id	sid	tid	token	lemma	pos
ejpam-642	1	1	2_642_zhu.dvi	2_642_zhu.dvi	NUM
ejpam-642	1	2	european	european	ADJ
ejpam-642	1	3	journal	journal	PROPN
ejpam-642	1	4	of	of	ADP
ejpam-642	1	5	pure	pure	ADJ
ejpam-642	1	6	and	and	CCONJ
ejpam-642	1	7	applied	apply	VERB
ejpam-642	1	8	mathematics	mathematic	NOUN
ejpam-642	1	9	vol	vol	NOUN
ejpam-642	1	10	.	.	PUNCT
ejpam-642	2	1	3	3	NUM
ejpam-642	2	2	,	,	PUNCT
ejpam-642	2	3	no	no	INTJ
ejpam-642	2	4	.	.	NOUN
ejpam-642	2	5	2	2	NUM
ejpam-642	2	6	,	,	PUNCT
ejpam-642	2	7	2010	2010	NUM
ejpam-642	2	8	,	,	PUNCT
ejpam-642	2	9	141	141	NUM
ejpam-642	2	10	-	-	SYM
ejpam-642	2	11	162	162	NUM
ejpam-642	2	12	issn	issn	PROPN
ejpam-642	2	13	1307	1307	NUM
ejpam-642	2	14	-	-	SYM
ejpam-642	2	15	5543	5543	NUM
ejpam-642	2	16	–	–	PUNCT
ejpam-642	2	17	www.ejpam.com	www.ejpam.com	X
ejpam-642	2	18	stepwise	stepwise	PROPN
ejpam-642	2	19	variable	variable	ADJ
ejpam-642	2	20	selection	selection	NOUN
ejpam-642	2	21	for	for	ADP
ejpam-642	2	22	loglinear	loglinear	NOUN
ejpam-642	2	23	mixtures	mixture	NOUN
ejpam-642	2	24	in	in	ADP
ejpam-642	2	25	record	record	NOUN
ejpam-642	2	26	linkage	linkage	NOUN
ejpam-642	2	27	rong	rong	PROPN
ejpam-642	2	28	zhu	zhu	PROPN
ejpam-642	2	29	1,∗	1,∗	PROPN
ejpam-642	2	30	,	,	PUNCT
ejpam-642	2	31	jin	jin	PROPN
ejpam-642	2	32	zhang	zhang	PROPN
ejpam-642	2	33	2	2	NUM
ejpam-642	2	34	,	,	PUNCT
ejpam-642	2	35	da	da	PROPN
ejpam-642	2	36	zhang	zhang	PROPN
ejpam-642	2	37	3	3	NUM
ejpam-642	2	38	,	,	PUNCT
ejpam-642	2	39	guohua	guohua	PROPN
ejpam-642	2	40	yan	yan	PROPN
ejpam-642	2	41	4	4	NUM
ejpam-642	2	42	1	1	NUM
ejpam-642	2	43	department	department	NOUN
ejpam-642	2	44	of	of	ADP
ejpam-642	2	45	mathematics	mathematic	NOUN
ejpam-642	2	46	and	and	CCONJ
ejpam-642	2	47	statistics	statistic	NOUN
ejpam-642	2	48	,	,	PUNCT
ejpam-642	2	49	mcmaster	mcmaster	PROPN
ejpam-642	2	50	university	university	PROPN
ejpam-642	2	51	,	,	PUNCT
ejpam-642	2	52	canada	canada	PROPN
ejpam-642	2	53	2	2	NUM
ejpam-642	2	54	office	office	NOUN
ejpam-642	2	55	of	of	ADP
ejpam-642	2	56	institutional	institutional	ADJ
ejpam-642	2	57	research	research	NOUN
ejpam-642	2	58	and	and	CCONJ
ejpam-642	2	59	analysis	analysis	NOUN
ejpam-642	2	60	,	,	PUNCT
ejpam-642	2	61	mcmaster	mcmaster	PROPN
ejpam-642	2	62	university	university	PROPN
ejpam-642	2	63	,	,	PUNCT
ejpam-642	2	64	canada	canada	PROPN
ejpam-642	2	65	3	3	NUM
ejpam-642	2	66	department	department	PROPN
ejpam-642	2	67	of	of	ADP
ejpam-642	2	68	computer	computer	NOUN
ejpam-642	2	69	science	science	NOUN
ejpam-642	2	70	technology	technology	NOUN
ejpam-642	2	71	,	,	PUNCT
ejpam-642	2	72	ohio	ohio	PROPN
ejpam-642	2	73	university	university	PROPN
ejpam-642	2	74	lancaster	lancaster	PROPN
ejpam-642	2	75	,	,	PUNCT
ejpam-642	2	76	usa	usa	PROPN
ejpam-642	2	77	4	4	NUM
ejpam-642	2	78	department	department	NOUN
ejpam-642	2	79	of	of	ADP
ejpam-642	2	80	mathematics	mathematic	NOUN
ejpam-642	2	81	and	and	CCONJ
ejpam-642	2	82	statistics	statistic	NOUN
ejpam-642	2	83	,	,	PUNCT
ejpam-642	2	84	university	university	NOUN
ejpam-642	2	85	of	of	ADP
ejpam-642	2	86	new	new	PROPN
ejpam-642	2	87	brunswick	brunswick	PROPN
ejpam-642	2	88	,	,	PUNCT
ejpam-642	2	89	canada	canada	PROPN
ejpam-642	2	90	abstract	abstract	PROPN
ejpam-642	2	91	.	.	PUNCT
ejpam-642	3	1	a	a	DET
ejpam-642	3	2	model	model	NOUN
ejpam-642	3	3	building	building	NOUN
ejpam-642	3	4	strategy	strategy	NOUN
ejpam-642	3	5	is	be	AUX
ejpam-642	3	6	proposed	propose	VERB
ejpam-642	3	7	to	to	PART
ejpam-642	3	8	improve	improve	VERB
ejpam-642	3	9	the	the	DET
ejpam-642	3	10	probabilistic	probabilistic	ADJ
ejpam-642	3	11	match	match	NOUN
ejpam-642	3	12	in	in	ADP
ejpam-642	3	13	record	record	NOUN
ejpam-642	3	14	linkage	linkage	NOUN
ejpam-642	3	15	with	with	ADP
ejpam-642	3	16	focus	focus	NOUN
ejpam-642	3	17	on	on	ADP
ejpam-642	3	18	the	the	DET
ejpam-642	3	19	loglinear	loglinear	NOUN
ejpam-642	3	20	mixture	mixture	NOUN
ejpam-642	3	21	model	model	NOUN
ejpam-642	3	22	of	of	ADP
ejpam-642	3	23	two	two	NUM
ejpam-642	3	24	components	component	NOUN
ejpam-642	3	25	,	,	PUNCT
ejpam-642	3	26	each	each	PRON
ejpam-642	3	27	for	for	ADP
ejpam-642	3	28	the	the	DET
ejpam-642	3	29	matched	match	VERB
ejpam-642	3	30	and	and	CCONJ
ejpam-642	3	31	unmatched	unmatched	ADJ
ejpam-642	3	32	pairs	pair	NOUN
ejpam-642	3	33	respectively	respectively	ADV
ejpam-642	3	34	.	.	PUNCT
ejpam-642	4	1	in	in	ADP
ejpam-642	4	2	reality	reality	NOUN
ejpam-642	4	3	,	,	PUNCT
ejpam-642	4	4	comparison	comparison	NOUN
ejpam-642	4	5	attributes	attribute	NOUN
ejpam-642	4	6	(	(	PUNCT
ejpam-642	4	7	i.e.	i.e.	X
ejpam-642	4	8	,	,	PUNCT
ejpam-642	4	9	covariates	covariate	NOUN
ejpam-642	4	10	)	)	PUNCT
ejpam-642	4	11	often	often	ADV
ejpam-642	4	12	interact	interact	VERB
ejpam-642	4	13	with	with	ADP
ejpam-642	4	14	each	each	DET
ejpam-642	4	15	other	other	ADJ
ejpam-642	4	16	,	,	PUNCT
ejpam-642	4	17	leading	lead	VERB
ejpam-642	4	18	to	to	ADP
ejpam-642	4	19	more	more	ADJ
ejpam-642	4	20	or	or	CCONJ
ejpam-642	4	21	less	less	ADJ
ejpam-642	4	22	interactions	interaction	NOUN
ejpam-642	4	23	in	in	ADP
ejpam-642	4	24	the	the	DET
ejpam-642	4	25	loglinear	loglinear	NOUN
ejpam-642	4	26	models	model	NOUN
ejpam-642	4	27	for	for	ADP
ejpam-642	4	28	both	both	CCONJ
ejpam-642	4	29	the	the	DET
ejpam-642	4	30	matched	match	VERB
ejpam-642	4	31	and	and	CCONJ
ejpam-642	4	32	unmatched	unmatched	ADJ
ejpam-642	4	33	pairs	pair	NOUN
ejpam-642	4	34	.	.	PUNCT
ejpam-642	5	1	however	however	ADV
ejpam-642	5	2	,	,	PUNCT
ejpam-642	5	3	the	the	DET
ejpam-642	5	4	interactions	interaction	NOUN
ejpam-642	5	5	patterns	pattern	NOUN
ejpam-642	5	6	are	be	AUX
ejpam-642	5	7	often	often	ADV
ejpam-642	5	8	not	not	PART
ejpam-642	5	9	the	the	DET
ejpam-642	5	10	same	same	ADJ
ejpam-642	5	11	for	for	ADP
ejpam-642	5	12	both	both	DET
ejpam-642	5	13	components	component	NOUN
ejpam-642	5	14	.	.	PUNCT
ejpam-642	6	1	particularly	particularly	ADV
ejpam-642	6	2	,	,	PUNCT
ejpam-642	6	3	because	because	SCONJ
ejpam-642	6	4	the	the	DET
ejpam-642	6	5	number	number	NOUN
ejpam-642	6	6	of	of	ADP
ejpam-642	6	7	matched	match	VERB
ejpam-642	6	8	pairs	pair	NOUN
ejpam-642	6	9	is	be	AUX
ejpam-642	6	10	usually	usually	ADV
ejpam-642	6	11	very	very	ADV
ejpam-642	6	12	small	small	ADJ
ejpam-642	6	13	compared	compare	VERB
ejpam-642	6	14	with	with	ADP
ejpam-642	6	15	that	that	PRON
ejpam-642	6	16	of	of	ADP
ejpam-642	6	17	unmatched	unmatched	ADJ
ejpam-642	6	18	pairs	pair	NOUN
ejpam-642	6	19	in	in	ADP
ejpam-642	6	20	practice	practice	NOUN
ejpam-642	6	21	,	,	PUNCT
ejpam-642	6	22	the	the	DET
ejpam-642	6	23	model	model	NOUN
ejpam-642	6	24	for	for	ADP
ejpam-642	6	25	matched	match	VERB
ejpam-642	6	26	pairs	pair	NOUN
ejpam-642	6	27	can	can	AUX
ejpam-642	6	28	not	not	PART
ejpam-642	6	29	be	be	AUX
ejpam-642	6	30	fitted	fit	VERB
ejpam-642	6	31	with	with	ADP
ejpam-642	6	32	the	the	DET
ejpam-642	6	33	same	same	ADJ
ejpam-642	6	34	higher	high	ADJ
ejpam-642	6	35	order	order	NOUN
ejpam-642	6	36	interactions	interaction	NOUN
ejpam-642	6	37	as	as	ADP
ejpam-642	6	38	that	that	PRON
ejpam-642	6	39	for	for	ADP
ejpam-642	6	40	the	the	DET
ejpam-642	6	41	unmatched	unmatched	ADJ
ejpam-642	6	42	pairs	pair	NOUN
ejpam-642	6	43	.	.	PUNCT
ejpam-642	7	1	the	the	DET
ejpam-642	7	2	proposed	propose	VERB
ejpam-642	7	3	strategy	strategy	NOUN
ejpam-642	7	4	is	be	AUX
ejpam-642	7	5	data	data	NOUN
ejpam-642	7	6	-	-	PUNCT
ejpam-642	7	7	driven	drive	VERB
ejpam-642	7	8	,	,	PUNCT
ejpam-642	7	9	and	and	CCONJ
ejpam-642	7	10	attempts	attempt	VERB
ejpam-642	7	11	to	to	PART
ejpam-642	7	12	avoid	avoid	VERB
ejpam-642	7	13	both	both	CCONJ
ejpam-642	7	14	underfitting	underfitte	VERB
ejpam-642	7	15	and	and	CCONJ
ejpam-642	7	16	overfitting	overfitte	VERB
ejpam-642	7	17	due	due	ADP
ejpam-642	7	18	to	to	ADP
ejpam-642	7	19	subjective	subjective	ADJ
ejpam-642	7	20	model	model	NOUN
ejpam-642	7	21	specification	specification	NOUN
ejpam-642	7	22	for	for	ADP
ejpam-642	7	23	the	the	DET
ejpam-642	7	24	data	datum	NOUN
ejpam-642	7	25	.	.	PUNCT
ejpam-642	8	1	starting	start	VERB
ejpam-642	8	2	from	from	ADP
ejpam-642	8	3	the	the	DET
ejpam-642	8	4	situation	situation	NOUN
ejpam-642	8	5	of	of	ADP
ejpam-642	8	6	no	no	DET
ejpam-642	8	7	interaction	interaction	NOUN
ejpam-642	8	8	,	,	PUNCT
ejpam-642	8	9	we	we	PRON
ejpam-642	8	10	add	add	VERB
ejpam-642	8	11	interactions	interaction	NOUN
ejpam-642	8	12	sequentially	sequentially	ADV
ejpam-642	8	13	in	in	ADP
ejpam-642	8	14	two	two	NUM
ejpam-642	8	15	loglinear	loglinear	NOUN
ejpam-642	8	16	components	component	NOUN
ejpam-642	8	17	using	use	VERB
ejpam-642	8	18	the	the	DET
ejpam-642	8	19	forward	forward	ADJ
ejpam-642	8	20	selection	selection	NOUN
ejpam-642	8	21	approach	approach	NOUN
ejpam-642	8	22	.	.	PUNCT
ejpam-642	9	1	specifically	specifically	ADV
ejpam-642	9	2	,	,	PUNCT
ejpam-642	9	3	we	we	PRON
ejpam-642	9	4	define	define	VERB
ejpam-642	9	5	the	the	DET
ejpam-642	9	6	alternatively	alternatively	ADV
ejpam-642	9	7	climbing	climb	VERB
ejpam-642	9	8	pathways	pathway	NOUN
ejpam-642	9	9	through	through	ADP
ejpam-642	9	10	mixture	mixture	NOUN
ejpam-642	9	11	families	family	NOUN
ejpam-642	9	12	of	of	ADP
ejpam-642	9	13	two	two	NUM
ejpam-642	9	14	components	component	NOUN
ejpam-642	9	15	with	with	ADP
ejpam-642	9	16	higher	high	ADJ
ejpam-642	9	17	order	order	NOUN
ejpam-642	9	18	interactions	interaction	NOUN
ejpam-642	9	19	.	.	PUNCT
ejpam-642	10	1	the	the	DET
ejpam-642	10	2	mixture	mixture	NOUN
ejpam-642	10	3	models	model	NOUN
ejpam-642	10	4	expanded	expand	VERB
ejpam-642	10	5	along	along	ADP
ejpam-642	10	6	a	a	DET
ejpam-642	10	7	pathway	pathway	NOUN
ejpam-642	10	8	are	be	AUX
ejpam-642	10	9	nested	nest	VERB
ejpam-642	10	10	successively	successively	ADV
ejpam-642	10	11	.	.	PUNCT
ejpam-642	11	1	thus	thus	ADV
ejpam-642	11	2	,	,	PUNCT
ejpam-642	11	3	conventional	conventional	ADJ
ejpam-642	11	4	tests	test	NOUN
ejpam-642	11	5	used	use	VERB
ejpam-642	11	6	for	for	ADP
ejpam-642	11	7	comparison	comparison	NOUN
ejpam-642	11	8	of	of	ADP
ejpam-642	11	9	nested	nested	ADJ
ejpam-642	11	10	models	model	NOUN
ejpam-642	11	11	can	can	AUX
ejpam-642	11	12	be	be	AUX
ejpam-642	11	13	applied	apply	VERB
ejpam-642	11	14	.	.	PUNCT
ejpam-642	12	1	regarding	regard	VERB
ejpam-642	12	2	parameter	parameter	NOUN
ejpam-642	12	3	estimation	estimation	NOUN
ejpam-642	12	4	for	for	ADP
ejpam-642	12	5	the	the	DET
ejpam-642	12	6	mixture	mixture	NOUN
ejpam-642	12	7	,	,	PUNCT
ejpam-642	12	8	a	a	DET
ejpam-642	12	9	simplified	simplified	ADJ
ejpam-642	12	10	method	method	NOUN
ejpam-642	12	11	(	(	PUNCT
ejpam-642	12	12	including	include	VERB
ejpam-642	12	13	the	the	DET
ejpam-642	12	14	choice	choice	NOUN
ejpam-642	12	15	of	of	ADP
ejpam-642	12	16	initial	initial	ADJ
ejpam-642	12	17	values	value	NOUN
ejpam-642	12	18	of	of	ADP
ejpam-642	12	19	parameters	parameter	NOUN
ejpam-642	12	20	)	)	PUNCT
ejpam-642	12	21	for	for	ADP
ejpam-642	12	22	the	the	DET
ejpam-642	12	23	em	em	PROPN
ejpam-642	12	24	algorithm	algorithm	NOUN
ejpam-642	12	25	is	be	AUX
ejpam-642	12	26	developed	develop	VERB
ejpam-642	12	27	,	,	PUNCT
ejpam-642	12	28	which	which	PRON
ejpam-642	12	29	facilitates	facilitate	VERB
ejpam-642	12	30	the	the	DET
ejpam-642	12	31	mixture	mixture	NOUN
ejpam-642	12	32	model	model	NOUN
ejpam-642	12	33	fitting	fit	VERB
ejpam-642	12	34	using	use	VERB
ejpam-642	12	35	existing	exist	VERB
ejpam-642	12	36	packages	package	NOUN
ejpam-642	12	37	and	and	CCONJ
ejpam-642	12	38	functions	function	NOUN
ejpam-642	12	39	in	in	ADP
ejpam-642	12	40	sophisticated	sophisticated	ADJ
ejpam-642	12	41	statistical	statistical	ADJ
ejpam-642	12	42	software	software	NOUN
ejpam-642	12	43	like	like	ADP
ejpam-642	12	44	r.	r.	PROPN
ejpam-642	12	45	simulation	simulation	PROPN
ejpam-642	12	46	studies	study	NOUN
ejpam-642	12	47	have	have	AUX
ejpam-642	12	48	then	then	ADV
ejpam-642	12	49	been	be	AUX
ejpam-642	12	50	conducted	conduct	VERB
ejpam-642	12	51	for	for	ADP
ejpam-642	12	52	various	various	ADJ
ejpam-642	12	53	situations	situation	NOUN
ejpam-642	12	54	to	to	PART
ejpam-642	12	55	assess	assess	VERB
ejpam-642	12	56	the	the	DET
ejpam-642	12	57	model	model	NOUN
ejpam-642	12	58	selection	selection	NOUN
ejpam-642	12	59	approach	approach	NOUN
ejpam-642	12	60	,	,	PUNCT
ejpam-642	12	61	and	and	CCONJ
ejpam-642	12	62	comparisons	comparison	NOUN
ejpam-642	12	63	of	of	ADP
ejpam-642	12	64	the	the	DET
ejpam-642	12	65	selected	select	VERB
ejpam-642	12	66	models	model	NOUN
ejpam-642	12	67	with	with	ADP
ejpam-642	12	68	the	the	DET
ejpam-642	12	69	naive	naive	ADJ
ejpam-642	12	70	model	model	NOUN
ejpam-642	12	71	assuming	assume	VERB
ejpam-642	12	72	field	field	NOUN
ejpam-642	12	73	independence	independence	NOUN
ejpam-642	12	74	have	have	AUX
ejpam-642	12	75	been	be	AUX
ejpam-642	12	76	made	make	VERB
ejpam-642	12	77	.	.	PUNCT
ejpam-642	13	1	we	we	PRON
ejpam-642	13	2	have	have	AUX
ejpam-642	13	3	applied	apply	VERB
ejpam-642	13	4	this	this	DET
ejpam-642	13	5	strategy	strategy	NOUN
ejpam-642	13	6	to	to	ADP
ejpam-642	13	7	the	the	DET
ejpam-642	13	8	record	record	NOUN
ejpam-642	13	9	linkage	linkage	NOUN
ejpam-642	13	10	case	case	NOUN
ejpam-642	13	11	study	study	NOUN
ejpam-642	13	12	in	in	ADP
ejpam-642	13	13	2006	2006	NUM
ejpam-642	13	14	annual	annual	ADJ
ejpam-642	13	15	meeting	meeting	NOUN
ejpam-642	13	16	of	of	ADP
ejpam-642	13	17	statistical	statistical	ADJ
ejpam-642	13	18	society	society	NOUN
ejpam-642	13	19	of	of	ADP
ejpam-642	13	20	canada	canada	PROPN
ejpam-642	13	21	(	(	PUNCT
ejpam-642	13	22	ssc	ssc	PROPN
ejpam-642	13	23	)	)	PUNCT
ejpam-642	13	24	and	and	CCONJ
ejpam-642	13	25	identified	identify	VERB
ejpam-642	13	26	interactions	interaction	NOUN
ejpam-642	13	27	among	among	ADP
ejpam-642	13	28	certain	certain	ADJ
ejpam-642	13	29	comparison	comparison	NOUN
ejpam-642	13	30	attributes	attribute	VERB
ejpam-642	13	31	for	for	ADP
ejpam-642	13	32	both	both	DET
ejpam-642	13	33	matched	match	VERB
ejpam-642	13	34	and	and	CCONJ
ejpam-642	13	35	unmatched	unmatched	ADJ
ejpam-642	13	36	pairs	pair	NOUN
ejpam-642	13	37	;	;	PUNCT
ejpam-642	13	38	these	these	DET
ejpam-642	13	39	interactions	interaction	NOUN
ejpam-642	13	40	are	be	AUX
ejpam-642	13	41	not	not	PART
ejpam-642	13	42	always	always	ADV
ejpam-642	13	43	the	the	DET
ejpam-642	13	44	same	same	ADJ
ejpam-642	13	45	for	for	ADP
ejpam-642	13	46	both	both	DET
ejpam-642	13	47	mixture	mixture	NOUN
ejpam-642	13	48	components	component	NOUN
ejpam-642	13	49	.	.	PUNCT
ejpam-642	14	1	2000	2000	NUM
ejpam-642	14	2	mathematics	mathematic	NOUN
ejpam-642	14	3	subject	subject	NOUN
ejpam-642	14	4	classifications	classification	NOUN
ejpam-642	14	5	:	:	PUNCT
ejpam-642	14	6	62f07	62f07	NOUN
ejpam-642	14	7	,	,	PUNCT
ejpam-642	14	8	62f10	62f10	NUM
ejpam-642	14	9	,	,	PUNCT
ejpam-642	14	10	62j99	62j99	NUM
ejpam-642	14	11	,	,	PUNCT
ejpam-642	14	12	62p25	62p25	NUM
ejpam-642	14	13	key	key	ADJ
ejpam-642	14	14	words	word	NOUN
ejpam-642	14	15	and	and	CCONJ
ejpam-642	14	16	phrases	phrase	NOUN
ejpam-642	14	17	:	:	PUNCT
ejpam-642	14	18	record	record	NOUN
ejpam-642	14	19	linkage	linkage	NOUN
ejpam-642	14	20	,	,	PUNCT
ejpam-642	14	21	loglinear	loglinear	NOUN
ejpam-642	14	22	mixture	mixture	NOUN
ejpam-642	14	23	models	model	NOUN
ejpam-642	14	24	,	,	PUNCT
ejpam-642	14	25	em	em	PRON
ejpam-642	14	26	algorithm	algorithm	NOUN
ejpam-642	14	27	,	,	PUNCT
ejpam-642	14	28	model	model	NOUN
ejpam-642	14	29	selection	selection	NOUN
ejpam-642	14	30	,	,	PUNCT
ejpam-642	14	31	alternatively	alternatively	ADV
ejpam-642	14	32	climbing	climb	VERB
ejpam-642	14	33	pathways	pathway	NOUN
ejpam-642	14	34	∗corresponding	∗corresponde	VERB
ejpam-642	14	35	author	author	NOUN
ejpam-642	14	36	.	.	PUNCT
ejpam-642	15	1	email	email	NOUN
ejpam-642	15	2	addresses	address	NOUN
ejpam-642	15	3	:	:	PUNCT
ejpam-642	15	4	rzhu	rzhu	PROPN
ejpam-642	15	5	�	�	PROPN
ejpam-642	15	6	math.m	math.m	PROPN
ejpam-642	15	7	master	master	NOUN
ejpam-642	15	8	.	.	PUNCT
ejpam-642	16	1	a	a	DET
ejpam-642	16	2	(	(	PUNCT
ejpam-642	16	3	r.	r.	PROPN
ejpam-642	16	4	zhu	zhu	PROPN
ejpam-642	16	5	)	)	PUNCT
ejpam-642	16	6	,	,	PUNCT
ejpam-642	16	7	zhangjn	zhangjn	PROPN
ejpam-642	16	8	�	�	PROPN
ejpam-642	16	9	m	m	PROPN
ejpam-642	16	10	master	master	NOUN
ejpam-642	16	11	.	.	PUNCT
ejpam-642	17	1	a	a	DET
ejpam-642	17	2	(	(	PUNCT
ejpam-642	17	3	j.	j.	PROPN
ejpam-642	17	4	zhang),zhangd1	zhang),zhangd1	PROPN
ejpam-642	17	5	�	�	PROPN
ejpam-642	17	6	ohio.edu	ohio.edu	X
ejpam-642	17	7	(	(	PUNCT
ejpam-642	17	8	d.	d.	PROPN
ejpam-642	17	9	zhang	zhang	PROPN
ejpam-642	17	10	)	)	PUNCT
ejpam-642	17	11	,	,	PUNCT
ejpam-642	17	12	gyan	gyan	PROPN
ejpam-642	17	13	�	�	PROPN
ejpam-642	17	14	unb	unb	NOUN
ejpam-642	17	15	.	.	PUNCT
ejpam-642	18	1	a	a	DET
ejpam-642	18	2	(	(	PUNCT
ejpam-642	18	3	g.	g.	PROPN
ejpam-642	18	4	yan	yan	PROPN
ejpam-642	18	5	)	)	PUNCT
ejpam-642	18	6	http://www.ejpam.com	http://www.ejpam.com	X
ejpam-642	19	1	141	141	NUM
ejpam-642	19	2	c	c	X
ejpam-642	19	3	©	©	PROPN
ejpam-642	19	4	2010	2010	NUM
ejpam-642	19	5	ejpam	ejpam	NOUN
ejpam-642	19	6	all	all	DET
ejpam-642	19	7	rights	right	NOUN
ejpam-642	19	8	reserved	reserve	VERB
ejpam-642	19	9	.	.	PUNCT
ejpam-642	20	1	r.	r.	PROPN
ejpam-642	20	2	zhu	zhu	PROPN
ejpam-642	20	3	,	,	PUNCT
ejpam-642	20	4	j.	j.	PROPN
ejpam-642	20	5	zhang	zhang	PROPN
ejpam-642	20	6	,	,	PUNCT
ejpam-642	20	7	d.	d.	PROPN
ejpam-642	20	8	zhang	zhang	PROPN
ejpam-642	20	9	,	,	PUNCT
ejpam-642	20	10	g.	g.	PROPN
ejpam-642	20	11	yan	yan	PROPN
ejpam-642	20	12	/	/	SYM
ejpam-642	20	13	eur	eur	PROPN
ejpam-642	20	14	.	.	PUNCT
ejpam-642	21	1	j.	j.	PROPN
ejpam-642	21	2	pure	pure	PROPN
ejpam-642	21	3	appl	appl	PROPN
ejpam-642	21	4	.	.	PROPN
ejpam-642	21	5	math	math	PROPN
ejpam-642	21	6	,	,	PUNCT
ejpam-642	21	7	3	3	NUM
ejpam-642	21	8	(	(	PUNCT
ejpam-642	21	9	2010	2010	NUM
ejpam-642	21	10	)	)	PUNCT
ejpam-642	21	11	,	,	PUNCT
ejpam-642	21	12	141	141	NUM
ejpam-642	21	13	-	-	SYM
ejpam-642	21	14	162	162	NUM
ejpam-642	21	15	142	142	NUM
ejpam-642	21	16	1	1	NUM
ejpam-642	21	17	.	.	PUNCT
ejpam-642	22	1	introduction	introduction	NOUN
ejpam-642	22	2	information	information	NOUN
ejpam-642	22	3	often	often	ADV
ejpam-642	22	4	comes	come	VERB
ejpam-642	22	5	from	from	ADP
ejpam-642	22	6	various	various	ADJ
ejpam-642	22	7	sources	source	NOUN
ejpam-642	22	8	regarding	regard	VERB
ejpam-642	22	9	different	different	ADJ
ejpam-642	22	10	aspects	aspect	NOUN
ejpam-642	22	11	of	of	ADP
ejpam-642	22	12	objects	object	NOUN
ejpam-642	22	13	,	,	PUNCT
ejpam-642	22	14	and	and	CCONJ
ejpam-642	22	15	could	could	AUX
ejpam-642	22	16	be	be	AUX
ejpam-642	22	17	recorded	record	VERB
ejpam-642	22	18	at	at	ADP
ejpam-642	22	19	different	different	ADJ
ejpam-642	22	20	time	time	NOUN
ejpam-642	22	21	points	point	NOUN
ejpam-642	22	22	.	.	PUNCT
ejpam-642	23	1	for	for	ADP
ejpam-642	23	2	example	example	NOUN
ejpam-642	23	3	,	,	PUNCT
ejpam-642	23	4	the	the	DET
ejpam-642	23	5	provincial	provincial	ADJ
ejpam-642	23	6	resident	resident	ADJ
ejpam-642	23	7	registry	registry	NOUN
ejpam-642	23	8	information	information	NOUN
ejpam-642	23	9	is	be	AUX
ejpam-642	23	10	comprised	comprise	VERB
ejpam-642	23	11	of	of	ADP
ejpam-642	23	12	several	several	ADJ
ejpam-642	23	13	data	datum	NOUN
ejpam-642	23	14	sources	source	NOUN
ejpam-642	23	15	from	from	ADP
ejpam-642	23	16	different	different	ADJ
ejpam-642	23	17	departments	department	NOUN
ejpam-642	23	18	such	such	ADJ
ejpam-642	23	19	as	as	ADP
ejpam-642	23	20	transportation	transportation	NOUN
ejpam-642	23	21	,	,	PUNCT
ejpam-642	23	22	health	health	NOUN
ejpam-642	23	23	,	,	PUNCT
ejpam-642	23	24	license	license	NOUN
ejpam-642	23	25	and	and	CCONJ
ejpam-642	23	26	registration	registration	NOUN
ejpam-642	23	27	.	.	PUNCT
ejpam-642	24	1	these	these	DET
ejpam-642	24	2	data	datum	NOUN
ejpam-642	24	3	sets	set	NOUN
ejpam-642	24	4	include	include	VERB
ejpam-642	24	5	specific	specific	ADJ
ejpam-642	24	6	personal	personal	ADJ
ejpam-642	24	7	information	information	NOUN
ejpam-642	24	8	such	such	ADJ
ejpam-642	24	9	as	as	ADP
ejpam-642	24	10	name	name	NOUN
ejpam-642	24	11	,	,	PUNCT
ejpam-642	24	12	date	date	NOUN
ejpam-642	24	13	of	of	ADP
ejpam-642	24	14	birth	birth	NOUN
ejpam-642	24	15	and	and	CCONJ
ejpam-642	24	16	address	address	NOUN
ejpam-642	24	17	.	.	PUNCT
ejpam-642	25	1	typically	typically	ADV
ejpam-642	25	2	,	,	PUNCT
ejpam-642	25	3	there	there	PRON
ejpam-642	25	4	is	be	VERB
ejpam-642	25	5	no	no	DET
ejpam-642	25	6	unique	unique	ADJ
ejpam-642	25	7	i	i	PRON
ejpam-642	25	8	d	d	PROPN
ejpam-642	25	9	assigned	assign	VERB
ejpam-642	25	10	to	to	ADP
ejpam-642	25	11	individual	individual	ADJ
ejpam-642	25	12	objects	object	NOUN
ejpam-642	25	13	among	among	ADP
ejpam-642	25	14	all	all	DET
ejpam-642	25	15	data	datum	NOUN
ejpam-642	25	16	sources	source	NOUN
ejpam-642	25	17	.	.	PUNCT
ejpam-642	26	1	however	however	ADV
ejpam-642	26	2	,	,	PUNCT
ejpam-642	26	3	for	for	ADP
ejpam-642	26	4	some	some	DET
ejpam-642	26	5	reasons	reason	NOUN
ejpam-642	26	6	,	,	PUNCT
ejpam-642	26	7	like	like	ADP
ejpam-642	26	8	administration	administration	NOUN
ejpam-642	26	9	or	or	CCONJ
ejpam-642	26	10	detection	detection	NOUN
ejpam-642	26	11	of	of	ADP
ejpam-642	26	12	particular	particular	ADJ
ejpam-642	26	13	interested	interested	ADJ
ejpam-642	26	14	individuals	individual	NOUN
ejpam-642	26	15	(	(	PUNCT
ejpam-642	26	16	say	say	INTJ
ejpam-642	26	17	,	,	PUNCT
ejpam-642	26	18	terrorists	terrorist	NOUN
ejpam-642	26	19	)	)	PUNCT
ejpam-642	26	20	,	,	PUNCT
ejpam-642	26	21	the	the	DET
ejpam-642	26	22	need	need	NOUN
ejpam-642	26	23	arises	arise	VERB
ejpam-642	26	24	to	to	PART
ejpam-642	26	25	link	link	VERB
ejpam-642	26	26	two	two	NUM
ejpam-642	26	27	or	or	CCONJ
ejpam-642	26	28	more	more	ADJ
ejpam-642	26	29	data	datum	NOUN
ejpam-642	26	30	sources	source	NOUN
ejpam-642	26	31	together	together	ADV
ejpam-642	26	32	so	so	SCONJ
ejpam-642	26	33	that	that	SCONJ
ejpam-642	26	34	the	the	DET
ejpam-642	26	35	records	record	NOUN
ejpam-642	26	36	for	for	ADP
ejpam-642	26	37	the	the	DET
ejpam-642	26	38	same	same	ADJ
ejpam-642	26	39	individual	individual	ADJ
ejpam-642	26	40	object	object	NOUN
ejpam-642	26	41	across	across	ADP
ejpam-642	26	42	different	different	ADJ
ejpam-642	26	43	files	file	NOUN
ejpam-642	26	44	can	can	AUX
ejpam-642	26	45	be	be	AUX
ejpam-642	26	46	identified	identify	VERB
ejpam-642	26	47	for	for	ADP
ejpam-642	26	48	further	further	ADJ
ejpam-642	26	49	processing	processing	NOUN
ejpam-642	26	50	.	.	PUNCT
ejpam-642	27	1	this	this	DET
ejpam-642	27	2	task	task	NOUN
ejpam-642	27	3	also	also	ADV
ejpam-642	27	4	identifies	identify	VERB
ejpam-642	27	5	duplicate	duplicate	ADJ
ejpam-642	27	6	records	record	NOUN
ejpam-642	27	7	in	in	ADP
ejpam-642	27	8	one	one	NUM
ejpam-642	27	9	file	file	NOUN
ejpam-642	27	10	where	where	SCONJ
ejpam-642	27	11	a	a	DET
ejpam-642	27	12	second	second	ADJ
ejpam-642	27	13	virtual	virtual	ADJ
ejpam-642	27	14	file	file	NOUN
ejpam-642	27	15	is	be	AUX
ejpam-642	27	16	the	the	DET
ejpam-642	27	17	copy	copy	NOUN
ejpam-642	27	18	of	of	ADP
ejpam-642	27	19	the	the	DET
ejpam-642	27	20	original	original	ADJ
ejpam-642	27	21	one	one	NUM
ejpam-642	27	22	.	.	PUNCT
ejpam-642	28	1	the	the	DET
ejpam-642	28	2	way	way	NOUN
ejpam-642	28	3	to	to	PART
ejpam-642	28	4	decide	decide	VERB
ejpam-642	28	5	whether	whether	SCONJ
ejpam-642	28	6	two	two	NUM
ejpam-642	28	7	records	record	NOUN
ejpam-642	28	8	are	be	AUX
ejpam-642	28	9	the	the	DET
ejpam-642	28	10	same	same	ADJ
ejpam-642	28	11	or	or	CCONJ
ejpam-642	28	12	whether	whether	SCONJ
ejpam-642	28	13	two	two	NUM
ejpam-642	28	14	records	record	NOUN
ejpam-642	28	15	describe	describe	VERB
ejpam-642	28	16	the	the	DET
ejpam-642	28	17	same	same	ADJ
ejpam-642	28	18	object	object	NOUN
ejpam-642	28	19	is	be	AUX
ejpam-642	28	20	called	call	VERB
ejpam-642	28	21	record	record	NOUN
ejpam-642	28	22	linkage	linkage	NOUN
ejpam-642	28	23	.	.	PUNCT
ejpam-642	29	1	records	record	NOUN
ejpam-642	29	2	in	in	ADP
ejpam-642	29	3	different	different	ADJ
ejpam-642	29	4	data	data	NOUN
ejpam-642	29	5	files	file	NOUN
ejpam-642	29	6	consist	consist	VERB
ejpam-642	29	7	of	of	ADP
ejpam-642	29	8	common	common	ADJ
ejpam-642	29	9	attributes	attribute	NOUN
ejpam-642	29	10	(	(	PUNCT
ejpam-642	29	11	such	such	ADJ
ejpam-642	29	12	as	as	ADP
ejpam-642	29	13	name	name	NOUN
ejpam-642	29	14	,	,	PUNCT
ejpam-642	29	15	address	address	NOUN
ejpam-642	29	16	and	and	CCONJ
ejpam-642	29	17	date	date	NOUN
ejpam-642	29	18	of	of	ADP
ejpam-642	29	19	birth	birth	NOUN
ejpam-642	29	20	)	)	PUNCT
ejpam-642	29	21	and	and	CCONJ
ejpam-642	29	22	different	different	ADJ
ejpam-642	29	23	attributes	attribute	NOUN
ejpam-642	29	24	(	(	PUNCT
ejpam-642	29	25	such	such	ADJ
ejpam-642	29	26	as	as	ADP
ejpam-642	29	27	weight	weight	NOUN
ejpam-642	29	28	in	in	ADP
ejpam-642	29	29	one	one	NUM
ejpam-642	29	30	file	file	NOUN
ejpam-642	29	31	and	and	CCONJ
ejpam-642	29	32	height	height	NOUN
ejpam-642	29	33	in	in	ADP
ejpam-642	29	34	another	another	DET
ejpam-642	29	35	file	file	NOUN
ejpam-642	29	36	)	)	PUNCT
ejpam-642	29	37	for	for	ADP
ejpam-642	29	38	individual	individual	ADJ
ejpam-642	29	39	objects	object	NOUN
ejpam-642	29	40	.	.	PUNCT
ejpam-642	30	1	usually	usually	ADV
ejpam-642	30	2	,	,	PUNCT
ejpam-642	30	3	the	the	DET
ejpam-642	30	4	common	common	ADJ
ejpam-642	30	5	attributes	attribute	NOUN
ejpam-642	30	6	are	be	AUX
ejpam-642	30	7	chosen	choose	VERB
ejpam-642	30	8	to	to	PART
ejpam-642	30	9	be	be	AUX
ejpam-642	30	10	identifiers	identifier	NOUN
ejpam-642	30	11	for	for	ADP
ejpam-642	30	12	record	record	NOUN
ejpam-642	30	13	matching	matching	NOUN
ejpam-642	30	14	.	.	PUNCT
ejpam-642	31	1	however	however	ADV
ejpam-642	31	2	,	,	PUNCT
ejpam-642	31	3	the	the	DET
ejpam-642	31	4	different	different	ADJ
ejpam-642	31	5	attributes	attribute	NOUN
ejpam-642	31	6	in	in	ADP
ejpam-642	31	7	two	two	NUM
ejpam-642	31	8	files	file	NOUN
ejpam-642	31	9	may	may	AUX
ejpam-642	31	10	provide	provide	VERB
ejpam-642	31	11	partial	partial	ADJ
ejpam-642	31	12	information	information	NOUN
ejpam-642	31	13	for	for	ADP
ejpam-642	31	14	the	the	DET
ejpam-642	31	15	matching	matching	NOUN
ejpam-642	31	16	if	if	SCONJ
ejpam-642	31	17	these	these	DET
ejpam-642	31	18	attributes	attribute	NOUN
ejpam-642	31	19	are	be	AUX
ejpam-642	31	20	highly	highly	ADV
ejpam-642	31	21	associated	associate	VERB
ejpam-642	31	22	.	.	PUNCT
ejpam-642	32	1	in	in	ADP
ejpam-642	32	2	this	this	DET
ejpam-642	32	3	study	study	NOUN
ejpam-642	32	4	,	,	PUNCT
ejpam-642	32	5	we	we	PRON
ejpam-642	32	6	only	only	ADV
ejpam-642	32	7	consider	consider	VERB
ejpam-642	32	8	the	the	DET
ejpam-642	32	9	common	common	ADJ
ejpam-642	32	10	attributes	attribute	NOUN
ejpam-642	32	11	as	as	ADP
ejpam-642	32	12	the	the	DET
ejpam-642	32	13	identifiers	identifier	NOUN
ejpam-642	32	14	.	.	PUNCT
ejpam-642	33	1	comparison	comparison	NOUN
ejpam-642	33	2	vectors	vector	NOUN
ejpam-642	33	3	are	be	AUX
ejpam-642	33	4	then	then	ADV
ejpam-642	33	5	obtained	obtain	VERB
ejpam-642	33	6	for	for	ADP
ejpam-642	33	7	each	each	DET
ejpam-642	33	8	record	record	NOUN
ejpam-642	33	9	pair	pair	NOUN
ejpam-642	33	10	by	by	ADP
ejpam-642	33	11	comparing	compare	VERB
ejpam-642	33	12	the	the	DET
ejpam-642	33	13	values	value	NOUN
ejpam-642	33	14	of	of	ADP
ejpam-642	33	15	identifiers	identifier	NOUN
ejpam-642	33	16	.	.	PUNCT
ejpam-642	34	1	the	the	DET
ejpam-642	34	2	status	status	NOUN
ejpam-642	34	3	of	of	ADP
ejpam-642	34	4	a	a	DET
ejpam-642	34	5	pair	pair	NOUN
ejpam-642	34	6	of	of	ADP
ejpam-642	34	7	records	record	NOUN
ejpam-642	34	8	for	for	ADP
ejpam-642	34	9	comparison	comparison	NOUN
ejpam-642	34	10	is	be	AUX
ejpam-642	34	11	either	either	CCONJ
ejpam-642	34	12	“	"	PUNCT
ejpam-642	34	13	matched	match	VERB
ejpam-642	34	14	”	"	PUNCT
ejpam-642	34	15	or	or	CCONJ
ejpam-642	34	16	“	"	PUNCT
ejpam-642	34	17	unmatched	unmatched	ADJ
ejpam-642	34	18	”	"	PUNCT
ejpam-642	34	19	.	.	PUNCT
ejpam-642	35	1	ideally	ideally	ADV
ejpam-642	35	2	,	,	PUNCT
ejpam-642	35	3	if	if	SCONJ
ejpam-642	35	4	the	the	DET
ejpam-642	35	5	values	value	NOUN
ejpam-642	35	6	of	of	ADP
ejpam-642	35	7	identifiers	identifier	NOUN
ejpam-642	35	8	match	match	VERB
ejpam-642	35	9	exactly	exactly	ADV
ejpam-642	35	10	for	for	ADP
ejpam-642	35	11	a	a	DET
ejpam-642	35	12	record	record	NOUN
ejpam-642	35	13	pair	pair	NOUN
ejpam-642	35	14	,	,	PUNCT
ejpam-642	35	15	then	then	ADV
ejpam-642	35	16	the	the	DET
ejpam-642	35	17	two	two	NUM
ejpam-642	35	18	records	record	NOUN
ejpam-642	35	19	are	be	AUX
ejpam-642	35	20	from	from	ADP
ejpam-642	35	21	the	the	DET
ejpam-642	35	22	same	same	ADJ
ejpam-642	35	23	object	object	NOUN
ejpam-642	35	24	.	.	PUNCT
ejpam-642	36	1	however	however	ADV
ejpam-642	36	2	,	,	PUNCT
ejpam-642	36	3	in	in	ADP
ejpam-642	36	4	reality	reality	NOUN
ejpam-642	36	5	,	,	PUNCT
ejpam-642	36	6	there	there	PRON
ejpam-642	36	7	are	be	VERB
ejpam-642	36	8	many	many	ADJ
ejpam-642	36	9	situations	situation	NOUN
ejpam-642	36	10	where	where	SCONJ
ejpam-642	36	11	the	the	DET
ejpam-642	36	12	values	value	NOUN
ejpam-642	36	13	of	of	ADP
ejpam-642	36	14	identifiers	identifier	NOUN
ejpam-642	36	15	may	may	AUX
ejpam-642	36	16	not	not	PART
ejpam-642	36	17	be	be	AUX
ejpam-642	36	18	the	the	DET
ejpam-642	36	19	same	same	ADJ
ejpam-642	36	20	even	even	ADV
ejpam-642	36	21	if	if	SCONJ
ejpam-642	36	22	they	they	PRON
ejpam-642	36	23	are	be	AUX
ejpam-642	36	24	from	from	ADP
ejpam-642	36	25	the	the	DET
ejpam-642	36	26	same	same	ADJ
ejpam-642	36	27	object	object	NOUN
ejpam-642	36	28	.	.	PUNCT
ejpam-642	37	1	for	for	ADP
ejpam-642	37	2	example	example	NOUN
ejpam-642	37	3	,	,	PUNCT
ejpam-642	37	4	the	the	DET
ejpam-642	37	5	surname	surname	NOUN
ejpam-642	37	6	of	of	ADP
ejpam-642	37	7	a	a	DET
ejpam-642	37	8	female	female	NOUN
ejpam-642	37	9	may	may	AUX
ejpam-642	37	10	change	change	VERB
ejpam-642	37	11	when	when	SCONJ
ejpam-642	37	12	her	her	PRON
ejpam-642	37	13	marriage	marriage	NOUN
ejpam-642	37	14	status	status	NOUN
ejpam-642	37	15	changes	change	NOUN
ejpam-642	37	16	.	.	PUNCT
ejpam-642	38	1	hence	hence	ADV
ejpam-642	38	2	,	,	PUNCT
ejpam-642	38	3	the	the	DET
ejpam-642	38	4	method	method	NOUN
ejpam-642	38	5	of	of	ADP
ejpam-642	38	6	exact	exact	ADJ
ejpam-642	38	7	match	match	NOUN
ejpam-642	38	8	or	or	CCONJ
ejpam-642	38	9	deterministic	deterministic	ADJ
ejpam-642	38	10	match	match	NOUN
ejpam-642	38	11	may	may	AUX
ejpam-642	38	12	miss	miss	VERB
ejpam-642	38	13	many	many	ADJ
ejpam-642	38	14	matched	match	VERB
ejpam-642	38	15	records	record	NOUN
ejpam-642	38	16	.	.	PUNCT
ejpam-642	39	1	on	on	ADP
ejpam-642	39	2	the	the	DET
ejpam-642	39	3	other	other	ADJ
ejpam-642	39	4	hand	hand	NOUN
ejpam-642	39	5	,	,	PUNCT
ejpam-642	39	6	exact	exact	ADJ
ejpam-642	39	7	match	match	NOUN
ejpam-642	39	8	may	may	AUX
ejpam-642	39	9	not	not	PART
ejpam-642	39	10	always	always	ADV
ejpam-642	39	11	imply	imply	VERB
ejpam-642	39	12	the	the	DET
ejpam-642	39	13	same	same	ADJ
ejpam-642	39	14	object	object	NOUN
ejpam-642	39	15	.	.	PUNCT
ejpam-642	40	1	for	for	ADP
ejpam-642	40	2	instance	instance	NOUN
ejpam-642	40	3	,	,	PUNCT
ejpam-642	40	4	if	if	SCONJ
ejpam-642	40	5	the	the	DET
ejpam-642	40	6	identifiers	identifier	NOUN
ejpam-642	40	7	are	be	AUX
ejpam-642	40	8	just	just	ADV
ejpam-642	40	9	name	name	NOUN
ejpam-642	40	10	and	and	CCONJ
ejpam-642	40	11	date	date	NOUN
ejpam-642	40	12	of	of	ADP
ejpam-642	40	13	birth	birth	NOUN
ejpam-642	40	14	,	,	PUNCT
ejpam-642	40	15	it	it	PRON
ejpam-642	40	16	could	could	AUX
ejpam-642	40	17	occur	occur	VERB
ejpam-642	40	18	that	that	SCONJ
ejpam-642	40	19	two	two	NUM
ejpam-642	40	20	different	different	ADJ
ejpam-642	40	21	persons	person	NOUN
ejpam-642	40	22	have	have	VERB
ejpam-642	40	23	the	the	DET
ejpam-642	40	24	same	same	ADJ
ejpam-642	40	25	name	name	NOUN
ejpam-642	40	26	and	and	CCONJ
ejpam-642	40	27	date	date	NOUN
ejpam-642	40	28	of	of	ADP
ejpam-642	40	29	birth	birth	NOUN
ejpam-642	40	30	although	although	SCONJ
ejpam-642	40	31	the	the	DET
ejpam-642	40	32	chance	chance	NOUN
ejpam-642	40	33	is	be	AUX
ejpam-642	40	34	small	small	ADJ
ejpam-642	40	35	.	.	PUNCT
ejpam-642	41	1	for	for	ADP
ejpam-642	41	2	this	this	DET
ejpam-642	41	3	reason	reason	NOUN
ejpam-642	41	4	,	,	PUNCT
ejpam-642	41	5	probabilistic	probabilistic	ADJ
ejpam-642	41	6	match	match	NOUN
ejpam-642	41	7	is	be	AUX
ejpam-642	41	8	employed	employ	VERB
ejpam-642	41	9	where	where	SCONJ
ejpam-642	41	10	the	the	DET
ejpam-642	41	11	pair	pair	NOUN
ejpam-642	41	12	of	of	ADP
ejpam-642	41	13	records	record	NOUN
ejpam-642	41	14	is	be	AUX
ejpam-642	41	15	regarded	regard	VERB
ejpam-642	41	16	as	as	ADP
ejpam-642	41	17	from	from	ADP
ejpam-642	41	18	the	the	DET
ejpam-642	41	19	same	same	ADJ
ejpam-642	41	20	object	object	NOUN
ejpam-642	41	21	if	if	SCONJ
ejpam-642	41	22	the	the	DET
ejpam-642	41	23	weight	weight	NOUN
ejpam-642	41	24	involving	involve	VERB
ejpam-642	41	25	the	the	DET
ejpam-642	41	26	matched	match	VERB
ejpam-642	41	27	and	and	CCONJ
ejpam-642	41	28	unmatched	unmatched	ADJ
ejpam-642	41	29	probabilities	probability	NOUN
ejpam-642	41	30	is	be	AUX
ejpam-642	41	31	large	large	ADJ
ejpam-642	41	32	enough	enough	ADV
ejpam-642	41	33	.	.	PUNCT
ejpam-642	42	1	historically	historically	ADV
ejpam-642	42	2	,	,	PUNCT
ejpam-642	42	3	newcombe	newcombe	PROPN
ejpam-642	42	4	et	et	PROPN
ejpam-642	42	5	al	al	PROPN
ejpam-642	42	6	.	.	PROPN
ejpam-642	42	7	(	(	PUNCT
ejpam-642	42	8	1959	1959	NUM
ejpam-642	42	9	)	)	PUNCT
ejpam-642	42	10	studied	study	VERB
ejpam-642	42	11	probabilistic	probabilistic	ADJ
ejpam-642	42	12	linkage	linkage	NOUN
ejpam-642	42	13	on	on	ADP
ejpam-642	42	14	vital	vital	ADJ
ejpam-642	42	15	records	record	NOUN
ejpam-642	42	16	;	;	PUNCT
ejpam-642	42	17	refer	refer	VERB
ejpam-642	42	18	to	to	ADP
ejpam-642	42	19	[	[	X
ejpam-642	42	20	13	13	NUM
ejpam-642	42	21	]	]	PUNCT
ejpam-642	42	22	.	.	PUNCT
ejpam-642	43	1	fellegi	fellegi	PROPN
ejpam-642	43	2	and	and	CCONJ
ejpam-642	43	3	sunter	sunter	NOUN
ejpam-642	43	4	(	(	PUNCT
ejpam-642	43	5	1969	1969	NUM
ejpam-642	43	6	)	)	PUNCT
ejpam-642	43	7	developed	develop	VERB
ejpam-642	43	8	the	the	DET
ejpam-642	43	9	formal	formal	ADJ
ejpam-642	43	10	mathematical	mathematical	ADJ
ejpam-642	43	11	framework	framework	NOUN
ejpam-642	43	12	for	for	ADP
ejpam-642	43	13	probabilistic	probabilistic	ADJ
ejpam-642	43	14	record	record	NOUN
ejpam-642	43	15	linkage	linkage	NOUN
ejpam-642	43	16	,	,	PUNCT
ejpam-642	43	17	and	and	CCONJ
ejpam-642	43	18	they	they	PRON
ejpam-642	43	19	also	also	ADV
ejpam-642	43	20	specified	specify	VERB
ejpam-642	43	21	models	model	NOUN
ejpam-642	43	22	in	in	ADP
ejpam-642	43	23	a	a	DET
ejpam-642	43	24	hierarchical	hierarchical	ADJ
ejpam-642	43	25	way	way	NOUN
ejpam-642	43	26	according	accord	VERB
ejpam-642	43	27	to	to	ADP
ejpam-642	43	28	a	a	DET
ejpam-642	43	29	possible	possible	ADJ
ejpam-642	43	30	error	error	NOUN
ejpam-642	43	31	-	-	PUNCT
ejpam-642	43	32	making	make	VERB
ejpam-642	43	33	mechanism	mechanism	NOUN
ejpam-642	43	34	;	;	PUNCT
ejpam-642	43	35	see	see	VERB
ejpam-642	43	36	[	[	X
ejpam-642	43	37	6	6	NUM
ejpam-642	43	38	]	]	PUNCT
ejpam-642	43	39	.	.	PUNCT
ejpam-642	44	1	after	after	ADP
ejpam-642	44	2	that	that	PRON
ejpam-642	44	3	,	,	PUNCT
ejpam-642	44	4	many	many	ADJ
ejpam-642	44	5	techniques	technique	NOUN
ejpam-642	44	6	and	and	CCONJ
ejpam-642	44	7	software	software	NOUN
ejpam-642	44	8	have	have	AUX
ejpam-642	44	9	been	be	AUX
ejpam-642	44	10	developed	develop	VERB
ejpam-642	44	11	for	for	ADP
ejpam-642	44	12	record	record	NOUN
ejpam-642	44	13	linkage	linkage	NOUN
ejpam-642	44	14	.	.	PUNCT
ejpam-642	45	1	for	for	ADP
ejpam-642	45	2	example	example	NOUN
ejpam-642	45	3	,	,	PUNCT
ejpam-642	45	4	blocking	blocking	NOUN
ejpam-642	45	5	is	be	AUX
ejpam-642	45	6	used	use	VERB
ejpam-642	45	7	to	to	PART
ejpam-642	45	8	reduce	reduce	VERB
ejpam-642	45	9	not	not	PART
ejpam-642	45	10	only	only	ADV
ejpam-642	45	11	the	the	DET
ejpam-642	45	12	size	size	NOUN
ejpam-642	45	13	of	of	ADP
ejpam-642	45	14	comparison	comparison	NOUN
ejpam-642	45	15	pairs	pair	NOUN
ejpam-642	45	16	,	,	PUNCT
ejpam-642	45	17	but	but	CCONJ
ejpam-642	45	18	also	also	ADV
ejpam-642	45	19	the	the	DET
ejpam-642	45	20	dependence	dependence	NOUN
ejpam-642	45	21	effects	effect	NOUN
ejpam-642	45	22	(	(	PUNCT
ejpam-642	45	23	i.e.	i.e.	X
ejpam-642	45	24	,	,	PUNCT
ejpam-642	45	25	interaction	interaction	NOUN
ejpam-642	45	26	between	between	ADP
ejpam-642	45	27	identifiers	identifier	NOUN
ejpam-642	45	28	)	)	PUNCT
ejpam-642	45	29	.	.	PUNCT
ejpam-642	46	1	nowadays	nowadays	ADV
ejpam-642	46	2	record	record	NOUN
ejpam-642	46	3	linkage	linkage	NOUN
ejpam-642	46	4	methodology	methodology	NOUN
ejpam-642	46	5	is	be	AUX
ejpam-642	46	6	used	use	VERB
ejpam-642	46	7	in	in	ADP
ejpam-642	46	8	many	many	ADJ
ejpam-642	46	9	areas	area	NOUN
ejpam-642	46	10	such	such	ADJ
ejpam-642	46	11	as	as	ADP
ejpam-642	46	12	census	census	NOUN
ejpam-642	46	13	,	,	PUNCT
ejpam-642	46	14	survey	survey	NOUN
ejpam-642	46	15	,	,	PUNCT
ejpam-642	46	16	administration	administration	NOUN
ejpam-642	46	17	and	and	CCONJ
ejpam-642	46	18	medical	medical	ADJ
ejpam-642	46	19	research	research	NOUN
ejpam-642	46	20	;	;	PUNCT
ejpam-642	46	21	see	see	VERB
ejpam-642	46	22	[	[	X
ejpam-642	46	23	1	1	NUM
ejpam-642	46	24	,	,	PUNCT
ejpam-642	46	25	3	3	NUM
ejpam-642	46	26	,	,	PUNCT
ejpam-642	46	27	8	8	NUM
ejpam-642	46	28	,	,	PUNCT
ejpam-642	46	29	12	12	NUM
ejpam-642	46	30	]	]	PUNCT
ejpam-642	46	31	.	.	PUNCT
ejpam-642	47	1	regarding	regard	VERB
ejpam-642	47	2	modelling	modelling	NOUN
ejpam-642	47	3	,	,	PUNCT
ejpam-642	47	4	the	the	DET
ejpam-642	47	5	approach	approach	NOUN
ejpam-642	47	6	applying	apply	VERB
ejpam-642	47	7	a	a	DET
ejpam-642	47	8	mixture	mixture	NOUN
ejpam-642	47	9	of	of	ADP
ejpam-642	47	10	two	two	NUM
ejpam-642	47	11	components	component	NOUN
ejpam-642	47	12	to	to	ADP
ejpam-642	47	13	the	the	DET
ejpam-642	47	14	matched	match	VERB
ejpam-642	47	15	and	and	CCONJ
ejpam-642	47	16	unmatched	unmatched	ADJ
ejpam-642	47	17	record	record	NOUN
ejpam-642	47	18	pairs	pair	NOUN
ejpam-642	47	19	seems	seem	VERB
ejpam-642	47	20	to	to	PART
ejpam-642	47	21	be	be	AUX
ejpam-642	47	22	more	more	ADV
ejpam-642	47	23	convenient	convenient	ADJ
ejpam-642	47	24	,	,	PUNCT
ejpam-642	47	25	because	because	SCONJ
ejpam-642	47	26	it	it	PRON
ejpam-642	47	27	does	do	AUX
ejpam-642	47	28	not	not	PART
ejpam-642	47	29	require	require	VERB
ejpam-642	47	30	the	the	DET
ejpam-642	47	31	details	detail	NOUN
ejpam-642	47	32	of	of	ADP
ejpam-642	47	33	the	the	DET
ejpam-642	47	34	error	error	NOUN
ejpam-642	47	35	-	-	PUNCT
ejpam-642	47	36	making	make	VERB
ejpam-642	47	37	mechanism	mechanism	NOUN
ejpam-642	47	38	.	.	PUNCT
ejpam-642	48	1	jaro	jaro	PROPN
ejpam-642	48	2	(	(	PUNCT
ejpam-642	48	3	1989	1989	NUM
ejpam-642	48	4	)	)	PUNCT
ejpam-642	48	5	considered	consider	VERB
ejpam-642	48	6	a	a	DET
ejpam-642	48	7	mixture	mixture	NOUN
ejpam-642	48	8	model	model	NOUN
ejpam-642	48	9	for	for	ADP
ejpam-642	48	10	the	the	DET
ejpam-642	48	11	simplest	simple	ADJ
ejpam-642	48	12	contingency	contingency	NOUN
ejpam-642	48	13	table	table	NOUN
ejpam-642	48	14	derived	derive	VERB
ejpam-642	48	15	from	from	ADP
ejpam-642	48	16	comparison	comparison	NOUN
ejpam-642	48	17	vectors	vector	NOUN
ejpam-642	48	18	where	where	SCONJ
ejpam-642	48	19	fields	field	NOUN
ejpam-642	48	20	are	be	AUX
ejpam-642	48	21	binary	binary	ADJ
ejpam-642	48	22	and	and	CCONJ
ejpam-642	48	23	independent	independent	ADJ
ejpam-642	48	24	,	,	PUNCT
ejpam-642	48	25	r.	r.	PROPN
ejpam-642	48	26	zhu	zhu	PROPN
ejpam-642	48	27	,	,	PUNCT
ejpam-642	48	28	j.	j.	PROPN
ejpam-642	48	29	zhang	zhang	PROPN
ejpam-642	48	30	,	,	PUNCT
ejpam-642	48	31	d.	d.	PROPN
ejpam-642	48	32	zhang	zhang	PROPN
ejpam-642	48	33	,	,	PUNCT
ejpam-642	48	34	g.	g.	PROPN
ejpam-642	48	35	yan	yan	PROPN
ejpam-642	48	36	/	/	SYM
ejpam-642	48	37	eur	eur	PROPN
ejpam-642	48	38	.	.	PUNCT
ejpam-642	49	1	j.	j.	PROPN
ejpam-642	49	2	pure	pure	PROPN
ejpam-642	49	3	appl	appl	PROPN
ejpam-642	49	4	.	.	PROPN
ejpam-642	49	5	math	math	PROPN
ejpam-642	49	6	,	,	PUNCT
ejpam-642	49	7	3	3	NUM
ejpam-642	49	8	(	(	PUNCT
ejpam-642	49	9	2010	2010	NUM
ejpam-642	49	10	)	)	PUNCT
ejpam-642	49	11	,	,	PUNCT
ejpam-642	49	12	141	141	NUM
ejpam-642	49	13	-	-	SYM
ejpam-642	49	14	162	162	NUM
ejpam-642	49	15	143	143	NUM
ejpam-642	49	16	and	and	CCONJ
ejpam-642	49	17	used	use	VERB
ejpam-642	49	18	the	the	DET
ejpam-642	49	19	em	em	PROPN
ejpam-642	49	20	algorithm	algorithm	NOUN
ejpam-642	49	21	for	for	ADP
ejpam-642	49	22	parameter	parameter	NOUN
ejpam-642	49	23	estimation	estimation	NOUN
ejpam-642	49	24	with	with	ADP
ejpam-642	49	25	application	application	NOUN
ejpam-642	49	26	in	in	ADP
ejpam-642	49	27	a	a	DET
ejpam-642	49	28	test	test	NOUN
ejpam-642	49	29	census	census	NOUN
ejpam-642	49	30	of	of	ADP
ejpam-642	49	31	tampa	tampa	PROPN
ejpam-642	49	32	in	in	ADP
ejpam-642	49	33	1985	1985	NUM
ejpam-642	49	34	;	;	PUNCT
ejpam-642	49	35	see	see	VERB
ejpam-642	49	36	[	[	X
ejpam-642	49	37	7	7	NUM
ejpam-642	49	38	]	]	PUNCT
ejpam-642	49	39	.	.	PUNCT
ejpam-642	50	1	winkler	winkler	NOUN
ejpam-642	50	2	(	(	PUNCT
ejpam-642	50	3	1989	1989	NUM
ejpam-642	50	4	)	)	PUNCT
ejpam-642	50	5	presented	present	VERB
ejpam-642	50	6	a	a	DET
ejpam-642	50	7	loglinear	loglinear	NOUN
ejpam-642	50	8	model	model	NOUN
ejpam-642	50	9	to	to	PART
ejpam-642	50	10	adjust	adjust	VERB
ejpam-642	50	11	the	the	DET
ejpam-642	50	12	lack	lack	NOUN
ejpam-642	50	13	of	of	ADP
ejpam-642	50	14	independence	independence	NOUN
ejpam-642	50	15	;	;	PUNCT
ejpam-642	50	16	refer	refer	VERB
ejpam-642	50	17	to	to	ADP
ejpam-642	50	18	[	[	X
ejpam-642	50	19	16	16	NUM
ejpam-642	50	20	]	]	PUNCT
ejpam-642	50	21	.	.	PUNCT
ejpam-642	51	1	more	more	ADJ
ejpam-642	51	2	work	work	NOUN
ejpam-642	51	3	on	on	ADP
ejpam-642	51	4	loglinear	loglinear	NOUN
ejpam-642	51	5	models	model	NOUN
ejpam-642	51	6	can	can	AUX
ejpam-642	51	7	be	be	AUX
ejpam-642	51	8	found	find	VERB
ejpam-642	51	9	in	in	ADP
ejpam-642	51	10	[	[	X
ejpam-642	51	11	8	8	NUM
ejpam-642	51	12	,	,	PUNCT
ejpam-642	51	13	9	9	NUM
ejpam-642	51	14	,	,	PUNCT
ejpam-642	51	15	10	10	NUM
ejpam-642	51	16	]	]	PUNCT
ejpam-642	51	17	,	,	PUNCT
ejpam-642	51	18	and	and	CCONJ
ejpam-642	51	19	references	reference	NOUN
ejpam-642	51	20	therein	therein	ADV
ejpam-642	51	21	.	.	PUNCT
ejpam-642	52	1	we	we	PRON
ejpam-642	52	2	focus	focus	VERB
ejpam-642	52	3	on	on	ADP
ejpam-642	52	4	mixture	mixture	NOUN
ejpam-642	52	5	models	model	NOUN
ejpam-642	52	6	.	.	PUNCT
ejpam-642	53	1	for	for	ADP
ejpam-642	53	2	a	a	DET
ejpam-642	53	3	given	give	VERB
ejpam-642	53	4	record	record	NOUN
ejpam-642	53	5	linkage	linkage	NOUN
ejpam-642	53	6	task	task	NOUN
ejpam-642	53	7	,	,	PUNCT
ejpam-642	53	8	many	many	ADJ
ejpam-642	53	9	mixture	mixture	NOUN
ejpam-642	53	10	models	model	NOUN
ejpam-642	53	11	can	can	AUX
ejpam-642	53	12	be	be	AUX
ejpam-642	53	13	specified	specify	VERB
ejpam-642	53	14	.	.	PUNCT
ejpam-642	54	1	however	however	ADV
ejpam-642	54	2	,	,	PUNCT
ejpam-642	54	3	which	which	DET
ejpam-642	54	4	model	model	NOUN
ejpam-642	54	5	fits	fit	VERB
ejpam-642	54	6	the	the	DET
ejpam-642	54	7	observed	observed	ADJ
ejpam-642	54	8	comparison	comparison	NOUN
ejpam-642	54	9	pairs	pair	VERB
ejpam-642	54	10	better	well	ADV
ejpam-642	54	11	,	,	PUNCT
ejpam-642	54	12	and	and	CCONJ
ejpam-642	54	13	how	how	SCONJ
ejpam-642	54	14	can	can	AUX
ejpam-642	54	15	underfitting	underfitte	VERB
ejpam-642	54	16	or	or	CCONJ
ejpam-642	54	17	overfitting	overfitting	NOUN
ejpam-642	54	18	be	be	AUX
ejpam-642	54	19	avoided	avoid	VERB
ejpam-642	54	20	?	?	PUNCT
ejpam-642	55	1	these	these	PRON
ejpam-642	55	2	are	be	AUX
ejpam-642	55	3	questions	question	NOUN
ejpam-642	55	4	for	for	ADP
ejpam-642	55	5	record	record	NOUN
ejpam-642	55	6	linkage	linkage	NOUN
ejpam-642	55	7	practitioners	practitioner	NOUN
ejpam-642	55	8	.	.	PUNCT
ejpam-642	56	1	our	our	PRON
ejpam-642	56	2	study	study	NOUN
ejpam-642	56	3	is	be	AUX
ejpam-642	56	4	motivated	motivate	VERB
ejpam-642	56	5	by	by	ADP
ejpam-642	56	6	the	the	DET
ejpam-642	56	7	record	record	NOUN
ejpam-642	56	8	linkage	linkage	NOUN
ejpam-642	56	9	case	case	NOUN
ejpam-642	56	10	study	study	NOUN
ejpam-642	56	11	in	in	ADP
ejpam-642	56	12	2006	2006	NUM
ejpam-642	56	13	annual	annual	ADJ
ejpam-642	56	14	meeting	meeting	NOUN
ejpam-642	56	15	of	of	ADP
ejpam-642	56	16	statistical	statistical	ADJ
ejpam-642	56	17	society	society	NOUN
ejpam-642	56	18	of	of	ADP
ejpam-642	56	19	canada	canada	PROPN
ejpam-642	56	20	(	(	PUNCT
ejpam-642	56	21	ssc	ssc	PROPN
ejpam-642	56	22	2006	2006	NUM
ejpam-642	56	23	)	)	PUNCT
ejpam-642	56	24	.	.	PUNCT
ejpam-642	57	1	we	we	PRON
ejpam-642	57	2	tried	try	VERB
ejpam-642	57	3	some	some	DET
ejpam-642	57	4	free	free	ADJ
ejpam-642	57	5	software	software	NOUN
ejpam-642	57	6	for	for	ADP
ejpam-642	57	7	record	record	NOUN
ejpam-642	57	8	linkage	linkage	NOUN
ejpam-642	57	9	and	and	CCONJ
ejpam-642	57	10	found	find	VERB
ejpam-642	57	11	that	that	SCONJ
ejpam-642	57	12	their	their	PRON
ejpam-642	57	13	models	model	NOUN
ejpam-642	57	14	assume	assume	VERB
ejpam-642	57	15	field	field	NOUN
ejpam-642	57	16	independence	independence	NOUN
ejpam-642	57	17	without	without	ADP
ejpam-642	57	18	interactions	interaction	NOUN
ejpam-642	57	19	.	.	PUNCT
ejpam-642	58	1	the	the	DET
ejpam-642	58	2	possible	possible	ADJ
ejpam-642	58	3	reason	reason	NOUN
ejpam-642	58	4	for	for	ADP
ejpam-642	58	5	this	this	DET
ejpam-642	58	6	limitation	limitation	NOUN
ejpam-642	58	7	could	could	AUX
ejpam-642	58	8	be	be	AUX
ejpam-642	58	9	the	the	DET
ejpam-642	58	10	computational	computational	ADJ
ejpam-642	58	11	burden	burden	NOUN
ejpam-642	58	12	for	for	ADP
ejpam-642	58	13	arbitrary	arbitrary	ADJ
ejpam-642	58	14	mixture	mixture	NOUN
ejpam-642	58	15	models	model	NOUN
ejpam-642	58	16	.	.	PUNCT
ejpam-642	59	1	however	however	ADV
ejpam-642	59	2	,	,	PUNCT
ejpam-642	59	3	the	the	DET
ejpam-642	59	4	lack	lack	NOUN
ejpam-642	59	5	of	of	ADP
ejpam-642	59	6	fit	fit	NOUN
ejpam-642	59	7	is	be	AUX
ejpam-642	59	8	obvious	obvious	ADJ
ejpam-642	59	9	in	in	ADP
ejpam-642	59	10	this	this	DET
ejpam-642	59	11	simple	simple	ADJ
ejpam-642	59	12	model	model	NOUN
ejpam-642	59	13	if	if	SCONJ
ejpam-642	59	14	fields	field	NOUN
ejpam-642	59	15	are	be	AUX
ejpam-642	59	16	dependent	dependent	ADJ
ejpam-642	59	17	.	.	PUNCT
ejpam-642	60	1	on	on	ADP
ejpam-642	60	2	the	the	DET
ejpam-642	60	3	other	other	ADJ
ejpam-642	60	4	hand	hand	NOUN
ejpam-642	60	5	,	,	PUNCT
ejpam-642	60	6	subjectively	subjectively	ADV
ejpam-642	60	7	specifying	specify	VERB
ejpam-642	60	8	models	model	NOUN
ejpam-642	60	9	with	with	ADP
ejpam-642	60	10	higher	high	ADJ
ejpam-642	60	11	order	order	NOUN
ejpam-642	60	12	interactions	interaction	NOUN
ejpam-642	60	13	may	may	AUX
ejpam-642	60	14	lead	lead	VERB
ejpam-642	60	15	to	to	ADP
ejpam-642	60	16	overfitting	overfitte	VERB
ejpam-642	60	17	for	for	ADP
ejpam-642	60	18	one	one	NUM
ejpam-642	60	19	or	or	CCONJ
ejpam-642	60	20	both	both	PRON
ejpam-642	60	21	components	component	NOUN
ejpam-642	60	22	.	.	PUNCT
ejpam-642	61	1	both	both	CCONJ
ejpam-642	61	2	underfitting	underfitte	VERB
ejpam-642	61	3	and	and	CCONJ
ejpam-642	61	4	overfitting	overfitting	NOUN
ejpam-642	61	5	will	will	AUX
ejpam-642	61	6	affect	affect	VERB
ejpam-642	61	7	the	the	DET
ejpam-642	61	8	calculation	calculation	NOUN
ejpam-642	61	9	of	of	ADP
ejpam-642	61	10	matched	match	VERB
ejpam-642	61	11	and	and	CCONJ
ejpam-642	61	12	unmatched	unmatched	ADJ
ejpam-642	61	13	probabilities	probability	NOUN
ejpam-642	61	14	,	,	PUNCT
ejpam-642	61	15	and	and	CCONJ
ejpam-642	61	16	thus	thus	ADV
ejpam-642	61	17	affect	affect	VERB
ejpam-642	61	18	the	the	DET
ejpam-642	61	19	partition	partition	NOUN
ejpam-642	61	20	of	of	ADP
ejpam-642	61	21	record	record	NOUN
ejpam-642	61	22	pairs	pair	NOUN
ejpam-642	61	23	.	.	PUNCT
ejpam-642	62	1	these	these	DET
ejpam-642	62	2	problems	problem	NOUN
ejpam-642	62	3	lead	lead	VERB
ejpam-642	62	4	to	to	ADP
ejpam-642	62	5	the	the	DET
ejpam-642	62	6	model	model	NOUN
ejpam-642	62	7	selection	selection	NOUN
ejpam-642	62	8	in	in	ADP
ejpam-642	62	9	the	the	DET
ejpam-642	62	10	mixture	mixture	NOUN
ejpam-642	62	11	setting	setting	NOUN
ejpam-642	62	12	.	.	PUNCT
ejpam-642	63	1	in	in	ADP
ejpam-642	63	2	our	our	PRON
ejpam-642	63	3	study	study	NOUN
ejpam-642	63	4	,	,	PUNCT
ejpam-642	63	5	we	we	PRON
ejpam-642	63	6	propose	propose	VERB
ejpam-642	63	7	a	a	DET
ejpam-642	63	8	strategy	strategy	NOUN
ejpam-642	63	9	for	for	ADP
ejpam-642	63	10	model	model	NOUN
ejpam-642	63	11	selection	selection	NOUN
ejpam-642	63	12	in	in	ADP
ejpam-642	63	13	loglinear	loglinear	NOUN
ejpam-642	63	14	mixtures	mixture	NOUN
ejpam-642	63	15	,	,	PUNCT
ejpam-642	63	16	where	where	SCONJ
ejpam-642	63	17	parameter	parameter	NOUN
ejpam-642	63	18	estimates	estimate	NOUN
ejpam-642	63	19	are	be	AUX
ejpam-642	63	20	obtained	obtain	VERB
ejpam-642	63	21	via	via	ADP
ejpam-642	63	22	a	a	DET
ejpam-642	63	23	simplified	simplified	ADJ
ejpam-642	63	24	method	method	NOUN
ejpam-642	63	25	for	for	ADP
ejpam-642	63	26	the	the	DET
ejpam-642	63	27	em	em	PROPN
ejpam-642	63	28	algorithm	algorithm	NOUN
ejpam-642	63	29	for	for	ADP
ejpam-642	63	30	the	the	DET
ejpam-642	63	31	mixture	mixture	NOUN
ejpam-642	63	32	of	of	ADP
ejpam-642	63	33	two	two	NUM
ejpam-642	63	34	components	component	NOUN
ejpam-642	63	35	.	.	PUNCT
ejpam-642	64	1	such	such	DET
ejpam-642	64	2	a	a	DET
ejpam-642	64	3	simplified	simplified	ADJ
ejpam-642	64	4	method	method	NOUN
ejpam-642	64	5	can	can	AUX
ejpam-642	64	6	take	take	VERB
ejpam-642	64	7	advantage	advantage	NOUN
ejpam-642	64	8	of	of	ADP
ejpam-642	64	9	existing	exist	VERB
ejpam-642	64	10	packages	package	NOUN
ejpam-642	64	11	and	and	CCONJ
ejpam-642	64	12	functions	function	NOUN
ejpam-642	64	13	in	in	ADP
ejpam-642	64	14	sophisticated	sophisticated	ADJ
ejpam-642	64	15	statistical	statistical	ADJ
ejpam-642	64	16	software	software	NOUN
ejpam-642	64	17	like	like	ADP
ejpam-642	64	18	r	r	NOUN
ejpam-642	64	19	,	,	PUNCT
ejpam-642	64	20	so	so	ADV
ejpam-642	64	21	computationally	computationally	ADV
ejpam-642	64	22	it	it	PRON
ejpam-642	64	23	is	be	AUX
ejpam-642	64	24	tractable	tractable	ADJ
ejpam-642	64	25	.	.	PUNCT
ejpam-642	65	1	we	we	PRON
ejpam-642	65	2	define	define	VERB
ejpam-642	65	3	the	the	DET
ejpam-642	65	4	alternatively	alternatively	ADV
ejpam-642	65	5	climbing	climb	VERB
ejpam-642	65	6	pathways	pathway	NOUN
ejpam-642	65	7	to	to	PART
ejpam-642	65	8	indicate	indicate	VERB
ejpam-642	65	9	a	a	DET
ejpam-642	65	10	sequence	sequence	NOUN
ejpam-642	65	11	of	of	ADP
ejpam-642	65	12	nested	nested	ADJ
ejpam-642	65	13	models	model	NOUN
ejpam-642	65	14	in	in	ADP
ejpam-642	65	15	the	the	DET
ejpam-642	65	16	mixture	mixture	NOUN
ejpam-642	65	17	family	family	NOUN
ejpam-642	65	18	.	.	PUNCT
ejpam-642	66	1	for	for	ADP
ejpam-642	66	2	comparison	comparison	NOUN
ejpam-642	66	3	,	,	PUNCT
ejpam-642	66	4	we	we	PRON
ejpam-642	66	5	adopt	adopt	VERB
ejpam-642	66	6	the	the	DET
ejpam-642	66	7	pearson	pearson	PROPN
ejpam-642	66	8	chi	chi	PROPN
ejpam-642	66	9	-	-	PUNCT
ejpam-642	66	10	square	square	ADJ
ejpam-642	66	11	test	test	NOUN
ejpam-642	66	12	for	for	ADP
ejpam-642	66	13	nested	nested	ADJ
ejpam-642	66	14	models	model	NOUN
ejpam-642	66	15	on	on	ADP
ejpam-642	66	16	alternatively	alternatively	ADV
ejpam-642	66	17	climbing	climb	VERB
ejpam-642	66	18	pathways	pathway	NOUN
ejpam-642	66	19	.	.	PUNCT
ejpam-642	67	1	the	the	DET
ejpam-642	67	2	feature	feature	NOUN
ejpam-642	67	3	of	of	ADP
ejpam-642	67	4	this	this	DET
ejpam-642	67	5	strategy	strategy	NOUN
ejpam-642	67	6	is	be	AUX
ejpam-642	67	7	data	data	NOUN
ejpam-642	67	8	-	-	PUNCT
ejpam-642	67	9	driven	drive	VERB
ejpam-642	67	10	,	,	PUNCT
ejpam-642	67	11	and	and	CCONJ
ejpam-642	67	12	thus	thus	ADV
ejpam-642	67	13	,	,	PUNCT
ejpam-642	67	14	the	the	DET
ejpam-642	67	15	model	model	NOUN
ejpam-642	67	16	building	building	NOUN
ejpam-642	67	17	avoids	avoid	VERB
ejpam-642	67	18	subjective	subjective	ADJ
ejpam-642	67	19	specification	specification	NOUN
ejpam-642	67	20	.	.	PUNCT
ejpam-642	68	1	simulation	simulation	NOUN
ejpam-642	68	2	studies	study	NOUN
ejpam-642	68	3	have	have	AUX
ejpam-642	68	4	been	be	AUX
ejpam-642	68	5	conducted	conduct	VERB
ejpam-642	68	6	to	to	PART
ejpam-642	68	7	verify	verify	VERB
ejpam-642	68	8	the	the	DET
ejpam-642	68	9	strategy	strategy	NOUN
ejpam-642	68	10	,	,	PUNCT
ejpam-642	68	11	and	and	CCONJ
ejpam-642	68	12	to	to	PART
ejpam-642	68	13	acquire	acquire	VERB
ejpam-642	68	14	experience	experience	NOUN
ejpam-642	68	15	of	of	ADP
ejpam-642	68	16	model	model	NOUN
ejpam-642	68	17	building	building	NOUN
ejpam-642	68	18	.	.	PUNCT
ejpam-642	69	1	we	we	PRON
ejpam-642	69	2	compare	compare	VERB
ejpam-642	69	3	the	the	DET
ejpam-642	69	4	results	result	NOUN
ejpam-642	69	5	from	from	ADP
ejpam-642	69	6	this	this	DET
ejpam-642	69	7	strategy	strategy	NOUN
ejpam-642	69	8	with	with	ADP
ejpam-642	69	9	those	those	PRON
ejpam-642	69	10	from	from	ADP
ejpam-642	69	11	the	the	DET
ejpam-642	69	12	simplest	simple	ADJ
ejpam-642	69	13	model	model	NOUN
ejpam-642	69	14	where	where	SCONJ
ejpam-642	69	15	all	all	DET
ejpam-642	69	16	factors	factor	NOUN
ejpam-642	69	17	are	be	AUX
ejpam-642	69	18	binary	binary	ADJ
ejpam-642	69	19	and	and	CCONJ
ejpam-642	69	20	independent	independent	ADJ
ejpam-642	69	21	.	.	PUNCT
ejpam-642	70	1	it	it	PRON
ejpam-642	70	2	shows	show	VERB
ejpam-642	70	3	that	that	SCONJ
ejpam-642	70	4	the	the	DET
ejpam-642	70	5	data	data	NOUN
ejpam-642	70	6	-	-	PUNCT
ejpam-642	70	7	driven	drive	VERB
ejpam-642	70	8	strategy	strategy	NOUN
ejpam-642	70	9	does	do	AUX
ejpam-642	70	10	give	give	VERB
ejpam-642	70	11	a	a	DET
ejpam-642	70	12	refinement	refinement	NOUN
ejpam-642	70	13	on	on	ADP
ejpam-642	70	14	modelling	modelling	NOUN
ejpam-642	70	15	and	and	CCONJ
ejpam-642	70	16	lead	lead	VERB
ejpam-642	70	17	to	to	ADP
ejpam-642	70	18	more	more	ADV
ejpam-642	70	19	accurate	accurate	ADJ
ejpam-642	70	20	estimation	estimation	NOUN
ejpam-642	70	21	of	of	ADP
ejpam-642	70	22	probabilities	probability	NOUN
ejpam-642	70	23	compared	compare	VERB
ejpam-642	70	24	with	with	ADP
ejpam-642	70	25	the	the	DET
ejpam-642	70	26	simplest	simple	ADJ
ejpam-642	70	27	model	model	NOUN
ejpam-642	70	28	.	.	PUNCT
ejpam-642	71	1	this	this	DET
ejpam-642	71	2	paper	paper	NOUN
ejpam-642	71	3	is	be	AUX
ejpam-642	71	4	organized	organize	VERB
ejpam-642	71	5	as	as	SCONJ
ejpam-642	71	6	follows	follow	VERB
ejpam-642	71	7	.	.	PUNCT
ejpam-642	72	1	we	we	PRON
ejpam-642	72	2	outline	outline	VERB
ejpam-642	72	3	statistical	statistical	ADJ
ejpam-642	72	4	modelling	modelling	NOUN
ejpam-642	72	5	and	and	CCONJ
ejpam-642	72	6	a	a	DET
ejpam-642	72	7	simplified	simplified	ADJ
ejpam-642	72	8	method	method	NOUN
ejpam-642	72	9	for	for	ADP
ejpam-642	72	10	the	the	DET
ejpam-642	72	11	em	em	PROPN
ejpam-642	72	12	algorithm	algorithm	NOUN
ejpam-642	72	13	in	in	ADP
ejpam-642	72	14	record	record	NOUN
ejpam-642	72	15	linkage	linkage	NOUN
ejpam-642	72	16	in	in	ADP
ejpam-642	72	17	section	section	NOUN
ejpam-642	72	18	2	2	NUM
ejpam-642	72	19	.	.	PUNCT
ejpam-642	73	1	in	in	ADP
ejpam-642	73	2	section	section	NOUN
ejpam-642	73	3	3	3	NUM
ejpam-642	73	4	,	,	PUNCT
ejpam-642	73	5	we	we	PRON
ejpam-642	73	6	propose	propose	VERB
ejpam-642	73	7	the	the	DET
ejpam-642	73	8	data	data	NOUN
ejpam-642	73	9	-	-	PUNCT
ejpam-642	73	10	driven	drive	VERB
ejpam-642	73	11	strategy	strategy	NOUN
ejpam-642	73	12	for	for	ADP
ejpam-642	73	13	modelling	model	VERB
ejpam-642	73	14	loglinear	loglinear	NOUN
ejpam-642	73	15	mixtures	mixture	NOUN
ejpam-642	73	16	.	.	PUNCT
ejpam-642	74	1	then	then	ADV
ejpam-642	74	2	we	we	PRON
ejpam-642	74	3	conduct	conduct	VERB
ejpam-642	74	4	simulation	simulation	NOUN
ejpam-642	74	5	studies	study	NOUN
ejpam-642	74	6	to	to	PART
ejpam-642	74	7	assess	assess	VERB
ejpam-642	74	8	the	the	DET
ejpam-642	74	9	proposed	propose	VERB
ejpam-642	74	10	strategy	strategy	NOUN
ejpam-642	74	11	in	in	ADP
ejpam-642	74	12	section	section	NOUN
ejpam-642	74	13	4	4	NUM
ejpam-642	74	14	,	,	PUNCT
ejpam-642	74	15	and	and	CCONJ
ejpam-642	74	16	apply	apply	VERB
ejpam-642	74	17	it	it	PRON
ejpam-642	74	18	to	to	ADP
ejpam-642	74	19	a	a	DET
ejpam-642	74	20	case	case	NOUN
ejpam-642	74	21	study	study	NOUN
ejpam-642	74	22	in	in	ADP
ejpam-642	74	23	section	section	NOUN
ejpam-642	74	24	5	5	NUM
ejpam-642	74	25	.	.	PUNCT
ejpam-642	75	1	finally	finally	ADV
ejpam-642	75	2	,	,	PUNCT
ejpam-642	75	3	we	we	PRON
ejpam-642	75	4	make	make	VERB
ejpam-642	75	5	brief	brief	ADJ
ejpam-642	75	6	concluding	concluding	NOUN
ejpam-642	75	7	remarks	remark	NOUN
ejpam-642	75	8	in	in	ADP
ejpam-642	75	9	section	section	NOUN
ejpam-642	75	10	6	6	NUM
ejpam-642	75	11	.	.	NOUN
ejpam-642	75	12	2	2	NUM
ejpam-642	75	13	.	.	PUNCT
ejpam-642	75	14	statistical	statistical	ADJ
ejpam-642	75	15	framework	framework	NOUN
ejpam-642	75	16	and	and	CCONJ
ejpam-642	75	17	parameter	parameter	NOUN
ejpam-642	75	18	estimation	estimation	NOUN
ejpam-642	75	19	for	for	ADP
ejpam-642	75	20	record	record	NOUN
ejpam-642	75	21	linkage	linkage	NOUN
ejpam-642	75	22	in	in	ADP
ejpam-642	75	23	this	this	DET
ejpam-642	75	24	section	section	NOUN
ejpam-642	75	25	,	,	PUNCT
ejpam-642	75	26	we	we	PRON
ejpam-642	75	27	first	first	ADV
ejpam-642	75	28	outline	outline	VERB
ejpam-642	75	29	the	the	DET
ejpam-642	75	30	statistical	statistical	ADJ
ejpam-642	75	31	framework	framework	NOUN
ejpam-642	75	32	for	for	ADP
ejpam-642	75	33	the	the	DET
ejpam-642	75	34	record	record	NOUN
ejpam-642	75	35	linkage	linkage	NOUN
ejpam-642	75	36	,	,	PUNCT
ejpam-642	75	37	where	where	SCONJ
ejpam-642	75	38	the	the	DET
ejpam-642	75	39	population	population	NOUN
ejpam-642	75	40	of	of	ADP
ejpam-642	75	41	all	all	DET
ejpam-642	75	42	pairs	pair	NOUN
ejpam-642	75	43	is	be	AUX
ejpam-642	75	44	the	the	DET
ejpam-642	75	45	mixture	mixture	NOUN
ejpam-642	75	46	of	of	ADP
ejpam-642	75	47	two	two	NUM
ejpam-642	75	48	components	component	NOUN
ejpam-642	75	49	or	or	CCONJ
ejpam-642	75	50	subpopulations	subpopulation	NOUN
ejpam-642	75	51	:	:	PUNCT
ejpam-642	75	52	matched	match	VERB
ejpam-642	75	53	pairs	pair	NOUN
ejpam-642	75	54	and	and	CCONJ
ejpam-642	75	55	unmatched	unmatched	ADJ
ejpam-642	75	56	pairs	pair	NOUN
ejpam-642	75	57	.	.	PUNCT
ejpam-642	76	1	then	then	ADV
ejpam-642	76	2	we	we	PRON
ejpam-642	76	3	give	give	VERB
ejpam-642	76	4	a	a	DET
ejpam-642	76	5	simplified	simplified	ADJ
ejpam-642	76	6	method	method	NOUN
ejpam-642	76	7	for	for	ADP
ejpam-642	76	8	the	the	DET
ejpam-642	76	9	em	em	PROPN
ejpam-642	76	10	algorithm	algorithm	NOUN
ejpam-642	76	11	for	for	ADP
ejpam-642	76	12	this	this	DET
ejpam-642	76	13	specific	specific	ADJ
ejpam-642	76	14	type	type	NOUN
ejpam-642	76	15	of	of	ADP
ejpam-642	76	16	mixture	mixture	NOUN
ejpam-642	76	17	models	model	NOUN
ejpam-642	76	18	.	.	PUNCT
ejpam-642	77	1	r.	r.	PROPN
ejpam-642	77	2	zhu	zhu	PROPN
ejpam-642	77	3	,	,	PUNCT
ejpam-642	77	4	j.	j.	PROPN
ejpam-642	77	5	zhang	zhang	PROPN
ejpam-642	77	6	,	,	PUNCT
ejpam-642	77	7	d.	d.	PROPN
ejpam-642	77	8	zhang	zhang	PROPN
ejpam-642	77	9	,	,	PUNCT
ejpam-642	77	10	g.	g.	PROPN
ejpam-642	77	11	yan	yan	PROPN
ejpam-642	77	12	/	/	SYM
ejpam-642	77	13	eur	eur	PROPN
ejpam-642	77	14	.	.	PUNCT
ejpam-642	78	1	j.	j.	PROPN
ejpam-642	78	2	pure	pure	PROPN
ejpam-642	78	3	appl	appl	PROPN
ejpam-642	78	4	.	.	PROPN
ejpam-642	78	5	math	math	PROPN
ejpam-642	78	6	,	,	PUNCT
ejpam-642	78	7	3	3	NUM
ejpam-642	78	8	(	(	PUNCT
ejpam-642	78	9	2010	2010	NUM
ejpam-642	78	10	)	)	PUNCT
ejpam-642	78	11	,	,	PUNCT
ejpam-642	78	12	141	141	NUM
ejpam-642	78	13	-	-	SYM
ejpam-642	78	14	162	162	NUM
ejpam-642	78	15	144	144	NUM
ejpam-642	78	16	2.1	2.1	NUM
ejpam-642	78	17	.	.	PUNCT
ejpam-642	79	1	notation	notation	NOUN
ejpam-642	79	2	and	and	CCONJ
ejpam-642	79	3	framework	framework	NOUN
ejpam-642	79	4	consider	consider	VERB
ejpam-642	79	5	two	two	NUM
ejpam-642	79	6	files	file	NOUN
ejpam-642	79	7	where	where	SCONJ
ejpam-642	79	8	file	file	NOUN
ejpam-642	79	9	a	a	PRON
ejpam-642	79	10	has	have	VERB
ejpam-642	79	11	a	a	DET
ejpam-642	79	12	records	record	NOUN
ejpam-642	79	13	and	and	CCONJ
ejpam-642	79	14	file	file	NOUN
ejpam-642	79	15	b	b	PROPN
ejpam-642	79	16	has	have	VERB
ejpam-642	79	17	b	b	NOUN
ejpam-642	79	18	records	record	NOUN
ejpam-642	79	19	.	.	PUNCT
ejpam-642	80	1	the	the	DET
ejpam-642	80	2	total	total	ADJ
ejpam-642	80	3	number	number	NOUN
ejpam-642	80	4	of	of	ADP
ejpam-642	80	5	possible	possible	ADJ
ejpam-642	80	6	record	record	NOUN
ejpam-642	80	7	pairs	pair	NOUN
ejpam-642	80	8	for	for	ADP
ejpam-642	80	9	comparison	comparison	NOUN
ejpam-642	80	10	is	be	AUX
ejpam-642	80	11	n	n	NOUN
ejpam-642	80	12	=	=	SYM
ejpam-642	80	13	a×	a×	PROPN
ejpam-642	80	14	b.	b.	NOUN
ejpam-642	81	1	the	the	DET
ejpam-642	81	2	records	record	NOUN
ejpam-642	81	3	consist	consist	VERB
ejpam-642	81	4	of	of	ADP
ejpam-642	81	5	attributes	attribute	NOUN
ejpam-642	81	6	which	which	PRON
ejpam-642	81	7	can	can	AUX
ejpam-642	81	8	be	be	AUX
ejpam-642	81	9	categorical	categorical	ADJ
ejpam-642	81	10	variables	variable	NOUN
ejpam-642	81	11	such	such	ADJ
ejpam-642	81	12	as	as	ADP
ejpam-642	81	13	name	name	NOUN
ejpam-642	81	14	,	,	PUNCT
ejpam-642	81	15	address	address	NOUN
ejpam-642	81	16	and	and	CCONJ
ejpam-642	81	17	date	date	NOUN
ejpam-642	81	18	of	of	ADP
ejpam-642	81	19	birth	birth	NOUN
ejpam-642	81	20	,	,	PUNCT
ejpam-642	81	21	or	or	CCONJ
ejpam-642	81	22	numerical	numerical	ADJ
ejpam-642	81	23	variables	variable	NOUN
ejpam-642	81	24	such	such	ADJ
ejpam-642	81	25	as	as	ADP
ejpam-642	81	26	age	age	NOUN
ejpam-642	81	27	.	.	PUNCT
ejpam-642	82	1	select	select	VERB
ejpam-642	82	2	n	n	CCONJ
ejpam-642	82	3	common	common	ADJ
ejpam-642	82	4	attributes	attribute	NOUN
ejpam-642	82	5	as	as	ADP
ejpam-642	82	6	the	the	DET
ejpam-642	82	7	identifiers	identifier	NOUN
ejpam-642	82	8	for	for	ADP
ejpam-642	82	9	comparing	compare	VERB
ejpam-642	82	10	pairs	pair	NOUN
ejpam-642	82	11	of	of	ADP
ejpam-642	82	12	records	record	NOUN
ejpam-642	82	13	from	from	ADP
ejpam-642	82	14	both	both	DET
ejpam-642	82	15	files	file	NOUN
ejpam-642	82	16	a	a	PRON
ejpam-642	82	17	and	and	CCONJ
ejpam-642	82	18	b.	b.	PROPN
ejpam-642	83	1	then	then	ADV
ejpam-642	83	2	the	the	DET
ejpam-642	83	3	comparison	comparison	NOUN
ejpam-642	83	4	for	for	ADP
ejpam-642	83	5	each	each	DET
ejpam-642	83	6	record	record	NOUN
ejpam-642	83	7	pair	pair	NOUN
ejpam-642	83	8	is	be	AUX
ejpam-642	83	9	made	make	VERB
ejpam-642	83	10	for	for	ADP
ejpam-642	83	11	all	all	DET
ejpam-642	83	12	n	n	PRON
ejpam-642	83	13	identifiers	identifier	NOUN
ejpam-642	83	14	by	by	ADP
ejpam-642	83	15	matching	match	VERB
ejpam-642	83	16	rules	rule	NOUN
ejpam-642	83	17	defined	define	VERB
ejpam-642	83	18	according	accord	VERB
ejpam-642	83	19	to	to	ADP
ejpam-642	83	20	the	the	DET
ejpam-642	83	21	needs	need	NOUN
ejpam-642	83	22	for	for	ADP
ejpam-642	83	23	these	these	DET
ejpam-642	83	24	attributes	attribute	NOUN
ejpam-642	83	25	.	.	PUNCT
ejpam-642	84	1	this	this	PRON
ejpam-642	84	2	will	will	AUX
ejpam-642	84	3	result	result	VERB
ejpam-642	84	4	in	in	ADP
ejpam-642	84	5	a	a	DET
ejpam-642	84	6	comparison	comparison	NOUN
ejpam-642	84	7	vector	vector	NOUN
ejpam-642	84	8	of	of	ADP
ejpam-642	84	9	length	length	NOUN
ejpam-642	84	10	n	n	NOUN
ejpam-642	84	11	with	with	ADP
ejpam-642	84	12	each	each	DET
ejpam-642	84	13	component	component	NOUN
ejpam-642	84	14	being	be	AUX
ejpam-642	84	15	categorical	categorical	ADJ
ejpam-642	84	16	.	.	PUNCT
ejpam-642	85	1	each	each	DET
ejpam-642	85	2	component	component	NOUN
ejpam-642	85	3	associated	associate	VERB
ejpam-642	85	4	with	with	ADP
ejpam-642	85	5	an	an	DET
ejpam-642	85	6	identifier	identifier	NOUN
ejpam-642	85	7	(	(	PUNCT
ejpam-642	85	8	attribute	attribute	NOUN
ejpam-642	85	9	)	)	PUNCT
ejpam-642	85	10	in	in	ADP
ejpam-642	85	11	the	the	DET
ejpam-642	85	12	comparison	comparison	NOUN
ejpam-642	85	13	vector	vector	NOUN
ejpam-642	85	14	is	be	AUX
ejpam-642	85	15	called	call	VERB
ejpam-642	85	16	a	a	DET
ejpam-642	85	17	field	field	NOUN
ejpam-642	85	18	.	.	PUNCT
ejpam-642	86	1	the	the	DET
ejpam-642	86	2	field	field	NOUN
ejpam-642	86	3	values	value	NOUN
ejpam-642	86	4	of	of	ADP
ejpam-642	86	5	comparison	comparison	NOUN
ejpam-642	86	6	vector	vector	NOUN
ejpam-642	86	7	could	could	AUX
ejpam-642	86	8	be	be	AUX
ejpam-642	86	9	binary	binary	ADJ
ejpam-642	86	10	like	like	ADP
ejpam-642	86	11	0	0	NUM
ejpam-642	86	12	and	and	CCONJ
ejpam-642	86	13	1	1	NUM
ejpam-642	86	14	,	,	PUNCT
ejpam-642	86	15	or	or	CCONJ
ejpam-642	86	16	any	any	DET
ejpam-642	86	17	others	other	NOUN
ejpam-642	86	18	defined	define	VERB
ejpam-642	86	19	by	by	ADP
ejpam-642	86	20	some	some	DET
ejpam-642	86	21	rules	rule	NOUN
ejpam-642	86	22	.	.	PUNCT
ejpam-642	87	1	for	for	ADP
ejpam-642	87	2	example	example	NOUN
ejpam-642	87	3	,	,	PUNCT
ejpam-642	87	4	the	the	DET
ejpam-642	87	5	simplest	simple	ADJ
ejpam-642	87	6	rule	rule	NOUN
ejpam-642	87	7	is	be	AUX
ejpam-642	87	8	agreement	agreement	NOUN
ejpam-642	87	9	/	/	SYM
ejpam-642	87	10	disagreement	disagreement	NOUN
ejpam-642	87	11	on	on	ADP
ejpam-642	87	12	an	an	DET
ejpam-642	87	13	attribute	attribute	NOUN
ejpam-642	87	14	.	.	PUNCT
ejpam-642	88	1	under	under	ADP
ejpam-642	88	2	this	this	DET
ejpam-642	88	3	simplest	simple	ADJ
ejpam-642	88	4	rule	rule	NOUN
ejpam-642	88	5	,	,	PUNCT
ejpam-642	88	6	if	if	SCONJ
ejpam-642	88	7	comparison	comparison	NOUN
ejpam-642	88	8	attributes	attribute	NOUN
ejpam-642	88	9	are	be	AUX
ejpam-642	88	10	name	name	NOUN
ejpam-642	88	11	,	,	PUNCT
ejpam-642	88	12	address	address	NOUN
ejpam-642	88	13	,	,	PUNCT
ejpam-642	88	14	gender	gender	NOUN
ejpam-642	88	15	and	and	CCONJ
ejpam-642	88	16	date	date	NOUN
ejpam-642	88	17	of	of	ADP
ejpam-642	88	18	birth	birth	NOUN
ejpam-642	88	19	,	,	PUNCT
ejpam-642	88	20	an	an	DET
ejpam-642	88	21	observed	observed	ADJ
ejpam-642	88	22	comparison	comparison	NOUN
ejpam-642	88	23	vector	vector	NOUN
ejpam-642	88	24	(	(	PUNCT
ejpam-642	88	25	0,1,1,0	0,1,1,0	NUM
ejpam-642	88	26	)	)	PUNCT
ejpam-642	88	27	means	mean	VERB
ejpam-642	88	28	disagreement	disagreement	NOUN
ejpam-642	88	29	on	on	ADP
ejpam-642	88	30	name	name	NOUN
ejpam-642	88	31	and	and	CCONJ
ejpam-642	88	32	date	date	NOUN
ejpam-642	88	33	of	of	ADP
ejpam-642	88	34	birth	birth	NOUN
ejpam-642	88	35	,	,	PUNCT
ejpam-642	88	36	but	but	CCONJ
ejpam-642	88	37	agreement	agreement	NOUN
ejpam-642	88	38	on	on	ADP
ejpam-642	88	39	address	address	NOUN
ejpam-642	88	40	and	and	CCONJ
ejpam-642	88	41	gender	gender	NOUN
ejpam-642	88	42	.	.	PUNCT
ejpam-642	89	1	thus	thus	ADV
ejpam-642	89	2	,	,	PUNCT
ejpam-642	89	3	we	we	PRON
ejpam-642	89	4	obtain	obtain	VERB
ejpam-642	89	5	n	n	PRON
ejpam-642	89	6	comparison	comparison	NOUN
ejpam-642	89	7	vectors	vector	NOUN
ejpam-642	89	8	for	for	ADP
ejpam-642	89	9	all	all	DET
ejpam-642	89	10	record	record	NOUN
ejpam-642	89	11	pairs	pair	NOUN
ejpam-642	89	12	denoted	denote	VERB
ejpam-642	89	13	as	as	SCONJ
ejpam-642	89	14	follows	follow	VERB
ejpam-642	89	15	:	:	PUNCT
ejpam-642	89	16	γ	γ	X
ejpam-642	89	17	j	j	PROPN
ejpam-642	90	1	=	=	PUNCT
ejpam-642	90	2	(	(	PUNCT
ejpam-642	90	3	γ	γ	X
ejpam-642	90	4	j	j	PROPN
ejpam-642	90	5	1	1	NUM
ejpam-642	90	6	,	,	PUNCT
ejpam-642	90	7	.	.	PUNCT
ejpam-642	90	8	.	.	PUNCT
ejpam-642	90	9	.	.	PUNCT
ejpam-642	91	1	,	,	PUNCT
ejpam-642	91	2	γ	γ	PROPN
ejpam-642	91	3	j	j	PROPN
ejpam-642	91	4	n	n	CCONJ
ejpam-642	91	5	)	)	PUNCT
ejpam-642	91	6	,	,	PUNCT
ejpam-642	91	7	j	j	PROPN
ejpam-642	91	8	=	=	SYM
ejpam-642	91	9	1	1	NUM
ejpam-642	91	10	,	,	PUNCT
ejpam-642	91	11	.	.	PUNCT
ejpam-642	91	12	.	.	PUNCT
ejpam-642	92	1	.	.	PUNCT
ejpam-642	93	1	,	,	PUNCT
ejpam-642	93	2	n	n	X
ejpam-642	93	3	,	,	PUNCT
ejpam-642	93	4	where	where	SCONJ
ejpam-642	93	5	the	the	DET
ejpam-642	93	6	subscript	subscript	NOUN
ejpam-642	93	7	denotes	denote	VERB
ejpam-642	93	8	the	the	DET
ejpam-642	93	9	comparison	comparison	NOUN
ejpam-642	93	10	field	field	NOUN
ejpam-642	93	11	while	while	SCONJ
ejpam-642	93	12	the	the	DET
ejpam-642	93	13	superscript	superscript	NOUN
ejpam-642	93	14	indicates	indicate	VERB
ejpam-642	93	15	the	the	DET
ejpam-642	93	16	comparison	comparison	NOUN
ejpam-642	93	17	pair	pair	NOUN
ejpam-642	93	18	.	.	PUNCT
ejpam-642	94	1	note	note	VERB
ejpam-642	94	2	that	that	SCONJ
ejpam-642	94	3	all	all	DET
ejpam-642	94	4	comparison	comparison	NOUN
ejpam-642	94	5	pairs	pair	NOUN
ejpam-642	94	6	include	include	VERB
ejpam-642	94	7	two	two	NUM
ejpam-642	94	8	types	type	NOUN
ejpam-642	94	9	of	of	ADP
ejpam-642	94	10	pairs	pair	NOUN
ejpam-642	94	11	:	:	PUNCT
ejpam-642	94	12	matched	match	VERB
ejpam-642	94	13	and	and	CCONJ
ejpam-642	94	14	unmatched	unmatched	ADJ
ejpam-642	94	15	.	.	PUNCT
ejpam-642	95	1	therefore	therefore	ADV
ejpam-642	95	2	,	,	PUNCT
ejpam-642	95	3	we	we	PRON
ejpam-642	95	4	can	can	AUX
ejpam-642	95	5	view	view	VERB
ejpam-642	95	6	the	the	DET
ejpam-642	95	7	entire	entire	ADJ
ejpam-642	95	8	population	population	NOUN
ejpam-642	95	9	of	of	ADP
ejpam-642	95	10	all	all	DET
ejpam-642	95	11	possible	possible	ADJ
ejpam-642	95	12	comparison	comparison	NOUN
ejpam-642	95	13	pairs	pair	NOUN
ejpam-642	95	14	as	as	ADP
ejpam-642	95	15	a	a	DET
ejpam-642	95	16	mixture	mixture	NOUN
ejpam-642	95	17	of	of	ADP
ejpam-642	95	18	two	two	NUM
ejpam-642	95	19	components	component	NOUN
ejpam-642	95	20	or	or	CCONJ
ejpam-642	95	21	subpopulations	subpopulation	NOUN
ejpam-642	95	22	:	:	PUNCT
ejpam-642	95	23	matched	match	VERB
ejpam-642	95	24	pairs	pair	NOUN
ejpam-642	95	25	and	and	CCONJ
ejpam-642	95	26	unmatched	unmatched	ADJ
ejpam-642	95	27	pairs	pair	NOUN
ejpam-642	95	28	.	.	PUNCT
ejpam-642	96	1	however	however	ADV
ejpam-642	96	2	,	,	PUNCT
ejpam-642	96	3	in	in	ADP
ejpam-642	96	4	this	this	DET
ejpam-642	96	5	mixture	mixture	NOUN
ejpam-642	96	6	,	,	PUNCT
ejpam-642	96	7	the	the	DET
ejpam-642	96	8	membership	membership	NOUN
ejpam-642	96	9	for	for	ADP
ejpam-642	96	10	each	each	DET
ejpam-642	96	11	pair	pair	NOUN
ejpam-642	96	12	is	be	AUX
ejpam-642	96	13	missing	miss	VERB
ejpam-642	96	14	or	or	CCONJ
ejpam-642	96	15	unknown	unknown	ADJ
ejpam-642	96	16	.	.	PUNCT
ejpam-642	97	1	let	let	VERB
ejpam-642	97	2	m	m	PRON
ejpam-642	97	3	denote	denote	VERB
ejpam-642	97	4	the	the	DET
ejpam-642	97	5	matched	match	VERB
ejpam-642	97	6	subpopulation	subpopulation	NOUN
ejpam-642	97	7	and	and	CCONJ
ejpam-642	97	8	u	u	NOUN
ejpam-642	97	9	denote	denote	VERB
ejpam-642	97	10	the	the	DET
ejpam-642	97	11	unmatched	unmatched	ADJ
ejpam-642	97	12	subpopulation	subpopulation	NOUN
ejpam-642	97	13	.	.	PUNCT
ejpam-642	98	1	each	each	DET
ejpam-642	98	2	pair	pair	NOUN
ejpam-642	98	3	is	be	AUX
ejpam-642	98	4	either	either	CCONJ
ejpam-642	98	5	from	from	ADP
ejpam-642	98	6	m	m	PROPN
ejpam-642	98	7	or	or	CCONJ
ejpam-642	98	8	u	u	PROPN
ejpam-642	98	9	.	.	PUNCT
ejpam-642	99	1	let	let	VERB
ejpam-642	99	2	m	m	PRON
ejpam-642	99	3	j	j	NOUN
ejpam-642	99	4	=	=	SYM
ejpam-642	99	5	p(γ	p(γ	PROPN
ejpam-642	99	6	j|m	j|m	NUM
ejpam-642	99	7	)	)	PUNCT
ejpam-642	99	8	,	,	PUNCT
ejpam-642	99	9	u	u	NOUN
ejpam-642	99	10	j	j	PROPN
ejpam-642	99	11	=	=	SYM
ejpam-642	99	12	p(γ	p(γ	NOUN
ejpam-642	99	13	j	j	NOUN
ejpam-642	99	14	|u	|u	ADJ
ejpam-642	99	15	)	)	PUNCT
ejpam-642	99	16	,	,	PUNCT
ejpam-642	99	17	j	j	PROPN
ejpam-642	99	18	=	=	SYM
ejpam-642	99	19	1	1	NUM
ejpam-642	99	20	,	,	PUNCT
ejpam-642	99	21	.	.	PUNCT
ejpam-642	99	22	.	.	PUNCT
ejpam-642	100	1	.	.	PUNCT
ejpam-642	101	1	,	,	PUNCT
ejpam-642	101	2	n	n	X
ejpam-642	101	3	.	.	PUNCT
ejpam-642	102	1	define	define	VERB
ejpam-642	102	2	the	the	DET
ejpam-642	102	3	matching	matching	NOUN
ejpam-642	102	4	weight	weight	NOUN
ejpam-642	102	5	as	as	SCONJ
ejpam-642	102	6	follows	follow	VERB
ejpam-642	102	7	:	:	PUNCT
ejpam-642	102	8	w	w	PROPN
ejpam-642	102	9	j	j	PROPN
ejpam-642	102	10	=	=	PRON
ejpam-642	102	11	log	log	PROPN
ejpam-642	102	12	p(γ	p(γ	NUM
ejpam-642	102	13	j|m	j|m	NUM
ejpam-642	102	14	)	)	PUNCT
ejpam-642	102	15	p(γ	p(γ	NOUN
ejpam-642	102	16	j|u	j|u	ADJ
ejpam-642	102	17	)	)	PUNCT
ejpam-642	103	1	=	=	PRON
ejpam-642	103	2	log	log	VERB
ejpam-642	103	3	p(γ	p(γ	NOUN
ejpam-642	103	4	j|m)−	j|m)−	PROPN
ejpam-642	103	5	log	log	VERB
ejpam-642	103	6	p(γ	p(γ	NOUN
ejpam-642	103	7	j|u	j|u	ADJ
ejpam-642	103	8	)	)	PUNCT
ejpam-642	103	9	,	,	PUNCT
ejpam-642	103	10	j	j	PROPN
ejpam-642	103	11	=	=	SYM
ejpam-642	103	12	1	1	NUM
ejpam-642	103	13	,	,	PUNCT
ejpam-642	103	14	.	.	PUNCT
ejpam-642	103	15	.	.	PUNCT
ejpam-642	104	1	.	.	PUNCT
ejpam-642	105	1	,	,	PUNCT
ejpam-642	105	2	n	n	X
ejpam-642	105	3	,	,	PUNCT
ejpam-642	105	4	the	the	DET
ejpam-642	105	5	logarithm	logarithm	NOUN
ejpam-642	105	6	of	of	ADP
ejpam-642	105	7	the	the	DET
ejpam-642	105	8	likelihood	likelihood	NOUN
ejpam-642	105	9	ratio	ratio	NOUN
ejpam-642	105	10	between	between	ADP
ejpam-642	105	11	the	the	DET
ejpam-642	105	12	matched	match	VERB
ejpam-642	105	13	and	and	CCONJ
ejpam-642	105	14	unmatched	unmatched	ADJ
ejpam-642	105	15	.	.	PUNCT
ejpam-642	106	1	probabilistic	probabilistic	ADJ
ejpam-642	106	2	models	model	NOUN
ejpam-642	106	3	are	be	AUX
ejpam-642	106	4	specified	specify	VERB
ejpam-642	106	5	to	to	PART
ejpam-642	106	6	obtain	obtain	VERB
ejpam-642	106	7	the	the	DET
ejpam-642	106	8	associated	associated	ADJ
ejpam-642	106	9	probabilities	probability	NOUN
ejpam-642	106	10	.	.	PUNCT
ejpam-642	107	1	fellegi	fellegi	NOUN
ejpam-642	107	2	and	and	CCONJ
ejpam-642	107	3	sunter	sunter	NOUN
ejpam-642	107	4	(	(	PUNCT
ejpam-642	107	5	1969	1969	NUM
ejpam-642	107	6	)	)	PUNCT
ejpam-642	107	7	set	set	VERB
ejpam-642	107	8	the	the	DET
ejpam-642	107	9	fundamental	fundamental	ADJ
ejpam-642	107	10	probabilistic	probabilistic	ADJ
ejpam-642	107	11	framework	framework	NOUN
ejpam-642	107	12	for	for	ADP
ejpam-642	107	13	record	record	NOUN
ejpam-642	107	14	linkage	linkage	NOUN
ejpam-642	107	15	;	;	PUNCT
ejpam-642	107	16	see	see	VERB
ejpam-642	107	17	[	[	X
ejpam-642	107	18	6	6	NUM
ejpam-642	107	19	]	]	PUNCT
ejpam-642	107	20	.	.	PUNCT
ejpam-642	108	1	assume	assume	VERB
ejpam-642	108	2	that	that	SCONJ
ejpam-642	108	3	m	m	VERB
ejpam-642	108	4	j	j	NOUN
ejpam-642	108	5	,	,	PUNCT
ejpam-642	108	6	u	u	PROPN
ejpam-642	108	7	j	j	PROPN
ejpam-642	108	8	,	,	PUNCT
ejpam-642	108	9	w	w	PROPN
ejpam-642	108	10	j	j	PROPN
ejpam-642	108	11	are	be	AUX
ejpam-642	108	12	estimated	estimate	VERB
ejpam-642	108	13	from	from	ADP
ejpam-642	108	14	data	data	PROPN
ejpam-642	108	15	.	.	PUNCT
ejpam-642	109	1	without	without	ADP
ejpam-642	109	2	loss	loss	NOUN
ejpam-642	109	3	of	of	ADP
ejpam-642	109	4	generality	generality	NOUN
ejpam-642	109	5	,	,	PUNCT
ejpam-642	109	6	the	the	DET
ejpam-642	109	7	descending	descend	VERB
ejpam-642	109	8	matching	matching	NOUN
ejpam-642	109	9	weight	weight	NOUN
ejpam-642	109	10	sequence	sequence	NOUN
ejpam-642	109	11	is	be	AUX
ejpam-642	109	12	assumed	assume	VERB
ejpam-642	109	13	to	to	PART
ejpam-642	109	14	be	be	AUX
ejpam-642	109	15	w1	w1	PROPN
ejpam-642	109	16	≥	≥	NUM
ejpam-642	109	17	.	.	PUNCT
ejpam-642	109	18	.	.	PUNCT
ejpam-642	110	1	.	.	PUNCT
ejpam-642	111	1	≥	≥	PROPN
ejpam-642	111	2	wk	wk	INTJ
ejpam-642	111	3	≥	≥	PROPN
ejpam-642	111	4	.	.	PUNCT
ejpam-642	111	5	.	.	PUNCT
ejpam-642	112	1	.	.	PUNCT
ejpam-642	113	1	≥	≥	PROPN
ejpam-642	113	2	wl	wl	PROPN
ejpam-642	113	3	≥	≥	PROPN
ejpam-642	113	4	.	.	PUNCT
ejpam-642	113	5	.	.	PUNCT
ejpam-642	114	1	.	.	PUNCT
ejpam-642	115	1	≥	≥	PROPN
ejpam-642	115	2	wn	wn	INTJ
ejpam-642	115	3	,	,	PUNCT
ejpam-642	115	4	and	and	CCONJ
ejpam-642	115	5	the	the	DET
ejpam-642	115	6	corresponding	correspond	VERB
ejpam-642	115	7	matched	match	VERB
ejpam-642	115	8	and	and	CCONJ
ejpam-642	115	9	unmatched	unmatched	ADJ
ejpam-642	115	10	probability	probability	NOUN
ejpam-642	115	11	sequences	sequence	NOUN
ejpam-642	115	12	are	be	AUX
ejpam-642	115	13	m1	m1	NOUN
ejpam-642	115	14	,	,	PUNCT
ejpam-642	115	15	.	.	PUNCT
ejpam-642	115	16	.	.	PUNCT
ejpam-642	115	17	.	.	PUNCT
ejpam-642	116	1	,	,	PUNCT
ejpam-642	116	2	mk	mk	X
ejpam-642	116	3	,	,	PUNCT
ejpam-642	116	4	|	|	ADV
ejpam-642	116	5	.	.	PUNCT
ejpam-642	116	6	.	.	PUNCT
ejpam-642	116	7	.	.	PUNCT
ejpam-642	117	1	,	,	PUNCT
ejpam-642	117	2	|	|	ADV
ejpam-642	117	3	ml	ml	INTJ
ejpam-642	117	4	,	,	PUNCT
ejpam-642	117	5	.	.	PUNCT
ejpam-642	117	6	.	.	PUNCT
ejpam-642	118	1	.	.	PUNCT
ejpam-642	119	1	,	,	PUNCT
ejpam-642	119	2	mn	mn	PROPN
ejpam-642	119	3	,	,	PUNCT
ejpam-642	119	4	r.	r.	PROPN
ejpam-642	119	5	zhu	zhu	PROPN
ejpam-642	119	6	,	,	PUNCT
ejpam-642	119	7	j.	j.	PROPN
ejpam-642	119	8	zhang	zhang	PROPN
ejpam-642	119	9	,	,	PUNCT
ejpam-642	119	10	d.	d.	PROPN
ejpam-642	119	11	zhang	zhang	PROPN
ejpam-642	119	12	,	,	PUNCT
ejpam-642	119	13	g.	g.	PROPN
ejpam-642	119	14	yan	yan	PROPN
ejpam-642	119	15	/	/	SYM
ejpam-642	119	16	eur	eur	PROPN
ejpam-642	119	17	.	.	PUNCT
ejpam-642	120	1	j.	j.	PROPN
ejpam-642	120	2	pure	pure	PROPN
ejpam-642	120	3	appl	appl	PROPN
ejpam-642	120	4	.	.	PROPN
ejpam-642	120	5	math	math	PROPN
ejpam-642	120	6	,	,	PUNCT
ejpam-642	120	7	3	3	NUM
ejpam-642	120	8	(	(	PUNCT
ejpam-642	120	9	2010	2010	NUM
ejpam-642	120	10	)	)	PUNCT
ejpam-642	120	11	,	,	PUNCT
ejpam-642	120	12	141	141	NUM
ejpam-642	120	13	-	-	SYM
ejpam-642	120	14	162	162	NUM
ejpam-642	120	15	145	145	NUM
ejpam-642	120	16	u1	u1	NOUN
ejpam-642	120	17	,	,	PUNCT
ejpam-642	120	18	.	.	PUNCT
ejpam-642	120	19	.	.	PUNCT
ejpam-642	120	20	.	.	PUNCT
ejpam-642	121	1	,	,	PUNCT
ejpam-642	121	2	uk	uk	PROPN
ejpam-642	121	3	,	,	PUNCT
ejpam-642	121	4	|	|	ADV
ejpam-642	121	5	.	.	PUNCT
ejpam-642	121	6	.	.	PUNCT
ejpam-642	121	7	.	.	PUNCT
ejpam-642	122	1	,	,	PUNCT
ejpam-642	122	2	|	|	ADV
ejpam-642	122	3	ul	ul	INTJ
ejpam-642	122	4	,	,	PUNCT
ejpam-642	122	5	.	.	PUNCT
ejpam-642	122	6	.	.	PUNCT
ejpam-642	122	7	.	.	PUNCT
ejpam-642	123	1	,	,	PUNCT
ejpam-642	123	2	un	un	PROPN
ejpam-642	123	3	,	,	PUNCT
ejpam-642	123	4	where	where	SCONJ
ejpam-642	123	5	k	k	PROPN
ejpam-642	123	6	and	and	CCONJ
ejpam-642	123	7	l	l	NOUN
ejpam-642	123	8	are	be	AUX
ejpam-642	123	9	the	the	DET
ejpam-642	123	10	cutting	cut	VERB
ejpam-642	123	11	points	point	NOUN
ejpam-642	123	12	which	which	PRON
ejpam-642	123	13	partition	partition	NOUN
ejpam-642	123	14	all	all	DET
ejpam-642	123	15	comparison	comparison	NOUN
ejpam-642	123	16	pairs	pair	NOUN
ejpam-642	123	17	into	into	ADP
ejpam-642	123	18	three	three	NUM
ejpam-642	123	19	parts	part	NOUN
ejpam-642	123	20	:	:	PUNCT
ejpam-642	123	21	matched	match	VERB
ejpam-642	123	22	,	,	PUNCT
ejpam-642	123	23	uncertain	uncertain	ADJ
ejpam-642	123	24	and	and	CCONJ
ejpam-642	123	25	unmatched	unmatched	ADJ
ejpam-642	123	26	,	,	PUNCT
ejpam-642	123	27	as	as	SCONJ
ejpam-642	123	28	follows	follow	VERB
ejpam-642	123	29	:	:	PUNCT
ejpam-642	123	30	c1	c1	NOUN
ejpam-642	123	31	:	:	PUNCT
ejpam-642	123	32	j	j	PROPN
ejpam-642	123	33	=	=	SYM
ejpam-642	123	34	1	1	NUM
ejpam-642	123	35	,	,	PUNCT
ejpam-642	123	36	.	.	PUNCT
ejpam-642	123	37	.	.	PUNCT
ejpam-642	123	38	.	.	PUNCT
ejpam-642	124	1	,	,	PUNCT
ejpam-642	124	2	k	k	X
ejpam-642	124	3	;	;	PUNCT
ejpam-642	124	4	matched	match	VERB
ejpam-642	124	5	,	,	PUNCT
ejpam-642	124	6	c2	c2	PROPN
ejpam-642	124	7	:	:	PUNCT
ejpam-642	125	1	j	j	PROPN
ejpam-642	125	2	=	=	PUNCT
ejpam-642	125	3	k+	k+	PROPN
ejpam-642	125	4	1	1	NUM
ejpam-642	125	5	,	,	PUNCT
ejpam-642	125	6	.	.	PUNCT
ejpam-642	125	7	.	.	PUNCT
ejpam-642	125	8	.	.	PUNCT
ejpam-642	126	1	,	,	PUNCT
ejpam-642	127	1	l	l	NOUN
ejpam-642	127	2	−	−	NOUN
ejpam-642	127	3	1	1	NUM
ejpam-642	127	4	;	;	PUNCT
ejpam-642	127	5	uncertain	uncertain	ADJ
ejpam-642	127	6	for	for	ADP
ejpam-642	127	7	clerical	clerical	ADJ
ejpam-642	127	8	review	review	NOUN
ejpam-642	127	9	,	,	PUNCT
ejpam-642	127	10	c3	c3	X
ejpam-642	127	11	:	:	PUNCT
ejpam-642	127	12	j	j	PROPN
ejpam-642	127	13	=	=	SYM
ejpam-642	127	14	l	l	PROPN
ejpam-642	127	15	,	,	PUNCT
ejpam-642	127	16	.	.	PUNCT
ejpam-642	127	17	.	.	PUNCT
ejpam-642	127	18	.	.	PUNCT
ejpam-642	128	1	,	,	PUNCT
ejpam-642	128	2	n	n	PROPN
ejpam-642	128	3	;	;	PUNCT
ejpam-642	128	4	unmatched	unmatched	ADJ
ejpam-642	128	5	.	.	PUNCT
ejpam-642	129	1	the	the	DET
ejpam-642	129	2	cutting	cut	VERB
ejpam-642	129	3	conditions	condition	NOUN
ejpam-642	129	4	are	be	AUX
ejpam-642	129	5	set	set	VERB
ejpam-642	129	6	as	as	ADP
ejpam-642	129	7	k	k	PROPN
ejpam-642	129	8	∑	∑	PROPN
ejpam-642	129	9	j=1	j=1	PROPN
ejpam-642	129	10	u	u	PROPN
ejpam-642	129	11	j	j	PROPN
ejpam-642	129	12	≤	≤	PROPN
ejpam-642	129	13	α1	α1	PROPN
ejpam-642	129	14	,	,	PUNCT
ejpam-642	129	15	n	n	PROPN
ejpam-642	129	16	∑	∑	ADV
ejpam-642	129	17	j	j	PROPN
ejpam-642	129	18	=	=	PROPN
ejpam-642	129	19	l	l	NOUN
ejpam-642	129	20	m	m	VERB
ejpam-642	129	21	j	j	PROPN
ejpam-642	129	22	≤	≤	PROPN
ejpam-642	129	23	α2	α2	PROPN
ejpam-642	129	24	,	,	PUNCT
ejpam-642	129	25	where	where	SCONJ
ejpam-642	129	26	α1	α1	PROPN
ejpam-642	129	27	and	and	CCONJ
ejpam-642	129	28	α2	α2	NOUN
ejpam-642	129	29	are	be	AUX
ejpam-642	129	30	prespecified	prespecifie	VERB
ejpam-642	129	31	.	.	PUNCT
ejpam-642	130	1	a	a	DET
ejpam-642	130	2	good	good	ADJ
ejpam-642	130	3	tutorial	tutorial	NOUN
ejpam-642	130	4	was	be	AUX
ejpam-642	130	5	given	give	VERB
ejpam-642	130	6	by	by	ADP
ejpam-642	130	7	fair	fair	ADJ
ejpam-642	130	8	and	and	CCONJ
ejpam-642	130	9	whitridge	whitridge	NOUN
ejpam-642	130	10	(	(	PUNCT
ejpam-642	130	11	1997	1997	NUM
ejpam-642	130	12	)	)	PUNCT
ejpam-642	130	13	;	;	PUNCT
ejpam-642	130	14	see	see	VERB
ejpam-642	130	15	[	[	X
ejpam-642	130	16	5	5	NUM
ejpam-642	130	17	]	]	PUNCT
ejpam-642	130	18	.	.	PUNCT
ejpam-642	131	1	in	in	ADP
ejpam-642	131	2	addition	addition	NOUN
ejpam-642	131	3	,	,	PUNCT
ejpam-642	131	4	winkler	winkler	NOUN
ejpam-642	131	5	(	(	PUNCT
ejpam-642	131	6	2005	2005	NUM
ejpam-642	131	7	)	)	PUNCT
ejpam-642	131	8	gave	give	VERB
ejpam-642	131	9	an	an	DET
ejpam-642	131	10	overview	overview	NOUN
ejpam-642	131	11	on	on	ADP
ejpam-642	131	12	record	record	NOUN
ejpam-642	131	13	linkage	linkage	NOUN
ejpam-642	131	14	;	;	PUNCT
ejpam-642	131	15	refer	refer	VERB
ejpam-642	131	16	to	to	ADP
ejpam-642	131	17	[	[	X
ejpam-642	131	18	17	17	NUM
ejpam-642	131	19	]	]	PUNCT
ejpam-642	131	20	.	.	PUNCT
ejpam-642	132	1	in	in	ADP
ejpam-642	132	2	the	the	DET
ejpam-642	132	3	literature	literature	NOUN
ejpam-642	132	4	,	,	PUNCT
ejpam-642	132	5	various	various	ADJ
ejpam-642	132	6	models	model	NOUN
ejpam-642	132	7	have	have	AUX
ejpam-642	132	8	been	be	AUX
ejpam-642	132	9	proposed	propose	VERB
ejpam-642	132	10	for	for	ADP
ejpam-642	132	11	the	the	DET
ejpam-642	132	12	mixture	mixture	NOUN
ejpam-642	132	13	population	population	NOUN
ejpam-642	132	14	from	from	ADP
ejpam-642	132	15	different	different	ADJ
ejpam-642	132	16	angles	angle	NOUN
ejpam-642	132	17	.	.	PUNCT
ejpam-642	133	1	we	we	PRON
ejpam-642	133	2	have	have	AUX
ejpam-642	133	3	mentioned	mention	VERB
ejpam-642	133	4	a	a	DET
ejpam-642	133	5	few	few	ADJ
ejpam-642	133	6	in	in	ADP
ejpam-642	133	7	section	section	NOUN
ejpam-642	133	8	1	1	NUM
ejpam-642	133	9	.	.	PUNCT
ejpam-642	134	1	in	in	ADP
ejpam-642	134	2	this	this	DET
ejpam-642	134	3	study	study	NOUN
ejpam-642	134	4	,	,	PUNCT
ejpam-642	134	5	we	we	PRON
ejpam-642	134	6	focus	focus	VERB
ejpam-642	134	7	on	on	ADP
ejpam-642	134	8	loglinear	loglinear	NOUN
ejpam-642	134	9	mixture	mixture	NOUN
ejpam-642	134	10	models	model	NOUN
ejpam-642	134	11	.	.	PUNCT
ejpam-642	135	1	the	the	DET
ejpam-642	135	2	advantage	advantage	NOUN
ejpam-642	135	3	of	of	ADP
ejpam-642	135	4	using	use	VERB
ejpam-642	135	5	a	a	DET
ejpam-642	135	6	parametric	parametric	ADJ
ejpam-642	135	7	model	model	NOUN
ejpam-642	135	8	is	be	AUX
ejpam-642	135	9	that	that	SCONJ
ejpam-642	135	10	we	we	PRON
ejpam-642	135	11	can	can	AUX
ejpam-642	135	12	ignore	ignore	VERB
ejpam-642	135	13	the	the	DET
ejpam-642	135	14	underlying	underlie	VERB
ejpam-642	135	15	error	error	NOUN
ejpam-642	135	16	-	-	PUNCT
ejpam-642	135	17	making	make	VERB
ejpam-642	135	18	mechanism	mechanism	NOUN
ejpam-642	135	19	.	.	PUNCT
ejpam-642	136	1	note	note	VERB
ejpam-642	136	2	that	that	SCONJ
ejpam-642	136	3	the	the	DET
ejpam-642	136	4	mixture	mixture	NOUN
ejpam-642	136	5	models	model	NOUN
ejpam-642	136	6	for	for	ADP
ejpam-642	136	7	record	record	NOUN
ejpam-642	136	8	linkage	linkage	NOUN
ejpam-642	136	9	consist	consist	NOUN
ejpam-642	136	10	of	of	ADP
ejpam-642	136	11	only	only	ADV
ejpam-642	136	12	two	two	NUM
ejpam-642	136	13	components	component	NOUN
ejpam-642	136	14	.	.	PUNCT
ejpam-642	137	1	in	in	ADP
ejpam-642	137	2	next	next	ADJ
ejpam-642	137	3	subsection	subsection	NOUN
ejpam-642	137	4	,	,	PUNCT
ejpam-642	137	5	we	we	PRON
ejpam-642	137	6	will	will	AUX
ejpam-642	137	7	outline	outline	VERB
ejpam-642	137	8	the	the	DET
ejpam-642	137	9	general	general	ADJ
ejpam-642	137	10	framework	framework	NOUN
ejpam-642	137	11	for	for	ADP
ejpam-642	137	12	this	this	DET
ejpam-642	137	13	particular	particular	ADJ
ejpam-642	137	14	type	type	NOUN
ejpam-642	137	15	of	of	ADP
ejpam-642	137	16	mixture	mixture	NOUN
ejpam-642	137	17	models	model	NOUN
ejpam-642	137	18	and	and	CCONJ
ejpam-642	137	19	propose	propose	VERB
ejpam-642	137	20	a	a	DET
ejpam-642	137	21	simplified	simplified	ADJ
ejpam-642	137	22	method	method	NOUN
ejpam-642	137	23	for	for	ADP
ejpam-642	137	24	the	the	DET
ejpam-642	137	25	em	em	PROPN
ejpam-642	137	26	algorithm	algorithm	NOUN
ejpam-642	137	27	as	as	ADV
ejpam-642	137	28	well	well	ADV
ejpam-642	137	29	as	as	ADP
ejpam-642	137	30	the	the	DET
ejpam-642	137	31	choice	choice	NOUN
ejpam-642	137	32	of	of	ADP
ejpam-642	137	33	initial	initial	ADJ
ejpam-642	137	34	parameter	parameter	NOUN
ejpam-642	137	35	values	value	NOUN
ejpam-642	137	36	for	for	ADP
ejpam-642	137	37	this	this	DET
ejpam-642	137	38	method	method	NOUN
ejpam-642	137	39	.	.	PUNCT
ejpam-642	138	1	2.2	2.2	NUM
ejpam-642	138	2	.	.	PUNCT
ejpam-642	138	3	mixture	mixture	NOUN
ejpam-642	138	4	models	model	NOUN
ejpam-642	138	5	and	and	CCONJ
ejpam-642	138	6	the	the	DET
ejpam-642	138	7	simplified	simplified	ADJ
ejpam-642	138	8	method	method	NOUN
ejpam-642	138	9	of	of	ADP
ejpam-642	138	10	the	the	DET
ejpam-642	138	11	em	em	PROPN
ejpam-642	138	12	algorithm	algorithm	NOUN
ejpam-642	138	13	for	for	ADP
ejpam-642	138	14	parameter	parameter	NOUN
ejpam-642	138	15	estimation	estimation	NOUN
ejpam-642	138	16	we	we	PRON
ejpam-642	138	17	consider	consider	VERB
ejpam-642	138	18	parametric	parametric	ADJ
ejpam-642	138	19	models	model	NOUN
ejpam-642	138	20	for	for	ADP
ejpam-642	138	21	both	both	CCONJ
ejpam-642	138	22	the	the	DET
ejpam-642	138	23	matched	match	VERB
ejpam-642	138	24	and	and	CCONJ
ejpam-642	138	25	unmatched	unmatched	ADJ
ejpam-642	138	26	subpopulations	subpopulation	NOUN
ejpam-642	138	27	.	.	PUNCT
ejpam-642	139	1	assume	assume	VERB
ejpam-642	139	2	that	that	SCONJ
ejpam-642	139	3	the	the	DET
ejpam-642	139	4	model	model	NOUN
ejpam-642	139	5	for	for	ADP
ejpam-642	139	6	m	m	PROPN
ejpam-642	139	7	has	have	VERB
ejpam-642	139	8	pmf	pmf	NOUN
ejpam-642	139	9	or	or	CCONJ
ejpam-642	139	10	pdf	pdf	PROPN
ejpam-642	139	11	fm	fm	PROPN
ejpam-642	139	12	(	(	PUNCT
ejpam-642	139	13	γ;βm	γ;βm	PROPN
ejpam-642	139	14	)	)	PUNCT
ejpam-642	139	15	while	while	SCONJ
ejpam-642	139	16	the	the	DET
ejpam-642	139	17	model	model	NOUN
ejpam-642	139	18	for	for	ADP
ejpam-642	139	19	u	u	PROPN
ejpam-642	139	20	has	have	VERB
ejpam-642	139	21	pmf	pmf	NOUN
ejpam-642	139	22	or	or	CCONJ
ejpam-642	139	23	pdf	pdf	NOUN
ejpam-642	139	24	fu(γ;βu	fu(γ;βu	NOUN
ejpam-642	139	25	)	)	PUNCT
ejpam-642	139	26	.	.	PUNCT
ejpam-642	140	1	here	here	ADV
ejpam-642	140	2	γ	γ	PROPN
ejpam-642	140	3	is	be	AUX
ejpam-642	140	4	the	the	DET
ejpam-642	140	5	comparison	comparison	NOUN
ejpam-642	140	6	vector	vector	NOUN
ejpam-642	140	7	,	,	PUNCT
ejpam-642	140	8	and	and	CCONJ
ejpam-642	140	9	βm	βm	VERB
ejpam-642	140	10	and	and	CCONJ
ejpam-642	140	11	βu	βu	ADP
ejpam-642	140	12	are	be	AUX
ejpam-642	140	13	parameter	parameter	NOUN
ejpam-642	140	14	vectors	vector	NOUN
ejpam-642	140	15	associated	associate	VERB
ejpam-642	140	16	with	with	ADP
ejpam-642	140	17	the	the	DET
ejpam-642	140	18	models	model	NOUN
ejpam-642	140	19	for	for	ADP
ejpam-642	140	20	m	m	PROPN
ejpam-642	140	21	and	and	CCONJ
ejpam-642	140	22	u	u	PRON
ejpam-642	140	23	respectively	respectively	ADV
ejpam-642	140	24	.	.	PUNCT
ejpam-642	141	1	suppose	suppose	VERB
ejpam-642	141	2	the	the	DET
ejpam-642	141	3	chance	chance	NOUN
ejpam-642	141	4	of	of	ADP
ejpam-642	141	5	a	a	DET
ejpam-642	141	6	comparison	comparison	NOUN
ejpam-642	141	7	pair	pair	NOUN
ejpam-642	141	8	from	from	ADP
ejpam-642	141	9	m	m	PROPN
ejpam-642	141	10	is	be	AUX
ejpam-642	141	11	π	π	PROPN
ejpam-642	141	12	.	.	PUNCT
ejpam-642	142	1	then	then	ADV
ejpam-642	142	2	the	the	DET
ejpam-642	142	3	pmf	pmf	NOUN
ejpam-642	142	4	or	or	CCONJ
ejpam-642	142	5	pdf	pdf	NOUN
ejpam-642	142	6	for	for	ADP
ejpam-642	142	7	the	the	DET
ejpam-642	142	8	j	j	PROPN
ejpam-642	142	9	-	-	PUNCT
ejpam-642	142	10	th	th	VERB
ejpam-642	142	11	comparison	comparison	NOUN
ejpam-642	142	12	pair	pair	NOUN
ejpam-642	142	13	is	be	AUX
ejpam-642	142	14	fγ(γ	fγ(γ	NOUN
ejpam-642	142	15	j;βm	j;βm	PROPN
ejpam-642	142	16	,	,	PUNCT
ejpam-642	142	17	βu	βu	NOUN
ejpam-642	142	18	,	,	PUNCT
ejpam-642	142	19	π	π	NOUN
ejpam-642	142	20	)	)	PUNCT
ejpam-642	142	21	=	=	PUNCT
ejpam-642	143	1	π	π	PROPN
ejpam-642	143	2	fm	fm	X
ejpam-642	143	3	(	(	PUNCT
ejpam-642	143	4	γ	γ	X
ejpam-642	143	5	j;βm	j;βm	PROPN
ejpam-642	143	6	)	)	PUNCT
ejpam-642	144	1	+	+	CCONJ
ejpam-642	144	2	(	(	PUNCT
ejpam-642	144	3	1−π	1−π	NUM
ejpam-642	144	4	)	)	PUNCT
ejpam-642	144	5	fu(γ	fu(γ	PUNCT
ejpam-642	144	6	j;βu	j;βu	NOUN
ejpam-642	144	7	)	)	PUNCT
ejpam-642	144	8	.	.	PUNCT
ejpam-642	145	1	(	(	PUNCT
ejpam-642	145	2	1	1	X
ejpam-642	145	3	)	)	PUNCT
ejpam-642	145	4	on	on	ADP
ejpam-642	145	5	the	the	DET
ejpam-642	145	6	other	other	ADJ
ejpam-642	145	7	hand	hand	NOUN
ejpam-642	145	8	,	,	PUNCT
ejpam-642	145	9	denote	denote	VERB
ejpam-642	145	10	the	the	DET
ejpam-642	145	11	membership	membership	NOUN
ejpam-642	145	12	for	for	ADP
ejpam-642	145	13	the	the	DET
ejpam-642	145	14	j	j	PROPN
ejpam-642	145	15	-	-	PUNCT
ejpam-642	145	16	th	th	VERB
ejpam-642	145	17	comparison	comparison	NOUN
ejpam-642	145	18	pair	pair	NOUN
ejpam-642	145	19	as	as	SCONJ
ejpam-642	145	20	follows	follow	VERB
ejpam-642	145	21	g	g	PROPN
ejpam-642	145	22	j	j	PROPN
ejpam-642	145	23	=	=	PUNCT
ejpam-642	145	24	(	(	PUNCT
ejpam-642	145	25	1	1	NUM
ejpam-642	145	26	,	,	PUNCT
ejpam-642	145	27	if	if	SCONJ
ejpam-642	145	28	γ	γ	PROPN
ejpam-642	145	29	j	j	PROPN
ejpam-642	145	30	is	be	AUX
ejpam-642	145	31	from	from	ADP
ejpam-642	145	32	the	the	DET
ejpam-642	145	33	matched	match	VERB
ejpam-642	145	34	population	population	NOUN
ejpam-642	145	35	m	m	NOUN
ejpam-642	145	36	,	,	PUNCT
ejpam-642	145	37	0	0	NUM
ejpam-642	145	38	,	,	PUNCT
ejpam-642	145	39	if	if	SCONJ
ejpam-642	145	40	γ	γ	PROPN
ejpam-642	145	41	j	j	PROPN
ejpam-642	145	42	is	be	AUX
ejpam-642	145	43	from	from	ADP
ejpam-642	145	44	the	the	DET
ejpam-642	145	45	unmatched	unmatched	ADJ
ejpam-642	145	46	u	u	PROPN
ejpam-642	145	47	,	,	PUNCT
ejpam-642	145	48	j	j	PROPN
ejpam-642	145	49	=	=	SYM
ejpam-642	145	50	1	1	NUM
ejpam-642	145	51	,	,	PUNCT
ejpam-642	145	52	.	.	PUNCT
ejpam-642	145	53	.	.	PUNCT
ejpam-642	146	1	.	.	PUNCT
ejpam-642	147	1	,	,	PUNCT
ejpam-642	147	2	n	n	X
ejpam-642	147	3	.	.	PUNCT
ejpam-642	148	1	for	for	ADP
ejpam-642	148	2	the	the	DET
ejpam-642	148	3	j	j	PROPN
ejpam-642	148	4	-	-	PUNCT
ejpam-642	148	5	th	th	VERB
ejpam-642	148	6	pair	pair	NOUN
ejpam-642	148	7	,	,	PUNCT
ejpam-642	148	8	the	the	DET
ejpam-642	148	9	complete	complete	ADJ
ejpam-642	148	10	data	data	NOUN
ejpam-642	148	11	is	be	AUX
ejpam-642	148	12	x	x	X
ejpam-642	148	13	j	j	PROPN
ejpam-642	148	14	=	=	PUNCT
ejpam-642	148	15	(	(	PUNCT
ejpam-642	148	16	γ	γ	X
ejpam-642	148	17	j	j	PROPN
ejpam-642	148	18	,	,	PUNCT
ejpam-642	148	19	g	g	PROPN
ejpam-642	148	20	j	j	PROPN
ejpam-642	148	21	)	)	PUNCT
ejpam-642	148	22	.	.	PUNCT
ejpam-642	149	1	thus	thus	ADV
ejpam-642	149	2	,	,	PUNCT
ejpam-642	149	3	the	the	DET
ejpam-642	149	4	pmf	pmf	NOUN
ejpam-642	149	5	or	or	CCONJ
ejpam-642	149	6	pdf	pdf	NOUN
ejpam-642	149	7	for	for	ADP
ejpam-642	149	8	the	the	DET
ejpam-642	149	9	complete	complete	ADJ
ejpam-642	149	10	data	datum	NOUN
ejpam-642	149	11	of	of	ADP
ejpam-642	149	12	the	the	DET
ejpam-642	149	13	j	j	PROPN
ejpam-642	149	14	-	-	PUNCT
ejpam-642	149	15	th	th	VERB
ejpam-642	149	16	pair	pair	NOUN
ejpam-642	149	17	is	be	AUX
ejpam-642	149	18	fx	fx	ADP
ejpam-642	149	19	�	�	PROPN
ejpam-642	149	20	(	(	PUNCT
ejpam-642	149	21	γ	γ	PROPN
ejpam-642	149	22	j	j	PROPN
ejpam-642	149	23	,	,	PUNCT
ejpam-642	149	24	g	g	PROPN
ejpam-642	149	25	j);βm	j);βm	PROPN
ejpam-642	149	26	,	,	PUNCT
ejpam-642	149	27	βu	βu	PROPN
ejpam-642	149	28	,	,	PUNCT
ejpam-642	149	29	π	π	PROPN
ejpam-642	149	30	�	�	PROPN
ejpam-642	149	31	=	=	SYM
ejpam-642	149	32	�	�	PROPN
ejpam-642	149	33	π	π	PROPN
ejpam-642	149	34	fm	fm	PROPN
ejpam-642	149	35	(	(	PUNCT
ejpam-642	149	36	γ	γ	X
ejpam-642	149	37	j;βm	j;βm	PROPN
ejpam-642	149	38	)	)	PUNCT
ejpam-642	149	39	�	�	PROPN
ejpam-642	149	40	g	g	PROPN
ejpam-642	149	41	j	j	PROPN
ejpam-642	149	42	�	�	PROPN
ejpam-642	149	43	(	(	PUNCT
ejpam-642	149	44	1−π	1−π	NUM
ejpam-642	149	45	)	)	PUNCT
ejpam-642	149	46	fu(γ	fu(γ	PUNCT
ejpam-642	149	47	j;βu	j;βu	VERB
ejpam-642	149	48	)	)	PUNCT
ejpam-642	149	49	�	�	PROPN
ejpam-642	149	50	(	(	PUNCT
ejpam-642	149	51	1−g	1−g	NUM
ejpam-642	149	52	j	j	PROPN
ejpam-642	149	53	)	)	PUNCT
ejpam-642	149	54	.	.	PUNCT
ejpam-642	150	1	(	(	PUNCT
ejpam-642	150	2	2	2	X
ejpam-642	150	3	)	)	PUNCT
ejpam-642	150	4	r.	r.	PROPN
ejpam-642	150	5	zhu	zhu	PROPN
ejpam-642	150	6	,	,	PUNCT
ejpam-642	150	7	j.	j.	PROPN
ejpam-642	150	8	zhang	zhang	PROPN
ejpam-642	150	9	,	,	PUNCT
ejpam-642	150	10	d.	d.	PROPN
ejpam-642	150	11	zhang	zhang	PROPN
ejpam-642	150	12	,	,	PUNCT
ejpam-642	150	13	g.	g.	PROPN
ejpam-642	150	14	yan	yan	PROPN
ejpam-642	150	15	/	/	SYM
ejpam-642	150	16	eur	eur	PROPN
ejpam-642	150	17	.	.	PUNCT
ejpam-642	151	1	j.	j.	PROPN
ejpam-642	151	2	pure	pure	PROPN
ejpam-642	151	3	appl	appl	PROPN
ejpam-642	151	4	.	.	PROPN
ejpam-642	151	5	math	math	PROPN
ejpam-642	151	6	,	,	PUNCT
ejpam-642	151	7	3	3	NUM
ejpam-642	151	8	(	(	PUNCT
ejpam-642	151	9	2010	2010	NUM
ejpam-642	151	10	)	)	PUNCT
ejpam-642	151	11	,	,	PUNCT
ejpam-642	151	12	141	141	NUM
ejpam-642	151	13	-	-	SYM
ejpam-642	151	14	162	162	NUM
ejpam-642	151	15	146	146	NUM
ejpam-642	151	16	assume	assume	VERB
ejpam-642	151	17	that	that	SCONJ
ejpam-642	151	18	the	the	DET
ejpam-642	151	19	comparison	comparison	NOUN
ejpam-642	151	20	pairs	pair	NOUN
ejpam-642	151	21	are	be	AUX
ejpam-642	151	22	independent	independent	ADJ
ejpam-642	151	23	.	.	PUNCT
ejpam-642	152	1	the	the	DET
ejpam-642	152	2	logarithm	logarithm	NOUN
ejpam-642	152	3	of	of	ADP
ejpam-642	152	4	the	the	DET
ejpam-642	152	5	joint	joint	ADJ
ejpam-642	152	6	pmf	pmf	NOUN
ejpam-642	152	7	or	or	CCONJ
ejpam-642	152	8	pdf	pdf	NOUN
ejpam-642	152	9	for	for	ADP
ejpam-642	152	10	all	all	DET
ejpam-642	152	11	complete	complete	ADJ
ejpam-642	152	12	data	datum	NOUN
ejpam-642	152	13	is	be	AUX
ejpam-642	152	14	lc	lc	PROPN
ejpam-642	152	15	�	�	PROPN
ejpam-642	152	16	(	(	PUNCT
ejpam-642	152	17	γ1	γ1	PROPN
ejpam-642	152	18	,	,	PUNCT
ejpam-642	152	19	g1	g1	PROPN
ejpam-642	152	20	)	)	PUNCT
ejpam-642	152	21	,	,	PUNCT
ejpam-642	152	22	.	.	PUNCT
ejpam-642	152	23	.	.	PUNCT
ejpam-642	152	24	.	.	PUNCT
ejpam-642	153	1	,	,	PUNCT
ejpam-642	153	2	(	(	PUNCT
ejpam-642	153	3	γn	γn	NOUN
ejpam-642	153	4	,	,	PUNCT
ejpam-642	153	5	gn	gn	PROPN
ejpam-642	153	6	)	)	PUNCT
ejpam-642	153	7	;	;	PUNCT
ejpam-642	153	8	βm	βm	VERB
ejpam-642	153	9	,	,	PUNCT
ejpam-642	153	10	βu	βu	PRON
ejpam-642	153	11	,	,	PUNCT
ejpam-642	153	12	π	π	PROPN
ejpam-642	153	13	�	�	PROPN
ejpam-642	153	14	=	=	SYM
ejpam-642	153	15	log	log	NOUN
ejpam-642	153	16			NOUN
ejpam-642	153	17			NOUN
ejpam-642	154	1			PROPN
ejpam-642	154	2	n	n	CCONJ
ejpam-642	154	3	∏	∏	NOUN
ejpam-642	154	4	j=1	j=1	PROPN
ejpam-642	154	5	fx	fx	PROPN
ejpam-642	154	6	�	�	PROPN
ejpam-642	154	7	(	(	PUNCT
ejpam-642	154	8	γ	γ	PROPN
ejpam-642	154	9	j	j	PROPN
ejpam-642	154	10	,	,	PUNCT
ejpam-642	154	11	g	g	PROPN
ejpam-642	154	12	j);βm	j);βm	PROPN
ejpam-642	154	13	,	,	PUNCT
ejpam-642	154	14	βu	βu	PROPN
ejpam-642	154	15	,	,	PUNCT
ejpam-642	154	16	π	π	PROPN
ejpam-642	154	17	�	�	PROPN
ejpam-642	154	18			PROPN
ejpam-642	154	19			VERB
ejpam-642	154	20			PUNCT
ejpam-642	155	1	=	=	SYM
ejpam-642	155	2	n	n	X
ejpam-642	155	3	∑	∑	PROPN
ejpam-642	155	4	j=1	j=1	PROPN
ejpam-642	155	5	�	�	PROPN
ejpam-642	155	6	g	g	PROPN
ejpam-642	155	7	j	j	PROPN
ejpam-642	155	8	log	log	PROPN
ejpam-642	155	9	fm	fm	PROPN
ejpam-642	155	10	(	(	PUNCT
ejpam-642	155	11	γ	γ	X
ejpam-642	155	12	j;βm	j;βm	PROPN
ejpam-642	155	13	)	)	PUNCT
ejpam-642	156	1	+	+	CCONJ
ejpam-642	156	2	(	(	PUNCT
ejpam-642	156	3	1−	1−	NUM
ejpam-642	156	4	g	g	PROPN
ejpam-642	156	5	j	j	PROPN
ejpam-642	156	6	)	)	PUNCT
ejpam-642	156	7	log	log	NOUN
ejpam-642	156	8	fu(γ	fu(γ	PUNCT
ejpam-642	156	9	j;βu	j;βu	PROPN
ejpam-642	156	10	)	)	PUNCT
ejpam-642	157	1	+	+	ADP
ejpam-642	157	2	g	g	PROPN
ejpam-642	157	3	j	j	PROPN
ejpam-642	157	4	logπ+	logπ+	X
ejpam-642	157	5	(	(	PUNCT
ejpam-642	157	6	1−	1−	NUM
ejpam-642	157	7	g	g	PROPN
ejpam-642	157	8	j	j	PROPN
ejpam-642	157	9	)	)	PUNCT
ejpam-642	157	10	log(1−π	log(1−π	PROPN
ejpam-642	157	11	)	)	PUNCT
ejpam-642	157	12	�	�	PROPN
ejpam-642	157	13	.	.	PUNCT
ejpam-642	158	1	(	(	PUNCT
ejpam-642	158	2	3	3	X
ejpam-642	158	3	)	)	PUNCT
ejpam-642	158	4	the	the	DET
ejpam-642	158	5	conditional	conditional	ADJ
ejpam-642	158	6	pmf	pmf	NOUN
ejpam-642	158	7	for	for	ADP
ejpam-642	158	8	the	the	DET
ejpam-642	158	9	membership	membership	NOUN
ejpam-642	158	10	g	g	NOUN
ejpam-642	158	11	is	be	AUX
ejpam-642	158	12	then	then	ADV
ejpam-642	158	13	fg|γ	fg|γ	PROPN
ejpam-642	158	14	�	�	PROPN
ejpam-642	159	1	g	g	PROPN
ejpam-642	159	2	j	j	PROPN
ejpam-642	159	3	|	|	ADV
ejpam-642	159	4	γ	γ	PROPN
ejpam-642	159	5	j;βm	j;βm	PROPN
ejpam-642	159	6	,	,	PUNCT
ejpam-642	159	7	βu	βu	SYM
ejpam-642	159	8	,	,	PUNCT
ejpam-642	159	9	π	π	PROPN
ejpam-642	159	10	�	�	PROPN
ejpam-642	159	11	=	=	SYM
ejpam-642	159	12	fx	fx	PROPN
ejpam-642	159	13	�	�	PROPN
ejpam-642	159	14	(	(	PUNCT
ejpam-642	159	15	γ	γ	PROPN
ejpam-642	159	16	j	j	PROPN
ejpam-642	159	17	,	,	PUNCT
ejpam-642	159	18	g	g	PROPN
ejpam-642	159	19	j);βm	j);βm	PROPN
ejpam-642	159	20	,	,	PUNCT
ejpam-642	159	21	βu	βu	PROPN
ejpam-642	159	22	,	,	PUNCT
ejpam-642	159	23	π	π	PROPN
ejpam-642	159	24	�	�	PROPN
ejpam-642	159	25	fγ(γ	fγ(γ	VERB
ejpam-642	159	26	j;βm	j;βm	PROPN
ejpam-642	159	27	,	,	PUNCT
ejpam-642	159	28	βu	βu	NOUN
ejpam-642	159	29	,	,	PUNCT
ejpam-642	159	30	π	π	NOUN
ejpam-642	159	31	)	)	PUNCT
ejpam-642	159	32	=	=	SYM
ejpam-642	159	33	fx	fx	PROPN
ejpam-642	159	34	�	�	PROPN
ejpam-642	159	35	(	(	PUNCT
ejpam-642	159	36	γ	γ	PROPN
ejpam-642	159	37	j	j	PROPN
ejpam-642	159	38	,	,	PUNCT
ejpam-642	159	39	g	g	PROPN
ejpam-642	159	40	j);βm	j);βm	PROPN
ejpam-642	159	41	,	,	PUNCT
ejpam-642	159	42	βu	βu	PRON
ejpam-642	159	43	,	,	PUNCT
ejpam-642	159	44	π	π	PROPN
ejpam-642	159	45	�	�	PROPN
ejpam-642	159	46	π	π	PROPN
ejpam-642	159	47	fm	fm	PROPN
ejpam-642	159	48	(	(	PUNCT
ejpam-642	159	49	γ	γ	X
ejpam-642	159	50	j;βm	j;βm	PROPN
ejpam-642	159	51	)	)	PUNCT
ejpam-642	160	1	+	+	CCONJ
ejpam-642	160	2	(	(	PUNCT
ejpam-642	160	3	1−π	1−π	NUM
ejpam-642	160	4	)	)	PUNCT
ejpam-642	160	5	fu(γ	fu(γ	PUNCT
ejpam-642	160	6	j;βu	j;βu	VERB
ejpam-642	160	7	)	)	PUNCT
ejpam-642	160	8	.	.	PUNCT
ejpam-642	161	1	(	(	PUNCT
ejpam-642	161	2	4	4	X
ejpam-642	161	3	)	)	PUNCT
ejpam-642	161	4	however	however	ADV
ejpam-642	161	5	,	,	PUNCT
ejpam-642	161	6	the	the	DET
ejpam-642	161	7	membership	membership	NOUN
ejpam-642	161	8	g	g	PROPN
ejpam-642	161	9	j	j	PROPN
ejpam-642	161	10	is	be	AUX
ejpam-642	161	11	missing	miss	VERB
ejpam-642	161	12	in	in	ADP
ejpam-642	161	13	the	the	DET
ejpam-642	161	14	mixture	mixture	NOUN
ejpam-642	161	15	population	population	NOUN
ejpam-642	161	16	.	.	PUNCT
ejpam-642	162	1	therefore	therefore	ADV
ejpam-642	162	2	,	,	PUNCT
ejpam-642	162	3	the	the	DET
ejpam-642	162	4	mixture	mixture	NOUN
ejpam-642	162	5	can	can	AUX
ejpam-642	162	6	be	be	AUX
ejpam-642	162	7	regarded	regard	VERB
ejpam-642	162	8	as	as	ADP
ejpam-642	162	9	a	a	DET
ejpam-642	162	10	missing	missing	ADJ
ejpam-642	162	11	value	value	NOUN
ejpam-642	162	12	problem	problem	NOUN
ejpam-642	162	13	and	and	CCONJ
ejpam-642	162	14	the	the	DET
ejpam-642	162	15	em	em	PROPN
ejpam-642	162	16	algorithm	algorithm	NOUN
ejpam-642	162	17	can	can	AUX
ejpam-642	162	18	be	be	AUX
ejpam-642	162	19	utilized	utilize	VERB
ejpam-642	162	20	to	to	PART
ejpam-642	162	21	estimate	estimate	VERB
ejpam-642	162	22	the	the	DET
ejpam-642	162	23	model	model	NOUN
ejpam-642	162	24	parameters	parameter	NOUN
ejpam-642	162	25	βm	βm	VERB
ejpam-642	162	26	,	,	PUNCT
ejpam-642	162	27	βu	βu	PUNCT
ejpam-642	162	28	and	and	CCONJ
ejpam-642	162	29	π	π	X
ejpam-642	162	30	.	.	PROPN
ejpam-642	163	1	for	for	ADP
ejpam-642	163	2	references	reference	NOUN
ejpam-642	163	3	on	on	ADP
ejpam-642	163	4	the	the	DET
ejpam-642	163	5	em	em	PROPN
ejpam-642	163	6	algorithm	algorithm	NOUN
ejpam-642	163	7	,	,	PUNCT
ejpam-642	163	8	see	see	VERB
ejpam-642	163	9	[	[	X
ejpam-642	163	10	4	4	NUM
ejpam-642	163	11	]	]	PUNCT
ejpam-642	163	12	,	,	PUNCT
ejpam-642	163	13	[	[	X
ejpam-642	163	14	11	11	NUM
ejpam-642	163	15	]	]	PUNCT
ejpam-642	163	16	and	and	CCONJ
ejpam-642	163	17	references	reference	NOUN
ejpam-642	163	18	therein	therein	ADV
ejpam-642	163	19	.	.	PUNCT
ejpam-642	164	1	the	the	DET
ejpam-642	164	2	em	em	PROPN
ejpam-642	164	3	algorithm	algorithm	NOUN
ejpam-642	164	4	is	be	AUX
ejpam-642	164	5	an	an	DET
ejpam-642	164	6	iterative	iterative	NOUN
ejpam-642	164	7	approach	approach	NOUN
ejpam-642	164	8	which	which	PRON
ejpam-642	164	9	alternates	alternate	VERB
ejpam-642	164	10	between	between	ADP
ejpam-642	164	11	two	two	NUM
ejpam-642	164	12	steps	step	NOUN
ejpam-642	164	13	,	,	PUNCT
ejpam-642	164	14	an	an	DET
ejpam-642	164	15	e	e	NOUN
ejpam-642	164	16	-	-	NOUN
ejpam-642	164	17	step	step	NOUN
ejpam-642	164	18	and	and	CCONJ
ejpam-642	164	19	an	an	DET
ejpam-642	164	20	m	m	NOUN
ejpam-642	164	21	-	-	NOUN
ejpam-642	164	22	step	step	NOUN
ejpam-642	164	23	.	.	PUNCT
ejpam-642	165	1	the	the	DET
ejpam-642	165	2	e	e	NOUN
ejpam-642	165	3	-	-	NOUN
ejpam-642	165	4	step	step	NOUN
ejpam-642	165	5	is	be	AUX
ejpam-642	165	6	conducted	conduct	VERB
ejpam-642	165	7	as	as	SCONJ
ejpam-642	165	8	follows	follow	VERB
ejpam-642	165	9	.	.	PUNCT
ejpam-642	166	1	conditional	conditional	ADJ
ejpam-642	166	2	on	on	ADP
ejpam-642	166	3	values	value	NOUN
ejpam-642	166	4	of	of	ADP
ejpam-642	166	5	β	β	X
ejpam-642	166	6	(	(	PUNCT
ejpam-642	166	7	k)m	k)m	X
ejpam-642	166	8	,	,	PUNCT
ejpam-642	166	9	β	β	X
ejpam-642	166	10	(	(	PUNCT
ejpam-642	166	11	k)u	k)u	PROPN
ejpam-642	166	12	and	and	CCONJ
ejpam-642	166	13	π(k	π(k	NUM
ejpam-642	166	14	)	)	PUNCT
ejpam-642	166	15	,	,	PUNCT
ejpam-642	166	16	as	as	ADV
ejpam-642	166	17	well	well	ADV
ejpam-642	166	18	as	as	ADP
ejpam-642	166	19	the	the	DET
ejpam-642	166	20	observed	observed	ADJ
ejpam-642	166	21	comparison	comparison	NOUN
ejpam-642	166	22	vectors	vector	NOUN
ejpam-642	166	23	,	,	PUNCT
ejpam-642	166	24	the	the	DET
ejpam-642	166	25	expectation	expectation	NOUN
ejpam-642	166	26	of	of	ADP
ejpam-642	166	27	the	the	DET
ejpam-642	166	28	logarithm	logarithm	NOUN
ejpam-642	166	29	of	of	ADP
ejpam-642	166	30	the	the	DET
ejpam-642	166	31	joint	joint	ADJ
ejpam-642	166	32	pmf	pmf	NOUN
ejpam-642	166	33	or	or	CCONJ
ejpam-642	166	34	pdf	pdf	NOUN
ejpam-642	166	35	for	for	ADP
ejpam-642	166	36	all	all	DET
ejpam-642	166	37	complete	complete	ADJ
ejpam-642	166	38	data	datum	NOUN
ejpam-642	166	39	is	be	AUX
ejpam-642	166	40	l	l	NOUN
ejpam-642	166	41	=	=	PUNCT
ejpam-642	166	42	e	e	PART
ejpam-642	166	43	�	�	PROPN
ejpam-642	166	44	lc	lc	PROPN
ejpam-642	166	45	�	�	PROPN
ejpam-642	166	46	(	(	PUNCT
ejpam-642	166	47	γ1	γ1	PROPN
ejpam-642	166	48	,	,	PUNCT
ejpam-642	166	49	g1	g1	PROPN
ejpam-642	166	50	)	)	PUNCT
ejpam-642	166	51	,	,	PUNCT
ejpam-642	166	52	.	.	PUNCT
ejpam-642	166	53	.	.	PUNCT
ejpam-642	167	1	.	.	PUNCT
ejpam-642	168	1	,	,	PUNCT
ejpam-642	168	2	(	(	PUNCT
ejpam-642	168	3	γn	γn	NOUN
ejpam-642	168	4	,	,	PUNCT
ejpam-642	168	5	gn	gn	PROPN
ejpam-642	168	6	)	)	PUNCT
ejpam-642	168	7	;	;	PUNCT
ejpam-642	168	8	βm	βm	VERB
ejpam-642	168	9	,	,	PUNCT
ejpam-642	168	10	βu	βu	PRON
ejpam-642	168	11	,	,	PUNCT
ejpam-642	168	12	π	π	PROPN
ejpam-642	168	13	�	�	PROPN
ejpam-642	168	14	|	|	ADV
ejpam-642	168	15	γ1	γ1	PROPN
ejpam-642	168	16	,	,	PUNCT
ejpam-642	168	17	.	.	PUNCT
ejpam-642	168	18	.	.	PUNCT
ejpam-642	169	1	.	.	PUNCT
ejpam-642	170	1	,	,	PUNCT
ejpam-642	170	2	γn	γn	X
ejpam-642	170	3	;	;	PUNCT
ejpam-642	170	4	β	β	X
ejpam-642	170	5	(	(	PUNCT
ejpam-642	170	6	k)m	k)m	X
ejpam-642	170	7	,	,	PUNCT
ejpam-642	170	8	β	β	X
ejpam-642	170	9	(	(	PUNCT
ejpam-642	170	10	k)u	k)u	X
ejpam-642	170	11	,	,	PUNCT
ejpam-642	170	12	π(k	π(k	PROPN
ejpam-642	170	13	)	)	PUNCT
ejpam-642	170	14	�	�	PROPN
ejpam-642	170	15	=	=	SYM
ejpam-642	170	16	n	n	PROPN
ejpam-642	170	17	∑	∑	PROPN
ejpam-642	170	18	j=1	j=1	PROPN
ejpam-642	170	19	e	e	PROPN
ejpam-642	170	20	�	�	PROPN
ejpam-642	170	21	log	log	VERB
ejpam-642	170	22	�	�	PROPN
ejpam-642	170	23	fx	fx	PROPN
ejpam-642	170	24	�	�	PROPN
ejpam-642	170	25	(	(	PUNCT
ejpam-642	170	26	γ	γ	PROPN
ejpam-642	170	27	j	j	PROPN
ejpam-642	170	28	,	,	PUNCT
ejpam-642	170	29	g	g	PROPN
ejpam-642	170	30	j);βm	j);βm	PROPN
ejpam-642	170	31	,	,	PUNCT
ejpam-642	170	32	βu	βu	PROPN
ejpam-642	170	33	,	,	PUNCT
ejpam-642	170	34	π	π	PROPN
ejpam-642	170	35	�	�	PROPN
ejpam-642	170	36	�	�	PROPN
ejpam-642	170	37	|	|	ADV
ejpam-642	170	38	γ	γ	X
ejpam-642	170	39	j;β	j;β	PROPN
ejpam-642	170	40	(	(	PUNCT
ejpam-642	170	41	k)m	k)m	PROPN
ejpam-642	170	42	,	,	PUNCT
ejpam-642	170	43	β	β	X
ejpam-642	170	44	(	(	PUNCT
ejpam-642	170	45	k)u	k)u	X
ejpam-642	170	46	,	,	PUNCT
ejpam-642	170	47	π(k	π(k	PROPN
ejpam-642	170	48	)	)	PUNCT
ejpam-642	170	49	�	�	PROPN
ejpam-642	170	50	=	=	SYM
ejpam-642	170	51	n	n	CCONJ
ejpam-642	170	52	∑	∑	PUNCT
ejpam-642	170	53	j=1	j=1	PROPN
ejpam-642	170	54	∫	∫	PROPN
ejpam-642	170	55	log	log	PROPN
ejpam-642	170	56	�	�	PROPN
ejpam-642	170	57	fx	fx	PROPN
ejpam-642	170	58	�	�	PROPN
ejpam-642	170	59	(	(	PUNCT
ejpam-642	170	60	γ	γ	PROPN
ejpam-642	170	61	j	j	PROPN
ejpam-642	170	62	,	,	PUNCT
ejpam-642	170	63	g	g	PROPN
ejpam-642	170	64	j);βm	j);βm	PROPN
ejpam-642	170	65	,	,	PUNCT
ejpam-642	170	66	βu	βu	PROPN
ejpam-642	170	67	,	,	PUNCT
ejpam-642	170	68	π	π	PROPN
ejpam-642	170	69	�	�	PROPN
ejpam-642	170	70	�	�	PROPN
ejpam-642	170	71	fg|γ	fg|γ	PROPN
ejpam-642	170	72	�	�	PROPN
ejpam-642	171	1	g	g	PROPN
ejpam-642	171	2	j	j	PROPN
ejpam-642	171	3	|	|	ADV
ejpam-642	171	4	γ	γ	X
ejpam-642	171	5	j;β	j;β	PROPN
ejpam-642	171	6	(	(	PUNCT
ejpam-642	171	7	k)m	k)m	PROPN
ejpam-642	171	8	,	,	PUNCT
ejpam-642	171	9	β	β	X
ejpam-642	171	10	(	(	PUNCT
ejpam-642	171	11	k)u	k)u	X
ejpam-642	171	12	,	,	PUNCT
ejpam-642	171	13	π(k	π(k	PROPN
ejpam-642	171	14	)	)	PUNCT
ejpam-642	171	15	�	�	PROPN
ejpam-642	172	1	d	d	X
ejpam-642	172	2	g	g	PROPN
ejpam-642	172	3	j	j	PROPN
ejpam-642	172	4	=	=	SYM
ejpam-642	173	1	n	n	PROPN
ejpam-642	173	2	∑	∑	ADV
ejpam-642	173	3	j=1	j=1	NOUN
ejpam-642	173	4	1	1	NUM
ejpam-642	173	5	∑	∑	PROPN
ejpam-642	173	6	g	g	PROPN
ejpam-642	173	7	j=0	j=0	PROPN
ejpam-642	173	8	¦	¦	PROPN
ejpam-642	173	9	�	�	PROPN
ejpam-642	173	10	g	g	PROPN
ejpam-642	173	11	j	j	PROPN
ejpam-642	173	12	log	log	PROPN
ejpam-642	173	13	fm	fm	PROPN
ejpam-642	173	14	(	(	PUNCT
ejpam-642	173	15	γ	γ	X
ejpam-642	173	16	j;βm	j;βm	PROPN
ejpam-642	173	17	)	)	PUNCT
ejpam-642	174	1	+	+	CCONJ
ejpam-642	174	2	(	(	PUNCT
ejpam-642	174	3	1−	1−	NUM
ejpam-642	174	4	g	g	PROPN
ejpam-642	174	5	j	j	PROPN
ejpam-642	174	6	)	)	PUNCT
ejpam-642	174	7	log	log	NOUN
ejpam-642	174	8	fu(γ	fu(γ	PUNCT
ejpam-642	174	9	j;βu	j;βu	PROPN
ejpam-642	174	10	)	)	PUNCT
ejpam-642	175	1	+	+	ADP
ejpam-642	175	2	g	g	PROPN
ejpam-642	175	3	j	j	PROPN
ejpam-642	175	4	logπ+	logπ+	X
ejpam-642	175	5	(	(	PUNCT
ejpam-642	175	6	1−	1−	NUM
ejpam-642	175	7	g	g	PROPN
ejpam-642	175	8	j	j	PROPN
ejpam-642	175	9	)	)	PUNCT
ejpam-642	175	10	log(1−π	log(1−π	PROPN
ejpam-642	175	11	)	)	PUNCT
ejpam-642	175	12	�	�	PROPN
ejpam-642	175	13	×	×	PROPN
ejpam-642	175	14	fg|γ	fg|γ	PROPN
ejpam-642	175	15	�	�	PROPN
ejpam-642	175	16	g	g	PROPN
ejpam-642	175	17	j	j	PROPN
ejpam-642	175	18	|	|	ADV
ejpam-642	175	19	γ	γ	X
ejpam-642	175	20	j;β	j;β	PROPN
ejpam-642	175	21	(	(	PUNCT
ejpam-642	175	22	k)m	k)m	PROPN
ejpam-642	175	23	,	,	PUNCT
ejpam-642	175	24	β	β	X
ejpam-642	175	25	(	(	PUNCT
ejpam-642	175	26	k)u	k)u	X
ejpam-642	175	27	,	,	PUNCT
ejpam-642	175	28	π(k	π(k	PROPN
ejpam-642	175	29	)	)	PUNCT
ejpam-642	175	30	�	�	PROPN
ejpam-642	175	31	©	©	NOUN
ejpam-642	175	32	.	.	PUNCT
ejpam-642	176	1	let	let	VERB
ejpam-642	176	2	g(k)m	g(k)m	NOUN
ejpam-642	176	3	(	(	PUNCT
ejpam-642	176	4	γ	γ	X
ejpam-642	176	5	j	j	PROPN
ejpam-642	176	6	)	)	PUNCT
ejpam-642	176	7	=	=	SYM
ejpam-642	177	1	fg|γ	fg|γ	PROPN
ejpam-642	177	2	�	�	PROPN
ejpam-642	177	3	1	1	NUM
ejpam-642	178	1	|	|	ADV
ejpam-642	178	2	γ	γ	X
ejpam-642	178	3	j;β	j;β	PROPN
ejpam-642	178	4	(	(	PUNCT
ejpam-642	178	5	k)m	k)m	PROPN
ejpam-642	178	6	,	,	PUNCT
ejpam-642	178	7	β	β	X
ejpam-642	178	8	(	(	PUNCT
ejpam-642	178	9	k)u	k)u	X
ejpam-642	178	10	,	,	PUNCT
ejpam-642	178	11	π(k	π(k	PROPN
ejpam-642	178	12	)	)	PUNCT
ejpam-642	178	13	�	�	PROPN
ejpam-642	178	14	=	=	SYM
ejpam-642	178	15	π(k	π(k	PROPN
ejpam-642	178	16	)	)	PUNCT
ejpam-642	178	17	fm	fm	NOUN
ejpam-642	178	18	(	(	PUNCT
ejpam-642	178	19	γ	γ	X
ejpam-642	178	20	j;β	j;β	PROPN
ejpam-642	178	21	(	(	PUNCT
ejpam-642	178	22	k)m	k)m	PROPN
ejpam-642	178	23	)	)	PUNCT
ejpam-642	178	24	π(k	π(k	PROPN
ejpam-642	178	25	)	)	PUNCT
ejpam-642	178	26	fm	fm	NOUN
ejpam-642	178	27	(	(	PUNCT
ejpam-642	178	28	γ	γ	X
ejpam-642	178	29	j;β	j;β	PROPN
ejpam-642	178	30	(	(	PUNCT
ejpam-642	178	31	k)m	k)m	PROPN
ejpam-642	178	32	)	)	PUNCT
ejpam-642	179	1	+	+	CCONJ
ejpam-642	179	2	(	(	PUNCT
ejpam-642	179	3	1−π	1−π	NUM
ejpam-642	179	4	(	(	PUNCT
ejpam-642	179	5	k	k	NOUN
ejpam-642	179	6	)	)	PUNCT
ejpam-642	179	7	)	)	PUNCT
ejpam-642	179	8	fu(γ	fu(γ	PROPN
ejpam-642	179	9	j;β	j;β	PROPN
ejpam-642	179	10	(	(	PUNCT
ejpam-642	179	11	k)u	k)u	PROPN
ejpam-642	179	12	)	)	PUNCT
ejpam-642	179	13	,	,	PUNCT
ejpam-642	179	14	r.	r.	PROPN
ejpam-642	179	15	zhu	zhu	PROPN
ejpam-642	179	16	,	,	PUNCT
ejpam-642	179	17	j.	j.	PROPN
ejpam-642	179	18	zhang	zhang	PROPN
ejpam-642	179	19	,	,	PUNCT
ejpam-642	179	20	d.	d.	PROPN
ejpam-642	179	21	zhang	zhang	PROPN
ejpam-642	179	22	,	,	PUNCT
ejpam-642	179	23	g.	g.	PROPN
ejpam-642	179	24	yan	yan	PROPN
ejpam-642	179	25	/	/	SYM
ejpam-642	179	26	eur	eur	PROPN
ejpam-642	179	27	.	.	PUNCT
ejpam-642	180	1	j.	j.	PROPN
ejpam-642	180	2	pure	pure	PROPN
ejpam-642	180	3	appl	appl	PROPN
ejpam-642	180	4	.	.	PROPN
ejpam-642	180	5	math	math	PROPN
ejpam-642	180	6	,	,	PUNCT
ejpam-642	180	7	3	3	NUM
ejpam-642	180	8	(	(	PUNCT
ejpam-642	180	9	2010	2010	NUM
ejpam-642	180	10	)	)	PUNCT
ejpam-642	180	11	,	,	PUNCT
ejpam-642	180	12	141	141	NUM
ejpam-642	180	13	-	-	SYM
ejpam-642	180	14	162	162	NUM
ejpam-642	180	15	147	147	NUM
ejpam-642	180	16	g(k)u	g(k)u	PROPN
ejpam-642	180	17	(	(	PUNCT
ejpam-642	180	18	γ	γ	X
ejpam-642	180	19	j	j	PROPN
ejpam-642	180	20	)	)	PUNCT
ejpam-642	180	21	=	=	SYM
ejpam-642	180	22	fg|γ	fg|γ	PROPN
ejpam-642	180	23	�	�	PROPN
ejpam-642	180	24	0	0	NUM
ejpam-642	181	1	|	|	ADV
ejpam-642	181	2	γ	γ	X
ejpam-642	181	3	j;β	j;β	PROPN
ejpam-642	181	4	(	(	PUNCT
ejpam-642	181	5	k)m	k)m	PROPN
ejpam-642	181	6	,	,	PUNCT
ejpam-642	181	7	β	β	X
ejpam-642	181	8	(	(	PUNCT
ejpam-642	181	9	k)u	k)u	X
ejpam-642	181	10	,	,	PUNCT
ejpam-642	181	11	π(k	π(k	PROPN
ejpam-642	181	12	)	)	PUNCT
ejpam-642	181	13	�	�	PROPN
ejpam-642	181	14	=	=	SYM
ejpam-642	181	15	(	(	PUNCT
ejpam-642	181	16	1−π(k	1−π(k	NUM
ejpam-642	181	17	)	)	PUNCT
ejpam-642	181	18	)	)	PUNCT
ejpam-642	182	1	fu(γ	fu(γ	PROPN
ejpam-642	182	2	j;β	j;β	PROPN
ejpam-642	182	3	(	(	PUNCT
ejpam-642	182	4	k)u	k)u	PROPN
ejpam-642	182	5	)	)	PUNCT
ejpam-642	182	6	π(k	π(k	PROPN
ejpam-642	182	7	)	)	PUNCT
ejpam-642	182	8	fm	fm	NOUN
ejpam-642	182	9	(	(	PUNCT
ejpam-642	182	10	γ	γ	X
ejpam-642	182	11	j;β	j;β	PROPN
ejpam-642	182	12	(	(	PUNCT
ejpam-642	182	13	k)m	k)m	PROPN
ejpam-642	182	14	)	)	PUNCT
ejpam-642	183	1	+	+	CCONJ
ejpam-642	183	2	(	(	PUNCT
ejpam-642	183	3	1−π	1−π	NUM
ejpam-642	183	4	(	(	PUNCT
ejpam-642	183	5	k	k	NOUN
ejpam-642	183	6	)	)	PUNCT
ejpam-642	183	7	)	)	PUNCT
ejpam-642	183	8	fu(γ	fu(γ	PROPN
ejpam-642	183	9	j;β	j;β	PROPN
ejpam-642	183	10	(	(	PUNCT
ejpam-642	183	11	k)u	k)u	X
ejpam-642	183	12	)	)	PUNCT
ejpam-642	184	1	=	=	SYM
ejpam-642	184	2	1−	1−	NUM
ejpam-642	184	3	gm(γ	gm(γ	NUM
ejpam-642	184	4	j	j	NOUN
ejpam-642	184	5	)	)	PUNCT
ejpam-642	184	6	.	.	PUNCT
ejpam-642	185	1	then	then	ADV
ejpam-642	185	2	,	,	PUNCT
ejpam-642	185	3	by	by	ADP
ejpam-642	185	4	algebra	algebra	NOUN
ejpam-642	185	5	,	,	PUNCT
ejpam-642	185	6	l	l	NOUN
ejpam-642	185	7	=	=	SYM
ejpam-642	185	8	n	n	CCONJ
ejpam-642	185	9	∑	∑	ADP
ejpam-642	185	10	j=1	j=1	ADJ
ejpam-642	185	11	g(k)m	g(k)m	NOUN
ejpam-642	185	12	(	(	PUNCT
ejpam-642	185	13	γ	γ	X
ejpam-642	185	14	j	j	PROPN
ejpam-642	185	15	)	)	PUNCT
ejpam-642	185	16	log	log	PROPN
ejpam-642	185	17	fm	fm	PROPN
ejpam-642	185	18	(	(	PUNCT
ejpam-642	185	19	γ	γ	X
ejpam-642	185	20	j;βm	j;βm	PROPN
ejpam-642	185	21	)	)	PUNCT
ejpam-642	186	1	+	+	CCONJ
ejpam-642	186	2	n	n	CCONJ
ejpam-642	186	3	∑	∑	ADP
ejpam-642	186	4	j=1	j=1	PROPN
ejpam-642	186	5	g(k)u	g(k)u	PROPN
ejpam-642	186	6	(	(	PUNCT
ejpam-642	186	7	γ	γ	PROPN
ejpam-642	186	8	j	j	PROPN
ejpam-642	186	9	)	)	PUNCT
ejpam-642	186	10	log	log	PROPN
ejpam-642	186	11	fu	fu	NOUN
ejpam-642	186	12	(	(	PUNCT
ejpam-642	186	13	γ	γ	PROPN
ejpam-642	186	14	j;βu	j;βu	PROPN
ejpam-642	186	15	)	)	PUNCT
ejpam-642	186	16	+	+	CCONJ
ejpam-642	186	17	n	n	CCONJ
ejpam-642	186	18	∑	∑	ADP
ejpam-642	186	19	j=1	j=1	PROPN
ejpam-642	186	20	�	�	PROPN
ejpam-642	186	21	g(k)m	g(k)m	PROPN
ejpam-642	186	22	(	(	PUNCT
ejpam-642	186	23	γ	γ	X
ejpam-642	186	24	j	j	PROPN
ejpam-642	186	25	)	)	PUNCT
ejpam-642	186	26	logπ+	logπ+	PART
ejpam-642	186	27	g(k)u	g(k)u	PROPN
ejpam-642	186	28	(	(	PUNCT
ejpam-642	186	29	γ	γ	X
ejpam-642	186	30	j	j	PROPN
ejpam-642	186	31	)	)	PUNCT
ejpam-642	186	32	log(1−π	log(1−π	PROPN
ejpam-642	186	33	)	)	PUNCT
ejpam-642	186	34	�	�	PROPN
ejpam-642	186	35	=	=	SYM
ejpam-642	186	36	l1	l1	PROPN
ejpam-642	186	37	+	+	CCONJ
ejpam-642	186	38	l2	l2	NOUN
ejpam-642	186	39	+	+	CCONJ
ejpam-642	186	40	l3	l3	NOUN
ejpam-642	186	41	,	,	PUNCT
ejpam-642	186	42	where	where	SCONJ
ejpam-642	186	43	l1	l1	PROPN
ejpam-642	186	44	=	=	PROPN
ejpam-642	186	45	n	n	CCONJ
ejpam-642	186	46	∑	∑	ADP
ejpam-642	186	47	j=1	j=1	ADJ
ejpam-642	186	48	g(k)m	g(k)m	NOUN
ejpam-642	186	49	(	(	PUNCT
ejpam-642	186	50	γ	γ	X
ejpam-642	186	51	j	j	PROPN
ejpam-642	186	52	)	)	PUNCT
ejpam-642	186	53	log	log	PROPN
ejpam-642	186	54	fm	fm	PROPN
ejpam-642	186	55	(	(	PUNCT
ejpam-642	186	56	γ	γ	X
ejpam-642	186	57	j;βm	j;βm	PROPN
ejpam-642	186	58	)	)	PUNCT
ejpam-642	186	59	,	,	PUNCT
ejpam-642	186	60	l2	l2	NOUN
ejpam-642	186	61	=	=	SYM
ejpam-642	186	62	n	n	CCONJ
ejpam-642	186	63	∑	∑	PROPN
ejpam-642	186	64	j=1	j=1	PROPN
ejpam-642	186	65	g(k)u	g(k)u	PROPN
ejpam-642	186	66	(	(	PUNCT
ejpam-642	186	67	γ	γ	PROPN
ejpam-642	186	68	j	j	PROPN
ejpam-642	186	69	)	)	PUNCT
ejpam-642	186	70	log	log	PROPN
ejpam-642	186	71	fu	fu	NOUN
ejpam-642	186	72	(	(	PUNCT
ejpam-642	186	73	γ	γ	PROPN
ejpam-642	186	74	j;βu	j;βu	PROPN
ejpam-642	186	75	)	)	PUNCT
ejpam-642	186	76	,	,	PUNCT
ejpam-642	186	77	and	and	CCONJ
ejpam-642	186	78	l3	l3	PROPN
ejpam-642	186	79	=	=	SYM
ejpam-642	186	80	n	n	PROPN
ejpam-642	186	81	∑	∑	PROPN
ejpam-642	186	82	j=1	j=1	PROPN
ejpam-642	186	83	�	�	PROPN
ejpam-642	186	84	g(k)m	g(k)m	PROPN
ejpam-642	186	85	(	(	PUNCT
ejpam-642	186	86	γ	γ	X
ejpam-642	186	87	j	j	PROPN
ejpam-642	186	88	)	)	PUNCT
ejpam-642	186	89	logπ+	logπ+	PART
ejpam-642	186	90	g(k)u	g(k)u	PROPN
ejpam-642	186	91	(	(	PUNCT
ejpam-642	186	92	γ	γ	X
ejpam-642	186	93	j	j	PROPN
ejpam-642	186	94	)	)	PUNCT
ejpam-642	186	95	log(1−π	log(1−π	PROPN
ejpam-642	186	96	)	)	PUNCT
ejpam-642	186	97	�	�	PROPN
ejpam-642	186	98	.	.	PUNCT
ejpam-642	187	1	this	this	PRON
ejpam-642	187	2	splits	split	VERB
ejpam-642	187	3	the	the	DET
ejpam-642	187	4	conditional	conditional	ADJ
ejpam-642	187	5	expectation	expectation	NOUN
ejpam-642	187	6	l	l	NOUN
ejpam-642	187	7	into	into	ADP
ejpam-642	187	8	three	three	NUM
ejpam-642	187	9	separate	separate	ADJ
ejpam-642	187	10	parts	part	NOUN
ejpam-642	187	11	corresponding	correspond	VERB
ejpam-642	187	12	to	to	ADP
ejpam-642	187	13	the	the	DET
ejpam-642	187	14	matched	match	VERB
ejpam-642	187	15	pairs	pair	NOUN
ejpam-642	187	16	,	,	PUNCT
ejpam-642	187	17	the	the	DET
ejpam-642	187	18	unmatched	unmatched	ADJ
ejpam-642	187	19	pairs	pair	NOUN
ejpam-642	187	20	and	and	CCONJ
ejpam-642	187	21	the	the	DET
ejpam-642	187	22	mixing	mix	VERB
ejpam-642	187	23	proportion	proportion	NOUN
ejpam-642	187	24	π	π	PROPN
ejpam-642	187	25	respectively	respectively	ADV
ejpam-642	187	26	.	.	PUNCT
ejpam-642	188	1	the	the	DET
ejpam-642	188	2	m	m	NOUN
ejpam-642	188	3	-	-	PUNCT
ejpam-642	188	4	step	step	NOUN
ejpam-642	188	5	maximizes	maximize	VERB
ejpam-642	188	6	the	the	DET
ejpam-642	188	7	conditional	conditional	ADJ
ejpam-642	188	8	expectation	expectation	NOUN
ejpam-642	188	9	l	l	NOUN
ejpam-642	188	10	,	,	PUNCT
ejpam-642	188	11	which	which	PRON
ejpam-642	188	12	is	be	AUX
ejpam-642	188	13	in	in	ADP
ejpam-642	188	14	turn	turn	NOUN
ejpam-642	188	15	equivalent	equivalent	ADJ
ejpam-642	188	16	to	to	ADP
ejpam-642	188	17	maximization	maximization	NOUN
ejpam-642	188	18	of	of	ADP
ejpam-642	188	19	l1	l1	PROPN
ejpam-642	188	20	,	,	PUNCT
ejpam-642	188	21	l2	l2	NOUN
ejpam-642	188	22	and	and	CCONJ
ejpam-642	188	23	l3	l3	NOUN
ejpam-642	188	24	separately	separately	ADV
ejpam-642	188	25	.	.	PUNCT
ejpam-642	189	1	note	note	VERB
ejpam-642	189	2	that	that	SCONJ
ejpam-642	189	3	l1	l1	PROPN
ejpam-642	189	4	and	and	CCONJ
ejpam-642	189	5	l2	l2	NOUN
ejpam-642	189	6	are	be	AUX
ejpam-642	189	7	weighted	weight	VERB
ejpam-642	189	8	log	log	NOUN
ejpam-642	189	9	-	-	PUNCT
ejpam-642	189	10	likelihoods	likelihood	NOUN
ejpam-642	189	11	for	for	ADP
ejpam-642	189	12	the	the	DET
ejpam-642	189	13	matched	match	VERB
ejpam-642	189	14	and	and	CCONJ
ejpam-642	189	15	unmatched	unmatched	ADJ
ejpam-642	189	16	pairs	pair	NOUN
ejpam-642	189	17	respectively	respectively	ADV
ejpam-642	189	18	.	.	PUNCT
ejpam-642	190	1	thus	thus	ADV
ejpam-642	190	2	,	,	PUNCT
ejpam-642	190	3	for	for	ADP
ejpam-642	190	4	models	model	NOUN
ejpam-642	190	5	which	which	PRON
ejpam-642	190	6	can	can	AUX
ejpam-642	190	7	be	be	AUX
ejpam-642	190	8	fitted	fit	VERB
ejpam-642	190	9	by	by	ADP
ejpam-642	190	10	weighted	weight	VERB
ejpam-642	190	11	mle	mle	PROPN
ejpam-642	190	12	using	use	VERB
ejpam-642	190	13	available	available	ADJ
ejpam-642	190	14	software	software	NOUN
ejpam-642	190	15	,	,	PUNCT
ejpam-642	190	16	their	their	PRON
ejpam-642	190	17	maximization	maximization	NOUN
ejpam-642	190	18	can	can	AUX
ejpam-642	190	19	be	be	AUX
ejpam-642	190	20	readily	readily	ADV
ejpam-642	190	21	obtained	obtain	VERB
ejpam-642	190	22	by	by	ADP
ejpam-642	190	23	that	that	DET
ejpam-642	190	24	software	software	NOUN
ejpam-642	190	25	.	.	PUNCT
ejpam-642	191	1	the	the	DET
ejpam-642	191	2	maximization	maximization	NOUN
ejpam-642	191	3	of	of	ADP
ejpam-642	191	4	l3	l3	PROPN
ejpam-642	191	5	is	be	AUX
ejpam-642	191	6	straightforward	straightforward	ADJ
ejpam-642	191	7	:	:	PUNCT
ejpam-642	191	8	∂	∂	NUM
ejpam-642	191	9	l3	l3	PROPN
ejpam-642	191	10	∂	∂	NOUN
ejpam-642	191	11	π	π	PROPN
ejpam-642	191	12	=	=	SYM
ejpam-642	191	13	n	n	PROPN
ejpam-642	191	14	∑	∑	ADP
ejpam-642	191	15	j=1	j=1	PROPN
ejpam-642	191	16	�	�	PROPN
ejpam-642	191	17	g(k)m	g(k)m	PROPN
ejpam-642	191	18	(	(	PUNCT
ejpam-642	191	19	γ	γ	X
ejpam-642	191	20	j	j	PROPN
ejpam-642	191	21	)	)	PUNCT
ejpam-642	191	22	1	1	NUM
ejpam-642	191	23	π	π	NOUN
ejpam-642	191	24	−	−	PROPN
ejpam-642	191	25	g(k)u	g(k)u	PROPN
ejpam-642	191	26	(	(	PUNCT
ejpam-642	191	27	γ	γ	X
ejpam-642	191	28	j	j	PROPN
ejpam-642	191	29	)	)	PUNCT
ejpam-642	191	30	1	1	NUM
ejpam-642	191	31	1−π	1−π	NUM
ejpam-642	191	32	�	�	X
ejpam-642	191	33	=	=	SYM
ejpam-642	191	34	0	0	NUM
ejpam-642	191	35	yields	yield	NOUN
ejpam-642	191	36	1	1	NUM
ejpam-642	191	37	π	π	PROPN
ejpam-642	191	38	n	n	CCONJ
ejpam-642	191	39	∑	∑	ADV
ejpam-642	191	40	j=1	j=1	ADJ
ejpam-642	191	41	g(k)m	g(k)m	NOUN
ejpam-642	191	42	(	(	PUNCT
ejpam-642	191	43	γ	γ	X
ejpam-642	191	44	j	j	PROPN
ejpam-642	191	45	)	)	PUNCT
ejpam-642	191	46	=	=	SYM
ejpam-642	191	47	1	1	NUM
ejpam-642	191	48	1−π	1−π	NUM
ejpam-642	191	49	n	n	CCONJ
ejpam-642	191	50	∑	∑	PROPN
ejpam-642	191	51	j=1	j=1	PROPN
ejpam-642	191	52	g(k)u	g(k)u	PROPN
ejpam-642	191	53	(	(	PUNCT
ejpam-642	191	54	γ	γ	PROPN
ejpam-642	191	55	j	j	PROPN
ejpam-642	191	56	)	)	PUNCT
ejpam-642	191	57	.	.	PUNCT
ejpam-642	192	1	thus	thus	ADV
ejpam-642	192	2	,	,	PUNCT
ejpam-642	192	3	π=	π=	PROPN
ejpam-642	192	4	∑n	∑n	PROPN
ejpam-642	192	5	j=1	j=1	ADJ
ejpam-642	192	6	g(k)m	g(k)m	PROPN
ejpam-642	192	7	(	(	PUNCT
ejpam-642	192	8	γ	γ	X
ejpam-642	192	9	j	j	PROPN
ejpam-642	192	10	)	)	PUNCT
ejpam-642	192	11	∑n	∑n	PROPN
ejpam-642	192	12	j=1	j=1	PROPN
ejpam-642	192	13	g	g	PROPN
ejpam-642	192	14	(	(	PUNCT
ejpam-642	192	15	k	k	NOUN
ejpam-642	192	16	)	)	PUNCT
ejpam-642	192	17	m	m	VERB
ejpam-642	192	18	(	(	PUNCT
ejpam-642	192	19	γ	γ	X
ejpam-642	192	20	j	j	PROPN
ejpam-642	192	21	)	)	PUNCT
ejpam-642	192	22	+	+	CCONJ
ejpam-642	192	23	∑n	∑n	PROPN
ejpam-642	192	24	j=1	j=1	PROPN
ejpam-642	192	25	g	g	PROPN
ejpam-642	192	26	(	(	PUNCT
ejpam-642	192	27	k	k	NOUN
ejpam-642	192	28	)	)	PUNCT
ejpam-642	192	29	u	u	NOUN
ejpam-642	192	30	(	(	PUNCT
ejpam-642	192	31	γ	γ	X
ejpam-642	192	32	j	j	PROPN
ejpam-642	192	33	)	)	PUNCT
ejpam-642	192	34	=	=	SYM
ejpam-642	192	35	1	1	NUM
ejpam-642	192	36	n	n	CCONJ
ejpam-642	192	37	n	n	NOUN
ejpam-642	192	38	∑	∑	ADP
ejpam-642	192	39	j=1	j=1	ADJ
ejpam-642	192	40	g(k)m	g(k)m	NOUN
ejpam-642	192	41	(	(	PUNCT
ejpam-642	192	42	γ	γ	X
ejpam-642	192	43	j	j	PROPN
ejpam-642	192	44	)	)	PUNCT
ejpam-642	192	45	.	.	PUNCT
ejpam-642	193	1	r.	r.	PROPN
ejpam-642	193	2	zhu	zhu	PROPN
ejpam-642	193	3	,	,	PUNCT
ejpam-642	193	4	j.	j.	PROPN
ejpam-642	193	5	zhang	zhang	PROPN
ejpam-642	193	6	,	,	PUNCT
ejpam-642	193	7	d.	d.	PROPN
ejpam-642	193	8	zhang	zhang	PROPN
ejpam-642	193	9	,	,	PUNCT
ejpam-642	193	10	g.	g.	PROPN
ejpam-642	193	11	yan	yan	PROPN
ejpam-642	193	12	/	/	SYM
ejpam-642	193	13	eur	eur	PROPN
ejpam-642	193	14	.	.	PUNCT
ejpam-642	194	1	j.	j.	PROPN
ejpam-642	194	2	pure	pure	PROPN
ejpam-642	194	3	appl	appl	PROPN
ejpam-642	194	4	.	.	PROPN
ejpam-642	194	5	math	math	PROPN
ejpam-642	194	6	,	,	PUNCT
ejpam-642	194	7	3	3	NUM
ejpam-642	194	8	(	(	PUNCT
ejpam-642	194	9	2010	2010	NUM
ejpam-642	194	10	)	)	PUNCT
ejpam-642	194	11	,	,	PUNCT
ejpam-642	194	12	141	141	NUM
ejpam-642	194	13	-	-	SYM
ejpam-642	194	14	162	162	NUM
ejpam-642	194	15	148	148	NUM
ejpam-642	194	16	we	we	PRON
ejpam-642	194	17	obtain	obtain	VERB
ejpam-642	194	18	π(k+1	π(k+1	NOUN
ejpam-642	194	19	)	)	PUNCT
ejpam-642	194	20	=	=	SYM
ejpam-642	194	21	1	1	NUM
ejpam-642	194	22	n	n	CCONJ
ejpam-642	194	23	n	n	NOUN
ejpam-642	194	24	∑	∑	ADP
ejpam-642	194	25	j=1	j=1	ADJ
ejpam-642	194	26	g(k)m	g(k)m	NOUN
ejpam-642	194	27	(	(	PUNCT
ejpam-642	194	28	γ	γ	X
ejpam-642	194	29	j	j	PROPN
ejpam-642	194	30	)	)	PUNCT
ejpam-642	194	31	,	,	PUNCT
ejpam-642	194	32	β	β	X
ejpam-642	194	33	(	(	PUNCT
ejpam-642	194	34	k+1	k+1	NOUN
ejpam-642	194	35	)	)	PUNCT
ejpam-642	194	36	m	m	PROPN
ejpam-642	194	37	=	=	NOUN
ejpam-642	194	38	arg	arg	NOUN
ejpam-642	194	39	max	max	PROPN
ejpam-642	194	40	l1	l1	PROPN
ejpam-642	194	41	,	,	PUNCT
ejpam-642	194	42	β	β	X
ejpam-642	194	43	(	(	PUNCT
ejpam-642	194	44	k+1	k+1	NOUN
ejpam-642	194	45	)	)	PUNCT
ejpam-642	194	46	u	u	NOUN
ejpam-642	194	47	=	=	NOUN
ejpam-642	194	48	arg	arg	NOUN
ejpam-642	194	49	max	max	PROPN
ejpam-642	194	50	l2	l2	NOUN
ejpam-642	194	51	.	.	PUNCT
ejpam-642	195	1	(	(	PUNCT
ejpam-642	195	2	5	5	X
ejpam-642	195	3	)	)	PUNCT
ejpam-642	195	4	this	this	DET
ejpam-642	195	5	approach	approach	NOUN
ejpam-642	195	6	is	be	AUX
ejpam-642	195	7	regarded	regard	VERB
ejpam-642	195	8	as	as	ADP
ejpam-642	195	9	a	a	DET
ejpam-642	195	10	simplified	simplified	ADJ
ejpam-642	195	11	method	method	NOUN
ejpam-642	195	12	of	of	ADP
ejpam-642	195	13	the	the	DET
ejpam-642	195	14	em	em	PROPN
ejpam-642	195	15	algorithm	algorithm	NOUN
ejpam-642	195	16	for	for	ADP
ejpam-642	195	17	the	the	DET
ejpam-642	195	18	mixture	mixture	NOUN
ejpam-642	195	19	situation	situation	NOUN
ejpam-642	195	20	.	.	PUNCT
ejpam-642	196	1	it	it	PRON
ejpam-642	196	2	can	can	AUX
ejpam-642	196	3	take	take	VERB
ejpam-642	196	4	advantage	advantage	NOUN
ejpam-642	196	5	of	of	ADP
ejpam-642	196	6	existing	exist	VERB
ejpam-642	196	7	packages	package	NOUN
ejpam-642	196	8	or	or	CCONJ
ejpam-642	196	9	functions	function	NOUN
ejpam-642	196	10	in	in	ADP
ejpam-642	196	11	statistical	statistical	ADJ
ejpam-642	196	12	software	software	NOUN
ejpam-642	196	13	without	without	ADP
ejpam-642	196	14	developing	develop	VERB
ejpam-642	196	15	particular	particular	ADJ
ejpam-642	196	16	code	code	NOUN
ejpam-642	196	17	for	for	ADP
ejpam-642	196	18	maximization	maximization	NOUN
ejpam-642	196	19	of	of	ADP
ejpam-642	196	20	the	the	DET
ejpam-642	196	21	complicated	complicated	ADJ
ejpam-642	196	22	model	model	NOUN
ejpam-642	196	23	components	component	NOUN
ejpam-642	196	24	,	,	PUNCT
ejpam-642	196	25	especially	especially	ADV
ejpam-642	196	26	with	with	ADP
ejpam-642	196	27	interactions	interaction	NOUN
ejpam-642	196	28	.	.	PUNCT
ejpam-642	197	1	note	note	VERB
ejpam-642	197	2	that	that	SCONJ
ejpam-642	197	3	winkler	winkler	NOUN
ejpam-642	197	4	(	(	PUNCT
ejpam-642	197	5	1988	1988	NUM
ejpam-642	197	6	)	)	PUNCT
ejpam-642	197	7	investigated	investigate	VERB
ejpam-642	197	8	the	the	DET
ejpam-642	197	9	em	em	PROPN
ejpam-642	197	10	algorithm	algorithm	NOUN
ejpam-642	197	11	for	for	ADP
ejpam-642	197	12	fellegi	fellegi	NOUN
ejpam-642	197	13	-	-	PUNCT
ejpam-642	197	14	sunter	sunter	NOUN
ejpam-642	197	15	model	model	NOUN
ejpam-642	197	16	;	;	PUNCT
ejpam-642	197	17	see	see	VERB
ejpam-642	197	18	[	[	X
ejpam-642	197	19	15	15	NUM
ejpam-642	197	20	]	]	PUNCT
ejpam-642	197	21	.	.	PUNCT
ejpam-642	198	1	we	we	PRON
ejpam-642	198	2	are	be	AUX
ejpam-642	198	3	suggestting	suggestte	VERB
ejpam-642	198	4	a	a	DET
ejpam-642	198	5	computational	computational	ADJ
ejpam-642	198	6	method	method	NOUN
ejpam-642	198	7	for	for	ADP
ejpam-642	198	8	fitting	fit	VERB
ejpam-642	198	9	a	a	DET
ejpam-642	198	10	mixture	mixture	NOUN
ejpam-642	198	11	with	with	ADP
ejpam-642	198	12	two	two	NUM
ejpam-642	198	13	components	component	NOUN
ejpam-642	198	14	where	where	SCONJ
ejpam-642	198	15	each	each	DET
ejpam-642	198	16	component	component	NOUN
ejpam-642	198	17	is	be	AUX
ejpam-642	198	18	readily	readily	ADV
ejpam-642	198	19	fitted	fit	VERB
ejpam-642	198	20	by	by	ADP
ejpam-642	198	21	a	a	DET
ejpam-642	198	22	model	model	NOUN
ejpam-642	198	23	like	like	ADP
ejpam-642	198	24	glm	glm	PROPN
ejpam-642	198	25	.	.	PUNCT
ejpam-642	199	1	3	3	X
ejpam-642	199	2	.	.	X
ejpam-642	199	3	model	model	NOUN
ejpam-642	199	4	building	building	NOUN
ejpam-642	199	5	strategy	strategy	NOUN
ejpam-642	199	6	for	for	ADP
ejpam-642	199	7	loglinear	loglinear	NOUN
ejpam-642	199	8	mixtures	mixture	NOUN
ejpam-642	199	9	of	of	ADP
ejpam-642	199	10	two	two	NUM
ejpam-642	199	11	components	component	NOUN
ejpam-642	199	12	we	we	PRON
ejpam-642	199	13	consider	consider	VERB
ejpam-642	199	14	all	all	DET
ejpam-642	199	15	fields	field	NOUN
ejpam-642	199	16	of	of	ADP
ejpam-642	199	17	the	the	DET
ejpam-642	199	18	comparison	comparison	NOUN
ejpam-642	199	19	vector	vector	NOUN
ejpam-642	199	20	γ	γ	X
ejpam-642	199	21	=	=	SYM
ejpam-642	199	22	(	(	PUNCT
ejpam-642	199	23	γ1	γ1	PROPN
ejpam-642	199	24	,	,	PUNCT
ejpam-642	199	25	.	.	PUNCT
ejpam-642	199	26	.	.	PUNCT
ejpam-642	199	27	.	.	PUNCT
ejpam-642	200	1	,	,	PUNCT
ejpam-642	200	2	γn	γn	X
ejpam-642	200	3	)	)	PUNCT
ejpam-642	200	4	are	be	AUX
ejpam-642	200	5	categorical	categorical	ADJ
ejpam-642	200	6	.	.	PUNCT
ejpam-642	201	1	for	for	ADP
ejpam-642	201	2	example	example	NOUN
ejpam-642	201	3	,	,	PUNCT
ejpam-642	201	4	γ1	γ1	PROPN
ejpam-642	201	5	could	could	AUX
ejpam-642	201	6	be	be	AUX
ejpam-642	201	7	binary	binary	ADJ
ejpam-642	201	8	taking	taking	NOUN
ejpam-642	201	9	value	value	NOUN
ejpam-642	201	10	0	0	NUM
ejpam-642	201	11	or	or	CCONJ
ejpam-642	201	12	1	1	NUM
ejpam-642	201	13	,	,	PUNCT
ejpam-642	201	14	meaning	mean	VERB
ejpam-642	201	15	unmatching	unmatching	NOUN
ejpam-642	201	16	or	or	CCONJ
ejpam-642	201	17	matching	match	VERB
ejpam-642	201	18	respectively	respectively	ADV
ejpam-642	201	19	for	for	ADP
ejpam-642	201	20	a	a	DET
ejpam-642	201	21	pair	pair	NOUN
ejpam-642	201	22	of	of	ADP
ejpam-642	201	23	records	record	NOUN
ejpam-642	201	24	in	in	ADP
ejpam-642	201	25	field	field	NOUN
ejpam-642	201	26	1	1	NUM
ejpam-642	201	27	.	.	PUNCT
ejpam-642	202	1	then	then	ADV
ejpam-642	202	2	all	all	PRON
ejpam-642	202	3	of	of	ADP
ejpam-642	202	4	such	such	ADJ
ejpam-642	202	5	observed	observe	VERB
ejpam-642	202	6	comparison	comparison	NOUN
ejpam-642	202	7	vectors	vector	NOUN
ejpam-642	202	8	form	form	VERB
ejpam-642	202	9	a	a	DET
ejpam-642	202	10	multi	multi	ADJ
ejpam-642	202	11	-	-	ADJ
ejpam-642	202	12	way	way	ADJ
ejpam-642	202	13	contingency	contingency	NOUN
ejpam-642	202	14	table	table	NOUN
ejpam-642	202	15	with	with	ADP
ejpam-642	202	16	n	n	ADP
ejpam-642	202	17	factors	factor	NOUN
ejpam-642	202	18	(	(	PUNCT
ejpam-642	202	19	each	each	DET
ejpam-642	202	20	field	field	NOUN
ejpam-642	202	21	corresponds	correspond	VERB
ejpam-642	202	22	to	to	ADP
ejpam-642	202	23	a	a	DET
ejpam-642	202	24	factor	factor	NOUN
ejpam-642	202	25	)	)	PUNCT
ejpam-642	202	26	.	.	PUNCT
ejpam-642	203	1	thus	thus	ADV
ejpam-642	203	2	,	,	PUNCT
ejpam-642	203	3	it	it	PRON
ejpam-642	203	4	is	be	AUX
ejpam-642	203	5	natural	natural	ADJ
ejpam-642	203	6	to	to	PART
ejpam-642	203	7	utilize	utilize	VERB
ejpam-642	203	8	the	the	DET
ejpam-642	203	9	loglinear	loglinear	NOUN
ejpam-642	203	10	model	model	NOUN
ejpam-642	203	11	for	for	ADP
ejpam-642	203	12	analyzing	analyze	VERB
ejpam-642	203	13	such	such	DET
ejpam-642	203	14	a	a	DET
ejpam-642	203	15	contingency	contingency	NOUN
ejpam-642	203	16	table	table	NOUN
ejpam-642	203	17	.	.	PUNCT
ejpam-642	204	1	a	a	DET
ejpam-642	204	2	loglinear	loglinear	NOUN
ejpam-642	204	3	model	model	NOUN
ejpam-642	204	4	states	state	VERB
ejpam-642	204	5	that	that	SCONJ
ejpam-642	204	6	the	the	DET
ejpam-642	204	7	logarithm	logarithm	NOUN
ejpam-642	204	8	of	of	ADP
ejpam-642	204	9	the	the	DET
ejpam-642	204	10	expected	expect	VERB
ejpam-642	204	11	number	number	NOUN
ejpam-642	204	12	of	of	ADP
ejpam-642	204	13	a	a	DET
ejpam-642	204	14	cell	cell	NOUN
ejpam-642	204	15	in	in	ADP
ejpam-642	204	16	the	the	DET
ejpam-642	204	17	contingency	contingency	NOUN
ejpam-642	204	18	table	table	NOUN
ejpam-642	204	19	can	can	AUX
ejpam-642	204	20	be	be	AUX
ejpam-642	204	21	expressed	express	VERB
ejpam-642	204	22	as	as	ADP
ejpam-642	204	23	the	the	DET
ejpam-642	204	24	additive	additive	ADJ
ejpam-642	204	25	function	function	NOUN
ejpam-642	204	26	of	of	ADP
ejpam-642	204	27	main	main	ADJ
ejpam-642	204	28	effects	effect	NOUN
ejpam-642	204	29	and	and	CCONJ
ejpam-642	204	30	interactions	interaction	NOUN
ejpam-642	204	31	of	of	ADP
ejpam-642	204	32	factors	factor	NOUN
ejpam-642	204	33	.	.	PUNCT
ejpam-642	205	1	for	for	ADP
ejpam-642	205	2	instance	instance	NOUN
ejpam-642	205	3	,	,	PUNCT
ejpam-642	205	4	we	we	PRON
ejpam-642	205	5	consider	consider	VERB
ejpam-642	205	6	a	a	DET
ejpam-642	205	7	four	four	NUM
ejpam-642	205	8	factor	factor	NOUN
ejpam-642	206	1	i	i	PRON
ejpam-642	206	2	×	×	VERB
ejpam-642	206	3	j	j	PROPN
ejpam-642	206	4	×k×	×k×	X
ejpam-642	206	5	l	l	NOUN
ejpam-642	206	6	contingency	contingency	NOUN
ejpam-642	206	7	table	table	NOUN
ejpam-642	206	8	.	.	PUNCT
ejpam-642	207	1	two	two	NUM
ejpam-642	207	2	loglinear	loglinear	NOUN
ejpam-642	207	3	models	model	NOUN
ejpam-642	207	4	we	we	PRON
ejpam-642	207	5	will	will	AUX
ejpam-642	207	6	use	use	VERB
ejpam-642	207	7	in	in	ADP
ejpam-642	207	8	the	the	DET
ejpam-642	207	9	remaining	remain	VERB
ejpam-642	207	10	are	be	AUX
ejpam-642	207	11	logµi	logµi	NOUN
ejpam-642	207	12	jkl	jkl	NOUN
ejpam-642	207	13	=	=	PUNCT
ejpam-642	207	14	λ+λ	λ+λ	X
ejpam-642	207	15	(	(	PUNCT
ejpam-642	207	16	1	1	NUM
ejpam-642	207	17	)	)	PUNCT
ejpam-642	207	18	i	i	PRON
ejpam-642	208	1	+	+	PROPN
ejpam-642	208	2	λ	λ	X
ejpam-642	208	3	(	(	PUNCT
ejpam-642	208	4	2	2	NUM
ejpam-642	208	5	)	)	PUNCT
ejpam-642	208	6	j	j	NOUN
ejpam-642	209	1	+	+	PROPN
ejpam-642	209	2	λ	λ	X
ejpam-642	209	3	(	(	PUNCT
ejpam-642	209	4	3	3	NUM
ejpam-642	209	5	)	)	PUNCT
ejpam-642	209	6	k	k	NOUN
ejpam-642	210	1	+	+	PROPN
ejpam-642	210	2	λ	λ	X
ejpam-642	210	3	(	(	PUNCT
ejpam-642	210	4	4	4	NUM
ejpam-642	210	5	)	)	PUNCT
ejpam-642	210	6	l	l	NOUN
ejpam-642	210	7	,	,	PUNCT
ejpam-642	210	8	(	(	PUNCT
ejpam-642	210	9	6	6	NUM
ejpam-642	210	10	)	)	PUNCT
ejpam-642	210	11	or	or	CCONJ
ejpam-642	210	12	logµi	logµi	NOUN
ejpam-642	210	13	jkl	jkl	VERB
ejpam-642	210	14	=	=	PUNCT
ejpam-642	210	15	λ+λ	λ+λ	X
ejpam-642	210	16	(	(	PUNCT
ejpam-642	210	17	1	1	NUM
ejpam-642	210	18	)	)	PUNCT
ejpam-642	210	19	i	i	PRON
ejpam-642	211	1	+	+	PROPN
ejpam-642	211	2	λ	λ	X
ejpam-642	211	3	(	(	PUNCT
ejpam-642	211	4	2	2	NUM
ejpam-642	211	5	)	)	PUNCT
ejpam-642	211	6	j	j	NOUN
ejpam-642	212	1	+	+	PROPN
ejpam-642	212	2	λ	λ	X
ejpam-642	212	3	(	(	PUNCT
ejpam-642	212	4	3	3	NUM
ejpam-642	212	5	)	)	PUNCT
ejpam-642	212	6	k	k	NOUN
ejpam-642	213	1	+	+	PROPN
ejpam-642	213	2	λ	λ	X
ejpam-642	213	3	(	(	PUNCT
ejpam-642	213	4	4	4	NUM
ejpam-642	213	5	)	)	PUNCT
ejpam-642	213	6	l	l	NOUN
ejpam-642	214	1	+	+	PUNCT
ejpam-642	214	2	λ	λ	X
ejpam-642	214	3	(	(	PUNCT
ejpam-642	214	4	12	12	NUM
ejpam-642	214	5	)	)	PUNCT
ejpam-642	214	6	i	i	PRON
ejpam-642	214	7	j	j	X
ejpam-642	215	1	+	+	NOUN
ejpam-642	215	2	λ	λ	X
ejpam-642	215	3	(	(	PUNCT
ejpam-642	215	4	13	13	NUM
ejpam-642	215	5	)	)	PUNCT
ejpam-642	215	6	ik	ik	PROPN
ejpam-642	216	1	+	+	PROPN
ejpam-642	216	2	λ	λ	X
ejpam-642	216	3	(	(	PUNCT
ejpam-642	216	4	14	14	NUM
ejpam-642	216	5	)	)	PUNCT
ejpam-642	216	6	il	il	PROPN
ejpam-642	217	1	+	+	PROPN
ejpam-642	217	2	λ	λ	X
ejpam-642	217	3	(	(	PUNCT
ejpam-642	217	4	23	23	NUM
ejpam-642	217	5	)	)	PUNCT
ejpam-642	217	6	jk	jk	NOUN
ejpam-642	218	1	+	+	PROPN
ejpam-642	218	2	λ	λ	X
ejpam-642	218	3	(	(	PUNCT
ejpam-642	218	4	24	24	NUM
ejpam-642	218	5	)	)	PUNCT
ejpam-642	218	6	jl	jl	NOUN
ejpam-642	218	7	+	+	PROPN
ejpam-642	218	8	λ	λ	X
ejpam-642	218	9	(	(	PUNCT
ejpam-642	218	10	34	34	NUM
ejpam-642	218	11	)	)	PUNCT
ejpam-642	218	12	kl	kl	NOUN
ejpam-642	218	13	,	,	PUNCT
ejpam-642	218	14	(	(	PUNCT
ejpam-642	218	15	7	7	X
ejpam-642	218	16	)	)	PUNCT
ejpam-642	218	17	i	i	NOUN
ejpam-642	218	18	=	=	NOUN
ejpam-642	218	19	1	1	NUM
ejpam-642	218	20	,	,	PUNCT
ejpam-642	218	21	.	.	PUNCT
ejpam-642	218	22	.	.	PUNCT
ejpam-642	219	1	.	.	PUNCT
ejpam-642	220	1	,	,	PUNCT
ejpam-642	220	2	i	i	PRON
ejpam-642	220	3	;	;	PUNCT
ejpam-642	220	4	j	j	PROPN
ejpam-642	220	5	=	=	SYM
ejpam-642	220	6	1	1	NUM
ejpam-642	220	7	,	,	PUNCT
ejpam-642	220	8	.	.	PUNCT
ejpam-642	220	9	.	.	PUNCT
ejpam-642	220	10	.	.	PUNCT
ejpam-642	221	1	,	,	PUNCT
ejpam-642	221	2	j	j	PROPN
ejpam-642	221	3	;	;	PUNCT
ejpam-642	221	4	k	k	PROPN
ejpam-642	221	5	=	=	SYM
ejpam-642	221	6	1	1	NUM
ejpam-642	221	7	,	,	PUNCT
ejpam-642	221	8	.	.	PUNCT
ejpam-642	221	9	.	.	PUNCT
ejpam-642	221	10	.	.	PUNCT
ejpam-642	222	1	,	,	PUNCT
ejpam-642	222	2	k	k	NOUN
ejpam-642	222	3	,	,	PUNCT
ejpam-642	222	4	l	l	NOUN
ejpam-642	222	5	=	=	SYM
ejpam-642	222	6	1	1	NUM
ejpam-642	222	7	,	,	PUNCT
ejpam-642	222	8	.	.	PUNCT
ejpam-642	222	9	.	.	PUNCT
ejpam-642	222	10	.	.	PUNCT
ejpam-642	223	1	,	,	PUNCT
ejpam-642	223	2	l.	l.	PROPN
ejpam-642	223	3	here	here	ADV
ejpam-642	223	4	µi	µi	PROPN
ejpam-642	223	5	jkl	jkl	PROPN
ejpam-642	223	6	denotes	denote	VERB
ejpam-642	223	7	the	the	DET
ejpam-642	223	8	expected	expect	VERB
ejpam-642	223	9	cell	cell	NOUN
ejpam-642	223	10	count	count	NOUN
ejpam-642	223	11	,	,	PUNCT
ejpam-642	223	12	λ	λ	PROPN
ejpam-642	223	13	is	be	AUX
ejpam-642	223	14	the	the	DET
ejpam-642	223	15	overall	overall	ADJ
ejpam-642	223	16	mean	mean	NOUN
ejpam-642	223	17	,	,	PUNCT
ejpam-642	223	18	λ	λ	X
ejpam-642	223	19	(	(	PUNCT
ejpam-642	223	20	1	1	NUM
ejpam-642	223	21	)	)	PUNCT
ejpam-642	223	22	i	i	PRON
ejpam-642	223	23	,	,	PUNCT
ejpam-642	223	24	λ	λ	X
ejpam-642	223	25	(	(	PUNCT
ejpam-642	223	26	2	2	NUM
ejpam-642	223	27	)	)	PUNCT
ejpam-642	223	28	j	j	NOUN
ejpam-642	223	29	,	,	PUNCT
ejpam-642	223	30	λ	λ	X
ejpam-642	223	31	(	(	PUNCT
ejpam-642	223	32	3	3	NUM
ejpam-642	223	33	)	)	PUNCT
ejpam-642	223	34	k	k	NOUN
ejpam-642	223	35	and	and	CCONJ
ejpam-642	223	36	λ	λ	X
ejpam-642	223	37	(	(	PUNCT
ejpam-642	223	38	4	4	NUM
ejpam-642	223	39	)	)	PUNCT
ejpam-642	223	40	l	l	NOUN
ejpam-642	223	41	are	be	AUX
ejpam-642	223	42	the	the	DET
ejpam-642	223	43	main	main	ADJ
ejpam-642	223	44	effects	effect	NOUN
ejpam-642	223	45	of	of	ADP
ejpam-642	223	46	each	each	DET
ejpam-642	223	47	factor	factor	NOUN
ejpam-642	223	48	at	at	ADP
ejpam-642	223	49	corresponding	corresponding	ADJ
ejpam-642	223	50	levels	level	NOUN
ejpam-642	223	51	i	i	PRON
ejpam-642	223	52	,	,	PUNCT
ejpam-642	223	53	j	j	PROPN
ejpam-642	223	54	,	,	PUNCT
ejpam-642	223	55	k	k	PROPN
ejpam-642	223	56	and	and	CCONJ
ejpam-642	223	57	l	l	NOUN
ejpam-642	223	58	,	,	PUNCT
ejpam-642	223	59	λ	λ	X
ejpam-642	223	60	(	(	PUNCT
ejpam-642	223	61	12	12	NUM
ejpam-642	223	62	)	)	PUNCT
ejpam-642	224	1	i	i	PRON
ejpam-642	225	1	j	j	PROPN
ejpam-642	225	2	,	,	PUNCT
ejpam-642	225	3	λ	λ	X
ejpam-642	225	4	(	(	PUNCT
ejpam-642	225	5	13	13	NUM
ejpam-642	225	6	)	)	PUNCT
ejpam-642	225	7	ik	ik	PROPN
ejpam-642	225	8	,	,	PUNCT
ejpam-642	225	9	λ	λ	X
ejpam-642	225	10	(	(	PUNCT
ejpam-642	225	11	14	14	NUM
ejpam-642	225	12	)	)	PUNCT
ejpam-642	225	13	il	il	PROPN
ejpam-642	225	14	,	,	PUNCT
ejpam-642	225	15	λ	λ	X
ejpam-642	225	16	(	(	PUNCT
ejpam-642	225	17	23	23	NUM
ejpam-642	225	18	)	)	PUNCT
ejpam-642	225	19	jk	jk	PROPN
ejpam-642	225	20	,	,	PUNCT
ejpam-642	225	21	λ	λ	X
ejpam-642	225	22	(	(	PUNCT
ejpam-642	225	23	24	24	NUM
ejpam-642	225	24	)	)	PUNCT
ejpam-642	225	25	jl	jl	NOUN
ejpam-642	225	26	,	,	PUNCT
ejpam-642	225	27	λ	λ	X
ejpam-642	225	28	(	(	PUNCT
ejpam-642	225	29	34	34	NUM
ejpam-642	225	30	)	)	PUNCT
ejpam-642	225	31	kl	kl	NOUN
ejpam-642	225	32	are	be	AUX
ejpam-642	225	33	two	two	NUM
ejpam-642	225	34	-	-	PUNCT
ejpam-642	225	35	factor	factor	NOUN
ejpam-642	225	36	interactions	interaction	NOUN
ejpam-642	225	37	at	at	ADP
ejpam-642	225	38	specified	specified	ADJ
ejpam-642	225	39	level	level	NOUN
ejpam-642	225	40	pairs	pair	NOUN
ejpam-642	225	41	.	.	PUNCT
ejpam-642	226	1	model	model	NOUN
ejpam-642	226	2	(	(	PUNCT
ejpam-642	226	3	6	6	NUM
ejpam-642	226	4	)	)	PUNCT
ejpam-642	226	5	corresponds	correspond	VERB
ejpam-642	226	6	to	to	ADP
ejpam-642	226	7	the	the	DET
ejpam-642	226	8	situation	situation	NOUN
ejpam-642	226	9	of	of	ADP
ejpam-642	226	10	independence	independence	NOUN
ejpam-642	226	11	among	among	ADP
ejpam-642	226	12	factors	factor	NOUN
ejpam-642	226	13	,	,	PUNCT
ejpam-642	226	14	while	while	SCONJ
ejpam-642	226	15	model	model	NOUN
ejpam-642	226	16	(	(	PUNCT
ejpam-642	226	17	7	7	X
ejpam-642	226	18	)	)	PUNCT
ejpam-642	226	19	corresponds	correspond	VERB
ejpam-642	226	20	to	to	ADP
ejpam-642	226	21	a	a	DET
ejpam-642	226	22	situation	situation	NOUN
ejpam-642	226	23	of	of	ADP
ejpam-642	226	24	dependence	dependence	NOUN
ejpam-642	226	25	among	among	ADP
ejpam-642	226	26	factors	factor	NOUN
ejpam-642	226	27	.	.	PUNCT
ejpam-642	227	1	note	note	VERB
ejpam-642	227	2	that	that	DET
ejpam-642	227	3	model	model	NOUN
ejpam-642	227	4	(	(	PUNCT
ejpam-642	227	5	6	6	NUM
ejpam-642	227	6	)	)	PUNCT
ejpam-642	227	7	is	be	AUX
ejpam-642	227	8	nested	nest	VERB
ejpam-642	227	9	in	in	ADP
ejpam-642	227	10	model	model	NOUN
ejpam-642	227	11	(	(	PUNCT
ejpam-642	227	12	7	7	NUM
ejpam-642	227	13	)	)	PUNCT
ejpam-642	227	14	.	.	PUNCT
ejpam-642	228	1	apart	apart	ADV
ejpam-642	228	2	from	from	ADP
ejpam-642	228	3	the	the	DET
ejpam-642	228	4	above	above	ADJ
ejpam-642	228	5	common	common	ADJ
ejpam-642	228	6	anova	anova	PROPN
ejpam-642	228	7	-	-	ADJ
ejpam-642	228	8	like	like	ADJ
ejpam-642	228	9	formulation	formulation	NOUN
ejpam-642	228	10	,	,	PUNCT
ejpam-642	228	11	there	there	PRON
ejpam-642	228	12	is	be	VERB
ejpam-642	228	13	a	a	DET
ejpam-642	228	14	generalized	generalized	ADJ
ejpam-642	228	15	linear	linear	ADJ
ejpam-642	228	16	model	model	NOUN
ejpam-642	228	17	formulation	formulation	NOUN
ejpam-642	228	18	,	,	PUNCT
ejpam-642	228	19	where	where	SCONJ
ejpam-642	228	20	the	the	DET
ejpam-642	228	21	cell	cell	NOUN
ejpam-642	228	22	counts	count	NOUN
ejpam-642	228	23	are	be	AUX
ejpam-642	228	24	assumed	assume	VERB
ejpam-642	228	25	to	to	PART
ejpam-642	228	26	be	be	AUX
ejpam-642	228	27	distributed	distribute	VERB
ejpam-642	228	28	in	in	ADP
ejpam-642	228	29	poisson	poisson	NOUN
ejpam-642	228	30	.	.	PUNCT
ejpam-642	229	1	for	for	ADP
ejpam-642	229	2	instance	instance	NOUN
ejpam-642	229	3	,	,	PUNCT
ejpam-642	229	4	model	model	NOUN
ejpam-642	229	5	(	(	PUNCT
ejpam-642	229	6	6	6	NUM
ejpam-642	229	7	)	)	PUNCT
ejpam-642	229	8	and	and	CCONJ
ejpam-642	229	9	(	(	PUNCT
ejpam-642	229	10	7	7	X
ejpam-642	229	11	)	)	PUNCT
ejpam-642	229	12	have	have	VERB
ejpam-642	229	13	the	the	DET
ejpam-642	229	14	following	follow	VERB
ejpam-642	229	15	equivalent	equivalent	ADJ
ejpam-642	229	16	poisson	poisson	PROPN
ejpam-642	229	17	glm	glm	PROPN
ejpam-642	229	18	formulations	formulation	NOUN
ejpam-642	229	19	:	:	PUNCT
ejpam-642	229	20	logµ	logµ	NOUN
ejpam-642	229	21	=	=	SYM
ejpam-642	230	1	α+α1γ1+α2γ2	α+α1γ1+α2γ2	NOUN
ejpam-642	230	2	+	+	NOUN
ejpam-642	230	3	α3γ3	α3γ3	PROPN
ejpam-642	230	4	+	+	ADJ
ejpam-642	230	5	α4γ4	α4γ4	NUM
ejpam-642	230	6	,	,	PUNCT
ejpam-642	230	7	(	(	PUNCT
ejpam-642	230	8	8)	8)	NUM
ejpam-642	230	9	and	and	CCONJ
ejpam-642	230	10	logµ	logµ	NOUN
ejpam-642	230	11	=	=	SYM
ejpam-642	230	12	α+α1γ1	α+α1γ1	NUM
ejpam-642	230	13	+	+	NOUN
ejpam-642	230	14	α2γ2	α2γ2	NUM
ejpam-642	230	15	+	+	ADJ
ejpam-642	230	16	α3γ3	α3γ3	PROPN
ejpam-642	230	17	+	+	ADJ
ejpam-642	230	18	α4γ4	α4γ4	PROPN
ejpam-642	230	19	r.	r.	PROPN
ejpam-642	230	20	zhu	zhu	PROPN
ejpam-642	230	21	,	,	PUNCT
ejpam-642	230	22	j.	j.	PROPN
ejpam-642	230	23	zhang	zhang	PROPN
ejpam-642	230	24	,	,	PUNCT
ejpam-642	230	25	d.	d.	PROPN
ejpam-642	230	26	zhang	zhang	PROPN
ejpam-642	230	27	,	,	PUNCT
ejpam-642	230	28	g.	g.	PROPN
ejpam-642	230	29	yan	yan	PROPN
ejpam-642	230	30	/	/	SYM
ejpam-642	230	31	eur	eur	PROPN
ejpam-642	230	32	.	.	PUNCT
ejpam-642	231	1	j.	j.	PROPN
ejpam-642	231	2	pure	pure	PROPN
ejpam-642	231	3	appl	appl	PROPN
ejpam-642	231	4	.	.	PROPN
ejpam-642	231	5	math	math	PROPN
ejpam-642	231	6	,	,	PUNCT
ejpam-642	231	7	3	3	NUM
ejpam-642	231	8	(	(	PUNCT
ejpam-642	231	9	2010	2010	NUM
ejpam-642	231	10	)	)	PUNCT
ejpam-642	231	11	,	,	PUNCT
ejpam-642	231	12	141	141	NUM
ejpam-642	231	13	-	-	SYM
ejpam-642	231	14	162	162	NUM
ejpam-642	231	15	149	149	NUM
ejpam-642	232	1	+	+	ADV
ejpam-642	232	2	α12γ1γ2	α12γ1γ2	ADJ
ejpam-642	232	3	+	+	ADJ
ejpam-642	232	4	α13γ1γ3	α13γ1γ3	PROPN
ejpam-642	232	5	+	+	NOUN
ejpam-642	232	6	α14γ1γ4	α14γ1γ4	PROPN
ejpam-642	232	7	+	+	PROPN
ejpam-642	232	8	α23γ2γ3	α23γ2γ3	PROPN
ejpam-642	232	9	+	+	PROPN
ejpam-642	232	10	α24γ2γ4	α24γ2γ4	PROPN
ejpam-642	232	11	+	+	PROPN
ejpam-642	232	12	α34γ3γ4	α34γ3γ4	PROPN
ejpam-642	232	13	.	.	PUNCT
ejpam-642	233	1	(	(	PUNCT
ejpam-642	233	2	9	9	NUM
ejpam-642	233	3	)	)	PUNCT
ejpam-642	233	4	here	here	ADV
ejpam-642	233	5	α	α	PRON
ejpam-642	233	6	’s	’	VERB
ejpam-642	233	7	are	be	AUX
ejpam-642	233	8	the	the	DET
ejpam-642	233	9	intercept	intercept	NOUN
ejpam-642	233	10	for	for	ADP
ejpam-642	233	11	the	the	DET
ejpam-642	233	12	model	model	NOUN
ejpam-642	233	13	and	and	CCONJ
ejpam-642	233	14	coefficients	coefficient	NOUN
ejpam-642	233	15	for	for	ADP
ejpam-642	233	16	variable	variable	ADJ
ejpam-642	233	17	γ	γ	PROPN
ejpam-642	233	18	’s	’s	X
ejpam-642	233	19	and	and	CCONJ
ejpam-642	233	20	their	their	PRON
ejpam-642	233	21	interactions	interaction	NOUN
ejpam-642	233	22	,	,	PUNCT
ejpam-642	233	23	µ	µ	X
ejpam-642	233	24	is	be	AUX
ejpam-642	233	25	the	the	DET
ejpam-642	233	26	expected	expect	VERB
ejpam-642	233	27	cell	cell	NOUN
ejpam-642	233	28	count	count	NOUN
ejpam-642	233	29	associated	associate	VERB
ejpam-642	233	30	with	with	ADP
ejpam-642	233	31	covariates	covariates	PROPN
ejpam-642	233	32	γ=	γ=	PROPN
ejpam-642	233	33	(	(	PUNCT
ejpam-642	233	34	γ1,γ2,γ3,γ4	γ1,γ2,γ3,γ4	PROPN
ejpam-642	233	35	)	)	PUNCT
ejpam-642	233	36	.	.	PUNCT
ejpam-642	234	1	r	r	NOUN
ejpam-642	234	2	has	have	VERB
ejpam-642	234	3	function	function	NOUN
ejpam-642	234	4	loglm	loglm	NOUN
ejpam-642	234	5	(	(	PUNCT
ejpam-642	234	6	)	)	PUNCT
ejpam-642	234	7	for	for	ADP
ejpam-642	234	8	fitting	fitting	ADJ
ejpam-642	234	9	loglinear	loglinear	NOUN
ejpam-642	234	10	models	model	NOUN
ejpam-642	234	11	in	in	ADP
ejpam-642	234	12	the	the	DET
ejpam-642	234	13	form	form	NOUN
ejpam-642	234	14	of	of	ADP
ejpam-642	234	15	anova	anova	PROPN
ejpam-642	234	16	-	-	PUNCT
ejpam-642	234	17	like	like	ADJ
ejpam-642	234	18	specification	specification	NOUN
ejpam-642	234	19	,	,	PUNCT
ejpam-642	234	20	and	and	CCONJ
ejpam-642	234	21	function	function	VERB
ejpam-642	234	22	glm	glm	PROPN
ejpam-642	234	23	(	(	PUNCT
ejpam-642	234	24	)	)	PUNCT
ejpam-642	234	25	for	for	ADP
ejpam-642	234	26	fitting	fitting	ADJ
ejpam-642	234	27	poisson	poisson	NOUN
ejpam-642	234	28	glm	glm	PROPN
ejpam-642	234	29	models	model	NOUN
ejpam-642	234	30	.	.	PUNCT
ejpam-642	235	1	each	each	PRON
ejpam-642	235	2	has	have	VERB
ejpam-642	235	3	advantages	advantage	NOUN
ejpam-642	235	4	and	and	CCONJ
ejpam-642	235	5	disadvantages	disadvantage	NOUN
ejpam-642	235	6	.	.	PUNCT
ejpam-642	236	1	we	we	PRON
ejpam-642	236	2	use	use	VERB
ejpam-642	236	3	both	both	PRON
ejpam-642	236	4	in	in	ADP
ejpam-642	236	5	our	our	PRON
ejpam-642	236	6	computation	computation	NOUN
ejpam-642	236	7	in	in	ADP
ejpam-642	236	8	case	case	NOUN
ejpam-642	236	9	one	one	NUM
ejpam-642	236	10	encounters	encounter	VERB
ejpam-642	236	11	unexpected	unexpected	ADJ
ejpam-642	236	12	problems	problem	NOUN
ejpam-642	236	13	.	.	PUNCT
ejpam-642	237	1	in	in	ADP
ejpam-642	237	2	record	record	NOUN
ejpam-642	237	3	linkage	linkage	NOUN
ejpam-642	237	4	,	,	PUNCT
ejpam-642	237	5	the	the	DET
ejpam-642	237	6	matched	match	VERB
ejpam-642	237	7	and	and	CCONJ
ejpam-642	237	8	unmatched	unmatched	ADJ
ejpam-642	237	9	subpopulations	subpopulation	NOUN
ejpam-642	237	10	could	could	AUX
ejpam-642	237	11	have	have	VERB
ejpam-642	237	12	their	their	PRON
ejpam-642	237	13	own	own	ADJ
ejpam-642	237	14	model	model	NOUN
ejpam-642	237	15	specifications	specification	NOUN
ejpam-642	237	16	,	,	PUNCT
ejpam-642	237	17	thus	thus	ADV
ejpam-642	237	18	,	,	PUNCT
ejpam-642	237	19	m	m	VERB
ejpam-642	237	20	and	and	CCONJ
ejpam-642	237	21	u	u	PROPN
ejpam-642	237	22	are	be	AUX
ejpam-642	237	23	assumed	assume	VERB
ejpam-642	237	24	to	to	PART
ejpam-642	237	25	have	have	VERB
ejpam-642	237	26	different	different	ADJ
ejpam-642	237	27	loglinear	loglinear	NOUN
ejpam-642	237	28	models	model	NOUN
ejpam-642	237	29	respectively	respectively	ADV
ejpam-642	237	30	.	.	PUNCT
ejpam-642	238	1	however	however	ADV
ejpam-642	238	2	,	,	PUNCT
ejpam-642	238	3	the	the	DET
ejpam-642	238	4	membership	membership	NOUN
ejpam-642	238	5	is	be	AUX
ejpam-642	238	6	unknown	unknown	ADJ
ejpam-642	238	7	.	.	PUNCT
ejpam-642	239	1	so	so	ADV
ejpam-642	239	2	the	the	DET
ejpam-642	239	3	entire	entire	ADJ
ejpam-642	239	4	population	population	NOUN
ejpam-642	239	5	of	of	ADP
ejpam-642	239	6	possible	possible	ADJ
ejpam-642	239	7	pairs	pair	NOUN
ejpam-642	239	8	is	be	AUX
ejpam-642	239	9	a	a	DET
ejpam-642	239	10	mixture	mixture	NOUN
ejpam-642	239	11	of	of	ADP
ejpam-642	239	12	two	two	NUM
ejpam-642	239	13	loglinear	loglinear	NOUN
ejpam-642	239	14	models	model	NOUN
ejpam-642	239	15	for	for	ADP
ejpam-642	239	16	m	m	PROPN
ejpam-642	239	17	and	and	CCONJ
ejpam-642	239	18	u	u	NOUN
ejpam-642	239	19	.	.	PUNCT
ejpam-642	240	1	the	the	DET
ejpam-642	240	2	observed	observe	VERB
ejpam-642	240	3	contingency	contingency	NOUN
ejpam-642	240	4	table	table	NOUN
ejpam-642	240	5	is	be	AUX
ejpam-642	240	6	a	a	DET
ejpam-642	240	7	sample	sample	NOUN
ejpam-642	240	8	from	from	ADP
ejpam-642	240	9	this	this	DET
ejpam-642	240	10	mixture	mixture	NOUN
ejpam-642	240	11	population	population	NOUN
ejpam-642	240	12	.	.	PUNCT
ejpam-642	241	1	jaro	jaro	PROPN
ejpam-642	241	2	(	(	PUNCT
ejpam-642	241	3	1989	1989	NUM
ejpam-642	241	4	)	)	PUNCT
ejpam-642	241	5	considered	consider	VERB
ejpam-642	241	6	the	the	DET
ejpam-642	241	7	simplest	simple	ADJ
ejpam-642	241	8	case	case	NOUN
ejpam-642	241	9	for	for	ADP
ejpam-642	241	10	the	the	DET
ejpam-642	241	11	contingency	contingency	NOUN
ejpam-642	241	12	table	table	NOUN
ejpam-642	241	13	where	where	SCONJ
ejpam-642	241	14	all	all	DET
ejpam-642	241	15	factors	factor	NOUN
ejpam-642	241	16	are	be	AUX
ejpam-642	241	17	binary	binary	ADJ
ejpam-642	241	18	,	,	PUNCT
ejpam-642	241	19	and	and	CCONJ
ejpam-642	241	20	thus	thus	ADV
ejpam-642	241	21	,	,	PUNCT
ejpam-642	241	22	the	the	DET
ejpam-642	241	23	contingency	contingency	NOUN
ejpam-642	241	24	table	table	NOUN
ejpam-642	241	25	is	be	AUX
ejpam-642	241	26	2×	2×	NUM
ejpam-642	241	27	2×	2×	NUM
ejpam-642	241	28	·	·	PUNCT
ejpam-642	241	29	·	·	PUNCT
ejpam-642	241	30	·	·	PUNCT
ejpam-642	242	1	×	×	NOUN
ejpam-642	242	2	2	2	NUM
ejpam-642	242	3	,	,	PUNCT
ejpam-642	242	4	and	and	CCONJ
ejpam-642	242	5	the	the	DET
ejpam-642	242	6	number	number	NOUN
ejpam-642	242	7	of	of	ADP
ejpam-642	242	8	all	all	DET
ejpam-642	242	9	cells	cell	NOUN
ejpam-642	242	10	of	of	ADP
ejpam-642	242	11	the	the	DET
ejpam-642	242	12	table	table	NOUN
ejpam-642	242	13	is	be	AUX
ejpam-642	242	14	2n	2n	NUM
ejpam-642	242	15	;	;	PUNCT
ejpam-642	242	16	see	see	VERB
ejpam-642	242	17	[	[	X
ejpam-642	242	18	7	7	NUM
ejpam-642	242	19	]	]	PUNCT
ejpam-642	242	20	.	.	PUNCT
ejpam-642	243	1	he	he	PRON
ejpam-642	243	2	assumed	assume	VERB
ejpam-642	243	3	field	field	NOUN
ejpam-642	243	4	independence	independence	NOUN
ejpam-642	243	5	,	,	PUNCT
ejpam-642	243	6	and	and	CCONJ
ejpam-642	243	7	employed	employ	VERB
ejpam-642	243	8	the	the	DET
ejpam-642	243	9	em	em	PROPN
ejpam-642	243	10	algorithm	algorithm	NOUN
ejpam-642	243	11	for	for	ADP
ejpam-642	243	12	the	the	DET
ejpam-642	243	13	estimation	estimation	NOUN
ejpam-642	243	14	of	of	ADP
ejpam-642	243	15	all	all	DET
ejpam-642	243	16	probabilities	probability	NOUN
ejpam-642	243	17	related	relate	VERB
ejpam-642	243	18	to	to	ADP
ejpam-642	243	19	fields	field	NOUN
ejpam-642	243	20	and	and	CCONJ
ejpam-642	243	21	the	the	DET
ejpam-642	243	22	mixing	mix	VERB
ejpam-642	243	23	proportion	proportion	NOUN
ejpam-642	243	24	.	.	PUNCT
ejpam-642	244	1	from	from	ADP
ejpam-642	244	2	the	the	DET
ejpam-642	244	3	prospective	prospective	ADJ
ejpam-642	244	4	of	of	ADP
ejpam-642	244	5	loglinear	loglinear	NOUN
ejpam-642	244	6	model	model	NOUN
ejpam-642	244	7	formulation	formulation	NOUN
ejpam-642	244	8	,	,	PUNCT
ejpam-642	244	9	the	the	DET
ejpam-642	244	10	models	model	NOUN
ejpam-642	244	11	for	for	ADP
ejpam-642	244	12	m	m	PROPN
ejpam-642	244	13	and	and	CCONJ
ejpam-642	244	14	u	u	PROPN
ejpam-642	244	15	in	in	ADP
ejpam-642	244	16	jaro	jaro	PROPN
ejpam-642	244	17	(	(	PUNCT
ejpam-642	244	18	1989	1989	NUM
ejpam-642	244	19	)	)	PUNCT
ejpam-642	244	20	were	be	AUX
ejpam-642	244	21	loglinear	loglinear	NOUN
ejpam-642	244	22	models	model	NOUN
ejpam-642	244	23	without	without	ADP
ejpam-642	244	24	interactions	interaction	NOUN
ejpam-642	244	25	.	.	PUNCT
ejpam-642	245	1	the	the	DET
ejpam-642	245	2	field	field	NOUN
ejpam-642	245	3	independence	independence	NOUN
ejpam-642	245	4	may	may	AUX
ejpam-642	245	5	not	not	PART
ejpam-642	245	6	hold	hold	VERB
ejpam-642	245	7	in	in	ADP
ejpam-642	245	8	reality	reality	NOUN
ejpam-642	245	9	.	.	PUNCT
ejpam-642	246	1	for	for	ADP
ejpam-642	246	2	instance	instance	NOUN
ejpam-642	246	3	,	,	PUNCT
ejpam-642	246	4	strong	strong	ADJ
ejpam-642	246	5	dependencies	dependency	NOUN
ejpam-642	246	6	were	be	AUX
ejpam-642	246	7	observed	observe	VERB
ejpam-642	246	8	by	by	ADP
ejpam-642	246	9	thibaudeau	thibaudeau	NOUN
ejpam-642	246	10	(	(	PUNCT
ejpam-642	246	11	1993	1993	NUM
ejpam-642	246	12	)	)	PUNCT
ejpam-642	246	13	in	in	ADP
ejpam-642	246	14	census	census	NOUN
ejpam-642	246	15	data	datum	NOUN
ejpam-642	246	16	;	;	PUNCT
ejpam-642	246	17	see	see	VERB
ejpam-642	246	18	[	[	X
ejpam-642	246	19	14	14	NUM
ejpam-642	246	20	]	]	PUNCT
ejpam-642	246	21	.	.	PUNCT
ejpam-642	247	1	prior	prior	ADV
ejpam-642	247	2	to	to	ADP
ejpam-642	247	3	that	that	DET
ejpam-642	247	4	report	report	NOUN
ejpam-642	247	5	,	,	PUNCT
ejpam-642	247	6	winkler	winkler	X
ejpam-642	247	7	(	(	PUNCT
ejpam-642	247	8	1989	1989	NUM
ejpam-642	247	9	)	)	PUNCT
ejpam-642	247	10	extended	extend	VERB
ejpam-642	247	11	the	the	DET
ejpam-642	247	12	simplest	simple	ADJ
ejpam-642	247	13	case	case	NOUN
ejpam-642	247	14	to	to	ADP
ejpam-642	247	15	a	a	DET
ejpam-642	247	16	general	general	ADJ
ejpam-642	247	17	one	one	NOUN
ejpam-642	247	18	,	,	PUNCT
ejpam-642	247	19	where	where	SCONJ
ejpam-642	247	20	interactions	interaction	NOUN
ejpam-642	247	21	were	be	AUX
ejpam-642	247	22	taken	take	VERB
ejpam-642	247	23	into	into	ADP
ejpam-642	247	24	account	account	NOUN
ejpam-642	247	25	in	in	ADP
ejpam-642	247	26	the	the	DET
ejpam-642	247	27	models	model	NOUN
ejpam-642	247	28	for	for	ADP
ejpam-642	247	29	m	m	PROPN
ejpam-642	247	30	and	and	CCONJ
ejpam-642	247	31	u	u	X
ejpam-642	247	32	,	,	PUNCT
ejpam-642	247	33	specifically	specifically	ADV
ejpam-642	247	34	,	,	PUNCT
ejpam-642	247	35	two	two	NUM
ejpam-642	247	36	-	-	PUNCT
ejpam-642	247	37	class	class	NOUN
ejpam-642	247	38	and	and	CCONJ
ejpam-642	247	39	three	three	NUM
ejpam-642	247	40	-	-	PUNCT
ejpam-642	247	41	class	class	NOUN
ejpam-642	247	42	interactions	interaction	NOUN
ejpam-642	247	43	were	be	AUX
ejpam-642	247	44	considered	consider	VERB
ejpam-642	247	45	,	,	PUNCT
ejpam-642	247	46	and	and	CCONJ
ejpam-642	247	47	models	model	NOUN
ejpam-642	247	48	were	be	AUX
ejpam-642	247	49	assessed	assess	VERB
ejpam-642	247	50	using	use	VERB
ejpam-642	247	51	goodness	goodness	NOUN
ejpam-642	247	52	-	-	PUNCT
ejpam-642	247	53	of	of	ADP
ejpam-642	247	54	-	-	PUNCT
ejpam-642	247	55	fit	fit	ADJ
ejpam-642	247	56	tests	test	NOUN
ejpam-642	247	57	(	(	PUNCT
ejpam-642	247	58	see	see	VERB
ejpam-642	247	59	[	[	X
ejpam-642	247	60	16	16	NUM
ejpam-642	247	61	]	]	SYM
ejpam-642	247	62	)	)	PUNCT
ejpam-642	247	63	.	.	PUNCT
ejpam-642	248	1	larsen	larsen	PROPN
ejpam-642	248	2	(	(	PUNCT
ejpam-642	248	3	1997	1997	NUM
ejpam-642	248	4	)	)	PUNCT
ejpam-642	248	5	considered	consider	VERB
ejpam-642	248	6	a	a	DET
ejpam-642	248	7	loglinear	loglinear	NOUN
ejpam-642	248	8	model	model	NOUN
ejpam-642	248	9	with	with	ADP
ejpam-642	248	10	main	main	ADJ
ejpam-642	248	11	effects	effect	NOUN
ejpam-642	248	12	,	,	PUNCT
ejpam-642	248	13	all	all	DET
ejpam-642	248	14	two	two	NUM
ejpam-642	248	15	-	-	PUNCT
ejpam-642	248	16	way	way	NOUN
ejpam-642	248	17	interactions	interaction	NOUN
ejpam-642	248	18	,	,	PUNCT
ejpam-642	248	19	all	all	DET
ejpam-642	248	20	three	three	NUM
ejpam-642	248	21	-	-	PUNCT
ejpam-642	248	22	way	way	NOUN
ejpam-642	248	23	interactions	interaction	NOUN
ejpam-642	248	24	and	and	CCONJ
ejpam-642	248	25	two	two	NUM
ejpam-642	248	26	five	five	NUM
ejpam-642	248	27	-	-	PUNCT
ejpam-642	248	28	way	way	NOUN
ejpam-642	248	29	interactions	interaction	NOUN
ejpam-642	248	30	in	in	ADP
ejpam-642	248	31	a	a	DET
ejpam-642	248	32	trial	trial	NOUN
ejpam-642	248	33	census	census	NOUN
ejpam-642	248	34	and	and	CCONJ
ejpam-642	248	35	post	post	ADJ
ejpam-642	248	36	-	-	ADJ
ejpam-642	248	37	enumeration	enumeration	ADJ
ejpam-642	248	38	survey	survey	NOUN
ejpam-642	248	39	(	(	PUNCT
ejpam-642	248	40	see	see	VERB
ejpam-642	248	41	[	[	X
ejpam-642	248	42	9	9	NUM
ejpam-642	248	43	]	]	NUM
ejpam-642	248	44	)	)	PUNCT
ejpam-642	248	45	.	.	PUNCT
ejpam-642	249	1	any	any	PRON
ejpam-642	249	2	of	of	ADP
ejpam-642	249	3	the	the	DET
ejpam-642	249	4	specified	specify	VERB
ejpam-642	249	5	loglinear	loglinear	NOUN
ejpam-642	249	6	mixture	mixture	NOUN
ejpam-642	249	7	models	model	NOUN
ejpam-642	249	8	can	can	AUX
ejpam-642	249	9	be	be	AUX
ejpam-642	249	10	estimated	estimate	VERB
ejpam-642	249	11	using	use	VERB
ejpam-642	249	12	the	the	DET
ejpam-642	249	13	simplified	simplified	ADJ
ejpam-642	249	14	method	method	NOUN
ejpam-642	249	15	for	for	ADP
ejpam-642	249	16	the	the	DET
ejpam-642	249	17	em	em	PROPN
ejpam-642	249	18	algorithm	algorithm	NOUN
ejpam-642	249	19	in	in	ADP
ejpam-642	249	20	section	section	NOUN
ejpam-642	249	21	2	2	NUM
ejpam-642	249	22	.	.	PUNCT
ejpam-642	250	1	for	for	ADP
ejpam-642	250	2	a	a	DET
ejpam-642	250	3	real	real	ADJ
ejpam-642	250	4	case	case	NOUN
ejpam-642	250	5	,	,	PUNCT
ejpam-642	250	6	the	the	DET
ejpam-642	250	7	specification	specification	NOUN
ejpam-642	250	8	can	can	AUX
ejpam-642	250	9	be	be	AUX
ejpam-642	250	10	given	give	VERB
ejpam-642	250	11	from	from	ADP
ejpam-642	250	12	background	background	NOUN
ejpam-642	250	13	information	information	NOUN
ejpam-642	250	14	or	or	CCONJ
ejpam-642	250	15	subjective	subjective	ADJ
ejpam-642	250	16	opinion	opinion	NOUN
ejpam-642	250	17	.	.	PUNCT
ejpam-642	251	1	thus	thus	ADV
ejpam-642	251	2	,	,	PUNCT
ejpam-642	251	3	different	different	ADJ
ejpam-642	251	4	opinions	opinion	NOUN
ejpam-642	251	5	or	or	CCONJ
ejpam-642	251	6	understanding	understanding	NOUN
ejpam-642	251	7	about	about	ADP
ejpam-642	251	8	the	the	DET
ejpam-642	251	9	real	real	ADJ
ejpam-642	251	10	case	case	NOUN
ejpam-642	251	11	can	can	AUX
ejpam-642	251	12	result	result	VERB
ejpam-642	251	13	in	in	ADP
ejpam-642	251	14	different	different	ADJ
ejpam-642	251	15	model	model	NOUN
ejpam-642	251	16	specifications	specification	NOUN
ejpam-642	251	17	.	.	PUNCT
ejpam-642	252	1	however	however	ADV
ejpam-642	252	2	,	,	PUNCT
ejpam-642	252	3	among	among	ADP
ejpam-642	252	4	those	those	DET
ejpam-642	252	5	subjectively	subjectively	ADV
ejpam-642	252	6	specified	specify	VERB
ejpam-642	252	7	models	model	NOUN
ejpam-642	252	8	,	,	PUNCT
ejpam-642	252	9	which	which	DET
ejpam-642	252	10	one	one	PRON
ejpam-642	252	11	fits	fit	VERB
ejpam-642	252	12	the	the	DET
ejpam-642	252	13	observed	observe	VERB
ejpam-642	252	14	contingency	contingency	NOUN
ejpam-642	252	15	table	table	NOUN
ejpam-642	252	16	better	well	ADV
ejpam-642	252	17	,	,	PUNCT
ejpam-642	252	18	or	or	CCONJ
ejpam-642	252	19	captures	capture	VERB
ejpam-642	252	20	the	the	DET
ejpam-642	252	21	data	datum	NOUN
ejpam-642	252	22	feature	feature	VERB
ejpam-642	252	23	better	well	ADV
ejpam-642	252	24	?	?	PUNCT
ejpam-642	253	1	how	how	SCONJ
ejpam-642	253	2	can	can	AUX
ejpam-642	253	3	we	we	PRON
ejpam-642	253	4	avoid	avoid	VERB
ejpam-642	253	5	overfitting	overfitte	VERB
ejpam-642	253	6	or	or	CCONJ
ejpam-642	253	7	underfitting	underfitting	NOUN
ejpam-642	253	8	?	?	PUNCT
ejpam-642	254	1	these	these	DET
ejpam-642	254	2	questions	question	NOUN
ejpam-642	254	3	lead	lead	VERB
ejpam-642	254	4	to	to	ADP
ejpam-642	254	5	the	the	DET
ejpam-642	254	6	model	model	NOUN
ejpam-642	254	7	selection	selection	NOUN
ejpam-642	254	8	for	for	ADP
ejpam-642	254	9	loglinear	loglinear	NOUN
ejpam-642	254	10	mixture	mixture	NOUN
ejpam-642	254	11	models	model	NOUN
ejpam-642	254	12	in	in	ADP
ejpam-642	254	13	record	record	NOUN
ejpam-642	254	14	linkage	linkage	NOUN
ejpam-642	254	15	.	.	PUNCT
ejpam-642	255	1	usually	usually	ADV
ejpam-642	255	2	,	,	PUNCT
ejpam-642	255	3	a	a	DET
ejpam-642	255	4	better	well	ADJ
ejpam-642	255	5	model	model	NOUN
ejpam-642	255	6	will	will	AUX
ejpam-642	255	7	give	give	VERB
ejpam-642	255	8	a	a	DET
ejpam-642	255	9	better	well	ADJ
ejpam-642	255	10	estimation	estimation	NOUN
ejpam-642	255	11	of	of	ADP
ejpam-642	255	12	cell	cell	NOUN
ejpam-642	255	13	probabilities	probability	NOUN
ejpam-642	255	14	for	for	ADP
ejpam-642	255	15	matched	match	VERB
ejpam-642	255	16	and	and	CCONJ
ejpam-642	255	17	unmatched	unmatched	ADJ
ejpam-642	255	18	pairs	pair	NOUN
ejpam-642	255	19	,	,	PUNCT
ejpam-642	255	20	which	which	PRON
ejpam-642	255	21	in	in	ADP
ejpam-642	255	22	turn	turn	NOUN
ejpam-642	255	23	increases	increase	VERB
ejpam-642	255	24	the	the	DET
ejpam-642	255	25	accuracy	accuracy	NOUN
ejpam-642	255	26	of	of	ADP
ejpam-642	255	27	record	record	NOUN
ejpam-642	255	28	pair	pair	NOUN
ejpam-642	255	29	partition	partition	NOUN
ejpam-642	255	30	.	.	PUNCT
ejpam-642	256	1	when	when	SCONJ
ejpam-642	256	2	we	we	PRON
ejpam-642	256	3	search	search	VERB
ejpam-642	256	4	among	among	ADP
ejpam-642	256	5	a	a	DET
ejpam-642	256	6	series	series	NOUN
ejpam-642	256	7	of	of	ADP
ejpam-642	256	8	models	model	NOUN
ejpam-642	256	9	,	,	PUNCT
ejpam-642	256	10	a	a	DET
ejpam-642	256	11	method	method	NOUN
ejpam-642	256	12	of	of	ADP
ejpam-642	256	13	statistical	statistical	ADJ
ejpam-642	256	14	testing	testing	NOUN
ejpam-642	256	15	is	be	AUX
ejpam-642	256	16	required	require	VERB
ejpam-642	256	17	for	for	ADP
ejpam-642	256	18	model	model	NOUN
ejpam-642	256	19	comparison	comparison	NOUN
ejpam-642	256	20	at	at	ADP
ejpam-642	256	21	each	each	DET
ejpam-642	256	22	step	step	NOUN
ejpam-642	256	23	,	,	PUNCT
ejpam-642	256	24	and	and	CCONJ
ejpam-642	256	25	a	a	DET
ejpam-642	256	26	strategy	strategy	NOUN
ejpam-642	256	27	is	be	AUX
ejpam-642	256	28	needed	need	VERB
ejpam-642	256	29	in	in	ADP
ejpam-642	256	30	the	the	DET
ejpam-642	256	31	iterative	iterative	NOUN
ejpam-642	256	32	selection	selection	NOUN
ejpam-642	256	33	process	process	NOUN
ejpam-642	256	34	.	.	PUNCT
ejpam-642	257	1	suppose	suppose	VERB
ejpam-642	257	2	the	the	DET
ejpam-642	257	3	loglinear	loglinear	NOUN
ejpam-642	257	4	mixture	mixture	NOUN
ejpam-642	257	5	modelm	modelm	NOUN
ejpam-642	257	6	(	(	PUNCT
ejpam-642	257	7	0	0	NUM
ejpam-642	257	8	)	)	PUNCT
ejpam-642	257	9	is	be	AUX
ejpam-642	257	10	nested	nest	VERB
ejpam-642	257	11	in	in	ADP
ejpam-642	257	12	another	another	DET
ejpam-642	257	13	loglinear	loglinear	NOUN
ejpam-642	257	14	mixture	mixture	NOUN
ejpam-642	257	15	modelm	modelm	NOUN
ejpam-642	257	16	(	(	PUNCT
ejpam-642	257	17	1	1	NUM
ejpam-642	257	18	)	)	PUNCT
ejpam-642	257	19	,	,	PUNCT
ejpam-642	257	20	namelym	namelym	NOUN
ejpam-642	257	21	(	(	PUNCT
ejpam-642	257	22	0	0	NUM
ejpam-642	257	23	)	)	PUNCT
ejpam-642	257	24	⊂	⊂	PRON
ejpam-642	257	25	m	m	VERB
ejpam-642	257	26	(	(	PUNCT
ejpam-642	257	27	1	1	NUM
ejpam-642	257	28	)	)	PUNCT
ejpam-642	257	29	.	.	PUNCT
ejpam-642	258	1	we	we	PRON
ejpam-642	258	2	adopt	adopt	VERB
ejpam-642	258	3	a	a	DET
ejpam-642	258	4	general	general	ADJ
ejpam-642	258	5	method	method	NOUN
ejpam-642	258	6	to	to	PART
ejpam-642	258	7	compare	compare	VERB
ejpam-642	258	8	these	these	DET
ejpam-642	258	9	two	two	NUM
ejpam-642	258	10	loglinear	loglinear	NOUN
ejpam-642	258	11	mixture	mixture	NOUN
ejpam-642	258	12	models	model	NOUN
ejpam-642	258	13	,	,	PUNCT
ejpam-642	258	14	i.e.	i.e.	X
ejpam-642	258	15	,	,	PUNCT
ejpam-642	258	16	testing	testing	NOUN
ejpam-642	258	17	h0	h0	NOUN
ejpam-642	258	18	:	:	PUNCT
ejpam-642	258	19	m	m	PROPN
ejpam-642	258	20	(	(	PUNCT
ejpam-642	258	21	0	0	NUM
ejpam-642	258	22	)	)	PUNCT
ejpam-642	259	1	=	=	NOUN
ejpam-642	259	2	m	m	NOUN
ejpam-642	259	3	(	(	PUNCT
ejpam-642	259	4	1	1	NUM
ejpam-642	259	5	)	)	PUNCT
ejpam-642	259	6	versus	versus	ADP
ejpam-642	259	7	h1	h1	PROPN
ejpam-642	259	8	:	:	PUNCT
ejpam-642	259	9	m	m	VERB
ejpam-642	259	10	(	(	PUNCT
ejpam-642	259	11	0	0	NUM
ejpam-642	259	12	)	)	PUNCT
ejpam-642	259	13	⊂m	⊂m	PROPN
ejpam-642	259	14	(	(	PUNCT
ejpam-642	259	15	1	1	NUM
ejpam-642	259	16	)	)	PUNCT
ejpam-642	259	17	.	.	PUNCT
ejpam-642	260	1	if	if	SCONJ
ejpam-642	260	2	h0	h0	PROPN
ejpam-642	260	3	is	be	AUX
ejpam-642	260	4	not	not	PART
ejpam-642	260	5	rejected	reject	VERB
ejpam-642	260	6	,	,	PUNCT
ejpam-642	260	7	we	we	PRON
ejpam-642	260	8	do	do	AUX
ejpam-642	260	9	not	not	PART
ejpam-642	260	10	need	need	VERB
ejpam-642	260	11	to	to	PART
ejpam-642	260	12	improvem	improvem	VERB
ejpam-642	260	13	(	(	PUNCT
ejpam-642	260	14	0	0	NUM
ejpam-642	260	15	)	)	PUNCT
ejpam-642	260	16	by	by	ADP
ejpam-642	260	17	considering	consider	VERB
ejpam-642	260	18	the	the	DET
ejpam-642	260	19	larger	large	ADJ
ejpam-642	260	20	modelm	modelm	NOUN
ejpam-642	260	21	(	(	PUNCT
ejpam-642	260	22	1	1	NUM
ejpam-642	260	23	)	)	PUNCT
ejpam-642	260	24	.	.	PUNCT
ejpam-642	261	1	otherwise	otherwise	ADV
ejpam-642	261	2	,	,	PUNCT
ejpam-642	261	3	model	model	NOUN
ejpam-642	261	4	refinement	refinement	NOUN
ejpam-642	261	5	is	be	AUX
ejpam-642	261	6	necessary	necessary	ADJ
ejpam-642	261	7	.	.	PUNCT
ejpam-642	262	1	for	for	ADP
ejpam-642	262	2	this	this	DET
ejpam-642	262	3	testing	testing	NOUN
ejpam-642	262	4	,	,	PUNCT
ejpam-642	262	5	the	the	DET
ejpam-642	262	6	pearson	pearson	PROPN
ejpam-642	262	7	chi	chi	PROPN
ejpam-642	262	8	-	-	PUNCT
ejpam-642	262	9	squared	square	VERB
ejpam-642	262	10	statistic	statistic	NOUN
ejpam-642	262	11	is	be	AUX
ejpam-642	262	12	r.	r.	PROPN
ejpam-642	262	13	zhu	zhu	PROPN
ejpam-642	262	14	,	,	PUNCT
ejpam-642	262	15	j.	j.	PROPN
ejpam-642	262	16	zhang	zhang	PROPN
ejpam-642	262	17	,	,	PUNCT
ejpam-642	262	18	d.	d.	PROPN
ejpam-642	262	19	zhang	zhang	PROPN
ejpam-642	262	20	,	,	PUNCT
ejpam-642	262	21	g.	g.	PROPN
ejpam-642	262	22	yan	yan	PROPN
ejpam-642	262	23	/	/	SYM
ejpam-642	262	24	eur	eur	PROPN
ejpam-642	262	25	.	.	PUNCT
ejpam-642	263	1	j.	j.	PROPN
ejpam-642	263	2	pure	pure	PROPN
ejpam-642	263	3	appl	appl	PROPN
ejpam-642	263	4	.	.	PROPN
ejpam-642	263	5	math	math	PROPN
ejpam-642	263	6	,	,	PUNCT
ejpam-642	263	7	3	3	NUM
ejpam-642	263	8	(	(	PUNCT
ejpam-642	263	9	2010	2010	NUM
ejpam-642	263	10	)	)	PUNCT
ejpam-642	263	11	,	,	PUNCT
ejpam-642	263	12	141	141	NUM
ejpam-642	263	13	-	-	SYM
ejpam-642	263	14	162	162	NUM
ejpam-642	263	15	150	150	NUM
ejpam-642	263	16	employed	employ	VERB
ejpam-642	263	17	as	as	SCONJ
ejpam-642	263	18	follows	follow	VERB
ejpam-642	263	19	χ2	χ2	PROPN
ejpam-642	263	20	�	�	PROPN
ejpam-642	263	21	m	m	PROPN
ejpam-642	263	22	(	(	PUNCT
ejpam-642	263	23	0)|m	0)|m	PROPN
ejpam-642	263	24	(	(	PUNCT
ejpam-642	263	25	1	1	NUM
ejpam-642	263	26	)	)	PUNCT
ejpam-642	263	27	�	�	PROPN
ejpam-642	263	28	=	=	SYM
ejpam-642	263	29	∑	∑	PROPN
ejpam-642	263	30	(	(	PUNCT
ejpam-642	263	31	µ̂1i	µ̂1i	NUM
ejpam-642	263	32	−	−	PROPN
ejpam-642	263	33	µ̂0i	µ̂0i	NUM
ejpam-642	263	34	)	)	PUNCT
ejpam-642	263	35	2/µ̂0i	2/µ̂0i	NUM
ejpam-642	263	36	,	,	PUNCT
ejpam-642	263	37	(	(	PUNCT
ejpam-642	263	38	10	10	NUM
ejpam-642	263	39	)	)	PUNCT
ejpam-642	263	40	where	where	SCONJ
ejpam-642	263	41	µ̂1i	µ̂1i	NUM
ejpam-642	263	42	and	and	CCONJ
ejpam-642	263	43	µ̂0i	µ̂0i	VERB
ejpam-642	263	44	are	be	AUX
ejpam-642	263	45	fitted	fit	VERB
ejpam-642	263	46	values	value	NOUN
ejpam-642	263	47	of	of	ADP
ejpam-642	263	48	individual	individual	ADJ
ejpam-642	263	49	cell	cell	NOUN
ejpam-642	263	50	expectations	expectation	NOUN
ejpam-642	263	51	form	form	NOUN
ejpam-642	263	52	(	(	PUNCT
ejpam-642	263	53	0	0	NUM
ejpam-642	263	54	)	)	PUNCT
ejpam-642	263	55	andm	andm	NOUN
ejpam-642	263	56	(	(	PUNCT
ejpam-642	263	57	1	1	NUM
ejpam-642	263	58	)	)	PUNCT
ejpam-642	263	59	respectively	respectively	ADV
ejpam-642	263	60	.	.	PUNCT
ejpam-642	264	1	because	because	SCONJ
ejpam-642	264	2	one	one	NUM
ejpam-642	264	3	model	model	NOUN
ejpam-642	264	4	is	be	AUX
ejpam-642	264	5	nested	nest	VERB
ejpam-642	264	6	in	in	ADP
ejpam-642	264	7	the	the	DET
ejpam-642	264	8	other	other	ADJ
ejpam-642	264	9	,	,	PUNCT
ejpam-642	264	10	the	the	DET
ejpam-642	264	11	asymptotic	asymptotic	ADJ
ejpam-642	264	12	distribution	distribution	NOUN
ejpam-642	264	13	of	of	ADP
ejpam-642	264	14	χ2	χ2	PROPN
ejpam-642	264	15	�	�	PROPN
ejpam-642	264	16	m	m	PROPN
ejpam-642	264	17	(	(	PUNCT
ejpam-642	264	18	0)|m	0)|m	PROPN
ejpam-642	264	19	(	(	PUNCT
ejpam-642	264	20	1	1	NUM
ejpam-642	264	21	)	)	PUNCT
ejpam-642	264	22	�	�	PROPN
ejpam-642	264	23	under	under	ADP
ejpam-642	264	24	h0	h0	PROPN
ejpam-642	264	25	is	be	AUX
ejpam-642	264	26	χ2(d	χ2(d	PROPN
ejpam-642	264	27	f	f	PROPN
ejpam-642	264	28	)	)	PUNCT
ejpam-642	264	29	,	,	PUNCT
ejpam-642	264	30	where	where	SCONJ
ejpam-642	264	31	the	the	DET
ejpam-642	264	32	degrees	degree	NOUN
ejpam-642	264	33	of	of	ADP
ejpam-642	264	34	freedom	freedom	NOUN
ejpam-642	264	35	d	d	X
ejpam-642	264	36	f	f	PROPN
ejpam-642	264	37	is	be	AUX
ejpam-642	264	38	the	the	DET
ejpam-642	264	39	difference	difference	NOUN
ejpam-642	264	40	of	of	ADP
ejpam-642	264	41	parameter	parameter	NOUN
ejpam-642	264	42	numbers	number	NOUN
ejpam-642	264	43	betweenm	betweenm	NOUN
ejpam-642	264	44	(	(	PUNCT
ejpam-642	264	45	0	0	NUM
ejpam-642	264	46	)	)	PUNCT
ejpam-642	264	47	andm	andm	NOUN
ejpam-642	264	48	(	(	PUNCT
ejpam-642	264	49	1	1	NUM
ejpam-642	264	50	)	)	PUNCT
ejpam-642	264	51	.	.	PUNCT
ejpam-642	265	1	for	for	ADP
ejpam-642	265	2	a	a	DET
ejpam-642	265	3	reference	reference	NOUN
ejpam-642	265	4	,	,	PUNCT
ejpam-642	265	5	see	see	VERB
ejpam-642	265	6	[	[	X
ejpam-642	265	7	2	2	NUM
ejpam-642	265	8	]	]	PUNCT
ejpam-642	265	9	,	,	PUNCT
ejpam-642	265	10	p.	p.	NOUN
ejpam-642	265	11	364	364	NUM
ejpam-642	265	12	.	.	PUNCT
ejpam-642	266	1	the	the	DET
ejpam-642	266	2	loglinear	loglinear	NOUN
ejpam-642	266	3	models	model	NOUN
ejpam-642	266	4	for	for	ADP
ejpam-642	266	5	components	component	NOUN
ejpam-642	266	6	u	u	NOUN
ejpam-642	266	7	and	and	CCONJ
ejpam-642	266	8	m	m	VERB
ejpam-642	266	9	in	in	ADP
ejpam-642	266	10	the	the	DET
ejpam-642	266	11	mixture	mixture	NOUN
ejpam-642	266	12	we	we	PRON
ejpam-642	266	13	consider	consider	VERB
ejpam-642	266	14	for	for	ADP
ejpam-642	266	15	record	record	NOUN
ejpam-642	266	16	linkage	linkage	NOUN
ejpam-642	266	17	may	may	AUX
ejpam-642	266	18	or	or	CCONJ
ejpam-642	266	19	may	may	AUX
ejpam-642	266	20	not	not	PART
ejpam-642	266	21	include	include	VERB
ejpam-642	266	22	higher	high	ADJ
ejpam-642	266	23	order	order	NOUN
ejpam-642	266	24	interactions	interaction	NOUN
ejpam-642	266	25	among	among	ADP
ejpam-642	266	26	fields	field	NOUN
ejpam-642	266	27	.	.	PUNCT
ejpam-642	267	1	we	we	PRON
ejpam-642	267	2	denotemi	denotemi	VERB
ejpam-642	267	3	j	j	PROPN
ejpam-642	267	4	for	for	ADP
ejpam-642	267	5	a	a	DET
ejpam-642	267	6	loglinear	loglinear	NOUN
ejpam-642	267	7	mixture	mixture	NOUN
ejpam-642	267	8	model	model	NOUN
ejpam-642	267	9	where	where	SCONJ
ejpam-642	267	10	i	i	PRON
ejpam-642	267	11	indicates	indicate	VERB
ejpam-642	267	12	that	that	SCONJ
ejpam-642	267	13	the	the	DET
ejpam-642	267	14	loglinear	loglinear	NOUN
ejpam-642	267	15	model	model	NOUN
ejpam-642	267	16	for	for	ADP
ejpam-642	267	17	u	u	PROPN
ejpam-642	267	18	has	have	VERB
ejpam-642	267	19	up	up	ADP
ejpam-642	267	20	to	to	ADP
ejpam-642	267	21	the	the	DET
ejpam-642	267	22	(	(	PUNCT
ejpam-642	267	23	i+1)th	i+1)th	NOUN
ejpam-642	267	24	order	order	NOUN
ejpam-642	267	25	interactions	interaction	NOUN
ejpam-642	267	26	(	(	PUNCT
ejpam-642	267	27	i.e.	i.e.	X
ejpam-642	267	28	,	,	PUNCT
ejpam-642	267	29	(	(	PUNCT
ejpam-642	267	30	i+1)-factor	i+1)-factor	NOUN
ejpam-642	267	31	-	-	PUNCT
ejpam-642	267	32	interactions	interaction	NOUN
ejpam-642	267	33	)	)	PUNCT
ejpam-642	267	34	and	and	CCONJ
ejpam-642	267	35	j	j	PROPN
ejpam-642	267	36	indicates	indicate	VERB
ejpam-642	267	37	that	that	SCONJ
ejpam-642	267	38	the	the	DET
ejpam-642	267	39	loglinear	loglinear	NOUN
ejpam-642	267	40	model	model	NOUN
ejpam-642	267	41	for	for	ADP
ejpam-642	267	42	m	m	PROPN
ejpam-642	267	43	has	have	VERB
ejpam-642	267	44	up	up	ADP
ejpam-642	267	45	to	to	ADP
ejpam-642	267	46	the	the	DET
ejpam-642	267	47	(	(	PUNCT
ejpam-642	267	48	j+1)-th	j+1)-th	NOUN
ejpam-642	267	49	order	order	NOUN
ejpam-642	267	50	interactions	interaction	NOUN
ejpam-642	267	51	(	(	PUNCT
ejpam-642	267	52	i.e.	i.e.	X
ejpam-642	267	53	,	,	PUNCT
ejpam-642	267	54	(	(	PUNCT
ejpam-642	267	55	j+1)-factor	j+1)-factor	NOUN
ejpam-642	267	56	-	-	PUNCT
ejpam-642	267	57	interactions	interaction	NOUN
ejpam-642	267	58	)	)	PUNCT
ejpam-642	267	59	,	,	PUNCT
ejpam-642	267	60	i	i	PRON
ejpam-642	267	61	,	,	PUNCT
ejpam-642	267	62	j	j	PROPN
ejpam-642	267	63	≥	≥	PROPN
ejpam-642	267	64	0	0	NUM
ejpam-642	267	65	.	.	PUNCT
ejpam-642	268	1	for	for	ADP
ejpam-642	268	2	instance	instance	NOUN
ejpam-642	268	3	,	,	PUNCT
ejpam-642	268	4	m00	m00	PROPN
ejpam-642	268	5	means	mean	VERB
ejpam-642	268	6	that	that	SCONJ
ejpam-642	268	7	the	the	DET
ejpam-642	268	8	models	model	NOUN
ejpam-642	268	9	for	for	ADP
ejpam-642	268	10	both	both	DET
ejpam-642	268	11	m	m	PROPN
ejpam-642	268	12	and	and	CCONJ
ejpam-642	268	13	u	u	NOUN
ejpam-642	268	14	in	in	ADP
ejpam-642	268	15	the	the	DET
ejpam-642	268	16	mixture	mixture	NOUN
ejpam-642	268	17	have	have	VERB
ejpam-642	268	18	no	no	DET
ejpam-642	268	19	interactions	interaction	NOUN
ejpam-642	268	20	,	,	PUNCT
ejpam-642	268	21	i.e.	i.e.	X
ejpam-642	268	22	,	,	PUNCT
ejpam-642	268	23	model	model	NOUN
ejpam-642	268	24	(	(	PUNCT
ejpam-642	268	25	6).m10	6).m10	NUM
ejpam-642	268	26	means	mean	VERB
ejpam-642	268	27	that	that	SCONJ
ejpam-642	268	28	the	the	DET
ejpam-642	268	29	model	model	NOUN
ejpam-642	268	30	for	for	ADP
ejpam-642	268	31	u	u	PROPN
ejpam-642	268	32	has	have	VERB
ejpam-642	268	33	two	two	NUM
ejpam-642	268	34	-	-	PUNCT
ejpam-642	268	35	factor	factor	NOUN
ejpam-642	268	36	interactions	interaction	NOUN
ejpam-642	268	37	,	,	PUNCT
ejpam-642	268	38	while	while	SCONJ
ejpam-642	268	39	the	the	DET
ejpam-642	268	40	model	model	NOUN
ejpam-642	268	41	for	for	ADP
ejpam-642	268	42	m	m	PROPN
ejpam-642	268	43	has	have	VERB
ejpam-642	268	44	no	no	DET
ejpam-642	268	45	interactions	interaction	NOUN
ejpam-642	268	46	.	.	PUNCT
ejpam-642	269	1	m02	m02	PROPN
ejpam-642	269	2	means	mean	VERB
ejpam-642	269	3	that	that	SCONJ
ejpam-642	269	4	the	the	DET
ejpam-642	269	5	model	model	NOUN
ejpam-642	269	6	for	for	ADP
ejpam-642	269	7	u	u	PROPN
ejpam-642	269	8	has	have	VERB
ejpam-642	269	9	no	no	DET
ejpam-642	269	10	interactions	interaction	NOUN
ejpam-642	269	11	,	,	PUNCT
ejpam-642	269	12	while	while	SCONJ
ejpam-642	269	13	the	the	DET
ejpam-642	269	14	model	model	NOUN
ejpam-642	269	15	for	for	ADP
ejpam-642	269	16	m	m	PROPN
ejpam-642	269	17	has	have	VERB
ejpam-642	269	18	two	two	NUM
ejpam-642	269	19	-	-	PUNCT
ejpam-642	269	20	factor	factor	NOUN
ejpam-642	269	21	and	and	CCONJ
ejpam-642	269	22	three	three	NUM
ejpam-642	269	23	-	-	PUNCT
ejpam-642	269	24	factor	factor	NOUN
ejpam-642	269	25	interactions	interaction	NOUN
ejpam-642	269	26	.	.	PUNCT
ejpam-642	270	1	the	the	DET
ejpam-642	270	2	nested	nested	ADJ
ejpam-642	270	3	situation	situation	NOUN
ejpam-642	270	4	of	of	ADP
ejpam-642	270	5	two	two	NUM
ejpam-642	270	6	loglinear	loglinear	NOUN
ejpam-642	270	7	mixture	mixture	NOUN
ejpam-642	270	8	models	model	NOUN
ejpam-642	270	9	is	be	AUX
ejpam-642	270	10	totally	totally	ADV
ejpam-642	270	11	determined	determine	VERB
ejpam-642	270	12	by	by	ADP
ejpam-642	270	13	the	the	DET
ejpam-642	270	14	nested	nested	ADJ
ejpam-642	270	15	situations	situation	NOUN
ejpam-642	270	16	of	of	ADP
ejpam-642	270	17	their	their	PRON
ejpam-642	270	18	both	both	DET
ejpam-642	270	19	components	component	NOUN
ejpam-642	270	20	.	.	PUNCT
ejpam-642	271	1	therefore	therefore	ADV
ejpam-642	271	2	,	,	PUNCT
ejpam-642	271	3	we	we	PRON
ejpam-642	271	4	can	can	AUX
ejpam-642	271	5	build	build	VERB
ejpam-642	271	6	nested	nested	ADJ
ejpam-642	271	7	loglinear	loglinear	NOUN
ejpam-642	271	8	mixture	mixture	NOUN
ejpam-642	271	9	models	model	NOUN
ejpam-642	271	10	by	by	ADP
ejpam-642	271	11	building	build	VERB
ejpam-642	271	12	nested	nest	VERB
ejpam-642	271	13	loglinear	loglinear	NOUN
ejpam-642	271	14	models	model	NOUN
ejpam-642	271	15	for	for	ADP
ejpam-642	271	16	either	either	CCONJ
ejpam-642	271	17	u	u	NOUN
ejpam-642	271	18	or	or	CCONJ
ejpam-642	271	19	m	m	VERB
ejpam-642	271	20	,	,	PUNCT
ejpam-642	271	21	or	or	CCONJ
ejpam-642	271	22	both	both	DET
ejpam-642	271	23	u	u	NOUN
ejpam-642	271	24	and	and	CCONJ
ejpam-642	271	25	m	m	PROPN
ejpam-642	271	26	.	.	PUNCT
ejpam-642	272	1	that	that	PRON
ejpam-642	272	2	is	be	AUX
ejpam-642	272	3	,	,	PUNCT
ejpam-642	272	4	mi	mi	PROPN
ejpam-642	272	5	j	j	PROPN
ejpam-642	272	6	⊂mi′	⊂mi′	PROPN
ejpam-642	272	7	j′	j′	PROPN
ejpam-642	272	8	if	if	SCONJ
ejpam-642	272	9	i	i	PRON
ejpam-642	272	10	≤	≤	VERB
ejpam-642	272	11	i′	i′	NOUN
ejpam-642	272	12	and	and	CCONJ
ejpam-642	272	13	j	j	PROPN
ejpam-642	272	14	≤	≤	ADJ
ejpam-642	272	15	j′.	j′.	NOUN
ejpam-642	272	16	for	for	ADP
ejpam-642	272	17	instance	instance	NOUN
ejpam-642	272	18	,	,	PUNCT
ejpam-642	272	19	m00	m00	VERB
ejpam-642	272	20	⊂m01	⊂m01	PROPN
ejpam-642	272	21	⊂m11	⊂m11	PROPN
ejpam-642	272	22	andm00	andm00	PROPN
ejpam-642	272	23	⊂m10	⊂m10	PROPN
ejpam-642	272	24	⊂m11	⊂m11	NOUN
ejpam-642	272	25	,	,	PUNCT
ejpam-642	272	26	butm01	butm01	NOUN
ejpam-642	272	27	andm10	andm10	NOUN
ejpam-642	272	28	are	be	AUX
ejpam-642	272	29	not	not	PART
ejpam-642	272	30	nested	nest	VERB
ejpam-642	272	31	in	in	ADP
ejpam-642	272	32	each	each	DET
ejpam-642	272	33	other	other	ADJ
ejpam-642	272	34	.	.	PUNCT
ejpam-642	273	1	see	see	VERB
ejpam-642	273	2	the	the	DET
ejpam-642	273	3	illustration	illustration	NOUN
ejpam-642	273	4	in	in	ADP
ejpam-642	273	5	(	(	PUNCT
ejpam-642	273	6	11	11	NUM
ejpam-642	273	7	)	)	PUNCT
ejpam-642	273	8	where	where	SCONJ
ejpam-642	273	9	mixture	mixture	NOUN
ejpam-642	273	10	models	model	NOUN
ejpam-642	273	11	of	of	ADP
ejpam-642	273	12	two	two	NUM
ejpam-642	273	13	components	component	NOUN
ejpam-642	273	14	are	be	AUX
ejpam-642	273	15	arranged	arrange	VERB
ejpam-642	273	16	in	in	ADP
ejpam-642	273	17	a	a	DET
ejpam-642	273	18	matrix	matrix	NOUN
ejpam-642	273	19	format	format	NOUN
ejpam-642	273	20	.	.	PUNCT
ejpam-642	273	21	...	...	PUNCT
ejpam-642	273	22	...	...	PUNCT
ejpam-642	273	23	...	...	PUNCT
ejpam-642	273	24	...	...	PUNCT
ejpam-642	274	1	m30	m30	PROPN
ejpam-642	274	2	m31	m31	PROPN
ejpam-642	274	3	m32	m32	PROPN
ejpam-642	274	4	→	→	SYM
ejpam-642	274	5	m33	m33	PROPN
ejpam-642	274	6	·	·	PUNCT
ejpam-642	274	7	·	·	PUNCT
ejpam-642	274	8	·	·	PUNCT
ejpam-642	275	1	↑	↑	PROPN
ejpam-642	275	2	⇑	⇑	PROPN
ejpam-642	275	3	m20	m20	PROPN
ejpam-642	275	4	m21	m21	PROPN
ejpam-642	275	5	→	→	SYM
ejpam-642	275	6	m22	m22	PROPN
ejpam-642	275	7	⇒	⇒	PROPN
ejpam-642	275	8	m23	m23	X
ejpam-642	275	9	·	·	PUNCT
ejpam-642	275	10	·	·	PUNCT
ejpam-642	275	11	·	·	PUNCT
ejpam-642	276	1	↑	↑	PROPN
ejpam-642	276	2	⇑	⇑	PROPN
ejpam-642	276	3	m10	m10	PROPN
ejpam-642	276	4	→	→	SYM
ejpam-642	276	5	m11	m11	ADJ
ejpam-642	276	6	⇒	⇒	PROPN
ejpam-642	276	7	m12	m12	PROPN
ejpam-642	276	8	m13	m13	X
ejpam-642	276	9	·	·	PUNCT
ejpam-642	276	10	·	·	PUNCT
ejpam-642	276	11	·	·	PUNCT
ejpam-642	277	1	↑	↑	PROPN
ejpam-642	277	2	⇑	⇑	PROPN
ejpam-642	277	3	m00	m00	PROPN
ejpam-642	277	4	⇒	⇒	PROPN
ejpam-642	277	5	m01	m01	PROPN
ejpam-642	277	6	m02	m02	PROPN
ejpam-642	277	7	m03	m03	NOUN
ejpam-642	277	8	·	·	PUNCT
ejpam-642	277	9	·	·	PUNCT
ejpam-642	277	10	·	·	PUNCT
ejpam-642	278	1	(	(	PUNCT
ejpam-642	278	2	11	11	X
ejpam-642	278	3	)	)	PUNCT
ejpam-642	278	4	we	we	PRON
ejpam-642	278	5	make	make	VERB
ejpam-642	278	6	comparison	comparison	NOUN
ejpam-642	278	7	for	for	ADP
ejpam-642	278	8	models	model	NOUN
ejpam-642	278	9	on	on	ADP
ejpam-642	278	10	the	the	DET
ejpam-642	278	11	nested	nested	ADJ
ejpam-642	278	12	pathways	pathway	NOUN
ejpam-642	278	13	.	.	PUNCT
ejpam-642	279	1	however	however	ADV
ejpam-642	279	2	,	,	PUNCT
ejpam-642	279	3	there	there	PRON
ejpam-642	279	4	are	be	VERB
ejpam-642	279	5	many	many	ADJ
ejpam-642	279	6	nested	nested	ADJ
ejpam-642	279	7	paths	path	NOUN
ejpam-642	279	8	in	in	ADP
ejpam-642	279	9	the	the	DET
ejpam-642	279	10	family	family	NOUN
ejpam-642	279	11	of	of	ADP
ejpam-642	279	12	loglinear	loglinear	NOUN
ejpam-642	279	13	mixtures	mixture	NOUN
ejpam-642	279	14	of	of	ADP
ejpam-642	279	15	two	two	NUM
ejpam-642	279	16	components	component	NOUN
ejpam-642	279	17	.	.	PUNCT
ejpam-642	280	1	for	for	ADP
ejpam-642	280	2	example	example	NOUN
ejpam-642	280	3	,	,	PUNCT
ejpam-642	280	4	fromm00	fromm00	PROPN
ejpam-642	280	5	tom11	tom11	PROPN
ejpam-642	280	6	,	,	PUNCT
ejpam-642	280	7	there	there	PRON
ejpam-642	280	8	are	be	VERB
ejpam-642	280	9	two	two	NUM
ejpam-642	280	10	nested	nested	ADJ
ejpam-642	280	11	pathways	pathway	NOUN
ejpam-642	280	12	:	:	PUNCT
ejpam-642	280	13	m00	m00	PROPN
ejpam-642	280	14	⊂	⊂	PROPN
ejpam-642	280	15	m10	m10	PROPN
ejpam-642	280	16	⊂	⊂	PROPN
ejpam-642	280	17	m11	m11	PROPN
ejpam-642	280	18	and	and	CCONJ
ejpam-642	280	19	m00	m00	PROPN
ejpam-642	280	20	⊂	⊂	PROPN
ejpam-642	280	21	m01	m01	PROPN
ejpam-642	280	22	⊂	⊂	PROPN
ejpam-642	280	23	m11	m11	PROPN
ejpam-642	280	24	.	.	PUNCT
ejpam-642	281	1	in	in	ADP
ejpam-642	281	2	general	general	ADJ
ejpam-642	281	3	,	,	PUNCT
ejpam-642	281	4	from	from	ADP
ejpam-642	281	5	one	one	NUM
ejpam-642	281	6	diagonal	diagonal	ADJ
ejpam-642	281	7	modelmii	modelmii	NOUN
ejpam-642	281	8	to	to	ADP
ejpam-642	281	9	the	the	DET
ejpam-642	281	10	next	next	ADJ
ejpam-642	281	11	diagonal	diagonal	ADJ
ejpam-642	281	12	model	model	NOUN
ejpam-642	281	13	m(i+1)(i+1	m(i+1)(i+1	PROPN
ejpam-642	281	14	)	)	PUNCT
ejpam-642	281	15	,	,	PUNCT
ejpam-642	281	16	there	there	PRON
ejpam-642	281	17	are	be	VERB
ejpam-642	281	18	two	two	NUM
ejpam-642	281	19	nested	nested	ADJ
ejpam-642	281	20	pathways	pathway	NOUN
ejpam-642	281	21	similar	similar	ADJ
ejpam-642	281	22	to	to	ADP
ejpam-642	281	23	this	this	DET
ejpam-642	281	24	illustrated	illustrate	VERB
ejpam-642	281	25	example	example	NOUN
ejpam-642	281	26	.	.	PUNCT
ejpam-642	282	1	refer	refer	VERB
ejpam-642	282	2	to	to	ADP
ejpam-642	282	3	(	(	PUNCT
ejpam-642	282	4	11	11	NUM
ejpam-642	282	5	)	)	PUNCT
ejpam-642	282	6	.	.	PUNCT
ejpam-642	283	1	this	this	DET
ejpam-642	283	2	pair	pair	NOUN
ejpam-642	283	3	of	of	ADP
ejpam-642	283	4	nested	nested	ADJ
ejpam-642	283	5	pathways	pathway	NOUN
ejpam-642	283	6	is	be	AUX
ejpam-642	283	7	called	call	VERB
ejpam-642	283	8	alternatively	alternatively	ADV
ejpam-642	283	9	climbing	climb	VERB
ejpam-642	283	10	pathways	pathway	NOUN
ejpam-642	283	11	(	(	PUNCT
ejpam-642	283	12	denoted	denote	VERB
ejpam-642	283	13	by	by	ADP
ejpam-642	283	14	“	"	PUNCT
ejpam-642	283	15	→	→	SYM
ejpam-642	283	16	”	"	PUNCT
ejpam-642	283	17	and	and	CCONJ
ejpam-642	283	18	“	"	PUNCT
ejpam-642	283	19	⇒	⇒	NOUN
ejpam-642	283	20	”	"	PUNCT
ejpam-642	283	21	respectively	respectively	ADV
ejpam-642	283	22	)	)	PUNCT
ejpam-642	283	23	.	.	PUNCT
ejpam-642	284	1	we	we	PRON
ejpam-642	284	2	propose	propose	VERB
ejpam-642	284	3	the	the	DET
ejpam-642	284	4	following	follow	VERB
ejpam-642	284	5	forward	forward	ADJ
ejpam-642	284	6	selection	selection	NOUN
ejpam-642	284	7	strategy	strategy	NOUN
ejpam-642	284	8	along	along	ADP
ejpam-642	284	9	the	the	DET
ejpam-642	284	10	alternatively	alternatively	ADV
ejpam-642	284	11	climbing	climb	VERB
ejpam-642	284	12	pathways	pathway	NOUN
ejpam-642	284	13	:	:	PUNCT
ejpam-642	284	14	•	•	NUM
ejpam-642	284	15	step	step	NOUN
ejpam-642	284	16	1	1	NUM
ejpam-642	284	17	:	:	PUNCT
ejpam-642	284	18	start	start	VERB
ejpam-642	284	19	from	from	ADP
ejpam-642	284	20	the	the	DET
ejpam-642	284	21	loglinear	loglinear	NOUN
ejpam-642	284	22	mixture	mixture	NOUN
ejpam-642	284	23	modelm00	modelm00	NOUN
ejpam-642	284	24	(	(	PUNCT
ejpam-642	284	25	the	the	DET
ejpam-642	284	26	one	one	NOUN
ejpam-642	284	27	on	on	ADP
ejpam-642	284	28	the	the	DET
ejpam-642	284	29	bottom	bottom	NOUN
ejpam-642	284	30	left	leave	VERB
ejpam-642	284	31	corner	corner	NOUN
ejpam-642	284	32	in	in	ADP
ejpam-642	284	33	(	(	PUNCT
ejpam-642	284	34	11	11	NUM
ejpam-642	284	35	)	)	PUNCT
ejpam-642	284	36	)	)	PUNCT
ejpam-642	284	37	,	,	PUNCT
ejpam-642	284	38	where	where	SCONJ
ejpam-642	284	39	the	the	DET
ejpam-642	284	40	loglinear	loglinear	NOUN
ejpam-642	284	41	models	model	VERB
ejpam-642	284	42	for	for	ADP
ejpam-642	284	43	both	both	CCONJ
ejpam-642	284	44	u	u	NOUN
ejpam-642	284	45	and	and	CCONJ
ejpam-642	284	46	m	m	VERB
ejpam-642	284	47	have	have	VERB
ejpam-642	284	48	only	only	ADJ
ejpam-642	284	49	main	main	ADJ
ejpam-642	284	50	factor	factor	NOUN
ejpam-642	284	51	effects	effect	NOUN
ejpam-642	284	52	with	with	ADP
ejpam-642	284	53	no	no	DET
ejpam-642	284	54	interactions	interaction	NOUN
ejpam-642	284	55	.	.	PUNCT
ejpam-642	285	1	r.	r.	PROPN
ejpam-642	285	2	zhu	zhu	PROPN
ejpam-642	285	3	,	,	PUNCT
ejpam-642	285	4	j.	j.	PROPN
ejpam-642	285	5	zhang	zhang	PROPN
ejpam-642	285	6	,	,	PUNCT
ejpam-642	285	7	d.	d.	PROPN
ejpam-642	285	8	zhang	zhang	PROPN
ejpam-642	285	9	,	,	PUNCT
ejpam-642	285	10	g.	g.	PROPN
ejpam-642	285	11	yan	yan	PROPN
ejpam-642	285	12	/	/	SYM
ejpam-642	285	13	eur	eur	PROPN
ejpam-642	285	14	.	.	PUNCT
ejpam-642	286	1	j.	j.	PROPN
ejpam-642	286	2	pure	pure	PROPN
ejpam-642	286	3	appl	appl	PROPN
ejpam-642	286	4	.	.	PROPN
ejpam-642	286	5	math	math	PROPN
ejpam-642	286	6	,	,	PUNCT
ejpam-642	286	7	3	3	NUM
ejpam-642	286	8	(	(	PUNCT
ejpam-642	286	9	2010	2010	NUM
ejpam-642	286	10	)	)	PUNCT
ejpam-642	286	11	,	,	PUNCT
ejpam-642	286	12	141	141	NUM
ejpam-642	286	13	-	-	SYM
ejpam-642	286	14	162	162	NUM
ejpam-642	286	15	151	151	NUM
ejpam-642	286	16	•	•	NOUN
ejpam-642	286	17	step	step	NOUN
ejpam-642	286	18	2	2	NUM
ejpam-642	286	19	:	:	PUNCT
ejpam-642	286	20	proceed	proceed	VERB
ejpam-642	286	21	on	on	ADP
ejpam-642	286	22	the	the	DET
ejpam-642	286	23	pathwaym00	pathwaym00	NOUN
ejpam-642	286	24	⊂m10	⊂m10	PROPN
ejpam-642	286	25	⊂m11	⊂m11	NOUN
ejpam-642	286	26	,	,	PUNCT
ejpam-642	286	27	and	and	CCONJ
ejpam-642	286	28	compare	compare	VERB
ejpam-642	286	29	two	two	NUM
ejpam-642	286	30	adjacent	adjacent	ADJ
ejpam-642	286	31	models	model	NOUN
ejpam-642	286	32	successively	successively	ADV
ejpam-642	286	33	using	use	VERB
ejpam-642	286	34	a	a	DET
ejpam-642	286	35	prespecified	prespecifie	VERB
ejpam-642	286	36	significance	significance	NOUN
ejpam-642	286	37	level	level	NOUN
ejpam-642	286	38	α	α	NOUN
ejpam-642	286	39	.	.	PUNCT
ejpam-642	287	1	this	this	DET
ejpam-642	287	2	pathway	pathway	NOUN
ejpam-642	287	3	considers	consider	VERB
ejpam-642	287	4	the	the	DET
ejpam-642	287	5	model	model	NOUN
ejpam-642	287	6	with	with	ADP
ejpam-642	287	7	higher	high	ADJ
ejpam-642	287	8	order	order	NOUN
ejpam-642	287	9	interactions	interaction	NOUN
ejpam-642	287	10	for	for	ADP
ejpam-642	287	11	u	u	NOUN
ejpam-642	287	12	first	first	ADJ
ejpam-642	287	13	.	.	PUNCT
ejpam-642	288	1	–	–	PUNCT
ejpam-642	288	2	if	if	SCONJ
ejpam-642	288	3	the	the	DET
ejpam-642	288	4	series	series	NOUN
ejpam-642	288	5	testing	testing	NOUN
ejpam-642	288	6	stops	stop	VERB
ejpam-642	288	7	atm00	atm00	ADV
ejpam-642	288	8	,	,	PUNCT
ejpam-642	288	9	then	then	ADV
ejpam-642	288	10	go	go	VERB
ejpam-642	288	11	to	to	PART
ejpam-642	288	12	step	step	VERB
ejpam-642	288	13	3	3	NUM
ejpam-642	288	14	.	.	PUNCT
ejpam-642	289	1	–	–	PUNCT
ejpam-642	289	2	if	if	SCONJ
ejpam-642	289	3	the	the	DET
ejpam-642	289	4	series	series	NOUN
ejpam-642	289	5	testing	testing	NOUN
ejpam-642	289	6	stops	stop	VERB
ejpam-642	289	7	atm10	atm10	PROPN
ejpam-642	289	8	,	,	PUNCT
ejpam-642	289	9	then	then	ADV
ejpam-642	289	10	the	the	DET
ejpam-642	289	11	final	final	ADJ
ejpam-642	289	12	model	model	NOUN
ejpam-642	289	13	ism10	ism10	PROPN
ejpam-642	289	14	.	.	PUNCT
ejpam-642	290	1	–	–	PUNCT
ejpam-642	290	2	if	if	SCONJ
ejpam-642	290	3	the	the	DET
ejpam-642	290	4	series	series	NOUN
ejpam-642	290	5	testing	testing	NOUN
ejpam-642	290	6	stops	stop	VERB
ejpam-642	290	7	atm11	atm11	PROPN
ejpam-642	290	8	,	,	PUNCT
ejpam-642	290	9	then	then	ADV
ejpam-642	290	10	go	go	VERB
ejpam-642	290	11	to	to	PART
ejpam-642	290	12	step	step	VERB
ejpam-642	290	13	4	4	NUM
ejpam-642	290	14	.	.	NOUN
ejpam-642	291	1	•	•	NUM
ejpam-642	291	2	step	step	NOUN
ejpam-642	291	3	3	3	NUM
ejpam-642	291	4	:	:	PUNCT
ejpam-642	291	5	proceed	proceed	VERB
ejpam-642	291	6	on	on	ADP
ejpam-642	291	7	the	the	DET
ejpam-642	291	8	alternative	alternative	ADJ
ejpam-642	291	9	pathway	pathway	NOUN
ejpam-642	291	10	m00	m00	PROPN
ejpam-642	291	11	⊂	⊂	PROPN
ejpam-642	291	12	m01	m01	PROPN
ejpam-642	291	13	⊂	⊂	PROPN
ejpam-642	291	14	m11	m11	PROPN
ejpam-642	291	15	,	,	PUNCT
ejpam-642	291	16	and	and	CCONJ
ejpam-642	291	17	compare	compare	VERB
ejpam-642	291	18	two	two	NUM
ejpam-642	291	19	adjacent	adjacent	ADJ
ejpam-642	291	20	models	model	NOUN
ejpam-642	291	21	successively	successively	ADV
ejpam-642	291	22	using	use	VERB
ejpam-642	291	23	the	the	DET
ejpam-642	291	24	significance	significance	NOUN
ejpam-642	291	25	level	level	NOUN
ejpam-642	291	26	α	α	NOUN
ejpam-642	291	27	.	.	PUNCT
ejpam-642	292	1	this	this	DET
ejpam-642	292	2	pathway	pathway	NOUN
ejpam-642	292	3	considers	consider	VERB
ejpam-642	292	4	the	the	DET
ejpam-642	292	5	model	model	NOUN
ejpam-642	292	6	with	with	ADP
ejpam-642	292	7	higher	high	ADJ
ejpam-642	292	8	order	order	NOUN
ejpam-642	292	9	interactions	interaction	NOUN
ejpam-642	292	10	for	for	ADP
ejpam-642	292	11	m	m	NOUN
ejpam-642	292	12	first	first	ADJ
ejpam-642	292	13	.	.	PUNCT
ejpam-642	293	1	–	–	PUNCT
ejpam-642	293	2	if	if	SCONJ
ejpam-642	293	3	the	the	DET
ejpam-642	293	4	series	series	NOUN
ejpam-642	293	5	testing	testing	NOUN
ejpam-642	293	6	stops	stop	VERB
ejpam-642	293	7	atm00	atm00	PROPN
ejpam-642	293	8	,	,	PUNCT
ejpam-642	293	9	then	then	ADV
ejpam-642	293	10	the	the	DET
ejpam-642	293	11	final	final	ADJ
ejpam-642	293	12	model	model	NOUN
ejpam-642	293	13	ism00	ism00	NOUN
ejpam-642	293	14	.	.	PUNCT
ejpam-642	294	1	–	–	PUNCT
ejpam-642	294	2	if	if	SCONJ
ejpam-642	294	3	the	the	DET
ejpam-642	294	4	series	series	NOUN
ejpam-642	294	5	testing	testing	NOUN
ejpam-642	294	6	stops	stop	VERB
ejpam-642	294	7	atm01	atm01	ADV
ejpam-642	294	8	,	,	PUNCT
ejpam-642	294	9	then	then	ADV
ejpam-642	294	10	the	the	DET
ejpam-642	294	11	final	final	ADJ
ejpam-642	294	12	model	model	NOUN
ejpam-642	294	13	ism01	ism01	ADJ
ejpam-642	294	14	.	.	PUNCT
ejpam-642	295	1	–	–	PUNCT
ejpam-642	295	2	if	if	SCONJ
ejpam-642	295	3	the	the	DET
ejpam-642	295	4	series	series	NOUN
ejpam-642	295	5	testing	testing	NOUN
ejpam-642	295	6	stops	stop	VERB
ejpam-642	295	7	atm11	atm11	PROPN
ejpam-642	295	8	,	,	PUNCT
ejpam-642	295	9	then	then	ADV
ejpam-642	295	10	go	go	VERB
ejpam-642	295	11	to	to	PART
ejpam-642	295	12	step	step	VERB
ejpam-642	295	13	4	4	NUM
ejpam-642	295	14	.	.	NOUN
ejpam-642	296	1	•	•	NUM
ejpam-642	296	2	step	step	NOUN
ejpam-642	296	3	4	4	NUM
ejpam-642	296	4	:	:	PUNCT
ejpam-642	296	5	treat	treat	VERB
ejpam-642	296	6	m11	m11	NOUN
ejpam-642	296	7	like	like	ADP
ejpam-642	296	8	m00	m00	NOUN
ejpam-642	296	9	,	,	PUNCT
ejpam-642	296	10	and	and	CCONJ
ejpam-642	296	11	repeat	repeat	NOUN
ejpam-642	296	12	steps	step	NOUN
ejpam-642	296	13	1	1	NUM
ejpam-642	296	14	to	to	PART
ejpam-642	296	15	3	3	NUM
ejpam-642	296	16	along	along	ADP
ejpam-642	296	17	the	the	DET
ejpam-642	296	18	alternatively	alternatively	ADV
ejpam-642	296	19	climbing	climb	VERB
ejpam-642	296	20	pathways	pathway	NOUN
ejpam-642	296	21	fromm11	fromm11	NOUN
ejpam-642	296	22	tom22	tom22	PROPN
ejpam-642	296	23	,	,	PUNCT
ejpam-642	296	24	i.e.	i.e.	X
ejpam-642	296	25	,	,	PUNCT
ejpam-642	296	26	start	start	VERB
ejpam-642	296	27	a	a	DET
ejpam-642	296	28	second	second	ADJ
ejpam-642	296	29	round	round	NOUN
ejpam-642	296	30	of	of	ADP
ejpam-642	296	31	search	search	NOUN
ejpam-642	296	32	from	from	ADP
ejpam-642	296	33	the	the	DET
ejpam-642	296	34	next	next	ADJ
ejpam-642	296	35	diagonal	diagonal	ADJ
ejpam-642	296	36	entry	entry	NOUN
ejpam-642	296	37	.	.	PUNCT
ejpam-642	297	1	•	•	NUM
ejpam-642	297	2	step	step	NOUN
ejpam-642	297	3	5	5	NUM
ejpam-642	297	4	:	:	PUNCT
ejpam-642	297	5	continue	continue	VERB
ejpam-642	297	6	the	the	DET
ejpam-642	297	7	iteration	iteration	NOUN
ejpam-642	297	8	until	until	SCONJ
ejpam-642	297	9	the	the	DET
ejpam-642	297	10	series	series	NOUN
ejpam-642	297	11	testing	testing	NOUN
ejpam-642	297	12	stops	stop	VERB
ejpam-642	297	13	.	.	PUNCT
ejpam-642	298	1	the	the	DET
ejpam-642	298	2	above	above	ADJ
ejpam-642	298	3	search	search	NOUN
ejpam-642	298	4	on	on	ADP
ejpam-642	298	5	alternatively	alternatively	ADV
ejpam-642	298	6	climbing	climb	VERB
ejpam-642	298	7	pathways	pathway	NOUN
ejpam-642	298	8	guarantees	guarantee	VERB
ejpam-642	298	9	the	the	DET
ejpam-642	298	10	nesting	nesting	NOUN
ejpam-642	298	11	between	between	ADP
ejpam-642	298	12	two	two	NUM
ejpam-642	298	13	loglinear	loglinear	NOUN
ejpam-642	298	14	mixture	mixture	NOUN
ejpam-642	298	15	models	model	NOUN
ejpam-642	298	16	under	under	ADP
ejpam-642	298	17	comparison	comparison	NOUN
ejpam-642	298	18	.	.	PUNCT
ejpam-642	299	1	we	we	PRON
ejpam-642	299	2	place	place	VERB
ejpam-642	299	3	priority	priority	NOUN
ejpam-642	299	4	on	on	ADP
ejpam-642	299	5	the	the	DET
ejpam-642	299	6	pathway	pathway	NOUN
ejpam-642	299	7	where	where	SCONJ
ejpam-642	299	8	the	the	DET
ejpam-642	299	9	model	model	NOUN
ejpam-642	299	10	for	for	ADP
ejpam-642	299	11	u	u	NOUN
ejpam-642	299	12	is	be	AUX
ejpam-642	299	13	added	add	VERB
ejpam-642	299	14	higher	high	ADJ
ejpam-642	299	15	order	order	NOUN
ejpam-642	299	16	interactions	interaction	NOUN
ejpam-642	299	17	(	(	PUNCT
ejpam-642	299	18	the	the	DET
ejpam-642	299	19	route	route	NOUN
ejpam-642	299	20	denoted	denote	VERB
ejpam-642	299	21	by	by	ADP
ejpam-642	299	22	“	"	PUNCT
ejpam-642	299	23	→	→	VERB
ejpam-642	299	24	”	"	PUNCT
ejpam-642	299	25	)	)	PUNCT
ejpam-642	299	26	first	first	ADV
ejpam-642	299	27	,	,	PUNCT
ejpam-642	299	28	because	because	SCONJ
ejpam-642	299	29	in	in	ADP
ejpam-642	299	30	record	record	NOUN
ejpam-642	299	31	linkage	linkage	NOUN
ejpam-642	299	32	,	,	PUNCT
ejpam-642	299	33	the	the	DET
ejpam-642	299	34	unmatched	unmatched	ADJ
ejpam-642	299	35	pairs	pair	NOUN
ejpam-642	299	36	are	be	AUX
ejpam-642	299	37	the	the	DET
ejpam-642	299	38	majority	majority	NOUN
ejpam-642	299	39	.	.	PUNCT
ejpam-642	300	1	if	if	SCONJ
ejpam-642	300	2	this	this	PRON
ejpam-642	300	3	fails	fail	VERB
ejpam-642	300	4	,	,	PUNCT
ejpam-642	300	5	then	then	ADV
ejpam-642	300	6	we	we	PRON
ejpam-642	300	7	consider	consider	VERB
ejpam-642	300	8	the	the	DET
ejpam-642	300	9	other	other	ADJ
ejpam-642	300	10	pathway	pathway	NOUN
ejpam-642	300	11	where	where	SCONJ
ejpam-642	300	12	the	the	DET
ejpam-642	300	13	model	model	NOUN
ejpam-642	300	14	for	for	ADP
ejpam-642	300	15	m	m	PROPN
ejpam-642	300	16	is	be	AUX
ejpam-642	300	17	added	add	VERB
ejpam-642	300	18	higher	high	ADJ
ejpam-642	300	19	order	order	NOUN
ejpam-642	300	20	interactions	interaction	NOUN
ejpam-642	300	21	(	(	PUNCT
ejpam-642	300	22	the	the	DET
ejpam-642	300	23	route	route	NOUN
ejpam-642	300	24	denoted	denote	VERB
ejpam-642	300	25	by	by	ADP
ejpam-642	300	26	“	"	PUNCT
ejpam-642	300	27	⇒	⇒	NOUN
ejpam-642	300	28	”	"	PUNCT
ejpam-642	300	29	)	)	PUNCT
ejpam-642	300	30	.	.	PUNCT
ejpam-642	301	1	note	note	VERB
ejpam-642	301	2	that	that	SCONJ
ejpam-642	301	3	for	for	ADP
ejpam-642	301	4	step	step	NOUN
ejpam-642	301	5	1	1	NUM
ejpam-642	301	6	,	,	PUNCT
ejpam-642	301	7	if	if	SCONJ
ejpam-642	301	8	the	the	DET
ejpam-642	301	9	contingency	contingency	NOUN
ejpam-642	301	10	table	table	NOUN
ejpam-642	301	11	degenerates	degenerate	NOUN
ejpam-642	301	12	to	to	ADP
ejpam-642	301	13	the	the	DET
ejpam-642	301	14	case	case	NOUN
ejpam-642	301	15	where	where	SCONJ
ejpam-642	301	16	all	all	DET
ejpam-642	301	17	factors	factor	NOUN
ejpam-642	301	18	are	be	AUX
ejpam-642	301	19	binary	binary	ADJ
ejpam-642	301	20	,	,	PUNCT
ejpam-642	301	21	the	the	DET
ejpam-642	301	22	computation	computation	NOUN
ejpam-642	301	23	will	will	AUX
ejpam-642	301	24	become	become	VERB
ejpam-642	301	25	the	the	DET
ejpam-642	301	26	jaro	jaro	NOUN
ejpam-642	301	27	’s	’s	PART
ejpam-642	301	28	method	method	NOUN
ejpam-642	301	29	.	.	PUNCT
ejpam-642	302	1	the	the	DET
ejpam-642	302	2	simplest	simple	ADJ
ejpam-642	302	3	loglinear	loglinear	NOUN
ejpam-642	302	4	mixture	mixture	NOUN
ejpam-642	302	5	model	model	NOUN
ejpam-642	302	6	m00	m00	PROPN
ejpam-642	302	7	captures	capture	VERB
ejpam-642	302	8	the	the	DET
ejpam-642	302	9	main	main	ADJ
ejpam-642	302	10	effects	effect	NOUN
ejpam-642	302	11	explained	explain	VERB
ejpam-642	302	12	by	by	ADP
ejpam-642	302	13	factors	factor	NOUN
ejpam-642	302	14	alone	alone	ADV
ejpam-642	302	15	.	.	PUNCT
ejpam-642	303	1	note	note	VERB
ejpam-642	303	2	that	that	SCONJ
ejpam-642	303	3	the	the	DET
ejpam-642	303	4	attributes	attribute	NOUN
ejpam-642	303	5	chosen	choose	VERB
ejpam-642	303	6	by	by	ADP
ejpam-642	303	7	investigators	investigator	NOUN
ejpam-642	303	8	are	be	AUX
ejpam-642	303	9	typically	typically	ADV
ejpam-642	303	10	important	important	ADJ
ejpam-642	303	11	variables	variable	NOUN
ejpam-642	303	12	in	in	ADP
ejpam-642	303	13	record	record	NOUN
ejpam-642	303	14	matching	matching	NOUN
ejpam-642	303	15	according	accord	VERB
ejpam-642	303	16	to	to	ADP
ejpam-642	303	17	background	background	NOUN
ejpam-642	303	18	knowledge	knowledge	NOUN
ejpam-642	303	19	.	.	PUNCT
ejpam-642	304	1	thus	thus	ADV
ejpam-642	304	2	,	,	PUNCT
ejpam-642	304	3	their	their	PRON
ejpam-642	304	4	main	main	ADJ
ejpam-642	304	5	effects	effect	NOUN
ejpam-642	304	6	practically	practically	ADV
ejpam-642	304	7	exist	exist	VERB
ejpam-642	304	8	.	.	PUNCT
ejpam-642	305	1	on	on	ADP
ejpam-642	305	2	the	the	DET
ejpam-642	305	3	other	other	ADJ
ejpam-642	305	4	hand	hand	NOUN
ejpam-642	305	5	,	,	PUNCT
ejpam-642	305	6	we	we	PRON
ejpam-642	305	7	may	may	AUX
ejpam-642	305	8	not	not	PART
ejpam-642	305	9	know	know	VERB
ejpam-642	305	10	the	the	DET
ejpam-642	305	11	joint	joint	ADJ
ejpam-642	305	12	effects	effect	NOUN
ejpam-642	305	13	of	of	ADP
ejpam-642	305	14	attributes	attribute	NOUN
ejpam-642	305	15	in	in	ADP
ejpam-642	305	16	advance	advance	NOUN
ejpam-642	305	17	.	.	PUNCT
ejpam-642	306	1	thus	thus	ADV
ejpam-642	306	2	,	,	PUNCT
ejpam-642	306	3	it	it	PRON
ejpam-642	306	4	is	be	AUX
ejpam-642	306	5	safe	safe	ADJ
ejpam-642	306	6	and	and	CCONJ
ejpam-642	306	7	convenient	convenient	ADJ
ejpam-642	306	8	to	to	PART
ejpam-642	306	9	start	start	VERB
ejpam-642	306	10	from	from	ADP
ejpam-642	306	11	the	the	DET
ejpam-642	306	12	simplest	simple	ADJ
ejpam-642	306	13	mixture	mixture	NOUN
ejpam-642	306	14	modelm00	modelm00	NOUN
ejpam-642	306	15	.	.	PUNCT
ejpam-642	307	1	however	however	ADV
ejpam-642	307	2	,	,	PUNCT
ejpam-642	307	3	if	if	SCONJ
ejpam-642	307	4	the	the	DET
ejpam-642	307	5	joint	joint	ADJ
ejpam-642	307	6	effects	effect	NOUN
ejpam-642	307	7	have	have	AUX
ejpam-642	307	8	been	be	AUX
ejpam-642	307	9	known	know	VERB
ejpam-642	307	10	to	to	PART
ejpam-642	307	11	be	be	AUX
ejpam-642	307	12	strong	strong	ADJ
ejpam-642	307	13	by	by	ADP
ejpam-642	307	14	the	the	DET
ejpam-642	307	15	background	background	NOUN
ejpam-642	307	16	knowledge	knowledge	NOUN
ejpam-642	307	17	,	,	PUNCT
ejpam-642	307	18	we	we	PRON
ejpam-642	307	19	can	can	AUX
ejpam-642	307	20	start	start	VERB
ejpam-642	307	21	from	from	ADP
ejpam-642	307	22	a	a	DET
ejpam-642	307	23	particular	particular	ADJ
ejpam-642	307	24	mixture	mixture	NOUN
ejpam-642	307	25	model	model	NOUN
ejpam-642	307	26	which	which	PRON
ejpam-642	307	27	may	may	AUX
ejpam-642	307	28	not	not	PART
ejpam-642	307	29	be	be	AUX
ejpam-642	307	30	a	a	DET
ejpam-642	307	31	diagonal	diagonal	ADJ
ejpam-642	307	32	entry	entry	NOUN
ejpam-642	307	33	in	in	ADP
ejpam-642	307	34	(	(	PUNCT
ejpam-642	307	35	11	11	NUM
ejpam-642	307	36	)	)	PUNCT
ejpam-642	307	37	.	.	PUNCT
ejpam-642	308	1	a	a	DET
ejpam-642	308	2	modification	modification	NOUN
ejpam-642	308	3	of	of	ADP
ejpam-642	308	4	the	the	DET
ejpam-642	308	5	model	model	NOUN
ejpam-642	308	6	search	search	NOUN
ejpam-642	308	7	is	be	AUX
ejpam-642	308	8	then	then	ADV
ejpam-642	308	9	that	that	SCONJ
ejpam-642	308	10	we	we	PRON
ejpam-642	308	11	proceed	proceed	VERB
ejpam-642	308	12	with	with	ADP
ejpam-642	308	13	the	the	DET
ejpam-642	308	14	testings	testing	NOUN
ejpam-642	308	15	on	on	ADP
ejpam-642	308	16	an	an	DET
ejpam-642	308	17	upward	upward	ADJ
ejpam-642	308	18	or	or	CCONJ
ejpam-642	308	19	a	a	DET
ejpam-642	308	20	rightward	rightward	ADJ
ejpam-642	308	21	pathway	pathway	NOUN
ejpam-642	308	22	to	to	ADP
ejpam-642	308	23	the	the	DET
ejpam-642	308	24	model	model	NOUN
ejpam-642	308	25	in	in	ADP
ejpam-642	308	26	the	the	DET
ejpam-642	308	27	closest	close	ADJ
ejpam-642	308	28	diagonal	diagonal	ADJ
ejpam-642	308	29	entry	entry	NOUN
ejpam-642	308	30	,	,	PUNCT
ejpam-642	308	31	then	then	ADV
ejpam-642	308	32	continue	continue	VERB
ejpam-642	308	33	the	the	DET
ejpam-642	308	34	search	search	NOUN
ejpam-642	308	35	using	use	VERB
ejpam-642	308	36	the	the	DET
ejpam-642	308	37	above	above	ADJ
ejpam-642	308	38	strategy	strategy	NOUN
ejpam-642	308	39	.	.	PUNCT
ejpam-642	309	1	note	note	VERB
ejpam-642	309	2	that	that	SCONJ
ejpam-642	309	3	the	the	DET
ejpam-642	309	4	backward	backward	ADJ
ejpam-642	309	5	selection	selection	NOUN
ejpam-642	309	6	could	could	AUX
ejpam-642	309	7	be	be	AUX
ejpam-642	309	8	feasible	feasible	ADJ
ejpam-642	309	9	.	.	PUNCT
ejpam-642	310	1	however	however	ADV
ejpam-642	310	2	,	,	PUNCT
ejpam-642	310	3	it	it	PRON
ejpam-642	310	4	starts	start	VERB
ejpam-642	310	5	from	from	ADP
ejpam-642	310	6	the	the	DET
ejpam-642	310	7	most	most	ADV
ejpam-642	310	8	complicated	complicated	ADJ
ejpam-642	310	9	model	model	NOUN
ejpam-642	310	10	with	with	ADP
ejpam-642	310	11	all	all	DET
ejpam-642	310	12	possible	possible	ADJ
ejpam-642	310	13	interactions	interaction	NOUN
ejpam-642	310	14	,	,	PUNCT
ejpam-642	310	15	and	and	CCONJ
ejpam-642	310	16	thus	thus	ADV
ejpam-642	310	17	,	,	PUNCT
ejpam-642	310	18	the	the	DET
ejpam-642	310	19	computational	computational	ADJ
ejpam-642	310	20	cost	cost	NOUN
ejpam-642	310	21	is	be	AUX
ejpam-642	310	22	typically	typically	ADV
ejpam-642	310	23	large	large	ADJ
ejpam-642	310	24	.	.	PUNCT
ejpam-642	311	1	in	in	ADP
ejpam-642	311	2	record	record	NOUN
ejpam-642	311	3	linkage	linkage	NOUN
ejpam-642	311	4	,	,	PUNCT
ejpam-642	311	5	the	the	DET
ejpam-642	311	6	number	number	NOUN
ejpam-642	311	7	of	of	ADP
ejpam-642	311	8	matched	match	VERB
ejpam-642	311	9	pairs	pair	NOUN
ejpam-642	311	10	is	be	AUX
ejpam-642	311	11	extremely	extremely	ADV
ejpam-642	311	12	small	small	ADJ
ejpam-642	311	13	relative	relative	ADJ
ejpam-642	311	14	to	to	ADP
ejpam-642	311	15	the	the	DET
ejpam-642	311	16	number	number	NOUN
ejpam-642	311	17	of	of	ADP
ejpam-642	311	18	unmatched	unmatched	ADJ
ejpam-642	311	19	pairs	pair	NOUN
ejpam-642	311	20	.	.	PUNCT
ejpam-642	312	1	so	so	ADV
ejpam-642	312	2	usually	usually	ADV
ejpam-642	312	3	the	the	DET
ejpam-642	312	4	model	model	NOUN
ejpam-642	312	5	for	for	ADP
ejpam-642	312	6	m	m	PROPN
ejpam-642	312	7	has	have	VERB
ejpam-642	312	8	lower	low	ADJ
ejpam-642	312	9	order	order	NOUN
ejpam-642	312	10	interactions	interaction	NOUN
ejpam-642	312	11	than	than	ADP
ejpam-642	312	12	that	that	PRON
ejpam-642	312	13	for	for	ADP
ejpam-642	312	14	u	u	PRON
ejpam-642	312	15	.	.	PUNCT
ejpam-642	313	1	that	that	PRON
ejpam-642	313	2	is	be	AUX
ejpam-642	313	3	why	why	SCONJ
ejpam-642	313	4	we	we	PRON
ejpam-642	313	5	consider	consider	VERB
ejpam-642	313	6	to	to	PART
ejpam-642	313	7	enlarge	enlarge	VERB
ejpam-642	313	8	the	the	DET
ejpam-642	313	9	component	component	NOUN
ejpam-642	313	10	for	for	ADP
ejpam-642	313	11	u	u	NOUN
ejpam-642	313	12	first	first	ADV
ejpam-642	313	13	.	.	PUNCT
ejpam-642	314	1	the	the	DET
ejpam-642	314	2	restriction	restriction	NOUN
ejpam-642	314	3	of	of	ADP
ejpam-642	314	4	searching	search	VERB
ejpam-642	314	5	r.	r.	PROPN
ejpam-642	314	6	zhu	zhu	PROPN
ejpam-642	314	7	,	,	PUNCT
ejpam-642	314	8	j.	j.	PROPN
ejpam-642	314	9	zhang	zhang	PROPN
ejpam-642	314	10	,	,	PUNCT
ejpam-642	314	11	d.	d.	PROPN
ejpam-642	314	12	zhang	zhang	PROPN
ejpam-642	314	13	,	,	PUNCT
ejpam-642	314	14	g.	g.	PROPN
ejpam-642	314	15	yan	yan	PROPN
ejpam-642	314	16	/	/	SYM
ejpam-642	314	17	eur	eur	PROPN
ejpam-642	314	18	.	.	PUNCT
ejpam-642	315	1	j.	j.	PROPN
ejpam-642	315	2	pure	pure	PROPN
ejpam-642	315	3	appl	appl	PROPN
ejpam-642	315	4	.	.	PROPN
ejpam-642	315	5	math	math	PROPN
ejpam-642	315	6	,	,	PUNCT
ejpam-642	315	7	3	3	NUM
ejpam-642	315	8	(	(	PUNCT
ejpam-642	315	9	2010	2010	NUM
ejpam-642	315	10	)	)	PUNCT
ejpam-642	315	11	,	,	PUNCT
ejpam-642	315	12	141	141	NUM
ejpam-642	315	13	-	-	SYM
ejpam-642	315	14	162	162	NUM
ejpam-642	315	15	152	152	NUM
ejpam-642	315	16	along	along	ADP
ejpam-642	315	17	the	the	DET
ejpam-642	315	18	alternatively	alternatively	ADV
ejpam-642	315	19	climbing	climb	VERB
ejpam-642	315	20	pathways	pathway	NOUN
ejpam-642	315	21	prevents	prevent	VERB
ejpam-642	315	22	over	over	ADP
ejpam-642	315	23	-	-	PUNCT
ejpam-642	315	24	extension	extension	NOUN
ejpam-642	315	25	of	of	ADP
ejpam-642	315	26	model	model	NOUN
ejpam-642	315	27	for	for	ADP
ejpam-642	315	28	either	either	PRON
ejpam-642	315	29	component	component	NOUN
ejpam-642	315	30	u	u	NOUN
ejpam-642	315	31	or	or	CCONJ
ejpam-642	315	32	m	m	PROPN
ejpam-642	315	33	.	.	PUNCT
ejpam-642	316	1	this	this	DET
ejpam-642	316	2	conservative	conservative	ADJ
ejpam-642	316	3	procedure	procedure	NOUN
ejpam-642	316	4	is	be	AUX
ejpam-642	316	5	based	base	VERB
ejpam-642	316	6	on	on	ADP
ejpam-642	316	7	the	the	DET
ejpam-642	316	8	reasoning	reasoning	NOUN
ejpam-642	316	9	that	that	SCONJ
ejpam-642	316	10	interactions	interaction	NOUN
ejpam-642	316	11	among	among	ADP
ejpam-642	316	12	attributes	attribute	NOUN
ejpam-642	316	13	usually	usually	ADV
ejpam-642	316	14	affect	affect	VERB
ejpam-642	316	15	both	both	DET
ejpam-642	316	16	subpopulations	subpopulation	NOUN
ejpam-642	316	17	u	u	NOUN
ejpam-642	316	18	and	and	CCONJ
ejpam-642	316	19	m	m	PROPN
ejpam-642	316	20	.	.	PUNCT
ejpam-642	317	1	thus	thus	ADV
ejpam-642	317	2	,	,	PUNCT
ejpam-642	317	3	we	we	PRON
ejpam-642	317	4	do	do	AUX
ejpam-642	317	5	not	not	PART
ejpam-642	317	6	wish	wish	VERB
ejpam-642	317	7	the	the	DET
ejpam-642	317	8	model	model	NOUN
ejpam-642	317	9	for	for	ADP
ejpam-642	317	10	u	u	PROPN
ejpam-642	317	11	has	have	VERB
ejpam-642	317	12	higher	high	ADJ
ejpam-642	317	13	order	order	NOUN
ejpam-642	317	14	interactions	interaction	NOUN
ejpam-642	317	15	while	while	SCONJ
ejpam-642	317	16	the	the	DET
ejpam-642	317	17	model	model	NOUN
ejpam-642	317	18	for	for	ADP
ejpam-642	317	19	m	m	PROPN
ejpam-642	317	20	does	do	VERB
ejpam-642	317	21	not	not	PART
ejpam-642	317	22	.	.	PUNCT
ejpam-642	318	1	if	if	SCONJ
ejpam-642	318	2	this	this	DET
ejpam-642	318	3	reasoning	reasoning	NOUN
ejpam-642	318	4	is	be	AUX
ejpam-642	318	5	not	not	PART
ejpam-642	318	6	appropriate	appropriate	ADJ
ejpam-642	318	7	,	,	PUNCT
ejpam-642	318	8	the	the	DET
ejpam-642	318	9	restriction	restriction	NOUN
ejpam-642	318	10	can	can	AUX
ejpam-642	318	11	be	be	AUX
ejpam-642	318	12	relaxed	relax	VERB
ejpam-642	318	13	,	,	PUNCT
ejpam-642	318	14	say	say	VERB
ejpam-642	318	15	we	we	PRON
ejpam-642	318	16	can	can	AUX
ejpam-642	318	17	further	far	ADV
ejpam-642	318	18	extend	extend	VERB
ejpam-642	318	19	the	the	DET
ejpam-642	318	20	model	model	NOUN
ejpam-642	318	21	for	for	ADP
ejpam-642	318	22	u	u	PRON
ejpam-642	318	23	when	when	SCONJ
ejpam-642	318	24	a	a	DET
ejpam-642	318	25	final	final	ADJ
ejpam-642	318	26	model	model	NOUN
ejpam-642	318	27	is	be	AUX
ejpam-642	318	28	m(i+1)i	m(i+1)i	PROPN
ejpam-642	318	29	.	.	PUNCT
ejpam-642	319	1	however	however	ADV
ejpam-642	319	2	,	,	PUNCT
ejpam-642	319	3	our	our	PRON
ejpam-642	319	4	simulation	simulation	NOUN
ejpam-642	319	5	studies	study	NOUN
ejpam-642	319	6	show	show	VERB
ejpam-642	319	7	that	that	SCONJ
ejpam-642	319	8	it	it	PRON
ejpam-642	319	9	is	be	AUX
ejpam-642	319	10	hard	hard	ADJ
ejpam-642	319	11	to	to	PART
ejpam-642	319	12	capture	capture	VERB
ejpam-642	319	13	higher	high	ADJ
ejpam-642	319	14	order	order	NOUN
ejpam-642	319	15	interactions	interaction	NOUN
ejpam-642	319	16	from	from	ADP
ejpam-642	319	17	data	datum	NOUN
ejpam-642	319	18	even	even	ADV
ejpam-642	319	19	if	if	SCONJ
ejpam-642	319	20	they	they	PRON
ejpam-642	319	21	are	be	AUX
ejpam-642	319	22	truly	truly	ADV
ejpam-642	319	23	generated	generate	VERB
ejpam-642	319	24	from	from	ADP
ejpam-642	319	25	a	a	DET
ejpam-642	319	26	model	model	NOUN
ejpam-642	319	27	with	with	ADP
ejpam-642	319	28	higher	high	ADJ
ejpam-642	319	29	order	order	NOUN
ejpam-642	319	30	interactions	interaction	NOUN
ejpam-642	319	31	.	.	PUNCT
ejpam-642	320	1	therefore	therefore	ADV
ejpam-642	320	2	,	,	PUNCT
ejpam-642	320	3	from	from	ADP
ejpam-642	320	4	the	the	DET
ejpam-642	320	5	practical	practical	ADJ
ejpam-642	320	6	viewpoint	viewpoint	NOUN
ejpam-642	320	7	,	,	PUNCT
ejpam-642	320	8	we	we	PRON
ejpam-642	320	9	suggest	suggest	VERB
ejpam-642	320	10	not	not	PART
ejpam-642	320	11	to	to	PART
ejpam-642	320	12	seek	seek	VERB
ejpam-642	320	13	very	very	ADV
ejpam-642	320	14	higher	high	ADJ
ejpam-642	320	15	order	order	NOUN
ejpam-642	320	16	interactions	interaction	NOUN
ejpam-642	320	17	in	in	ADP
ejpam-642	320	18	record	record	NOUN
ejpam-642	320	19	linkage	linkage	NOUN
ejpam-642	320	20	.	.	PUNCT
ejpam-642	321	1	the	the	DET
ejpam-642	321	2	final	final	ADJ
ejpam-642	321	3	model	model	NOUN
ejpam-642	321	4	obtained	obtain	VERB
ejpam-642	321	5	from	from	ADP
ejpam-642	321	6	the	the	DET
ejpam-642	321	7	proposed	propose	VERB
ejpam-642	321	8	strategy	strategy	NOUN
ejpam-642	321	9	will	will	AUX
ejpam-642	321	10	include	include	VERB
ejpam-642	321	11	all	all	DET
ejpam-642	321	12	interactions	interaction	NOUN
ejpam-642	321	13	of	of	ADP
ejpam-642	321	14	the	the	DET
ejpam-642	321	15	same	same	ADJ
ejpam-642	321	16	order	order	NOUN
ejpam-642	321	17	,	,	PUNCT
ejpam-642	321	18	say	say	VERB
ejpam-642	321	19	all	all	DET
ejpam-642	321	20	three	three	NUM
ejpam-642	321	21	-	-	PUNCT
ejpam-642	321	22	factor	factor	NOUN
ejpam-642	321	23	-	-	PUNCT
ejpam-642	321	24	interactions	interaction	NOUN
ejpam-642	321	25	or	or	CCONJ
ejpam-642	321	26	third	third	ADJ
ejpam-642	321	27	order	order	NOUN
ejpam-642	321	28	interactions	interaction	NOUN
ejpam-642	321	29	.	.	PUNCT
ejpam-642	322	1	it	it	PRON
ejpam-642	322	2	may	may	AUX
ejpam-642	322	3	not	not	PART
ejpam-642	322	4	be	be	AUX
ejpam-642	322	5	necessary	necessary	ADJ
ejpam-642	322	6	to	to	PART
ejpam-642	322	7	do	do	VERB
ejpam-642	322	8	so	so	ADV
ejpam-642	322	9	because	because	SCONJ
ejpam-642	322	10	some	some	PRON
ejpam-642	322	11	of	of	ADP
ejpam-642	322	12	them	they	PRON
ejpam-642	322	13	could	could	AUX
ejpam-642	322	14	be	be	AUX
ejpam-642	322	15	very	very	ADV
ejpam-642	322	16	weak	weak	ADJ
ejpam-642	322	17	.	.	PUNCT
ejpam-642	323	1	thus	thus	ADV
ejpam-642	323	2	,	,	PUNCT
ejpam-642	323	3	we	we	PRON
ejpam-642	323	4	can	can	AUX
ejpam-642	323	5	apply	apply	VERB
ejpam-642	323	6	testing	testing	NOUN
ejpam-642	323	7	in	in	ADP
ejpam-642	323	8	(	(	PUNCT
ejpam-642	323	9	10	10	NUM
ejpam-642	323	10	)	)	PUNCT
ejpam-642	323	11	again	again	ADV
ejpam-642	323	12	to	to	PART
ejpam-642	323	13	prune	prune	VERB
ejpam-642	323	14	unnecessary	unnecessary	ADJ
ejpam-642	323	15	interactions	interaction	NOUN
ejpam-642	323	16	for	for	ADP
ejpam-642	323	17	either	either	CCONJ
ejpam-642	323	18	u	u	PROPN
ejpam-642	323	19	or	or	CCONJ
ejpam-642	323	20	m	m	VERB
ejpam-642	323	21	in	in	ADP
ejpam-642	323	22	the	the	DET
ejpam-642	323	23	final	final	ADJ
ejpam-642	323	24	mixture	mixture	NOUN
ejpam-642	323	25	model	model	NOUN
ejpam-642	323	26	.	.	PUNCT
ejpam-642	324	1	the	the	DET
ejpam-642	324	2	choice	choice	NOUN
ejpam-642	324	3	of	of	ADP
ejpam-642	324	4	initial	initial	ADJ
ejpam-642	324	5	values	value	NOUN
ejpam-642	324	6	of	of	ADP
ejpam-642	324	7	parameters	parameter	NOUN
ejpam-642	324	8	for	for	ADP
ejpam-642	324	9	the	the	DET
ejpam-642	324	10	em	em	PROPN
ejpam-642	324	11	algorithm	algorithm	NOUN
ejpam-642	324	12	is	be	AUX
ejpam-642	324	13	important	important	ADJ
ejpam-642	324	14	for	for	ADP
ejpam-642	324	15	the	the	DET
ejpam-642	324	16	final	final	ADJ
ejpam-642	324	17	solution	solution	NOUN
ejpam-642	324	18	.	.	PUNCT
ejpam-642	325	1	if	if	SCONJ
ejpam-642	325	2	the	the	DET
ejpam-642	325	3	contingency	contingency	NOUN
ejpam-642	325	4	table	table	NOUN
ejpam-642	325	5	is	be	AUX
ejpam-642	325	6	the	the	DET
ejpam-642	325	7	one	one	NUM
ejpam-642	325	8	with	with	ADP
ejpam-642	325	9	binary	binary	ADJ
ejpam-642	325	10	factors	factor	NOUN
ejpam-642	325	11	,	,	PUNCT
ejpam-642	325	12	then	then	ADV
ejpam-642	325	13	we	we	PRON
ejpam-642	325	14	can	can	AUX
ejpam-642	325	15	use	use	VERB
ejpam-642	325	16	the	the	DET
ejpam-642	325	17	jaro	jaro	NOUN
ejpam-642	325	18	’s	’s	PART
ejpam-642	325	19	method	method	NOUN
ejpam-642	325	20	to	to	PART
ejpam-642	325	21	determine	determine	VERB
ejpam-642	325	22	the	the	DET
ejpam-642	325	23	initial	initial	ADJ
ejpam-642	325	24	values	value	NOUN
ejpam-642	325	25	.	.	PUNCT
ejpam-642	326	1	note	note	VERB
ejpam-642	326	2	that	that	SCONJ
ejpam-642	326	3	the	the	DET
ejpam-642	326	4	maximum	maximum	ADJ
ejpam-642	326	5	number	number	NOUN
ejpam-642	326	6	of	of	ADP
ejpam-642	326	7	matched	match	VERB
ejpam-642	326	8	pairs	pair	NOUN
ejpam-642	326	9	is	be	AUX
ejpam-642	326	10	min(a	min(a	PROPN
ejpam-642	326	11	,	,	PUNCT
ejpam-642	326	12	b	b	NOUN
ejpam-642	326	13	)	)	PUNCT
ejpam-642	326	14	.	.	PUNCT
ejpam-642	327	1	in	in	ADP
ejpam-642	327	2	a	a	DET
ejpam-642	327	3	general	general	ADJ
ejpam-642	327	4	case	case	NOUN
ejpam-642	327	5	,	,	PUNCT
ejpam-642	327	6	we	we	PRON
ejpam-642	327	7	propose	propose	VERB
ejpam-642	327	8	the	the	DET
ejpam-642	327	9	following	follow	VERB
ejpam-642	327	10	data	data	NOUN
ejpam-642	327	11	-	-	PUNCT
ejpam-642	327	12	driven	drive	VERB
ejpam-642	327	13	method	method	NOUN
ejpam-642	327	14	named	name	VERB
ejpam-642	327	15	as	as	ADP
ejpam-642	327	16	the	the	DET
ejpam-642	327	17	j	j	PROPN
ejpam-642	327	18	method	method	NOUN
ejpam-642	327	19	for	for	ADP
ejpam-642	327	20	the	the	DET
ejpam-642	327	21	choice	choice	NOUN
ejpam-642	327	22	of	of	ADP
ejpam-642	327	23	initial	initial	ADJ
ejpam-642	327	24	values	value	NOUN
ejpam-642	327	25	regarding	regard	VERB
ejpam-642	327	26	parameters	parameter	NOUN
ejpam-642	327	27	π	π	PROPN
ejpam-642	327	28	,	,	PUNCT
ejpam-642	327	29	βu	βu	PUNCT
ejpam-642	327	30	and	and	CCONJ
ejpam-642	327	31	βm	βm	VERB
ejpam-642	327	32	:	:	PUNCT
ejpam-642	327	33	•	•	ADP
ejpam-642	327	34	the	the	DET
ejpam-642	327	35	initial	initial	ADJ
ejpam-642	327	36	value	value	NOUN
ejpam-642	327	37	for	for	ADP
ejpam-642	327	38	the	the	DET
ejpam-642	327	39	mixing	mix	VERB
ejpam-642	327	40	proportion	proportion	NOUN
ejpam-642	327	41	is	be	AUX
ejpam-642	327	42	chosen	choose	VERB
ejpam-642	327	43	asπ(0	asπ(0	VERB
ejpam-642	327	44	)	)	PUNCT
ejpam-642	328	1	=	=	SYM
ejpam-642	328	2	min(a	min(a	PROPN
ejpam-642	328	3	,	,	PUNCT
ejpam-642	328	4	b)/ab	b)/ab	NOUN
ejpam-642	328	5	=	=	SYM
ejpam-642	328	6	1	1	X
ejpam-642	328	7	/	/	SYM
ejpam-642	328	8	max(a	max(a	PROPN
ejpam-642	328	9	,	,	PUNCT
ejpam-642	328	10	b	b	NOUN
ejpam-642	328	11	)	)	PUNCT
ejpam-642	328	12	,	,	PUNCT
ejpam-642	328	13	the	the	DET
ejpam-642	328	14	upper	upper	ADJ
ejpam-642	328	15	bound	bound	NOUN
ejpam-642	328	16	of	of	ADP
ejpam-642	328	17	the	the	DET
ejpam-642	328	18	mixing	mix	VERB
ejpam-642	328	19	proportion	proportion	NOUN
ejpam-642	328	20	.	.	PUNCT
ejpam-642	329	1	this	this	PRON
ejpam-642	329	2	can	can	AUX
ejpam-642	329	3	be	be	AUX
ejpam-642	329	4	modified	modify	VERB
ejpam-642	329	5	as	as	ADP
ejpam-642	329	6	π(0)/10	π(0)/10	NOUN
ejpam-642	329	7	,	,	PUNCT
ejpam-642	329	8	π(0)/5	π(0)/5	NOUN
ejpam-642	329	9	or	or	CCONJ
ejpam-642	329	10	π(0)/2	π(0)/2	VERB
ejpam-642	329	11	depending	depend	VERB
ejpam-642	329	12	on	on	ADP
ejpam-642	329	13	the	the	DET
ejpam-642	329	14	blocking	block	VERB
ejpam-642	329	15	situation	situation	NOUN
ejpam-642	329	16	.	.	PUNCT
ejpam-642	330	1	•	•	NUM
ejpam-642	330	2	β	β	X
ejpam-642	330	3	(	(	PUNCT
ejpam-642	330	4	0)u	0)u	INTJ
ejpam-642	330	5	,	,	PUNCT
ejpam-642	330	6	the	the	DET
ejpam-642	330	7	vector	vector	NOUN
ejpam-642	330	8	of	of	ADP
ejpam-642	330	9	initial	initial	ADJ
ejpam-642	330	10	values	value	NOUN
ejpam-642	330	11	in	in	ADP
ejpam-642	330	12	the	the	DET
ejpam-642	330	13	loglinear	loglinear	NOUN
ejpam-642	330	14	model	model	NOUN
ejpam-642	330	15	for	for	ADP
ejpam-642	330	16	u	u	PROPN
ejpam-642	330	17	,	,	PUNCT
ejpam-642	330	18	is	be	AUX
ejpam-642	330	19	obtained	obtain	VERB
ejpam-642	330	20	from	from	ADP
ejpam-642	330	21	fitting	fit	VERB
ejpam-642	330	22	that	that	DET
ejpam-642	330	23	model	model	NOUN
ejpam-642	330	24	to	to	ADP
ejpam-642	330	25	a	a	DET
ejpam-642	330	26	modified	modify	VERB
ejpam-642	330	27	table	table	NOUN
ejpam-642	330	28	which	which	PRON
ejpam-642	330	29	has	have	VERB
ejpam-642	330	30	more	more	ADJ
ejpam-642	330	31	unmatched	unmatched	ADJ
ejpam-642	330	32	pairs	pair	NOUN
ejpam-642	330	33	and	and	CCONJ
ejpam-642	330	34	fewer	few	ADJ
ejpam-642	330	35	matched	matched	ADJ
ejpam-642	330	36	pairs	pair	NOUN
ejpam-642	330	37	.	.	PUNCT
ejpam-642	331	1	the	the	DET
ejpam-642	331	2	cell	cell	NOUN
ejpam-642	331	3	values	value	NOUN
ejpam-642	331	4	of	of	ADP
ejpam-642	331	5	this	this	DET
ejpam-642	331	6	table	table	NOUN
ejpam-642	331	7	are	be	AUX
ejpam-642	331	8	the	the	DET
ejpam-642	331	9	original	original	ADJ
ejpam-642	331	10	cell	cell	NOUN
ejpam-642	331	11	counts	count	NOUN
ejpam-642	331	12	divided	divide	VERB
ejpam-642	331	13	by	by	ADP
ejpam-642	331	14	min(a	min(a	PROPN
ejpam-642	331	15	,	,	PUNCT
ejpam-642	331	16	b	b	NOUN
ejpam-642	331	17	)	)	PUNCT
ejpam-642	331	18	(	(	PUNCT
ejpam-642	331	19	thus	thus	ADV
ejpam-642	331	20	may	may	AUX
ejpam-642	331	21	not	not	PART
ejpam-642	331	22	be	be	AUX
ejpam-642	331	23	integers	integer	NOUN
ejpam-642	331	24	)	)	PUNCT
ejpam-642	331	25	.	.	PUNCT
ejpam-642	332	1	•	•	NUM
ejpam-642	332	2	β	β	X
ejpam-642	332	3	(	(	PUNCT
ejpam-642	332	4	0)m	0)m	NOUN
ejpam-642	332	5	,	,	PUNCT
ejpam-642	332	6	the	the	DET
ejpam-642	332	7	vector	vector	NOUN
ejpam-642	332	8	of	of	ADP
ejpam-642	332	9	initial	initial	ADJ
ejpam-642	332	10	values	value	NOUN
ejpam-642	332	11	in	in	ADP
ejpam-642	332	12	the	the	DET
ejpam-642	332	13	loglinear	loglinear	NOUN
ejpam-642	332	14	model	model	NOUN
ejpam-642	332	15	for	for	ADP
ejpam-642	332	16	m	m	PROPN
ejpam-642	332	17	,	,	PUNCT
ejpam-642	332	18	is	be	AUX
ejpam-642	332	19	obtained	obtain	VERB
ejpam-642	332	20	from	from	ADP
ejpam-642	332	21	fitting	fit	VERB
ejpam-642	332	22	that	that	DET
ejpam-642	332	23	model	model	NOUN
ejpam-642	332	24	to	to	ADP
ejpam-642	332	25	another	another	DET
ejpam-642	332	26	modified	modify	VERB
ejpam-642	332	27	table	table	NOUN
ejpam-642	332	28	which	which	PRON
ejpam-642	332	29	is	be	AUX
ejpam-642	332	30	created	create	VERB
ejpam-642	332	31	by	by	ADP
ejpam-642	332	32	picking	pick	VERB
ejpam-642	332	33	counts	count	NOUN
ejpam-642	332	34	for	for	ADP
ejpam-642	332	35	cells	cell	NOUN
ejpam-642	332	36	with	with	ADP
ejpam-642	332	37	higher	high	ADJ
ejpam-642	332	38	matching	matching	NOUN
ejpam-642	332	39	chance	chance	NOUN
ejpam-642	332	40	,	,	PUNCT
ejpam-642	332	41	and	and	CCONJ
ejpam-642	332	42	assigning	assign	VERB
ejpam-642	332	43	zeroes	zero	NOUN
ejpam-642	332	44	for	for	ADP
ejpam-642	332	45	others	other	NOUN
ejpam-642	332	46	cells	cell	NOUN
ejpam-642	332	47	.	.	PUNCT
ejpam-642	333	1	intuitively	intuitively	ADV
ejpam-642	333	2	,	,	PUNCT
ejpam-642	333	3	if	if	SCONJ
ejpam-642	333	4	the	the	DET
ejpam-642	333	5	table	table	NOUN
ejpam-642	333	6	is	be	AUX
ejpam-642	333	7	derived	derive	VERB
ejpam-642	333	8	from	from	ADP
ejpam-642	333	9	more	more	ADJ
ejpam-642	333	10	unmatched	unmatched	ADJ
ejpam-642	333	11	pairs	pair	NOUN
ejpam-642	333	12	or	or	CCONJ
ejpam-642	333	13	from	from	ADP
ejpam-642	333	14	more	more	ADJ
ejpam-642	333	15	matched	match	VERB
ejpam-642	333	16	pairs	pair	NOUN
ejpam-642	333	17	,	,	PUNCT
ejpam-642	333	18	then	then	ADV
ejpam-642	333	19	the	the	DET
ejpam-642	333	20	fitted	fit	VERB
ejpam-642	333	21	model	model	NOUN
ejpam-642	333	22	is	be	AUX
ejpam-642	333	23	closer	close	ADJ
ejpam-642	333	24	to	to	ADP
ejpam-642	333	25	the	the	DET
ejpam-642	333	26	true	true	ADJ
ejpam-642	333	27	model	model	NOUN
ejpam-642	333	28	for	for	ADP
ejpam-642	333	29	u	u	NOUN
ejpam-642	333	30	or	or	CCONJ
ejpam-642	333	31	for	for	ADP
ejpam-642	333	32	m	m	NOUN
ejpam-642	333	33	respectively	respectively	ADV
ejpam-642	333	34	.	.	PUNCT
ejpam-642	334	1	it	it	PRON
ejpam-642	334	2	works	work	VERB
ejpam-642	334	3	quite	quite	ADV
ejpam-642	334	4	well	well	ADV
ejpam-642	334	5	in	in	ADP
ejpam-642	334	6	the	the	DET
ejpam-642	334	7	simulation	simulation	NOUN
ejpam-642	334	8	studies	study	NOUN
ejpam-642	334	9	.	.	PUNCT
ejpam-642	335	1	note	note	VERB
ejpam-642	335	2	that	that	SCONJ
ejpam-642	335	3	fitting	fitting	ADJ
ejpam-642	335	4	a	a	DET
ejpam-642	335	5	loglinear	loglinear	NOUN
ejpam-642	335	6	model	model	NOUN
ejpam-642	335	7	to	to	ADP
ejpam-642	335	8	a	a	DET
ejpam-642	335	9	non	non	ADJ
ejpam-642	335	10	-	-	ADJ
ejpam-642	335	11	integer	integer	ADJ
ejpam-642	335	12	-	-	PUNCT
ejpam-642	335	13	valued	value	VERB
ejpam-642	335	14	table	table	NOUN
ejpam-642	335	15	is	be	AUX
ejpam-642	335	16	the	the	DET
ejpam-642	335	17	same	same	ADJ
ejpam-642	335	18	as	as	ADP
ejpam-642	335	19	fitting	fitting	ADJ
ejpam-642	335	20	to	to	ADP
ejpam-642	335	21	an	an	DET
ejpam-642	335	22	integer	integer	NOUN
ejpam-642	335	23	-	-	PUNCT
ejpam-642	335	24	valued	value	VERB
ejpam-642	335	25	table	table	NOUN
ejpam-642	335	26	in	in	ADP
ejpam-642	335	27	r.	r.	PROPN
ejpam-642	335	28	4	4	NUM
ejpam-642	335	29	.	.	PUNCT
ejpam-642	336	1	simulation	simulation	NOUN
ejpam-642	336	2	studies	studies	PROPN
ejpam-642	336	3	simulation	simulation	NOUN
ejpam-642	336	4	studies	study	NOUN
ejpam-642	336	5	are	be	AUX
ejpam-642	336	6	necessary	necessary	ADJ
ejpam-642	336	7	for	for	ADP
ejpam-642	336	8	verifying	verify	VERB
ejpam-642	336	9	the	the	DET
ejpam-642	336	10	proposed	propose	VERB
ejpam-642	336	11	data	data	NOUN
ejpam-642	336	12	-	-	PUNCT
ejpam-642	336	13	driven	drive	VERB
ejpam-642	336	14	strategy	strategy	NOUN
ejpam-642	336	15	of	of	ADP
ejpam-642	336	16	model	model	NOUN
ejpam-642	336	17	building	building	NOUN
ejpam-642	336	18	.	.	PUNCT
ejpam-642	337	1	from	from	ADP
ejpam-642	337	2	simulation	simulation	NOUN
ejpam-642	337	3	,	,	PUNCT
ejpam-642	337	4	we	we	PRON
ejpam-642	337	5	can	can	AUX
ejpam-642	337	6	obtain	obtain	VERB
ejpam-642	337	7	experience	experience	NOUN
ejpam-642	337	8	on	on	ADP
ejpam-642	337	9	model	model	NOUN
ejpam-642	337	10	fitting	fitting	ADJ
ejpam-642	337	11	and	and	CCONJ
ejpam-642	337	12	building	build	VERB
ejpam-642	337	13	in	in	ADP
ejpam-642	337	14	various	various	ADJ
ejpam-642	337	15	settings	setting	NOUN
ejpam-642	337	16	such	such	ADJ
ejpam-642	337	17	as	as	ADP
ejpam-642	337	18	file	file	NOUN
ejpam-642	337	19	sizes	size	NOUN
ejpam-642	337	20	,	,	PUNCT
ejpam-642	337	21	mixing	mix	VERB
ejpam-642	337	22	proportions	proportion	NOUN
ejpam-642	337	23	and	and	CCONJ
ejpam-642	337	24	interactions	interaction	NOUN
ejpam-642	337	25	patterns	pattern	NOUN
ejpam-642	337	26	.	.	PUNCT
ejpam-642	338	1	besides	besides	SCONJ
ejpam-642	338	2	,	,	PUNCT
ejpam-642	338	3	we	we	PRON
ejpam-642	338	4	can	can	AUX
ejpam-642	338	5	also	also	ADV
ejpam-642	338	6	identify	identify	VERB
ejpam-642	338	7	potential	potential	ADJ
ejpam-642	338	8	drawbacks	drawback	NOUN
ejpam-642	338	9	for	for	ADP
ejpam-642	338	10	necessary	necessary	ADJ
ejpam-642	338	11	adjustment	adjustment	NOUN
ejpam-642	338	12	.	.	PUNCT
ejpam-642	339	1	we	we	PRON
ejpam-642	339	2	consider	consider	VERB
ejpam-642	339	3	the	the	DET
ejpam-642	339	4	comparison	comparison	NOUN
ejpam-642	339	5	vector	vector	NOUN
ejpam-642	339	6	consisting	consist	VERB
ejpam-642	339	7	of	of	ADP
ejpam-642	339	8	four	four	NUM
ejpam-642	339	9	fields	field	NOUN
ejpam-642	339	10	,	,	PUNCT
ejpam-642	339	11	γ	γ	X
ejpam-642	339	12	=	=	SYM
ejpam-642	339	13	(	(	PUNCT
ejpam-642	339	14	γ1,γ2,γ3,γ4	γ1,γ2,γ3,γ4	PROPN
ejpam-642	339	15	)	)	PUNCT
ejpam-642	339	16	,	,	PUNCT
ejpam-642	339	17	which	which	PRON
ejpam-642	339	18	is	be	AUX
ejpam-642	339	19	fairly	fairly	ADV
ejpam-642	339	20	moderate	moderate	ADJ
ejpam-642	339	21	in	in	ADP
ejpam-642	339	22	reality	reality	NOUN
ejpam-642	339	23	.	.	PUNCT
ejpam-642	340	1	each	each	DET
ejpam-642	340	2	field	field	NOUN
ejpam-642	340	3	is	be	AUX
ejpam-642	340	4	binary	binary	ADJ
ejpam-642	340	5	,	,	PUNCT
ejpam-642	340	6	where	where	SCONJ
ejpam-642	340	7	0	0	NUM
ejpam-642	340	8	means	mean	VERB
ejpam-642	340	9	the	the	DET
ejpam-642	340	10	field	field	NOUN
ejpam-642	340	11	unmatched	unmatched	ADJ
ejpam-642	340	12	while	while	SCONJ
ejpam-642	340	13	1	1	NUM
ejpam-642	340	14	r.	r.	PROPN
ejpam-642	340	15	zhu	zhu	PROPN
ejpam-642	340	16	,	,	PUNCT
ejpam-642	340	17	j.	j.	PROPN
ejpam-642	340	18	zhang	zhang	PROPN
ejpam-642	340	19	,	,	PUNCT
ejpam-642	340	20	d.	d.	PROPN
ejpam-642	340	21	zhang	zhang	PROPN
ejpam-642	340	22	,	,	PUNCT
ejpam-642	340	23	g.	g.	PROPN
ejpam-642	340	24	yan	yan	PROPN
ejpam-642	340	25	/	/	SYM
ejpam-642	340	26	eur	eur	PROPN
ejpam-642	340	27	.	.	PUNCT
ejpam-642	341	1	j.	j.	PROPN
ejpam-642	341	2	pure	pure	PROPN
ejpam-642	341	3	appl	appl	PROPN
ejpam-642	341	4	.	.	PROPN
ejpam-642	341	5	math	math	PROPN
ejpam-642	341	6	,	,	PUNCT
ejpam-642	341	7	3	3	NUM
ejpam-642	341	8	(	(	PUNCT
ejpam-642	341	9	2010	2010	NUM
ejpam-642	341	10	)	)	PUNCT
ejpam-642	341	11	,	,	PUNCT
ejpam-642	341	12	141	141	NUM
ejpam-642	341	13	-	-	SYM
ejpam-642	341	14	162	162	NUM
ejpam-642	341	15	153	153	NUM
ejpam-642	341	16	indicates	indicate	VERB
ejpam-642	341	17	the	the	DET
ejpam-642	341	18	field	field	NOUN
ejpam-642	341	19	matched	match	VERB
ejpam-642	341	20	.	.	PUNCT
ejpam-642	342	1	thus	thus	ADV
ejpam-642	342	2	,	,	PUNCT
ejpam-642	342	3	the	the	DET
ejpam-642	342	4	total	total	ADJ
ejpam-642	342	5	number	number	NOUN
ejpam-642	342	6	of	of	ADP
ejpam-642	342	7	cells	cell	NOUN
ejpam-642	342	8	is	be	AUX
ejpam-642	342	9	24	24	NUM
ejpam-642	342	10	=	=	SYM
ejpam-642	342	11	16	16	NUM
ejpam-642	342	12	.	.	PUNCT
ejpam-642	343	1	table	table	NOUN
ejpam-642	343	2	1	1	NUM
ejpam-642	343	3	labels	label	NOUN
ejpam-642	343	4	all	all	PRON
ejpam-642	343	5	of	of	ADP
ejpam-642	343	6	these	these	DET
ejpam-642	343	7	sixteen	sixteen	NUM
ejpam-642	343	8	cells	cell	NOUN
ejpam-642	343	9	in	in	ADP
ejpam-642	343	10	the	the	DET
ejpam-642	343	11	way	way	NOUN
ejpam-642	343	12	of	of	ADP
ejpam-642	343	13	a	a	DET
ejpam-642	343	14	24	24	NUM
ejpam-642	343	15	factorial	factorial	ADJ
ejpam-642	343	16	design	design	NOUN
ejpam-642	343	17	matrix	matrix	NOUN
ejpam-642	343	18	for	for	ADP
ejpam-642	343	19	later	later	ADJ
ejpam-642	343	20	reference	reference	NOUN
ejpam-642	343	21	.	.	PUNCT
ejpam-642	344	1	commonly	commonly	ADV
ejpam-642	344	2	,	,	PUNCT
ejpam-642	344	3	table	table	NOUN
ejpam-642	344	4	1	1	NUM
ejpam-642	344	5	:	:	PUNCT
ejpam-642	344	6	cell	cell	NOUN
ejpam-642	344	7	labels	label	NOUN
ejpam-642	344	8	for	for	ADP
ejpam-642	344	9	the	the	DET
ejpam-642	344	10	2×	2×	NUM
ejpam-642	344	11	2×	2×	NUM
ejpam-642	344	12	2×	2×	NUM
ejpam-642	344	13	2	2	NUM
ejpam-642	344	14	ontingen	ontingen	PROPN
ejpam-642	344	15	y	y	PROPN
ejpam-642	344	16	table	table	NOUN
ejpam-642	344	17	.	.	PUNCT
ejpam-642	345	1	cell	cell	NOUN
ejpam-642	345	2	comparison	comparison	NOUN
ejpam-642	345	3	vector	vector	NOUN
ejpam-642	345	4	cell	cell	NOUN
ejpam-642	345	5	comparison	comparison	NOUN
ejpam-642	345	6	vector	vector	NOUN
ejpam-642	345	7	1	1	NUM
ejpam-642	345	8	(	(	PUNCT
ejpam-642	345	9	0,0,0,0	0,0,0,0	NOUN
ejpam-642	345	10	)	)	NUM
ejpam-642	345	11	9	9	NUM
ejpam-642	345	12	(	(	PUNCT
ejpam-642	345	13	0,0,0,1	0,0,0,1	NOUN
ejpam-642	345	14	)	)	PUNCT
ejpam-642	345	15	2	2	NUM
ejpam-642	345	16	(	(	PUNCT
ejpam-642	345	17	1,0,0,0	1,0,0,0	NUM
ejpam-642	345	18	)	)	SYM
ejpam-642	345	19	10	10	NUM
ejpam-642	345	20	(	(	PUNCT
ejpam-642	345	21	1,0,0,1	1,0,0,1	NUM
ejpam-642	345	22	)	)	SYM
ejpam-642	345	23	3	3	NUM
ejpam-642	345	24	(	(	PUNCT
ejpam-642	345	25	0,1,0,0	0,1,0,0	NUM
ejpam-642	345	26	)	)	PUNCT
ejpam-642	345	27	11	11	NUM
ejpam-642	345	28	(	(	PUNCT
ejpam-642	345	29	0,1,0,1	0,1,0,1	NOUN
ejpam-642	345	30	)	)	PUNCT
ejpam-642	345	31	4	4	NUM
ejpam-642	345	32	(	(	PUNCT
ejpam-642	345	33	1,1,0,0	1,1,0,0	NUM
ejpam-642	345	34	)	)	PUNCT
ejpam-642	345	35	12	12	NUM
ejpam-642	345	36	(	(	PUNCT
ejpam-642	345	37	1,1,0,1	1,1,0,1	NUM
ejpam-642	345	38	)	)	SYM
ejpam-642	345	39	5	5	NUM
ejpam-642	345	40	(	(	PUNCT
ejpam-642	345	41	0,0,1,0	0,0,1,0	NUM
ejpam-642	345	42	)	)	PUNCT
ejpam-642	345	43	13	13	NUM
ejpam-642	345	44	(	(	PUNCT
ejpam-642	345	45	0,0,1,1	0,0,1,1	NUM
ejpam-642	345	46	)	)	PUNCT
ejpam-642	345	47	6	6	NUM
ejpam-642	345	48	(	(	PUNCT
ejpam-642	345	49	1,0,1,0	1,0,1,0	NUM
ejpam-642	345	50	)	)	SYM
ejpam-642	345	51	14	14	NUM
ejpam-642	345	52	(	(	PUNCT
ejpam-642	345	53	1,0,1,1	1,0,1,1	NUM
ejpam-642	345	54	)	)	PUNCT
ejpam-642	345	55	7	7	NUM
ejpam-642	345	56	(	(	PUNCT
ejpam-642	345	57	0,1,1,0	0,1,1,0	NUM
ejpam-642	345	58	)	)	PUNCT
ejpam-642	345	59	15	15	NUM
ejpam-642	345	60	(	(	PUNCT
ejpam-642	345	61	0,1,1,1	0,1,1,1	NUM
ejpam-642	345	62	)	)	PUNCT
ejpam-642	345	63	8	8	NUM
ejpam-642	345	64	(	(	PUNCT
ejpam-642	345	65	1,1,1,0	1,1,1,0	NUM
ejpam-642	345	66	)	)	PUNCT
ejpam-642	345	67	16	16	NUM
ejpam-642	345	68	(	(	PUNCT
ejpam-642	345	69	1,1,1,1	1,1,1,1	NUM
ejpam-642	345	70	)	)	PUNCT
ejpam-642	345	71	a	a	DET
ejpam-642	345	72	cell	cell	NOUN
ejpam-642	345	73	with	with	ADP
ejpam-642	345	74	more	more	ADJ
ejpam-642	345	75	1	1	NUM
ejpam-642	345	76	’s	’s	PART
ejpam-642	345	77	has	have	VERB
ejpam-642	345	78	larger	large	ADJ
ejpam-642	345	79	matching	matching	NOUN
ejpam-642	345	80	chance	chance	NOUN
ejpam-642	345	81	for	for	ADP
ejpam-642	345	82	record	record	NOUN
ejpam-642	345	83	pairs	pair	NOUN
ejpam-642	345	84	with	with	ADP
ejpam-642	345	85	comparison	comparison	NOUN
ejpam-642	345	86	vectors	vector	NOUN
ejpam-642	345	87	in	in	ADP
ejpam-642	345	88	that	that	DET
ejpam-642	345	89	cell	cell	NOUN
ejpam-642	345	90	.	.	PUNCT
ejpam-642	346	1	thus	thus	ADV
ejpam-642	346	2	,	,	PUNCT
ejpam-642	346	3	cells	cell	NOUN
ejpam-642	346	4	corresponding	correspond	VERB
ejpam-642	346	5	to	to	ADP
ejpam-642	346	6	comparison	comparison	NOUN
ejpam-642	346	7	vectors	vector	NOUN
ejpam-642	346	8	being	be	AUX
ejpam-642	346	9	(	(	PUNCT
ejpam-642	346	10	1,1,1,1	1,1,1,1	NUM
ejpam-642	346	11	)	)	PUNCT
ejpam-642	346	12	,	,	PUNCT
ejpam-642	346	13	(	(	PUNCT
ejpam-642	346	14	1,1,0,1	1,1,0,1	NUM
ejpam-642	346	15	)	)	PUNCT
ejpam-642	346	16	and	and	CCONJ
ejpam-642	346	17	(	(	PUNCT
ejpam-642	346	18	1,1,1,0	1,1,1,0	NUM
ejpam-642	346	19	)	)	PUNCT
ejpam-642	346	20	have	have	VERB
ejpam-642	346	21	more	more	ADJ
ejpam-642	346	22	matching	matching	NOUN
ejpam-642	346	23	possibilities	possibility	NOUN
ejpam-642	346	24	than	than	ADP
ejpam-642	346	25	cells	cell	NOUN
ejpam-642	346	26	corresponding	correspond	VERB
ejpam-642	346	27	to	to	ADP
ejpam-642	346	28	(	(	PUNCT
ejpam-642	346	29	0,0,0,0	0,0,0,0	NOUN
ejpam-642	346	30	)	)	PUNCT
ejpam-642	346	31	,	,	PUNCT
ejpam-642	346	32	(	(	PUNCT
ejpam-642	346	33	0,0,1,0	0,0,1,0	NUM
ejpam-642	346	34	)	)	PUNCT
ejpam-642	346	35	and	and	CCONJ
ejpam-642	346	36	(	(	PUNCT
ejpam-642	346	37	0,0,0,1	0,0,0,1	NOUN
ejpam-642	346	38	)	)	PUNCT
ejpam-642	346	39	.	.	PUNCT
ejpam-642	347	1	this	this	PRON
ejpam-642	347	2	is	be	AUX
ejpam-642	347	3	the	the	DET
ejpam-642	347	4	type	type	NOUN
ejpam-642	347	5	of	of	ADP
ejpam-642	347	6	contingency	contingency	NOUN
ejpam-642	347	7	table	table	NOUN
ejpam-642	347	8	considered	consider	VERB
ejpam-642	347	9	in	in	ADP
ejpam-642	347	10	jaro	jaro	PROPN
ejpam-642	347	11	(	(	PUNCT
ejpam-642	347	12	1989	1989	NUM
ejpam-642	347	13	)	)	PUNCT
ejpam-642	347	14	;	;	PUNCT
ejpam-642	347	15	see	see	VERB
ejpam-642	347	16	[	[	X
ejpam-642	347	17	7	7	NUM
ejpam-642	347	18	]	]	PUNCT
ejpam-642	347	19	.	.	PUNCT
ejpam-642	348	1	for	for	ADP
ejpam-642	348	2	this	this	DET
ejpam-642	348	3	situation	situation	NOUN
ejpam-642	348	4	,	,	PUNCT
ejpam-642	348	5	a	a	DET
ejpam-642	348	6	byproduct	byproduct	NOUN
ejpam-642	348	7	is	be	AUX
ejpam-642	348	8	that	that	SCONJ
ejpam-642	348	9	we	we	PRON
ejpam-642	348	10	can	can	AUX
ejpam-642	348	11	compare	compare	VERB
ejpam-642	348	12	the	the	DET
ejpam-642	348	13	estimated	estimate	VERB
ejpam-642	348	14	16	16	NUM
ejpam-642	348	15	cell	cell	NOUN
ejpam-642	348	16	probabilities	probability	NOUN
ejpam-642	348	17	from	from	ADP
ejpam-642	348	18	the	the	DET
ejpam-642	348	19	model	model	NOUN
ejpam-642	348	20	selected	select	VERB
ejpam-642	348	21	via	via	ADP
ejpam-642	348	22	the	the	DET
ejpam-642	348	23	data	data	NOUN
ejpam-642	348	24	-	-	PUNCT
ejpam-642	348	25	driven	drive	VERB
ejpam-642	348	26	strategy	strategy	NOUN
ejpam-642	348	27	with	with	ADP
ejpam-642	348	28	those	those	PRON
ejpam-642	348	29	from	from	ADP
ejpam-642	348	30	the	the	DET
ejpam-642	348	31	simplest	simple	ADJ
ejpam-642	348	32	starting	starting	NOUN
ejpam-642	348	33	model	model	NOUN
ejpam-642	348	34	m00	m00	PROPN
ejpam-642	348	35	.	.	PUNCT
ejpam-642	349	1	for	for	ADP
ejpam-642	349	2	each	each	DET
ejpam-642	349	3	individual	individual	ADJ
ejpam-642	349	4	simulation	simulation	NOUN
ejpam-642	349	5	,	,	PUNCT
ejpam-642	349	6	we	we	PRON
ejpam-642	349	7	set	set	VERB
ejpam-642	349	8	the	the	DET
ejpam-642	349	9	lengths	length	NOUN
ejpam-642	349	10	of	of	ADP
ejpam-642	349	11	two	two	NUM
ejpam-642	349	12	data	datum	NOUN
ejpam-642	349	13	sets	set	NOUN
ejpam-642	349	14	as	as	ADP
ejpam-642	349	15	a	a	PRON
ejpam-642	349	16	and	and	CCONJ
ejpam-642	349	17	b	b	NOUN
ejpam-642	349	18	,	,	PUNCT
ejpam-642	349	19	as	as	ADV
ejpam-642	349	20	well	well	ADV
ejpam-642	349	21	as	as	ADP
ejpam-642	349	22	the	the	DET
ejpam-642	349	23	true	true	ADJ
ejpam-642	349	24	matched	match	VERB
ejpam-642	349	25	pairs	pair	NOUN
ejpam-642	349	26	nm	nm	PRON
ejpam-642	349	27	(	(	PUNCT
ejpam-642	349	28	thus	thus	ADV
ejpam-642	349	29	,	,	PUNCT
ejpam-642	349	30	leading	lead	VERB
ejpam-642	349	31	to	to	ADP
ejpam-642	349	32	the	the	DET
ejpam-642	349	33	mixing	mix	VERB
ejpam-642	349	34	proportion	proportion	NOUN
ejpam-642	349	35	π	π	NOUN
ejpam-642	349	36	)	)	PUNCT
ejpam-642	349	37	.	.	PUNCT
ejpam-642	350	1	for	for	ADP
ejpam-642	350	2	the	the	DET
ejpam-642	350	3	sake	sake	NOUN
ejpam-642	350	4	of	of	ADP
ejpam-642	350	5	simplicity	simplicity	NOUN
ejpam-642	350	6	,	,	PUNCT
ejpam-642	350	7	we	we	PRON
ejpam-642	350	8	do	do	AUX
ejpam-642	350	9	not	not	PART
ejpam-642	350	10	consider	consider	VERB
ejpam-642	350	11	the	the	DET
ejpam-642	350	12	blocking	blocking	NOUN
ejpam-642	350	13	technique	technique	NOUN
ejpam-642	350	14	,	,	PUNCT
ejpam-642	350	15	thus	thus	ADV
ejpam-642	350	16	,	,	PUNCT
ejpam-642	350	17	there	there	PRON
ejpam-642	350	18	are	be	VERB
ejpam-642	350	19	n	n	PROPN
ejpam-642	350	20	=	=	SYM
ejpam-642	350	21	ab	ab	PROPN
ejpam-642	350	22	record	record	NOUN
ejpam-642	350	23	pairs	pair	NOUN
ejpam-642	350	24	.	.	PUNCT
ejpam-642	351	1	six	six	NUM
ejpam-642	351	2	settings	setting	NOUN
ejpam-642	351	3	are	be	AUX
ejpam-642	351	4	then	then	ADV
ejpam-642	351	5	chosen	choose	VERB
ejpam-642	351	6	for	for	ADP
ejpam-642	351	7	the	the	DET
ejpam-642	351	8	simulation	simulation	NOUN
ejpam-642	351	9	(	(	PUNCT
ejpam-642	351	10	see	see	VERB
ejpam-642	351	11	table	table	NOUN
ejpam-642	351	12	2	2	NUM
ejpam-642	351	13	)	)	PUNCT
ejpam-642	351	14	.	.	PUNCT
ejpam-642	352	1	the	the	DET
ejpam-642	352	2	combination	combination	NOUN
ejpam-642	352	3	of	of	ADP
ejpam-642	352	4	lengths	length	NOUN
ejpam-642	352	5	a	a	PRON
ejpam-642	352	6	and	and	CCONJ
ejpam-642	352	7	b	b	NOUN
ejpam-642	352	8	covers	cover	VERB
ejpam-642	352	9	threetable	threetable	ADJ
ejpam-642	352	10	2	2	NUM
ejpam-642	352	11	:	:	PUNCT
ejpam-642	352	12	lengths	length	NOUN
ejpam-642	352	13	of	of	ADP
ejpam-642	352	14	two	two	NUM
ejpam-642	352	15	data	datum	NOUN
ejpam-642	352	16	sets	set	NOUN
ejpam-642	352	17	and	and	CCONJ
ejpam-642	352	18	true	true	ADJ
ejpam-642	352	19	mat	mat	NOUN
ejpam-642	352	20	he	he	PRON
ejpam-642	352	21	s	s	VERB
ejpam-642	352	22	in	in	ADP
ejpam-642	352	23	six	six	NUM
ejpam-642	352	24	setting	set	VERB
ejpam-642	352	25	ases	as	NOUN
ejpam-642	352	26	.	.	PUNCT
ejpam-642	353	1	setting	set	VERB
ejpam-642	353	2	1	1	NUM
ejpam-642	353	3	2	2	NUM
ejpam-642	353	4	3	3	NUM
ejpam-642	353	5	4	4	NUM
ejpam-642	353	6	5	5	NUM
ejpam-642	353	7	6	6	NUM
ejpam-642	353	8	a	a	DET
ejpam-642	353	9	200	200	NUM
ejpam-642	353	10	200	200	NUM
ejpam-642	353	11	400	400	NUM
ejpam-642	353	12	400	400	NUM
ejpam-642	353	13	1,000	1,000	NUM
ejpam-642	353	14	1,000	1,000	NUM
ejpam-642	353	15	b	b	NOUN
ejpam-642	353	16	1,000	1,000	NUM
ejpam-642	353	17	1,000	1,000	NUM
ejpam-642	353	18	400	400	NUM
ejpam-642	353	19	400	400	NUM
ejpam-642	353	20	2,000	2,000	NUM
ejpam-642	353	21	2,000	2,000	NUM
ejpam-642	353	22	n	n	NOUN
ejpam-642	353	23	=	=	SYM
ejpam-642	353	24	ab	ab	PROPN
ejpam-642	353	25	200,000	200,000	NUM
ejpam-642	353	26	200,000	200,000	NUM
ejpam-642	353	27	160,000	160,000	NUM
ejpam-642	353	28	160,000	160,000	NUM
ejpam-642	353	29	2,000,000	2,000,000	NUM
ejpam-642	353	30	2,000,000	2,000,000	NUM
ejpam-642	353	31	nm	nm	NOUN
ejpam-642	353	32	20	20	NUM
ejpam-642	353	33	150	150	NUM
ejpam-642	353	34	40	40	NUM
ejpam-642	353	35	200	200	NUM
ejpam-642	353	36	50	50	NUM
ejpam-642	353	37	600	600	NUM
ejpam-642	353	38	situations	situation	NOUN
ejpam-642	353	39	:	:	PUNCT
ejpam-642	353	40	•	•	NUM
ejpam-642	353	41	one	one	NOUN
ejpam-642	353	42	is	be	AUX
ejpam-642	353	43	small	small	ADJ
ejpam-642	353	44	,	,	PUNCT
ejpam-642	353	45	and	and	CCONJ
ejpam-642	353	46	one	one	NUM
ejpam-642	353	47	is	be	AUX
ejpam-642	353	48	large	large	ADJ
ejpam-642	353	49	;	;	PUNCT
ejpam-642	353	50	•	•	ADP
ejpam-642	353	51	both	both	PRON
ejpam-642	353	52	are	be	AUX
ejpam-642	353	53	of	of	ADP
ejpam-642	353	54	equal	equal	ADJ
ejpam-642	353	55	lengths	length	NOUN
ejpam-642	353	56	;	;	PUNCT
ejpam-642	353	57	•	•	NOUN
ejpam-642	353	58	both	both	PRON
ejpam-642	353	59	are	be	AUX
ejpam-642	353	60	large	large	ADJ
ejpam-642	353	61	,	,	PUNCT
ejpam-642	353	62	but	but	CCONJ
ejpam-642	353	63	of	of	ADP
ejpam-642	353	64	different	different	ADJ
ejpam-642	353	65	lengths	length	NOUN
ejpam-642	353	66	.	.	PUNCT
ejpam-642	354	1	r.	r.	PROPN
ejpam-642	354	2	zhu	zhu	PROPN
ejpam-642	354	3	,	,	PUNCT
ejpam-642	354	4	j.	j.	PROPN
ejpam-642	354	5	zhang	zhang	PROPN
ejpam-642	354	6	,	,	PUNCT
ejpam-642	354	7	d.	d.	PROPN
ejpam-642	354	8	zhang	zhang	PROPN
ejpam-642	354	9	,	,	PUNCT
ejpam-642	354	10	g.	g.	PROPN
ejpam-642	354	11	yan	yan	PROPN
ejpam-642	354	12	/	/	SYM
ejpam-642	354	13	eur	eur	PROPN
ejpam-642	354	14	.	.	PUNCT
ejpam-642	355	1	j.	j.	PROPN
ejpam-642	355	2	pure	pure	PROPN
ejpam-642	355	3	appl	appl	PROPN
ejpam-642	355	4	.	.	PROPN
ejpam-642	355	5	math	math	PROPN
ejpam-642	355	6	,	,	PUNCT
ejpam-642	355	7	3	3	NUM
ejpam-642	355	8	(	(	PUNCT
ejpam-642	355	9	2010	2010	NUM
ejpam-642	355	10	)	)	PUNCT
ejpam-642	355	11	,	,	PUNCT
ejpam-642	355	12	141	141	NUM
ejpam-642	355	13	-	-	SYM
ejpam-642	355	14	162	162	NUM
ejpam-642	355	15	154	154	NUM
ejpam-642	355	16	two	two	NUM
ejpam-642	355	17	situations	situation	NOUN
ejpam-642	355	18	regarding	regard	VERB
ejpam-642	355	19	the	the	DET
ejpam-642	355	20	mixing	mix	VERB
ejpam-642	355	21	proportion	proportion	NOUN
ejpam-642	355	22	are	be	AUX
ejpam-642	355	23	considered	consider	VERB
ejpam-642	355	24	:	:	PUNCT
ejpam-642	355	25	one	one	NUM
ejpam-642	355	26	with	with	ADP
ejpam-642	355	27	a	a	DET
ejpam-642	355	28	small	small	ADJ
ejpam-642	355	29	mixing	mixing	NOUN
ejpam-642	355	30	proportion	proportion	NOUN
ejpam-642	355	31	corresponding	correspond	VERB
ejpam-642	355	32	to	to	ADP
ejpam-642	355	33	fewer	few	ADJ
ejpam-642	355	34	matched	match	VERB
ejpam-642	355	35	pairs	pair	NOUN
ejpam-642	355	36	,	,	PUNCT
ejpam-642	355	37	and	and	CCONJ
ejpam-642	355	38	one	one	NUM
ejpam-642	355	39	with	with	ADP
ejpam-642	355	40	a	a	DET
ejpam-642	355	41	fairly	fairly	ADV
ejpam-642	355	42	large	large	ADJ
ejpam-642	355	43	proportion	proportion	NOUN
ejpam-642	355	44	corresponding	correspond	VERB
ejpam-642	355	45	to	to	ADP
ejpam-642	355	46	more	more	ADV
ejpam-642	355	47	matched	match	VERB
ejpam-642	355	48	pairs	pair	NOUN
ejpam-642	355	49	.	.	PUNCT
ejpam-642	356	1	regarding	regard	VERB
ejpam-642	356	2	the	the	DET
ejpam-642	356	3	loglinear	loglinear	NOUN
ejpam-642	356	4	mixture	mixture	NOUN
ejpam-642	356	5	model	model	NOUN
ejpam-642	356	6	,	,	PUNCT
ejpam-642	356	7	we	we	PRON
ejpam-642	356	8	choose	choose	VERB
ejpam-642	356	9	two	two	NUM
ejpam-642	356	10	types	type	NOUN
ejpam-642	356	11	of	of	ADP
ejpam-642	356	12	loglinear	loglinear	NOUN
ejpam-642	356	13	models	model	NOUN
ejpam-642	356	14	for	for	ADP
ejpam-642	356	15	both	both	CCONJ
ejpam-642	356	16	u	u	NOUN
ejpam-642	356	17	and	and	CCONJ
ejpam-642	356	18	m	m	VERB
ejpam-642	356	19	as	as	SCONJ
ejpam-642	356	20	follows	follow	VERB
ejpam-642	356	21	.	.	PUNCT
ejpam-642	357	1	(	(	PUNCT
ejpam-642	357	2	1	1	X
ejpam-642	357	3	)	)	PUNCT
ejpam-642	357	4	models	model	NOUN
ejpam-642	357	5	with	with	ADP
ejpam-642	357	6	no	no	DET
ejpam-642	357	7	interactions	interaction	NOUN
ejpam-642	357	8	for	for	ADP
ejpam-642	357	9	both	both	CCONJ
ejpam-642	357	10	u	u	NOUN
ejpam-642	357	11	and	and	CCONJ
ejpam-642	357	12	m	m	VERB
ejpam-642	357	13	,	,	PUNCT
ejpam-642	357	14	i.e.	i.e.	X
ejpam-642	357	15	,	,	PUNCT
ejpam-642	357	16	modelm00	modelm00	NOUN
ejpam-642	357	17	.	.	PUNCT
ejpam-642	358	1	specifications	specification	NOUN
ejpam-642	358	2	of	of	ADP
ejpam-642	358	3	main	main	ADJ
ejpam-642	358	4	effects	effect	NOUN
ejpam-642	358	5	are	be	AUX
ejpam-642	358	6	in	in	ADP
ejpam-642	358	7	table	table	NOUN
ejpam-642	358	8	3.table	3.table	NUM
ejpam-642	358	9	3	3	NUM
ejpam-642	358	10	:	:	PUNCT
ejpam-642	358	11	spe	spe	PROPN
ejpam-642	358	12	i	i	NUM
ejpam-642	358	13	�	�	PROPN
ejpam-642	358	14	ations	ation	NOUN
ejpam-642	358	15	of	of	ADP
ejpam-642	358	16	main	main	ADJ
ejpam-642	358	17	e	e	NOUN
ejpam-642	358	18	�	�	NOUN
ejpam-642	358	19	e	e	NOUN
ejpam-642	358	20	ts	ts	ADP
ejpam-642	358	21	of	of	ADP
ejpam-642	358	22	models	model	NOUN
ejpam-642	358	23	with	with	ADP
ejpam-642	358	24	no	no	DET
ejpam-642	358	25	intera	intera	NOUN
ejpam-642	358	26	tions	tion	NOUN
ejpam-642	358	27	.	.	PUNCT
ejpam-642	359	1	component	component	NOUN
ejpam-642	359	2	λ	λ	PROPN
ejpam-642	359	3	λ	λ	PROPN
ejpam-642	359	4	(	(	PUNCT
ejpam-642	359	5	1	1	NUM
ejpam-642	359	6	)	)	PUNCT
ejpam-642	359	7	0	0	NUM
ejpam-642	360	1	λ	λ	NOUN
ejpam-642	360	2	(	(	PUNCT
ejpam-642	360	3	1	1	NUM
ejpam-642	360	4	)	)	SYM
ejpam-642	360	5	1	1	NUM
ejpam-642	360	6	λ	λ	X
ejpam-642	360	7	(	(	PUNCT
ejpam-642	360	8	2	2	NUM
ejpam-642	360	9	)	)	PUNCT
ejpam-642	360	10	0	0	NUM
ejpam-642	361	1	λ	λ	NOUN
ejpam-642	361	2	(	(	PUNCT
ejpam-642	361	3	2	2	NUM
ejpam-642	361	4	)	)	SYM
ejpam-642	361	5	1	1	NUM
ejpam-642	361	6	λ	λ	X
ejpam-642	361	7	(	(	PUNCT
ejpam-642	361	8	3	3	NUM
ejpam-642	361	9	)	)	PUNCT
ejpam-642	361	10	0	0	NUM
ejpam-642	362	1	λ	λ	NOUN
ejpam-642	362	2	(	(	PUNCT
ejpam-642	362	3	3	3	NUM
ejpam-642	362	4	)	)	SYM
ejpam-642	362	5	1	1	NUM
ejpam-642	362	6	λ	λ	X
ejpam-642	362	7	(	(	PUNCT
ejpam-642	362	8	4	4	NUM
ejpam-642	362	9	)	)	PUNCT
ejpam-642	362	10	0	0	NUM
ejpam-642	363	1	λ	λ	NOUN
ejpam-642	363	2	(	(	PUNCT
ejpam-642	363	3	4	4	NUM
ejpam-642	363	4	)	)	SYM
ejpam-642	363	5	1	1	NUM
ejpam-642	363	6	u	u	NOUN
ejpam-642	363	7	0.5	0.5	NUM
ejpam-642	363	8	2.30	2.30	NUM
ejpam-642	363	9	-2.30	-2.30	NUM
ejpam-642	363	10	1.59	1.59	NUM
ejpam-642	363	11	-1.59	-1.59	NUM
ejpam-642	363	12	1.74	1.74	NUM
ejpam-642	363	13	-1.74	-1.74	NUM
ejpam-642	363	14	2.09	2.09	NUM
ejpam-642	363	15	-2.09	-2.09	NUM
ejpam-642	363	16	m	m	NOUN
ejpam-642	363	17	0.3	0.3	NUM
ejpam-642	363	18	-1.10	-1.10	NOUN
ejpam-642	363	19	1.10	1.10	NUM
ejpam-642	363	20	-0.87	-0.87	NUM
ejpam-642	363	21	0.87	0.87	NUM
ejpam-642	363	22	-0.69	-0.69	NUM
ejpam-642	363	23	0.69	0.69	NUM
ejpam-642	363	24	-1.10	-1.10	NOUN
ejpam-642	363	25	1.10	1.10	NUM
ejpam-642	363	26	(	(	PUNCT
ejpam-642	363	27	2	2	NUM
ejpam-642	363	28	)	)	PUNCT
ejpam-642	363	29	models	model	NOUN
ejpam-642	363	30	with	with	ADP
ejpam-642	363	31	two	two	NUM
ejpam-642	363	32	-	-	PUNCT
ejpam-642	363	33	factor	factor	NOUN
ejpam-642	363	34	interactions	interaction	NOUN
ejpam-642	363	35	for	for	ADP
ejpam-642	363	36	both	both	CCONJ
ejpam-642	363	37	u	u	NOUN
ejpam-642	363	38	and	and	CCONJ
ejpam-642	363	39	m	m	VERB
ejpam-642	363	40	,	,	PUNCT
ejpam-642	363	41	i.e.	i.e.	X
ejpam-642	363	42	,	,	PUNCT
ejpam-642	363	43	modelm11	modelm11	NOUN
ejpam-642	363	44	.	.	PUNCT
ejpam-642	364	1	specifications	specification	NOUN
ejpam-642	364	2	of	of	ADP
ejpam-642	364	3	main	main	ADJ
ejpam-642	364	4	effects	effect	NOUN
ejpam-642	364	5	and	and	CCONJ
ejpam-642	364	6	interactions	interaction	NOUN
ejpam-642	364	7	are	be	AUX
ejpam-642	364	8	in	in	ADP
ejpam-642	364	9	table	table	NOUN
ejpam-642	364	10	4	4	NUM
ejpam-642	364	11	.	.	PUNCT
ejpam-642	365	1	note	note	VERB
ejpam-642	365	2	that	that	SCONJ
ejpam-642	366	1	λ	λ	PROPN
ejpam-642	366	2	(	(	PUNCT
ejpam-642	366	3	i	i	NOUN
ejpam-642	366	4	)	)	PUNCT
ejpam-642	366	5	1	1	NUM
ejpam-642	366	6	=	=	NOUN
ejpam-642	366	7	−λ	−λ	PROPN
ejpam-642	366	8	(	(	PUNCT
ejpam-642	366	9	i	i	NOUN
ejpam-642	366	10	)	)	PUNCT
ejpam-642	366	11	0	0	PUNCT
ejpam-642	366	12	,	,	PUNCT
ejpam-642	366	13	λ	λ	INTJ
ejpam-642	366	14	(	(	PUNCT
ejpam-642	366	15	i	i	NOUN
ejpam-642	366	16	j	j	PROPN
ejpam-642	366	17	)	)	PUNCT
ejpam-642	366	18	01	01	NUM
ejpam-642	367	1	=	=	SYM
ejpam-642	367	2	λ	λ	INTJ
ejpam-642	367	3	(	(	PUNCT
ejpam-642	367	4	i	i	NOUN
ejpam-642	367	5	j	j	PROPN
ejpam-642	367	6	)	)	PUNCT
ejpam-642	367	7	01	01	NUM
ejpam-642	368	1	=	=	NOUN
ejpam-642	368	2	−λ	−λ	PROPN
ejpam-642	368	3	(	(	PUNCT
ejpam-642	368	4	i	i	NOUN
ejpam-642	368	5	j	j	PROPN
ejpam-642	368	6	)	)	PUNCT
ejpam-642	368	7	00	00	PUNCT
ejpam-642	368	8	,	,	PUNCT
ejpam-642	369	1	λ	λ	INTJ
ejpam-642	369	2	(	(	PUNCT
ejpam-642	369	3	i	i	PRON
ejpam-642	369	4	j	j	PROPN
ejpam-642	369	5	)	)	PUNCT
ejpam-642	369	6	11	11	NUM
ejpam-642	370	1	=	=	SYM
ejpam-642	370	2	λ	λ	INTJ
ejpam-642	370	3	(	(	PUNCT
ejpam-642	370	4	i	i	NOUN
ejpam-642	370	5	j	j	PROPN
ejpam-642	370	6	)	)	PUNCT
ejpam-642	370	7	00	00	PUNCT
ejpam-642	370	8	,	,	PUNCT
ejpam-642	370	9	i	i	PRON
ejpam-642	370	10	,	,	PUNCT
ejpam-642	370	11	j	j	PROPN
ejpam-642	370	12	=	=	NOUN
ejpam-642	370	13	1,2,3,4	1,2,3,4	NUM
ejpam-642	370	14	.	.	PUNCT
ejpam-642	371	1	thus	thus	ADV
ejpam-642	371	2	,	,	PUNCT
ejpam-642	371	3	effects	effect	NOUN
ejpam-642	371	4	and	and	CCONJ
ejpam-642	371	5	interactions	interaction	NOUN
ejpam-642	371	6	at	at	ADP
ejpam-642	371	7	other	other	ADJ
ejpam-642	371	8	levels	level	NOUN
ejpam-642	371	9	can	can	AUX
ejpam-642	371	10	be	be	AUX
ejpam-642	371	11	readily	readily	ADV
ejpam-642	371	12	derived	derive	VERB
ejpam-642	371	13	from	from	ADP
ejpam-642	371	14	the	the	DET
ejpam-642	371	15	one	one	NUM
ejpam-642	371	16	at	at	ADP
ejpam-642	371	17	level	level	NOUN
ejpam-642	371	18	0.table	0.table	NUM
ejpam-642	371	19	4	4	NUM
ejpam-642	371	20	:	:	PUNCT
ejpam-642	371	21	spe	spe	PROPN
ejpam-642	371	22	i	i	NUM
ejpam-642	371	23	�	�	PROPN
ejpam-642	371	24	ations	ation	NOUN
ejpam-642	371	25	of	of	ADP
ejpam-642	371	26	main	main	ADJ
ejpam-642	371	27	e	e	NOUN
ejpam-642	371	28	�	�	NOUN
ejpam-642	371	29	e	e	NOUN
ejpam-642	371	30	ts	ts	NOUN
ejpam-642	371	31	and	and	CCONJ
ejpam-642	371	32	intera	intera	NOUN
ejpam-642	371	33	tions	tion	NOUN
ejpam-642	371	34	of	of	ADP
ejpam-642	371	35	models	model	NOUN
ejpam-642	371	36	with	with	ADP
ejpam-642	371	37	intera	intera	NOUN
ejpam-642	371	38	tions	tion	NOUN
ejpam-642	371	39	at	at	ADP
ejpam-642	371	40	level	level	NOUN
ejpam-642	371	41	0	0	NUM
ejpam-642	371	42	.	.	PUNCT
ejpam-642	372	1	component	component	NOUN
ejpam-642	372	2	λ	λ	PROPN
ejpam-642	372	3	λ	λ	PROPN
ejpam-642	372	4	(	(	PUNCT
ejpam-642	372	5	1	1	NUM
ejpam-642	372	6	)	)	PUNCT
ejpam-642	372	7	0	0	NUM
ejpam-642	373	1	λ	λ	NOUN
ejpam-642	373	2	(	(	PUNCT
ejpam-642	373	3	2	2	NUM
ejpam-642	373	4	)	)	PUNCT
ejpam-642	373	5	0	0	NUM
ejpam-642	374	1	λ	λ	NOUN
ejpam-642	374	2	(	(	PUNCT
ejpam-642	374	3	3	3	NUM
ejpam-642	374	4	)	)	PUNCT
ejpam-642	374	5	0	0	NUM
ejpam-642	375	1	λ	λ	NOUN
ejpam-642	375	2	(	(	PUNCT
ejpam-642	375	3	4	4	NUM
ejpam-642	375	4	)	)	PUNCT
ejpam-642	375	5	0	0	NUM
ejpam-642	376	1	λ	λ	NOUN
ejpam-642	376	2	(	(	PUNCT
ejpam-642	376	3	12	12	NUM
ejpam-642	376	4	)	)	PUNCT
ejpam-642	376	5	00	00	PUNCT
ejpam-642	377	1	λ	λ	INTJ
ejpam-642	377	2	(	(	PUNCT
ejpam-642	377	3	13	13	NUM
ejpam-642	377	4	)	)	PUNCT
ejpam-642	377	5	00	00	PUNCT
ejpam-642	378	1	λ	λ	INTJ
ejpam-642	378	2	(	(	PUNCT
ejpam-642	378	3	14	14	NUM
ejpam-642	378	4	)	)	PUNCT
ejpam-642	378	5	00	00	PUNCT
ejpam-642	379	1	λ	λ	INTJ
ejpam-642	379	2	(	(	PUNCT
ejpam-642	379	3	23	23	NUM
ejpam-642	379	4	)	)	PUNCT
ejpam-642	379	5	00	00	PUNCT
ejpam-642	380	1	λ	λ	INTJ
ejpam-642	380	2	(	(	PUNCT
ejpam-642	380	3	24	24	NUM
ejpam-642	380	4	)	)	PUNCT
ejpam-642	380	5	00	00	NUM
ejpam-642	381	1	λ	λ	INTJ
ejpam-642	381	2	(	(	PUNCT
ejpam-642	381	3	34	34	NUM
ejpam-642	381	4	)	)	PUNCT
ejpam-642	381	5	00	00	PUNCT
ejpam-642	382	1	u	u	NOUN
ejpam-642	382	2	0.5	0.5	NUM
ejpam-642	382	3	2.3	2.3	NUM
ejpam-642	382	4	0.6	0.6	NUM
ejpam-642	382	5	1.00	1.00	NUM
ejpam-642	382	6	3.0	3.0	NUM
ejpam-642	382	7	1.04	1.04	NUM
ejpam-642	382	8	0.40	0.40	NUM
ejpam-642	382	9	1.13	1.13	NUM
ejpam-642	382	10	0.65	0.65	NUM
ejpam-642	382	11	0.20	0.20	NUM
ejpam-642	382	12	1.40	1.40	NUM
ejpam-642	382	13	m	m	NOUN
ejpam-642	382	14	0.3	0.3	NUM
ejpam-642	382	15	-1.2	-1.2	NOUN
ejpam-642	382	16	-0.9	-0.9	PROPN
ejpam-642	382	17	-0.66	-0.66	NUM
ejpam-642	382	18	-0.5	-0.5	NUM
ejpam-642	382	19	0.44	0.44	NUM
ejpam-642	382	20	-0.64	-0.64	NUM
ejpam-642	382	21	-0.16	-0.16	ADP
ejpam-642	382	22	-1.08	-1.08	NUM
ejpam-642	382	23	1.08	1.08	NUM
ejpam-642	382	24	0.32	0.32	NUM
ejpam-642	382	25	with	with	ADP
ejpam-642	382	26	two	two	NUM
ejpam-642	382	27	different	different	ADJ
ejpam-642	382	28	mixture	mixture	NOUN
ejpam-642	382	29	models	model	NOUN
ejpam-642	382	30	assigned	assign	VERB
ejpam-642	382	31	for	for	ADP
ejpam-642	382	32	each	each	PRON
ejpam-642	382	33	of	of	ADP
ejpam-642	382	34	the	the	DET
ejpam-642	382	35	six	six	NUM
ejpam-642	382	36	cases	case	NOUN
ejpam-642	382	37	of	of	ADP
ejpam-642	382	38	data	datum	NOUN
ejpam-642	382	39	settings	setting	NOUN
ejpam-642	382	40	,	,	PUNCT
ejpam-642	382	41	we	we	PRON
ejpam-642	382	42	have	have	AUX
ejpam-642	382	43	investigated	investigate	VERB
ejpam-642	382	44	12	12	NUM
ejpam-642	382	45	simulation	simulation	NOUN
ejpam-642	382	46	cases	case	NOUN
ejpam-642	382	47	in	in	ADP
ejpam-642	382	48	total	total	NOUN
ejpam-642	382	49	.	.	PUNCT
ejpam-642	383	1	since	since	SCONJ
ejpam-642	383	2	the	the	DET
ejpam-642	383	3	mixture	mixture	NOUN
ejpam-642	383	4	model	model	NOUN
ejpam-642	383	5	is	be	AUX
ejpam-642	383	6	for	for	ADP
ejpam-642	383	7	comparison	comparison	NOUN
ejpam-642	383	8	vectors	vector	NOUN
ejpam-642	383	9	,	,	PUNCT
ejpam-642	383	10	we	we	PRON
ejpam-642	383	11	do	do	AUX
ejpam-642	383	12	not	not	PART
ejpam-642	383	13	need	need	VERB
ejpam-642	383	14	to	to	PART
ejpam-642	383	15	generate	generate	VERB
ejpam-642	383	16	two	two	NUM
ejpam-642	383	17	original	original	ADJ
ejpam-642	383	18	record	record	NOUN
ejpam-642	383	19	files	file	NOUN
ejpam-642	383	20	in	in	ADP
ejpam-642	383	21	each	each	DET
ejpam-642	383	22	simulation	simulation	NOUN
ejpam-642	383	23	case	case	NOUN
ejpam-642	383	24	.	.	PUNCT
ejpam-642	384	1	actually	actually	ADV
ejpam-642	384	2	,	,	PUNCT
ejpam-642	384	3	what	what	PRON
ejpam-642	384	4	we	we	PRON
ejpam-642	384	5	need	need	VERB
ejpam-642	384	6	to	to	PART
ejpam-642	384	7	generate	generate	VERB
ejpam-642	384	8	are	be	AUX
ejpam-642	384	9	(	(	PUNCT
ejpam-642	384	10	n	n	CCONJ
ejpam-642	384	11	−	−	PROPN
ejpam-642	384	12	nm	nm	NOUN
ejpam-642	384	13	)	)	PUNCT
ejpam-642	384	14	comparison	comparison	NOUN
ejpam-642	384	15	vectors	vector	NOUN
ejpam-642	384	16	from	from	ADP
ejpam-642	384	17	the	the	DET
ejpam-642	384	18	loglinear	loglinear	NOUN
ejpam-642	384	19	model	model	NOUN
ejpam-642	384	20	for	for	ADP
ejpam-642	384	21	u	u	PROPN
ejpam-642	384	22	and	and	CCONJ
ejpam-642	384	23	nm	nm	ADJ
ejpam-642	384	24	comparison	comparison	NOUN
ejpam-642	384	25	vectors	vector	NOUN
ejpam-642	384	26	from	from	ADP
ejpam-642	384	27	the	the	DET
ejpam-642	384	28	loglinear	loglinear	NOUN
ejpam-642	384	29	model	model	NOUN
ejpam-642	384	30	for	for	ADP
ejpam-642	384	31	m	m	PROPN
ejpam-642	384	32	in	in	ADP
ejpam-642	384	33	each	each	DET
ejpam-642	384	34	simulation	simulation	NOUN
ejpam-642	384	35	case	case	NOUN
ejpam-642	384	36	.	.	PUNCT
ejpam-642	385	1	these	these	DET
ejpam-642	385	2	comparison	comparison	NOUN
ejpam-642	385	3	vectors	vector	NOUN
ejpam-642	385	4	yield	yield	VERB
ejpam-642	385	5	the	the	DET
ejpam-642	385	6	2	2	NUM
ejpam-642	385	7	×	×	NOUN
ejpam-642	385	8	2	2	NUM
ejpam-642	385	9	×	×	NOUN
ejpam-642	385	10	2	2	NUM
ejpam-642	385	11	×	×	NOUN
ejpam-642	385	12	2	2	NUM
ejpam-642	385	13	tables	table	NOUN
ejpam-642	385	14	of	of	ADP
ejpam-642	385	15	16	16	NUM
ejpam-642	385	16	cells	cell	NOUN
ejpam-642	385	17	(	(	PUNCT
ejpam-642	385	18	see	see	VERB
ejpam-642	385	19	table	table	NOUN
ejpam-642	385	20	1	1	NUM
ejpam-642	385	21	)	)	PUNCT
ejpam-642	385	22	.	.	PUNCT
ejpam-642	386	1	therefore	therefore	ADV
ejpam-642	386	2	,	,	PUNCT
ejpam-642	386	3	we	we	PRON
ejpam-642	386	4	finally	finally	ADV
ejpam-642	386	5	need	need	VERB
ejpam-642	386	6	the	the	DET
ejpam-642	386	7	contingency	contingency	NOUN
ejpam-642	386	8	tables	table	NOUN
ejpam-642	386	9	for	for	ADP
ejpam-642	386	10	u	u	NOUN
ejpam-642	386	11	and	and	CCONJ
ejpam-642	386	12	m	m	VERB
ejpam-642	386	13	respectively	respectively	ADV
ejpam-642	386	14	.	.	PUNCT
ejpam-642	387	1	note	note	VERB
ejpam-642	387	2	that	that	SCONJ
ejpam-642	387	3	conditional	conditional	ADJ
ejpam-642	387	4	on	on	ADP
ejpam-642	387	5	the	the	DET
ejpam-642	387	6	total	total	ADJ
ejpam-642	387	7	count	count	NOUN
ejpam-642	387	8	,	,	PUNCT
ejpam-642	387	9	the	the	DET
ejpam-642	387	10	joint	joint	ADJ
ejpam-642	387	11	counts	count	NOUN
ejpam-642	387	12	of	of	ADP
ejpam-642	387	13	all	all	DET
ejpam-642	387	14	cells	cell	NOUN
ejpam-642	387	15	follow	follow	VERB
ejpam-642	387	16	a	a	DET
ejpam-642	387	17	multinomial	multinomial	ADJ
ejpam-642	387	18	distribution	distribution	NOUN
ejpam-642	387	19	.	.	PUNCT
ejpam-642	388	1	suppose	suppose	VERB
ejpam-642	388	2	yi	yi	PROPN
ejpam-642	388	3	and	and	CCONJ
ejpam-642	388	4	µi	µi	PROPN
ejpam-642	388	5	are	be	AUX
ejpam-642	388	6	the	the	DET
ejpam-642	388	7	observed	observed	ADJ
ejpam-642	388	8	count	count	NOUN
ejpam-642	388	9	and	and	CCONJ
ejpam-642	388	10	expected	expect	VERB
ejpam-642	388	11	count	count	NOUN
ejpam-642	388	12	in	in	ADP
ejpam-642	388	13	cell	cell	NOUN
ejpam-642	388	14	i	i	PRON
ejpam-642	388	15	respectively	respectively	ADV
ejpam-642	388	16	.	.	PUNCT
ejpam-642	389	1	let	let	VERB
ejpam-642	389	2	n0	n0	X
ejpam-642	389	3	be	be	AUX
ejpam-642	389	4	the	the	DET
ejpam-642	389	5	total	total	ADJ
ejpam-642	389	6	count	count	NOUN
ejpam-642	389	7	.	.	PUNCT
ejpam-642	390	1	then	then	ADV
ejpam-642	390	2	,	,	PUNCT
ejpam-642	390	3	conditional	conditional	ADJ
ejpam-642	390	4	on	on	ADP
ejpam-642	390	5	∑16	∑16	PROPN
ejpam-642	390	6	i=1	i=1	PROPN
ejpam-642	390	7	yi	yi	PROPN
ejpam-642	390	8	=	=	SYM
ejpam-642	390	9	n0	n0	PROPN
ejpam-642	390	10	,	,	PUNCT
ejpam-642	390	11	(	(	PUNCT
ejpam-642	390	12	y1	y1	INTJ
ejpam-642	390	13	,	,	PUNCT
ejpam-642	390	14	y2	y2	PROPN
ejpam-642	390	15	,	,	PUNCT
ejpam-642	390	16	.	.	PUNCT
ejpam-642	390	17	.	.	PUNCT
ejpam-642	390	18	.	.	PUNCT
ejpam-642	391	1	,	,	PUNCT
ejpam-642	391	2	y16)∼multinomial(n0	y16)∼multinomial(n0	PROPN
ejpam-642	391	3	;	;	PUNCT
ejpam-642	391	4	p1	p1	NOUN
ejpam-642	391	5	,	,	PUNCT
ejpam-642	391	6	p2	p2	NOUN
ejpam-642	391	7	,	,	PUNCT
ejpam-642	391	8	.	.	PUNCT
ejpam-642	391	9	.	.	PUNCT
ejpam-642	391	10	.	.	PUNCT
ejpam-642	392	1	,	,	PUNCT
ejpam-642	392	2	p16	p16	NOUN
ejpam-642	392	3	)	)	PUNCT
ejpam-642	392	4	,	,	PUNCT
ejpam-642	392	5	(	(	PUNCT
ejpam-642	392	6	12	12	NUM
ejpam-642	392	7	)	)	PUNCT
ejpam-642	393	1	r.	r.	PROPN
ejpam-642	393	2	zhu	zhu	PROPN
ejpam-642	393	3	,	,	PUNCT
ejpam-642	393	4	j.	j.	PROPN
ejpam-642	393	5	zhang	zhang	PROPN
ejpam-642	393	6	,	,	PUNCT
ejpam-642	393	7	d.	d.	PROPN
ejpam-642	393	8	zhang	zhang	PROPN
ejpam-642	393	9	,	,	PUNCT
ejpam-642	393	10	g.	g.	PROPN
ejpam-642	393	11	yan	yan	PROPN
ejpam-642	393	12	/	/	SYM
ejpam-642	393	13	eur	eur	PROPN
ejpam-642	393	14	.	.	PUNCT
ejpam-642	394	1	j.	j.	PROPN
ejpam-642	394	2	pure	pure	PROPN
ejpam-642	394	3	appl	appl	PROPN
ejpam-642	394	4	.	.	PROPN
ejpam-642	394	5	math	math	PROPN
ejpam-642	394	6	,	,	PUNCT
ejpam-642	394	7	3	3	NUM
ejpam-642	394	8	(	(	PUNCT
ejpam-642	394	9	2010	2010	NUM
ejpam-642	394	10	)	)	PUNCT
ejpam-642	394	11	,	,	PUNCT
ejpam-642	394	12	141	141	NUM
ejpam-642	394	13	-	-	SYM
ejpam-642	394	14	162	162	NUM
ejpam-642	394	15	155	155	NUM
ejpam-642	395	1	where	where	SCONJ
ejpam-642	395	2	pi	pi	NOUN
ejpam-642	395	3	=	=	PUNCT
ejpam-642	395	4	µi	µi	PROPN
ejpam-642	395	5	�	�	PROPN
ejpam-642	395	6	16	16	NUM
ejpam-642	395	7	∑	∑	PROPN
ejpam-642	395	8	j=1	j=1	PROPN
ejpam-642	395	9	µ	µ	PROPN
ejpam-642	395	10	j	j	PROPN
ejpam-642	395	11	,	,	PUNCT
ejpam-642	395	12	the	the	DET
ejpam-642	395	13	probability	probability	NOUN
ejpam-642	395	14	that	that	SCONJ
ejpam-642	395	15	an	an	DET
ejpam-642	395	16	observed	observed	ADJ
ejpam-642	395	17	vector	vector	NOUN
ejpam-642	395	18	occurs	occur	VERB
ejpam-642	395	19	in	in	ADP
ejpam-642	395	20	cell	cell	NOUN
ejpam-642	395	21	i	i	PRON
ejpam-642	395	22	,	,	PUNCT
ejpam-642	395	23	i	i	NOUN
ejpam-642	395	24	=	=	NOUN
ejpam-642	395	25	1,2	1,2	NUM
ejpam-642	395	26	,	,	PUNCT
ejpam-642	395	27	.	.	PUNCT
ejpam-642	395	28	.	.	PUNCT
ejpam-642	396	1	.	.	PUNCT
ejpam-642	397	1	,	,	PUNCT
ejpam-642	397	2	16	16	NUM
ejpam-642	397	3	.	.	PUNCT
ejpam-642	398	1	for	for	ADP
ejpam-642	398	2	a	a	DET
ejpam-642	398	3	reference	reference	NOUN
ejpam-642	398	4	,	,	PUNCT
ejpam-642	398	5	see	see	VERB
ejpam-642	398	6	[	[	X
ejpam-642	398	7	2	2	NUM
ejpam-642	398	8	]	]	PUNCT
ejpam-642	398	9	,	,	PUNCT
ejpam-642	398	10	p.	p.	NOUN
ejpam-642	398	11	317	317	NUM
ejpam-642	398	12	-	-	SYM
ejpam-642	398	13	318	318	NUM
ejpam-642	398	14	.	.	PUNCT
ejpam-642	399	1	in	in	ADP
ejpam-642	399	2	general	general	ADJ
ejpam-642	399	3	,	,	PUNCT
ejpam-642	399	4	we	we	PRON
ejpam-642	399	5	can	can	AUX
ejpam-642	399	6	simulate	simulate	VERB
ejpam-642	399	7	a	a	DET
ejpam-642	399	8	contingency	contingency	NOUN
ejpam-642	399	9	table	table	NOUN
ejpam-642	399	10	according	accord	VERB
ejpam-642	399	11	to	to	ADP
ejpam-642	399	12	(	(	PUNCT
ejpam-642	399	13	12	12	NUM
ejpam-642	399	14	)	)	PUNCT
ejpam-642	399	15	for	for	ADP
ejpam-642	399	16	a	a	DET
ejpam-642	399	17	specified	specify	VERB
ejpam-642	399	18	total	total	ADJ
ejpam-642	399	19	count	count	NOUN
ejpam-642	399	20	.	.	PUNCT
ejpam-642	400	1	this	this	PRON
ejpam-642	400	2	is	be	AUX
ejpam-642	400	3	the	the	DET
ejpam-642	400	4	method	method	NOUN
ejpam-642	400	5	we	we	PRON
ejpam-642	400	6	use	use	VERB
ejpam-642	400	7	to	to	PART
ejpam-642	400	8	simulate	simulate	VERB
ejpam-642	400	9	data	data	NOUN
ejpam-642	400	10	sets	set	NOUN
ejpam-642	400	11	for	for	ADP
ejpam-642	400	12	six	six	NUM
ejpam-642	400	13	settings	setting	NOUN
ejpam-642	400	14	with	with	ADP
ejpam-642	400	15	mixture	mixture	NOUN
ejpam-642	400	16	modelm11	modelm11	NOUN
ejpam-642	400	17	.	.	PUNCT
ejpam-642	401	1	however	however	ADV
ejpam-642	401	2	,	,	PUNCT
ejpam-642	401	3	if	if	SCONJ
ejpam-642	401	4	fields	field	NOUN
ejpam-642	401	5	are	be	AUX
ejpam-642	401	6	independent	independent	ADJ
ejpam-642	401	7	,	,	PUNCT
ejpam-642	401	8	we	we	PRON
ejpam-642	401	9	can	can	AUX
ejpam-642	401	10	generate	generate	VERB
ejpam-642	401	11	four	four	NUM
ejpam-642	401	12	independent	independent	ADJ
ejpam-642	401	13	binary	binary	ADJ
ejpam-642	401	14	columns	column	NOUN
ejpam-642	401	15	of	of	ADP
ejpam-642	401	16	length	length	NOUN
ejpam-642	401	17	being	be	AUX
ejpam-642	401	18	the	the	DET
ejpam-642	401	19	specified	specify	VERB
ejpam-642	401	20	total	total	ADJ
ejpam-642	401	21	count	count	NOUN
ejpam-642	401	22	n0	n0	PROPN
ejpam-642	401	23	.	.	PUNCT
ejpam-642	402	1	each	each	DET
ejpam-642	402	2	column	column	NOUN
ejpam-642	402	3	is	be	AUX
ejpam-642	402	4	a	a	DET
ejpam-642	402	5	series	series	NOUN
ejpam-642	402	6	of	of	ADP
ejpam-642	402	7	n0	n0	ADJ
ejpam-642	402	8	bernoulli	bernoulli	PROPN
ejpam-642	402	9	trials	trial	NOUN
ejpam-642	402	10	with	with	ADP
ejpam-642	402	11	success	success	NOUN
ejpam-642	402	12	probability	probability	NOUN
ejpam-642	402	13	which	which	PRON
ejpam-642	402	14	can	can	AUX
ejpam-642	402	15	be	be	AUX
ejpam-642	402	16	readily	readily	ADV
ejpam-642	402	17	computed	compute	VERB
ejpam-642	402	18	according	accord	VERB
ejpam-642	402	19	to	to	ADP
ejpam-642	402	20	(	(	PUNCT
ejpam-642	402	21	12	12	NUM
ejpam-642	402	22	)	)	PUNCT
ejpam-642	402	23	.	.	PUNCT
ejpam-642	403	1	information	information	NOUN
ejpam-642	403	2	of	of	ADP
ejpam-642	403	3	a	a	PRON
ejpam-642	403	4	and	and	CCONJ
ejpam-642	403	5	b	b	NOUN
ejpam-642	403	6	will	will	AUX
ejpam-642	403	7	be	be	AUX
ejpam-642	403	8	used	use	VERB
ejpam-642	403	9	in	in	ADP
ejpam-642	403	10	the	the	DET
ejpam-642	403	11	choice	choice	NOUN
ejpam-642	403	12	of	of	ADP
ejpam-642	403	13	initial	initial	ADJ
ejpam-642	403	14	values	value	NOUN
ejpam-642	403	15	in	in	ADP
ejpam-642	403	16	the	the	DET
ejpam-642	403	17	simplified	simplified	ADJ
ejpam-642	403	18	method	method	NOUN
ejpam-642	403	19	for	for	ADP
ejpam-642	403	20	the	the	DET
ejpam-642	403	21	em	em	PROPN
ejpam-642	403	22	algorithm	algorithm	NOUN
ejpam-642	403	23	.	.	PUNCT
ejpam-642	404	1	we	we	PRON
ejpam-642	404	2	apply	apply	VERB
ejpam-642	404	3	the	the	DET
ejpam-642	404	4	proposed	propose	VERB
ejpam-642	404	5	building	building	NOUN
ejpam-642	404	6	strategy	strategy	NOUN
ejpam-642	404	7	to	to	PART
ejpam-642	404	8	select	select	VERB
ejpam-642	404	9	a	a	DET
ejpam-642	404	10	loglinear	loglinear	NOUN
ejpam-642	404	11	mixture	mixture	NOUN
ejpam-642	404	12	model	model	NOUN
ejpam-642	404	13	for	for	ADP
ejpam-642	404	14	each	each	DET
ejpam-642	404	15	simulation	simulation	NOUN
ejpam-642	404	16	case	case	NOUN
ejpam-642	404	17	.	.	PUNCT
ejpam-642	405	1	the	the	DET
ejpam-642	405	2	significance	significance	NOUN
ejpam-642	405	3	level	level	NOUN
ejpam-642	405	4	is	be	AUX
ejpam-642	405	5	set	set	VERB
ejpam-642	405	6	as	as	ADP
ejpam-642	405	7	α	α	NOUN
ejpam-642	405	8	=	=	SYM
ejpam-642	405	9	5	5	NUM
ejpam-642	405	10	%	%	NOUN
ejpam-642	405	11	.	.	PUNCT
ejpam-642	406	1	the	the	DET
ejpam-642	406	2	following	follow	VERB
ejpam-642	406	3	lists	list	VERB
ejpam-642	406	4	the	the	DET
ejpam-642	406	5	model	model	NOUN
ejpam-642	406	6	building	building	NOUN
ejpam-642	406	7	processes	process	NOUN
ejpam-642	406	8	and	and	CCONJ
ejpam-642	406	9	the	the	DET
ejpam-642	406	10	finally	finally	ADV
ejpam-642	406	11	selected	select	VERB
ejpam-642	406	12	models	model	NOUN
ejpam-642	406	13	for	for	ADP
ejpam-642	406	14	all	all	DET
ejpam-642	406	15	the	the	DET
ejpam-642	406	16	12	12	NUM
ejpam-642	406	17	simulation	simulation	NOUN
ejpam-642	406	18	cases	case	NOUN
ejpam-642	406	19	.	.	PUNCT
ejpam-642	407	1	due	due	ADP
ejpam-642	407	2	to	to	ADP
ejpam-642	407	3	space	space	NOUN
ejpam-642	407	4	limitation	limitation	NOUN
ejpam-642	407	5	,	,	PUNCT
ejpam-642	407	6	we	we	PRON
ejpam-642	407	7	omit	omit	VERB
ejpam-642	407	8	the	the	DET
ejpam-642	407	9	details	detail	NOUN
ejpam-642	407	10	of	of	ADP
ejpam-642	407	11	the	the	DET
ejpam-642	407	12	estimated	estimate	VERB
ejpam-642	407	13	models	model	NOUN
ejpam-642	407	14	.	.	PUNCT
ejpam-642	408	1	•	•	NUM
ejpam-642	408	2	for	for	ADP
ejpam-642	408	3	the	the	DET
ejpam-642	408	4	true	true	ADJ
ejpam-642	408	5	loglinear	loglinear	NOUN
ejpam-642	408	6	mixture	mixture	NOUN
ejpam-642	408	7	model	model	NOUN
ejpam-642	408	8	beingm00	beingm00	NOUN
ejpam-642	408	9	(	(	PUNCT
ejpam-642	408	10	refer	refer	VERB
ejpam-642	408	11	to	to	ADP
ejpam-642	408	12	table	table	NOUN
ejpam-642	408	13	3	3	NUM
ejpam-642	408	14	)	)	PUNCT
ejpam-642	408	15	,	,	PUNCT
ejpam-642	408	16	settings	setting	NOUN
ejpam-642	408	17	1	1	NUM
ejpam-642	408	18	,	,	PUNCT
ejpam-642	408	19	2	2	NUM
ejpam-642	408	20	,	,	PUNCT
ejpam-642	408	21	3	3	NUM
ejpam-642	408	22	,	,	PUNCT
ejpam-642	408	23	5	5	NUM
ejpam-642	408	24	and	and	CCONJ
ejpam-642	408	25	6	6	NUM
ejpam-642	408	26	lead	lead	NOUN
ejpam-642	408	27	tom00	tom00	NOUN
ejpam-642	408	28	while	while	SCONJ
ejpam-642	408	29	setting	set	VERB
ejpam-642	408	30	4	4	NUM
ejpam-642	408	31	yieldsm00	yieldsm00	NOUN
ejpam-642	408	32	→m10	→m10	NUM
ejpam-642	408	33	.	.	PUNCT
ejpam-642	409	1	thus	thus	ADV
ejpam-642	409	2	,	,	PUNCT
ejpam-642	409	3	from	from	ADP
ejpam-642	409	4	the	the	DET
ejpam-642	409	5	viewpoint	viewpoint	NOUN
ejpam-642	409	6	of	of	ADP
ejpam-642	409	7	model	model	NOUN
ejpam-642	409	8	fitting	fit	VERB
ejpam-642	409	9	,	,	PUNCT
ejpam-642	409	10	setting	set	VERB
ejpam-642	409	11	4	4	NUM
ejpam-642	409	12	is	be	AUX
ejpam-642	409	13	a	a	DET
ejpam-642	409	14	little	little	ADJ
ejpam-642	409	15	bit	bit	NOUN
ejpam-642	409	16	overfitted	overfitte	VERB
ejpam-642	409	17	with	with	ADP
ejpam-642	409	18	false	false	ADJ
ejpam-642	409	19	interactions	interaction	NOUN
ejpam-642	409	20	for	for	ADP
ejpam-642	409	21	u	u	PROPN
ejpam-642	409	22	.	.	PUNCT
ejpam-642	410	1	•	•	NUM
ejpam-642	410	2	for	for	ADP
ejpam-642	410	3	the	the	DET
ejpam-642	410	4	true	true	ADJ
ejpam-642	410	5	loglinear	loglinear	NOUN
ejpam-642	410	6	mixture	mixture	NOUN
ejpam-642	410	7	model	model	NOUN
ejpam-642	410	8	beingm11	beingm11	NOUN
ejpam-642	410	9	(	(	PUNCT
ejpam-642	410	10	refer	refer	VERB
ejpam-642	410	11	to	to	ADP
ejpam-642	410	12	table	table	NOUN
ejpam-642	410	13	4	4	NUM
ejpam-642	410	14	)	)	PUNCT
ejpam-642	410	15	,	,	PUNCT
ejpam-642	410	16	settings	setting	NOUN
ejpam-642	410	17	1	1	NUM
ejpam-642	410	18	,	,	PUNCT
ejpam-642	410	19	2	2	NUM
ejpam-642	410	20	,	,	PUNCT
ejpam-642	410	21	4	4	NUM
ejpam-642	410	22	and	and	CCONJ
ejpam-642	410	23	6	6	NUM
ejpam-642	410	24	proceed	proceed	VERB
ejpam-642	410	25	asm00→m10→m11	asm00→m10→m11	PROPN
ejpam-642	410	26	while	while	SCONJ
ejpam-642	410	27	settings	setting	NOUN
ejpam-642	410	28	3	3	NUM
ejpam-642	410	29	and	and	CCONJ
ejpam-642	410	30	5	5	NUM
ejpam-642	410	31	result	result	NOUN
ejpam-642	410	32	inm00→m10	inm00→m10	NOUN
ejpam-642	410	33	.	.	PUNCT
ejpam-642	411	1	hence	hence	ADV
ejpam-642	411	2	,	,	PUNCT
ejpam-642	411	3	settings	setting	NOUN
ejpam-642	411	4	3	3	NUM
ejpam-642	411	5	and	and	CCONJ
ejpam-642	411	6	5	5	NUM
ejpam-642	411	7	are	be	AUX
ejpam-642	411	8	a	a	DET
ejpam-642	411	9	little	little	ADJ
ejpam-642	411	10	bit	bit	NOUN
ejpam-642	411	11	underfitted	underfitte	VERB
ejpam-642	411	12	without	without	ADP
ejpam-642	411	13	capturing	capture	VERB
ejpam-642	411	14	the	the	DET
ejpam-642	411	15	two	two	NUM
ejpam-642	411	16	-	-	PUNCT
ejpam-642	411	17	factor	factor	NOUN
ejpam-642	411	18	interactions	interaction	NOUN
ejpam-642	411	19	for	for	ADP
ejpam-642	411	20	m	m	PRON
ejpam-642	411	21	.	.	PUNCT
ejpam-642	412	1	regarding	regard	VERB
ejpam-642	412	2	the	the	DET
ejpam-642	412	3	true	true	ADJ
ejpam-642	412	4	mixture	mixture	NOUN
ejpam-642	412	5	model	model	NOUN
ejpam-642	412	6	being	be	AUX
ejpam-642	412	7	m00	m00	NOUN
ejpam-642	412	8	,	,	PUNCT
ejpam-642	412	9	five	five	NUM
ejpam-642	412	10	out	out	ADP
ejpam-642	412	11	of	of	ADP
ejpam-642	412	12	six	six	NUM
ejpam-642	412	13	simulation	simulation	NOUN
ejpam-642	412	14	cases	case	NOUN
ejpam-642	412	15	yield	yield	VERB
ejpam-642	412	16	the	the	DET
ejpam-642	412	17	selected	select	VERB
ejpam-642	412	18	models	model	NOUN
ejpam-642	412	19	consistent	consistent	ADJ
ejpam-642	412	20	with	with	ADP
ejpam-642	412	21	the	the	DET
ejpam-642	412	22	true	true	ADJ
ejpam-642	412	23	one	one	NOUN
ejpam-642	412	24	,	,	PUNCT
ejpam-642	412	25	however	however	ADV
ejpam-642	412	26	,	,	PUNCT
ejpam-642	412	27	one	one	NUM
ejpam-642	412	28	case	case	NOUN
ejpam-642	412	29	results	result	VERB
ejpam-642	412	30	in	in	ADP
ejpam-642	412	31	overfitting	overfitte	VERB
ejpam-642	412	32	.	.	PUNCT
ejpam-642	413	1	further	further	ADJ
ejpam-642	413	2	investigation	investigation	NOUN
ejpam-642	413	3	shows	show	VERB
ejpam-642	413	4	that	that	SCONJ
ejpam-642	413	5	the	the	DET
ejpam-642	413	6	p	p	NOUN
ejpam-642	413	7	-	-	PUNCT
ejpam-642	413	8	values	value	NOUN
ejpam-642	413	9	in	in	ADP
ejpam-642	413	10	testing	testing	NOUN
ejpam-642	413	11	h0	h0	NOUN
ejpam-642	413	12	:	:	PUNCT
ejpam-642	413	13	m00	m00	VERB
ejpam-642	414	1	=	=	NOUN
ejpam-642	414	2	m10	m10	PROPN
ejpam-642	414	3	for	for	ADP
ejpam-642	414	4	the	the	DET
ejpam-642	414	5	setting	set	VERB
ejpam-642	414	6	a	a	PRON
ejpam-642	414	7	=	=	SYM
ejpam-642	414	8	400	400	NUM
ejpam-642	414	9	and	and	CCONJ
ejpam-642	414	10	b	b	NOUN
ejpam-642	415	1	=	=	SYM
ejpam-642	415	2	400	400	NUM
ejpam-642	415	3	are	be	AUX
ejpam-642	415	4	0.047	0.047	NUM
ejpam-642	415	5	(	(	PUNCT
ejpam-642	415	6	<	<	X
ejpam-642	415	7	0.05	0.05	NUM
ejpam-642	415	8	)	)	PUNCT
ejpam-642	415	9	when	when	SCONJ
ejpam-642	415	10	nm	nm	ADV
ejpam-642	415	11	=	=	NOUN
ejpam-642	415	12	40	40	NUM
ejpam-642	415	13	(	(	PUNCT
ejpam-642	415	14	setting	set	VERB
ejpam-642	415	15	3	3	NUM
ejpam-642	415	16	)	)	PUNCT
ejpam-642	415	17	and	and	CCONJ
ejpam-642	415	18	0.105	0.105	NUM
ejpam-642	415	19	(	(	PUNCT
ejpam-642	415	20	>	>	X
ejpam-642	415	21	0.05	0.05	NUM
ejpam-642	415	22	)	)	PUNCT
ejpam-642	415	23	when	when	SCONJ
ejpam-642	415	24	nm	nm	ADV
ejpam-642	415	25	=	=	VERB
ejpam-642	415	26	200	200	NUM
ejpam-642	415	27	(	(	PUNCT
ejpam-642	415	28	setting	set	VERB
ejpam-642	415	29	4	4	NUM
ejpam-642	415	30	)	)	PUNCT
ejpam-642	415	31	.	.	PUNCT
ejpam-642	416	1	note	note	VERB
ejpam-642	416	2	that	that	SCONJ
ejpam-642	416	3	setting	set	VERB
ejpam-642	416	4	4	4	NUM
ejpam-642	416	5	has	have	VERB
ejpam-642	416	6	the	the	DET
ejpam-642	416	7	highest	high	ADJ
ejpam-642	416	8	proportion	proportion	NOUN
ejpam-642	416	9	of	of	ADP
ejpam-642	416	10	matched	match	VERB
ejpam-642	416	11	pairs	pair	NOUN
ejpam-642	416	12	,	,	PUNCT
ejpam-642	416	13	with	with	ADP
ejpam-642	416	14	half	half	NOUN
ejpam-642	416	15	of	of	ADP
ejpam-642	416	16	records	record	NOUN
ejpam-642	416	17	in	in	ADP
ejpam-642	416	18	each	each	DET
ejpam-642	416	19	file	file	NOUN
ejpam-642	416	20	corresponding	correspond	VERB
ejpam-642	416	21	to	to	ADP
ejpam-642	416	22	the	the	DET
ejpam-642	416	23	same	same	ADJ
ejpam-642	416	24	objects	object	NOUN
ejpam-642	416	25	.	.	PUNCT
ejpam-642	417	1	this	this	DET
ejpam-642	417	2	highly	highly	ADV
ejpam-642	417	3	matching	matching	ADJ
ejpam-642	417	4	feature	feature	NOUN
ejpam-642	417	5	might	might	AUX
ejpam-642	417	6	cause	cause	VERB
ejpam-642	417	7	model	model	NOUN
ejpam-642	417	8	inflation	inflation	NOUN
ejpam-642	417	9	for	for	ADP
ejpam-642	417	10	component	component	NOUN
ejpam-642	417	11	u	u	PROPN
ejpam-642	417	12	when	when	SCONJ
ejpam-642	417	13	fields	field	NOUN
ejpam-642	417	14	are	be	AUX
ejpam-642	417	15	actually	actually	ADV
ejpam-642	417	16	independent	independent	ADJ
ejpam-642	417	17	.	.	PUNCT
ejpam-642	418	1	as	as	ADP
ejpam-642	418	2	to	to	ADP
ejpam-642	418	3	the	the	DET
ejpam-642	418	4	true	true	ADJ
ejpam-642	418	5	mixture	mixture	NOUN
ejpam-642	418	6	model	model	NOUN
ejpam-642	418	7	beingm11	beingm11	NOUN
ejpam-642	418	8	,	,	PUNCT
ejpam-642	418	9	four	four	NUM
ejpam-642	418	10	out	out	ADP
ejpam-642	418	11	of	of	ADP
ejpam-642	418	12	six	six	NUM
ejpam-642	418	13	cases	case	NOUN
ejpam-642	418	14	lead	lead	VERB
ejpam-642	418	15	to	to	ADP
ejpam-642	418	16	models	model	NOUN
ejpam-642	418	17	consistent	consistent	ADJ
ejpam-642	418	18	with	with	ADP
ejpam-642	418	19	the	the	DET
ejpam-642	418	20	true	true	ADJ
ejpam-642	418	21	ones	one	NOUN
ejpam-642	418	22	,	,	PUNCT
ejpam-642	418	23	but	but	CCONJ
ejpam-642	418	24	two	two	NUM
ejpam-642	418	25	cases	case	NOUN
ejpam-642	418	26	fail	fail	VERB
ejpam-642	418	27	to	to	PART
ejpam-642	418	28	capture	capture	VERB
ejpam-642	418	29	the	the	DET
ejpam-642	418	30	two	two	NUM
ejpam-642	418	31	-	-	PUNCT
ejpam-642	418	32	factor	factor	NOUN
ejpam-642	418	33	interactions	interaction	NOUN
ejpam-642	418	34	in	in	ADP
ejpam-642	418	35	the	the	DET
ejpam-642	418	36	loglinear	loglinear	NOUN
ejpam-642	418	37	model	model	NOUN
ejpam-642	418	38	for	for	ADP
ejpam-642	418	39	m	m	PROPN
ejpam-642	418	40	,	,	PUNCT
ejpam-642	418	41	leading	lead	VERB
ejpam-642	418	42	to	to	ADP
ejpam-642	418	43	underfitting	underfitte	VERB
ejpam-642	418	44	for	for	ADP
ejpam-642	418	45	the	the	DET
ejpam-642	418	46	mixture	mixture	NOUN
ejpam-642	418	47	.	.	PUNCT
ejpam-642	419	1	note	note	VERB
ejpam-642	419	2	that	that	SCONJ
ejpam-642	419	3	both	both	PRON
ejpam-642	419	4	overfitting	overfitte	VERB
ejpam-642	419	5	and	and	CCONJ
ejpam-642	419	6	underfitting	underfitting	NOUN
ejpam-642	419	7	happen	happen	VERB
ejpam-642	419	8	when	when	SCONJ
ejpam-642	419	9	the	the	DET
ejpam-642	419	10	number	number	NOUN
ejpam-642	419	11	of	of	ADP
ejpam-642	419	12	matched	match	VERB
ejpam-642	419	13	pairs	pair	NOUN
ejpam-642	419	14	nm	nm	ADV
ejpam-642	419	15	is	be	AUX
ejpam-642	419	16	small	small	ADJ
ejpam-642	419	17	.	.	PUNCT
ejpam-642	420	1	hence	hence	ADV
ejpam-642	420	2	,	,	PUNCT
ejpam-642	420	3	empirically	empirically	ADV
ejpam-642	420	4	,	,	PUNCT
ejpam-642	420	5	it	it	PRON
ejpam-642	420	6	may	may	AUX
ejpam-642	420	7	be	be	AUX
ejpam-642	420	8	more	more	ADV
ejpam-642	420	9	difficult	difficult	ADJ
ejpam-642	420	10	to	to	PART
ejpam-642	420	11	find	find	VERB
ejpam-642	420	12	higher	high	ADJ
ejpam-642	420	13	order	order	NOUN
ejpam-642	420	14	interactions	interaction	NOUN
ejpam-642	420	15	for	for	ADP
ejpam-642	420	16	the	the	DET
ejpam-642	420	17	m	m	NOUN
ejpam-642	420	18	component	component	NOUN
ejpam-642	420	19	than	than	ADP
ejpam-642	420	20	for	for	ADP
ejpam-642	420	21	the	the	DET
ejpam-642	420	22	u	u	PROPN
ejpam-642	420	23	component	component	NOUN
ejpam-642	420	24	in	in	ADP
ejpam-642	420	25	the	the	DET
ejpam-642	420	26	mixture	mixture	NOUN
ejpam-642	420	27	.	.	PUNCT
ejpam-642	421	1	curiously	curiously	ADV
ejpam-642	421	2	,	,	PUNCT
ejpam-642	421	3	we	we	PRON
ejpam-642	421	4	have	have	AUX
ejpam-642	421	5	not	not	PART
ejpam-642	421	6	observed	observe	VERB
ejpam-642	421	7	any	any	DET
ejpam-642	421	8	overfitting	overfitting	NOUN
ejpam-642	421	9	when	when	SCONJ
ejpam-642	421	10	the	the	DET
ejpam-642	421	11	true	true	ADJ
ejpam-642	421	12	mixture	mixture	NOUN
ejpam-642	421	13	model	model	NOUN
ejpam-642	421	14	ism11	ism11	PROPN
ejpam-642	421	15	in	in	ADP
ejpam-642	421	16	a	a	DET
ejpam-642	421	17	larger	large	ADJ
ejpam-642	421	18	scale	scale	NOUN
ejpam-642	421	19	investigation	investigation	NOUN
ejpam-642	421	20	.	.	PUNCT
ejpam-642	422	1	these	these	DET
ejpam-642	422	2	observed	observe	VERB
ejpam-642	422	3	phenomena	phenomenon	NOUN
ejpam-642	422	4	indicate	indicate	VERB
ejpam-642	422	5	that	that	SCONJ
ejpam-642	422	6	the	the	DET
ejpam-642	422	7	proposed	propose	VERB
ejpam-642	422	8	model	model	NOUN
ejpam-642	422	9	building	building	NOUN
ejpam-642	422	10	strategy	strategy	NOUN
ejpam-642	422	11	may	may	AUX
ejpam-642	422	12	not	not	PART
ejpam-642	422	13	capture	capture	VERB
ejpam-642	422	14	the	the	DET
ejpam-642	422	15	true	true	ADJ
ejpam-642	422	16	underlying	underlie	VERB
ejpam-642	422	17	mixture	mixture	NOUN
ejpam-642	422	18	model	model	NOUN
ejpam-642	422	19	perfectly	perfectly	ADV
ejpam-642	422	20	.	.	PUNCT
ejpam-642	423	1	sometimes	sometimes	ADV
ejpam-642	423	2	it	it	PRON
ejpam-642	423	3	yields	yield	VERB
ejpam-642	423	4	a	a	DET
ejpam-642	423	5	slight	slight	ADJ
ejpam-642	423	6	overfitting	overfitting	NOUN
ejpam-642	423	7	or	or	CCONJ
ejpam-642	423	8	underfitting	underfitting	NOUN
ejpam-642	423	9	for	for	ADP
ejpam-642	423	10	one	one	NUM
ejpam-642	423	11	of	of	ADP
ejpam-642	423	12	the	the	DET
ejpam-642	423	13	mixture	mixture	NOUN
ejpam-642	423	14	components	component	NOUN
ejpam-642	423	15	.	.	PUNCT
ejpam-642	424	1	since	since	SCONJ
ejpam-642	424	2	in	in	ADP
ejpam-642	424	3	some	some	DET
ejpam-642	424	4	cases	case	NOUN
ejpam-642	424	5	the	the	DET
ejpam-642	424	6	selected	select	VERB
ejpam-642	424	7	models	model	NOUN
ejpam-642	424	8	deviate	deviate	VERB
ejpam-642	424	9	from	from	ADP
ejpam-642	424	10	the	the	DET
ejpam-642	424	11	true	true	ADJ
ejpam-642	424	12	models	model	NOUN
ejpam-642	424	13	,	,	PUNCT
ejpam-642	424	14	a	a	DET
ejpam-642	424	15	natural	natural	ADJ
ejpam-642	424	16	question	question	NOUN
ejpam-642	424	17	is	be	AUX
ejpam-642	424	18	how	how	SCONJ
ejpam-642	424	19	the	the	DET
ejpam-642	424	20	corresponding	correspond	VERB
ejpam-642	424	21	cell	cell	NOUN
ejpam-642	424	22	probabilities	probability	NOUN
ejpam-642	424	23	deviate	deviate	VERB
ejpam-642	424	24	from	from	ADP
ejpam-642	424	25	the	the	DET
ejpam-642	424	26	truth	truth	NOUN
ejpam-642	424	27	.	.	PUNCT
ejpam-642	425	1	this	this	PRON
ejpam-642	425	2	is	be	AUX
ejpam-642	425	3	because	because	SCONJ
ejpam-642	425	4	the	the	DET
ejpam-642	425	5	ultimate	ultimate	ADJ
ejpam-642	425	6	concern	concern	NOUN
ejpam-642	425	7	in	in	ADP
ejpam-642	425	8	record	record	NOUN
ejpam-642	425	9	linkage	linkage	NOUN
ejpam-642	425	10	is	be	AUX
ejpam-642	425	11	the	the	DET
ejpam-642	425	12	probabilities	probability	NOUN
ejpam-642	425	13	from	from	ADP
ejpam-642	425	14	u	u	NOUN
ejpam-642	425	15	and	and	CCONJ
ejpam-642	425	16	m	m	VERB
ejpam-642	425	17	for	for	ADP
ejpam-642	425	18	all	all	DET
ejpam-642	425	19	record	record	NOUN
ejpam-642	425	20	pairs	pair	NOUN
ejpam-642	425	21	,	,	PUNCT
ejpam-642	425	22	as	as	ADV
ejpam-642	425	23	well	well	ADV
ejpam-642	425	24	as	as	ADP
ejpam-642	425	25	their	their	PRON
ejpam-642	425	26	matching	matching	NOUN
ejpam-642	425	27	weights	weight	NOUN
ejpam-642	425	28	.	.	PUNCT
ejpam-642	426	1	thus	thus	ADV
ejpam-642	426	2	,	,	PUNCT
ejpam-642	426	3	it	it	PRON
ejpam-642	426	4	is	be	AUX
ejpam-642	426	5	necessary	necessary	ADJ
ejpam-642	426	6	to	to	PART
ejpam-642	426	7	investigate	investigate	VERB
ejpam-642	426	8	the	the	DET
ejpam-642	426	9	cell	cell	NOUN
ejpam-642	426	10	probability	probability	NOUN
ejpam-642	426	11	differences	difference	NOUN
ejpam-642	426	12	.	.	PUNCT
ejpam-642	427	1	in	in	ADP
ejpam-642	427	2	addition	addition	NOUN
ejpam-642	427	3	,	,	PUNCT
ejpam-642	427	4	we	we	PRON
ejpam-642	427	5	can	can	AUX
ejpam-642	427	6	compare	compare	VERB
ejpam-642	427	7	the	the	DET
ejpam-642	427	8	selected	select	VERB
ejpam-642	427	9	model	model	NOUN
ejpam-642	427	10	with	with	ADP
ejpam-642	427	11	the	the	DET
ejpam-642	427	12	fitted	fit	VERB
ejpam-642	427	13	starting	starting	NOUN
ejpam-642	427	14	modelm00	modelm00	NOUN
ejpam-642	427	15	from	from	ADP
ejpam-642	427	16	the	the	DET
ejpam-642	427	17	r.	r.	PROPN
ejpam-642	427	18	zhu	zhu	PROPN
ejpam-642	427	19	,	,	PUNCT
ejpam-642	427	20	j.	j.	PROPN
ejpam-642	427	21	zhang	zhang	PROPN
ejpam-642	427	22	,	,	PUNCT
ejpam-642	427	23	d.	d.	PROPN
ejpam-642	427	24	zhang	zhang	PROPN
ejpam-642	427	25	,	,	PUNCT
ejpam-642	427	26	g.	g.	PROPN
ejpam-642	427	27	yan	yan	PROPN
ejpam-642	427	28	/	/	SYM
ejpam-642	427	29	eur	eur	PROPN
ejpam-642	427	30	.	.	PUNCT
ejpam-642	428	1	j.	j.	PROPN
ejpam-642	428	2	pure	pure	PROPN
ejpam-642	428	3	appl	appl	PROPN
ejpam-642	428	4	.	.	PROPN
ejpam-642	428	5	math	math	PROPN
ejpam-642	428	6	,	,	PUNCT
ejpam-642	428	7	3	3	NUM
ejpam-642	428	8	(	(	PUNCT
ejpam-642	428	9	2010	2010	NUM
ejpam-642	428	10	)	)	PUNCT
ejpam-642	428	11	,	,	PUNCT
ejpam-642	428	12	141	141	NUM
ejpam-642	428	13	-	-	SYM
ejpam-642	428	14	162	162	NUM
ejpam-642	428	15	156	156	NUM
ejpam-642	428	16	viewpoint	viewpoint	NOUN
ejpam-642	428	17	of	of	ADP
ejpam-642	428	18	cell	cell	NOUN
ejpam-642	428	19	probabilities	probability	NOUN
ejpam-642	428	20	.	.	PUNCT
ejpam-642	429	1	this	this	PRON
ejpam-642	429	2	is	be	AUX
ejpam-642	429	3	to	to	PART
ejpam-642	429	4	see	see	VERB
ejpam-642	429	5	whether	whether	SCONJ
ejpam-642	429	6	or	or	CCONJ
ejpam-642	429	7	not	not	PART
ejpam-642	429	8	our	our	PRON
ejpam-642	429	9	proposed	propose	VERB
ejpam-642	429	10	data	data	NOUN
ejpam-642	429	11	-	-	PUNCT
ejpam-642	429	12	driven	drive	VERB
ejpam-642	429	13	strategy	strategy	NOUN
ejpam-642	429	14	can	can	AUX
ejpam-642	429	15	beat	beat	VERB
ejpam-642	429	16	or	or	CCONJ
ejpam-642	429	17	at	at	ADP
ejpam-642	429	18	least	least	ADJ
ejpam-642	429	19	is	be	AUX
ejpam-642	429	20	equivalent	equivalent	ADJ
ejpam-642	429	21	to	to	ADP
ejpam-642	429	22	the	the	DET
ejpam-642	429	23	subjective	subjective	ADJ
ejpam-642	429	24	specification	specification	NOUN
ejpam-642	429	25	with	with	ADP
ejpam-642	429	26	naive	naive	ADJ
ejpam-642	429	27	field	field	NOUN
ejpam-642	429	28	independence	independence	NOUN
ejpam-642	429	29	used	use	VERB
ejpam-642	429	30	in	in	ADP
ejpam-642	429	31	some	some	DET
ejpam-642	429	32	record	record	NOUN
ejpam-642	429	33	linkage	linkage	NOUN
ejpam-642	429	34	software	software	NOUN
ejpam-642	429	35	.	.	PUNCT
ejpam-642	430	1	we	we	PRON
ejpam-642	430	2	show	show	VERB
ejpam-642	430	3	the	the	DET
ejpam-642	430	4	details	detail	NOUN
ejpam-642	430	5	in	in	ADP
ejpam-642	430	6	table	table	NOUN
ejpam-642	430	7	5	5	NUM
ejpam-642	430	8	for	for	ADP
ejpam-642	430	9	setting	set	VERB
ejpam-642	430	10	5	5	NUM
ejpam-642	430	11	when	when	SCONJ
ejpam-642	430	12	the	the	DET
ejpam-642	430	13	true	true	ADJ
ejpam-642	430	14	mixture	mixture	NOUN
ejpam-642	430	15	model	model	NOUN
ejpam-642	430	16	is	be	AUX
ejpam-642	430	17	m11	m11	ADJ
ejpam-642	430	18	.	.	PUNCT
ejpam-642	431	1	in	in	ADP
ejpam-642	431	2	this	this	DET
ejpam-642	431	3	simulation	simulation	NOUN
ejpam-642	431	4	case	case	NOUN
ejpam-642	431	5	,	,	PUNCT
ejpam-642	431	6	the	the	DET
ejpam-642	431	7	selected	select	VERB
ejpam-642	431	8	model	model	NOUN
ejpam-642	431	9	by	by	ADP
ejpam-642	431	10	the	the	DET
ejpam-642	431	11	data	data	NOUN
ejpam-642	431	12	-	-	PUNCT
ejpam-642	431	13	driven	drive	VERB
ejpam-642	431	14	strategy	strategy	NOUN
ejpam-642	431	15	is	be	AUX
ejpam-642	431	16	underfitting	underfitte	VERB
ejpam-642	431	17	in	in	ADP
ejpam-642	431	18	the	the	DET
ejpam-642	431	19	m	m	NOUN
ejpam-642	431	20	component	component	NOUN
ejpam-642	431	21	,	,	PUNCT
ejpam-642	431	22	not	not	PART
ejpam-642	431	23	perfectly	perfectly	ADV
ejpam-642	431	24	consistent	consistent	ADJ
ejpam-642	431	25	with	with	ADP
ejpam-642	431	26	the	the	DET
ejpam-642	431	27	true	true	ADJ
ejpam-642	431	28	underlying	underlie	VERB
ejpam-642	431	29	model	model	NOUN
ejpam-642	431	30	.	.	PUNCT
ejpam-642	432	1	in	in	ADP
ejpam-642	432	2	table	table	NOUN
ejpam-642	432	3	5	5	NUM
ejpam-642	432	4	,	,	PUNCT
ejpam-642	432	5	•	•	ADP
ejpam-642	432	6	pd(s	pd(s	NUM
ejpam-642	432	7	,	,	PUNCT
ejpam-642	432	8	t	t	PROPN
ejpam-642	432	9	|u	|u	NOUN
ejpam-642	432	10	)	)	PUNCT
ejpam-642	432	11	and	and	CCONJ
ejpam-642	432	12	pd(s	pd(s	NUM
ejpam-642	432	13	,	,	PUNCT
ejpam-642	432	14	t	t	PROPN
ejpam-642	432	15	|m	|m	NOUN
ejpam-642	432	16	)	)	PUNCT
ejpam-642	432	17	denote	denote	NOUN
ejpam-642	432	18	differeences	differeence	NOUN
ejpam-642	432	19	in	in	ADP
ejpam-642	432	20	cell	cell	NOUN
ejpam-642	432	21	probabilities	probability	NOUN
ejpam-642	432	22	between	between	ADP
ejpam-642	432	23	the	the	DET
ejpam-642	432	24	selected	select	VERB
ejpam-642	432	25	model	model	NOUN
ejpam-642	432	26	by	by	ADP
ejpam-642	432	27	the	the	DET
ejpam-642	432	28	data	data	NOUN
ejpam-642	432	29	-	-	PUNCT
ejpam-642	432	30	driven	drive	VERB
ejpam-642	432	31	strategy	strategy	NOUN
ejpam-642	432	32	and	and	CCONJ
ejpam-642	432	33	the	the	DET
ejpam-642	432	34	true	true	ADJ
ejpam-642	432	35	model	model	NOUN
ejpam-642	432	36	for	for	ADP
ejpam-642	432	37	component	component	NOUN
ejpam-642	432	38	u	u	NOUN
ejpam-642	432	39	and	and	CCONJ
ejpam-642	432	40	m	m	VERB
ejpam-642	432	41	respectively	respectively	ADV
ejpam-642	432	42	,	,	PUNCT
ejpam-642	432	43	•	•	PRON
ejpam-642	432	44	pd(i	pd(i	X
ejpam-642	432	45	,	,	PUNCT
ejpam-642	432	46	t	t	PROPN
ejpam-642	432	47	|u	|u	NOUN
ejpam-642	432	48	)	)	PUNCT
ejpam-642	432	49	and	and	CCONJ
ejpam-642	432	50	pd(i	pd(i	X
ejpam-642	432	51	,	,	PUNCT
ejpam-642	432	52	t	t	PROPN
ejpam-642	432	53	|m	|m	NOUN
ejpam-642	432	54	)	)	PUNCT
ejpam-642	432	55	denote	denote	NOUN
ejpam-642	432	56	differeences	differeence	NOUN
ejpam-642	432	57	in	in	ADP
ejpam-642	432	58	cell	cell	NOUN
ejpam-642	432	59	probabilities	probability	NOUN
ejpam-642	432	60	between	between	ADP
ejpam-642	432	61	the	the	DET
ejpam-642	432	62	simplest	simple	ADJ
ejpam-642	432	63	starting	starting	NOUN
ejpam-642	432	64	modelm00	modelm00	NOUN
ejpam-642	432	65	and	and	CCONJ
ejpam-642	432	66	the	the	DET
ejpam-642	432	67	true	true	ADJ
ejpam-642	432	68	model	model	NOUN
ejpam-642	432	69	for	for	ADP
ejpam-642	432	70	component	component	NOUN
ejpam-642	432	71	u	u	NOUN
ejpam-642	432	72	and	and	CCONJ
ejpam-642	432	73	m	m	PROPN
ejpam-642	432	74	respectively.table	respectively.table	ADJ
ejpam-642	432	75	5	5	NUM
ejpam-642	432	76	:	:	PUNCT
ejpam-642	432	77	comparisons	comparison	NOUN
ejpam-642	432	78	of	of	ADP
ejpam-642	432	79	ell	ell	PROPN
ejpam-642	432	80	probabilities	probability	NOUN
ejpam-642	432	81	between	between	ADP
ejpam-642	432	82	two	two	NUM
ejpam-642	432	83	�	�	NOUN
ejpam-642	432	84	tted	tte	VERB
ejpam-642	432	85	models	model	NOUN
ejpam-642	432	86	and	and	CCONJ
ejpam-642	432	87	the	the	DET
ejpam-642	432	88	true	true	ADJ
ejpam-642	432	89	model	model	NOUN
ejpam-642	432	90	,	,	PUNCT
ejpam-642	432	91	m11	m11	NOUN
ejpam-642	432	92	,	,	PUNCT
ejpam-642	432	93	for	for	ADP
ejpam-642	432	94	setting5	setting5	NOUN
ejpam-642	432	95	.	.	PUNCT
ejpam-642	433	1	cell	cell	NOUN
ejpam-642	433	2	pd(s	pd(s	PROPN
ejpam-642	433	3	,	,	PUNCT
ejpam-642	433	4	t	t	PROPN
ejpam-642	433	5	|u	|u	NOUN
ejpam-642	433	6	)	)	PUNCT
ejpam-642	433	7	pd(i	pd(i	X
ejpam-642	433	8	,	,	PUNCT
ejpam-642	433	9	t	t	PROPN
ejpam-642	433	10	|u	|u	ADJ
ejpam-642	433	11	)	)	PUNCT
ejpam-642	433	12	pd(s	pd(s	NUM
ejpam-642	433	13	,	,	PUNCT
ejpam-642	433	14	t	t	PROPN
ejpam-642	433	15	|m	|m	NOUN
ejpam-642	433	16	)	)	PUNCT
ejpam-642	433	17	pd(i	pd(i	NOUN
ejpam-642	433	18	,	,	PUNCT
ejpam-642	433	19	t	t	PROPN
ejpam-642	433	20	|m	|m	NOUN
ejpam-642	433	21	)	)	PUNCT
ejpam-642	433	22	1	1	NUM
ejpam-642	433	23	-2.710e-05	-2.710e-05	SYM
ejpam-642	433	24	2.130e-04	2.130e-04	NUM
ejpam-642	433	25	-2.233e-04	-2.233e-04	PUNCT
ejpam-642	433	26	4.487e-02	4.487e-02	NUM
ejpam-642	433	27	2	2	NUM
ejpam-642	433	28	-2.547e-07	-2.547e-07	PROPN
ejpam-642	433	29	-3.767e-06	-3.767e-06	INTJ
ejpam-642	433	30	-5.058e-03	-5.058e-03	PROPN
ejpam-642	433	31	6.675e-03	6.675e-03	NUM
ejpam-642	433	32	3	3	NUM
ejpam-642	433	33	2.973e-05	2.973e-05	NUM
ejpam-642	433	34	-8.666e-05	-8.666e-05	PUNCT
ejpam-642	433	35	-5.604e-04	-5.604e-04	NUM
ejpam-642	433	36	4.272e-01	4.272e-01	NUM
ejpam-642	433	37	4	4	NUM
ejpam-642	433	38	2.606e-06	2.606e-06	NUM
ejpam-642	433	39	-2.535e-05	-2.535e-05	SYM
ejpam-642	433	40	-7.308e-02	-7.308e-02	PROPN
ejpam-642	433	41	3.753e-02	3.753e-02	NUM
ejpam-642	433	42	5	5	NUM
ejpam-642	433	43	2.865e-06	2.865e-06	NUM
ejpam-642	433	44	-1.356e-06	-1.356e-06	NOUN
ejpam-642	434	1	-1.375e-02	-1.375e-02	PROPN
ejpam-642	435	1	9.176e-03	9.176e-03	NUM
ejpam-642	435	2	6	6	NUM
ejpam-642	435	3	5.579e-08	5.579e-08	NUM
ejpam-642	435	4	-2.366e-07	-2.366e-07	SYM
ejpam-642	436	1	-2.407e-02	-2.407e-02	NUM
ejpam-642	436	2	1.810e-02	1.810e-02	NUM
ejpam-642	436	3	7	7	NUM
ejpam-642	436	4	-4.573e-06	-4.573e-06	NOUN
ejpam-642	436	5	-8.575e-05	-8.575e-05	PUNCT
ejpam-642	436	6	-4.588e-04	-4.588e-04	PROPN
ejpam-642	436	7	2.170e-01	2.170e-01	NUM
ejpam-642	436	8	8	8	NUM
ejpam-642	436	9	4.212e-07	4.212e-07	NUM
ejpam-642	436	10	-1.728e-06	-1.728e-06	PUNCT
ejpam-642	437	1	-4.552e-03	-4.552e-03	PRON
ejpam-642	438	1	5.192e-02	5.192e-02	NUM
ejpam-642	438	2	9	9	NUM
ejpam-642	438	3	-2.060e-06	-2.060e-06	PUNCT
ejpam-642	438	4	-3.412e-06	-3.412e-06	NOUN
ejpam-642	438	5	-5.084e-05	-5.084e-05	PUNCT
ejpam-642	439	1	5.027e-03	5.027e-03	NUM
ejpam-642	439	2	10	10	NUM
ejpam-642	439	3	-5.656e-08	-5.656e-08	NOUN
ejpam-642	439	4	-5.617e-08	-5.617e-08	INTJ
ejpam-642	439	5	-6.071e-04	-6.071e-04	ADJ
ejpam-642	439	6	7.143e-04	7.143e-04	NUM
ejpam-642	439	7	11	11	NUM
ejpam-642	439	8	-5.870e-08	-5.870e-08	INTJ
ejpam-642	439	9	-1.122e-07	-1.122e-07	PROPN
ejpam-642	439	10	-9.592e-03	-9.592e-03	PROPN
ejpam-642	440	1	3.858e-02	3.858e-02	NUM
ejpam-642	440	2	12	12	NUM
ejpam-642	440	3	-5.544e-08	-5.544e-08	NOUN
ejpam-642	440	4	-5.543e-08	-5.543e-08	PUNCT
ejpam-642	440	5	1.877e-01	1.877e-01	NUM
ejpam-642	440	6	-6.532e-01	-6.532e-01	ADP
ejpam-642	440	7	13	13	NUM
ejpam-642	440	8	-2.490e-07	-2.490e-07	PROPN
ejpam-642	440	9	-2.843e-06	-2.843e-06	PUNCT
ejpam-642	440	10	-1.126e-02	-1.126e-02	PROPN
ejpam-642	440	11	-8.674e-03	-8.674e-03	PROPN
ejpam-642	440	12	14	14	NUM
ejpam-642	440	13	-7.634e-08	-7.634e-08	PROPN
ejpam-642	440	14	-7.634e-08	-7.634e-08	INTJ
ejpam-642	440	15	-1.039e-02	-1.039e-02	PROPN
ejpam-642	440	16	-9.719e-03	-9.719e-03	SYM
ejpam-642	440	17	15	15	NUM
ejpam-642	440	18	-1.884e-07	-1.884e-07	PROPN
ejpam-642	440	19	-5.871e-07	-5.871e-07	PROPN
ejpam-642	440	20	-2.825e-02	-2.825e-02	PROPN
ejpam-642	440	21	-3.754e-03	-3.754e-03	PROPN
ejpam-642	440	22	16	16	NUM
ejpam-642	440	23	-1.007e-06	-1.007e-06	INTJ
ejpam-642	440	24	-1.007e-06	-1.007e-06	INTJ
ejpam-642	440	25	-5.842e-03	-5.842e-03	INTJ
ejpam-642	440	26	-1.452e-01	-1.452e-01	NOUN
ejpam-642	440	27	the	the	PRON
ejpam-642	440	28	smaller	small	ADJ
ejpam-642	440	29	the	the	DET
ejpam-642	440	30	absolute	absolute	ADJ
ejpam-642	440	31	value	value	NOUN
ejpam-642	440	32	in	in	ADP
ejpam-642	440	33	the	the	DET
ejpam-642	440	34	table	table	NOUN
ejpam-642	440	35	,	,	PUNCT
ejpam-642	440	36	the	the	PRON
ejpam-642	440	37	closer	close	ADV
ejpam-642	440	38	the	the	DET
ejpam-642	440	39	corresponding	correspond	VERB
ejpam-642	440	40	cell	cell	NOUN
ejpam-642	440	41	probability	probability	NOUN
ejpam-642	440	42	is	be	AUX
ejpam-642	440	43	to	to	ADP
ejpam-642	440	44	the	the	DET
ejpam-642	440	45	truth	truth	NOUN
ejpam-642	440	46	.	.	PUNCT
ejpam-642	441	1	table	table	NOUN
ejpam-642	441	2	5	5	NUM
ejpam-642	441	3	shows	show	VERB
ejpam-642	441	4	that	that	SCONJ
ejpam-642	441	5	compared	compare	VERB
ejpam-642	441	6	with	with	ADP
ejpam-642	441	7	the	the	DET
ejpam-642	441	8	simplest	simple	ADJ
ejpam-642	441	9	starting	starting	NOUN
ejpam-642	441	10	modelm00	modelm00	NOUN
ejpam-642	441	11	,	,	PUNCT
ejpam-642	441	12	the	the	DET
ejpam-642	441	13	model	model	NOUN
ejpam-642	441	14	selected	select	VERB
ejpam-642	441	15	by	by	ADP
ejpam-642	441	16	the	the	DET
ejpam-642	441	17	data	data	NOUN
ejpam-642	441	18	-	-	PUNCT
ejpam-642	441	19	driven	drive	VERB
ejpam-642	441	20	strategy	strategy	NOUN
ejpam-642	441	21	has	have	VERB
ejpam-642	441	22	smaller	small	ADJ
ejpam-642	441	23	absolute	absolute	ADJ
ejpam-642	441	24	values	value	NOUN
ejpam-642	441	25	of	of	ADP
ejpam-642	441	26	probability	probability	NOUN
ejpam-642	441	27	differences	difference	NOUN
ejpam-642	441	28	in	in	ADP
ejpam-642	441	29	almost	almost	ADV
ejpam-642	441	30	all	all	PRON
ejpam-642	441	31	cells	cell	NOUN
ejpam-642	441	32	for	for	ADP
ejpam-642	441	33	the	the	DET
ejpam-642	441	34	estimated	estimate	VERB
ejpam-642	441	35	u	u	NOUN
ejpam-642	441	36	component	component	NOUN
ejpam-642	441	37	,	,	PUNCT
ejpam-642	441	38	and	and	CCONJ
ejpam-642	441	39	in	in	ADP
ejpam-642	441	40	the	the	DET
ejpam-642	441	41	majority	majority	NOUN
ejpam-642	441	42	of	of	ADP
ejpam-642	441	43	cells	cell	NOUN
ejpam-642	441	44	for	for	ADP
ejpam-642	441	45	the	the	DET
ejpam-642	441	46	estimated	estimate	VERB
ejpam-642	441	47	m	m	NOUN
ejpam-642	441	48	component	component	NOUN
ejpam-642	441	49	,	,	PUNCT
ejpam-642	441	50	thus	thus	ADV
ejpam-642	441	51	,	,	PUNCT
ejpam-642	441	52	leading	lead	VERB
ejpam-642	441	53	to	to	ADP
ejpam-642	441	54	a	a	DET
ejpam-642	441	55	better	well	ADJ
ejpam-642	441	56	overall	overall	ADJ
ejpam-642	441	57	performance	performance	NOUN
ejpam-642	441	58	in	in	ADP
ejpam-642	441	59	this	this	DET
ejpam-642	441	60	particular	particular	ADJ
ejpam-642	441	61	simulation	simulation	NOUN
ejpam-642	441	62	case	case	NOUN
ejpam-642	441	63	.	.	PUNCT
ejpam-642	442	1	this	this	PRON
ejpam-642	442	2	means	mean	VERB
ejpam-642	442	3	that	that	SCONJ
ejpam-642	442	4	the	the	DET
ejpam-642	442	5	selected	select	VERB
ejpam-642	442	6	model	model	NOUN
ejpam-642	442	7	is	be	AUX
ejpam-642	442	8	closer	close	ADJ
ejpam-642	442	9	to	to	ADP
ejpam-642	442	10	the	the	DET
ejpam-642	442	11	true	true	ADJ
ejpam-642	442	12	underlying	underlie	VERB
ejpam-642	442	13	model	model	NOUN
ejpam-642	442	14	than	than	ADP
ejpam-642	442	15	the	the	DET
ejpam-642	442	16	fittedm00	fittedm00	NOUN
ejpam-642	442	17	.	.	PUNCT
ejpam-642	443	1	similar	similar	ADJ
ejpam-642	443	2	patterns	pattern	NOUN
ejpam-642	443	3	have	have	AUX
ejpam-642	443	4	been	be	AUX
ejpam-642	443	5	found	find	VERB
ejpam-642	443	6	in	in	ADP
ejpam-642	443	7	other	other	ADJ
ejpam-642	443	8	simulation	simulation	NOUN
ejpam-642	443	9	cases	case	NOUN
ejpam-642	443	10	when	when	SCONJ
ejpam-642	443	11	the	the	DET
ejpam-642	443	12	true	true	ADJ
ejpam-642	443	13	model	model	NOUN
ejpam-642	443	14	is	be	AUX
ejpam-642	443	15	m11	m11	ADJ
ejpam-642	443	16	.	.	PUNCT
ejpam-642	444	1	therefore	therefore	ADV
ejpam-642	444	2	,	,	PUNCT
ejpam-642	444	3	the	the	DET
ejpam-642	444	4	proposed	propose	VERB
ejpam-642	444	5	strategy	strategy	NOUN
ejpam-642	444	6	does	do	AUX
ejpam-642	444	7	have	have	VERB
ejpam-642	444	8	improvement	improvement	NOUN
ejpam-642	444	9	in	in	ADP
ejpam-642	444	10	the	the	DET
ejpam-642	444	11	model	model	NOUN
ejpam-642	444	12	building	building	NOUN
ejpam-642	444	13	compared	compare	VERB
ejpam-642	444	14	r.	r.	PROPN
ejpam-642	444	15	zhu	zhu	PROPN
ejpam-642	444	16	,	,	PUNCT
ejpam-642	444	17	j.	j.	PROPN
ejpam-642	444	18	zhang	zhang	PROPN
ejpam-642	444	19	,	,	PUNCT
ejpam-642	444	20	d.	d.	PROPN
ejpam-642	444	21	zhang	zhang	PROPN
ejpam-642	444	22	,	,	PUNCT
ejpam-642	444	23	g.	g.	PROPN
ejpam-642	444	24	yan	yan	PROPN
ejpam-642	444	25	/	/	SYM
ejpam-642	444	26	eur	eur	PROPN
ejpam-642	444	27	.	.	PUNCT
ejpam-642	445	1	j.	j.	PROPN
ejpam-642	445	2	pure	pure	PROPN
ejpam-642	445	3	appl	appl	PROPN
ejpam-642	445	4	.	.	PROPN
ejpam-642	445	5	math	math	PROPN
ejpam-642	445	6	,	,	PUNCT
ejpam-642	445	7	3	3	NUM
ejpam-642	445	8	(	(	PUNCT
ejpam-642	445	9	2010	2010	NUM
ejpam-642	445	10	)	)	PUNCT
ejpam-642	445	11	,	,	PUNCT
ejpam-642	445	12	141	141	NUM
ejpam-642	445	13	-	-	SYM
ejpam-642	445	14	162	162	NUM
ejpam-642	445	15	157	157	NUM
ejpam-642	445	16	with	with	ADP
ejpam-642	445	17	the	the	DET
ejpam-642	445	18	naive	naive	ADJ
ejpam-642	445	19	specification	specification	NOUN
ejpam-642	445	20	of	of	ADP
ejpam-642	445	21	field	field	NOUN
ejpam-642	445	22	independence	independence	NOUN
ejpam-642	445	23	(	(	PUNCT
ejpam-642	445	24	i.e.	i.e.	X
ejpam-642	445	25	,	,	PUNCT
ejpam-642	445	26	starting	start	VERB
ejpam-642	445	27	model	model	NOUN
ejpam-642	445	28	m00	m00	PROPN
ejpam-642	445	29	)	)	PUNCT
ejpam-642	445	30	.	.	PUNCT
ejpam-642	446	1	due	due	ADP
ejpam-642	446	2	to	to	ADP
ejpam-642	446	3	space	space	NOUN
ejpam-642	446	4	limitation	limitation	NOUN
ejpam-642	446	5	,	,	PUNCT
ejpam-642	446	6	we	we	PRON
ejpam-642	446	7	summarize	summarize	VERB
ejpam-642	446	8	only	only	ADV
ejpam-642	446	9	the	the	DET
ejpam-642	446	10	cell	cell	NOUN
ejpam-642	446	11	probability	probability	NOUN
ejpam-642	446	12	differences	difference	NOUN
ejpam-642	446	13	between	between	ADP
ejpam-642	446	14	the	the	DET
ejpam-642	446	15	selected	select	VERB
ejpam-642	446	16	and	and	CCONJ
ejpam-642	446	17	the	the	DET
ejpam-642	446	18	true	true	ADJ
ejpam-642	446	19	models	model	NOUN
ejpam-642	446	20	,	,	PUNCT
ejpam-642	446	21	as	as	ADV
ejpam-642	446	22	well	well	ADV
ejpam-642	446	23	as	as	ADP
ejpam-642	446	24	those	those	PRON
ejpam-642	446	25	between	between	ADP
ejpam-642	446	26	the	the	DET
ejpam-642	446	27	fittedm00	fittedm00	NOUN
ejpam-642	446	28	and	and	CCONJ
ejpam-642	446	29	the	the	DET
ejpam-642	446	30	true	true	ADJ
ejpam-642	446	31	model	model	NOUN
ejpam-642	446	32	in	in	ADP
ejpam-642	446	33	table	table	NOUN
ejpam-642	446	34	6	6	NUM
ejpam-642	446	35	for	for	ADP
ejpam-642	446	36	u	u	NOUN
ejpam-642	446	37	and	and	CCONJ
ejpam-642	446	38	m	m	VERB
ejpam-642	446	39	respectively	respectively	ADV
ejpam-642	446	40	,	,	PUNCT
ejpam-642	446	41	using	use	VERB
ejpam-642	446	42	the	the	DET
ejpam-642	446	43	following	follow	VERB
ejpam-642	446	44	measures	measure	NOUN
ejpam-642	446	45	for	for	ADP
ejpam-642	446	46	two	two	NUM
ejpam-642	446	47	general	general	ADJ
ejpam-642	446	48	loglinear	loglinear	NOUN
ejpam-642	446	49	mixture	mixture	NOUN
ejpam-642	446	50	models	model	NOUN
ejpam-642	446	51	:	:	PUNCT
ejpam-642	446	52	d(m	d(m	NOUN
ejpam-642	446	53	(	(	PUNCT
ejpam-642	446	54	0),m	0),m	PROPN
ejpam-642	446	55	(	(	PUNCT
ejpam-642	446	56	1	1	NUM
ejpam-642	446	57	)	)	PUNCT
ejpam-642	446	58	|	|	ADV
ejpam-642	446	59	u	u	NOUN
ejpam-642	446	60	)	)	PUNCT
ejpam-642	446	61	=	=	SYM
ejpam-642	446	62	∑	∑	PUNCT
ejpam-642	446	63	i∈all	i∈all	PROPN
ejpam-642	446	64	cells	cell	NOUN
ejpam-642	446	65	|p(0	|p(0	VERB
ejpam-642	446	66	)	)	PUNCT
ejpam-642	446	67	i	i	PRON
ejpam-642	446	68	(	(	PUNCT
ejpam-642	446	69	u)−	u)−	PROPN
ejpam-642	446	70	p	p	NOUN
ejpam-642	446	71	(	(	PUNCT
ejpam-642	446	72	1	1	NUM
ejpam-642	446	73	)	)	PUNCT
ejpam-642	446	74	i	i	PRON
ejpam-642	446	75	(	(	PUNCT
ejpam-642	446	76	u)|	u)|	NOUN
ejpam-642	446	77	,	,	PUNCT
ejpam-642	446	78	d(m	d(m	PROPN
ejpam-642	446	79	(	(	PUNCT
ejpam-642	446	80	0),m	0),m	PROPN
ejpam-642	446	81	(	(	PUNCT
ejpam-642	446	82	1	1	NUM
ejpam-642	446	83	)	)	PUNCT
ejpam-642	446	84	|	|	ADV
ejpam-642	446	85	m	m	NOUN
ejpam-642	446	86	)	)	PUNCT
ejpam-642	446	87	=	=	PUNCT
ejpam-642	446	88	∑	∑	PUNCT
ejpam-642	446	89	i∈all	i∈all	PROPN
ejpam-642	446	90	cells	cell	NOUN
ejpam-642	446	91	|p(0	|p(0	VERB
ejpam-642	446	92	)	)	PUNCT
ejpam-642	446	93	i	i	PRON
ejpam-642	446	94	(	(	PUNCT
ejpam-642	446	95	m)−	m)−	PROPN
ejpam-642	446	96	p	p	X
ejpam-642	446	97	(	(	PUNCT
ejpam-642	446	98	1	1	NUM
ejpam-642	446	99	)	)	PUNCT
ejpam-642	446	100	i	i	PRON
ejpam-642	446	101	(	(	PUNCT
ejpam-642	446	102	m)|	m)|	NOUN
ejpam-642	446	103	.	.	PUNCT
ejpam-642	447	1	here	here	ADV
ejpam-642	447	2	p	p	X
ejpam-642	447	3	(	(	PUNCT
ejpam-642	447	4	j	j	NOUN
ejpam-642	447	5	)	)	PUNCT
ejpam-642	448	1	i	i	PRON
ejpam-642	448	2	is	be	AUX
ejpam-642	448	3	the	the	DET
ejpam-642	448	4	probability	probability	NOUN
ejpam-642	448	5	of	of	ADP
ejpam-642	448	6	cell	cell	NOUN
ejpam-642	448	7	i	i	PRON
ejpam-642	448	8	in	in	ADP
ejpam-642	448	9	modelm	modelm	PROPN
ejpam-642	448	10	(	(	PUNCT
ejpam-642	448	11	j	j	PROPN
ejpam-642	448	12	)	)	PUNCT
ejpam-642	448	13	.	.	PUNCT
ejpam-642	449	1	in	in	ADP
ejpam-642	449	2	addition	addition	NOUN
ejpam-642	449	3	,	,	PUNCT
ejpam-642	449	4	we	we	PRON
ejpam-642	449	5	also	also	ADV
ejpam-642	449	6	compare	compare	VERB
ejpam-642	449	7	the	the	DET
ejpam-642	449	8	estimated	estimate	VERB
ejpam-642	449	9	mixing	mix	VERB
ejpam-642	449	10	proportion	proportion	NOUN
ejpam-642	449	11	π̂(s	π̂(	NOUN
ejpam-642	449	12	)	)	PUNCT
ejpam-642	449	13	from	from	ADP
ejpam-642	449	14	the	the	DET
ejpam-642	449	15	data	data	NOUN
ejpam-642	449	16	-	-	PUNCT
ejpam-642	449	17	driven	drive	VERB
ejpam-642	449	18	strategy	strategy	NOUN
ejpam-642	449	19	and	and	CCONJ
ejpam-642	449	20	π̂(i	π̂(i	NOUN
ejpam-642	449	21	)	)	PUNCT
ejpam-642	449	22	from	from	ADP
ejpam-642	449	23	the	the	DET
ejpam-642	449	24	field	field	NOUN
ejpam-642	449	25	independence	independence	NOUN
ejpam-642	449	26	specification	specification	NOUN
ejpam-642	449	27	with	with	ADP
ejpam-642	449	28	the	the	DET
ejpam-642	449	29	true	true	ADJ
ejpam-642	449	30	π	π	NOUN
ejpam-642	449	31	:	:	PUNCT
ejpam-642	449	32	π̂(s)−π	π̂(s)−π	VERB
ejpam-642	449	33	and	and	CCONJ
ejpam-642	449	34	π̂(i)−π.table	π̂(i)−π.table	ADJ
ejpam-642	449	35	6	6	NUM
ejpam-642	449	36	:	:	PUNCT
ejpam-642	449	37	overall	overall	ADJ
ejpam-642	449	38	�	�	PROPN
ejpam-642	449	39	tting	tte	VERB
ejpam-642	449	40	omparisons	omparison	NOUN
ejpam-642	449	41	between	between	ADP
ejpam-642	449	42	two	two	NUM
ejpam-642	449	43	�	�	NOUN
ejpam-642	449	44	tted	tte	VERB
ejpam-642	449	45	models	model	NOUN
ejpam-642	449	46	and	and	CCONJ
ejpam-642	449	47	the	the	DET
ejpam-642	449	48	true	true	ADJ
ejpam-642	449	49	model	model	NOUN
ejpam-642	449	50	.	.	PUNCT
ejpam-642	450	1	simulation	simulation	NOUN
ejpam-642	450	2	case	case	NOUN
ejpam-642	450	3	d(s	d(s	PROPN
ejpam-642	450	4	,	,	PUNCT
ejpam-642	450	5	t	t	PROPN
ejpam-642	450	6	|u	|u	ADJ
ejpam-642	450	7	)	)	PUNCT
ejpam-642	450	8	d(i	d(i	PROPN
ejpam-642	450	9	,	,	PUNCT
ejpam-642	450	10	t	t	PROPN
ejpam-642	450	11	|u	|u	ADJ
ejpam-642	450	12	)	)	PUNCT
ejpam-642	450	13	d(s	d(s	PROPN
ejpam-642	450	14	,	,	PUNCT
ejpam-642	450	15	t	t	PROPN
ejpam-642	450	16	|m	|m	NOUN
ejpam-642	450	17	)	)	PUNCT
ejpam-642	450	18	d(i	d(i	PROPN
ejpam-642	450	19	,	,	PUNCT
ejpam-642	450	20	t	t	PROPN
ejpam-642	450	21	|m	|m	NOUN
ejpam-642	450	22	)	)	PUNCT
ejpam-642	450	23	π̂(s)−π	π̂(s)−π	VERB
ejpam-642	450	24	π̂(i)−π	π̂(i)−π	NOUN
ejpam-642	450	25	setting	set	VERB
ejpam-642	450	26	1,m00	1,m00	PROPN
ejpam-642	450	27	1.668e-03	1.668e-03	NUM
ejpam-642	450	28	1.668e-03	1.668e-03	NUM
ejpam-642	450	29	1.967e-01	1.967e-01	NUM
ejpam-642	450	30	1.967e-01	1.967e-01	NUM
ejpam-642	451	1	-7.0e-06	-7.0e-06	PROPN
ejpam-642	451	2	-7.0e-06	-7.0e-06	PUNCT
ejpam-642	451	3	setting	set	VERB
ejpam-642	451	4	2,m00	2,m00	NUM
ejpam-642	451	5	1.947e-03	1.947e-03	NUM
ejpam-642	451	6	1.947e-03	1.947e-03	NUM
ejpam-642	451	7	2.200e-01	2.200e-01	NUM
ejpam-642	451	8	2.200e-01	2.200e-01	NUM
ejpam-642	451	9	4.3e-05	4.3e-05	NUM
ejpam-642	451	10	4.3e-05	4.3e-05	NUM
ejpam-642	451	11	setting	set	VERB
ejpam-642	451	12	3,m00	3,m00	NUM
ejpam-642	451	13	2.059e-03	2.059e-03	NUM
ejpam-642	452	1	2.059e-03	2.059e-03	NUM
ejpam-642	452	2	3.287e-01	3.287e-01	NUM
ejpam-642	452	3	3.287e-01	3.287e-01	NUM
ejpam-642	452	4	-2.4e-05	-2.4e-05	PROPN
ejpam-642	452	5	-2.4e-05	-2.4e-05	PUNCT
ejpam-642	452	6	setting	set	VERB
ejpam-642	452	7	4,m00	4,m00	NUM
ejpam-642	452	8	2.066e-03	2.066e-03	NUM
ejpam-642	452	9	1.487e-03	1.487e-03	NUM
ejpam-642	452	10	1.124e-01	1.124e-01	NUM
ejpam-642	452	11	1.009e-01	1.009e-01	NUM
ejpam-642	452	12	5.7e-06	5.7e-06	NUM
ejpam-642	452	13	-4.0e-06	-4.0e-06	NOUN
ejpam-642	452	14	setting	set	VERB
ejpam-642	452	15	5,m00	5,m00	NUM
ejpam-642	452	16	6.018e-04	6.018e-04	NUM
ejpam-642	452	17	6.018e-04	6.018e-04	NUM
ejpam-642	452	18	2.037e-01	2.037e-01	NUM
ejpam-642	452	19	2.037e-01	2.037e-01	NUM
ejpam-642	452	20	2.2e-04	2.2e-04	NUM
ejpam-642	452	21	2.2e-04	2.2e-04	NUM
ejpam-642	452	22	setting	set	VERB
ejpam-642	452	23	6,m00	6,m00	NUM
ejpam-642	452	24	5.984e-04	5.984e-04	NUM
ejpam-642	452	25	5.984e-04	5.984e-04	NUM
ejpam-642	452	26	1.293e-01	1.293e-01	NUM
ejpam-642	452	27	1.293e-01	1.293e-01	NUM
ejpam-642	452	28	1.2e-05	1.2e-05	NUM
ejpam-642	452	29	1.2e-05	1.2e-05	NUM
ejpam-642	452	30	setting	set	VERB
ejpam-642	452	31	1,m11	1,m11	NUM
ejpam-642	452	32	3.794e-04	3.794e-04	NUM
ejpam-642	452	33	2.400e-04	2.400e-04	NUM
ejpam-642	452	34	4.097e-01	4.097e-01	NUM
ejpam-642	452	35	1.482e-00	1.482e-00	NUM
ejpam-642	452	36	-1.5e-05	-1.5e-05	PROPN
ejpam-642	453	1	2.1e-04	2.1e-04	NUM
ejpam-642	453	2	setting	set	VERB
ejpam-642	453	3	2,m11	2,m11	NUM
ejpam-642	453	4	6.460e-04	6.460e-04	NUM
ejpam-642	453	5	7.581e-04	7.581e-04	NUM
ejpam-642	453	6	4.143e-01	4.143e-01	NUM
ejpam-642	453	7	5.985e-01	5.985e-01	NUM
ejpam-642	453	8	-3.0e-04	-3.0e-04	PUNCT
ejpam-642	453	9	1.1e-04	1.1e-04	NUM
ejpam-642	453	10	setting	set	VERB
ejpam-642	453	11	3,m11	3,m11	NUM
ejpam-642	453	12	4.167e-04	4.167e-04	NUM
ejpam-642	453	13	7.049e-04	7.049e-04	NUM
ejpam-642	453	14	3.239e-01	3.239e-01	NUM
ejpam-642	453	15	1.103e-00	1.103e-00	NUM
ejpam-642	453	16	-4.5e-05	-4.5e-05	PUNCT
ejpam-642	453	17	1.7e-04	1.7e-04	NUM
ejpam-642	453	18	setting	set	VERB
ejpam-642	453	19	4,m11	4,m11	NUM
ejpam-642	453	20	4.722e-04	4.722e-04	NUM
ejpam-642	453	21	4.678e-04	4.678e-04	NUM
ejpam-642	453	22	2.751e-01	2.751e-01	NUM
ejpam-642	453	23	3.372e-01	3.372e-01	NUM
ejpam-642	453	24	1.4e-04	1.4e-04	NUM
ejpam-642	453	25	4.6e-05	4.6e-05	NUM
ejpam-642	453	26	setting	set	VERB
ejpam-642	453	27	5,m11	5,m11	NUM
ejpam-642	453	28	7.136e-05	7.136e-05	NUM
ejpam-642	453	29	4.260e-04	4.260e-04	NUM
ejpam-642	453	30	3.755e-01	3.755e-01	NUM
ejpam-642	453	31	1.677e-00	1.677e-00	NUM
ejpam-642	453	32	-9.8e-07	-9.8e-07	NOUN
ejpam-642	453	33	2.6e-04	2.6e-04	NUM
ejpam-642	453	34	setting	set	VERB
ejpam-642	453	35	6,m11	6,m11	NUM
ejpam-642	453	36	2.453e-04	2.453e-04	NUM
ejpam-642	453	37	2.525e-04	2.525e-04	NUM
ejpam-642	453	38	4.145e-01	4.145e-01	NUM
ejpam-642	453	39	7.516e-01	7.516e-01	NUM
ejpam-642	453	40	-7.1e-05	-7.1e-05	SYM
ejpam-642	453	41	1.1e-04	1.1e-04	NUM
ejpam-642	453	42	table	table	NOUN
ejpam-642	453	43	6	6	NUM
ejpam-642	453	44	shows	show	VERB
ejpam-642	453	45	that	that	SCONJ
ejpam-642	453	46	when	when	SCONJ
ejpam-642	453	47	the	the	DET
ejpam-642	453	48	underlying	underlie	VERB
ejpam-642	453	49	true	true	ADJ
ejpam-642	453	50	mixture	mixture	NOUN
ejpam-642	453	51	model	model	NOUN
ejpam-642	453	52	ism00	ism00	NOUN
ejpam-642	453	53	,	,	PUNCT
ejpam-642	453	54	the	the	DET
ejpam-642	453	55	data	data	NOUN
ejpam-642	453	56	-	-	PUNCT
ejpam-642	453	57	driven	drive	VERB
ejpam-642	453	58	strategy	strategy	NOUN
ejpam-642	453	59	yields	yield	VERB
ejpam-642	453	60	the	the	DET
ejpam-642	453	61	same	same	ADJ
ejpam-642	453	62	results	result	NOUN
ejpam-642	453	63	as	as	ADP
ejpam-642	453	64	the	the	DET
ejpam-642	453	65	jaro	jaro	PROPN
ejpam-642	453	66	’s	’s	PART
ejpam-642	453	67	method	method	NOUN
ejpam-642	453	68	if	if	SCONJ
ejpam-642	453	69	the	the	DET
ejpam-642	453	70	selected	select	VERB
ejpam-642	453	71	model	model	NOUN
ejpam-642	453	72	is	be	AUX
ejpam-642	453	73	consistent	consistent	ADJ
ejpam-642	453	74	withm00	withm00	NOUN
ejpam-642	453	75	.	.	PUNCT
ejpam-642	454	1	this	this	PRON
ejpam-642	454	2	means	mean	VERB
ejpam-642	454	3	that	that	SCONJ
ejpam-642	454	4	the	the	DET
ejpam-642	454	5	simplified	simplified	ADJ
ejpam-642	454	6	method	method	NOUN
ejpam-642	454	7	for	for	ADP
ejpam-642	454	8	the	the	DET
ejpam-642	454	9	em	em	PROPN
ejpam-642	454	10	algorithm	algorithm	NOUN
ejpam-642	454	11	is	be	AUX
ejpam-642	454	12	equivalent	equivalent	ADJ
ejpam-642	454	13	to	to	ADP
ejpam-642	454	14	the	the	DET
ejpam-642	454	15	jaro	jaro	NOUN
ejpam-642	454	16	’s	’s	PART
ejpam-642	454	17	method	method	NOUN
ejpam-642	454	18	when	when	SCONJ
ejpam-642	454	19	the	the	DET
ejpam-642	454	20	field	field	NOUN
ejpam-642	454	21	independence	independence	NOUN
ejpam-642	454	22	exists	exist	VERB
ejpam-642	454	23	,	,	PUNCT
ejpam-642	454	24	and	and	CCONJ
ejpam-642	454	25	is	be	AUX
ejpam-642	454	26	thus	thus	ADV
ejpam-642	454	27	an	an	DET
ejpam-642	454	28	alternative	alternative	ADJ
ejpam-642	454	29	approach	approach	NOUN
ejpam-642	454	30	.	.	PUNCT
ejpam-642	455	1	however	however	ADV
ejpam-642	455	2	,	,	PUNCT
ejpam-642	455	3	the	the	DET
ejpam-642	455	4	advantage	advantage	NOUN
ejpam-642	455	5	of	of	ADP
ejpam-642	455	6	this	this	DET
ejpam-642	455	7	approach	approach	NOUN
ejpam-642	455	8	is	be	AUX
ejpam-642	455	9	that	that	SCONJ
ejpam-642	455	10	it	it	PRON
ejpam-642	455	11	can	can	AUX
ejpam-642	455	12	handle	handle	VERB
ejpam-642	455	13	general	general	ADJ
ejpam-642	455	14	contingency	contingency	NOUN
ejpam-642	455	15	tables	table	NOUN
ejpam-642	455	16	with	with	ADP
ejpam-642	455	17	factor	factor	NOUN
ejpam-642	455	18	levels	level	NOUN
ejpam-642	455	19	more	more	ADJ
ejpam-642	455	20	than	than	ADP
ejpam-642	455	21	two	two	NUM
ejpam-642	455	22	.	.	PUNCT
ejpam-642	456	1	regarding	regard	VERB
ejpam-642	456	2	the	the	DET
ejpam-642	456	3	performance	performance	NOUN
ejpam-642	456	4	of	of	ADP
ejpam-642	456	5	two	two	NUM
ejpam-642	456	6	fitted	fit	VERB
ejpam-642	456	7	models	model	NOUN
ejpam-642	456	8	in	in	ADP
ejpam-642	456	9	table	table	NOUN
ejpam-642	456	10	6	6	NUM
ejpam-642	456	11	,	,	PUNCT
ejpam-642	456	12	when	when	SCONJ
ejpam-642	456	13	the	the	DET
ejpam-642	456	14	field	field	NOUN
ejpam-642	456	15	independence	independence	NOUN
ejpam-642	456	16	holds	hold	VERB
ejpam-642	456	17	,	,	PUNCT
ejpam-642	456	18	the	the	DET
ejpam-642	456	19	selected	select	VERB
ejpam-642	456	20	models	model	NOUN
ejpam-642	456	21	by	by	ADP
ejpam-642	456	22	the	the	DET
ejpam-642	456	23	data	data	NOUN
ejpam-642	456	24	-	-	PUNCT
ejpam-642	456	25	driven	drive	VERB
ejpam-642	456	26	strategy	strategy	NOUN
ejpam-642	456	27	are	be	AUX
ejpam-642	456	28	almost	almost	ADV
ejpam-642	456	29	the	the	DET
ejpam-642	456	30	same	same	ADJ
ejpam-642	456	31	as	as	ADP
ejpam-642	456	32	the	the	DET
ejpam-642	456	33	fittedm00	fittedm00	NOUN
ejpam-642	456	34	except	except	SCONJ
ejpam-642	456	35	one	one	NUM
ejpam-642	456	36	overfitting	overfitte	VERB
ejpam-642	456	37	case	case	NOUN
ejpam-642	456	38	.	.	PUNCT
ejpam-642	457	1	however	however	ADV
ejpam-642	457	2	,	,	PUNCT
ejpam-642	457	3	they	they	PRON
ejpam-642	457	4	perform	perform	VERB
ejpam-642	457	5	better	well	ADV
ejpam-642	457	6	than	than	ADP
ejpam-642	457	7	the	the	DET
ejpam-642	457	8	fitted	fit	VERB
ejpam-642	457	9	m00	m00	NOUN
ejpam-642	457	10	when	when	SCONJ
ejpam-642	457	11	field	field	NOUN
ejpam-642	457	12	independence	independence	NOUN
ejpam-642	457	13	is	be	AUX
ejpam-642	457	14	violated	violate	VERB
ejpam-642	457	15	.	.	PUNCT
ejpam-642	458	1	as	as	ADP
ejpam-642	458	2	to	to	ADP
ejpam-642	458	3	detecting	detect	VERB
ejpam-642	458	4	interactions	interaction	NOUN
ejpam-642	458	5	from	from	ADP
ejpam-642	458	6	the	the	DET
ejpam-642	458	7	table	table	NOUN
ejpam-642	458	8	data	datum	NOUN
ejpam-642	458	9	,	,	PUNCT
ejpam-642	458	10	our	our	PRON
ejpam-642	458	11	experience	experience	NOUN
ejpam-642	458	12	shows	show	VERB
ejpam-642	458	13	that	that	SCONJ
ejpam-642	458	14	it	it	PRON
ejpam-642	458	15	is	be	AUX
ejpam-642	458	16	not	not	PART
ejpam-642	458	17	easy	easy	ADJ
ejpam-642	458	18	to	to	PART
ejpam-642	458	19	capture	capture	VERB
ejpam-642	458	20	the	the	DET
ejpam-642	458	21	higher	high	ADJ
ejpam-642	458	22	order	order	NOUN
ejpam-642	458	23	interactions	interaction	NOUN
ejpam-642	458	24	accurately	accurately	ADV
ejpam-642	458	25	.	.	PUNCT
ejpam-642	459	1	for	for	ADP
ejpam-642	459	2	this	this	DET
ejpam-642	459	3	reason	reason	NOUN
ejpam-642	459	4	,	,	PUNCT
ejpam-642	459	5	we	we	PRON
ejpam-642	459	6	do	do	AUX
ejpam-642	459	7	not	not	PART
ejpam-642	459	8	investigate	investigate	VERB
ejpam-642	459	9	further	further	ADJ
ejpam-642	459	10	r.	r.	PROPN
ejpam-642	459	11	zhu	zhu	PROPN
ejpam-642	459	12	,	,	PUNCT
ejpam-642	459	13	j.	j.	PROPN
ejpam-642	459	14	zhang	zhang	PROPN
ejpam-642	459	15	,	,	PUNCT
ejpam-642	459	16	d.	d.	PROPN
ejpam-642	459	17	zhang	zhang	PROPN
ejpam-642	459	18	,	,	PUNCT
ejpam-642	459	19	g.	g.	PROPN
ejpam-642	459	20	yan	yan	PROPN
ejpam-642	459	21	/	/	SYM
ejpam-642	459	22	eur	eur	PROPN
ejpam-642	459	23	.	.	PUNCT
ejpam-642	460	1	j.	j.	PROPN
ejpam-642	460	2	pure	pure	PROPN
ejpam-642	460	3	appl	appl	PROPN
ejpam-642	460	4	.	.	PROPN
ejpam-642	460	5	math	math	PROPN
ejpam-642	460	6	,	,	PUNCT
ejpam-642	460	7	3	3	NUM
ejpam-642	460	8	(	(	PUNCT
ejpam-642	460	9	2010	2010	NUM
ejpam-642	460	10	)	)	PUNCT
ejpam-642	460	11	,	,	PUNCT
ejpam-642	460	12	141	141	NUM
ejpam-642	460	13	-	-	SYM
ejpam-642	460	14	162	162	NUM
ejpam-642	460	15	158	158	NUM
ejpam-642	460	16	the	the	DET
ejpam-642	460	17	individual	individual	ADJ
ejpam-642	460	18	interactions	interaction	NOUN
ejpam-642	460	19	in	in	ADP
ejpam-642	460	20	a	a	DET
ejpam-642	460	21	given	give	VERB
ejpam-642	460	22	loglinear	loglinear	NOUN
ejpam-642	460	23	model	model	NOUN
ejpam-642	460	24	in	in	ADP
ejpam-642	460	25	the	the	DET
ejpam-642	460	26	proposed	propose	VERB
ejpam-642	460	27	model	model	NOUN
ejpam-642	460	28	building	building	NOUN
ejpam-642	460	29	strategy	strategy	NOUN
ejpam-642	460	30	,	,	PUNCT
ejpam-642	460	31	and	and	CCONJ
ejpam-642	460	32	it	it	PRON
ejpam-642	460	33	is	be	AUX
ejpam-642	460	34	also	also	ADV
ejpam-642	460	35	the	the	DET
ejpam-642	460	36	partial	partial	ADJ
ejpam-642	460	37	reason	reason	NOUN
ejpam-642	460	38	why	why	SCONJ
ejpam-642	460	39	models	model	NOUN
ejpam-642	460	40	are	be	AUX
ejpam-642	460	41	enlarged	enlarge	VERB
ejpam-642	460	42	with	with	ADP
ejpam-642	460	43	a	a	DET
ejpam-642	460	44	group	group	NOUN
ejpam-642	460	45	of	of	ADP
ejpam-642	460	46	interactions	interaction	NOUN
ejpam-642	460	47	in	in	ADP
ejpam-642	460	48	the	the	DET
ejpam-642	460	49	building	building	NOUN
ejpam-642	460	50	strategy	strategy	NOUN
ejpam-642	460	51	in	in	ADP
ejpam-642	460	52	(	(	PUNCT
ejpam-642	460	53	11	11	NUM
ejpam-642	460	54	)	)	PUNCT
ejpam-642	460	55	.	.	PUNCT
ejpam-642	461	1	even	even	ADV
ejpam-642	461	2	more	more	ADV
ejpam-642	461	3	,	,	PUNCT
ejpam-642	461	4	for	for	ADP
ejpam-642	461	5	a	a	DET
ejpam-642	461	6	data	datum	NOUN
ejpam-642	461	7	set	set	VERB
ejpam-642	461	8	simulated	simulate	VERB
ejpam-642	461	9	from	from	ADP
ejpam-642	461	10	a	a	DET
ejpam-642	461	11	known	know	VERB
ejpam-642	461	12	loglinear	loglinear	NOUN
ejpam-642	461	13	model	model	NOUN
ejpam-642	461	14	with	with	ADP
ejpam-642	461	15	three	three	NUM
ejpam-642	461	16	-	-	PUNCT
ejpam-642	461	17	factor	factor	NOUN
ejpam-642	461	18	interactions	interaction	NOUN
ejpam-642	461	19	,	,	PUNCT
ejpam-642	461	20	the	the	DET
ejpam-642	461	21	fitted	fit	VERB
ejpam-642	461	22	model	model	NOUN
ejpam-642	461	23	does	do	AUX
ejpam-642	461	24	not	not	PART
ejpam-642	461	25	include	include	VERB
ejpam-642	461	26	these	these	DET
ejpam-642	461	27	three	three	NUM
ejpam-642	461	28	-	-	PUNCT
ejpam-642	461	29	factor	factor	NOUN
ejpam-642	461	30	interactions	interaction	NOUN
ejpam-642	461	31	although	although	SCONJ
ejpam-642	461	32	we	we	PRON
ejpam-642	461	33	increase	increase	VERB
ejpam-642	461	34	the	the	DET
ejpam-642	461	35	total	total	ADJ
ejpam-642	461	36	count	count	NOUN
ejpam-642	461	37	as	as	ADV
ejpam-642	461	38	many	many	ADJ
ejpam-642	461	39	as	as	ADP
ejpam-642	461	40	possible	possible	ADJ
ejpam-642	461	41	in	in	ADP
ejpam-642	461	42	r.	r.	PROPN
ejpam-642	461	43	therefore	therefore	ADV
ejpam-642	461	44	,	,	PUNCT
ejpam-642	461	45	in	in	ADP
ejpam-642	461	46	practice	practice	NOUN
ejpam-642	461	47	,	,	PUNCT
ejpam-642	461	48	it	it	PRON
ejpam-642	461	49	is	be	AUX
ejpam-642	461	50	likely	likely	ADJ
ejpam-642	461	51	to	to	PART
ejpam-642	461	52	stop	stop	VERB
ejpam-642	461	53	at	at	ADP
ejpam-642	461	54	a	a	DET
ejpam-642	461	55	model	model	NOUN
ejpam-642	461	56	with	with	ADP
ejpam-642	461	57	lower	low	ADJ
ejpam-642	461	58	order	order	NOUN
ejpam-642	461	59	interactions	interaction	NOUN
ejpam-642	461	60	such	such	ADJ
ejpam-642	461	61	as	as	ADP
ejpam-642	461	62	two	two	NUM
ejpam-642	461	63	-	-	PUNCT
ejpam-642	461	64	factor	factor	NOUN
ejpam-642	461	65	interactions	interaction	NOUN
ejpam-642	461	66	,	,	PUNCT
ejpam-642	461	67	and	and	CCONJ
ejpam-642	461	68	this	this	DET
ejpam-642	461	69	selected	select	VERB
ejpam-642	461	70	model	model	NOUN
ejpam-642	461	71	may	may	AUX
ejpam-642	461	72	be	be	AUX
ejpam-642	461	73	a	a	DET
ejpam-642	461	74	good	good	ADJ
ejpam-642	461	75	approximation	approximation	NOUN
ejpam-642	461	76	to	to	ADP
ejpam-642	461	77	the	the	DET
ejpam-642	461	78	true	true	ADJ
ejpam-642	461	79	model	model	NOUN
ejpam-642	461	80	under	under	ADP
ejpam-642	461	81	the	the	DET
ejpam-642	461	82	given	give	VERB
ejpam-642	461	83	sample	sample	NOUN
ejpam-642	461	84	size	size	NOUN
ejpam-642	461	85	.	.	PUNCT
ejpam-642	462	1	5	5	X
ejpam-642	462	2	.	.	X
ejpam-642	462	3	application	application	NOUN
ejpam-642	462	4	to	to	ADP
ejpam-642	462	5	a	a	DET
ejpam-642	462	6	case	case	NOUN
ejpam-642	462	7	study	study	NOUN
ejpam-642	462	8	we	we	PRON
ejpam-642	462	9	apply	apply	VERB
ejpam-642	462	10	the	the	DET
ejpam-642	462	11	proposed	propose	VERB
ejpam-642	462	12	model	model	NOUN
ejpam-642	462	13	building	building	NOUN
ejpam-642	462	14	strategy	strategy	NOUN
ejpam-642	462	15	to	to	ADP
ejpam-642	462	16	the	the	DET
ejpam-642	462	17	record	record	NOUN
ejpam-642	462	18	linkage	linkage	NOUN
ejpam-642	462	19	case	case	NOUN
ejpam-642	462	20	study	study	NOUN
ejpam-642	462	21	at	at	ADP
ejpam-642	462	22	ssc	ssc	PROPN
ejpam-642	462	23	2006	2006	NUM
ejpam-642	462	24	(	(	PUNCT
ejpam-642	462	25	www.ss	www.ss	PROPN
ejpam-642	462	26	.	.	PUNCT
ejpam-642	463	1	a	a	DET
ejpam-642	463	2	/	/	SYM
ejpam-642	463	3	do	do	AUX
ejpam-642	463	4	uments/	uments/	NUM
ejpam-642	463	5	ase_studies/2006	ase_studies/2006	PROPN
ejpam-642	463	6	/	/	SYM
ejpam-642	463	7	re	re	NOUN
ejpam-642	463	8	ord_index_e.html	ord_index_e.html	NOUN
ejpam-642	463	9	)	)	PUNCT
ejpam-642	463	10	.	.	PUNCT
ejpam-642	464	1	the	the	DET
ejpam-642	464	2	objective	objective	NOUN
ejpam-642	464	3	of	of	ADP
ejpam-642	464	4	the	the	DET
ejpam-642	464	5	case	case	NOUN
ejpam-642	464	6	study	study	NOUN
ejpam-642	464	7	is	be	AUX
ejpam-642	464	8	to	to	PART
ejpam-642	464	9	maintain	maintain	VERB
ejpam-642	464	10	a	a	DET
ejpam-642	464	11	registry	registry	NOUN
ejpam-642	464	12	.	.	PUNCT
ejpam-642	465	1	four	four	NUM
ejpam-642	465	2	completely	completely	ADV
ejpam-642	465	3	synthetic	synthetic	ADJ
ejpam-642	465	4	data	data	NOUN
ejpam-642	465	5	files	file	NOUN
ejpam-642	465	6	named	name	VERB
ejpam-642	465	7	as	as	ADP
ejpam-642	465	8	“	"	PUNCT
ejpam-642	465	9	register	register	VERB
ejpam-642	465	10	”	"	PUNCT
ejpam-642	465	11	,	,	PUNCT
ejpam-642	465	12	“	"	PUNCT
ejpam-642	465	13	births	birth	NOUN
ejpam-642	465	14	”	"	PUNCT
ejpam-642	465	15	,	,	PUNCT
ejpam-642	465	16	“	"	PUNCT
ejpam-642	465	17	drivers	driver	NOUN
ejpam-642	465	18	”	"	PUNCT
ejpam-642	465	19	and	and	CCONJ
ejpam-642	465	20	“	"	PUNCT
ejpam-642	465	21	deaths	death	NOUN
ejpam-642	465	22	”	"	PUNCT
ejpam-642	465	23	were	be	AUX
ejpam-642	465	24	constructed	construct	VERB
ejpam-642	465	25	to	to	PART
ejpam-642	465	26	simulate	simulate	VERB
ejpam-642	465	27	the	the	DET
ejpam-642	465	28	registry	registry	NOUN
ejpam-642	465	29	for	for	ADP
ejpam-642	465	30	a	a	DET
ejpam-642	465	31	sample	sample	NOUN
ejpam-642	465	32	of	of	ADP
ejpam-642	465	33	residents	resident	NOUN
ejpam-642	465	34	of	of	ADP
ejpam-642	465	35	the	the	DET
ejpam-642	465	36	province	province	NOUN
ejpam-642	465	37	of	of	ADP
ejpam-642	465	38	prince	prince	PROPN
ejpam-642	465	39	edward	edward	PROPN
ejpam-642	465	40	island	island	PROPN
ejpam-642	465	41	(	(	PUNCT
ejpam-642	465	42	pei	pei	PROPN
ejpam-642	465	43	)	)	PUNCT
ejpam-642	465	44	,	,	PUNCT
ejpam-642	465	45	canada	canada	PROPN
ejpam-642	465	46	.	.	PUNCT
ejpam-642	466	1	the	the	DET
ejpam-642	466	2	“	"	PUNCT
ejpam-642	466	3	register	register	NOUN
ejpam-642	466	4	”	"	PUNCT
ejpam-642	466	5	file	file	NOUN
ejpam-642	466	6	represents	represent	VERB
ejpam-642	466	7	the	the	DET
ejpam-642	466	8	population	population	NOUN
ejpam-642	466	9	of	of	ADP
ejpam-642	466	10	pei	pei	PROPN
ejpam-642	466	11	,	,	PUNCT
ejpam-642	466	12	containing	contain	VERB
ejpam-642	466	13	a	a	DET
ejpam-642	466	14	unique	unique	ADJ
ejpam-642	466	15	i	i	NOUN
ejpam-642	466	16	d	d	NOUN
ejpam-642	466	17	,	,	PUNCT
ejpam-642	466	18	name	name	NOUN
ejpam-642	466	19	,	,	PUNCT
ejpam-642	466	20	date	date	NOUN
ejpam-642	466	21	of	of	ADP
ejpam-642	466	22	birth	birth	NOUN
ejpam-642	466	23	,	,	PUNCT
ejpam-642	466	24	and	and	CCONJ
ejpam-642	466	25	address	address	VERB
ejpam-642	466	26	information	information	NOUN
ejpam-642	466	27	for	for	ADP
ejpam-642	466	28	residents	resident	NOUN
ejpam-642	466	29	.	.	PUNCT
ejpam-642	467	1	this	this	PRON
ejpam-642	467	2	is	be	AUX
ejpam-642	467	3	the	the	DET
ejpam-642	467	4	file	file	NOUN
ejpam-642	467	5	that	that	SCONJ
ejpam-642	467	6	we	we	PRON
ejpam-642	467	7	are	be	AUX
ejpam-642	467	8	working	work	VERB
ejpam-642	467	9	to	to	PART
ejpam-642	467	10	maintain	maintain	VERB
ejpam-642	467	11	.	.	PUNCT
ejpam-642	468	1	the	the	DET
ejpam-642	468	2	“	"	PUNCT
ejpam-642	468	3	births	birth	NOUN
ejpam-642	468	4	”	"	PUNCT
ejpam-642	468	5	file	file	NOUN
ejpam-642	468	6	records	record	VERB
ejpam-642	468	7	new	new	ADJ
ejpam-642	468	8	entrants	entrant	NOUN
ejpam-642	468	9	into	into	ADP
ejpam-642	468	10	the	the	DET
ejpam-642	468	11	population	population	NOUN
ejpam-642	468	12	.	.	PUNCT
ejpam-642	469	1	they	they	PRON
ejpam-642	469	2	are	be	AUX
ejpam-642	469	3	persons	person	NOUN
ejpam-642	469	4	who	who	PRON
ejpam-642	469	5	enter	enter	VERB
ejpam-642	469	6	the	the	DET
ejpam-642	469	7	population	population	NOUN
ejpam-642	469	8	of	of	ADP
ejpam-642	469	9	interest	interest	NOUN
ejpam-642	469	10	,	,	PUNCT
ejpam-642	469	11	for	for	ADP
ejpam-642	469	12	example	example	NOUN
ejpam-642	469	13	by	by	ADP
ejpam-642	469	14	moving	move	VERB
ejpam-642	469	15	into	into	ADP
ejpam-642	469	16	an	an	DET
ejpam-642	469	17	area	area	NOUN
ejpam-642	469	18	of	of	ADP
ejpam-642	469	19	interest	interest	NOUN
ejpam-642	469	20	,	,	PUNCT
ejpam-642	469	21	or	or	CCONJ
ejpam-642	469	22	attaining	attain	VERB
ejpam-642	469	23	a	a	DET
ejpam-642	469	24	certain	certain	ADJ
ejpam-642	469	25	age	age	NOUN
ejpam-642	469	26	.	.	PUNCT
ejpam-642	470	1	the	the	DET
ejpam-642	470	2	data	datum	NOUN
ejpam-642	470	3	file	file	NOUN
ejpam-642	470	4	on	on	ADP
ejpam-642	470	5	births	birth	NOUN
ejpam-642	470	6	contains	contain	VERB
ejpam-642	470	7	both	both	CCONJ
ejpam-642	470	8	present	present	ADJ
ejpam-642	470	9	and	and	CCONJ
ejpam-642	470	10	previous	previous	ADJ
ejpam-642	470	11	address	address	NOUN
ejpam-642	470	12	information	information	NOUN
ejpam-642	470	13	as	as	ADV
ejpam-642	470	14	well	well	ADV
ejpam-642	470	15	as	as	ADP
ejpam-642	470	16	complete	complete	ADJ
ejpam-642	470	17	name	name	NOUN
ejpam-642	470	18	and	and	CCONJ
ejpam-642	470	19	date	date	NOUN
ejpam-642	470	20	of	of	ADP
ejpam-642	470	21	birth	birth	NOUN
ejpam-642	470	22	information	information	NOUN
ejpam-642	470	23	.	.	PUNCT
ejpam-642	471	1	the	the	DET
ejpam-642	471	2	“	"	PUNCT
ejpam-642	471	3	drivers	driver	NOUN
ejpam-642	471	4	”	"	PUNCT
ejpam-642	471	5	file	file	NOUN
ejpam-642	471	6	provides	provide	VERB
ejpam-642	471	7	moving	move	VERB
ejpam-642	471	8	information	information	NOUN
ejpam-642	471	9	for	for	ADP
ejpam-642	471	10	registry	registry	NOUN
ejpam-642	471	11	.	.	PUNCT
ejpam-642	472	1	people	people	NOUN
ejpam-642	472	2	living	live	VERB
ejpam-642	472	3	in	in	ADP
ejpam-642	472	4	canada	canada	PROPN
ejpam-642	472	5	may	may	AUX
ejpam-642	472	6	take	take	VERB
ejpam-642	472	7	out	out	ADP
ejpam-642	472	8	their	their	PRON
ejpam-642	472	9	first	first	ADJ
ejpam-642	472	10	license	license	NOUN
ejpam-642	472	11	,	,	PUNCT
ejpam-642	472	12	get	get	VERB
ejpam-642	472	13	a	a	DET
ejpam-642	472	14	new	new	ADJ
ejpam-642	472	15	license	license	NOUN
ejpam-642	472	16	(	(	PUNCT
ejpam-642	472	17	out	out	ADP
ejpam-642	472	18	of	of	ADP
ejpam-642	472	19	province	province	NOUN
ejpam-642	472	20	)	)	PUNCT
ejpam-642	472	21	or	or	CCONJ
ejpam-642	472	22	update	update	VERB
ejpam-642	472	23	their	their	PRON
ejpam-642	472	24	present	present	ADJ
ejpam-642	472	25	license	license	NOUN
ejpam-642	472	26	(	(	PUNCT
ejpam-642	472	27	within	within	ADP
ejpam-642	472	28	a	a	DET
ejpam-642	472	29	province	province	NOUN
ejpam-642	472	30	)	)	PUNCT
ejpam-642	472	31	.	.	PUNCT
ejpam-642	473	1	this	this	DET
ejpam-642	473	2	file	file	NOUN
ejpam-642	473	3	has	have	VERB
ejpam-642	473	4	information	information	NOUN
ejpam-642	473	5	on	on	ADP
ejpam-642	473	6	name	name	NOUN
ejpam-642	473	7	,	,	PUNCT
ejpam-642	473	8	address	address	NOUN
ejpam-642	473	9	and	and	CCONJ
ejpam-642	473	10	date	date	NOUN
ejpam-642	473	11	of	of	ADP
ejpam-642	473	12	birth	birth	NOUN
ejpam-642	473	13	.	.	PUNCT
ejpam-642	474	1	the	the	DET
ejpam-642	474	2	“	"	PUNCT
ejpam-642	474	3	deaths	death	NOUN
ejpam-642	474	4	”	"	PUNCT
ejpam-642	474	5	file	file	NOUN
ejpam-642	474	6	contains	contain	VERB
ejpam-642	474	7	name	name	NOUN
ejpam-642	474	8	,	,	PUNCT
ejpam-642	474	9	address	address	NOUN
ejpam-642	474	10	,	,	PUNCT
ejpam-642	474	11	date	date	NOUN
ejpam-642	474	12	of	of	ADP
ejpam-642	474	13	birth	birth	NOUN
ejpam-642	474	14	and	and	CCONJ
ejpam-642	474	15	date	date	NOUN
ejpam-642	474	16	of	of	ADP
ejpam-642	474	17	death	death	NOUN
ejpam-642	474	18	.	.	PUNCT
ejpam-642	475	1	note	note	VERB
ejpam-642	475	2	that	that	SCONJ
ejpam-642	475	3	data	datum	NOUN
ejpam-642	475	4	sources	source	NOUN
ejpam-642	475	5	are	be	AUX
ejpam-642	475	6	updated	update	VERB
ejpam-642	475	7	independently	independently	ADV
ejpam-642	475	8	.	.	PUNCT
ejpam-642	476	1	the	the	DET
ejpam-642	476	2	main	main	ADJ
ejpam-642	476	3	file	file	NOUN
ejpam-642	476	4	is	be	AUX
ejpam-642	476	5	the	the	DET
ejpam-642	476	6	“	"	PUNCT
ejpam-642	476	7	register	register	NOUN
ejpam-642	476	8	”	"	PUNCT
ejpam-642	476	9	file	file	NOUN
ejpam-642	476	10	which	which	PRON
ejpam-642	476	11	needs	need	VERB
ejpam-642	476	12	to	to	PART
ejpam-642	476	13	be	be	AUX
ejpam-642	476	14	updated	update	VERB
ejpam-642	476	15	according	accord	VERB
ejpam-642	476	16	to	to	ADP
ejpam-642	476	17	the	the	DET
ejpam-642	476	18	other	other	ADJ
ejpam-642	476	19	three	three	NUM
ejpam-642	476	20	files	file	NOUN
ejpam-642	476	21	.	.	PUNCT
ejpam-642	477	1	in	in	ADP
ejpam-642	477	2	summary	summary	NOUN
ejpam-642	477	3	,	,	PUNCT
ejpam-642	477	4	the	the	DET
ejpam-642	477	5	specific	specific	ADJ
ejpam-642	477	6	tasks	task	NOUN
ejpam-642	477	7	include	include	VERB
ejpam-642	477	8	(	(	PUNCT
ejpam-642	477	9	1	1	X
ejpam-642	477	10	)	)	PUNCT
ejpam-642	477	11	add	add	VERB
ejpam-642	477	12	new	new	ADJ
ejpam-642	477	13	entrant	entrant	NOUN
ejpam-642	477	14	records	record	NOUN
ejpam-642	477	15	to	to	PART
ejpam-642	477	16	“	"	PUNCT
ejpam-642	477	17	register	register	VERB
ejpam-642	477	18	”	"	PUNCT
ejpam-642	477	19	file	file	NOUN
ejpam-642	477	20	from	from	ADP
ejpam-642	477	21	“	"	PUNCT
ejpam-642	477	22	births	birth	NOUN
ejpam-642	477	23	”	"	PUNCT
ejpam-642	477	24	file	file	NOUN
ejpam-642	477	25	,	,	PUNCT
ejpam-642	477	26	(	(	PUNCT
ejpam-642	477	27	2	2	X
ejpam-642	477	28	)	)	PUNCT
ejpam-642	477	29	update	update	NOUN
ejpam-642	477	30	moving	moving	NOUN
ejpam-642	477	31	records	record	NOUN
ejpam-642	477	32	to	to	PART
ejpam-642	477	33	“	"	PUNCT
ejpam-642	477	34	register	register	VERB
ejpam-642	477	35	”	"	PUNCT
ejpam-642	477	36	file	file	NOUN
ejpam-642	477	37	from	from	ADP
ejpam-642	477	38	“	"	PUNCT
ejpam-642	477	39	drivers	driver	NOUN
ejpam-642	477	40	”	"	PUNCT
ejpam-642	477	41	file	file	NOUN
ejpam-642	477	42	,	,	PUNCT
ejpam-642	477	43	(	(	PUNCT
ejpam-642	477	44	3	3	X
ejpam-642	477	45	)	)	PUNCT
ejpam-642	477	46	remove	remove	VERB
ejpam-642	477	47	death	death	NOUN
ejpam-642	477	48	records	record	NOUN
ejpam-642	477	49	from	from	ADP
ejpam-642	477	50	“	"	PUNCT
ejpam-642	477	51	register	register	VERB
ejpam-642	477	52	”	"	PUNCT
ejpam-642	477	53	file	file	NOUN
ejpam-642	477	54	according	accord	VERB
ejpam-642	477	55	to	to	ADP
ejpam-642	477	56	“	"	PUNCT
ejpam-642	477	57	deaths	death	NOUN
ejpam-642	477	58	”	"	PUNCT
ejpam-642	477	59	file	file	NOUN
ejpam-642	477	60	.	.	PUNCT
ejpam-642	478	1	to	to	PART
ejpam-642	478	2	achieve	achieve	VERB
ejpam-642	478	3	these	these	DET
ejpam-642	478	4	goals	goal	NOUN
ejpam-642	478	5	,	,	PUNCT
ejpam-642	478	6	we	we	PRON
ejpam-642	478	7	first	first	ADV
ejpam-642	478	8	need	need	VERB
ejpam-642	478	9	to	to	PART
ejpam-642	478	10	identify	identify	VERB
ejpam-642	478	11	common	common	ADJ
ejpam-642	478	12	individuals	individual	NOUN
ejpam-642	478	13	in	in	ADP
ejpam-642	478	14	two	two	NUM
ejpam-642	478	15	compared	compare	VERB
ejpam-642	478	16	files	file	NOUN
ejpam-642	478	17	.	.	PUNCT
ejpam-642	479	1	thus	thus	ADV
ejpam-642	479	2	,	,	PUNCT
ejpam-642	479	3	record	record	NOUN
ejpam-642	479	4	linkage	linkage	NOUN
ejpam-642	479	5	is	be	AUX
ejpam-642	479	6	employed	employ	VERB
ejpam-642	479	7	in	in	ADP
ejpam-642	479	8	these	these	DET
ejpam-642	479	9	tasks	task	NOUN
ejpam-642	479	10	.	.	PUNCT
ejpam-642	480	1	basically	basically	ADV
ejpam-642	480	2	,	,	PUNCT
ejpam-642	480	3	the	the	DET
ejpam-642	480	4	records	record	NOUN
ejpam-642	480	5	in	in	ADP
ejpam-642	480	6	these	these	DET
ejpam-642	480	7	files	file	NOUN
ejpam-642	480	8	contain	contain	VERB
ejpam-642	480	9	common	common	ADJ
ejpam-642	480	10	personal	personal	ADJ
ejpam-642	480	11	information	information	NOUN
ejpam-642	480	12	such	such	ADJ
ejpam-642	480	13	as	as	ADP
ejpam-642	480	14	name	name	NOUN
ejpam-642	480	15	(	(	PUNCT
ejpam-642	480	16	first	first	ADJ
ejpam-642	480	17	,	,	PUNCT
ejpam-642	480	18	middle	middle	ADJ
ejpam-642	480	19	and	and	CCONJ
ejpam-642	480	20	last	last	ADJ
ejpam-642	480	21	)	)	PUNCT
ejpam-642	480	22	,	,	PUNCT
ejpam-642	480	23	date	date	NOUN
ejpam-642	480	24	of	of	ADP
ejpam-642	480	25	birth	birth	NOUN
ejpam-642	480	26	(	(	PUNCT
ejpam-642	480	27	day	day	NOUN
ejpam-642	480	28	,	,	PUNCT
ejpam-642	480	29	month	month	NOUN
ejpam-642	480	30	and	and	CCONJ
ejpam-642	480	31	year	year	NOUN
ejpam-642	480	32	)	)	PUNCT
ejpam-642	480	33	,	,	PUNCT
ejpam-642	480	34	gender	gender	NOUN
ejpam-642	480	35	,	,	PUNCT
ejpam-642	480	36	address	address	NOUN
ejpam-642	480	37	,	,	PUNCT
ejpam-642	480	38	and	and	CCONJ
ejpam-642	480	39	so	so	ADV
ejpam-642	480	40	on	on	ADV
ejpam-642	480	41	.	.	PUNCT
ejpam-642	481	1	we	we	PRON
ejpam-642	481	2	adopt	adopt	VERB
ejpam-642	481	3	the	the	DET
ejpam-642	481	4	gender	gender	NOUN
ejpam-642	481	5	as	as	ADP
ejpam-642	481	6	the	the	DET
ejpam-642	481	7	blocking	block	VERB
ejpam-642	481	8	factor	factor	NOUN
ejpam-642	481	9	to	to	PART
ejpam-642	481	10	reduce	reduce	VERB
ejpam-642	481	11	both	both	DET
ejpam-642	481	12	the	the	DET
ejpam-642	481	13	number	number	NOUN
ejpam-642	481	14	of	of	ADP
ejpam-642	481	15	record	record	NOUN
ejpam-642	481	16	pairs	pair	NOUN
ejpam-642	481	17	and	and	CCONJ
ejpam-642	481	18	the	the	DET
ejpam-642	481	19	number	number	NOUN
ejpam-642	481	20	of	of	ADP
ejpam-642	481	21	factors	factor	NOUN
ejpam-642	481	22	in	in	ADP
ejpam-642	481	23	the	the	DET
ejpam-642	481	24	fitted	fit	VERB
ejpam-642	481	25	models	model	NOUN
ejpam-642	481	26	.	.	PUNCT
ejpam-642	482	1	table	table	NOUN
ejpam-642	482	2	7	7	NUM
ejpam-642	482	3	summarizes	summarize	NOUN
ejpam-642	482	4	the	the	DET
ejpam-642	482	5	total	total	ADJ
ejpam-642	482	6	pairs	pair	NOUN
ejpam-642	482	7	in	in	ADP
ejpam-642	482	8	each	each	PRON
ejpam-642	482	9	of	of	ADP
ejpam-642	482	10	three	three	NUM
ejpam-642	482	11	tasks	task	NOUN
ejpam-642	482	12	after	after	ADP
ejpam-642	482	13	the	the	DET
ejpam-642	482	14	gender	gender	NOUN
ejpam-642	482	15	blocking	blocking	NOUN
ejpam-642	482	16	.	.	PUNCT
ejpam-642	483	1	after	after	ADP
ejpam-642	483	2	a	a	DET
ejpam-642	483	3	preliminary	preliminary	ADJ
ejpam-642	483	4	data	datum	NOUN
ejpam-642	483	5	analysis	analysis	NOUN
ejpam-642	483	6	,	,	PUNCT
ejpam-642	483	7	we	we	PRON
ejpam-642	483	8	choose	choose	VERB
ejpam-642	483	9	first	first	ADJ
ejpam-642	483	10	name	name	NOUN
ejpam-642	483	11	,	,	PUNCT
ejpam-642	483	12	middle	middle	ADJ
ejpam-642	483	13	name	name	NOUN
ejpam-642	483	14	,	,	PUNCT
ejpam-642	483	15	last	last	ADJ
ejpam-642	483	16	name	name	NOUN
ejpam-642	483	17	and	and	CCONJ
ejpam-642	483	18	date	date	NOUN
ejpam-642	483	19	of	of	ADP
ejpam-642	483	20	birth	birth	NOUN
ejpam-642	483	21	(	(	PUNCT
ejpam-642	483	22	day	day	NOUN
ejpam-642	483	23	/	/	SYM
ejpam-642	483	24	month	month	NOUN
ejpam-642	483	25	/	/	SYM
ejpam-642	483	26	year	year	NOUN
ejpam-642	483	27	)	)	PUNCT
ejpam-642	483	28	as	as	ADP
ejpam-642	483	29	identifiers	identifier	NOUN
ejpam-642	483	30	for	for	ADP
ejpam-642	483	31	the	the	DET
ejpam-642	483	32	tasks	task	NOUN
ejpam-642	483	33	of	of	ADP
ejpam-642	483	34	adding	add	VERB
ejpam-642	483	35	new	new	ADJ
ejpam-642	483	36	entrant	entrant	ADJ
ejpam-642	483	37	records	record	NOUN
ejpam-642	483	38	and	and	CCONJ
ejpam-642	483	39	removing	remove	VERB
ejpam-642	483	40	death	death	NOUN
ejpam-642	483	41	records	record	NOUN
ejpam-642	483	42	.	.	PUNCT
ejpam-642	484	1	for	for	ADP
ejpam-642	484	2	updating	update	VERB
ejpam-642	484	3	moving	move	VERB
ejpam-642	484	4	records	record	NOUN
ejpam-642	484	5	,	,	PUNCT
ejpam-642	484	6	we	we	PRON
ejpam-642	484	7	only	only	ADV
ejpam-642	484	8	adopt	adopt	VERB
ejpam-642	484	9	first	first	ADJ
ejpam-642	484	10	name	name	NOUN
ejpam-642	484	11	,	,	PUNCT
ejpam-642	484	12	middle	middle	PROPN
ejpam-642	484	13	r.	r.	PROPN
ejpam-642	484	14	zhu	zhu	PROPN
ejpam-642	484	15	,	,	PUNCT
ejpam-642	484	16	j.	j.	PROPN
ejpam-642	484	17	zhang	zhang	PROPN
ejpam-642	484	18	,	,	PUNCT
ejpam-642	484	19	d.	d.	PROPN
ejpam-642	484	20	zhang	zhang	PROPN
ejpam-642	484	21	,	,	PUNCT
ejpam-642	484	22	g.	g.	PROPN
ejpam-642	484	23	yan	yan	PROPN
ejpam-642	484	24	/	/	SYM
ejpam-642	484	25	eur	eur	PROPN
ejpam-642	484	26	.	.	PUNCT
ejpam-642	485	1	j.	j.	PROPN
ejpam-642	485	2	pure	pure	PROPN
ejpam-642	485	3	appl	appl	PROPN
ejpam-642	485	4	.	.	PROPN
ejpam-642	485	5	math	math	PROPN
ejpam-642	485	6	,	,	PUNCT
ejpam-642	485	7	3	3	NUM
ejpam-642	485	8	(	(	PUNCT
ejpam-642	485	9	2010	2010	NUM
ejpam-642	485	10	)	)	PUNCT
ejpam-642	485	11	,	,	PUNCT
ejpam-642	485	12	141	141	NUM
ejpam-642	485	13	-	-	SYM
ejpam-642	485	14	162	162	NUM
ejpam-642	485	15	159table	159table	PROPN
ejpam-642	485	16	7	7	NUM
ejpam-642	485	17	:	:	PUNCT
ejpam-642	485	18	numbers	number	NOUN
ejpam-642	485	19	of	of	ADP
ejpam-642	485	20	re	re	NOUN
ejpam-642	485	21	ord	ord	NOUN
ejpam-642	485	22	pairs	pair	NOUN
ejpam-642	485	23	between	between	ADP
ejpam-642	485	24	ompared	ompared	ADJ
ejpam-642	485	25	�	�	PROPN
ejpam-642	485	26	les	le	NOUN
ejpam-642	485	27	after	after	ADP
ejpam-642	485	28	blo	blo	NOUN
ejpam-642	485	29	king	king	NOUN
ejpam-642	485	30	using	use	VERB
ejpam-642	485	31	gender	gender	NOUN
ejpam-642	485	32	.	.	PUNCT
ejpam-642	486	1	file	file	NOUN
ejpam-642	486	2	1	1	NUM
ejpam-642	486	3	file	file	NOUN
ejpam-642	486	4	2	2	NUM
ejpam-642	486	5	gender	gender	NOUN
ejpam-642	486	6	pair	pair	NOUN
ejpam-642	486	7	number	number	NOUN
ejpam-642	486	8	total	total	ADJ
ejpam-642	486	9	pair	pair	NOUN
ejpam-642	486	10	male	male	NOUN
ejpam-642	486	11	2000×	2000×	NUM
ejpam-642	486	12	102=	102=	NUM
ejpam-642	486	13	204,000register	204,000register	NUM
ejpam-642	486	14	births	birth	NOUN
ejpam-642	486	15	female	female	ADJ
ejpam-642	486	16	2000×	2000×	NUM
ejpam-642	486	17	72=	72=	NUM
ejpam-642	486	18	144,000	144,000	NUM
ejpam-642	486	19	348,000	348,000	NUM
ejpam-642	486	20	male	male	NOUN
ejpam-642	486	21	2000×	2000×	NUM
ejpam-642	486	22	470=	470=	X
ejpam-642	486	23	940,000register	940,000register	NUM
ejpam-642	487	1	drivers	driver	NOUN
ejpam-642	487	2	female	female	VERB
ejpam-642	487	3	2000×	2000×	NUM
ejpam-642	487	4	514=	514=	NOUN
ejpam-642	487	5	1,028,000	1,028,000	NUM
ejpam-642	487	6	1968,000	1968,000	NUM
ejpam-642	487	7	male	male	NOUN
ejpam-642	487	8	2000×	2000×	NUM
ejpam-642	487	9	69=	69=	NUM
ejpam-642	487	10	138,000register	138,000register	NUM
ejpam-642	487	11	deaths	death	NOUN
ejpam-642	487	12	female	female	VERB
ejpam-642	487	13	2000×	2000×	NUM
ejpam-642	487	14	96=	96=	NUM
ejpam-642	487	15	192,000	192,000	NUM
ejpam-642	487	16	330,000	330,000	NUM
ejpam-642	487	17	name	name	NOUN
ejpam-642	487	18	,	,	PUNCT
ejpam-642	487	19	last	last	ADJ
ejpam-642	487	20	name	name	NOUN
ejpam-642	487	21	and	and	CCONJ
ejpam-642	487	22	postal	postal	ADJ
ejpam-642	487	23	code	code	NOUN
ejpam-642	487	24	as	as	ADP
ejpam-642	487	25	identifiers	identifier	NOUN
ejpam-642	487	26	,	,	PUNCT
ejpam-642	487	27	because	because	SCONJ
ejpam-642	487	28	the	the	DET
ejpam-642	487	29	date	date	NOUN
ejpam-642	487	30	of	of	ADP
ejpam-642	487	31	birth	birth	NOUN
ejpam-642	487	32	seems	seem	VERB
ejpam-642	487	33	wrong	wrong	ADJ
ejpam-642	487	34	by	by	ADP
ejpam-642	487	35	visual	visual	ADJ
ejpam-642	487	36	check	check	NOUN
ejpam-642	487	37	.	.	PUNCT
ejpam-642	488	1	furthermore	furthermore	ADV
ejpam-642	488	2	,	,	PUNCT
ejpam-642	488	3	we	we	PRON
ejpam-642	488	4	exclude	exclude	VERB
ejpam-642	488	5	the	the	DET
ejpam-642	488	6	records	record	NOUN
ejpam-642	488	7	with	with	ADP
ejpam-642	488	8	missing	miss	VERB
ejpam-642	488	9	values	value	NOUN
ejpam-642	488	10	or	or	CCONJ
ejpam-642	488	11	error	error	NOUN
ejpam-642	488	12	values	value	NOUN
ejpam-642	488	13	on	on	ADP
ejpam-642	488	14	identifiers	identifier	NOUN
ejpam-642	488	15	and	and	CCONJ
ejpam-642	488	16	gender	gender	NOUN
ejpam-642	488	17	in	in	ADP
ejpam-642	488	18	each	each	DET
ejpam-642	488	19	file	file	NOUN
ejpam-642	488	20	.	.	PUNCT
ejpam-642	489	1	these	these	PRON
ejpam-642	489	2	excluded	exclude	VERB
ejpam-642	489	3	records	record	NOUN
ejpam-642	489	4	will	will	AUX
ejpam-642	489	5	be	be	AUX
ejpam-642	489	6	manually	manually	ADV
ejpam-642	489	7	reviewed	review	VERB
ejpam-642	489	8	.	.	PUNCT
ejpam-642	490	1	the	the	DET
ejpam-642	490	2	simple	simple	ADJ
ejpam-642	490	3	matching	matching	NOUN
ejpam-642	490	4	of	of	ADP
ejpam-642	490	5	“	"	PUNCT
ejpam-642	490	6	agreement	agreement	NOUN
ejpam-642	490	7	”	"	PUNCT
ejpam-642	490	8	or	or	CCONJ
ejpam-642	490	9	“	"	PUNCT
ejpam-642	490	10	disagreement	disagreement	NOUN
ejpam-642	490	11	”	"	PUNCT
ejpam-642	490	12	is	be	AUX
ejpam-642	490	13	chosen	choose	VERB
ejpam-642	490	14	as	as	ADP
ejpam-642	490	15	the	the	DET
ejpam-642	490	16	comparison	comparison	NOUN
ejpam-642	490	17	rule	rule	NOUN
ejpam-642	490	18	.	.	PUNCT
ejpam-642	491	1	therefore	therefore	ADV
ejpam-642	491	2	,	,	PUNCT
ejpam-642	491	3	for	for	ADP
ejpam-642	491	4	the	the	DET
ejpam-642	491	5	comparison	comparison	NOUN
ejpam-642	491	6	vector	vector	NOUN
ejpam-642	491	7	γ	γ	X
ejpam-642	491	8	=	=	SYM
ejpam-642	491	9	(	(	PUNCT
ejpam-642	491	10	γ1,γ2,γ3,γ4	γ1,γ2,γ3,γ4	PROPN
ejpam-642	491	11	)	)	PUNCT
ejpam-642	491	12	,	,	PUNCT
ejpam-642	491	13	all	all	DET
ejpam-642	491	14	four	four	NUM
ejpam-642	491	15	fields	field	NOUN
ejpam-642	491	16	are	be	AUX
ejpam-642	491	17	binary	binary	ADJ
ejpam-642	491	18	,	,	PUNCT
ejpam-642	491	19	indicating	indicate	VERB
ejpam-642	491	20	match	match	NOUN
ejpam-642	491	21	or	or	CCONJ
ejpam-642	491	22	nonmatch	nonmatch	VERB
ejpam-642	491	23	on	on	ADP
ejpam-642	491	24	identifiers	identifier	NOUN
ejpam-642	491	25	.	.	PUNCT
ejpam-642	492	1	each	each	PRON
ejpam-642	492	2	of	of	ADP
ejpam-642	492	3	the	the	DET
ejpam-642	492	4	record	record	NOUN
ejpam-642	492	5	pairs	pair	NOUN
ejpam-642	492	6	in	in	ADP
ejpam-642	492	7	table	table	NOUN
ejpam-642	492	8	7	7	NUM
ejpam-642	492	9	yields	yield	NOUN
ejpam-642	492	10	an	an	DET
ejpam-642	492	11	observed	observed	ADJ
ejpam-642	492	12	comparison	comparison	NOUN
ejpam-642	492	13	vector	vector	NOUN
ejpam-642	492	14	.	.	PUNCT
ejpam-642	493	1	three	three	NUM
ejpam-642	493	2	contingency	contingency	NOUN
ejpam-642	493	3	tables	table	NOUN
ejpam-642	493	4	are	be	AUX
ejpam-642	493	5	then	then	ADV
ejpam-642	493	6	created	create	VERB
ejpam-642	493	7	for	for	ADP
ejpam-642	493	8	the	the	DET
ejpam-642	493	9	comparison	comparison	NOUN
ejpam-642	493	10	vectors	vector	NOUN
ejpam-642	493	11	,	,	PUNCT
ejpam-642	493	12	one	one	NUM
ejpam-642	493	13	for	for	ADP
ejpam-642	493	14	each	each	DET
ejpam-642	493	15	task	task	NOUN
ejpam-642	493	16	.	.	PUNCT
ejpam-642	494	1	next	next	ADV
ejpam-642	494	2	we	we	PRON
ejpam-642	494	3	apply	apply	VERB
ejpam-642	494	4	the	the	DET
ejpam-642	494	5	data	data	NOUN
ejpam-642	494	6	-	-	PUNCT
ejpam-642	494	7	driven	drive	VERB
ejpam-642	494	8	strategy	strategy	NOUN
ejpam-642	494	9	to	to	ADP
ejpam-642	494	10	the	the	DET
ejpam-642	494	11	three	three	NUM
ejpam-642	494	12	obtained	obtain	VERB
ejpam-642	494	13	contingency	contingency	NOUN
ejpam-642	494	14	tables	table	NOUN
ejpam-642	494	15	.	.	PUNCT
ejpam-642	495	1	the	the	DET
ejpam-642	495	2	significance	significance	NOUN
ejpam-642	495	3	level	level	NOUN
ejpam-642	495	4	is	be	AUX
ejpam-642	495	5	set	set	VERB
ejpam-642	495	6	as	as	ADP
ejpam-642	495	7	5	5	NUM
ejpam-642	495	8	%	%	NOUN
ejpam-642	495	9	.	.	PUNCT
ejpam-642	496	1	the	the	DET
ejpam-642	496	2	selected	select	VERB
ejpam-642	496	3	models	model	NOUN
ejpam-642	496	4	(	(	PUNCT
ejpam-642	496	5	in	in	ADP
ejpam-642	496	6	poisson	poisson	PROPN
ejpam-642	496	7	glm	glm	PROPN
ejpam-642	496	8	formulation	formulation	PROPN
ejpam-642	496	9	)	)	PUNCT
ejpam-642	496	10	are	be	AUX
ejpam-642	496	11	listed	list	VERB
ejpam-642	496	12	below	below	ADV
ejpam-642	496	13	.	.	PUNCT
ejpam-642	497	1	(	(	PUNCT
ejpam-642	497	2	1	1	X
ejpam-642	497	3	)	)	PUNCT
ejpam-642	497	4	for	for	ADP
ejpam-642	497	5	adding	add	VERB
ejpam-642	497	6	new	new	ADJ
ejpam-642	497	7	entrant	entrant	ADJ
ejpam-642	497	8	records	record	NOUN
ejpam-642	497	9	from	from	ADP
ejpam-642	497	10	the	the	DET
ejpam-642	497	11	“	"	PUNCT
ejpam-642	497	12	births	birth	NOUN
ejpam-642	497	13	”	"	PUNCT
ejpam-642	497	14	file	file	NOUN
ejpam-642	497	15	,	,	PUNCT
ejpam-642	497	16	γ1	γ1	NOUN
ejpam-642	497	17	,	,	PUNCT
ejpam-642	497	18	γ2	γ2	PROPN
ejpam-642	497	19	,	,	PUNCT
ejpam-642	497	20	γ3	γ3	NOUN
ejpam-642	497	21	and	and	CCONJ
ejpam-642	497	22	γ4	γ4	PROPN
ejpam-642	497	23	mean	mean	NOUN
ejpam-642	497	24	match	match	NOUN
ejpam-642	497	25	or	or	CCONJ
ejpam-642	497	26	non	non	ADJ
ejpam-642	497	27	-	-	NOUN
ejpam-642	497	28	match	match	NOUN
ejpam-642	497	29	on	on	ADP
ejpam-642	497	30	first	first	ADJ
ejpam-642	497	31	name	name	NOUN
ejpam-642	497	32	,	,	PUNCT
ejpam-642	497	33	middle	middle	ADJ
ejpam-642	497	34	name	name	NOUN
ejpam-642	497	35	,	,	PUNCT
ejpam-642	497	36	last	last	ADJ
ejpam-642	497	37	name	name	NOUN
ejpam-642	497	38	and	and	CCONJ
ejpam-642	497	39	date	date	NOUN
ejpam-642	497	40	of	of	ADP
ejpam-642	497	41	birth	birth	NOUN
ejpam-642	497	42	respectively	respectively	ADV
ejpam-642	497	43	.	.	PUNCT
ejpam-642	498	1	the	the	DET
ejpam-642	498	2	fitted	fit	VERB
ejpam-642	498	3	model	model	NOUN
ejpam-642	498	4	ism10	ism10	PROPN
ejpam-642	498	5	with	with	ADP
ejpam-642	498	6	an	an	DET
ejpam-642	498	7	estimated	estimate	VERB
ejpam-642	498	8	mixing	mix	VERB
ejpam-642	498	9	proportion	proportion	NOUN
ejpam-642	498	10	π̂	π̂	NUM
ejpam-642	498	11	=	=	SYM
ejpam-642	498	12	4.885e−	4.885e−	PROPN
ejpam-642	498	13	4	4	NUM
ejpam-642	498	14	,	,	PUNCT
ejpam-642	498	15	where	where	SCONJ
ejpam-642	498	16	the	the	DET
ejpam-642	498	17	u	u	PROPN
ejpam-642	498	18	component	component	NOUN
ejpam-642	498	19	is	be	AUX
ejpam-642	498	20	α	α	NOUN
ejpam-642	498	21	=	=	SYM
ejpam-642	498	22	12.704	12.704	NUM
ejpam-642	498	23	,	,	PUNCT
ejpam-642	498	24	α1	α1	PROPN
ejpam-642	498	25	=	=	SYM
ejpam-642	498	26	−6.586	−6.586	ADJ
ejpam-642	498	27	,	,	PUNCT
ejpam-642	498	28	α2	α2	NOUN
ejpam-642	498	29	=	=	SYM
ejpam-642	498	30	−3.046	−3.046	NOUN
ejpam-642	498	31	,	,	PUNCT
ejpam-642	498	32	α3	α3	NOUN
ejpam-642	498	33	=	=	SYM
ejpam-642	498	34	−5.998	−5.998	PROPN
ejpam-642	498	35	,	,	PUNCT
ejpam-642	498	36	α4	α4	NOUN
ejpam-642	498	37	=	=	SYM
ejpam-642	498	38	−10.219	−10.219	PROPN
ejpam-642	498	39	,	,	PUNCT
ejpam-642	498	40	α12	α12	NOUN
ejpam-642	498	41	=	=	SYM
ejpam-642	498	42	4.245	4.245	NUM
ejpam-642	498	43	,	,	PUNCT
ejpam-642	498	44	α13	α13	NOUN
ejpam-642	498	45	=	=	PUNCT
ejpam-642	498	46	1.819	1.819	NUM
ejpam-642	498	47	,	,	PUNCT
ejpam-642	498	48	α14	α14	NUM
ejpam-642	498	49	=	=	SYM
ejpam-642	498	50	−13.911	−13.911	PROPN
ejpam-642	498	51	,	,	PUNCT
ejpam-642	498	52	α23	α23	NOUN
ejpam-642	498	53	=	=	SYM
ejpam-642	498	54	1.804	1.804	NUM
ejpam-642	498	55	,	,	PUNCT
ejpam-642	498	56	α24	α24	NOUN
ejpam-642	498	57	=	=	SYM
ejpam-642	498	58	0.561	0.561	NUM
ejpam-642	498	59	,	,	PUNCT
ejpam-642	498	60	α34	α34	NOUN
ejpam-642	498	61	=	=	SYM
ejpam-642	498	62	−14.031	−14.031	NOUN
ejpam-642	498	63	,	,	PUNCT
ejpam-642	498	64	and	and	CCONJ
ejpam-642	498	65	the	the	DET
ejpam-642	498	66	m	m	PROPN
ejpam-642	498	67	component	component	NOUN
ejpam-642	498	68	is	be	AUX
ejpam-642	498	69	α=	α=	PROPN
ejpam-642	498	70	−77.456	−77.456	NOUN
ejpam-642	498	71	,	,	PUNCT
ejpam-642	498	72	α1	α1	PROPN
ejpam-642	498	73	=	=	SYM
ejpam-642	498	74	27.503	27.503	NUM
ejpam-642	498	75	,	,	PUNCT
ejpam-642	498	76	α2	α2	PROPN
ejpam-642	498	77	=	=	SYM
ejpam-642	498	78	27.805	27.805	NUM
ejpam-642	498	79	,	,	PUNCT
ejpam-642	498	80	α3	α3	NOUN
ejpam-642	498	81	=	=	SYM
ejpam-642	498	82	0.332	0.332	NUM
ejpam-642	498	83	,	,	PUNCT
ejpam-642	498	84	α4	α4	NOUN
ejpam-642	498	85	=	=	SYM
ejpam-642	498	86	26.410	26.410	NUM
ejpam-642	498	87	.	.	PUNCT
ejpam-642	499	1	(	(	PUNCT
ejpam-642	499	2	2	2	NUM
ejpam-642	499	3	)	)	PUNCT
ejpam-642	499	4	for	for	ADP
ejpam-642	499	5	updating	update	VERB
ejpam-642	499	6	moving	move	VERB
ejpam-642	499	7	records	record	NOUN
ejpam-642	499	8	from	from	ADP
ejpam-642	499	9	the	the	DET
ejpam-642	499	10	“	"	PUNCT
ejpam-642	499	11	drivers	driver	NOUN
ejpam-642	499	12	”	"	PUNCT
ejpam-642	499	13	file	file	NOUN
ejpam-642	499	14	,	,	PUNCT
ejpam-642	499	15	γ1	γ1	NOUN
ejpam-642	499	16	,	,	PUNCT
ejpam-642	499	17	γ2	γ2	PROPN
ejpam-642	499	18	,	,	PUNCT
ejpam-642	499	19	γ3	γ3	NOUN
ejpam-642	499	20	and	and	CCONJ
ejpam-642	499	21	γ4	γ4	PROPN
ejpam-642	499	22	mean	mean	NOUN
ejpam-642	499	23	match	match	NOUN
ejpam-642	499	24	or	or	CCONJ
ejpam-642	499	25	non	non	ADJ
ejpam-642	499	26	-	-	NOUN
ejpam-642	499	27	match	match	NOUN
ejpam-642	499	28	on	on	ADP
ejpam-642	499	29	first	first	ADJ
ejpam-642	499	30	name	name	NOUN
ejpam-642	499	31	,	,	PUNCT
ejpam-642	499	32	middle	middle	ADJ
ejpam-642	499	33	name	name	NOUN
ejpam-642	499	34	,	,	PUNCT
ejpam-642	499	35	last	last	ADJ
ejpam-642	499	36	name	name	NOUN
ejpam-642	499	37	and	and	CCONJ
ejpam-642	499	38	postal	postal	ADJ
ejpam-642	499	39	code	code	NOUN
ejpam-642	499	40	respectively	respectively	ADV
ejpam-642	499	41	.	.	PUNCT
ejpam-642	500	1	the	the	DET
ejpam-642	500	2	fitted	fit	VERB
ejpam-642	500	3	model	model	NOUN
ejpam-642	500	4	ism11	ism11	PROPN
ejpam-642	500	5	with	with	ADP
ejpam-642	500	6	an	an	DET
ejpam-642	500	7	estimated	estimate	VERB
ejpam-642	500	8	mixing	mix	VERB
ejpam-642	500	9	proportion	proportion	NOUN
ejpam-642	500	10	π̂	π̂	VERB
ejpam-642	500	11	=	=	SYM
ejpam-642	500	12	4.032e−	4.032e−	PROPN
ejpam-642	500	13	4	4	NUM
ejpam-642	500	14	,	,	PUNCT
ejpam-642	500	15	where	where	SCONJ
ejpam-642	500	16	the	the	DET
ejpam-642	500	17	u	u	PROPN
ejpam-642	500	18	component	component	NOUN
ejpam-642	500	19	is	be	AUX
ejpam-642	500	20	α=	α=	ADJ
ejpam-642	500	21	14.439	14.439	NUM
ejpam-642	500	22	,	,	PUNCT
ejpam-642	500	23	α1	α1	PROPN
ejpam-642	500	24	=	=	SYM
ejpam-642	500	25	−6.432	−6.432	PROPN
ejpam-642	500	26	,	,	PUNCT
ejpam-642	500	27	α2	α2	NOUN
ejpam-642	500	28	=	=	SYM
ejpam-642	500	29	−3.128	−3.128	PROPN
ejpam-642	500	30	,	,	PUNCT
ejpam-642	500	31	α3	α3	PROPN
ejpam-642	500	32	=	=	SYM
ejpam-642	500	33	−5.849	−5.849	PROPN
ejpam-642	500	34	,	,	PUNCT
ejpam-642	500	35	α4	α4	NOUN
ejpam-642	500	36	=	=	SYM
ejpam-642	500	37	−8.882	−8.882	PROPN
ejpam-642	500	38	,	,	PUNCT
ejpam-642	500	39	α12	α12	NOUN
ejpam-642	500	40	=	=	NUM
ejpam-642	500	41	4.205	4.205	NUM
ejpam-642	500	42	,	,	PUNCT
ejpam-642	500	43	α13	α13	X
ejpam-642	500	44	=	=	SYM
ejpam-642	500	45	1.427	1.427	NUM
ejpam-642	500	46	,	,	PUNCT
ejpam-642	500	47	α14	α14	NUM
ejpam-642	500	48	=	=	SYM
ejpam-642	500	49	1.569	1.569	NUM
ejpam-642	500	50	,	,	PUNCT
ejpam-642	500	51	α23	α23	NOUN
ejpam-642	500	52	=	=	NUM
ejpam-642	500	53	1.879	1.879	NUM
ejpam-642	500	54	,	,	PUNCT
ejpam-642	500	55	α24	α24	NOUN
ejpam-642	500	56	=	=	NOUN
ejpam-642	500	57	0.427	0.427	NUM
ejpam-642	500	58	,	,	PUNCT
ejpam-642	500	59	α34	α34	NOUN
ejpam-642	500	60	=	=	PUNCT
ejpam-642	500	61	1.807	1.807	NUM
ejpam-642	500	62	,	,	PUNCT
ejpam-642	500	63	and	and	CCONJ
ejpam-642	500	64	the	the	DET
ejpam-642	500	65	m	m	PROPN
ejpam-642	500	66	component	component	NOUN
ejpam-642	500	67	is	be	AUX
ejpam-642	500	68	α=	α=	NOUN
ejpam-642	500	69	−21.502	−21.502	ADJ
ejpam-642	500	70	,	,	PUNCT
ejpam-642	500	71	α1	α1	PROPN
ejpam-642	500	72	=	=	SYM
ejpam-642	500	73	−20.617	−20.617	PROPN
ejpam-642	500	74	,	,	PUNCT
ejpam-642	500	75	α2	α2	PROPN
ejpam-642	500	76	=	=	SYM
ejpam-642	500	77	−5.895	−5.895	PROPN
ejpam-642	500	78	,	,	PUNCT
ejpam-642	500	79	α3	α3	NOUN
ejpam-642	500	80	=	=	NOUN
ejpam-642	500	81	0.557	0.557	NUM
ejpam-642	500	82	,	,	PUNCT
ejpam-642	500	83	α4	α4	NOUN
ejpam-642	500	84	=	=	SYM
ejpam-642	500	85	0.249	0.249	NUM
ejpam-642	500	86	,	,	PUNCT
ejpam-642	500	87	r.	r.	PROPN
ejpam-642	500	88	zhu	zhu	PROPN
ejpam-642	500	89	,	,	PUNCT
ejpam-642	500	90	j.	j.	PROPN
ejpam-642	500	91	zhang	zhang	PROPN
ejpam-642	500	92	,	,	PUNCT
ejpam-642	500	93	d.	d.	PROPN
ejpam-642	500	94	zhang	zhang	PROPN
ejpam-642	500	95	,	,	PUNCT
ejpam-642	500	96	g.	g.	PROPN
ejpam-642	500	97	yan	yan	PROPN
ejpam-642	500	98	/	/	SYM
ejpam-642	500	99	eur	eur	PROPN
ejpam-642	500	100	.	.	PUNCT
ejpam-642	501	1	j.	j.	PROPN
ejpam-642	501	2	pure	pure	PROPN
ejpam-642	501	3	appl	appl	PROPN
ejpam-642	501	4	.	.	PROPN
ejpam-642	501	5	math	math	PROPN
ejpam-642	501	6	,	,	PUNCT
ejpam-642	501	7	3	3	NUM
ejpam-642	501	8	(	(	PUNCT
ejpam-642	501	9	2010	2010	NUM
ejpam-642	501	10	)	)	PUNCT
ejpam-642	501	11	,	,	PUNCT
ejpam-642	501	12	141	141	NUM
ejpam-642	501	13	-	-	SYM
ejpam-642	501	14	162	162	NUM
ejpam-642	501	15	160	160	NUM
ejpam-642	501	16	α12	α12	NOUN
ejpam-642	501	17	=	=	SYM
ejpam-642	501	18	27.088	27.088	NUM
ejpam-642	501	19	,	,	PUNCT
ejpam-642	501	20	α13	α13	X
ejpam-642	501	21	=	=	SYM
ejpam-642	501	22	45.199	45.199	NUM
ejpam-642	501	23	,	,	PUNCT
ejpam-642	501	24	α14	α14	NUM
ejpam-642	501	25	=	=	SYM
ejpam-642	501	26	−27.411	−27.411	NOUN
ejpam-642	501	27	,	,	PUNCT
ejpam-642	501	28	α23	α23	NOUN
ejpam-642	501	29	=	=	SYM
ejpam-642	501	30	−43.852	−43.852	NOUN
ejpam-642	501	31	,	,	PUNCT
ejpam-642	501	32	α24	α24	NOUN
ejpam-642	501	33	=	=	SYM
ejpam-642	501	34	29.412	29.412	NUM
ejpam-642	501	35	,	,	PUNCT
ejpam-642	501	36	α34	α34	NOUN
ejpam-642	501	37	=	=	SYM
ejpam-642	501	38	23.367	23.367	NUM
ejpam-642	501	39	.	.	PUNCT
ejpam-642	502	1	(	(	PUNCT
ejpam-642	502	2	3	3	X
ejpam-642	502	3	)	)	PUNCT
ejpam-642	502	4	for	for	ADP
ejpam-642	502	5	removing	remove	VERB
ejpam-642	502	6	death	death	NOUN
ejpam-642	502	7	records	record	NOUN
ejpam-642	502	8	according	accord	VERB
ejpam-642	502	9	to	to	ADP
ejpam-642	502	10	the	the	DET
ejpam-642	502	11	“	"	PUNCT
ejpam-642	502	12	deaths	death	NOUN
ejpam-642	502	13	”	"	PUNCT
ejpam-642	502	14	file	file	NOUN
ejpam-642	502	15	,	,	PUNCT
ejpam-642	502	16	γ1	γ1	NOUN
ejpam-642	502	17	,	,	PUNCT
ejpam-642	502	18	γ2	γ2	PROPN
ejpam-642	502	19	,	,	PUNCT
ejpam-642	502	20	γ3	γ3	NOUN
ejpam-642	502	21	and	and	CCONJ
ejpam-642	502	22	γ4	γ4	PROPN
ejpam-642	502	23	mean	mean	NOUN
ejpam-642	502	24	match	match	NOUN
ejpam-642	502	25	or	or	CCONJ
ejpam-642	502	26	non	non	ADJ
ejpam-642	502	27	-	-	NOUN
ejpam-642	502	28	match	match	NOUN
ejpam-642	502	29	on	on	ADP
ejpam-642	502	30	first	first	ADJ
ejpam-642	502	31	name	name	NOUN
ejpam-642	502	32	,	,	PUNCT
ejpam-642	502	33	middle	middle	ADJ
ejpam-642	502	34	name	name	NOUN
ejpam-642	502	35	,	,	PUNCT
ejpam-642	502	36	last	last	ADJ
ejpam-642	502	37	name	name	NOUN
ejpam-642	502	38	and	and	CCONJ
ejpam-642	502	39	date	date	NOUN
ejpam-642	502	40	of	of	ADP
ejpam-642	502	41	birth	birth	NOUN
ejpam-642	502	42	respectively	respectively	ADV
ejpam-642	502	43	.	.	PUNCT
ejpam-642	503	1	the	the	DET
ejpam-642	503	2	fitted	fit	VERB
ejpam-642	503	3	model	model	NOUN
ejpam-642	503	4	ism10	ism10	PROPN
ejpam-642	503	5	with	with	ADP
ejpam-642	503	6	an	an	DET
ejpam-642	503	7	estimated	estimate	VERB
ejpam-642	503	8	mixing	mix	VERB
ejpam-642	503	9	proportion	proportion	NOUN
ejpam-642	503	10	π̂	π̂	NUM
ejpam-642	503	11	=	=	SYM
ejpam-642	503	12	4.455e−	4.455e−	PROPN
ejpam-642	503	13	4	4	NUM
ejpam-642	503	14	,	,	PUNCT
ejpam-642	503	15	where	where	SCONJ
ejpam-642	503	16	the	the	DET
ejpam-642	503	17	u	u	PROPN
ejpam-642	503	18	component	component	NOUN
ejpam-642	503	19	is	be	AUX
ejpam-642	503	20	α=	α=	NOUN
ejpam-642	503	21	12.652	12.652	NUM
ejpam-642	503	22	,	,	PUNCT
ejpam-642	503	23	α1	α1	PROPN
ejpam-642	503	24	=	=	SYM
ejpam-642	503	25	−6.725	−6.725	ADJ
ejpam-642	503	26	,	,	PUNCT
ejpam-642	503	27	α2	α2	PROPN
ejpam-642	503	28	=	=	SYM
ejpam-642	503	29	−3.058	−3.058	PROPN
ejpam-642	503	30	,	,	PUNCT
ejpam-642	503	31	α3	α3	NOUN
ejpam-642	503	32	=	=	SYM
ejpam-642	503	33	−5.531	−5.531	PROPN
ejpam-642	503	34	,	,	PUNCT
ejpam-642	503	35	α4	α4	NOUN
ejpam-642	503	36	=	=	SYM
ejpam-642	503	37	−9.320	−9.320	PROPN
ejpam-642	503	38	,	,	PUNCT
ejpam-642	503	39	α12	α12	NOUN
ejpam-642	503	40	=	=	SYM
ejpam-642	503	41	3.794	3.794	NUM
ejpam-642	503	42	,	,	PUNCT
ejpam-642	503	43	α13	α13	NOUN
ejpam-642	503	44	=	=	SYM
ejpam-642	503	45	2.159	2.159	NUM
ejpam-642	503	46	,	,	PUNCT
ejpam-642	503	47	α14	α14	NUM
ejpam-642	503	48	=	=	SYM
ejpam-642	503	49	−20.402	−20.402	NOUN
ejpam-642	503	50	,	,	PUNCT
ejpam-642	503	51	α23	α23	NOUN
ejpam-642	503	52	=	=	SYM
ejpam-642	503	53	1.391	1.391	NUM
ejpam-642	503	54	,	,	PUNCT
ejpam-642	503	55	α24	α24	NOUN
ejpam-642	503	56	=	=	SYM
ejpam-642	503	57	−0.302	−0.302	NOUN
ejpam-642	503	58	,	,	PUNCT
ejpam-642	503	59	α34	α34	NOUN
ejpam-642	503	60	=	=	PUNCT
ejpam-642	503	61	−20.899	−20.899	NOUN
ejpam-642	503	62	,	,	PUNCT
ejpam-642	503	63	and	and	CCONJ
ejpam-642	503	64	the	the	DET
ejpam-642	503	65	m	m	PROPN
ejpam-642	503	66	component	component	NOUN
ejpam-642	503	67	is	be	AUX
ejpam-642	503	68	α=	α=	NOUN
ejpam-642	503	69	−28.828	−28.828	NOUN
ejpam-642	503	70	,	,	PUNCT
ejpam-642	503	71	α1	α1	PROPN
ejpam-642	503	72	=	=	SYM
ejpam-642	503	73	3.570	3.570	NUM
ejpam-642	503	74	,	,	PUNCT
ejpam-642	503	75	α2	α2	NOUN
ejpam-642	503	76	=	=	SYM
ejpam-642	503	77	3.871	3.871	NUM
ejpam-642	503	78	,	,	PUNCT
ejpam-642	503	79	α3	α3	NOUN
ejpam-642	503	80	=	=	PROPN
ejpam-642	503	81	4.956	4.956	NUM
ejpam-642	503	82	,	,	PUNCT
ejpam-642	503	83	α4	α4	NOUN
ejpam-642	503	84	=	=	SYM
ejpam-642	503	85	21.366	21.366	NUM
ejpam-642	503	86	.	.	PUNCT
ejpam-642	504	1	in	in	ADP
ejpam-642	504	2	these	these	DET
ejpam-642	504	3	fitted	fit	VERB
ejpam-642	504	4	models	model	NOUN
ejpam-642	504	5	,	,	PUNCT
ejpam-642	504	6	we	we	PRON
ejpam-642	504	7	see	see	VERB
ejpam-642	504	8	interactions	interaction	NOUN
ejpam-642	504	9	among	among	ADP
ejpam-642	504	10	names	name	NOUN
ejpam-642	504	11	(	(	PUNCT
ejpam-642	504	12	first	first	ADJ
ejpam-642	504	13	,	,	PUNCT
ejpam-642	504	14	middle	middle	ADJ
ejpam-642	504	15	and	and	CCONJ
ejpam-642	504	16	last	last	ADJ
ejpam-642	504	17	)	)	PUNCT
ejpam-642	504	18	,	,	PUNCT
ejpam-642	504	19	date	date	NOUN
ejpam-642	504	20	of	of	ADP
ejpam-642	504	21	birth	birth	NOUN
ejpam-642	504	22	,	,	PUNCT
ejpam-642	504	23	and	and	CCONJ
ejpam-642	504	24	postal	postal	PROPN
ejpam-642	504	25	code	code	PROPN
ejpam-642	504	26	.	.	PUNCT
ejpam-642	505	1	note	note	VERB
ejpam-642	505	2	that	that	SCONJ
ejpam-642	505	3	these	these	DET
ejpam-642	505	4	four	four	NUM
ejpam-642	505	5	synthetic	synthetic	ADJ
ejpam-642	505	6	files	file	NOUN
ejpam-642	505	7	were	be	AUX
ejpam-642	505	8	actually	actually	ADV
ejpam-642	505	9	constructed	construct	VERB
ejpam-642	505	10	from	from	ADP
ejpam-642	505	11	other	other	ADJ
ejpam-642	505	12	survey	survey	NOUN
ejpam-642	505	13	sources	source	NOUN
ejpam-642	505	14	about	about	ADP
ejpam-642	505	15	pei	pei	PROPN
ejpam-642	505	16	.	.	PUNCT
ejpam-642	506	1	so	so	ADV
ejpam-642	506	2	more	more	ADV
ejpam-642	506	3	or	or	CCONJ
ejpam-642	506	4	less	less	ADV
ejpam-642	506	5	they	they	PRON
ejpam-642	506	6	reflect	reflect	VERB
ejpam-642	506	7	the	the	DET
ejpam-642	506	8	local	local	ADJ
ejpam-642	506	9	information	information	NOUN
ejpam-642	506	10	.	.	PUNCT
ejpam-642	507	1	the	the	DET
ejpam-642	507	2	reason	reason	NOUN
ejpam-642	507	3	for	for	ADP
ejpam-642	507	4	the	the	DET
ejpam-642	507	5	interactions	interaction	NOUN
ejpam-642	507	6	among	among	ADP
ejpam-642	507	7	names	name	NOUN
ejpam-642	507	8	may	may	AUX
ejpam-642	507	9	be	be	AUX
ejpam-642	507	10	traced	trace	VERB
ejpam-642	507	11	back	back	ADP
ejpam-642	507	12	historically	historically	ADV
ejpam-642	507	13	on	on	ADP
ejpam-642	507	14	the	the	DET
ejpam-642	507	15	residents	resident	NOUN
ejpam-642	507	16	’	’	PART
ejpam-642	507	17	family	family	NOUN
ejpam-642	507	18	,	,	PUNCT
ejpam-642	507	19	origin	origin	NOUN
ejpam-642	507	20	and	and	CCONJ
ejpam-642	507	21	culture	culture	NOUN
ejpam-642	507	22	.	.	PUNCT
ejpam-642	508	1	in	in	ADP
ejpam-642	508	2	addition	addition	NOUN
ejpam-642	508	3	,	,	PUNCT
ejpam-642	508	4	pei	pei	PROPN
ejpam-642	508	5	is	be	AUX
ejpam-642	508	6	a	a	DET
ejpam-642	508	7	relatively	relatively	ADV
ejpam-642	508	8	closed	closed	ADJ
ejpam-642	508	9	island	island	NOUN
ejpam-642	508	10	and	and	CCONJ
ejpam-642	508	11	the	the	DET
ejpam-642	508	12	economy	economy	NOUN
ejpam-642	508	13	is	be	AUX
ejpam-642	508	14	not	not	PART
ejpam-642	508	15	as	as	ADV
ejpam-642	508	16	active	active	ADJ
ejpam-642	508	17	as	as	ADP
ejpam-642	508	18	most	most	ADJ
ejpam-642	508	19	of	of	ADP
ejpam-642	508	20	other	other	ADJ
ejpam-642	508	21	canadian	canadian	ADJ
ejpam-642	508	22	provinces	province	NOUN
ejpam-642	508	23	.	.	PUNCT
ejpam-642	509	1	families	family	NOUN
ejpam-642	509	2	who	who	PRON
ejpam-642	509	3	are	be	AUX
ejpam-642	509	4	relatives	relative	NOUN
ejpam-642	509	5	might	might	AUX
ejpam-642	509	6	be	be	AUX
ejpam-642	509	7	more	more	ADV
ejpam-642	509	8	likely	likely	ADJ
ejpam-642	509	9	to	to	PART
ejpam-642	509	10	live	live	VERB
ejpam-642	509	11	in	in	ADP
ejpam-642	509	12	a	a	DET
ejpam-642	509	13	closed	closed	ADJ
ejpam-642	509	14	regions	region	NOUN
ejpam-642	509	15	due	due	ADJ
ejpam-642	509	16	to	to	ADP
ejpam-642	509	17	some	some	DET
ejpam-642	509	18	reasons	reason	NOUN
ejpam-642	509	19	like	like	ADP
ejpam-642	509	20	traditional	traditional	ADJ
ejpam-642	509	21	farming	farming	NOUN
ejpam-642	509	22	or	or	CCONJ
ejpam-642	509	23	other	other	ADJ
ejpam-642	509	24	social	social	ADJ
ejpam-642	509	25	activies	activie	NOUN
ejpam-642	509	26	,	,	PUNCT
ejpam-642	509	27	which	which	PRON
ejpam-642	509	28	could	could	AUX
ejpam-642	509	29	lead	lead	VERB
ejpam-642	509	30	to	to	ADP
ejpam-642	509	31	the	the	DET
ejpam-642	509	32	spatial	spatial	ADJ
ejpam-642	509	33	patterns	pattern	NOUN
ejpam-642	509	34	for	for	ADP
ejpam-642	509	35	surnames	surname	NOUN
ejpam-642	509	36	.	.	PUNCT
ejpam-642	510	1	hence	hence	ADV
ejpam-642	510	2	,	,	PUNCT
ejpam-642	510	3	these	these	DET
ejpam-642	510	4	interactions	interaction	NOUN
ejpam-642	510	5	are	be	AUX
ejpam-642	510	6	caused	cause	VERB
ejpam-642	510	7	or	or	CCONJ
ejpam-642	510	8	confounded	confound	VERB
ejpam-642	510	9	by	by	ADP
ejpam-642	510	10	the	the	DET
ejpam-642	510	11	latent	latent	NOUN
ejpam-642	510	12	factors	factor	NOUN
ejpam-642	510	13	.	.	PUNCT
ejpam-642	511	1	of	of	ADP
ejpam-642	511	2	course	course	NOUN
ejpam-642	511	3	,	,	PUNCT
ejpam-642	511	4	these	these	DET
ejpam-642	511	5	reasons	reason	NOUN
ejpam-642	511	6	could	could	AUX
ejpam-642	511	7	lead	lead	VERB
ejpam-642	511	8	to	to	ADP
ejpam-642	511	9	higher	high	ADJ
ejpam-642	511	10	order	order	NOUN
ejpam-642	511	11	interactions	interaction	NOUN
ejpam-642	511	12	,	,	PUNCT
ejpam-642	511	13	say	say	VERB
ejpam-642	511	14	three	three	NUM
ejpam-642	511	15	-	-	PUNCT
ejpam-642	511	16	factor	factor	NOUN
ejpam-642	511	17	interactions	interaction	NOUN
ejpam-642	511	18	.	.	PUNCT
ejpam-642	512	1	however	however	ADV
ejpam-642	512	2	,	,	PUNCT
ejpam-642	512	3	with	with	ADP
ejpam-642	512	4	the	the	DET
ejpam-642	512	5	current	current	ADJ
ejpam-642	512	6	sample	sample	NOUN
ejpam-642	512	7	sizes	size	NOUN
ejpam-642	512	8	,	,	PUNCT
ejpam-642	512	9	only	only	ADV
ejpam-642	512	10	twofactor	twofactor	NOUN
ejpam-642	512	11	interactions	interaction	NOUN
ejpam-642	512	12	are	be	AUX
ejpam-642	512	13	significant	significant	ADJ
ejpam-642	512	14	enough	enough	ADV
ejpam-642	512	15	to	to	PART
ejpam-642	512	16	be	be	AUX
ejpam-642	512	17	included	include	VERB
ejpam-642	512	18	in	in	ADP
ejpam-642	512	19	the	the	DET
ejpam-642	512	20	selected	select	VERB
ejpam-642	512	21	models	model	NOUN
ejpam-642	512	22	.	.	PUNCT
ejpam-642	513	1	we	we	PRON
ejpam-642	513	2	also	also	ADV
ejpam-642	513	3	build	build	VERB
ejpam-642	513	4	models	model	NOUN
ejpam-642	513	5	for	for	ADP
ejpam-642	513	6	male	male	ADJ
ejpam-642	513	7	and	and	CCONJ
ejpam-642	513	8	female	female	ADJ
ejpam-642	513	9	block	block	NOUN
ejpam-642	513	10	pairs	pair	NOUN
ejpam-642	513	11	separately	separately	ADV
ejpam-642	513	12	in	in	ADP
ejpam-642	513	13	each	each	DET
ejpam-642	513	14	task	task	NOUN
ejpam-642	513	15	.	.	PUNCT
ejpam-642	514	1	results	result	NOUN
ejpam-642	514	2	are	be	AUX
ejpam-642	514	3	very	very	ADV
ejpam-642	514	4	similar	similar	ADJ
ejpam-642	514	5	and	and	CCONJ
ejpam-642	514	6	hence	hence	ADV
ejpam-642	514	7	omitted	omit	VERB
ejpam-642	514	8	.	.	PUNCT
ejpam-642	515	1	it	it	PRON
ejpam-642	515	2	is	be	AUX
ejpam-642	515	3	straightforward	straightforward	ADJ
ejpam-642	515	4	to	to	PART
ejpam-642	515	5	calculate	calculate	VERB
ejpam-642	515	6	matching	match	VERB
ejpam-642	515	7	weights	weight	NOUN
ejpam-642	515	8	and	and	CCONJ
ejpam-642	515	9	perform	perform	VERB
ejpam-642	515	10	record	record	NOUN
ejpam-642	515	11	pair	pair	NOUN
ejpam-642	515	12	partition	partition	NOUN
ejpam-642	515	13	for	for	ADP
ejpam-642	515	14	each	each	DET
ejpam-642	515	15	task	task	NOUN
ejpam-642	515	16	according	accord	VERB
ejpam-642	515	17	to	to	ADP
ejpam-642	515	18	the	the	DET
ejpam-642	515	19	corresponding	corresponding	ADJ
ejpam-642	515	20	selected	select	VERB
ejpam-642	515	21	model	model	NOUN
ejpam-642	515	22	;	;	PUNCT
ejpam-642	515	23	this	this	DET
ejpam-642	515	24	part	part	NOUN
ejpam-642	515	25	is	be	AUX
ejpam-642	515	26	also	also	ADV
ejpam-642	515	27	omitted	omit	VERB
ejpam-642	515	28	as	as	SCONJ
ejpam-642	515	29	it	it	PRON
ejpam-642	515	30	is	be	AUX
ejpam-642	515	31	not	not	PART
ejpam-642	515	32	the	the	DET
ejpam-642	515	33	focus	focus	NOUN
ejpam-642	515	34	of	of	ADP
ejpam-642	515	35	this	this	DET
ejpam-642	515	36	paper	paper	NOUN
ejpam-642	515	37	.	.	PUNCT
ejpam-642	516	1	6	6	NUM
ejpam-642	516	2	.	.	X
ejpam-642	516	3	discussion	discussion	NOUN
ejpam-642	516	4	the	the	DET
ejpam-642	516	5	simplified	simplified	ADJ
ejpam-642	516	6	method	method	NOUN
ejpam-642	516	7	for	for	ADP
ejpam-642	516	8	the	the	DET
ejpam-642	516	9	em	em	PROPN
ejpam-642	516	10	algorithm	algorithm	NOUN
ejpam-642	516	11	is	be	AUX
ejpam-642	516	12	computationally	computationally	ADV
ejpam-642	516	13	effective	effective	ADJ
ejpam-642	516	14	.	.	PUNCT
ejpam-642	517	1	it	it	PRON
ejpam-642	517	2	also	also	ADV
ejpam-642	517	3	extends	extend	VERB
ejpam-642	517	4	the	the	DET
ejpam-642	517	5	jaro	jaro	PROPN
ejpam-642	517	6	’s	’s	PART
ejpam-642	517	7	method	method	NOUN
ejpam-642	517	8	for	for	ADP
ejpam-642	517	9	a	a	DET
ejpam-642	517	10	general	general	ADJ
ejpam-642	517	11	contingency	contingency	NOUN
ejpam-642	517	12	table	table	NOUN
ejpam-642	517	13	in	in	ADP
ejpam-642	517	14	fitting	fitting	ADJ
ejpam-642	517	15	loglinear	loglinear	NOUN
ejpam-642	517	16	mixture	mixture	NOUN
ejpam-642	517	17	models	model	NOUN
ejpam-642	517	18	.	.	PUNCT
ejpam-642	518	1	this	this	DET
ejpam-642	518	2	method	method	NOUN
ejpam-642	518	3	can	can	AUX
ejpam-642	518	4	be	be	AUX
ejpam-642	518	5	easily	easily	ADV
ejpam-642	518	6	implemented	implement	VERB
ejpam-642	518	7	in	in	ADP
ejpam-642	518	8	a	a	DET
ejpam-642	518	9	sophisticated	sophisticated	ADJ
ejpam-642	518	10	statistical	statistical	ADJ
ejpam-642	518	11	software	software	NOUN
ejpam-642	518	12	using	use	VERB
ejpam-642	518	13	existing	exist	VERB
ejpam-642	518	14	packages	package	NOUN
ejpam-642	518	15	or	or	CCONJ
ejpam-642	518	16	functions	function	NOUN
ejpam-642	518	17	for	for	ADP
ejpam-642	518	18	loglinear	loglinear	NOUN
ejpam-642	518	19	model	model	NOUN
ejpam-642	518	20	fitting	fitting	ADJ
ejpam-642	518	21	.	.	PUNCT
ejpam-642	519	1	we	we	PRON
ejpam-642	519	2	utilize	utilize	VERB
ejpam-642	519	3	r	r	NOUN
ejpam-642	519	4	in	in	ADP
ejpam-642	519	5	this	this	DET
ejpam-642	519	6	study	study	NOUN
ejpam-642	519	7	.	.	PUNCT
ejpam-642	520	1	compared	compare	VERB
ejpam-642	520	2	with	with	ADP
ejpam-642	520	3	the	the	DET
ejpam-642	520	4	jaro	jaro	NOUN
ejpam-642	520	5	’s	’s	PART
ejpam-642	520	6	method	method	NOUN
ejpam-642	520	7	,	,	PUNCT
ejpam-642	520	8	we	we	PRON
ejpam-642	520	9	extend	extend	VERB
ejpam-642	520	10	not	not	PART
ejpam-642	520	11	only	only	ADV
ejpam-642	520	12	the	the	DET
ejpam-642	520	13	factor	factor	NOUN
ejpam-642	520	14	interactions	interaction	NOUN
ejpam-642	520	15	but	but	CCONJ
ejpam-642	520	16	also	also	ADV
ejpam-642	520	17	the	the	DET
ejpam-642	520	18	number	number	NOUN
ejpam-642	520	19	of	of	ADP
ejpam-642	520	20	levels	level	NOUN
ejpam-642	520	21	of	of	ADP
ejpam-642	520	22	each	each	DET
ejpam-642	520	23	factor	factor	NOUN
ejpam-642	520	24	.	.	PUNCT
ejpam-642	521	1	the	the	DET
ejpam-642	521	2	factor	factor	NOUN
ejpam-642	521	3	is	be	AUX
ejpam-642	521	4	not	not	PART
ejpam-642	521	5	necessarily	necessarily	ADV
ejpam-642	521	6	binary	binary	ADJ
ejpam-642	521	7	.	.	PUNCT
ejpam-642	522	1	it	it	PRON
ejpam-642	522	2	can	can	AUX
ejpam-642	522	3	be	be	AUX
ejpam-642	522	4	more	more	ADJ
ejpam-642	522	5	than	than	ADP
ejpam-642	522	6	two	two	NUM
ejpam-642	522	7	levels	level	NOUN
ejpam-642	522	8	for	for	ADP
ejpam-642	522	9	each	each	DET
ejpam-642	522	10	factor	factor	NOUN
ejpam-642	522	11	.	.	PUNCT
ejpam-642	523	1	this	this	DET
ejpam-642	523	2	data	data	NOUN
ejpam-642	523	3	-	-	PUNCT
ejpam-642	523	4	driven	drive	VERB
ejpam-642	523	5	strategy	strategy	NOUN
ejpam-642	523	6	of	of	ADP
ejpam-642	523	7	model	model	NOUN
ejpam-642	523	8	building	building	NOUN
ejpam-642	523	9	avoids	avoid	VERB
ejpam-642	523	10	subjective	subjective	ADJ
ejpam-642	523	11	specification	specification	NOUN
ejpam-642	523	12	of	of	ADP
ejpam-642	523	13	the	the	DET
ejpam-642	523	14	underlying	underlying	ADJ
ejpam-642	523	15	loglinear	loglinear	NOUN
ejpam-642	523	16	mixture	mixture	NOUN
ejpam-642	523	17	model	model	NOUN
ejpam-642	523	18	,	,	PUNCT
ejpam-642	523	19	and	and	CCONJ
ejpam-642	523	20	thus	thus	ADV
ejpam-642	523	21	minimizes	minimize	VERB
ejpam-642	523	22	the	the	DET
ejpam-642	523	23	chance	chance	NOUN
ejpam-642	523	24	of	of	ADP
ejpam-642	523	25	underfitting	underfitte	VERB
ejpam-642	523	26	or	or	CCONJ
ejpam-642	523	27	overfitting	overfitte	VERB
ejpam-642	523	28	references	reference	NOUN
ejpam-642	523	29	161	161	NUM
ejpam-642	523	30	caused	cause	VERB
ejpam-642	523	31	by	by	ADP
ejpam-642	523	32	biased	biased	ADJ
ejpam-642	523	33	opinions	opinion	NOUN
ejpam-642	523	34	.	.	PUNCT
ejpam-642	524	1	often	often	ADV
ejpam-642	524	2	it	it	PRON
ejpam-642	524	3	yields	yield	VERB
ejpam-642	524	4	a	a	DET
ejpam-642	524	5	model	model	NOUN
ejpam-642	524	6	consistent	consistent	ADJ
ejpam-642	524	7	with	with	ADP
ejpam-642	524	8	the	the	DET
ejpam-642	524	9	true	true	ADJ
ejpam-642	524	10	one	one	NOUN
ejpam-642	524	11	or	or	CCONJ
ejpam-642	524	12	close	close	ADJ
ejpam-642	524	13	to	to	ADP
ejpam-642	524	14	the	the	DET
ejpam-642	524	15	true	true	ADJ
ejpam-642	524	16	one	one	NUM
ejpam-642	524	17	.	.	PUNCT
ejpam-642	525	1	therefore	therefore	ADV
ejpam-642	525	2	,	,	PUNCT
ejpam-642	525	3	it	it	PRON
ejpam-642	525	4	is	be	AUX
ejpam-642	525	5	more	more	ADV
ejpam-642	525	6	likely	likely	ADJ
ejpam-642	525	7	to	to	PART
ejpam-642	525	8	result	result	VERB
ejpam-642	525	9	in	in	ADP
ejpam-642	525	10	a	a	DET
ejpam-642	525	11	better	well	ADJ
ejpam-642	525	12	partition	partition	NOUN
ejpam-642	525	13	for	for	ADP
ejpam-642	525	14	record	record	NOUN
ejpam-642	525	15	linkage	linkage	NOUN
ejpam-642	525	16	.	.	PUNCT
ejpam-642	526	1	however	however	ADV
ejpam-642	526	2	,	,	PUNCT
ejpam-642	526	3	since	since	SCONJ
ejpam-642	526	4	the	the	DET
ejpam-642	526	5	proposed	propose	VERB
ejpam-642	526	6	strategy	strategy	NOUN
ejpam-642	526	7	is	be	AUX
ejpam-642	526	8	not	not	PART
ejpam-642	526	9	a	a	DET
ejpam-642	526	10	global	global	ADJ
ejpam-642	526	11	search	search	NOUN
ejpam-642	526	12	,	,	PUNCT
ejpam-642	526	13	the	the	DET
ejpam-642	526	14	selected	select	VERB
ejpam-642	526	15	model	model	NOUN
ejpam-642	526	16	may	may	AUX
ejpam-642	526	17	not	not	PART
ejpam-642	526	18	be	be	AUX
ejpam-642	526	19	the	the	DET
ejpam-642	526	20	best	good	ADJ
ejpam-642	526	21	.	.	PUNCT
ejpam-642	527	1	future	future	ADJ
ejpam-642	527	2	investigations	investigation	NOUN
ejpam-642	527	3	regarding	regard	VERB
ejpam-642	527	4	other	other	ADJ
ejpam-642	527	5	testing	testing	NOUN
ejpam-642	527	6	methods	method	NOUN
ejpam-642	527	7	and	and	CCONJ
ejpam-642	527	8	building	building	NOUN
ejpam-642	527	9	strategies	strategy	NOUN
ejpam-642	527	10	,	,	PUNCT
ejpam-642	527	11	as	as	ADV
ejpam-642	527	12	well	well	ADV
ejpam-642	527	13	as	as	ADP
ejpam-642	527	14	error	error	NOUN
ejpam-642	527	15	rate	rate	NOUN
ejpam-642	527	16	investigation	investigation	NOUN
ejpam-642	527	17	are	be	AUX
ejpam-642	527	18	under	under	ADP
ejpam-642	527	19	way	way	NOUN
ejpam-642	527	20	.	.	PUNCT
ejpam-642	528	1	we	we	PRON
ejpam-642	528	2	welcome	welcome	VERB
ejpam-642	528	3	any	any	DET
ejpam-642	528	4	real	real	ADJ
ejpam-642	528	5	case	case	NOUN
ejpam-642	528	6	collaboration	collaboration	NOUN
ejpam-642	528	7	.	.	PUNCT
ejpam-642	529	1	acknowledgements	acknowledgement	NOUN
ejpam-642	529	2	this	this	DET
ejpam-642	529	3	research	research	NOUN
ejpam-642	529	4	has	have	AUX
ejpam-642	529	5	been	be	AUX
ejpam-642	529	6	supported	support	VERB
ejpam-642	529	7	by	by	ADP
ejpam-642	529	8	an	an	DET
ejpam-642	529	9	nserc	nserc	NOUN
ejpam-642	529	10	discovery	discovery	NOUN
ejpam-642	529	11	grant	grant	NOUN
ejpam-642	529	12	.	.	PUNCT
ejpam-642	530	1	we	we	PRON
ejpam-642	530	2	thank	thank	VERB
ejpam-642	530	3	professor	professor	NOUN
ejpam-642	530	4	román	román	NOUN
ejpam-642	530	5	viveros	vivero	NOUN
ejpam-642	530	6	-	-	PUNCT
ejpam-642	530	7	aguilera	aguilera	PROPN
ejpam-642	530	8	at	at	ADP
ejpam-642	530	9	mcmaster	mcmaster	PROPN
ejpam-642	530	10	university	university	PROPN
ejpam-642	530	11	,	,	PUNCT
ejpam-642	530	12	canada	canada	PROPN
ejpam-642	530	13	for	for	ADP
ejpam-642	530	14	his	his	PRON
ejpam-642	530	15	helpful	helpful	ADJ
ejpam-642	530	16	discussion	discussion	NOUN
ejpam-642	530	17	,	,	PUNCT
ejpam-642	530	18	and	and	CCONJ
ejpam-642	530	19	miss	miss	VERB
ejpam-642	530	20	yunna	yunna	NOUN
ejpam-642	530	21	song	song	NOUN
ejpam-642	530	22	for	for	ADP
ejpam-642	530	23	her	her	PRON
ejpam-642	530	24	participation	participation	NOUN
ejpam-642	530	25	as	as	ADP
ejpam-642	530	26	a	a	DET
ejpam-642	530	27	team	team	NOUN
ejpam-642	530	28	member	member	NOUN
ejpam-642	530	29	in	in	ADP
ejpam-642	530	30	the	the	DET
ejpam-642	530	31	record	record	NOUN
ejpam-642	530	32	linkage	linkage	NOUN
ejpam-642	530	33	case	case	NOUN
ejpam-642	530	34	study	study	NOUN
ejpam-642	530	35	at	at	ADP
ejpam-642	530	36	ssc	ssc	PROPN
ejpam-642	530	37	2006	2006	NUM
ejpam-642	530	38	.	.	PUNCT
ejpam-642	531	1	in	in	ADP
ejpam-642	531	2	addition	addition	NOUN
ejpam-642	531	3	,	,	PUNCT
ejpam-642	531	4	we	we	PRON
ejpam-642	531	5	are	be	AUX
ejpam-642	531	6	grateful	grateful	ADJ
ejpam-642	531	7	to	to	ADP
ejpam-642	531	8	a	a	DET
ejpam-642	531	9	referee	referee	NOUN
ejpam-642	531	10	for	for	ADP
ejpam-642	531	11	comments	comment	NOUN
ejpam-642	531	12	leading	lead	VERB
ejpam-642	531	13	to	to	ADP
ejpam-642	531	14	an	an	DET
ejpam-642	531	15	improved	improved	ADJ
ejpam-642	531	16	presentation	presentation	NOUN
ejpam-642	531	17	.	.	PUNCT
ejpam-642	532	1	references	reference	NOUN
ejpam-642	532	2	[	[	X
ejpam-642	532	3	1	1	X
ejpam-642	532	4	]	]	PUNCT
ejpam-642	532	5	e.	e.	PROPN
ejpam-642	532	6	d.	d.	PROPN
ejpam-642	532	7	acheson	acheson	PROPN
ejpam-642	532	8	.	.	PUNCT
ejpam-642	533	1	record	record	NOUN
ejpam-642	533	2	linkage	linkage	NOUN
ejpam-642	533	3	in	in	ADP
ejpam-642	533	4	medicine	medicine	NOUN
ejpam-642	533	5	.	.	PUNCT
ejpam-642	534	1	e.	e.	PROPN
ejpam-642	534	2	&	&	CCONJ
ejpam-642	534	3	s.	s.	PROPN
ejpam-642	534	4	livingstone	livingstone	PROPN
ejpam-642	534	5	ltd	ltd	PROPN
ejpam-642	534	6	.	.	PROPN
ejpam-642	534	7	,	,	PUNCT
ejpam-642	534	8	edinburgh	edinburgh	PROPN
ejpam-642	534	9	and	and	CCONJ
ejpam-642	534	10	london	london	PROPN
ejpam-642	534	11	,	,	PUNCT
ejpam-642	534	12	1968	1968	NUM
ejpam-642	534	13	.	.	PUNCT
ejpam-642	535	1	[	[	X
ejpam-642	535	2	2	2	NUM
ejpam-642	535	3	]	]	PUNCT
ejpam-642	535	4	a.	a.	NOUN
ejpam-642	535	5	agresti	agresti	PROPN
ejpam-642	535	6	.	.	PUNCT
ejpam-642	536	1	categorical	categorical	ADJ
ejpam-642	536	2	data	datum	NOUN
ejpam-642	536	3	analysis	analysis	NOUN
ejpam-642	536	4	,	,	PUNCT
ejpam-642	536	5	second	second	ADJ
ejpam-642	536	6	edition	edition	NOUN
ejpam-642	536	7	.	.	PUNCT
ejpam-642	537	1	john	john	PROPN
ejpam-642	537	2	wiley	wiley	PROPN
ejpam-642	537	3	&	&	CCONJ
ejpam-642	537	4	sons	sons	PROPN
ejpam-642	537	5	,	,	PUNCT
ejpam-642	537	6	inc	inc	PROPN
ejpam-642	537	7	.	.	PROPN
ejpam-642	537	8	,	,	PUNCT
ejpam-642	537	9	hoboken	hoboken	PROPN
ejpam-642	537	10	,	,	PUNCT
ejpam-642	537	11	new	new	PROPN
ejpam-642	537	12	jersey	jersey	PROPN
ejpam-642	537	13	,	,	PUNCT
ejpam-642	537	14	2002	2002	NUM
ejpam-642	537	15	.	.	PUNCT
ejpam-642	538	1	[	[	X
ejpam-642	538	2	3	3	X
ejpam-642	538	3	]	]	X
ejpam-642	538	4	j.	j.	PROPN
ejpam-642	538	5	b.	b.	PROPN
ejpam-642	538	6	copas	copas	PROPN
ejpam-642	538	7	and	and	CCONJ
ejpam-642	538	8	f.	f.	PROPN
ejpam-642	538	9	j.	j.	PROPN
ejpam-642	538	10	hilton	hilton	PROPN
ejpam-642	538	11	.	.	PUNCT
ejpam-642	539	1	record	record	NOUN
ejpam-642	539	2	linkage	linkage	NOUN
ejpam-642	539	3	:	:	PUNCT
ejpam-642	539	4	statistical	statistical	ADJ
ejpam-642	539	5	models	model	NOUN
ejpam-642	539	6	for	for	ADP
ejpam-642	539	7	matching	match	VERB
ejpam-642	539	8	computer	computer	NOUN
ejpam-642	539	9	records	record	NOUN
ejpam-642	539	10	.	.	PUNCT
ejpam-642	540	1	journal	journal	NOUN
ejpam-642	540	2	of	of	ADP
ejpam-642	540	3	the	the	DET
ejpam-642	540	4	royal	royal	ADJ
ejpam-642	540	5	statistical	statistical	ADJ
ejpam-642	540	6	society	society	NOUN
ejpam-642	540	7	,	,	PUNCT
ejpam-642	540	8	a	a	DET
ejpam-642	540	9	,	,	PUNCT
ejpam-642	540	10	153:287–320	153:287–320	NUM
ejpam-642	540	11	,	,	PUNCT
ejpam-642	540	12	1990	1990	NUM
ejpam-642	540	13	.	.	PUNCT
ejpam-642	541	1	[	[	X
ejpam-642	541	2	4	4	NUM
ejpam-642	541	3	]	]	PUNCT
ejpam-642	541	4	a.	a.	NOUN
ejpam-642	541	5	p.	p.	PROPN
ejpam-642	541	6	dempster	dempster	PROPN
ejpam-642	541	7	,	,	PUNCT
ejpam-642	541	8	n.	n.	PROPN
ejpam-642	541	9	m.	m.	PROPN
ejpam-642	541	10	laird	laird	PROPN
ejpam-642	541	11	,	,	PUNCT
ejpam-642	541	12	and	and	CCONJ
ejpam-642	542	1	d.	d.	PROPN
ejpam-642	542	2	b.	b.	PROPN
ejpam-642	542	3	rubin	rubin	PROPN
ejpam-642	542	4	.	.	PUNCT
ejpam-642	543	1	maximum	maximum	ADJ
ejpam-642	543	2	likelihood	likelihood	NOUN
ejpam-642	543	3	from	from	ADP
ejpam-642	543	4	incomplete	incomplete	ADJ
ejpam-642	543	5	data	datum	NOUN
ejpam-642	543	6	via	via	ADP
ejpam-642	543	7	the	the	DET
ejpam-642	543	8	em	em	PROPN
ejpam-642	543	9	algorithm	algorithm	NOUN
ejpam-642	543	10	.	.	PUNCT
ejpam-642	544	1	journal	journal	NOUN
ejpam-642	544	2	of	of	ADP
ejpam-642	544	3	the	the	DET
ejpam-642	544	4	royal	royal	ADJ
ejpam-642	544	5	statistical	statistical	ADJ
ejpam-642	544	6	society	society	NOUN
ejpam-642	544	7	,	,	PUNCT
ejpam-642	544	8	b	b	NOUN
ejpam-642	544	9	,	,	PUNCT
ejpam-642	544	10	39:1–38	39:1–38	NUM
ejpam-642	544	11	,	,	PUNCT
ejpam-642	544	12	1977	1977	NUM
ejpam-642	544	13	.	.	PUNCT
ejpam-642	545	1	[	[	X
ejpam-642	545	2	5	5	NUM
ejpam-642	545	3	]	]	PUNCT
ejpam-642	545	4	m.	m.	NOUN
ejpam-642	545	5	e.	e.	PROPN
ejpam-642	545	6	fair	fair	PROPN
ejpam-642	545	7	and	and	CCONJ
ejpam-642	545	8	p.	p.	NOUN
ejpam-642	545	9	whitridge	whitridge	NOUN
ejpam-642	545	10	.	.	PUNCT
ejpam-642	546	1	tutorial	tutorial	NOUN
ejpam-642	546	2	on	on	ADP
ejpam-642	546	3	record	record	NOUN
ejpam-642	546	4	linkage	linkage	NOUN
ejpam-642	546	5	.	.	PUNCT
ejpam-642	547	1	in	in	ADP
ejpam-642	547	2	w.	w.	PROPN
ejpam-642	547	3	alvey	alvey	PROPN
ejpam-642	547	4	and	and	CCONJ
ejpam-642	547	5	b.	b.	PROPN
ejpam-642	547	6	jamerson	jamerson	PROPN
ejpam-642	547	7	,	,	PUNCT
ejpam-642	547	8	editors	editor	NOUN
ejpam-642	547	9	,	,	PUNCT
ejpam-642	547	10	record	record	NOUN
ejpam-642	547	11	linkage	linkage	NOUN
ejpam-642	547	12	techniques	technique	NOUN
ejpam-642	547	13	1997	1997	NUM
ejpam-642	547	14	:	:	PUNCT
ejpam-642	547	15	proceedings	proceeding	NOUN
ejpam-642	547	16	of	of	ADP
ejpam-642	547	17	an	an	DET
ejpam-642	547	18	international	international	ADJ
ejpam-642	547	19	workshop	workshop	NOUN
ejpam-642	547	20	and	and	CCONJ
ejpam-642	547	21	exposition	exposition	NOUN
ejpam-642	547	22	.	.	PUNCT
ejpam-642	547	23	,	,	PUNCT
ejpam-642	547	24	pages	page	NOUN
ejpam-642	547	25	457–482	457–482	PROPN
ejpam-642	547	26	,	,	PUNCT
ejpam-642	547	27	arlington	arlington	PROPN
ejpam-642	547	28	,	,	PUNCT
ejpam-642	547	29	va	va	PROPN
ejpam-642	547	30	,	,	PUNCT
ejpam-642	547	31	1997	1997	NUM
ejpam-642	547	32	.	.	PUNCT
ejpam-642	548	1	federal	federal	ADJ
ejpam-642	548	2	committee	committee	PROPN
ejpam-642	548	3	on	on	ADP
ejpam-642	548	4	statistical	statistical	ADJ
ejpam-642	548	5	methodology	methodology	NOUN
ejpam-642	548	6	,	,	PUNCT
ejpam-642	548	7	and	and	CCONJ
ejpam-642	548	8	office	office	NOUN
ejpam-642	548	9	of	of	ADP
ejpam-642	548	10	management	management	NOUN
ejpam-642	548	11	and	and	CCONJ
ejpam-642	548	12	budget	budget	NOUN
ejpam-642	548	13	.	.	PUNCT
ejpam-642	549	1	[	[	X
ejpam-642	549	2	6	6	NUM
ejpam-642	549	3	]	]	PUNCT
ejpam-642	550	1	i.	i.	PROPN
ejpam-642	550	2	p.	p.	PROPN
ejpam-642	550	3	fellegi	fellegi	PROPN
ejpam-642	551	1	and	and	CCONJ
ejpam-642	551	2	a.	a.	PROPN
ejpam-642	551	3	b.	b.	PROPN
ejpam-642	551	4	sunter	sunter	PROPN
ejpam-642	551	5	.	.	PUNCT
ejpam-642	552	1	a	a	DET
ejpam-642	552	2	theory	theory	NOUN
ejpam-642	552	3	for	for	ADP
ejpam-642	552	4	record	record	NOUN
ejpam-642	552	5	linkage	linkage	NOUN
ejpam-642	552	6	.	.	PUNCT
ejpam-642	553	1	journal	journal	NOUN
ejpam-642	553	2	of	of	ADP
ejpam-642	553	3	the	the	DET
ejpam-642	553	4	american	american	PROPN
ejpam-642	553	5	statistical	statistical	PROPN
ejpam-642	553	6	association	association	PROPN
ejpam-642	553	7	,	,	PUNCT
ejpam-642	553	8	64:1183–1210	64:1183–1210	NUM
ejpam-642	553	9	,	,	PUNCT
ejpam-642	553	10	1969	1969	NUM
ejpam-642	553	11	.	.	PUNCT
ejpam-642	554	1	[	[	X
ejpam-642	554	2	7	7	X
ejpam-642	554	3	]	]	PUNCT
ejpam-642	554	4	m.	m.	NOUN
ejpam-642	554	5	a.	a.	NOUN
ejpam-642	554	6	jaro	jaro	PROPN
ejpam-642	554	7	.	.	PUNCT
ejpam-642	555	1	advances	advance	NOUN
ejpam-642	555	2	in	in	ADP
ejpam-642	555	3	record	record	NOUN
ejpam-642	555	4	-	-	PUNCT
ejpam-642	555	5	linkage	linkage	NOUN
ejpam-642	555	6	methodology	methodology	NOUN
ejpam-642	555	7	as	as	SCONJ
ejpam-642	555	8	applied	apply	VERB
ejpam-642	555	9	to	to	ADP
ejpam-642	555	10	matching	match	VERB
ejpam-642	555	11	the	the	DET
ejpam-642	555	12	1985	1985	NUM
ejpam-642	555	13	census	census	NOUN
ejpam-642	555	14	of	of	ADP
ejpam-642	555	15	tampa	tampa	PROPN
ejpam-642	555	16	florida	florida	PROPN
ejpam-642	555	17	.	.	PUNCT
ejpam-642	556	1	journal	journal	PROPN
ejpam-642	556	2	of	of	ADP
ejpam-642	556	3	the	the	DET
ejpam-642	556	4	american	american	PROPN
ejpam-642	556	5	statistical	statistical	PROPN
ejpam-642	556	6	association	association	PROPN
ejpam-642	556	7	,	,	PUNCT
ejpam-642	556	8	84:414–420	84:414–420	NUM
ejpam-642	556	9	,	,	PUNCT
ejpam-642	556	10	1989	1989	NUM
ejpam-642	556	11	.	.	PUNCT
ejpam-642	557	1	[	[	X
ejpam-642	557	2	8	8	NUM
ejpam-642	557	3	]	]	PUNCT
ejpam-642	557	4	m.	m.	NOUN
ejpam-642	557	5	a.	a.	NOUN
ejpam-642	557	6	jaro	jaro	PROPN
ejpam-642	557	7	.	.	PUNCT
ejpam-642	558	1	probabilistic	probabilistic	ADJ
ejpam-642	558	2	linkage	linkage	NOUN
ejpam-642	558	3	of	of	ADP
ejpam-642	558	4	large	large	ADJ
ejpam-642	558	5	public	public	ADJ
ejpam-642	558	6	health	health	NOUN
ejpam-642	558	7	datafiles	datafile	NOUN
ejpam-642	558	8	.	.	PUNCT
ejpam-642	559	1	statistics	statistic	NOUN
ejpam-642	559	2	in	in	ADP
ejpam-642	559	3	medicine	medicine	NOUN
ejpam-642	559	4	,	,	PUNCT
ejpam-642	559	5	14:491–498	14:491–498	PROPN
ejpam-642	559	6	,	,	PUNCT
ejpam-642	559	7	1995	1995	NUM
ejpam-642	559	8	.	.	PUNCT
ejpam-642	560	1	[	[	X
ejpam-642	560	2	9	9	NUM
ejpam-642	560	3	]	]	PUNCT
ejpam-642	560	4	m.	m.	NOUN
ejpam-642	560	5	d.	d.	PROPN
ejpam-642	560	6	larsen	larsen	PROPN
ejpam-642	560	7	.	.	PUNCT
ejpam-642	561	1	modeling	model	VERB
ejpam-642	561	2	issues	issue	NOUN
ejpam-642	561	3	and	and	CCONJ
ejpam-642	561	4	the	the	DET
ejpam-642	561	5	use	use	NOUN
ejpam-642	561	6	of	of	ADP
ejpam-642	561	7	experience	experience	NOUN
ejpam-642	561	8	in	in	ADP
ejpam-642	561	9	record	record	NOUN
ejpam-642	561	10	linkage	linkage	NOUN
ejpam-642	561	11	.	.	PUNCT
ejpam-642	562	1	in	in	ADP
ejpam-642	562	2	w.	w.	PROPN
ejpam-642	562	3	alvey	alvey	PROPN
ejpam-642	562	4	and	and	CCONJ
ejpam-642	562	5	b.	b.	PROPN
ejpam-642	562	6	jamerson	jamerson	PROPN
ejpam-642	562	7	,	,	PUNCT
ejpam-642	562	8	editors	editor	NOUN
ejpam-642	562	9	,	,	PUNCT
ejpam-642	562	10	record	record	NOUN
ejpam-642	562	11	linkage	linkage	NOUN
ejpam-642	562	12	techniques	technique	NOUN
ejpam-642	562	13	1997	1997	NUM
ejpam-642	562	14	:	:	PUNCT
ejpam-642	562	15	proceedings	proceeding	NOUN
ejpam-642	562	16	of	of	ADP
ejpam-642	562	17	an	an	DET
ejpam-642	562	18	international	international	ADJ
ejpam-642	562	19	workshop	workshop	NOUN
ejpam-642	562	20	and	and	CCONJ
ejpam-642	562	21	exposition	exposition	NOUN
ejpam-642	562	22	.	.	PUNCT
ejpam-642	562	23	,	,	PUNCT
ejpam-642	562	24	pages	page	NOUN
ejpam-642	562	25	95–105	95–105	PROPN
ejpam-642	562	26	,	,	PUNCT
ejpam-642	562	27	arlington	arlington	PROPN
ejpam-642	562	28	,	,	PUNCT
ejpam-642	562	29	va	va	PROPN
ejpam-642	562	30	,	,	PUNCT
ejpam-642	562	31	1997	1997	NUM
ejpam-642	562	32	.	.	PUNCT
ejpam-642	563	1	federal	federal	ADJ
ejpam-642	563	2	committee	committee	PROPN
ejpam-642	563	3	on	on	ADP
ejpam-642	563	4	statistical	statistical	ADJ
ejpam-642	563	5	methodology	methodology	NOUN
ejpam-642	563	6	,	,	PUNCT
ejpam-642	563	7	and	and	CCONJ
ejpam-642	563	8	office	office	NOUN
ejpam-642	563	9	of	of	ADP
ejpam-642	563	10	management	management	NOUN
ejpam-642	563	11	and	and	CCONJ
ejpam-642	563	12	budget	budget	NOUN
ejpam-642	563	13	.	.	PUNCT
ejpam-642	564	1	references	reference	NOUN
ejpam-642	564	2	162	162	NUM
ejpam-642	565	1	[	[	X
ejpam-642	565	2	10	10	NUM
ejpam-642	565	3	]	]	PUNCT
ejpam-642	565	4	m.	m.	PROPN
ejpam-642	565	5	d.	d.	PROPN
ejpam-642	565	6	larsen	larsen	PROPN
ejpam-642	565	7	and	and	CCONJ
ejpam-642	566	1	d.	d.	PROPN
ejpam-642	566	2	b.	b.	PROPN
ejpam-642	566	3	rubin	rubin	PROPN
ejpam-642	566	4	.	.	PROPN
ejpam-642	567	1	iterative	iterative	PROPN
ejpam-642	567	2	automated	automate	VERB
ejpam-642	567	3	record	record	NOUN
ejpam-642	567	4	linkage	linkage	NOUN
ejpam-642	567	5	using	use	VERB
ejpam-642	567	6	mixture	mixture	NOUN
ejpam-642	567	7	models	model	NOUN
ejpam-642	567	8	.	.	PUNCT
ejpam-642	568	1	journal	journal	NOUN
ejpam-642	568	2	of	of	ADP
ejpam-642	568	3	the	the	DET
ejpam-642	568	4	american	american	PROPN
ejpam-642	568	5	statistical	statistical	PROPN
ejpam-642	568	6	association	association	NOUN
ejpam-642	568	7	,	,	PUNCT
ejpam-642	568	8	96:32–41	96:32–41	NUM
ejpam-642	568	9	,	,	PUNCT
ejpam-642	568	10	2001	2001	NUM
ejpam-642	568	11	.	.	PUNCT
ejpam-642	569	1	[	[	X
ejpam-642	569	2	11	11	NUM
ejpam-642	569	3	]	]	X
ejpam-642	569	4	g.	g.	PROPN
ejpam-642	569	5	j.	j.	PROPN
ejpam-642	569	6	mclachlan	mclachlan	PROPN
ejpam-642	569	7	and	and	CCONJ
ejpam-642	569	8	t.	t.	PROPN
ejpam-642	569	9	krishnan	krishnan	PROPN
ejpam-642	569	10	.	.	PUNCT
ejpam-642	570	1	the	the	DET
ejpam-642	570	2	em	em	PROPN
ejpam-642	570	3	algorithm	algorithm	NOUN
ejpam-642	570	4	and	and	CCONJ
ejpam-642	570	5	extensions	extension	NOUN
ejpam-642	570	6	.	.	PUNCT
ejpam-642	571	1	wiley	wiley	PROPN
ejpam-642	571	2	,	,	PUNCT
ejpam-642	571	3	new	new	PROPN
ejpam-642	571	4	york	york	PROPN
ejpam-642	571	5	,	,	PUNCT
ejpam-642	571	6	1997	1997	NUM
ejpam-642	571	7	.	.	PUNCT
ejpam-642	572	1	[	[	X
ejpam-642	572	2	12	12	NUM
ejpam-642	572	3	]	]	PUNCT
ejpam-642	572	4	h.	h.	PROPN
ejpam-642	572	5	b.	b.	PROPN
ejpam-642	572	6	newcombe	newcombe	PROPN
ejpam-642	572	7	.	.	PUNCT
ejpam-642	573	1	handbook	handbook	NOUN
ejpam-642	573	2	of	of	ADP
ejpam-642	573	3	record	record	NOUN
ejpam-642	573	4	linkage	linkage	NOUN
ejpam-642	573	5	:	:	PUNCT
ejpam-642	573	6	methods	method	NOUN
ejpam-642	573	7	for	for	ADP
ejpam-642	573	8	health	health	NOUN
ejpam-642	573	9	and	and	CCONJ
ejpam-642	573	10	statistical	statistical	ADJ
ejpam-642	573	11	studies	study	NOUN
ejpam-642	573	12	,	,	PUNCT
ejpam-642	573	13	administration	administration	NOUN
ejpam-642	573	14	,	,	PUNCT
ejpam-642	573	15	and	and	CCONJ
ejpam-642	573	16	business	business	NOUN
ejpam-642	573	17	.	.	PUNCT
ejpam-642	574	1	oxford	oxford	PROPN
ejpam-642	574	2	university	university	PROPN
ejpam-642	574	3	press	press	PROPN
ejpam-642	574	4	,	,	PUNCT
ejpam-642	574	5	inc	inc	PROPN
ejpam-642	574	6	.	.	PROPN
ejpam-642	574	7	,	,	PUNCT
ejpam-642	574	8	new	new	PROPN
ejpam-642	574	9	york	york	PROPN
ejpam-642	574	10	,	,	PUNCT
ejpam-642	574	11	1988	1988	NUM
ejpam-642	574	12	.	.	PUNCT
ejpam-642	575	1	[	[	X
ejpam-642	575	2	13	13	NUM
ejpam-642	575	3	]	]	X
ejpam-642	575	4	h.	h.	PROPN
ejpam-642	575	5	b.	b.	PROPN
ejpam-642	575	6	newcombe	newcombe	PROPN
ejpam-642	575	7	,	,	PUNCT
ejpam-642	575	8	j.	j.	PROPN
ejpam-642	575	9	m.	m.	PROPN
ejpam-642	575	10	kennedy	kennedy	PROPN
ejpam-642	575	11	,	,	PUNCT
ejpam-642	575	12	s.	s.	PROPN
ejpam-642	575	13	j.	j.	PROPN
ejpam-642	575	14	axford	axford	PROPN
ejpam-642	575	15	,	,	PUNCT
ejpam-642	575	16	and	and	CCONJ
ejpam-642	575	17	a.	a.	PROPN
ejpam-642	575	18	p.	p.	PROPN
ejpam-642	575	19	james	james	PROPN
ejpam-642	575	20	.	.	PUNCT
ejpam-642	576	1	automatic	automatic	ADJ
ejpam-642	576	2	linkage	linkage	NOUN
ejpam-642	576	3	of	of	ADP
ejpam-642	576	4	vital	vital	ADJ
ejpam-642	576	5	records	record	NOUN
ejpam-642	576	6	.	.	PUNCT
ejpam-642	577	1	science	science	NOUN
ejpam-642	577	2	,	,	PUNCT
ejpam-642	577	3	130:954–959	130:954–959	NUM
ejpam-642	577	4	,	,	PUNCT
ejpam-642	577	5	1959	1959	NUM
ejpam-642	577	6	.	.	PUNCT
ejpam-642	578	1	[	[	X
ejpam-642	578	2	14	14	NUM
ejpam-642	578	3	]	]	X
ejpam-642	578	4	y.	y.	PROPN
ejpam-642	578	5	thibaudeau	thibaudeau	PROPN
ejpam-642	578	6	.	.	PUNCT
ejpam-642	579	1	the	the	DET
ejpam-642	579	2	discrimination	discrimination	NOUN
ejpam-642	579	3	power	power	NOUN
ejpam-642	579	4	of	of	ADP
ejpam-642	579	5	dependency	dependency	NOUN
ejpam-642	579	6	structures	structure	NOUN
ejpam-642	579	7	in	in	ADP
ejpam-642	579	8	record	record	NOUN
ejpam-642	579	9	linkage	linkage	NOUN
ejpam-642	579	10	.	.	PUNCT
ejpam-642	580	1	survey	survey	NOUN
ejpam-642	580	2	methodology	methodology	NOUN
ejpam-642	580	3	,	,	PUNCT
ejpam-642	580	4	19:31–38	19:31–38	NUM
ejpam-642	580	5	,	,	PUNCT
ejpam-642	580	6	1993	1993	NUM
ejpam-642	580	7	.	.	PUNCT
ejpam-642	581	1	[	[	X
ejpam-642	581	2	15	15	NUM
ejpam-642	581	3	]	]	X
ejpam-642	581	4	w.	w.	PROPN
ejpam-642	581	5	e.	e.	PROPN
ejpam-642	581	6	winkler	winkler	PROPN
ejpam-642	581	7	.	.	PUNCT
ejpam-642	582	1	using	use	VERB
ejpam-642	582	2	the	the	DET
ejpam-642	582	3	em	em	PROPN
ejpam-642	582	4	algorithm	algorithm	NOUN
ejpam-642	582	5	for	for	ADP
ejpam-642	582	6	weight	weight	NOUN
ejpam-642	582	7	computation	computation	NOUN
ejpam-642	582	8	in	in	ADP
ejpam-642	582	9	the	the	DET
ejpam-642	582	10	fellegi	fellegi	ADJ
ejpam-642	582	11	-	-	PUNCT
ejpam-642	582	12	sunter	sunter	NOUN
ejpam-642	582	13	model	model	NOUN
ejpam-642	582	14	of	of	ADP
ejpam-642	582	15	record	record	NOUN
ejpam-642	582	16	linkage	linkage	NOUN
ejpam-642	582	17	.	.	PUNCT
ejpam-642	583	1	in	in	ADP
ejpam-642	583	2	proceedings	proceeding	NOUN
ejpam-642	583	3	of	of	ADP
ejpam-642	583	4	the	the	DET
ejpam-642	583	5	section	section	NOUN
ejpam-642	583	6	on	on	ADP
ejpam-642	583	7	survey	survey	NOUN
ejpam-642	583	8	research	research	NOUN
ejpam-642	583	9	methods	method	NOUN
ejpam-642	583	10	.	.	PUNCT
ejpam-642	583	11	,	,	PUNCT
ejpam-642	583	12	pages	page	NOUN
ejpam-642	583	13	667–671	667–671	NUM
ejpam-642	583	14	.	.	PUNCT
ejpam-642	584	1	american	american	ADJ
ejpam-642	584	2	statistical	statistical	PROPN
ejpam-642	584	3	association	association	PROPN
ejpam-642	584	4	,	,	PUNCT
ejpam-642	584	5	1988	1988	NUM
ejpam-642	584	6	.	.	PUNCT
ejpam-642	585	1	[	[	X
ejpam-642	585	2	16	16	NUM
ejpam-642	585	3	]	]	X
ejpam-642	585	4	w.	w.	PROPN
ejpam-642	585	5	e.	e.	PROPN
ejpam-642	585	6	winkler	winkler	PROPN
ejpam-642	585	7	.	.	PUNCT
ejpam-642	586	1	method	method	NOUN
ejpam-642	586	2	for	for	ADP
ejpam-642	586	3	adjusting	adjust	VERB
ejpam-642	586	4	for	for	ADP
ejpam-642	586	5	lack	lack	NOUN
ejpam-642	586	6	of	of	ADP
ejpam-642	586	7	independence	independence	NOUN
ejpam-642	586	8	in	in	ADP
ejpam-642	586	9	an	an	DET
ejpam-642	586	10	application	application	NOUN
ejpam-642	586	11	of	of	ADP
ejpam-642	586	12	the	the	DET
ejpam-642	586	13	fellegi	fellegi	ADJ
ejpam-642	586	14	-	-	PUNCT
ejpam-642	586	15	sunter	sunter	NOUN
ejpam-642	586	16	model	model	NOUN
ejpam-642	586	17	of	of	ADP
ejpam-642	586	18	record	record	NOUN
ejpam-642	586	19	linkage	linkage	NOUN
ejpam-642	586	20	.	.	PUNCT
ejpam-642	587	1	survey	survey	NOUN
ejpam-642	587	2	methodology	methodology	NOUN
ejpam-642	587	3	,	,	PUNCT
ejpam-642	587	4	15:101–107	15:101–107	NUM
ejpam-642	587	5	,	,	PUNCT
ejpam-642	587	6	1989	1989	NUM
ejpam-642	587	7	.	.	PUNCT
ejpam-642	588	1	[	[	X
ejpam-642	588	2	17	17	NUM
ejpam-642	588	3	]	]	X
ejpam-642	588	4	w.	w.	PROPN
ejpam-642	588	5	e.	e.	PROPN
ejpam-642	588	6	winkler	winkler	PROPN
ejpam-642	588	7	.	.	PUNCT
ejpam-642	589	1	overview	overview	NOUN
ejpam-642	589	2	of	of	ADP
ejpam-642	589	3	record	record	NOUN
ejpam-642	589	4	linkage	linkage	NOUN
ejpam-642	589	5	and	and	CCONJ
ejpam-642	589	6	current	current	ADJ
ejpam-642	589	7	research	research	NOUN
ejpam-642	589	8	directions	direction	NOUN
ejpam-642	589	9	.	.	PUNCT
ejpam-642	590	1	research	research	NOUN
ejpam-642	590	2	report	report	NOUN
ejpam-642	590	3	series	series	NOUN
ejpam-642	590	4	statistics	statistic	VERB
ejpam-642	590	5	#	#	SYM
ejpam-642	590	6	2006	2006	NUM
ejpam-642	590	7	-	-	SYM
ejpam-642	590	8	2	2	NUM
ejpam-642	590	9	,	,	PUNCT
ejpam-642	590	10	u.s	u.s	PROPN
ejpam-642	590	11	.	.	PROPN
ejpam-642	590	12	bureau	bureau	PROPN
ejpam-642	590	13	of	of	ADP
ejpam-642	590	14	the	the	DET
ejpam-642	590	15	census	census	NOUN
ejpam-642	590	16	,	,	PUNCT
ejpam-642	590	17	2005	2005	NUM
ejpam-642	590	18	.	.	PUNCT