id	sid	tid	token	lemma	pos
ejpam-1184	1	1	9_1184_qian.dvi	9_1184_qian.dvi	PROPN
ejpam-1184	1	2	european	european	PROPN
ejpam-1184	1	3	journal	journal	PROPN
ejpam-1184	1	4	of	of	ADP
ejpam-1184	1	5	pure	pure	ADJ
ejpam-1184	1	6	and	and	CCONJ
ejpam-1184	1	7	applied	apply	VERB
ejpam-1184	1	8	mathematics	mathematic	NOUN
ejpam-1184	1	9	vol	vol	NOUN
ejpam-1184	1	10	.	.	PROPN
ejpam-1184	2	1	4	4	NUM
ejpam-1184	2	2	,	,	PUNCT
ejpam-1184	2	3	no	no	INTJ
ejpam-1184	2	4	.	.	NOUN
ejpam-1184	2	5	4	4	NUM
ejpam-1184	2	6	,	,	PUNCT
ejpam-1184	2	7	2011	2011	NUM
ejpam-1184	2	8	,	,	PUNCT
ejpam-1184	2	9	455	455	NUM
ejpam-1184	2	10	-	-	SYM
ejpam-1184	2	11	466	466	NUM
ejpam-1184	2	12	issn	issn	PROPN
ejpam-1184	2	13	1307	1307	NUM
ejpam-1184	2	14	-	-	SYM
ejpam-1184	2	15	5543	5543	NUM
ejpam-1184	2	16	–	–	PUNCT
ejpam-1184	2	17	www.ejpam.com	www.ejpam.com	X
ejpam-1184	2	18	estimation	estimation	NOUN
ejpam-1184	2	19	and	and	CCONJ
ejpam-1184	2	20	selection	selection	NOUN
ejpam-1184	2	21	in	in	ADP
ejpam-1184	2	22	regression	regression	NOUN
ejpam-1184	2	23	clustering	cluster	VERB
ejpam-1184	2	24	guoqi	guoqi	PROPN
ejpam-1184	2	25	qian1,∗	qian1,∗	NOUN
ejpam-1184	2	26	,	,	PUNCT
ejpam-1184	2	27	yuehua	yuehua	PROPN
ejpam-1184	2	28	wu2	wu2	NOUN
ejpam-1184	2	29	1	1	NUM
ejpam-1184	2	30	department	department	NOUN
ejpam-1184	2	31	of	of	ADP
ejpam-1184	2	32	mathematics	mathematic	NOUN
ejpam-1184	2	33	and	and	CCONJ
ejpam-1184	2	34	statistics	statistic	NOUN
ejpam-1184	2	35	,	,	PUNCT
ejpam-1184	2	36	university	university	NOUN
ejpam-1184	2	37	of	of	ADP
ejpam-1184	2	38	melbourne	melbourne	PROPN
ejpam-1184	2	39	,	,	PUNCT
ejpam-1184	2	40	melbourne	melbourne	PROPN
ejpam-1184	2	41	,	,	PUNCT
ejpam-1184	2	42	australia	australia	PROPN
ejpam-1184	2	43	2	2	NUM
ejpam-1184	2	44	department	department	NOUN
ejpam-1184	2	45	of	of	ADP
ejpam-1184	2	46	mathematics	mathematic	NOUN
ejpam-1184	2	47	and	and	CCONJ
ejpam-1184	2	48	statistics	statistic	NOUN
ejpam-1184	2	49	,	,	PUNCT
ejpam-1184	2	50	york	york	PROPN
ejpam-1184	2	51	university	university	PROPN
ejpam-1184	2	52	,	,	PUNCT
ejpam-1184	2	53	toronto	toronto	PROPN
ejpam-1184	2	54	,	,	PUNCT
ejpam-1184	2	55	canada	canada	PROPN
ejpam-1184	2	56	abstract	abstract	NOUN
ejpam-1184	2	57	.	.	PUNCT
ejpam-1184	3	1	regression	regression	NOUN
ejpam-1184	3	2	clustering	cluster	VERB
ejpam-1184	3	3	is	be	AUX
ejpam-1184	3	4	an	an	DET
ejpam-1184	3	5	important	important	ADJ
ejpam-1184	3	6	model	model	NOUN
ejpam-1184	3	7	-	-	PUNCT
ejpam-1184	3	8	based	base	VERB
ejpam-1184	3	9	clustering	clustering	ADJ
ejpam-1184	3	10	tool	tool	NOUN
ejpam-1184	3	11	having	have	VERB
ejpam-1184	3	12	applications	application	NOUN
ejpam-1184	3	13	in	in	ADP
ejpam-1184	3	14	a	a	DET
ejpam-1184	3	15	variety	variety	NOUN
ejpam-1184	3	16	of	of	ADP
ejpam-1184	3	17	disciplines	discipline	NOUN
ejpam-1184	3	18	.	.	PUNCT
ejpam-1184	4	1	it	it	PRON
ejpam-1184	4	2	discovers	discover	VERB
ejpam-1184	4	3	and	and	CCONJ
ejpam-1184	4	4	reconstructs	reconstruct	VERB
ejpam-1184	4	5	the	the	DET
ejpam-1184	4	6	hidden	hide	VERB
ejpam-1184	4	7	structure	structure	NOUN
ejpam-1184	4	8	for	for	ADP
ejpam-1184	4	9	a	a	DET
ejpam-1184	4	10	data	datum	NOUN
ejpam-1184	4	11	set	set	VERB
ejpam-1184	4	12	which	which	PRON
ejpam-1184	4	13	is	be	AUX
ejpam-1184	4	14	a	a	DET
ejpam-1184	4	15	random	random	ADJ
ejpam-1184	4	16	sample	sample	NOUN
ejpam-1184	4	17	from	from	ADP
ejpam-1184	4	18	a	a	DET
ejpam-1184	4	19	population	population	NOUN
ejpam-1184	4	20	comprising	comprise	VERB
ejpam-1184	4	21	a	a	DET
ejpam-1184	4	22	fixed	fix	VERB
ejpam-1184	4	23	,	,	PUNCT
ejpam-1184	4	24	but	but	CCONJ
ejpam-1184	4	25	unknown	unknown	ADJ
ejpam-1184	4	26	,	,	PUNCT
ejpam-1184	4	27	number	number	NOUN
ejpam-1184	4	28	of	of	ADP
ejpam-1184	4	29	sub	sub	NOUN
ejpam-1184	4	30	-	-	NOUN
ejpam-1184	4	31	populations	population	NOUN
ejpam-1184	4	32	,	,	PUNCT
ejpam-1184	4	33	each	each	PRON
ejpam-1184	4	34	of	of	ADP
ejpam-1184	4	35	which	which	PRON
ejpam-1184	4	36	is	be	AUX
ejpam-1184	4	37	characterized	characterize	VERB
ejpam-1184	4	38	by	by	ADP
ejpam-1184	4	39	a	a	DET
ejpam-1184	4	40	class	class	NOUN
ejpam-1184	4	41	-	-	PUNCT
ejpam-1184	4	42	specific	specific	ADJ
ejpam-1184	4	43	regression	regression	NOUN
ejpam-1184	4	44	hyperplane	hyperplane	NOUN
ejpam-1184	4	45	.	.	PUNCT
ejpam-1184	5	1	an	an	DET
ejpam-1184	5	2	essential	essential	ADJ
ejpam-1184	5	3	objective	objective	NOUN
ejpam-1184	5	4	,	,	PUNCT
ejpam-1184	5	5	as	as	ADV
ejpam-1184	5	6	well	well	ADV
ejpam-1184	5	7	as	as	ADP
ejpam-1184	5	8	a	a	DET
ejpam-1184	5	9	preliminary	preliminary	ADJ
ejpam-1184	5	10	step	step	NOUN
ejpam-1184	5	11	,	,	PUNCT
ejpam-1184	5	12	in	in	ADP
ejpam-1184	5	13	most	most	ADJ
ejpam-1184	5	14	clustering	clustering	ADJ
ejpam-1184	5	15	techniques	technique	NOUN
ejpam-1184	5	16	including	include	VERB
ejpam-1184	5	17	regression	regression	NOUN
ejpam-1184	5	18	clustering	clustering	NOUN
ejpam-1184	5	19	,	,	PUNCT
ejpam-1184	5	20	is	be	AUX
ejpam-1184	5	21	to	to	PART
ejpam-1184	5	22	determine	determine	VERB
ejpam-1184	5	23	the	the	DET
ejpam-1184	5	24	underlying	underlie	VERB
ejpam-1184	5	25	number	number	NOUN
ejpam-1184	5	26	of	of	ADP
ejpam-1184	5	27	clusters	cluster	NOUN
ejpam-1184	5	28	in	in	ADP
ejpam-1184	5	29	the	the	DET
ejpam-1184	5	30	data	datum	NOUN
ejpam-1184	5	31	.	.	PUNCT
ejpam-1184	6	1	in	in	ADP
ejpam-1184	6	2	this	this	DET
ejpam-1184	6	3	paper	paper	NOUN
ejpam-1184	6	4	,	,	PUNCT
ejpam-1184	6	5	we	we	PRON
ejpam-1184	6	6	briefly	briefly	ADV
ejpam-1184	6	7	review	review	VERB
ejpam-1184	6	8	regression	regression	NOUN
ejpam-1184	6	9	clustering	cluster	VERB
ejpam-1184	6	10	methods	method	NOUN
ejpam-1184	6	11	and	and	CCONJ
ejpam-1184	6	12	discuss	discuss	VERB
ejpam-1184	6	13	how	how	SCONJ
ejpam-1184	6	14	to	to	PART
ejpam-1184	6	15	determine	determine	VERB
ejpam-1184	6	16	the	the	DET
ejpam-1184	6	17	underlying	underlie	VERB
ejpam-1184	6	18	number	number	NOUN
ejpam-1184	6	19	of	of	ADP
ejpam-1184	6	20	clusters	cluster	NOUN
ejpam-1184	6	21	by	by	ADP
ejpam-1184	6	22	using	use	VERB
ejpam-1184	6	23	model	model	NOUN
ejpam-1184	6	24	selection	selection	NOUN
ejpam-1184	6	25	techniques	technique	NOUN
ejpam-1184	6	26	,	,	PUNCT
ejpam-1184	6	27	in	in	ADP
ejpam-1184	6	28	particular	particular	ADJ
ejpam-1184	6	29	,	,	PUNCT
ejpam-1184	6	30	the	the	DET
ejpam-1184	6	31	information	information	NOUN
ejpam-1184	6	32	-	-	PUNCT
ejpam-1184	6	33	based	base	VERB
ejpam-1184	6	34	technique	technique	NOUN
ejpam-1184	6	35	.	.	PUNCT
ejpam-1184	7	1	a	a	DET
ejpam-1184	7	2	computing	compute	VERB
ejpam-1184	7	3	algorithm	algorithm	NOUN
ejpam-1184	7	4	is	be	AUX
ejpam-1184	7	5	developed	develop	VERB
ejpam-1184	7	6	for	for	ADP
ejpam-1184	7	7	estimating	estimate	VERB
ejpam-1184	7	8	the	the	DET
ejpam-1184	7	9	number	number	NOUN
ejpam-1184	7	10	of	of	ADP
ejpam-1184	7	11	clusters	cluster	NOUN
ejpam-1184	7	12	and	and	CCONJ
ejpam-1184	7	13	other	other	ADJ
ejpam-1184	7	14	parameters	parameter	NOUN
ejpam-1184	7	15	in	in	ADP
ejpam-1184	7	16	regression	regression	NOUN
ejpam-1184	7	17	clustering	cluster	VERB
ejpam-1184	7	18	.	.	PUNCT
ejpam-1184	8	1	simulation	simulation	NOUN
ejpam-1184	8	2	studies	study	NOUN
ejpam-1184	8	3	are	be	AUX
ejpam-1184	8	4	also	also	ADV
ejpam-1184	8	5	provided	provide	VERB
ejpam-1184	8	6	to	to	PART
ejpam-1184	8	7	show	show	VERB
ejpam-1184	8	8	the	the	DET
ejpam-1184	8	9	performance	performance	NOUN
ejpam-1184	8	10	of	of	ADP
ejpam-1184	8	11	the	the	DET
ejpam-1184	8	12	algorithm	algorithm	NOUN
ejpam-1184	8	13	.	.	PUNCT
ejpam-1184	9	1	2000	2000	NUM
ejpam-1184	9	2	mathematics	mathematic	NOUN
ejpam-1184	9	3	subject	subject	NOUN
ejpam-1184	9	4	classifications	classification	NOUN
ejpam-1184	9	5	:	:	PUNCT
ejpam-1184	9	6	62h30	62h30	NUM
ejpam-1184	9	7	,	,	PUNCT
ejpam-1184	9	8	68t10	68t10	NUM
ejpam-1184	9	9	,	,	PUNCT
ejpam-1184	9	10	91c20	91c20	NUM
ejpam-1184	9	11	key	key	ADJ
ejpam-1184	9	12	words	word	NOUN
ejpam-1184	9	13	and	and	CCONJ
ejpam-1184	9	14	phrases	phrase	NOUN
ejpam-1184	9	15	:	:	PUNCT
ejpam-1184	9	16	regression	regression	NOUN
ejpam-1184	9	17	clustering	clustering	NOUN
ejpam-1184	9	18	,	,	PUNCT
ejpam-1184	9	19	least	least	ADJ
ejpam-1184	9	20	squares	square	NOUN
ejpam-1184	9	21	estimation	estimation	NOUN
ejpam-1184	9	22	,	,	PUNCT
ejpam-1184	9	23	model	model	NOUN
ejpam-1184	9	24	selection	selection	NOUN
ejpam-1184	9	25	1	1	NUM
ejpam-1184	9	26	.	.	PUNCT
ejpam-1184	10	1	introduction	introduction	NOUN
ejpam-1184	10	2	cluster	cluster	NOUN
ejpam-1184	10	3	analysis	analysis	NOUN
ejpam-1184	10	4	is	be	AUX
ejpam-1184	10	5	an	an	DET
ejpam-1184	10	6	important	important	ADJ
ejpam-1184	10	7	scientific	scientific	ADJ
ejpam-1184	10	8	tool	tool	NOUN
ejpam-1184	10	9	for	for	ADP
ejpam-1184	10	10	examining	examine	VERB
ejpam-1184	10	11	multivariate	multivariate	NOUN
ejpam-1184	10	12	data	datum	NOUN
ejpam-1184	10	13	with	with	ADP
ejpam-1184	10	14	a	a	DET
ejpam-1184	10	15	view	view	NOUN
ejpam-1184	10	16	to	to	ADP
ejpam-1184	10	17	uncovering	uncover	VERB
ejpam-1184	10	18	or	or	CCONJ
ejpam-1184	10	19	discovering	discover	VERB
ejpam-1184	10	20	clusters	cluster	NOUN
ejpam-1184	10	21	or	or	CCONJ
ejpam-1184	10	22	groups	group	NOUN
ejpam-1184	10	23	of	of	ADP
ejpam-1184	10	24	homogeneous	homogeneous	ADJ
ejpam-1184	10	25	observations	observation	NOUN
ejpam-1184	10	26	.	.	PUNCT
ejpam-1184	11	1	it	it	PRON
ejpam-1184	11	2	finds	find	VERB
ejpam-1184	11	3	clusters	cluster	NOUN
ejpam-1184	11	4	in	in	ADP
ejpam-1184	11	5	the	the	DET
ejpam-1184	11	6	data	datum	NOUN
ejpam-1184	11	7	such	such	ADJ
ejpam-1184	11	8	that	that	SCONJ
ejpam-1184	11	9	observations	observation	NOUN
ejpam-1184	11	10	are	be	AUX
ejpam-1184	11	11	as	as	ADP
ejpam-1184	11	12	“	"	PUNCT
ejpam-1184	11	13	similar	similar	ADJ
ejpam-1184	11	14	”	"	PUNCT
ejpam-1184	11	15	as	as	ADP
ejpam-1184	11	16	possible	possible	ADJ
ejpam-1184	11	17	within	within	ADP
ejpam-1184	11	18	clusters	cluster	NOUN
ejpam-1184	11	19	(	(	PUNCT
ejpam-1184	11	20	internal	internal	ADJ
ejpam-1184	11	21	cohesion	cohesion	NOUN
ejpam-1184	11	22	or	or	CCONJ
ejpam-1184	11	23	homogeneity	homogeneity	NOUN
ejpam-1184	11	24	)	)	PUNCT
ejpam-1184	11	25	,	,	PUNCT
ejpam-1184	11	26	and	and	CCONJ
ejpam-1184	11	27	as	as	ADP
ejpam-1184	11	28	“	"	PUNCT
ejpam-1184	11	29	dissimilar	dissimilar	ADJ
ejpam-1184	11	30	”	"	PUNCT
ejpam-1184	11	31	as	as	SCONJ
ejpam-1184	11	32	it	it	PRON
ejpam-1184	11	33	could	could	AUX
ejpam-1184	11	34	be	be	AUX
ejpam-1184	11	35	between	between	ADP
ejpam-1184	11	36	clusters	cluster	NOUN
ejpam-1184	11	37	(	(	PUNCT
ejpam-1184	11	38	external	external	ADJ
ejpam-1184	11	39	separation	separation	NOUN
ejpam-1184	11	40	or	or	CCONJ
ejpam-1184	11	41	heterogeneity	heterogeneity	NOUN
ejpam-1184	11	42	)	)	PUNCT
ejpam-1184	11	43	.	.	PUNCT
ejpam-1184	12	1	cluster	cluster	NOUN
ejpam-1184	12	2	analysis	analysis	NOUN
ejpam-1184	12	3	should	should	AUX
ejpam-1184	12	4	be	be	AUX
ejpam-1184	12	5	distinguished	distinguish	VERB
ejpam-1184	12	6	from	from	ADP
ejpam-1184	12	7	the	the	DET
ejpam-1184	12	8	related	relate	VERB
ejpam-1184	12	9	problem	problem	NOUN
ejpam-1184	12	10	of	of	ADP
ejpam-1184	12	11	discriminant	discriminant	ADJ
ejpam-1184	12	12	analysis	analysis	NOUN
ejpam-1184	12	13	in	in	SCONJ
ejpam-1184	12	14	that	that	SCONJ
ejpam-1184	12	15	it	it	PRON
ejpam-1184	12	16	actually	actually	ADV
ejpam-1184	12	17	establishes	establish	VERB
ejpam-1184	12	18	the	the	DET
ejpam-1184	12	19	clusters	cluster	NOUN
ejpam-1184	12	20	,	,	PUNCT
ejpam-1184	12	21	whereas	whereas	SCONJ
ejpam-1184	12	22	in	in	ADP
ejpam-1184	12	23	discriminant	discriminant	ADJ
ejpam-1184	12	24	analysis	analysis	NOUN
ejpam-1184	12	25	,	,	PUNCT
ejpam-1184	12	26	known	know	VERB
ejpam-1184	12	27	clusterings	clustering	NOUN
ejpam-1184	12	28	(	(	PUNCT
ejpam-1184	12	29	or	or	CCONJ
ejpam-1184	12	30	groupings	grouping	NOUN
ejpam-1184	12	31	)	)	PUNCT
ejpam-1184	12	32	of	of	ADP
ejpam-1184	12	33	some	some	DET
ejpam-1184	12	34	observations	observation	NOUN
ejpam-1184	12	35	are	be	AUX
ejpam-1184	12	36	used	use	VERB
ejpam-1184	12	37	to	to	PART
ejpam-1184	12	38	categorize	categorize	VERB
ejpam-1184	12	39	others	other	NOUN
ejpam-1184	12	40	and	and	CCONJ
ejpam-1184	12	41	infer	infer	VERB
ejpam-1184	12	42	the	the	DET
ejpam-1184	12	43	structure	structure	NOUN
ejpam-1184	12	44	of	of	ADP
ejpam-1184	12	45	the	the	DET
ejpam-1184	12	46	data	datum	NOUN
ejpam-1184	12	47	as	as	ADP
ejpam-1184	12	48	a	a	DET
ejpam-1184	12	49	whole	whole	NOUN
ejpam-1184	12	50	.	.	PUNCT
ejpam-1184	13	1	clustering	cluster	VERB
ejpam-1184	13	2	techniques	technique	NOUN
ejpam-1184	13	3	range	range	VERB
ejpam-1184	13	4	from	from	ADP
ejpam-1184	13	5	those	those	PRON
ejpam-1184	13	6	that	that	PRON
ejpam-1184	13	7	are	be	AUX
ejpam-1184	13	8	largely	largely	ADV
ejpam-1184	13	9	heuristic	heuristic	ADJ
ejpam-1184	13	10	and	and	CCONJ
ejpam-1184	13	11	descriptive	descriptive	ADJ
ejpam-1184	13	12	to	to	ADP
ejpam-1184	13	13	more	more	ADV
ejpam-1184	13	14	formal	formal	ADJ
ejpam-1184	13	15	procedures	procedure	NOUN
ejpam-1184	13	16	based	base	VERB
ejpam-1184	13	17	on	on	ADP
ejpam-1184	13	18	statistical	statistical	ADJ
ejpam-1184	13	19	models	model	NOUN
ejpam-1184	13	20	.	.	PUNCT
ejpam-1184	14	1	in	in	ADP
ejpam-1184	14	2	general	general	ADJ
ejpam-1184	14	3	,	,	PUNCT
ejpam-1184	14	4	they	they	PRON
ejpam-1184	14	5	follow	follow	VERB
ejpam-1184	14	6	either	either	CCONJ
ejpam-1184	14	7	a	a	DET
ejpam-1184	14	8	hierarchical	hierarchical	ADJ
ejpam-1184	14	9	strategy	strategy	NOUN
ejpam-1184	14	10	or	or	CCONJ
ejpam-1184	14	11	partitioning	partition	VERB
ejpam-1184	14	12	type	type	NOUN
ejpam-1184	14	13	of	of	ADP
ejpam-1184	14	14	methods	method	NOUN
ejpam-1184	14	15	.	.	PUNCT
ejpam-1184	15	1	hierarchical	hierarchical	ADJ
ejpam-1184	15	2	methods	method	NOUN
ejpam-1184	15	3	proceed	proceed	VERB
ejpam-1184	15	4	by	by	ADP
ejpam-1184	15	5	stages	stage	NOUN
ejpam-1184	15	6	producing	produce	VERB
ejpam-1184	15	7	a	a	DET
ejpam-1184	15	8	series	series	NOUN
ejpam-1184	15	9	of	of	ADP
ejpam-1184	15	10	partitions	partition	NOUN
ejpam-1184	15	11	,	,	PUNCT
ejpam-1184	15	12	which	which	PRON
ejpam-1184	15	13	may	may	AUX
ejpam-1184	15	14	run	run	VERB
ejpam-1184	15	15	from	from	ADP
ejpam-1184	15	16	a	a	DET
ejpam-1184	15	17	single	single	ADJ
ejpam-1184	15	18	cluster	cluster	NOUN
ejpam-1184	15	19	containing	contain	VERB
ejpam-1184	15	20	all	all	DET
ejpam-1184	15	21	objects	object	NOUN
ejpam-1184	15	22	to	to	ADP
ejpam-1184	15	23	as	as	ADV
ejpam-1184	15	24	many	many	ADJ
ejpam-1184	15	25	∗corresponding	∗corresponde	VERB
ejpam-1184	15	26	author	author	NOUN
ejpam-1184	15	27	.	.	PUNCT
ejpam-1184	16	1	email	email	NOUN
ejpam-1184	16	2	addresses	address	NOUN
ejpam-1184	16	3	:	:	PUNCT
ejpam-1184	16	4	g.qian�ms.unimelb.edu.au	g.qian�ms.unimelb.edu.au	PROPN
ejpam-1184	16	5	(	(	PUNCT
ejpam-1184	16	6	g.	g.	PROPN
ejpam-1184	16	7	qian	qian	PROPN
ejpam-1184	16	8	)	)	PUNCT
ejpam-1184	16	9	,	,	PUNCT
ejpam-1184	16	10	wuyh�mathstat.yorku	wuyh�mathstat.yorku	PROPN
ejpam-1184	16	11	.	.	PUNCT
ejpam-1184	17	1	a	a	DET
ejpam-1184	17	2	(	(	PUNCT
ejpam-1184	17	3	y.	y.	PROPN
ejpam-1184	17	4	wu	wu	PROPN
ejpam-1184	17	5	)	)	PUNCT
ejpam-1184	18	1	http://www.ejpam.com	http://www.ejpam.com	X
ejpam-1184	19	1	455	455	NUM
ejpam-1184	20	1	c	c	X
ejpam-1184	20	2	©	©	PROPN
ejpam-1184	20	3	2011	2011	NUM
ejpam-1184	20	4	ejpam	ejpam	VERB
ejpam-1184	20	5	all	all	DET
ejpam-1184	20	6	rights	right	NOUN
ejpam-1184	20	7	reserved	reserve	VERB
ejpam-1184	20	8	.	.	PUNCT
ejpam-1184	21	1	g.	g.	PROPN
ejpam-1184	21	2	qian	qian	PROPN
ejpam-1184	21	3	,	,	PUNCT
ejpam-1184	21	4	y.	y.	PROPN
ejpam-1184	21	5	wu	wu	PROPN
ejpam-1184	21	6	/	/	SYM
ejpam-1184	21	7	eur	eur	PROPN
ejpam-1184	21	8	.	.	PUNCT
ejpam-1184	22	1	j.	j.	PROPN
ejpam-1184	22	2	pure	pure	PROPN
ejpam-1184	22	3	appl	appl	PROPN
ejpam-1184	22	4	.	.	PROPN
ejpam-1184	22	5	math	math	PROPN
ejpam-1184	22	6	,	,	PUNCT
ejpam-1184	22	7	4	4	NUM
ejpam-1184	22	8	(	(	PUNCT
ejpam-1184	22	9	2011	2011	NUM
ejpam-1184	22	10	)	)	PUNCT
ejpam-1184	22	11	,	,	PUNCT
ejpam-1184	22	12	455	455	NUM
ejpam-1184	22	13	-	-	SYM
ejpam-1184	22	14	466	466	NUM
ejpam-1184	22	15	456	456	NUM
ejpam-1184	22	16	clusters	cluster	NOUN
ejpam-1184	22	17	as	as	ADP
ejpam-1184	22	18	the	the	DET
ejpam-1184	22	19	total	total	ADJ
ejpam-1184	22	20	number	number	NOUN
ejpam-1184	22	21	of	of	ADP
ejpam-1184	22	22	objects	object	NOUN
ejpam-1184	22	23	,	,	PUNCT
ejpam-1184	22	24	with	with	ADP
ejpam-1184	22	25	each	each	PRON
ejpam-1184	22	26	containing	contain	VERB
ejpam-1184	22	27	a	a	DET
ejpam-1184	22	28	single	single	ADJ
ejpam-1184	22	29	object	object	NOUN
ejpam-1184	22	30	.	.	PUNCT
ejpam-1184	23	1	they	they	PRON
ejpam-1184	23	2	can	can	AUX
ejpam-1184	23	3	be	be	AUX
ejpam-1184	23	4	either	either	CCONJ
ejpam-1184	23	5	“	"	PUNCT
ejpam-1184	23	6	agglomerative	agglomerative	ADJ
ejpam-1184	23	7	”	"	PUNCT
ejpam-1184	23	8	,	,	PUNCT
ejpam-1184	23	9	meaning	mean	VERB
ejpam-1184	23	10	that	that	SCONJ
ejpam-1184	23	11	groups	group	NOUN
ejpam-1184	23	12	are	be	AUX
ejpam-1184	23	13	merged	merge	VERB
ejpam-1184	23	14	,	,	PUNCT
ejpam-1184	23	15	or	or	CCONJ
ejpam-1184	23	16	“	"	PUNCT
ejpam-1184	23	17	divisive	divisive	ADJ
ejpam-1184	23	18	”	"	PUNCT
ejpam-1184	23	19	,	,	PUNCT
ejpam-1184	23	20	in	in	ADP
ejpam-1184	23	21	which	which	PRON
ejpam-1184	23	22	one	one	NUM
ejpam-1184	23	23	or	or	CCONJ
ejpam-1184	23	24	more	more	ADJ
ejpam-1184	23	25	groups	group	NOUN
ejpam-1184	23	26	are	be	AUX
ejpam-1184	23	27	split	split	VERB
ejpam-1184	23	28	at	at	ADP
ejpam-1184	23	29	each	each	DET
ejpam-1184	23	30	stage	stage	NOUN
ejpam-1184	23	31	.	.	PUNCT
ejpam-1184	24	1	at	at	ADP
ejpam-1184	24	2	each	each	DET
ejpam-1184	24	3	stage	stage	NOUN
ejpam-1184	24	4	of	of	ADP
ejpam-1184	24	5	hierarchical	hierarchical	ADJ
ejpam-1184	24	6	clustering	clustering	NOUN
ejpam-1184	24	7	,	,	PUNCT
ejpam-1184	24	8	the	the	DET
ejpam-1184	24	9	splitting	splitting	NOUN
ejpam-1184	24	10	or	or	CCONJ
ejpam-1184	24	11	merging	merging	NOUN
ejpam-1184	24	12	is	be	AUX
ejpam-1184	24	13	chosen	choose	VERB
ejpam-1184	24	14	so	so	SCONJ
ejpam-1184	24	15	as	as	SCONJ
ejpam-1184	24	16	to	to	PART
ejpam-1184	24	17	optimize	optimize	VERB
ejpam-1184	24	18	some	some	DET
ejpam-1184	24	19	criterion	criterion	NOUN
ejpam-1184	24	20	.	.	PUNCT
ejpam-1184	25	1	conventional	conventional	ADJ
ejpam-1184	25	2	agglomerative	agglomerative	ADJ
ejpam-1184	25	3	hierarchical	hierarchical	ADJ
ejpam-1184	25	4	methods	method	NOUN
ejpam-1184	25	5	use	use	VERB
ejpam-1184	25	6	heuristic	heuristic	ADJ
ejpam-1184	25	7	criteria	criterion	NOUN
ejpam-1184	25	8	,	,	PUNCT
ejpam-1184	25	9	such	such	ADJ
ejpam-1184	25	10	as	as	ADP
ejpam-1184	25	11	single	single	ADJ
ejpam-1184	25	12	linkage	linkage	NOUN
ejpam-1184	25	13	(	(	PUNCT
ejpam-1184	25	14	nearest	near	ADJ
ejpam-1184	25	15	neighbor	neighbor	NOUN
ejpam-1184	25	16	)	)	PUNCT
ejpam-1184	25	17	,	,	PUNCT
ejpam-1184	25	18	complete	complete	ADJ
ejpam-1184	25	19	linkage	linkage	NOUN
ejpam-1184	25	20	(	(	PUNCT
ejpam-1184	25	21	furthest	furth	ADJ
ejpam-1184	25	22	neighbor	neighbor	NOUN
ejpam-1184	25	23	)	)	PUNCT
ejpam-1184	25	24	,	,	PUNCT
ejpam-1184	25	25	centroid	centroid	NOUN
ejpam-1184	25	26	clustering	clustering	NOUN
ejpam-1184	25	27	,	,	PUNCT
ejpam-1184	25	28	or	or	CCONJ
ejpam-1184	25	29	sum	sum	NOUN
ejpam-1184	25	30	of	of	ADP
ejpam-1184	25	31	squares	square	NOUN
ejpam-1184	25	32	etc	etc	X
ejpam-1184	25	33	.	.	PUNCT
ejpam-1184	26	1	[	[	X
ejpam-1184	26	2	11	11	NUM
ejpam-1184	26	3	]	]	PUNCT
ejpam-1184	26	4	.	.	PUNCT
ejpam-1184	27	1	in	in	ADP
ejpam-1184	27	2	applications	application	NOUN
ejpam-1184	27	3	,	,	PUNCT
ejpam-1184	27	4	divisive	divisive	ADJ
ejpam-1184	27	5	methods	method	NOUN
ejpam-1184	27	6	are	be	AUX
ejpam-1184	27	7	less	less	ADV
ejpam-1184	27	8	commonly	commonly	ADV
ejpam-1184	27	9	used	use	VERB
ejpam-1184	27	10	than	than	ADP
ejpam-1184	27	11	agglomerative	agglomerative	ADJ
ejpam-1184	27	12	procedures	procedure	NOUN
ejpam-1184	27	13	since	since	SCONJ
ejpam-1184	27	14	they	they	PRON
ejpam-1184	27	15	are	be	AUX
ejpam-1184	27	16	computationally	computationally	ADV
ejpam-1184	27	17	demanding	demanding	ADJ
ejpam-1184	27	18	.	.	PUNCT
ejpam-1184	28	1	yet	yet	CCONJ
ejpam-1184	28	2	a	a	DET
ejpam-1184	28	3	significant	significant	ADJ
ejpam-1184	28	4	drawback	drawback	NOUN
ejpam-1184	28	5	of	of	ADP
ejpam-1184	28	6	hierarchical	hierarchical	ADJ
ejpam-1184	28	7	clustering	clustering	ADJ
ejpam-1184	28	8	methods	method	NOUN
ejpam-1184	28	9	is	be	AUX
ejpam-1184	28	10	that	that	SCONJ
ejpam-1184	28	11	the	the	DET
ejpam-1184	28	12	divisions	division	NOUN
ejpam-1184	28	13	or	or	CCONJ
ejpam-1184	28	14	fusions	fusion	NOUN
ejpam-1184	28	15	,	,	PUNCT
ejpam-1184	28	16	once	once	ADV
ejpam-1184	28	17	made	make	VERB
ejpam-1184	28	18	,	,	PUNCT
ejpam-1184	28	19	are	be	AUX
ejpam-1184	28	20	irrevocable	irrevocable	ADJ
ejpam-1184	28	21	.	.	PUNCT
ejpam-1184	29	1	when	when	SCONJ
ejpam-1184	29	2	an	an	DET
ejpam-1184	29	3	agglomerative	agglomerative	ADJ
ejpam-1184	29	4	algorithm	algorithm	NOUN
ejpam-1184	29	5	has	have	AUX
ejpam-1184	29	6	joined	join	VERB
ejpam-1184	29	7	two	two	NUM
ejpam-1184	29	8	objects	object	NOUN
ejpam-1184	29	9	into	into	ADP
ejpam-1184	29	10	a	a	DET
ejpam-1184	29	11	cluster	cluster	NOUN
ejpam-1184	29	12	they	they	PRON
ejpam-1184	29	13	can	can	AUX
ejpam-1184	29	14	not	not	PART
ejpam-1184	29	15	subsequently	subsequently	ADV
ejpam-1184	29	16	be	be	AUX
ejpam-1184	29	17	separated	separate	VERB
ejpam-1184	29	18	,	,	PUNCT
ejpam-1184	29	19	and	and	CCONJ
ejpam-1184	29	20	when	when	SCONJ
ejpam-1184	29	21	a	a	DET
ejpam-1184	29	22	divisive	divisive	ADJ
ejpam-1184	29	23	algorithm	algorithm	NOUN
ejpam-1184	29	24	has	have	AUX
ejpam-1184	29	25	made	make	VERB
ejpam-1184	29	26	a	a	DET
ejpam-1184	29	27	split	split	NOUN
ejpam-1184	29	28	,	,	PUNCT
ejpam-1184	29	29	the	the	DET
ejpam-1184	29	30	objects	object	NOUN
ejpam-1184	29	31	can	can	AUX
ejpam-1184	29	32	not	not	PART
ejpam-1184	29	33	be	be	AUX
ejpam-1184	29	34	recombined	recombine	VERB
ejpam-1184	29	35	.	.	PUNCT
ejpam-1184	30	1	as	as	SCONJ
ejpam-1184	30	2	kaufman	kaufman	PROPN
ejpam-1184	30	3	and	and	CCONJ
ejpam-1184	30	4	rousseeuw	rousseeuw	VERB
ejpam-1184	30	5	[	[	X
ejpam-1184	30	6	11	11	NUM
ejpam-1184	30	7	]	]	PUNCT
ejpam-1184	30	8	comment	comment	NOUN
ejpam-1184	30	9	:	:	PUNCT
ejpam-1184	30	10	“	"	PUNCT
ejpam-1184	30	11	a	a	DET
ejpam-1184	30	12	hierarchical	hierarchical	ADJ
ejpam-1184	30	13	method	method	NOUN
ejpam-1184	30	14	suffers	suffer	VERB
ejpam-1184	30	15	from	from	ADP
ejpam-1184	30	16	the	the	DET
ejpam-1184	30	17	defect	defect	NOUN
ejpam-1184	30	18	that	that	PRON
ejpam-1184	30	19	it	it	PRON
ejpam-1184	30	20	can	can	AUX
ejpam-1184	30	21	never	never	ADV
ejpam-1184	30	22	repair	repair	VERB
ejpam-1184	30	23	what	what	PRON
ejpam-1184	30	24	was	be	AUX
ejpam-1184	30	25	done	do	VERB
ejpam-1184	30	26	in	in	ADP
ejpam-1184	30	27	previous	previous	ADJ
ejpam-1184	30	28	steps	step	NOUN
ejpam-1184	30	29	”	"	PUNCT
ejpam-1184	30	30	.	.	PUNCT
ejpam-1184	31	1	in	in	ADP
ejpam-1184	31	2	contrast	contrast	NOUN
ejpam-1184	31	3	,	,	PUNCT
ejpam-1184	31	4	a	a	DET
ejpam-1184	31	5	partitioning	partitioning	NOUN
ejpam-1184	31	6	method	method	NOUN
ejpam-1184	31	7	constructs	construct	VERB
ejpam-1184	31	8	a	a	DET
ejpam-1184	31	9	fixed	fix	VERB
ejpam-1184	31	10	number	number	NOUN
ejpam-1184	31	11	of	of	ADP
ejpam-1184	31	12	clusters	cluster	NOUN
ejpam-1184	31	13	,	,	PUNCT
ejpam-1184	31	14	say	say	VERB
ejpam-1184	31	15	k.	k.	PROPN
ejpam-1184	32	1	it	it	PRON
ejpam-1184	32	2	classifies	classify	VERB
ejpam-1184	32	3	the	the	DET
ejpam-1184	32	4	data	datum	NOUN
ejpam-1184	32	5	into	into	ADP
ejpam-1184	32	6	k	k	PROPN
ejpam-1184	32	7	clusters	cluster	NOUN
ejpam-1184	32	8	,	,	PUNCT
ejpam-1184	32	9	which	which	PRON
ejpam-1184	32	10	together	together	ADV
ejpam-1184	32	11	satisfy	satisfy	VERB
ejpam-1184	32	12	two	two	NUM
ejpam-1184	32	13	requirements	requirement	NOUN
ejpam-1184	32	14	of	of	ADP
ejpam-1184	32	15	a	a	DET
ejpam-1184	32	16	partition	partition	NOUN
ejpam-1184	32	17	:	:	PUNCT
ejpam-1184	32	18	(	(	PUNCT
ejpam-1184	32	19	i	i	NOUN
ejpam-1184	32	20	)	)	PUNCT
ejpam-1184	32	21	each	each	DET
ejpam-1184	32	22	cluster	cluster	NOUN
ejpam-1184	32	23	must	must	AUX
ejpam-1184	32	24	contain	contain	VERB
ejpam-1184	32	25	at	at	ADV
ejpam-1184	32	26	least	least	ADV
ejpam-1184	32	27	one	one	NUM
ejpam-1184	32	28	object	object	NOUN
ejpam-1184	32	29	;	;	PUNCT
ejpam-1184	32	30	(	(	PUNCT
ejpam-1184	32	31	ii	ii	NOUN
ejpam-1184	32	32	)	)	PUNCT
ejpam-1184	32	33	each	each	DET
ejpam-1184	32	34	object	object	NOUN
ejpam-1184	32	35	must	must	AUX
ejpam-1184	32	36	belong	belong	VERB
ejpam-1184	32	37	to	to	ADP
ejpam-1184	32	38	exactly	exactly	ADV
ejpam-1184	32	39	one	one	NUM
ejpam-1184	32	40	cluster	cluster	NOUN
ejpam-1184	32	41	.	.	PUNCT
ejpam-1184	33	1	usually	usually	ADV
ejpam-1184	33	2	,	,	PUNCT
ejpam-1184	33	3	partitioning	partition	VERB
ejpam-1184	33	4	methods	method	NOUN
ejpam-1184	33	5	move	move	VERB
ejpam-1184	33	6	observations	observation	NOUN
ejpam-1184	33	7	iteratively	iteratively	ADV
ejpam-1184	33	8	from	from	ADP
ejpam-1184	33	9	one	one	NUM
ejpam-1184	33	10	cluster	cluster	NOUN
ejpam-1184	33	11	to	to	ADP
ejpam-1184	33	12	another	another	PRON
ejpam-1184	33	13	,	,	PUNCT
ejpam-1184	33	14	starting	start	VERB
ejpam-1184	33	15	from	from	ADP
ejpam-1184	33	16	an	an	DET
ejpam-1184	33	17	initial	initial	ADJ
ejpam-1184	33	18	partition	partition	NOUN
ejpam-1184	33	19	,	,	PUNCT
ejpam-1184	33	20	to	to	PART
ejpam-1184	33	21	achieve	achieve	VERB
ejpam-1184	33	22	some	some	DET
ejpam-1184	33	23	pre	pre	ADJ
ejpam-1184	33	24	-	-	ADJ
ejpam-1184	33	25	chosen	choose	VERB
ejpam-1184	33	26	optimization	optimization	NOUN
ejpam-1184	33	27	.	.	PUNCT
ejpam-1184	34	1	in	in	ADP
ejpam-1184	34	2	most	most	ADJ
ejpam-1184	34	3	circumstances	circumstance	NOUN
ejpam-1184	34	4	,	,	PUNCT
ejpam-1184	34	5	the	the	DET
ejpam-1184	34	6	number	number	NOUN
ejpam-1184	34	7	of	of	ADP
ejpam-1184	34	8	clusters	cluster	NOUN
ejpam-1184	34	9	has	have	VERB
ejpam-1184	34	10	to	to	PART
ejpam-1184	34	11	be	be	AUX
ejpam-1184	34	12	specified	specify	VERB
ejpam-1184	34	13	in	in	ADP
ejpam-1184	34	14	advance	advance	NOUN
ejpam-1184	34	15	and	and	CCONJ
ejpam-1184	34	16	typically	typically	ADV
ejpam-1184	34	17	does	do	AUX
ejpam-1184	34	18	not	not	PART
ejpam-1184	34	19	change	change	VERB
ejpam-1184	34	20	during	during	ADP
ejpam-1184	34	21	the	the	DET
ejpam-1184	34	22	course	course	NOUN
ejpam-1184	34	23	of	of	ADP
ejpam-1184	34	24	the	the	DET
ejpam-1184	34	25	iteration	iteration	NOUN
ejpam-1184	34	26	.	.	PUNCT
ejpam-1184	35	1	for	for	ADP
ejpam-1184	35	2	instance	instance	NOUN
ejpam-1184	35	3	,	,	PUNCT
ejpam-1184	35	4	the	the	DET
ejpam-1184	35	5	most	most	ADV
ejpam-1184	35	6	commonly	commonly	ADV
ejpam-1184	35	7	used	use	VERB
ejpam-1184	35	8	relocation	relocation	NOUN
ejpam-1184	35	9	methods	method	NOUN
ejpam-1184	35	10	–	–	PUNCT
ejpam-1184	35	11	the	the	DET
ejpam-1184	35	12	k	k	NOUN
ejpam-1184	35	13	-	-	PUNCT
ejpam-1184	35	14	means	means	NOUN
ejpam-1184	35	15	type	type	NOUN
ejpam-1184	35	16	of	of	ADP
ejpam-1184	35	17	methods	method	NOUN
ejpam-1184	35	18	:	:	PUNCT
ejpam-1184	35	19	k	k	X
ejpam-1184	35	20	-	-	PUNCT
ejpam-1184	35	21	means	means	PROPN
ejpam-1184	35	22	,	,	PUNCT
ejpam-1184	35	23	k	k	NOUN
ejpam-1184	35	24	-	-	PUNCT
ejpam-1184	35	25	modes	mode	NOUN
ejpam-1184	35	26	,	,	PUNCT
ejpam-1184	35	27	k	k	NOUN
ejpam-1184	35	28	-	-	PUNCT
ejpam-1184	35	29	medians	median	NOUN
ejpam-1184	35	30	and	and	CCONJ
ejpam-1184	35	31	k	k	NOUN
ejpam-1184	35	32	-	-	NOUN
ejpam-1184	35	33	mediods	mediod	NOUN
ejpam-1184	35	34	[	[	X
ejpam-1184	35	35	9	9	NUM
ejpam-1184	35	36	,	,	PUNCT
ejpam-1184	35	37	10	10	NUM
ejpam-1184	35	38	]	]	PUNCT
ejpam-1184	35	39	–	–	PUNCT
ejpam-1184	35	40	reduce	reduce	VERB
ejpam-1184	35	41	the	the	DET
ejpam-1184	35	42	average	average	ADJ
ejpam-1184	35	43	within	within	ADP
ejpam-1184	35	44	-	-	PUNCT
ejpam-1184	35	45	group	group	NOUN
ejpam-1184	35	46	distance	distance	NOUN
ejpam-1184	35	47	of	of	ADP
ejpam-1184	35	48	objects	object	NOUN
ejpam-1184	35	49	to	to	ADP
ejpam-1184	35	50	their	their	PRON
ejpam-1184	35	51	nearest	near	ADJ
ejpam-1184	35	52	representatives	representative	NOUN
ejpam-1184	35	53	(	(	PUNCT
ejpam-1184	35	54	means	mean	NOUN
ejpam-1184	35	55	,	,	PUNCT
ejpam-1184	35	56	modes	mode	NOUN
ejpam-1184	35	57	,	,	PUNCT
ejpam-1184	35	58	medians	median	NOUN
ejpam-1184	35	59	or	or	CCONJ
ejpam-1184	35	60	medoids	medoid	NOUN
ejpam-1184	35	61	)	)	PUNCT
ejpam-1184	35	62	.	.	PUNCT
ejpam-1184	36	1	we	we	PRON
ejpam-1184	36	2	can	can	AUX
ejpam-1184	36	3	easily	easily	ADV
ejpam-1184	36	4	envision	envision	VERB
ejpam-1184	36	5	that	that	PRON
ejpam-1184	36	6	to	to	PART
ejpam-1184	36	7	identify	identify	VERB
ejpam-1184	36	8	possible	possible	ADJ
ejpam-1184	36	9	clusters	cluster	NOUN
ejpam-1184	36	10	of	of	ADP
ejpam-1184	36	11	observations	observation	NOUN
ejpam-1184	36	12	in	in	ADP
ejpam-1184	36	13	data	datum	NOUN
ejpam-1184	36	14	,	,	PUNCT
ejpam-1184	36	15	it	it	PRON
ejpam-1184	36	16	is	be	AUX
ejpam-1184	36	17	of	of	ADP
ejpam-1184	36	18	essential	essential	ADJ
ejpam-1184	36	19	importance	importance	NOUN
ejpam-1184	36	20	to	to	PART
ejpam-1184	36	21	have	have	VERB
ejpam-1184	36	22	the	the	DET
ejpam-1184	36	23	knowledge	knowledge	NOUN
ejpam-1184	36	24	of	of	ADP
ejpam-1184	36	25	how	how	SCONJ
ejpam-1184	36	26	“	"	PUNCT
ejpam-1184	36	27	close	close	ADJ
ejpam-1184	36	28	”	"	PUNCT
ejpam-1184	36	29	individuals	individual	NOUN
ejpam-1184	36	30	are	be	AUX
ejpam-1184	36	31	to	to	ADP
ejpam-1184	36	32	each	each	DET
ejpam-1184	36	33	other	other	ADJ
ejpam-1184	36	34	,	,	PUNCT
ejpam-1184	36	35	or	or	CCONJ
ejpam-1184	36	36	how	how	SCONJ
ejpam-1184	36	37	far	far	ADV
ejpam-1184	36	38	apart	apart	ADV
ejpam-1184	36	39	they	they	PRON
ejpam-1184	36	40	are	be	AUX
ejpam-1184	36	41	.	.	PUNCT
ejpam-1184	37	1	the	the	DET
ejpam-1184	37	2	aforementioned	aforementioned	ADJ
ejpam-1184	37	3	methods	method	NOUN
ejpam-1184	37	4	such	such	ADJ
ejpam-1184	37	5	as	as	ADP
ejpam-1184	37	6	the	the	DET
ejpam-1184	37	7	single	single	ADJ
ejpam-1184	37	8	linkage	linkage	NOUN
ejpam-1184	37	9	,	,	PUNCT
ejpam-1184	37	10	complete	complete	ADJ
ejpam-1184	37	11	linkage	linkage	NOUN
ejpam-1184	37	12	,	,	PUNCT
ejpam-1184	37	13	k	k	NOUN
ejpam-1184	37	14	-	-	PUNCT
ejpam-1184	37	15	means	means	PROPN
ejpam-1184	37	16	,	,	PUNCT
ejpam-1184	37	17	etc	etc	X
ejpam-1184	37	18	.	.	X
ejpam-1184	37	19	are	be	AUX
ejpam-1184	37	20	usually	usually	ADV
ejpam-1184	37	21	considered	consider	VERB
ejpam-1184	37	22	as	as	ADP
ejpam-1184	37	23	descriptive	descriptive	ADJ
ejpam-1184	37	24	methods	method	NOUN
ejpam-1184	37	25	since	since	SCONJ
ejpam-1184	37	26	they	they	PRON
ejpam-1184	37	27	are	be	AUX
ejpam-1184	37	28	mainly	mainly	ADV
ejpam-1184	37	29	heuristically	heuristically	ADV
ejpam-1184	37	30	motivated	motivate	VERB
ejpam-1184	37	31	and	and	CCONJ
ejpam-1184	37	32	use	use	VERB
ejpam-1184	37	33	descriptive	descriptive	ADJ
ejpam-1184	37	34	statistics	statistic	NOUN
ejpam-1184	37	35	as	as	ADP
ejpam-1184	37	36	the	the	DET
ejpam-1184	37	37	measures	measure	NOUN
ejpam-1184	37	38	of	of	ADP
ejpam-1184	37	39	similarity	similarity	NOUN
ejpam-1184	37	40	or	or	CCONJ
ejpam-1184	37	41	dissimilarity	dissimilarity	NOUN
ejpam-1184	37	42	between	between	ADP
ejpam-1184	37	43	observations	observation	NOUN
ejpam-1184	37	44	.	.	PUNCT
ejpam-1184	38	1	for	for	ADP
ejpam-1184	38	2	instance	instance	NOUN
ejpam-1184	38	3	,	,	PUNCT
ejpam-1184	38	4	the	the	DET
ejpam-1184	38	5	k	k	NOUN
ejpam-1184	38	6	-	-	PUNCT
ejpam-1184	38	7	means	means	NOUN
ejpam-1184	38	8	type	type	NOUN
ejpam-1184	38	9	of	of	ADP
ejpam-1184	38	10	methods	method	NOUN
ejpam-1184	38	11	are	be	AUX
ejpam-1184	38	12	characterized	characterize	VERB
ejpam-1184	38	13	by	by	ADP
ejpam-1184	38	14	taking	take	VERB
ejpam-1184	38	15	the	the	DET
ejpam-1184	38	16	distance	distance	NOUN
ejpam-1184	38	17	(	(	PUNCT
ejpam-1184	38	18	euclidean	euclidean	NOUN
ejpam-1184	38	19	or	or	CCONJ
ejpam-1184	38	20	manhattan	manhattan	PROPN
ejpam-1184	38	21	or	or	CCONJ
ejpam-1184	38	22	minkowski	minkowski	ADJ
ejpam-1184	38	23	distances	distance	NOUN
ejpam-1184	38	24	)	)	PUNCT
ejpam-1184	38	25	of	of	ADP
ejpam-1184	38	26	each	each	DET
ejpam-1184	38	27	object	object	NOUN
ejpam-1184	38	28	to	to	ADP
ejpam-1184	38	29	the	the	DET
ejpam-1184	38	30	cluster	cluster	NOUN
ejpam-1184	38	31	centres	centre	NOUN
ejpam-1184	38	32	(	(	PUNCT
ejpam-1184	38	33	mean	mean	ADJ
ejpam-1184	38	34	,	,	PUNCT
ejpam-1184	38	35	median	median	ADJ
ejpam-1184	38	36	,	,	PUNCT
ejpam-1184	38	37	mode	mode	NOUN
ejpam-1184	38	38	,	,	PUNCT
ejpam-1184	38	39	medoid	medoid	PROPN
ejpam-1184	38	40	)	)	PUNCT
ejpam-1184	38	41	as	as	ADP
ejpam-1184	38	42	the	the	DET
ejpam-1184	38	43	similarity	similarity	NOUN
ejpam-1184	38	44	or	or	CCONJ
ejpam-1184	38	45	dissimilarity	dissimilarity	NOUN
ejpam-1184	38	46	measure	measure	NOUN
ejpam-1184	38	47	.	.	PUNCT
ejpam-1184	39	1	on	on	ADP
ejpam-1184	39	2	the	the	DET
ejpam-1184	39	3	other	other	ADJ
ejpam-1184	39	4	hand	hand	NOUN
ejpam-1184	39	5	,	,	PUNCT
ejpam-1184	39	6	model	model	NOUN
ejpam-1184	39	7	-	-	PUNCT
ejpam-1184	39	8	based	base	VERB
ejpam-1184	39	9	clustering	clustering	NOUN
ejpam-1184	39	10	uses	use	VERB
ejpam-1184	39	11	a	a	DET
ejpam-1184	39	12	probability	probability	NOUN
ejpam-1184	39	13	model	model	NOUN
ejpam-1184	39	14	as	as	ADP
ejpam-1184	39	15	the	the	DET
ejpam-1184	39	16	similarity	similarity	NOUN
ejpam-1184	39	17	or	or	CCONJ
ejpam-1184	39	18	dissimilarity	dissimilarity	NOUN
ejpam-1184	39	19	measure	measure	NOUN
ejpam-1184	39	20	,	,	PUNCT
ejpam-1184	39	21	i.e.	i.e.	X
ejpam-1184	39	22	objects	object	NOUN
ejpam-1184	39	23	have	have	VERB
ejpam-1184	39	24	the	the	DET
ejpam-1184	39	25	same	same	ADJ
ejpam-1184	39	26	model	model	NOUN
ejpam-1184	39	27	specification	specification	NOUN
ejpam-1184	39	28	within	within	ADP
ejpam-1184	39	29	clusters	cluster	NOUN
ejpam-1184	39	30	.	.	PUNCT
ejpam-1184	40	1	furthermore	furthermore	ADV
ejpam-1184	40	2	,	,	PUNCT
ejpam-1184	40	3	model	model	NOUN
ejpam-1184	40	4	-	-	PUNCT
ejpam-1184	40	5	based	base	VERB
ejpam-1184	40	6	clustering	clustering	NOUN
ejpam-1184	40	7	techniques	technique	NOUN
ejpam-1184	40	8	use	use	VERB
ejpam-1184	40	9	inferential	inferential	ADJ
ejpam-1184	40	10	statistics	statistic	NOUN
ejpam-1184	40	11	by	by	ADP
ejpam-1184	40	12	means	mean	NOUN
ejpam-1184	40	13	of	of	ADP
ejpam-1184	40	14	probabilistic	probabilistic	ADJ
ejpam-1184	40	15	models	model	NOUN
ejpam-1184	40	16	,	,	PUNCT
ejpam-1184	40	17	not	not	PART
ejpam-1184	40	18	only	only	ADV
ejpam-1184	40	19	for	for	ADP
ejpam-1184	40	20	checking	check	VERB
ejpam-1184	40	21	the	the	DET
ejpam-1184	40	22	significance	significance	NOUN
ejpam-1184	40	23	of	of	ADP
ejpam-1184	40	24	clusters	cluster	NOUN
ejpam-1184	40	25	and	and	CCONJ
ejpam-1184	40	26	clustering	clustering	NOUN
ejpam-1184	40	27	,	,	PUNCT
ejpam-1184	40	28	but	but	CCONJ
ejpam-1184	40	29	also	also	ADV
ejpam-1184	40	30	for	for	ADP
ejpam-1184	40	31	providing	provide	VERB
ejpam-1184	40	32	a	a	DET
ejpam-1184	40	33	firm	firm	ADJ
ejpam-1184	40	34	theoretical	theoretical	ADJ
ejpam-1184	40	35	basis	basis	NOUN
ejpam-1184	40	36	for	for	ADP
ejpam-1184	40	37	clustering	clustering	ADJ
ejpam-1184	40	38	methods	method	NOUN
ejpam-1184	40	39	and	and	CCONJ
ejpam-1184	40	40	strategies	strategy	NOUN
ejpam-1184	40	41	.	.	PUNCT
ejpam-1184	41	1	model	model	NOUN
ejpam-1184	41	2	-	-	PUNCT
ejpam-1184	41	3	based	base	VERB
ejpam-1184	41	4	methods	method	NOUN
ejpam-1184	41	5	can	can	AUX
ejpam-1184	41	6	be	be	AUX
ejpam-1184	41	7	applied	apply	VERB
ejpam-1184	41	8	both	both	CCONJ
ejpam-1184	41	9	in	in	ADP
ejpam-1184	41	10	hierarchical	hierarchical	ADJ
ejpam-1184	41	11	clustering	clustering	NOUN
ejpam-1184	41	12	and	and	CCONJ
ejpam-1184	41	13	partitioning	partitioning	ADJ
ejpam-1184	41	14	-	-	PUNCT
ejpam-1184	41	15	type	type	NOUN
ejpam-1184	41	16	clustering	clustering	NOUN
ejpam-1184	41	17	.	.	PUNCT
ejpam-1184	42	1	this	this	DET
ejpam-1184	42	2	research	research	NOUN
ejpam-1184	42	3	focuses	focus	VERB
ejpam-1184	42	4	on	on	ADP
ejpam-1184	42	5	partitioning	partition	VERB
ejpam-1184	42	6	-	-	PUNCT
ejpam-1184	42	7	type	type	NOUN
ejpam-1184	42	8	model	model	NOUN
ejpam-1184	42	9	-	-	PUNCT
ejpam-1184	42	10	based	base	VERB
ejpam-1184	42	11	approaches	approach	NOUN
ejpam-1184	42	12	.	.	PUNCT
ejpam-1184	43	1	it	it	PRON
ejpam-1184	43	2	is	be	AUX
ejpam-1184	43	3	noted	note	VERB
ejpam-1184	43	4	that	that	SCONJ
ejpam-1184	43	5	there	there	PRON
ejpam-1184	43	6	is	be	VERB
ejpam-1184	43	7	no	no	DET
ejpam-1184	43	8	absolute	absolute	ADJ
ejpam-1184	43	9	boundary	boundary	NOUN
ejpam-1184	43	10	between	between	ADP
ejpam-1184	43	11	descriptive	descriptive	ADJ
ejpam-1184	43	12	and	and	CCONJ
ejpam-1184	43	13	model	model	ADV
ejpam-1184	43	14	-	-	PUNCT
ejpam-1184	43	15	based	base	VERB
ejpam-1184	43	16	clustering	clustering	ADJ
ejpam-1184	43	17	methods	method	NOUN
ejpam-1184	43	18	.	.	PUNCT
ejpam-1184	44	1	some	some	DET
ejpam-1184	44	2	clustering	clustering	ADJ
ejpam-1184	44	3	methods	method	NOUN
ejpam-1184	44	4	were	be	AUX
ejpam-1184	44	5	heuristically	heuristically	ADV
ejpam-1184	44	6	motivated	motivate	VERB
ejpam-1184	44	7	,	,	PUNCT
ejpam-1184	44	8	but	but	CCONJ
ejpam-1184	44	9	later	later	ADV
ejpam-1184	44	10	on	on	ADP
ejpam-1184	44	11	statisticians	statistician	NOUN
ejpam-1184	44	12	studied	study	VERB
ejpam-1184	44	13	their	their	PRON
ejpam-1184	44	14	performance	performance	NOUN
ejpam-1184	44	15	from	from	ADP
ejpam-1184	44	16	a	a	DET
ejpam-1184	44	17	probabilistic	probabilistic	ADJ
ejpam-1184	44	18	perspective	perspective	NOUN
ejpam-1184	44	19	.	.	PUNCT
ejpam-1184	45	1	for	for	ADP
ejpam-1184	45	2	instance	instance	NOUN
ejpam-1184	45	3	,	,	PUNCT
ejpam-1184	45	4	[	[	X
ejpam-1184	45	5	10	10	NUM
ejpam-1184	45	6	,	,	PUNCT
ejpam-1184	45	7	3	3	NUM
ejpam-1184	45	8	]	]	PUNCT
ejpam-1184	45	9	studied	study	VERB
ejpam-1184	45	10	the	the	DET
ejpam-1184	45	11	asymptotic	asymptotic	ADJ
ejpam-1184	45	12	behaviour	behaviour	NOUN
ejpam-1184	45	13	of	of	ADP
ejpam-1184	45	14	k	k	NOUN
ejpam-1184	45	15	-	-	PUNCT
ejpam-1184	45	16	means	mean	VERB
ejpam-1184	45	17	using	use	VERB
ejpam-1184	45	18	a	a	DET
ejpam-1184	45	19	model	model	NOUN
ejpam-1184	45	20	-	-	PUNCT
ejpam-1184	45	21	based	base	VERB
ejpam-1184	45	22	approach	approach	NOUN
ejpam-1184	45	23	;	;	PUNCT
ejpam-1184	45	24	[	[	X
ejpam-1184	45	25	8	8	NUM
ejpam-1184	45	26	,	,	PUNCT
ejpam-1184	45	27	12	12	NUM
ejpam-1184	45	28	]	]	PUNCT
ejpam-1184	45	29	investigated	investigate	VERB
ejpam-1184	45	30	g.	g.	PROPN
ejpam-1184	45	31	qian	qian	PROPN
ejpam-1184	45	32	,	,	PUNCT
ejpam-1184	45	33	y.	y.	PROPN
ejpam-1184	45	34	wu	wu	PROPN
ejpam-1184	45	35	/	/	SYM
ejpam-1184	45	36	eur	eur	PROPN
ejpam-1184	45	37	.	.	PUNCT
ejpam-1184	46	1	j.	j.	PROPN
ejpam-1184	46	2	pure	pure	PROPN
ejpam-1184	46	3	appl	appl	PROPN
ejpam-1184	46	4	.	.	PROPN
ejpam-1184	46	5	math	math	PROPN
ejpam-1184	46	6	,	,	PUNCT
ejpam-1184	46	7	4	4	NUM
ejpam-1184	46	8	(	(	PUNCT
ejpam-1184	46	9	2011	2011	NUM
ejpam-1184	46	10	)	)	PUNCT
ejpam-1184	46	11	,	,	PUNCT
ejpam-1184	46	12	455	455	NUM
ejpam-1184	46	13	-	-	SYM
ejpam-1184	46	14	466	466	NUM
ejpam-1184	46	15	457	457	NUM
ejpam-1184	46	16	the	the	DET
ejpam-1184	46	17	mathematical	mathematical	ADJ
ejpam-1184	46	18	relationship	relationship	NOUN
ejpam-1184	46	19	between	between	ADP
ejpam-1184	46	20	high	high	ADJ
ejpam-1184	46	21	-	-	PUNCT
ejpam-1184	46	22	density	density	NOUN
ejpam-1184	46	23	clusters	cluster	NOUN
ejpam-1184	46	24	and	and	CCONJ
ejpam-1184	46	25	the	the	DET
ejpam-1184	46	26	single	single	ADJ
ejpam-1184	46	27	-	-	PUNCT
ejpam-1184	46	28	linkage	linkage	NOUN
ejpam-1184	46	29	clustering	clustering	NOUN
ejpam-1184	46	30	method	method	NOUN
ejpam-1184	46	31	.	.	PUNCT
ejpam-1184	47	1	consider	consider	VERB
ejpam-1184	47	2	a	a	DET
ejpam-1184	47	3	finite	finite	ADJ
ejpam-1184	47	4	set	set	NOUN
ejpam-1184	47	5	of	of	ADP
ejpam-1184	47	6	n	n	X
ejpam-1184	47	7	objects	object	VERB
ejpam-1184	47	8	o	o	NOUN
ejpam-1184	47	9	=	=	PUNCT
ejpam-1184	47	10	{	{	PUNCT
ejpam-1184	47	11	1	1	NUM
ejpam-1184	47	12	,	,	PUNCT
ejpam-1184	47	13	.	.	PUNCT
ejpam-1184	47	14	.	.	PUNCT
ejpam-1184	48	1	.	.	PUNCT
ejpam-1184	49	1	,	,	PUNCT
ejpam-1184	50	1	n	n	CCONJ
ejpam-1184	50	2	}	}	PUNCT
ejpam-1184	50	3	together	together	ADV
ejpam-1184	50	4	with	with	ADP
ejpam-1184	50	5	data	datum	NOUN
ejpam-1184	50	6	y1	y1	NOUN
ejpam-1184	50	7	,	,	PUNCT
ejpam-1184	50	8	.	.	PUNCT
ejpam-1184	50	9	.	.	PUNCT
ejpam-1184	51	1	.	.	PUNCT
ejpam-1184	52	1	,	,	PUNCT
ejpam-1184	52	2	yn	yn	PROPN
ejpam-1184	52	3	∈	∈	PROPN
ejpam-1184	52	4	r	r	NOUN
ejpam-1184	52	5	p	p	NOUN
ejpam-1184	52	6	describing	describe	VERB
ejpam-1184	52	7	the	the	DET
ejpam-1184	52	8	properties	property	NOUN
ejpam-1184	52	9	of	of	ADP
ejpam-1184	52	10	these	these	DET
ejpam-1184	52	11	objects	object	NOUN
ejpam-1184	52	12	.	.	PUNCT
ejpam-1184	53	1	based	base	VERB
ejpam-1184	53	2	on	on	ADP
ejpam-1184	53	3	these	these	DET
ejpam-1184	53	4	data	datum	NOUN
ejpam-1184	53	5	,	,	PUNCT
ejpam-1184	53	6	our	our	PRON
ejpam-1184	53	7	problem	problem	NOUN
ejpam-1184	53	8	is	be	AUX
ejpam-1184	53	9	to	to	PART
ejpam-1184	53	10	recover	recover	VERB
ejpam-1184	53	11	the	the	DET
ejpam-1184	53	12	latent	latent	NOUN
ejpam-1184	53	13	partitioning	partition	VERB
ejpam-1184	53	14	π	π	PROPN
ejpam-1184	53	15	=	=	SYM
ejpam-1184	53	16	(	(	PUNCT
ejpam-1184	53	17	c1	c1	PROPN
ejpam-1184	53	18	,	,	PUNCT
ejpam-1184	53	19	.	.	PUNCT
ejpam-1184	53	20	.	.	PUNCT
ejpam-1184	54	1	.	.	PUNCT
ejpam-1184	55	1	,	,	PUNCT
ejpam-1184	55	2	ck	ck	PROPN
ejpam-1184	55	3	)	)	PUNCT
ejpam-1184	55	4	of	of	ADP
ejpam-1184	55	5	o	o	PROPN
ejpam-1184	55	6	and	and	CCONJ
ejpam-1184	55	7	to	to	PART
ejpam-1184	55	8	construct	construct	VERB
ejpam-1184	55	9	a	a	DET
ejpam-1184	55	10	clustering	clustering	NOUN
ejpam-1184	55	11	of	of	ADP
ejpam-1184	55	12	the	the	DET
ejpam-1184	55	13	corresponding	corresponding	ADJ
ejpam-1184	55	14	objects	object	NOUN
ejpam-1184	55	15	.	.	PUNCT
ejpam-1184	56	1	a	a	DET
ejpam-1184	56	2	model	model	NOUN
ejpam-1184	56	3	-	-	PUNCT
ejpam-1184	56	4	based	base	VERB
ejpam-1184	56	5	or	or	CCONJ
ejpam-1184	56	6	probabilistic	probabilistic	ADJ
ejpam-1184	56	7	clustering	clustering	ADJ
ejpam-1184	56	8	approach	approach	NOUN
ejpam-1184	56	9	assumes	assume	VERB
ejpam-1184	56	10	that	that	SCONJ
ejpam-1184	56	11	the	the	DET
ejpam-1184	56	12	observed	observe	VERB
ejpam-1184	56	13	data	data	NOUN
ejpam-1184	56	14	y1	y1	NOUN
ejpam-1184	56	15	,	,	PUNCT
ejpam-1184	56	16	.	.	PUNCT
ejpam-1184	56	17	.	.	PUNCT
ejpam-1184	57	1	.	.	PUNCT
ejpam-1184	58	1	,	,	PUNCT
ejpam-1184	58	2	y	y	PROPN
ejpam-1184	58	3	n	n	ADV
ejpam-1184	58	4	are	be	AUX
ejpam-1184	58	5	a	a	DET
ejpam-1184	58	6	sample	sample	NOUN
ejpam-1184	58	7	of	of	ADP
ejpam-1184	58	8	random	random	ADJ
ejpam-1184	58	9	vectors	vector	NOUN
ejpam-1184	58	10	y	y	PROPN
ejpam-1184	58	11	1	1	NUM
ejpam-1184	58	12	,	,	PUNCT
ejpam-1184	58	13	.	.	PUNCT
ejpam-1184	58	14	.	.	PUNCT
ejpam-1184	59	1	.	.	PUNCT
ejpam-1184	60	1	,	,	PUNCT
ejpam-1184	60	2	y	y	PROPN
ejpam-1184	60	3	n	n	PROPN
ejpam-1184	60	4	that	that	PRON
ejpam-1184	60	5	belong	belong	VERB
ejpam-1184	60	6	to	to	ADP
ejpam-1184	60	7	a	a	DET
ejpam-1184	60	8	structured	structured	ADJ
ejpam-1184	60	9	population	population	NOUN
ejpam-1184	60	10	.	.	PUNCT
ejpam-1184	61	1	it	it	PRON
ejpam-1184	61	2	characterizes	characterize	VERB
ejpam-1184	61	3	the	the	DET
ejpam-1184	61	4	clusters	cluster	NOUN
ejpam-1184	61	5	by	by	ADP
ejpam-1184	61	6	specifications	specification	NOUN
ejpam-1184	61	7	for	for	ADP
ejpam-1184	61	8	the	the	DET
ejpam-1184	61	9	probability	probability	NOUN
ejpam-1184	61	10	distribution	distribution	NOUN
ejpam-1184	61	11	of	of	ADP
ejpam-1184	61	12	the	the	DET
ejpam-1184	61	13	random	random	ADJ
ejpam-1184	61	14	vectors	vector	NOUN
ejpam-1184	61	15	y1	y1	NOUN
ejpam-1184	61	16	,	,	PUNCT
ejpam-1184	61	17	.	.	PUNCT
ejpam-1184	61	18	.	.	PUNCT
ejpam-1184	62	1	.	.	PUNCT
ejpam-1184	63	1	,	,	PUNCT
ejpam-1184	63	2	y	y	PROPN
ejpam-1184	63	3	n	n	PROPN
ejpam-1184	63	4	which	which	PRON
ejpam-1184	63	5	may	may	AUX
ejpam-1184	63	6	differ	differ	VERB
ejpam-1184	63	7	from	from	ADP
ejpam-1184	63	8	cluster	cluster	NOUN
ejpam-1184	63	9	to	to	ADP
ejpam-1184	63	10	cluster	cluster	NOUN
ejpam-1184	63	11	.	.	PUNCT
ejpam-1184	64	1	roughly	roughly	ADV
ejpam-1184	64	2	speaking	speak	VERB
ejpam-1184	64	3	,	,	PUNCT
ejpam-1184	64	4	stochastic	stochastic	ADJ
ejpam-1184	64	5	model	model	NOUN
ejpam-1184	64	6	-	-	PUNCT
ejpam-1184	64	7	based	base	VERB
ejpam-1184	64	8	clustering	clustering	NOUN
ejpam-1184	64	9	techniques	technique	NOUN
ejpam-1184	64	10	can	can	AUX
ejpam-1184	64	11	be	be	AUX
ejpam-1184	64	12	divided	divide	VERB
ejpam-1184	64	13	into	into	ADP
ejpam-1184	64	14	the	the	DET
ejpam-1184	64	15	following	follow	VERB
ejpam-1184	64	16	two	two	NUM
ejpam-1184	64	17	categories	category	NOUN
ejpam-1184	64	18	:	:	PUNCT
ejpam-1184	64	19	(	(	PUNCT
ejpam-1184	64	20	1	1	X
ejpam-1184	64	21	)	)	PUNCT
ejpam-1184	64	22	parametric	parametric	ADJ
ejpam-1184	64	23	approach	approach	NOUN
ejpam-1184	64	24	in	in	ADP
ejpam-1184	64	25	which	which	PRON
ejpam-1184	64	26	the	the	DET
ejpam-1184	64	27	probability	probability	NOUN
ejpam-1184	64	28	distribution	distribution	NOUN
ejpam-1184	64	29	of	of	ADP
ejpam-1184	64	30	y	y	PROPN
ejpam-1184	64	31	1	1	NUM
ejpam-1184	64	32	,	,	PUNCT
ejpam-1184	64	33	.	.	PUNCT
ejpam-1184	64	34	.	.	PUNCT
ejpam-1184	65	1	.	.	PUNCT
ejpam-1184	66	1	,	,	PUNCT
ejpam-1184	66	2	y	y	PROPN
ejpam-1184	66	3	n	n	ADV
ejpam-1184	66	4	is	be	AUX
ejpam-1184	66	5	assumed	assume	VERB
ejpam-1184	66	6	to	to	PART
ejpam-1184	66	7	have	have	VERB
ejpam-1184	66	8	a	a	DET
ejpam-1184	66	9	known	know	VERB
ejpam-1184	66	10	parametric	parametric	ADJ
ejpam-1184	66	11	form	form	NOUN
ejpam-1184	66	12	but	but	CCONJ
ejpam-1184	66	13	with	with	ADP
ejpam-1184	66	14	unknown	unknown	ADJ
ejpam-1184	66	15	parameters	parameter	NOUN
ejpam-1184	66	16	;	;	PUNCT
ejpam-1184	66	17	(	(	PUNCT
ejpam-1184	66	18	2	2	X
ejpam-1184	66	19	)	)	PUNCT
ejpam-1184	66	20	non	non	ADJ
ejpam-1184	66	21	-	-	ADJ
ejpam-1184	66	22	parametric	parametric	ADJ
ejpam-1184	66	23	approach	approach	NOUN
ejpam-1184	66	24	in	in	ADP
ejpam-1184	66	25	which	which	PRON
ejpam-1184	66	26	no	no	DET
ejpam-1184	66	27	distributional	distributional	ADJ
ejpam-1184	66	28	assumption	assumption	NOUN
ejpam-1184	66	29	is	be	AUX
ejpam-1184	66	30	explicitly	explicitly	ADV
ejpam-1184	66	31	made	make	VERB
ejpam-1184	66	32	for	for	ADP
ejpam-1184	66	33	the	the	DET
ejpam-1184	66	34	individual	individual	ADJ
ejpam-1184	66	35	clusters	cluster	NOUN
ejpam-1184	66	36	.	.	PUNCT
ejpam-1184	67	1	this	this	DET
ejpam-1184	67	2	paper	paper	NOUN
ejpam-1184	67	3	will	will	AUX
ejpam-1184	67	4	focus	focus	VERB
ejpam-1184	67	5	on	on	ADP
ejpam-1184	67	6	parametric	parametric	ADJ
ejpam-1184	67	7	model	model	NOUN
ejpam-1184	67	8	-	-	PUNCT
ejpam-1184	67	9	based	base	VERB
ejpam-1184	67	10	partitioning	partitioning	NOUN
ejpam-1184	67	11	-	-	PUNCT
ejpam-1184	67	12	type	type	NOUN
ejpam-1184	67	13	clustering	clustering	ADJ
ejpam-1184	67	14	methods	method	NOUN
ejpam-1184	67	15	,	,	PUNCT
ejpam-1184	67	16	in	in	ADP
ejpam-1184	67	17	particular	particular	ADJ
ejpam-1184	67	18	,	,	PUNCT
ejpam-1184	67	19	the	the	DET
ejpam-1184	67	20	likelihood	likelihood	NOUN
ejpam-1184	67	21	method	method	NOUN
ejpam-1184	67	22	.	.	PUNCT
ejpam-1184	68	1	further	far	ADV
ejpam-1184	68	2	,	,	PUNCT
ejpam-1184	68	3	we	we	PRON
ejpam-1184	68	4	study	study	VERB
ejpam-1184	68	5	only	only	ADV
ejpam-1184	68	6	the	the	DET
ejpam-1184	68	7	regression	regression	NOUN
ejpam-1184	68	8	clustering	clustering	NOUN
ejpam-1184	68	9	problem	problem	NOUN
ejpam-1184	68	10	,	,	PUNCT
ejpam-1184	68	11	which	which	PRON
ejpam-1184	68	12	means	mean	VERB
ejpam-1184	68	13	the	the	DET
ejpam-1184	68	14	data	data	NOUN
ejpam-1184	68	15	y1	y1	NOUN
ejpam-1184	68	16	,	,	PUNCT
ejpam-1184	68	17	.	.	PUNCT
ejpam-1184	68	18	.	.	PUNCT
ejpam-1184	69	1	.	.	PUNCT
ejpam-1184	70	1	,	,	PUNCT
ejpam-1184	70	2	yn	yn	PRON
ejpam-1184	70	3	contain	contain	VERB
ejpam-1184	70	4	observations	observation	NOUN
ejpam-1184	70	5	of	of	ADP
ejpam-1184	70	6	both	both	CCONJ
ejpam-1184	70	7	dependent	dependent	ADJ
ejpam-1184	70	8	and	and	CCONJ
ejpam-1184	70	9	explanatory	explanatory	ADJ
ejpam-1184	70	10	variables	variable	NOUN
ejpam-1184	70	11	and	and	CCONJ
ejpam-1184	70	12	their	their	PRON
ejpam-1184	70	13	relationship	relationship	NOUN
ejpam-1184	70	14	is	be	AUX
ejpam-1184	70	15	of	of	ADP
ejpam-1184	70	16	our	our	PRON
ejpam-1184	70	17	interest	interest	NOUN
ejpam-1184	70	18	.	.	PUNCT
ejpam-1184	71	1	in	in	ADP
ejpam-1184	71	2	section	section	NOUN
ejpam-1184	71	3	2	2	NUM
ejpam-1184	71	4	,	,	PUNCT
ejpam-1184	71	5	we	we	PRON
ejpam-1184	71	6	review	review	VERB
ejpam-1184	71	7	regression	regression	NOUN
ejpam-1184	71	8	clustering	cluster	VERB
ejpam-1184	71	9	.	.	PUNCT
ejpam-1184	72	1	in	in	ADP
ejpam-1184	72	2	section	section	NOUN
ejpam-1184	72	3	3	3	NUM
ejpam-1184	72	4	,	,	PUNCT
ejpam-1184	72	5	we	we	PRON
ejpam-1184	72	6	discuss	discuss	VERB
ejpam-1184	72	7	a	a	DET
ejpam-1184	72	8	procedures	procedure	NOUN
ejpam-1184	72	9	for	for	ADP
ejpam-1184	72	10	estimating	estimate	VERB
ejpam-1184	72	11	the	the	DET
ejpam-1184	72	12	parameters	parameter	NOUN
ejpam-1184	72	13	and	and	CCONJ
ejpam-1184	72	14	the	the	DET
ejpam-1184	72	15	number	number	NOUN
ejpam-1184	72	16	of	of	ADP
ejpam-1184	72	17	clusters	cluster	NOUN
ejpam-1184	72	18	in	in	ADP
ejpam-1184	72	19	linear	linear	PROPN
ejpam-1184	72	20	regression	regression	NOUN
ejpam-1184	72	21	clustering	cluster	VERB
ejpam-1184	72	22	under	under	ADP
ejpam-1184	72	23	the	the	DET
ejpam-1184	72	24	classification	classification	NOUN
ejpam-1184	72	25	likelihood	likelihood	NOUN
ejpam-1184	72	26	framework	framework	NOUN
ejpam-1184	72	27	.	.	PUNCT
ejpam-1184	73	1	in	in	ADP
ejpam-1184	73	2	section	section	NOUN
ejpam-1184	73	3	4	4	NUM
ejpam-1184	73	4	,	,	PUNCT
ejpam-1184	73	5	an	an	DET
ejpam-1184	73	6	algorithm	algorithm	NOUN
ejpam-1184	73	7	is	be	AUX
ejpam-1184	73	8	given	give	VERB
ejpam-1184	73	9	for	for	ADP
ejpam-1184	73	10	selecting	select	VERB
ejpam-1184	73	11	the	the	DET
ejpam-1184	73	12	clustering	clustering	NOUN
ejpam-1184	73	13	and	and	CCONJ
ejpam-1184	73	14	the	the	DET
ejpam-1184	73	15	number	number	NOUN
ejpam-1184	73	16	of	of	ADP
ejpam-1184	73	17	clusters	cluster	NOUN
ejpam-1184	73	18	in	in	ADP
ejpam-1184	73	19	regression	regression	NOUN
ejpam-1184	73	20	clustering	cluster	VERB
ejpam-1184	73	21	.	.	PUNCT
ejpam-1184	74	1	the	the	DET
ejpam-1184	74	2	simulation	simulation	NOUN
ejpam-1184	74	3	study	study	NOUN
ejpam-1184	74	4	is	be	AUX
ejpam-1184	74	5	presented	present	VERB
ejpam-1184	74	6	in	in	ADP
ejpam-1184	74	7	section	section	NOUN
ejpam-1184	74	8	5	5	NUM
ejpam-1184	74	9	.	.	PUNCT
ejpam-1184	75	1	finally	finally	ADV
ejpam-1184	75	2	,	,	PUNCT
ejpam-1184	75	3	section	section	NOUN
ejpam-1184	75	4	6	6	NUM
ejpam-1184	75	5	provides	provide	VERB
ejpam-1184	75	6	some	some	DET
ejpam-1184	75	7	discussions	discussion	NOUN
ejpam-1184	75	8	and	and	CCONJ
ejpam-1184	75	9	concludes	conclude	VERB
ejpam-1184	75	10	the	the	DET
ejpam-1184	75	11	paper	paper	NOUN
ejpam-1184	75	12	.	.	PUNCT
ejpam-1184	76	1	2	2	X
ejpam-1184	76	2	.	.	X
ejpam-1184	76	3	regression	regression	NOUN
ejpam-1184	76	4	clustering	cluster	VERB
ejpam-1184	76	5	regression	regression	NOUN
ejpam-1184	76	6	clustering	cluster	VERB
ejpam-1184	76	7	refers	refer	VERB
ejpam-1184	76	8	to	to	ADP
ejpam-1184	76	9	estimating	estimate	VERB
ejpam-1184	76	10	the	the	DET
ejpam-1184	76	11	class	class	NOUN
ejpam-1184	76	12	-	-	PUNCT
ejpam-1184	76	13	specific	specific	ADJ
ejpam-1184	76	14	regression	regression	NOUN
ejpam-1184	76	15	hyperplanes	hyperplane	NOUN
ejpam-1184	76	16	underlying	underlie	VERB
ejpam-1184	76	17	the	the	DET
ejpam-1184	76	18	data	datum	NOUN
ejpam-1184	76	19	that	that	PRON
ejpam-1184	76	20	randomly	randomly	ADV
ejpam-1184	76	21	come	come	VERB
ejpam-1184	76	22	from	from	ADP
ejpam-1184	76	23	a	a	DET
ejpam-1184	76	24	population	population	NOUN
ejpam-1184	76	25	consisting	consist	VERB
ejpam-1184	76	26	of	of	ADP
ejpam-1184	76	27	distinct	distinct	ADJ
ejpam-1184	76	28	classes	class	NOUN
ejpam-1184	76	29	.	.	PUNCT
ejpam-1184	77	1	note	note	VERB
ejpam-1184	77	2	that	that	SCONJ
ejpam-1184	77	3	the	the	DET
ejpam-1184	77	4	notion	notion	NOUN
ejpam-1184	77	5	hyperplane	hyperplane	NOUN
ejpam-1184	77	6	used	use	VERB
ejpam-1184	77	7	here	here	ADV
ejpam-1184	77	8	is	be	AUX
ejpam-1184	77	9	a	a	DET
ejpam-1184	77	10	generic	generic	ADJ
ejpam-1184	77	11	one	one	NOUN
ejpam-1184	77	12	,	,	PUNCT
ejpam-1184	77	13	which	which	PRON
ejpam-1184	77	14	means	mean	VERB
ejpam-1184	77	15	it	it	PRON
ejpam-1184	77	16	does	do	AUX
ejpam-1184	77	17	not	not	PART
ejpam-1184	77	18	necessarily	necessarily	ADV
ejpam-1184	77	19	pass	pass	VERB
ejpam-1184	77	20	through	through	ADP
ejpam-1184	77	21	the	the	DET
ejpam-1184	77	22	origin	origin	NOUN
ejpam-1184	77	23	in	in	ADP
ejpam-1184	77	24	the	the	DET
ejpam-1184	77	25	space	space	NOUN
ejpam-1184	77	26	.	.	PUNCT
ejpam-1184	78	1	it	it	PRON
ejpam-1184	78	2	should	should	AUX
ejpam-1184	78	3	be	be	AUX
ejpam-1184	78	4	more	more	ADV
ejpam-1184	78	5	correctly	correctly	ADV
ejpam-1184	78	6	called	call	VERB
ejpam-1184	78	7	an	an	DET
ejpam-1184	78	8	affine	affine	NOUN
ejpam-1184	78	9	set	set	NOUN
ejpam-1184	78	10	.	.	PUNCT
ejpam-1184	79	1	but	but	CCONJ
ejpam-1184	79	2	we	we	PRON
ejpam-1184	79	3	do	do	AUX
ejpam-1184	79	4	not	not	PART
ejpam-1184	79	5	distinguish	distinguish	VERB
ejpam-1184	79	6	them	they	PRON
ejpam-1184	79	7	in	in	ADP
ejpam-1184	79	8	this	this	DET
ejpam-1184	79	9	paper	paper	NOUN
ejpam-1184	79	10	.	.	PUNCT
ejpam-1184	80	1	for	for	ADP
ejpam-1184	80	2	the	the	DET
ejpam-1184	80	3	regression	regression	NOUN
ejpam-1184	80	4	clustering	clustering	NOUN
ejpam-1184	80	5	problem	problem	NOUN
ejpam-1184	80	6	,	,	PUNCT
ejpam-1184	80	7	the	the	DET
ejpam-1184	80	8	data	datum	NOUN
ejpam-1184	80	9	have	have	VERB
ejpam-1184	80	10	the	the	DET
ejpam-1184	80	11	form	form	NOUN
ejpam-1184	80	12	(	(	PUNCT
ejpam-1184	80	13	y	y	PROPN
ejpam-1184	80	14	j	j	PROPN
ejpam-1184	80	15	,	,	PUNCT
ejpam-1184	80	16	x	x	ADJ
ejpam-1184	80	17	′j	′j	NOUN
ejpam-1184	80	18	)	)	PUNCT
ejpam-1184	80	19	,	,	PUNCT
ejpam-1184	80	20	j	j	PROPN
ejpam-1184	80	21	=	=	SYM
ejpam-1184	80	22	1	1	NUM
ejpam-1184	80	23	,	,	PUNCT
ejpam-1184	80	24	.	.	PUNCT
ejpam-1184	80	25	.	.	PUNCT
ejpam-1184	81	1	.	.	PUNCT
ejpam-1184	82	1	,	,	PUNCT
ejpam-1184	83	1	n	n	CCONJ
ejpam-1184	83	2	,	,	PUNCT
ejpam-1184	83	3	where	where	SCONJ
ejpam-1184	83	4	x	x	PUNCT
ejpam-1184	83	5	j	j	PROPN
ejpam-1184	83	6	∈	∈	PROPN
ejpam-1184	84	1	r	r	NOUN
ejpam-1184	84	2	p	p	NOUN
ejpam-1184	84	3	is	be	AUX
ejpam-1184	84	4	a	a	DET
ejpam-1184	84	5	(	(	PUNCT
ejpam-1184	84	6	non	non	ADJ
ejpam-1184	84	7	-	-	ADJ
ejpam-1184	84	8	random	random	ADJ
ejpam-1184	84	9	)	)	PUNCT
ejpam-1184	84	10	explanatory	explanatory	ADJ
ejpam-1184	84	11	column	column	NOUN
ejpam-1184	84	12	vector	vector	NOUN
ejpam-1184	84	13	and	and	CCONJ
ejpam-1184	84	14	y	y	PROPN
ejpam-1184	84	15	j	j	PROPN
ejpam-1184	84	16	∈	∈	PROPN
ejpam-1184	84	17	r	r	NOUN
ejpam-1184	84	18	a	a	DET
ejpam-1184	84	19	random	random	ADJ
ejpam-1184	84	20	dependent	dependent	ADJ
ejpam-1184	84	21	variable	variable	NOUN
ejpam-1184	84	22	for	for	ADP
ejpam-1184	84	23	the	the	DET
ejpam-1184	84	24	j	j	PROPN
ejpam-1184	84	25	-	-	PUNCT
ejpam-1184	84	26	th	th	VERB
ejpam-1184	84	27	object	object	NOUN
ejpam-1184	84	28	.	.	PUNCT
ejpam-1184	85	1	as	as	ADP
ejpam-1184	85	2	in	in	ADP
ejpam-1184	85	3	the	the	DET
ejpam-1184	85	4	general	general	ADJ
ejpam-1184	85	5	setting	setting	NOUN
ejpam-1184	85	6	of	of	ADP
ejpam-1184	85	7	model	model	NOUN
ejpam-1184	85	8	-	-	PUNCT
ejpam-1184	85	9	based	base	VERB
ejpam-1184	85	10	clustering	clustering	NOUN
ejpam-1184	85	11	,	,	PUNCT
ejpam-1184	85	12	there	there	PRON
ejpam-1184	85	13	are	be	VERB
ejpam-1184	85	14	also	also	ADV
ejpam-1184	85	15	two	two	NUM
ejpam-1184	85	16	different	different	ADJ
ejpam-1184	85	17	approaches	approach	NOUN
ejpam-1184	85	18	for	for	ADP
ejpam-1184	85	19	regression	regression	NOUN
ejpam-1184	85	20	clustering	cluster	VERB
ejpam-1184	85	21	in	in	ADP
ejpam-1184	85	22	the	the	DET
ejpam-1184	85	23	literature	literature	NOUN
ejpam-1184	85	24	.	.	PUNCT
ejpam-1184	86	1	one	one	NUM
ejpam-1184	86	2	is	be	AUX
ejpam-1184	86	3	the	the	DET
ejpam-1184	86	4	random	random	ADJ
ejpam-1184	86	5	partition	partition	NOUN
ejpam-1184	86	6	regression	regression	NOUN
ejpam-1184	86	7	clustering	cluster	VERB
ejpam-1184	86	8	.	.	PUNCT
ejpam-1184	87	1	the	the	DET
ejpam-1184	87	2	discussion	discussion	NOUN
ejpam-1184	87	3	can	can	AUX
ejpam-1184	87	4	be	be	AUX
ejpam-1184	87	5	found	find	VERB
ejpam-1184	87	6	in	in	ADP
ejpam-1184	87	7	[	[	X
ejpam-1184	87	8	16	16	NUM
ejpam-1184	87	9	,	,	PUNCT
ejpam-1184	87	10	15	15	NUM
ejpam-1184	87	11	]	]	PUNCT
ejpam-1184	87	12	among	among	ADP
ejpam-1184	87	13	others	other	NOUN
ejpam-1184	87	14	.	.	PUNCT
ejpam-1184	88	1	another	another	DET
ejpam-1184	88	2	one	one	NOUN
ejpam-1184	88	3	is	be	AUX
ejpam-1184	88	4	the	the	DET
ejpam-1184	88	5	fixed	fix	VERB
ejpam-1184	88	6	partition	partition	NOUN
ejpam-1184	88	7	regression	regression	NOUN
ejpam-1184	88	8	clustering	cluster	VERB
ejpam-1184	88	9	.	.	PUNCT
ejpam-1184	89	1	as	as	SCONJ
ejpam-1184	89	2	discussed	discuss	VERB
ejpam-1184	89	3	in	in	ADP
ejpam-1184	89	4	[	[	X
ejpam-1184	89	5	4	4	NUM
ejpam-1184	89	6	,	,	PUNCT
ejpam-1184	89	7	5	5	NUM
ejpam-1184	89	8	,	,	PUNCT
ejpam-1184	89	9	6	6	NUM
ejpam-1184	89	10	,	,	PUNCT
ejpam-1184	89	11	7	7	NUM
ejpam-1184	89	12	]	]	PUNCT
ejpam-1184	89	13	,	,	PUNCT
ejpam-1184	89	14	the	the	DET
ejpam-1184	89	15	classification	classification	NOUN
ejpam-1184	89	16	likelihood	likelihood	NOUN
ejpam-1184	89	17	model	model	NOUN
ejpam-1184	89	18	or	or	CCONJ
ejpam-1184	89	19	the	the	DET
ejpam-1184	89	20	fixed	fix	VERB
ejpam-1184	89	21	partition	partition	NOUN
ejpam-1184	89	22	regression	regression	NOUN
ejpam-1184	89	23	clustering	clustering	NOUN
ejpam-1184	89	24	model	model	NOUN
ejpam-1184	89	25	for	for	ADP
ejpam-1184	89	26	any	any	DET
ejpam-1184	89	27	partition	partition	NOUN
ejpam-1184	89	28	π	π	NOUN
ejpam-1184	89	29	=	=	SYM
ejpam-1184	89	30	(	(	PUNCT
ejpam-1184	89	31	c1	c1	PROPN
ejpam-1184	89	32	,	,	PUNCT
ejpam-1184	89	33	.	.	PUNCT
ejpam-1184	89	34	.	.	PUNCT
ejpam-1184	89	35	.	.	PUNCT
ejpam-1184	90	1	,	,	PUNCT
ejpam-1184	90	2	ck	ck	PROPN
ejpam-1184	90	3	)	)	PUNCT
ejpam-1184	90	4	of	of	ADP
ejpam-1184	90	5	o	o	PROPN
ejpam-1184	90	6	is	be	AUX
ejpam-1184	90	7	:	:	PUNCT
ejpam-1184	90	8	yj	yj	PROPN
ejpam-1184	90	9	∼	∼	PROPN
ejpam-1184	90	10	f	f	PROPN
ejpam-1184	90	11	(	(	PUNCT
ejpam-1184	90	12	·	·	PUNCT
ejpam-1184	90	13	;	;	PUNCT
ejpam-1184	90	14	β	β	X
ejpam-1184	90	15	i	i	PRON
ejpam-1184	90	16	,	,	PUNCT
ejpam-1184	90	17	σi)∼	σi)∼	PROPN
ejpam-1184	90	18	φ(x	φ(x	PROPN
ejpam-1184	90	19	′	′	NUM
ejpam-1184	90	20	jβ	jβ	PROPN
ejpam-1184	91	1	i	i	PRON
ejpam-1184	91	2	,	,	PUNCT
ejpam-1184	91	3	σi	σi	NOUN
ejpam-1184	91	4	)	)	PUNCT
ejpam-1184	91	5	for	for	ADP
ejpam-1184	91	6	all	all	DET
ejpam-1184	91	7	j	j	PROPN
ejpam-1184	91	8	∈	∈	PROPN
ejpam-1184	91	9	ci	ci	PROPN
ejpam-1184	91	10	,	,	PUNCT
ejpam-1184	91	11	i	i	PRON
ejpam-1184	91	12	=	=	NOUN
ejpam-1184	91	13	1	1	NUM
ejpam-1184	91	14	,	,	PUNCT
ejpam-1184	91	15	.	.	PUNCT
ejpam-1184	91	16	.	.	PUNCT
ejpam-1184	91	17	.	.	PUNCT
ejpam-1184	92	1	,	,	PUNCT
ejpam-1184	92	2	k.	k.	PROPN
ejpam-1184	92	3	equivalently	equivalently	PROPN
ejpam-1184	92	4	,	,	PUNCT
ejpam-1184	92	5	it	it	PRON
ejpam-1184	92	6	can	can	AUX
ejpam-1184	92	7	be	be	AUX
ejpam-1184	92	8	written	write	VERB
ejpam-1184	92	9	in	in	ADP
ejpam-1184	92	10	the	the	DET
ejpam-1184	92	11	form	form	NOUN
ejpam-1184	92	12	of	of	ADP
ejpam-1184	92	13	a	a	DET
ejpam-1184	92	14	group	group	NOUN
ejpam-1184	92	15	of	of	ADP
ejpam-1184	92	16	linear	linear	PROPN
ejpam-1184	92	17	models	model	NOUN
ejpam-1184	92	18	:	:	PUNCT
ejpam-1184	92	19	y	y	PROPN
ejpam-1184	92	20	j	j	PROPN
ejpam-1184	92	21	=	=	PUNCT
ejpam-1184	92	22	x	x	SYM
ejpam-1184	92	23	′jβ	′jβ	NOUN
ejpam-1184	92	24	i	i	PRON
ejpam-1184	92	25	+	+	CCONJ
ejpam-1184	92	26	e	e	PROPN
ejpam-1184	92	27	j	j	PROPN
ejpam-1184	92	28	,	,	PUNCT
ejpam-1184	92	29	e	e	X
ejpam-1184	92	30	j	j	X
ejpam-1184	92	31	∼	∼	NOUN
ejpam-1184	92	32	n(0,σ2	n(0,σ2	PROPN
ejpam-1184	92	33	i	i	PROPN
ejpam-1184	92	34	)	)	PUNCT
ejpam-1184	92	35	for	for	ADP
ejpam-1184	92	36	all	all	DET
ejpam-1184	92	37	j	j	PROPN
ejpam-1184	92	38	∈	∈	PROPN
ejpam-1184	92	39	ci	ci	PROPN
ejpam-1184	92	40	,	,	PUNCT
ejpam-1184	92	41	i	i	PRON
ejpam-1184	92	42	=	=	NOUN
ejpam-1184	92	43	1	1	NUM
ejpam-1184	92	44	,	,	PUNCT
ejpam-1184	92	45	.	.	PUNCT
ejpam-1184	92	46	.	.	PUNCT
ejpam-1184	92	47	.	.	PUNCT
ejpam-1184	93	1	,	,	PUNCT
ejpam-1184	93	2	k.	k.	PROPN
ejpam-1184	93	3	(	(	PUNCT
ejpam-1184	93	4	1	1	X
ejpam-1184	93	5	)	)	PUNCT
ejpam-1184	93	6	g.	g.	PROPN
ejpam-1184	93	7	qian	qian	PROPN
ejpam-1184	93	8	,	,	PUNCT
ejpam-1184	93	9	y.	y.	PROPN
ejpam-1184	93	10	wu	wu	PROPN
ejpam-1184	93	11	/	/	SYM
ejpam-1184	93	12	eur	eur	PROPN
ejpam-1184	93	13	.	.	PUNCT
ejpam-1184	94	1	j.	j.	PROPN
ejpam-1184	94	2	pure	pure	PROPN
ejpam-1184	94	3	appl	appl	PROPN
ejpam-1184	94	4	.	.	PROPN
ejpam-1184	94	5	math	math	PROPN
ejpam-1184	94	6	,	,	PUNCT
ejpam-1184	94	7	4	4	NUM
ejpam-1184	94	8	(	(	PUNCT
ejpam-1184	94	9	2011	2011	NUM
ejpam-1184	94	10	)	)	PUNCT
ejpam-1184	94	11	,	,	PUNCT
ejpam-1184	94	12	455	455	NUM
ejpam-1184	94	13	-	-	SYM
ejpam-1184	94	14	466	466	NUM
ejpam-1184	94	15	458	458	NUM
ejpam-1184	94	16	under	under	ADP
ejpam-1184	94	17	the	the	DET
ejpam-1184	94	18	fixed	fix	VERB
ejpam-1184	94	19	-	-	PUNCT
ejpam-1184	94	20	partition	partition	NOUN
ejpam-1184	94	21	model	model	NOUN
ejpam-1184	94	22	(	(	PUNCT
ejpam-1184	94	23	1	1	NUM
ejpam-1184	94	24	)	)	PUNCT
ejpam-1184	94	25	,	,	PUNCT
ejpam-1184	94	26	the	the	DET
ejpam-1184	94	27	log	log	NOUN
ejpam-1184	94	28	-	-	PUNCT
ejpam-1184	94	29	likelihood	likelihood	NOUN
ejpam-1184	94	30	function	function	NOUN
ejpam-1184	94	31	is	be	AUX
ejpam-1184	94	32	given	give	VERB
ejpam-1184	94	33	by	by	ADP
ejpam-1184	94	34	log	log	NOUN
ejpam-1184	94	35	ln(k	ln(k	NOUN
ejpam-1184	94	36	,	,	PUNCT
ejpam-1184	94	37	(	(	PUNCT
ejpam-1184	94	38	β	β	X
ejpam-1184	94	39	i	i	PROPN
ejpam-1184	94	40	,	,	PUNCT
ejpam-1184	94	41	σ	σ	PROPN
ejpam-1184	94	42	2	2	NUM
ejpam-1184	94	43	i	i	NOUN
ejpam-1184	94	44	)	)	PUNCT
ejpam-1184	94	45	i=1,	i=1,	PROPN
ejpam-1184	94	46	...	...	PUNCT
ejpam-1184	94	47	,k	,k	PUNCT
ejpam-1184	94	48	)	)	PUNCT
ejpam-1184	95	1	=	=	SYM
ejpam-1184	95	2	−	−	PROPN
ejpam-1184	95	3	1	1	NUM
ejpam-1184	95	4	2	2	NUM
ejpam-1184	95	5	k∑	k∑	NOUN
ejpam-1184	95	6	i=1	i=1	PROPN
ejpam-1184	95	7	∑	∑	ADV
ejpam-1184	95	8	j∈ci	j∈ci	PROPN
ejpam-1184	95	9	�	�	PROPN
ejpam-1184	95	10	log2π+	log2π+	PROPN
ejpam-1184	95	11	logσ2	logσ2	PROPN
ejpam-1184	96	1	i	i	PRON
ejpam-1184	96	2	+	+	CCONJ
ejpam-1184	96	3	(	(	PUNCT
ejpam-1184	96	4	y	y	PROPN
ejpam-1184	96	5	j	j	PROPN
ejpam-1184	96	6	−β	−β	PROPN
ejpam-1184	96	7	′	′	VERB
ejpam-1184	97	1	i	i	PRON
ejpam-1184	97	2	x	x	SYM
ejpam-1184	97	3	j	j	NOUN
ejpam-1184	97	4	)	)	PUNCT
ejpam-1184	97	5	2	2	NUM
ejpam-1184	97	6	σ2	σ2	PROPN
ejpam-1184	97	7	i	i	PRON
ejpam-1184	97	8	�	�	PROPN
ejpam-1184	97	9	.	.	PUNCT
ejpam-1184	98	1	(	(	PUNCT
ejpam-1184	98	2	2	2	X
ejpam-1184	98	3	)	)	PUNCT
ejpam-1184	98	4	for	for	ADP
ejpam-1184	98	5	given	give	VERB
ejpam-1184	98	6	(	(	PUNCT
ejpam-1184	98	7	β̂	β̂	ADP
ejpam-1184	98	8	i	i	PRON
ejpam-1184	98	9	,	,	PUNCT
ejpam-1184	98	10	σ̂	σ̂	PROPN
ejpam-1184	98	11	2	2	NUM
ejpam-1184	98	12	i	i	NOUN
ejpam-1184	98	13	)	)	PUNCT
ejpam-1184	98	14	i=1,	i=1,	PROPN
ejpam-1184	98	15	...	...	PUNCT
ejpam-1184	98	16	,k	,k	PUNCT
ejpam-1184	98	17	,	,	PUNCT
ejpam-1184	98	18	(	(	PUNCT
ejpam-1184	98	19	2	2	X
ejpam-1184	98	20	)	)	PUNCT
ejpam-1184	98	21	is	be	AUX
ejpam-1184	98	22	maximized	maximize	VERB
ejpam-1184	98	23	at	at	ADP
ejpam-1184	98	24	setting	set	VERB
ejpam-1184	98	25	ci	ci	NOUN
ejpam-1184	98	26	to	to	ADP
ejpam-1184	98	27	ĉi	ĉi	NOUN
ejpam-1184	98	28	=	=	SYM
ejpam-1184	98	29	arg	arg	NOUN
ejpam-1184	98	30	min	min	NOUN
ejpam-1184	99	1	i	i	PRON
ejpam-1184	99	2	log	log	VERB
ejpam-1184	99	3	σ̂2	σ̂2	PROPN
ejpam-1184	100	1	i	i	PRON
ejpam-1184	100	2	+	+	PUNCT
ejpam-1184	100	3	(	(	PUNCT
ejpam-1184	100	4	y	y	PROPN
ejpam-1184	100	5	j	j	PROPN
ejpam-1184	100	6	−	−	PROPN
ejpam-1184	101	1	β̂	β̂	ADP
ejpam-1184	101	2	′	′	NUM
ejpam-1184	102	1	ix	ix	PROPN
ejpam-1184	102	2	j	j	NOUN
ejpam-1184	102	3	)	)	PUNCT
ejpam-1184	102	4	2	2	NUM
ejpam-1184	102	5	σ̂2	σ̂2	NOUN
ejpam-1184	102	6	i	i	PRON
ejpam-1184	102	7	!	!	PUNCT
ejpam-1184	102	8	.	.	PUNCT
ejpam-1184	103	1	(	(	PUNCT
ejpam-1184	103	2	3	3	X
ejpam-1184	103	3	)	)	PUNCT
ejpam-1184	103	4	for	for	ADP
ejpam-1184	103	5	given	give	VERB
ejpam-1184	103	6	ĉi	ĉi	NOUN
ejpam-1184	103	7	,	,	PUNCT
ejpam-1184	103	8	(	(	PUNCT
ejpam-1184	103	9	2	2	X
ejpam-1184	103	10	)	)	PUNCT
ejpam-1184	103	11	is	be	AUX
ejpam-1184	103	12	the	the	DET
ejpam-1184	103	13	sum	sum	NOUN
ejpam-1184	103	14	of	of	ADP
ejpam-1184	103	15	the	the	DET
ejpam-1184	103	16	usual	usual	ADJ
ejpam-1184	103	17	log	log	NOUN
ejpam-1184	103	18	-	-	PUNCT
ejpam-1184	103	19	likelihood	likelihood	NOUN
ejpam-1184	103	20	functions	function	NOUN
ejpam-1184	103	21	for	for	ADP
ejpam-1184	103	22	homogeneous	homogeneous	ADJ
ejpam-1184	103	23	linear	linear	NOUN
ejpam-1184	103	24	regressions	regression	NOUN
ejpam-1184	103	25	within	within	ADP
ejpam-1184	103	26	clusters	cluster	NOUN
ejpam-1184	103	27	.	.	PUNCT
ejpam-1184	104	1	hence	hence	ADV
ejpam-1184	104	2	,	,	PUNCT
ejpam-1184	104	3	it	it	PRON
ejpam-1184	104	4	is	be	AUX
ejpam-1184	104	5	maximized	maximize	VERB
ejpam-1184	104	6	by	by	ADP
ejpam-1184	104	7	the	the	DET
ejpam-1184	104	8	ls	ls	PROPN
ejpam-1184	104	9	-	-	PUNCT
ejpam-1184	104	10	estimator	estimator	NOUN
ejpam-1184	104	11	β̂	β̂	PUNCT
ejpam-1184	105	1	i	i	PRON
ejpam-1184	105	2	from	from	ADP
ejpam-1184	105	3	the	the	DET
ejpam-1184	105	4	data	data	NOUN
ejpam-1184	105	5	points	point	NOUN
ejpam-1184	105	6	(	(	PUNCT
ejpam-1184	105	7	y	y	PROPN
ejpam-1184	105	8	j	j	PROPN
ejpam-1184	105	9	,	,	PUNCT
ejpam-1184	105	10	x	x	PROPN
ejpam-1184	105	11	j	j	NOUN
ejpam-1184	105	12	)	)	PUNCT
ejpam-1184	105	13	with	with	ADP
ejpam-1184	105	14	j	j	PROPN
ejpam-1184	105	15	∈	∈	PROPN
ejpam-1184	105	16	ci	ci	PROPN
ejpam-1184	105	17	and	and	CCONJ
ejpam-1184	105	18	σ̂2	σ̂2	PROPN
ejpam-1184	106	1	i	i	PRON
ejpam-1184	106	2	=	=	PUNCT
ejpam-1184	106	3	∑	∑	PUNCT
ejpam-1184	106	4	j∈ĉi	j∈ĉi	PROPN
ejpam-1184	106	5	(	(	PUNCT
ejpam-1184	106	6	y	y	PROPN
ejpam-1184	106	7	j	j	PROPN
ejpam-1184	106	8	−	−	PROPN
ejpam-1184	107	1	β̂	β̂	ADP
ejpam-1184	107	2	′	′	NUM
ejpam-1184	107	3	ix	ix	PROPN
ejpam-1184	107	4	j	j	PROPN
ejpam-1184	107	5	)	)	PUNCT
ejpam-1184	107	6	2	2	NUM
ejpam-1184	107	7	n̂i	n̂i	NOUN
ejpam-1184	107	8	,	,	PUNCT
ejpam-1184	107	9	i	i	PRON
ejpam-1184	107	10	=	=	NOUN
ejpam-1184	107	11	1	1	NUM
ejpam-1184	107	12	,	,	PUNCT
ejpam-1184	107	13	.	.	PUNCT
ejpam-1184	107	14	.	.	PUNCT
ejpam-1184	107	15	.	.	PUNCT
ejpam-1184	108	1	,	,	PUNCT
ejpam-1184	108	2	k	k	NOUN
ejpam-1184	108	3	,	,	PUNCT
ejpam-1184	108	4	(	(	PUNCT
ejpam-1184	108	5	4	4	X
ejpam-1184	108	6	)	)	PUNCT
ejpam-1184	108	7	where	where	SCONJ
ejpam-1184	108	8	n̂i	n̂i	NOUN
ejpam-1184	108	9	=	=	SYM
ejpam-1184	108	10	|ĉi|	|ĉi|	PROPN
ejpam-1184	108	11	is	be	AUX
ejpam-1184	108	12	the	the	DET
ejpam-1184	108	13	number	number	NOUN
ejpam-1184	108	14	of	of	ADP
ejpam-1184	108	15	data	datum	NOUN
ejpam-1184	108	16	points	point	NOUN
ejpam-1184	108	17	in	in	ADP
ejpam-1184	108	18	ci	ci	PROPN
ejpam-1184	108	19	.	.	PUNCT
ejpam-1184	109	1	then	then	ADV
ejpam-1184	109	2	log	log	VERB
ejpam-1184	109	3	l̂n	l̂n	NUM
ejpam-1184	109	4	is	be	AUX
ejpam-1184	109	5	monotonically	monotonically	ADV
ejpam-1184	109	6	increased	increase	VERB
ejpam-1184	109	7	if	if	SCONJ
ejpam-1184	109	8	the	the	DET
ejpam-1184	109	9	steps	step	NOUN
ejpam-1184	109	10	(	(	PUNCT
ejpam-1184	109	11	3	3	NUM
ejpam-1184	109	12	)	)	PUNCT
ejpam-1184	109	13	and	and	CCONJ
ejpam-1184	109	14	(	(	PUNCT
ejpam-1184	109	15	4	4	X
ejpam-1184	109	16	)	)	PUNCT
ejpam-1184	109	17	are	be	AUX
ejpam-1184	109	18	carried	carry	VERB
ejpam-1184	109	19	out	out	ADP
ejpam-1184	109	20	alternately	alternately	ADV
ejpam-1184	109	21	.	.	PUNCT
ejpam-1184	110	1	this	this	DET
ejpam-1184	110	2	algorithm	algorithm	NOUN
ejpam-1184	110	3	leads	lead	VERB
ejpam-1184	110	4	to	to	ADP
ejpam-1184	110	5	a	a	DET
ejpam-1184	110	6	local	local	ADJ
ejpam-1184	110	7	maximum	maximum	NOUN
ejpam-1184	110	8	(	(	PUNCT
ejpam-1184	110	9	one	one	PRON
ejpam-1184	110	10	would	would	AUX
ejpam-1184	110	11	hope	hope	VERB
ejpam-1184	110	12	,	,	PUNCT
ejpam-1184	110	13	to	to	ADP
ejpam-1184	110	14	an	an	DET
ejpam-1184	110	15	approximation	approximation	NOUN
ejpam-1184	110	16	of	of	ADP
ejpam-1184	110	17	the	the	DET
ejpam-1184	110	18	global	global	ADJ
ejpam-1184	110	19	maximum	maximum	NOUN
ejpam-1184	110	20	by	by	ADP
ejpam-1184	110	21	proper	proper	ADJ
ejpam-1184	110	22	initialization	initialization	NOUN
ejpam-1184	110	23	)	)	PUNCT
ejpam-1184	110	24	in	in	ADP
ejpam-1184	110	25	finitely	finitely	ADV
ejpam-1184	110	26	many	many	ADJ
ejpam-1184	110	27	steps	step	NOUN
ejpam-1184	110	28	.	.	PUNCT
ejpam-1184	111	1	the	the	DET
ejpam-1184	111	2	fixed	fix	VERB
ejpam-1184	111	3	partition	partition	NOUN
ejpam-1184	111	4	approach	approach	NOUN
ejpam-1184	111	5	has	have	VERB
ejpam-1184	111	6	a	a	DET
ejpam-1184	111	7	particular	particular	ADJ
ejpam-1184	111	8	advantage	advantage	NOUN
ejpam-1184	111	9	over	over	ADP
ejpam-1184	111	10	the	the	DET
ejpam-1184	111	11	random	random	ADJ
ejpam-1184	111	12	partitioning	partitioning	NOUN
ejpam-1184	111	13	in	in	ADP
ejpam-1184	111	14	the	the	DET
ejpam-1184	111	15	context	context	NOUN
ejpam-1184	111	16	of	of	ADP
ejpam-1184	111	17	regression	regression	NOUN
ejpam-1184	111	18	clustering	cluster	VERB
ejpam-1184	111	19	.	.	PUNCT
ejpam-1184	112	1	as	as	SCONJ
ejpam-1184	112	2	observed	observe	VERB
ejpam-1184	112	3	by	by	ADP
ejpam-1184	112	4	hennig	hennig	PROPN
ejpam-1184	112	5	[	[	X
ejpam-1184	112	6	1	1	NUM
ejpam-1184	112	7	]	]	PUNCT
ejpam-1184	112	8	,	,	PUNCT
ejpam-1184	112	9	the	the	DET
ejpam-1184	112	10	mixture	mixture	NOUN
ejpam-1184	112	11	model	model	NOUN
ejpam-1184	112	12	presumes	presume	VERB
ejpam-1184	112	13	implicitly	implicitly	ADV
ejpam-1184	112	14	an	an	DET
ejpam-1184	112	15	assignment	assignment	ADJ
ejpam-1184	112	16	independence	independence	NOUN
ejpam-1184	112	17	of	of	ADP
ejpam-1184	112	18	each	each	DET
ejpam-1184	112	19	object	object	NOUN
ejpam-1184	112	20	to	to	ADP
ejpam-1184	112	21	clusters	cluster	NOUN
ejpam-1184	112	22	with	with	ADP
ejpam-1184	112	23	respect	respect	NOUN
ejpam-1184	112	24	to	to	ADP
ejpam-1184	112	25	the	the	DET
ejpam-1184	112	26	covariate	covariate	ADJ
ejpam-1184	112	27	vectors	vector	NOUN
ejpam-1184	112	28	x	x	PUNCT
ejpam-1184	112	29	j.	j.	PROPN
ejpam-1184	112	30	that	that	ADV
ejpam-1184	112	31	is	be	AUX
ejpam-1184	112	32	,	,	PUNCT
ejpam-1184	112	33	the	the	DET
ejpam-1184	112	34	clusters	cluster	NOUN
ejpam-1184	112	35	keep	keep	VERB
ejpam-1184	112	36	the	the	DET
ejpam-1184	112	37	same	same	ADJ
ejpam-1184	112	38	conditional	conditional	ADJ
ejpam-1184	112	39	proportions	proportion	NOUN
ejpam-1184	112	40	πi	πi	ADV
ejpam-1184	112	41	,	,	PUNCT
ejpam-1184	112	42	i	i	PRON
ejpam-1184	112	43	=	=	NOUN
ejpam-1184	112	44	1	1	NUM
ejpam-1184	112	45	,	,	PUNCT
ejpam-1184	112	46	.	.	PUNCT
ejpam-1184	112	47	.	.	PUNCT
ejpam-1184	113	1	.	.	PUNCT
ejpam-1184	114	1	,	,	PUNCT
ejpam-1184	114	2	k	k	PROPN
ejpam-1184	114	3	for	for	ADP
ejpam-1184	114	4	every	every	DET
ejpam-1184	114	5	fixed	fix	VERB
ejpam-1184	114	6	covariate	covariate	ADJ
ejpam-1184	114	7	vector	vector	NOUN
ejpam-1184	114	8	x	x	X
ejpam-1184	114	9	j	j	PROPN
ejpam-1184	114	10	.	.	PUNCT
ejpam-1184	115	1	in	in	ADP
ejpam-1184	115	2	other	other	ADJ
ejpam-1184	115	3	words	word	NOUN
ejpam-1184	115	4	,	,	PUNCT
ejpam-1184	115	5	the	the	DET
ejpam-1184	115	6	probability	probability	NOUN
ejpam-1184	115	7	of	of	ADP
ejpam-1184	115	8	a	a	DET
ejpam-1184	115	9	point	point	NOUN
ejpam-1184	115	10	(	(	PUNCT
ejpam-1184	115	11	y	y	PROPN
ejpam-1184	115	12	j	j	PROPN
ejpam-1184	115	13	,	,	PUNCT
ejpam-1184	115	14	x	x	PROPN
ejpam-1184	115	15	′j	′j	NOUN
ejpam-1184	115	16	)	)	PUNCT
ejpam-1184	115	17	to	to	PART
ejpam-1184	115	18	be	be	AUX
ejpam-1184	115	19	generated	generate	VERB
ejpam-1184	115	20	by	by	ADP
ejpam-1184	115	21	cluster	cluster	NOUN
ejpam-1184	115	22	i	i	PRON
ejpam-1184	115	23	is	be	AUX
ejpam-1184	115	24	independent	independent	ADJ
ejpam-1184	115	25	of	of	ADP
ejpam-1184	115	26	x	x	X
ejpam-1184	115	27	and	and	CCONJ
ejpam-1184	115	28	j.	j.	PROPN
ejpam-1184	115	29	this	this	PRON
ejpam-1184	115	30	is	be	AUX
ejpam-1184	115	31	generally	generally	ADV
ejpam-1184	115	32	not	not	PART
ejpam-1184	115	33	true	true	ADJ
ejpam-1184	115	34	as	as	SCONJ
ejpam-1184	115	35	shown	show	VERB
ejpam-1184	115	36	in	in	ADP
ejpam-1184	115	37	figure	figure	NOUN
ejpam-1184	115	38	1	1	NUM
ejpam-1184	115	39	,	,	PUNCT
ejpam-1184	115	40	which	which	PRON
ejpam-1184	115	41	is	be	AUX
ejpam-1184	115	42	adapted	adapt	VERB
ejpam-1184	115	43	from	from	ADP
ejpam-1184	115	44	[	[	X
ejpam-1184	115	45	1	1	NUM
ejpam-1184	115	46	]	]	PUNCT
ejpam-1184	115	47	.	.	PUNCT
ejpam-1184	116	1	on	on	ADP
ejpam-1184	116	2	the	the	DET
ejpam-1184	116	3	other	other	ADJ
ejpam-1184	116	4	hand	hand	NOUN
ejpam-1184	116	5	,	,	PUNCT
ejpam-1184	116	6	the	the	DET
ejpam-1184	116	7	fixed	fix	VERB
ejpam-1184	116	8	partition	partition	NOUN
ejpam-1184	116	9	model	model	NOUN
ejpam-1184	116	10	(	(	PUNCT
ejpam-1184	116	11	1	1	X
ejpam-1184	116	12	)	)	PUNCT
ejpam-1184	116	13	supposes	suppose	VERB
ejpam-1184	116	14	that	that	SCONJ
ejpam-1184	116	15	the	the	DET
ejpam-1184	116	16	cluster	cluster	NOUN
ejpam-1184	116	17	membership	membership	NOUN
ejpam-1184	116	18	of	of	ADP
ejpam-1184	116	19	each	each	DET
ejpam-1184	116	20	object	object	NOUN
ejpam-1184	116	21	or	or	CCONJ
ejpam-1184	116	22	cluster	cluster	NOUN
ejpam-1184	116	23	labels	label	NOUN
ejpam-1184	116	24	are	be	AUX
ejpam-1184	116	25	explicitly	explicitly	ADV
ejpam-1184	116	26	parametrized	parametrized	ADJ
ejpam-1184	116	27	and	and	CCONJ
ejpam-1184	116	28	are	be	AUX
ejpam-1184	116	29	determined	determine	VERB
ejpam-1184	116	30	by	by	ADP
ejpam-1184	116	31	the	the	DET
ejpam-1184	116	32	estimation	estimation	NOUN
ejpam-1184	116	33	of	of	ADP
ejpam-1184	116	34	bβ	bβ	NOUN
ejpam-1184	116	35	i	i	PRON
ejpam-1184	116	36	and	and	CCONJ
ejpam-1184	116	37	σ̂2	σ̂2	PROPN
ejpam-1184	116	38	i	i	PRON
ejpam-1184	116	39	through	through	ADP
ejpam-1184	116	40	the	the	DET
ejpam-1184	116	41	points	point	NOUN
ejpam-1184	116	42	(	(	PUNCT
ejpam-1184	116	43	y	y	PROPN
ejpam-1184	116	44	j	j	PROPN
ejpam-1184	116	45	,	,	PUNCT
ejpam-1184	116	46	x	x	ADJ
ejpam-1184	116	47	′j	′j	NOUN
ejpam-1184	116	48	)	)	PUNCT
ejpam-1184	116	49	(	(	PUNCT
ejpam-1184	116	50	j	j	PROPN
ejpam-1184	116	51	∈	∈	PROPN
ejpam-1184	116	52	ci	ci	PROPN
ejpam-1184	116	53	)	)	PUNCT
ejpam-1184	116	54	.	.	PUNCT
ejpam-1184	117	1	hence	hence	ADV
ejpam-1184	117	2	the	the	DET
ejpam-1184	117	3	fixed	fixed	ADJ
ejpam-1184	117	4	partition	partition	NOUN
ejpam-1184	117	5	model	model	NOUN
ejpam-1184	117	6	does	do	AUX
ejpam-1184	117	7	make	make	VERB
ejpam-1184	117	8	allowance	allowance	NOUN
ejpam-1184	117	9	of	of	ADP
ejpam-1184	117	10	possible	possible	ADJ
ejpam-1184	117	11	assignment	assignment	ADJ
ejpam-1184	117	12	dependence	dependence	NOUN
ejpam-1184	117	13	between	between	ADP
ejpam-1184	117	14	the	the	DET
ejpam-1184	117	15	j	j	PROPN
ejpam-1184	117	16	-	-	PUNCT
ejpam-1184	117	17	th	th	VERB
ejpam-1184	117	18	object	object	NOUN
ejpam-1184	117	19	and	and	CCONJ
ejpam-1184	117	20	the	the	DET
ejpam-1184	117	21	associated	associated	ADJ
ejpam-1184	117	22	covariate	covariate	NOUN
ejpam-1184	117	23	x	x	PROPN
ejpam-1184	117	24	j.	j.	PROPN
ejpam-1184	117	25	3	3	PROPN
ejpam-1184	117	26	.	.	PUNCT
ejpam-1184	117	27	procedures	procedure	NOUN
ejpam-1184	117	28	for	for	ADP
ejpam-1184	117	29	estimating	estimate	VERB
ejpam-1184	117	30	the	the	DET
ejpam-1184	117	31	parameters	parameter	NOUN
ejpam-1184	117	32	and	and	CCONJ
ejpam-1184	117	33	the	the	DET
ejpam-1184	117	34	number	number	NOUN
ejpam-1184	117	35	of	of	ADP
ejpam-1184	117	36	clusters	cluster	NOUN
ejpam-1184	117	37	in	in	ADP
ejpam-1184	117	38	regression	regression	NOUN
ejpam-1184	117	39	clustering	cluster	VERB
ejpam-1184	117	40	suppose	suppose	VERB
ejpam-1184	117	41	that	that	SCONJ
ejpam-1184	117	42	we	we	PRON
ejpam-1184	117	43	have	have	VERB
ejpam-1184	117	44	n	n	PRON
ejpam-1184	117	45	objects	object	VERB
ejpam-1184	117	46	o	o	X
ejpam-1184	117	47	(	(	PUNCT
ejpam-1184	117	48	n	n	CCONJ
ejpam-1184	117	49	)	)	PUNCT
ejpam-1184	117	50	=	=	SYM
ejpam-1184	117	51	{	{	PUNCT
ejpam-1184	117	52	1,2	1,2	NUM
ejpam-1184	117	53	,	,	PUNCT
ejpam-1184	117	54	.	.	PUNCT
ejpam-1184	117	55	.	.	PUNCT
ejpam-1184	118	1	.	.	PUNCT
ejpam-1184	119	1	,	,	PUNCT
ejpam-1184	119	2	n	n	CCONJ
ejpam-1184	119	3	}	}	PUNCT
ejpam-1184	119	4	with	with	ADP
ejpam-1184	119	5	the	the	DET
ejpam-1184	119	6	associated	associated	ADJ
ejpam-1184	119	7	data	data	NOUN
ejpam-1184	119	8	points	point	NOUN
ejpam-1184	119	9	(	(	PUNCT
ejpam-1184	119	10	x	x	SYM
ejpam-1184	119	11	1	1	NUM
ejpam-1184	119	12	,	,	PUNCT
ejpam-1184	119	13	y1	y1	NOUN
ejpam-1184	119	14	)	)	PUNCT
ejpam-1184	119	15	,	,	PUNCT
ejpam-1184	119	16	.	.	PUNCT
ejpam-1184	120	1	.	.	PUNCT
ejpam-1184	121	1	.	.	PUNCT
ejpam-1184	122	1	,	,	PUNCT
ejpam-1184	122	2	(	(	PUNCT
ejpam-1184	122	3	x	x	SYM
ejpam-1184	122	4	n	n	CCONJ
ejpam-1184	122	5	,	,	PUNCT
ejpam-1184	122	6	yn	yn	PROPN
ejpam-1184	122	7	)	)	PUNCT
ejpam-1184	122	8	.	.	PUNCT
ejpam-1184	123	1	here	here	ADV
ejpam-1184	123	2	x	x	PUNCT
ejpam-1184	123	3	j	j	PROPN
ejpam-1184	123	4	∈	∈	PROPN
ejpam-1184	123	5	r	r	NOUN
ejpam-1184	123	6	p	p	NOUN
ejpam-1184	123	7	is	be	AUX
ejpam-1184	123	8	a	a	DET
ejpam-1184	123	9	non	non	ADJ
ejpam-1184	123	10	-	-	ADJ
ejpam-1184	123	11	random	random	ADJ
ejpam-1184	123	12	explanatory	explanatory	ADJ
ejpam-1184	123	13	p	p	NOUN
ejpam-1184	123	14	-	-	PUNCT
ejpam-1184	123	15	vector	vector	NOUN
ejpam-1184	123	16	and	and	CCONJ
ejpam-1184	124	1	y	y	PROPN
ejpam-1184	124	2	j	j	PROPN
ejpam-1184	124	3	∈	∈	PROPN
ejpam-1184	124	4	r	r	NOUN
ejpam-1184	124	5	is	be	AUX
ejpam-1184	124	6	a	a	DET
ejpam-1184	124	7	random	random	ADJ
ejpam-1184	124	8	dependent	dependent	ADJ
ejpam-1184	124	9	variable	variable	NOUN
ejpam-1184	124	10	for	for	ADP
ejpam-1184	124	11	the	the	DET
ejpam-1184	124	12	j	j	PROPN
ejpam-1184	124	13	-	-	PUNCT
ejpam-1184	124	14	th	th	VERB
ejpam-1184	124	15	object	object	NOUN
ejpam-1184	124	16	(	(	PUNCT
ejpam-1184	124	17	j	j	NOUN
ejpam-1184	124	18	=	=	SYM
ejpam-1184	124	19	1,2	1,2	NUM
ejpam-1184	124	20	,	,	PUNCT
ejpam-1184	124	21	.	.	PUNCT
ejpam-1184	124	22	.	.	PUNCT
ejpam-1184	125	1	.	.	PUNCT
ejpam-1184	125	2	,	,	PUNCT
ejpam-1184	125	3	n	n	CCONJ
ejpam-1184	125	4	)	)	PUNCT
ejpam-1184	125	5	.	.	PUNCT
ejpam-1184	126	1	these	these	DET
ejpam-1184	126	2	n	n	NUM
ejpam-1184	126	3	objects	object	NOUN
ejpam-1184	126	4	are	be	AUX
ejpam-1184	126	5	assumed	assume	VERB
ejpam-1184	126	6	to	to	PART
ejpam-1184	126	7	be	be	AUX
ejpam-1184	126	8	a	a	DET
ejpam-1184	126	9	random	random	ADJ
ejpam-1184	126	10	sample	sample	NOUN
ejpam-1184	126	11	coming	come	VERB
ejpam-1184	126	12	from	from	ADP
ejpam-1184	126	13	a	a	DET
ejpam-1184	126	14	structured	structured	ADJ
ejpam-1184	126	15	population	population	NOUN
ejpam-1184	126	16	,	,	PUNCT
ejpam-1184	126	17	which	which	PRON
ejpam-1184	126	18	consists	consist	VERB
ejpam-1184	126	19	of	of	ADP
ejpam-1184	126	20	a	a	DET
ejpam-1184	126	21	fixed	fix	VERB
ejpam-1184	126	22	(	(	PUNCT
ejpam-1184	126	23	but	but	CCONJ
ejpam-1184	126	24	unknown	unknown	ADJ
ejpam-1184	126	25	)	)	PUNCT
ejpam-1184	126	26	number	number	NOUN
ejpam-1184	126	27	,	,	PUNCT
ejpam-1184	126	28	say	say	VERB
ejpam-1184	126	29	k0	k0	PROPN
ejpam-1184	126	30	,	,	PUNCT
ejpam-1184	126	31	of	of	ADP
ejpam-1184	126	32	sub	sub	NOUN
ejpam-1184	126	33	-	-	NOUN
ejpam-1184	126	34	populations	population	NOUN
ejpam-1184	126	35	each	each	PRON
ejpam-1184	126	36	of	of	ADP
ejpam-1184	126	37	which	which	PRON
ejpam-1184	126	38	is	be	AUX
ejpam-1184	126	39	characterized	characterize	VERB
ejpam-1184	126	40	by	by	ADP
ejpam-1184	126	41	a	a	DET
ejpam-1184	126	42	regression	regression	NOUN
ejpam-1184	126	43	hyperplane	hyperplane	NOUN
ejpam-1184	126	44	with	with	ADP
ejpam-1184	126	45	class	class	NOUN
ejpam-1184	126	46	-	-	PUNCT
ejpam-1184	126	47	specific	specific	ADJ
ejpam-1184	126	48	unknown	unknown	ADJ
ejpam-1184	126	49	parameters	parameter	NOUN
ejpam-1184	126	50	.	.	PUNCT
ejpam-1184	127	1	therefore	therefore	ADV
ejpam-1184	127	2	for	for	ADP
ejpam-1184	127	3	the	the	DET
ejpam-1184	127	4	n	n	PRON
ejpam-1184	127	5	observations	observation	NOUN
ejpam-1184	127	6	from	from	ADP
ejpam-1184	127	7	g.	g.	PROPN
ejpam-1184	127	8	qian	qian	PROPN
ejpam-1184	127	9	,	,	PUNCT
ejpam-1184	127	10	y.	y.	PROPN
ejpam-1184	127	11	wu	wu	PROPN
ejpam-1184	127	12	/	/	SYM
ejpam-1184	127	13	eur	eur	PROPN
ejpam-1184	127	14	.	.	PUNCT
ejpam-1184	128	1	j.	j.	PROPN
ejpam-1184	128	2	pure	pure	PROPN
ejpam-1184	128	3	appl	appl	PROPN
ejpam-1184	128	4	.	.	PROPN
ejpam-1184	128	5	math	math	PROPN
ejpam-1184	128	6	,	,	PUNCT
ejpam-1184	128	7	4	4	NUM
ejpam-1184	128	8	(	(	PUNCT
ejpam-1184	128	9	2011	2011	NUM
ejpam-1184	128	10	)	)	PUNCT
ejpam-1184	128	11	,	,	PUNCT
ejpam-1184	128	12	455	455	NUM
ejpam-1184	128	13	-	-	SYM
ejpam-1184	128	14	466	466	NUM
ejpam-1184	128	15	459	459	NUM
ejpam-1184	128	16	x	x	SYM
ejpam-1184	128	17	y	y	NOUN
ejpam-1184	128	18	x	x	PUNCT
ejpam-1184	128	19	y	y	NOUN
ejpam-1184	128	20	figure	figure	NOUN
ejpam-1184	128	21	1	1	NUM
ejpam-1184	128	22	:	:	PUNCT
ejpam-1184	128	23	assignment	assignment	NOUN
ejpam-1184	128	24	independence	independence	NOUN
ejpam-1184	128	25	–	–	PUNCT
ejpam-1184	128	26	assignment	assignment	NOUN
ejpam-1184	128	27	dependence	dependence	NOUN
ejpam-1184	128	28	this	this	DET
ejpam-1184	128	29	population	population	NOUN
ejpam-1184	129	1	,	,	PUNCT
ejpam-1184	129	2	there	there	PRON
ejpam-1184	129	3	exists	exist	VERB
ejpam-1184	129	4	an	an	DET
ejpam-1184	129	5	underlying	underlie	VERB
ejpam-1184	129	6	partition	partition	NOUN
ejpam-1184	129	7	π	π	X
ejpam-1184	129	8	(	(	PUNCT
ejpam-1184	129	9	n	n	CCONJ
ejpam-1184	129	10	)	)	PUNCT
ejpam-1184	129	11	k0	k0	PROPN
ejpam-1184	129	12	=	=	PUNCT
ejpam-1184	129	13	{	{	PUNCT
ejpam-1184	129	14	o	o	X
ejpam-1184	129	15	(	(	PUNCT
ejpam-1184	129	16	n	n	CCONJ
ejpam-1184	129	17	)	)	PUNCT
ejpam-1184	129	18	1	1	NUM
ejpam-1184	129	19	,	,	PUNCT
ejpam-1184	129	20	.	.	PUNCT
ejpam-1184	129	21	.	.	PUNCT
ejpam-1184	130	1	.	.	PUNCT
ejpam-1184	131	1	,	,	PUNCT
ejpam-1184	131	2	o	o	X
ejpam-1184	131	3	(	(	PUNCT
ejpam-1184	131	4	n	n	CCONJ
ejpam-1184	131	5	)	)	PUNCT
ejpam-1184	131	6	k0	k0	PROPN
ejpam-1184	131	7	}	}	PUNCT
ejpam-1184	131	8	,	,	PUNCT
ejpam-1184	131	9	and	and	CCONJ
ejpam-1184	131	10	each	each	DET
ejpam-1184	131	11	cluster	cluster	NOUN
ejpam-1184	131	12	o	o	NOUN
ejpam-1184	131	13	(	(	PUNCT
ejpam-1184	131	14	n	n	CCONJ
ejpam-1184	131	15	)	)	PUNCT
ejpam-1184	131	16	i	i	PRON
ejpam-1184	131	17	¬	¬	VERB
ejpam-1184	131	18	{	{	PUNCT
ejpam-1184	131	19	i1	i1	PROPN
ejpam-1184	131	20	,	,	PUNCT
ejpam-1184	131	21	.	.	PUNCT
ejpam-1184	131	22	.	.	PUNCT
ejpam-1184	132	1	.	.	PUNCT
ejpam-1184	133	1	,	,	PUNCT
ejpam-1184	133	2	ini	ini	PROPN
ejpam-1184	133	3	}	}	PUNCT
ejpam-1184	133	4	⊆	⊆	NUM
ejpam-1184	133	5	o	o	NOUN
ejpam-1184	133	6	(	(	PUNCT
ejpam-1184	133	7	n	n	CCONJ
ejpam-1184	133	8	)	)	PUNCT
ejpam-1184	133	9	is	be	AUX
ejpam-1184	133	10	represented	represent	VERB
ejpam-1184	133	11	by	by	ADP
ejpam-1184	133	12	yoi	yoi	PROPN
ejpam-1184	133	13	=	=	SYM
ejpam-1184	134	1	xoi	xoi	X
ejpam-1184	134	2	β0i	β0i	SYM
ejpam-1184	134	3	+	+	NUM
ejpam-1184	134	4	eoi	eoi	X
ejpam-1184	134	5	,	,	PUNCT
ejpam-1184	134	6	eoi	eoi	X
ejpam-1184	134	7	∼	∼	NOUN
ejpam-1184	134	8	n(0,σ2	n(0,σ2	NOUN
ejpam-1184	134	9	i	i	PROPN
ejpam-1184	134	10	ini	ini	PROPN
ejpam-1184	134	11	)	)	PUNCT
ejpam-1184	134	12	,	,	PUNCT
ejpam-1184	134	13	(	(	PUNCT
ejpam-1184	134	14	5	5	X
ejpam-1184	134	15	)	)	PUNCT
ejpam-1184	134	16	where	where	SCONJ
ejpam-1184	134	17	yoi	yoi	PROPN
ejpam-1184	134	18	=	=	SYM
ejpam-1184	134	19	(	(	PUNCT
ejpam-1184	134	20	yi1	yi1	INTJ
ejpam-1184	134	21	,	,	PUNCT
ejpam-1184	134	22	.	.	PUNCT
ejpam-1184	134	23	.	.	PUNCT
ejpam-1184	134	24	.	.	PUNCT
ejpam-1184	135	1	,	,	PUNCT
ejpam-1184	135	2	yini	yini	NOUN
ejpam-1184	135	3	)	)	PUNCT
ejpam-1184	135	4	′	′	PROPN
ejpam-1184	135	5	,	,	PUNCT
ejpam-1184	135	6	xoi	xoi	X
ejpam-1184	136	1	=	=	PUNCT
ejpam-1184	136	2	(	(	PUNCT
ejpam-1184	136	3	x	x	PROPN
ejpam-1184	136	4	i1	i1	PROPN
ejpam-1184	136	5	,	,	PUNCT
ejpam-1184	136	6	.	.	PUNCT
ejpam-1184	136	7	.	.	PUNCT
ejpam-1184	136	8	.	.	PUNCT
ejpam-1184	137	1	,	,	PUNCT
ejpam-1184	137	2	x	x	X
ejpam-1184	137	3	ini	ini	PROPN
ejpam-1184	137	4	)	)	PUNCT
ejpam-1184	137	5	′	′	NUM
ejpam-1184	137	6	is	be	AUX
ejpam-1184	137	7	an	an	DET
ejpam-1184	137	8	ni×p	ni×p	NOUN
ejpam-1184	137	9	design	design	NOUN
ejpam-1184	137	10	matrix	matrix	NOUN
ejpam-1184	137	11	in	in	ADP
ejpam-1184	137	12	the	the	DET
ejpam-1184	137	13	cluster	cluster	NOUN
ejpam-1184	137	14	oi	oi	NOUN
ejpam-1184	137	15	,	,	PUNCT
ejpam-1184	137	16	eoi	eoi	PROPN
ejpam-1184	137	17	is	be	AUX
ejpam-1184	137	18	an	an	DET
ejpam-1184	137	19	ni	ni	NOUN
ejpam-1184	137	20	-	-	NOUN
ejpam-1184	137	21	vector	vector	NOUN
ejpam-1184	137	22	of	of	ADP
ejpam-1184	137	23	random	random	ADJ
ejpam-1184	137	24	errors	error	NOUN
ejpam-1184	137	25	,	,	PUNCT
ejpam-1184	137	26	ini	ini	PROPN
ejpam-1184	137	27	is	be	AUX
ejpam-1184	137	28	an	an	DET
ejpam-1184	137	29	ni×ni	ni×ni	NOUN
ejpam-1184	137	30	identity	identity	NOUN
ejpam-1184	137	31	matrix	matrix	NOUN
ejpam-1184	137	32	,	,	PUNCT
ejpam-1184	137	33	and	and	CCONJ
ejpam-1184	137	34	ni	ni	PROPN
ejpam-1184	137	35	=	=	PUNCT
ejpam-1184	137	36	|oi|	|oi|	PROPN
ejpam-1184	137	37	for	for	ADP
ejpam-1184	137	38	i	i	PROPN
ejpam-1184	137	39	=	=	NOUN
ejpam-1184	137	40	1	1	NUM
ejpam-1184	137	41	,	,	PUNCT
ejpam-1184	137	42	.	.	PUNCT
ejpam-1184	137	43	.	.	PUNCT
ejpam-1184	138	1	.	.	PUNCT
ejpam-1184	139	1	,	,	PUNCT
ejpam-1184	139	2	k0	k0	PROPN
ejpam-1184	139	3	.	.	PUNCT
ejpam-1184	140	1	here	here	ADV
ejpam-1184	140	2	,	,	PUNCT
ejpam-1184	140	3	(	(	PUNCT
ejpam-1184	140	4	β	β	X
ejpam-1184	140	5	′0i	′0i	NUM
ejpam-1184	140	6	,	,	PUNCT
ejpam-1184	140	7	σi	σi	NOUN
ejpam-1184	140	8	)	)	PUNCT
ejpam-1184	140	9	′	′	NUM
ejpam-1184	140	10	∈	∈	PROPN
ejpam-1184	140	11	rp×r+	rp×r+	NOUN
ejpam-1184	140	12	,	,	PUNCT
ejpam-1184	140	13	1≤	1≤	NUM
ejpam-1184	140	14	i	i	PROPN
ejpam-1184	140	15	≤	≤	PROPN
ejpam-1184	140	16	k0	k0	PROPN
ejpam-1184	140	17	,	,	PUNCT
ejpam-1184	140	18	are	be	AUX
ejpam-1184	140	19	k0	k0	PROPN
ejpam-1184	140	20	unknown	unknown	ADJ
ejpam-1184	140	21	parameter	parameter	NOUN
ejpam-1184	140	22	vectors	vector	NOUN
ejpam-1184	140	23	.	.	PUNCT
ejpam-1184	141	1	β0i	β0i	PUNCT
ejpam-1184	141	2	,	,	PUNCT
ejpam-1184	141	3	1≤	1≤	NUM
ejpam-1184	141	4	i	i	PROPN
ejpam-1184	141	5	≤	≤	PROPN
ejpam-1184	141	6	k0	k0	PROPN
ejpam-1184	141	7	,	,	PUNCT
ejpam-1184	141	8	are	be	AUX
ejpam-1184	141	9	assumed	assume	VERB
ejpam-1184	141	10	to	to	PART
ejpam-1184	141	11	be	be	AUX
ejpam-1184	141	12	distinct	distinct	ADJ
ejpam-1184	141	13	from	from	ADP
ejpam-1184	141	14	one	one	NUM
ejpam-1184	141	15	another	another	DET
ejpam-1184	141	16	.	.	PUNCT
ejpam-1184	142	1	it	it	PRON
ejpam-1184	142	2	is	be	AUX
ejpam-1184	142	3	clear	clear	ADJ
ejpam-1184	142	4	that	that	SCONJ
ejpam-1184	142	5	n=	n=	ADJ
ejpam-1184	142	6	n1	n1	PROPN
ejpam-1184	142	7	+	+	X
ejpam-1184	142	8	.	.	PUNCT
ejpam-1184	142	9	.	.	PUNCT
ejpam-1184	143	1	.+nk0	.+nk0	PROPN
ejpam-1184	143	2	.	.	PUNCT
ejpam-1184	144	1	in	in	ADP
ejpam-1184	144	2	the	the	DET
ejpam-1184	144	3	following	following	NOUN
ejpam-1184	144	4	,	,	PUNCT
ejpam-1184	144	5	we	we	PRON
ejpam-1184	144	6	assume	assume	VERB
ejpam-1184	144	7	that	that	SCONJ
ejpam-1184	144	8	k0	k0	PROPN
ejpam-1184	144	9	≤	≤	PROPN
ejpam-1184	144	10	k	k	PROPN
ejpam-1184	144	11	,	,	PUNCT
ejpam-1184	144	12	where	where	SCONJ
ejpam-1184	144	13	k	k	PROPN
ejpam-1184	144	14	is	be	AUX
ejpam-1184	144	15	a	a	DET
ejpam-1184	144	16	known	know	VERB
ejpam-1184	144	17	positive	positive	ADJ
ejpam-1184	144	18	integer	integer	NOUN
ejpam-1184	144	19	.	.	PUNCT
ejpam-1184	145	1	note	note	VERB
ejpam-1184	145	2	that	that	SCONJ
ejpam-1184	145	3	in	in	ADP
ejpam-1184	145	4	(	(	PUNCT
ejpam-1184	145	5	5	5	X
ejpam-1184	145	6	)	)	PUNCT
ejpam-1184	145	7	we	we	PRON
ejpam-1184	145	8	have	have	AUX
ejpam-1184	145	9	suppressed	suppress	VERB
ejpam-1184	145	10	the	the	DET
ejpam-1184	145	11	n	n	NOUN
ejpam-1184	145	12	in	in	ADP
ejpam-1184	145	13	o	o	PROPN
ejpam-1184	145	14	(	(	PUNCT
ejpam-1184	145	15	n	n	CCONJ
ejpam-1184	145	16	)	)	PUNCT
ejpam-1184	145	17	i	i	PRON
ejpam-1184	145	18	for	for	ADP
ejpam-1184	145	19	convenience	convenience	NOUN
ejpam-1184	145	20	.	.	PUNCT
ejpam-1184	146	1	the	the	DET
ejpam-1184	146	2	objective	objective	NOUN
ejpam-1184	146	3	is	be	AUX
ejpam-1184	146	4	to	to	PART
ejpam-1184	146	5	reconstruct	reconstruct	VERB
ejpam-1184	146	6	the	the	DET
ejpam-1184	146	7	underlying	underlie	VERB
ejpam-1184	146	8	structure	structure	NOUN
ejpam-1184	146	9	(	(	PUNCT
ejpam-1184	146	10	5	5	NUM
ejpam-1184	146	11	)	)	PUNCT
ejpam-1184	146	12	from	from	ADP
ejpam-1184	146	13	the	the	DET
ejpam-1184	146	14	observed	observe	VERB
ejpam-1184	146	15	data	datum	NOUN
ejpam-1184	146	16	by	by	ADP
ejpam-1184	146	17	estimating	estimate	VERB
ejpam-1184	146	18	the	the	DET
ejpam-1184	146	19	number	number	NOUN
ejpam-1184	146	20	of	of	ADP
ejpam-1184	146	21	clusters	cluster	NOUN
ejpam-1184	146	22	k0	k0	PROPN
ejpam-1184	146	23	and	and	CCONJ
ejpam-1184	146	24	then	then	ADV
ejpam-1184	146	25	classifying	classify	VERB
ejpam-1184	146	26	the	the	DET
ejpam-1184	146	27	data	datum	NOUN
ejpam-1184	146	28	and	and	CCONJ
ejpam-1184	146	29	estimating	estimate	VERB
ejpam-1184	146	30	the	the	DET
ejpam-1184	146	31	classspecific	classspecific	NOUN
ejpam-1184	146	32	parameters	parameter	NOUN
ejpam-1184	146	33	accordingly	accordingly	ADV
ejpam-1184	146	34	.	.	PUNCT
ejpam-1184	147	1	what	what	PRON
ejpam-1184	147	2	can	can	AUX
ejpam-1184	147	3	be	be	AUX
ejpam-1184	147	4	done	do	VERB
ejpam-1184	147	5	in	in	ADP
ejpam-1184	147	6	practice	practice	NOUN
ejpam-1184	147	7	,	,	PUNCT
ejpam-1184	147	8	however	however	ADV
ejpam-1184	147	9	,	,	PUNCT
ejpam-1184	147	10	is	be	AUX
ejpam-1184	147	11	to	to	PART
ejpam-1184	147	12	first	first	ADV
ejpam-1184	147	13	consider	consider	VERB
ejpam-1184	147	14	every	every	DET
ejpam-1184	147	15	given	give	VERB
ejpam-1184	147	16	partition	partition	NOUN
ejpam-1184	147	17	of	of	ADP
ejpam-1184	147	18	these	these	DET
ejpam-1184	147	19	n	n	PRON
ejpam-1184	147	20	observations	observation	NOUN
ejpam-1184	147	21	:	:	PUNCT
ejpam-1184	148	1	π	π	PROPN
ejpam-1184	148	2	(	(	PUNCT
ejpam-1184	148	3	n	n	CCONJ
ejpam-1184	148	4	)	)	PUNCT
ejpam-1184	148	5	k	k	NOUN
ejpam-1184	149	1	=	=	PUNCT
ejpam-1184	149	2	{	{	PUNCT
ejpam-1184	149	3	c	c	NOUN
ejpam-1184	149	4	(	(	PUNCT
ejpam-1184	149	5	n)1	n)1	NOUN
ejpam-1184	149	6	,	,	PUNCT
ejpam-1184	149	7	.	.	PUNCT
ejpam-1184	149	8	.	.	PUNCT
ejpam-1184	149	9	.	.	PUNCT
ejpam-1184	150	1	,	,	PUNCT
ejpam-1184	150	2	c	c	NOUN
ejpam-1184	150	3	(	(	PUNCT
ejpam-1184	150	4	n	n	CCONJ
ejpam-1184	150	5	)	)	PUNCT
ejpam-1184	150	6	k	k	NOUN
ejpam-1184	151	1	}	}	PUNCT
ejpam-1184	151	2	,	,	PUNCT
ejpam-1184	151	3	where	where	SCONJ
ejpam-1184	151	4	k	k	PROPN
ejpam-1184	151	5	≤	≤	PROPN
ejpam-1184	151	6	k	k	PROPN
ejpam-1184	151	7	is	be	AUX
ejpam-1184	151	8	a	a	DET
ejpam-1184	151	9	positive	positive	ADJ
ejpam-1184	151	10	integer	integer	NOUN
ejpam-1184	151	11	.	.	PUNCT
ejpam-1184	152	1	for	for	ADP
ejpam-1184	152	2	such	such	DET
ejpam-1184	152	3	a	a	DET
ejpam-1184	152	4	partition	partition	NOUN
ejpam-1184	152	5	,	,	PUNCT
ejpam-1184	152	6	one	one	NUM
ejpam-1184	152	7	then	then	ADV
ejpam-1184	152	8	fits	fit	VERB
ejpam-1184	152	9	k	k	PROPN
ejpam-1184	152	10	clusterwise	clusterwise	PROPN
ejpam-1184	152	11	regression	regression	NOUN
ejpam-1184	152	12	models	model	NOUN
ejpam-1184	152	13	and	and	CCONJ
ejpam-1184	152	14	obtain	obtain	VERB
ejpam-1184	152	15	k	k	PROPN
ejpam-1184	152	16	least	least	ADJ
ejpam-1184	152	17	squares	square	NOUN
ejpam-1184	152	18	(	(	PUNCT
ejpam-1184	152	19	ls	ls	PROPN
ejpam-1184	152	20	)	)	PUNCT
ejpam-1184	152	21	estimates	estimate	NOUN
ejpam-1184	152	22	bβ	bβ	VERB
ejpam-1184	152	23	i	i	PRON
ejpam-1184	152	24	,	,	PUNCT
ejpam-1184	152	25	i	i	PRON
ejpam-1184	152	26	=	=	NOUN
ejpam-1184	152	27	1	1	NUM
ejpam-1184	152	28	,	,	PUNCT
ejpam-1184	152	29	.	.	PUNCT
ejpam-1184	152	30	.	.	PUNCT
ejpam-1184	153	1	.	.	PUNCT
ejpam-1184	154	1	,	,	PUNCT
ejpam-1184	154	2	k.	k.	PROPN
ejpam-1184	154	3	by	by	ADP
ejpam-1184	154	4	this	this	DET
ejpam-1184	154	5	stage	stage	NOUN
ejpam-1184	154	6	,	,	PUNCT
ejpam-1184	154	7	one	one	PRON
ejpam-1184	154	8	can	can	AUX
ejpam-1184	154	9	use	use	VERB
ejpam-1184	154	10	a	a	DET
ejpam-1184	154	11	criterion	criterion	NOUN
ejpam-1184	154	12	to	to	PART
ejpam-1184	154	13	select	select	VERB
ejpam-1184	154	14	the	the	DET
ejpam-1184	154	15	best	good	ADJ
ejpam-1184	154	16	k	k	NOUN
ejpam-1184	154	17	and	and	CCONJ
ejpam-1184	154	18	the	the	DET
ejpam-1184	154	19	associated	associated	ADJ
ejpam-1184	154	20	partition	partition	NOUN
ejpam-1184	154	21	.	.	PUNCT
ejpam-1184	155	1	shao	shao	PROPN
ejpam-1184	155	2	and	and	CCONJ
ejpam-1184	155	3	wu	wu	PROPN
ejpam-1184	156	1	[	[	X
ejpam-1184	156	2	14	14	NUM
ejpam-1184	156	3	]	]	X
ejpam-1184	156	4	propose	propose	VERB
ejpam-1184	156	5	an	an	DET
ejpam-1184	156	6	information	information	NOUN
ejpam-1184	156	7	-	-	PUNCT
ejpam-1184	156	8	based	base	VERB
ejpam-1184	156	9	criterion	criterion	NOUN
ejpam-1184	156	10	for	for	ADP
ejpam-1184	156	11	determining	determine	VERB
ejpam-1184	156	12	the	the	DET
ejpam-1184	156	13	number	number	NOUN
ejpam-1184	156	14	of	of	ADP
ejpam-1184	156	15	clusters	cluster	NOUN
ejpam-1184	156	16	as	as	ADP
ejpam-1184	156	17	following	follow	VERB
ejpam-1184	156	18	:	:	PUNCT
ejpam-1184	156	19	let	let	VERB
ejpam-1184	156	20	q(k	q(k	PRON
ejpam-1184	156	21	)	)	PUNCT
ejpam-1184	156	22	be	be	AUX
ejpam-1184	156	23	a	a	DET
ejpam-1184	156	24	strictly	strictly	ADV
ejpam-1184	156	25	increasing	increase	VERB
ejpam-1184	156	26	positive	positive	ADJ
ejpam-1184	156	27	function	function	NOUN
ejpam-1184	156	28	of	of	ADP
ejpam-1184	156	29	k	k	PROPN
ejpam-1184	156	30	,	,	PUNCT
ejpam-1184	156	31	and	and	CCONJ
ejpam-1184	156	32	an	an	DET
ejpam-1184	156	33	be	be	AUX
ejpam-1184	156	34	a	a	DET
ejpam-1184	156	35	sequence	sequence	NOUN
ejpam-1184	156	36	of	of	ADP
ejpam-1184	156	37	positive	positive	ADJ
ejpam-1184	156	38	constants	constant	NOUN
ejpam-1184	156	39	.	.	PUNCT
ejpam-1184	157	1	define	define	VERB
ejpam-1184	157	2	dn(π	dn(π	PUNCT
ejpam-1184	157	3	(	(	PUNCT
ejpam-1184	157	4	n	n	CCONJ
ejpam-1184	157	5	)	)	PUNCT
ejpam-1184	157	6	k	k	NOUN
ejpam-1184	157	7	)	)	PUNCT
ejpam-1184	158	1	=	=	PRON
ejpam-1184	158	2	k∑	k∑	VERB
ejpam-1184	158	3	i=1	i=1	PROPN
ejpam-1184	158	4	||yc	||yc	NOUN
ejpam-1184	158	5	(	(	PUNCT
ejpam-1184	158	6	n	n	CCONJ
ejpam-1184	158	7	)	)	PUNCT
ejpam-1184	159	1	i	i	PRON
ejpam-1184	159	2	−	−	PROPN
ejpam-1184	159	3	xc	xc	PROPN
ejpam-1184	159	4	(	(	PUNCT
ejpam-1184	159	5	n	n	CCONJ
ejpam-1184	159	6	)	)	PUNCT
ejpam-1184	159	7	i	i	PRON
ejpam-1184	159	8	β̂	β̂	ADP
ejpam-1184	159	9	i||	i||	VERB
ejpam-1184	159	10	2	2	NUM
ejpam-1184	159	11	+	+	NOUN
ejpam-1184	159	12	q(k)an	q(k)an	PROPN
ejpam-1184	159	13	,	,	PUNCT
ejpam-1184	159	14	(	(	PUNCT
ejpam-1184	159	15	6	6	NUM
ejpam-1184	159	16	)	)	PUNCT
ejpam-1184	159	17	and	and	CCONJ
ejpam-1184	159	18	k̂n	k̂n	PROPN
ejpam-1184	159	19	,	,	PUNCT
ejpam-1184	159	20	the	the	DET
ejpam-1184	159	21	estimate	estimate	NOUN
ejpam-1184	159	22	of	of	ADP
ejpam-1184	159	23	k0	k0	PROPN
ejpam-1184	159	24	,	,	PUNCT
ejpam-1184	159	25	is	be	AUX
ejpam-1184	159	26	the	the	DET
ejpam-1184	159	27	integer	integer	NOUN
ejpam-1184	159	28	that	that	PRON
ejpam-1184	159	29	minimizes	minimize	VERB
ejpam-1184	159	30	this	this	DET
ejpam-1184	159	31	criterion	criterion	NOUN
ejpam-1184	159	32	,	,	PUNCT
ejpam-1184	159	33	i.e.	i.e.	X
ejpam-1184	159	34	dn(k̂n	dn(k̂n	NOUN
ejpam-1184	159	35	)	)	PUNCT
ejpam-1184	159	36	=	=	SYM
ejpam-1184	160	1	min	min	PROPN
ejpam-1184	160	2	1≤k≤k	1≤k≤k	NUM
ejpam-1184	160	3	min	min	PROPN
ejpam-1184	160	4	π	π	PROPN
ejpam-1184	160	5	(	(	PUNCT
ejpam-1184	160	6	n	n	CCONJ
ejpam-1184	160	7	)	)	PUNCT
ejpam-1184	160	8	k	k	NOUN
ejpam-1184	160	9	dn(π	dn(π	X
ejpam-1184	160	10	(	(	PUNCT
ejpam-1184	160	11	n	n	CCONJ
ejpam-1184	160	12	)	)	PUNCT
ejpam-1184	160	13	k	k	NOUN
ejpam-1184	160	14	)	)	PUNCT
ejpam-1184	160	15	,	,	PUNCT
ejpam-1184	160	16	(	(	PUNCT
ejpam-1184	160	17	7	7	X
ejpam-1184	160	18	)	)	PUNCT
ejpam-1184	160	19	g.	g.	PROPN
ejpam-1184	160	20	qian	qian	PROPN
ejpam-1184	160	21	,	,	PUNCT
ejpam-1184	160	22	y.	y.	PROPN
ejpam-1184	160	23	wu	wu	PROPN
ejpam-1184	160	24	/	/	SYM
ejpam-1184	160	25	eur	eur	PROPN
ejpam-1184	160	26	.	.	PUNCT
ejpam-1184	161	1	j.	j.	PROPN
ejpam-1184	161	2	pure	pure	PROPN
ejpam-1184	161	3	appl	appl	PROPN
ejpam-1184	161	4	.	.	PROPN
ejpam-1184	161	5	math	math	PROPN
ejpam-1184	161	6	,	,	PUNCT
ejpam-1184	161	7	4	4	NUM
ejpam-1184	161	8	(	(	PUNCT
ejpam-1184	161	9	2011	2011	NUM
ejpam-1184	161	10	)	)	PUNCT
ejpam-1184	161	11	,	,	PUNCT
ejpam-1184	161	12	455	455	NUM
ejpam-1184	161	13	-	-	SYM
ejpam-1184	161	14	466	466	NUM
ejpam-1184	161	15	460	460	NUM
ejpam-1184	161	16	where	where	SCONJ
ejpam-1184	161	17	1≤	1≤	NOUN
ejpam-1184	162	1	k	k	PROPN
ejpam-1184	162	2	≤	≤	PROPN
ejpam-1184	163	1	k	k	X
ejpam-1184	163	2	.	.	PUNCT
ejpam-1184	164	1	it	it	PRON
ejpam-1184	164	2	can	can	AUX
ejpam-1184	164	3	be	be	AUX
ejpam-1184	164	4	seen	see	VERB
ejpam-1184	164	5	that	that	SCONJ
ejpam-1184	164	6	in	in	ADP
ejpam-1184	164	7	(	(	PUNCT
ejpam-1184	164	8	6	6	NUM
ejpam-1184	164	9	)	)	PUNCT
ejpam-1184	164	10	,	,	PUNCT
ejpam-1184	164	11	the	the	DET
ejpam-1184	164	12	first	first	ADJ
ejpam-1184	164	13	term	term	NOUN
ejpam-1184	164	14	is	be	AUX
ejpam-1184	164	15	the	the	DET
ejpam-1184	164	16	residual	residual	ADJ
ejpam-1184	164	17	sum	sum	NOUN
ejpam-1184	164	18	of	of	ADP
ejpam-1184	164	19	squares	square	NOUN
ejpam-1184	164	20	which	which	PRON
ejpam-1184	164	21	measures	measure	VERB
ejpam-1184	164	22	the	the	DET
ejpam-1184	164	23	goodness	goodness	NOUN
ejpam-1184	164	24	of	of	ADP
ejpam-1184	164	25	fit	fit	NOUN
ejpam-1184	164	26	of	of	ADP
ejpam-1184	164	27	the	the	DET
ejpam-1184	164	28	model	model	NOUN
ejpam-1184	164	29	and	and	CCONJ
ejpam-1184	164	30	the	the	DET
ejpam-1184	164	31	second	second	ADJ
ejpam-1184	164	32	term	term	NOUN
ejpam-1184	164	33	is	be	AUX
ejpam-1184	164	34	the	the	DET
ejpam-1184	164	35	penalty	penalty	NOUN
ejpam-1184	164	36	for	for	ADP
ejpam-1184	164	37	over	over	ADV
ejpam-1184	164	38	-	-	PUNCT
ejpam-1184	164	39	fitting	fitting	NOUN
ejpam-1184	164	40	.	.	PUNCT
ejpam-1184	165	1	furthermore	furthermore	ADV
ejpam-1184	165	2	,	,	PUNCT
ejpam-1184	165	3	the	the	DET
ejpam-1184	165	4	criterion	criterion	NOUN
ejpam-1184	165	5	(	(	PUNCT
ejpam-1184	165	6	7	7	X
ejpam-1184	165	7	)	)	PUNCT
ejpam-1184	165	8	shows	show	VERB
ejpam-1184	165	9	that	that	SCONJ
ejpam-1184	165	10	one	one	NOUN
ejpam-1184	165	11	determines	determine	VERB
ejpam-1184	165	12	the	the	DET
ejpam-1184	165	13	optimal	optimal	ADJ
ejpam-1184	165	14	number	number	NOUN
ejpam-1184	165	15	of	of	ADP
ejpam-1184	165	16	clusters	cluster	NOUN
ejpam-1184	165	17	and	and	CCONJ
ejpam-1184	165	18	the	the	DET
ejpam-1184	165	19	corresponding	correspond	VERB
ejpam-1184	165	20	partitioning	partitioning	NOUN
ejpam-1184	165	21	simultaneously	simultaneously	ADV
ejpam-1184	165	22	.	.	PUNCT
ejpam-1184	166	1	we	we	PRON
ejpam-1184	166	2	shall	shall	AUX
ejpam-1184	166	3	call	call	VERB
ejpam-1184	166	4	(	(	PUNCT
ejpam-1184	166	5	7	7	NUM
ejpam-1184	166	6	)	)	PUNCT
ejpam-1184	166	7	criterion	criterion	NOUN
ejpam-1184	166	8	ls	ls	PROPN
ejpam-1184	166	9	-	-	PUNCT
ejpam-1184	166	10	c	c	NOUN
ejpam-1184	166	11	in	in	ADP
ejpam-1184	166	12	the	the	DET
ejpam-1184	166	13	sequel	sequel	NOUN
ejpam-1184	166	14	,	,	PUNCT
ejpam-1184	166	15	which	which	PRON
ejpam-1184	166	16	stands	stand	VERB
ejpam-1184	166	17	for	for	ADP
ejpam-1184	166	18	clustering	cluster	VERB
ejpam-1184	166	19	by	by	ADP
ejpam-1184	166	20	the	the	DET
ejpam-1184	166	21	ls	ls	ADJ
ejpam-1184	166	22	method	method	NOUN
ejpam-1184	166	23	.	.	PUNCT
ejpam-1184	167	1	under	under	ADP
ejpam-1184	167	2	some	some	DET
ejpam-1184	167	3	mild	mild	ADJ
ejpam-1184	167	4	conditions	condition	NOUN
ejpam-1184	167	5	,	,	PUNCT
ejpam-1184	167	6	it	it	PRON
ejpam-1184	167	7	is	be	AUX
ejpam-1184	167	8	shown	show	VERB
ejpam-1184	167	9	in	in	ADP
ejpam-1184	167	10	shao	shao	PROPN
ejpam-1184	167	11	and	and	CCONJ
ejpam-1184	167	12	wu	wu	PROPN
ejpam-1184	168	1	[	[	X
ejpam-1184	168	2	14	14	NUM
ejpam-1184	168	3	]	]	PUNCT
ejpam-1184	168	4	that	that	SCONJ
ejpam-1184	168	5	the	the	DET
ejpam-1184	168	6	proposed	propose	VERB
ejpam-1184	168	7	criterion	criterion	NOUN
ejpam-1184	168	8	selects	select	VERB
ejpam-1184	168	9	the	the	DET
ejpam-1184	168	10	true	true	ADJ
ejpam-1184	168	11	number	number	NOUN
ejpam-1184	168	12	of	of	ADP
ejpam-1184	168	13	regression	regression	NOUN
ejpam-1184	168	14	hyperplanes	hyperplane	NOUN
ejpam-1184	168	15	with	with	ADP
ejpam-1184	168	16	probability	probability	NOUN
ejpam-1184	168	17	one	one	NOUN
ejpam-1184	168	18	among	among	ADP
ejpam-1184	168	19	all	all	DET
ejpam-1184	168	20	classgrowing	classgrowe	VERB
ejpam-1184	168	21	sequences	sequence	NOUN
ejpam-1184	168	22	of	of	ADP
ejpam-1184	168	23	classifications	classification	NOUN
ejpam-1184	168	24	,	,	PUNCT
ejpam-1184	168	25	when	when	SCONJ
ejpam-1184	168	26	the	the	DET
ejpam-1184	168	27	number	number	NOUN
ejpam-1184	168	28	of	of	ADP
ejpam-1184	168	29	observations	observation	NOUN
ejpam-1184	168	30	n	n	CCONJ
ejpam-1184	168	31	from	from	ADP
ejpam-1184	168	32	the	the	DET
ejpam-1184	168	33	population	population	NOUN
ejpam-1184	168	34	increases	increase	VERB
ejpam-1184	168	35	to	to	ADP
ejpam-1184	168	36	infinity	infinity	NOUN
ejpam-1184	168	37	.	.	PUNCT
ejpam-1184	169	1	note	note	VERB
ejpam-1184	169	2	that	that	SCONJ
ejpam-1184	169	3	the	the	DET
ejpam-1184	169	4	assumption	assumption	NOUN
ejpam-1184	169	5	eoi	eoi	NOUN
ejpam-1184	169	6	∼	∼	NOUN
ejpam-1184	169	7	n(0,σ2	n(0,σ2	NOUN
ejpam-1184	169	8	i	i	PROPN
ejpam-1184	169	9	ini	ini	PROPN
ejpam-1184	169	10	)	)	PUNCT
ejpam-1184	169	11	in	in	ADP
ejpam-1184	169	12	(	(	PUNCT
ejpam-1184	169	13	5	5	NUM
ejpam-1184	169	14	)	)	PUNCT
ejpam-1184	169	15	is	be	AUX
ejpam-1184	169	16	not	not	PART
ejpam-1184	169	17	required	require	VERB
ejpam-1184	169	18	in	in	ADP
ejpam-1184	169	19	computing	compute	VERB
ejpam-1184	169	20	the	the	DET
ejpam-1184	169	21	lsestimates	lsestimates	PROPN
ejpam-1184	169	22	bβ	bβ	NOUN
ejpam-1184	169	23	i	i	PRON
ejpam-1184	169	24	and	and	CCONJ
ejpam-1184	169	25	the	the	DET
ejpam-1184	169	26	criterion	criterion	NOUN
ejpam-1184	169	27	function	function	NOUN
ejpam-1184	169	28	dn(π	dn(π	X
ejpam-1184	169	29	(	(	PUNCT
ejpam-1184	169	30	n	n	CCONJ
ejpam-1184	169	31	)	)	PUNCT
ejpam-1184	169	32	k	k	NOUN
ejpam-1184	169	33	)	)	PUNCT
ejpam-1184	169	34	.	.	PUNCT
ejpam-1184	170	1	but	but	CCONJ
ejpam-1184	170	2	the	the	DET
ejpam-1184	170	3	least	least	ADJ
ejpam-1184	170	4	squares	square	NOUN
ejpam-1184	170	5	estimates	estimate	NOUN
ejpam-1184	170	6	are	be	AUX
ejpam-1184	170	7	known	know	VERB
ejpam-1184	170	8	to	to	PART
ejpam-1184	170	9	be	be	AUX
ejpam-1184	170	10	sensitive	sensitive	ADJ
ejpam-1184	170	11	to	to	ADP
ejpam-1184	170	12	outliers	outlier	NOUN
ejpam-1184	170	13	and	and	CCONJ
ejpam-1184	170	14	violation	violation	NOUN
ejpam-1184	170	15	of	of	ADP
ejpam-1184	170	16	the	the	DET
ejpam-1184	170	17	normality	normality	NOUN
ejpam-1184	170	18	assumption	assumption	NOUN
ejpam-1184	170	19	in	in	ADP
ejpam-1184	170	20	the	the	DET
ejpam-1184	170	21	data	datum	NOUN
ejpam-1184	170	22	.	.	PUNCT
ejpam-1184	171	1	this	this	PRON
ejpam-1184	171	2	implies	imply	VERB
ejpam-1184	171	3	that	that	SCONJ
ejpam-1184	171	4	the	the	DET
ejpam-1184	171	5	ls	ls	PROPN
ejpam-1184	171	6	-	-	PUNCT
ejpam-1184	171	7	c	c	NOUN
ejpam-1184	171	8	criterion	criterion	NOUN
ejpam-1184	171	9	is	be	AUX
ejpam-1184	171	10	expected	expect	VERB
ejpam-1184	171	11	to	to	PART
ejpam-1184	171	12	work	work	VERB
ejpam-1184	171	13	well	well	ADV
ejpam-1184	171	14	for	for	ADP
ejpam-1184	171	15	selecting	select	VERB
ejpam-1184	171	16	the	the	DET
ejpam-1184	171	17	number	number	NOUN
ejpam-1184	171	18	of	of	ADP
ejpam-1184	171	19	clusters	cluster	NOUN
ejpam-1184	171	20	and	and	CCONJ
ejpam-1184	171	21	estimating	estimate	VERB
ejpam-1184	171	22	the	the	DET
ejpam-1184	171	23	partition	partition	NOUN
ejpam-1184	171	24	in	in	ADP
ejpam-1184	171	25	linear	linear	PROPN
ejpam-1184	171	26	regression	regression	NOUN
ejpam-1184	171	27	clustering	cluster	VERB
ejpam-1184	171	28	only	only	ADV
ejpam-1184	171	29	when	when	SCONJ
ejpam-1184	171	30	the	the	DET
ejpam-1184	171	31	normality	normality	NOUN
ejpam-1184	171	32	assumption	assumption	NOUN
ejpam-1184	171	33	is	be	AUX
ejpam-1184	171	34	not	not	PART
ejpam-1184	171	35	seriously	seriously	ADV
ejpam-1184	171	36	violated	violate	VERB
ejpam-1184	171	37	.	.	PUNCT
ejpam-1184	172	1	recently	recently	ADV
ejpam-1184	172	2	,	,	PUNCT
ejpam-1184	172	3	a	a	DET
ejpam-1184	172	4	consistent	consistent	ADJ
ejpam-1184	172	5	robust	robust	ADJ
ejpam-1184	172	6	procedure	procedure	NOUN
ejpam-1184	172	7	for	for	ADP
ejpam-1184	172	8	determining	determine	VERB
ejpam-1184	172	9	the	the	DET
ejpam-1184	172	10	number	number	NOUN
ejpam-1184	172	11	of	of	ADP
ejpam-1184	172	12	clusters	cluster	NOUN
ejpam-1184	172	13	in	in	ADP
ejpam-1184	172	14	regression	regression	NOUN
ejpam-1184	172	15	clustering	clustering	NOUN
ejpam-1184	172	16	is	be	AUX
ejpam-1184	172	17	proposed	propose	VERB
ejpam-1184	172	18	in	in	ADP
ejpam-1184	172	19	[	[	X
ejpam-1184	172	20	2	2	NUM
ejpam-1184	172	21	]	]	PUNCT
ejpam-1184	172	22	.	.	PUNCT
ejpam-1184	173	1	however	however	ADV
ejpam-1184	173	2	,	,	PUNCT
ejpam-1184	173	3	we	we	PRON
ejpam-1184	173	4	will	will	AUX
ejpam-1184	173	5	not	not	PART
ejpam-1184	173	6	get	get	VERB
ejpam-1184	173	7	into	into	ADP
ejpam-1184	173	8	its	its	PRON
ejpam-1184	173	9	theoretic	theoretic	ADJ
ejpam-1184	173	10	detail	detail	NOUN
ejpam-1184	173	11	here	here	ADV
ejpam-1184	173	12	to	to	PART
ejpam-1184	173	13	keep	keep	VERB
ejpam-1184	173	14	this	this	DET
ejpam-1184	173	15	paper	paper	NOUN
ejpam-1184	173	16	into	into	ADP
ejpam-1184	173	17	reasonable	reasonable	ADJ
ejpam-1184	173	18	length	length	NOUN
ejpam-1184	173	19	.	.	PUNCT
ejpam-1184	174	1	we	we	PRON
ejpam-1184	174	2	will	will	AUX
ejpam-1184	174	3	use	use	VERB
ejpam-1184	174	4	only	only	ADV
ejpam-1184	174	5	the	the	DET
ejpam-1184	174	6	simulation	simulation	NOUN
ejpam-1184	174	7	in	in	ADP
ejpam-1184	174	8	section	section	NOUN
ejpam-1184	174	9	5	5	NUM
ejpam-1184	174	10	to	to	PART
ejpam-1184	174	11	illustrate	illustrate	VERB
ejpam-1184	174	12	the	the	DET
ejpam-1184	174	13	sensitivity	sensitivity	NOUN
ejpam-1184	174	14	of	of	ADP
ejpam-1184	174	15	the	the	DET
ejpam-1184	174	16	ls	ls	PROPN
ejpam-1184	174	17	-	-	PUNCT
ejpam-1184	174	18	c	c	NOUN
ejpam-1184	174	19	criterion	criterion	NOUN
ejpam-1184	174	20	against	against	ADP
ejpam-1184	174	21	normality	normality	NOUN
ejpam-1184	174	22	.	.	PUNCT
ejpam-1184	175	1	finally	finally	ADV
ejpam-1184	175	2	,	,	PUNCT
ejpam-1184	175	3	it	it	PRON
ejpam-1184	175	4	seems	seem	VERB
ejpam-1184	175	5	that	that	SCONJ
ejpam-1184	175	6	each	each	DET
ejpam-1184	175	7	squared	square	VERB
ejpam-1184	175	8	residual	residual	ADJ
ejpam-1184	175	9	sum	sum	NOUN
ejpam-1184	175	10	in	in	ADP
ejpam-1184	175	11	the	the	DET
ejpam-1184	175	12	first	first	ADJ
ejpam-1184	175	13	term	term	NOUN
ejpam-1184	175	14	of	of	ADP
ejpam-1184	175	15	(	(	PUNCT
ejpam-1184	175	16	6	6	NUM
ejpam-1184	175	17	)	)	PUNCT
ejpam-1184	175	18	should	should	AUX
ejpam-1184	175	19	be	be	AUX
ejpam-1184	175	20	scaled	scale	VERB
ejpam-1184	175	21	by	by	ADP
ejpam-1184	175	22	the	the	DET
ejpam-1184	175	23	corresponding	corresponding	ADJ
ejpam-1184	175	24	variance	variance	NOUN
ejpam-1184	175	25	estimate	estimate	NOUN
ejpam-1184	175	26	σ̂2	σ̂2	NOUN
ejpam-1184	176	1	i	i	PRON
ejpam-1184	176	2	.	.	PUNCT
ejpam-1184	177	1	actually	actually	ADV
ejpam-1184	177	2	ignoring	ignore	VERB
ejpam-1184	177	3	this	this	DET
ejpam-1184	177	4	scaling	scaling	NOUN
ejpam-1184	177	5	does	do	AUX
ejpam-1184	177	6	not	not	PART
ejpam-1184	177	7	affect	affect	VERB
ejpam-1184	177	8	the	the	DET
ejpam-1184	177	9	asymptotic	asymptotic	ADJ
ejpam-1184	177	10	properties	property	NOUN
ejpam-1184	177	11	of	of	ADP
ejpam-1184	177	12	the	the	DET
ejpam-1184	177	13	criterion	criterion	NOUN
ejpam-1184	177	14	function	function	NOUN
ejpam-1184	177	15	dn(π	dn(π	X
ejpam-1184	177	16	(	(	PUNCT
ejpam-1184	177	17	n	n	CCONJ
ejpam-1184	177	18	)	)	PUNCT
ejpam-1184	177	19	k	k	NOUN
ejpam-1184	177	20	)	)	PUNCT
ejpam-1184	177	21	.	.	PUNCT
ejpam-1184	178	1	it	it	PRON
ejpam-1184	178	2	turns	turn	VERB
ejpam-1184	178	3	out	out	ADP
ejpam-1184	178	4	that	that	SCONJ
ejpam-1184	178	5	ignoring	ignore	VERB
ejpam-1184	178	6	the	the	DET
ejpam-1184	178	7	scaling	scaling	NOUN
ejpam-1184	178	8	would	would	AUX
ejpam-1184	178	9	improve	improve	VERB
ejpam-1184	178	10	the	the	DET
ejpam-1184	178	11	robustness	robustness	NOUN
ejpam-1184	178	12	of	of	ADP
ejpam-1184	178	13	the	the	DET
ejpam-1184	178	14	clustering	cluster	VERB
ejpam-1184	178	15	procedure	procedure	NOUN
ejpam-1184	178	16	.	.	PUNCT
ejpam-1184	179	1	this	this	PRON
ejpam-1184	179	2	is	be	AUX
ejpam-1184	179	3	because	because	SCONJ
ejpam-1184	179	4	a	a	DET
ejpam-1184	179	5	large	large	ADJ
ejpam-1184	179	6	σ2	σ2	NOUN
ejpam-1184	179	7	i	i	PRON
ejpam-1184	179	8	estimate	estimate	VERB
ejpam-1184	179	9	is	be	AUX
ejpam-1184	179	10	more	more	ADV
ejpam-1184	179	11	likely	likely	ADJ
ejpam-1184	179	12	to	to	PART
ejpam-1184	179	13	be	be	AUX
ejpam-1184	179	14	associated	associate	VERB
ejpam-1184	179	15	with	with	ADP
ejpam-1184	179	16	a	a	DET
ejpam-1184	179	17	cluster	cluster	NOUN
ejpam-1184	179	18	with	with	ADP
ejpam-1184	179	19	large	large	ADJ
ejpam-1184	179	20	variability	variability	NOUN
ejpam-1184	179	21	,	,	PUNCT
ejpam-1184	179	22	thus	thus	ADV
ejpam-1184	179	23	being	be	AUX
ejpam-1184	179	24	less	less	ADV
ejpam-1184	179	25	separable	separable	ADJ
ejpam-1184	179	26	from	from	ADP
ejpam-1184	179	27	the	the	DET
ejpam-1184	179	28	other	other	ADJ
ejpam-1184	179	29	clusters	cluster	NOUN
ejpam-1184	179	30	.	.	PUNCT
ejpam-1184	180	1	ignoring	ignore	VERB
ejpam-1184	180	2	the	the	DET
ejpam-1184	180	3	scaling	scaling	NOUN
ejpam-1184	180	4	would	would	AUX
ejpam-1184	180	5	favor	favor	VERB
ejpam-1184	180	6	not	not	PART
ejpam-1184	180	7	including	include	VERB
ejpam-1184	180	8	the	the	DET
ejpam-1184	180	9	outlaying	outlaye	VERB
ejpam-1184	180	10	data	data	NOUN
ejpam-1184	180	11	points	point	NOUN
ejpam-1184	180	12	in	in	ADP
ejpam-1184	180	13	the	the	DET
ejpam-1184	180	14	current	current	ADJ
ejpam-1184	180	15	cluster	cluster	NOUN
ejpam-1184	180	16	.	.	PUNCT
ejpam-1184	181	1	4	4	X
ejpam-1184	181	2	.	.	X
ejpam-1184	181	3	an	an	DET
ejpam-1184	181	4	algorithm	algorithm	NOUN
ejpam-1184	181	5	for	for	ADP
ejpam-1184	181	6	estimation	estimation	NOUN
ejpam-1184	181	7	and	and	CCONJ
ejpam-1184	181	8	selection	selection	NOUN
ejpam-1184	181	9	in	in	ADP
ejpam-1184	181	10	regression	regression	NOUN
ejpam-1184	181	11	clustering	cluster	VERB
ejpam-1184	181	12	we	we	PRON
ejpam-1184	181	13	give	give	VERB
ejpam-1184	181	14	an	an	DET
ejpam-1184	181	15	iterative	iterative	NOUN
ejpam-1184	181	16	algorithm	algorithm	NOUN
ejpam-1184	181	17	in	in	ADP
ejpam-1184	181	18	this	this	DET
ejpam-1184	181	19	section	section	NOUN
ejpam-1184	181	20	to	to	PART
ejpam-1184	181	21	implement	implement	VERB
ejpam-1184	181	22	the	the	DET
ejpam-1184	181	23	procedures	procedure	NOUN
ejpam-1184	181	24	in	in	ADP
ejpam-1184	181	25	the	the	DET
ejpam-1184	181	26	previous	previous	ADJ
ejpam-1184	181	27	section	section	NOUN
ejpam-1184	181	28	for	for	ADP
ejpam-1184	181	29	selecting	select	VERB
ejpam-1184	181	30	the	the	DET
ejpam-1184	181	31	optimal	optimal	ADJ
ejpam-1184	181	32	clustering	clustering	NOUN
ejpam-1184	181	33	and	and	CCONJ
ejpam-1184	181	34	estimating	estimate	VERB
ejpam-1184	181	35	the	the	DET
ejpam-1184	181	36	number	number	NOUN
ejpam-1184	181	37	of	of	ADP
ejpam-1184	181	38	clusters	cluster	NOUN
ejpam-1184	181	39	in	in	ADP
ejpam-1184	181	40	regression	regression	NOUN
ejpam-1184	181	41	clustering	cluster	VERB
ejpam-1184	181	42	.	.	PUNCT
ejpam-1184	182	1	for	for	ADP
ejpam-1184	182	2	each	each	DET
ejpam-1184	182	3	fixed	fix	VERB
ejpam-1184	182	4	k	k	PROPN
ejpam-1184	182	5	,	,	PUNCT
ejpam-1184	182	6	we	we	PRON
ejpam-1184	182	7	obtain	obtain	VERB
ejpam-1184	182	8	the	the	DET
ejpam-1184	182	9	optimal	optimal	ADJ
ejpam-1184	182	10	clustering	clustering	NOUN
ejpam-1184	182	11	of	of	ADP
ejpam-1184	182	12	the	the	DET
ejpam-1184	182	13	data	datum	NOUN
ejpam-1184	182	14	πk	πk	X
ejpam-1184	182	15	=	=	SYM
ejpam-1184	182	16	{	{	PUNCT
ejpam-1184	182	17	c1	c1	NOUN
ejpam-1184	182	18	,	,	PUNCT
ejpam-1184	182	19	.	.	PUNCT
ejpam-1184	182	20	.	.	PUNCT
ejpam-1184	183	1	.	.	PUNCT
ejpam-1184	184	1	,	,	PUNCT
ejpam-1184	184	2	ck	ck	PROPN
ejpam-1184	184	3	}	}	PUNCT
ejpam-1184	184	4	by	by	ADP
ejpam-1184	184	5	minimizing	minimize	VERB
ejpam-1184	184	6	the	the	DET
ejpam-1184	184	7	within	within	ADP
ejpam-1184	184	8	-	-	PUNCT
ejpam-1184	184	9	cluster	cluster	NOUN
ejpam-1184	184	10	sum	sum	NOUN
ejpam-1184	184	11	of	of	ADP
ejpam-1184	184	12	residual	residual	ADJ
ejpam-1184	184	13	squares	square	NOUN
ejpam-1184	184	14	.	.	PUNCT
ejpam-1184	185	1	the	the	DET
ejpam-1184	185	2	quantity	quantity	NOUN
ejpam-1184	185	3	to	to	PART
ejpam-1184	185	4	be	be	AUX
ejpam-1184	185	5	minimized	minimize	VERB
ejpam-1184	185	6	is	be	AUX
ejpam-1184	185	7	then	then	ADV
ejpam-1184	185	8	srss(πk	srss(πk	NOUN
ejpam-1184	185	9	)	)	PUNCT
ejpam-1184	185	10	=	=	VERB
ejpam-1184	186	1	k∑	k∑	VERB
ejpam-1184	186	2	i=1	i=1	X
ejpam-1184	187	1	||yci	||yci	ADJ
ejpam-1184	187	2	−	−	NOUN
ejpam-1184	187	3	x	x	SYM
ejpam-1184	187	4	′ci	′ci	NOUN
ejpam-1184	187	5	bβ	bβ	NOUN
ejpam-1184	187	6	i	i	PRON
ejpam-1184	187	7	||	||	NOUN
ejpam-1184	187	8	2	2	NUM
ejpam-1184	187	9	(	(	PUNCT
ejpam-1184	187	10	8)	8)	NUM
ejpam-1184	187	11	where	where	SCONJ
ejpam-1184	187	12	bβ	bβ	NOUN
ejpam-1184	187	13	i	i	PRON
ejpam-1184	187	14	,	,	PUNCT
ejpam-1184	187	15	i	i	PRON
ejpam-1184	187	16	=	=	NOUN
ejpam-1184	187	17	1	1	NUM
ejpam-1184	187	18	,	,	PUNCT
ejpam-1184	187	19	·	·	PUNCT
ejpam-1184	187	20	·	·	PUNCT
ejpam-1184	187	21	·	·	PUNCT
ejpam-1184	187	22	,	,	PUNCT
ejpam-1184	187	23	k	k	X
ejpam-1184	187	24	,	,	PUNCT
ejpam-1184	187	25	are	be	AUX
ejpam-1184	187	26	the	the	DET
ejpam-1184	187	27	least	least	ADJ
ejpam-1184	187	28	squares	square	NOUN
ejpam-1184	187	29	estimators	estimator	NOUN
ejpam-1184	187	30	based	base	VERB
ejpam-1184	187	31	on	on	ADP
ejpam-1184	187	32	given	give	VERB
ejpam-1184	187	33	{	{	PUNCT
ejpam-1184	187	34	c1	c1	NOUN
ejpam-1184	187	35	,	,	PUNCT
ejpam-1184	187	36	.	.	PUNCT
ejpam-1184	187	37	.	.	PUNCT
ejpam-1184	188	1	.	.	PUNCT
ejpam-1184	189	1	,	,	PUNCT
ejpam-1184	189	2	ck	ck	ADJ
ejpam-1184	189	3	}	}	PUNCT
ejpam-1184	189	4	.	.	PUNCT
ejpam-1184	190	1	this	this	DET
ejpam-1184	190	2	minimization	minimization	NOUN
ejpam-1184	190	3	can	can	AUX
ejpam-1184	190	4	be	be	AUX
ejpam-1184	190	5	accomplished	accomplish	VERB
ejpam-1184	190	6	according	accord	VERB
ejpam-1184	190	7	to	to	ADP
ejpam-1184	190	8	the	the	DET
ejpam-1184	190	9	following	following	ADJ
ejpam-1184	190	10	algorithm	algorithm	NOUN
ejpam-1184	190	11	:	:	PUNCT
ejpam-1184	190	12	(	(	PUNCT
ejpam-1184	190	13	i	i	NOUN
ejpam-1184	190	14	)	)	PUNCT
ejpam-1184	190	15	label	label	VERB
ejpam-1184	190	16	all	all	DET
ejpam-1184	190	17	the	the	DET
ejpam-1184	190	18	data	data	NOUN
ejpam-1184	190	19	points	point	NOUN
ejpam-1184	190	20	in	in	ADP
ejpam-1184	190	21	the	the	DET
ejpam-1184	190	22	sample	sample	NOUN
ejpam-1184	190	23	as	as	ADP
ejpam-1184	190	24	1	1	NUM
ejpam-1184	190	25	to	to	PART
ejpam-1184	190	26	n.	n.	NOUN
ejpam-1184	190	27	given	give	VERB
ejpam-1184	190	28	an	an	DET
ejpam-1184	190	29	initial	initial	ADJ
ejpam-1184	190	30	partition	partition	NOUN
ejpam-1184	190	31	πk	πk	X
ejpam-1184	190	32	=	=	SYM
ejpam-1184	190	33	{	{	PUNCT
ejpam-1184	190	34	c1	c1	NOUN
ejpam-1184	190	35	,	,	PUNCT
ejpam-1184	190	36	.	.	PUNCT
ejpam-1184	190	37	.	.	PUNCT
ejpam-1184	191	1	.	.	PUNCT
ejpam-1184	192	1	,	,	PUNCT
ejpam-1184	192	2	ck	ck	NOUN
ejpam-1184	192	3	}	}	PUNCT
ejpam-1184	192	4	of	of	ADP
ejpam-1184	192	5	o	o	NOUN
ejpam-1184	192	6	=	=	PUNCT
ejpam-1184	192	7	{	{	PUNCT
ejpam-1184	192	8	1	1	NUM
ejpam-1184	192	9	,	,	PUNCT
ejpam-1184	192	10	.	.	PUNCT
ejpam-1184	192	11	.	.	PUNCT
ejpam-1184	193	1	.	.	PUNCT
ejpam-1184	193	2	,	,	PUNCT
ejpam-1184	194	1	n	n	CCONJ
ejpam-1184	194	2	}	}	PUNCT
ejpam-1184	194	3	,	,	PUNCT
ejpam-1184	194	4	fit	fit	ADJ
ejpam-1184	194	5	regression	regression	NOUN
ejpam-1184	194	6	models	model	NOUN
ejpam-1184	194	7	for	for	ADP
ejpam-1184	194	8	each	each	PRON
ejpam-1184	194	9	of	of	ADP
ejpam-1184	194	10	the	the	DET
ejpam-1184	194	11	k	k	PROPN
ejpam-1184	194	12	clusters	cluster	NOUN
ejpam-1184	194	13	and	and	CCONJ
ejpam-1184	194	14	obtain	obtain	VERB
ejpam-1184	194	15	the	the	DET
ejpam-1184	194	16	overall	overall	ADJ
ejpam-1184	194	17	sum	sum	NOUN
ejpam-1184	194	18	of	of	ADP
ejpam-1184	194	19	the	the	DET
ejpam-1184	194	20	squared	square	VERB
ejpam-1184	194	21	residuals	residual	NOUN
ejpam-1184	194	22	srss0	srss0	NOUN
ejpam-1184	194	23	for	for	ADP
ejpam-1184	194	24	this	this	DET
ejpam-1184	194	25	partition	partition	NOUN
ejpam-1184	194	26	.	.	PUNCT
ejpam-1184	195	1	initialize	initialize	VERB
ejpam-1184	195	2	i	i	PRON
ejpam-1184	195	3	=	=	NOUN
ejpam-1184	195	4	0	0	PROPN
ejpam-1184	195	5	.	.	PUNCT
ejpam-1184	196	1	g.	g.	PROPN
ejpam-1184	196	2	qian	qian	PROPN
ejpam-1184	196	3	,	,	PUNCT
ejpam-1184	196	4	y.	y.	PROPN
ejpam-1184	196	5	wu	wu	PROPN
ejpam-1184	196	6	/	/	SYM
ejpam-1184	196	7	eur	eur	PROPN
ejpam-1184	196	8	.	.	PUNCT
ejpam-1184	197	1	j.	j.	PROPN
ejpam-1184	197	2	pure	pure	PROPN
ejpam-1184	197	3	appl	appl	PROPN
ejpam-1184	197	4	.	.	PROPN
ejpam-1184	197	5	math	math	PROPN
ejpam-1184	197	6	,	,	PUNCT
ejpam-1184	197	7	4	4	NUM
ejpam-1184	197	8	(	(	PUNCT
ejpam-1184	197	9	2011	2011	NUM
ejpam-1184	197	10	)	)	PUNCT
ejpam-1184	197	11	,	,	PUNCT
ejpam-1184	197	12	455	455	NUM
ejpam-1184	197	13	-	-	SYM
ejpam-1184	197	14	466	466	NUM
ejpam-1184	197	15	461	461	NUM
ejpam-1184	197	16	(	(	PUNCT
ejpam-1184	197	17	ii	ii	NOUN
ejpam-1184	197	18	)	)	PUNCT
ejpam-1184	197	19	set	set	NOUN
ejpam-1184	197	20	i	i	NOUN
ejpam-1184	197	21	=	=	PUNCT
ejpam-1184	197	22	i+	i+	PUNCT
ejpam-1184	197	23	1	1	NUM
ejpam-1184	197	24	and	and	CCONJ
ejpam-1184	197	25	reset	reset	VERB
ejpam-1184	197	26	i	i	NOUN
ejpam-1184	197	27	=	=	NOUN
ejpam-1184	197	28	1	1	NUM
ejpam-1184	197	29	if	if	SCONJ
ejpam-1184	197	30	i	i	PRON
ejpam-1184	197	31	>	>	X
ejpam-1184	197	32	n.	n.	PROPN
ejpam-1184	197	33	suppose	suppose	VERB
ejpam-1184	197	34	i	i	PRON
ejpam-1184	197	35	∈	∈	PROPN
ejpam-1184	198	1	c	c	PROPN
ejpam-1184	198	2	j.	j.	PROPN
ejpam-1184	198	3	then	then	ADV
ejpam-1184	198	4	move	move	VERB
ejpam-1184	198	5	i	i	PRON
ejpam-1184	198	6	into	into	ADP
ejpam-1184	198	7	ch	ch	PROPN
ejpam-1184	198	8	,	,	PUNCT
ejpam-1184	198	9	h=	h=	NOUN
ejpam-1184	198	10	1	1	NUM
ejpam-1184	198	11	,	,	PUNCT
ejpam-1184	198	12	.	.	PUNCT
ejpam-1184	198	13	.	.	PUNCT
ejpam-1184	199	1	.	.	PUNCT
ejpam-1184	200	1	,	,	PUNCT
ejpam-1184	200	2	k	k	NOUN
ejpam-1184	200	3	,	,	PUNCT
ejpam-1184	200	4	h	h	PROPN
ejpam-1184	200	5	6=	6=	PROPN
ejpam-1184	200	6	j	j	PROPN
ejpam-1184	200	7	respectively	respectively	ADV
ejpam-1184	200	8	.	.	PUNCT
ejpam-1184	201	1	for	for	ADP
ejpam-1184	201	2	each	each	PRON
ejpam-1184	201	3	of	of	ADP
ejpam-1184	201	4	these	these	PRON
ejpam-1184	201	5	k	k	NOUN
ejpam-1184	201	6	−	−	NUM
ejpam-1184	201	7	1	1	NUM
ejpam-1184	201	8	relocations	relocation	NOUN
ejpam-1184	201	9	,	,	PUNCT
ejpam-1184	201	10	re	re	VERB
ejpam-1184	201	11	-	-	VERB
ejpam-1184	201	12	fit	fit	VERB
ejpam-1184	201	13	the	the	DET
ejpam-1184	201	14	regression	regression	NOUN
ejpam-1184	201	15	models	model	NOUN
ejpam-1184	201	16	for	for	ADP
ejpam-1184	201	17	the	the	DET
ejpam-1184	201	18	changed	change	VERB
ejpam-1184	201	19	clusters	cluster	NOUN
ejpam-1184	201	20	and	and	CCONJ
ejpam-1184	201	21	calculate	calculate	VERB
ejpam-1184	201	22	the	the	DET
ejpam-1184	201	23	overall	overall	ADJ
ejpam-1184	201	24	sum	sum	NOUN
ejpam-1184	201	25	of	of	ADP
ejpam-1184	201	26	the	the	DET
ejpam-1184	201	27	squared	square	VERB
ejpam-1184	201	28	residuals	residual	NOUN
ejpam-1184	201	29	accordingly	accordingly	ADV
ejpam-1184	201	30	.	.	PUNCT
ejpam-1184	202	1	denote	denote	VERB
ejpam-1184	202	2	the	the	DET
ejpam-1184	202	3	smallest	small	ADJ
ejpam-1184	202	4	one	one	NUM
ejpam-1184	202	5	by	by	ADP
ejpam-1184	202	6	srssh	srssh	NOUN
ejpam-1184	202	7	.	.	PUNCT
ejpam-1184	203	1	if	if	SCONJ
ejpam-1184	203	2	srssh	srssh	PROPN
ejpam-1184	203	3	<	<	X
ejpam-1184	203	4	srss0	srss0	PROPN
ejpam-1184	203	5	,	,	PUNCT
ejpam-1184	203	6	redefine	redefine	VERB
ejpam-1184	203	7	c	c	PROPN
ejpam-1184	203	8	j	j	PROPN
ejpam-1184	203	9	=	=	PUNCT
ejpam-1184	203	10	c	c	PROPN
ejpam-1184	203	11	j	j	PROPN
ejpam-1184	203	12	−	−	PROPN
ejpam-1184	203	13	{	{	PUNCT
ejpam-1184	203	14	i	i	NOUN
ejpam-1184	203	15	}	}	PUNCT
ejpam-1184	203	16	,	,	PUNCT
ejpam-1184	203	17	ch	ch	NOUN
ejpam-1184	203	18	=	=	PROPN
ejpam-1184	203	19	ch+	ch+	NOUN
ejpam-1184	203	20	{	{	PUNCT
ejpam-1184	203	21	i	i	NOUN
ejpam-1184	203	22	}	}	PUNCT
ejpam-1184	203	23	,	,	PUNCT
ejpam-1184	203	24	and	and	CCONJ
ejpam-1184	203	25	set	set	VERB
ejpam-1184	203	26	srss0	srss0	PROPN
ejpam-1184	203	27	=	=	SYM
ejpam-1184	203	28	srssh	srssh	PROPN
ejpam-1184	203	29	.	.	PUNCT
ejpam-1184	204	1	otherwise	otherwise	ADV
ejpam-1184	204	2	keep	keep	VERB
ejpam-1184	204	3	i	i	PRON
ejpam-1184	204	4	in	in	ADP
ejpam-1184	204	5	c	c	PROPN
ejpam-1184	204	6	j	j	PROPN
ejpam-1184	204	7	.	.	PUNCT
ejpam-1184	205	1	(	(	PUNCT
ejpam-1184	205	2	iii	iii	NOUN
ejpam-1184	205	3	)	)	PUNCT
ejpam-1184	205	4	repeat	repeat	NOUN
ejpam-1184	205	5	(	(	PUNCT
ejpam-1184	205	6	ii	ii	NOUN
ejpam-1184	205	7	)	)	PUNCT
ejpam-1184	205	8	until	until	SCONJ
ejpam-1184	205	9	the	the	DET
ejpam-1184	205	10	objective	objective	ADJ
ejpam-1184	205	11	function	function	NOUN
ejpam-1184	205	12	(	(	PUNCT
ejpam-1184	205	13	8)	8)	NOUN
ejpam-1184	205	14	could	could	AUX
ejpam-1184	205	15	not	not	PART
ejpam-1184	205	16	be	be	AUX
ejpam-1184	205	17	reduced	reduce	VERB
ejpam-1184	205	18	any	any	PRON
ejpam-1184	205	19	further	far	ADV
ejpam-1184	205	20	,	,	PUNCT
ejpam-1184	205	21	which	which	PRON
ejpam-1184	205	22	means	mean	VERB
ejpam-1184	205	23	no	no	DET
ejpam-1184	205	24	observation	observation	NOUN
ejpam-1184	205	25	relocation	relocation	NOUN
ejpam-1184	205	26	is	be	AUX
ejpam-1184	205	27	necessary	necessary	ADJ
ejpam-1184	205	28	and	and	CCONJ
ejpam-1184	205	29	the	the	DET
ejpam-1184	205	30	optimal	optimal	ADJ
ejpam-1184	205	31	clustering	clustering	NOUN
ejpam-1184	205	32	is	be	AUX
ejpam-1184	205	33	achieved	achieve	VERB
ejpam-1184	205	34	for	for	ADP
ejpam-1184	205	35	this	this	DET
ejpam-1184	205	36	k.	k.	NOUN
ejpam-1184	206	1	the	the	DET
ejpam-1184	206	2	idea	idea	NOUN
ejpam-1184	206	3	behind	behind	ADP
ejpam-1184	206	4	the	the	DET
ejpam-1184	206	5	above	above	ADJ
ejpam-1184	206	6	algorithm	algorithm	NOUN
ejpam-1184	206	7	comes	come	VERB
ejpam-1184	206	8	from	from	ADP
ejpam-1184	206	9	[	[	X
ejpam-1184	206	10	7	7	NUM
ejpam-1184	206	11	]	]	PUNCT
ejpam-1184	206	12	.	.	PUNCT
ejpam-1184	207	1	once	once	ADV
ejpam-1184	207	2	the	the	DET
ejpam-1184	207	3	optimal	optimal	ADJ
ejpam-1184	207	4	clustering	clustering	NOUN
ejpam-1184	207	5	is	be	AUX
ejpam-1184	207	6	done	do	VERB
ejpam-1184	207	7	for	for	ADP
ejpam-1184	207	8	each	each	DET
ejpam-1184	207	9	possible	possible	ADJ
ejpam-1184	207	10	k	k	PROPN
ejpam-1184	207	11	,	,	PUNCT
ejpam-1184	207	12	the	the	DET
ejpam-1184	207	13	criterion	criterion	NOUN
ejpam-1184	207	14	ls	ls	PROPN
ejpam-1184	207	15	-	-	PUNCT
ejpam-1184	207	16	c	c	PROPN
ejpam-1184	207	17	is	be	AUX
ejpam-1184	207	18	used	use	VERB
ejpam-1184	207	19	as	as	ADP
ejpam-1184	207	20	a	a	DET
ejpam-1184	207	21	rule	rule	NOUN
ejpam-1184	207	22	to	to	PART
ejpam-1184	207	23	select	select	VERB
ejpam-1184	207	24	the	the	DET
ejpam-1184	207	25	best	good	ADJ
ejpam-1184	207	26	number	number	NOUN
ejpam-1184	207	27	of	of	ADP
ejpam-1184	207	28	clusters	cluster	NOUN
ejpam-1184	207	29	.	.	PUNCT
ejpam-1184	208	1	it	it	PRON
ejpam-1184	208	2	is	be	AUX
ejpam-1184	208	3	noted	note	VERB
ejpam-1184	208	4	that	that	SCONJ
ejpam-1184	208	5	the	the	DET
ejpam-1184	208	6	initial	initial	ADJ
ejpam-1184	208	7	partition	partition	NOUN
ejpam-1184	208	8	of	of	ADP
ejpam-1184	208	9	o	o	NOUN
ejpam-1184	208	10	=	=	PUNCT
ejpam-1184	208	11	{	{	PUNCT
ejpam-1184	208	12	1	1	NUM
ejpam-1184	208	13	,	,	PUNCT
ejpam-1184	208	14	.	.	PUNCT
ejpam-1184	208	15	.	.	PUNCT
ejpam-1184	209	1	.	.	PUNCT
ejpam-1184	210	1	,	,	PUNCT
ejpam-1184	210	2	n	n	CCONJ
ejpam-1184	210	3	}	}	PUNCT
ejpam-1184	210	4	,	,	PUNCT
ejpam-1184	210	5	if	if	SCONJ
ejpam-1184	210	6	properly	properly	ADV
ejpam-1184	210	7	set	set	VERB
ejpam-1184	210	8	,	,	PUNCT
ejpam-1184	210	9	will	will	AUX
ejpam-1184	210	10	facilitate	facilitate	VERB
ejpam-1184	210	11	the	the	DET
ejpam-1184	210	12	convergence	convergence	NOUN
ejpam-1184	210	13	and	and	CCONJ
ejpam-1184	210	14	performance	performance	NOUN
ejpam-1184	210	15	of	of	ADP
ejpam-1184	210	16	the	the	DET
ejpam-1184	210	17	algorithm	algorithm	NOUN
ejpam-1184	210	18	above	above	ADV
ejpam-1184	210	19	.	.	PUNCT
ejpam-1184	211	1	denote	denote	VERB
ejpam-1184	211	2	the	the	DET
ejpam-1184	211	3	complement	complement	NOUN
ejpam-1184	211	4	set	set	NOUN
ejpam-1184	211	5	of	of	ADP
ejpam-1184	211	6	a	a	DET
ejpam-1184	211	7	set	set	NOUN
ejpam-1184	211	8	c	c	NOUN
ejpam-1184	211	9	by	by	ADP
ejpam-1184	211	10	c	c	PROPN
ejpam-1184	211	11	c.	c.	PROPN
ejpam-1184	211	12	we	we	PRON
ejpam-1184	211	13	propose	propose	VERB
ejpam-1184	211	14	to	to	PART
ejpam-1184	211	15	generate	generate	VERB
ejpam-1184	211	16	an	an	DET
ejpam-1184	211	17	initial	initial	ADJ
ejpam-1184	211	18	partition	partition	NOUN
ejpam-1184	211	19	of	of	ADP
ejpam-1184	211	20	a	a	DET
ejpam-1184	211	21	dataset	dataset	NOUN
ejpam-1184	211	22	as	as	SCONJ
ejpam-1184	211	23	follows	follow	VERB
ejpam-1184	211	24	:	:	PUNCT
ejpam-1184	211	25	step	step	NOUN
ejpam-1184	211	26	1	1	NUM
ejpam-1184	211	27	.	.	PUNCT
ejpam-1184	212	1	consider	consider	VERB
ejpam-1184	212	2	the	the	DET
ejpam-1184	212	3	linear	linear	ADJ
ejpam-1184	212	4	model	model	NOUN
ejpam-1184	212	5	yi	yi	PROPN
ejpam-1184	213	1	=	=	PUNCT
ejpam-1184	213	2	x	x	SYM
ejpam-1184	213	3	′iβ	′iβ	PROPN
ejpam-1184	213	4	+	+	CCONJ
ejpam-1184	213	5	ei	ei	X
ejpam-1184	213	6	.	.	PROPN
ejpam-1184	213	7	(	(	PUNCT
ejpam-1184	213	8	9	9	NUM
ejpam-1184	213	9	)	)	PUNCT
ejpam-1184	213	10	based	base	VERB
ejpam-1184	213	11	on	on	ADP
ejpam-1184	213	12	the	the	DET
ejpam-1184	213	13	whole	whole	ADJ
ejpam-1184	213	14	dataset	dataset	NOUN
ejpam-1184	213	15	,	,	PUNCT
ejpam-1184	213	16	one	one	NUM
ejpam-1184	213	17	estimates	estimate	VERB
ejpam-1184	213	18	β	β	X
ejpam-1184	213	19	by	by	ADP
ejpam-1184	213	20	a	a	DET
ejpam-1184	213	21	robust	robust	ADJ
ejpam-1184	213	22	method	method	NOUN
ejpam-1184	213	23	,	,	PUNCT
ejpam-1184	213	24	e.g.	e.g.	ADV
ejpam-1184	213	25	least	least	ADJ
ejpam-1184	213	26	median	median	ADJ
ejpam-1184	213	27	squares	square	NOUN
ejpam-1184	213	28	method	method	NOUN
ejpam-1184	213	29	or	or	CCONJ
ejpam-1184	213	30	least	least	ADV
ejpam-1184	213	31	trimmed	trim	VERB
ejpam-1184	213	32	squares	square	NOUN
ejpam-1184	213	33	method	method	NOUN
ejpam-1184	213	34	[	[	X
ejpam-1184	213	35	13	13	NUM
ejpam-1184	213	36	]	]	PUNCT
ejpam-1184	213	37	.	.	PUNCT
ejpam-1184	214	1	step	step	NOUN
ejpam-1184	214	2	2	2	NUM
ejpam-1184	214	3	.	.	PUNCT
ejpam-1184	215	1	put	put	VERB
ejpam-1184	215	2	all	all	DET
ejpam-1184	215	3	data	datum	NOUN
ejpam-1184	215	4	points	point	NOUN
ejpam-1184	215	5	,	,	PUNCT
ejpam-1184	215	6	whose	whose	DET
ejpam-1184	215	7	distances	distance	NOUN
ejpam-1184	215	8	to	to	ADP
ejpam-1184	215	9	the	the	DET
ejpam-1184	215	10	regression	regression	NOUN
ejpam-1184	215	11	hyperplane	hyperplane	NOUN
ejpam-1184	215	12	estimated	estimate	VERB
ejpam-1184	215	13	in	in	ADP
ejpam-1184	215	14	step	step	NOUN
ejpam-1184	215	15	1	1	NUM
ejpam-1184	215	16	is	be	AUX
ejpam-1184	215	17	less	less	ADJ
ejpam-1184	215	18	than	than	ADP
ejpam-1184	215	19	a	a	DET
ejpam-1184	215	20	predetermined	predetermine	VERB
ejpam-1184	215	21	number	number	NOUN
ejpam-1184	215	22	,	,	PUNCT
ejpam-1184	215	23	say	say	VERB
ejpam-1184	215	24	δ	δ	PROPN
ejpam-1184	215	25	,	,	PUNCT
ejpam-1184	215	26	into	into	ADP
ejpam-1184	215	27	a	a	DET
ejpam-1184	215	28	set	set	ADJ
ejpam-1184	215	29	c1	c1	NOUN
ejpam-1184	215	30	.	.	PUNCT
ejpam-1184	216	1	if	if	SCONJ
ejpam-1184	216	2	|c1|	|c1|	PROPN
ejpam-1184	216	3	and	and	CCONJ
ejpam-1184	216	4	|c	|c	PROPN
ejpam-1184	216	5	c	c	NOUN
ejpam-1184	216	6	1|	1|	PRON
ejpam-1184	216	7	are	be	AUX
ejpam-1184	216	8	both	both	ADV
ejpam-1184	216	9	larger	large	ADJ
ejpam-1184	216	10	than	than	ADP
ejpam-1184	216	11	a	a	DET
ejpam-1184	216	12	predetermined	predetermine	VERB
ejpam-1184	216	13	integer	integer	NOUN
ejpam-1184	216	14	,	,	PUNCT
ejpam-1184	216	15	say	say	VERB
ejpam-1184	216	16	m	m	PRON
ejpam-1184	216	17	,	,	PUNCT
ejpam-1184	216	18	set	set	VERB
ejpam-1184	216	19	ℓ	ℓ	NOUN
ejpam-1184	216	20	=	=	SYM
ejpam-1184	216	21	1	1	NUM
ejpam-1184	216	22	and	and	CCONJ
ejpam-1184	216	23	go	go	VERB
ejpam-1184	216	24	to	to	ADP
ejpam-1184	216	25	the	the	DET
ejpam-1184	216	26	next	next	ADJ
ejpam-1184	216	27	step	step	NOUN
ejpam-1184	216	28	;	;	PUNCT
ejpam-1184	216	29	otherwise	otherwise	ADV
ejpam-1184	216	30	,	,	PUNCT
ejpam-1184	216	31	set	set	NOUN
ejpam-1184	216	32	ℓ	ℓ	NOUN
ejpam-1184	216	33	=	=	SYM
ejpam-1184	216	34	0	0	PUNCT
ejpam-1184	217	1	and	and	CCONJ
ejpam-1184	217	2	go	go	VERB
ejpam-1184	217	3	to	to	PART
ejpam-1184	217	4	step	step	VERB
ejpam-1184	217	5	5	5	NUM
ejpam-1184	217	6	.	.	PUNCT
ejpam-1184	218	1	step	step	NOUN
ejpam-1184	218	2	3	3	NUM
ejpam-1184	218	3	.	.	PUNCT
ejpam-1184	218	4	based	base	VERB
ejpam-1184	218	5	on	on	ADP
ejpam-1184	218	6	the	the	DET
ejpam-1184	218	7	dataset	dataset	NOUN
ejpam-1184	218	8	⋂ℓ	⋂ℓ	NOUN
ejpam-1184	219	1	i=1	i=1	PROPN
ejpam-1184	220	1	c	c	PROPN
ejpam-1184	220	2	c	c	NOUN
ejpam-1184	221	1	i	i	PRON
ejpam-1184	221	2	,	,	PUNCT
ejpam-1184	221	3	one	one	NUM
ejpam-1184	221	4	estimates	estimate	VERB
ejpam-1184	221	5	β	β	X
ejpam-1184	221	6	in	in	ADP
ejpam-1184	221	7	(	(	PUNCT
ejpam-1184	221	8	9	9	NUM
ejpam-1184	221	9	)	)	PUNCT
ejpam-1184	221	10	by	by	ADP
ejpam-1184	221	11	the	the	DET
ejpam-1184	221	12	same	same	ADJ
ejpam-1184	221	13	robust	robust	ADJ
ejpam-1184	221	14	method	method	NOUN
ejpam-1184	221	15	used	use	VERB
ejpam-1184	221	16	in	in	ADP
ejpam-1184	221	17	step	step	NOUN
ejpam-1184	221	18	1	1	NUM
ejpam-1184	221	19	.	.	PUNCT
ejpam-1184	221	20	step	step	NOUN
ejpam-1184	221	21	4	4	NUM
ejpam-1184	221	22	.	.	PUNCT
ejpam-1184	222	1	put	put	VERB
ejpam-1184	222	2	all	all	DET
ejpam-1184	222	3	data	datum	NOUN
ejpam-1184	222	4	points	point	NOUN
ejpam-1184	222	5	in	in	ADP
ejpam-1184	222	6	⋂ℓ	⋂ℓ	NOUN
ejpam-1184	222	7	i=1	i=1	PROPN
ejpam-1184	223	1	c	c	PROPN
ejpam-1184	223	2	c	c	NOUN
ejpam-1184	224	1	i	i	PRON
ejpam-1184	224	2	,	,	PUNCT
ejpam-1184	224	3	whose	whose	DET
ejpam-1184	224	4	distances	distance	NOUN
ejpam-1184	224	5	to	to	ADP
ejpam-1184	224	6	the	the	DET
ejpam-1184	224	7	regression	regression	NOUN
ejpam-1184	224	8	hyperplane	hyperplane	NOUN
ejpam-1184	224	9	estimated	estimate	VERB
ejpam-1184	224	10	in	in	ADP
ejpam-1184	224	11	step	step	NOUN
ejpam-1184	224	12	3	3	NUM
ejpam-1184	224	13	is	be	AUX
ejpam-1184	224	14	less	less	ADJ
ejpam-1184	224	15	than	than	ADP
ejpam-1184	224	16	δ	δ	PROPN
ejpam-1184	224	17	,	,	PUNCT
ejpam-1184	224	18	into	into	ADP
ejpam-1184	224	19	a	a	DET
ejpam-1184	224	20	set	set	NOUN
ejpam-1184	224	21	cℓ+1	cℓ+1	X
ejpam-1184	224	22	.	.	PUNCT
ejpam-1184	225	1	if	if	SCONJ
ejpam-1184	225	2	|cℓ+1|	|cℓ+1|	PROPN
ejpam-1184	225	3	and	and	CCONJ
ejpam-1184	225	4	|	|	ADV
ejpam-1184	225	5	⋂ℓ+1	⋂ℓ+1	ADJ
ejpam-1184	225	6	i=1	i=1	X
ejpam-1184	226	1	c	c	NOUN
ejpam-1184	227	1	c	c	NOUN
ejpam-1184	227	2	i	i	PRON
ejpam-1184	227	3	|	|	ADV
ejpam-1184	227	4	are	be	AUX
ejpam-1184	227	5	both	both	ADV
ejpam-1184	227	6	larger	large	ADJ
ejpam-1184	227	7	than	than	ADP
ejpam-1184	227	8	m	m	PRON
ejpam-1184	227	9	,	,	PUNCT
ejpam-1184	227	10	set	set	VERB
ejpam-1184	227	11	ℓ=	ℓ=	NOUN
ejpam-1184	227	12	ℓ+	ℓ+	PUNCT
ejpam-1184	227	13	1	1	NUM
ejpam-1184	227	14	and	and	CCONJ
ejpam-1184	227	15	repeat	repeat	VERB
ejpam-1184	227	16	step	step	NOUN
ejpam-1184	227	17	3	3	NUM
ejpam-1184	227	18	;	;	PUNCT
ejpam-1184	227	19	otherwise	otherwise	ADV
ejpam-1184	227	20	,	,	PUNCT
ejpam-1184	227	21	go	go	VERB
ejpam-1184	227	22	to	to	PART
ejpam-1184	227	23	step	step	VERB
ejpam-1184	227	24	5	5	NUM
ejpam-1184	227	25	.	.	PUNCT
ejpam-1184	228	1	step	step	NOUN
ejpam-1184	228	2	5	5	NUM
ejpam-1184	228	3	.	.	PUNCT
ejpam-1184	229	1	the	the	DET
ejpam-1184	229	2	initial	initial	ADJ
ejpam-1184	229	3	partition	partition	NOUN
ejpam-1184	229	4	is	be	AUX
ejpam-1184	229	5	{	{	PUNCT
ejpam-1184	229	6	c1	c1	NOUN
ejpam-1184	229	7	,	,	PUNCT
ejpam-1184	229	8	.	.	PUNCT
ejpam-1184	229	9	.	.	PUNCT
ejpam-1184	230	1	.	.	PUNCT
ejpam-1184	231	1	,	,	PUNCT
ejpam-1184	231	2	cℓ	cℓ	PROPN
ejpam-1184	231	3	,	,	PUNCT
ejpam-1184	231	4	⋂ℓ	⋂ℓ	NOUN
ejpam-1184	231	5	i=1	i=1	PROPN
ejpam-1184	232	1	c	c	PROPN
ejpam-1184	232	2	c	c	NOUN
ejpam-1184	233	1	i	i	PRON
ejpam-1184	233	2	}	}	PUNCT
ejpam-1184	233	3	if	if	SCONJ
ejpam-1184	233	4	ℓ	ℓ	NOUN
ejpam-1184	233	5	>	>	X
ejpam-1184	233	6	1	1	NUM
ejpam-1184	233	7	or	or	CCONJ
ejpam-1184	233	8	just	just	ADV
ejpam-1184	233	9	the	the	DET
ejpam-1184	233	10	whole	whole	ADJ
ejpam-1184	233	11	dataset	dataset	VERB
ejpam-1184	233	12	itself	itself	PRON
ejpam-1184	233	13	if	if	SCONJ
ejpam-1184	233	14	ℓ	ℓ	PROPN
ejpam-1184	233	15	=	=	SYM
ejpam-1184	233	16	0	0	NUM
ejpam-1184	233	17	.	.	NOUN
ejpam-1184	233	18	5	5	NUM
ejpam-1184	233	19	.	.	X
ejpam-1184	233	20	simulation	simulation	NOUN
ejpam-1184	233	21	study	study	NOUN
ejpam-1184	233	22	in	in	ADP
ejpam-1184	233	23	this	this	DET
ejpam-1184	233	24	section	section	NOUN
ejpam-1184	233	25	we	we	PRON
ejpam-1184	233	26	assess	assess	VERB
ejpam-1184	233	27	the	the	DET
ejpam-1184	233	28	finite	finite	ADJ
ejpam-1184	233	29	sample	sample	NOUN
ejpam-1184	233	30	performance	performance	NOUN
ejpam-1184	233	31	of	of	ADP
ejpam-1184	233	32	criterion	criterion	NOUN
ejpam-1184	233	33	ls	ls	PROPN
ejpam-1184	233	34	-	-	PUNCT
ejpam-1184	233	35	c	c	NOUN
ejpam-1184	233	36	together	together	ADV
ejpam-1184	233	37	with	with	ADP
ejpam-1184	233	38	the	the	DET
ejpam-1184	233	39	use	use	NOUN
ejpam-1184	233	40	of	of	ADP
ejpam-1184	233	41	the	the	DET
ejpam-1184	233	42	algorithm	algorithm	NOUN
ejpam-1184	233	43	in	in	ADP
ejpam-1184	233	44	the	the	DET
ejpam-1184	233	45	previous	previous	ADJ
ejpam-1184	233	46	section	section	NOUN
ejpam-1184	233	47	.	.	PUNCT
ejpam-1184	234	1	simulated	simulate	VERB
ejpam-1184	234	2	data	data	NOUN
ejpam-1184	234	3	sets	set	NOUN
ejpam-1184	234	4	are	be	AUX
ejpam-1184	234	5	to	to	PART
ejpam-1184	234	6	be	be	AUX
ejpam-1184	234	7	used	use	VERB
ejpam-1184	234	8	to	to	PART
ejpam-1184	234	9	perform	perform	VERB
ejpam-1184	234	10	regression	regression	NOUN
ejpam-1184	234	11	clustering	cluster	VERB
ejpam-1184	234	12	for	for	ADP
ejpam-1184	234	13	the	the	DET
ejpam-1184	234	14	assessment	assessment	NOUN
ejpam-1184	234	15	.	.	PUNCT
ejpam-1184	235	1	while	while	SCONJ
ejpam-1184	235	2	many	many	ADJ
ejpam-1184	235	3	types	type	NOUN
ejpam-1184	235	4	of	of	ADP
ejpam-1184	235	5	data	datum	NOUN
ejpam-1184	235	6	sets	set	NOUN
ejpam-1184	235	7	can	can	AUX
ejpam-1184	235	8	be	be	AUX
ejpam-1184	235	9	simulated	simulate	VERB
ejpam-1184	235	10	,	,	PUNCT
ejpam-1184	235	11	we	we	PRON
ejpam-1184	235	12	consider	consider	VERB
ejpam-1184	235	13	only	only	ADV
ejpam-1184	235	14	two	two	NUM
ejpam-1184	235	15	factors	factor	NOUN
ejpam-1184	235	16	in	in	ADP
ejpam-1184	235	17	determining	determine	VERB
ejpam-1184	235	18	the	the	DET
ejpam-1184	235	19	type	type	NOUN
ejpam-1184	235	20	:	:	PUNCT
ejpam-1184	235	21	number	number	NOUN
ejpam-1184	235	22	of	of	ADP
ejpam-1184	235	23	clusters	cluster	NOUN
ejpam-1184	235	24	(	(	PUNCT
ejpam-1184	235	25	2	2	NUM
ejpam-1184	235	26	or	or	CCONJ
ejpam-1184	235	27	3	3	NUM
ejpam-1184	235	28	)	)	PUNCT
ejpam-1184	235	29	,	,	PUNCT
ejpam-1184	235	30	and	and	CCONJ
ejpam-1184	235	31	error	error	NOUN
ejpam-1184	235	32	distributions	distribution	NOUN
ejpam-1184	235	33	(	(	PUNCT
ejpam-1184	235	34	standard	standard	ADJ
ejpam-1184	235	35	normal	normal	ADJ
ejpam-1184	235	36	n(0,1	n(0,1	NOUN
ejpam-1184	235	37	)	)	PUNCT
ejpam-1184	235	38	or	or	CCONJ
ejpam-1184	235	39	t(3	t(3	PROPN
ejpam-1184	235	40	)	)	PUNCT
ejpam-1184	235	41	)	)	PUNCT
ejpam-1184	235	42	,	,	PUNCT
ejpam-1184	235	43	so	so	CCONJ
ejpam-1184	235	44	there	there	PRON
ejpam-1184	235	45	are	be	VERB
ejpam-1184	235	46	in	in	ADP
ejpam-1184	235	47	total	total	ADJ
ejpam-1184	235	48	4	4	NUM
ejpam-1184	235	49	cases	case	NOUN
ejpam-1184	235	50	of	of	ADP
ejpam-1184	235	51	data	datum	NOUN
ejpam-1184	235	52	to	to	PART
ejpam-1184	235	53	be	be	AUX
ejpam-1184	235	54	considered	consider	VERB
ejpam-1184	235	55	,	,	PUNCT
ejpam-1184	235	56	which	which	PRON
ejpam-1184	235	57	are	be	AUX
ejpam-1184	235	58	summarized	summarize	VERB
ejpam-1184	235	59	in	in	ADP
ejpam-1184	235	60	table	table	NOUN
ejpam-1184	235	61	1	1	NUM
ejpam-1184	235	62	.	.	PUNCT
ejpam-1184	236	1	g.	g.	PROPN
ejpam-1184	236	2	qian	qian	PROPN
ejpam-1184	236	3	,	,	PUNCT
ejpam-1184	236	4	y.	y.	PROPN
ejpam-1184	236	5	wu	wu	PROPN
ejpam-1184	236	6	/	/	SYM
ejpam-1184	236	7	eur	eur	PROPN
ejpam-1184	236	8	.	.	PUNCT
ejpam-1184	237	1	j.	j.	PROPN
ejpam-1184	237	2	pure	pure	PROPN
ejpam-1184	237	3	appl	appl	PROPN
ejpam-1184	237	4	.	.	PROPN
ejpam-1184	237	5	math	math	PROPN
ejpam-1184	237	6	,	,	PUNCT
ejpam-1184	237	7	4	4	NUM
ejpam-1184	237	8	(	(	PUNCT
ejpam-1184	237	9	2011	2011	NUM
ejpam-1184	237	10	)	)	PUNCT
ejpam-1184	237	11	,	,	PUNCT
ejpam-1184	237	12	455	455	NUM
ejpam-1184	237	13	-	-	SYM
ejpam-1184	237	14	466	466	NUM
ejpam-1184	237	15	462	462	NUM
ejpam-1184	237	16	there	there	PRON
ejpam-1184	237	17	will	will	AUX
ejpam-1184	237	18	be	be	AUX
ejpam-1184	237	19	only	only	ADV
ejpam-1184	237	20	one	one	NUM
ejpam-1184	237	21	covariate	covariate	NOUN
ejpam-1184	237	22	involved	involve	VERB
ejpam-1184	237	23	in	in	ADP
ejpam-1184	237	24	the	the	DET
ejpam-1184	237	25	regression	regression	NOUN
ejpam-1184	237	26	in	in	ADP
ejpam-1184	237	27	each	each	DET
ejpam-1184	237	28	cluster	cluster	NOUN
ejpam-1184	237	29	,	,	PUNCT
ejpam-1184	237	30	and	and	CCONJ
ejpam-1184	237	31	the	the	DET
ejpam-1184	237	32	covariate	covariate	NOUN
ejpam-1184	237	33	is	be	AUX
ejpam-1184	237	34	generated	generate	VERB
ejpam-1184	237	35	from	from	ADP
ejpam-1184	237	36	n(0,1	n(0,1	NOUN
ejpam-1184	237	37	)	)	PUNCT
ejpam-1184	237	38	.	.	PUNCT
ejpam-1184	238	1	the	the	DET
ejpam-1184	238	2	parameters	parameter	NOUN
ejpam-1184	238	3	used	use	VERB
ejpam-1184	238	4	for	for	ADP
ejpam-1184	238	5	each	each	DET
ejpam-1184	238	6	case	case	NOUN
ejpam-1184	238	7	are	be	AUX
ejpam-1184	238	8	given	give	VERB
ejpam-1184	238	9	in	in	ADP
ejpam-1184	238	10	table	table	NOUN
ejpam-1184	238	11	2	2	NUM
ejpam-1184	238	12	.	.	PUNCT
ejpam-1184	239	1	then	then	ADV
ejpam-1184	239	2	the	the	DET
ejpam-1184	239	3	fixed	fixed	ADJ
ejpam-1184	239	4	partition	partition	NOUN
ejpam-1184	239	5	regression	regression	NOUN
ejpam-1184	239	6	clustering	cluster	VERB
ejpam-1184	239	7	model	model	NOUN
ejpam-1184	239	8	y	y	PROPN
ejpam-1184	239	9	ji	ji	PROPN
ejpam-1184	240	1	=	=	PUNCT
ejpam-1184	240	2	x	x	PUNCT
ejpam-1184	240	3	′	′	NUM
ejpam-1184	241	1	ji	ji	INTJ
ejpam-1184	241	2	β0i	β0i	PUNCT
ejpam-1184	242	1	+	+	CCONJ
ejpam-1184	242	2	e	e	X
ejpam-1184	242	3	ji	ji	PROPN
ejpam-1184	242	4	,	,	PUNCT
ejpam-1184	242	5	j	j	PROPN
ejpam-1184	242	6	=	=	SYM
ejpam-1184	242	7	1	1	NUM
ejpam-1184	242	8	,	,	PUNCT
ejpam-1184	242	9	.	.	PUNCT
ejpam-1184	242	10	.	.	PUNCT
ejpam-1184	243	1	.	.	PUNCT
ejpam-1184	244	1	,	,	PUNCT
ejpam-1184	244	2	ni	ni	PROPN
ejpam-1184	244	3	,	,	PUNCT
ejpam-1184	244	4	i	i	NOUN
ejpam-1184	244	5	=	=	NOUN
ejpam-1184	244	6	1	1	NUM
ejpam-1184	244	7	,	,	PUNCT
ejpam-1184	244	8	.	.	PUNCT
ejpam-1184	244	9	.	.	PUNCT
ejpam-1184	245	1	.	.	PUNCT
ejpam-1184	246	1	,	,	PUNCT
ejpam-1184	246	2	k0	k0	PROPN
ejpam-1184	246	3	is	be	AUX
ejpam-1184	246	4	applied	apply	VERB
ejpam-1184	246	5	to	to	PART
ejpam-1184	246	6	generate	generate	VERB
ejpam-1184	246	7	the	the	DET
ejpam-1184	246	8	response	response	NOUN
ejpam-1184	246	9	values	value	VERB
ejpam-1184	246	10	y	y	PROPN
ejpam-1184	246	11	ji	ji	PROPN
ejpam-1184	246	12	,	,	PUNCT
ejpam-1184	246	13	where	where	SCONJ
ejpam-1184	246	14	e	e	X
ejpam-1184	246	15	ji	ji	PROPN
ejpam-1184	246	16	is	be	AUX
ejpam-1184	246	17	a	a	DET
ejpam-1184	246	18	random	random	ADJ
ejpam-1184	246	19	number	number	NOUN
ejpam-1184	246	20	originating	originate	VERB
ejpam-1184	246	21	from	from	ADP
ejpam-1184	246	22	n(0,1	n(0,1	NOUN
ejpam-1184	246	23	)	)	PUNCT
ejpam-1184	246	24	or	or	CCONJ
ejpam-1184	246	25	t(3	t(3	PROPN
ejpam-1184	246	26	)	)	PUNCT
ejpam-1184	246	27	,	,	PUNCT
ejpam-1184	246	28	and	and	CCONJ
ejpam-1184	246	29	the	the	DET
ejpam-1184	246	30	first	first	ADJ
ejpam-1184	246	31	element	element	NOUN
ejpam-1184	246	32	of	of	ADP
ejpam-1184	246	33	x	x	INTJ
ejpam-1184	246	34	ji	ji	PROPN
ejpam-1184	246	35	is	be	AUX
ejpam-1184	246	36	the	the	DET
ejpam-1184	246	37	constant	constant	ADJ
ejpam-1184	246	38	1	1	NUM
ejpam-1184	246	39	corresponding	correspond	VERB
ejpam-1184	246	40	to	to	ADP
ejpam-1184	246	41	the	the	DET
ejpam-1184	246	42	intercept	intercept	NOUN
ejpam-1184	246	43	term	term	NOUN
ejpam-1184	246	44	in	in	ADP
ejpam-1184	246	45	the	the	DET
ejpam-1184	246	46	model	model	NOUN
ejpam-1184	246	47	.	.	PUNCT
ejpam-1184	247	1	table	table	NOUN
ejpam-1184	247	2	1	1	NUM
ejpam-1184	247	3	:	:	PUNCT
ejpam-1184	247	4	shorthand	shorthand	NOUN
ejpam-1184	247	5	notation	notation	NOUN
ejpam-1184	247	6	for	for	ADP
ejpam-1184	247	7	the	the	DET
ejpam-1184	247	8	four	four	NUM
ejpam-1184	247	9	cases	case	NOUN
ejpam-1184	247	10	.	.	PUNCT
ejpam-1184	248	1	n1c2	n1c2	INTJ
ejpam-1184	248	2	case	case	NOUN
ejpam-1184	248	3	1	1	NUM
ejpam-1184	248	4	,	,	PUNCT
ejpam-1184	248	5	two	two	NUM
ejpam-1184	248	6	regression	regression	NOUN
ejpam-1184	248	7	lines	line	NOUN
ejpam-1184	248	8	normal	normal	ADJ
ejpam-1184	248	9	error	error	NOUN
ejpam-1184	248	10	t1c2	t1c2	NOUN
ejpam-1184	248	11	case	case	NOUN
ejpam-1184	248	12	2	2	NUM
ejpam-1184	248	13	,	,	PUNCT
ejpam-1184	248	14	two	two	NUM
ejpam-1184	248	15	regression	regression	NOUN
ejpam-1184	248	16	lines	line	NOUN
ejpam-1184	248	17	t(3	t(3	PROPN
ejpam-1184	248	18	)	)	PUNCT
ejpam-1184	248	19	error	error	NOUN
ejpam-1184	248	20	n1c3	n1c3	NOUN
ejpam-1184	248	21	case	case	NOUN
ejpam-1184	248	22	3	3	NUM
ejpam-1184	248	23	,	,	PUNCT
ejpam-1184	248	24	three	three	NUM
ejpam-1184	248	25	regression	regression	NOUN
ejpam-1184	248	26	lines	line	NOUN
ejpam-1184	248	27	normal	normal	ADJ
ejpam-1184	248	28	error	error	NOUN
ejpam-1184	248	29	t1c3	t1c3	NOUN
ejpam-1184	248	30	case	case	NOUN
ejpam-1184	248	31	4	4	NUM
ejpam-1184	248	32	,	,	PUNCT
ejpam-1184	248	33	three	three	NUM
ejpam-1184	248	34	regression	regression	NOUN
ejpam-1184	248	35	lines	line	NOUN
ejpam-1184	248	36	t(3	t(3	PROPN
ejpam-1184	248	37	)	)	PUNCT
ejpam-1184	248	38	error	error	NOUN
ejpam-1184	248	39	table	table	NOUN
ejpam-1184	248	40	2	2	NUM
ejpam-1184	248	41	:	:	PUNCT
ejpam-1184	248	42	parameter	parameter	NOUN
ejpam-1184	248	43	values	value	NOUN
ejpam-1184	248	44	used	use	VERB
ejpam-1184	248	45	in	in	ADP
ejpam-1184	248	46	the	the	DET
ejpam-1184	248	47	simulation	simulation	NOUN
ejpam-1184	248	48	study	study	NOUN
ejpam-1184	248	49	of	of	ADP
ejpam-1184	248	50	regression	regression	NOUN
ejpam-1184	248	51	clustering	cluster	VERB
ejpam-1184	248	52	.	.	PUNCT
ejpam-1184	249	1	case	case	NOUN
ejpam-1184	249	2	k0	k0	PROPN
ejpam-1184	249	3	regression	regression	PROPN
ejpam-1184	249	4	coefficients	coefficient	VERB
ejpam-1184	249	5	no	no	INTJ
ejpam-1184	249	6	.	.	PUNCT
ejpam-1184	250	1	of	of	ADP
ejpam-1184	250	2	obs	obs	PROPN
ejpam-1184	250	3	.	.	PUNCT
ejpam-1184	251	1	1–2	1–2	NUM
ejpam-1184	251	2	2	2	NUM
ejpam-1184	251	3	β01	β01	NOUN
ejpam-1184	251	4	=	=	SYM
ejpam-1184	251	5	�	�	PROPN
ejpam-1184	251	6	2	2	NUM
ejpam-1184	251	7	8	8	NUM
ejpam-1184	251	8	�	�	PROPN
ejpam-1184	251	9	,	,	PUNCT
ejpam-1184	251	10	β02	β02	PROPN
ejpam-1184	251	11	=	=	SYM
ejpam-1184	251	12	�	�	PROPN
ejpam-1184	251	13	1	1	NUM
ejpam-1184	251	14	5	5	NUM
ejpam-1184	251	15	�	�	PROPN
ejpam-1184	251	16	n1	n1	PROPN
ejpam-1184	251	17	=	=	SYM
ejpam-1184	251	18	70	70	NUM
ejpam-1184	251	19	n2	n2	NOUN
ejpam-1184	251	20	=	=	NOUN
ejpam-1184	251	21	50	50	NUM
ejpam-1184	251	22	3–4	3–4	NUM
ejpam-1184	251	23	3	3	NUM
ejpam-1184	251	24	β01	β01	NOUN
ejpam-1184	251	25	=	=	SYM
ejpam-1184	251	26	�	�	PROPN
ejpam-1184	251	27	18	18	NUM
ejpam-1184	251	28	6	6	NUM
ejpam-1184	251	29	�	�	PROPN
ejpam-1184	251	30	,	,	PUNCT
ejpam-1184	251	31	β02	β02	PROPN
ejpam-1184	251	32	=	=	SYM
ejpam-1184	251	33	�	�	PROPN
ejpam-1184	251	34	12	12	NUM
ejpam-1184	251	35	8	8	NUM
ejpam-1184	251	36	�	�	PROPN
ejpam-1184	251	37	,	,	PUNCT
ejpam-1184	251	38	β03	β03	NUM
ejpam-1184	251	39	=	=	SYM
ejpam-1184	251	40	�	�	PROPN
ejpam-1184	251	41	15	15	NUM
ejpam-1184	251	42	−2	−2	PROPN
ejpam-1184	251	43	�	�	PROPN
ejpam-1184	251	44	n1	n1	NOUN
ejpam-1184	251	45	=	=	SYM
ejpam-1184	251	46	35	35	NUM
ejpam-1184	251	47	n2	n2	NOUN
ejpam-1184	251	48	=	=	SYM
ejpam-1184	251	49	35	35	NUM
ejpam-1184	251	50	n3	n3	NOUN
ejpam-1184	251	51	=	=	SYM
ejpam-1184	251	52	50	50	NUM
ejpam-1184	251	53	figures	figure	NOUN
ejpam-1184	251	54	2	2	NUM
ejpam-1184	251	55	and	and	CCONJ
ejpam-1184	251	56	3	3	NUM
ejpam-1184	251	57	illustrate	illustrate	VERB
ejpam-1184	251	58	what	what	PRON
ejpam-1184	251	59	the	the	DET
ejpam-1184	251	60	data	datum	NOUN
ejpam-1184	251	61	typically	typically	ADV
ejpam-1184	251	62	would	would	AUX
ejpam-1184	251	63	look	look	VERB
ejpam-1184	251	64	like	like	ADP
ejpam-1184	251	65	for	for	ADP
ejpam-1184	251	66	cases	case	NOUN
ejpam-1184	251	67	1	1	NUM
ejpam-1184	251	68	to	to	PART
ejpam-1184	251	69	4	4	NUM
ejpam-1184	251	70	with	with	ADP
ejpam-1184	251	71	normal	normal	ADJ
ejpam-1184	251	72	or	or	CCONJ
ejpam-1184	251	73	t(3	t(3	PROPN
ejpam-1184	251	74	)	)	PUNCT
ejpam-1184	251	75	errors	error	NOUN
ejpam-1184	251	76	.	.	PUNCT
ejpam-1184	252	1	these	these	DET
ejpam-1184	252	2	figures	figure	NOUN
ejpam-1184	252	3	show	show	VERB
ejpam-1184	252	4	that	that	SCONJ
ejpam-1184	252	5	the	the	DET
ejpam-1184	252	6	groupings	grouping	NOUN
ejpam-1184	252	7	of	of	ADP
ejpam-1184	252	8	the	the	DET
ejpam-1184	252	9	linear	linear	ADJ
ejpam-1184	252	10	patterns	pattern	NOUN
ejpam-1184	252	11	are	be	AUX
ejpam-1184	252	12	visible	visible	ADJ
ejpam-1184	252	13	with	with	ADP
ejpam-1184	252	14	standard	standard	ADJ
ejpam-1184	252	15	normal	normal	ADJ
ejpam-1184	252	16	random	random	ADJ
ejpam-1184	252	17	errors	error	NOUN
ejpam-1184	252	18	and	and	CCONJ
ejpam-1184	252	19	getting	get	VERB
ejpam-1184	252	20	worse	bad	ADJ
ejpam-1184	252	21	with	with	ADP
ejpam-1184	252	22	t(3	t(3	PROPN
ejpam-1184	252	23	)	)	PUNCT
ejpam-1184	252	24	random	random	ADJ
ejpam-1184	252	25	errors	error	NOUN
ejpam-1184	252	26	.	.	PUNCT
ejpam-1184	253	1	−2	−2	X
ejpam-1184	253	2	−1	−1	NOUN
ejpam-1184	253	3	0	0	NUM
ejpam-1184	253	4	1	1	NUM
ejpam-1184	253	5	2	2	NUM
ejpam-1184	253	6	−	−	NUM
ejpam-1184	253	7	20	20	NUM
ejpam-1184	253	8	−	−	NOUN
ejpam-1184	253	9	10	10	NUM
ejpam-1184	253	10	0	0	NUM
ejpam-1184	253	11	10	10	NUM
ejpam-1184	253	12	20	20	NUM
ejpam-1184	253	13	x	x	SYM
ejpam-1184	253	14	y	y	PROPN
ejpam-1184	253	15	(	(	PUNCT
ejpam-1184	253	16	a	a	PRON
ejpam-1184	253	17	)	)	PUNCT
ejpam-1184	253	18	case	case	NOUN
ejpam-1184	253	19	1	1	NUM
ejpam-1184	253	20	−2	−2	NOUN
ejpam-1184	253	21	−1	−1	NOUN
ejpam-1184	253	22	0	0	NUM
ejpam-1184	253	23	1	1	NUM
ejpam-1184	253	24	2	2	NUM
ejpam-1184	253	25	−	−	NUM
ejpam-1184	253	26	15	15	NUM
ejpam-1184	253	27	−	−	NOUN
ejpam-1184	253	28	5	5	NUM
ejpam-1184	253	29	0	0	NUM
ejpam-1184	253	30	5	5	NUM
ejpam-1184	253	31	10	10	NUM
ejpam-1184	253	32	15	15	NUM
ejpam-1184	253	33	20	20	NUM
ejpam-1184	253	34	x	x	SYM
ejpam-1184	253	35	y	y	PROPN
ejpam-1184	253	36	(	(	PUNCT
ejpam-1184	253	37	b	b	NOUN
ejpam-1184	253	38	)	)	PUNCT
ejpam-1184	253	39	case	case	NOUN
ejpam-1184	253	40	2	2	NUM
ejpam-1184	253	41	figure	figure	NOUN
ejpam-1184	253	42	2	2	NUM
ejpam-1184	253	43	:	:	PUNCT
ejpam-1184	253	44	simulated	simulate	VERB
ejpam-1184	253	45	data	datum	NOUN
ejpam-1184	253	46	with	with	ADP
ejpam-1184	253	47	two	two	NUM
ejpam-1184	253	48	clusters	cluster	NOUN
ejpam-1184	253	49	.	.	PUNCT
ejpam-1184	254	1	g.	g.	PROPN
ejpam-1184	254	2	qian	qian	PROPN
ejpam-1184	254	3	,	,	PUNCT
ejpam-1184	254	4	y.	y.	PROPN
ejpam-1184	254	5	wu	wu	PROPN
ejpam-1184	254	6	/	/	SYM
ejpam-1184	254	7	eur	eur	PROPN
ejpam-1184	254	8	.	.	PUNCT
ejpam-1184	255	1	j.	j.	PROPN
ejpam-1184	255	2	pure	pure	PROPN
ejpam-1184	255	3	appl	appl	PROPN
ejpam-1184	255	4	.	.	PROPN
ejpam-1184	255	5	math	math	PROPN
ejpam-1184	255	6	,	,	PUNCT
ejpam-1184	255	7	4	4	NUM
ejpam-1184	255	8	(	(	PUNCT
ejpam-1184	255	9	2011	2011	NUM
ejpam-1184	255	10	)	)	PUNCT
ejpam-1184	255	11	,	,	PUNCT
ejpam-1184	255	12	455	455	NUM
ejpam-1184	255	13	-	-	SYM
ejpam-1184	255	14	466	466	NUM
ejpam-1184	255	15	463	463	NUM
ejpam-1184	255	16	−2	−2	NOUN
ejpam-1184	255	17	−1	−1	NOUN
ejpam-1184	255	18	0	0	NUM
ejpam-1184	255	19	1	1	NUM
ejpam-1184	255	20	2	2	NUM
ejpam-1184	255	21	−	−	NUM
ejpam-1184	255	22	10	10	NUM
ejpam-1184	255	23	0	0	NUM
ejpam-1184	255	24	10	10	NUM
ejpam-1184	255	25	20	20	NUM
ejpam-1184	255	26	30	30	NUM
ejpam-1184	255	27	x	x	SYM
ejpam-1184	255	28	y	y	PROPN
ejpam-1184	255	29	*	*	PUNCT
ejpam-1184	256	1	*	*	PUNCT
ejpam-1184	257	1	*	*	PUNCT
ejpam-1184	258	1	*	*	PUNCT
ejpam-1184	259	1	*	*	PUNCT
ejpam-1184	260	1	*	*	PUNCT
ejpam-1184	261	1	*	*	PUNCT
ejpam-1184	262	1	*	*	PUNCT
ejpam-1184	263	1	*	*	PUNCT
ejpam-1184	263	2	*	*	PUNCT
ejpam-1184	264	1	*	*	PUNCT
ejpam-1184	265	1	*	*	PUNCT
ejpam-1184	266	1	*	*	PUNCT
ejpam-1184	267	1	*	*	PUNCT
ejpam-1184	268	1	*	*	PUNCT
ejpam-1184	269	1	*	*	PUNCT
ejpam-1184	270	1	*	*	PUNCT
ejpam-1184	270	2	*	*	PUNCT
ejpam-1184	271	1	*	*	PUNCT
ejpam-1184	272	1	*	*	PUNCT
ejpam-1184	272	2	*	*	PUNCT
ejpam-1184	273	1	*	*	PUNCT
ejpam-1184	274	1	*	*	PUNCT
ejpam-1184	275	1	*	*	PUNCT
ejpam-1184	276	1	*	*	PUNCT
ejpam-1184	276	2	*	*	PUNCT
ejpam-1184	277	1	*	*	PUNCT
ejpam-1184	277	2	*	*	PUNCT
ejpam-1184	278	1	*	*	PUNCT
ejpam-1184	279	1	*	*	PUNCT
ejpam-1184	280	1	*	*	PUNCT
ejpam-1184	281	1	*	*	PUNCT
ejpam-1184	282	1	*	*	PUNCT
ejpam-1184	283	1	*	*	PUNCT
ejpam-1184	284	1	*	*	PUNCT
ejpam-1184	285	1	*	*	PUNCT
ejpam-1184	286	1	*	*	PUNCT
ejpam-1184	286	2	*	*	PUNCT
ejpam-1184	287	1	*	*	PUNCT
ejpam-1184	288	1	*	*	PUNCT
ejpam-1184	289	1	*	*	PUNCT
ejpam-1184	290	1	*	*	PUNCT
ejpam-1184	291	1	*	*	PUNCT
ejpam-1184	291	2	*	*	PUNCT
ejpam-1184	292	1	*	*	PUNCT
ejpam-1184	293	1	*	*	PUNCT
ejpam-1184	294	1	*	*	PUNCT
ejpam-1184	295	1	*	*	PUNCT
ejpam-1184	296	1	*	*	PUNCT
ejpam-1184	296	2	*	*	PUNCT
ejpam-1184	296	3	(	(	PUNCT
ejpam-1184	296	4	a	a	PRON
ejpam-1184	296	5	)	)	PUNCT
ejpam-1184	296	6	case	case	NOUN
ejpam-1184	296	7	3	3	NUM
ejpam-1184	296	8	−2	−2	NOUN
ejpam-1184	296	9	−1	−1	NOUN
ejpam-1184	296	10	0	0	NUM
ejpam-1184	296	11	1	1	NUM
ejpam-1184	296	12	2	2	NUM
ejpam-1184	296	13	−	−	NUM
ejpam-1184	296	14	10	10	NUM
ejpam-1184	296	15	0	0	NUM
ejpam-1184	296	16	10	10	NUM
ejpam-1184	296	17	20	20	NUM
ejpam-1184	296	18	30	30	NUM
ejpam-1184	296	19	x	x	SYM
ejpam-1184	296	20	y	y	PROPN
ejpam-1184	296	21	*	*	PUNCT
ejpam-1184	296	22	*	*	PUNCT
ejpam-1184	297	1	*	*	PUNCT
ejpam-1184	298	1	*	*	PUNCT
ejpam-1184	299	1	*	*	PUNCT
ejpam-1184	300	1	*	*	PUNCT
ejpam-1184	301	1	*	*	PUNCT
ejpam-1184	302	1	*	*	PUNCT
ejpam-1184	303	1	*	*	PUNCT
ejpam-1184	304	1	*	*	PUNCT
ejpam-1184	305	1	*	*	PUNCT
ejpam-1184	306	1	*	*	PUNCT
ejpam-1184	307	1	*	*	PUNCT
ejpam-1184	308	1	*	*	PUNCT
ejpam-1184	309	1	*	*	PUNCT
ejpam-1184	310	1	*	*	PUNCT
ejpam-1184	311	1	*	*	PUNCT
ejpam-1184	311	2	*	*	PUNCT
ejpam-1184	312	1	*	*	PUNCT
ejpam-1184	313	1	*	*	PUNCT
ejpam-1184	314	1	*	*	PUNCT
ejpam-1184	315	1	*	*	PUNCT
ejpam-1184	316	1	*	*	PUNCT
ejpam-1184	317	1	*	*	PUNCT
ejpam-1184	318	1	*	*	PUNCT
ejpam-1184	319	1	*	*	PUNCT
ejpam-1184	320	1	*	*	PUNCT
ejpam-1184	321	1	*	*	PUNCT
ejpam-1184	322	1	*	*	PUNCT
ejpam-1184	323	1	*	*	PUNCT
ejpam-1184	324	1	*	*	PUNCT
ejpam-1184	325	1	*	*	PUNCT
ejpam-1184	326	1	*	*	PUNCT
ejpam-1184	327	1	*	*	PUNCT
ejpam-1184	328	1	*	*	PUNCT
ejpam-1184	329	1	*	*	PUNCT
ejpam-1184	330	1	*	*	PUNCT
ejpam-1184	331	1	*	*	PUNCT
ejpam-1184	331	2	*	*	PUNCT
ejpam-1184	332	1	*	*	PUNCT
ejpam-1184	333	1	*	*	PUNCT
ejpam-1184	334	1	*	*	PUNCT
ejpam-1184	335	1	*	*	PUNCT
ejpam-1184	335	2	*	*	PUNCT
ejpam-1184	336	1	*	*	PUNCT
ejpam-1184	337	1	*	*	PUNCT
ejpam-1184	338	1	*	*	PUNCT
ejpam-1184	338	2	*	*	PUNCT
ejpam-1184	338	3	*	*	PUNCT
ejpam-1184	338	4	*	*	PUNCT
ejpam-1184	338	5	(	(	PUNCT
ejpam-1184	338	6	b	b	NOUN
ejpam-1184	338	7	)	)	PUNCT
ejpam-1184	338	8	case	case	NOUN
ejpam-1184	338	9	4	4	NUM
ejpam-1184	338	10	figure	figure	NOUN
ejpam-1184	338	11	3	3	NUM
ejpam-1184	338	12	:	:	PUNCT
ejpam-1184	338	13	simulated	simulate	VERB
ejpam-1184	338	14	data	datum	NOUN
ejpam-1184	338	15	with	with	ADP
ejpam-1184	338	16	three	three	NUM
ejpam-1184	338	17	clusters	cluster	NOUN
ejpam-1184	338	18	.	.	PUNCT
ejpam-1184	339	1	in	in	ADP
ejpam-1184	339	2	this	this	DET
ejpam-1184	339	3	study	study	NOUN
ejpam-1184	339	4	,	,	PUNCT
ejpam-1184	339	5	we	we	PRON
ejpam-1184	339	6	set	set	VERB
ejpam-1184	339	7	q(k	q(k	NOUN
ejpam-1184	339	8	)	)	PUNCT
ejpam-1184	340	1	=	=	SYM
ejpam-1184	340	2	kp	kp	NOUN
ejpam-1184	340	3	in	in	ADP
ejpam-1184	340	4	criterion	criterion	NOUN
ejpam-1184	340	5	ls	ls	PROPN
ejpam-1184	340	6	-	-	PROPN
ejpam-1184	340	7	c	c	NOUN
ejpam-1184	340	8	(	(	PUNCT
ejpam-1184	340	9	6	6	NUM
ejpam-1184	340	10	)	)	PUNCT
ejpam-1184	340	11	,	,	PUNCT
ejpam-1184	340	12	where	where	SCONJ
ejpam-1184	340	13	p	p	NOUN
ejpam-1184	340	14	is	be	AUX
ejpam-1184	340	15	the	the	DET
ejpam-1184	340	16	number	number	NOUN
ejpam-1184	340	17	of	of	ADP
ejpam-1184	340	18	regression	regression	NOUN
ejpam-1184	340	19	coefficients	coefficient	NOUN
ejpam-1184	340	20	in	in	ADP
ejpam-1184	340	21	the	the	DET
ejpam-1184	340	22	model	model	NOUN
ejpam-1184	340	23	and	and	CCONJ
ejpam-1184	340	24	is	be	AUX
ejpam-1184	340	25	a	a	DET
ejpam-1184	340	26	constant	constant	ADJ
ejpam-1184	340	27	in	in	ADP
ejpam-1184	340	28	our	our	PRON
ejpam-1184	340	29	study	study	NOUN
ejpam-1184	340	30	;	;	PUNCT
ejpam-1184	340	31	and	and	CCONJ
ejpam-1184	340	32	k	k	PROPN
ejpam-1184	340	33	is	be	AUX
ejpam-1184	340	34	the	the	DET
ejpam-1184	340	35	number	number	NOUN
ejpam-1184	340	36	of	of	ADP
ejpam-1184	340	37	clusters	cluster	NOUN
ejpam-1184	340	38	used	use	VERB
ejpam-1184	340	39	in	in	ADP
ejpam-1184	340	40	the	the	DET
ejpam-1184	340	41	regression	regression	NOUN
ejpam-1184	340	42	clustering	cluster	VERB
ejpam-1184	340	43	under	under	ADP
ejpam-1184	340	44	assessment	assessment	NOUN
ejpam-1184	340	45	.	.	PUNCT
ejpam-1184	341	1	it	it	PRON
ejpam-1184	341	2	is	be	AUX
ejpam-1184	341	3	noted	note	VERB
ejpam-1184	341	4	that	that	SCONJ
ejpam-1184	341	5	in	in	ADP
ejpam-1184	341	6	an	an	DET
ejpam-1184	341	7	information	information	NOUN
ejpam-1184	341	8	model	model	NOUN
ejpam-1184	341	9	selection	selection	NOUN
ejpam-1184	341	10	criterion	criterion	NOUN
ejpam-1184	341	11	,	,	PUNCT
ejpam-1184	341	12	a	a	DET
ejpam-1184	341	13	penalty	penalty	NOUN
ejpam-1184	341	14	function	function	NOUN
ejpam-1184	341	15	,	,	PUNCT
ejpam-1184	341	16	which	which	PRON
ejpam-1184	341	17	is	be	AUX
ejpam-1184	341	18	an	an	DET
ejpam-1184	341	19	in	in	ADP
ejpam-1184	341	20	(	(	PUNCT
ejpam-1184	341	21	6	6	NUM
ejpam-1184	341	22	)	)	PUNCT
ejpam-1184	341	23	,	,	PUNCT
ejpam-1184	341	24	is	be	AUX
ejpam-1184	341	25	usually	usually	ADV
ejpam-1184	341	26	chosen	choose	VERB
ejpam-1184	341	27	as	as	ADP
ejpam-1184	341	28	c	c	PROPN
ejpam-1184	341	29	log(n	log(n	PROPN
ejpam-1184	341	30	)	)	PUNCT
ejpam-1184	341	31	or	or	CCONJ
ejpam-1184	341	32	c	c	AUX
ejpam-1184	341	33	log	log	VERB
ejpam-1184	341	34	log(n	log(n	NOUN
ejpam-1184	341	35	)	)	PUNCT
ejpam-1184	341	36	with	with	ADP
ejpam-1184	341	37	a	a	DET
ejpam-1184	341	38	constant	constant	ADJ
ejpam-1184	341	39	c	c	NOUN
ejpam-1184	341	40	>	>	X
ejpam-1184	341	41	0	0	X
ejpam-1184	341	42	.	.	PUNCT
ejpam-1184	342	1	in	in	ADP
ejpam-1184	342	2	light	light	NOUN
ejpam-1184	342	3	of	of	ADP
ejpam-1184	342	4	the	the	DET
ejpam-1184	342	5	fact	fact	NOUN
ejpam-1184	342	6	that	that	SCONJ
ejpam-1184	342	7	limλ→0	limλ→0	PROPN
ejpam-1184	342	8	�	�	PROPN
ejpam-1184	342	9	(	(	PUNCT
ejpam-1184	342	10	log	log	VERB
ejpam-1184	342	11	n)λ−	n)λ−	PROPN
ejpam-1184	342	12	1	1	NUM
ejpam-1184	342	13	�	�	NOUN
ejpam-1184	342	14	/λ	/λ	PUNCT
ejpam-1184	343	1	=	=	NOUN
ejpam-1184	343	2	log	log	VERB
ejpam-1184	343	3	log	log	NOUN
ejpam-1184	343	4	n	n	CCONJ
ejpam-1184	343	5	,	,	PUNCT
ejpam-1184	343	6	we	we	PRON
ejpam-1184	343	7	set	set	VERB
ejpam-1184	343	8	an	an	DET
ejpam-1184	343	9	=	=	SYM
ejpam-1184	343	10	�	�	PROPN
ejpam-1184	343	11	(	(	PUNCT
ejpam-1184	343	12	log	log	VERB
ejpam-1184	343	13	n)3	n)3	PROPN
ejpam-1184	343	14	−	−	PROPN
ejpam-1184	343	15	1	1	NUM
ejpam-1184	343	16	�	�	PROPN
ejpam-1184	343	17	/3	/3	PROPN
ejpam-1184	343	18	.	.	PUNCT
ejpam-1184	344	1	for	for	ADP
ejpam-1184	344	2	each	each	PRON
ejpam-1184	344	3	of	of	ADP
ejpam-1184	344	4	the	the	DET
ejpam-1184	344	5	four	four	NUM
ejpam-1184	344	6	cases	case	NOUN
ejpam-1184	344	7	,	,	PUNCT
ejpam-1184	344	8	we	we	PRON
ejpam-1184	344	9	conduct	conduct	VERB
ejpam-1184	344	10	1000	1000	NUM
ejpam-1184	344	11	simulations	simulation	NOUN
ejpam-1184	344	12	using	use	VERB
ejpam-1184	344	13	criteria	criterion	NOUN
ejpam-1184	344	14	ls	ls	PROPN
ejpam-1184	344	15	-	-	PUNCT
ejpam-1184	344	16	c.	c.	NOUN
ejpam-1184	344	17	to	to	PART
ejpam-1184	344	18	apply	apply	VERB
ejpam-1184	344	19	the	the	DET
ejpam-1184	344	20	algorithm	algorithm	NOUN
ejpam-1184	344	21	given	give	VERB
ejpam-1184	344	22	in	in	ADP
ejpam-1184	344	23	section	section	NOUN
ejpam-1184	344	24	4	4	NUM
ejpam-1184	344	25	,	,	PUNCT
ejpam-1184	344	26	we	we	PRON
ejpam-1184	344	27	set	set	VERB
ejpam-1184	344	28	δ	δ	X
ejpam-1184	344	29	=	=	PUNCT
ejpam-1184	344	30	0.2	0.2	NUM
ejpam-1184	344	31	and	and	CCONJ
ejpam-1184	344	32	m	m	VERB
ejpam-1184	344	33	=	=	ADJ
ejpam-1184	344	34	2p	2p	NOUN
ejpam-1184	344	35	.	.	PUNCT
ejpam-1184	345	1	the	the	DET
ejpam-1184	345	2	algorithm	algorithm	NOUN
ejpam-1184	345	3	is	be	AUX
ejpam-1184	345	4	then	then	ADV
ejpam-1184	345	5	used	use	VERB
ejpam-1184	345	6	to	to	PART
ejpam-1184	345	7	estimate	estimate	VERB
ejpam-1184	345	8	the	the	DET
ejpam-1184	345	9	number	number	NOUN
ejpam-1184	345	10	of	of	ADP
ejpam-1184	345	11	clusters	cluster	NOUN
ejpam-1184	345	12	in	in	ADP
ejpam-1184	345	13	linear	linear	ADJ
ejpam-1184	345	14	regression	regression	NOUN
ejpam-1184	345	15	clustering	cluster	VERB
ejpam-1184	345	16	.	.	PUNCT
ejpam-1184	346	1	in	in	ADP
ejpam-1184	346	2	table	table	NOUN
ejpam-1184	346	3	3	3	NUM
ejpam-1184	346	4	we	we	PRON
ejpam-1184	346	5	summarize	summarize	VERB
ejpam-1184	346	6	the	the	DET
ejpam-1184	346	7	results	result	NOUN
ejpam-1184	346	8	from	from	ADP
ejpam-1184	346	9	the	the	DET
ejpam-1184	346	10	simulation	simulation	NOUN
ejpam-1184	346	11	study	study	NOUN
ejpam-1184	346	12	,	,	PUNCT
ejpam-1184	346	13	where	where	SCONJ
ejpam-1184	346	14	each	each	DET
ejpam-1184	346	15	number	number	NOUN
ejpam-1184	346	16	represents	represent	VERB
ejpam-1184	346	17	the	the	DET
ejpam-1184	346	18	relative	relative	ADJ
ejpam-1184	346	19	frequency	frequency	NOUN
ejpam-1184	346	20	of	of	ADP
ejpam-1184	346	21	selecting	select	VERB
ejpam-1184	346	22	a	a	DET
ejpam-1184	346	23	given	give	VERB
ejpam-1184	346	24	number	number	NOUN
ejpam-1184	346	25	k	k	PROPN
ejpam-1184	346	26	clusters	cluster	NOUN
ejpam-1184	346	27	in	in	ADP
ejpam-1184	346	28	regression	regression	NOUN
ejpam-1184	346	29	clustering	cluster	VERB
ejpam-1184	346	30	out	out	ADP
ejpam-1184	346	31	of	of	ADP
ejpam-1184	346	32	the	the	DET
ejpam-1184	346	33	1000	1000	NUM
ejpam-1184	346	34	replications	replication	NOUN
ejpam-1184	346	35	.	.	PUNCT
ejpam-1184	347	1	from	from	ADP
ejpam-1184	347	2	table	table	NOUN
ejpam-1184	347	3	3	3	NUM
ejpam-1184	347	4	we	we	PRON
ejpam-1184	347	5	see	see	VERB
ejpam-1184	347	6	that	that	DET
ejpam-1184	347	7	criterion	criterion	NOUN
ejpam-1184	347	8	ls	ls	PROPN
ejpam-1184	347	9	-	-	PUNCT
ejpam-1184	347	10	c	c	NOUN
ejpam-1184	347	11	performs	perform	VERB
ejpam-1184	347	12	almost	almost	ADV
ejpam-1184	347	13	perfectly	perfectly	ADV
ejpam-1184	347	14	in	in	ADP
ejpam-1184	347	15	cases	case	NOUN
ejpam-1184	347	16	1	1	NUM
ejpam-1184	347	17	and	and	CCONJ
ejpam-1184	347	18	3	3	NUM
ejpam-1184	347	19	,	,	PUNCT
ejpam-1184	347	20	which	which	PRON
ejpam-1184	347	21	is	be	AUX
ejpam-1184	347	22	expected	expect	VERB
ejpam-1184	347	23	since	since	SCONJ
ejpam-1184	347	24	the	the	DET
ejpam-1184	347	25	errors	error	NOUN
ejpam-1184	347	26	are	be	AUX
ejpam-1184	347	27	standard	standard	ADJ
ejpam-1184	347	28	normal	normal	ADJ
ejpam-1184	347	29	distributed	distributed	NOUN
ejpam-1184	347	30	.	.	PUNCT
ejpam-1184	348	1	however	however	ADV
ejpam-1184	348	2	,	,	PUNCT
ejpam-1184	348	3	the	the	DET
ejpam-1184	348	4	criterion	criterion	NOUN
ejpam-1184	348	5	tends	tend	VERB
ejpam-1184	348	6	to	to	AUX
ejpam-1184	348	7	over	over	ADV
ejpam-1184	348	8	-	-	PUNCT
ejpam-1184	348	9	estimate	estimate	VERB
ejpam-1184	348	10	the	the	DET
ejpam-1184	348	11	number	number	NOUN
ejpam-1184	348	12	of	of	ADP
ejpam-1184	348	13	clusters	cluster	NOUN
ejpam-1184	348	14	when	when	SCONJ
ejpam-1184	348	15	the	the	DET
ejpam-1184	348	16	error	error	NOUN
ejpam-1184	348	17	distribution	distribution	NOUN
ejpam-1184	348	18	becomes	become	VERB
ejpam-1184	348	19	heavytailed	heavytailed	ADJ
ejpam-1184	348	20	,	,	PUNCT
ejpam-1184	348	21	as	as	SCONJ
ejpam-1184	348	22	shown	show	VERB
ejpam-1184	348	23	in	in	ADP
ejpam-1184	348	24	cases	case	NOUN
ejpam-1184	348	25	2	2	NUM
ejpam-1184	348	26	and	and	CCONJ
ejpam-1184	348	27	4	4	NUM
ejpam-1184	348	28	.	.	PUNCT
ejpam-1184	349	1	this	this	PRON
ejpam-1184	349	2	is	be	AUX
ejpam-1184	349	3	also	also	ADV
ejpam-1184	349	4	expected	expect	VERB
ejpam-1184	349	5	but	but	CCONJ
ejpam-1184	349	6	it	it	PRON
ejpam-1184	349	7	indicates	indicate	VERB
ejpam-1184	349	8	that	that	SCONJ
ejpam-1184	349	9	the	the	DET
ejpam-1184	349	10	direction	direction	NOUN
ejpam-1184	349	11	of	of	ADP
ejpam-1184	349	12	non	non	ADJ
ejpam-1184	349	13	-	-	NOUN
ejpam-1184	349	14	robustness	robustness	NOUN
ejpam-1184	349	15	of	of	ADP
ejpam-1184	349	16	ls	ls	PROPN
ejpam-1184	349	17	-	-	PUNCT
ejpam-1184	349	18	c	c	NOUN
ejpam-1184	349	19	against	against	ADP
ejpam-1184	349	20	normality	normality	NOUN
ejpam-1184	349	21	is	be	AUX
ejpam-1184	349	22	more	more	ADV
ejpam-1184	349	23	likely	likely	ADJ
ejpam-1184	349	24	to	to	PART
ejpam-1184	349	25	be	be	AUX
ejpam-1184	349	26	over	over	ADV
ejpam-1184	349	27	-	-	PUNCT
ejpam-1184	349	28	clustering	clustering	NOUN
ejpam-1184	349	29	rather	rather	ADV
ejpam-1184	349	30	than	than	ADP
ejpam-1184	349	31	under	under	ADV
ejpam-1184	349	32	-	-	PUNCT
ejpam-1184	349	33	clustering	clustering	NOUN
ejpam-1184	349	34	.	.	PUNCT
ejpam-1184	350	1	the	the	DET
ejpam-1184	350	2	cluster	cluster	NOUN
ejpam-1184	350	3	-	-	PUNCT
ejpam-1184	350	4	specific	specific	ADJ
ejpam-1184	350	5	regression	regression	NOUN
ejpam-1184	350	6	lines	line	NOUN
ejpam-1184	350	7	can	can	AUX
ejpam-1184	350	8	also	also	ADV
ejpam-1184	350	9	be	be	AUX
ejpam-1184	350	10	estimated	estimate	VERB
ejpam-1184	350	11	during	during	ADP
ejpam-1184	350	12	applying	apply	VERB
ejpam-1184	350	13	the	the	DET
ejpam-1184	350	14	criterion	criterion	NOUN
ejpam-1184	350	15	ls	ls	PROPN
ejpam-1184	350	16	-	-	PUNCT
ejpam-1184	350	17	c.	c.	NOUN
ejpam-1184	350	18	table	table	NOUN
ejpam-1184	350	19	4	4	NUM
ejpam-1184	350	20	presents	present	VERB
ejpam-1184	350	21	the	the	DET
ejpam-1184	350	22	estimation	estimation	NOUN
ejpam-1184	350	23	of	of	ADP
ejpam-1184	350	24	the	the	DET
ejpam-1184	350	25	regression	regression	NOUN
ejpam-1184	350	26	parameters	parameter	NOUN
ejpam-1184	350	27	by	by	ADP
ejpam-1184	350	28	applying	apply	VERB
ejpam-1184	350	29	ls	ls	PROPN
ejpam-1184	350	30	-	-	PUNCT
ejpam-1184	350	31	c	c	NOUN
ejpam-1184	350	32	to	to	ADP
ejpam-1184	350	33	the	the	DET
ejpam-1184	350	34	data	datum	NOUN
ejpam-1184	350	35	shown	show	VERB
ejpam-1184	350	36	in	in	ADP
ejpam-1184	350	37	figures	figure	NOUN
ejpam-1184	350	38	2	2	NUM
ejpam-1184	350	39	and	and	CCONJ
ejpam-1184	350	40	3	3	NUM
ejpam-1184	350	41	.	.	NOUN
ejpam-1184	350	42	from	from	ADP
ejpam-1184	350	43	the	the	DET
ejpam-1184	350	44	table	table	NOUN
ejpam-1184	350	45	,	,	PUNCT
ejpam-1184	350	46	one	one	PRON
ejpam-1184	350	47	can	can	AUX
ejpam-1184	350	48	conclude	conclude	VERB
ejpam-1184	350	49	that	that	SCONJ
ejpam-1184	350	50	when	when	SCONJ
ejpam-1184	350	51	the	the	DET
ejpam-1184	350	52	errors	error	NOUN
ejpam-1184	350	53	are	be	AUX
ejpam-1184	350	54	t(3	t(3	PROPN
ejpam-1184	350	55	)	)	PUNCT
ejpam-1184	350	56	distributed	distribute	VERB
ejpam-1184	350	57	,	,	PUNCT
ejpam-1184	350	58	the	the	DET
ejpam-1184	350	59	least	least	ADJ
ejpam-1184	350	60	squares	square	NOUN
ejpam-1184	350	61	regression	regression	NOUN
ejpam-1184	350	62	clustering	clustering	NOUN
ejpam-1184	350	63	method	method	NOUN
ejpam-1184	350	64	is	be	AUX
ejpam-1184	350	65	not	not	PART
ejpam-1184	350	66	able	able	ADJ
ejpam-1184	350	67	to	to	PART
ejpam-1184	350	68	capture	capture	VERB
ejpam-1184	350	69	the	the	DET
ejpam-1184	350	70	underlying	underlie	VERB
ejpam-1184	350	71	groupings	grouping	NOUN
ejpam-1184	350	72	,	,	PUNCT
ejpam-1184	350	73	while	while	SCONJ
ejpam-1184	350	74	it	it	PRON
ejpam-1184	350	75	can	can	AUX
ejpam-1184	350	76	when	when	SCONJ
ejpam-1184	350	77	the	the	DET
ejpam-1184	350	78	errors	error	NOUN
ejpam-1184	350	79	are	be	AUX
ejpam-1184	350	80	normal	normal	ADJ
ejpam-1184	350	81	.	.	PUNCT
ejpam-1184	351	1	g.	g.	PROPN
ejpam-1184	351	2	qian	qian	PROPN
ejpam-1184	351	3	,	,	PUNCT
ejpam-1184	351	4	y.	y.	PROPN
ejpam-1184	351	5	wu	wu	PROPN
ejpam-1184	351	6	/	/	SYM
ejpam-1184	351	7	eur	eur	PROPN
ejpam-1184	351	8	.	.	PUNCT
ejpam-1184	352	1	j.	j.	PROPN
ejpam-1184	352	2	pure	pure	PROPN
ejpam-1184	352	3	appl	appl	PROPN
ejpam-1184	352	4	.	.	PROPN
ejpam-1184	352	5	math	math	PROPN
ejpam-1184	352	6	,	,	PUNCT
ejpam-1184	352	7	4	4	NUM
ejpam-1184	352	8	(	(	PUNCT
ejpam-1184	352	9	2011	2011	NUM
ejpam-1184	352	10	)	)	PUNCT
ejpam-1184	352	11	,	,	PUNCT
ejpam-1184	352	12	455	455	NUM
ejpam-1184	352	13	-	-	SYM
ejpam-1184	352	14	466	466	NUM
ejpam-1184	352	15	464	464	NUM
ejpam-1184	352	16	table	table	NOUN
ejpam-1184	352	17	3	3	NUM
ejpam-1184	352	18	:	:	PUNCT
ejpam-1184	352	19	relative	relative	ADJ
ejpam-1184	352	20	frequencies	frequency	NOUN
ejpam-1184	352	21	of	of	ADP
ejpam-1184	352	22	selecting	select	VERB
ejpam-1184	352	23	k	k	PROPN
ejpam-1184	352	24	based	base	VERB
ejpam-1184	352	25	on	on	ADP
ejpam-1184	352	26	1000	1000	NUM
ejpam-1184	352	27	simulations	simulation	NOUN
ejpam-1184	352	28	for	for	ADP
ejpam-1184	352	29	cases	case	NOUN
ejpam-1184	352	30	1	1	NUM
ejpam-1184	352	31	-	-	SYM
ejpam-1184	352	32	4	4	NUM
ejpam-1184	352	33	.	.	PUNCT
ejpam-1184	352	34	case	case	NOUN
ejpam-1184	352	35	1	1	NUM
ejpam-1184	352	36	case	case	NOUN
ejpam-1184	352	37	2	2	NUM
ejpam-1184	352	38	case	case	NOUN
ejpam-1184	352	39	3	3	NUM
ejpam-1184	352	40	case	case	NOUN
ejpam-1184	352	41	4	4	NUM
ejpam-1184	352	42	n(0,1	n(0,1	NOUN
ejpam-1184	352	43	)	)	PUNCT
ejpam-1184	352	44	error	error	NOUN
ejpam-1184	352	45	t(3	t(3	PROPN
ejpam-1184	352	46	)	)	PUNCT
ejpam-1184	352	47	error	error	NOUN
ejpam-1184	352	48	n(0,1	n(0,1	NOUN
ejpam-1184	352	49	)	)	PUNCT
ejpam-1184	352	50	error	error	NOUN
ejpam-1184	352	51	t(3	t(3	PROPN
ejpam-1184	352	52	)	)	PUNCT
ejpam-1184	352	53	error	error	NOUN
ejpam-1184	352	54	k0	k0	NOUN
ejpam-1184	352	55	=	=	PROPN
ejpam-1184	352	56	2	2	NUM
ejpam-1184	352	57	k0	k0	PROPN
ejpam-1184	352	58	=	=	PROPN
ejpam-1184	352	59	2	2	NUM
ejpam-1184	352	60	k0	k0	PROPN
ejpam-1184	352	61	=	=	PROPN
ejpam-1184	352	62	3	3	NUM
ejpam-1184	352	63	k0	k0	PROPN
ejpam-1184	352	64	=	=	PROPN
ejpam-1184	352	65	3	3	NUM
ejpam-1184	352	66	k	k	NOUN
ejpam-1184	352	67	=	=	SYM
ejpam-1184	352	68	1	1	NUM
ejpam-1184	352	69	.000	.000	NUM
ejpam-1184	352	70	.001	.001	NUM
ejpam-1184	352	71	.000	.000	NUM
ejpam-1184	352	72	.000	.000	NUM
ejpam-1184	353	1	k	k	X
ejpam-1184	353	2	=	=	SYM
ejpam-1184	353	3	2	2	NUM
ejpam-1184	353	4	.986	.986	NUM
ejpam-1184	353	5	.422	.422	NUM
ejpam-1184	353	6	.000	.000	NUM
ejpam-1184	353	7	.000	.000	NUM
ejpam-1184	353	8	k	k	X
ejpam-1184	353	9	=	=	SYM
ejpam-1184	353	10	3	3	NUM
ejpam-1184	353	11	.014	.014	NUM
ejpam-1184	353	12	.488	.488	NUM
ejpam-1184	353	13	.999	.999	NUM
ejpam-1184	353	14	.791	.791	NUM
ejpam-1184	354	1	k	k	X
ejpam-1184	354	2	=	=	SYM
ejpam-1184	354	3	4	4	NUM
ejpam-1184	354	4	.000	.000	NUM
ejpam-1184	354	5	.087	.087	NUM
ejpam-1184	354	6	.001	.001	NUM
ejpam-1184	354	7	.207	.207	NUM
ejpam-1184	355	1	k	k	X
ejpam-1184	355	2	=	=	SYM
ejpam-1184	355	3	5	5	NUM
ejpam-1184	355	4	.000	.000	NUM
ejpam-1184	355	5	.002	.002	NUM
ejpam-1184	355	6	.000	.000	NUM
ejpam-1184	355	7	.002	.002	NUM
ejpam-1184	355	8	table	table	NOUN
ejpam-1184	355	9	4	4	NUM
ejpam-1184	355	10	:	:	PUNCT
ejpam-1184	355	11	the	the	DET
ejpam-1184	355	12	estimation	estimation	NOUN
ejpam-1184	355	13	of	of	ADP
ejpam-1184	355	14	the	the	DET
ejpam-1184	355	15	regression	regression	NOUN
ejpam-1184	355	16	parameters	parameter	NOUN
ejpam-1184	355	17	by	by	ADP
ejpam-1184	355	18	applying	apply	VERB
ejpam-1184	355	19	ls	ls	PROPN
ejpam-1184	355	20	-	-	PUNCT
ejpam-1184	355	21	c	c	NOUN
ejpam-1184	355	22	to	to	ADP
ejpam-1184	355	23	the	the	DET
ejpam-1184	355	24	data	datum	NOUN
ejpam-1184	355	25	shown	show	VERB
ejpam-1184	355	26	in	in	ADP
ejpam-1184	355	27	figures	figure	NOUN
ejpam-1184	355	28	2	2	NUM
ejpam-1184	355	29	and	and	CCONJ
ejpam-1184	355	30	3	3	NUM
ejpam-1184	355	31	k0	k0	PROPN
ejpam-1184	355	32	case	case	NOUN
ejpam-1184	355	33	clusters	cluster	VERB
ejpam-1184	355	34	β1	β1	PROPN
ejpam-1184	355	35	β2	β2	PROPN
ejpam-1184	355	36	β3	β3	VERB
ejpam-1184	355	37	β4	β4	ADJ
ejpam-1184	355	38	2	2	NUM
ejpam-1184	355	39	true	true	ADJ
ejpam-1184	355	40	�	�	PROPN
ejpam-1184	355	41	2	2	NUM
ejpam-1184	355	42	8	8	NUM
ejpam-1184	355	43	�	�	PROPN
ejpam-1184	355	44	�	�	PROPN
ejpam-1184	355	45	1	1	NUM
ejpam-1184	355	46	5	5	NUM
ejpam-1184	355	47	�	�	PROPN
ejpam-1184	355	48	1	1	NUM
ejpam-1184	355	49	ls	ls	PROPN
ejpam-1184	355	50	-	-	PUNCT
ejpam-1184	355	51	c	c	NOUN
ejpam-1184	355	52	�	�	PROPN
ejpam-1184	355	53	2.12	2.12	NUM
ejpam-1184	355	54	8.02	8.02	NUM
ejpam-1184	355	55	�	�	PROPN
ejpam-1184	355	56	�	�	PROPN
ejpam-1184	355	57	0.76	0.76	NUM
ejpam-1184	355	58	5.11	5.11	NUM
ejpam-1184	355	59	�	�	PROPN
ejpam-1184	355	60	2	2	NUM
ejpam-1184	355	61	ls	ls	PROPN
ejpam-1184	355	62	-	-	PUNCT
ejpam-1184	355	63	c	c	NOUN
ejpam-1184	355	64	�	�	PROPN
ejpam-1184	355	65	1.48	1.48	NUM
ejpam-1184	355	66	5.56	5.56	NUM
ejpam-1184	355	67	�	�	PROPN
ejpam-1184	355	68	�	�	PROPN
ejpam-1184	355	69	−1.13	−1.13	NUM
ejpam-1184	355	70	5.87	5.87	NUM
ejpam-1184	355	71	�	�	PROPN
ejpam-1184	355	72	�	�	PROPN
ejpam-1184	355	73	4.46	4.46	NUM
ejpam-1184	355	74	6.18	6.18	NUM
ejpam-1184	355	75	�	�	PROPN
ejpam-1184	355	76	3	3	NUM
ejpam-1184	355	77	true	true	ADJ
ejpam-1184	355	78	�	�	PROPN
ejpam-1184	355	79	18	18	NUM
ejpam-1184	355	80	6	6	NUM
ejpam-1184	355	81	�	�	PROPN
ejpam-1184	355	82	�	�	PROPN
ejpam-1184	355	83	12	12	NUM
ejpam-1184	355	84	8	8	NUM
ejpam-1184	355	85	�	�	PROPN
ejpam-1184	355	86	�	�	PROPN
ejpam-1184	355	87	15	15	NUM
ejpam-1184	355	88	−2	−2	NOUN
ejpam-1184	355	89	�	�	PROPN
ejpam-1184	355	90	3	3	NUM
ejpam-1184	355	91	ls	ls	PROPN
ejpam-1184	355	92	-	-	PUNCT
ejpam-1184	355	93	c	c	NOUN
ejpam-1184	355	94	�	�	PROPN
ejpam-1184	355	95	18.05	18.05	NUM
ejpam-1184	355	96	6.06	6.06	NUM
ejpam-1184	355	97	�	�	PROPN
ejpam-1184	355	98	�	�	PROPN
ejpam-1184	355	99	11.97	11.97	NUM
ejpam-1184	355	100	8.02	8.02	NUM
ejpam-1184	355	101	�	�	PROPN
ejpam-1184	355	102	�	�	PROPN
ejpam-1184	355	103	14.66	14.66	NUM
ejpam-1184	355	104	−1.85	−1.85	PROPN
ejpam-1184	355	105	�	�	PROPN
ejpam-1184	355	106	4	4	NUM
ejpam-1184	355	107	ls	ls	PROPN
ejpam-1184	355	108	-	-	PUNCT
ejpam-1184	355	109	c	c	NOUN
ejpam-1184	355	110	�	�	PROPN
ejpam-1184	355	111	17.74	17.74	NUM
ejpam-1184	355	112	6.14	6.14	NUM
ejpam-1184	355	113	�	�	PROPN
ejpam-1184	355	114	�	�	PROPN
ejpam-1184	355	115	12.02	12.02	NUM
ejpam-1184	355	116	8.16	8.16	NUM
ejpam-1184	355	117	�	�	PROPN
ejpam-1184	355	118	�	�	PROPN
ejpam-1184	355	119	10.73	10.73	NUM
ejpam-1184	355	120	−2.87	−2.87	PROPN
ejpam-1184	355	121	�	�	PROPN
ejpam-1184	355	122	�	�	PROPN
ejpam-1184	355	123	15.54	15.54	NUM
ejpam-1184	355	124	−1.70	−1.70	DET
ejpam-1184	355	125	�	�	PROPN
ejpam-1184	355	126	references	reference	VERB
ejpam-1184	355	127	465	465	NUM
ejpam-1184	355	128	6	6	NUM
ejpam-1184	355	129	.	.	PUNCT
ejpam-1184	356	1	discussion	discussion	NOUN
ejpam-1184	356	2	in	in	ADP
ejpam-1184	356	3	this	this	DET
ejpam-1184	356	4	paper	paper	NOUN
ejpam-1184	356	5	we	we	PRON
ejpam-1184	356	6	review	review	VERB
ejpam-1184	356	7	the	the	DET
ejpam-1184	356	8	general	general	ADJ
ejpam-1184	356	9	cluster	cluster	NOUN
ejpam-1184	356	10	analysis	analysis	NOUN
ejpam-1184	356	11	methods	method	NOUN
ejpam-1184	356	12	,	,	PUNCT
ejpam-1184	356	13	then	then	ADV
ejpam-1184	356	14	focus	focus	VERB
ejpam-1184	356	15	on	on	ADP
ejpam-1184	356	16	regression	regression	NOUN
ejpam-1184	356	17	clustering	clustering	NOUN
ejpam-1184	356	18	which	which	PRON
ejpam-1184	356	19	uses	use	VERB
ejpam-1184	356	20	the	the	DET
ejpam-1184	356	21	model	model	NOUN
ejpam-1184	356	22	-	-	PUNCT
ejpam-1184	356	23	based	base	VERB
ejpam-1184	356	24	fixed	fix	VERB
ejpam-1184	356	25	partition	partition	NOUN
ejpam-1184	356	26	method	method	NOUN
ejpam-1184	356	27	and	and	CCONJ
ejpam-1184	356	28	also	also	ADV
ejpam-1184	356	29	takes	take	VERB
ejpam-1184	356	30	into	into	ADP
ejpam-1184	356	31	account	account	NOUN
ejpam-1184	356	32	the	the	DET
ejpam-1184	356	33	dependence	dependence	NOUN
ejpam-1184	356	34	between	between	ADP
ejpam-1184	356	35	the	the	DET
ejpam-1184	356	36	response	response	NOUN
ejpam-1184	356	37	and	and	CCONJ
ejpam-1184	356	38	explanatory	explanatory	ADJ
ejpam-1184	356	39	variables	variable	NOUN
ejpam-1184	356	40	.	.	PUNCT
ejpam-1184	357	1	regression	regression	NOUN
ejpam-1184	357	2	clustering	clustering	NOUN
ejpam-1184	357	3	has	have	AUX
ejpam-1184	357	4	not	not	PART
ejpam-1184	357	5	been	be	AUX
ejpam-1184	357	6	widely	widely	ADV
ejpam-1184	357	7	used	use	VERB
ejpam-1184	357	8	in	in	ADP
ejpam-1184	357	9	practice	practice	NOUN
ejpam-1184	357	10	even	even	ADV
ejpam-1184	357	11	though	though	SCONJ
ejpam-1184	357	12	it	it	PRON
ejpam-1184	357	13	has	have	VERB
ejpam-1184	357	14	a	a	DET
ejpam-1184	357	15	great	great	ADJ
ejpam-1184	357	16	potential	potential	NOUN
ejpam-1184	357	17	.	.	PUNCT
ejpam-1184	358	1	a	a	DET
ejpam-1184	358	2	possible	possible	ADJ
ejpam-1184	358	3	reason	reason	NOUN
ejpam-1184	358	4	is	be	AUX
ejpam-1184	358	5	the	the	DET
ejpam-1184	358	6	computing	compute	VERB
ejpam-1184	358	7	complexity	complexity	NOUN
ejpam-1184	358	8	involved	involve	VERB
ejpam-1184	358	9	in	in	ADP
ejpam-1184	358	10	the	the	DET
ejpam-1184	358	11	method	method	NOUN
ejpam-1184	358	12	.	.	PUNCT
ejpam-1184	359	1	this	this	DET
ejpam-1184	359	2	paper	paper	NOUN
ejpam-1184	359	3	provides	provide	VERB
ejpam-1184	359	4	a	a	DET
ejpam-1184	359	5	computing	compute	VERB
ejpam-1184	359	6	procedure	procedure	NOUN
ejpam-1184	359	7	and	and	CCONJ
ejpam-1184	359	8	a	a	DET
ejpam-1184	359	9	feasible	feasible	ADJ
ejpam-1184	359	10	algorithm	algorithm	NOUN
ejpam-1184	359	11	for	for	ADP
ejpam-1184	359	12	estimation	estimation	NOUN
ejpam-1184	359	13	and	and	CCONJ
ejpam-1184	359	14	selection	selection	NOUN
ejpam-1184	359	15	involved	involve	VERB
ejpam-1184	359	16	in	in	ADP
ejpam-1184	359	17	regression	regression	NOUN
ejpam-1184	359	18	clustering	cluster	VERB
ejpam-1184	359	19	.	.	PUNCT
ejpam-1184	360	1	the	the	DET
ejpam-1184	360	2	simulation	simulation	NOUN
ejpam-1184	360	3	study	study	NOUN
ejpam-1184	360	4	concludes	conclude	VERB
ejpam-1184	360	5	a	a	DET
ejpam-1184	360	6	satisfactory	satisfactory	ADJ
ejpam-1184	360	7	finite	finite	NOUN
ejpam-1184	360	8	sample	sample	NOUN
ejpam-1184	360	9	performance	performance	NOUN
ejpam-1184	360	10	of	of	ADP
ejpam-1184	360	11	the	the	DET
ejpam-1184	360	12	algorithm	algorithm	NOUN
ejpam-1184	360	13	when	when	SCONJ
ejpam-1184	360	14	the	the	DET
ejpam-1184	360	15	error	error	NOUN
ejpam-1184	360	16	distribution	distribution	NOUN
ejpam-1184	360	17	involved	involve	VERB
ejpam-1184	360	18	is	be	AUX
ejpam-1184	360	19	close	close	ADJ
ejpam-1184	360	20	to	to	ADP
ejpam-1184	360	21	normal	normal	ADJ
ejpam-1184	360	22	.	.	PUNCT
ejpam-1184	361	1	it	it	PRON
ejpam-1184	361	2	also	also	ADV
ejpam-1184	361	3	suggests	suggest	VERB
ejpam-1184	361	4	the	the	DET
ejpam-1184	361	5	need	need	NOUN
ejpam-1184	361	6	to	to	PART
ejpam-1184	361	7	use	use	VERB
ejpam-1184	361	8	a	a	DET
ejpam-1184	361	9	robust	robust	ADJ
ejpam-1184	361	10	clustering	clustering	NOUN
ejpam-1184	361	11	method	method	NOUN
ejpam-1184	361	12	when	when	SCONJ
ejpam-1184	361	13	the	the	DET
ejpam-1184	361	14	error	error	NOUN
ejpam-1184	361	15	distribution	distribution	NOUN
ejpam-1184	361	16	strays	stray	VERB
ejpam-1184	361	17	away	away	ADV
ejpam-1184	361	18	from	from	ADP
ejpam-1184	361	19	the	the	DET
ejpam-1184	361	20	normal	normal	ADJ
ejpam-1184	361	21	.	.	PUNCT
ejpam-1184	362	1	references	reference	NOUN
ejpam-1184	362	2	[	[	X
ejpam-1184	362	3	1	1	NUM
ejpam-1184	362	4	]	]	X
ejpam-1184	362	5	c	c	X
ejpam-1184	362	6	hennig	hennig	PROPN
ejpam-1184	362	7	.	.	PUNCT
ejpam-1184	363	1	identifiability	identifiability	NOUN
ejpam-1184	363	2	of	of	ADP
ejpam-1184	363	3	models	model	NOUN
ejpam-1184	363	4	for	for	ADP
ejpam-1184	363	5	clusterwise	clusterwise	ADJ
ejpam-1184	363	6	linear	linear	PROPN
ejpam-1184	363	7	regression	regression	NOUN
ejpam-1184	363	8	.	.	PUNCT
ejpam-1184	364	1	journal	journal	NOUN
ejpam-1184	364	2	of	of	ADP
ejpam-1184	364	3	classification	classification	NOUN
ejpam-1184	364	4	,	,	PUNCT
ejpam-1184	364	5	17:273–296	17:273–296	NUM
ejpam-1184	364	6	,	,	PUNCT
ejpam-1184	364	7	2000	2000	NUM
ejpam-1184	364	8	.	.	PUNCT
ejpam-1184	365	1	[	[	X
ejpam-1184	365	2	2	2	NUM
ejpam-1184	365	3	]	]	X
ejpam-1184	365	4	c	c	NOUN
ejpam-1184	365	5	rao	rao	NOUN
ejpam-1184	365	6	and	and	CCONJ
ejpam-1184	365	7	y	y	PROPN
ejpam-1184	365	8	wu	wu	PROPN
ejpam-1184	365	9	and	and	CCONJ
ejpam-1184	365	10	q	q	PROPN
ejpam-1184	365	11	shao	shao	PROPN
ejpam-1184	365	12	.	.	PUNCT
ejpam-1184	366	1	an	an	DET
ejpam-1184	366	2	m	m	NOUN
ejpam-1184	366	3	-	-	PUNCT
ejpam-1184	366	4	estimation	estimation	NOUN
ejpam-1184	366	5	-	-	PUNCT
ejpam-1184	366	6	based	base	VERB
ejpam-1184	366	7	procedure	procedure	NOUN
ejpam-1184	366	8	for	for	ADP
ejpam-1184	366	9	determining	determine	VERB
ejpam-1184	366	10	the	the	DET
ejpam-1184	366	11	number	number	NOUN
ejpam-1184	366	12	of	of	ADP
ejpam-1184	366	13	regression	regression	NOUN
ejpam-1184	366	14	models	model	NOUN
ejpam-1184	366	15	in	in	ADP
ejpam-1184	366	16	regression	regression	NOUN
ejpam-1184	366	17	clustering	cluster	VERB
ejpam-1184	366	18	.	.	PUNCT
ejpam-1184	367	1	journal	journal	NOUN
ejpam-1184	367	2	of	of	ADP
ejpam-1184	367	3	applied	apply	VERB
ejpam-1184	367	4	mathematics	mathematic	NOUN
ejpam-1184	367	5	and	and	CCONJ
ejpam-1184	367	6	decision	decision	NOUN
ejpam-1184	367	7	sciences	science	NOUN
ejpam-1184	367	8	,	,	PUNCT
ejpam-1184	367	9	2007	2007	NUM
ejpam-1184	367	10	,	,	PUNCT
ejpam-1184	367	11	2007	2007	NUM
ejpam-1184	367	12	.	.	PUNCT
ejpam-1184	368	1	[	[	X
ejpam-1184	368	2	3	3	X
ejpam-1184	368	3	]	]	X
ejpam-1184	368	4	d	d	X
ejpam-1184	368	5	pollard	pollard	NOUN
ejpam-1184	368	6	.	.	PUNCT
ejpam-1184	369	1	strong	strong	ADJ
ejpam-1184	369	2	consistency	consistency	NOUN
ejpam-1184	369	3	of	of	ADP
ejpam-1184	369	4	k	k	PROPN
ejpam-1184	369	5	-	-	PUNCT
ejpam-1184	369	6	means	means	NOUN
ejpam-1184	369	7	clustering	clustering	NOUN
ejpam-1184	369	8	.	.	PUNCT
ejpam-1184	370	1	the	the	DET
ejpam-1184	370	2	annals	annal	NOUN
ejpam-1184	370	3	of	of	ADP
ejpam-1184	370	4	statistics	statistic	NOUN
ejpam-1184	370	5	,	,	PUNCT
ejpam-1184	370	6	9:135–140	9:135–140	NUM
ejpam-1184	370	7	,	,	PUNCT
ejpam-1184	370	8	1981	1981	NUM
ejpam-1184	370	9	.	.	PUNCT
ejpam-1184	371	1	[	[	X
ejpam-1184	371	2	4	4	X
ejpam-1184	371	3	]	]	PUNCT
ejpam-1184	371	4	h	h	NOUN
ejpam-1184	371	5	bock	bock	NOUN
ejpam-1184	371	6	.	.	PUNCT
ejpam-1184	372	1	the	the	DET
ejpam-1184	372	2	equivalence	equivalence	NOUN
ejpam-1184	372	3	of	of	ADP
ejpam-1184	372	4	two	two	NUM
ejpam-1184	372	5	extremal	extremal	ADJ
ejpam-1184	372	6	problems	problem	NOUN
ejpam-1184	372	7	and	and	CCONJ
ejpam-1184	372	8	its	its	PRON
ejpam-1184	372	9	application	application	NOUN
ejpam-1184	372	10	to	to	ADP
ejpam-1184	372	11	the	the	DET
ejpam-1184	372	12	iterative	iterative	ADJ
ejpam-1184	372	13	classification	classification	NOUN
ejpam-1184	372	14	of	of	ADP
ejpam-1184	372	15	multivariate	multivariate	NOUN
ejpam-1184	372	16	data	datum	NOUN
ejpam-1184	372	17	.	.	PUNCT
ejpam-1184	373	1	manuscript	manuscript	NOUN
ejpam-1184	373	2	for	for	ADP
ejpam-1184	373	3	the	the	DET
ejpam-1184	373	4	medizinische	medizinische	PROPN
ejpam-1184	373	5	statistik	statistik	PROPN
ejpam-1184	373	6	conference	conference	PROPN
ejpam-1184	373	7	,	,	PUNCT
ejpam-1184	373	8	forschungsinstitut	forschungsinstitut	VERB
ejpam-1184	373	9	oberworfachl	oberworfachl	ADJ
ejpam-1184	373	10	,	,	PUNCT
ejpam-1184	373	11	1969	1969	NUM
ejpam-1184	373	12	.	.	PUNCT
ejpam-1184	374	1	[	[	X
ejpam-1184	374	2	5	5	X
ejpam-1184	374	3	]	]	PUNCT
ejpam-1184	374	4	h	h	NOUN
ejpam-1184	374	5	bock	bock	NOUN
ejpam-1184	374	6	.	.	PUNCT
ejpam-1184	375	1	probability	probability	NOUN
ejpam-1184	375	2	models	model	NOUN
ejpam-1184	375	3	and	and	CCONJ
ejpam-1184	375	4	hypotheses	hypothesis	NOUN
ejpam-1184	375	5	testing	test	VERB
ejpam-1184	375	6	in	in	ADP
ejpam-1184	375	7	partitioning	partition	VERB
ejpam-1184	375	8	cluster	cluster	NOUN
ejpam-1184	375	9	analysis	analysis	NOUN
ejpam-1184	375	10	.	.	PUNCT
ejpam-1184	376	1	in	in	ADP
ejpam-1184	376	2	p	p	X
ejpam-1184	376	3	arabie	arabie	NOUN
ejpam-1184	376	4	and	and	CCONJ
ejpam-1184	376	5	l	l	PROPN
ejpam-1184	376	6	hubert	hubert	PROPN
ejpam-1184	376	7	and	and	CCONJ
ejpam-1184	376	8	g	g	PROPN
ejpam-1184	376	9	de	de	X
ejpam-1184	376	10	soete	soete	NOUN
ejpam-1184	376	11	,	,	PUNCT
ejpam-1184	376	12	editor	editor	NOUN
ejpam-1184	376	13	,	,	PUNCT
ejpam-1184	376	14	clustering	clustering	NOUN
ejpam-1184	376	15	and	and	CCONJ
ejpam-1184	376	16	classification	classification	NOUN
ejpam-1184	376	17	.	.	PUNCT
ejpam-1184	376	18	,	,	PUNCT
ejpam-1184	376	19	pages	page	NOUN
ejpam-1184	376	20	377	377	NUM
ejpam-1184	376	21	–	–	PUNCT
ejpam-1184	376	22	453	453	NUM
ejpam-1184	376	23	,	,	PUNCT
ejpam-1184	376	24	river	river	NOUN
ejpam-1184	376	25	edge	edge	NOUN
ejpam-1184	376	26	,	,	PUNCT
ejpam-1184	376	27	new	new	PROPN
ejpam-1184	376	28	jersey	jersey	PROPN
ejpam-1184	376	29	.	.	PROPN
ejpam-1184	376	30	,	,	PUNCT
ejpam-1184	376	31	1996	1996	NUM
ejpam-1184	376	32	.	.	PUNCT
ejpam-1184	377	1	world	world	NOUN
ejpam-1184	377	2	scientific	scientific	ADJ
ejpam-1184	377	3	publishing	publishing	NOUN
ejpam-1184	377	4	.	.	PUNCT
ejpam-1184	378	1	[	[	X
ejpam-1184	378	2	6	6	NUM
ejpam-1184	378	3	]	]	PUNCT
ejpam-1184	378	4	h	h	NOUN
ejpam-1184	378	5	späth	späth	NOUN
ejpam-1184	378	6	.	.	PUNCT
ejpam-1184	379	1	clusterwise	clusterwise	PROPN
ejpam-1184	379	2	linear	linear	PROPN
ejpam-1184	379	3	regression	regression	NOUN
ejpam-1184	379	4	.	.	PUNCT
ejpam-1184	380	1	computing	computing	NOUN
ejpam-1184	380	2	,	,	PUNCT
ejpam-1184	380	3	22:367–373	22:367–373	NUM
ejpam-1184	380	4	,	,	PUNCT
ejpam-1184	380	5	1979	1979	NUM
ejpam-1184	380	6	.	.	PUNCT
ejpam-1184	381	1	[	[	X
ejpam-1184	381	2	7	7	X
ejpam-1184	381	3	]	]	X
ejpam-1184	381	4	h	h	NOUN
ejpam-1184	381	5	späth	späth	NOUN
ejpam-1184	381	6	.	.	PUNCT
ejpam-1184	382	1	algorithm	algorithm	NOUN
ejpam-1184	382	2	48	48	NUM
ejpam-1184	382	3	:	:	PUNCT
ejpam-1184	382	4	a	a	DET
ejpam-1184	382	5	fast	fast	ADJ
ejpam-1184	382	6	algorithm	algorithm	NOUN
ejpam-1184	382	7	for	for	ADP
ejpam-1184	382	8	clusterwise	clusterwise	ADJ
ejpam-1184	382	9	linear	linear	PROPN
ejpam-1184	382	10	regression	regression	NOUN
ejpam-1184	382	11	.	.	PUNCT
ejpam-1184	383	1	computing	computing	NOUN
ejpam-1184	383	2	,	,	PUNCT
ejpam-1184	383	3	29:175–181	29:175–181	PROPN
ejpam-1184	383	4	,	,	PUNCT
ejpam-1184	383	5	1982	1982	NUM
ejpam-1184	383	6	.	.	PUNCT
ejpam-1184	384	1	[	[	X
ejpam-1184	384	2	8	8	NUM
ejpam-1184	384	3	]	]	X
ejpam-1184	384	4	j	j	PROPN
ejpam-1184	384	5	hartigan	hartigan	PROPN
ejpam-1184	384	6	.	.	PUNCT
ejpam-1184	385	1	consistency	consistency	NOUN
ejpam-1184	385	2	of	of	ADP
ejpam-1184	385	3	single	single	ADJ
ejpam-1184	385	4	linkage	linkage	NOUN
ejpam-1184	385	5	for	for	ADP
ejpam-1184	385	6	high	high	ADJ
ejpam-1184	385	7	-	-	PUNCT
ejpam-1184	385	8	density	density	NOUN
ejpam-1184	385	9	clusters	cluster	NOUN
ejpam-1184	385	10	.	.	PUNCT
ejpam-1184	386	1	journal	journal	NOUN
ejpam-1184	386	2	of	of	ADP
ejpam-1184	386	3	the	the	DET
ejpam-1184	386	4	american	american	PROPN
ejpam-1184	386	5	statistical	statistical	PROPN
ejpam-1184	386	6	association	association	NOUN
ejpam-1184	386	7	,	,	PUNCT
ejpam-1184	386	8	76:388–394	76:388–394	NUM
ejpam-1184	386	9	,	,	PUNCT
ejpam-1184	386	10	1981	1981	NUM
ejpam-1184	386	11	.	.	PUNCT
ejpam-1184	387	1	[	[	X
ejpam-1184	387	2	9	9	NUM
ejpam-1184	387	3	]	]	SYM
ejpam-1184	387	4	j	j	PROPN
ejpam-1184	387	5	hartigan	hartigan	PROPN
ejpam-1184	387	6	and	and	CCONJ
ejpam-1184	387	7	m	m	PROPN
ejpam-1184	387	8	wong	wong	PROPN
ejpam-1184	387	9	.	.	PUNCT
ejpam-1184	387	10	algorithm	algorithm	PROPN
ejpam-1184	387	11	as	as	ADP
ejpam-1184	387	12	136	136	NUM
ejpam-1184	387	13	:	:	PUNCT
ejpam-1184	387	14	a	a	DET
ejpam-1184	387	15	k	k	X
ejpam-1184	387	16	-	-	PUNCT
ejpam-1184	387	17	means	mean	VERB
ejpam-1184	387	18	clustering	clustering	ADJ
ejpam-1184	387	19	algorithm	algorithm	NOUN
ejpam-1184	387	20	.	.	PUNCT
ejpam-1184	388	1	applied	apply	VERB
ejpam-1184	388	2	statistics	statistic	NOUN
ejpam-1184	388	3	,	,	PUNCT
ejpam-1184	388	4	28:100–108	28:100–108	NUM
ejpam-1184	388	5	,	,	PUNCT
ejpam-1184	388	6	1978	1978	NUM
ejpam-1184	388	7	.	.	PUNCT
ejpam-1184	389	1	[	[	X
ejpam-1184	389	2	10	10	NUM
ejpam-1184	389	3	]	]	X
ejpam-1184	389	4	j	j	PROPN
ejpam-1184	389	5	macqueen	macqueen	PROPN
ejpam-1184	389	6	.	.	PUNCT
ejpam-1184	390	1	some	some	DET
ejpam-1184	390	2	methods	method	NOUN
ejpam-1184	390	3	for	for	ADP
ejpam-1184	390	4	classification	classification	NOUN
ejpam-1184	390	5	and	and	CCONJ
ejpam-1184	390	6	analysis	analysis	NOUN
ejpam-1184	390	7	of	of	ADP
ejpam-1184	390	8	multivariate	multivariate	NOUN
ejpam-1184	390	9	observations	observation	NOUN
ejpam-1184	390	10	.	.	PUNCT
ejpam-1184	391	1	in	in	ADP
ejpam-1184	391	2	n	n	PRON
ejpam-1184	391	3	le	le	X
ejpam-1184	391	4	cam	cam	VERB
ejpam-1184	391	5	and	and	CCONJ
ejpam-1184	391	6	j	j	PROPN
ejpam-1184	391	7	neyman	neyman	PROPN
ejpam-1184	391	8	,	,	PUNCT
ejpam-1184	391	9	editors	editor	NOUN
ejpam-1184	391	10	,	,	PUNCT
ejpam-1184	391	11	proceedings	proceeding	NOUN
ejpam-1184	391	12	of	of	ADP
ejpam-1184	391	13	the	the	DET
ejpam-1184	391	14	5th	5th	ADJ
ejpam-1184	391	15	berkeley	berkeley	PROPN
ejpam-1184	391	16	symposium	symposium	NOUN
ejpam-1184	391	17	on	on	ADP
ejpam-1184	391	18	mathematical	mathematical	ADJ
ejpam-1184	391	19	statistics	statistic	NOUN
ejpam-1184	391	20	and	and	CCONJ
ejpam-1184	391	21	probability	probability	NOUN
ejpam-1184	391	22	.	.	PUNCT
ejpam-1184	391	23	,	,	PUNCT
ejpam-1184	391	24	volume	volume	NOUN
ejpam-1184	391	25	1	1	NUM
ejpam-1184	391	26	,	,	PUNCT
ejpam-1184	391	27	pages	page	NOUN
ejpam-1184	391	28	281–297	281–297	NUM
ejpam-1184	391	29	.	.	PUNCT
ejpam-1184	392	1	university	university	PROPN
ejpam-1184	392	2	of	of	ADP
ejpam-1184	392	3	california	california	PROPN
ejpam-1184	392	4	press	press	PROPN
ejpam-1184	392	5	.	.	PUNCT
ejpam-1184	392	6	,	,	PUNCT
ejpam-1184	392	7	1967	1967	NUM
ejpam-1184	392	8	.	.	PUNCT
ejpam-1184	393	1	references	reference	NOUN
ejpam-1184	393	2	466	466	NUM
ejpam-1184	393	3	[	[	X
ejpam-1184	393	4	11	11	NUM
ejpam-1184	393	5	]	]	PUNCT
ejpam-1184	393	6	l	l	NOUN
ejpam-1184	393	7	kaufman	kaufman	NOUN
ejpam-1184	393	8	and	and	CCONJ
ejpam-1184	393	9	p	p	NOUN
ejpam-1184	393	10	rousseeuw	rousseeuw	NOUN
ejpam-1184	393	11	.	.	PUNCT
ejpam-1184	394	1	finding	find	VERB
ejpam-1184	394	2	groups	group	NOUN
ejpam-1184	394	3	in	in	ADP
ejpam-1184	394	4	data	data	PROPN
ejpam-1184	394	5	.	.	PUNCT
ejpam-1184	395	1	wiley	wiley	PROPN
ejpam-1184	395	2	-	-	PUNCT
ejpam-1184	395	3	interscience	interscience	PROPN
ejpam-1184	395	4	,	,	PUNCT
ejpam-1184	395	5	new	new	PROPN
ejpam-1184	395	6	york	york	PROPN
ejpam-1184	395	7	,	,	PUNCT
ejpam-1184	395	8	1990	1990	NUM
ejpam-1184	395	9	.	.	PUNCT
ejpam-1184	396	1	[	[	X
ejpam-1184	396	2	12	12	NUM
ejpam-1184	396	3	]	]	X
ejpam-1184	396	4	m	m	PROPN
ejpam-1184	396	5	wong	wong	PROPN
ejpam-1184	396	6	.	.	PUNCT
ejpam-1184	397	1	a	a	DET
ejpam-1184	397	2	hybrid	hybrid	ADJ
ejpam-1184	397	3	clustering	clustering	NOUN
ejpam-1184	397	4	method	method	NOUN
ejpam-1184	397	5	for	for	ADP
ejpam-1184	397	6	identifying	identify	VERB
ejpam-1184	397	7	high	high	ADJ
ejpam-1184	397	8	-	-	PUNCT
ejpam-1184	397	9	density	density	NOUN
ejpam-1184	397	10	clusters	cluster	NOUN
ejpam-1184	397	11	.	.	PUNCT
ejpam-1184	398	1	journal	journal	NOUN
ejpam-1184	398	2	of	of	ADP
ejpam-1184	398	3	the	the	DET
ejpam-1184	398	4	american	american	PROPN
ejpam-1184	398	5	statistical	statistical	PROPN
ejpam-1184	398	6	association	association	NOUN
ejpam-1184	398	7	,	,	PUNCT
ejpam-1184	398	8	77:841–847	77:841–847	PROPN
ejpam-1184	398	9	,	,	PUNCT
ejpam-1184	398	10	1982	1982	NUM
ejpam-1184	398	11	.	.	PUNCT
ejpam-1184	399	1	[	[	X
ejpam-1184	399	2	13	13	NUM
ejpam-1184	399	3	]	]	X
ejpam-1184	399	4	p	p	X
ejpam-1184	399	5	rousseeuw	rousseeuw	NOUN
ejpam-1184	399	6	and	and	CCONJ
ejpam-1184	399	7	a	a	DET
ejpam-1184	399	8	leroy	leroy	PROPN
ejpam-1184	399	9	.	.	PUNCT
ejpam-1184	400	1	robust	robust	ADJ
ejpam-1184	400	2	regression	regression	NOUN
ejpam-1184	400	3	and	and	CCONJ
ejpam-1184	400	4	outlier	outlier	NOUN
ejpam-1184	400	5	detection	detection	NOUN
ejpam-1184	400	6	.	.	PUNCT
ejpam-1184	401	1	wiley	wiley	PROPN
ejpam-1184	401	2	,	,	PUNCT
ejpam-1184	401	3	new	new	PROPN
ejpam-1184	401	4	york	york	PROPN
ejpam-1184	401	5	,	,	PUNCT
ejpam-1184	401	6	1987	1987	NUM
ejpam-1184	401	7	.	.	PUNCT
ejpam-1184	402	1	[	[	X
ejpam-1184	402	2	14	14	NUM
ejpam-1184	402	3	]	]	X
ejpam-1184	402	4	q	q	PROPN
ejpam-1184	402	5	shao	shao	PROPN
ejpam-1184	402	6	and	and	CCONJ
ejpam-1184	402	7	y	y	PROPN
ejpam-1184	402	8	wu	wu	PROPN
ejpam-1184	402	9	.	.	PUNCT
ejpam-1184	403	1	a	a	DET
ejpam-1184	403	2	consistent	consistent	ADJ
ejpam-1184	403	3	procedure	procedure	NOUN
ejpam-1184	403	4	for	for	ADP
ejpam-1184	403	5	determining	determine	VERB
ejpam-1184	403	6	the	the	DET
ejpam-1184	403	7	number	number	NOUN
ejpam-1184	403	8	of	of	ADP
ejpam-1184	403	9	clusters	cluster	NOUN
ejpam-1184	403	10	in	in	ADP
ejpam-1184	403	11	regression	regression	NOUN
ejpam-1184	403	12	clustering	cluster	VERB
ejpam-1184	403	13	.	.	PUNCT
ejpam-1184	404	1	journal	journal	NOUN
ejpam-1184	404	2	of	of	ADP
ejpam-1184	404	3	statistical	statistical	ADJ
ejpam-1184	404	4	planning	planning	NOUN
ejpam-1184	404	5	and	and	CCONJ
ejpam-1184	404	6	inference	inference	NOUN
ejpam-1184	404	7	,	,	PUNCT
ejpam-1184	404	8	135:461–476	135:461–476	NUM
ejpam-1184	404	9	,	,	PUNCT
ejpam-1184	404	10	2005	2005	NUM
ejpam-1184	404	11	.	.	PUNCT
ejpam-1184	405	1	[	[	X
ejpam-1184	405	2	15	15	NUM
ejpam-1184	405	3	]	]	X
ejpam-1184	405	4	r	r	NOUN
ejpam-1184	405	5	quandt	quandt	NOUN
ejpam-1184	405	6	and	and	CCONJ
ejpam-1184	405	7	j	j	PROPN
ejpam-1184	405	8	ramsey	ramsey	PROPN
ejpam-1184	405	9	.	.	PUNCT
ejpam-1184	406	1	estimating	estimate	VERB
ejpam-1184	406	2	mixtures	mixture	NOUN
ejpam-1184	406	3	of	of	ADP
ejpam-1184	406	4	normal	normal	ADJ
ejpam-1184	406	5	distributions	distribution	NOUN
ejpam-1184	406	6	and	and	CCONJ
ejpam-1184	406	7	switching	switch	VERB
ejpam-1184	406	8	regressions	regression	NOUN
ejpam-1184	406	9	.	.	PUNCT
ejpam-1184	407	1	journal	journal	NOUN
ejpam-1184	407	2	of	of	ADP
ejpam-1184	407	3	the	the	DET
ejpam-1184	407	4	american	american	PROPN
ejpam-1184	407	5	statistical	statistical	ADJ
ejpam-1184	407	6	association	association	PROPN
ejpam-1184	407	7	.	.	PUNCT
ejpam-1184	407	8	,	,	PUNCT
ejpam-1184	407	9	73:730–752	73:730–752	PROPN
ejpam-1184	407	10	,	,	PUNCT
ejpam-1184	407	11	1978	1978	NUM
ejpam-1184	407	12	.	.	PUNCT
ejpam-1184	408	1	[	[	X
ejpam-1184	408	2	16	16	NUM
ejpam-1184	408	3	]	]	X
ejpam-1184	408	4	w	w	PROPN
ejpam-1184	408	5	desarbo	desarbo	PROPN
ejpam-1184	408	6	and	and	CCONJ
ejpam-1184	408	7	w	w	PROPN
ejpam-1184	408	8	cron	cron	PROPN
ejpam-1184	408	9	.	.	PUNCT
ejpam-1184	409	1	a	a	DET
ejpam-1184	409	2	maximum	maximum	ADJ
ejpam-1184	409	3	likelihood	likelihood	NOUN
ejpam-1184	409	4	methodology	methodology	NOUN
ejpam-1184	409	5	for	for	ADP
ejpam-1184	409	6	clusterwise	clusterwise	ADJ
ejpam-1184	409	7	linear	linear	PROPN
ejpam-1184	409	8	regression	regression	NOUN
ejpam-1184	409	9	.	.	PUNCT
ejpam-1184	410	1	journal	journal	NOUN
ejpam-1184	410	2	of	of	ADP
ejpam-1184	410	3	classification	classification	NOUN
ejpam-1184	410	4	,	,	PUNCT
ejpam-1184	410	5	5:249–282	5:249–282	PROPN
ejpam-1184	410	6	,	,	PUNCT
ejpam-1184	410	7	1988	1988	NUM
ejpam-1184	410	8	.	.	PUNCT