	ADP
ejpam-86	373	10	a	a	DET
ejpam-86	373	11	binary	binary	ADJ
ejpam-86	373	12	string	string	NOUN
ejpam-86	373	13	with	with	ADP
ejpam-86	373	14	a	a	DET
ejpam-86	373	15	fixed	fix	VERB
ejpam-86	373	16	length	length	NOUN
ejpam-86	373	17	based	base	VERB
ejpam-86	373	18	on	on	ADP
ejpam-86	373	19	the	the	DET
ejpam-86	373	20	number	number	NOUN
ejpam-86	373	21	of	of	ADP
ejpam-86	373	22	total	total	ADJ
ejpam-86	373	23	available	available	ADJ
ejpam-86	373	24	independent	independent	ADJ
ejpam-86	373	25	variables	variable	NOUN
ejpam-86	373	26	(	(	PUNCT
ejpam-86	373	27	including	include	VERB
ejpam-86	373	28	the	the	DET
ejpam-86	373	29	constant	constant	ADJ
ejpam-86	373	30	term	term	NOUN
ejpam-86	373	31	)	)	PUNCT
ejpam-86	373	32	.	.	PUNCT
ejpam-86	374	1	each	each	DET
ejpam-86	374	2	bit	bit	NOUN
ejpam-86	374	3	in	in	ADP
ejpam-86	374	4	the	the	DET
ejpam-86	374	5	string	string	NOUN
ejpam-86	374	6	is	be	AUX
ejpam-86	374	7	a	a	DET
ejpam-86	374	8	binary	binary	ADJ
ejpam-86	374	9	code	code	NOUN
ejpam-86	374	10	indicating	indicate	VERB
ejpam-86	374	11	the	the	DET
ejpam-86	374	12	presence	presence	NOUN
ejpam-86	374	13	(	(	PUNCT
ejpam-86	374	14	1	1	NUM
ejpam-86	374	15	)	)	PUNCT
ejpam-86	374	16	or	or	CCONJ
ejpam-86	374	17	absence	absence	NOUN
ejpam-86	374	18	(	(	PUNCT
ejpam-86	374	19	0	0	NUM
ejpam-86	374	20	)	)	PUNCT
ejpam-86	374	21	of	of	ADP
ejpam-86	374	22	a	a	DET
ejpam-86	374	23	given	give	VERB
ejpam-86	374	24	predictor	predictor	NOUN
ejpam-86	374	25	variable	variable	NOUN
ejpam-86	374	26	in	in	ADP
ejpam-86	374	27	the	the	DET
ejpam-86	374	28	model	model	NOUN
ejpam-86	374	29	.	.	PUNCT
ejpam-86	375	1	for	for	ADP
ejpam-86	375	2	example	example	NOUN
ejpam-86	375	3	,	,	PUNCT
ejpam-86	375	4	if	if	SCONJ
ejpam-86	375	5	there	there	PRON
ejpam-86	375	6	are	be	VERB
ejpam-86	375	7	10	10	NUM
ejpam-86	375	8	predictor	predictor	NOUN
ejpam-86	375	9	variables	variable	NOUN
ejpam-86	375	10	available	available	ADJ
ejpam-86	375	11	in	in	ADP
ejpam-86	375	12	a	a	DET
ejpam-86	375	13	given	give	VERB
ejpam-86	375	14	data	datum	NOUN
ejpam-86	375	15	set	set	NOUN
ejpam-86	375	16	,	,	PUNCT
ejpam-86	375	17	then	then	ADV
ejpam-86	375	18	the	the	DET
ejpam-86	375	19	string	string	NOUN
ejpam-86	375	20	1010110110	1010110110	NUM
ejpam-86	375	21	represents	represent	VERB
ejpam-86	375	22	a	a	DET
ejpam-86	375	23	model	model	NOUN
ejpam-86	375	24	,	,	PUNCT
ejpam-86	375	25	where	where	SCONJ
ejpam-86	375	26	constant	constant	ADJ
ejpam-86	375	27	term	term	NOUN
ejpam-86	375	28	is	be	AUX
ejpam-86	375	29	included	include	VERB
ejpam-86	375	30	in	in	ADP
ejpam-86	375	31	the	the	DET
ejpam-86	375	32	model	model	NOUN
ejpam-86	375	33	,	,	PUNCT
ejpam-86	375	34	variable	variable	ADJ
ejpam-86	375	35	1	1	NUM
ejpam-86	375	36	is	be	AUX
ejpam-86	375	37	excluded	exclude	VERB
ejpam-86	375	38	from	from	ADP
ejpam-86	375	39	the	the	DET
ejpam-86	375	40	model	model	NOUN
ejpam-86	375	41	,	,	PUNCT
ejpam-86	375	42	variable	variable	ADJ
ejpam-86	375	43	2	2	NUM
ejpam-86	375	44	is	be	AUX
ejpam-86	375	45	included	include	VERB
ejpam-86	375	46	in	in	ADP
ejpam-86	375	47	the	the	DET
ejpam-86	375	48	model	model	NOUN
ejpam-86	375	49	,	,	PUNCT
ejpam-86	375	50	and	and	CCONJ
ejpam-86	375	51	so	so	ADV
ejpam-86	375	52	on	on	ADV
ejpam-86	375	53	.	.	PUNCT
ejpam-86	376	1	6.3	6.3	NUM
ejpam-86	376	2	.	.	PUNCT
ejpam-86	376	3	ga	ga	PROPN
ejpam-86	376	4	on	on	ADP
ejpam-86	376	5	ga	ga	PROPN
ejpam-86	376	6	hybridization	hybridization	NOUN
ejpam-86	376	7	we	we	PRON
ejpam-86	376	8	combine	combine	VERB
ejpam-86	376	9	and	and	CCONJ
ejpam-86	376	10	hybridize	hybridize	VERB
ejpam-86	376	11	the	the	DET
ejpam-86	376	12	ga	ga	PROPN
ejpam-86	376	13	for	for	ADP
ejpam-86	376	14	parameter	parameter	NOUN
ejpam-86	376	15	estimation	estimation	NOUN
ejpam-86	376	16	and	and	CCONJ
ejpam-86	376	17	the	the	DET
ejpam-86	376	18	ga	ga	PROPN
ejpam-86	376	19	for	for	ADP
ejpam-86	376	20	model	model	NOUN
ejpam-86	376	21	selection	selection	NOUN
ejpam-86	376	22	as	as	SCONJ
ejpam-86	376	23	follows	follow	VERB
ejpam-86	376	24	:	:	PUNCT
ejpam-86	376	25	•	•	NUM
ejpam-86	376	26	first	first	ADV
ejpam-86	376	27	,	,	PUNCT
ejpam-86	376	28	the	the	DET
ejpam-86	376	29	ga	ga	PROPN
ejpam-86	376	30	for	for	ADP
ejpam-86	376	31	model	model	NOUN
ejpam-86	376	32	selection	selection	NOUN
ejpam-86	376	33	is	be	AUX
ejpam-86	376	34	called	call	VERB
ejpam-86	376	35	to	to	PART
ejpam-86	376	36	select	select	VERB
ejpam-86	376	37	the	the	DET
ejpam-86	376	38	best	good	ADJ
ejpam-86	376	39	subset	subset	NOUN
ejpam-86	376	40	of	of	ADP
ejpam-86	376	41	predictor	predictor	NOUN
ejpam-86	376	42	variables	variable	NOUN
ejpam-86	376	43	.	.	PUNCT
ejpam-86	377	1	•	•	NOUN
ejpam-86	377	2	at	at	ADP
ejpam-86	377	3	each	each	DET
ejpam-86	377	4	step	step	NOUN
ejpam-86	377	5	a	a	DET
ejpam-86	377	6	model	model	NOUN
ejpam-86	377	7	is	be	AUX
ejpam-86	377	8	chosen	choose	VERB
ejpam-86	377	9	to	to	PART
ejpam-86	377	10	be	be	AUX
ejpam-86	377	11	evaluated	evaluate	VERB
ejpam-86	377	12	.	.	PUNCT
ejpam-86	378	1	then	then	ADV
ejpam-86	378	2	the	the	DET
ejpam-86	378	3	ga	ga	PROPN
ejpam-86	378	4	for	for	ADP
ejpam-86	378	5	model	model	NOUN
ejpam-86	378	6	parameter	parameter	PROPN
ejpam-86	378	7	estimation	estimation	NOUN
ejpam-86	378	8	is	be	AUX
ejpam-86	378	9	called	call	VERB
ejpam-86	378	10	to	to	PART
ejpam-86	378	11	obtain	obtain	VERB
ejpam-86	378	12	the	the	DET
ejpam-86	378	13	mles	mle	NOUN
ejpam-86	378	14	of	of	ADP
ejpam-86	378	15	the	the	DET
ejpam-86	378	16	model	model	NOUN
ejpam-86	378	17	parameters	parameter	NOUN
ejpam-86	378	18	,	,	PUNCT
ejpam-86	378	19	or	or	CCONJ
ejpam-86	378	20	the	the	DET
ejpam-86	378	21	mles	mle	NOUN
ejpam-86	378	22	are	be	AUX
ejpam-86	378	23	retrieved	retrieve	VERB
ejpam-86	378	24	if	if	SCONJ
ejpam-86	378	25	the	the	DET
ejpam-86	378	26	model	model	NOUN
ejpam-86	378	27	has	have	AUX
ejpam-86	378	28	been	be	AUX
ejpam-86	378	29	evaluated	evaluate	VERB
ejpam-86	378	30	before	before	ADV
ejpam-86	378	31	.	.	PUNCT
ejpam-86	379	1	the	the	DET
ejpam-86	379	2	fitness	fitness	NOUN
ejpam-86	379	3	,	,	PUNCT
ejpam-86	379	4	i.e.	i.e.	X
ejpam-86	379	5	,	,	PUNCT
ejpam-86	379	6	icomp(ifim	icomp(ifim	NOUN
ejpam-86	379	7	)	)	PUNCT
ejpam-86	379	8	of	of	ADP
ejpam-86	379	9	the	the	DET
ejpam-86	379	10	model	model	NOUN
ejpam-86	379	11	is	be	AUX
ejpam-86	379	12	computed	compute	VERB
ejpam-86	379	13	using	use	VERB
ejpam-86	379	14	the	the	DET
ejpam-86	379	15	mles	mle	NOUN
ejpam-86	379	16	of	of	ADP
ejpam-86	379	17	the	the	DET
ejpam-86	379	18	model	model	NOUN
ejpam-86	379	19	parameters	parameter	NOUN
ejpam-86	379	20	.	.	PUNCT
ejpam-86	380	1	with	with	ADP
ejpam-86	380	2	the	the	DET
ejpam-86	380	3	ga	ga	PROPN
ejpam-86	380	4	on	on	ADP
ejpam-86	380	5	ga	ga	PROPN
ejpam-86	380	6	approach	approach	NOUN
ejpam-86	380	7	,	,	PUNCT
ejpam-86	380	8	the	the	DET
ejpam-86	380	9	output	output	NOUN
ejpam-86	380	10	of	of	ADP
ejpam-86	380	11	the	the	DET
ejpam-86	380	12	ga	ga	PROPN
ejpam-86	380	13	for	for	ADP
ejpam-86	380	14	estimation	estimation	NOUN
ejpam-86	380	15	is	be	AUX
ejpam-86	380	16	used	use	VERB
ejpam-86	380	17	in	in	ADP
ejpam-86	380	18	the	the	DET
ejpam-86	380	19	fitness	fitness	NOUN
ejpam-86	380	20	function	function	NOUN
ejpam-86	380	21	(	(	PUNCT
ejpam-86	380	22	icomp	icomp	PROPN
ejpam-86	380	23	)	)	PUNCT
ejpam-86	380	24	to	to	PART
ejpam-86	380	25	evaluate	evaluate	VERB
ejpam-86	380	26	the	the	DET
ejpam-86	380	27	best	good	ADJ
ejpam-86	380	28	subset	subset	ADJ
ejpam-86	380	29	candidate	candidate	NOUN
ejpam-86	380	30	models	model	NOUN
ejpam-86	380	31	.	.	PUNCT
ejpam-86	381	1	the	the	DET
ejpam-86	381	2	estimation	estimation	NOUN
ejpam-86	381	3	in	in	ADP
ejpam-86	381	4	fact	fact	NOUN
ejpam-86	381	5	initially	initially	ADV
ejpam-86	381	6	acts	act	VERB
ejpam-86	381	7	as	as	ADP
ejpam-86	381	8	the	the	DET
ejpam-86	381	9	fitness	fitness	NOUN
ejpam-86	381	10	function	function	NOUN
ejpam-86	381	11	of	of	ADP
ejpam-86	381	12	the	the	DET
ejpam-86	381	13	model	model	NOUN
ejpam-86	381	14	selection	selection	NOUN
ejpam-86	381	15	in	in	ADP
ejpam-86	381	16	ga	ga	PROPN
ejpam-86	381	17	.	.	PUNCT
ejpam-86	382	1	the	the	DET
ejpam-86	382	2	ga	ga	PROPN
ejpam-86	382	3	on	on	ADP
ejpam-86	382	4	ga	ga	PROPN
ejpam-86	382	5	approach	approach	NOUN
ejpam-86	382	6	inherits	inherit	VERB
ejpam-86	382	7	both	both	DET
ejpam-86	382	8	advantages	advantage	NOUN
ejpam-86	382	9	and	and	CCONJ
ejpam-86	382	10	disadvantages	disadvantage	NOUN
ejpam-86	382	11	of	of	ADP
ejpam-86	382	12	the	the	DET
ejpam-86	382	13	general	general	PROPN
ejpam-86	382	14	ga	ga	PROPN
ejpam-86	382	15	.	.	PROPN
ejpam-86	383	1	but	but	CCONJ
ejpam-86	383	2	the	the	DET
ejpam-86	383	3	major	major	ADJ
ejpam-86	383	4	convenience	convenience	NOUN
ejpam-86	383	5	of	of	ADP
ejpam-86	383	6	the	the	DET
ejpam-86	383	7	ga	ga	PROPN
ejpam-86	383	8	on	on	ADP
ejpam-86	383	9	ga	ga	PROPN
ejpam-86	383	10	approach	approach	NOUN
ejpam-86	383	11	is	be	AUX
ejpam-86	383	12	that	that	SCONJ
ejpam-86	383	13	both	both	CCONJ
ejpam-86	383	14	the	the	DET
ejpam-86	383	15	estimation	estimation	NOUN
ejpam-86	383	16	and	and	CCONJ
ejpam-86	383	17	model	model	NOUN
ejpam-86	383	18	selection	selection	NOUN
ejpam-86	383	19	in	in	ADP
ejpam-86	383	20	ga	ga	PROPN
ejpam-86	383	21	can	can	AUX
ejpam-86	383	22	share	share	VERB
ejpam-86	383	23	the	the	DET
ejpam-86	383	24	same	same	ADJ
ejpam-86	383	25	ga	ga	NOUN
ejpam-86	383	26	procedure	procedure	NOUN
ejpam-86	383	27	and	and	CCONJ
ejpam-86	383	28	code	code	NOUN
ejpam-86	383	29	.	.	PUNCT
ejpam-86	384	1	only	only	ADV
ejpam-86	384	2	the	the	DET
ejpam-86	384	3	fitness	fitness	NOUN
ejpam-86	384	4	functions	function	NOUN
ejpam-86	384	5	and	and	CCONJ
ejpam-86	384	6	representations	representation	NOUN
ejpam-86	384	7	of	of	ADP
ejpam-86	384	8	the	the	DET
ejpam-86	384	9	solutions	solution	NOUN
ejpam-86	384	10	need	need	VERB
ejpam-86	384	11	to	to	PART
ejpam-86	384	12	be	be	AUX
ejpam-86	384	13	changed	change	VERB
ejpam-86	384	14	correspondingly	correspondingly	ADV
ejpam-86	384	15	.	.	PUNCT
ejpam-86	385	1	the	the	DET
ejpam-86	385	2	disadvantage	disadvantage	NOUN
ejpam-86	385	3	of	of	ADP
ejpam-86	385	4	the	the	DET
ejpam-86	385	5	ga	ga	PROPN
ejpam-86	385	6	on	on	ADP
ejpam-86	385	7	ga	ga	PROPN
ejpam-86	385	8	is	be	AUX
ejpam-86	385	9	that	that	SCONJ
ejpam-86	385	10	,	,	PUNCT
ejpam-86	385	11	since	since	SCONJ
ejpam-86	385	12	the	the	DET
ejpam-86	385	13	output	output	NOUN
ejpam-86	385	14	of	of	ADP
ejpam-86	385	15	the	the	DET
ejpam-86	385	16	ga	ga	PROPN
ejpam-86	385	17	for	for	ADP
ejpam-86	385	18	model	model	NOUN
ejpam-86	385	19	parameter	parameter	PROPN
ejpam-86	385	20	estimation	estimation	NOUN
ejpam-86	385	21	is	be	AUX
ejpam-86	385	22	random	random	ADJ
ejpam-86	385	23	,	,	PUNCT
ejpam-86	385	24	the	the	DET
ejpam-86	385	25	fitness	fitness	NOUN
ejpam-86	385	26	function	function	NOUN
ejpam-86	385	27	of	of	ADP
ejpam-86	385	28	the	the	DET
ejpam-86	385	29	ga	ga	PROPN
ejpam-86	385	30	for	for	ADP
ejpam-86	385	31	model	model	NOUN
ejpam-86	385	32	selection	selection	NOUN
ejpam-86	385	33	is	be	AUX
ejpam-86	385	34	also	also	ADV
ejpam-86	385	35	random	random	ADJ
ejpam-86	385	36	.	.	PUNCT
ejpam-86	386	1	in	in	ADP
ejpam-86	386	2	this	this	DET
ejpam-86	386	3	case	case	NOUN
ejpam-86	386	4	,	,	PUNCT
ejpam-86	386	5	the	the	DET
ejpam-86	386	6	evaluation	evaluation	NOUN
ejpam-86	386	7	of	of	ADP
ejpam-86	386	8	models	model	NOUN
ejpam-86	386	9	will	will	AUX
ejpam-86	386	10	be	be	AUX
ejpam-86	386	11	inconsistent	inconsistent	ADJ
ejpam-86	386	12	during	during	ADP
ejpam-86	386	13	the	the	DET
ejpam-86	386	14	model	model	NOUN
ejpam-86	386	15	selection	selection	PROPN
ejpam-86	386	16	ga	ga	PROPN
ejpam-86	386	17	process	process	NOUN
ejpam-86	386	18	.	.	PUNCT
ejpam-86	387	1	for	for	ADP
ejpam-86	387	2	example	example	NOUN
ejpam-86	387	3	,	,	PUNCT
ejpam-86	387	4	in	in	ADP
ejpam-86	387	5	generation	generation	NOUN
ejpam-86	387	6	i	i	PRON
ejpam-86	387	7	,	,	PUNCT
ejpam-86	387	8	if	if	SCONJ
ejpam-86	387	9	model	model	NOUN
ejpam-86	387	10	a	a	PRON
ejpam-86	387	11	is	be	AUX
ejpam-86	387	12	better	well	ADJ
ejpam-86	387	13	than	than	ADP
ejpam-86	387	14	model	model	NOUN
ejpam-86	387	15	b	b	PROPN
ejpam-86	387	16	according	accord	VERB
ejpam-86	387	17	to	to	ADP
ejpam-86	387	18	their	their	PRON
ejpam-86	387	19	fitness	fitness	NOUN
ejpam-86	387	20	values	value	NOUN
ejpam-86	387	21	,	,	PUNCT
ejpam-86	387	22	but	but	CCONJ
ejpam-86	387	23	in	in	ADP
ejpam-86	387	24	generation	generation	NOUN
ejpam-86	387	25	j	j	PROPN
ejpam-86	387	26	,	,	PUNCT
ejpam-86	387	27	model	model	PROPN
ejpam-86	387	28	b	b	PROPN
ejpam-86	387	29	can	can	AUX
ejpam-86	387	30	be	be	AUX
ejpam-86	387	31	better	well	ADJ
ejpam-86	387	32	than	than	ADP
ejpam-86	387	33	model	model	VERB
ejpam-86	387	34	a	a	PRON
ejpam-86	387	35	since	since	SCONJ
ejpam-86	387	36	their	their	PRON
ejpam-86	387	37	fitness	fitness	NOUN
ejpam-86	387	38	changes	change	NOUN
ejpam-86	387	39	.	.	PUNCT
ejpam-86	388	1	to	to	PART
ejpam-86	388	2	solve	solve	VERB
ejpam-86	388	3	this	this	DET
ejpam-86	388	4	problem	problem	NOUN
ejpam-86	388	5	,	,	PUNCT
ejpam-86	388	6	we	we	PRON
ejpam-86	388	7	remember	remember	VERB
ejpam-86	388	8	the	the	DET
ejpam-86	388	9	fitness	fitness	NOUN
ejpam-86	388	10	and	and	CCONJ
ejpam-86	388	11	parameter	parameter	NOUN
ejpam-86	388	12	estimation	estimation	NOUN
ejpam-86	388	13	of	of	ADP
ejpam-86	388	14	all	all	DET
ejpam-86	388	15	evaluated	evaluated	ADJ
ejpam-86	388	16	models	model	NOUN
ejpam-86	388	17	and	and	CCONJ
ejpam-86	388	18	retrieve	retrieve	VERB
ejpam-86	388	19	them	they	PRON
ejpam-86	388	20	when	when	SCONJ
ejpam-86	388	21	they	they	PRON
ejpam-86	388	22	are	be	AUX
ejpam-86	388	23	needed	need	VERB
ejpam-86	388	24	to	to	PART
ejpam-86	388	25	keep	keep	VERB
ejpam-86	388	26	the	the	DET
ejpam-86	388	27	evaluation	evaluation	NOUN
ejpam-86	388	28	process	process	NOUN
ejpam-86	388	29	to	to	PART
ejpam-86	388	30	be	be	AUX
ejpam-86	388	31	consistent	consistent	ADJ
ejpam-86	388	32	.	.	PUNCT
ejpam-86	389	1	our	our	PRON
ejpam-86	389	2	simulation	simulation	NOUN
ejpam-86	389	3	results	result	NOUN
ejpam-86	389	4	show	show	VERB
ejpam-86	389	5	that	that	SCONJ
ejpam-86	389	6	this	this	DET
ejpam-86	389	7	approach	approach	NOUN
ejpam-86	389	8	is	be	AUX
ejpam-86	389	9	practical	practical	ADJ
ejpam-86	389	10	with	with	ADP
ejpam-86	389	11	ga	ga	PROPN
ejpam-86	389	12	properly	properly	ADV
ejpam-86	389	13	set	set	VERB
ejpam-86	389	14	up	up	ADP
ejpam-86	389	15	and	and	CCONJ
ejpam-86	389	16	engineered	engineer	VERB
ejpam-86	389	17	(	(	PUNCT
ejpam-86	389	18	or	or	CCONJ
ejpam-86	389	19	cloned	clone	VERB
ejpam-86	389	20	)	)	PUNCT
ejpam-86	389	21	.	.	PUNCT
ejpam-86	390	1	we	we	PRON
ejpam-86	390	2	further	far	ADV
ejpam-86	390	3	note	note	VERB
ejpam-86	390	4	that	that	SCONJ
ejpam-86	390	5	the	the	DET
ejpam-86	390	6	ga	ga	PROPN
ejpam-86	390	7	on	on	ADP
ejpam-86	390	8	ga	ga	PROPN
ejpam-86	390	9	approach	approach	NOUN
ejpam-86	390	10	improves	improve	VERB
ejpam-86	390	11	the	the	DET
ejpam-86	390	12	computational	computational	ADJ
ejpam-86	390	13	efficiency	efficiency	NOUN
ejpam-86	390	14	by	by	ADP
ejpam-86	390	15	eliminating	eliminate	VERB
ejpam-86	390	16	repeated	repeat	VERB
ejpam-86	390	17	model	model	NOUN
ejpam-86	390	18	parameter	parameter	NOUN
ejpam-86	390	19	estimation	estimation	NOUN
ejpam-86	390	20	process	process	NOUN
ejpam-86	390	21	when	when	SCONJ
ejpam-86	390	22	we	we	PRON
ejpam-86	390	23	evaluate	evaluate	VERB
ejpam-86	390	24	and	and	CCONJ
ejpam-86	390	25	fit	fit	VERB
ejpam-86	390	26	the	the	DET
ejpam-86	390	27	models	model	NOUN
ejpam-86	390	28	.	.	PUNCT
ejpam-86	391	1	m.	m.	NOUN
ejpam-86	391	2	liu	liu	PROPN
ejpam-86	391	3	and	and	CCONJ
ejpam-86	391	4	h.	h.	PROPN
ejpam-86	391	5	bozdogan	bozdogan	PROPN
ejpam-86	391	6	/	/	SYM
ejpam-86	391	7	eur	eur	PROPN
ejpam-86	391	8	.	.	PUNCT
ejpam-86	392	1	j.	j.	PROPN
ejpam-86	392	2	pure	pure	PROPN
ejpam-86	392	3	appl	appl	PROPN
ejpam-86	392	4	.	.	PROPN
ejpam-86	392	5	math	math	PROPN
ejpam-86	392	6	,	,	PUNCT
ejpam-86	392	7	1	1	NUM
ejpam-86	392	8	(	(	PUNCT
ejpam-86	392	9	2008	2008	NUM
ejpam-86	392	10	)	)	PUNCT
ejpam-86	392	11	,	,	PUNCT
ejpam-86	392	12	(	(	PUNCT
ejpam-86	392	13	4	4	NUM
ejpam-86	392	14	-	-	SYM
ejpam-86	392	15	37	37	NUM
ejpam-86	392	16	)	)	PUNCT
ejpam-86	392	17	20	20	NUM
ejpam-86	392	18	6.4	6.4	NUM
ejpam-86	392	19	.	.	PUNCT
ejpam-86	393	1	a	a	DET
ejpam-86	393	2	new	new	ADJ
ejpam-86	393	3	ga	ga	NOUN
ejpam-86	393	4	operator	operator	NOUN
ejpam-86	393	5	:	:	PUNCT
ejpam-86	393	6	ga	ga	PROPN
ejpam-86	393	7	engineering	engineering	PROPN
ejpam-86	393	8	ga	ga	PROPN
ejpam-86	393	9	engineering	engineering	NOUN
ejpam-86	393	10	is	be	AUX
ejpam-86	393	11	a	a	DET
ejpam-86	393	12	new	new	ADJ
ejpam-86	393	13	operator	operator	NOUN
ejpam-86	393	14	we	we	PRON
ejpam-86	393	15	developed	develop	VERB
ejpam-86	393	16	and	and	CCONJ
ejpam-86	393	17	introduced	introduce	VERB
ejpam-86	393	18	to	to	PART
ejpam-86	393	19	improve	improve	VERB
ejpam-86	393	20	the	the	DET
ejpam-86	393	21	evolution	evolution	NOUN
ejpam-86	393	22	of	of	ADP
ejpam-86	393	23	the	the	DET
ejpam-86	393	24	ga	ga	PROPN
ejpam-86	393	25	process	process	NOUN
ejpam-86	393	26	.	.	PUNCT
ejpam-86	394	1	since	since	SCONJ
ejpam-86	394	2	ga	ga	PROPN
ejpam-86	394	3	is	be	AUX
ejpam-86	394	4	a	a	DET
ejpam-86	394	5	“	"	PUNCT
ejpam-86	394	6	random	random	ADJ
ejpam-86	394	7	”	"	PUNCT
ejpam-86	394	8	or	or	CCONJ
ejpam-86	394	9	“	"	PUNCT
ejpam-86	394	10	stochastic	stochastic	ADJ
ejpam-86	394	11	”	"	PUNCT
ejpam-86	394	12	search	search	NOUN
ejpam-86	394	13	method	method	NOUN
ejpam-86	394	14	,	,	PUNCT
ejpam-86	394	15	it	it	PRON
ejpam-86	394	16	is	be	AUX
ejpam-86	394	17	not	not	PART
ejpam-86	394	18	guaranteed	guarantee	VERB
ejpam-86	394	19	that	that	SCONJ
ejpam-86	394	20	each	each	DET
ejpam-86	394	21	run	run	NOUN
ejpam-86	394	22	of	of	ADP
ejpam-86	394	23	the	the	DET
ejpam-86	394	24	ga	ga	PROPN
ejpam-86	394	25	with	with	ADP
ejpam-86	394	26	the	the	DET
ejpam-86	394	27	same	same	ADJ
ejpam-86	394	28	settings	setting	NOUN
ejpam-86	394	29	will	will	AUX
ejpam-86	394	30	converge	converge	VERB
ejpam-86	394	31	to	to	ADP
ejpam-86	394	32	the	the	DET
ejpam-86	394	33	same	same	ADJ
ejpam-86	394	34	optimal	optimal	ADJ
ejpam-86	394	35	solution	solution	NOUN
ejpam-86	394	36	.	.	PUNCT
ejpam-86	395	1	with	with	ADP
ejpam-86	395	2	classic	classic	ADJ
ejpam-86	395	3	ga	ga	PROPN
ejpam-86	395	4	operators	operator	NOUN
ejpam-86	395	5	and	and	CCONJ
ejpam-86	395	6	the	the	DET
ejpam-86	395	7	proper	proper	ADJ
ejpam-86	395	8	setting	setting	NOUN
ejpam-86	395	9	of	of	ADP
ejpam-86	395	10	the	the	DET
ejpam-86	395	11	parameters	parameter	NOUN
ejpam-86	395	12	,	,	PUNCT
ejpam-86	395	13	the	the	DET
ejpam-86	395	14	bias	bias	NOUN
ejpam-86	395	15	and	and	CCONJ
ejpam-86	395	16	variance	variance	NOUN
ejpam-86	395	17	of	of	ADP
ejpam-86	395	18	mles	mle	NOUN
ejpam-86	395	19	caused	cause	VERB
ejpam-86	395	20	by	by	ADP
ejpam-86	395	21	the	the	DET
ejpam-86	395	22	ga	ga	PROPN
ejpam-86	395	23	are	be	AUX
ejpam-86	395	24	usually	usually	ADV
ejpam-86	395	25	acceptable	acceptable	ADJ
ejpam-86	395	26	according	accord	VERB
ejpam-86	395	27	to	to	ADP
ejpam-86	395	28	chatterjee	chatterjee	PROPN
ejpam-86	395	29	(	(	PUNCT
ejpam-86	395	30	[	[	X
ejpam-86	395	31	15	15	NUM
ejpam-86	395	32	]	]	NUM
ejpam-86	395	33	)	)	PUNCT
ejpam-86	395	34	.	.	PUNCT
ejpam-86	396	1	but	but	CCONJ
ejpam-86	396	2	for	for	ADP
ejpam-86	396	3	the	the	DET
ejpam-86	396	4	ga	ga	PROPN
ejpam-86	396	5	on	on	ADP
ejpam-86	396	6	ga	ga	PROPN
ejpam-86	396	7	approach	approach	NOUN
ejpam-86	396	8	,	,	PUNCT
ejpam-86	396	9	the	the	DET
ejpam-86	396	10	bias	bias	NOUN
ejpam-86	396	11	and	and	CCONJ
ejpam-86	396	12	variance	variance	NOUN
ejpam-86	396	13	caused	cause	VERB
ejpam-86	396	14	by	by	ADP
ejpam-86	396	15	ga	ga	PROPN
ejpam-86	396	16	during	during	ADP
ejpam-86	396	17	the	the	DET
ejpam-86	396	18	estimation	estimation	NOUN
ejpam-86	396	19	of	of	ADP
ejpam-86	396	20	the	the	DET
ejpam-86	396	21	model	model	NOUN
ejpam-86	396	22	parameters	parameter	NOUN
ejpam-86	396	23	will	will	AUX
ejpam-86	396	24	affect	affect	VERB
ejpam-86	396	25	the	the	DET
ejpam-86	396	26	ga	ga	PROPN
ejpam-86	396	27	for	for	ADP
ejpam-86	396	28	model	model	NOUN
ejpam-86	396	29	selection	selection	NOUN
ejpam-86	396	30	.	.	PUNCT
ejpam-86	397	1	for	for	ADP
ejpam-86	397	2	this	this	DET
ejpam-86	397	3	reason	reason	NOUN
ejpam-86	397	4	,	,	PUNCT
ejpam-86	397	5	we	we	PRON
ejpam-86	397	6	introduce	introduce	VERB
ejpam-86	397	7	the	the	DET
ejpam-86	397	8	ga	ga	PROPN
ejpam-86	397	9	engineering	engineering	NOUN
ejpam-86	397	10	(	(	PUNCT
ejpam-86	397	11	or	or	CCONJ
ejpam-86	397	12	ga	ga	NOUN
ejpam-86	397	13	cloning	cloning	NOUN
ejpam-86	397	14	)	)	PUNCT
ejpam-86	397	15	operator	operator	NOUN
ejpam-86	397	16	to	to	PART
ejpam-86	397	17	improve	improve	VERB
ejpam-86	397	18	the	the	DET
ejpam-86	397	19	estimation	estimation	NOUN
ejpam-86	397	20	further	far	ADV
ejpam-86	397	21	.	.	PUNCT
ejpam-86	398	1	if	if	SCONJ
ejpam-86	398	2	the	the	DET
ejpam-86	398	3	population	population	NOUN
ejpam-86	398	4	evolves	evolve	VERB
ejpam-86	398	5	“	"	PUNCT
ejpam-86	398	6	naturally	naturally	ADV
ejpam-86	398	7	”	"	PUNCT
ejpam-86	398	8	with	with	ADP
ejpam-86	398	9	classic	classic	ADJ
ejpam-86	398	10	ga	ga	PROPN
ejpam-86	398	11	operators	operator	NOUN
ejpam-86	398	12	,	,	PUNCT
ejpam-86	398	13	ga	ga	PROPN
ejpam-86	398	14	engineering	engineering	NOUN
ejpam-86	398	15	means	mean	VERB
ejpam-86	398	16	to	to	PART
ejpam-86	398	17	improve	improve	VERB
ejpam-86	398	18	the	the	DET
ejpam-86	398	19	quality	quality	NOUN
ejpam-86	398	20	of	of	ADP
ejpam-86	398	21	the	the	DET
ejpam-86	398	22	population	population	NOUN
ejpam-86	398	23	“	"	PUNCT
ejpam-86	398	24	artificially	artificially	ADV
ejpam-86	398	25	”	"	PUNCT
ejpam-86	398	26	.	.	PUNCT
ejpam-86	399	1	the	the	DET
ejpam-86	399	2	idea	idea	NOUN
ejpam-86	399	3	of	of	ADP
ejpam-86	399	4	ga	ga	PROPN
ejpam-86	399	5	engineering	engineering	NOUN
ejpam-86	399	6	comes	come	VERB
ejpam-86	399	7	from	from	ADP
ejpam-86	399	8	the	the	DET
ejpam-86	399	9	fact	fact	NOUN
ejpam-86	399	10	that	that	SCONJ
ejpam-86	399	11	the	the	DET
ejpam-86	399	12	difference	difference	NOUN
ejpam-86	399	13	of	of	ADP
ejpam-86	399	14	the	the	DET
ejpam-86	399	15	fitness	fitness	NOUN
ejpam-86	399	16	values	value	NOUN
ejpam-86	399	17	between	between	ADP
ejpam-86	399	18	the	the	DET
ejpam-86	399	19	two	two	NUM
ejpam-86	399	20	chromosomes	chromosome	NOUN
ejpam-86	399	21	is	be	AUX
ejpam-86	399	22	caused	cause	VERB
ejpam-86	399	23	by	by	ADP
ejpam-86	399	24	the	the	DET
ejpam-86	399	25	bits	bit	NOUN
ejpam-86	399	26	with	with	ADP
ejpam-86	399	27	different	different	ADJ
ejpam-86	399	28	binary	binary	ADJ
ejpam-86	399	29	codes	code	NOUN
ejpam-86	399	30	in	in	ADP
ejpam-86	399	31	these	these	DET
ejpam-86	399	32	two	two	NUM
ejpam-86	399	33	chromosomes	chromosome	NOUN
ejpam-86	399	34	and	and	CCONJ
ejpam-86	399	35	the	the	DET
ejpam-86	399	36	bit	bit	NOUN
ejpam-86	399	37	in	in	ADP
ejpam-86	399	38	the	the	DET
ejpam-86	399	39	chromosome	chromosome	NOUN
ejpam-86	399	40	with	with	ADP
ejpam-86	399	41	better	well	ADJ
ejpam-86	399	42	fitness	fitness	NOUN
ejpam-86	399	43	are	be	AUX
ejpam-86	399	44	supposed	suppose	VERB
ejpam-86	399	45	to	to	PART
ejpam-86	399	46	contain	contain	VERB
ejpam-86	399	47	better	well	ADJ
ejpam-86	399	48	genes	gene	NOUN
ejpam-86	399	49	/	/	SYM
ejpam-86	399	50	information	information	NOUN
ejpam-86	399	51	.	.	PUNCT
ejpam-86	400	1	for	for	ADP
ejpam-86	400	2	example	example	NOUN
ejpam-86	400	3	,	,	PUNCT
ejpam-86	400	4	given	give	VERB
ejpam-86	400	5	chromosome	chromosome	NOUN
ejpam-86	400	6	a	a	DET
ejpam-86	400	7	100101110	100101110	NUM
ejpam-86	400	8	(	(	PUNCT
ejpam-86	400	9	fitness	fitness	NOUN
ejpam-86	400	10	a=10	a=10	NOUN
ejpam-86	400	11	)	)	PUNCT
ejpam-86	400	12	chromosome	chromosome	NOUN
ejpam-86	400	13	b	b	NOUN
ejpam-86	400	14	101100101	101100101	NUM
ejpam-86	400	15	(	(	PUNCT
ejpam-86	400	16	fitness	fitness	NOUN
ejpam-86	400	17	b=20	b=20	NOUN
ejpam-86	400	18	)	)	PUNCT
ejpam-86	400	19	→	→	ADP
ejpam-86	400	20	different	different	ADJ
ejpam-86	400	21	bits	bit	NOUN
ejpam-86	400	22	in	in	ADP
ejpam-86	400	23	a	a	DET
ejpam-86	400	24	t	t	NOUN
ejpam-86	400	25	t	t	NOUN
ejpam-86	400	26	0	0	NUM
ejpam-86	400	27	t	t	NOUN
ejpam-86	400	28	t1	t1	NOUN
ejpam-86	400	29	t	t	PROPN
ejpam-86	400	30	10	10	NUM
ejpam-86	400	31	different	different	ADJ
ejpam-86	400	32	bits	bit	NOUN
ejpam-86	400	33	in	in	ADP
ejpam-86	400	34	b	b	PROPN
ejpam-86	400	35	t	t	PROPN
ejpam-86	400	36	t	t	PROPN
ejpam-86	400	37	1	1	NUM
ejpam-86	400	38	t	t	NOUN
ejpam-86	400	39	t0	t0	PROPN
ejpam-86	400	40	t	t	PROPN
ejpam-86	400	41	01	01	NUM
ejpam-86	400	42	,	,	PUNCT
ejpam-86	400	43	tt	tt	PROPN
ejpam-86	400	44	0tt1	0tt1	PROPN
ejpam-86	400	45	t	t	PROPN
ejpam-86	400	46	10	10	NUM
ejpam-86	400	47	is	be	AUX
ejpam-86	400	48	supposed	suppose	VERB
ejpam-86	400	49	to	to	PART
ejpam-86	400	50	have	have	VERB
ejpam-86	400	51	better	well	ADJ
ejpam-86	400	52	genes	gene	NOUN
ejpam-86	400	53	than	than	ADP
ejpam-86	400	54	tt	tt	PROPN
ejpam-86	400	55	1tt0	1tt0	NUM
ejpam-86	400	56	t	t	NOUN
ejpam-86	400	57	01	01	NUM
ejpam-86	400	58	if	if	SCONJ
ejpam-86	400	59	a	a	DET
ejpam-86	400	60	smaller	small	ADJ
ejpam-86	400	61	fitness	fitness	NOUN
ejpam-86	400	62	value	value	NOUN
ejpam-86	400	63	is	be	AUX
ejpam-86	400	64	preferred	prefer	VERB
ejpam-86	400	65	.	.	PUNCT
ejpam-86	401	1	so	so	ADV
ejpam-86	401	2	,	,	PUNCT
ejpam-86	401	3	if	if	SCONJ
ejpam-86	401	4	we	we	PRON
ejpam-86	401	5	compare	compare	VERB
ejpam-86	401	6	the	the	DET
ejpam-86	401	7	best	good	ADJ
ejpam-86	401	8	chromosome	chromosome	NOUN
ejpam-86	401	9	of	of	ADP
ejpam-86	401	10	generation	generation	NOUN
ejpam-86	401	11	i	i	PRON
ejpam-86	401	12	with	with	ADP
ejpam-86	401	13	that	that	PRON
ejpam-86	401	14	of	of	ADP
ejpam-86	401	15	generation	generation	NOUN
ejpam-86	401	16	i	i	PRON
ejpam-86	401	17	+	+	NOUN
ejpam-86	401	18	1	1	NUM
ejpam-86	401	19	,	,	PUNCT
ejpam-86	401	20	with	with	ADP
ejpam-86	401	21	probability	probability	NOUN
ejpam-86	401	22	pe	pe	NOUN
ejpam-86	401	23	,	,	PUNCT
ejpam-86	401	24	and	and	CCONJ
ejpam-86	401	25	find	find	VERB
ejpam-86	401	26	that	that	SCONJ
ejpam-86	401	27	they	they	PRON
ejpam-86	401	28	have	have	VERB
ejpam-86	401	29	different	different	ADJ
ejpam-86	401	30	binary	binary	ADJ
ejpam-86	401	31	codes	code	NOUN
ejpam-86	401	32	and	and	CCONJ
ejpam-86	401	33	bits	bit	NOUN
ejpam-86	401	34	,	,	PUNCT
ejpam-86	401	35	then	then	ADV
ejpam-86	401	36	we	we	PRON
ejpam-86	401	37	prefer	prefer	VERB
ejpam-86	401	38	the	the	DET
ejpam-86	401	39	bits	bit	NOUN
ejpam-86	401	40	corresponding	correspond	VERB
ejpam-86	401	41	to	to	ADP
ejpam-86	401	42	the	the	DET
ejpam-86	401	43	chromosome	chromosome	NOUN
ejpam-86	401	44	which	which	PRON
ejpam-86	401	45	has	have	VERB
ejpam-86	401	46	the	the	DET
ejpam-86	401	47	smaller	small	ADJ
ejpam-86	401	48	fitness	fitness	NOUN
ejpam-86	401	49	value	value	NOUN
ejpam-86	401	50	,	,	PUNCT
ejpam-86	401	51	since	since	SCONJ
ejpam-86	401	52	we	we	PRON
ejpam-86	401	53	are	be	AUX
ejpam-86	401	54	minimizing	minimize	VERB
ejpam-86	401	55	the	the	DET
ejpam-86	401	56	information	information	NOUN
ejpam-86	401	57	criteria	criterion	NOUN
ejpam-86	401	58	to	to	PART
ejpam-86	401	59	pick	pick	VERB
ejpam-86	401	60	the	the	DET
ejpam-86	401	61	best	good	ADJ
ejpam-86	401	62	model	model	NOUN
ejpam-86	401	63	.	.	PUNCT
ejpam-86	402	1	indeed	indeed	ADV
ejpam-86	402	2	,	,	PUNCT
ejpam-86	402	3	our	our	PRON
ejpam-86	402	4	simulation	simulation	NOUN
ejpam-86	402	5	results	result	NOUN
ejpam-86	402	6	show	show	VERB
ejpam-86	402	7	that	that	SCONJ
ejpam-86	402	8	this	this	DET
ejpam-86	402	9	new	new	ADJ
ejpam-86	402	10	operator	operator	NOUN
ejpam-86	402	11	does	do	AUX
ejpam-86	402	12	reduce	reduce	VERB
ejpam-86	402	13	the	the	DET
ejpam-86	402	14	bias	bias	NOUN
ejpam-86	402	15	and	and	CCONJ
ejpam-86	402	16	variance	variance	NOUN
ejpam-86	402	17	caused	cause	VERB
ejpam-86	402	18	by	by	ADP
ejpam-86	402	19	the	the	DET
ejpam-86	402	20	ga	ga	PROPN
ejpam-86	402	21	.	.	PUNCT
ejpam-86	403	1	this	this	PRON
ejpam-86	403	2	we	we	PRON
ejpam-86	403	3	like	like	VERB
ejpam-86	403	4	.	.	PUNCT
ejpam-86	404	1	7	7	X
ejpam-86	404	2	.	.	X
ejpam-86	404	3	numerical	numerical	ADJ
ejpam-86	404	4	examples	example	NOUN
ejpam-86	404	5	7.1	7.1	NUM
ejpam-86	404	6	.	.	PUNCT
ejpam-86	405	1	simulation	simulation	NOUN
ejpam-86	405	2	examples	example	NOUN
ejpam-86	405	3	in	in	ADP
ejpam-86	405	4	the	the	DET
ejpam-86	405	5	following	follow	VERB
ejpam-86	405	6	simulation	simulation	NOUN
ejpam-86	405	7	examples	example	NOUN
ejpam-86	405	8	,	,	PUNCT
ejpam-86	405	9	we	we	PRON
ejpam-86	405	10	use	use	VERB
ejpam-86	405	11	the	the	DET
ejpam-86	405	12	procedure	procedure	NOUN
ejpam-86	405	13	described	describe	VERB
ejpam-86	405	14	by	by	ADP
ejpam-86	405	15	gómez	gómez	NOUN
ejpam-86	405	16	-	-	PUNCT
ejpam-86	405	17	villegas	villegas	PROPN
ejpam-86	405	18	and	and	CCONJ
ejpam-86	405	19	sánchez	sánchez	PROPN
ejpam-86	405	20	-	-	PUNCT
ejpam-86	405	21	manzano	manzano	PROPN
ejpam-86	405	22	et	et	PROPN
ejpam-86	405	23	al	al	PROPN
ejpam-86	405	24	.	.	PUNCT
ejpam-86	406	1	(	(	PUNCT
ejpam-86	406	2	[	[	X
ejpam-86	406	3	20	20	NUM
ejpam-86	406	4	]	]	PUNCT
ejpam-86	406	5	)	)	PUNCT
ejpam-86	406	6	to	to	PART
ejpam-86	406	7	generate	generate	VERB
ejpam-86	406	8	pe	pe	ADP
ejpam-86	406	9	random	random	ADJ
ejpam-86	406	10	vectors	vector	NOUN
ejpam-86	406	11	.	.	PUNCT
ejpam-86	407	1	the	the	DET
ejpam-86	407	2	ga	ga	PROPN
ejpam-86	407	3	on	on	ADP
ejpam-86	407	4	ga	ga	PROPN
ejpam-86	407	5	approach	approach	NOUN
ejpam-86	407	6	developed	develop	VERB
ejpam-86	407	7	in	in	ADP
ejpam-86	407	8	bozdogan	bozdogan	NOUN
ejpam-86	407	9	and	and	CCONJ
ejpam-86	407	10	liu	liu	PROPN
ejpam-86	407	11	(	(	PUNCT
ejpam-86	407	12	[	[	X
ejpam-86	407	13	23	23	NUM
ejpam-86	407	14	]	]	PUNCT
ejpam-86	407	15	)	)	PUNCT
ejpam-86	407	16	is	be	AUX
ejpam-86	407	17	used	use	VERB
ejpam-86	407	18	to	to	PART
ejpam-86	407	19	select	select	VERB
ejpam-86	407	20	predictor	predictor	NOUN
ejpam-86	407	21	variables	variable	NOUN
ejpam-86	407	22	and	and	CCONJ
ejpam-86	407	23	to	to	PART
ejpam-86	407	24	estimate	estimate	VERB
ejpam-86	407	25	the	the	DET
ejpam-86	407	26	model	model	NOUN
ejpam-86	407	27	parameters	parameter	NOUN
ejpam-86	407	28	.	.	PUNCT
ejpam-86	408	1	predictor	predictor	NOUN
ejpam-86	408	2	variables	variable	NOUN
ejpam-86	408	3	are	be	AUX
ejpam-86	408	4	simulated	simulate	VERB
ejpam-86	408	5	using	use	VERB
ejpam-86	408	6	the	the	DET
ejpam-86	408	7	following	follow	VERB
ejpam-86	408	8	simulation	simulation	NOUN
ejpam-86	408	9	protocol	protocol	NOUN
ejpam-86	408	10	.	.	PUNCT
ejpam-86	409	1	the	the	DET
ejpam-86	409	2	first	first	ADJ
ejpam-86	409	3	three	three	NUM
ejpam-86	409	4	predictors	predictor	NOUN
ejpam-86	409	5	are	be	AUX
ejpam-86	409	6	simulated	simulate	VERB
ejpam-86	409	7	by	by	ADP
ejpam-86	409	8	x1	x1	PROPN
ejpam-86	409	9	=	=	SYM
ejpam-86	409	10	10	10	NUM
ejpam-86	409	11	+	+	CCONJ
ejpam-86	409	12	ε1	ε1	VERB
ejpam-86	409	13	x2	x2	NOUN
ejpam-86	409	14	=	=	PUNCT
ejpam-86	409	15	10	10	NUM
ejpam-86	409	16	+	+	CCONJ
ejpam-86	409	17	0.3ε1	0.3ε1	NUM
ejpam-86	410	1	+	+	CCONJ
ejpam-86	411	1	αε2,where	αε2,where	NUM
ejpam-86	411	2	α	α	NOUN
ejpam-86	411	3	=	=	PUNCT
ejpam-86	411	4	√	√	PROPN
ejpam-86	411	5	1−	1−	NUM
ejpam-86	411	6	0.32	0.32	NUM
ejpam-86	411	7	=	=	SYM
ejpam-86	411	8	0.9539	0.9539	NUM
ejpam-86	411	9	x3	x3	NOUN
ejpam-86	411	10	=	=	SYM
ejpam-86	411	11	10	10	NUM
ejpam-86	411	12	+	+	CCONJ
ejpam-86	411	13	0.3ε1	0.3ε1	NUM
ejpam-86	411	14	+	+	CCONJ
ejpam-86	411	15	0.5604αε2	0.5604αε2	PRON
ejpam-86	412	1	+	+	CCONJ
ejpam-86	412	2	0.8282αε3	0.8282αε3	NUM
ejpam-86	412	3	(	(	PUNCT
ejpam-86	412	4	7.1	7.1	NUM
ejpam-86	412	5	)	)	PUNCT
ejpam-86	412	6	where	where	SCONJ
ejpam-86	412	7	the	the	DET
ejpam-86	412	8	components	component	NOUN
ejpam-86	412	9	of	of	ADP
ejpam-86	412	10	ε1	ε1	PROPN
ejpam-86	412	11	,	,	PUNCT
ejpam-86	412	12	ε2	ε2	ADJ
ejpam-86	412	13	and	and	CCONJ
ejpam-86	412	14	ε3	ε3	PROPN
ejpam-86	412	15	are	be	AUX
ejpam-86	412	16	i.i.d	i.i.d	ADP
ejpam-86	412	17	.	.	PUNCT
ejpam-86	413	1	according	accord	VERB
ejpam-86	413	2	to	to	ADP
ejpam-86	413	3	n(0	n(0	PROPN
ejpam-86	413	4	,	,	PUNCT
ejpam-86	413	5	1	1	NUM
ejpam-86	413	6	)	)	PUNCT
ejpam-86	413	7	.	.	PUNCT
ejpam-86	414	1	the	the	DET
ejpam-86	414	2	parameter	parameter	NOUN
ejpam-86	414	3	α	α	PROPN
ejpam-86	414	4	controls	control	VERB
ejpam-86	414	5	the	the	DET
ejpam-86	414	6	degree	degree	NOUN
ejpam-86	414	7	of	of	ADP
ejpam-86	414	8	collinearity	collinearity	NOUN
ejpam-86	414	9	in	in	ADP
ejpam-86	414	10	the	the	DET
ejpam-86	414	11	predictors	predictor	NOUN
ejpam-86	414	12	.	.	PUNCT
ejpam-86	415	1	then	then	ADV
ejpam-86	415	2	,	,	PUNCT
ejpam-86	415	3	we	we	PRON
ejpam-86	415	4	include	include	VERB
ejpam-86	415	5	some	some	DET
ejpam-86	415	6	redundant	redundant	ADJ
ejpam-86	415	7	variables	variable	NOUN
ejpam-86	415	8	,	,	PUNCT
ejpam-86	415	9	x4	x4	PROPN
ejpam-86	415	10	,	,	PUNCT
ejpam-86	415	11	·	·	PUNCT
ejpam-86	415	12	·	·	PUNCT
ejpam-86	415	13	·	·	PUNCT
ejpam-86	415	14	,	,	PUNCT
ejpam-86	415	15	xi	xi	PROPN
ejpam-86	415	16	m.	m.	PROPN
ejpam-86	415	17	liu	liu	PROPN
ejpam-86	415	18	and	and	CCONJ
ejpam-86	415	19	h.	h.	PROPN
ejpam-86	415	20	bozdogan	bozdogan	PROPN
ejpam-86	415	21	/	/	SYM
ejpam-86	415	22	eur	eur	PROPN
ejpam-86	415	23	.	.	PUNCT
ejpam-86	416	1	j.	j.	PROPN
ejpam-86	416	2	pure	pure	PROPN
ejpam-86	416	3	appl	appl	PROPN
ejpam-86	416	4	.	.	PROPN
ejpam-86	416	5	math	math	PROPN
ejpam-86	416	6	,	,	PUNCT
ejpam-86	416	7	1	1	NUM
ejpam-86	416	8	(	(	PUNCT
ejpam-86	416	9	2008	2008	NUM
ejpam-86	416	10	)	)	PUNCT
ejpam-86	416	11	,	,	PUNCT
ejpam-86	416	12	(	(	PUNCT
ejpam-86	416	13	4	4	NUM
ejpam-86	416	14	-	-	SYM
ejpam-86	416	15	37	37	NUM
ejpam-86	416	16	)	)	PUNCT
ejpam-86	416	17	21	21	NUM
ejpam-86	416	18	which	which	PRON
ejpam-86	416	19	are	be	AUX
ejpam-86	416	20	simulated	simulate	VERB
ejpam-86	416	21	by	by	ADP
ejpam-86	416	22	:	:	PUNCT
ejpam-86	416	23	x4	x4	PROPN
ejpam-86	416	24	=	=	SYM
ejpam-86	416	25	4×	4×	NOUN
ejpam-86	416	26	rand(0	rand(0	NOUN
ejpam-86	416	27	,	,	PUNCT
ejpam-86	416	28	1	1	NUM
ejpam-86	416	29	)	)	PUNCT
ejpam-86	416	30	,	,	PUNCT
ejpam-86	416	31	·	·	PUNCT
ejpam-86	416	32	·	·	PUNCT
ejpam-86	416	33	·	·	PUNCT
ejpam-86	416	34	,	,	PUNCT
ejpam-86	416	35	xi	xi	X
ejpam-86	416	36	=	=	PUNCT
ejpam-86	416	37	i×	i×	PROPN
ejpam-86	417	1	rand(0	rand(0	PROPN
ejpam-86	417	2	,	,	PUNCT
ejpam-86	417	3	1	1	NUM
ejpam-86	417	4	)	)	PUNCT
ejpam-86	417	5	(	(	PUNCT
ejpam-86	417	6	7.2	7.2	NUM
ejpam-86	417	7	)	)	PUNCT
ejpam-86	417	8	where	where	SCONJ
ejpam-86	417	9	rand(0	rand(0	PROPN
ejpam-86	417	10	,	,	PUNCT
ejpam-86	417	11	1	1	NUM
ejpam-86	417	12	)	)	PUNCT
ejpam-86	417	13	generates	generate	VERB
ejpam-86	417	14	the	the	DET
ejpam-86	417	15	uniform	uniform	ADJ
ejpam-86	417	16	random	random	ADJ
ejpam-86	417	17	numbers	number	NOUN
ejpam-86	417	18	in	in	ADP
ejpam-86	417	19	(	(	PUNCT
ejpam-86	417	20	0	0	NUM
ejpam-86	417	21	,	,	PUNCT
ejpam-86	417	22	1	1	NUM
ejpam-86	417	23	)	)	PUNCT
ejpam-86	417	24	.	.	PUNCT
ejpam-86	418	1	the	the	DET
ejpam-86	418	2	response	response	NOUN
ejpam-86	418	3	variables	variable	NOUN
ejpam-86	418	4	are	be	AUX
ejpam-86	418	5	generated	generate	VERB
ejpam-86	418	6	from	from	ADP
ejpam-86	418	7	:	:	PUNCT
ejpam-86	418	8	yn×2	yn×2	PROPN
ejpam-86	419	1	=	=	PUNCT
ejpam-86	420	1	[	[	X
ejpam-86	420	2	1	1	NUM
ejpam-86	420	3	,	,	PUNCT
ejpam-86	420	4	xn×3]b4×2	xn×3]b4×2	PROPN
ejpam-86	421	1	+	+	CCONJ
ejpam-86	422	1	en×2	en×2	PROPN
ejpam-86	422	2	(	(	PUNCT
ejpam-86	422	3	7.3	7.3	NUM
ejpam-86	422	4	)	)	PUNCT
ejpam-86	422	5	with	with	ADP
ejpam-86	422	6	x	x	X
ejpam-86	422	7	=	=	PUNCT
ejpam-86	423	1	[	[	X
ejpam-86	423	2	x1,x2,x3	x1,x2,x3	X
ejpam-86	423	3	]	]	X
ejpam-86	423	4	and	and	CCONJ
ejpam-86	423	5	b	b	X
ejpam-86	423	6	=	=	NOUN
ejpam-86	423	7			NOUN
ejpam-86	423	8			VERB
ejpam-86	423	9	8	8	NUM
ejpam-86	423	10	−5	−5	NOUN
ejpam-86	423	11	1	1	NUM
ejpam-86	423	12	0.5	0.5	NUM
ejpam-86	423	13	0.5	0.5	NUM
ejpam-86	423	14	0	0	NUM
ejpam-86	423	15	0.3	0.3	NUM
ejpam-86	423	16	0.3	0.3	NUM
ejpam-86	423	17			PROPN
ejpam-86	423	18			NOUN
ejpam-86	423	19	.	.	PUNCT
ejpam-86	424	1	the	the	DET
ejpam-86	424	2	error	error	NOUN
ejpam-86	424	3	terms	term	NOUN
ejpam-86	424	4	in	in	ADP
ejpam-86	424	5	(	(	PUNCT
ejpam-86	424	6	7.3	7.3	NUM
ejpam-86	424	7	)	)	PUNCT
ejpam-86	424	8	are	be	AUX
ejpam-86	424	9	now	now	ADV
ejpam-86	424	10	multivariate	multivariate	NOUN
ejpam-86	424	11	pe	pe	INTJ
ejpam-86	424	12	(	(	PUNCT
ejpam-86	424	13	mpe	mpe	NOUN
ejpam-86	424	14	)	)	PUNCT
ejpam-86	424	15	distributed	distribute	VERB
ejpam-86	424	16	rather	rather	ADV
ejpam-86	424	17	than	than	ADP
ejpam-86	424	18	multivariate	multivariate	VERB
ejpam-86	424	19	normal	normal	ADJ
ejpam-86	424	20	(	(	PUNCT
ejpam-86	424	21	mn	mn	PROPN
ejpam-86	424	22	)	)	PUNCT
ejpam-86	424	23	.	.	PUNCT
ejpam-86	425	1	7.1.1	7.1.1	X
ejpam-86	425	2	.	.	PUNCT
ejpam-86	425	3	example	example	NOUN
ejpam-86	425	4	1	1	NUM
ejpam-86	425	5	:	:	PUNCT
ejpam-86	425	6	model	model	NOUN
ejpam-86	425	7	parameter	parameter	PROPN
ejpam-86	425	8	estimation	estimation	NOUN
ejpam-86	425	9	and	and	CCONJ
ejpam-86	425	10	model	model	NOUN
ejpam-86	425	11	selection	selection	NOUN
ejpam-86	425	12	in	in	ADP
ejpam-86	425	13	this	this	DET
ejpam-86	425	14	example	example	NOUN
ejpam-86	425	15	,	,	PUNCT
ejpam-86	425	16	we	we	PRON
ejpam-86	425	17	generate	generate	VERB
ejpam-86	425	18	x1,x2,and	x1,x2,and	PROPN
ejpam-86	425	19	x3	x3	ADJ
ejpam-86	425	20	from	from	ADP
ejpam-86	425	21	(	(	PUNCT
ejpam-86	425	22	7.1	7.1	NUM
ejpam-86	425	23	)	)	PUNCT
ejpam-86	425	24	and	and	CCONJ
ejpam-86	425	25	generate	generate	VERB
ejpam-86	425	26	y	y	PROPN
ejpam-86	425	27	from	from	ADP
ejpam-86	425	28	(	(	PUNCT
ejpam-86	425	29	7.3	7.3	NUM
ejpam-86	425	30	)	)	PUNCT
ejpam-86	425	31	with	with	ADP
ejpam-86	425	32	sample	sample	NOUN
ejpam-86	425	33	size	size	NOUN
ejpam-86	425	34	n	n	NOUN
ejpam-86	425	35	=	=	SYM
ejpam-86	425	36	500	500	NUM
ejpam-86	425	37	.	.	PUNCT
ejpam-86	426	1	the	the	DET
ejpam-86	426	2	rows	row	NOUN
ejpam-86	426	3	of	of	ADP
ejpam-86	426	4	e	e	NOUN
ejpam-86	426	5	are	be	AUX
ejpam-86	426	6	i.i.d	i.i.d	ADJ
ejpam-86	426	7	.	.	PUNCT
ejpam-86	426	8	pe2(0,σ	pe2(0,σ	PROPN
ejpam-86	426	9	,	,	PUNCT
ejpam-86	426	10	β	β	NOUN
ejpam-86	426	11	)	)	PUNCT
ejpam-86	426	12	with	with	ADP
ejpam-86	426	13	σ	σ	PROPN
ejpam-86	426	14	=	=	PUNCT
ejpam-86	426	15	[	[	PUNCT
ejpam-86	426	16	1	1	NUM
ejpam-86	426	17	0.5	0.5	NUM
ejpam-86	426	18	0.5	0.5	NUM
ejpam-86	426	19	2	2	NUM
ejpam-86	426	20	]	]	PUNCT
ejpam-86	426	21	which	which	PRON
ejpam-86	426	22	is	be	AUX
ejpam-86	426	23	positive	positive	ADJ
ejpam-86	426	24	definite	definite	ADJ
ejpam-86	426	25	symmetric	symmetric	NOUN
ejpam-86	426	26	.	.	PUNCT
ejpam-86	427	1	we	we	PRON
ejpam-86	427	2	simulate	simulate	VERB
ejpam-86	427	3	two	two	NUM
ejpam-86	427	4	cases	case	NOUN
ejpam-86	427	5	,	,	PUNCT
ejpam-86	427	6	β	β	X
ejpam-86	427	7	=	=	PUNCT
ejpam-86	427	8	0.3	0.3	NUM
ejpam-86	427	9	and	and	CCONJ
ejpam-86	427	10	β	β	X
ejpam-86	427	11	=	=	SYM
ejpam-86	427	12	2	2	NUM
ejpam-86	427	13	with	with	ADP
ejpam-86	427	14	the	the	DET
ejpam-86	427	15	two	two	NUM
ejpam-86	427	16	dimensional	dimensional	ADJ
ejpam-86	427	17	plots	plot	NOUN
ejpam-86	427	18	given	give	VERB
ejpam-86	427	19	in	in	ADP
ejpam-86	427	20	figure	figure	NOUN
ejpam-86	427	21	1	1	NUM
ejpam-86	427	22	.	.	PUNCT
ejpam-86	427	23	to	to	ADP
ejpam-86	427	24	these	these	DET
ejpam-86	427	25	simulated	simulate	VERB
ejpam-86	427	26	data	data	NOUN
ejpam-86	427	27	sets	set	NOUN
ejpam-86	427	28	,	,	PUNCT
ejpam-86	427	29	we	we	PRON
ejpam-86	427	30	fit	fit	VERB
ejpam-86	427	31	both	both	DET
ejpam-86	427	32	multivariate	multivariate	NOUN
ejpam-86	427	33	regression	regression	NOUN
ejpam-86	427	34	model	model	NOUN
ejpam-86	427	35	under	under	ADP
ejpam-86	427	36	normality	normality	NOUN
ejpam-86	427	37	assumption	assumption	NOUN
ejpam-86	427	38	and	and	CCONJ
ejpam-86	427	39	type	type	NOUN
ejpam-86	427	40	i	i	PROPN
ejpam-86	427	41	mvper	mvper	PROPN
ejpam-86	427	42	model	model	NOUN
ejpam-86	427	43	.	.	PUNCT
ejpam-86	428	1	in	in	ADP
ejpam-86	428	2	this	this	DET
ejpam-86	428	3	case	case	NOUN
ejpam-86	428	4	,	,	PUNCT
ejpam-86	428	5	we	we	PRON
ejpam-86	428	6	expect	expect	VERB
ejpam-86	428	7	that	that	SCONJ
ejpam-86	428	8	the	the	DET
ejpam-86	428	9	type	type	NOUN
ejpam-86	428	10	i	i	PRON
ejpam-86	428	11	mvper	mvper	NOUN
ejpam-86	428	12	model	model	NOUN
ejpam-86	428	13	would	would	AUX
ejpam-86	428	14	be	be	AUX
ejpam-86	428	15	chosen	choose	VERB
ejpam-86	428	16	as	as	ADP
ejpam-86	428	17	our	our	PRON
ejpam-86	428	18	best	good	ADJ
ejpam-86	428	19	model	model	NOUN
ejpam-86	428	20	according	accord	VERB
ejpam-86	428	21	to	to	ADP
ejpam-86	428	22	the	the	DET
ejpam-86	428	23	minimum	minimum	NOUN
ejpam-86	428	24	of	of	ADP
ejpam-86	428	25	aic	aic	PROPN
ejpam-86	428	26	or	or	CCONJ
ejpam-86	428	27	icomp	icomp	NOUN
ejpam-86	428	28	,	,	PUNCT
ejpam-86	428	29	and	and	CCONJ
ejpam-86	428	30	that	that	SCONJ
ejpam-86	428	31	the	the	DET
ejpam-86	428	32	model	model	NOUN
ejpam-86	428	33	parameters	parameter	NOUN
ejpam-86	428	34	would	would	AUX
ejpam-86	428	35	be	be	AUX
ejpam-86	428	36	estimated	estimate	VERB
ejpam-86	428	37	correctly	correctly	ADV
ejpam-86	428	38	,	,	PUNCT
ejpam-86	428	39	since	since	SCONJ
ejpam-86	428	40	our	our	PRON
ejpam-86	428	41	true	true	ADJ
ejpam-86	428	42	model	model	NOUN
ejpam-86	428	43	is	be	AUX
ejpam-86	428	44	generated	generate	VERB
ejpam-86	428	45	under	under	ADP
ejpam-86	428	46	the	the	DET
ejpam-86	428	47	mpe	mpe	PROPN
ejpam-86	428	48	assumption	assumption	NOUN
ejpam-86	428	49	.	.	PUNCT
ejpam-86	429	1	we	we	PRON
ejpam-86	429	2	are	be	AUX
ejpam-86	429	3	especially	especially	ADV
ejpam-86	429	4	interested	interested	ADJ
ejpam-86	429	5	in	in	ADP
ejpam-86	429	6	the	the	DET
ejpam-86	429	7	estimates	estimate	NOUN
ejpam-86	429	8	of	of	ADP
ejpam-86	429	9	β	β	X
ejpam-86	429	10	.	.	PUNCT
ejpam-86	430	1	the	the	DET
ejpam-86	430	2	results	result	NOUN
ejpam-86	430	3	of	of	ADP
ejpam-86	430	4	200	200	NUM
ejpam-86	430	5	runs	run	NOUN
ejpam-86	430	6	of	of	ADP
ejpam-86	430	7	both	both	DET
ejpam-86	430	8	simulation	simulation	NOUN
ejpam-86	430	9	cases	case	NOUN
ejpam-86	430	10	are	be	AUX
ejpam-86	430	11	reported	report	VERB
ejpam-86	430	12	in	in	ADP
ejpam-86	430	13	table	table	NOUN
ejpam-86	430	14	1	1	NUM
ejpam-86	430	15	.	.	PUNCT
ejpam-86	430	16	table	table	NOUN
ejpam-86	430	17	1	1	NUM
ejpam-86	430	18	:	:	PUNCT
ejpam-86	430	19	results	result	NOUN
ejpam-86	430	20	of	of	ADP
ejpam-86	430	21	the	the	DET
ejpam-86	430	22	simulation	simulation	NOUN
ejpam-86	430	23	example	example	NOUN
ejpam-86	430	24	1	1	NUM
ejpam-86	430	25	.	.	PUNCT
ejpam-86	430	26	model	model	PROPN
ejpam-86	430	27	avg	avg	PROPN
ejpam-86	430	28	.	.	PROPN
ejpam-86	431	1	βmom	βmom	PROPN
ejpam-86	431	2	avg	avg	PROPN
ejpam-86	431	3	.	.	PROPN
ejpam-86	431	4	βmle	βmle	PROPN
ejpam-86	431	5	avg	avg	PROPN
ejpam-86	431	6	.	.	PUNCT
ejpam-86	432	1	aic	aic	PROPN
ejpam-86	432	2	avg	avg	PROPN
ejpam-86	432	3	.	.	PROPN
ejpam-86	433	1	icomp	icomp	PROPN
ejpam-86	433	2	real	real	ADJ
ejpam-86	433	3	β	β	X
ejpam-86	434	1	=	=	NOUN
ejpam-86	434	2	0.3	0.3	NUM
ejpam-86	434	3	normal	normal	ADJ
ejpam-86	434	4	1	1	NUM
ejpam-86	434	5	1	1	NUM
ejpam-86	434	6	10487	10487	NUM
ejpam-86	434	7	10529	10529	NUM
ejpam-86	434	8	type	type	NOUN
ejpam-86	434	9	i	i	PRON
ejpam-86	434	10	mvper	mvper	VERB
ejpam-86	434	11	0.3312	0.3312	NUM
ejpam-86	434	12	0.3618	0.3618	NUM
ejpam-86	434	13	10154	10154	NUM
ejpam-86	434	14	10193	10193	NUM
ejpam-86	434	15	real	real	ADJ
ejpam-86	434	16	β	β	X
ejpam-86	434	17	=	=	SYM
ejpam-86	434	18	2	2	NUM
ejpam-86	434	19	normal	normal	ADJ
ejpam-86	434	20	1	1	NUM
ejpam-86	434	21	1	1	NUM
ejpam-86	434	22	3015.7	3015.7	NUM
ejpam-86	434	23	3021.3	3021.3	NUM
ejpam-86	434	24	type	type	NOUN
ejpam-86	434	25	i	i	PRON
ejpam-86	434	26	mvper	mvper	VERB
ejpam-86	434	27	1.9668	1.9668	NUM
ejpam-86	434	28	2.4140	2.4140	NUM
ejpam-86	434	29	2840.3	2840.3	NUM
ejpam-86	434	30	2860.0	2860.0	NUM
ejpam-86	434	31	from	from	ADP
ejpam-86	434	32	table	table	NOUN
ejpam-86	434	33	1	1	NUM
ejpam-86	434	34	,	,	PUNCT
ejpam-86	434	35	we	we	PRON
ejpam-86	434	36	see	see	VERB
ejpam-86	434	37	that	that	SCONJ
ejpam-86	434	38	in	in	ADP
ejpam-86	434	39	all	all	DET
ejpam-86	434	40	200	200	NUM
ejpam-86	434	41	runs	run	NOUN
ejpam-86	434	42	of	of	ADP
ejpam-86	434	43	the	the	DET
ejpam-86	434	44	simulation	simulation	NOUN
ejpam-86	434	45	,	,	PUNCT
ejpam-86	434	46	the	the	DET
ejpam-86	434	47	true	true	ADJ
ejpam-86	434	48	model	model	NOUN
ejpam-86	434	49	,	,	PUNCT
ejpam-86	434	50	i.e.	i.e.	X
ejpam-86	434	51	,	,	PUNCT
ejpam-86	434	52	the	the	DET
ejpam-86	434	53	type	type	NOUN
ejpam-86	434	54	i	i	PROPN
ejpam-86	434	55	mvper	mvper	PROPN
ejpam-86	434	56	model	model	NOUN
ejpam-86	434	57	,	,	PUNCT
ejpam-86	434	58	is	be	AUX
ejpam-86	434	59	chosen	choose	VERB
ejpam-86	434	60	by	by	ADP
ejpam-86	434	61	both	both	DET
ejpam-86	434	62	aic	aic	PROPN
ejpam-86	434	63	and	and	CCONJ
ejpam-86	434	64	icomp	icomp	PROPN
ejpam-86	434	65	criteria	criterion	NOUN
ejpam-86	434	66	.	.	PUNCT
ejpam-86	435	1	further	far	ADV
ejpam-86	435	2	,	,	PUNCT
ejpam-86	435	3	the	the	DET
ejpam-86	435	4	estimates	estimate	NOUN
ejpam-86	435	5	of	of	ADP
ejpam-86	435	6	the	the	DET
ejpam-86	435	7	shape	shape	NOUN
ejpam-86	435	8	m.	m.	NOUN
ejpam-86	435	9	liu	liu	PROPN
ejpam-86	435	10	and	and	CCONJ
ejpam-86	435	11	h.	h.	PROPN
ejpam-86	435	12	bozdogan	bozdogan	PROPN
ejpam-86	435	13	/	/	SYM
ejpam-86	435	14	eur	eur	PROPN
ejpam-86	435	15	.	.	PUNCT
ejpam-86	436	1	j.	j.	PROPN
ejpam-86	436	2	pure	pure	PROPN
ejpam-86	436	3	appl	appl	PROPN
ejpam-86	436	4	.	.	PROPN
ejpam-86	436	5	math	math	PROPN
ejpam-86	436	6	,	,	PUNCT
ejpam-86	436	7	1	1	NUM
ejpam-86	436	8	(	(	PUNCT
ejpam-86	436	9	2008	2008	NUM
ejpam-86	436	10	)	)	PUNCT
ejpam-86	436	11	,	,	PUNCT
ejpam-86	436	12	(	(	PUNCT
ejpam-86	436	13	4	4	NUM
ejpam-86	436	14	-	-	SYM
ejpam-86	436	15	37	37	NUM
ejpam-86	436	16	)	)	PUNCT
ejpam-86	436	17	22	22	NUM
ejpam-86	436	18	0.1	0.1	NUM
ejpam-86	436	19	0.2	0.2	NUM
ejpam-86	436	20	0.3	0.3	NUM
ejpam-86	436	21	0.4	0.4	NUM
ejpam-86	436	22	0.5	0.5	NUM
ejpam-86	436	23	0.6	0.6	NUM
ejpam-86	436	24	0	0	NUM
ejpam-86	436	25	10	10	NUM
ejpam-86	436	26	20	20	NUM
ejpam-86	436	27	30	30	NUM
ejpam-86	436	28	histogram	histogram	NOUN
ejpam-86	436	29	of	of	ADP
ejpam-86	436	30	estimated	estimate	VERB
ejpam-86	436	31	β	β	X
ejpam-86	436	32	mom	mom	NOUN
ejpam-86	436	33	with	with	ADP
ejpam-86	436	34	real	real	ADJ
ejpam-86	436	35	β=0.3	β=0.3	NOUN
ejpam-86	436	36	0.1	0.1	NUM
ejpam-86	436	37	0.2	0.2	NUM
ejpam-86	436	38	0.3	0.3	NUM
ejpam-86	436	39	0.4	0.4	NUM
ejpam-86	436	40	0.5	0.5	NUM
ejpam-86	436	41	0.6	0.6	NUM
ejpam-86	436	42	0	0	NUM
ejpam-86	436	43	10	10	NUM
ejpam-86	436	44	20	20	NUM
ejpam-86	436	45	30	30	NUM
ejpam-86	436	46	histogram	histogram	NOUN
ejpam-86	436	47	of	of	ADP
ejpam-86	436	48	estimated	estimate	VERB
ejpam-86	436	49	β	β	X
ejpam-86	436	50	mle	mle	PROPN
ejpam-86	436	51	with	with	ADP
ejpam-86	436	52	real	real	ADJ
ejpam-86	436	53	β=0.3	β=0.3	NOUN
ejpam-86	436	54	1	1	NUM
ejpam-86	436	55	1.5	1.5	NUM
ejpam-86	436	56	2	2	NUM
ejpam-86	436	57	2.5	2.5	NUM
ejpam-86	436	58	3	3	NUM
ejpam-86	436	59	0	0	NUM
ejpam-86	436	60	5	5	NUM
ejpam-86	436	61	10	10	NUM
ejpam-86	436	62	15	15	NUM
ejpam-86	436	63	20	20	NUM
ejpam-86	436	64	25	25	NUM
ejpam-86	436	65	histogram	histogram	NOUN
ejpam-86	436	66	of	of	ADP
ejpam-86	436	67	estimated	estimate	VERB
ejpam-86	436	68	β	β	X
ejpam-86	436	69	mom	mom	NOUN
ejpam-86	436	70	with	with	ADP
ejpam-86	436	71	real	real	ADJ
ejpam-86	436	72	β=2	β=2	SYM
ejpam-86	436	73	0	0	NUM
ejpam-86	436	74	2	2	NUM
ejpam-86	436	75	4	4	NUM
ejpam-86	436	76	6	6	NUM
ejpam-86	436	77	8	8	NUM
ejpam-86	436	78	0	0	NUM
ejpam-86	436	79	10	10	NUM
ejpam-86	436	80	20	20	NUM
ejpam-86	436	81	30	30	NUM
ejpam-86	436	82	40	40	NUM
ejpam-86	436	83	50	50	NUM
ejpam-86	436	84	histogram	histogram	NOUN
ejpam-86	436	85	of	of	ADP
ejpam-86	436	86	estimated	estimate	VERB
ejpam-86	436	87	β	β	X
ejpam-86	436	88	mle	mle	PROPN
ejpam-86	436	89	with	with	ADP
ejpam-86	436	90	real	real	ADJ
ejpam-86	436	91	β=2	β=2	SYM
ejpam-86	436	92	figure	figure	NOUN
ejpam-86	436	93	3	3	NUM
ejpam-86	436	94	:	:	PUNCT
ejpam-86	436	95	the	the	DET
ejpam-86	436	96	mom	mom	NOUN
ejpam-86	436	97	estimates	estimate	NOUN
ejpam-86	436	98	and	and	CCONJ
ejpam-86	436	99	mles	mle	NOUN
ejpam-86	436	100	of	of	ADP
ejpam-86	436	101	β	β	PRON
ejpam-86	436	102	in	in	ADP
ejpam-86	436	103	200	200	NUM
ejpam-86	436	104	runs	run	NOUN
ejpam-86	436	105	of	of	ADP
ejpam-86	436	106	simulation	simulation	NOUN
ejpam-86	436	107	example	example	NOUN
ejpam-86	437	1	1	1	NUM
ejpam-86	437	2	.	.	PUNCT
ejpam-86	438	1	parameter	parameter	PROPN
ejpam-86	438	2	β	β	PROPN
ejpam-86	438	3	are	be	AUX
ejpam-86	438	4	close	close	ADJ
ejpam-86	438	5	to	to	ADP
ejpam-86	438	6	the	the	DET
ejpam-86	438	7	true	true	ADJ
ejpam-86	438	8	values	value	NOUN
ejpam-86	438	9	.	.	PUNCT
ejpam-86	439	1	the	the	DET
ejpam-86	439	2	multivariate	multivariate	NOUN
ejpam-86	439	3	normal	normal	ADJ
ejpam-86	439	4	regression	regression	NOUN
ejpam-86	439	5	model	model	NOUN
ejpam-86	439	6	has	have	AUX
ejpam-86	439	7	never	never	ADV
ejpam-86	439	8	been	be	AUX
ejpam-86	439	9	chosen	choose	VERB
ejpam-86	439	10	by	by	ADP
ejpam-86	439	11	looking	look	VERB
ejpam-86	439	12	at	at	ADP
ejpam-86	439	13	the	the	DET
ejpam-86	439	14	average	average	ADJ
ejpam-86	439	15	values	value	NOUN
ejpam-86	439	16	of	of	ADP
ejpam-86	439	17	aic	aic	PROPN
ejpam-86	439	18	and	and	CCONJ
ejpam-86	439	19	icomp	icomp	PROPN
ejpam-86	439	20	across	across	ADP
ejpam-86	439	21	the	the	DET
ejpam-86	439	22	200	200	NUM
ejpam-86	439	23	runs	run	NOUN
ejpam-86	439	24	of	of	ADP
ejpam-86	439	25	the	the	DET
ejpam-86	439	26	simulation	simulation	NOUN
ejpam-86	439	27	.	.	PUNCT
ejpam-86	440	1	the	the	DET
ejpam-86	440	2	distributions	distribution	NOUN
ejpam-86	440	3	of	of	ADP
ejpam-86	440	4	mom	mom	NOUN
ejpam-86	440	5	estimates	estimate	NOUN
ejpam-86	440	6	and	and	CCONJ
ejpam-86	440	7	mles	mle	NOUN
ejpam-86	440	8	of	of	ADP
ejpam-86	440	9	β	β	PRON
ejpam-86	440	10	in	in	ADP
ejpam-86	440	11	the	the	DET
ejpam-86	440	12	200	200	NUM
ejpam-86	440	13	runs	run	NOUN
ejpam-86	440	14	of	of	ADP
ejpam-86	440	15	the	the	DET
ejpam-86	440	16	simulation	simulation	NOUN
ejpam-86	440	17	are	be	AUX
ejpam-86	440	18	shown	show	VERB
ejpam-86	440	19	in	in	ADP
ejpam-86	440	20	figure	figure	NOUN
ejpam-86	440	21	3	3	NUM
ejpam-86	440	22	.	.	PUNCT
ejpam-86	441	1	the	the	DET
ejpam-86	441	2	q	q	ADJ
ejpam-86	441	3	-	-	PUNCT
ejpam-86	441	4	q	q	NOUN
ejpam-86	441	5	plots	plot	NOUN
ejpam-86	441	6	for	for	ADP
ejpam-86	441	7	the	the	DET
ejpam-86	441	8	type	type	NOUN
ejpam-86	441	9	i	i	PROPN
ejpam-86	441	10	mvper	mvper	NOUN
ejpam-86	441	11	model	model	NOUN
ejpam-86	441	12	in	in	ADP
ejpam-86	441	13	one	one	NUM
ejpam-86	441	14	run	run	NOUN
ejpam-86	441	15	of	of	ADP
ejpam-86	441	16	the	the	DET
ejpam-86	441	17	simulation	simulation	NOUN
ejpam-86	441	18	example	example	NOUN
ejpam-86	441	19	are	be	AUX
ejpam-86	441	20	shown	show	VERB
ejpam-86	441	21	in	in	ADP
ejpam-86	441	22	figure	figure	NOUN
ejpam-86	441	23	4	4	NUM
ejpam-86	441	24	.	.	PUNCT
ejpam-86	441	25	from	from	ADP
ejpam-86	441	26	the	the	DET
ejpam-86	441	27	q	q	ADJ
ejpam-86	441	28	-	-	PUNCT
ejpam-86	441	29	q	q	NOUN
ejpam-86	441	30	plots	plot	NOUN
ejpam-86	441	31	,	,	PUNCT
ejpam-86	441	32	we	we	PRON
ejpam-86	441	33	see	see	VERB
ejpam-86	441	34	that	that	SCONJ
ejpam-86	441	35	the	the	DET
ejpam-86	441	36	type	type	NOUN
ejpam-86	441	37	i	i	PRON
ejpam-86	441	38	models	model	NOUN
ejpam-86	441	39	are	be	AUX
ejpam-86	441	40	far	far	ADV
ejpam-86	441	41	from	from	ADP
ejpam-86	441	42	normal	normal	ADJ
ejpam-86	441	43	models	model	NOUN
ejpam-86	441	44	for	for	ADP
ejpam-86	441	45	both	both	PRON
ejpam-86	441	46	β	β	X
ejpam-86	441	47	=	=	PUNCT
ejpam-86	441	48	0.3	0.3	NUM
ejpam-86	441	49	and	and	CCONJ
ejpam-86	441	50	β	β	X
ejpam-86	441	51	=	=	SYM
ejpam-86	441	52	2	2	NUM
ejpam-86	441	53	,	,	PUNCT
ejpam-86	441	54	which	which	PRON
ejpam-86	441	55	we	we	PRON
ejpam-86	441	56	expected	expect	VERB
ejpam-86	441	57	.	.	PUNCT
ejpam-86	442	1	figure	figure	NOUN
ejpam-86	442	2	5	5	NUM
ejpam-86	442	3	shows	show	VERB
ejpam-86	442	4	a	a	DET
ejpam-86	442	5	typical	typical	ADJ
ejpam-86	442	6	ga	ga	NOUN
ejpam-86	442	7	process	process	NOUN
ejpam-86	442	8	used	use	VERB
ejpam-86	442	9	to	to	PART
ejpam-86	442	10	estimated	estimate	VERB
ejpam-86	442	11	the	the	DET
ejpam-86	442	12	model	model	NOUN
ejpam-86	442	13	parameters	parameter	NOUN
ejpam-86	442	14	in	in	ADP
ejpam-86	442	15	one	one	NUM
ejpam-86	442	16	run	run	NOUN
ejpam-86	442	17	of	of	ADP
ejpam-86	442	18	the	the	DET
ejpam-86	442	19	simulation	simulation	NOUN
ejpam-86	442	20	.	.	PUNCT
ejpam-86	443	1	the	the	DET
ejpam-86	443	2	ga	ga	PROPN
ejpam-86	443	3	parameters	parameter	NOUN
ejpam-86	443	4	are	be	AUX
ejpam-86	443	5	given	give	VERB
ejpam-86	443	6	in	in	ADP
ejpam-86	443	7	table	table	NOUN
ejpam-86	443	8	2	2	NUM
ejpam-86	443	9	,	,	PUNCT
ejpam-86	443	10	where	where	SCONJ
ejpam-86	443	11	ng	ng	PROPN
ejpam-86	443	12	is	be	AUX
ejpam-86	443	13	the	the	DET
ejpam-86	443	14	number	number	NOUN
ejpam-86	443	15	of	of	ADP
ejpam-86	443	16	generations	generation	NOUN
ejpam-86	443	17	,	,	PUNCT
ejpam-86	443	18	ps	ps	PROPN
ejpam-86	443	19	is	be	AUX
ejpam-86	443	20	the	the	DET
ejpam-86	443	21	population	population	NOUN
ejpam-86	443	22	size	size	NOUN
ejpam-86	443	23	,	,	PUNCT
ejpam-86	443	24	pc	pc	NOUN
ejpam-86	443	25	is	be	AUX
ejpam-86	443	26	the	the	DET
ejpam-86	443	27	crossover	crossover	NOUN
ejpam-86	443	28	probability	probability	NOUN
ejpam-86	443	29	,	,	PUNCT
ejpam-86	443	30	pm	pm	NOUN
ejpam-86	443	31	is	be	AUX
ejpam-86	443	32	the	the	DET
ejpam-86	443	33	mutation	mutation	NOUN
ejpam-86	443	34	probability	probability	NOUN
ejpam-86	443	35	,	,	PUNCT
ejpam-86	443	36	pe	pe	PROPN
ejpam-86	443	37	is	be	AUX
ejpam-86	443	38	the	the	DET
ejpam-86	443	39	ga	ga	PROPN
ejpam-86	443	40	engineering	engineering	NOUN
ejpam-86	443	41	probability	probability	NOUN
ejpam-86	443	42	,	,	PUNCT
ejpam-86	443	43	ct	ct	PROPN
ejpam-86	443	44	is	be	AUX
ejpam-86	443	45	the	the	DET
ejpam-86	443	46	crossover	crossover	NOUN
ejpam-86	443	47	type	type	NOUN
ejpam-86	443	48	,	,	PUNCT
ejpam-86	443	49	l	l	NOUN
ejpam-86	443	50	is	be	AUX
ejpam-86	443	51	the	the	DET
ejpam-86	443	52	length	length	NOUN
ejpam-86	443	53	of	of	ADP
ejpam-86	443	54	the	the	DET
ejpam-86	443	55	binary	binary	PROPN
ejpam-86	443	56	string	string	NOUN
ejpam-86	443	57	used	use	VERB
ejpam-86	443	58	to	to	PART
ejpam-86	443	59	encode	encode	VERB
ejpam-86	443	60	real	real	ADJ
ejpam-86	443	61	numbers	number	NOUN
ejpam-86	443	62	,	,	PUNCT
ejpam-86	443	63	and	and	CCONJ
ejpam-86	443	64	nr	nr	PRON
ejpam-86	443	65	is	be	AUX
ejpam-86	443	66	the	the	DET
ejpam-86	443	67	number	number	NOUN
ejpam-86	443	68	of	of	ADP
ejpam-86	443	69	ga	ga	PROPN
ejpam-86	443	70	runs	run	NOUN
ejpam-86	443	71	.	.	PUNCT
ejpam-86	444	1	ct	ct	PRON
ejpam-86	444	2	=	=	SYM
ejpam-86	444	3	3	3	NUM
ejpam-86	444	4	means	mean	VERB
ejpam-86	444	5	uniform	uniform	NOUN
ejpam-86	444	6	crossover	crossover	ADP
ejpam-86	444	7	method	method	NOUN
ejpam-86	444	8	is	be	AUX
ejpam-86	444	9	used	use	VERB
ejpam-86	444	10	.	.	PUNCT
ejpam-86	445	1	table	table	NOUN
ejpam-86	445	2	2	2	NUM
ejpam-86	445	3	:	:	PUNCT
ejpam-86	445	4	ga	ga	NOUN
ejpam-86	445	5	parameters	parameter	NOUN
ejpam-86	445	6	of	of	ADP
ejpam-86	445	7	the	the	DET
ejpam-86	445	8	simulation	simulation	NOUN
ejpam-86	445	9	example	example	NOUN
ejpam-86	445	10	1	1	NUM
ejpam-86	445	11	.	.	PUNCT
ejpam-86	445	12	ga	ga	PROPN
ejpam-86	445	13	parameters	parameter	NOUN
ejpam-86	445	14	ng	ng	PROPN
ejpam-86	445	15	ps	ps	PROPN
ejpam-86	445	16	pc	pc	NOUN
ejpam-86	445	17	pm	pm	NOUN
ejpam-86	446	1	pe	pe	INTJ
ejpam-86	446	2	ct	ct	NUM
ejpam-86	446	3	l	l	NOUN
ejpam-86	447	1	nr	nr	NOUN
ejpam-86	447	2	elitisim	elitisim	PROPN
ejpam-86	447	3	values	value	NOUN
ejpam-86	447	4	50	50	NUM
ejpam-86	447	5	100	100	NUM
ejpam-86	447	6	0.7	0.7	NUM
ejpam-86	447	7	0.1	0.1	NUM
ejpam-86	447	8	0.5	0.5	NUM
ejpam-86	447	9	3	3	NUM
ejpam-86	447	10	32	32	NUM
ejpam-86	447	11	200	200	NUM
ejpam-86	448	1	yes	yes	INTJ
ejpam-86	448	2	7.1.2	7.1.2	NUM
ejpam-86	448	3	.	.	PUNCT
ejpam-86	448	4	example	example	NOUN
ejpam-86	448	5	2	2	NUM
ejpam-86	448	6	:	:	PUNCT
ejpam-86	448	7	subset	subset	NOUN
ejpam-86	448	8	selection	selection	NOUN
ejpam-86	448	9	in	in	ADP
ejpam-86	448	10	this	this	DET
ejpam-86	448	11	example	example	NOUN
ejpam-86	449	1	,	,	PUNCT
ejpam-86	449	2	we	we	PRON
ejpam-86	449	3	show	show	VERB
ejpam-86	449	4	the	the	DET
ejpam-86	449	5	subset	subset	NOUN
ejpam-86	449	6	selection	selection	NOUN
ejpam-86	449	7	of	of	ADP
ejpam-86	449	8	the	the	DET
ejpam-86	449	9	best	good	ADJ
ejpam-86	449	10	predictors	predictor	NOUN
ejpam-86	449	11	under	under	ADP
ejpam-86	449	12	type	type	NOUN
ejpam-86	449	13	i	i	PROPN
ejpam-86	449	14	mvper	mvper	NOUN
ejpam-86	449	15	model	model	NOUN
ejpam-86	449	16	with	with	ADP
ejpam-86	449	17	aic	aic	PROPN
ejpam-86	449	18	and	and	CCONJ
ejpam-86	449	19	icomp(ifim	icomp(ifim	NOUN
ejpam-86	449	20	)	)	PUNCT
ejpam-86	449	21	.	.	PUNCT
ejpam-86	450	1	in	in	ADP
ejpam-86	450	2	this	this	DET
ejpam-86	450	3	simulation	simulation	NOUN
ejpam-86	450	4	,	,	PUNCT
ejpam-86	450	5	three	three	NUM
ejpam-86	450	6	correct	correct	ADJ
ejpam-86	450	7	predictors	predictor	NOUN
ejpam-86	450	8	need	need	VERB
ejpam-86	450	9	to	to	PART
ejpam-86	450	10	be	be	AUX
ejpam-86	450	11	selected	select	VERB
ejpam-86	450	12	from	from	ADP
ejpam-86	450	13	total	total	NOUN
ejpam-86	450	14	of	of	ADP
ejpam-86	450	15	ten	ten	NUM
ejpam-86	450	16	available	available	ADJ
ejpam-86	450	17	predictor	predictor	NOUN
ejpam-86	450	18	variables	variable	NOUN
ejpam-86	450	19	,	,	PUNCT
ejpam-86	450	20	where	where	SCONJ
ejpam-86	450	21	x4	x4	X
ejpam-86	450	22	,	,	PUNCT
ejpam-86	450	23	·	·	PUNCT
ejpam-86	450	24	·	·	PUNCT
ejpam-86	450	25	·	·	PUNCT
ejpam-86	450	26	,	,	PUNCT
ejpam-86	450	27	x10	x10	NOUN
ejpam-86	450	28	are	be	AUX
ejpam-86	450	29	considered	consider	VERB
ejpam-86	450	30	as	as	ADP
ejpam-86	450	31	redundant	redundant	ADJ
ejpam-86	450	32	variables	variable	NOUN
ejpam-86	450	33	.	.	PUNCT
ejpam-86	451	1	we	we	PRON
ejpam-86	451	2	generate	generate	VERB
ejpam-86	451	3	x1	x1	PROPN
ejpam-86	451	4	,	,	PUNCT
ejpam-86	451	5	·	·	PUNCT
ejpam-86	451	6	·	·	PUNCT
ejpam-86	451	7	·	·	PUNCT
ejpam-86	451	8	,	,	PUNCT
ejpam-86	451	9	x10	x10	ADP
ejpam-86	451	10	from	from	ADP
ejpam-86	451	11	(	(	PUNCT
ejpam-86	451	12	7.1	7.1	NUM
ejpam-86	451	13	)	)	PUNCT
ejpam-86	451	14	and	and	CCONJ
ejpam-86	451	15	(	(	PUNCT
ejpam-86	451	16	7.2	7.2	NUM
ejpam-86	451	17	)	)	PUNCT
ejpam-86	451	18	,	,	PUNCT
ejpam-86	451	19	and	and	CCONJ
ejpam-86	451	20	generate	generate	VERB
ejpam-86	451	21	y	y	PROPN
ejpam-86	451	22	from	from	ADP
ejpam-86	451	23	(	(	PUNCT
ejpam-86	451	24	7.3	7.3	NUM
ejpam-86	451	25	)	)	PUNCT
ejpam-86	451	26	with	with	ADP
ejpam-86	451	27	sample	sample	NOUN
ejpam-86	451	28	size	size	NOUN
ejpam-86	451	29	n	n	NOUN
ejpam-86	451	30	=	=	SYM
ejpam-86	451	31	500	500	NUM
ejpam-86	451	32	.	.	PUNCT
ejpam-86	452	1	the	the	DET
ejpam-86	452	2	rows	row	NOUN
ejpam-86	452	3	of	of	ADP
ejpam-86	452	4	e	e	NOUN
ejpam-86	452	5	are	be	AUX
ejpam-86	452	6	i.i.d	i.i.d	ADJ
ejpam-86	452	7	.	.	PUNCT
ejpam-86	452	8	pe2(0,σ	pe2(0,σ	PROPN
ejpam-86	452	9	,	,	PUNCT
ejpam-86	452	10	β	β	NOUN
ejpam-86	452	11	)	)	PUNCT
ejpam-86	452	12	with	with	ADP
ejpam-86	452	13	σ	σ	PROPN
ejpam-86	452	14	=	=	PUNCT
ejpam-86	452	15	[	[	PUNCT
ejpam-86	452	16	1	1	NUM
ejpam-86	452	17	0.5	0.5	NUM
ejpam-86	452	18	0.5	0.5	NUM
ejpam-86	452	19	2	2	NUM
ejpam-86	452	20	]	]	PUNCT
ejpam-86	452	21	,	,	PUNCT
ejpam-86	452	22	and	and	CCONJ
ejpam-86	452	23	we	we	PRON
ejpam-86	452	24	set	set	VERB
ejpam-86	452	25	β	β	X
ejpam-86	452	26	=	=	SYM
ejpam-86	452	27	2	2	X
ejpam-86	452	28	.	.	X
ejpam-86	452	29	we	we	PRON
ejpam-86	452	30	fit	fit	VERB
ejpam-86	452	31	a	a	DET
ejpam-86	452	32	type	type	NOUN
ejpam-86	452	33	i	i	PRON
ejpam-86	452	34	mvper	mvper	NOUN
ejpam-86	452	35	model	model	NOUN
ejpam-86	452	36	of	of	ADP
ejpam-86	452	37	y	y	PROPN
ejpam-86	452	38	on	on	ADP
ejpam-86	452	39	[	[	X
ejpam-86	452	40	1	1	NUM
ejpam-86	452	41	,	,	PUNCT
ejpam-86	452	42	xn×10	xn×10	PROPN
ejpam-86	452	43	]	]	PUNCT
ejpam-86	452	44	.	.	PUNCT
ejpam-86	453	1	we	we	PRON
ejpam-86	453	2	expect	expect	VERB
ejpam-86	453	3	the	the	DET
ejpam-86	453	4	algorithm	algorithm	NOUN
ejpam-86	453	5	to	to	PART
ejpam-86	453	6	pick	pick	VERB
ejpam-86	453	7	the	the	DET
ejpam-86	453	8	subset	subset	NOUN
ejpam-86	453	9	{	{	PUNCT
ejpam-86	453	10	x0	x0	PROPN
ejpam-86	453	11	,	,	PUNCT
ejpam-86	453	12	x1	x1	PROPN
ejpam-86	453	13	,	,	PUNCT
ejpam-86	453	14	x2	x2	PROPN
ejpam-86	453	15	,	,	PUNCT
ejpam-86	453	16	x3	x3	ADJ
ejpam-86	453	17	}	}	PUNCT
ejpam-86	453	18	to	to	PART
ejpam-86	453	19	be	be	AUX
ejpam-86	453	20	the	the	DET
ejpam-86	453	21	best	good	ADJ
ejpam-86	453	22	subset	subset	NOUN
ejpam-86	453	23	selected	select	VERB
ejpam-86	453	24	using	use	VERB
ejpam-86	453	25	the	the	DET
ejpam-86	453	26	minimum	minimum	NOUN
ejpam-86	453	27	aic	aic	PROPN
ejpam-86	453	28	or	or	CCONJ
ejpam-86	453	29	icomp(ifim	icomp(ifim	PROPN
ejpam-86	453	30	)	)	PUNCT
ejpam-86	453	31	criteria	criterion	NOUN
ejpam-86	453	32	,	,	PUNCT
ejpam-86	453	33	where	where	SCONJ
ejpam-86	453	34	x0	x0	PROPN
ejpam-86	453	35	denotes	denote	VERB
ejpam-86	453	36	the	the	DET
ejpam-86	453	37	constant	constant	ADJ
ejpam-86	453	38	term	term	NOUN
ejpam-86	453	39	.	.	PUNCT
ejpam-86	454	1	parameters	parameter	NOUN
ejpam-86	454	2	of	of	ADP
ejpam-86	454	3	ga	ga	PROPN
ejpam-86	454	4	are	be	AUX
ejpam-86	454	5	given	give	VERB
ejpam-86	454	6	in	in	ADP
ejpam-86	454	7	table	table	NOUN
ejpam-86	454	8	m.	m.	NOUN
ejpam-86	454	9	liu	liu	PROPN
ejpam-86	454	10	and	and	CCONJ
ejpam-86	454	11	h.	h.	PROPN
ejpam-86	454	12	bozdogan	bozdogan	PROPN
ejpam-86	454	13	/	/	SYM
ejpam-86	454	14	eur	eur	PROPN
ejpam-86	454	15	.	.	PUNCT
ejpam-86	455	1	j.	j.	PROPN
ejpam-86	455	2	pure	pure	PROPN
ejpam-86	455	3	appl	appl	PROPN
ejpam-86	455	4	.	.	PROPN
ejpam-86	455	5	math	math	PROPN
ejpam-86	455	6	,	,	PUNCT
ejpam-86	455	7	1	1	NUM
ejpam-86	455	8	(	(	PUNCT
ejpam-86	455	9	2008	2008	NUM
ejpam-86	455	10	)	)	PUNCT
ejpam-86	455	11	,	,	PUNCT
ejpam-86	455	12	(	(	PUNCT
ejpam-86	455	13	4	4	NUM
ejpam-86	455	14	-	-	SYM
ejpam-86	455	15	37	37	NUM
ejpam-86	455	16	)	)	PUNCT
ejpam-86	455	17	23	23	NUM
ejpam-86	455	18	0	0	NUM
ejpam-86	455	19	5	5	NUM
ejpam-86	455	20	10	10	NUM
ejpam-86	455	21	15	15	NUM
ejpam-86	455	22	−10	−10	NOUN
ejpam-86	455	23	0	0	NUM
ejpam-86	455	24	10	10	NUM
ejpam-86	455	25	20	20	NUM
ejpam-86	455	26	30	30	NUM
ejpam-86	455	27	40	40	NUM
ejpam-86	455	28	50	50	NUM
ejpam-86	455	29	chi−square	chi−square	ADP
ejpam-86	455	30	quantile	quantile	ADJ
ejpam-86	455	31	plot	plot	NOUN
ejpam-86	455	32	with	with	ADP
ejpam-86	455	33	real	real	ADJ
ejpam-86	455	34	β=0.3	β=0.3	NOUN
ejpam-86	456	1	o	o	NOUN
ejpam-86	456	2	rd	rd	NOUN
ejpam-86	456	3	er	er	INTJ
ejpam-86	456	4	ed	ed	NOUN
ejpam-86	456	5	s	s	PROPN
ejpam-86	456	6	qu	qu	PROPN
ejpam-86	456	7	ar	ar	PROPN
ejpam-86	456	8	ed	ed	PROPN
ejpam-86	456	9	m	m	VERB
ejpam-86	456	10	ah	ah	INTJ
ejpam-86	456	11	al	al	PROPN
ejpam-86	456	12	an	an	DET
ejpam-86	456	13	ob	ob	NOUN
ejpam-86	456	14	is	be	AUX
ejpam-86	456	15	d	d	NOUN
ejpam-86	456	16	is	be	AUX
ejpam-86	456	17	ta	ta	PROPN
ejpam-86	456	18	nc	nc	PROPN
ejpam-86	456	19	e	e	PROPN
ejpam-86	456	20	of	of	ADP
ejpam-86	456	21	r	r	NOUN
ejpam-86	456	22	es	es	X
ejpam-86	456	23	i	i	PROPN
ejpam-86	456	24	d	d	PROPN
ejpam-86	456	25	ua	ua	PROPN
ejpam-86	456	26	ls	ls	PROPN
ejpam-86	456	27	p−quantiles	p−quantiles	PROPN
ejpam-86	456	28	0	0	NUM
ejpam-86	456	29	5	5	NUM
ejpam-86	456	30	10	10	NUM
ejpam-86	456	31	15	15	NUM
ejpam-86	456	32	0	0	NUM
ejpam-86	456	33	5	5	NUM
ejpam-86	456	34	10	10	NUM
ejpam-86	456	35	15	15	NUM
ejpam-86	456	36	chi−square	chi−square	ADP
ejpam-86	456	37	quantile	quantile	ADJ
ejpam-86	456	38	plot	plot	NOUN
ejpam-86	456	39	with	with	ADP
ejpam-86	456	40	real	real	ADJ
ejpam-86	456	41	β=2	β=2	ADP
ejpam-86	457	1	o	o	X
ejpam-86	457	2	rd	rd	NOUN
ejpam-86	458	1	er	er	INTJ
ejpam-86	458	2	ed	ed	NOUN
ejpam-86	458	3	s	s	PROPN
ejpam-86	458	4	qu	qu	PROPN
ejpam-86	458	5	ar	ar	PROPN
ejpam-86	458	6	ed	ed	PROPN
ejpam-86	458	7	m	m	VERB
ejpam-86	458	8	ah	ah	INTJ
ejpam-86	458	9	al	al	PROPN
ejpam-86	458	10	an	an	DET
ejpam-86	458	11	ob	ob	NOUN
ejpam-86	458	12	is	be	AUX
ejpam-86	458	13	d	d	NOUN
ejpam-86	458	14	is	be	AUX
ejpam-86	458	15	ta	ta	PROPN
ejpam-86	458	16	nc	nc	PROPN
ejpam-86	458	17	e	e	PROPN
ejpam-86	458	18	of	of	ADP
ejpam-86	458	19	r	r	NOUN
ejpam-86	458	20	es	es	X
ejpam-86	458	21	i	i	PROPN
ejpam-86	458	22	d	d	PROPN
ejpam-86	458	23	ua	ua	PROPN
ejpam-86	458	24	ls	ls	PROPN
ejpam-86	458	25	p−quantiles	p−quantile	NOUN
ejpam-86	458	26	figure	figure	VERB
ejpam-86	458	27	4	4	NUM
ejpam-86	458	28	:	:	PUNCT
ejpam-86	458	29	the	the	DET
ejpam-86	458	30	q	q	NOUN
ejpam-86	458	31	-	-	PUNCT
ejpam-86	458	32	q	q	NOUN
ejpam-86	458	33	plots	plot	NOUN
ejpam-86	458	34	for	for	ADP
ejpam-86	458	35	type	type	NOUN
ejpam-86	458	36	i	i	PROPN
ejpam-86	458	37	mvper	mvper	NOUN
ejpam-86	458	38	models	model	NOUN
ejpam-86	458	39	in	in	ADP
ejpam-86	458	40	one	one	NUM
ejpam-86	458	41	run	run	NOUN
ejpam-86	458	42	of	of	ADP
ejpam-86	458	43	the	the	DET
ejpam-86	458	44	simulation	simulation	NOUN
ejpam-86	458	45	example	example	NOUN
ejpam-86	459	1	1	1	NUM
ejpam-86	459	2	.	.	SYM
ejpam-86	459	3	0	0	NUM
ejpam-86	460	1	10	10	NUM
ejpam-86	460	2	20	20	NUM
ejpam-86	460	3	30	30	NUM
ejpam-86	460	4	40	40	NUM
ejpam-86	460	5	50	50	NUM
ejpam-86	460	6	1060	1060	NUM
ejpam-86	460	7	1080	1080	NUM
ejpam-86	460	8	1100	1100	NUM
ejpam-86	460	9	1120	1120	NUM
ejpam-86	460	10	1140	1140	NUM
ejpam-86	460	11	1160	1160	NUM
ejpam-86	460	12	1180	1180	NUM
ejpam-86	460	13	1200	1200	NUM
ejpam-86	460	14	1220	1220	NUM
ejpam-86	460	15	1240	1240	NUM
ejpam-86	460	16	a	a	DET
ejpam-86	460	17	typical	typical	ADJ
ejpam-86	460	18	ga	ga	NOUN
ejpam-86	460	19	procedure	procedure	NOUN
ejpam-86	460	20	in	in	ADP
ejpam-86	460	21	one	one	NUM
ejpam-86	460	22	run	run	NOUN
ejpam-86	460	23	of	of	ADP
ejpam-86	460	24	simulation	simulation	NOUN
ejpam-86	460	25	example	example	NOUN
ejpam-86	460	26	1	1	NUM
ejpam-86	460	27	−	−	PROPN
ejpam-86	461	1	lo	lo	PROPN
ejpam-86	461	2	gl	gl	PROPN
ejpam-86	461	3	ik	ik	PROPN
ejpam-86	461	4	el	el	PROPN
ejpam-86	461	5	ih	ih	PROPN
ejpam-86	462	1	oo	oo	INTJ
ejpam-86	462	2	d	d	PROPN
ejpam-86	462	3	of	of	ADP
ejpam-86	462	4	th	th	X
ejpam-86	462	5	e	e	NOUN
ejpam-86	462	6	fit	fit	NOUN
ejpam-86	462	7	te	te	PROPN
ejpam-86	462	8	d	d	PROPN
ejpam-86	462	9	m	m	PROPN
ejpam-86	462	10	od	od	PROPN
ejpam-86	462	11	el	el	PROPN
ejpam-86	462	12	ga	ga	PROPN
ejpam-86	462	13	generations	generation	NOUN
ejpam-86	462	14	figure	figure	VERB
ejpam-86	462	15	5	5	NUM
ejpam-86	462	16	:	:	PUNCT
ejpam-86	462	17	a	a	DET
ejpam-86	462	18	typical	typical	ADJ
ejpam-86	462	19	ga	ga	NOUN
ejpam-86	462	20	process	process	NOUN
ejpam-86	462	21	in	in	ADP
ejpam-86	462	22	one	one	NUM
ejpam-86	462	23	run	run	NOUN
ejpam-86	462	24	of	of	ADP
ejpam-86	462	25	the	the	DET
ejpam-86	462	26	simulation	simulation	NOUN
ejpam-86	462	27	example	example	NOUN
ejpam-86	462	28	1	1	NUM
ejpam-86	462	29	.	.	NOUN
ejpam-86	462	30	3	3	X
ejpam-86	462	31	.	.	NOUN
ejpam-86	462	32	method	method	NOUN
ejpam-86	462	33	of	of	ADP
ejpam-86	462	34	moments	moment	NOUN
ejpam-86	462	35	(	(	PUNCT
ejpam-86	462	36	mom	mom	NOUN
ejpam-86	462	37	)	)	PUNCT
ejpam-86	462	38	estimates	estimate	NOUN
ejpam-86	462	39	are	be	AUX
ejpam-86	462	40	used	use	VERB
ejpam-86	462	41	as	as	ADP
ejpam-86	462	42	the	the	DET
ejpam-86	462	43	starting	starting	NOUN
ejpam-86	462	44	values	value	NOUN
ejpam-86	462	45	for	for	ADP
ejpam-86	462	46	the	the	DET
ejpam-86	462	47	ga	ga	PROPN
ejpam-86	462	48	process	process	NOUN
ejpam-86	462	49	.	.	PUNCT
ejpam-86	463	1	for	for	ADP
ejpam-86	463	2	11	11	NUM
ejpam-86	463	3	independent	independent	ADJ
ejpam-86	463	4	variables	variable	NOUN
ejpam-86	463	5	(	(	PUNCT
ejpam-86	463	6	including	include	VERB
ejpam-86	463	7	the	the	DET
ejpam-86	463	8	constant	constant	ADJ
ejpam-86	463	9	term	term	NOUN
ejpam-86	463	10	)	)	PUNCT
ejpam-86	463	11	,	,	PUNCT
ejpam-86	463	12	there	there	PRON
ejpam-86	463	13	are	be	VERB
ejpam-86	463	14	211	211	NUM
ejpam-86	463	15	=	=	SYM
ejpam-86	463	16	2048	2048	NUM
ejpam-86	463	17	possible	possible	ADJ
ejpam-86	463	18	subsets	subset	NOUN
ejpam-86	463	19	.	.	PUNCT
ejpam-86	464	1	each	each	DET
ejpam-86	464	2	subset	subset	NOUN
ejpam-86	464	3	,	,	PUNCT
ejpam-86	464	4	or	or	CCONJ
ejpam-86	464	5	model	model	NOUN
ejpam-86	464	6	,	,	PUNCT
ejpam-86	464	7	is	be	AUX
ejpam-86	464	8	encoded	encode	VERB
ejpam-86	464	9	as	as	ADP
ejpam-86	464	10	a	a	DET
ejpam-86	464	11	binary	binary	ADJ
ejpam-86	464	12	string	string	NOUN
ejpam-86	464	13	with	with	ADP
ejpam-86	464	14	the	the	DET
ejpam-86	464	15	fixed	fix	VERB
ejpam-86	464	16	length	length	NOUN
ejpam-86	464	17	as	as	ADP
ejpam-86	464	18	the	the	DET
ejpam-86	464	19	total	total	ADJ
ejpam-86	464	20	number	number	NOUN
ejpam-86	464	21	of	of	ADP
ejpam-86	464	22	available	available	ADJ
ejpam-86	464	23	independent	independent	ADJ
ejpam-86	464	24	or	or	CCONJ
ejpam-86	464	25	predictor	predictor	NOUN
ejpam-86	464	26	variables	variable	NOUN
ejpam-86	464	27	.	.	PUNCT
ejpam-86	465	1	each	each	DET
ejpam-86	465	2	locus	locus	NOUN
ejpam-86	465	3	in	in	ADP
ejpam-86	465	4	the	the	DET
ejpam-86	465	5	string	string	NOUN
ejpam-86	465	6	is	be	AUX
ejpam-86	465	7	a	a	DET
ejpam-86	465	8	binary	binary	ADJ
ejpam-86	465	9	code	code	NOUN
ejpam-86	465	10	indicating	indicate	VERB
ejpam-86	465	11	the	the	DET
ejpam-86	465	12	presence	presence	NOUN
ejpam-86	465	13	(	(	PUNCT
ejpam-86	465	14	1	1	NUM
ejpam-86	465	15	)	)	PUNCT
ejpam-86	465	16	or	or	CCONJ
ejpam-86	465	17	absence	absence	NOUN
ejpam-86	465	18	(	(	PUNCT
ejpam-86	465	19	0	0	NUM
ejpam-86	465	20	)	)	PUNCT
ejpam-86	465	21	of	of	ADP
ejpam-86	465	22	a	a	DET
ejpam-86	465	23	given	give	VERB
ejpam-86	465	24	predictor	predictor	NOUN
ejpam-86	465	25	variable	variable	NOUN
ejpam-86	465	26	in	in	ADP
ejpam-86	465	27	the	the	DET
ejpam-86	465	28	model	model	NOUN
ejpam-86	465	29	.	.	PUNCT
ejpam-86	466	1	for	for	ADP
ejpam-86	466	2	example	example	NOUN
ejpam-86	466	3	,	,	PUNCT
ejpam-86	466	4	the	the	DET
ejpam-86	466	5	string	string	NOUN
ejpam-86	466	6	101011	101011	NUM
ejpam-86	466	7	represents	represent	VERB
ejpam-86	466	8	a	a	DET
ejpam-86	466	9	model	model	NOUN
ejpam-86	466	10	,	,	PUNCT
ejpam-86	466	11	where	where	SCONJ
ejpam-86	466	12	constant	constant	ADJ
ejpam-86	466	13	term	term	NOUN
ejpam-86	466	14	is	be	AUX
ejpam-86	466	15	included	include	VERB
ejpam-86	466	16	in	in	ADP
ejpam-86	466	17	the	the	DET
ejpam-86	466	18	model	model	NOUN
ejpam-86	466	19	,	,	PUNCT
ejpam-86	466	20	variable	variable	NOUN
ejpam-86	466	21	x1	x1	PROPN
ejpam-86	466	22	is	be	AUX
ejpam-86	466	23	excluded	exclude	VERB
ejpam-86	466	24	from	from	ADP
ejpam-86	466	25	the	the	DET
ejpam-86	466	26	model	model	NOUN
ejpam-86	466	27	,	,	PUNCT
ejpam-86	466	28	variable	variable	ADJ
ejpam-86	466	29	x2	x2	PROPN
ejpam-86	466	30	is	be	AUX
ejpam-86	466	31	included	include	VERB
ejpam-86	466	32	in	in	ADP
ejpam-86	466	33	the	the	DET
ejpam-86	466	34	model	model	NOUN
ejpam-86	466	35	,	,	PUNCT
ejpam-86	466	36	and	and	CCONJ
ejpam-86	466	37	so	so	ADV
ejpam-86	466	38	on	on	ADV
ejpam-86	466	39	.	.	PUNCT
ejpam-86	467	1	figure	figure	NOUN
ejpam-86	467	2	6	6	NUM
ejpam-86	467	3	shows	show	VERB
ejpam-86	467	4	the	the	DET
ejpam-86	467	5	plots	plot	NOUN
ejpam-86	467	6	of	of	ADP
ejpam-86	467	7	all	all	DET
ejpam-86	467	8	subsets	subset	NOUN
ejpam-86	467	9	evaluated	evaluate	VERB
ejpam-86	467	10	in	in	ADP
ejpam-86	467	11	the	the	DET
ejpam-86	467	12	ga	ga	PROPN
ejpam-86	467	13	process	process	NOUN
ejpam-86	467	14	.	.	PUNCT
ejpam-86	468	1	both	both	DET
ejpam-86	468	2	ga	ga	NOUN
ejpam-86	468	3	processes	process	NOUN
ejpam-86	468	4	converge	converge	VERB
ejpam-86	468	5	to	to	ADP
ejpam-86	468	6	the	the	DET
ejpam-86	468	7	true	true	ADJ
ejpam-86	468	8	model	model	NOUN
ejpam-86	468	9	.	.	PUNCT
ejpam-86	469	1	the	the	DET
ejpam-86	469	2	top	top	ADJ
ejpam-86	469	3	5	5	NUM
ejpam-86	469	4	subsets	subset	NOUN
ejpam-86	469	5	selected	select	VERB
ejpam-86	469	6	by	by	ADP
ejpam-86	469	7	the	the	DET
ejpam-86	469	8	minimum	minimum	ADJ
ejpam-86	469	9	icomp(ifim	icomp(ifim	NOUN
ejpam-86	469	10	)	)	PUNCT
ejpam-86	469	11	and	and	CCONJ
ejpam-86	469	12	aic	aic	PROPN
ejpam-86	469	13	are	be	AUX
ejpam-86	469	14	summarized	summarize	VERB
ejpam-86	469	15	in	in	ADP
ejpam-86	469	16	table	table	NOUN
ejpam-86	469	17	4	4	NUM
ejpam-86	469	18	.	.	PUNCT
ejpam-86	469	19	from	from	ADP
ejpam-86	469	20	table	table	NOUN
ejpam-86	469	21	4	4	NUM
ejpam-86	469	22	,	,	PUNCT
ejpam-86	469	23	we	we	PRON
ejpam-86	469	24	see	see	VERB
ejpam-86	469	25	that	that	SCONJ
ejpam-86	469	26	the	the	DET
ejpam-86	469	27	true	true	ADJ
ejpam-86	469	28	model	model	NOUN
ejpam-86	469	29	is	be	AUX
ejpam-86	469	30	selected	select	VERB
ejpam-86	469	31	as	as	ADP
ejpam-86	469	32	the	the	DET
ejpam-86	469	33	best	good	ADJ
ejpam-86	469	34	subset	subset	NOUN
ejpam-86	469	35	according	accord	VERB
ejpam-86	469	36	to	to	ADP
ejpam-86	469	37	minimum	minimum	ADJ
ejpam-86	469	38	icomp(ifim	icomp(ifim	NOUN
ejpam-86	469	39	)	)	PUNCT
ejpam-86	469	40	and	and	CCONJ
ejpam-86	469	41	aic	aic	PROPN
ejpam-86	469	42	.	.	PUNCT
ejpam-86	470	1	the	the	DET
ejpam-86	470	2	mles	mle	NOUN
ejpam-86	470	3	of	of	ADP
ejpam-86	470	4	the	the	DET
ejpam-86	470	5	best	good	ADJ
ejpam-86	470	6	model	model	NOUN
ejpam-86	470	7	parameters	parameter	NOUN
ejpam-86	470	8	chosen	choose	VERB
ejpam-86	470	9	by	by	ADP
ejpam-86	470	10	icomp	icomp	PROPN
ejpam-86	470	11	are	be	AUX
ejpam-86	470	12	:	:	PUNCT
ejpam-86	470	13	β̂mle	β̂mle	PUNCT
ejpam-86	470	14	=	=	SYM
ejpam-86	471	1	1.7488	1.7488	NUM
ejpam-86	471	2	,	,	PUNCT
ejpam-86	471	3	m.	m.	NOUN
ejpam-86	471	4	liu	liu	PROPN
ejpam-86	471	5	and	and	CCONJ
ejpam-86	471	6	h.	h.	PROPN
ejpam-86	471	7	bozdogan	bozdogan	PROPN
ejpam-86	471	8	/	/	SYM
ejpam-86	471	9	eur	eur	PROPN
ejpam-86	471	10	.	.	PUNCT
ejpam-86	472	1	j.	j.	PROPN
ejpam-86	472	2	pure	pure	PROPN
ejpam-86	472	3	appl	appl	PROPN
ejpam-86	472	4	.	.	PROPN
ejpam-86	472	5	math	math	PROPN
ejpam-86	472	6	,	,	PUNCT
ejpam-86	472	7	1	1	NUM
ejpam-86	472	8	(	(	PUNCT
ejpam-86	472	9	2008	2008	NUM
ejpam-86	472	10	)	)	PUNCT
ejpam-86	472	11	,	,	PUNCT
ejpam-86	472	12	(	(	PUNCT
ejpam-86	472	13	4	4	NUM
ejpam-86	472	14	-	-	SYM
ejpam-86	472	15	37	37	NUM
ejpam-86	472	16	)	)	PUNCT
ejpam-86	472	17	24	24	NUM
ejpam-86	472	18	0	0	NUM
ejpam-86	472	19	10	10	NUM
ejpam-86	472	20	20	20	NUM
ejpam-86	472	21	30	30	NUM
ejpam-86	472	22	0	0	NUM
ejpam-86	472	23	10	10	NUM
ejpam-86	472	24	20	20	NUM
ejpam-86	472	25	30	30	NUM
ejpam-86	472	26	2000	2000	NUM
ejpam-86	472	27	2500	2500	NUM
ejpam-86	472	28	3000	3000	NUM
ejpam-86	472	29	3500	3500	NUM
ejpam-86	472	30	4000	4000	NUM
ejpam-86	472	31	ga	ga	NOUN
ejpam-86	472	32	generation	generation	NOUN
ejpam-86	472	33	icomp(ifim	icomp(ifim	NOUN
ejpam-86	472	34	)	)	PUNCT
ejpam-86	472	35	ga	ga	PROPN
ejpam-86	472	36	process	process	NOUN
ejpam-86	472	37	ga	ga	PROPN
ejpam-86	472	38	population	population	PROPN
ejpam-86	473	1	ic	ic	INTJ
ejpam-86	473	2	o	o	NOUN
ejpam-86	473	3	m	m	VERB
ejpam-86	473	4	p	p	X
ejpam-86	473	5	(	(	PUNCT
ejpam-86	473	6	i	i	PRON
ejpam-86	473	7	f	f	PROPN
ejpam-86	474	1	i	i	NOUN
ejpam-86	474	2	m	m	VERB
ejpam-86	474	3	)	)	PUNCT
ejpam-86	474	4	0	0	NUM
ejpam-86	475	1	10	10	NUM
ejpam-86	475	2	20	20	NUM
ejpam-86	475	3	30	30	NUM
ejpam-86	475	4	0	0	NUM
ejpam-86	475	5	10	10	NUM
ejpam-86	475	6	20	20	NUM
ejpam-86	475	7	30	30	NUM
ejpam-86	475	8	2000	2000	NUM
ejpam-86	475	9	2500	2500	NUM
ejpam-86	475	10	3000	3000	NUM
ejpam-86	475	11	3500	3500	NUM
ejpam-86	475	12	4000	4000	NUM
ejpam-86	475	13	4500	4500	NUM
ejpam-86	475	14	5000	5000	NUM
ejpam-86	475	15	ga	ga	NOUN
ejpam-86	475	16	generation	generation	NOUN
ejpam-86	475	17	aic	aic	PROPN
ejpam-86	475	18	ga	ga	PROPN
ejpam-86	475	19	process	process	NOUN
ejpam-86	475	20	ga	ga	PROPN
ejpam-86	475	21	population	population	PROPN
ejpam-86	475	22	a	a	DET
ejpam-86	475	23	ic	ic	PROPN
ejpam-86	475	24	figure	figure	NOUN
ejpam-86	475	25	6	6	NUM
ejpam-86	475	26	:	:	PUNCT
ejpam-86	475	27	the	the	DET
ejpam-86	475	28	ga	ga	NOUN
ejpam-86	475	29	process	process	NOUN
ejpam-86	475	30	for	for	ADP
ejpam-86	475	31	the	the	DET
ejpam-86	475	32	simulation	simulation	NOUN
ejpam-86	475	33	example	example	NOUN
ejpam-86	475	34	2	2	NUM
ejpam-86	475	35	for	for	ADP
ejpam-86	475	36	icomp	icomp	NOUN
ejpam-86	475	37	and	and	CCONJ
ejpam-86	475	38	aic	aic	PROPN
ejpam-86	475	39	.	.	PUNCT
ejpam-86	476	1	σ̂mle	σ̂mle	NOUN
ejpam-86	476	2	=	=	PUNCT
ejpam-86	477	1	[	[	PUNCT
ejpam-86	477	2	0.8788	0.8788	NUM
ejpam-86	477	3	0.4007	0.4007	NUM
ejpam-86	477	4	0.4007	0.4007	NUM
ejpam-86	477	5	1.7192	1.7192	NUM
ejpam-86	477	6	]	]	PUNCT
ejpam-86	477	7	and	and	CCONJ
ejpam-86	477	8	b̂mle	b̂mle	PROPN
ejpam-86	477	9	=	=	SYM
ejpam-86	477	10			NOUN
ejpam-86	477	11			VERB
ejpam-86	477	12	7.2275	7.2275	NUM
ejpam-86	477	13	−5.4046	−5.4046	NOUN
ejpam-86	477	14	1.0004	1.0004	NUM
ejpam-86	477	15	0.5128	0.5128	NUM
ejpam-86	477	16	0.5142	0.5142	NUM
ejpam-86	477	17	0.0951	0.0951	NUM
ejpam-86	477	18	0.3587	0.3587	NUM
ejpam-86	477	19	0.2335	0.2335	NUM
ejpam-86	477	20			PROPN
ejpam-86	477	21			NOUN
ejpam-86	477	22	.	.	PUNCT
ejpam-86	478	1	our	our	PRON
ejpam-86	478	2	results	result	NOUN
ejpam-86	478	3	show	show	VERB
ejpam-86	478	4	that	that	SCONJ
ejpam-86	478	5	the	the	DET
ejpam-86	478	6	ga	ga	PROPN
ejpam-86	478	7	on	on	ADP
ejpam-86	478	8	ga	ga	PROPN
ejpam-86	478	9	approach	approach	NOUN
ejpam-86	478	10	with	with	ADP
ejpam-86	478	11	icomp(ifim	icomp(ifim	NOUN
ejpam-86	478	12	)	)	PUNCT
ejpam-86	478	13	or	or	CCONJ
ejpam-86	478	14	aic	aic	PROPN
ejpam-86	478	15	as	as	SCONJ
ejpam-86	478	16	the	the	DET
ejpam-86	478	17	fitness	fitness	NOUN
ejpam-86	478	18	function	function	NOUN
ejpam-86	478	19	can	can	AUX
ejpam-86	478	20	detect	detect	VERB
ejpam-86	478	21	the	the	DET
ejpam-86	478	22	true	true	ADJ
ejpam-86	478	23	relationship	relationship	NOUN
ejpam-86	478	24	and	and	CCONJ
ejpam-86	478	25	pick	pick	VERB
ejpam-86	478	26	the	the	DET
ejpam-86	478	27	correct	correct	ADJ
ejpam-86	478	28	models	model	NOUN
ejpam-86	478	29	.	.	PUNCT
ejpam-86	479	1	we	we	PRON
ejpam-86	479	2	notice	notice	VERB
ejpam-86	479	3	that	that	SCONJ
ejpam-86	479	4	the	the	DET
ejpam-86	479	5	coefficients	coefficient	NOUN
ejpam-86	479	6	of	of	ADP
ejpam-86	479	7	the	the	DET
ejpam-86	479	8	redundant	redundant	ADJ
ejpam-86	479	9	variables	variable	NOUN
ejpam-86	479	10	in	in	ADP
ejpam-86	479	11	the	the	DET
ejpam-86	479	12	top	top	ADJ
ejpam-86	479	13	5	5	NUM
ejpam-86	479	14	best	good	ADJ
ejpam-86	479	15	subset	subset	NOUN
ejpam-86	479	16	models	model	NOUN
ejpam-86	479	17	selected	select	VERB
ejpam-86	479	18	are	be	AUX
ejpam-86	479	19	small	small	ADJ
ejpam-86	479	20	,	,	PUNCT
ejpam-86	479	21	which	which	PRON
ejpam-86	479	22	means	mean	VERB
ejpam-86	479	23	that	that	SCONJ
ejpam-86	479	24	these	these	DET
ejpam-86	479	25	redundant	redundant	ADJ
ejpam-86	479	26	variables	variable	NOUN
ejpam-86	479	27	can	can	AUX
ejpam-86	479	28	be	be	AUX
ejpam-86	479	29	ignored	ignore	VERB
ejpam-86	479	30	.	.	PUNCT
ejpam-86	480	1	table	table	NOUN
ejpam-86	480	2	3	3	NUM
ejpam-86	480	3	:	:	PUNCT
ejpam-86	480	4	ga	ga	NOUN
ejpam-86	480	5	parameters	parameter	NOUN
ejpam-86	480	6	of	of	ADP
ejpam-86	480	7	the	the	DET
ejpam-86	480	8	simulation	simulation	NOUN
ejpam-86	480	9	example	example	NOUN
ejpam-86	480	10	2	2	NUM
ejpam-86	480	11	and	and	CCONJ
ejpam-86	480	12	the	the	DET
ejpam-86	480	13	real	real	ADJ
ejpam-86	480	14	data	datum	NOUN
ejpam-86	480	15	example	example	NOUN
ejpam-86	480	16	.	.	PUNCT
ejpam-86	481	1	ng	ng	PROPN
ejpam-86	481	2	ps	ps	PROPN
ejpam-86	481	3	pc	pc	PROPN
ejpam-86	481	4	pm	pm	NOUN
ejpam-86	481	5	pe	pe	INTJ
ejpam-86	481	6	ct	ct	NUM
ejpam-86	481	7	l	l	NOUN
ejpam-86	481	8	elitism	elitism	NOUN
ejpam-86	481	9	subset	subset	VERB
ejpam-86	481	10	ga	ga	PROPN
ejpam-86	481	11	30	30	NUM
ejpam-86	481	12	30	30	NUM
ejpam-86	481	13	0.7	0.7	NUM
ejpam-86	481	14	0.01	0.01	NUM
ejpam-86	481	15	0.5	0.5	NUM
ejpam-86	481	16	3	3	NUM
ejpam-86	481	17	yes	yes	NOUN
ejpam-86	481	18	model	model	PROPN
ejpam-86	481	19	estimation	estimation	PROPN
ejpam-86	481	20	ga	ga	PROPN
ejpam-86	482	1	50	50	NUM
ejpam-86	482	2	100	100	NUM
ejpam-86	482	3	0.7	0.7	NUM
ejpam-86	482	4	0.01	0.01	NUM
ejpam-86	482	5	0.5	0.5	NUM
ejpam-86	482	6	3	3	NUM
ejpam-86	482	7	32	32	NUM
ejpam-86	482	8	yes	yes	NOUN
ejpam-86	482	9	7.2	7.2	NUM
ejpam-86	482	10	.	.	PUNCT
ejpam-86	483	1	a	a	DET
ejpam-86	483	2	real	real	ADJ
ejpam-86	483	3	example	example	NOUN
ejpam-86	483	4	:	:	PUNCT
ejpam-86	483	5	a	a	DET
ejpam-86	483	6	macroeconomic	macroeconomic	ADJ
ejpam-86	483	7	time	time	NOUN
ejpam-86	483	8	series	series	PROPN
ejpam-86	483	9	data	datum	NOUN
ejpam-86	483	10	in	in	ADP
ejpam-86	483	11	this	this	DET
ejpam-86	483	12	last	last	ADJ
ejpam-86	483	13	example	example	NOUN
ejpam-86	483	14	,	,	PUNCT
ejpam-86	483	15	we	we	PRON
ejpam-86	483	16	use	use	VERB
ejpam-86	483	17	the	the	DET
ejpam-86	483	18	famous	famous	ADJ
ejpam-86	483	19	quarterly	quarterly	ADJ
ejpam-86	483	20	macroeconomic	macroeconomic	ADJ
ejpam-86	483	21	time	time	NOUN
ejpam-86	483	22	series	series	PROPN
ejpam-86	483	23	data	data	PROPN
ejpam-86	483	24	for	for	ADP
ejpam-86	483	25	the	the	DET
ejpam-86	483	26	united	united	PROPN
ejpam-86	483	27	kingdom	kingdom	NOUN
ejpam-86	483	28	during	during	ADP
ejpam-86	483	29	1948	1948	NUM
ejpam-86	483	30	-	-	SYM
ejpam-86	483	31	1956	1956	NUM
ejpam-86	483	32	.	.	PUNCT
ejpam-86	484	1	this	this	DET
ejpam-86	484	2	data	datum	NOUN
ejpam-86	484	3	set	set	VERB
ejpam-86	484	4	consists	consist	NOUN
ejpam-86	484	5	of	of	ADP
ejpam-86	484	6	n	n	NOUN
ejpam-86	484	7	=	=	NUM
ejpam-86	484	8	36	36	NUM
ejpam-86	484	9	quarterly	quarterly	ADJ
ejpam-86	484	10	observations	observation	NOUN
ejpam-86	484	11	,	,	PUNCT
ejpam-86	484	12	starting	start	VERB
ejpam-86	484	13	with	with	ADP
ejpam-86	484	14	the	the	DET
ejpam-86	484	15	first	first	ADJ
ejpam-86	484	16	quarter	quarter	NOUN
ejpam-86	484	17	of	of	ADP
ejpam-86	484	18	1948	1948	NUM
ejpam-86	484	19	and	and	CCONJ
ejpam-86	484	20	ending	end	VERB
ejpam-86	484	21	with	with	ADP
ejpam-86	484	22	the	the	DET
ejpam-86	484	23	last	last	ADJ
ejpam-86	484	24	quarter	quarter	NOUN
ejpam-86	484	25	of	of	ADP
ejpam-86	484	26	1956	1956	NUM
ejpam-86	484	27	.	.	PUNCT
ejpam-86	485	1	all	all	DET
ejpam-86	485	2	the	the	DET
ejpam-86	485	3	n	n	NOUN
ejpam-86	485	4	=	=	SYM
ejpam-86	485	5	36	36	NUM
ejpam-86	485	6	observations	observation	NOUN
ejpam-86	485	7	are	be	AUX
ejpam-86	485	8	used	use	VERB
ejpam-86	485	9	in	in	ADP
ejpam-86	485	10	our	our	PRON
ejpam-86	485	11	analysis	analysis	NOUN
ejpam-86	485	12	.	.	PUNCT
ejpam-86	486	1	the	the	DET
ejpam-86	486	2	descriptions	description	NOUN
ejpam-86	486	3	of	of	ADP
ejpam-86	486	4	5	5	NUM
ejpam-86	486	5	response	response	NOUN
ejpam-86	486	6	variables	variable	NOUN
ejpam-86	486	7	and	and	CCONJ
ejpam-86	486	8	5	5	NUM
ejpam-86	486	9	independent	independent	ADJ
ejpam-86	486	10	variables	variable	NOUN
ejpam-86	486	11	from	from	ADP
ejpam-86	486	12	klein	klein	PROPN
ejpam-86	486	13	et	et	PROPN
ejpam-86	486	14	al	al	PROPN
ejpam-86	486	15	.	.	PROPN
ejpam-86	487	1	(	(	PUNCT
ejpam-86	487	2	1961	1961	NUM
ejpam-86	487	3	)	)	PUNCT
ejpam-86	487	4	are	be	AUX
ejpam-86	487	5	given	give	VERB
ejpam-86	487	6	in	in	ADP
ejpam-86	487	7	table	table	NOUN
ejpam-86	487	8	5	5	NUM
ejpam-86	487	9	.	.	PUNCT
ejpam-86	488	1	now	now	ADV
ejpam-86	488	2	for	for	ADP
ejpam-86	488	3	this	this	DET
ejpam-86	488	4	data	datum	NOUN
ejpam-86	488	5	set	set	NOUN
ejpam-86	488	6	,	,	PUNCT
ejpam-86	488	7	we	we	PRON
ejpam-86	488	8	fit	fit	VERB
ejpam-86	488	9	both	both	DET
ejpam-86	488	10	type	type	NOUN
ejpam-86	488	11	i	i	PRON
ejpam-86	488	12	and	and	CCONJ
ejpam-86	488	13	type	type	PROPN
ejpam-86	488	14	ii	ii	PROPN
ejpam-86	488	15	mvper	mvper	NOUN
ejpam-86	488	16	models	model	NOUN
ejpam-86	488	17	to	to	PART
ejpam-86	488	18	determine	determine	VERB
ejpam-86	488	19	the	the	DET
ejpam-86	488	20	best	good	ADJ
ejpam-86	488	21	fitting	fitting	ADJ
ejpam-86	488	22	model	model	NOUN
ejpam-86	488	23	.	.	PUNCT
ejpam-86	489	1	for	for	ADP
ejpam-86	489	2	type	type	NOUN
ejpam-86	489	3	ii	ii	PROPN
ejpam-86	489	4	mvper	mvper	NOUN
ejpam-86	489	5	model	model	NOUN
ejpam-86	489	6	,	,	PUNCT
ejpam-86	489	7	we	we	PRON
ejpam-86	489	8	assume	assume	VERB
ejpam-86	489	9	φ	φ	PROPN
ejpam-86	489	10	=	=	PUNCT
ejpam-86	489	11	in	in	ADP
ejpam-86	489	12	with	with	ADP
ejpam-86	489	13	sample	sample	NOUN
ejpam-86	489	14	size	size	NOUN
ejpam-86	489	15	n	n	PROPN
ejpam-86	489	16	=	=	SYM
ejpam-86	489	17	36	36	NUM
ejpam-86	489	18	.	.	PUNCT
ejpam-86	489	19	method	method	NOUN
ejpam-86	489	20	of	of	ADP
ejpam-86	489	21	moments	moment	NOUN
ejpam-86	489	22	(	(	PUNCT
ejpam-86	489	23	mom	mom	NOUN
ejpam-86	489	24	)	)	PUNCT
ejpam-86	489	25	estimates	estimate	NOUN
ejpam-86	489	26	are	be	AUX
ejpam-86	489	27	used	use	VERB
ejpam-86	489	28	m.	m.	NOUN
ejpam-86	489	29	liu	liu	PROPN
ejpam-86	489	30	and	and	CCONJ
ejpam-86	489	31	h.	h.	PROPN
ejpam-86	489	32	bozdogan	bozdogan	PROPN
ejpam-86	489	33	/	/	SYM
ejpam-86	489	34	eur	eur	PROPN
ejpam-86	489	35	.	.	PUNCT
ejpam-86	490	1	j.	j.	PROPN
ejpam-86	490	2	pure	pure	PROPN
ejpam-86	490	3	appl	appl	PROPN
ejpam-86	490	4	.	.	PROPN
ejpam-86	490	5	math	math	PROPN
ejpam-86	490	6	,	,	PUNCT
ejpam-86	490	7	1	1	NUM
ejpam-86	490	8	(	(	PUNCT
ejpam-86	490	9	2008	2008	NUM
ejpam-86	490	10	)	)	PUNCT
ejpam-86	490	11	,	,	PUNCT
ejpam-86	490	12	(	(	PUNCT
ejpam-86	490	13	4	4	NUM
ejpam-86	490	14	-	-	SYM
ejpam-86	490	15	37	37	NUM
ejpam-86	490	16	)	)	PUNCT
ejpam-86	490	17	25	25	NUM
ejpam-86	490	18	table	table	NOUN
ejpam-86	490	19	4	4	NUM
ejpam-86	490	20	:	:	PUNCT
ejpam-86	490	21	top	top	ADJ
ejpam-86	490	22	5	5	NUM
ejpam-86	490	23	subsets	subset	NOUN
ejpam-86	490	24	according	accord	VERB
ejpam-86	490	25	to	to	ADP
ejpam-86	490	26	minimum	minimum	NOUN
ejpam-86	490	27	aic	aic	PROPN
ejpam-86	490	28	and	and	CCONJ
ejpam-86	490	29	icomp(ifim	icomp(ifim	NOUN
ejpam-86	490	30	)	)	PUNCT
ejpam-86	490	31	.	.	PUNCT
ejpam-86	491	1	ranking	rank	VERB
ejpam-86	491	2	1	1	NUM
ejpam-86	491	3	2	2	NUM
ejpam-86	491	4	3	3	NUM
ejpam-86	491	5	4	4	NUM
ejpam-86	491	6	5	5	NUM
ejpam-86	491	7	subsets	subset	NOUN
ejpam-86	491	8	selected	select	VERB
ejpam-86	491	9	by	by	ADP
ejpam-86	491	10	icomp(ifim	icomp(ifim	NOUN
ejpam-86	491	11	)	)	PUNCT
ejpam-86	491	12	subset	subset	VERB
ejpam-86	491	13	11110000000	11110000000	NUM
ejpam-86	491	14	11110100000	11110100000	NUM
ejpam-86	491	15	11111000000	11111000000	NUM
ejpam-86	491	16	11110000100	11110000100	NUM
ejpam-86	491	17	11110001000	11110001000	NUM
ejpam-86	491	18	β̂mle	β̂mle	PUNCT
ejpam-86	491	19	1.7488	1.7488	NUM
ejpam-86	491	20	1.7472	1.7472	NUM
ejpam-86	491	21	1.7464	1.7464	NUM
ejpam-86	491	22	1.7456	1.7456	NUM
ejpam-86	491	23	1.7408	1.7408	NUM
ejpam-86	491	24	icomp	icomp	NOUN
ejpam-86	491	25	2182.4	2182.4	NUM
ejpam-86	491	26	2187.7	2187.7	NUM
ejpam-86	491	27	2188.4	2188.4	NUM
ejpam-86	491	28	2188.8	2188.8	NUM
ejpam-86	491	29	2190.1	2190.1	NUM
ejpam-86	491	30	subsets	subset	NOUN
ejpam-86	491	31	selected	select	VERB
ejpam-86	491	32	by	by	ADP
ejpam-86	491	33	aic	aic	PROPN
ejpam-86	491	34	subset	subset	VERB
ejpam-86	491	35	11110000000	11110000000	NUM
ejpam-86	491	36	11110000100	11110000100	NUM
ejpam-86	491	37	11110000010	11110000010	NUM
ejpam-86	491	38	11110000110	11110000110	NUM
ejpam-86	491	39	11111000000	11111000000	NUM
ejpam-86	491	40	β̂mle	β̂mle	PUNCT
ejpam-86	491	41	1.9263	1.9263	NUM
ejpam-86	491	42	1.8944	1.8944	NUM
ejpam-86	491	43	1.9408	1.9408	NUM
ejpam-86	491	44	1.9055	1.9055	NUM
ejpam-86	491	45	1.9082	1.9082	NUM
ejpam-86	491	46	aic	aic	PROPN
ejpam-86	491	47	2118.5	2118.5	NUM
ejpam-86	491	48	2119.2	2119.2	NUM
ejpam-86	491	49	2120.0	2120.0	NUM
ejpam-86	491	50	2120.8	2120.8	NUM
ejpam-86	491	51	2121.1	2121.1	NUM
ejpam-86	491	52	as	as	ADP
ejpam-86	491	53	the	the	DET
ejpam-86	491	54	starting	starting	NOUN
ejpam-86	491	55	values	value	NOUN
ejpam-86	491	56	for	for	ADP
ejpam-86	491	57	the	the	DET
ejpam-86	491	58	ga	ga	PROPN
ejpam-86	491	59	process	process	NOUN
ejpam-86	491	60	.	.	PUNCT
ejpam-86	492	1	aic	aic	PROPN
ejpam-86	492	2	and	and	CCONJ
ejpam-86	492	3	icomp(ifim	icomp(ifim	PROPN
ejpam-86	492	4	)	)	PUNCT
ejpam-86	492	5	criteria	criterion	NOUN
ejpam-86	492	6	developed	develop	VERB
ejpam-86	492	7	in	in	ADP
ejpam-86	492	8	section	section	NOUN
ejpam-86	492	9	5	5	NUM
ejpam-86	492	10	are	be	AUX
ejpam-86	492	11	used	use	VERB
ejpam-86	492	12	for	for	ADP
ejpam-86	492	13	subset	subset	ADJ
ejpam-86	492	14	selection	selection	NOUN
ejpam-86	492	15	of	of	ADP
ejpam-86	492	16	the	the	DET
ejpam-86	492	17	best	good	ADJ
ejpam-86	492	18	predictors	predictor	NOUN
ejpam-86	492	19	.	.	PUNCT
ejpam-86	493	1	the	the	DET
ejpam-86	493	2	top	top	ADJ
ejpam-86	493	3	5	5	NUM
ejpam-86	493	4	best	good	ADJ
ejpam-86	493	5	subsets	subset	NOUN
ejpam-86	493	6	selected	select	VERB
ejpam-86	493	7	according	accord	VERB
ejpam-86	493	8	to	to	ADP
ejpam-86	493	9	aic	aic	PROPN
ejpam-86	493	10	and	and	CCONJ
ejpam-86	493	11	icomp(ifim	icomp(ifim	PROPN
ejpam-86	493	12	)	)	PUNCT
ejpam-86	493	13	scores	score	NOUN
ejpam-86	493	14	by	by	ADP
ejpam-86	493	15	fitting	fitting	ADJ
ejpam-86	493	16	type	type	NOUN
ejpam-86	493	17	i	i	PRON
ejpam-86	493	18	and	and	CCONJ
ejpam-86	493	19	type	type	PROPN
ejpam-86	493	20	ii	ii	PROPN
ejpam-86	493	21	mvper	mvper	NOUN
ejpam-86	493	22	models	model	NOUN
ejpam-86	493	23	on	on	ADP
ejpam-86	493	24	the	the	DET
ejpam-86	493	25	klein	klein	PROPN
ejpam-86	493	26	data	data	PROPN
ejpam-86	493	27	set	set	PROPN
ejpam-86	493	28	.	.	PUNCT
ejpam-86	494	1	the	the	DET
ejpam-86	494	2	results	result	NOUN
ejpam-86	494	3	are	be	AUX
ejpam-86	494	4	summarized	summarize	VERB
ejpam-86	494	5	in	in	ADP
ejpam-86	494	6	tables	table	NOUN
ejpam-86	494	7	6	6	NUM
ejpam-86	494	8	and	and	CCONJ
ejpam-86	494	9	7	7	NUM
ejpam-86	494	10	,	,	PUNCT
ejpam-86	494	11	respectively	respectively	ADV
ejpam-86	494	12	.	.	PUNCT
ejpam-86	495	1	we	we	PRON
ejpam-86	495	2	can	can	AUX
ejpam-86	495	3	see	see	VERB
ejpam-86	495	4	that	that	SCONJ
ejpam-86	495	5	the	the	DET
ejpam-86	495	6	results	result	NOUN
ejpam-86	495	7	of	of	ADP
ejpam-86	495	8	aic	aic	PROPN
ejpam-86	495	9	and	and	CCONJ
ejpam-86	495	10	icomp(ifim	icomp(ifim	PROPN
ejpam-86	495	11	)	)	PUNCT
ejpam-86	495	12	are	be	AUX
ejpam-86	495	13	slightly	slightly	ADV
ejpam-86	495	14	different	different	ADJ
ejpam-86	495	15	for	for	ADP
ejpam-86	495	16	each	each	DET
ejpam-86	495	17	type	type	NOUN
ejpam-86	495	18	of	of	ADP
ejpam-86	495	19	mvper	mvper	NOUN
ejpam-86	495	20	models	model	NOUN
ejpam-86	495	21	for	for	ADP
ejpam-86	495	22	this	this	DET
ejpam-86	495	23	data	datum	NOUN
ejpam-86	495	24	set	set	VERB
ejpam-86	495	25	.	.	PUNCT
ejpam-86	496	1	for	for	ADP
ejpam-86	496	2	type	type	NOUN
ejpam-86	496	3	i	i	PROPN
ejpam-86	496	4	mvper	mvper	PROPN
ejpam-86	496	5	model	model	NOUN
ejpam-86	496	6	,	,	PUNCT
ejpam-86	496	7	both	both	CCONJ
ejpam-86	496	8	aic	aic	PROPN
ejpam-86	496	9	and	and	CCONJ
ejpam-86	496	10	icomp(ifim	icomp(ifim	PROPN
ejpam-86	496	11	)	)	PUNCT
ejpam-86	496	12	criteria	criterion	NOUN
ejpam-86	496	13	select	select	VERB
ejpam-86	496	14	the	the	DET
ejpam-86	496	15	binary	binary	PROPN
ejpam-86	496	16	string	string	PROPN
ejpam-86	496	17	000001	000001	NUM
ejpam-86	496	18	as	as	ADP
ejpam-86	496	19	the	the	DET
ejpam-86	496	20	optimal	optimal	ADJ
ejpam-86	496	21	model	model	NOUN
ejpam-86	496	22	.	.	PUNCT
ejpam-86	497	1	in	in	ADP
ejpam-86	497	2	other	other	ADJ
ejpam-86	497	3	words	word	NOUN
ejpam-86	497	4	,	,	PUNCT
ejpam-86	497	5	the	the	DET
ejpam-86	497	6	predictor	predictor	NOUN
ejpam-86	497	7	variable	variable	PROPN
ejpam-86	497	8	x5	x5	PROPN
ejpam-86	497	9	=	=	SYM
ejpam-86	497	10	price	price	NOUN
ejpam-86	497	11	index	index	NOUN
ejpam-86	497	12	of	of	ADP
ejpam-86	497	13	consumption	consumption	NOUN
ejpam-86	497	14	is	be	AUX
ejpam-86	497	15	the	the	DET
ejpam-86	497	16	best	good	ADJ
ejpam-86	497	17	predictor	predictor	NOUN
ejpam-86	497	18	to	to	PART
ejpam-86	497	19	predict	predict	VERB
ejpam-86	497	20	all	all	DET
ejpam-86	497	21	the	the	DET
ejpam-86	497	22	response	response	NOUN
ejpam-86	497	23	variable	variable	ADJ
ejpam-86	497	24	y	y	PROPN
ejpam-86	497	25	with	with	ADP
ejpam-86	497	26	β̂mle	β̂mle	PUNCT
ejpam-86	497	27	=	=	PUNCT
ejpam-86	497	28	1.3099	1.3099	NUM
ejpam-86	497	29	which	which	PRON
ejpam-86	497	30	indicates	indicate	VERB
ejpam-86	497	31	that	that	SCONJ
ejpam-86	497	32	this	this	DET
ejpam-86	497	33	data	datum	NOUN
ejpam-86	497	34	set	set	VERB
ejpam-86	497	35	is	be	AUX
ejpam-86	497	36	not	not	PART
ejpam-86	497	37	normal	normal	ADJ
ejpam-86	497	38	.	.	PUNCT
ejpam-86	498	1	the	the	DET
ejpam-86	498	2	mles	mle	NOUN
ejpam-86	498	3	of	of	ADP
ejpam-86	498	4	b	b	PROPN
ejpam-86	498	5	and	and	CCONJ
ejpam-86	498	6	σ	σ	NOUN
ejpam-86	498	7	for	for	ADP
ejpam-86	498	8	the	the	DET
ejpam-86	498	9	optimal	optimal	ADJ
ejpam-86	498	10	model	model	NOUN
ejpam-86	498	11	are	be	AUX
ejpam-86	498	12	:	:	PUNCT
ejpam-86	498	13	b̂mle	b̂mle	NOUN
ejpam-86	498	14	,	,	PUNCT
ejpam-86	498	15	model	model	NOUN
ejpam-86	498	16	i	i	NOUN
ejpam-86	498	17	=	=	PUNCT
ejpam-86	498	18	(	(	PUNCT
ejpam-86	498	19	1.0084	1.0084	NUM
ejpam-86	498	20	0.9151	0.9151	NUM
ejpam-86	498	21	0.8590	0.8590	NUM
ejpam-86	498	22	0.9793	0.9793	NUM
ejpam-86	498	23	1.0597	1.0597	NUM
ejpam-86	498	24	)	)	PUNCT
ejpam-86	498	25	and	and	CCONJ
ejpam-86	498	26	σ̂mle	σ̂mle	NOUN
ejpam-86	498	27	,	,	PUNCT
ejpam-86	498	28	model	model	NOUN
ejpam-86	499	1	i	i	NOUN
ejpam-86	499	2	=	=	PUNCT
ejpam-86	499	3			X
ejpam-86	499	4			PROPN
ejpam-86	499	5	84.9514	84.9514	NUM
ejpam-86	499	6	9.1643	9.1643	NUM
ejpam-86	499	7	6.1158	6.1158	NUM
ejpam-86	499	8	9.5936	9.5936	NUM
ejpam-86	499	9	15.5155	15.5155	NUM
ejpam-86	499	10	9.1643	9.1643	NUM
ejpam-86	499	11	91.5875	91.5875	NUM
ejpam-86	499	12	138.4841	138.4841	NUM
ejpam-86	499	13	3.8669	3.8669	NUM
ejpam-86	499	14	−18.4478	−18.4478	PROPN
ejpam-86	499	15	6.1158	6.1158	NUM
ejpam-86	499	16	138.4841	138.4841	NUM
ejpam-86	499	17	576.3684	576.3684	NUM
ejpam-86	499	18	−55.7333	−55.7333	PROPN
ejpam-86	499	19	−54.1991	−54.1991	PROPN
ejpam-86	499	20	9.5936	9.5936	NUM
ejpam-86	499	21	3.8669	3.8669	NUM
ejpam-86	499	22	−55.7333	−55.7333	NOUN
ejpam-86	499	23	34.2577	34.2577	NUM
ejpam-86	499	24	−4.7476	−4.7476	X
ejpam-86	499	25	15.5155	15.5155	NUM
ejpam-86	499	26	−18.4478	−18.4478	PROPN
ejpam-86	499	27	−54.1991	−54.1991	PROPN
ejpam-86	499	28	−4.7476	−4.7476	X
ejpam-86	499	29	81.4006	81.4006	NUM
ejpam-86	499	30			NOUN
ejpam-86	500	1			NOUN
ejpam-86	500	2	.	.	PUNCT
ejpam-86	501	1	for	for	ADP
ejpam-86	501	2	type	type	NOUN
ejpam-86	501	3	ii	ii	PROPN
ejpam-86	501	4	mvper	mvper	PROPN
ejpam-86	501	5	model	model	NOUN
ejpam-86	501	6	,	,	PUNCT
ejpam-86	501	7	both	both	CCONJ
ejpam-86	501	8	aic	aic	PROPN
ejpam-86	501	9	and	and	CCONJ
ejpam-86	501	10	icomp(ifim	icomp(ifim	PROPN
ejpam-86	501	11	)	)	PUNCT
ejpam-86	501	12	select	select	VERB
ejpam-86	501	13	110000	110000	NUM
ejpam-86	501	14	as	as	ADP
ejpam-86	501	15	the	the	DET
ejpam-86	501	16	optimal	optimal	ADJ
ejpam-86	501	17	model	model	NOUN
ejpam-86	501	18	.	.	PUNCT
ejpam-86	502	1	that	that	PRON
ejpam-86	502	2	is	is	ADV
ejpam-86	502	3	,	,	PUNCT
ejpam-86	502	4	the	the	DET
ejpam-86	502	5	constant	constant	ADJ
ejpam-86	502	6	term	term	NOUN
ejpam-86	502	7	x0	x0	PROPN
ejpam-86	502	8	and	and	CCONJ
ejpam-86	502	9	the	the	DET
ejpam-86	502	10	predictor	predictor	NOUN
ejpam-86	502	11	variable	variable	NOUN
ejpam-86	502	12	x1	x1	PROPN
ejpam-86	503	1	=	=	PUNCT
ejpam-86	503	2	total	total	ADJ
ejpam-86	503	3	labor	labor	NOUN
ejpam-86	503	4	force	force	NOUN
ejpam-86	503	5	are	be	AUX
ejpam-86	503	6	chosen	choose	VERB
ejpam-86	503	7	as	as	ADP
ejpam-86	503	8	the	the	DET
ejpam-86	503	9	best	good	ADJ
ejpam-86	503	10	subset	subset	NOUN
ejpam-86	503	11	of	of	ADP
ejpam-86	503	12	predictors	predictor	NOUN
ejpam-86	503	13	.	.	PUNCT
ejpam-86	504	1	the	the	DET
ejpam-86	504	2	mles	mle	NOUN
ejpam-86	504	3	of	of	ADP
ejpam-86	504	4	b	b	PROPN
ejpam-86	504	5	and	and	CCONJ
ejpam-86	504	6	σ	σ	NOUN
ejpam-86	504	7	for	for	ADP
ejpam-86	504	8	the	the	DET
ejpam-86	504	9	optimal	optimal	ADJ
ejpam-86	504	10	model	model	NOUN
ejpam-86	504	11	are	be	AUX
ejpam-86	504	12	b̂mle	b̂mle	NOUN
ejpam-86	504	13	,	,	PUNCT
ejpam-86	504	14	model	model	NOUN
ejpam-86	504	15	ii	ii	PROPN
ejpam-86	505	1	=	=	PRON
ejpam-86	505	2	(	(	PUNCT
ejpam-86	505	3	−462.5338	−462.5338	NOUN
ejpam-86	505	4	−115.9112	−115.9112	NOUN
ejpam-86	505	5	473.4929	473.4929	NUM
ejpam-86	505	6	−417.9496	−417.9496	PROPN
ejpam-86	505	7	−505.6549	−505.6549	NOUN
ejpam-86	505	8	5.6339	5.6339	NUM
ejpam-86	505	9	2.1539	2.1539	NUM
ejpam-86	505	10	−3.6246	−3.6246	NUM
ejpam-86	505	11	5.1773	5.1773	NUM
ejpam-86	505	12	6.0989	6.0989	NUM
ejpam-86	505	13	)	)	PUNCT
ejpam-86	505	14	and	and	CCONJ
ejpam-86	505	15	σ̂mle	σ̂mle	NOUN
ejpam-86	505	16	,	,	PUNCT
ejpam-86	505	17	model	model	NOUN
ejpam-86	505	18	ii	ii	PROPN
ejpam-86	505	19	=	=	SYM
ejpam-86	505	20			PUNCT
ejpam-86	505	21			PROPN
ejpam-86	505	22	496.1	496.1	NUM
ejpam-86	505	23	171.4	171.4	NUM
ejpam-86	505	24	−729.1	−729.1	PROPN
ejpam-86	505	25	263.0	263.0	NUM
ejpam-86	505	26	256.7	256.7	NUM
ejpam-86	505	27	171.4	171.4	NUM
ejpam-86	505	28	176.3	176.3	NUM
ejpam-86	505	29	−188.7	−188.7	NOUN
ejpam-86	505	30	25.6	25.6	NUM
ejpam-86	505	31	26.7	26.7	NUM
ejpam-86	505	32	−729.1	−729.1	PROPN
ejpam-86	505	33	−188.7	−188.7	PROPN
ejpam-86	505	34	9812.5	9812.5	NUM
ejpam-86	505	35	−1485.6	−1485.6	VERB
ejpam-86	505	36	−765.4	−765.4	PROPN
ejpam-86	505	37	263.0	263.0	NUM
ejpam-86	505	38	25.6	25.6	NUM
ejpam-86	505	39	−1485.6	−1485.6	NUM
ejpam-86	505	40	973.4	973.4	NUM
ejpam-86	505	41	275.0	275.0	NUM
ejpam-86	505	42	256.7	256.7	NUM
ejpam-86	505	43	26.7	26.7	NUM
ejpam-86	505	44	−765.4	−765.4	NOUN
ejpam-86	505	45	275.0	275.0	NUM
ejpam-86	505	46	1479.4	1479.4	NUM
ejpam-86	505	47			PUNCT
ejpam-86	506	1			NOUN
ejpam-86	506	2	.	.	PUNCT
ejpam-86	507	1	m.	m.	PROPN
ejpam-86	507	2	liu	liu	PROPN
ejpam-86	507	3	and	and	CCONJ
ejpam-86	507	4	h.	h.	PROPN
ejpam-86	507	5	bozdogan	bozdogan	PROPN
ejpam-86	507	6	/	/	SYM
ejpam-86	507	7	eur	eur	PROPN
ejpam-86	507	8	.	.	PUNCT
ejpam-86	508	1	j.	j.	PROPN
ejpam-86	508	2	pure	pure	PROPN
ejpam-86	508	3	appl	appl	PROPN
ejpam-86	508	4	.	.	PROPN
ejpam-86	508	5	math	math	PROPN
ejpam-86	508	6	,	,	PUNCT
ejpam-86	508	7	1	1	NUM
ejpam-86	508	8	(	(	PUNCT
ejpam-86	508	9	2008	2008	NUM
ejpam-86	508	10	)	)	PUNCT
ejpam-86	508	11	,	,	PUNCT
ejpam-86	508	12	(	(	PUNCT
ejpam-86	508	13	4	4	NUM
ejpam-86	508	14	-	-	SYM
ejpam-86	508	15	37	37	NUM
ejpam-86	508	16	)	)	PUNCT
ejpam-86	508	17	26	26	NUM
ejpam-86	508	18	table	table	NOUN
ejpam-86	508	19	5	5	NUM
ejpam-86	508	20	:	:	PUNCT
ejpam-86	508	21	variables	variable	NOUN
ejpam-86	508	22	of	of	ADP
ejpam-86	508	23	the	the	DET
ejpam-86	508	24	klein	klein	PROPN
ejpam-86	508	25	data	data	PROPN
ejpam-86	508	26	set	set	PROPN
ejpam-86	508	27	.	.	PUNCT
ejpam-86	509	1	response	response	NOUN
ejpam-86	509	2	variables	variable	VERB
ejpam-86	509	3	independent	independent	ADJ
ejpam-86	509	4	variables	variable	NOUN
ejpam-86	509	5	y1	y1	NOUN
ejpam-86	509	6	=	=	PUNCT
ejpam-86	509	7	industrial	industrial	ADJ
ejpam-86	509	8	production	production	NOUN
ejpam-86	510	1	x1	x1	NOUN
ejpam-86	510	2	=	=	PUNCT
ejpam-86	510	3	total	total	ADJ
ejpam-86	510	4	labor	labor	NOUN
ejpam-86	510	5	force	force	NOUN
ejpam-86	510	6	y2	y2	NOUN
ejpam-86	510	7	=	=	NOUN
ejpam-86	510	8	consumption	consumption	NOUN
ejpam-86	511	1	x2	x2	NOUN
ejpam-86	511	2	=	=	PUNCT
ejpam-86	511	3	weekly	weekly	ADJ
ejpam-86	511	4	wage	wage	NOUN
ejpam-86	511	5	rates	rate	NOUN
ejpam-86	511	6	y3	y3	NOUN
ejpam-86	511	7	=	=	PUNCT
ejpam-86	511	8	unemployment	unemployment	NOUN
ejpam-86	511	9	x3	x3	NOUN
ejpam-86	511	10	=	=	PUNCT
ejpam-86	511	11	price	price	NOUN
ejpam-86	511	12	index	index	NOUN
ejpam-86	511	13	of	of	ADP
ejpam-86	511	14	imports	import	NOUN
ejpam-86	511	15	y4	y4	NOUN
ejpam-86	511	16	=	=	PUNCT
ejpam-86	511	17	total	total	ADJ
ejpam-86	511	18	imports	import	NOUN
ejpam-86	511	19	x4	x4	NOUN
ejpam-86	511	20	=	=	NOUN
ejpam-86	511	21	price	price	NOUN
ejpam-86	511	22	index	index	NOUN
ejpam-86	511	23	of	of	ADP
ejpam-86	511	24	exports	export	NOUN
ejpam-86	511	25	y5	y5	NOUN
ejpam-86	511	26	=	=	SYM
ejpam-86	511	27	total	total	ADJ
ejpam-86	511	28	exports	export	NOUN
ejpam-86	511	29	x5	x5	NOUN
ejpam-86	511	30	=	=	SYM
ejpam-86	511	31	price	price	NOUN
ejpam-86	511	32	index	index	NOUN
ejpam-86	511	33	of	of	ADP
ejpam-86	511	34	consumption	consumption	NOUN
ejpam-86	511	35	table	table	NOUN
ejpam-86	511	36	6	6	NUM
ejpam-86	511	37	:	:	PUNCT
ejpam-86	511	38	top	top	NOUN
ejpam-86	511	39	5	5	NUM
ejpam-86	511	40	ranking	ranking	ADJ
ejpam-86	511	41	subsets	subset	NOUN
ejpam-86	511	42	selected	select	VERB
ejpam-86	511	43	under	under	ADP
ejpam-86	511	44	type	type	NOUN
ejpam-86	511	45	i	i	PROPN
ejpam-86	511	46	mvper	mvper	PROPN
ejpam-86	511	47	model	model	NOUN
ejpam-86	511	48	.	.	PUNCT
ejpam-86	512	1	ranking	rank	VERB
ejpam-86	512	2	1	1	NUM
ejpam-86	512	3	2	2	NUM
ejpam-86	512	4	3	3	NUM
ejpam-86	512	5	4	4	NUM
ejpam-86	512	6	5	5	NUM
ejpam-86	512	7	subsets	subset	NOUN
ejpam-86	512	8	selected	select	VERB
ejpam-86	512	9	by	by	ADP
ejpam-86	512	10	aic	aic	PROPN
ejpam-86	512	11	subset	subset	AUX
ejpam-86	512	12	000001	000001	NUM
ejpam-86	512	13	101000	101000	NUM
ejpam-86	512	14	010000	010000	NUM
ejpam-86	512	15	000100	000100	NUM
ejpam-86	512	16	000010	000010	NUM
ejpam-86	512	17	β̂mom	β̂mom	NUM
ejpam-86	513	1	1.0641	1.0641	NUM
ejpam-86	513	2	0.7963	0.7963	NUM
ejpam-86	513	3	0.7980	0.7980	NUM
ejpam-86	513	4	1.0177	1.0177	NUM
ejpam-86	513	5	1.0925	1.0925	NUM
ejpam-86	513	6	β̂mle	β̂mle	NUM
ejpam-86	513	7	1.3099	1.3099	NUM
ejpam-86	513	8	0.5446	0.5446	NUM
ejpam-86	513	9	0.5574	0.5574	NUM
ejpam-86	513	10	1.0446	1.0446	NUM
ejpam-86	513	11	0.7016	0.7016	NUM
ejpam-86	513	12	aic	aic	PROPN
ejpam-86	513	13	1379.2	1379.2	NUM
ejpam-86	513	14	1483.1	1483.1	NUM
ejpam-86	513	15	1498.3	1498.3	NUM
ejpam-86	513	16	1521.5	1521.5	NUM
ejpam-86	513	17	1527.6	1527.6	NUM
ejpam-86	513	18	subsets	subset	NOUN
ejpam-86	513	19	selected	select	VERB
ejpam-86	513	20	by	by	ADP
ejpam-86	513	21	icomp(ifim	icomp(ifim	NOUN
ejpam-86	513	22	)	)	PUNCT
ejpam-86	513	23	subset	subset	VERB
ejpam-86	513	24	000001	000001	NUM
ejpam-86	513	25	101000	101000	NUM
ejpam-86	513	26	000100	000100	NUM
ejpam-86	513	27	000010	000010	NUM
ejpam-86	513	28	010000	010000	NUM
ejpam-86	513	29	β̂mom	β̂mom	NUM
ejpam-86	514	1	1.0641	1.0641	NUM
ejpam-86	514	2	0.7963	0.7963	NUM
ejpam-86	514	3	1.0177	1.0177	NUM
ejpam-86	514	4	1.0925	1.0925	NUM
ejpam-86	514	5	0.7980	0.7980	NUM
ejpam-86	514	6	β̂mle	β̂mle	NUM
ejpam-86	514	7	1.3099	1.3099	NUM
ejpam-86	514	8	0.5446	0.5446	NUM
ejpam-86	514	9	1.0446	1.0446	NUM
ejpam-86	514	10	0.7016	0.7016	NUM
ejpam-86	514	11	0.5574	0.5574	NUM
ejpam-86	514	12	icomp	icomp	NOUN
ejpam-86	514	13	1495.0	1495.0	NUM
ejpam-86	514	14	1540.6	1540.6	NUM
ejpam-86	514	15	1616.0	1616.0	NUM
ejpam-86	514	16	1637.0	1637.0	NUM
ejpam-86	514	17	1655.5	1655.5	NUM
ejpam-86	514	18	from	from	ADP
ejpam-86	514	19	the	the	DET
ejpam-86	514	20	results	result	NOUN
ejpam-86	514	21	in	in	ADP
ejpam-86	514	22	tables	table	NOUN
ejpam-86	514	23	6	6	NUM
ejpam-86	514	24	and	and	CCONJ
ejpam-86	514	25	7	7	NUM
ejpam-86	514	26	,	,	PUNCT
ejpam-86	514	27	we	we	PRON
ejpam-86	514	28	see	see	VERB
ejpam-86	514	29	that	that	SCONJ
ejpam-86	514	30	the	the	DET
ejpam-86	514	31	aic	aic	PROPN
ejpam-86	514	32	and	and	CCONJ
ejpam-86	514	33	icomp(ifim	icomp(ifim	PROPN
ejpam-86	514	34	)	)	PUNCT
ejpam-86	514	35	values	value	NOUN
ejpam-86	514	36	for	for	ADP
ejpam-86	514	37	the	the	DET
ejpam-86	514	38	best	good	ADJ
ejpam-86	514	39	type	type	NOUN
ejpam-86	514	40	ii	ii	NOUN
ejpam-86	514	41	model	model	NOUN
ejpam-86	514	42	are	be	AUX
ejpam-86	514	43	much	much	ADV
ejpam-86	514	44	smaller	small	ADJ
ejpam-86	514	45	than	than	ADP
ejpam-86	514	46	those	those	PRON
ejpam-86	514	47	for	for	ADP
ejpam-86	514	48	the	the	DET
ejpam-86	514	49	best	good	ADJ
ejpam-86	514	50	type	type	NOUN
ejpam-86	514	51	i	i	PRON
ejpam-86	514	52	model	model	NOUN
ejpam-86	514	53	.	.	PUNCT
ejpam-86	515	1	so	so	ADV
ejpam-86	515	2	according	accord	VERB
ejpam-86	515	3	to	to	ADP
ejpam-86	515	4	the	the	DET
ejpam-86	515	5	minimum	minimum	ADJ
ejpam-86	515	6	value	value	NOUN
ejpam-86	515	7	of	of	ADP
ejpam-86	515	8	both	both	CCONJ
ejpam-86	515	9	aic	aic	PROPN
ejpam-86	515	10	and	and	CCONJ
ejpam-86	515	11	icomp(ifim	icomp(ifim	PROPN
ejpam-86	515	12	)	)	PUNCT
ejpam-86	515	13	criteria	criterion	NOUN
ejpam-86	515	14	,	,	PUNCT
ejpam-86	515	15	type	type	NOUN
ejpam-86	515	16	ii	ii	PROPN
ejpam-86	515	17	mvper	mvper	NOUN
ejpam-86	515	18	model	model	NOUN
ejpam-86	515	19	will	will	AUX
ejpam-86	515	20	be	be	AUX
ejpam-86	515	21	selected	select	VERB
ejpam-86	515	22	as	as	ADP
ejpam-86	515	23	the	the	DET
ejpam-86	515	24	best	good	ADJ
ejpam-86	515	25	fitting	fitting	ADJ
ejpam-86	515	26	model	model	NOUN
ejpam-86	515	27	for	for	ADP
ejpam-86	515	28	the	the	DET
ejpam-86	515	29	klein	klein	PROPN
ejpam-86	515	30	data	data	PROPN
ejpam-86	515	31	set	set	PROPN
ejpam-86	515	32	.	.	PUNCT
ejpam-86	516	1	we	we	PRON
ejpam-86	516	2	note	note	VERB
ejpam-86	516	3	that	that	SCONJ
ejpam-86	516	4	such	such	DET
ejpam-86	516	5	a	a	DET
ejpam-86	516	6	choice	choice	NOUN
ejpam-86	516	7	is	be	AUX
ejpam-86	516	8	in	in	ADP
ejpam-86	516	9	agreement	agreement	NOUN
ejpam-86	516	10	with	with	ADP
ejpam-86	516	11	our	our	PRON
ejpam-86	516	12	prior	prior	ADJ
ejpam-86	516	13	knowledge	knowledge	NOUN
ejpam-86	516	14	about	about	ADP
ejpam-86	516	15	this	this	DET
ejpam-86	516	16	data	datum	NOUN
ejpam-86	516	17	set	set	VERB
ejpam-86	516	18	in	in	ADP
ejpam-86	516	19	that	that	SCONJ
ejpam-86	516	20	the	the	DET
ejpam-86	516	21	observations	observation	NOUN
ejpam-86	516	22	actually	actually	ADV
ejpam-86	516	23	are	be	AUX
ejpam-86	516	24	not	not	PART
ejpam-86	516	25	independent	independent	ADJ
ejpam-86	516	26	.	.	PUNCT
ejpam-86	517	1	they	they	PRON
ejpam-86	517	2	are	be	AUX
ejpam-86	517	3	dependent	dependent	ADJ
ejpam-86	517	4	since	since	SCONJ
ejpam-86	517	5	this	this	DET
ejpam-86	517	6	data	datum	NOUN
ejpam-86	517	7	set	set	VERB
ejpam-86	517	8	is	be	AUX
ejpam-86	517	9	time	time	NOUN
ejpam-86	517	10	dependent	dependent	ADJ
ejpam-86	517	11	.	.	PUNCT
ejpam-86	518	1	we	we	PRON
ejpam-86	518	2	can	can	AUX
ejpam-86	518	3	also	also	ADV
ejpam-86	518	4	see	see	VERB
ejpam-86	518	5	that	that	SCONJ
ejpam-86	518	6	for	for	ADP
ejpam-86	518	7	both	both	DET
ejpam-86	518	8	type	type	NOUN
ejpam-86	518	9	i	i	PRON
ejpam-86	518	10	and	and	CCONJ
ejpam-86	518	11	type	type	NOUN
ejpam-86	518	12	ii	ii	PROPN
ejpam-86	518	13	mvper	mvper	NOUN
ejpam-86	518	14	models	model	NOUN
ejpam-86	518	15	,	,	PUNCT
ejpam-86	518	16	the	the	DET
ejpam-86	518	17	mles	mle	NOUN
ejpam-86	518	18	of	of	ADP
ejpam-86	518	19	β	β	X
ejpam-86	518	20	are	be	AUX
ejpam-86	518	21	different	different	ADJ
ejpam-86	518	22	from	from	ADP
ejpam-86	518	23	one	one	NUM
ejpam-86	518	24	another	another	DET
ejpam-86	518	25	,	,	PUNCT
ejpam-86	518	26	which	which	PRON
ejpam-86	518	27	means	mean	VERB
ejpam-86	518	28	that	that	SCONJ
ejpam-86	518	29	the	the	DET
ejpam-86	518	30	residuals	residual	NOUN
ejpam-86	518	31	are	be	AUX
ejpam-86	518	32	non	non	ADJ
ejpam-86	518	33	-	-	ADJ
ejpam-86	518	34	normal	normal	ADJ
ejpam-86	518	35	for	for	ADP
ejpam-86	518	36	the	the	DET
ejpam-86	518	37	klein	klein	PROPN
ejpam-86	518	38	data	data	PROPN
ejpam-86	518	39	set	set	PROPN
ejpam-86	518	40	.	.	PUNCT
ejpam-86	519	1	here	here	ADV
ejpam-86	519	2	we	we	PRON
ejpam-86	519	3	use	use	VERB
ejpam-86	519	4	q	q	ADJ
ejpam-86	519	5	-	-	PUNCT
ejpam-86	519	6	q	q	NOUN
ejpam-86	519	7	plots	plot	NOUN
ejpam-86	519	8	to	to	PART
ejpam-86	519	9	test	test	VERB
ejpam-86	519	10	the	the	DET
ejpam-86	519	11	normality	normality	NOUN
ejpam-86	519	12	of	of	ADP
ejpam-86	519	13	the	the	DET
ejpam-86	519	14	residuals	residual	NOUN
ejpam-86	519	15	.	.	PUNCT
ejpam-86	520	1	for	for	ADP
ejpam-86	520	2	the	the	DET
ejpam-86	520	3	type	type	NOUN
ejpam-86	520	4	i	i	PRON
ejpam-86	520	5	model	model	NOUN
ejpam-86	520	6	,	,	PUNCT
ejpam-86	520	7	if	if	SCONJ
ejpam-86	520	8	the	the	DET
ejpam-86	520	9	residuals	residual	NOUN
ejpam-86	520	10	are	be	AUX
ejpam-86	520	11	i.i.d	i.i.d	ADJ
ejpam-86	520	12	.	.	PUNCT
ejpam-86	520	13	multivariate	multivariate	NOUN
ejpam-86	520	14	normally	normally	ADV
ejpam-86	520	15	distributed	distribute	VERB
ejpam-86	520	16	,	,	PUNCT
ejpam-86	520	17	the	the	DET
ejpam-86	520	18	squared	square	VERB
ejpam-86	520	19	mahalanobis	mahalanobis	ADJ
ejpam-86	520	20	distance	distance	NOUN
ejpam-86	520	21	of	of	ADP
ejpam-86	520	22	each	each	DET
ejpam-86	520	23	residual	residual	ADJ
ejpam-86	520	24	vector	vector	NOUN
ejpam-86	520	25	,	,	PUNCT
ejpam-86	520	26	di	di	X
ejpam-86	520	27	=	=	PUNCT
ejpam-86	520	28	ε′−1	ε′−1	X
ejpam-86	520	29	(	(	PUNCT
ejpam-86	520	30	i	i	NOUN
ejpam-86	520	31	)	)	PUNCT
ejpam-86	520	32	s	s	PART
ejpam-86	520	33	−1ε(i	−1ε(i	PROPN
ejpam-86	520	34	)	)	PUNCT
ejpam-86	520	35	,	,	PUNCT
ejpam-86	520	36	will	will	AUX
ejpam-86	520	37	be	be	AUX
ejpam-86	520	38	distributed	distribute	VERB
ejpam-86	520	39	approximately	approximately	ADV
ejpam-86	520	40	as	as	SCONJ
ejpam-86	520	41	χ2	χ2	NOUN
ejpam-86	520	42	with	with	ADP
ejpam-86	520	43	p	p	NOUN
ejpam-86	520	44	degrees	degree	NOUN
ejpam-86	520	45	of	of	ADP
ejpam-86	520	46	freedom	freedom	NOUN
ejpam-86	520	47	,	,	PUNCT
ejpam-86	520	48	where	where	SCONJ
ejpam-86	520	49	p	p	NOUN
ejpam-86	520	50	=	=	SYM
ejpam-86	520	51	5	5	NUM
ejpam-86	520	52	is	be	AUX
ejpam-86	520	53	the	the	DET
ejpam-86	520	54	number	number	NOUN
ejpam-86	520	55	of	of	ADP
ejpam-86	520	56	dependent	dependent	ADJ
ejpam-86	520	57	variables	variable	NOUN
ejpam-86	520	58	in	in	ADP
ejpam-86	520	59	the	the	DET
ejpam-86	520	60	model	model	NOUN
ejpam-86	520	61	.	.	PUNCT
ejpam-86	521	1	so	so	ADV
ejpam-86	521	2	the	the	DET
ejpam-86	521	3	q	q	ADJ
ejpam-86	521	4	-	-	PUNCT
ejpam-86	521	5	q	q	NOUN
ejpam-86	521	6	plot	plot	NOUN
ejpam-86	521	7	for	for	ADP
ejpam-86	521	8	type	type	NOUN
ejpam-86	521	9	i	i	PRON
ejpam-86	521	10	model	model	NOUN
ejpam-86	521	11	is	be	AUX
ejpam-86	521	12	the	the	DET
ejpam-86	521	13	ordered	order	VERB
ejpam-86	521	14	distance	distance	NOUN
ejpam-86	521	15	values	value	NOUN
ejpam-86	521	16	di	di	VERB
ejpam-86	521	17	against	against	ADP
ejpam-86	521	18	the	the	DET
ejpam-86	521	19	corresponding	corresponding	ADJ
ejpam-86	521	20	theoretical	theoretical	ADJ
ejpam-86	521	21	quantiles	quantile	NOUN
ejpam-86	521	22	of	of	ADP
ejpam-86	521	23	χ2(5	χ2(5	NOUN
ejpam-86	521	24	)	)	PUNCT
ejpam-86	521	25	distribution	distribution	NOUN
ejpam-86	521	26	.	.	PUNCT
ejpam-86	522	1	for	for	ADP
ejpam-86	522	2	the	the	DET
ejpam-86	522	3	type	type	NOUN
ejpam-86	522	4	ii	ii	PROPN
ejpam-86	522	5	model	model	NOUN
ejpam-86	522	6	,	,	PUNCT
ejpam-86	522	7	we	we	PRON
ejpam-86	522	8	first	first	ADV
ejpam-86	522	9	vectorize	vectorize	VERB
ejpam-86	522	10	the	the	DET
ejpam-86	522	11	residuals	residual	NOUN
ejpam-86	522	12	and	and	CCONJ
ejpam-86	522	13	then	then	ADV
ejpam-86	522	14	do	do	VERB
ejpam-86	522	15	the	the	DET
ejpam-86	522	16	plot	plot	NOUN
ejpam-86	522	17	as	as	ADP
ejpam-86	522	18	the	the	DET
ejpam-86	522	19	classic	classic	ADJ
ejpam-86	522	20	normal	normal	ADJ
ejpam-86	522	21	q	q	ADJ
ejpam-86	522	22	-	-	PUNCT
ejpam-86	522	23	q	q	NOUN
ejpam-86	522	24	plot	plot	NOUN
ejpam-86	522	25	.	.	PUNCT
ejpam-86	523	1	the	the	DET
ejpam-86	523	2	q	q	NOUN
ejpam-86	523	3	-	-	PUNCT
ejpam-86	523	4	q	q	NOUN
ejpam-86	523	5	plots	plot	NOUN
ejpam-86	523	6	are	be	AUX
ejpam-86	523	7	shown	show	VERB
ejpam-86	523	8	in	in	ADP
ejpam-86	523	9	figures	figure	NOUN
ejpam-86	523	10	7	7	NUM
ejpam-86	523	11	and	and	CCONJ
ejpam-86	523	12	8	8	NUM
ejpam-86	523	13	,	,	PUNCT
ejpam-86	523	14	respectively	respectively	ADV
ejpam-86	523	15	.	.	PUNCT
ejpam-86	524	1	from	from	ADP
ejpam-86	524	2	the	the	DET
ejpam-86	524	3	graphs	graph	NOUN
ejpam-86	524	4	,	,	PUNCT
ejpam-86	524	5	we	we	PRON
ejpam-86	524	6	see	see	VERB
ejpam-86	524	7	that	that	SCONJ
ejpam-86	524	8	the	the	DET
ejpam-86	524	9	type	type	NOUN
ejpam-86	524	10	i	i	PRON
ejpam-86	524	11	model	model	VERB
ejpam-86	524	12	is	be	AUX
ejpam-86	524	13	closer	close	ADJ
ejpam-86	524	14	to	to	ADP
ejpam-86	524	15	the	the	DET
ejpam-86	524	16	normal	normal	ADJ
ejpam-86	524	17	model	model	NOUN
ejpam-86	524	18	according	accord	VERB
ejpam-86	524	19	to	to	ADP
ejpam-86	524	20	the	the	DET
ejpam-86	524	21	estimated	estimated	ADJ
ejpam-86	524	22	β	β	NOUN
ejpam-86	524	23	values	value	NOUN
ejpam-86	524	24	.	.	PUNCT
ejpam-86	525	1	however	however	ADV
ejpam-86	525	2	,	,	PUNCT
ejpam-86	525	3	we	we	PRON
ejpam-86	525	4	see	see	VERB
ejpam-86	525	5	that	that	SCONJ
ejpam-86	525	6	the	the	DET
ejpam-86	525	7	best	well	ADV
ejpam-86	525	8	fitting	fitting	ADJ
ejpam-86	525	9	type	type	NOUN
ejpam-86	525	10	ii	ii	NOUN
ejpam-86	525	11	model	model	NOUN
ejpam-86	525	12	does	do	AUX
ejpam-86	525	13	not	not	PART
ejpam-86	525	14	follow	follow	VERB
ejpam-86	525	15	the	the	DET
ejpam-86	525	16	normal	normal	ADJ
ejpam-86	525	17	distribution	distribution	NOUN
ejpam-86	525	18	based	base	VERB
ejpam-86	525	19	on	on	ADP
ejpam-86	525	20	the	the	DET
ejpam-86	525	21	estimated	estimated	ADJ
ejpam-86	525	22	β	β	NOUN
ejpam-86	525	23	values	value	NOUN
ejpam-86	525	24	.	.	PUNCT
ejpam-86	526	1	indeed	indeed	ADV
ejpam-86	526	2	,	,	PUNCT
ejpam-86	526	3	this	this	DET
ejpam-86	526	4	data	datum	NOUN
ejpam-86	526	5	set	set	NOUN
ejpam-86	526	6	has	have	VERB
ejpam-86	526	7	serial	serial	ADJ
ejpam-86	526	8	correlations	correlation	NOUN
ejpam-86	526	9	which	which	PRON
ejpam-86	526	10	causes	cause	VERB
ejpam-86	526	11	the	the	DET
ejpam-86	526	12	fact	fact	NOUN
ejpam-86	526	13	that	that	SCONJ
ejpam-86	526	14	the	the	DET
ejpam-86	526	15	probability	probability	NOUN
ejpam-86	526	16	distribution	distribution	NOUN
ejpam-86	526	17	of	of	ADP
ejpam-86	526	18	the	the	DET
ejpam-86	526	19	model	model	NOUN
ejpam-86	526	20	is	be	AUX
ejpam-86	526	21	misspecified	misspecifie	VERB
ejpam-86	526	22	.	.	PUNCT
ejpam-86	527	1	type	type	NOUN
ejpam-86	527	2	ii	ii	PROPN
ejpam-86	527	3	model	model	NOUN
ejpam-86	527	4	captures	capture	VERB
ejpam-86	527	5	such	such	ADJ
ejpam-86	527	6	misspecification	misspecification	NOUN
ejpam-86	527	7	as	as	ADP
ejpam-86	527	8	a	a	DET
ejpam-86	527	9	general	general	ADJ
ejpam-86	527	10	and	and	CCONJ
ejpam-86	527	11	flexible	flexible	ADJ
ejpam-86	527	12	model	model	NOUN
ejpam-86	527	13	which	which	PRON
ejpam-86	527	14	†the	†the	DET
ejpam-86	527	15	ga	ga	PROPN
ejpam-86	527	16	is	be	AUX
ejpam-86	527	17	setup	setup	NOUN
ejpam-86	527	18	to	to	PART
ejpam-86	527	19	search	search	VERB
ejpam-86	527	20	β	β	X
ejpam-86	527	21	in	in	ADP
ejpam-86	527	22	[	[	X
ejpam-86	527	23	0.001	0.001	NUM
ejpam-86	527	24	10	10	NUM
ejpam-86	527	25	]	]	PUNCT
ejpam-86	527	26	.	.	PUNCT
ejpam-86	528	1	if	if	SCONJ
ejpam-86	528	2	the	the	DET
ejpam-86	528	3	mle	mle	NOUN
ejpam-86	528	4	of	of	ADP
ejpam-86	528	5	β	β	PROPN
ejpam-86	528	6	is	be	AUX
ejpam-86	528	7	very	very	ADV
ejpam-86	528	8	near	near	ADV
ejpam-86	528	9	to	to	ADP
ejpam-86	528	10	10	10	NUM
ejpam-86	528	11	,	,	PUNCT
ejpam-86	528	12	we	we	PRON
ejpam-86	528	13	consider	consider	VERB
ejpam-86	528	14	the	the	DET
ejpam-86	528	15	estimate	estimate	NOUN
ejpam-86	528	16	is	be	AUX
ejpam-86	528	17	∞.	∞.	PROPN
ejpam-86	528	18	if	if	SCONJ
ejpam-86	528	19	the	the	DET
ejpam-86	528	20	mom	mom	NOUN
ejpam-86	528	21	of	of	ADP
ejpam-86	528	22	β	β	PROPN
ejpam-86	528	23	is	be	AUX
ejpam-86	528	24	very	very	ADV
ejpam-86	528	25	near	near	ADV
ejpam-86	528	26	to	to	ADP
ejpam-86	528	27	10	10	NUM
ejpam-86	528	28	,	,	PUNCT
ejpam-86	528	29	it	it	PRON
ejpam-86	528	30	means	mean	VERB
ejpam-86	528	31	the	the	DET
ejpam-86	528	32	mom	mom	NOUN
ejpam-86	528	33	of	of	ADP
ejpam-86	528	34	β	β	PROPN
ejpam-86	528	35	does	do	AUX
ejpam-86	528	36	not	not	PART
ejpam-86	528	37	exist	exist	VERB
ejpam-86	528	38	.	.	PUNCT
ejpam-86	529	1	m.	m.	PROPN
ejpam-86	529	2	liu	liu	PROPN
ejpam-86	529	3	and	and	CCONJ
ejpam-86	529	4	h.	h.	PROPN
ejpam-86	529	5	bozdogan	bozdogan	PROPN
ejpam-86	529	6	/	/	SYM
ejpam-86	529	7	eur	eur	PROPN
ejpam-86	529	8	.	.	PUNCT
ejpam-86	530	1	j.	j.	PROPN
ejpam-86	530	2	pure	pure	PROPN
ejpam-86	530	3	appl	appl	PROPN
ejpam-86	530	4	.	.	PROPN
ejpam-86	530	5	math	math	PROPN
ejpam-86	530	6	,	,	PUNCT
ejpam-86	530	7	1	1	NUM
ejpam-86	530	8	(	(	PUNCT
ejpam-86	530	9	2008	2008	NUM
ejpam-86	530	10	)	)	PUNCT
ejpam-86	530	11	,	,	PUNCT
ejpam-86	530	12	(	(	PUNCT
ejpam-86	530	13	4	4	NUM
ejpam-86	530	14	-	-	SYM
ejpam-86	530	15	37	37	NUM
ejpam-86	530	16	)	)	PUNCT
ejpam-86	530	17	27	27	NUM
ejpam-86	530	18	table	table	NOUN
ejpam-86	530	19	7	7	NUM
ejpam-86	530	20	:	:	PUNCT
ejpam-86	530	21	top	top	ADJ
ejpam-86	530	22	5	5	NUM
ejpam-86	530	23	ranking	ranking	ADJ
ejpam-86	530	24	subsets	subset	NOUN
ejpam-86	530	25	selected	select	VERB
ejpam-86	530	26	under	under	ADP
ejpam-86	530	27	type	type	NOUN
ejpam-86	530	28	ii	ii	PROPN
ejpam-86	530	29	mvper	mvper	PROPN
ejpam-86	530	30	model	model	NOUN
ejpam-86	530	31	.	.	PUNCT
ejpam-86	531	1	ranking	rank	VERB
ejpam-86	531	2	1	1	NUM
ejpam-86	531	3	2	2	NUM
ejpam-86	531	4	3	3	NUM
ejpam-86	531	5	4	4	NUM
ejpam-86	531	6	5	5	NUM
ejpam-86	531	7	subsets	subset	NOUN
ejpam-86	531	8	selected	select	VERB
ejpam-86	531	9	by	by	ADP
ejpam-86	531	10	aic	aic	PROPN
ejpam-86	531	11	subset	subset	VERB
ejpam-86	531	12	110000	110000	NUM
ejpam-86	531	13	110100	110100	NUM
ejpam-86	531	14	110001	110001	NUM
ejpam-86	531	15	110010	110010	NUM
ejpam-86	531	16	111001	111001	NUM
ejpam-86	531	17	β̂mom	β̂mom	NOUN
ejpam-86	532	1	0.0041	0.0041	NUM
ejpam-86	532	2	0.0083	0.0083	NUM
ejpam-86	532	3	0.0077	0.0077	NUM
ejpam-86	532	4	10.0000†	10.0000†	NUM
ejpam-86	532	5	9.9994	9.9994	NUM
ejpam-86	532	6	β̂mle	β̂mle	PUNCT
ejpam-86	532	7	2.5533	2.5533	NUM
ejpam-86	532	8	2.5115	2.5115	NUM
ejpam-86	532	9	2.2101	2.2101	NUM
ejpam-86	532	10	2.2664	2.2664	NUM
ejpam-86	532	11	2.1740	2.1740	NUM
ejpam-86	532	12	aic	aic	PROPN
ejpam-86	532	13	632.8	632.8	NUM
ejpam-86	532	14	638.9	638.9	NUM
ejpam-86	532	15	658.7	658.7	NUM
ejpam-86	532	16	661.4	661.4	NUM
ejpam-86	532	17	666.3	666.3	NUM
ejpam-86	532	18	subsets	subset	NOUN
ejpam-86	532	19	selected	select	VERB
ejpam-86	532	20	by	by	ADP
ejpam-86	532	21	icomp(ifim	icomp(ifim	NOUN
ejpam-86	532	22	)	)	PUNCT
ejpam-86	532	23	subset	subset	VERB
ejpam-86	532	24	110000	110000	NUM
ejpam-86	532	25	100000	100000	NUM
ejpam-86	532	26	110100	110100	NUM
ejpam-86	532	27	110010	110010	NUM
ejpam-86	532	28	110001	110001	NUM
ejpam-86	532	29	β̂mom	β̂mom	NOUN
ejpam-86	533	1	0.0041	0.0041	NUM
ejpam-86	533	2	4.1282	4.1282	NUM
ejpam-86	533	3	0.0083	0.0083	NUM
ejpam-86	533	4	10.0000	10.0000	NUM
ejpam-86	533	5	0.0077	0.0077	NUM
ejpam-86	533	6	β̂mle	β̂mle	PUNCT
ejpam-86	534	1	2.5533	2.5533	NUM
ejpam-86	534	2	2.2379	2.2379	NUM
ejpam-86	534	3	2.5115	2.5115	NUM
ejpam-86	534	4	2.2664	2.2664	NUM
ejpam-86	534	5	2.2101	2.2101	NUM
ejpam-86	534	6	icomp	icomp	NOUN
ejpam-86	534	7	939.3	939.3	NUM
ejpam-86	534	8	1018.0	1018.0	NUM
ejpam-86	534	9	1034.1	1034.1	NUM
ejpam-86	534	10	1043.8	1043.8	NUM
ejpam-86	534	11	1044.8	1044.8	NUM
ejpam-86	534	12	takes	take	VERB
ejpam-86	534	13	the	the	DET
ejpam-86	534	14	dependency	dependency	NOUN
ejpam-86	534	15	structure	structure	NOUN
ejpam-86	534	16	of	of	ADP
ejpam-86	534	17	the	the	DET
ejpam-86	534	18	data	datum	NOUN
ejpam-86	534	19	into	into	ADP
ejpam-86	534	20	account	account	NOUN
ejpam-86	534	21	.	.	PUNCT
ejpam-86	535	1	0	0	NUM
ejpam-86	536	1	5	5	NUM
ejpam-86	536	2	10	10	NUM
ejpam-86	536	3	15	15	NUM
ejpam-86	536	4	0	0	NUM
ejpam-86	536	5	5	5	NUM
ejpam-86	536	6	10	10	NUM
ejpam-86	536	7	15	15	NUM
ejpam-86	536	8	chi−square	chi−square	ADP
ejpam-86	536	9	quantile−quantile	quantile−quantile	ADJ
ejpam-86	536	10	plot	plot	NOUN
ejpam-86	536	11	o	o	PROPN
ejpam-86	536	12	rd	rd	NOUN
ejpam-86	537	1	er	er	INTJ
ejpam-86	537	2	ed	ed	NOUN
ejpam-86	537	3	s	s	PROPN
ejpam-86	537	4	qu	qu	PROPN
ejpam-86	537	5	ar	ar	PROPN
ejpam-86	537	6	ed	ed	PROPN
ejpam-86	537	7	m	m	VERB
ejpam-86	537	8	ah	ah	INTJ
ejpam-86	537	9	al	al	PROPN
ejpam-86	537	10	an	an	DET
ejpam-86	537	11	ob	ob	NOUN
ejpam-86	537	12	is	be	AUX
ejpam-86	537	13	d	d	NOUN
ejpam-86	537	14	is	be	AUX
ejpam-86	537	15	ta	ta	PROPN
ejpam-86	537	16	nc	nc	PROPN
ejpam-86	537	17	e	e	PROPN
ejpam-86	537	18	of	of	ADP
ejpam-86	537	19	r	r	NOUN
ejpam-86	537	20	es	es	X
ejpam-86	537	21	i	i	PROPN
ejpam-86	537	22	d	d	PROPN
ejpam-86	537	23	ua	ua	PROPN
ejpam-86	537	24	ls	ls	PROPN
ejpam-86	537	25	p−quantiles	p−quantiles	X
ejpam-86	537	26	figure	figure	VERB
ejpam-86	537	27	7	7	NUM
ejpam-86	537	28	:	:	PUNCT
ejpam-86	537	29	q	q	ADJ
ejpam-86	537	30	-	-	PUNCT
ejpam-86	537	31	q	q	NOUN
ejpam-86	537	32	plot	plot	NOUN
ejpam-86	537	33	under	under	ADP
ejpam-86	537	34	the	the	DET
ejpam-86	537	35	type	type	NOUN
ejpam-86	537	36	i	i	PRON
ejpam-86	537	37	model	model	NOUN
ejpam-86	537	38	.	.	PUNCT
ejpam-86	538	1	8	8	X
ejpam-86	538	2	.	.	X
ejpam-86	538	3	conclusion	conclusion	NOUN
ejpam-86	538	4	in	in	ADP
ejpam-86	538	5	this	this	DET
ejpam-86	538	6	paper	paper	NOUN
ejpam-86	538	7	we	we	PRON
ejpam-86	538	8	presented	present	VERB
ejpam-86	538	9	a	a	DET
ejpam-86	538	10	new	new	ADJ
ejpam-86	538	11	and	and	CCONJ
ejpam-86	538	12	novel	novel	ADJ
ejpam-86	538	13	model	model	NOUN
ejpam-86	538	14	selection	selection	NOUN
ejpam-86	538	15	technique	technique	NOUN
ejpam-86	538	16	to	to	PART
ejpam-86	538	17	deal	deal	VERB
ejpam-86	538	18	with	with	ADP
ejpam-86	538	19	nonnormality	nonnormality	NOUN
ejpam-86	538	20	in	in	ADP
ejpam-86	538	21	multivariate	multivariate	NOUN
ejpam-86	538	22	regression	regression	NOUN
ejpam-86	538	23	models	model	NOUN
ejpam-86	538	24	using	use	VERB
ejpam-86	538	25	information	information	NOUN
ejpam-86	538	26	-	-	PUNCT
ejpam-86	538	27	theoretic	theoretic	NOUN
ejpam-86	538	28	model	model	NOUN
ejpam-86	538	29	selection	selection	NOUN
ejpam-86	538	30	criteria	criterion	NOUN
ejpam-86	538	31	such	such	ADJ
ejpam-86	538	32	as	as	ADP
ejpam-86	538	33	aic	aic	PROPN
ejpam-86	538	34	and	and	CCONJ
ejpam-86	538	35	icomp	icomp	PROPN
ejpam-86	538	36	.	.	PUNCT
ejpam-86	539	1	we	we	PRON
ejpam-86	539	2	developed	develop	VERB
ejpam-86	539	3	two	two	NUM
ejpam-86	539	4	types	type	NOUN
ejpam-86	539	5	of	of	ADP
ejpam-86	539	6	mvper	mvper	NOUN
ejpam-86	539	7	models	model	NOUN
ejpam-86	539	8	.	.	PUNCT
ejpam-86	540	1	these	these	DET
ejpam-86	540	2	two	two	NUM
ejpam-86	540	3	types	type	NOUN
ejpam-86	540	4	of	of	ADP
ejpam-86	540	5	mvper	mvper	NOUN
ejpam-86	540	6	models	model	NOUN
ejpam-86	540	7	discussed	discuss	VERB
ejpam-86	540	8	in	in	ADP
ejpam-86	540	9	this	this	DET
ejpam-86	540	10	paper	paper	NOUN
ejpam-86	540	11	can	can	AUX
ejpam-86	540	12	be	be	AUX
ejpam-86	540	13	used	use	VERB
ejpam-86	540	14	to	to	PART
ejpam-86	540	15	model	model	VERB
ejpam-86	540	16	random	random	ADJ
ejpam-86	540	17	phenomena	phenomenon	NOUN
ejpam-86	540	18	whose	whose	DET
ejpam-86	540	19	observations	observation	NOUN
ejpam-86	540	20	are	be	AUX
ejpam-86	540	21	dependent	dependent	ADJ
ejpam-86	540	22	or	or	CCONJ
ejpam-86	540	23	independent	independent	ADJ
ejpam-86	540	24	when	when	SCONJ
ejpam-86	540	25	the	the	DET
ejpam-86	540	26	tails	tail	NOUN
ejpam-86	540	27	are	be	AUX
ejpam-86	540	28	thicker	thick	ADJ
ejpam-86	540	29	or	or	CCONJ
ejpam-86	540	30	thinner	thin	ADJ
ejpam-86	540	31	than	than	ADP
ejpam-86	540	32	those	those	PRON
ejpam-86	540	33	of	of	ADP
ejpam-86	540	34	multivariate	multivariate	NOUN
ejpam-86	540	35	normal	normal	ADJ
ejpam-86	540	36	distribution	distribution	NOUN
ejpam-86	540	37	which	which	PRON
ejpam-86	540	38	is	be	AUX
ejpam-86	540	39	used	use	VERB
ejpam-86	540	40	often	often	ADV
ejpam-86	540	41	in	in	ADP
ejpam-86	540	42	the	the	DET
ejpam-86	540	43	literature	literature	NOUN
ejpam-86	540	44	.	.	PUNCT
ejpam-86	541	1	as	as	ADP
ejpam-86	541	2	a	a	DET
ejpam-86	541	3	subfamily	subfamily	NOUN
ejpam-86	541	4	of	of	ADP
ejpam-86	541	5	matrix	matrix	NOUN
ejpam-86	541	6	ec	ec	NOUN
ejpam-86	541	7	distribution	distribution	NOUN
ejpam-86	541	8	,	,	PUNCT
ejpam-86	541	9	type	type	NOUN
ejpam-86	541	10	ii	ii	NOUN
ejpam-86	541	11	model	model	NOUN
ejpam-86	541	12	has	have	AUX
ejpam-86	541	13	been	be	AUX
ejpam-86	541	14	studied	study	VERB
ejpam-86	541	15	partly	partly	ADV
ejpam-86	541	16	in	in	ADP
ejpam-86	541	17	the	the	DET
ejpam-86	541	18	context	context	NOUN
ejpam-86	541	19	of	of	ADP
ejpam-86	541	20	multivariate	multivariate	NOUN
ejpam-86	541	21	ec	ec	PROPN
ejpam-86	541	22	regression	regression	NOUN
ejpam-86	541	23	model	model	NOUN
ejpam-86	541	24	such	such	ADJ
ejpam-86	541	25	as	as	ADP
ejpam-86	541	26	the	the	DET
ejpam-86	541	27	work	work	NOUN
ejpam-86	541	28	of	of	ADP
ejpam-86	541	29	fang	fang	PROPN
ejpam-86	541	30	and	and	CCONJ
ejpam-86	541	31	anderson	anderson	PROPN
ejpam-86	541	32	(	(	PUNCT
ejpam-86	542	1	[	[	X
ejpam-86	542	2	17	17	NUM
ejpam-86	542	3	,	,	PUNCT
ejpam-86	542	4	p.214	p.214	NUM
ejpam-86	542	5	]	]	PUNCT
ejpam-86	542	6	)	)	PUNCT
ejpam-86	542	7	and	and	CCONJ
ejpam-86	542	8	bozdogan	bozdogan	NOUN
ejpam-86	542	9	(	(	PUNCT
ejpam-86	542	10	[	[	X
ejpam-86	542	11	11	11	NUM
ejpam-86	542	12	]	]	NUM
ejpam-86	542	13	)	)	PUNCT
ejpam-86	542	14	,	,	PUNCT
ejpam-86	542	15	etc	etc	X
ejpam-86	542	16	.	.	X
ejpam-86	543	1	but	but	CCONJ
ejpam-86	543	2	the	the	DET
ejpam-86	543	3	type	type	NOUN
ejpam-86	543	4	i	i	PRON
ejpam-86	543	5	model	model	NOUN
ejpam-86	543	6	is	be	AUX
ejpam-86	543	7	seldom	seldom	ADV
ejpam-86	543	8	discussed	discuss	VERB
ejpam-86	543	9	in	in	ADP
ejpam-86	543	10	literature	literature	NOUN
ejpam-86	543	11	.	.	PUNCT
ejpam-86	544	1	one	one	NUM
ejpam-86	544	2	special	special	ADJ
ejpam-86	544	3	difficulty	difficulty	NOUN
ejpam-86	544	4	of	of	ADP
ejpam-86	544	5	mvper	mvper	NOUN
ejpam-86	544	6	models	model	NOUN
ejpam-86	544	7	is	be	AUX
ejpam-86	544	8	to	to	PART
ejpam-86	544	9	estimate	estimate	VERB
ejpam-86	544	10	the	the	DET
ejpam-86	544	11	shape	shape	NOUN
ejpam-86	544	12	parameter	parameter	NOUN
ejpam-86	544	13	β	β	NOUN
ejpam-86	544	14	.	.	PUNCT
ejpam-86	545	1	in	in	ADP
ejpam-86	545	2	this	this	DET
ejpam-86	545	3	paper	paper	NOUN
ejpam-86	545	4	,	,	PUNCT
ejpam-86	545	5	we	we	PRON
ejpam-86	545	6	provided	provide	VERB
ejpam-86	545	7	methods	method	NOUN
ejpam-86	545	8	to	to	PART
ejpam-86	545	9	obtain	obtain	VERB
ejpam-86	545	10	mom	mom	NOUN
ejpam-86	545	11	estimates	estimate	NOUN
ejpam-86	545	12	and	and	CCONJ
ejpam-86	545	13	mles	mle	NOUN
ejpam-86	545	14	for	for	ADP
ejpam-86	545	15	both	both	DET
ejpam-86	545	16	types	type	NOUN
ejpam-86	545	17	of	of	ADP
ejpam-86	545	18	models	model	NOUN
ejpam-86	545	19	.	.	PUNCT
ejpam-86	546	1	from	from	ADP
ejpam-86	546	2	the	the	DET
ejpam-86	546	3	above	above	ADJ
ejpam-86	546	4	results	result	NOUN
ejpam-86	546	5	,	,	PUNCT
ejpam-86	546	6	we	we	PRON
ejpam-86	546	7	see	see	VERB
ejpam-86	546	8	that	that	SCONJ
ejpam-86	546	9	the	the	DET
ejpam-86	546	10	mles	mle	NOUN
ejpam-86	546	11	of	of	ADP
ejpam-86	546	12	the	the	DET
ejpam-86	546	13	type	type	NOUN
ejpam-86	546	14	ii	ii	PROPN
ejpam-86	546	15	model	model	NOUN
ejpam-86	546	16	actually	actually	ADV
ejpam-86	546	17	can	can	AUX
ejpam-86	546	18	m.	m.	VERB
ejpam-86	546	19	liu	liu	PROPN
ejpam-86	546	20	and	and	CCONJ
ejpam-86	546	21	h.	h.	PROPN
ejpam-86	546	22	bozdogan	bozdogan	PROPN
ejpam-86	546	23	/	/	SYM
ejpam-86	546	24	eur	eur	PROPN
ejpam-86	546	25	.	.	PUNCT
ejpam-86	547	1	j.	j.	PROPN
ejpam-86	547	2	pure	pure	PROPN
ejpam-86	547	3	appl	appl	PROPN
ejpam-86	547	4	.	.	PROPN
ejpam-86	547	5	math	math	PROPN
ejpam-86	547	6	,	,	PUNCT
ejpam-86	547	7	1	1	NUM
ejpam-86	547	8	(	(	PUNCT
ejpam-86	547	9	2008	2008	NUM
ejpam-86	547	10	)	)	PUNCT
ejpam-86	547	11	,	,	PUNCT
ejpam-86	547	12	(	(	PUNCT
ejpam-86	547	13	4	4	NUM
ejpam-86	547	14	-	-	SYM
ejpam-86	547	15	37	37	NUM
ejpam-86	547	16	)	)	PUNCT
ejpam-86	547	17	28	28	NUM
ejpam-86	547	18	−3	−3	NOUN
ejpam-86	547	19	−2	−2	NOUN
ejpam-86	547	20	−1	−1	NOUN
ejpam-86	547	21	0	0	NUM
ejpam-86	547	22	1	1	NUM
ejpam-86	547	23	2	2	NUM
ejpam-86	547	24	3	3	NUM
ejpam-86	547	25	−30	−30	NOUN
ejpam-86	547	26	−20	−20	NOUN
ejpam-86	547	27	−10	−10	X
ejpam-86	547	28	0	0	NUM
ejpam-86	547	29	10	10	NUM
ejpam-86	547	30	20	20	NUM
ejpam-86	547	31	30	30	NUM
ejpam-86	547	32	40	40	NUM
ejpam-86	547	33	50	50	NUM
ejpam-86	547	34	60	60	NUM
ejpam-86	547	35	normal	normal	ADJ
ejpam-86	547	36	quantile−quantile	quantile−quantile	ADJ
ejpam-86	547	37	plot	plot	NOUN
ejpam-86	547	38	of	of	ADP
ejpam-86	547	39	residuals	residual	NOUN
ejpam-86	547	40	o	o	PROPN
ejpam-86	547	41	rd	rd	NOUN
ejpam-86	548	1	er	er	INTJ
ejpam-86	548	2	ed	ed	NOUN
ejpam-86	548	3	r	r	NOUN
ejpam-86	548	4	es	es	X
ejpam-86	548	5	i	i	PROPN
ejpam-86	548	6	d	d	PROPN
ejpam-86	548	7	ua	ua	PROPN
ejpam-86	548	8	ls	ls	PROPN
ejpam-86	548	9	p−quantiles	p−quantile	NOUN
ejpam-86	548	10	figure	figure	VERB
ejpam-86	548	11	8	8	NUM
ejpam-86	548	12	:	:	PUNCT
ejpam-86	548	13	q	q	ADJ
ejpam-86	548	14	-	-	PUNCT
ejpam-86	548	15	q	q	NOUN
ejpam-86	548	16	plot	plot	NOUN
ejpam-86	548	17	under	under	ADP
ejpam-86	548	18	the	the	DET
ejpam-86	548	19	type	type	NOUN
ejpam-86	548	20	ii	ii	PROPN
ejpam-86	548	21	model	model	NOUN
ejpam-86	548	22	.	.	PUNCT
ejpam-86	549	1	not	not	PART
ejpam-86	549	2	be	be	AUX
ejpam-86	549	3	obtained	obtain	VERB
ejpam-86	549	4	from	from	ADP
ejpam-86	549	5	a	a	DET
ejpam-86	549	6	single	single	ADJ
ejpam-86	549	7	sample	sample	NOUN
ejpam-86	549	8	.	.	PUNCT
ejpam-86	550	1	but	but	CCONJ
ejpam-86	550	2	a	a	DET
ejpam-86	550	3	two	two	NUM
ejpam-86	550	4	step	step	NOUN
ejpam-86	550	5	re	re	VERB
ejpam-86	550	6	-	-	ADJ
ejpam-86	550	7	sampling	sample	VERB
ejpam-86	550	8	method	method	NOUN
ejpam-86	550	9	is	be	AUX
ejpam-86	550	10	used	use	VERB
ejpam-86	550	11	and	and	CCONJ
ejpam-86	550	12	developed	develop	VERB
ejpam-86	550	13	to	to	PART
ejpam-86	550	14	solve	solve	VERB
ejpam-86	550	15	this	this	DET
ejpam-86	550	16	problem	problem	NOUN
ejpam-86	550	17	.	.	PUNCT
ejpam-86	551	1	our	our	PRON
ejpam-86	551	2	simulated	simulate	VERB
ejpam-86	551	3	as	as	ADV
ejpam-86	551	4	well	well	ADV
ejpam-86	551	5	as	as	ADP
ejpam-86	551	6	the	the	DET
ejpam-86	551	7	real	real	ADJ
ejpam-86	551	8	computational	computational	ADJ
ejpam-86	551	9	examples	example	NOUN
ejpam-86	551	10	show	show	VERB
ejpam-86	551	11	that	that	SCONJ
ejpam-86	551	12	the	the	DET
ejpam-86	551	13	hybrid	hybrid	NOUN
ejpam-86	551	14	of	of	ADP
ejpam-86	551	15	information	information	NOUN
ejpam-86	551	16	criteria	criterion	NOUN
ejpam-86	551	17	such	such	ADJ
ejpam-86	551	18	as	as	ADP
ejpam-86	551	19	aic	aic	PROPN
ejpam-86	551	20	and	and	CCONJ
ejpam-86	551	21	icomp(ifim	icomp(ifim	NOUN
ejpam-86	551	22	)	)	PUNCT
ejpam-86	551	23	and	and	CCONJ
ejpam-86	551	24	the	the	DET
ejpam-86	551	25	ga	ga	PROPN
ejpam-86	551	26	approach	approach	NOUN
ejpam-86	551	27	works	work	VERB
ejpam-86	551	28	well	well	ADV
ejpam-86	551	29	for	for	ADP
ejpam-86	551	30	model	model	NOUN
ejpam-86	551	31	selection	selection	NOUN
ejpam-86	551	32	in	in	ADP
ejpam-86	551	33	both	both	DET
ejpam-86	551	34	cases	case	NOUN
ejpam-86	551	35	.	.	PUNCT
ejpam-86	552	1	advantages	advantage	NOUN
ejpam-86	552	2	of	of	ADP
ejpam-86	552	3	this	this	DET
ejpam-86	552	4	approach	approach	NOUN
ejpam-86	552	5	introduced	introduce	VERB
ejpam-86	552	6	in	in	ADP
ejpam-86	552	7	this	this	DET
ejpam-86	552	8	paper	paper	NOUN
ejpam-86	552	9	are	be	AUX
ejpam-86	552	10	flexible	flexible	ADJ
ejpam-86	552	11	to	to	PART
ejpam-86	552	12	resolve	resolve	VERB
ejpam-86	552	13	many	many	ADJ
ejpam-86	552	14	problems	problem	NOUN
ejpam-86	552	15	in	in	ADP
ejpam-86	552	16	vector	vector	NOUN
ejpam-86	552	17	autoregressive	autoregressive	ADJ
ejpam-86	552	18	(	(	PUNCT
ejpam-86	552	19	var	var	NOUN
ejpam-86	552	20	)	)	PUNCT
ejpam-86	552	21	models	model	NOUN
ejpam-86	552	22	,	,	PUNCT
ejpam-86	552	23	in	in	ADP
ejpam-86	552	24	kernel	kernel	PROPN
ejpam-86	552	25	support	support	PROPN
ejpam-86	552	26	vector	vector	NOUN
ejpam-86	552	27	machines	machine	NOUN
ejpam-86	552	28	(	(	PUNCT
ejpam-86	552	29	svms	svms	NOUN
ejpam-86	552	30	)	)	PUNCT
ejpam-86	552	31	,	,	PUNCT
ejpam-86	552	32	etc	etc	X
ejpam-86	552	33	.	.	X
ejpam-86	552	34	by	by	ADP
ejpam-86	552	35	taking	take	VERB
ejpam-86	552	36	the	the	DET
ejpam-86	552	37	dependency	dependency	NOUN
ejpam-86	552	38	in	in	ADP
ejpam-86	552	39	the	the	DET
ejpam-86	552	40	data	datum	NOUN
ejpam-86	552	41	into	into	ADP
ejpam-86	552	42	account	account	NOUN
ejpam-86	552	43	.	.	PUNCT
ejpam-86	553	1	all	all	DET
ejpam-86	553	2	our	our	PRON
ejpam-86	553	3	computations	computation	NOUN
ejpam-86	553	4	are	be	AUX
ejpam-86	553	5	carried	carry	VERB
ejpam-86	553	6	out	out	ADP
ejpam-86	553	7	using	use	VERB
ejpam-86	553	8	a	a	DET
ejpam-86	553	9	newly	newly	ADV
ejpam-86	553	10	developed	develop	VERB
ejpam-86	553	11	computational	computational	ADJ
ejpam-86	553	12	matlab	matlab	NOUN
ejpam-86	553	13	modules	module	NOUN
ejpam-86	553	14	with	with	ADP
ejpam-86	553	15	ga	ga	PROPN
ejpam-86	553	16	.	.	PUNCT
ejpam-86	553	17	acknowledgements	acknowledgement	NOUN
ejpam-86	553	18	this	this	DET
ejpam-86	553	19	paper	paper	NOUN
ejpam-86	553	20	is	be	AUX
ejpam-86	553	21	based	base	VERB
ejpam-86	553	22	on	on	ADP
ejpam-86	553	23	the	the	DET
ejpam-86	553	24	results	result	NOUN
ejpam-86	553	25	of	of	ADP
ejpam-86	553	26	the	the	DET
ejpam-86	553	27	ph.d	ph.d	PROPN
ejpam-86	553	28	.	.	PUNCT
ejpam-86	554	1	thesis	thesis	NOUN
ejpam-86	554	2	of	of	ADP
ejpam-86	554	3	the	the	DET
ejpam-86	554	4	first	first	ADJ
ejpam-86	554	5	author	author	NOUN
ejpam-86	554	6	under	under	ADP
ejpam-86	554	7	the	the	DET
ejpam-86	554	8	supervision	supervision	NOUN
ejpam-86	554	9	of	of	ADP
ejpam-86	554	10	prof	prof	PROPN
ejpam-86	554	11	.	.	PUNCT
ejpam-86	555	1	bozdogan	bozdogan	PROPN
ejpam-86	555	2	.	.	PUNCT
ejpam-86	556	1	first	first	ADJ
ejpam-86	556	2	author	author	NOUN
ejpam-86	556	3	extends	extend	VERB
ejpam-86	556	4	his	his	PRON
ejpam-86	556	5	thanks	thank	NOUN
ejpam-86	556	6	and	and	CCONJ
ejpam-86	556	7	gratitude	gratitude	NOUN
ejpam-86	556	8	to	to	PART
ejpam-86	556	9	prof	prof	VERB
ejpam-86	556	10	.	.	PUNCT
ejpam-86	557	1	bozdogan	bozdogan	PROPN
ejpam-86	557	2	for	for	ADP
ejpam-86	557	3	all	all	DET
ejpam-86	557	4	the	the	DET
ejpam-86	557	5	help	help	NOUN
ejpam-86	557	6	provided	provide	VERB
ejpam-86	557	7	throughout	throughout	ADP
ejpam-86	557	8	.	.	PUNCT
ejpam-86	558	1	this	this	DET
ejpam-86	558	2	paper	paper	NOUN
ejpam-86	558	3	has	have	AUX
ejpam-86	558	4	been	be	AUX
ejpam-86	558	5	presented	present	VERB
ejpam-86	558	6	by	by	ADP
ejpam-86	558	7	prof	prof	PROPN
ejpam-86	558	8	.	.	PUNCT
ejpam-86	559	1	bozdogan	bozdogan	PROPN
ejpam-86	559	2	as	as	ADP
ejpam-86	559	3	an	an	DET
ejpam-86	559	4	invited	invite	VERB
ejpam-86	559	5	paper	paper	NOUN
ejpam-86	559	6	at	at	ADP
ejpam-86	559	7	bayes	bayes	NOUN
ejpam-86	559	8	,	,	PUNCT
ejpam-86	559	9	multivariate	multivariate	VERB
ejpam-86	559	10	analysis	analysis	NOUN
ejpam-86	559	11	,	,	PUNCT
ejpam-86	559	12	and	and	CCONJ
ejpam-86	559	13	casm	casm	NOUN
ejpam-86	559	14	conference	conference	NOUN
ejpam-86	559	15	in	in	ADP
ejpam-86	559	16	honor	honor	NOUN
ejpam-86	559	17	of	of	ADP
ejpam-86	559	18	professor	professor	PROPN
ejpam-86	559	19	s.	s.	PROPN
ejpam-86	559	20	jim	jim	PROPN
ejpam-86	559	21	press	press	PROPN
ejpam-86	559	22	,	,	PUNCT
ejpam-86	559	23	distinguished	distinguished	ADJ
ejpam-86	559	24	university	university	NOUN
ejpam-86	559	25	professor	professor	NOUN
ejpam-86	559	26	,	,	PUNCT
ejpam-86	559	27	at	at	ADP
ejpam-86	559	28	the	the	DET
ejpam-86	559	29	university	university	PROPN
ejpam-86	559	30	of	of	ADP
ejpam-86	559	31	california	california	PROPN
ejpam-86	559	32	at	at	ADP
ejpam-86	559	33	riverside	riverside	PROPN
ejpam-86	559	34	,	,	PUNCT
ejpam-86	559	35	california	california	PROPN
ejpam-86	559	36	during	during	ADP
ejpam-86	559	37	may	may	PROPN
ejpam-86	559	38	13	13	NUM
ejpam-86	559	39	-	-	SYM
ejpam-86	559	40	14	14	NUM
ejpam-86	559	41	,	,	PUNCT
ejpam-86	559	42	2005	2005	NUM
ejpam-86	559	43	.	.	PUNCT
ejpam-86	560	1	we	we	PRON
ejpam-86	560	2	are	be	AUX
ejpam-86	560	3	deeply	deeply	ADV
ejpam-86	560	4	honored	honor	VERB
ejpam-86	560	5	to	to	PART
ejpam-86	560	6	dedicate	dedicate	VERB
ejpam-86	560	7	this	this	DET
ejpam-86	560	8	paper	paper	NOUN
ejpam-86	560	9	to	to	ADP
ejpam-86	560	10	professor	professor	PROPN
ejpam-86	560	11	jim	jim	PROPN
ejpam-86	560	12	press	press	PROPN
ejpam-86	560	13	in	in	ADP
ejpam-86	560	14	honor	honor	NOUN
ejpam-86	560	15	of	of	ADP
ejpam-86	560	16	his	his	PRON
ejpam-86	560	17	50	50	NUM
ejpam-86	560	18	years	year	NOUN
ejpam-86	560	19	of	of	ADP
ejpam-86	560	20	remarkable	remarkable	ADJ
ejpam-86	560	21	career	career	NOUN
ejpam-86	560	22	and	and	CCONJ
ejpam-86	560	23	scientific	scientific	ADJ
ejpam-86	560	24	contributions	contribution	NOUN
ejpam-86	560	25	to	to	ADP
ejpam-86	560	26	bayesian	bayesian	NOUN
ejpam-86	560	27	and	and	CCONJ
ejpam-86	560	28	multivariate	multivariate	VERB
ejpam-86	560	29	statistics	statistic	NOUN
ejpam-86	560	30	on	on	ADP
ejpam-86	560	31	the	the	DET
ejpam-86	560	32	occasion	occasion	NOUN
ejpam-86	560	33	of	of	ADP
ejpam-86	560	34	his	his	PRON
ejpam-86	560	35	retirement	retirement	NOUN
ejpam-86	560	36	.	.	PUNCT
ejpam-86	561	1	professor	professor	NOUN
ejpam-86	561	2	bozdogan	bozdogan	PROPN
ejpam-86	561	3	gratefully	gratefully	ADV
ejpam-86	561	4	acknowledges	acknowledge	VERB
ejpam-86	561	5	funding	funding	NOUN
ejpam-86	561	6	for	for	ADP
ejpam-86	561	7	his	his	PRON
ejpam-86	561	8	research	research	NOUN
ejpam-86	561	9	from	from	ADP
ejpam-86	561	10	the	the	DET
ejpam-86	561	11	scholarly	scholarly	ADJ
ejpam-86	561	12	research	research	NOUN
ejpam-86	561	13	grant	grant	NOUN
ejpam-86	561	14	program	program	NOUN
ejpam-86	561	15	(	(	PUNCT
ejpam-86	561	16	srgp	srgp	ADJ
ejpam-86	561	17	)	)	PUNCT
ejpam-86	561	18	awards	award	NOUN
ejpam-86	561	19	of	of	ADP
ejpam-86	561	20	the	the	DET
ejpam-86	561	21	college	college	PROPN
ejpam-86	561	22	of	of	ADP
ejpam-86	561	23	business	business	PROPN
ejpam-86	561	24	administration	administration	NOUN
ejpam-86	561	25	at	at	ADP
ejpam-86	561	26	the	the	DET
ejpam-86	561	27	university	university	PROPN
ejpam-86	561	28	of	of	ADP
ejpam-86	561	29	tennessee	tennessee	PROPN
ejpam-86	561	30	in	in	ADP
ejpam-86	561	31	knoxville	knoxville	PROPN
ejpam-86	561	32	,	,	PUNCT
ejpam-86	561	33	during	during	ADP
ejpam-86	561	34	2002	2002	NUM
ejpam-86	561	35	-	-	SYM
ejpam-86	561	36	03	03	NUM
ejpam-86	561	37	under	under	ADP
ejpam-86	561	38	the	the	DET
ejpam-86	561	39	title	title	NOUN
ejpam-86	561	40	multivariate	multivariate	NOUN
ejpam-86	561	41	regression	regression	NOUN
ejpam-86	561	42	model	model	NOUN
ejpam-86	561	43	with	with	ADP
ejpam-86	561	44	nonnormal	nonnormal	ADJ
ejpam-86	561	45	error	error	NOUN
ejpam-86	561	46	terms	term	NOUN
ejpam-86	561	47	,	,	PUNCT
ejpam-86	561	48	genetic	genetic	ADJ
ejpam-86	561	49	algorithm	algorithm	NOUN
ejpam-86	561	50	and	and	CCONJ
ejpam-86	561	51	information	information	NOUN
ejpam-86	561	52	complexity	complexity	NOUN
ejpam-86	561	53	.	.	PUNCT
ejpam-86	562	1	we	we	PRON
ejpam-86	562	2	also	also	ADV
ejpam-86	562	3	extend	extend	VERB
ejpam-86	562	4	our	our	PRON
ejpam-86	562	5	thanks	thank	NOUN
ejpam-86	562	6	and	and	CCONJ
ejpam-86	562	7	deep	deep	ADJ
ejpam-86	562	8	gratitude	gratitude	NOUN
ejpam-86	562	9	to	to	PART
ejpam-86	562	10	prof	prof	PROPN
ejpam-86	562	11	.	.	PUNCT
ejpam-86	563	1	dr	dr	PROPN
ejpam-86	563	2	.	.	PROPN
ejpam-86	563	3	eyüp	eyüp	PROPN
ejpam-86	563	4	çetin	çetin	PROPN
ejpam-86	563	5	,	,	PUNCT
ejpam-86	563	6	editor	editor	NOUN
ejpam-86	563	7	-	-	PUNCT
ejpam-86	563	8	inchief	inchief	NOUN
ejpam-86	563	9	of	of	ADP
ejpam-86	563	10	ejpam	ejpam	NOUN
ejpam-86	563	11	,	,	PUNCT
ejpam-86	563	12	for	for	ADP
ejpam-86	563	13	inviting	invite	VERB
ejpam-86	563	14	prof	prof	NOUN
ejpam-86	563	15	.	.	PUNCT
ejpam-86	564	1	bozdogan	bozdogan	PROPN
ejpam-86	564	2	to	to	PART
ejpam-86	564	3	make	make	VERB
ejpam-86	564	4	an	an	DET
ejpam-86	564	5	contribution	contribution	NOUN
ejpam-86	564	6	to	to	ADP
ejpam-86	564	7	the	the	DET
ejpam-86	564	8	“	"	PUNCT
ejpam-86	564	9	honorary	honorary	ADJ
ejpam-86	564	10	invited	invite	VERB
ejpam-86	564	11	paper	paper	NOUN
ejpam-86	564	12	”	"	PUNCT
ejpam-86	564	13	issue	issue	NOUN
ejpam-86	564	14	of	of	ADP
ejpam-86	564	15	ejpam	ejpam	NOUN
ejpam-86	564	16	.	.	PUNCT
ejpam-86	565	1	we	we	PRON
ejpam-86	565	2	are	be	AUX
ejpam-86	565	3	privileged	privileged	ADJ
ejpam-86	565	4	to	to	PART
ejpam-86	565	5	make	make	VERB
ejpam-86	565	6	this	this	DET
ejpam-86	565	7	contribution	contribution	NOUN
ejpam-86	565	8	in	in	ADP
ejpam-86	565	9	the	the	DET
ejpam-86	565	10	opening	opening	NOUN
ejpam-86	565	11	issue	issue	NOUN
ejpam-86	565	12	and	and	CCONJ
ejpam-86	565	13	wish	wish	VERB
ejpam-86	565	14	the	the	DET
ejpam-86	565	15	success	success	NOUN
ejpam-86	565	16	of	of	ADP
ejpam-86	565	17	this	this	DET
ejpam-86	565	18	prestigious	prestigious	ADJ
ejpam-86	565	19	journal	journal	NOUN
ejpam-86	565	20	.	.	PUNCT
ejpam-86	566	1	m.	m.	PROPN
ejpam-86	566	2	liu	liu	PROPN
ejpam-86	566	3	and	and	CCONJ
ejpam-86	566	4	h.	h.	PROPN
ejpam-86	566	5	bozdogan	bozdogan	PROPN
ejpam-86	566	6	/	/	SYM
ejpam-86	566	7	eur	eur	PROPN
ejpam-86	566	8	.	.	PUNCT
ejpam-86	567	1	j.	j.	PROPN
ejpam-86	567	2	pure	pure	PROPN
ejpam-86	567	3	appl	appl	PROPN
ejpam-86	567	4	.	.	PROPN
ejpam-86	567	5	math	math	PROPN
ejpam-86	567	6	,	,	PUNCT
ejpam-86	567	7	1	1	NUM
ejpam-86	567	8	(	(	PUNCT
ejpam-86	567	9	2008	2008	NUM
ejpam-86	567	10	)	)	PUNCT
ejpam-86	567	11	,	,	PUNCT
ejpam-86	567	12	(	(	PUNCT
ejpam-86	567	13	4	4	NUM
ejpam-86	567	14	-	-	SYM
ejpam-86	567	15	37	37	NUM
ejpam-86	567	16	)	)	PUNCT
ejpam-86	567	17	29	29	NUM
ejpam-86	567	18	the	the	DET
ejpam-86	567	19	pseudo	pseudo	NOUN
ejpam-86	567	20	code	code	NOUN
ejpam-86	567	21	of	of	ADP
ejpam-86	567	22	genetic	genetic	ADJ
ejpam-86	567	23	algorithm	algorithm	NOUN
ejpam-86	567	24	algorithm	algorithm	NOUN
ejpam-86	567	25	1	1	NUM
ejpam-86	567	26	:	:	PUNCT
ejpam-86	567	27	the	the	DET
ejpam-86	567	28	pseudo	pseudo	NOUN
ejpam-86	567	29	code	code	NOUN
ejpam-86	567	30	of	of	ADP
ejpam-86	567	31	ga	ga	PROPN
ejpam-86	567	32	for	for	ADP
ejpam-86	567	33	regression	regression	NOUN
ejpam-86	567	34	model	model	NOUN
ejpam-86	567	35	selection	selection	PROPN
ejpam-86	567	36	ps	ps	PROPN
ejpam-86	567	37	=	=	NOUN
ejpam-86	567	38	population	population	NOUN
ejpam-86	567	39	size	size	NOUN
ejpam-86	567	40	;	;	PUNCT
ejpam-86	567	41	ng	ng	PROPN
ejpam-86	567	42	=	=	NOUN
ejpam-86	567	43	number	number	NOUN
ejpam-86	567	44	of	of	ADP
ejpam-86	567	45	generations	generation	NOUN
ejpam-86	567	46	of	of	ADP
ejpam-86	567	47	ga	ga	PROPN
ejpam-86	567	48	;	;	PUNCT
ejpam-86	567	49	pc	pc	NOUN
ejpam-86	567	50	=	=	NOUN
ejpam-86	567	51	crossover	crossover	NOUN
ejpam-86	567	52	probability	probability	NOUN
ejpam-86	567	53	;	;	PUNCT
ejpam-86	567	54	pm	pm	NOUN
ejpam-86	567	55	=	=	NOUN
ejpam-86	567	56	mutation	mutation	NOUN
ejpam-86	567	57	probability	probability	NOUN
ejpam-86	567	58	;	;	PUNCT
ejpam-86	567	59	pe	pe	PROPN
ejpam-86	567	60	=	=	PROPN
ejpam-86	567	61	ga	ga	NOUN
ejpam-86	567	62	engineering	engineering	NOUN
ejpam-86	567	63	probability	probability	NOUN
ejpam-86	567	64	;	;	PUNCT
ejpam-86	567	65	ct	ct	PROPN
ejpam-86	567	66	=	=	NOUN
ejpam-86	567	67	crossover	crossover	NOUN
ejpam-86	567	68	type	type	NOUN
ejpam-86	567	69	;	;	PUNCT
ejpam-86	567	70	i	i	NOUN
ejpam-86	567	71	=	=	NOUN
ejpam-86	567	72	1	1	NUM
ejpam-86	567	73	;	;	PUNCT
ejpam-86	567	74	generate	generate	VERB
ejpam-86	567	75	ps	ps	NOUN
ejpam-86	567	76	original	original	ADJ
ejpam-86	567	77	solutions	solution	NOUN
ejpam-86	567	78	randomly	randomly	ADV
ejpam-86	567	79	;	;	PUNCT
ejpam-86	567	80	encode	encode	VERB
ejpam-86	567	81	the	the	DET
ejpam-86	567	82	solutions	solution	NOUN
ejpam-86	567	83	to	to	ADP
ejpam-86	567	84	binary	binary	ADJ
ejpam-86	567	85	strings	string	NOUN
ejpam-86	567	86	(	(	PUNCT
ejpam-86	567	87	chromosomes	chromosome	NOUN
ejpam-86	567	88	)	)	PUNCT
ejpam-86	567	89	;	;	PUNCT
ejpam-86	567	90	for	for	SCONJ
ejpam-86	567	91	i	i	PROPN
ejpam-86	567	92	≤	≤	NOUN
ejpam-86	567	93	ng	ng	PROPN
ejpam-86	567	94	evaluate	evaluate	VERB
ejpam-86	567	95	the	the	DET
ejpam-86	567	96	fitness	fitness	NOUN
ejpam-86	567	97	of	of	ADP
ejpam-86	567	98	each	each	DET
ejpam-86	567	99	solution	solution	NOUN
ejpam-86	567	100	of	of	ADP
ejpam-86	567	101	generation	generation	NOUN
ejpam-86	567	102	i	i	PRON
ejpam-86	567	103	;	;	PUNCT
ejpam-86	567	104	find	find	VERB
ejpam-86	567	105	the	the	DET
ejpam-86	567	106	best	good	ADJ
ejpam-86	567	107	solution	solution	NOUN
ejpam-86	567	108	in	in	ADP
ejpam-86	567	109	generation	generation	NOUN
ejpam-86	567	110	i	i	PRON
ejpam-86	567	111	;	;	PUNCT
ejpam-86	567	112	if	if	SCONJ
ejpam-86	567	113	elitism	elitism	NOUN
ejpam-86	567	114	is	be	AUX
ejpam-86	567	115	true	true	ADJ
ejpam-86	567	116	then	then	ADV
ejpam-86	567	117	select	select	ADJ
ejpam-86	567	118	ng/2−	ng/2−	SYM
ejpam-86	567	119	1	1	NUM
ejpam-86	567	120	pairs	pair	NOUN
ejpam-86	567	121	of	of	ADP
ejpam-86	567	122	parent	parent	NOUN
ejpam-86	567	123	chromosomes	chromosome	NOUN
ejpam-86	567	124	from	from	ADP
ejpam-86	567	125	generation	generation	NOUN
ejpam-86	567	126	i	i	PRON
ejpam-86	567	127	;	;	PUNCT
ejpam-86	567	128	else	else	ADV
ejpam-86	567	129	select	select	ADJ
ejpam-86	567	130	ng/2	ng/2	NOUN
ejpam-86	567	131	pairs	pair	NOUN
ejpam-86	567	132	of	of	ADP
ejpam-86	567	133	parent	parent	NOUN
ejpam-86	567	134	chromosomes	chromosome	NOUN
ejpam-86	567	135	from	from	ADP
ejpam-86	567	136	generation	generation	NOUN
ejpam-86	567	137	i	i	PRON
ejpam-86	567	138	;	;	PUNCT
ejpam-86	567	139	end	end	NOUN
ejpam-86	567	140	cross	cross	NOUN
ejpam-86	567	141	over	over	ADP
ejpam-86	567	142	each	each	DET
ejpam-86	567	143	pair	pair	NOUN
ejpam-86	567	144	of	of	ADP
ejpam-86	567	145	parent	parent	NOUN
ejpam-86	567	146	chromosomes	chromosome	NOUN
ejpam-86	567	147	with	with	ADP
ejpam-86	567	148	probability	probability	NOUN
ejpam-86	567	149	pc	pc	NOUN
ejpam-86	567	150	;	;	PUNCT
ejpam-86	567	151	if	if	SCONJ
ejpam-86	567	152	ct	ct	NUM
ejpam-86	567	153	=	=	NOUN
ejpam-86	567	154	1	1	NUM
ejpam-86	567	155	then	then	ADV
ejpam-86	567	156	do	do	VERB
ejpam-86	567	157	single	single	ADJ
ejpam-86	567	158	point	point	NOUN
ejpam-86	567	159	crossover	crossover	NOUN
ejpam-86	567	160	;	;	PUNCT
ejpam-86	567	161	else	else	ADV
ejpam-86	567	162	if	if	SCONJ
ejpam-86	567	163	ct	ct	PROPN
ejpam-86	567	164	=	=	NOUN
ejpam-86	567	165	2	2	NUM
ejpam-86	567	166	then	then	ADV
ejpam-86	567	167	do	do	VERB
ejpam-86	567	168	two	two	NUM
ejpam-86	567	169	point	point	NOUN
ejpam-86	567	170	crossover	crossover	NOUN
ejpam-86	567	171	;	;	PUNCT
ejpam-86	567	172	else	else	ADV
ejpam-86	567	173	do	do	VERB
ejpam-86	567	174	uniform	uniform	NOUN
ejpam-86	567	175	crossover	crossover	NOUN
ejpam-86	567	176	;	;	PUNCT
ejpam-86	567	177	end	end	VERB
ejpam-86	567	178	mutate	mutate	ADJ
ejpam-86	567	179	new	new	ADJ
ejpam-86	567	180	offspring	offspring	NOUN
ejpam-86	567	181	at	at	ADP
ejpam-86	567	182	each	each	DET
ejpam-86	567	183	lotus	lotus	NOUN
ejpam-86	567	184	with	with	ADP
ejpam-86	567	185	probability	probability	NOUN
ejpam-86	567	186	pm	pm	NOUN
ejpam-86	567	187	;	;	PUNCT
ejpam-86	567	188	engineer	engineer	VERB
ejpam-86	567	189	the	the	DET
ejpam-86	567	190	new	new	ADJ
ejpam-86	567	191	offspring	offspring	NOUN
ejpam-86	567	192	with	with	ADP
ejpam-86	567	193	probability	probability	NOUN
ejpam-86	567	194	pe	pe	ADP
ejpam-86	567	195	;	;	PUNCT
ejpam-86	567	196	if	if	SCONJ
ejpam-86	567	197	elitism	elitism	NOUN
ejpam-86	567	198	is	be	AUX
ejpam-86	567	199	true	true	ADJ
ejpam-86	567	200	then	then	ADV
ejpam-86	567	201	add	add	VERB
ejpam-86	567	202	the	the	DET
ejpam-86	567	203	best	good	ADJ
ejpam-86	567	204	chromosome	chromosome	NOUN
ejpam-86	567	205	in	in	ADP
ejpam-86	567	206	generation	generation	NOUN
ejpam-86	567	207	i	i	PRON
ejpam-86	567	208	to	to	ADP
ejpam-86	567	209	new	new	ADJ
ejpam-86	567	210	offspring	offspring	NOUN
ejpam-86	567	211	;	;	PUNCT
ejpam-86	567	212	add	add	VERB
ejpam-86	567	213	a	a	DET
ejpam-86	567	214	randomly	randomly	ADV
ejpam-86	567	215	selected	select	VERB
ejpam-86	567	216	chromosome	chromosome	NOUN
ejpam-86	567	217	in	in	ADP
ejpam-86	567	218	generation	generation	NOUN
ejpam-86	567	219	i	i	PRON
ejpam-86	567	220	to	to	ADP
ejpam-86	567	221	new	new	ADJ
ejpam-86	567	222	offspring	offspring	NOUN
ejpam-86	567	223	;	;	PUNCT
ejpam-86	567	224	end	end	VERB
ejpam-86	567	225	i	i	NOUN
ejpam-86	567	226	=	=	PUNCT
ejpam-86	567	227	i+	i+	PROPN
ejpam-86	567	228	1	1	NUM
ejpam-86	567	229	;	;	PUNCT
ejpam-86	567	230	decode	decode	VERB
ejpam-86	567	231	the	the	DET
ejpam-86	567	232	new	new	ADJ
ejpam-86	567	233	offspring	offspring	NOUN
ejpam-86	567	234	to	to	ADP
ejpam-86	567	235	solutions	solution	NOUN
ejpam-86	567	236	;	;	PUNCT
ejpam-86	567	237	replace	replace	VERB
ejpam-86	567	238	the	the	DET
ejpam-86	567	239	chromosomes	chromosome	NOUN
ejpam-86	567	240	in	in	ADP
ejpam-86	567	241	generation	generation	NOUN
ejpam-86	567	242	i	i	PRON
ejpam-86	567	243	with	with	ADP
ejpam-86	567	244	new	new	ADJ
ejpam-86	567	245	offspring	offspring	NOUN
ejpam-86	567	246	;	;	PUNCT
ejpam-86	567	247	replace	replace	VERB
ejpam-86	567	248	the	the	DET
ejpam-86	567	249	solutions	solution	NOUN
ejpam-86	567	250	in	in	ADP
ejpam-86	567	251	generation	generation	NOUN
ejpam-86	567	252	i	i	PRON
ejpam-86	567	253	with	with	ADP
ejpam-86	567	254	new	new	ADJ
ejpam-86	567	255	solutions	solution	NOUN
ejpam-86	567	256	;	;	PUNCT
ejpam-86	567	257	if	if	SCONJ
ejpam-86	567	258	any	any	DET
ejpam-86	567	259	special	special	ADJ
ejpam-86	567	260	final	final	ADJ
ejpam-86	567	261	condition	condition	NOUN
ejpam-86	567	262	is	be	AUX
ejpam-86	567	263	satisfied	satisfied	ADJ
ejpam-86	567	264	then	then	ADV
ejpam-86	567	265	exit	exit	NOUN
ejpam-86	567	266	;	;	PUNCT
ejpam-86	567	267	end	end	NOUN
ejpam-86	567	268	m.	m.	NOUN
ejpam-86	567	269	liu	liu	PROPN
ejpam-86	567	270	and	and	CCONJ
ejpam-86	567	271	h.	h.	PROPN
ejpam-86	567	272	bozdogan	bozdogan	PROPN
ejpam-86	567	273	/	/	SYM
ejpam-86	567	274	eur	eur	PROPN
ejpam-86	567	275	.	.	PUNCT
ejpam-86	568	1	j.	j.	PROPN
ejpam-86	568	2	pure	pure	PROPN
ejpam-86	568	3	appl	appl	PROPN
ejpam-86	568	4	.	.	PROPN
ejpam-86	568	5	math	math	PROPN
ejpam-86	568	6	,	,	PUNCT
ejpam-86	568	7	1	1	NUM
ejpam-86	568	8	(	(	PUNCT
ejpam-86	568	9	2008	2008	NUM
ejpam-86	568	10	)	)	PUNCT
ejpam-86	568	11	,	,	PUNCT
ejpam-86	568	12	(	(	PUNCT
ejpam-86	568	13	4	4	NUM
ejpam-86	568	14	-	-	SYM
ejpam-86	568	15	37	37	NUM
ejpam-86	568	16	)	)	PUNCT
ejpam-86	568	17	30	30	NUM
ejpam-86	568	18	end	end	NOUN
ejpam-86	568	19	generating	generate	VERB
ejpam-86	568	20	univariate	univariate	ADJ
ejpam-86	568	21	pe	pe	ADP
ejpam-86	568	22	pseudo	pseudo	NOUN
ejpam-86	568	23	-	-	ADJ
ejpam-86	568	24	random	random	ADJ
ejpam-86	568	25	numbers	number	NOUN
ejpam-86	568	26	let	let	VERB
ejpam-86	568	27	y	y	NOUN
ejpam-86	568	28	=	=	NOUN
ejpam-86	568	29	1	1	NUM
ejpam-86	568	30	2	2	NUM
ejpam-86	568	31	|x|2β	|x|2β	NOUN
ejpam-86	568	32	,	,	PUNCT
ejpam-86	568	33	where	where	SCONJ
ejpam-86	568	34	x	x	PRON
ejpam-86	568	35	has	have	VERB
ejpam-86	568	36	a	a	DET
ejpam-86	568	37	standard	standard	ADJ
ejpam-86	568	38	pe	pe	ADJ
ejpam-86	568	39	distribution	distribution	NOUN
ejpam-86	568	40	with	with	ADP
ejpam-86	568	41	shape	shape	NOUN
ejpam-86	568	42	parameter	parameter	NOUN
ejpam-86	568	43	β	β	PROPN
ejpam-86	568	44	.	.	PUNCT
ejpam-86	569	1	then	then	ADV
ejpam-86	569	2	p	p	X
ejpam-86	569	3	(	(	PUNCT
ejpam-86	569	4	y	y	PROPN
ejpam-86	569	5	<	<	X
ejpam-86	569	6	y	y	PROPN
ejpam-86	569	7	)	)	PUNCT
ejpam-86	569	8	=	=	SYM
ejpam-86	570	1	p	p	NOUN
ejpam-86	570	2	(	(	PUNCT
ejpam-86	570	3	|x|2β	|x|2β	VERB
ejpam-86	570	4	<	<	X
ejpam-86	570	5	2y	2y	NUM
ejpam-86	570	6	)	)	PUNCT
ejpam-86	570	7	=	=	SYM
ejpam-86	570	8	p	p	X
ejpam-86	570	9	(	(	PUNCT
ejpam-86	570	10	−(2y)1/2β	−(2y)1/2β	X
ejpam-86	570	11	<	<	X
ejpam-86	570	12	x	x	X
ejpam-86	570	13	<	<	X
ejpam-86	570	14	(	(	PUNCT
ejpam-86	570	15	2y)1/2β	2y)1/2β	NUM
ejpam-86	570	16	)	)	PUNCT
ejpam-86	570	17	=	=	SYM
ejpam-86	570	18	∫	∫	PROPN
ejpam-86	570	19	(	(	PUNCT
ejpam-86	570	20	2y)1/2β	2y)1/2β	NUM
ejpam-86	570	21	−(2y)1/2β	−(2y)1/2β	NOUN
ejpam-86	570	22	1	1	NUM
ejpam-86	570	23	γ	γ	X
ejpam-86	570	24	“	"	PUNCT
ejpam-86	570	25	1	1	NUM
ejpam-86	570	26	+	+	NUM
ejpam-86	570	27	1	1	NUM
ejpam-86	570	28	2β	2β	NOUN
ejpam-86	570	29	”	"	PUNCT
ejpam-86	570	30	2	2	NUM
ejpam-86	570	31	1	1	NUM
ejpam-86	570	32	+	+	SYM
ejpam-86	570	33	1	1	NUM
ejpam-86	570	34	2β	2β	NOUN
ejpam-86	570	35	exp	exp	NOUN
ejpam-86	570	36	(	(	PUNCT
ejpam-86	570	37	−1	−1	NOUN
ejpam-86	570	38	2	2	NUM
ejpam-86	570	39	|x|2β	|x|2β	NOUN
ejpam-86	570	40	)	)	PUNCT
ejpam-86	570	41	dx	dx	PROPN
ejpam-86	571	1	=	=	SYM
ejpam-86	571	2	∫	∫	PROPN
ejpam-86	571	3	(	(	PUNCT
ejpam-86	571	4	2y)1/2β	2y)1/2β	NUM
ejpam-86	571	5	0	0	NUM
ejpam-86	571	6	1	1	NUM
ejpam-86	571	7	γ	γ	NOUN
ejpam-86	571	8	“	"	PUNCT
ejpam-86	571	9	1	1	NUM
ejpam-86	571	10	+	+	NUM
ejpam-86	571	11	1	1	NUM
ejpam-86	571	12	2β	2β	NOUN
ejpam-86	571	13	”	"	PUNCT
ejpam-86	571	14	2	2	NUM
ejpam-86	571	15	1	1	NUM
ejpam-86	571	16	2β	2β	NOUN
ejpam-86	571	17	exp	exp	NOUN
ejpam-86	571	18	(	(	PUNCT
ejpam-86	571	19	−1	−1	NOUN
ejpam-86	571	20	2x	2x	NUM
ejpam-86	571	21	2β	2β	NOUN
ejpam-86	571	22	)	)	PUNCT
ejpam-86	571	23	dx	dx	PROPN
ejpam-86	572	1	=	=	SYM
ejpam-86	572	2	∫	∫	PROPN
ejpam-86	572	3	y	y	PROPN
ejpam-86	572	4	0	0	NUM
ejpam-86	572	5	t1/2β−1	t1/2β−1	PROPN
ejpam-86	572	6	γ	γ	PROPN
ejpam-86	572	7	“	"	PUNCT
ejpam-86	572	8	1	1	NUM
ejpam-86	572	9	2β	2β	NOUN
ejpam-86	572	10	”	"	PUNCT
ejpam-86	572	11	exp	exp	NOUN
ejpam-86	572	12	(	(	PUNCT
ejpam-86	572	13	−t	−t	NOUN
ejpam-86	572	14	)	)	PUNCT
ejpam-86	572	15	dt	dt	PROPN
ejpam-86	572	16	.	.	PUNCT
ejpam-86	573	1	(	(	PUNCT
ejpam-86	573	2	8.1	8.1	NUM
ejpam-86	573	3	)	)	PUNCT
ejpam-86	573	4	so	so	CCONJ
ejpam-86	573	5	y	y	PROPN
ejpam-86	573	6	has	have	VERB
ejpam-86	573	7	a	a	DET
ejpam-86	573	8	distribution	distribution	NOUN
ejpam-86	573	9	with	with	ADP
ejpam-86	573	10	density	density	NOUN
ejpam-86	573	11	f(y	f(y	NOUN
ejpam-86	573	12	)	)	PUNCT
ejpam-86	574	1	=	=	SYM
ejpam-86	574	2	y(1/2β)−1e−y	y(1/2β)−1e−y	PROPN
ejpam-86	574	3	γ(1/2β	γ(1/2β	NUM
ejpam-86	574	4	)	)	PUNCT
ejpam-86	574	5	,	,	PUNCT
ejpam-86	574	6	(	(	PUNCT
ejpam-86	574	7	8.2	8.2	NUM
ejpam-86	574	8	)	)	PUNCT
ejpam-86	574	9	i.e.	i.e.	X
ejpam-86	574	10	,	,	PUNCT
ejpam-86	574	11	y	y	PROPN
ejpam-86	574	12	∼	∼	NOUN
ejpam-86	574	13	gamma(1/2β	gamma(1/2β	NOUN
ejpam-86	574	14	)	)	PUNCT
ejpam-86	574	15	.	.	PUNCT
ejpam-86	575	1	one	one	NUM
ejpam-86	575	2	method	method	NOUN
ejpam-86	575	3	generating	generate	VERB
ejpam-86	575	4	pseudo	pseudo	NOUN
ejpam-86	575	5	-	-	ADJ
ejpam-86	575	6	random	random	ADJ
ejpam-86	575	7	pe	pe	NOUN
ejpam-86	575	8	variables	variable	NOUN
ejpam-86	575	9	is	be	AUX
ejpam-86	575	10	given	give	VERB
ejpam-86	575	11	as	as	SCONJ
ejpam-86	575	12	follows	follow	VERB
ejpam-86	575	13	:	:	PUNCT
ejpam-86	575	14	1	1	X
ejpam-86	575	15	.	.	X
ejpam-86	575	16	generate	generate	VERB
ejpam-86	575	17	ordinates	ordinate	NOUN
ejpam-86	575	18	from	from	ADP
ejpam-86	575	19	a	a	DET
ejpam-86	575	20	gamma	gamma	NOUN
ejpam-86	575	21	distribution	distribution	NOUN
ejpam-86	575	22	with	with	ADP
ejpam-86	575	23	density	density	NOUN
ejpam-86	575	24	(	(	PUNCT
ejpam-86	575	25	8.2	8.2	NUM
ejpam-86	575	26	)	)	PUNCT
ejpam-86	575	27	;	;	PUNCT
ejpam-86	575	28	2	2	X
ejpam-86	575	29	.	.	X
ejpam-86	575	30	generate	generate	VERB
ejpam-86	575	31	b	b	NOUN
ejpam-86	575	32	from	from	ADP
ejpam-86	575	33	a	a	DET
ejpam-86	575	34	bernoulli	bernoulli	NOUN
ejpam-86	575	35	distribution	distribution	NOUN
ejpam-86	575	36	with	with	ADP
ejpam-86	575	37	p	p	NOUN
ejpam-86	575	38	=	=	NOUN
ejpam-86	575	39	1/2	1/2	NUM
ejpam-86	575	40	;	;	PUNCT
ejpam-86	575	41	3	3	X
ejpam-86	575	42	.	.	X
ejpam-86	576	1	if	if	SCONJ
ejpam-86	576	2	b	b	PROPN
ejpam-86	576	3	=	=	SYM
ejpam-86	576	4	0	0	PROPN
ejpam-86	576	5	,	,	PUNCT
ejpam-86	576	6	then	then	ADV
ejpam-86	576	7	generate	generate	VERB
ejpam-86	576	8	x	x	X
ejpam-86	576	9	=	=	SYM
ejpam-86	576	10	(	(	PUNCT
ejpam-86	576	11	2y	2y	PROPN
ejpam-86	576	12	)	)	PUNCT
ejpam-86	576	13	1/2β	1/2β	NUM
ejpam-86	576	14	,	,	PUNCT
ejpam-86	576	15	otherwise	otherwise	ADV
ejpam-86	576	16	,	,	PUNCT
ejpam-86	576	17	generate	generate	VERB
ejpam-86	576	18	x	x	X
ejpam-86	576	19	=	=	SYM
ejpam-86	576	20	−(2y	−(2y	PROPN
ejpam-86	576	21	)	)	PUNCT
ejpam-86	576	22	1/2β	1/2β	NUM
ejpam-86	576	23	;	;	PUNCT
ejpam-86	576	24	4	4	X
ejpam-86	576	25	.	.	X
ejpam-86	576	26	generate	generate	VERB
ejpam-86	576	27	z	z	NOUN
ejpam-86	576	28	=	=	SYM
ejpam-86	576	29	σx	σx	PROPN
ejpam-86	576	30	+	+	PROPN
ejpam-86	576	31	µ.	µ.	PROPN
ejpam-86	576	32	histograms	histogram	NOUN
ejpam-86	576	33	of	of	ADP
ejpam-86	576	34	random	random	ADJ
ejpam-86	576	35	samples	sample	NOUN
ejpam-86	576	36	of	of	ADP
ejpam-86	576	37	size	size	NOUN
ejpam-86	576	38	1000	1000	NUM
ejpam-86	576	39	generated	generate	VERB
ejpam-86	576	40	from	from	ADP
ejpam-86	576	41	above	above	ADP
ejpam-86	576	42	algorithm	algorithm	NOUN
ejpam-86	576	43	are	be	AUX
ejpam-86	576	44	given	give	VERB
ejpam-86	576	45	in	in	ADP
ejpam-86	576	46	figure	figure	NOUN
ejpam-86	576	47	9	9	NUM
ejpam-86	576	48	.	.	PUNCT
ejpam-86	576	49	matrix	matrix	NOUN
ejpam-86	576	50	calculus	calculus	NOUN
ejpam-86	576	51	for	for	ADP
ejpam-86	576	52	type	type	NOUN
ejpam-86	576	53	i	i	PROPN
ejpam-86	576	54	mvper	mvper	PROPN
ejpam-86	576	55	model	model	PROPN
ejpam-86	576	56	∂ε(i	∂ε(i	PROPN
ejpam-86	576	57	)	)	PUNCT
ejpam-86	576	58	∂b	∂b	PROPN
ejpam-86	576	59	=	=	SYM
ejpam-86	576	60	∂(y(i)−b′x(i	∂(y(i)−b′x(i	PROPN
ejpam-86	576	61	)	)	PUNCT
ejpam-86	576	62	)	)	PUNCT
ejpam-86	577	1	∂b	∂b	PROPN
ejpam-86	577	2	=	=	SYM
ejpam-86	577	3	∂y(i	∂y(i	PROPN
ejpam-86	577	4	)	)	PUNCT
ejpam-86	577	5	∂b	∂b	PROPN
ejpam-86	577	6	−	−	PROPN
ejpam-86	577	7	∂b′x(i	∂b′x(i	NOUN
ejpam-86	577	8	)	)	PUNCT
ejpam-86	577	9	∂b	∂b	PROPN
ejpam-86	577	10	=	=	SYM
ejpam-86	577	11	−∂b′x(i	−∂b′x(i	PROPN
ejpam-86	577	12	)	)	PUNCT
ejpam-86	577	13	∂b	∂b	PROPN
ejpam-86	577	14	=	=	SYM
ejpam-86	577	15	−i(q	−i(q	PROPN
ejpam-86	577	16	,	,	PUNCT
ejpam-86	577	17	p)(x(i	p)(x(i	ADJ
ejpam-86	577	18	)	)	PUNCT
ejpam-86	577	19	⊗	⊗	NUM
ejpam-86	577	20	ip	ip	NOUN
ejpam-86	577	21	)	)	PUNCT
ejpam-86	577	22	(	(	PUNCT
ejpam-86	577	23	8.3	8.3	NUM
ejpam-86	577	24	)	)	PUNCT
ejpam-86	577	25	where	where	SCONJ
ejpam-86	577	26	i(q	i(q	NOUN
ejpam-86	577	27	,	,	PUNCT
ejpam-86	577	28	p	p	NOUN
ejpam-86	577	29	)	)	PUNCT
ejpam-86	577	30	is	be	AUX
ejpam-86	577	31	the	the	DET
ejpam-86	577	32	permuted	permuted	ADJ
ejpam-86	577	33	identity	identity	NOUN
ejpam-86	577	34	macrae	macrae	NOUN
ejpam-86	577	35	(	(	PUNCT
ejpam-86	577	36	[	[	X
ejpam-86	577	37	24	24	NUM
ejpam-86	577	38	]	]	PUNCT
ejpam-86	577	39	)	)	PUNCT
ejpam-86	577	40	or	or	CCONJ
ejpam-86	577	41	commutation	commutation	NOUN
ejpam-86	577	42	matrix	matrix	NOUN
ejpam-86	577	43	magnus	magnus	PROPN
ejpam-86	577	44	and	and	CCONJ
ejpam-86	577	45	neudecker	neudecker	NOUN
ejpam-86	577	46	(	(	PUNCT
ejpam-86	577	47	[	[	X
ejpam-86	577	48	25	25	NUM
ejpam-86	577	49	]	]	NUM
ejpam-86	577	50	)	)	PUNCT
ejpam-86	577	51	.	.	PUNCT
ejpam-86	578	1	m.	m.	PROPN
ejpam-86	578	2	liu	liu	PROPN
ejpam-86	578	3	and	and	CCONJ
ejpam-86	578	4	h.	h.	PROPN
ejpam-86	578	5	bozdogan	bozdogan	PROPN
ejpam-86	578	6	/	/	SYM
ejpam-86	578	7	eur	eur	PROPN
ejpam-86	578	8	.	.	PUNCT
ejpam-86	579	1	j.	j.	PROPN
ejpam-86	579	2	pure	pure	PROPN
ejpam-86	579	3	appl	appl	PROPN
ejpam-86	579	4	.	.	PROPN
ejpam-86	579	5	math	math	PROPN
ejpam-86	579	6	,	,	PUNCT
ejpam-86	579	7	1	1	NUM
ejpam-86	579	8	(	(	PUNCT
ejpam-86	579	9	2008	2008	NUM
ejpam-86	579	10	)	)	PUNCT
ejpam-86	579	11	,	,	PUNCT
ejpam-86	579	12	(	(	PUNCT
ejpam-86	579	13	4	4	NUM
ejpam-86	579	14	-	-	SYM
ejpam-86	579	15	37	37	NUM
ejpam-86	579	16	)	)	PUNCT
ejpam-86	579	17	31	31	NUM
ejpam-86	580	1	−2000	−2000	PROPN
ejpam-86	580	2	−1000	−1000	NOUN
ejpam-86	580	3	0	0	NUM
ejpam-86	580	4	1000	1000	NUM
ejpam-86	580	5	2000	2000	NUM
ejpam-86	580	6	0	0	NUM
ejpam-86	580	7	200	200	NUM
ejpam-86	580	8	400	400	NUM
ejpam-86	580	9	600	600	NUM
ejpam-86	580	10	mu=0,sigma=1,beta=0.2,n=1000	mu=0,sigma=1,beta=0.2,n=1000	NUM
ejpam-86	580	11	−20	−20	PROPN
ejpam-86	580	12	−10	−10	X
ejpam-86	580	13	0	0	NUM
ejpam-86	580	14	10	10	NUM
ejpam-86	580	15	20	20	NUM
ejpam-86	580	16	0	0	NUM
ejpam-86	580	17	200	200	NUM
ejpam-86	580	18	400	400	NUM
ejpam-86	580	19	600	600	NUM
ejpam-86	580	20	mu=0,sigma=1,beta=0.5,n=1000	mu=0,sigma=1,beta=0.5,n=1000	NUM
ejpam-86	580	21	−4	−4	X
ejpam-86	581	1	−2	−2	NOUN
ejpam-86	581	2	0	0	NUM
ejpam-86	581	3	2	2	NUM
ejpam-86	581	4	4	4	NUM
ejpam-86	581	5	0	0	NUM
ejpam-86	581	6	100	100	NUM
ejpam-86	581	7	200	200	NUM
ejpam-86	581	8	300	300	NUM
ejpam-86	581	9	mu=0,sigma=1,beta=1,n=1000	mu=0,sigma=1,beta=1,n=1000	NOUN
ejpam-86	581	10	−2	−2	NOUN
ejpam-86	581	11	−1	−1	NOUN
ejpam-86	581	12	0	0	NUM
ejpam-86	581	13	1	1	NUM
ejpam-86	581	14	2	2	NUM
ejpam-86	581	15	0	0	NUM
ejpam-86	581	16	50	50	NUM
ejpam-86	581	17	100	100	NUM
ejpam-86	581	18	150	150	NUM
ejpam-86	581	19	200	200	NUM
ejpam-86	581	20	mu=0,sigma=1,beta=2,n=1000	mu=0,sigma=1,beta=2,n=1000	NOUN
ejpam-86	581	21	−2	−2	NOUN
ejpam-86	581	22	−1	−1	NOUN
ejpam-86	581	23	0	0	NUM
ejpam-86	581	24	1	1	NUM
ejpam-86	581	25	2	2	NUM
ejpam-86	581	26	0	0	NUM
ejpam-86	581	27	50	50	NUM
ejpam-86	581	28	100	100	NUM
ejpam-86	581	29	150	150	NUM
ejpam-86	581	30	mu=0,sigma=1,beta=10,n=1000	mu=0,sigma=1,beta=10,n=1000	NOUN
ejpam-86	581	31	figure	figure	NOUN
ejpam-86	581	32	9	9	NUM
ejpam-86	581	33	:	:	PUNCT
ejpam-86	581	34	histograms	histogram	NOUN
ejpam-86	581	35	of	of	ADP
ejpam-86	581	36	pe	pe	ADP
ejpam-86	581	37	pseudo	pseudo	NOUN
ejpam-86	581	38	-	-	ADJ
ejpam-86	581	39	random	random	ADJ
ejpam-86	581	40	samples	sample	NOUN
ejpam-86	581	41	.	.	PUNCT
ejpam-86	582	1	∂ε(i	∂ε(i	NOUN
ejpam-86	582	2	)	)	PUNCT
ejpam-86	582	3	∂b′	∂b′	NOUN
ejpam-86	582	4	=	=	SYM
ejpam-86	582	5	∂(y(i)−b′x(i	∂(y(i)−b′x(i	PROPN
ejpam-86	582	6	)	)	PUNCT
ejpam-86	582	7	)	)	PUNCT
ejpam-86	583	1	∂b′	∂b′	NOUN
ejpam-86	583	2	=	=	SYM
ejpam-86	583	3	∂y(i	∂y(i	PROPN
ejpam-86	583	4	)	)	PUNCT
ejpam-86	583	5	∂b′	∂b′	NOUN
ejpam-86	583	6	−	−	ADP
ejpam-86	583	7	∂b′x(i	∂b′x(i	NOUN
ejpam-86	583	8	)	)	PUNCT
ejpam-86	583	9	∂b′	∂b′	NOUN
ejpam-86	583	10	=	=	SYM
ejpam-86	583	11	−∂b′x(i	−∂b′x(i	PROPN
ejpam-86	583	12	)	)	PUNCT
ejpam-86	583	13	∂b′	∂b′	NOUN
ejpam-86	583	14	=	=	SYM
ejpam-86	583	15	−v	−v	NOUN
ejpam-86	583	16	ec(ip)v	ec(ip)v	NOUN
ejpam-86	583	17	ec′(x(i	ec′(x(i	PROPN
ejpam-86	583	18	)	)	PUNCT
ejpam-86	583	19	)	)	PUNCT
ejpam-86	583	20	(	(	PUNCT
ejpam-86	583	21	8.4	8.4	X
ejpam-86	583	22	)	)	PUNCT
ejpam-86	583	23	∂ti	∂ti	PROPN
ejpam-86	583	24	∂b	∂b	PROPN
ejpam-86	583	25	=	=	PUNCT
ejpam-86	583	26	∂ε′−1	∂ε′−1	NOUN
ejpam-86	583	27	(	(	PUNCT
ejpam-86	583	28	i	i	NOUN
ejpam-86	583	29	)	)	PUNCT
ejpam-86	583	30	ε(i	ε(i	PROPN
ejpam-86	583	31	)	)	PUNCT
ejpam-86	583	32	∂b	∂b	NOUN
ejpam-86	583	33	=	=	SYM
ejpam-86	583	34	∂ε′−1	∂ε′−1	NOUN
ejpam-86	583	35	(	(	PUNCT
ejpam-86	583	36	i	i	NOUN
ejpam-86	583	37	)	)	PUNCT
ejpam-86	583	38	ε(i	ε(i	PROPN
ejpam-86	583	39	)	)	PUNCT
ejpam-86	583	40	∂ε(i	∂ε(i	PROPN
ejpam-86	583	41	)	)	PUNCT
ejpam-86	583	42	∗	∗	PROPN
ejpam-86	583	43	∂ε(i	∂ε(i	PROPN
ejpam-86	583	44	)	)	PUNCT
ejpam-86	583	45	∂b	∂b	PROPN
ejpam-86	583	46	=	=	SYM
ejpam-86	583	47	2σ−1ε(i	2σ−1ε(i	X
ejpam-86	583	48	)	)	PUNCT
ejpam-86	583	49	∗	∗	NOUN
ejpam-86	583	50	(	(	PUNCT
ejpam-86	583	51	−i(q	−i(q	NOUN
ejpam-86	583	52	,	,	PUNCT
ejpam-86	583	53	p)(x(i	p)(x(i	ADJ
ejpam-86	583	54	)	)	PUNCT
ejpam-86	583	55	⊗	⊗	NUM
ejpam-86	583	56	ip	ip	NOUN
ejpam-86	583	57	)	)	PUNCT
ejpam-86	583	58	)	)	PUNCT
ejpam-86	584	1	=	=	SYM
ejpam-86	584	2	−2σ−1ε(i	−2σ−1ε(i	ADJ
ejpam-86	584	3	)	)	PUNCT
ejpam-86	584	4	∗	∗	NOUN
ejpam-86	584	5	(	(	PUNCT
ejpam-86	584	6	iqpi(q	iqpi(q	PROPN
ejpam-86	584	7	,	,	PUNCT
ejpam-86	584	8	p)(x(i	p)(x(i	ADJ
ejpam-86	584	9	)	)	PUNCT
ejpam-86	584	10	⊗	⊗	NUM
ejpam-86	584	11	ip	ip	NOUN
ejpam-86	584	12	)	)	PUNCT
ejpam-86	584	13	)	)	PUNCT
ejpam-86	585	1	=	=	SYM
ejpam-86	585	2	−2σ−1ε(i	−2σ−1ε(i	ADJ
ejpam-86	585	3	)	)	PUNCT
ejpam-86	585	4	∗	∗	NOUN
ejpam-86	585	5	(	(	PUNCT
ejpam-86	585	6	(	(	PUNCT
ejpam-86	585	7	ip	ip	ADJ
ejpam-86	585	8	⊗	⊗	PROPN
ejpam-86	585	9	iq)i(q	iq)i(q	PROPN
ejpam-86	585	10	,	,	PUNCT
ejpam-86	585	11	p)(x(i	p)(x(i	ADJ
ejpam-86	585	12	)	)	PUNCT
ejpam-86	585	13	⊗	⊗	NUM
ejpam-86	585	14	ip	ip	NOUN
ejpam-86	585	15	)	)	PUNCT
ejpam-86	585	16	)	)	PUNCT
ejpam-86	586	1	=	=	PRON
ejpam-86	586	2	−2x(i)(σ−1si)′ip	−2x(i)(σ−1si)′ip	PROPN
ejpam-86	587	1	=	=	PUNCT
ejpam-86	587	2	−2x(i)ε	−2x(i)ε	PROPN
ejpam-86	587	3	′−1	′−1	X
ejpam-86	587	4	(	(	PUNCT
ejpam-86	587	5	i	i	NOUN
ejpam-86	587	6	)	)	PUNCT
ejpam-86	587	7	(	(	PUNCT
ejpam-86	587	8	8.5	8.5	NUM
ejpam-86	587	9	)	)	PUNCT
ejpam-86	587	10	where	where	SCONJ
ejpam-86	587	11	∗	∗	NOUN
ejpam-86	587	12	is	be	AUX
ejpam-86	587	13	the	the	DET
ejpam-86	587	14	star	star	NOUN
ejpam-86	587	15	product	product	NOUN
ejpam-86	587	16	(	(	PUNCT
ejpam-86	587	17	[	[	X
ejpam-86	587	18	24	24	NUM
ejpam-86	587	19	]	]	NUM
ejpam-86	587	20	)	)	PUNCT
ejpam-86	587	21	.	.	PUNCT
ejpam-86	588	1	∂ti	∂ti	PROPN
ejpam-86	588	2	∂b′	∂b′	PROPN
ejpam-86	588	3	=	=	SYM
ejpam-86	588	4	∂ε′−1	∂ε′−1	NOUN
ejpam-86	588	5	(	(	PUNCT
ejpam-86	588	6	i	i	NOUN
ejpam-86	588	7	)	)	PUNCT
ejpam-86	588	8	ε(i	ε(i	PROPN
ejpam-86	588	9	)	)	PUNCT
ejpam-86	588	10	∂b′	∂b′	NOUN
ejpam-86	588	11	=	=	SYM
ejpam-86	588	12	∂ε′−1	∂ε′−1	NOUN
ejpam-86	588	13	(	(	PUNCT
ejpam-86	588	14	i	i	NOUN
ejpam-86	588	15	)	)	PUNCT
ejpam-86	588	16	ε(i	ε(i	PROPN
ejpam-86	588	17	)	)	PUNCT
ejpam-86	588	18	∂ε(i	∂ε(i	PROPN
ejpam-86	588	19	)	)	PUNCT
ejpam-86	588	20	∗	∗	PROPN
ejpam-86	588	21	∂ε(i	∂ε(i	PROPN
ejpam-86	588	22	)	)	PUNCT
ejpam-86	588	23	∂b′	∂b′	NOUN
ejpam-86	588	24	=	=	SYM
ejpam-86	588	25	2σ−1ε(i	2σ−1ε(i	NUM
ejpam-86	588	26	)	)	PUNCT
ejpam-86	588	27	∗	∗	NOUN
ejpam-86	588	28	(	(	PUNCT
ejpam-86	588	29	−v	−v	NOUN
ejpam-86	588	30	ec(ip)v	ec(ip)v	NOUN
ejpam-86	588	31	ec′(x(i	ec′(x(i	PROPN
ejpam-86	588	32	)	)	PUNCT
ejpam-86	588	33	)	)	PUNCT
ejpam-86	588	34	)	)	PUNCT
ejpam-86	589	1	=	=	SYM
ejpam-86	589	2	−2σ−1ε(i)x′(i	−2σ−1ε(i)x′(i	X
ejpam-86	589	3	)	)	PUNCT
ejpam-86	589	4	(	(	PUNCT
ejpam-86	589	5	8.6	8.6	NUM
ejpam-86	589	6	)	)	PUNCT
ejpam-86	589	7	m.	m.	NOUN
ejpam-86	589	8	liu	liu	PROPN
ejpam-86	589	9	and	and	CCONJ
ejpam-86	589	10	h.	h.	PROPN
ejpam-86	589	11	bozdogan	bozdogan	PROPN
ejpam-86	589	12	/	/	SYM
ejpam-86	589	13	eur	eur	PROPN
ejpam-86	589	14	.	.	PUNCT
ejpam-86	590	1	j.	j.	PROPN
ejpam-86	590	2	pure	pure	PROPN
ejpam-86	590	3	appl	appl	PROPN
ejpam-86	590	4	.	.	PROPN
ejpam-86	590	5	math	math	PROPN
ejpam-86	590	6	,	,	PUNCT
ejpam-86	590	7	1	1	NUM
ejpam-86	590	8	(	(	PUNCT
ejpam-86	590	9	2008	2008	NUM
ejpam-86	590	10	)	)	PUNCT
ejpam-86	590	11	,	,	PUNCT
ejpam-86	590	12	(	(	PUNCT
ejpam-86	590	13	4	4	NUM
ejpam-86	590	14	-	-	SYM
ejpam-86	590	15	37	37	NUM
ejpam-86	590	16	)	)	PUNCT
ejpam-86	590	17	32	32	NUM
ejpam-86	590	18	∂l(θ	∂l(θ	PROPN
ejpam-86	590	19	)	)	PUNCT
ejpam-86	590	20	∂b	∂b	NOUN
ejpam-86	590	21	=	=	SYM
ejpam-86	590	22	−1	−1	NOUN
ejpam-86	590	23	2	2	NUM
ejpam-86	590	24	n∑	n∑	NOUN
ejpam-86	590	25	i=1	i=1	PROPN
ejpam-86	590	26	∂tβi	∂tβi	NOUN
ejpam-86	590	27	∂b	∂b	NOUN
ejpam-86	590	28	=	=	PUNCT
ejpam-86	590	29	−1	−1	NOUN
ejpam-86	590	30	2	2	NUM
ejpam-86	590	31	n∑	n∑	NOUN
ejpam-86	590	32	i=1	i=1	PROPN
ejpam-86	590	33	∂tβi	∂tβi	ADJ
ejpam-86	590	34	∂ti	∂ti	PROPN
ejpam-86	590	35	∂ti	∂ti	PROPN
ejpam-86	590	36	∂b	∂b	PROPN
ejpam-86	591	1	=	=	PUNCT
ejpam-86	592	1	β	β	X
ejpam-86	592	2	n∑	n∑	X
ejpam-86	592	3	i=1	i=1	PROPN
ejpam-86	593	1	tβ−1	tβ−1	INTJ
ejpam-86	594	1	i	i	PRON
ejpam-86	594	2	x(i)ε	x(i)ε	PROPN
ejpam-86	594	3	′−1	′−1	PUNCT
ejpam-86	594	4	(	(	PUNCT
ejpam-86	594	5	i	i	NOUN
ejpam-86	594	6	)	)	PUNCT
ejpam-86	594	7	=	=	PUNCT
ejpam-86	595	1	β	β	X
ejpam-86	595	2	n∑	n∑	X
ejpam-86	595	3	i=1	i=1	PROPN
ejpam-86	595	4	(	(	PUNCT
ejpam-86	595	5	(	(	PUNCT
ejpam-86	595	6	(	(	PUNCT
ejpam-86	595	7	y(i)−b′x(i))′−1(y(i)−b′x(i)))β−1x(i)(y(i)−b′x(i))′−1	y(i)−b′x(i))′−1(y(i)−b′x(i)))β−1x(i)(y(i)−b′x(i))′−1	NOUN
ejpam-86	595	8	)	)	PUNCT
ejpam-86	595	9	,	,	PUNCT
ejpam-86	595	10	(	(	PUNCT
ejpam-86	595	11	8.7	8.7	NUM
ejpam-86	595	12	)	)	PUNCT
ejpam-86	595	13	∂l(θ	∂l(θ	PROPN
ejpam-86	595	14	)	)	PUNCT
ejpam-86	595	15	∂b	∂b	PROPN
ejpam-86	595	16	=	=	SYM
ejpam-86	595	17	∂l(θ	∂l(θ	PROPN
ejpam-86	595	18	)	)	PUNCT
ejpam-86	595	19	∂v	∂v	PROPN
ejpam-86	595	20	ec(b′	ec(b′	PROPN
ejpam-86	595	21	)	)	PUNCT
ejpam-86	595	22	=	=	SYM
ejpam-86	595	23	vec∂l(θ	vec∂l(θ	NOUN
ejpam-86	595	24	)	)	PUNCT
ejpam-86	595	25	∂b′	∂b′	NOUN
ejpam-86	595	26	=	=	SYM
ejpam-86	595	27	v	v	ADP
ejpam-86	595	28	ec(∂l(θ	ec(∂l(θ	PROPN
ejpam-86	595	29	)	)	PUNCT
ejpam-86	595	30	∂b	∂b	PROPN
ejpam-86	595	31	)	)	PUNCT
ejpam-86	595	32	′	′	NUM
ejpam-86	596	1	=	=	PUNCT
ejpam-86	597	1	β	β	X
ejpam-86	597	2	n∑	n∑	X
ejpam-86	597	3	i=1	i=1	PROPN
ejpam-86	598	1	tβ−1	tβ−1	INTJ
ejpam-86	598	2	i	i	PRON
ejpam-86	598	3	v	v	ADP
ejpam-86	598	4	ec(σ−1ε(i)x′(i	ec(σ−1ε(i)x′(i	NUM
ejpam-86	598	5	)	)	PUNCT
ejpam-86	598	6	)	)	PUNCT
ejpam-86	598	7	(	(	PUNCT
ejpam-86	598	8	8.8	8.8	NUM
ejpam-86	598	9	)	)	PUNCT
ejpam-86	598	10	∂2l(θ	∂2l(θ	NOUN
ejpam-86	598	11	)	)	PUNCT
ejpam-86	598	12	∂b′∂b	∂b′∂b	NUM
ejpam-86	599	1	=	=	SYM
ejpam-86	599	2	∂	∂	NUM
ejpam-86	599	3	∂b′	∂b′	NOUN
ejpam-86	599	4	(	(	PUNCT
ejpam-86	599	5	∂l(θ	∂l(θ	PROPN
ejpam-86	599	6	)	)	PUNCT
ejpam-86	599	7	∂b	∂b	PROPN
ejpam-86	599	8	)	)	PUNCT
ejpam-86	600	1	=	=	SYM
ejpam-86	600	2	∂	∂	NUM
ejpam-86	600	3	∂b′	∂b′	NOUN
ejpam-86	600	4	(	(	PUNCT
ejpam-86	600	5	β	β	X
ejpam-86	600	6	n∑	n∑	X
ejpam-86	600	7	i=1	i=1	PROPN
ejpam-86	600	8	tβ−1	tβ−1	INTJ
ejpam-86	601	1	i	i	PRON
ejpam-86	601	2	x(i)ε	x(i)ε	PROPN
ejpam-86	601	3	′−1	′−1	PUNCT
ejpam-86	601	4	(	(	PUNCT
ejpam-86	601	5	i	i	NOUN
ejpam-86	601	6	)	)	PUNCT
ejpam-86	601	7	)	)	PUNCT
ejpam-86	602	1	=	=	PUNCT
ejpam-86	603	1	β	β	X
ejpam-86	603	2	n∑	n∑	X
ejpam-86	603	3	i=1	i=1	PROPN
ejpam-86	603	4	∂(tβ−1	∂(tβ−1	PROPN
ejpam-86	604	1	i	i	PRON
ejpam-86	604	2	x(i)ε	x(i)ε	PROPN
ejpam-86	604	3	′−1	′−1	PUNCT
ejpam-86	604	4	(	(	PUNCT
ejpam-86	604	5	i	i	NOUN
ejpam-86	604	6	)	)	PUNCT
ejpam-86	604	7	)	)	PUNCT
ejpam-86	604	8	∂b′	∂b′	NOUN
ejpam-86	604	9	=	=	SYM
ejpam-86	605	1	β	β	X
ejpam-86	605	2	n∑	n∑	X
ejpam-86	605	3	i=1	i=1	PROPN
ejpam-86	606	1	(	(	PUNCT
ejpam-86	606	2	tβ−1	tβ−1	VERB
ejpam-86	606	3	i	i	PRON
ejpam-86	606	4	∂(x(i)ε	∂(x(i)ε	ADP
ejpam-86	606	5	′−1	′−1	X
ejpam-86	606	6	(	(	PUNCT
ejpam-86	606	7	i	i	NOUN
ejpam-86	606	8	)	)	PUNCT
ejpam-86	606	9	)	)	PUNCT
ejpam-86	606	10	∂b′	∂b′	NOUN
ejpam-86	607	1	+	+	CCONJ
ejpam-86	607	2	(	(	PUNCT
ejpam-86	607	3	x(i)ε	x(i)ε	PROPN
ejpam-86	607	4	′−1	′−1	SYM
ejpam-86	607	5	(	(	PUNCT
ejpam-86	607	6	i	i	NOUN
ejpam-86	607	7	)	)	PUNCT
ejpam-86	607	8	)	)	PUNCT
ejpam-86	608	1	⊗	⊗	PROPN
ejpam-86	608	2	∂tβ−1	∂tβ−1	NUM
ejpam-86	608	3	i	i	PRON
ejpam-86	608	4	∂b′	∂b′	NOUN
ejpam-86	608	5	)	)	PUNCT
ejpam-86	608	6	=	=	PUNCT
ejpam-86	609	1	β	β	X
ejpam-86	609	2	n∑	n∑	X
ejpam-86	609	3	i=1	i=1	PROPN
ejpam-86	610	1	(	(	PUNCT
ejpam-86	610	2	tβ−1	tβ−1	NOUN
ejpam-86	610	3	i	i	PRON
ejpam-86	610	4	(	(	PUNCT
ejpam-86	610	5	x(i	x(i	PROPN
ejpam-86	610	6	)	)	PUNCT
ejpam-86	610	7	⊗	⊗	NUM
ejpam-86	610	8	ip	ip	NOUN
ejpam-86	610	9	)	)	PUNCT
ejpam-86	610	10	∂ε′	∂ε′	NOUN
ejpam-86	610	11	(	(	PUNCT
ejpam-86	610	12	i	i	NOUN
ejpam-86	610	13	)	)	PUNCT
ejpam-86	610	14	∂b′	∂b′	PROPN
ejpam-86	610	15	(	(	PUNCT
ejpam-86	610	16	σ	σ	NUM
ejpam-86	610	17	−1	−1	NOUN
ejpam-86	610	18	⊗	⊗	PROPN
ejpam-86	610	19	iq	iq	PROPN
ejpam-86	610	20	)	)	PUNCT
ejpam-86	611	1	+	+	CCONJ
ejpam-86	611	2	(	(	PUNCT
ejpam-86	611	3	x(i)ε	x(i)ε	PROPN
ejpam-86	611	4	′−1	′−1	SYM
ejpam-86	611	5	(	(	PUNCT
ejpam-86	611	6	i	i	NOUN
ejpam-86	611	7	)	)	PUNCT
ejpam-86	611	8	)	)	PUNCT
ejpam-86	612	1	⊗	⊗	NOUN
ejpam-86	612	2	(	(	PUNCT
ejpam-86	612	3	∂tβ−1	∂tβ−1	NUM
ejpam-86	612	4	i	i	PRON
ejpam-86	612	5	∂ti	∂ti	PROPN
ejpam-86	612	6	∂ti	∂ti	PROPN
ejpam-86	612	7	∂b′	∂b′	NOUN
ejpam-86	612	8	)	)	PUNCT
ejpam-86	612	9	)	)	PUNCT
ejpam-86	613	1	=	=	PUNCT
ejpam-86	614	1	β	β	X
ejpam-86	614	2	n∑	n∑	X
ejpam-86	614	3	i=1	i=1	PROPN
ejpam-86	614	4	(	(	PUNCT
ejpam-86	614	5	−tβ−1	−tβ−1	X
ejpam-86	614	6	i	i	PRON
ejpam-86	614	7	(	(	PUNCT
ejpam-86	614	8	x(i	x(i	PROPN
ejpam-86	614	9	)	)	PUNCT
ejpam-86	614	10	⊗	⊗	PROPN
ejpam-86	614	11	ip)(x′(i	ip)(x′(i	PUNCT
ejpam-86	614	12	)	)	PUNCT
ejpam-86	614	13	⊗	⊗	PROPN
ejpam-86	614	14	ip)i(p	ip)i(p	NOUN
ejpam-86	614	15	,	,	PUNCT
ejpam-86	614	16	q)(σ−1	q)(σ−1	PROPN
ejpam-86	614	17	⊗	⊗	PROPN
ejpam-86	614	18	iq	iq	PROPN
ejpam-86	614	19	)	)	PUNCT
ejpam-86	614	20	−2(β	−2(β	ADP
ejpam-86	614	21	−	−	PROPN
ejpam-86	614	22	1)tβ−2	1)tβ−2	NUM
ejpam-86	615	1	i	i	INTJ
ejpam-86	615	2	(	(	PUNCT
ejpam-86	615	3	x(i)ε	x(i)ε	PROPN
ejpam-86	615	4	′−1	′−1	SYM
ejpam-86	615	5	(	(	PUNCT
ejpam-86	615	6	i	i	NOUN
ejpam-86	615	7	)	)	PUNCT
ejpam-86	615	8	)	)	PUNCT
ejpam-86	616	1	⊗	⊗	PROPN
ejpam-86	616	2	(	(	PUNCT
ejpam-86	616	3	σ−1ε(i)x′(i	σ−1ε(i)x′(i	PROPN
ejpam-86	616	4	)	)	PUNCT
ejpam-86	616	5	)	)	PUNCT
ejpam-86	616	6	)	)	PUNCT
ejpam-86	617	1	=	=	PUNCT
ejpam-86	618	1	−β	−β	PROPN
ejpam-86	618	2	n∑	n∑	PROPN
ejpam-86	618	3	i=1	i=1	PROPN
ejpam-86	619	1	(	(	PUNCT
ejpam-86	619	2	tβ−1	tβ−1	NOUN
ejpam-86	619	3	i	i	PRON
ejpam-86	619	4	(	(	PUNCT
ejpam-86	619	5	x(i)x	x(i)x	PROPN
ejpam-86	619	6	′−1	′−1	SYM
ejpam-86	619	7	(	(	PUNCT
ejpam-86	619	8	i	i	NOUN
ejpam-86	619	9	)	)	PUNCT
ejpam-86	619	10	)	)	PUNCT
ejpam-86	620	1	+2(β	+2(β	PROPN
ejpam-86	620	2	−	−	NUM
ejpam-86	621	1	1)tβ−2	1)tβ−2	NUM
ejpam-86	621	2	i	i	INTJ
ejpam-86	621	3	(	(	PUNCT
ejpam-86	621	4	x(i)ε	x(i)ε	PROPN
ejpam-86	621	5	′−1	′−1	SYM
ejpam-86	621	6	(	(	PUNCT
ejpam-86	621	7	i	i	NOUN
ejpam-86	621	8	)	)	PUNCT
ejpam-86	621	9	)	)	PUNCT
ejpam-86	622	1	⊗	⊗	PROPN
ejpam-86	622	2	(	(	PUNCT
ejpam-86	622	3	σ−1ε(i)x′(i	σ−1ε(i)x′(i	PROPN
ejpam-86	622	4	)	)	PUNCT
ejpam-86	622	5	)	)	PUNCT
ejpam-86	622	6	)	)	PUNCT
ejpam-86	622	7	(	(	PUNCT
ejpam-86	622	8	8.9	8.9	NUM
ejpam-86	622	9	)	)	PUNCT
ejpam-86	622	10	∂(tβ−1	∂(tβ−1	NUM
ejpam-86	623	1	i	i	PRON
ejpam-86	623	2	σ−1ε(i)x	σ−1ε(i)x	VERB
ejpam-86	623	3	′	′	NUM
ejpam-86	623	4	(	(	PUNCT
ejpam-86	623	5	i	i	NOUN
ejpam-86	623	6	)	)	PUNCT
ejpam-86	623	7	)	)	PUNCT
ejpam-86	623	8	∂b′	∂b′	NOUN
ejpam-86	623	9	=	=	SYM
ejpam-86	623	10	tβ−1	tβ−1	NOUN
ejpam-86	623	11	i	i	PRON
ejpam-86	623	12	∂(σ−1ε(i)x	∂(σ−1ε(i)x	VERB
ejpam-86	624	1	′	′	NUM
ejpam-86	624	2	(	(	PUNCT
ejpam-86	624	3	i	i	NOUN
ejpam-86	624	4	)	)	PUNCT
ejpam-86	624	5	)	)	PUNCT
ejpam-86	624	6	∂b′	∂b′	NOUN
ejpam-86	625	1	+	+	CCONJ
ejpam-86	625	2	(	(	PUNCT
ejpam-86	625	3	σ−1ε(i)x′(i))⊗	σ−1ε(i)x′(i))⊗	PROPN
ejpam-86	625	4	∂tβ−1	∂tβ−1	NUM
ejpam-86	625	5	i	i	PRON
ejpam-86	625	6	∂b′	∂b′	NOUN
ejpam-86	625	7	=	=	PUNCT
ejpam-86	625	8	tβ−1	tβ−1	NOUN
ejpam-86	625	9	i	i	PRON
ejpam-86	625	10	(	(	PUNCT
ejpam-86	625	11	σ−1	σ−1	PROPN
ejpam-86	625	12	⊗	⊗	PROPN
ejpam-86	625	13	ip	ip	PROPN
ejpam-86	625	14	)	)	PUNCT
ejpam-86	625	15	∂ε(i	∂ε(i	PROPN
ejpam-86	625	16	)	)	PUNCT
ejpam-86	625	17	∂b′	∂b′	NOUN
ejpam-86	625	18	(	(	PUNCT
ejpam-86	625	19	x	x	SYM
ejpam-86	625	20	′	′	NUM
ejpam-86	625	21	(	(	PUNCT
ejpam-86	625	22	i	i	NOUN
ejpam-86	625	23	)	)	PUNCT
ejpam-86	625	24	⊗	⊗	PROPN
ejpam-86	625	25	iq	iq	PROPN
ejpam-86	625	26	)	)	PUNCT
ejpam-86	625	27	+	+	CCONJ
ejpam-86	625	28	(	(	PUNCT
ejpam-86	625	29	σ−1ε(i)x′(i))⊗	σ−1ε(i)x′(i))⊗	PROPN
ejpam-86	625	30	∂tβ−1	∂tβ−1	NUM
ejpam-86	625	31	i	i	PRON
ejpam-86	625	32	∂b′	∂b′	NOUN
ejpam-86	625	33	=	=	PUNCT
ejpam-86	625	34	−tβ−1	−tβ−1	PUNCT
ejpam-86	625	35	i	i	PRON
ejpam-86	625	36	(	(	PUNCT
ejpam-86	625	37	σ−1	σ−1	PROPN
ejpam-86	625	38	⊗	⊗	PROPN
ejpam-86	625	39	ip)v	ip)v	PROPN
ejpam-86	625	40	ec(ip)v	ec(ip)v	PROPN
ejpam-86	625	41	ec′(x(i))(x′(i	ec′(x(i))(x′(i	NOUN
ejpam-86	625	42	)	)	PUNCT
ejpam-86	625	43	⊗	⊗	PROPN
ejpam-86	625	44	iq	iq	PROPN
ejpam-86	625	45	)	)	PUNCT
ejpam-86	625	46	−(σ−1ε(i)x′(i))⊗	−(σ−1ε(i)x′(i))⊗	PROPN
ejpam-86	625	47	(	(	PUNCT
ejpam-86	625	48	∂tβ−1	∂tβ−1	NUM
ejpam-86	625	49	i	i	PRON
ejpam-86	625	50	∂ti	∂ti	PROPN
ejpam-86	625	51	∂ti	∂ti	PROPN
ejpam-86	625	52	∂b′	∂b′	NOUN
ejpam-86	625	53	)	)	PUNCT
ejpam-86	626	1	=	=	PUNCT
ejpam-86	626	2	−tβ−1	−tβ−1	PUNCT
ejpam-86	626	3	i	i	PRON
ejpam-86	626	4	v	v	VERB
ejpam-86	626	5	ec(σ−1)v	ec(σ−1)v	PROPN
ejpam-86	626	6	ec′(x(i)x′(i	ec′(x(i)x′(i	NOUN
ejpam-86	626	7	)	)	PUNCT
ejpam-86	626	8	)	)	PUNCT
ejpam-86	626	9	−2(β	−2(β	ADP
ejpam-86	626	10	−	−	PROPN
ejpam-86	626	11	1)tβ−2	1)tβ−2	NUM
ejpam-86	627	1	i	i	INTJ
ejpam-86	627	2	(	(	PUNCT
ejpam-86	627	3	σ−1ε(i)x	σ−1ε(i)x	PROPN
ejpam-86	627	4	′−1	′−1	X
ejpam-86	627	5	(	(	PUNCT
ejpam-86	627	6	i	i	NOUN
ejpam-86	627	7	)	)	PUNCT
ejpam-86	627	8	ε(i)x	ε(i)x	NOUN
ejpam-86	627	9	′	′	NUM
ejpam-86	627	10	(	(	PUNCT
ejpam-86	627	11	i	i	NOUN
ejpam-86	627	12	)	)	PUNCT
ejpam-86	627	13	)	)	PUNCT
ejpam-86	627	14	(	(	PUNCT
ejpam-86	627	15	8.10	8.10	NUM
ejpam-86	627	16	)	)	PUNCT
ejpam-86	627	17	m.	m.	NOUN
ejpam-86	627	18	liu	liu	PROPN
ejpam-86	627	19	and	and	CCONJ
ejpam-86	627	20	h.	h.	PROPN
ejpam-86	627	21	bozdogan	bozdogan	PROPN
ejpam-86	627	22	/	/	SYM
ejpam-86	627	23	eur	eur	PROPN
ejpam-86	627	24	.	.	PUNCT
ejpam-86	628	1	j.	j.	PROPN
ejpam-86	628	2	pure	pure	PROPN
ejpam-86	628	3	appl	appl	PROPN
ejpam-86	628	4	.	.	PROPN
ejpam-86	628	5	math	math	PROPN
ejpam-86	628	6	,	,	PUNCT
ejpam-86	628	7	1	1	NUM
ejpam-86	628	8	(	(	PUNCT
ejpam-86	628	9	2008	2008	NUM
ejpam-86	628	10	)	)	PUNCT
ejpam-86	628	11	,	,	PUNCT
ejpam-86	628	12	(	(	PUNCT
ejpam-86	628	13	4	4	NUM
ejpam-86	628	14	-	-	SYM
ejpam-86	628	15	37	37	NUM
ejpam-86	628	16	)	)	PUNCT
ejpam-86	628	17	33	33	NUM
ejpam-86	628	18	∂2(θ	∂2(θ	NUM
ejpam-86	628	19	)	)	PUNCT
ejpam-86	628	20	∂b′∂b	∂b′∂b	PUNCT
ejpam-86	629	1	=	=	SYM
ejpam-86	629	2	∂	∂	NUM
ejpam-86	629	3	∂b′	∂b′	NOUN
ejpam-86	629	4	(	(	PUNCT
ejpam-86	629	5	∂l(θ	∂l(θ	PROPN
ejpam-86	629	6	)	)	PUNCT
ejpam-86	629	7	∂b	∂b	PROPN
ejpam-86	629	8	)	)	PUNCT
ejpam-86	630	1	=	=	PUNCT
ejpam-86	631	1	β	β	X
ejpam-86	631	2	n∑	n∑	NOUN
ejpam-86	631	3	i=1	i=1	PROPN
ejpam-86	632	1	∂v	∂v	PROPN
ejpam-86	632	2	ec(tβ−1	ec(tβ−1	VERB
ejpam-86	632	3	i	i	PRON
ejpam-86	632	4	σ−1ε(i)x	σ−1ε(i)x	NUM
ejpam-86	632	5	′	′	NUM
ejpam-86	633	1	(	(	PUNCT
ejpam-86	633	2	i	i	NOUN
ejpam-86	633	3	)	)	PUNCT
ejpam-86	633	4	)	)	PUNCT
ejpam-86	633	5	∂b′	∂b′	NOUN
ejpam-86	633	6	=	=	SYM
ejpam-86	634	1	β	β	X
ejpam-86	634	2	n∑	n∑	NOUN
ejpam-86	634	3	i=1	i=1	PROPN
ejpam-86	635	1	∂v	∂v	PROPN
ejpam-86	635	2	ec(tβ−1	ec(tβ−1	VERB
ejpam-86	635	3	i	i	PRON
ejpam-86	635	4	σ−1ε(i)x	σ−1ε(i)x	NUM
ejpam-86	635	5	′	′	NUM
ejpam-86	636	1	(	(	PUNCT
ejpam-86	636	2	i	i	NOUN
ejpam-86	636	3	)	)	PUNCT
ejpam-86	636	4	)	)	PUNCT
ejpam-86	637	1	∂v	∂v	PROPN
ejpam-86	637	2	ec′(b′	ec′(b′	NOUN
ejpam-86	637	3	)	)	PUNCT
ejpam-86	637	4	=	=	PUNCT
ejpam-86	637	5	β	β	PROPN
ejpam-86	637	6	n∑	n∑	X
ejpam-86	637	7	i=1	i=1	PROPN
ejpam-86	637	8	(	(	PUNCT
ejpam-86	637	9	ipq	ipq	PROPN
ejpam-86	637	10	⊗	⊗	PROPN
ejpam-86	637	11	v	v	NOUN
ejpam-86	637	12	ec′(ip))(iq	ec′(ip))(iq	NOUN
ejpam-86	638	1	⊗	⊗	PROPN
ejpam-86	638	2	∂(tβ−1	∂(tβ−1	NUM
ejpam-86	639	1	i	i	PRON
ejpam-86	639	2	σ−1ε(i)x	σ−1ε(i)x	VERB
ejpam-86	639	3	′	′	NUM
ejpam-86	639	4	(	(	PUNCT
ejpam-86	639	5	i	i	NOUN
ejpam-86	639	6	)	)	PUNCT
ejpam-86	639	7	)	)	PUNCT
ejpam-86	639	8	∂b′	∂b′	PROPN
ejpam-86	639	9	⊗	⊗	NUM
ejpam-86	639	10	ip)(v	ip)(v	PROPN
ejpam-86	639	11	ec(iq)⊗	ec(iq)⊗	NOUN
ejpam-86	639	12	ipq	ipq	NOUN
ejpam-86	639	13	)	)	PUNCT
ejpam-86	639	14	=	=	PUNCT
ejpam-86	640	1	β	β	PROPN
ejpam-86	640	2	n∑	n∑	X
ejpam-86	640	3	i=1	i=1	PROPN
ejpam-86	641	1	(	(	PUNCT
ejpam-86	641	2	ipq	ipq	NOUN
ejpam-86	641	3	⊗	⊗	PROPN
ejpam-86	641	4	v	v	NOUN
ejpam-86	641	5	ec′(ip))(iq	ec′(ip))(iq	NOUN
ejpam-86	641	6	⊗mi	⊗mi	NUM
ejpam-86	641	7	⊗	⊗	PROPN
ejpam-86	641	8	ip)(v	ip)(v	ADJ
ejpam-86	641	9	ec(iq)⊗	ec(iq)⊗	NOUN
ejpam-86	641	10	ipq	ipq	NOUN
ejpam-86	641	11	)	)	PUNCT
ejpam-86	641	12	(	(	PUNCT
ejpam-86	641	13	8.11	8.11	NUM
ejpam-86	641	14	)	)	PUNCT
ejpam-86	641	15	where	where	SCONJ
ejpam-86	641	16	mi	mi	NOUN
ejpam-86	642	1	=	=	PROPN
ejpam-86	642	2	−tβ−1	−tβ−1	PUNCT
ejpam-86	642	3	i	i	PRON
ejpam-86	642	4	v	v	X
ejpam-86	642	5	ec(σ−1)v	ec(σ−1)v	NOUN
ejpam-86	642	6	ec′(x(i)x	ec′(x(i)x	PUNCT
ejpam-86	642	7	′	′	NUM
ejpam-86	642	8	(	(	PUNCT
ejpam-86	642	9	i))−	i))−	NOUN
ejpam-86	642	10	2(β	2(β	NUM
ejpam-86	642	11	−	−	NOUN
ejpam-86	643	1	1)tβ−2	1)tβ−2	NUM
ejpam-86	643	2	i	i	INTJ
ejpam-86	643	3	(	(	PUNCT
ejpam-86	643	4	σ−1ε(i)x	σ−1ε(i)x	PROPN
ejpam-86	643	5	′−1	′−1	X
ejpam-86	643	6	(	(	PUNCT
ejpam-86	643	7	i	i	NOUN
ejpam-86	643	8	)	)	PUNCT
ejpam-86	643	9	ε(i)x	ε(i)x	NOUN
ejpam-86	643	10	′	′	NUM
ejpam-86	643	11	(	(	PUNCT
ejpam-86	643	12	i	i	NOUN
ejpam-86	643	13	)	)	PUNCT
ejpam-86	643	14	)	)	PUNCT
ejpam-86	643	15	.	.	PUNCT
ejpam-86	644	1	∂ti	∂ti	PROPN
ejpam-86	644	2	∂σ	∂σ	PROPN
ejpam-86	645	1	=	=	PUNCT
ejpam-86	645	2	∂(ε′−1	∂(ε′−1	PROPN
ejpam-86	645	3	(	(	PUNCT
ejpam-86	645	4	i	i	NOUN
ejpam-86	645	5	)	)	PUNCT
ejpam-86	645	6	ε(i	ε(i	NOUN
ejpam-86	645	7	)	)	PUNCT
ejpam-86	645	8	)	)	PUNCT
ejpam-86	645	9	∂σ	∂σ	PROPN
ejpam-86	646	1	=	=	PUNCT
ejpam-86	646	2	−σ−1ε(i)ε	−σ−1ε(i)ε	PROPN
ejpam-86	646	3	′−1	′−1	PROPN
ejpam-86	646	4	(	(	PUNCT
ejpam-86	646	5	i	i	NOUN
ejpam-86	646	6	)	)	PUNCT
ejpam-86	646	7	(	(	PUNCT
ejpam-86	646	8	8.12	8.12	NUM
ejpam-86	646	9	)	)	PUNCT
ejpam-86	646	10	∂l(θ	∂l(θ	PROPN
ejpam-86	646	11	)	)	PUNCT
ejpam-86	646	12	∂σ	∂σ	PROPN
ejpam-86	647	1	=	=	NOUN
ejpam-86	647	2	−n	−n	ADJ
ejpam-86	647	3	2	2	NUM
ejpam-86	647	4	∂	∂	NUM
ejpam-86	647	5	log	log	NOUN
ejpam-86	647	6	|σ|	|σ|	PROPN
ejpam-86	647	7	∂σ	∂σ	PROPN
ejpam-86	648	1	−	−	PROPN
ejpam-86	648	2	1	1	NUM
ejpam-86	648	3	2	2	NUM
ejpam-86	648	4	n∑	n∑	NOUN
ejpam-86	648	5	i=1	i=1	PROPN
ejpam-86	648	6	∂tβi	∂tβi	NOUN
ejpam-86	648	7	∂σ	∂σ	X
ejpam-86	649	1	=	=	SYM
ejpam-86	649	2	−n	−n	ADJ
ejpam-86	649	3	2	2	NUM
ejpam-86	650	1	σ−1	σ−1	DET
ejpam-86	650	2	−	−	NOUN
ejpam-86	650	3	1	1	NUM
ejpam-86	650	4	2	2	NUM
ejpam-86	650	5	n∑	n∑	NOUN
ejpam-86	650	6	i=1	i=1	PROPN
ejpam-86	650	7	∂tβ−1	∂tβ−1	PROPN
ejpam-86	651	1	i	i	PRON
ejpam-86	651	2	∂ti	∂ti	PROPN
ejpam-86	651	3	∂ti	∂ti	PROPN
ejpam-86	651	4	∂σ	∂σ	PROPN
ejpam-86	652	1	=	=	PUNCT
ejpam-86	652	2	−n	−n	ADJ
ejpam-86	652	3	2	2	NUM
ejpam-86	653	1	σ−1	σ−1	DET
ejpam-86	653	2	−	−	NOUN
ejpam-86	653	3	1	1	NUM
ejpam-86	653	4	2	2	NUM
ejpam-86	653	5	n∑	n∑	NOUN
ejpam-86	653	6	i=1	i=1	X
ejpam-86	654	1	βtβ−1	βtβ−1	PUNCT
ejpam-86	655	1	i	i	PRON
ejpam-86	655	2	∂ti	∂ti	PROPN
ejpam-86	655	3	∂σ	∂σ	PROPN
ejpam-86	656	1	=	=	PUNCT
ejpam-86	656	2	−n	−n	ADJ
ejpam-86	656	3	2	2	NUM
ejpam-86	657	1	σ−1	σ−1	PROPN
ejpam-86	657	2	+	+	CCONJ
ejpam-86	657	3	β	β	X
ejpam-86	657	4	2	2	NUM
ejpam-86	657	5	n∑	n∑	NOUN
ejpam-86	657	6	i=1	i=1	PRON
ejpam-86	658	1	tβ−1	tβ−1	ADV
ejpam-86	658	2	i	i	PRON
ejpam-86	658	3	σ−1ε(i)ε	σ−1ε(i)ε	NUM
ejpam-86	658	4	′−1	′−1	SYM
ejpam-86	658	5	(	(	PUNCT
ejpam-86	658	6	i	i	NOUN
ejpam-86	658	7	)	)	PUNCT
ejpam-86	658	8	.	.	PUNCT
ejpam-86	659	1	(	(	PUNCT
ejpam-86	659	2	8.13	8.13	NUM
ejpam-86	659	3	)	)	PUNCT
ejpam-86	659	4	∂(σ−1ε(i)ε	∂(σ−1ε(i)ε	PROPN
ejpam-86	659	5	′−1	′−1	X
ejpam-86	659	6	(	(	PUNCT
ejpam-86	659	7	i	i	NOUN
ejpam-86	659	8	)	)	PUNCT
ejpam-86	659	9	)	)	PUNCT
ejpam-86	660	1	∂σ	∂σ	PROPN
ejpam-86	661	1	=	=	PUNCT
ejpam-86	661	2	∂σ−1	∂σ−1	PROPN
ejpam-86	661	3	∂σ	∂σ	PROPN
ejpam-86	661	4	(	(	PUNCT
ejpam-86	661	5	ε(i)ε	ε(i)ε	PROPN
ejpam-86	661	6	′−1	′−1	SYM
ejpam-86	661	7	(	(	PUNCT
ejpam-86	661	8	i	i	NOUN
ejpam-86	661	9	)	)	PUNCT
ejpam-86	661	10	⊗	⊗	PROPN
ejpam-86	661	11	ip	ip	NOUN
ejpam-86	661	12	)	)	PUNCT
ejpam-86	662	1	+	+	CCONJ
ejpam-86	662	2	(	(	PUNCT
ejpam-86	662	3	σ−1	σ−1	PROPN
ejpam-86	662	4	⊗	⊗	PROPN
ejpam-86	662	5	ip)(ε(i)ε′(i	ip)(ε(i)ε′(i	PUNCT
ejpam-86	662	6	)	)	PUNCT
ejpam-86	662	7	⊗	⊗	NOUN
ejpam-86	662	8	ip)∂σ−1	ip)∂σ−1	PUNCT
ejpam-86	663	1	∂σ	∂σ	PROPN
ejpam-86	663	2	=	=	PUNCT
ejpam-86	663	3	−v	−v	NOUN
ejpam-86	663	4	ec(σ−1)v	ec(σ−1)v	VERB
ejpam-86	663	5	ec′−1)(ε(i)ε	ec′−1)(ε(i)ε	PROPN
ejpam-86	663	6	′−1	′−1	PUNCT
ejpam-86	663	7	(	(	PUNCT
ejpam-86	663	8	i	i	NOUN
ejpam-86	663	9	)	)	PUNCT
ejpam-86	663	10	⊗	⊗	PROPN
ejpam-86	663	11	ip	ip	NOUN
ejpam-86	663	12	)	)	PUNCT
ejpam-86	663	13	−(σ−1ε(i)ε	−(σ−1ε(i)ε	NUM
ejpam-86	663	14	′	′	NUM
ejpam-86	664	1	(	(	PUNCT
ejpam-86	664	2	i	i	NOUN
ejpam-86	664	3	)	)	PUNCT
ejpam-86	664	4	⊗	⊗	PROPN
ejpam-86	664	5	ip)v	ip)v	PROPN
ejpam-86	664	6	ec(σ−1)v	ec(σ−1)v	PROPN
ejpam-86	664	7	ec′−1	ec′−1	NOUN
ejpam-86	664	8	)	)	PUNCT
ejpam-86	664	9	=	=	NOUN
ejpam-86	664	10	−v	−v	NOUN
ejpam-86	664	11	ec(σ−1)v	ec(σ−1)v	VERB
ejpam-86	664	12	ec′−1ε(i)ε	ec′−1ε(i)ε	NOUN
ejpam-86	664	13	′−1	′−1	X
ejpam-86	664	14	(	(	PUNCT
ejpam-86	664	15	i	i	NOUN
ejpam-86	664	16	)	)	PUNCT
ejpam-86	664	17	)	)	PUNCT
ejpam-86	664	18	−v	−v	VERB
ejpam-86	664	19	ec(σ−1ε(i)ε	ec(σ−1ε(i)ε	NOUN
ejpam-86	664	20	′−1	′−1	X
ejpam-86	664	21	(	(	PUNCT
ejpam-86	664	22	i	i	NOUN
ejpam-86	664	23	)	)	PUNCT
ejpam-86	664	24	)	)	PUNCT
ejpam-86	664	25	v	v	NOUN
ejpam-86	664	26	ec′−1	ec′−1	NOUN
ejpam-86	664	27	)	)	PUNCT
ejpam-86	664	28	(	(	PUNCT
ejpam-86	664	29	8.14	8.14	NUM
ejpam-86	664	30	)	)	PUNCT
ejpam-86	664	31	∂2l(θ	∂2l(θ	NOUN
ejpam-86	664	32	)	)	PUNCT
ejpam-86	664	33	∂σ′∂σ	∂σ′∂σ	PUNCT
ejpam-86	665	1	=	=	SYM
ejpam-86	665	2	∂	∂	NUM
ejpam-86	665	3	∂σ′	∂σ′	PRON
ejpam-86	665	4	(	(	PUNCT
ejpam-86	665	5	∂l(θ	∂l(θ	PROPN
ejpam-86	665	6	)	)	PUNCT
ejpam-86	665	7	∂σ	∂σ	PROPN
ejpam-86	665	8	)	)	PUNCT
ejpam-86	666	1	=	=	PUNCT
ejpam-86	666	2	−n	−n	ADJ
ejpam-86	666	3	2	2	NUM
ejpam-86	666	4	∂σ−1	∂σ−1	VERB
ejpam-86	666	5	∂σ	∂σ	PROPN
ejpam-86	667	1	+	+	CCONJ
ejpam-86	667	2	β	β	X
ejpam-86	667	3	2	2	NUM
ejpam-86	667	4	n∑	n∑	NOUN
ejpam-86	667	5	i=1	i=1	PROPN
ejpam-86	667	6	∂(tβ−1	∂(tβ−1	PROPN
ejpam-86	668	1	i	i	PRON
ejpam-86	668	2	σ−1ε(i)ε	σ−1ε(i)ε	NUM
ejpam-86	668	3	′−1	′−1	SYM
ejpam-86	668	4	(	(	PUNCT
ejpam-86	668	5	i	i	NOUN
ejpam-86	668	6	)	)	PUNCT
ejpam-86	668	7	)	)	PUNCT
ejpam-86	669	1	∂σ	∂σ	PROPN
ejpam-86	670	1	=	=	PUNCT
ejpam-86	670	2	n	n	PROPN
ejpam-86	670	3	2v	2v	PROPN
ejpam-86	670	4	ec(σ	ec(σ	CCONJ
ejpam-86	670	5	−1)v	−1)v	X
ejpam-86	670	6	ec′−1	ec′−1	PROPN
ejpam-86	670	7	)	)	PUNCT
ejpam-86	670	8	+	+	NOUN
ejpam-86	670	9	β	β	X
ejpam-86	670	10	2	2	NUM
ejpam-86	670	11	n∑	n∑	NOUN
ejpam-86	670	12	i=1	i=1	PROPN
ejpam-86	670	13	(	(	PUNCT
ejpam-86	670	14	tβ−1	tβ−1	VERB
ejpam-86	670	15	i	i	PRON
ejpam-86	670	16	∂(σ−1ε(i)ε	∂(σ−1ε(i)ε	NUM
ejpam-86	670	17	′−1	′−1	X
ejpam-86	670	18	(	(	PUNCT
ejpam-86	670	19	i	i	NOUN
ejpam-86	670	20	)	)	PUNCT
ejpam-86	670	21	)	)	PUNCT
ejpam-86	670	22	∂σ	∂σ	PROPN
ejpam-86	671	1	+	+	CCONJ
ejpam-86	671	2	(	(	PUNCT
ejpam-86	671	3	σ−1ε(i)ε	σ−1ε(i)ε	NUM
ejpam-86	671	4	′−1	′−1	SYM
ejpam-86	671	5	(	(	PUNCT
ejpam-86	671	6	i	i	NOUN
ejpam-86	671	7	)	)	PUNCT
ejpam-86	671	8	)	)	PUNCT
ejpam-86	672	1	⊗	⊗	PROPN
ejpam-86	672	2	∂tβ−1	∂tβ−1	NUM
ejpam-86	673	1	i	i	PRON
ejpam-86	673	2	∂σ	∂σ	PROPN
ejpam-86	673	3	)	)	PUNCT
ejpam-86	674	1	=	=	SYM
ejpam-86	674	2	n	n	NUM
ejpam-86	674	3	2v	2v	PROPN
ejpam-86	674	4	ec(σ	ec(σ	CCONJ
ejpam-86	674	5	−1)v	−1)v	X
ejpam-86	674	6	ec′−1	ec′−1	PROPN
ejpam-86	674	7	)	)	PUNCT
ejpam-86	674	8	−β	−β	NOUN
ejpam-86	674	9	2	2	NUM
ejpam-86	674	10	n∑	n∑	NOUN
ejpam-86	674	11	i=1	i=1	PRON
ejpam-86	674	12	tβ−1	tβ−1	ADV
ejpam-86	675	1	i	i	NOUN
ejpam-86	675	2	(	(	PUNCT
ejpam-86	675	3	v	v	NOUN
ejpam-86	675	4	ec(σ−1)v	ec(σ−1)v	VERB
ejpam-86	675	5	ec′−1ε(i)ε	ec′−1ε(i)ε	NOUN
ejpam-86	675	6	′−1	′−1	SYM
ejpam-86	675	7	(	(	PUNCT
ejpam-86	675	8	i	i	NOUN
ejpam-86	675	9	)	)	PUNCT
ejpam-86	675	10	)	)	PUNCT
ejpam-86	676	1	+	+	ADP
ejpam-86	676	2	v	v	NUM
ejpam-86	676	3	ec(σ−1ε(i)ε	ec(σ−1ε(i)ε	NOUN
ejpam-86	676	4	′−1	′−1	X
ejpam-86	676	5	(	(	PUNCT
ejpam-86	676	6	i	i	NOUN
ejpam-86	676	7	)	)	PUNCT
ejpam-86	676	8	)	)	PUNCT
ejpam-86	676	9	v	v	NOUN
ejpam-86	676	10	ec′−1	ec′−1	NOUN
ejpam-86	676	11	)	)	PUNCT
ejpam-86	676	12	)	)	PUNCT
ejpam-86	676	13	−β(β−1	−β(β−1	NUM
ejpam-86	676	14	)	)	PUNCT
ejpam-86	676	15	2	2	NUM
ejpam-86	676	16	n∑	n∑	NOUN
ejpam-86	676	17	i=1	i=1	PROPN
ejpam-86	676	18	(	(	PUNCT
ejpam-86	676	19	tβ−2	tβ−2	PROPN
ejpam-86	676	20	i	i	PRON
ejpam-86	676	21	(	(	PUNCT
ejpam-86	676	22	σ−1ε(i)ε	σ−1ε(i)ε	NUM
ejpam-86	676	23	′−1	′−1	SYM
ejpam-86	676	24	(	(	PUNCT
ejpam-86	676	25	i	i	NOUN
ejpam-86	676	26	)	)	PUNCT
ejpam-86	676	27	)	)	PUNCT
ejpam-86	677	1	⊗	⊗	PROPN
ejpam-86	677	2	(	(	PUNCT
ejpam-86	677	3	σ−1ε(i)ε	σ−1ε(i)ε	NUM
ejpam-86	677	4	′−1	′−1	SYM
ejpam-86	677	5	(	(	PUNCT
ejpam-86	677	6	i	i	NOUN
ejpam-86	677	7	)	)	PUNCT
ejpam-86	677	8	)	)	PUNCT
ejpam-86	677	9	)	)	PUNCT
ejpam-86	678	1	(	(	PUNCT
ejpam-86	678	2	8.15	8.15	NUM
ejpam-86	678	3	)	)	PUNCT
ejpam-86	678	4	m.	m.	NOUN
ejpam-86	678	5	liu	liu	PROPN
ejpam-86	678	6	and	and	CCONJ
ejpam-86	678	7	h.	h.	PROPN
ejpam-86	678	8	bozdogan	bozdogan	PROPN
ejpam-86	678	9	/	/	SYM
ejpam-86	678	10	eur	eur	PROPN
ejpam-86	678	11	.	.	PUNCT
ejpam-86	679	1	j.	j.	PROPN
ejpam-86	679	2	pure	pure	PROPN
ejpam-86	679	3	appl	appl	PROPN
ejpam-86	679	4	.	.	PROPN
ejpam-86	679	5	math	math	PROPN
ejpam-86	679	6	,	,	PUNCT
ejpam-86	679	7	1	1	NUM
ejpam-86	679	8	(	(	PUNCT
ejpam-86	679	9	2008	2008	NUM
ejpam-86	679	10	)	)	PUNCT
ejpam-86	679	11	,	,	PUNCT
ejpam-86	679	12	(	(	PUNCT
ejpam-86	679	13	4	4	NUM
ejpam-86	679	14	-	-	SYM
ejpam-86	679	15	37	37	NUM
ejpam-86	679	16	)	)	PUNCT
ejpam-86	679	17	34	34	NUM
ejpam-86	679	18	∂l(θ	∂l(θ	PROPN
ejpam-86	679	19	)	)	PUNCT
ejpam-86	679	20	∂v	∂v	PROPN
ejpam-86	679	21	ec(σ	ec(σ	PRON
ejpam-86	679	22	)	)	PUNCT
ejpam-86	679	23	=	=	NOUN
ejpam-86	679	24	vec(∂l(θ	vec(∂l(θ	NOUN
ejpam-86	679	25	)	)	PUNCT
ejpam-86	679	26	∂σ	∂σ	PROPN
ejpam-86	679	27	)	)	PUNCT
ejpam-86	680	1	=	=	PUNCT
ejpam-86	680	2	−n	−n	PROPN
ejpam-86	680	3	2v	2v	PROPN
ejpam-86	680	4	ec(σ	ec(σ	CCONJ
ejpam-86	680	5	−1	−1	NOUN
ejpam-86	680	6	)	)	PUNCT
ejpam-86	681	1	+	+	CCONJ
ejpam-86	681	2	β	β	X
ejpam-86	681	3	2	2	NUM
ejpam-86	681	4	n∑	n∑	NOUN
ejpam-86	681	5	i=1	i=1	PRON
ejpam-86	682	1	tβ−1	tβ−1	NOUN
ejpam-86	683	1	i	i	PRON
ejpam-86	683	2	v	v	VERB
ejpam-86	683	3	ec(σ−1ε(i)ε	ec(σ−1ε(i)ε	NOUN
ejpam-86	683	4	′−1	′−1	X
ejpam-86	683	5	(	(	PUNCT
ejpam-86	683	6	i	i	NOUN
ejpam-86	683	7	)	)	PUNCT
ejpam-86	683	8	)	)	PUNCT
ejpam-86	683	9	.	.	PUNCT
ejpam-86	684	1	(	(	PUNCT
ejpam-86	684	2	8.16	8.16	NUM
ejpam-86	684	3	)	)	PUNCT
ejpam-86	684	4	∂2l(θ	∂2l(θ	NOUN
ejpam-86	684	5	)	)	PUNCT
ejpam-86	685	1	∂v	∂v	PROPN
ejpam-86	685	2	ec′(σ)∂v	ec′(σ)∂v	NOUN
ejpam-86	685	3	ec(σ	ec(σ	NUM
ejpam-86	685	4	)	)	PUNCT
ejpam-86	685	5	=	=	SYM
ejpam-86	685	6	(	(	PUNCT
ejpam-86	685	7	ip2	ip2	PROPN
ejpam-86	685	8	⊗	⊗	PROPN
ejpam-86	685	9	v	v	ADP
ejpam-86	685	10	ec′(ip))(ip	ec′(ip))(ip	PROPN
ejpam-86	685	11	⊗	⊗	PROPN
ejpam-86	685	12	∂2l(θ	∂2l(θ	NOUN
ejpam-86	685	13	)	)	PUNCT
ejpam-86	685	14	∂σ′∂σ	∂σ′∂σ	PUNCT
ejpam-86	686	1	⊗	⊗	PROPN
ejpam-86	686	2	ip)(v	ip)(v	PROPN
ejpam-86	686	3	ec(ip)⊗	ec(ip)⊗	PROPN
ejpam-86	686	4	ip2	ip2	PROPN
ejpam-86	686	5	)	)	PUNCT
ejpam-86	686	6	)	)	PUNCT
ejpam-86	686	7	(	(	PUNCT
ejpam-86	686	8	8.17	8.17	NUM
ejpam-86	686	9	)	)	PUNCT
ejpam-86	686	10	∂l(θ	∂l(θ	PROPN
ejpam-86	686	11	)	)	PUNCT
ejpam-86	687	1	∂β	∂β	PROPN
ejpam-86	687	2	=	=	PUNCT
ejpam-86	688	1	np	np	PROPN
ejpam-86	688	2	2β2	2β2	NUM
ejpam-86	688	3	ψ(1	ψ(1	VERB
ejpam-86	688	4	+	+	PUNCT
ejpam-86	688	5	p	p	NOUN
ejpam-86	688	6	2β	2β	NOUN
ejpam-86	688	7	)	)	PUNCT
ejpam-86	689	1	+	+	CCONJ
ejpam-86	689	2	np	np	PRON
ejpam-86	689	3	2β2	2β2	NUM
ejpam-86	689	4	log	log	NOUN
ejpam-86	689	5	2−	2−	NUM
ejpam-86	689	6	1	1	NUM
ejpam-86	689	7	2	2	NUM
ejpam-86	689	8	n∑	n∑	NOUN
ejpam-86	689	9	i=1	i=1	PROPN
ejpam-86	689	10	tβi	tβi	NOUN
ejpam-86	689	11	log	log	NOUN
ejpam-86	689	12	ti	ti	NOUN
ejpam-86	689	13	(	(	PUNCT
ejpam-86	689	14	8.18	8.18	NUM
ejpam-86	689	15	)	)	PUNCT
ejpam-86	689	16	where	where	SCONJ
ejpam-86	689	17	ψ	ψ	X
ejpam-86	689	18	(	(	PUNCT
ejpam-86	689	19	·	·	PUNCT
ejpam-86	689	20	)	)	PUNCT
ejpam-86	690	1	=	=	PUNCT
ejpam-86	691	1	d	d	X
ejpam-86	691	2	log	log	NOUN
ejpam-86	691	3	γ(x)/dx	γ(x)/dx	INTJ
ejpam-86	691	4	is	be	AUX
ejpam-86	691	5	called	call	VERB
ejpam-86	691	6	digamma	digamma	PROPN
ejpam-86	691	7	function	function	NOUN
ejpam-86	691	8	abramowitz	abramowitz	NOUN
ejpam-86	691	9	and	and	CCONJ
ejpam-86	691	10	stegun	stegun	NOUN
ejpam-86	691	11	(	(	PUNCT
ejpam-86	691	12	[	[	X
ejpam-86	691	13	1	1	NUM
ejpam-86	691	14	]	]	NUM
ejpam-86	691	15	)	)	PUNCT
ejpam-86	691	16	,	,	PUNCT
ejpam-86	691	17	or	or	CCONJ
ejpam-86	691	18	psi	psi	NOUN
ejpam-86	691	19	function	function	NOUN
ejpam-86	691	20	.	.	PUNCT
ejpam-86	692	1	∂2l(θ	∂2l(θ	NOUN
ejpam-86	692	2	)	)	PUNCT
ejpam-86	692	3	∂β2	∂β2	ADJ
ejpam-86	692	4	=	=	PUNCT
ejpam-86	692	5	−np	−np	X
ejpam-86	692	6	β3ψ(1	β3ψ(1	PUNCT
ejpam-86	692	7	+	+	PUNCT
ejpam-86	692	8	p	p	PRON
ejpam-86	692	9	2β	2β	NOUN
ejpam-86	692	10	)	)	PUNCT
ejpam-86	692	11	−	−	PROPN
ejpam-86	692	12	np2	np2	NOUN
ejpam-86	692	13	4β4ψ	4β4ψ	NOUN
ejpam-86	692	14	′(1	′(1	PROPN
ejpam-86	692	15	+	+	CCONJ
ejpam-86	692	16	p	p	NOUN
ejpam-86	692	17	2β	2β	NOUN
ejpam-86	692	18	)	)	PUNCT
ejpam-86	692	19	−	−	PROPN
ejpam-86	693	1	np	np	PRON
ejpam-86	693	2	β3	β3	INTJ
ejpam-86	693	3	log	log	VERB
ejpam-86	693	4	2−	2−	NUM
ejpam-86	693	5	1	1	NUM
ejpam-86	693	6	2	2	NUM
ejpam-86	693	7	n∑	n∑	NOUN
ejpam-86	693	8	i=1	i=1	PROPN
ejpam-86	693	9	tβi	tβi	NOUN
ejpam-86	693	10	log2	log2	PROPN
ejpam-86	693	11	ti	ti	PROPN
ejpam-86	693	12	(	(	PUNCT
ejpam-86	693	13	8.19	8.19	NUM
ejpam-86	693	14	)	)	PUNCT
ejpam-86	693	15	where	where	SCONJ
ejpam-86	693	16	ψ′	ψ′	PROPN
ejpam-86	693	17	(	(	PUNCT
ejpam-86	693	18	·	·	PUNCT
ejpam-86	693	19	)	)	PUNCT
ejpam-86	693	20	=	=	SYM
ejpam-86	693	21	d2	d2	PROPN
ejpam-86	693	22	log	log	PROPN
ejpam-86	693	23	γ(x)/dx2	γ(x)/dx2	PROPN
ejpam-86	693	24	is	be	AUX
ejpam-86	693	25	called	call	VERB
ejpam-86	693	26	trigamma	trigamma	PROPN
ejpam-86	693	27	function	function	PROPN
ejpam-86	693	28	.	.	PUNCT
ejpam-86	694	1	∂2l(θ	∂2l(θ	NOUN
ejpam-86	694	2	)	)	PUNCT
ejpam-86	694	3	∂β∂b	∂β∂b	PROPN
ejpam-86	694	4	=	=	SYM
ejpam-86	694	5	∂	∂	NUM
ejpam-86	694	6	∂β	∂β	PROPN
ejpam-86	694	7	(	(	PUNCT
ejpam-86	694	8	∂l(θ	∂l(θ	PROPN
ejpam-86	694	9	)	)	PUNCT
ejpam-86	694	10	∂b	∂b	PROPN
ejpam-86	694	11	)	)	PUNCT
ejpam-86	695	1	=	=	PUNCT
ejpam-86	695	2	n∑	n∑	NOUN
ejpam-86	695	3	i=1	i=1	PRON
ejpam-86	696	1	tβ−1	tβ−1	INTJ
ejpam-86	696	2	i	i	PRON
ejpam-86	696	3	v	v	ADP
ejpam-86	696	4	ec(σ−1ε(i)x′(i	ec(σ−1ε(i)x′(i	NUM
ejpam-86	696	5	)	)	PUNCT
ejpam-86	696	6	)	)	PUNCT
ejpam-86	697	1	+	+	CCONJ
ejpam-86	697	2	β	β	AUX
ejpam-86	697	3	n∑	n∑	X
ejpam-86	697	4	i=1	i=1	X
ejpam-86	698	1	tβ−1	tβ−1	VERB
ejpam-86	698	2	i	i	PRON
ejpam-86	698	3	log(ti)v	log(ti)v	VERB
ejpam-86	698	4	ec(σ−1ε(i)x′(i	ec(σ−1ε(i)x′(i	NOUN
ejpam-86	698	5	)	)	PUNCT
ejpam-86	698	6	)	)	PUNCT
ejpam-86	699	1	(	(	PUNCT
ejpam-86	699	2	8.20	8.20	NUM
ejpam-86	699	3	)	)	PUNCT
ejpam-86	699	4	∂2l(θ	∂2l(θ	NOUN
ejpam-86	699	5	)	)	PUNCT
ejpam-86	699	6	∂β∂v	∂β∂v	PROPN
ejpam-86	699	7	ec(σ	ec(σ	PUNCT
ejpam-86	699	8	)	)	PUNCT
ejpam-86	699	9	=	=	SYM
ejpam-86	699	10	∂	∂	NUM
ejpam-86	699	11	∂β	∂β	PROPN
ejpam-86	699	12	(	(	PUNCT
ejpam-86	699	13	∂l(θ	∂l(θ	PROPN
ejpam-86	699	14	)	)	PUNCT
ejpam-86	699	15	∂v	∂v	PROPN
ejpam-86	699	16	ec(σ	ec(σ	PUNCT
ejpam-86	699	17	)	)	PUNCT
ejpam-86	699	18	)	)	PUNCT
ejpam-86	700	1	=	=	SYM
ejpam-86	700	2	1	1	NUM
ejpam-86	700	3	2	2	NUM
ejpam-86	700	4	n∑	n∑	NOUN
ejpam-86	700	5	i=1	i=1	PRON
ejpam-86	701	1	tβ−1	tβ−1	NOUN
ejpam-86	702	1	i	i	PRON
ejpam-86	702	2	v	v	VERB
ejpam-86	702	3	ec(σ−1ε(i)ε	ec(σ−1ε(i)ε	NOUN
ejpam-86	702	4	′−1	′−1	X
ejpam-86	702	5	(	(	PUNCT
ejpam-86	702	6	i	i	NOUN
ejpam-86	702	7	)	)	PUNCT
ejpam-86	702	8	)	)	PUNCT
ejpam-86	703	1	+	+	VERB
ejpam-86	703	2	β	β	X
ejpam-86	703	3	2	2	NUM
ejpam-86	703	4	n∑	n∑	NOUN
ejpam-86	703	5	i=1	i=1	PRON
ejpam-86	703	6	tβ−1	tβ−1	VERB
ejpam-86	703	7	i	i	PRON
ejpam-86	703	8	log(ti)v	log(ti)v	VERB
ejpam-86	703	9	ec(σ−1ε(i)ε	ec(σ−1ε(i)ε	NOUN
ejpam-86	703	10	′−1	′−1	X
ejpam-86	703	11	(	(	PUNCT
ejpam-86	703	12	i	i	NOUN
ejpam-86	703	13	)	)	PUNCT
ejpam-86	703	14	)	)	PUNCT
ejpam-86	703	15	.	.	PUNCT
ejpam-86	704	1	(	(	PUNCT
ejpam-86	704	2	8.21	8.21	NUM
ejpam-86	704	3	)	)	PUNCT
ejpam-86	704	4	∂2l(θ	∂2l(θ	NOUN
ejpam-86	704	5	)	)	PUNCT
ejpam-86	705	1	∂v	∂v	PROPN
ejpam-86	705	2	ec′(σ)∂b	ec′(σ)∂b	PROPN
ejpam-86	705	3	=	=	SYM
ejpam-86	705	4	∂	∂	NUM
ejpam-86	705	5	∂∂v	∂∂v	PROPN
ejpam-86	705	6	ec′(σ	ec′(σ	PROPN
ejpam-86	705	7	)	)	PUNCT
ejpam-86	705	8	(	(	PUNCT
ejpam-86	705	9	∂l(θ	∂l(θ	PROPN
ejpam-86	705	10	)	)	PUNCT
ejpam-86	705	11	∂b	∂b	PROPN
ejpam-86	705	12	)	)	PUNCT
ejpam-86	706	1	=	=	PUNCT
ejpam-86	707	1	β	β	X
ejpam-86	707	2	n∑	n∑	NOUN
ejpam-86	707	3	i=1	i=1	PROPN
ejpam-86	708	1	∂v	∂v	PROPN
ejpam-86	708	2	ec(tβ−1	ec(tβ−1	VERB
ejpam-86	708	3	i	i	PRON
ejpam-86	708	4	σ−1ε(i)x	σ−1ε(i)x	NUM
ejpam-86	708	5	′	′	NUM
ejpam-86	709	1	(	(	PUNCT
ejpam-86	709	2	i	i	NOUN
ejpam-86	709	3	)	)	PUNCT
ejpam-86	709	4	)	)	PUNCT
ejpam-86	710	1	∂v	∂v	PROPN
ejpam-86	710	2	ec′(σ	ec′(σ	PROPN
ejpam-86	710	3	)	)	PUNCT
ejpam-86	710	4	=	=	PUNCT
ejpam-86	711	1	β	β	X
ejpam-86	711	2	n∑	n∑	X
ejpam-86	711	3	i=1	i=1	PROPN
ejpam-86	711	4	(	(	PUNCT
ejpam-86	711	5	ipq	ipq	NOUN
ejpam-86	711	6	⊗	⊗	PROPN
ejpam-86	711	7	v	v	PROPN
ejpam-86	711	8	ec′(ip))(iq	ec′(ip))(iq	PROPN
ejpam-86	711	9	⊗ni	⊗ni	NUM
ejpam-86	711	10	⊗	⊗	PROPN
ejpam-86	711	11	ip)(v	ip)(v	PROPN
ejpam-86	711	12	ec(iq)⊗	ec(iq)⊗	PROPN
ejpam-86	711	13	ip2	ip2	PROPN
ejpam-86	711	14	)	)	PUNCT
ejpam-86	711	15	(	(	PUNCT
ejpam-86	711	16	8.22	8.22	NUM
ejpam-86	711	17	)	)	PUNCT
ejpam-86	712	1	where	where	SCONJ
ejpam-86	712	2	ni	ni	PROPN
ejpam-86	712	3	=	=	PROPN
ejpam-86	712	4	∂(tβ−1	∂(tβ−1	PROPN
ejpam-86	713	1	i	i	PRON
ejpam-86	713	2	σ−1ε(i)x	σ−1ε(i)x	VERB
ejpam-86	713	3	′	′	NUM
ejpam-86	713	4	(	(	PUNCT
ejpam-86	713	5	i	i	NOUN
ejpam-86	713	6	)	)	PUNCT
ejpam-86	713	7	)	)	PUNCT
ejpam-86	714	1	∂σ	∂σ	PROPN
ejpam-86	715	1	=	=	PUNCT
ejpam-86	715	2	tβ−1	tβ−1	NOUN
ejpam-86	715	3	i	i	PRON
ejpam-86	715	4	∂σ−1	∂σ−1	VERB
ejpam-86	715	5	∂σ	∂σ	PROPN
ejpam-86	715	6	(	(	PUNCT
ejpam-86	715	7	ε(i)x′(i	ε(i)x′(i	PROPN
ejpam-86	715	8	)	)	PUNCT
ejpam-86	715	9	⊗	⊗	PROPN
ejpam-86	715	10	ip	ip	NOUN
ejpam-86	715	11	)	)	PUNCT
ejpam-86	716	1	+	+	CCONJ
ejpam-86	716	2	(	(	PUNCT
ejpam-86	716	3	σ−1ε(i)x′(i))⊗	σ−1ε(i)x′(i))⊗	PROPN
ejpam-86	716	4	∂tβ−1	∂tβ−1	NUM
ejpam-86	716	5	i	i	PRON
ejpam-86	716	6	∂σ	∂σ	PROPN
ejpam-86	717	1	=	=	PUNCT
ejpam-86	717	2	−tβ−1	−tβ−1	PUNCT
ejpam-86	717	3	i	i	PRON
ejpam-86	717	4	v	v	VERB
ejpam-86	717	5	ec(σ−1)v	ec(σ−1)v	ADJ
ejpam-86	717	6	ec′−1)(ε(i)x′(i	ec′−1)(ε(i)x′(i	NOUN
ejpam-86	717	7	)	)	PUNCT
ejpam-86	718	1	⊗	⊗	NUM
ejpam-86	718	2	ip	ip	NOUN
ejpam-86	718	3	)	)	PUNCT
ejpam-86	718	4	−(β	−(β	PROPN
ejpam-86	718	5	−	−	PROPN
ejpam-86	719	1	1)tβ−2	1)tβ−2	NUM
ejpam-86	719	2	i	i	INTJ
ejpam-86	719	3	(	(	PUNCT
ejpam-86	719	4	σ−1ε(i)x	σ−1ε(i)x	PROPN
ejpam-86	719	5	′−1	′−1	X
ejpam-86	719	6	(	(	PUNCT
ejpam-86	719	7	i	i	NOUN
ejpam-86	719	8	)	)	PUNCT
ejpam-86	719	9	ε(i)ε	ε(i)ε	PROPN
ejpam-86	719	10	′−1	′−1	PUNCT
ejpam-86	719	11	(	(	PUNCT
ejpam-86	719	12	i	i	NOUN
ejpam-86	719	13	)	)	PUNCT
ejpam-86	719	14	)	)	PUNCT
ejpam-86	720	1	=	=	PUNCT
ejpam-86	720	2	−tβ−1	−tβ−1	PUNCT
ejpam-86	720	3	i	i	PRON
ejpam-86	720	4	v	v	VERB
ejpam-86	720	5	ec(σ−1)v	ec(σ−1)v	ADJ
ejpam-86	720	6	ec′−1ε(i)x′(i	ec′−1ε(i)x′(i	NOUN
ejpam-86	720	7	)	)	PUNCT
ejpam-86	720	8	)	)	PUNCT
ejpam-86	721	1	−(β	−(β	VERB
ejpam-86	721	2	−	−	PROPN
ejpam-86	722	1	1)tβ−2	1)tβ−2	NUM
ejpam-86	722	2	i	i	INTJ
ejpam-86	722	3	(	(	PUNCT
ejpam-86	722	4	σ−1ε(i)x	σ−1ε(i)x	PROPN
ejpam-86	722	5	′−1	′−1	X
ejpam-86	722	6	(	(	PUNCT
ejpam-86	722	7	i	i	NOUN
ejpam-86	722	8	)	)	PUNCT
ejpam-86	722	9	ε(i)ε	ε(i)ε	PROPN
ejpam-86	722	10	′−1	′−1	PUNCT
ejpam-86	722	11	(	(	PUNCT
ejpam-86	722	12	i	i	NOUN
ejpam-86	722	13	)	)	PUNCT
ejpam-86	722	14	)	)	PUNCT
ejpam-86	722	15	references	reference	VERB
ejpam-86	722	16	35	35	NUM
ejpam-86	722	17	matrix	matrix	NOUN
ejpam-86	722	18	calculus	calculus	NOUN
ejpam-86	722	19	for	for	ADP
ejpam-86	722	20	type	type	NOUN
ejpam-86	722	21	ii	ii	PROPN
ejpam-86	722	22	mvper	mvper	PROPN
ejpam-86	722	23	model	model	NOUN
ejpam-86	722	24	∂e	∂e	PROPN
ejpam-86	722	25	∂b	∂b	PROPN
ejpam-86	722	26	=	=	SYM
ejpam-86	722	27	−∂(xb	−∂(xb	NOUN
ejpam-86	722	28	)	)	PUNCT
ejpam-86	722	29	∂b	∂b	NOUN
ejpam-86	722	30	=	=	SYM
ejpam-86	722	31	−v	−v	NOUN
ejpam-86	722	32	ec(x	ec(x	NOUN
ejpam-86	722	33	′)v	′)v	NOUN
ejpam-86	722	34	ec′(ip	ec′(ip	PROPN
ejpam-86	722	35	)	)	PUNCT
ejpam-86	722	36	.	.	PUNCT
ejpam-86	723	1	∂e′	∂e′	NOUN
ejpam-86	723	2	∂b	∂b	PROPN
ejpam-86	723	3	=	=	PROPN
ejpam-86	723	4	−∂(b′x	−∂(b′x	NOUN
ejpam-86	723	5	′	′	NUM
ejpam-86	723	6	)	)	PUNCT
ejpam-86	723	7	∂b	∂b	PROPN
ejpam-86	723	8	=	=	SYM
ejpam-86	723	9	−i(q	−i(q	PROPN
ejpam-86	723	10	,	,	PUNCT
ejpam-86	723	11	p)(x	p)(x	NOUN
ejpam-86	723	12	′	′	NUM
ejpam-86	723	13	⊗	⊗	NUM
ejpam-86	723	14	ip	ip	NOUN
ejpam-86	723	15	)	)	PUNCT
ejpam-86	723	16	.	.	PUNCT
ejpam-86	724	1	(	(	PUNCT
ejpam-86	724	2	σ−1	σ−1	PROPN
ejpam-86	724	3	⊗	⊗	PROPN
ejpam-86	724	4	iq)(e′−1	iq)(e′−1	PROPN
ejpam-86	724	5	⊗	⊗	PROPN
ejpam-86	724	6	iq)∂e	iq)∂e	PROPN
ejpam-86	724	7	∂b	∂b	PROPN
ejpam-86	724	8	=	=	SYM
ejpam-86	724	9	−(σ−1e′−1	−(σ−1e′−1	NOUN
ejpam-86	725	1	⊗	⊗	PROPN
ejpam-86	725	2	iq)v	iq)v	PROPN
ejpam-86	725	3	ec(x	ec(x	NOUN
ejpam-86	725	4	′)v	′)v	NOUN
ejpam-86	725	5	ec′(ip	ec′(ip	PROPN
ejpam-86	725	6	)	)	PUNCT
ejpam-86	726	1	=	=	PUNCT
ejpam-86	726	2	−v	−v	PROPN
ejpam-86	726	3	ec(x	ec(x	NUM
ejpam-86	726	4	′−1eς−1)v	′−1eς−1)v	X
ejpam-86	726	5	ec′(ip	ec′(ip	PROPN
ejpam-86	726	6	)	)	PUNCT
ejpam-86	726	7	.	.	PUNCT
ejpam-86	727	1	(	(	PUNCT
ejpam-86	727	2	σ−1	σ−1	PROPN
ejpam-86	727	3	⊗	⊗	PROPN
ejpam-86	727	4	iq	iq	PROPN
ejpam-86	727	5	)	)	PUNCT
ejpam-86	727	6	∂e′	∂e′	NOUN
ejpam-86	727	7	∂b	∂b	PROPN
ejpam-86	727	8	(	(	PUNCT
ejpam-86	727	9	e	e	PROPN
ejpam-86	727	10	⊗	⊗	PROPN
ejpam-86	727	11	ip	ip	NOUN
ejpam-86	727	12	)	)	PUNCT
ejpam-86	727	13	=	=	SYM
ejpam-86	728	1	−(σ−1	−(σ−1	PROPN
ejpam-86	728	2	⊗	⊗	PROPN
ejpam-86	728	3	iq)i(q	iq)i(q	PROPN
ejpam-86	728	4	,	,	PUNCT
ejpam-86	728	5	p)(x	p)(x	NOUN
ejpam-86	728	6	′−1e	′−1e	PROPN
ejpam-86	728	7	⊗	⊗	NUM
ejpam-86	728	8	ip	ip	NOUN
ejpam-86	728	9	)	)	PUNCT
ejpam-86	728	10	∂l(θ	∂l(θ	PROPN
ejpam-86	728	11	)	)	PUNCT
ejpam-86	728	12	∂b	∂b	NOUN
ejpam-86	728	13	=	=	SYM
ejpam-86	728	14	−β	−β	PROPN
ejpam-86	728	15	2	2	NUM
ejpam-86	728	16	(	(	PUNCT
ejpam-86	728	17	tr(σ−1e′−1e))β−1(ip	tr(σ−1e′−1e))β−1(ip	NOUN
ejpam-86	728	18	∗	∗	NOUN
ejpam-86	728	19	∂(σ−1e′−1e	∂(σ−1e′−1e	NOUN
ejpam-86	728	20	)	)	PUNCT
ejpam-86	728	21	∂b	∂b	PROPN
ejpam-86	728	22	)	)	PUNCT
ejpam-86	729	1	=	=	SYM
ejpam-86	729	2	−β	−β	NOUN
ejpam-86	729	3	2	2	NUM
ejpam-86	729	4	(	(	PUNCT
ejpam-86	729	5	tr(σ−1e′−1e))β−1(ip	tr(σ−1e′−1e))β−1(ip	ADJ
ejpam-86	729	6	∗	∗	NOUN
ejpam-86	729	7	(	(	PUNCT
ejpam-86	729	8	σ−1	σ−1	PROPN
ejpam-86	729	9	⊗	⊗	PROPN
ejpam-86	729	10	iq	iq	PROPN
ejpam-86	729	11	)	)	PUNCT
ejpam-86	729	12	∂(e′−1e	∂(e′−1e	PROPN
ejpam-86	729	13	)	)	PUNCT
ejpam-86	729	14	∂b	∂b	PROPN
ejpam-86	729	15	)	)	PUNCT
ejpam-86	730	1	=	=	SYM
ejpam-86	730	2	−β	−β	NOUN
ejpam-86	730	3	2	2	NUM
ejpam-86	730	4	(	(	PUNCT
ejpam-86	730	5	tr(σ−1e′−1e))β−1(ip	tr(σ−1e′−1e))β−1(ip	ADJ
ejpam-86	730	6	∗	∗	NOUN
ejpam-86	730	7	(	(	PUNCT
ejpam-86	730	8	σ−1	σ−1	PROPN
ejpam-86	730	9	⊗	⊗	PROPN
ejpam-86	730	10	iq	iq	PROPN
ejpam-86	730	11	)	)	PUNCT
ejpam-86	730	12	(	(	PUNCT
ejpam-86	730	13	∂e′	∂e′	NOUN
ejpam-86	730	14	∂b	∂b	PROPN
ejpam-86	730	15	(	(	PUNCT
ejpam-86	730	16	φ−1e	φ−1e	PROPN
ejpam-86	730	17	⊗	⊗	NUM
ejpam-86	730	18	ip	ip	PROPN
ejpam-86	730	19	)	)	PUNCT
ejpam-86	730	20	+	+	CCONJ
ejpam-86	730	21	(	(	PUNCT
ejpam-86	730	22	e′−1	e′−1	PROPN
ejpam-86	730	23	⊗	⊗	PROPN
ejpam-86	730	24	iq)∂e	iq)∂e	PROPN
ejpam-86	730	25	∂b	∂b	PROPN
ejpam-86	730	26	)	)	PUNCT
ejpam-86	730	27	)	)	PUNCT
ejpam-86	731	1	=	=	NOUN
ejpam-86	732	1	β(tr(σ−1e′β−1x	β(tr(σ−1e′β−1x	X
ejpam-86	733	1	′−1eς−1	′−1eς−1	X
ejpam-86	733	2	(	(	PUNCT
ejpam-86	733	3	8.23	8.23	NUM
ejpam-86	733	4	)	)	PUNCT
ejpam-86	733	5	where	where	SCONJ
ejpam-86	733	6	∗	∗	NOUN
ejpam-86	733	7	is	be	AUX
ejpam-86	733	8	the	the	DET
ejpam-86	733	9	star	star	PROPN
ejpam-86	733	10	product	product	PROPN
ejpam-86	733	11	macrae	macrae	PROPN
ejpam-86	733	12	(	(	PUNCT
ejpam-86	733	13	[	[	X
ejpam-86	733	14	24	24	NUM
ejpam-86	733	15	]	]	PUNCT
ejpam-86	733	16	)	)	PUNCT
ejpam-86	733	17	and	and	CCONJ
ejpam-86	733	18	i(q	i(q	PROPN
ejpam-86	733	19	,	,	PUNCT
ejpam-86	733	20	p	p	NOUN
ejpam-86	733	21	)	)	PUNCT
ejpam-86	733	22	is	be	AUX
ejpam-86	733	23	the	the	DET
ejpam-86	733	24	permuted	permuted	ADJ
ejpam-86	733	25	identity	identity	NOUN
ejpam-86	733	26	macrae	macrae	NOUN
ejpam-86	733	27	(	(	PUNCT
ejpam-86	734	1	[	[	X
ejpam-86	734	2	24	24	NUM
ejpam-86	734	3	]	]	PUNCT
ejpam-86	734	4	)	)	PUNCT
ejpam-86	734	5	or	or	CCONJ
ejpam-86	734	6	commutation	commutation	NOUN
ejpam-86	734	7	matrix	matrix	NOUN
ejpam-86	734	8	magnus	magnus	PROPN
ejpam-86	734	9	and	and	CCONJ
ejpam-86	734	10	neudecker	neudecker	NOUN
ejpam-86	735	1	(	(	PUNCT
ejpam-86	735	2	[	[	X
ejpam-86	735	3	25	25	NUM
ejpam-86	735	4	]	]	NUM
ejpam-86	735	5	)	)	PUNCT
ejpam-86	735	6	.	.	PUNCT
ejpam-86	736	1	∂l(θ	∂l(θ	PROPN
ejpam-86	736	2	)	)	PUNCT
ejpam-86	736	3	∂σ	∂σ	PROPN
ejpam-86	737	1	=	=	NOUN
ejpam-86	737	2	−n	−n	ADJ
ejpam-86	737	3	2	2	NUM
ejpam-86	738	1	σ−1	σ−1	PRON
ejpam-86	738	2	−	−	NOUN
ejpam-86	738	3	β	β	X
ejpam-86	738	4	2	2	NUM
ejpam-86	738	5	(	(	PUNCT
ejpam-86	738	6	tr(σ−1e′−1e))β−1(ip	tr(σ−1e′−1e))β−1(ip	NOUN
ejpam-86	738	7	∗	∗	NOUN
ejpam-86	738	8	∂(σ−1e′−1e	∂(σ−1e′−1e	NOUN
ejpam-86	738	9	)	)	PUNCT
ejpam-86	738	10	∂σ	∂σ	PROPN
ejpam-86	738	11	)	)	PUNCT
ejpam-86	739	1	=	=	PUNCT
ejpam-86	740	1	−n	−n	ADJ
ejpam-86	740	2	2	2	NUM
ejpam-86	741	1	σ−1	σ−1	PRON
ejpam-86	741	2	−	−	NOUN
ejpam-86	741	3	β	β	X
ejpam-86	741	4	2	2	NUM
ejpam-86	741	5	(	(	PUNCT
ejpam-86	741	6	tr(σ−1e′−1e))β−1(ip	tr(σ−1e′−1e))β−1(ip	ADJ
ejpam-86	741	7	∗	∗	NOUN
ejpam-86	741	8	∂σ−1	∂σ−1	PROPN
ejpam-86	741	9	∂σ	∂σ	PROPN
ejpam-86	742	1	(	(	PUNCT
ejpam-86	742	2	e′−1e	e′−1e	PROPN
ejpam-86	742	3	⊗	⊗	PROPN
ejpam-86	742	4	ip	ip	PROPN
ejpam-86	742	5	)	)	PUNCT
ejpam-86	742	6	)	)	PUNCT
ejpam-86	743	1	=	=	PUNCT
ejpam-86	744	1	−n	−n	ADJ
ejpam-86	744	2	2	2	NUM
ejpam-86	745	1	σ−1	σ−1	PROPN
ejpam-86	745	2	+	+	ADJ
ejpam-86	745	3	β	β	X
ejpam-86	745	4	2	2	NUM
ejpam-86	745	5	(	(	PUNCT
ejpam-86	745	6	tr(σ−1e′−1e))β−1(ip	tr(σ−1e′−1e))β−1(ip	NOUN
ejpam-86	745	7	∗	∗	NOUN
ejpam-86	745	8	v	v	ADP
ejpam-86	745	9	ec(σ−1)v	ec(σ−1)v	X
ejpam-86	745	10	ec′−1)(e′−1e	ec′−1)(e′−1e	PROPN
ejpam-86	745	11	⊗	⊗	PROPN
ejpam-86	745	12	ip	ip	NOUN
ejpam-86	745	13	)	)	PUNCT
ejpam-86	745	14	)	)	PUNCT
ejpam-86	746	1	=	=	PUNCT
ejpam-86	747	1	−n	−n	ADJ
ejpam-86	747	2	2	2	NUM
ejpam-86	748	1	σ−1	σ−1	PROPN
ejpam-86	748	2	+	+	CCONJ
ejpam-86	748	3	β	β	X
ejpam-86	748	4	2	2	NUM
ejpam-86	748	5	(	(	PUNCT
ejpam-86	748	6	tr(σ−1e′−1e))β−1(ip	tr(σ−1e′−1e))β−1(ip	NOUN
ejpam-86	748	7	∗	∗	NOUN
ejpam-86	748	8	v	v	ADP
ejpam-86	748	9	ec(σ−1)v	ec(σ−1)v	PROPN
ejpam-86	748	10	ec′−1e′−1e	ec′−1e′−1e	NOUN
ejpam-86	748	11	)	)	PUNCT
ejpam-86	748	12	)	)	PUNCT
ejpam-86	749	1	=	=	PUNCT
ejpam-86	749	2	−n	−n	ADJ
ejpam-86	749	3	2	2	NUM
ejpam-86	750	1	σ−1	σ−1	PROPN
ejpam-86	750	2	+	+	CCONJ
ejpam-86	750	3	β	β	X
ejpam-86	750	4	2	2	NUM
ejpam-86	750	5	(	(	PUNCT
ejpam-86	750	6	tr(σ−1e′−1e))β−1σ−1e′−1eς−1	tr(σ−1e′−1e))β−1σ−1e′−1eς−1	PROPN
ejpam-86	750	7	.	.	PUNCT
ejpam-86	751	1	(	(	PUNCT
ejpam-86	751	2	8.24	8.24	NUM
ejpam-86	751	3	)	)	PUNCT
ejpam-86	751	4	∂l2(θ	∂l2(θ	PROPN
ejpam-86	751	5	)	)	PUNCT
ejpam-86	751	6	∂b′∂b′	∂b′∂b′	ADV
ejpam-86	751	7	=	=	SYM
ejpam-86	751	8	∂(β(tr(σ−1e′−1e))β−1σ−1e′−1x	∂(β(tr(σ−1e′−1e))β−1σ−1e′−1x	NOUN
ejpam-86	751	9	)	)	PUNCT
ejpam-86	751	10	∂b′	∂b′	NOUN
ejpam-86	751	11	=	=	SYM
ejpam-86	751	12	β(tr(σ−1e′−1e))β−1	β(tr(σ−1e′−1e))β−1	PROPN
ejpam-86	751	13	∂(σ−1e′−1x	∂(σ−1e′−1x	NOUN
ejpam-86	751	14	)	)	PUNCT
ejpam-86	751	15	∂b′	∂b′	NOUN
ejpam-86	752	1	+	+	ADV
ejpam-86	752	2	(	(	PUNCT
ejpam-86	752	3	σ−1e′−1x)⊗	σ−1e′−1x)⊗	X
ejpam-86	752	4	∂(β(tr(σ−1e′−1e))β−1	∂(β(tr(σ−1e′−1e))β−1	NUM
ejpam-86	752	5	)	)	PUNCT
ejpam-86	752	6	∂b′	∂b′	NOUN
ejpam-86	752	7	,	,	PUNCT
ejpam-86	752	8	(	(	PUNCT
ejpam-86	752	9	8.25	8.25	NUM
ejpam-86	752	10	)	)	PUNCT
ejpam-86	752	11	∂(σ−1e′−1x	∂(σ−1e′−1x	NOUN
ejpam-86	752	12	)	)	PUNCT
ejpam-86	752	13	∂b′	∂b′	NOUN
ejpam-86	752	14	=	=	SYM
ejpam-86	752	15	(	(	PUNCT
ejpam-86	752	16	σ−1	σ−1	PROPN
ejpam-86	752	17	⊗	⊗	PROPN
ejpam-86	752	18	ip)(∂e	ip)(∂e	PROPN
ejpam-86	752	19	∂b	∂b	PROPN
ejpam-86	752	20	)	)	PUNCT
ejpam-86	752	21	′−1x	′−1x	PROPN
ejpam-86	752	22	⊗	⊗	PROPN
ejpam-86	752	23	iq	iq	PROPN
ejpam-86	752	24	)	)	PUNCT
ejpam-86	752	25	=	=	PUNCT
ejpam-86	753	1	−(σ−1	−(σ−1	PROPN
ejpam-86	753	2	⊗	⊗	PROPN
ejpam-86	753	3	ip)v	ip)v	PROPN
ejpam-86	753	4	ec(ip)v	ec(ip)v	PROPN
ejpam-86	753	5	ec′(x	ec′(x	PROPN
ejpam-86	753	6	′−1x	′−1x	PROPN
ejpam-86	753	7	⊗	⊗	PROPN
ejpam-86	753	8	iq	iq	PROPN
ejpam-86	753	9	)	)	PUNCT
ejpam-86	754	1	=	=	PUNCT
ejpam-86	754	2	−v	−v	NOUN
ejpam-86	754	3	ec(σ−1)v	ec(σ−1)v	PROPN
ejpam-86	754	4	ec′(x	ec′(x	NOUN
ejpam-86	754	5	′−1x	′−1x	PROPN
ejpam-86	754	6	)	)	PUNCT
ejpam-86	754	7	.	.	PUNCT
ejpam-86	755	1	(	(	PUNCT
ejpam-86	755	2	8.26	8.26	NUM
ejpam-86	755	3	)	)	PUNCT
ejpam-86	755	4	references	reference	NOUN
ejpam-86	755	5	[	[	X
ejpam-86	755	6	1	1	NUM
ejpam-86	755	7	]	]	X
ejpam-86	755	8	milton	milton	PROPN
ejpam-86	755	9	abramowitz	abramowitz	PROPN
ejpam-86	755	10	and	and	CCONJ
ejpam-86	755	11	irene	irene	PROPN
ejpam-86	755	12	a.	a.	PROPN
ejpam-86	755	13	stegun	stegun	PROPN
ejpam-86	755	14	.	.	PUNCT
ejpam-86	756	1	handbook	handbook	NOUN
ejpam-86	756	2	of	of	ADP
ejpam-86	756	3	mathematical	mathematical	ADJ
ejpam-86	756	4	functions	function	NOUN
ejpam-86	756	5	,	,	PUNCT
ejpam-86	756	6	with	with	ADP
ejpam-86	756	7	formulas	formula	NOUN
ejpam-86	756	8	,	,	PUNCT
ejpam-86	756	9	graphs	graph	NOUN
ejpam-86	756	10	,	,	PUNCT
ejpam-86	756	11	and	and	CCONJ
ejpam-86	756	12	mathematical	mathematical	ADJ
ejpam-86	756	13	tables	table	NOUN
ejpam-86	756	14	.	.	PUNCT
ejpam-86	757	1	dover	dover	PROPN
ejpam-86	757	2	publications	publication	NOUN
ejpam-86	757	3	,	,	PUNCT
ejpam-86	757	4	new	new	PROPN
ejpam-86	757	5	york	york	PROPN
ejpam-86	757	6	,	,	PUNCT
ejpam-86	757	7	,	,	PUNCT
ejpam-86	757	8	1965	1965	NUM
ejpam-86	757	9	.	.	PUNCT
ejpam-86	758	1	[	[	X
ejpam-86	758	2	2	2	X
ejpam-86	758	3	]	]	X
ejpam-86	758	4	h.	h.	NOUN
ejpam-86	758	5	akaike	akaike	PROPN
ejpam-86	758	6	.	.	PUNCT
ejpam-86	759	1	information	information	NOUN
ejpam-86	759	2	theory	theory	NOUN
ejpam-86	759	3	and	and	CCONJ
ejpam-86	759	4	an	an	DET
ejpam-86	759	5	extension	extension	NOUN
ejpam-86	759	6	of	of	ADP
ejpam-86	759	7	the	the	DET
ejpam-86	759	8	maximum	maximum	ADJ
ejpam-86	759	9	likelihood	likelihood	NOUN
ejpam-86	759	10	principle	principle	NOUN
ejpam-86	759	11	.	.	PUNCT
ejpam-86	760	1	in	in	ADP
ejpam-86	760	2	second	second	ADJ
ejpam-86	760	3	international	international	ADJ
ejpam-86	760	4	symposium	symposium	NOUN
ejpam-86	760	5	on	on	ADP
ejpam-86	760	6	information	information	NOUN
ejpam-86	760	7	theory	theory	NOUN
ejpam-86	760	8	(	(	PUNCT
ejpam-86	760	9	tsahkadsor	tsahkadsor	NOUN
ejpam-86	760	10	,	,	PUNCT
ejpam-86	760	11	1971	1971	NUM
ejpam-86	760	12	)	)	PUNCT
ejpam-86	760	13	,	,	PUNCT
ejpam-86	760	14	pages	page	NOUN
ejpam-86	760	15	267–281	267–281	NUM
ejpam-86	760	16	.	.	PUNCT
ejpam-86	761	1	akadémiai	akadémiai	VERB
ejpam-86	761	2	kiadó	kiadó	NOUN
ejpam-86	761	3	,	,	PUNCT
ejpam-86	761	4	budapest	budapest	NOUN
ejpam-86	761	5	,	,	PUNCT
ejpam-86	761	6	1973	1973	NUM
ejpam-86	761	7	.	.	PUNCT
ejpam-86	762	1	references	reference	NOUN
ejpam-86	762	2	36	36	NUM
ejpam-86	762	3	[	[	X
ejpam-86	762	4	3	3	NUM
ejpam-86	762	5	]	]	X
ejpam-86	762	6	h.	h.	NOUN
ejpam-86	762	7	akaike	akaike	PROPN
ejpam-86	762	8	.	.	PUNCT
ejpam-86	763	1	likelihood	likelihood	NOUN
ejpam-86	763	2	of	of	ADP
ejpam-86	763	3	a	a	DET
ejpam-86	763	4	model	model	NOUN
ejpam-86	763	5	and	and	CCONJ
ejpam-86	763	6	information	information	NOUN
ejpam-86	763	7	criteria	criterion	NOUN
ejpam-86	763	8	.	.	PUNCT
ejpam-86	764	1	journal	journal	NOUN
ejpam-86	764	2	of	of	ADP
ejpam-86	764	3	econometrics	econometric	NOUN
ejpam-86	764	4	,	,	PUNCT
ejpam-86	764	5	16(1):3–14	16(1):3–14	NUM
ejpam-86	764	6	,	,	PUNCT
ejpam-86	764	7	1981	1981	NUM
ejpam-86	764	8	.	.	PUNCT
ejpam-86	765	1	[	[	X
ejpam-86	765	2	4	4	NUM
ejpam-86	765	3	]	]	X
ejpam-86	765	4	hirotugu	hirotugu	NOUN
ejpam-86	765	5	akaike	akaike	ADJ
ejpam-86	765	6	.	.	PUNCT
ejpam-86	766	1	a	a	DET
ejpam-86	766	2	new	new	ADJ
ejpam-86	766	3	look	look	NOUN
ejpam-86	766	4	at	at	ADP
ejpam-86	766	5	the	the	DET
ejpam-86	766	6	statistical	statistical	ADJ
ejpam-86	766	7	model	model	NOUN
ejpam-86	766	8	identification	identification	NOUN
ejpam-86	766	9	.	.	PUNCT
ejpam-86	767	1	ieee	ieee	PROPN
ejpam-86	767	2	trans	trans	PROPN
ejpam-86	767	3	.	.	PUNCT
ejpam-86	768	1	automatic	automatic	ADJ
ejpam-86	768	2	control	control	NOUN
ejpam-86	768	3	,	,	PUNCT
ejpam-86	768	4	ac-19:716–723	ac-19:716–723	PROPN
ejpam-86	768	5	,	,	PUNCT
ejpam-86	768	6	1974	1974	NUM
ejpam-86	768	7	.	.	PUNCT
ejpam-86	769	1	system	system	NOUN
ejpam-86	769	2	identification	identification	NOUN
ejpam-86	769	3	and	and	CCONJ
ejpam-86	769	4	time	time	NOUN
ejpam-86	769	5	-	-	PUNCT
ejpam-86	769	6	series	series	NOUN
ejpam-86	769	7	analysis	analysis	NOUN
ejpam-86	769	8	.	.	PUNCT
ejpam-86	770	1	[	[	X
ejpam-86	770	2	5	5	NUM
ejpam-86	770	3	]	]	PUNCT
ejpam-86	770	4	hirotugu	hirotugu	NOUN
ejpam-86	770	5	akaike	akaike	ADJ
ejpam-86	770	6	.	.	PUNCT
ejpam-86	771	1	factor	factor	NOUN
ejpam-86	771	2	analysis	analysis	NOUN
ejpam-86	771	3	and	and	CCONJ
ejpam-86	771	4	aic	aic	PROPN
ejpam-86	771	5	.	.	PROPN
ejpam-86	771	6	psychometrika	psychometrika	PROPN
ejpam-86	771	7	,	,	PUNCT
ejpam-86	771	8	52(3):317–332	52(3):317–332	PROPN
ejpam-86	771	9	,	,	PUNCT
ejpam-86	771	10	1987	1987	NUM
ejpam-86	771	11	.	.	PUNCT
ejpam-86	772	1	[	[	X
ejpam-86	772	2	6	6	NUM
ejpam-86	772	3	]	]	PUNCT
ejpam-86	772	4	george	george	PROPN
ejpam-86	772	5	e.	e.	PROPN
ejpam-86	772	6	p.	p.	PROPN
ejpam-86	772	7	box	box	PROPN
ejpam-86	772	8	and	and	CCONJ
ejpam-86	772	9	george	george	PROPN
ejpam-86	772	10	c.	c.	PROPN
ejpam-86	772	11	tiao	tiao	PROPN
ejpam-86	772	12	.	.	PUNCT
ejpam-86	773	1	bayesian	bayesian	NOUN
ejpam-86	773	2	inference	inference	NOUN
ejpam-86	773	3	in	in	ADP
ejpam-86	773	4	statistical	statistical	ADJ
ejpam-86	773	5	analysis	analysis	NOUN
ejpam-86	773	6	.	.	PUNCT
ejpam-86	774	1	addison	addison	PROPN
ejpam-86	774	2	-	-	PUNCT
ejpam-86	774	3	wesley	wesley	PROPN
ejpam-86	774	4	publishing	publishing	PROPN
ejpam-86	774	5	co.	co.	PROPN
ejpam-86	774	6	,	,	PUNCT
ejpam-86	774	7	reading	reading	NOUN
ejpam-86	774	8	,	,	PUNCT
ejpam-86	774	9	mass.-london	mass.-london	ADJ
ejpam-86	774	10	-	-	ADJ
ejpam-86	774	11	don	don	ADJ
ejpam-86	774	12	mills	mill	NOUN
ejpam-86	774	13	,	,	PUNCT
ejpam-86	774	14	ont	ont	PROPN
ejpam-86	774	15	.	.	PROPN
ejpam-86	774	16	,	,	PUNCT
ejpam-86	774	17	1973	1973	NUM
ejpam-86	774	18	.	.	PUNCT
ejpam-86	775	1	[	[	X
ejpam-86	775	2	7	7	X
ejpam-86	775	3	]	]	X
ejpam-86	775	4	h.	h.	NOUN
ejpam-86	775	5	bozdogan	bozdogan	PROPN
ejpam-86	775	6	.	.	PUNCT
ejpam-86	776	1	icomp	icomp	PROPN
ejpam-86	776	2	:	:	PUNCT
ejpam-86	776	3	a	a	DET
ejpam-86	776	4	new	new	ADJ
ejpam-86	776	5	model	model	NOUN
ejpam-86	776	6	-	-	PUNCT
ejpam-86	776	7	selection	selection	NOUN
ejpam-86	776	8	criterion	criterion	NOUN
ejpam-86	776	9	.	.	PUNCT
ejpam-86	777	1	in	in	ADP
ejpam-86	777	2	classification	classification	NOUN
ejpam-86	777	3	and	and	CCONJ
ejpam-86	777	4	related	related	ADJ
ejpam-86	777	5	methods	method	NOUN
ejpam-86	777	6	of	of	ADP
ejpam-86	777	7	data	datum	NOUN
ejpam-86	777	8	analysis	analysis	NOUN
ejpam-86	777	9	,	,	PUNCT
ejpam-86	777	10	pages	page	NOUN
ejpam-86	777	11	599–608	599–608	NUM
ejpam-86	777	12	,	,	PUNCT
ejpam-86	777	13	1988	1988	NUM
ejpam-86	777	14	.	.	PUNCT
ejpam-86	778	1	[	[	X
ejpam-86	778	2	8	8	NUM
ejpam-86	778	3	]	]	X
ejpam-86	778	4	h.	h.	NOUN
ejpam-86	778	5	bozdogan	bozdogan	PROPN
ejpam-86	778	6	.	.	PUNCT
ejpam-86	779	1	statistical	statistical	ADJ
ejpam-86	779	2	data	datum	NOUN
ejpam-86	779	3	mining	mining	NOUN
ejpam-86	779	4	and	and	CCONJ
ejpam-86	779	5	knowledge	knowledge	NOUN
ejpam-86	779	6	discovery	discovery	NOUN
ejpam-86	779	7	.	.	PUNCT
ejpam-86	780	1	chapman	chapman	PROPN
ejpam-86	780	2	&	&	CCONJ
ejpam-86	780	3	hall	hall	PROPN
ejpam-86	780	4	/	/	SYM
ejpam-86	780	5	crc	crc	PROPN
ejpam-86	780	6	,	,	PUNCT
ejpam-86	780	7	boca	boca	PROPN
ejpam-86	780	8	raton	raton	PROPN
ejpam-86	780	9	,	,	PUNCT
ejpam-86	780	10	2004	2004	NUM
ejpam-86	780	11	.	.	PUNCT
ejpam-86	781	1	[	[	X
ejpam-86	781	2	9	9	NUM
ejpam-86	781	3	]	]	PUNCT
ejpam-86	781	4	hamparsum	hamparsum	NOUN
ejpam-86	781	5	bozdogan	bozdogan	NOUN
ejpam-86	781	6	.	.	PUNCT
ejpam-86	782	1	on	on	ADP
ejpam-86	782	2	the	the	DET
ejpam-86	782	3	information	information	NOUN
ejpam-86	782	4	-	-	PUNCT
ejpam-86	782	5	based	base	VERB
ejpam-86	782	6	measure	measure	NOUN
ejpam-86	782	7	of	of	ADP
ejpam-86	782	8	covariance	covariance	NOUN
ejpam-86	782	9	complexity	complexity	NOUN
ejpam-86	782	10	and	and	CCONJ
ejpam-86	782	11	its	its	PRON
ejpam-86	782	12	application	application	NOUN
ejpam-86	782	13	to	to	ADP
ejpam-86	782	14	the	the	DET
ejpam-86	782	15	evaluation	evaluation	NOUN
ejpam-86	782	16	of	of	ADP
ejpam-86	782	17	multivariate	multivariate	NOUN
ejpam-86	782	18	linear	linear	NOUN
ejpam-86	782	19	models	model	NOUN
ejpam-86	782	20	.	.	PUNCT
ejpam-86	783	1	communications	communication	NOUN
ejpam-86	783	2	in	in	ADP
ejpam-86	783	3	statistics	statistic	NOUN
ejpam-86	783	4	,	,	PUNCT
ejpam-86	783	5	part	part	NOUN
ejpam-86	783	6	a	a	DET
ejpam-86	783	7	–	–	PUNCT
ejpam-86	783	8	theory	theory	NOUN
ejpam-86	783	9	and	and	CCONJ
ejpam-86	783	10	methods	method	NOUN
ejpam-86	783	11	[	[	X
ejpam-86	783	12	split	split	VERB
ejpam-86	783	13	from	from	ADP
ejpam-86	783	14	:	:	PUNCT
ejpam-86	783	15	@j(commstat	@j(commstat	X
ejpam-86	783	16	)	)	PUNCT
ejpam-86	783	17	]	]	X
ejpam-86	783	18	,	,	PUNCT
ejpam-86	783	19	19:221–278	19:221–278	NOUN
ejpam-86	783	20	,	,	PUNCT
ejpam-86	783	21	1990	1990	NUM
ejpam-86	783	22	.	.	PUNCT
ejpam-86	784	1	[	[	X
ejpam-86	784	2	10	10	NUM
ejpam-86	784	3	]	]	X
ejpam-86	784	4	hamparsum	hamparsum	NOUN
ejpam-86	784	5	bozdogan	bozdogan	NOUN
ejpam-86	784	6	.	.	PUNCT
ejpam-86	785	1	choosing	choose	VERB
ejpam-86	785	2	the	the	DET
ejpam-86	785	3	number	number	NOUN
ejpam-86	785	4	of	of	ADP
ejpam-86	785	5	component	component	NOUN
ejpam-86	785	6	clusters	cluster	NOUN
ejpam-86	785	7	in	in	ADP
ejpam-86	785	8	the	the	DET
ejpam-86	785	9	mixture	mixture	NOUN
ejpam-86	785	10	-	-	PUNCT
ejpam-86	785	11	model	model	NOUN
ejpam-86	785	12	using	use	VERB
ejpam-86	785	13	a	a	DET
ejpam-86	785	14	new	new	ADJ
ejpam-86	785	15	informational	informational	ADJ
ejpam-86	785	16	complexity	complexity	NOUN
ejpam-86	785	17	criterion	criterion	NOUN
ejpam-86	785	18	of	of	ADP
ejpam-86	785	19	the	the	DET
ejpam-86	785	20	inverse	inverse	NOUN
ejpam-86	785	21	-	-	PUNCT
ejpam-86	785	22	fisher	fisher	PROPN
ejpam-86	785	23	information	information	NOUN
ejpam-86	785	24	matrix	matrix	NOUN
ejpam-86	785	25	.	.	PUNCT
ejpam-86	786	1	in	in	ADP
ejpam-86	786	2	information	information	NOUN
ejpam-86	786	3	and	and	CCONJ
ejpam-86	786	4	classification	classification	NOUN
ejpam-86	786	5	.	.	PUNCT
ejpam-86	787	1	concepts	concept	NOUN
ejpam-86	787	2	,	,	PUNCT
ejpam-86	787	3	methods	method	NOUN
ejpam-86	787	4	and	and	CCONJ
ejpam-86	787	5	applications	application	NOUN
ejpam-86	787	6	.	.	PUNCT
ejpam-86	788	1	proceedings	proceeding	NOUN
ejpam-86	788	2	of	of	ADP
ejpam-86	788	3	the	the	DET
ejpam-86	788	4	16th	16th	ADJ
ejpam-86	788	5	annual	annual	ADJ
ejpam-86	788	6	conference	conference	NOUN
ejpam-86	788	7	of	of	ADP
ejpam-86	788	8	the	the	DET
ejpam-86	788	9	gesellschaft	gesellschaft	NOUN
ejpam-86	788	10	fr	fr	PROPN
ejpam-86	788	11	klassifikation	klassifikation	PROPN
ejpam-86	788	12	e.	e.	PROPN
ejpam-86	788	13	v.	v.	PROPN
ejpam-86	788	14	,	,	PUNCT
ejpam-86	788	15	pages	page	NOUN
ejpam-86	788	16	40–54	40–54	NUM
ejpam-86	788	17	,	,	PUNCT
ejpam-86	788	18	1993	1993	NUM
ejpam-86	788	19	.	.	PUNCT
ejpam-86	789	1	[	[	X
ejpam-86	789	2	11	11	NUM
ejpam-86	789	3	]	]	PUNCT
ejpam-86	789	4	hamparsum	hamparsum	NOUN
ejpam-86	789	5	bozdogan	bozdogan	NOUN
ejpam-86	789	6	.	.	PUNCT
ejpam-86	790	1	multivariate	multivariate	NOUN
ejpam-86	790	2	regression	regression	NOUN
ejpam-86	790	3	models	model	NOUN
ejpam-86	790	4	for	for	ADP
ejpam-86	790	5	nonnormal	nonnormal	ADJ
ejpam-86	790	6	data	datum	NOUN
ejpam-86	790	7	:	:	PUNCT
ejpam-86	790	8	a	a	DET
ejpam-86	790	9	new	new	ADJ
ejpam-86	790	10	model	model	NOUN
ejpam-86	790	11	selection	selection	NOUN
ejpam-86	790	12	approach	approach	NOUN
ejpam-86	790	13	.	.	PUNCT
ejpam-86	791	1	in	in	ADP
ejpam-86	791	2	52nd	52nd	ADJ
ejpam-86	791	3	session	session	NOUN
ejpam-86	791	4	of	of	ADP
ejpam-86	791	5	the	the	DET
ejpam-86	791	6	international	international	ADJ
ejpam-86	791	7	statistical	statistical	ADJ
ejpam-86	791	8	institute	institute	NOUN
ejpam-86	791	9	,	,	PUNCT
ejpam-86	791	10	helsinki	helsinki	PROPN
ejpam-86	791	11	,	,	PUNCT
ejpam-86	791	12	finland	finland	PROPN
ejpam-86	791	13	,	,	PUNCT
ejpam-86	791	14	1999	1999	NUM
ejpam-86	791	15	.	.	PUNCT
ejpam-86	792	1	[	[	X
ejpam-86	792	2	12	12	NUM
ejpam-86	792	3	]	]	PUNCT
ejpam-86	792	4	hamparsum	hamparsum	NOUN
ejpam-86	792	5	bozdogan	bozdogan	NOUN
ejpam-86	792	6	.	.	PUNCT
ejpam-86	793	1	akaike	akaike	ADP
ejpam-86	793	2	’s	’s	PART
ejpam-86	793	3	information	information	NOUN
ejpam-86	793	4	criterion	criterion	NOUN
ejpam-86	793	5	and	and	CCONJ
ejpam-86	793	6	recent	recent	ADJ
ejpam-86	793	7	developments	development	NOUN
ejpam-86	793	8	in	in	ADP
ejpam-86	793	9	information	information	NOUN
ejpam-86	793	10	complexity	complexity	NOUN
ejpam-86	793	11	.	.	PUNCT
ejpam-86	794	1	journal	journal	PROPN
ejpam-86	794	2	of	of	ADP
ejpam-86	794	3	mathematical	mathematical	ADJ
ejpam-86	794	4	psychology	psychology	NOUN
ejpam-86	794	5	,	,	PUNCT
ejpam-86	794	6	44(1):62–91	44(1):62–91	NUM
ejpam-86	794	7	,	,	PUNCT
ejpam-86	794	8	2000	2000	NUM
ejpam-86	794	9	.	.	PUNCT
ejpam-86	795	1	[	[	X
ejpam-86	795	2	13	13	NUM
ejpam-86	795	3	]	]	SYM
ejpam-86	795	4	hamparsum	hamparsum	NOUN
ejpam-86	795	5	bozdogan	bozdogan	NOUN
ejpam-86	795	6	and	and	CCONJ
ejpam-86	795	7	dominique	dominique	PROPN
ejpam-86	795	8	m.	m.	PROPN
ejpam-86	795	9	a.	a.	PROPN
ejpam-86	795	10	haughton	haughton	PROPN
ejpam-86	795	11	.	.	PUNCT
ejpam-86	796	1	informational	informational	ADJ
ejpam-86	796	2	complexity	complexity	NOUN
ejpam-86	796	3	criteria	criterion	NOUN
ejpam-86	796	4	for	for	ADP
ejpam-86	796	5	regression	regression	NOUN
ejpam-86	796	6	models	model	NOUN
ejpam-86	796	7	.	.	PUNCT
ejpam-86	797	1	computational	computational	ADJ
ejpam-86	797	2	statistics	statistic	NOUN
ejpam-86	797	3	and	and	CCONJ
ejpam-86	797	4	data	datum	NOUN
ejpam-86	797	5	analysis	analysis	NOUN
ejpam-86	797	6	,	,	PUNCT
ejpam-86	797	7	28:51–76	28:51–76	NUM
ejpam-86	797	8	,	,	PUNCT
ejpam-86	797	9	1998	1998	NUM
ejpam-86	797	10	.	.	PUNCT
ejpam-86	798	1	[	[	X
ejpam-86	798	2	14	14	NUM
ejpam-86	798	3	]	]	X
ejpam-86	798	4	hamparsum	hamparsum	NOUN
ejpam-86	798	5	bozdogan	bozdogan	NOUN
ejpam-86	798	6	and	and	CCONJ
ejpam-86	798	7	kazuo	kazuo	PROPN
ejpam-86	798	8	shigemasu	shigemasu	PROPN
ejpam-86	798	9	.	.	PUNCT
ejpam-86	799	1	bayesian	bayesian	NOUN
ejpam-86	799	2	factor	factor	NOUN
ejpam-86	799	3	analysis	analysis	NOUN
ejpam-86	799	4	model	model	NOUN
ejpam-86	799	5	and	and	CCONJ
ejpam-86	799	6	choosing	choose	VERB
ejpam-86	799	7	the	the	DET
ejpam-86	799	8	number	number	NOUN
ejpam-86	799	9	of	of	ADP
ejpam-86	799	10	factors	factor	NOUN
ejpam-86	799	11	using	use	VERB
ejpam-86	799	12	a	a	DET
ejpam-86	799	13	new	new	ADJ
ejpam-86	799	14	informational	informational	ADJ
ejpam-86	799	15	complexity	complexity	NOUN
ejpam-86	799	16	criterion	criterion	NOUN
ejpam-86	799	17	.	.	PUNCT
ejpam-86	800	1	in	in	ADP
ejpam-86	800	2	advances	advance	NOUN
ejpam-86	800	3	in	in	ADP
ejpam-86	800	4	data	datum	NOUN
ejpam-86	800	5	science	science	NOUN
ejpam-86	800	6	and	and	CCONJ
ejpam-86	800	7	classification	classification	NOUN
ejpam-86	800	8	.	.	PUNCT
ejpam-86	800	9	proceedings	proceeding	NOUN
ejpam-86	800	10	of	of	ADP
ejpam-86	800	11	the	the	DET
ejpam-86	800	12	6th	6th	ADJ
ejpam-86	800	13	conference	conference	NOUN
ejpam-86	800	14	of	of	ADP
ejpam-86	800	15	the	the	DET
ejpam-86	800	16	international	international	PROPN
ejpam-86	800	17	federation	federation	PROPN
ejpam-86	800	18	of	of	ADP
ejpam-86	800	19	classification	classification	NOUN
ejpam-86	800	20	societies	society	NOUN
ejpam-86	800	21	,	,	PUNCT
ejpam-86	800	22	pages	page	NOUN
ejpam-86	800	23	335–342	335–342	NUM
ejpam-86	800	24	,	,	PUNCT
ejpam-86	800	25	1998	1998	NUM
ejpam-86	800	26	.	.	PUNCT
ejpam-86	801	1	[	[	X
ejpam-86	801	2	15	15	NUM
ejpam-86	801	3	]	]	X
ejpam-86	801	4	s.	s.	PROPN
ejpam-86	801	5	chatterjee	chatterjee	PROPN
ejpam-86	801	6	and	and	CCONJ
ejpam-86	801	7	m.	m.	NOUN
ejpam-86	801	8	laudatto	laudatto	PROPN
ejpam-86	801	9	.	.	PUNCT
ejpam-86	802	1	genetic	genetic	ADJ
ejpam-86	802	2	algorithms	algorithm	NOUN
ejpam-86	802	3	in	in	ADP
ejpam-86	802	4	statistics	statistic	NOUN
ejpam-86	802	5	:	:	PUNCT
ejpam-86	802	6	procedures	procedure	NOUN
ejpam-86	802	7	and	and	CCONJ
ejpam-86	802	8	applications	application	NOUN
ejpam-86	802	9	.	.	PUNCT
ejpam-86	803	1	communications	communication	NOUN
ejpam-86	803	2	in	in	ADP
ejpam-86	803	3	statistics	statistic	NOUN
ejpam-86	803	4	-	-	PUNCT
ejpam-86	803	5	simulation	simulation	NOUN
ejpam-86	803	6	and	and	CCONJ
ejpam-86	803	7	computation	computation	NOUN
ejpam-86	803	8	,	,	PUNCT
ejpam-86	803	9	26(4):1617–1630	26(4):1617–1630	NUM
ejpam-86	803	10	,	,	PUNCT
ejpam-86	803	11	1997	1997	NUM
ejpam-86	803	12	.	.	PUNCT
ejpam-86	804	1	[	[	X
ejpam-86	804	2	16	16	NUM
ejpam-86	804	3	]	]	PUNCT
ejpam-86	804	4	m.	m.	NOUN
ejpam-86	804	5	h.	h.	PROPN
ejpam-86	804	6	van	van	PROPN
ejpam-86	804	7	emden	emden	PROPN
ejpam-86	804	8	.	.	PUNCT
ejpam-86	805	1	an	an	DET
ejpam-86	805	2	analysis	analysis	NOUN
ejpam-86	805	3	of	of	ADP
ejpam-86	805	4	complexity	complexity	NOUN
ejpam-86	805	5	.	.	PUNCT
ejpam-86	806	1	mathematical	mathematical	ADJ
ejpam-86	806	2	centre	centre	NOUN
ejpam-86	806	3	tracts	tract	NOUN
ejpam-86	806	4	.	.	PUNCT
ejpam-86	807	1	35	35	NUM
ejpam-86	807	2	.	.	PUNCT
ejpam-86	807	3	mathematisch	mathematisch	PROPN
ejpam-86	807	4	centrum	centrum	PROPN
ejpam-86	807	5	,	,	PUNCT
ejpam-86	807	6	amsterdam	amsterdam	PROPN
ejpam-86	807	7	,	,	PUNCT
ejpam-86	807	8	1971	1971	NUM
ejpam-86	807	9	.	.	PUNCT
ejpam-86	808	1	[	[	X
ejpam-86	808	2	17	17	NUM
ejpam-86	808	3	]	]	X
ejpam-86	808	4	kai	kai	PROPN
ejpam-86	808	5	-	-	PUNCT
ejpam-86	808	6	tang	tang	NOUN
ejpam-86	808	7	fang	fang	NOUN
ejpam-86	808	8	and	and	CCONJ
ejpam-86	808	9	t.	t.	PROPN
ejpam-86	808	10	w.	w.	PROPN
ejpam-86	808	11	anderson	anderson	PROPN
ejpam-86	808	12	.	.	PUNCT
ejpam-86	809	1	statistical	statistical	ADJ
ejpam-86	809	2	inference	inference	NOUN
ejpam-86	809	3	in	in	ADP
ejpam-86	809	4	elliptically	elliptically	ADV
ejpam-86	809	5	contoured	contoured	ADJ
ejpam-86	809	6	and	and	CCONJ
ejpam-86	809	7	related	related	ADJ
ejpam-86	809	8	distributions	distribution	NOUN
ejpam-86	809	9	.	.	PUNCT
ejpam-86	810	1	allerton	allerton	PROPN
ejpam-86	810	2	press	press	PROPN
ejpam-86	810	3	,	,	PUNCT
ejpam-86	810	4	new	new	PROPN
ejpam-86	810	5	york	york	PROPN
ejpam-86	810	6	,	,	PUNCT
ejpam-86	810	7	1990	1990	NUM
ejpam-86	810	8	.	.	PUNCT
ejpam-86	811	1	[	[	X
ejpam-86	811	2	18	18	NUM
ejpam-86	811	3	]	]	X
ejpam-86	811	4	kai	kai	PROPN
ejpam-86	811	5	-	-	PUNCT
ejpam-86	811	6	tang	tang	PROPN
ejpam-86	811	7	fang	fang	PROPN
ejpam-86	811	8	,	,	PUNCT
ejpam-86	811	9	samuel	samuel	PROPN
ejpam-86	811	10	kotz	kotz	PROPN
ejpam-86	811	11	,	,	PUNCT
ejpam-86	811	12	and	and	CCONJ
ejpam-86	811	13	kai	kai	PROPN
ejpam-86	811	14	w.	w.	PROPN
ejpam-86	811	15	ng	ng	PROPN
ejpam-86	811	16	.	.	PUNCT
ejpam-86	811	17	symmetric	symmetric	ADJ
ejpam-86	811	18	multivariate	multivariate	NOUN
ejpam-86	811	19	and	and	CCONJ
ejpam-86	811	20	related	related	ADJ
ejpam-86	811	21	distributions	distribution	NOUN
ejpam-86	811	22	.	.	PUNCT
ejpam-86	812	1	monographs	monograph	NOUN
ejpam-86	812	2	on	on	ADP
ejpam-86	812	3	statistics	statistic	NOUN
ejpam-86	812	4	and	and	CCONJ
ejpam-86	812	5	applied	apply	VERB
ejpam-86	812	6	probability	probability	NOUN
ejpam-86	812	7	.	.	PUNCT
ejpam-86	813	1	chapman	chapman	NOUN
ejpam-86	813	2	and	and	CCONJ
ejpam-86	813	3	hall	hall	PROPN
ejpam-86	813	4	,	,	PUNCT
ejpam-86	813	5	new	new	PROPN
ejpam-86	813	6	york	york	PROPN
ejpam-86	813	7	,	,	PUNCT
ejpam-86	813	8	1990	1990	NUM
ejpam-86	813	9	.	.	PUNCT
ejpam-86	814	1	[	[	X
ejpam-86	814	2	19	19	NUM
ejpam-86	814	3	]	]	X
ejpam-86	814	4	david	david	PROPN
ejpam-86	814	5	e.	e.	PROPN
ejpam-86	814	6	goldberg	goldberg	PROPN
ejpam-86	814	7	.	.	PUNCT
ejpam-86	815	1	genetic	genetic	ADJ
ejpam-86	815	2	algorithms	algorithm	NOUN
ejpam-86	815	3	in	in	ADP
ejpam-86	815	4	search	search	NOUN
ejpam-86	815	5	,	,	PUNCT
ejpam-86	815	6	optimization	optimization	NOUN
ejpam-86	815	7	,	,	PUNCT
ejpam-86	815	8	and	and	CCONJ
ejpam-86	815	9	machine	machine	NOUN
ejpam-86	815	10	learning	learning	NOUN
ejpam-86	815	11	.	.	PUNCT
ejpam-86	816	1	addisonwesley	addisonwesley	ADJ
ejpam-86	816	2	pub	pub	PROPN
ejpam-86	816	3	.	.	PUNCT
ejpam-86	817	1	co.	co.	PROPN
ejpam-86	817	2	,	,	PUNCT
ejpam-86	817	3	reading	reading	NOUN
ejpam-86	817	4	,	,	PUNCT
ejpam-86	817	5	mass	mass	PROPN
ejpam-86	817	6	.	.	PROPN
ejpam-86	817	7	,	,	PUNCT
ejpam-86	817	8	1989	1989	NUM
ejpam-86	817	9	.	.	PUNCT
ejpam-86	818	1	[	[	X
ejpam-86	818	2	20	20	NUM
ejpam-86	818	3	]	]	PUNCT
ejpam-86	818	4	e.	e.	PROPN
ejpam-86	818	5	gomez	gomez	PROPN
ejpam-86	818	6	,	,	PUNCT
ejpam-86	818	7	m.	m.	NOUN
ejpam-86	818	8	a.	a.	PROPN
ejpam-86	818	9	gomez	gomez	PROPN
ejpam-86	818	10	-	-	PUNCT
ejpam-86	818	11	villegas	villegas	PROPN
ejpam-86	818	12	,	,	PUNCT
ejpam-86	818	13	and	and	CCONJ
ejpam-86	818	14	j.	j.	PROPN
ejpam-86	818	15	m.	m.	PROPN
ejpam-86	818	16	marin	marin	PROPN
ejpam-86	818	17	.	.	PUNCT
ejpam-86	819	1	a	a	DET
ejpam-86	819	2	multivariate	multivariate	NOUN
ejpam-86	819	3	generalization	generalization	NOUN
ejpam-86	819	4	of	of	ADP
ejpam-86	819	5	the	the	DET
ejpam-86	819	6	power	power	NOUN
ejpam-86	819	7	exponential	exponential	NOUN
ejpam-86	819	8	family	family	NOUN
ejpam-86	819	9	of	of	ADP
ejpam-86	819	10	distributions	distribution	NOUN
ejpam-86	819	11	.	.	PUNCT
ejpam-86	820	1	communications	communication	NOUN
ejpam-86	820	2	in	in	ADP
ejpam-86	820	3	statistics	statistic	NOUN
ejpam-86	820	4	-	-	PUNCT
ejpam-86	820	5	theory	theory	NOUN
ejpam-86	820	6	and	and	CCONJ
ejpam-86	820	7	methods	method	NOUN
ejpam-86	820	8	,	,	PUNCT
ejpam-86	820	9	27(3):589	27(3):589	NUM
ejpam-86	820	10	–	–	PUNCT
ejpam-86	820	11	600	600	NUM
ejpam-86	820	12	,	,	PUNCT
ejpam-86	820	13	1998	1998	NUM
ejpam-86	820	14	.	.	PUNCT
ejpam-86	821	1	[	[	X
ejpam-86	821	2	21	21	NUM
ejpam-86	821	3	]	]	PUNCT
ejpam-86	821	4	a.	a.	NOUN
ejpam-86	821	5	k.	k.	PROPN
ejpam-86	821	6	gupta	gupta	PROPN
ejpam-86	821	7	and	and	CCONJ
ejpam-86	821	8	t.	t.	PROPN
ejpam-86	821	9	varga	varga	PROPN
ejpam-86	821	10	.	.	PUNCT
ejpam-86	822	1	elliptically	elliptically	ADV
ejpam-86	822	2	contoured	contour	VERB
ejpam-86	822	3	models	model	NOUN
ejpam-86	822	4	in	in	ADP
ejpam-86	822	5	statistics	statistic	NOUN
ejpam-86	822	6	.	.	PUNCT
ejpam-86	823	1	kluwer	kluwer	PROPN
ejpam-86	823	2	academic	academic	PROPN
ejpam-86	823	3	,	,	PUNCT
ejpam-86	823	4	dordrecht	dordrecht	PROPN
ejpam-86	823	5	;	;	PUNCT
ejpam-86	823	6	boston	boston	PROPN
ejpam-86	823	7	,	,	PUNCT
ejpam-86	823	8	1993	1993	NUM
ejpam-86	823	9	.	.	PUNCT
ejpam-86	824	1	[	[	X
ejpam-86	824	2	22	22	NUM
ejpam-86	824	3	]	]	PUNCT
ejpam-86	824	4	j.	j.	PROPN
ejpam-86	824	5	h.	h.	PROPN
ejpam-86	824	6	holland	holland	PROPN
ejpam-86	824	7	.	.	PUNCT
ejpam-86	825	1	genetic	genetic	ADJ
ejpam-86	825	2	algorithms	algorithm	NOUN
ejpam-86	825	3	.	.	PUNCT
ejpam-86	826	1	scientific	scientific	ADJ
ejpam-86	826	2	american	american	PROPN
ejpam-86	826	3	,	,	PUNCT
ejpam-86	826	4	267(1):66–72	267(1):66–72	NOUN
ejpam-86	826	5	,	,	PUNCT
ejpam-86	826	6	1992	1992	NUM
ejpam-86	826	7	.	.	PUNCT
ejpam-86	827	1	[	[	X
ejpam-86	827	2	23	23	NUM
ejpam-86	827	3	]	]	SYM
ejpam-86	827	4	min	min	PROPN
ejpam-86	827	5	-	-	PROPN
ejpam-86	827	6	hui	hui	PROPN
ejpam-86	827	7	liu	liu	PROPN
ejpam-86	827	8	and	and	CCONJ
ejpam-86	827	9	h.	h.	PROPN
ejpam-86	827	10	bozdogan	bozdogan	PROPN
ejpam-86	827	11	.	.	PUNCT
ejpam-86	828	1	pe	pe	X
ejpam-86	828	2	multiple	multiple	ADJ
ejpam-86	828	3	regression	regression	NOUN
ejpam-86	828	4	model	model	NOUN
ejpam-86	828	5	selection	selection	NOUN
ejpam-86	828	6	with	with	ADP
ejpam-86	828	7	icomp	icomp	NOUN
ejpam-86	828	8	and	and	CCONJ
ejpam-86	828	9	genetic	genetic	ADJ
ejpam-86	828	10	algorithms	algorithm	NOUN
ejpam-86	828	11	.	.	PUNCT
ejpam-86	829	1	working	work	VERB
ejpam-86	829	2	paper	paper	NOUN
ejpam-86	829	3	,	,	PUNCT
ejpam-86	829	4	2004	2004	NUM
ejpam-86	829	5	.	.	PUNCT
ejpam-86	830	1	[	[	X
ejpam-86	830	2	24	24	NUM
ejpam-86	830	3	]	]	X
ejpam-86	830	4	e.	e.	PROPN
ejpam-86	830	5	c.	c.	PROPN
ejpam-86	830	6	macrae	macrae	PROPN
ejpam-86	830	7	.	.	PUNCT
ejpam-86	831	1	matrix	matrix	NOUN
ejpam-86	831	2	derivatives	derivative	NOUN
ejpam-86	831	3	with	with	ADP
ejpam-86	831	4	an	an	DET
ejpam-86	831	5	application	application	NOUN
ejpam-86	831	6	to	to	ADP
ejpam-86	831	7	an	an	DET
ejpam-86	831	8	adaptive	adaptive	ADJ
ejpam-86	831	9	linear	linear	ADJ
ejpam-86	831	10	decision	decision	NOUN
ejpam-86	831	11	problem	problem	NOUN
ejpam-86	831	12	.	.	PUNCT
ejpam-86	832	1	annals	annal	NOUN
ejpam-86	832	2	of	of	ADP
ejpam-86	832	3	statistics	statistic	NOUN
ejpam-86	832	4	,	,	PUNCT
ejpam-86	832	5	2(2):337–346	2(2):337–346	NUM
ejpam-86	832	6	,	,	PUNCT
ejpam-86	832	7	1974	1974	NUM
ejpam-86	832	8	.	.	PUNCT
ejpam-86	833	1	references	reference	NOUN
ejpam-86	833	2	37	37	NUM
ejpam-86	834	1	[	[	X
ejpam-86	834	2	25	25	NUM
ejpam-86	834	3	]	]	PUNCT
ejpam-86	834	4	j.	j.	PROPN
ejpam-86	834	5	r.	r.	PROPN
ejpam-86	834	6	magnus	magnus	PROPN
ejpam-86	834	7	and	and	CCONJ
ejpam-86	834	8	h.	h.	PROPN
ejpam-86	834	9	neudecker	neudecker	PROPN
ejpam-86	834	10	.	.	PUNCT
ejpam-86	835	1	commutation	commutation	NOUN
ejpam-86	835	2	matrix	matrix	VERB
ejpam-86	835	3	some	some	DET
ejpam-86	835	4	properties	property	NOUN
ejpam-86	835	5	and	and	CCONJ
ejpam-86	835	6	applications	application	NOUN
ejpam-86	835	7	.	.	PUNCT
ejpam-86	836	1	annals	annal	NOUN
ejpam-86	836	2	of	of	ADP
ejpam-86	836	3	statistics	statistic	NOUN
ejpam-86	836	4	,	,	PUNCT
ejpam-86	836	5	7(2):381–394	7(2):381–394	NUM
ejpam-86	836	6	,	,	PUNCT
ejpam-86	836	7	1979	1979	NUM
ejpam-86	836	8	.	.	PUNCT
ejpam-86	837	1	[	[	X
ejpam-86	837	2	26	26	NUM
ejpam-86	837	3	]	]	PUNCT
ejpam-86	837	4	k.	k.	PROPN
ejpam-86	838	1	v.	v.	PROPN
ejpam-86	838	2	mardia	mardia	PROPN
ejpam-86	838	3	,	,	PUNCT
ejpam-86	838	4	j.	j.	PROPN
ejpam-86	838	5	t.	t.	PROPN
ejpam-86	838	6	kent	kent	PROPN
ejpam-86	838	7	,	,	PUNCT
ejpam-86	838	8	and	and	CCONJ
ejpam-86	838	9	j.	j.	PROPN
ejpam-86	838	10	m.	m.	PROPN
ejpam-86	838	11	bibby	bibby	PROPN
ejpam-86	838	12	.	.	PUNCT
ejpam-86	839	1	multivariate	multivariate	NOUN
ejpam-86	839	2	analysis	analysis	NOUN
ejpam-86	839	3	.	.	PUNCT
ejpam-86	840	1	probability	probability	NOUN
ejpam-86	840	2	and	and	CCONJ
ejpam-86	840	3	mathematical	mathematical	ADJ
ejpam-86	840	4	statistics	statistic	NOUN
ejpam-86	840	5	.	.	PUNCT
ejpam-86	841	1	academic	academic	ADJ
ejpam-86	841	2	press	press	PROPN
ejpam-86	841	3	,	,	PUNCT
ejpam-86	841	4	london	london	PROPN
ejpam-86	841	5	;	;	PUNCT
ejpam-86	841	6	new	new	PROPN
ejpam-86	841	7	york	york	PROPN
ejpam-86	841	8	,	,	PUNCT
ejpam-86	841	9	1979	1979	NUM
ejpam-86	841	10	.	.	PUNCT
ejpam-86	842	1	[	[	X
ejpam-86	842	2	27	27	NUM
ejpam-86	842	3	]	]	PUNCT
ejpam-86	842	4	m.	m.	NOUN
ejpam-86	842	5	mitchell	mitchell	PROPN
ejpam-86	842	6	.	.	PUNCT
ejpam-86	843	1	an	an	DET
ejpam-86	843	2	introduction	introduction	NOUN
ejpam-86	843	3	to	to	ADP
ejpam-86	843	4	genetic	genetic	ADJ
ejpam-86	843	5	algorithms	algorithm	NOUN
ejpam-86	843	6	.	.	PUNCT
ejpam-86	844	1	mit	mit	PROPN
ejpam-86	844	2	press	press	PROPN
ejpam-86	844	3	,	,	PUNCT
ejpam-86	844	4	cambridge	cambridge	PROPN
ejpam-86	844	5	,	,	PUNCT
ejpam-86	844	6	ma	ma	PROPN
ejpam-86	844	7	,	,	PUNCT
ejpam-86	844	8	1996	1996	NUM
ejpam-86	844	9	.	.	PUNCT
ejpam-86	845	1	[	[	X
ejpam-86	845	2	28	28	NUM
ejpam-86	845	3	]	]	X
ejpam-86	845	4	svetlozar	svetlozar	PROPN
ejpam-86	845	5	t.	t.	PROPN
ejpam-86	845	6	rachev	rachev	PROPN
ejpam-86	845	7	and	and	CCONJ
ejpam-86	845	8	s.	s.	PROPN
ejpam-86	845	9	mittnik	mittnik	PROPN
ejpam-86	845	10	.	.	PUNCT
ejpam-86	846	1	stable	stable	ADJ
ejpam-86	846	2	paretian	paretian	NOUN
ejpam-86	846	3	models	model	NOUN
ejpam-86	846	4	in	in	ADP
ejpam-86	846	5	finance	finance	NOUN
ejpam-86	846	6	.	.	PUNCT
ejpam-86	847	1	series	series	PROPN
ejpam-86	847	2	in	in	ADP
ejpam-86	847	3	financial	financial	ADJ
ejpam-86	847	4	economics	economic	NOUN
ejpam-86	847	5	and	and	CCONJ
ejpam-86	847	6	quantitative	quantitative	ADJ
ejpam-86	847	7	analysis	analysis	NOUN
ejpam-86	847	8	.	.	PUNCT
ejpam-86	848	1	wiley	wiley	PROPN
ejpam-86	848	2	,	,	PUNCT
ejpam-86	848	3	chichester	chichester	PROPN
ejpam-86	848	4	,	,	PUNCT
ejpam-86	848	5	2000	2000	NUM
ejpam-86	848	6	.	.	PUNCT
ejpam-86	849	1	[	[	X
ejpam-86	849	2	29	29	NUM
ejpam-86	849	3	]	]	X
ejpam-86	849	4	j.	j.	PROPN
ejpam-86	849	5	rissanen	rissanen	PROPN
ejpam-86	849	6	.	.	PUNCT
ejpam-86	849	7	minimax	minimax	PROPN
ejpam-86	849	8	entropy	entropy	PROPN
ejpam-86	849	9	estimation	estimation	NOUN
ejpam-86	849	10	of	of	ADP
ejpam-86	849	11	models	model	NOUN
ejpam-86	849	12	for	for	ADP
ejpam-86	849	13	vector	vector	NOUN
ejpam-86	849	14	processes	process	NOUN
ejpam-86	849	15	.	.	PUNCT
ejpam-86	850	1	in	in	ADP
ejpam-86	850	2	system	system	NOUN
ejpam-86	850	3	identification	identification	NOUN
ejpam-86	850	4	:	:	PUNCT
ejpam-86	850	5	advances	advance	NOUN
ejpam-86	850	6	in	in	ADP
ejpam-86	850	7	case	case	NOUN
ejpam-86	850	8	studies	study	NOUN
ejpam-86	850	9	,	,	PUNCT
ejpam-86	850	10	pages	page	NOUN
ejpam-86	850	11	97–120	97–120	PROPN
ejpam-86	850	12	,	,	PUNCT
ejpam-86	850	13	1976	1976	NUM
ejpam-86	850	14	.	.	PUNCT
ejpam-86	851	1	[	[	X
ejpam-86	851	2	30	30	NUM
ejpam-86	851	3	]	]	X
ejpam-86	851	4	gerald	gerald	PROPN
ejpam-86	851	5	stanley	stanley	PROPN
ejpam-86	851	6	rogers	rogers	PROPN
ejpam-86	851	7	.	.	PUNCT
ejpam-86	852	1	matrix	matrix	NOUN
ejpam-86	852	2	derivatives	derivative	NOUN
ejpam-86	852	3	.	.	PUNCT
ejpam-86	853	1	m.	m.	NOUN
ejpam-86	853	2	dekker	dekker	PROPN
ejpam-86	853	3	,	,	PUNCT
ejpam-86	853	4	new	new	PROPN
ejpam-86	853	5	york	york	PROPN
ejpam-86	853	6	,	,	PUNCT
ejpam-86	853	7	1980	1980	NUM
ejpam-86	853	8	.	.	PUNCT
ejpam-86	854	1	[	[	X
ejpam-86	854	2	31	31	NUM
ejpam-86	854	3	]	]	PUNCT
ejpam-86	854	4	eusebio	eusebio	PROPN
ejpam-86	854	5	gmez	gmez	PROPN
ejpam-86	854	6	sanchez	sanchez	PROPN
ejpam-86	854	7	-	-	PUNCT
ejpam-86	854	8	manzano	manzano	PROPN
ejpam-86	854	9	,	,	PUNCT
ejpam-86	854	10	miguel	miguel	PROPN
ejpam-86	854	11	angel	angel	PROPN
ejpam-86	854	12	gomez	gomez	PROPN
ejpam-86	854	13	-	-	PUNCT
ejpam-86	854	14	villegas	villegas	PROPN
ejpam-86	854	15	,	,	PUNCT
ejpam-86	854	16	and	and	CCONJ
ejpam-86	854	17	juan	juan	PROPN
ejpam-86	854	18	-	-	PUNCT
ejpam-86	854	19	miguel	miguel	PROPN
ejpam-86	854	20	marndiazaraque	marndiazaraque	PROPN
ejpam-86	854	21	.	.	PUNCT
ejpam-86	855	1	a	a	DET
ejpam-86	855	2	matrix	matrix	NOUN
ejpam-86	855	3	variate	variate	NOUN
ejpam-86	855	4	generalization	generalization	NOUN
ejpam-86	855	5	of	of	ADP
ejpam-86	855	6	the	the	DET
ejpam-86	855	7	power	power	NOUN
ejpam-86	855	8	exponential	exponential	NOUN
ejpam-86	855	9	family	family	NOUN
ejpam-86	855	10	of	of	ADP
ejpam-86	855	11	distributions	distribution	NOUN
ejpam-86	855	12	.	.	PUNCT
ejpam-86	856	1	communications	communication	NOUN
ejpam-86	856	2	in	in	ADP
ejpam-86	856	3	statistics	statistic	NOUN
ejpam-86	856	4	,	,	PUNCT
ejpam-86	856	5	part	part	NOUN
ejpam-86	856	6	a	a	DET
ejpam-86	856	7	–	–	PUNCT
ejpam-86	856	8	theory	theory	NOUN
ejpam-86	856	9	and	and	CCONJ
ejpam-86	856	10	methods	method	NOUN
ejpam-86	856	11	[	[	X
ejpam-86	856	12	split	split	VERB
ejpam-86	856	13	from	from	ADP
ejpam-86	856	14	:	:	PUNCT
ejpam-86	856	15	@j(commstat	@j(commstat	X
ejpam-86	856	16	)	)	PUNCT
ejpam-86	856	17	]	]	X
ejpam-86	856	18	,	,	PUNCT
ejpam-86	856	19	31(12):2167–2182	31(12):2167–2182	PROPN
ejpam-86	856	20	,	,	PUNCT
ejpam-86	856	21	2002	2002	NUM
ejpam-86	856	22	.	.	PUNCT
ejpam-86	857	1	[	[	X
ejpam-86	857	2	32	32	NUM
ejpam-86	857	3	]	]	PUNCT
ejpam-86	857	4	m.	m.	NOUN
ejpam-86	857	5	t.	t.	NOUN
ejpam-86	857	6	subbotin	subbotin	NOUN
ejpam-86	857	7	.	.	PUNCT
ejpam-86	858	1	on	on	ADP
ejpam-86	858	2	the	the	DET
ejpam-86	858	3	law	law	NOUN
ejpam-86	858	4	of	of	ADP
ejpam-86	858	5	frequency	frequency	NOUN
ejpam-86	858	6	of	of	ADP
ejpam-86	858	7	errors	error	NOUN
ejpam-86	858	8	.	.	PUNCT
ejpam-86	859	1	matematicheskii	matematicheskii	PROPN
ejpam-86	859	2	sbornik	sbornik	PROPN
ejpam-86	859	3	,	,	PUNCT
ejpam-86	859	4	pages	page	NOUN
ejpam-86	859	5	296–300	296–300	NUM
ejpam-86	859	6	,	,	PUNCT
ejpam-86	859	7	1923	1923	NUM
ejpam-86	859	8	.	.	PUNCT
ejpam-86	860	1	[	[	X
ejpam-86	860	2	33	33	NUM
ejpam-86	860	3	]	]	PUNCT
ejpam-86	860	4	p.	p.	NOUN
ejpam-86	860	5	theodossiou	theodossiou	NOUN
ejpam-86	860	6	.	.	PUNCT
ejpam-86	861	1	financial	financial	ADJ
ejpam-86	861	2	data	datum	NOUN
ejpam-86	861	3	and	and	CCONJ
ejpam-86	861	4	the	the	DET
ejpam-86	861	5	skewed	skewed	ADJ
ejpam-86	861	6	generalized	generalize	VERB
ejpam-86	861	7	t	t	NOUN
ejpam-86	861	8	distribution	distribution	NOUN
ejpam-86	861	9	.	.	PUNCT
ejpam-86	862	1	management	management	NOUN
ejpam-86	862	2	science	science	NOUN
ejpam-86	862	3	,	,	PUNCT
ejpam-86	862	4	44(12):1650–1661	44(12):1650–1661	NUM
ejpam-86	862	5	,	,	PUNCT
ejpam-86	862	6	1998	1998	NUM
ejpam-86	862	7	.	.	PUNCT
ejpam-86	863	1	[	[	X
ejpam-86	863	2	34	34	NUM
ejpam-86	863	3	]	]	X
ejpam-86	863	4	j.	j.	PROPN
ejpam-86	863	5	toyli	toyli	PROPN
ejpam-86	863	6	,	,	PUNCT
ejpam-86	863	7	k.	k.	PROPN
ejpam-86	863	8	kaski	kaski	PROPN
ejpam-86	863	9	,	,	PUNCT
ejpam-86	863	10	and	and	CCONJ
ejpam-86	863	11	a.	a.	NOUN
ejpam-86	863	12	kanto	kanto	PROPN
ejpam-86	863	13	.	.	PUNCT
ejpam-86	864	1	on	on	ADP
ejpam-86	864	2	the	the	DET
ejpam-86	864	3	shape	shape	NOUN
ejpam-86	864	4	of	of	ADP
ejpam-86	864	5	asset	asset	NOUN
ejpam-86	864	6	return	return	NOUN
ejpam-86	864	7	distribution	distribution	NOUN
ejpam-86	864	8	.	.	PUNCT
ejpam-86	865	1	communications	communication	NOUN
ejpam-86	865	2	in	in	ADP
ejpam-86	865	3	statistics	statistic	NOUN
ejpam-86	865	4	-	-	PUNCT
ejpam-86	865	5	simulation	simulation	NOUN
ejpam-86	865	6	and	and	CCONJ
ejpam-86	865	7	computation	computation	NOUN
ejpam-86	865	8	,	,	PUNCT
ejpam-86	865	9	31(4):489–521	31(4):489–521	NOUN
ejpam-86	865	10	,	,	PUNCT
ejpam-86	865	11	2002	2002	NUM
ejpam-86	865	12	.	.	PUNCT
ejpam-86	866	1	[	[	X
ejpam-86	866	2	35	35	NUM
ejpam-86	866	3	]	]	PUNCT
ejpam-86	866	4	j.	j.	PROPN
ejpam-86	866	5	yang	yang	PROPN
ejpam-86	866	6	and	and	CCONJ
ejpam-86	866	7	v.	v.	ADP
ejpam-86	866	8	honavar	honavar	NOUN
ejpam-86	866	9	.	.	PUNCT
ejpam-86	867	1	feature	feature	NOUN
ejpam-86	867	2	subset	subset	NOUN
ejpam-86	867	3	selection	selection	NOUN
ejpam-86	867	4	using	use	VERB
ejpam-86	867	5	a	a	DET
ejpam-86	867	6	genetic	genetic	ADJ
ejpam-86	867	7	algorithm	algorithm	NOUN
ejpam-86	867	8	.	.	PUNCT
ejpam-86	868	1	intelligent	intelligent	ADJ
ejpam-86	868	2	systems	system	NOUN
ejpam-86	868	3	and	and	CCONJ
ejpam-86	868	4	their	their	PRON
ejpam-86	868	5	applications	application	NOUN
ejpam-86	868	6	,	,	PUNCT
ejpam-86	868	7	ieee	ieee	NOUN
ejpam-86	868	8	,	,	PUNCT
ejpam-86	868	9	13(2):44	13(2):44	NUM
ejpam-86	868	10	–	–	SYM
ejpam-86	868	11	49	49	NUM
ejpam-86	868	12	,	,	PUNCT
ejpam-86	868	13	1998	1998	NUM
ejpam-86	868	14	.	.	PUNCT
ejpam-86	869	1	[	[	X
ejpam-86	869	2	36	36	NUM
ejpam-86	869	3	]	]	X
ejpam-86	869	4	r.	r.	PROPN
ejpam-86	869	5	zeckhauser	zeckhauser	PROPN
ejpam-86	869	6	and	and	CCONJ
ejpam-86	869	7	m.	m.	PROPN
ejpam-86	869	8	thompson	thompson	PROPN
ejpam-86	869	9	.	.	PUNCT
ejpam-86	870	1	linear	linear	PROPN
ejpam-86	870	2	regression	regression	NOUN
ejpam-86	870	3	with	with	ADP
ejpam-86	870	4	non	non	ADJ
ejpam-86	870	5	-	-	ADJ
ejpam-86	870	6	normal	normal	ADJ
ejpam-86	870	7	error	error	NOUN
ejpam-86	870	8	terms	term	NOUN
ejpam-86	870	9	.	.	PUNCT
ejpam-86	871	1	review	review	NOUN
ejpam-86	871	2	of	of	ADP
ejpam-86	871	3	economics	economic	NOUN
ejpam-86	871	4	and	and	CCONJ
ejpam-86	871	5	statistics	statistic	NOUN
ejpam-86	871	6	,	,	PUNCT
ejpam-86	871	7	52(3):280–286	52(3):280–286	NUM
ejpam-86	871	8	,	,	PUNCT
ejpam-86	871	9	1970	1970	NUM
ejpam-86	871	10	.	.	PUNCT