id	sid	tid	token	lemma	pos
fcis-32254	1	1	frontiers	frontier	NOUN
fcis-32254	1	2	in	in	ADP
fcis-32254	1	3	computing	computing	NOUN
fcis-32254	1	4	and	and	CCONJ
fcis-32254	1	5	intelligent	intelligent	ADJ
fcis-32254	1	6	systems	system	NOUN
fcis-32254	1	7	issn	issn	VERB
fcis-32254	1	8	:	:	PUNCT
fcis-32254	1	9	2832	2832	NUM
fcis-32254	1	10	-	-	SYM
fcis-32254	1	11	6024	6024	NUM
fcis-32254	1	12	|	|	NOUN
fcis-32254	1	13	vol	vol	NOUN
fcis-32254	1	14	.	.	PUNCT
fcis-32254	2	1	14	14	NUM
fcis-32254	2	2	,	,	PUNCT
fcis-32254	2	3	no	no	INTJ
fcis-32254	2	4	.	.	NOUN
fcis-32254	2	5	2	2	NUM
fcis-32254	2	6	,	,	PUNCT
fcis-32254	2	7	2025	2025	NUM
fcis-32254	2	8	20	20	NUM
fcis-32254	2	9	application	application	NOUN
fcis-32254	2	10	of	of	ADP
fcis-32254	2	11	sparse	sparse	ADJ
fcis-32254	2	12	kernel	kernel	NOUN
fcis-32254	2	13	graph‐regularized	graph‐regularize	VERB
fcis-32254	2	14	discriminant	discriminant	ADJ
fcis-32254	2	15	non‐negative	non‐negative	ADJ
fcis-32254	2	16	matrix	matrix	NOUN
fcis-32254	2	17	factorization	factorization	NOUN
fcis-32254	2	18	in	in	ADP
fcis-32254	2	19	image	image	NOUN
fcis-32254	2	20	clustering	clustering	NOUN
fcis-32254	2	21	ran	run	VERB
fcis-32254	2	22	zhou	zhou	PROPN
fcis-32254	2	23	,	,	PUNCT
fcis-32254	2	24	jianlei	jianlei	VERB
fcis-32254	2	25	li	li	PROPN
fcis-32254	2	26	*	*	PUNCT
fcis-32254	2	27	electronic	electronic	ADJ
fcis-32254	2	28	information	information	NOUN
fcis-32254	2	29	major	major	NOUN
fcis-32254	2	30	,	,	PUNCT
fcis-32254	2	31	school	school	NOUN
fcis-32254	2	32	of	of	ADP
fcis-32254	2	33	mathematics	mathematic	NOUN
fcis-32254	2	34	and	and	CCONJ
fcis-32254	2	35	statistics	statistic	NOUN
fcis-32254	2	36	,	,	PUNCT
fcis-32254	2	37	north	north	PROPN
fcis-32254	2	38	china	china	PROPN
fcis-32254	2	39	university	university	PROPN
fcis-32254	2	40	of	of	ADP
fcis-32254	2	41	water	water	NOUN
fcis-32254	2	42	resources	resource	NOUN
fcis-32254	2	43	and	and	CCONJ
fcis-32254	2	44	electric	electric	ADJ
fcis-32254	2	45	power	power	NOUN
fcis-32254	2	46	,	,	PUNCT
fcis-32254	2	47	zhengzhou	zhengzhou	PROPN
fcis-32254	2	48	,	,	PUNCT
fcis-32254	2	49	henan	henan	PROPN
fcis-32254	2	50	province	province	PROPN
fcis-32254	2	51	,	,	PUNCT
fcis-32254	2	52	450046	450046	NUM
fcis-32254	2	53	,	,	PUNCT
fcis-32254	2	54	china	china	PROPN
fcis-32254	2	55	*	*	PUNCT
fcis-32254	2	56	corresponding	correspond	VERB
fcis-32254	2	57	author	author	NOUN
fcis-32254	2	58	:	:	PUNCT
fcis-32254	2	59	jianlei	jianlei	PROPN
fcis-32254	2	60	li	li	PROPN
fcis-32254	2	61	(	(	PUNCT
fcis-32254	2	62	email	email	NOUN
fcis-32254	2	63	:	:	PUNCT
fcis-32254	2	64	lijianlei@ncwu.edu.cn	lijianlei@ncwu.edu.cn	ADJ
fcis-32254	2	65	)	)	PUNCT
fcis-32254	2	66	abstract	abstract	NOUN
fcis-32254	2	67	:	:	PUNCT
fcis-32254	2	68	non	non	ADJ
fcis-32254	2	69	-	-	ADJ
fcis-32254	2	70	negative	negative	ADJ
fcis-32254	2	71	matrix	matrix	NOUN
fcis-32254	2	72	factorization	factorization	NOUN
fcis-32254	2	73	(	(	PUNCT
fcis-32254	2	74	nmf	nmf	NOUN
fcis-32254	2	75	)	)	PUNCT
fcis-32254	2	76	is	be	AUX
fcis-32254	2	77	widely	widely	ADV
fcis-32254	2	78	used	use	VERB
fcis-32254	2	79	in	in	ADP
fcis-32254	2	80	image	image	NOUN
fcis-32254	2	81	clustering	clustering	NOUN
fcis-32254	2	82	;	;	PUNCT
fcis-32254	2	83	however	however	ADV
fcis-32254	2	84	,	,	PUNCT
fcis-32254	2	85	it	it	PRON
fcis-32254	2	86	has	have	VERB
fcis-32254	2	87	inherent	inherent	ADJ
fcis-32254	2	88	limitations	limitation	NOUN
fcis-32254	2	89	,	,	PUNCT
fcis-32254	2	90	including	include	VERB
fcis-32254	2	91	its	its	PRON
fcis-32254	2	92	unsupervised	unsupervised	ADJ
fcis-32254	2	93	nature	nature	NOUN
fcis-32254	2	94	,	,	PUNCT
fcis-32254	2	95	lack	lack	NOUN
fcis-32254	2	96	of	of	ADP
fcis-32254	2	97	sparsity	sparsity	NOUN
fcis-32254	2	98	constraints	constraint	NOUN
fcis-32254	2	99	,	,	PUNCT
fcis-32254	2	100	inability	inability	NOUN
fcis-32254	2	101	to	to	PART
fcis-32254	2	102	leverage	leverage	VERB
fcis-32254	2	103	label	label	NOUN
fcis-32254	2	104	information	information	NOUN
fcis-32254	2	105	,	,	PUNCT
fcis-32254	2	106	and	and	CCONJ
fcis-32254	2	107	difficulty	difficulty	NOUN
fcis-32254	2	108	in	in	ADP
fcis-32254	2	109	capturing	capture	VERB
fcis-32254	2	110	the	the	DET
fcis-32254	2	111	geometric	geometric	ADJ
fcis-32254	2	112	structure	structure	NOUN
fcis-32254	2	113	and	and	CCONJ
fcis-32254	2	114	nonlinear	nonlinear	ADJ
fcis-32254	2	115	characteristics	characteristic	NOUN
fcis-32254	2	116	of	of	ADP
fcis-32254	2	117	data	datum	NOUN
fcis-32254	2	118	.	.	PUNCT
fcis-32254	3	1	to	to	PART
fcis-32254	3	2	address	address	VERB
fcis-32254	3	3	these	these	DET
fcis-32254	3	4	limitations	limitation	NOUN
fcis-32254	3	5	,	,	PUNCT
fcis-32254	3	6	this	this	DET
fcis-32254	3	7	paper	paper	NOUN
fcis-32254	3	8	proposes	propose	VERB
fcis-32254	3	9	a	a	DET
fcis-32254	3	10	sparse	sparse	ADJ
fcis-32254	3	11	kernel	kernel	NOUN
fcis-32254	3	12	graph	graph	NOUN
fcis-32254	3	13	-	-	PUNCT
fcis-32254	3	14	regularized	regularize	VERB
fcis-32254	3	15	discriminant	discriminant	ADJ
fcis-32254	3	16	non	non	ADJ
fcis-32254	3	17	-	-	ADJ
fcis-32254	3	18	negative	negative	ADJ
fcis-32254	3	19	matrix	matrix	NOUN
fcis-32254	3	20	factorization	factorization	NOUN
fcis-32254	3	21	(	(	PUNCT
fcis-32254	3	22	skgdnmf	skgdnmf	NOUN
fcis-32254	3	23	)	)	PUNCT
fcis-32254	3	24	algorithm	algorithm	NOUN
fcis-32254	3	25	.	.	PUNCT
fcis-32254	4	1	the	the	DET
fcis-32254	4	2	algorithm	algorithm	NOUN
fcis-32254	4	3	innovatively	innovatively	ADV
fcis-32254	4	4	adopts	adopt	VERB
fcis-32254	4	5	a	a	DET
fcis-32254	4	6	dual	dual	ADJ
fcis-32254	4	7	normalization	normalization	NOUN
fcis-32254	4	8	strategy	strategy	NOUN
fcis-32254	4	9	,	,	PUNCT
fcis-32254	4	10	which	which	PRON
fcis-32254	4	11	involves	involve	VERB
fcis-32254	4	12	column	column	NOUN
fcis-32254	4	13	normalization	normalization	NOUN
fcis-32254	4	14	for	for	ADP
fcis-32254	4	15	the	the	DET
fcis-32254	4	16	basis	basis	NOUN
fcis-32254	4	17	matrix	matrix	NOUN
fcis-32254	4	18	and	and	CCONJ
fcis-32254	4	19	row	row	NOUN
fcis-32254	4	20	normalization	normalization	NOUN
fcis-32254	4	21	for	for	ADP
fcis-32254	4	22	the	the	DET
fcis-32254	4	23	coefficient	coefficient	NOUN
fcis-32254	4	24	matrix	matrix	NOUN
fcis-32254	4	25	,	,	PUNCT
fcis-32254	4	26	and	and	CCONJ
fcis-32254	4	27	integrates	integrate	NOUN
fcis-32254	4	28	spectral	spectral	ADJ
fcis-32254	4	29	clustering	clustering	NOUN
fcis-32254	4	30	to	to	PART
fcis-32254	4	31	construct	construct	VERB
fcis-32254	4	32	an	an	DET
fcis-32254	4	33	end	end	NOUN
fcis-32254	4	34	-	-	PUNCT
fcis-32254	4	35	to	to	ADP
fcis-32254	4	36	-	-	PUNCT
fcis-32254	4	37	end	end	NOUN
fcis-32254	4	38	deep	deep	ADJ
fcis-32254	4	39	clustering	clustering	ADJ
fcis-32254	4	40	framework	framework	NOUN
fcis-32254	4	41	.	.	PUNCT
fcis-32254	5	1	by	by	ADP
fcis-32254	5	2	applying	apply	VERB
fcis-32254	5	3	sparse	sparse	ADJ
fcis-32254	5	4	regularization	regularization	NOUN
fcis-32254	5	5	to	to	ADP
fcis-32254	5	6	the	the	DET
fcis-32254	5	7	coefficient	coefficient	NOUN
fcis-32254	5	8	matrix	matrix	NOUN
fcis-32254	5	9	,	,	PUNCT
fcis-32254	5	10	the	the	DET
fcis-32254	5	11	model	model	NOUN
fcis-32254	5	12	’s	’s	PART
fcis-32254	5	13	robustness	robustness	NOUN
fcis-32254	5	14	is	be	AUX
fcis-32254	5	15	significantly	significantly	ADV
fcis-32254	5	16	improved	improve	VERB
fcis-32254	5	17	;	;	PUNCT
fcis-32254	5	18	this	this	DET
fcis-32254	5	19	regularization	regularization	NOUN
fcis-32254	5	20	forms	form	VERB
fcis-32254	5	21	complementary	complementary	ADJ
fcis-32254	5	22	optimization	optimization	NOUN
fcis-32254	5	23	with	with	ADP
fcis-32254	5	24	graph	graph	NOUN
fcis-32254	5	25	regularization	regularization	NOUN
fcis-32254	5	26	,	,	PUNCT
fcis-32254	5	27	thereby	thereby	ADV
fcis-32254	5	28	effectively	effectively	ADV
fcis-32254	5	29	alleviating	alleviate	VERB
fcis-32254	5	30	overfitting	overfitting	NOUN
fcis-32254	5	31	.	.	PUNCT
fcis-32254	6	1	a	a	DET
fcis-32254	6	2	three	three	NUM
fcis-32254	6	3	-	-	PUNCT
fcis-32254	6	4	dimensional	dimensional	ADJ
fcis-32254	6	5	golden	golden	ADJ
fcis-32254	6	6	-	-	PUNCT
fcis-32254	6	7	section	section	NOUN
fcis-32254	6	8	parameter	parameter	NOUN
fcis-32254	6	9	optimization	optimization	NOUN
fcis-32254	6	10	method	method	NOUN
fcis-32254	6	11	is	be	AUX
fcis-32254	6	12	employed	employ	VERB
fcis-32254	6	13	to	to	PART
fcis-32254	6	14	determine	determine	VERB
fcis-32254	6	15	key	key	ADJ
fcis-32254	6	16	parameters	parameter	NOUN
fcis-32254	6	17	,	,	PUNCT
fcis-32254	6	18	which	which	PRON
fcis-32254	6	19	enhances	enhance	VERB
fcis-32254	6	20	the	the	DET
fcis-32254	6	21	algorithm	algorithm	NOUN
fcis-32254	6	22	’s	’s	PART
fcis-32254	6	23	practicality	practicality	NOUN
fcis-32254	6	24	.	.	PUNCT
fcis-32254	7	1	comparative	comparative	ADJ
fcis-32254	7	2	experiments	experiment	NOUN
fcis-32254	7	3	conducted	conduct	VERB
fcis-32254	7	4	on	on	ADP
fcis-32254	7	5	multiple	multiple	ADJ
fcis-32254	7	6	datasets	dataset	NOUN
fcis-32254	7	7	show	show	VERB
fcis-32254	7	8	that	that	SCONJ
fcis-32254	7	9	skgdnmf	skgdnmf	NOUN
fcis-32254	7	10	significantly	significantly	ADV
fcis-32254	7	11	outperforms	outperform	VERB
fcis-32254	7	12	mainstream	mainstream	ADJ
fcis-32254	7	13	algorithms	algorithm	NOUN
fcis-32254	7	14	in	in	ADP
fcis-32254	7	15	terms	term	NOUN
fcis-32254	7	16	of	of	ADP
fcis-32254	7	17	sparsity	sparsity	NOUN
fcis-32254	7	18	,	,	PUNCT
fcis-32254	7	19	robustness	robustness	NOUN
fcis-32254	7	20	,	,	PUNCT
fcis-32254	7	21	and	and	CCONJ
fcis-32254	7	22	clustering	clustering	ADJ
fcis-32254	7	23	performance	performance	NOUN
fcis-32254	7	24	,	,	PUNCT
fcis-32254	7	25	indicating	indicate	VERB
fcis-32254	7	26	its	its	PRON
fcis-32254	7	27	superior	superior	ADJ
fcis-32254	7	28	effectiveness	effectiveness	NOUN
fcis-32254	7	29	for	for	ADP
fcis-32254	7	30	image	image	NOUN
fcis-32254	7	31	clustering	clustering	NOUN
fcis-32254	7	32	tasks	task	NOUN
fcis-32254	7	33	.	.	PUNCT
fcis-32254	8	1	keywords	keyword	NOUN
fcis-32254	8	2	:	:	PUNCT
fcis-32254	8	3	graph	graph	NOUN
fcis-32254	8	4	regularization	regularization	NOUN
fcis-32254	8	5	;	;	PUNCT
fcis-32254	8	6	image	image	NOUN
fcis-32254	8	7	clustering	clustering	NOUN
fcis-32254	8	8	;	;	PUNCT
fcis-32254	8	9	non	non	ADJ
fcis-32254	8	10	-	-	ADJ
fcis-32254	8	11	negative	negative	ADJ
fcis-32254	8	12	matrix	matrix	NOUN
fcis-32254	8	13	factorization	factorization	NOUN
fcis-32254	8	14	;	;	PUNCT
fcis-32254	8	15	sparsity	sparsity	NOUN
fcis-32254	8	16	;	;	PUNCT
fcis-32254	8	17	spectral	spectral	ADJ
fcis-32254	8	18	clustering	clustering	NOUN
fcis-32254	8	19	.	.	PUNCT
fcis-32254	9	1	1	1	X
fcis-32254	9	2	.	.	X
fcis-32254	9	3	introduction	introduction	NOUN
fcis-32254	9	4	in	in	ADP
fcis-32254	9	5	image	image	NOUN
fcis-32254	9	6	clustering	clustering	ADJ
fcis-32254	9	7	algorithms	algorithm	NOUN
fcis-32254	9	8	,	,	PUNCT
fcis-32254	9	9	data	datum	NOUN
fcis-32254	9	10	dimensionality	dimensionality	NOUN
fcis-32254	9	11	reduction	reduction	NOUN
fcis-32254	9	12	is	be	AUX
fcis-32254	9	13	a	a	DET
fcis-32254	9	14	key	key	ADJ
fcis-32254	9	15	focus	focus	NOUN
fcis-32254	9	16	of	of	ADP
fcis-32254	9	17	researchers	researcher	NOUN
fcis-32254	9	18	.	.	PUNCT
fcis-32254	10	1	data	datum	NOUN
fcis-32254	10	2	usually	usually	ADV
fcis-32254	10	3	changes	change	VERB
fcis-32254	10	4	before	before	ADV
fcis-32254	10	5	and	and	CCONJ
fcis-32254	10	6	after	after	ADP
fcis-32254	10	7	dimensionality	dimensionality	NOUN
fcis-32254	10	8	reduction	reduction	NOUN
fcis-32254	10	9	,	,	PUNCT
fcis-32254	10	10	and	and	CCONJ
fcis-32254	10	11	based	base	VERB
fcis-32254	10	12	on	on	ADP
fcis-32254	10	13	these	these	DET
fcis-32254	10	14	changes	change	NOUN
fcis-32254	10	15	,	,	PUNCT
fcis-32254	10	16	dimensionality	dimensionality	NOUN
fcis-32254	10	17	reduction	reduction	NOUN
fcis-32254	10	18	methods	method	NOUN
fcis-32254	10	19	are	be	AUX
fcis-32254	10	20	categorized	categorize	VERB
fcis-32254	10	21	into	into	ADP
fcis-32254	10	22	nonlinear	nonlinear	ADJ
fcis-32254	10	23	dimensionality	dimensionality	NOUN
fcis-32254	10	24	reduction	reduction	NOUN
fcis-32254	10	25	and	and	CCONJ
fcis-32254	10	26	linear	linear	ADJ
fcis-32254	10	27	dimensionality	dimensionality	NOUN
fcis-32254	10	28	reduction	reduction	NOUN
fcis-32254	10	29	.	.	PUNCT
fcis-32254	11	1	in	in	ADP
fcis-32254	11	2	the	the	DET
fcis-32254	11	3	real	real	ADJ
fcis-32254	11	4	world	world	NOUN
fcis-32254	11	5	,	,	PUNCT
fcis-32254	11	6	however	however	ADV
fcis-32254	11	7	,	,	PUNCT
fcis-32254	11	8	there	there	PRON
fcis-32254	11	9	exists	exist	VERB
fcis-32254	11	10	a	a	DET
fcis-32254	11	11	large	large	ADJ
fcis-32254	11	12	amount	amount	NOUN
fcis-32254	11	13	of	of	ADP
fcis-32254	11	14	data	datum	NOUN
fcis-32254	11	15	,	,	PUNCT
fcis-32254	11	16	and	and	CCONJ
fcis-32254	11	17	such	such	ADJ
fcis-32254	11	18	data	datum	NOUN
fcis-32254	11	19	is	be	AUX
fcis-32254	11	20	often	often	ADV
fcis-32254	11	21	nonnegative	nonnegative	ADJ
fcis-32254	11	22	;	;	PUNCT
fcis-32254	11	23	otherwise	otherwise	ADV
fcis-32254	11	24	,	,	PUNCT
fcis-32254	11	25	it	it	PRON
fcis-32254	11	26	would	would	AUX
fcis-32254	11	27	be	be	AUX
fcis-32254	11	28	meaningless	meaningless	ADJ
fcis-32254	11	29	.	.	PUNCT
fcis-32254	12	1	therefore	therefore	ADV
fcis-32254	12	2	,	,	PUNCT
fcis-32254	12	3	the	the	DET
fcis-32254	12	4	non	non	ADJ
fcis-32254	12	5	-	-	ADJ
fcis-32254	12	6	negative	negative	ADJ
fcis-32254	12	7	matrix	matrix	NOUN
fcis-32254	12	8	factorization	factorization	NOUN
fcis-32254	12	9	(	(	PUNCT
fcis-32254	12	10	nmf	nmf	NOUN
fcis-32254	12	11	)	)	PUNCT
fcis-32254	12	12	algorithm	algorithm	NOUN
fcis-32254	12	13	is	be	AUX
fcis-32254	12	14	widely	widely	ADV
fcis-32254	12	15	applied	apply	VERB
fcis-32254	12	16	in	in	ADP
fcis-32254	12	17	data	datum	NOUN
fcis-32254	12	18	feature	feature	NOUN
fcis-32254	12	19	extraction	extraction	NOUN
fcis-32254	12	20	and	and	CCONJ
fcis-32254	12	21	has	have	AUX
fcis-32254	12	22	also	also	ADV
fcis-32254	12	23	been	be	AUX
fcis-32254	12	24	proven	prove	VERB
fcis-32254	12	25	equivalent	equivalent	ADJ
fcis-32254	12	26	to	to	ADP
fcis-32254	12	27	many	many	ADJ
fcis-32254	12	28	clustering	clustering	ADJ
fcis-32254	12	29	methods	method	NOUN
fcis-32254	12	30	.	.	PUNCT
fcis-32254	13	1	inspired	inspire	VERB
fcis-32254	13	2	by	by	ADP
fcis-32254	13	3	the	the	DET
fcis-32254	13	4	non	non	ADJ
fcis-32254	13	5	-	-	ADJ
fcis-32254	13	6	negative	negative	ADJ
fcis-32254	13	7	matrix	matrix	NOUN
fcis-32254	13	8	factorization	factorization	NOUN
fcis-32254	13	9	algorithm	algorithm	NOUN
fcis-32254	13	10	,	,	PUNCT
fcis-32254	13	11	this	this	DET
fcis-32254	13	12	paper	paper	NOUN
fcis-32254	13	13	proposes	propose	VERB
fcis-32254	13	14	a	a	DET
fcis-32254	13	15	kernel	kernel	NOUN
fcis-32254	13	16	-	-	PUNCT
fcis-32254	13	17	based	base	VERB
fcis-32254	13	18	sparse	sparse	ADJ
fcis-32254	13	19	discriminative	discriminative	NOUN
fcis-32254	13	20	graph	graph	NOUN
fcis-32254	13	21	-	-	PUNCT
fcis-32254	13	22	regularized	regularize	VERB
fcis-32254	13	23	non	non	ADJ
fcis-32254	13	24	-	-	ADJ
fcis-32254	13	25	negative	negative	ADJ
fcis-32254	13	26	matrix	matrix	NOUN
fcis-32254	13	27	factorization	factorization	NOUN
fcis-32254	13	28	(	(	PUNCT
fcis-32254	13	29	skgdnmf	skgdnmf	NOUN
fcis-32254	13	30	)	)	PUNCT
fcis-32254	13	31	algorithm	algorithm	NOUN
fcis-32254	13	32	.	.	PUNCT
fcis-32254	14	1	this	this	DET
fcis-32254	14	2	algorithm	algorithm	NOUN
fcis-32254	14	3	is	be	AUX
fcis-32254	14	4	a	a	DET
fcis-32254	14	5	semi	semi	ADJ
fcis-32254	14	6	-	-	ADJ
fcis-32254	14	7	supervised	supervised	ADJ
fcis-32254	14	8	method	method	NOUN
fcis-32254	14	9	that	that	PRON
fcis-32254	14	10	mainly	mainly	ADV
fcis-32254	14	11	uses	use	VERB
fcis-32254	14	12	label	label	NOUN
fcis-32254	14	13	information	information	NOUN
fcis-32254	14	14	to	to	PART
fcis-32254	14	15	construct	construct	VERB
fcis-32254	14	16	a	a	DET
fcis-32254	14	17	label	label	NOUN
fcis-32254	14	18	matrix	matrix	NOUN
fcis-32254	14	19	.	.	PUNCT
fcis-32254	15	1	meanwhile	meanwhile	ADV
fcis-32254	15	2	,	,	PUNCT
fcis-32254	15	3	it	it	PRON
fcis-32254	15	4	incorporates	incorporate	VERB
fcis-32254	15	5	a	a	DET
fcis-32254	15	6	graph	graph	NOUN
fcis-32254	15	7	regularization	regularization	NOUN
fcis-32254	15	8	term	term	NOUN
fcis-32254	15	9	to	to	PART
fcis-32254	15	10	capture	capture	VERB
fcis-32254	15	11	the	the	DET
fcis-32254	15	12	geometric	geometric	ADJ
fcis-32254	15	13	structure	structure	NOUN
fcis-32254	15	14	of	of	ADP
fcis-32254	15	15	data	datum	NOUN
fcis-32254	15	16	,	,	PUNCT
fcis-32254	15	17	adopts	adopt	VERB
fcis-32254	15	18	a	a	DET
fcis-32254	15	19	kernel	kernel	NOUN
fcis-32254	15	20	method	method	NOUN
fcis-32254	15	21	to	to	PART
fcis-32254	15	22	handle	handle	VERB
fcis-32254	15	23	the	the	DET
fcis-32254	15	24	nonlinear	nonlinear	ADJ
fcis-32254	15	25	nature	nature	NOUN
fcis-32254	15	26	of	of	ADP
fcis-32254	15	27	data	datum	NOUN
fcis-32254	15	28	,	,	PUNCT
fcis-32254	15	29	and	and	CCONJ
fcis-32254	15	30	adds	add	VERB
fcis-32254	15	31	an	an	DET
fcis-32254	15	32	1l	1l	NUM
fcis-32254	15	33	regularization	regularization	NOUN
fcis-32254	15	34	term	term	NOUN
fcis-32254	15	35	to	to	PART
fcis-32254	15	36	reduce	reduce	VERB
fcis-32254	15	37	the	the	DET
fcis-32254	15	38	impact	impact	NOUN
fcis-32254	15	39	of	of	ADP
fcis-32254	15	40	noise	noise	NOUN
fcis-32254	15	41	on	on	ADP
fcis-32254	15	42	clustering	clustering	ADJ
fcis-32254	15	43	results	result	NOUN
fcis-32254	15	44	and	and	CCONJ
fcis-32254	15	45	enhance	enhance	VERB
fcis-32254	15	46	the	the	DET
fcis-32254	15	47	discriminative	discriminative	NOUN
fcis-32254	15	48	power	power	NOUN
fcis-32254	15	49	of	of	ADP
fcis-32254	15	50	clustering	clustering	NOUN
fcis-32254	15	51	.	.	PUNCT
fcis-32254	16	1	2	2	X
fcis-32254	16	2	.	.	X
fcis-32254	16	3	sparse	sparse	ADJ
fcis-32254	16	4	kernel	kernel	NOUN
fcis-32254	16	5	graph	graph	VERB
fcis-32254	16	6	discriminative	discriminative	ADJ
fcis-32254	16	7	non	non	ADJ
fcis-32254	16	8	-	-	ADJ
fcis-32254	16	9	negative	negative	ADJ
fcis-32254	16	10	matrix	matrix	NOUN
fcis-32254	16	11	factorization	factorization	NOUN
fcis-32254	16	12	algorithm	algorithm	NOUN
fcis-32254	16	13	2.1	2.1	NUM
fcis-32254	16	14	.	.	PUNCT
fcis-32254	17	1	skgdnmf	skgdnmf	PROPN
fcis-32254	17	2	algorithm	algorithm	NOUN
fcis-32254	17	3	this	this	DET
fcis-32254	17	4	paper	paper	NOUN
fcis-32254	17	5	combines	combine	VERB
fcis-32254	17	6	three	three	NUM
fcis-32254	17	7	algorithms	algorithm	NOUN
fcis-32254	17	8	,	,	PUNCT
fcis-32254	17	9	namely	namely	ADV
fcis-32254	17	10	gnmf	gnmf	NOUN
fcis-32254	17	11	[	[	X
fcis-32254	17	12	1	1	NUM
fcis-32254	17	13	]	]	PUNCT
fcis-32254	17	14	,	,	PUNCT
fcis-32254	17	15	dnmf	dnmf	NOUN
fcis-32254	17	16	[	[	X
fcis-32254	17	17	2	2	NUM
fcis-32254	17	18	]	]	PUNCT
fcis-32254	17	19	,	,	PUNCT
fcis-32254	17	20	and	and	CCONJ
fcis-32254	17	21	knmf	knmf	NOUN
fcis-32254	18	1	[	[	X
fcis-32254	18	2	3	3	NUM
fcis-32254	18	3	]	]	PUNCT
fcis-32254	18	4	,	,	PUNCT
fcis-32254	19	1	and	and	CCONJ
fcis-32254	19	2	incorporates	incorporate	VERB
fcis-32254	19	3	a	a	DET
fcis-32254	19	4	sparsity	sparsity	NOUN
fcis-32254	19	5	regularization	regularization	NOUN
fcis-32254	19	6	term	term	NOUN
fcis-32254	19	7	into	into	ADP
fcis-32254	19	8	the	the	DET
fcis-32254	19	9	model	model	NOUN
fcis-32254	19	10	to	to	PART
fcis-32254	19	11	propose	propose	VERB
fcis-32254	19	12	a	a	DET
fcis-32254	19	13	new	new	ADJ
fcis-32254	19	14	algorithm	algorithm	NOUN
fcis-32254	19	15	.	.	PUNCT
fcis-32254	20	1	since	since	SCONJ
fcis-32254	20	2	the	the	DET
fcis-32254	20	3	nmf	nmf	PROPN
fcis-32254	20	4	algorithm	algorithm	NOUN
fcis-32254	20	5	is	be	AUX
fcis-32254	20	6	solved	solve	VERB
fcis-32254	20	7	through	through	ADP
fcis-32254	20	8	an	an	DET
fcis-32254	20	9	iterative	iterative	NOUN
fcis-32254	20	10	update	update	NOUN
fcis-32254	20	11	approach	approach	NOUN
fcis-32254	20	12	,	,	PUNCT
fcis-32254	20	13	and	and	CCONJ
fcis-32254	20	14	the	the	DET
fcis-32254	20	15	matrix	matrix	NOUN
fcis-32254	20	16	for	for	ADP
fcis-32254	20	17	initial	initial	ADJ
fcis-32254	20	18	iteration	iteration	NOUN
fcis-32254	20	19	is	be	AUX
fcis-32254	20	20	randomly	randomly	ADV
fcis-32254	20	21	generated	generate	VERB
fcis-32254	20	22	,	,	PUNCT
fcis-32254	20	23	this	this	DET
fcis-32254	20	24	paper	paper	NOUN
fcis-32254	20	25	,	,	PUNCT
fcis-32254	20	26	inspired	inspire	VERB
fcis-32254	20	27	by	by	ADP
fcis-32254	20	28	reference	reference	NOUN
fcis-32254	20	29	[	[	X
fcis-32254	20	30	4	4	NUM
fcis-32254	20	31	]	]	PUNCT
fcis-32254	20	32	,	,	PUNCT
fcis-32254	20	33	adopts	adopt	VERB
fcis-32254	20	34	a	a	DET
fcis-32254	20	35	"	"	PUNCT
fcis-32254	20	36	warm	warm	ADJ
fcis-32254	20	37	-	-	PUNCT
fcis-32254	20	38	start	start	NOUN
fcis-32254	20	39	"	"	PUNCT
fcis-32254	20	40	strategy	strategy	NOUN
fcis-32254	20	41	[	[	X
fcis-32254	20	42	4	4	X
fcis-32254	20	43	]	]	PUNCT
fcis-32254	20	44	at	at	ADP
fcis-32254	20	45	the	the	DET
fcis-32254	20	46	initial	initial	ADJ
fcis-32254	20	47	stage	stage	NOUN
fcis-32254	20	48	of	of	ADP
fcis-32254	20	49	the	the	DET
fcis-32254	20	50	algorithm	algorithm	NOUN
fcis-32254	20	51	.	.	PUNCT
fcis-32254	21	1	the	the	DET
fcis-32254	21	2	warm	warm	ADJ
fcis-32254	21	3	-	-	PUNCT
fcis-32254	21	4	start	start	NOUN
fcis-32254	21	5	strategy	strategy	NOUN
fcis-32254	21	6	first	first	ADV
fcis-32254	21	7	performs	perform	VERB
fcis-32254	21	8	k	k	NOUN
fcis-32254	21	9	-	-	PUNCT
fcis-32254	21	10	means	means	NOUN
fcis-32254	21	11	clustering	cluster	VERB
fcis-32254	21	12	on	on	ADP
fcis-32254	21	13	the	the	DET
fcis-32254	21	14	data	datum	NOUN
fcis-32254	21	15	,	,	PUNCT
fcis-32254	21	16	and	and	CCONJ
fcis-32254	21	17	uses	use	VERB
fcis-32254	21	18	the	the	DET
fcis-32254	21	19	obtained	obtain	VERB
fcis-32254	21	20	cluster	cluster	NOUN
fcis-32254	21	21	centers	center	NOUN
fcis-32254	21	22	as	as	ADP
fcis-32254	21	23	the	the	DET
fcis-32254	21	24	initial	initial	ADJ
fcis-32254	21	25	basis	basis	NOUN
fcis-32254	21	26	matrix	matrix	NOUN
fcis-32254	21	27	of	of	ADP
fcis-32254	21	28	the	the	DET
fcis-32254	21	29	model	model	NOUN
fcis-32254	21	30	.	.	PUNCT
fcis-32254	22	1	based	base	VERB
fcis-32254	22	2	on	on	ADP
fcis-32254	22	3	the	the	DET
fcis-32254	22	4	above	above	ADJ
fcis-32254	22	5	,	,	PUNCT
fcis-32254	22	6	sparse	sparse	ADJ
fcis-32254	22	7	kernel	kernel	NOUN
fcis-32254	22	8	graph	graph	VERB
fcis-32254	22	9	discriminative	discriminative	ADJ
fcis-32254	22	10	non	non	ADJ
fcis-32254	22	11	-	-	ADJ
fcis-32254	22	12	negative	negative	ADJ
fcis-32254	22	13	matrix	matrix	NOUN
fcis-32254	22	14	factorization	factorization	NOUN
fcis-32254	22	15	(	(	PUNCT
fcis-32254	22	16	skgdnmf	skgdnmf	NOUN
fcis-32254	22	17	)	)	PUNCT
fcis-32254	22	18	is	be	AUX
fcis-32254	22	19	proposed	propose	VERB
fcis-32254	22	20	.	.	PUNCT
fcis-32254	23	1	the	the	DET
fcis-32254	23	2	model	model	NOUN
fcis-32254	23	3	of	of	ADP
fcis-32254	23	4	the	the	DET
fcis-32254	23	5	algorithm	algorithm	NOUN
fcis-32254	23	6	is	be	AUX
fcis-32254	23	7	presented	present	VERB
fcis-32254	23	8	as	as	SCONJ
fcis-32254	23	9	follows	follow	VERB
fcis-32254	23	10	:	:	PUNCT
fcis-32254	23	11	2	2	NUM
fcis-32254	23	12	2	2	NUM
fcis-32254	23	13	l	l	NOUN
fcis-32254	23	14	0	0	NUM
fcis-32254	23	15	,	,	PUNCT
fcis-32254	23	16	0	0	NUM
fcis-32254	23	17	2	2	NUM
fcis-32254	23	18	0	0	NUM
fcis-32254	23	19	1	1	NUM
fcis-32254	23	20	min	min	NOUN
fcis-32254	23	21	(	(	PUNCT
fcis-32254	23	22	)	)	PUNCT
fcis-32254	23	23	(	(	PUNCT
fcis-32254	23	24	)	)	PUNCT
fcis-32254	24	1	tr	tr	VERB
fcis-32254	24	2	(	(	PUNCT
fcis-32254	24	3	)	)	PUNCT
fcis-32254	24	4	t	t	NOUN
fcis-32254	24	5	t	t	PROPN
fcis-32254	24	6	f	f	PROPN
fcis-32254	24	7	fu	fu	PROPN
fcis-32254	24	8	v	v	ADP
fcis-32254	24	9	t	t	PROPN
fcis-32254	25	1	f	f	X
fcis-32254	25	2	x	x	SYM
fcis-32254	25	3	u	u	NOUN
fcis-32254	25	4	v	v	X
fcis-32254	25	5	q	q	X
fcis-32254	25	6	av	av	PROPN
fcis-32254	25	7	v	v	NUM
fcis-32254	25	8	lv	lv	PROPN
fcis-32254	25	9	u	u	NOUN
fcis-32254	25	10	u	u	NOUN
fcis-32254	25	11	v	v	ADP
fcis-32254	25	12			NOUN
fcis-32254	25	13			NOUN
fcis-32254	25	14			X
fcis-32254	25	15			PROPN
fcis-32254	25	16			PROPN
fcis-32254	25	17			ADJ
fcis-32254	25	18			NUM
fcis-32254	25	19			NUM
fcis-32254	25	20			PROPN
fcis-32254	25	21			PUNCT
fcis-32254	25	22			PROPN
fcis-32254	25	23			PUNCT
fcis-32254	25	24			PUNCT
fcis-32254	25	25			PROPN
fcis-32254	25	26			X
fcis-32254	25	27	(	(	PUNCT
fcis-32254	25	28	1	1	X
fcis-32254	25	29	)	)	PUNCT
fcis-32254	25	30	similar	similar	ADJ
fcis-32254	25	31	to	to	ADP
fcis-32254	25	32	the	the	DET
fcis-32254	25	33	standard	standard	ADJ
fcis-32254	25	34	nmf	nmf	PROPN
fcis-32254	25	35	model	model	NOUN
fcis-32254	25	36	,	,	PUNCT
fcis-32254	25	37	the	the	DET
fcis-32254	25	38	optimization	optimization	NOUN
fcis-32254	25	39	problem	problem	NOUN
fcis-32254	25	40	of	of	ADP
fcis-32254	25	41	solving	solve	VERB
fcis-32254	25	42	u	u	NOUN
fcis-32254	25	43	and	and	CCONJ
fcis-32254	25	44	v	v	NOUN
fcis-32254	25	45	simultaneously	simultaneously	ADV
fcis-32254	25	46	is	be	AUX
fcis-32254	25	47	non	non	ADJ
fcis-32254	25	48	-	-	ADJ
fcis-32254	25	49	convex	convex	ADJ
fcis-32254	25	50	.	.	PUNCT
fcis-32254	26	1	therefore	therefore	ADV
fcis-32254	26	2	,	,	PUNCT
fcis-32254	26	3	this	this	DET
fcis-32254	26	4	paper	paper	NOUN
fcis-32254	26	5	still	still	ADV
fcis-32254	26	6	adopts	adopt	VERB
fcis-32254	26	7	the	the	DET
fcis-32254	26	8	alternating	alternate	VERB
fcis-32254	26	9	iterative	iterative	NOUN
fcis-32254	26	10	update	update	NOUN
fcis-32254	26	11	approach	approach	NOUN
fcis-32254	26	12	to	to	PART
fcis-32254	26	13	solve	solve	VERB
fcis-32254	26	14	the	the	DET
fcis-32254	26	15	iterative	iterative	NOUN
fcis-32254	26	16	formulas	formula	NOUN
fcis-32254	26	17	.	.	PUNCT
fcis-32254	27	1	the	the	DET
fcis-32254	27	2	objective	objective	ADJ
fcis-32254	27	3	formula	formula	NOUN
fcis-32254	27	4	can	can	AUX
fcis-32254	27	5	be	be	AUX
fcis-32254	27	6	simplified	simplify	VERB
fcis-32254	27	7	to	to	ADP
fcis-32254	27	8	:	:	PUNCT
fcis-32254	27	9	t	t	PROPN
fcis-32254	27	10	t	t	PROPN
fcis-32254	27	11	t	t	PROPN
fcis-32254	27	12	t	t	PROPN
fcis-32254	27	13	t	t	PROPN
fcis-32254	27	14	t	t	PROPN
fcis-32254	27	15	0	0	NUM
fcis-32254	27	16	0	0	NUM
fcis-32254	27	17	0	0	NUM
fcis-32254	27	18	=	=	NOUN
fcis-32254	27	19	tr(k	tr(k	NOUN
fcis-32254	27	20	)	)	PUNCT
fcis-32254	27	21	-2tr(k	-2tr(k	PROPN
fcis-32254	27	22	v	v	NOUN
fcis-32254	27	23	)	)	PUNCT
fcis-32254	28	1	+	+	PROPN
fcis-32254	28	2	tr(vk	tr(vk	PROPN
fcis-32254	28	3	v	v	ADJ
fcis-32254	28	4	)	)	PUNCT
fcis-32254	28	5	αtr(qq	αtr(qq	NOUN
fcis-32254	28	6	)	)	PUNCT
fcis-32254	29	1	2αtr(qv	2αtr(qv	NUM
fcis-32254	29	2	)	)	PUNCT
fcis-32254	29	3	αtr	αtr	PROPN
fcis-32254	29	4	(	(	PUNCT
fcis-32254	29	5	v	v	NOUN
fcis-32254	29	6	v	v	NOUN
fcis-32254	29	7	)	)	PUNCT
fcis-32254	30	1	+	+	SYM
fcis-32254	30	2	βtr(v	βtr(v	PROPN
fcis-32254	30	3	lv	lv	PROPN
fcis-32254	30	4	)	)	PUNCT
fcis-32254	30	5	γtr(u	γtr(u	PROPN
fcis-32254	30	6	)	)	PUNCT
fcis-32254	30	7	2γtr(u	2γtr(u	NUM
fcis-32254	30	8	)	)	PUNCT
fcis-32254	30	9	γtr	γtr	PROPN
fcis-32254	30	10	(	(	PUNCT
fcis-32254	30	11	)	)	PUNCT
fcis-32254	30	12	xx	xx	NUM
fcis-32254	31	1	xu	xu	INTJ
fcis-32254	31	2	uu	uu	INTJ
fcis-32254	32	1	t	t	PROPN
fcis-32254	32	2	l	l	NOUN
fcis-32254	32	3	l	l	X
fcis-32254	32	4	l	l	NOUN
fcis-32254	32	5	t	t	NOUN
fcis-32254	32	6	t	t	PROPN
fcis-32254	32	7	t	t	PROPN
fcis-32254	32	8	a	a	DET
fcis-32254	32	9	a	a	PRON
fcis-32254	32	10	a	a	DET
fcis-32254	32	11	u	u	NOUN
fcis-32254	32	12	u	u	X
fcis-32254	32	13	u	u	NOUN
fcis-32254	32	14	u	u	X
fcis-32254	32	15			PROPN
fcis-32254	32	16			ADV
fcis-32254	32	17			PROPN
fcis-32254	32	18			PUNCT
fcis-32254	32	19			PUNCT
fcis-32254	32	20			PROPN
fcis-32254	32	21			X
fcis-32254	32	22	(	(	PUNCT
fcis-32254	32	23	2	2	NUM
fcis-32254	32	24	)	)	PUNCT
fcis-32254	32	25	by	by	ADP
fcis-32254	32	26	constructing	construct	VERB
fcis-32254	32	27	the	the	DET
fcis-32254	32	28	lagrangian	lagrangian	ADJ
fcis-32254	32	29	function	function	NOUN
fcis-32254	32	30	,	,	PUNCT
fcis-32254	32	31	we	we	PRON
fcis-32254	32	32	can	can	AUX
fcis-32254	32	33	obtain	obtain	VERB
fcis-32254	32	34	:	:	PUNCT
fcis-32254	33	1	t	t	PROPN
fcis-32254	33	2	t	t	PROPN
fcis-32254	33	3	t	t	PROPN
fcis-32254	33	4	t	t	PROPN
fcis-32254	33	5	t	t	PROPN
fcis-32254	33	6	t	t	PROPN
fcis-32254	33	7	0	0	NUM
fcis-32254	33	8	0	0	NUM
fcis-32254	33	9	0	0	NUM
fcis-32254	33	10	=	=	NOUN
fcis-32254	33	11	tr(k	tr(k	NOUN
fcis-32254	33	12	)	)	PUNCT
fcis-32254	33	13	-2tr(k	-2tr(k	PROPN
fcis-32254	33	14	v	v	NOUN
fcis-32254	33	15	)	)	PUNCT
fcis-32254	34	1	+	+	PROPN
fcis-32254	34	2	tr(vk	tr(vk	PROPN
fcis-32254	34	3	v	v	ADJ
fcis-32254	34	4	)	)	PUNCT
fcis-32254	34	5	αtr(qq	αtr(qq	NOUN
fcis-32254	34	6	)	)	PUNCT
fcis-32254	35	1	2αtr(qv	2αtr(qv	NUM
fcis-32254	35	2	)	)	PUNCT
fcis-32254	35	3	αtr	αtr	PROPN
fcis-32254	35	4	(	(	PUNCT
fcis-32254	35	5	v	v	NOUN
fcis-32254	35	6	v	v	NOUN
fcis-32254	35	7	)	)	PUNCT
fcis-32254	36	1	+	+	PROPN
fcis-32254	36	2	βtr(v	βtr(v	PUNCT
fcis-32254	36	3	lv	lv	ADJ
fcis-32254	36	4	)	)	PUNCT
fcis-32254	36	5	γtr(u	γtr(u	PROPN
fcis-32254	36	6	)	)	PUNCT
fcis-32254	36	7	2γtr(u	2γtr(u	NUM
fcis-32254	36	8	)	)	PUNCT
fcis-32254	36	9	γtr	γtr	PROPN
fcis-32254	36	10	(	(	PUNCT
fcis-32254	36	11	)	)	PUNCT
fcis-32254	36	12	tr	tr	VERB
fcis-32254	36	13	(	(	PUNCT
fcis-32254	36	14	u	u	NOUN
fcis-32254	36	15	)	)	PUNCT
fcis-32254	36	16	+	+	NOUN
fcis-32254	36	17	tr	tr	VERB
fcis-32254	36	18	(	(	PUNCT
fcis-32254	36	19	v	v	NOUN
fcis-32254	36	20	)	)	PUNCT
fcis-32254	36	21	xx	xx	NUM
fcis-32254	37	1	xu	xu	INTJ
fcis-32254	37	2	uu	uu	INTJ
fcis-32254	38	1	t	t	PROPN
fcis-32254	38	2	l	l	NOUN
fcis-32254	38	3	l	l	X
fcis-32254	39	1	l	l	NOUN
fcis-32254	39	2	t	t	NOUN
fcis-32254	39	3	t	t	PROPN
fcis-32254	39	4	t	t	PROPN
fcis-32254	39	5	t	t	PROPN
fcis-32254	39	6	t	t	PROPN
fcis-32254	39	7	a	a	DET
fcis-32254	39	8	a	a	PRON
fcis-32254	39	9	a	a	DET
fcis-32254	39	10	u	u	NOUN
fcis-32254	39	11	u	u	NOUN
fcis-32254	39	12	u	u	X
fcis-32254	39	13	u	u	PROPN
fcis-32254	39	14			PROPN
fcis-32254	39	15			PROPN
fcis-32254	39	16			ADV
fcis-32254	39	17			PUNCT
fcis-32254	39	18			PROPN
fcis-32254	39	19			ADV
fcis-32254	39	20			PUNCT
fcis-32254	39	21			PROPN
fcis-32254	39	22			PROPN
fcis-32254	39	23	l	l	NOUN
fcis-32254	39	24	(	(	PUNCT
fcis-32254	39	25	3	3	NUM
fcis-32254	39	26	)	)	SYM
fcis-32254	39	27	21	21	NUM
fcis-32254	39	28	herein	herein	NOUN
fcis-32254	39	29	,	,	PUNCT
fcis-32254	39	30	ik	ik	ADP
fcis-32254	39	31	and	and	CCONJ
fcis-32254	39	32	ik	ik	PROPN
fcis-32254	39	33	are	be	AUX
fcis-32254	39	34	the	the	DET
fcis-32254	39	35	lagrange	lagrange	NOUN
fcis-32254	39	36	multipliers	multiplier	NOUN
fcis-32254	39	37	corresponding	correspond	VERB
fcis-32254	39	38	to	to	ADP
fcis-32254	39	39	constraints	constraint	NOUN
fcis-32254	39	40	0iku	0iku	PROPN
fcis-32254	39	41			NUM
fcis-32254	39	42	and	and	CCONJ
fcis-32254	39	43	0ikv	0ikv	PROPN
fcis-32254	39	44			NUM
fcis-32254	39	45	.	.	PUNCT
fcis-32254	40	1	therefore	therefore	ADV
fcis-32254	40	2	,	,	PUNCT
fcis-32254	40	3	by	by	ADP
fcis-32254	40	4	differentiating	differentiate	VERB
fcis-32254	40	5	the	the	DET
fcis-32254	40	6	lagrangian	lagrangian	ADJ
fcis-32254	40	7	function	function	NOUN
fcis-32254	40	8	and	and	CCONJ
fcis-32254	40	9	based	base	VERB
fcis-32254	40	10	on	on	ADP
fcis-32254	40	11	the	the	DET
fcis-32254	40	12	kkt	kkt	PROPN
fcis-32254	40	13	conditions	condition	NOUN
fcis-32254	40	14	(	(	PUNCT
fcis-32254	40	15	karush	karush	PROPN
fcis-32254	40	16	-	-	PUNCT
fcis-32254	40	17	kuhn	kuhn	PROPN
fcis-32254	40	18	-	-	PUNCT
fcis-32254	40	19	tucker	tucker	PROPN
fcis-32254	40	20	conditions	condition	NOUN
fcis-32254	40	21	):	):	PUNCT
fcis-32254	40	22	0ik	0ik	NOUN
fcis-32254	40	23	iku	iku	ADP
fcis-32254	40	24			PROPN
fcis-32254	40	25	,	,	PUNCT
fcis-32254	40	26	0ik	0ik	NOUN
fcis-32254	40	27	ikv	ikv	PUNCT
fcis-32254	41	1			NUM
fcis-32254	41	2	,	,	PUNCT
fcis-32254	41	3	u	u	PROPN
fcis-32254	41	4			ADJ
fcis-32254	41	5			ADJ
fcis-32254	41	6	=	=	SYM
fcis-32254	41	7	0	0	NUM
fcis-32254	41	8	l	l	NOUN
fcis-32254	41	9	,	,	PUNCT
fcis-32254	41	10	0	0	PROPN
fcis-32254	41	11	v	v	ADP
fcis-32254	41	12			ADJ
fcis-32254	41	13			NUM
fcis-32254	41	14			ADJ
fcis-32254	41	15	l	l	NOUN
fcis-32254	41	16	,	,	PUNCT
fcis-32254	41	17	and	and	CCONJ
fcis-32254	41	18	0	0	NUM
fcis-32254	41	19	a	a	DET
fcis-32254	41	20			ADJ
fcis-32254	41	21			PROPN
fcis-32254	41	22			PROPN
fcis-32254	41	23	l	l	NOUN
fcis-32254	41	24	,	,	PUNCT
fcis-32254	41	25	the	the	DET
fcis-32254	41	26	update	update	NOUN
fcis-32254	41	27	rules	rule	NOUN
fcis-32254	41	28	can	can	AUX
fcis-32254	41	29	be	be	AUX
fcis-32254	41	30	derived	derive	VERB
fcis-32254	41	31	as	as	SCONJ
fcis-32254	41	32	follows	follow	VERB
fcis-32254	41	33	:	:	PUNCT
fcis-32254	41	34	'	'	PUNCT
fcis-32254	41	35	0	0	NUM
fcis-32254	41	36	'	'	PUNCT
fcis-32254	41	37	(	(	PUNCT
fcis-32254	41	38	)	)	PUNCT
fcis-32254	41	39	(	(	PUNCT
fcis-32254	41	40	)	)	PUNCT
fcis-32254	41	41	xu	xu	PROPN
fcis-32254	42	1	ik	ik	PROPN
fcis-32254	42	2	ik	ik	PROPN
fcis-32254	42	3	ik	ik	PROPN
fcis-32254	42	4	uu	uu	PROPN
fcis-32254	42	5	ik	ik	PROPN
fcis-32254	42	6	k	k	PROPN
fcis-32254	42	7	v	v	NUM
fcis-32254	42	8	u	u	PROPN
fcis-32254	42	9	u	u	X
fcis-32254	42	10	u	u	NOUN
fcis-32254	42	11	ubk	ubk	NOUN
fcis-32254	42	12	u	u	NOUN
fcis-32254	42	13			NUM
fcis-32254	42	14			NUM
fcis-32254	42	15			ADV
fcis-32254	42	16			X
fcis-32254	42	17			PROPN
fcis-32254	42	18	(	(	PUNCT
fcis-32254	42	19	4	4	NUM
fcis-32254	42	20	)	)	PUNCT
fcis-32254	42	21	(	(	PUNCT
fcis-32254	42	22	(	(	PUNCT
fcis-32254	42	23	)	)	PUNCT
fcis-32254	42	24	(	(	PUNCT
fcis-32254	42	25	)	)	PUNCT
fcis-32254	42	26	)	)	PUNCT
fcis-32254	43	1	(	(	PUNCT
fcis-32254	43	2	(	(	PUNCT
fcis-32254	43	3	)	)	PUNCT
fcis-32254	43	4	(	(	PUNCT
fcis-32254	43	5	)	)	PUNCT
fcis-32254	43	6	)	)	PUNCT
fcis-32254	43	7	t	t	PROPN
fcis-32254	43	8	t	t	PROPN
fcis-32254	44	1	xu	xu	PROPN
fcis-32254	44	2	l	l	PROPN
fcis-32254	44	3	ik	ik	PROPN
fcis-32254	44	4	ik	ik	PROPN
fcis-32254	44	5	ik	ik	PROPN
fcis-32254	44	6	t	t	PROPN
fcis-32254	44	7	t	t	PROPN
fcis-32254	44	8	uu	uu	PROPN
fcis-32254	44	9	l	l	PROPN
fcis-32254	44	10	ik	ik	PROPN
fcis-32254	44	11	k	k	X
fcis-32254	44	12	q	q	PROPN
fcis-32254	44	13	a	a	DET
fcis-32254	44	14	v	v	NOUN
fcis-32254	44	15	a	a	DET
fcis-32254	44	16	a	a	DET
fcis-32254	44	17	wv	wv	PROPN
fcis-32254	44	18	v	v	NUM
fcis-32254	44	19	v	v	NOUN
fcis-32254	44	20	vk	vk	ADP
fcis-32254	44	21	q	q	PROPN
fcis-32254	45	1	a	a	DET
fcis-32254	45	2	v	v	NOUN
fcis-32254	45	3	a	a	DET
fcis-32254	45	4	a	a	DET
fcis-32254	45	5	dv	dv	PROPN
fcis-32254	45	6			NOUN
fcis-32254	45	7			X
fcis-32254	45	8			PROPN
fcis-32254	45	9			NOUN
fcis-32254	45	10			X
fcis-32254	45	11			PROPN
fcis-32254	45	12			ADJ
fcis-32254	45	13			ADJ
fcis-32254	45	14			PROPN
fcis-32254	45	15			PROPN
fcis-32254	45	16			PUNCT
fcis-32254	45	17			CCONJ
fcis-32254	45	18			PUNCT
fcis-32254	45	19			PROPN
fcis-32254	45	20			PROPN
fcis-32254	45	21			PROPN
fcis-32254	45	22			PROPN
fcis-32254	45	23			PUNCT
fcis-32254	45	24			X
fcis-32254	45	25	(	(	PUNCT
fcis-32254	45	26	5	5	NUM
fcis-32254	45	27	)	)	PUNCT
fcis-32254	45	28	1	1	NUM
fcis-32254	45	29	(	(	PUNCT
fcis-32254	45	30	)	)	PUNCT
fcis-32254	45	31	t	t	NOUN
fcis-32254	45	32	l	l	NOUN
fcis-32254	45	33	l	l	X
fcis-32254	45	34	la	la	X
fcis-32254	45	35	qv	qv	X
fcis-32254	45	36	v	v	ADP
fcis-32254	45	37	v	v	NOUN
fcis-32254	45	38	regai	regai	NOUN
fcis-32254	46	1			PROPN
fcis-32254	46	2			X
fcis-32254	46	3	(	(	PUNCT
fcis-32254	46	4	6	6	NUM
fcis-32254	46	5	)	)	PUNCT
fcis-32254	46	6	here	here	ADV
fcis-32254	46	7	,	,	PUNCT
fcis-32254	46	8	a	a	DET
fcis-32254	46	9	regularization	regularization	NOUN
fcis-32254	46	10	parameter	parameter	NOUN
fcis-32254	46	11	rega	rega	PROPN
fcis-32254	46	12	is	be	AUX
fcis-32254	46	13	incorporated	incorporate	VERB
fcis-32254	46	14	into	into	ADP
fcis-32254	46	15	the	the	DET
fcis-32254	46	16	iterative	iterative	NOUN
fcis-32254	46	17	formula	formula	NOUN
fcis-32254	46	18	of	of	ADP
fcis-32254	46	19	a	a	PRON
fcis-32254	46	20	to	to	PART
fcis-32254	46	21	avoid	avoid	VERB
fcis-32254	46	22	matrix	matrix	NOUN
fcis-32254	46	23	singularity	singularity	NOUN
fcis-32254	46	24	.	.	PUNCT
fcis-32254	47	1	regarding	regard	VERB
fcis-32254	47	2	the	the	DET
fcis-32254	47	3	update	update	NOUN
fcis-32254	47	4	rules	rule	NOUN
fcis-32254	47	5	,	,	PUNCT
fcis-32254	47	6	theorem	theorem	VERB
fcis-32254	47	7	1	1	NUM
fcis-32254	47	8	holds	hold	NOUN
fcis-32254	47	9	as	as	SCONJ
fcis-32254	47	10	follows	follow	VERB
fcis-32254	47	11	:	:	PUNCT
fcis-32254	47	12	the	the	DET
fcis-32254	47	13	optimization	optimization	NOUN
fcis-32254	47	14	problem	problem	NOUN
fcis-32254	47	15	(	(	PUNCT
fcis-32254	47	16	1	1	X
fcis-32254	47	17	)	)	PUNCT
fcis-32254	47	18	is	be	AUX
fcis-32254	47	19	non	non	ADJ
fcis-32254	47	20	-	-	ADJ
fcis-32254	47	21	increasing	increase	VERB
fcis-32254	47	22	under	under	ADP
fcis-32254	47	23	the	the	DET
fcis-32254	47	24	update	update	NOUN
fcis-32254	47	25	rules	rule	NOUN
fcis-32254	47	26	(	(	PUNCT
fcis-32254	47	27	4	4	NUM
fcis-32254	47	28	)	)	PUNCT
fcis-32254	47	29	,	,	PUNCT
fcis-32254	47	30	(	(	PUNCT
fcis-32254	47	31	5	5	NUM
fcis-32254	47	32	)	)	PUNCT
fcis-32254	47	33	,	,	PUNCT
fcis-32254	47	34	and	and	CCONJ
fcis-32254	47	35	(	(	PUNCT
fcis-32254	47	36	6	6	NUM
fcis-32254	47	37	)	)	PUNCT
fcis-32254	47	38	.	.	PUNCT
fcis-32254	48	1	the	the	DET
fcis-32254	48	2	value	value	NOUN
fcis-32254	48	3	of	of	ADP
fcis-32254	48	4	the	the	DET
fcis-32254	48	5	objective	objective	ADJ
fcis-32254	48	6	function	function	NOUN
fcis-32254	48	7	remains	remain	VERB
fcis-32254	48	8	unchanged	unchanged	ADJ
fcis-32254	48	9	if	if	SCONJ
fcis-32254	48	10	and	and	CCONJ
fcis-32254	48	11	only	only	ADV
fcis-32254	48	12	if	if	SCONJ
fcis-32254	48	13	the	the	DET
fcis-32254	48	14	matrices	matrix	NOUN
fcis-32254	48	15	u	u	NOUN
fcis-32254	48	16	,	,	PUNCT
fcis-32254	48	17	v	v	NOUN
fcis-32254	48	18	,	,	PUNCT
fcis-32254	48	19	and	and	CCONJ
fcis-32254	48	20	a	a	DET
fcis-32254	48	21	reach	reach	NOUN
fcis-32254	48	22	a	a	DET
fcis-32254	48	23	stable	stable	ADJ
fcis-32254	48	24	point	point	NOUN
fcis-32254	48	25	.	.	PUNCT
fcis-32254	49	1	2.2	2.2	NUM
fcis-32254	49	2	.	.	PUNCT
fcis-32254	50	1	convergence	convergence	NOUN
fcis-32254	50	2	analysis	analysis	NOUN
fcis-32254	50	3	the	the	DET
fcis-32254	50	4	following	follow	VERB
fcis-32254	50	5	proof	proof	NOUN
fcis-32254	50	6	refers	refer	VERB
fcis-32254	50	7	to	to	AUX
fcis-32254	50	8	reference	reference	NOUN
fcis-32254	50	9	[	[	X
fcis-32254	50	10	8].let	8].let	PROPN
fcis-32254	50	11	model	model	NOUN
fcis-32254	50	12	(	(	PUNCT
fcis-32254	50	13	1.1	1.1	NUM
fcis-32254	50	14	)	)	PUNCT
fcis-32254	50	15	be	be	AUX
fcis-32254	50	16	defined	define	VERB
fcis-32254	50	17	as	as	ADP
fcis-32254	50	18	the	the	DET
fcis-32254	50	19	following	follow	VERB
fcis-32254	50	20	objective	objective	ADJ
fcis-32254	50	21	function	function	NOUN
fcis-32254	50	22	:	:	PUNCT
fcis-32254	50	23	2	2	NUM
fcis-32254	50	24	2	2	NUM
fcis-32254	50	25	l	l	NOUN
fcis-32254	50	26	2	2	NUM
fcis-32254	50	27	0	0	NUM
fcis-32254	50	28	1	1	NUM
fcis-32254	50	29	(	(	PUNCT
fcis-32254	50	30	)	)	PUNCT
fcis-32254	50	31	(	(	PUNCT
fcis-32254	50	32	)	)	PUNCT
fcis-32254	51	1	tr	tr	VERB
fcis-32254	51	2	(	(	PUNCT
fcis-32254	51	3	)	)	PUNCT
fcis-32254	51	4	t	t	NOUN
fcis-32254	51	5	t	t	X
fcis-32254	52	1	f	f	PROPN
fcis-32254	52	2	f	f	PROPN
fcis-32254	53	1	t	t	PROPN
fcis-32254	53	2	f	f	NOUN
fcis-32254	53	3	m	m	VERB
fcis-32254	53	4	x	x	X
fcis-32254	53	5	u	u	NOUN
fcis-32254	53	6	v	v	X
fcis-32254	53	7	q	q	X
fcis-32254	53	8	av	av	PROPN
fcis-32254	53	9	v	v	NUM
fcis-32254	53	10	lv	lv	PROPN
fcis-32254	53	11	u	u	NOUN
fcis-32254	53	12	u	u	NOUN
fcis-32254	53	13	v	v	ADP
fcis-32254	53	14			NOUN
fcis-32254	53	15			NOUN
fcis-32254	53	16			X
fcis-32254	53	17			PROPN
fcis-32254	53	18			PROPN
fcis-32254	53	19			ADJ
fcis-32254	53	20			PROPN
fcis-32254	53	21			PROPN
fcis-32254	53	22			VERB
fcis-32254	53	23			PROPN
fcis-32254	53	24			PUNCT
fcis-32254	53	25			PUNCT
fcis-32254	53	26			PROPN
fcis-32254	53	27			X
fcis-32254	53	28	(	(	PUNCT
fcis-32254	53	29	7	7	X
fcis-32254	53	30	)	)	PUNCT
fcis-32254	53	31	definition	definition	NOUN
fcis-32254	53	32	1	1	NUM
fcis-32254	53	33	:	:	PUNCT
fcis-32254	53	34	function	function	NOUN
fcis-32254	53	35	(	(	PUNCT
fcis-32254	53	36	,	,	PUNCT
fcis-32254	53	37	)	)	PUNCT
fcis-32254	53	38	tg	tg	PROPN
fcis-32254	53	39	h	h	NOUN
fcis-32254	53	40	h	h	NOUN
fcis-32254	53	41	,	,	PUNCT
fcis-32254	53	42	which	which	PRON
fcis-32254	53	43	is	be	AUX
fcis-32254	53	44	the	the	DET
fcis-32254	53	45	auxiliary	auxiliary	ADJ
fcis-32254	53	46	function	function	NOUN
fcis-32254	53	47	of	of	ADP
fcis-32254	53	48	function	function	NOUN
fcis-32254	53	49	(	(	PUNCT
fcis-32254	53	50	)	)	PUNCT
fcis-32254	53	51	m	m	NOUN
fcis-32254	53	52	h	h	NOUN
fcis-32254	53	53	,	,	PUNCT
fcis-32254	53	54	satisfies	satisfy	VERB
fcis-32254	53	55	the	the	DET
fcis-32254	53	56	following	follow	VERB
fcis-32254	53	57	two	two	NUM
fcis-32254	53	58	conditions	condition	NOUN
fcis-32254	53	59	:	:	PUNCT
fcis-32254	53	60	(	(	PUNCT
fcis-32254	53	61	,	,	PUNCT
fcis-32254	53	62	)	)	PUNCT
fcis-32254	53	63	(	(	PUNCT
fcis-32254	53	64	)	)	PUNCT
fcis-32254	53	65	,	,	PUNCT
fcis-32254	53	66	(	(	PUNCT
fcis-32254	53	67	,	,	PUNCT
fcis-32254	53	68	)	)	PUNCT
fcis-32254	53	69	(	(	PUNCT
fcis-32254	53	70	)	)	PUNCT
fcis-32254	54	1	tg	tg	PROPN
fcis-32254	54	2	h	h	NOUN
fcis-32254	55	1	h	h	NOUN
fcis-32254	55	2	m	m	VERB
fcis-32254	55	3	h	h	NOUN
fcis-32254	56	1	g	g	PROPN
fcis-32254	56	2	h	h	NOUN
fcis-32254	57	1	h	h	NOUN
fcis-32254	57	2	m	m	VERB
fcis-32254	57	3	h	h	SCONJ
fcis-32254	57	4			PRON
fcis-32254	57	5	(	(	PUNCT
fcis-32254	57	6	8)	8)	NUM
fcis-32254	57	7	lemma	lemma	PROPN
fcis-32254	57	8	1	1	NUM
fcis-32254	57	9	:	:	PUNCT
fcis-32254	57	10	if	if	SCONJ
fcis-32254	57	11	function	function	NOUN
fcis-32254	57	12	(	(	PUNCT
fcis-32254	57	13	,	,	PUNCT
fcis-32254	57	14	)	)	PUNCT
fcis-32254	57	15	tg	tg	PROPN
fcis-32254	57	16	h	h	NOUN
fcis-32254	57	17	h	h	NOUN
fcis-32254	57	18	is	be	AUX
fcis-32254	57	19	the	the	DET
fcis-32254	57	20	auxiliary	auxiliary	ADJ
fcis-32254	57	21	function	function	NOUN
fcis-32254	57	22	of	of	ADP
fcis-32254	57	23	function	function	NOUN
fcis-32254	57	24	(	(	PUNCT
fcis-32254	57	25	)	)	PUNCT
fcis-32254	57	26	m	m	VERB
fcis-32254	57	27	h	h	NOUN
fcis-32254	57	28	,	,	PUNCT
fcis-32254	57	29	then	then	ADV
fcis-32254	57	30	(	(	PUNCT
fcis-32254	57	31	)	)	PUNCT
fcis-32254	57	32	m	m	NOUN
fcis-32254	57	33	h	h	NOUN
fcis-32254	57	34	is	be	AUX
fcis-32254	57	35	non	non	ADJ
fcis-32254	57	36	-	-	ADJ
fcis-32254	57	37	increasing	increase	VERB
fcis-32254	57	38	under	under	ADP
fcis-32254	57	39	the	the	DET
fcis-32254	57	40	iterative	iterative	NOUN
fcis-32254	57	41	rule	rule	NOUN
fcis-32254	57	42	:	:	PUNCT
fcis-32254	57	43	1	1	NUM
fcis-32254	57	44	arg	arg	NOUN
fcis-32254	57	45	min	min	NOUN
fcis-32254	57	46	(	(	PUNCT
fcis-32254	57	47	,	,	PUNCT
fcis-32254	57	48	)	)	PUNCT
fcis-32254	57	49	t	t	NOUN
fcis-32254	57	50	t	t	PROPN
fcis-32254	58	1	h	h	NOUN
fcis-32254	58	2	m	m	VERB
fcis-32254	58	3	g	g	PROPN
fcis-32254	58	4	h	h	PROPN
fcis-32254	58	5	h	h	PROPN
fcis-32254	59	1			NOUN
fcis-32254	59	2	(	(	PUNCT
fcis-32254	59	3	9	9	NUM
fcis-32254	59	4	)	)	PUNCT
fcis-32254	59	5	it	it	PRON
fcis-32254	59	6	can	can	AUX
fcis-32254	59	7	be	be	AUX
fcis-32254	59	8	proven	prove	VERB
fcis-32254	59	9	as	as	SCONJ
fcis-32254	59	10	follows	follow	VERB
fcis-32254	59	11	:	:	PUNCT
fcis-32254	59	12	1	1	NUM
fcis-32254	59	13	1	1	NUM
fcis-32254	59	14	(	(	PUNCT
fcis-32254	59	15	)	)	PUNCT
fcis-32254	59	16	(	(	PUNCT
fcis-32254	59	17	,	,	PUNCT
fcis-32254	59	18	)	)	PUNCT
fcis-32254	59	19	(	(	PUNCT
fcis-32254	59	20	,	,	PUNCT
fcis-32254	59	21	)	)	PUNCT
fcis-32254	59	22	(	(	PUNCT
fcis-32254	59	23	)	)	PUNCT
fcis-32254	59	24	t	t	NOUN
fcis-32254	59	25	t	t	PROPN
fcis-32254	59	26	t	t	PROPN
fcis-32254	59	27	t	t	PROPN
fcis-32254	59	28	t	t	PROPN
fcis-32254	60	1	tm	tm	PROPN
fcis-32254	60	2	h	h	PROPN
fcis-32254	61	1	g	g	PROPN
fcis-32254	62	1	h	h	NOUN
fcis-32254	63	1	h	h	NOUN
fcis-32254	64	1	g	g	PROPN
fcis-32254	65	1	h	h	PROPN
fcis-32254	66	1	h	h	PROPN
fcis-32254	66	2	m	m	PROPN
fcis-32254	66	3	h	h	PROPN
fcis-32254	67	1			PROPN
fcis-32254	67	2			NOUN
fcis-32254	67	3			NUM
fcis-32254	67	4	(	(	PUNCT
fcis-32254	67	5	10	10	NUM
fcis-32254	67	6	)	)	PUNCT
fcis-32254	67	7	moreover	moreover	ADV
fcis-32254	67	8	,	,	PUNCT
fcis-32254	67	9	1	1	NUM
fcis-32254	67	10	(	(	PUNCT
fcis-32254	67	11	)	)	PUNCT
fcis-32254	67	12	(	(	PUNCT
fcis-32254	67	13	)	)	PUNCT
fcis-32254	67	14	t	t	PROPN
fcis-32254	67	15	tm	tm	PROPN
fcis-32254	67	16	h	h	PROPN
fcis-32254	67	17	m	m	PROPN
fcis-32254	67	18	h	h	PROPN
fcis-32254	68	1			PRON
fcis-32254	68	2	holds	hold	VERB
fcis-32254	68	3	if	if	SCONJ
fcis-32254	68	4	and	and	CCONJ
fcis-32254	68	5	only	only	ADV
fcis-32254	68	6	if	if	SCONJ
fcis-32254	68	7	tm	tm	PROPN
fcis-32254	68	8	is	be	AUX
fcis-32254	68	9	a	a	DET
fcis-32254	68	10	local	local	ADJ
fcis-32254	68	11	minimum	minimum	NOUN
fcis-32254	68	12	point	point	NOUN
fcis-32254	68	13	of	of	ADP
fcis-32254	68	14	(	(	PUNCT
fcis-32254	68	15	,	,	PUNCT
fcis-32254	68	16	)	)	PUNCT
fcis-32254	68	17	tg	tg	PROPN
fcis-32254	68	18	h	h	PROPN
fcis-32254	68	19	h	h	NOUN
fcis-32254	68	20	.	.	PUNCT
fcis-32254	69	1	if	if	SCONJ
fcis-32254	69	2	the	the	DET
fcis-32254	69	3	derivative	derivative	NOUN
fcis-32254	69	4	of	of	ADP
fcis-32254	69	5	m	m	PROPN
fcis-32254	69	6	exists	exist	VERB
fcis-32254	69	7	and	and	CCONJ
fcis-32254	69	8	is	be	AUX
fcis-32254	69	9	continuous	continuous	ADJ
fcis-32254	69	10	within	within	ADP
fcis-32254	69	11	a	a	DET
fcis-32254	69	12	small	small	ADJ
fcis-32254	69	13	neighborhood	neighborhood	NOUN
fcis-32254	69	14	of	of	ADP
fcis-32254	69	15	'	'	PUNCT
fcis-32254	69	16	m	m	VERB
fcis-32254	69	17	,	,	PUNCT
fcis-32254	69	18	then	then	ADV
fcis-32254	69	19	the	the	DET
fcis-32254	69	20	differential	differential	NOUN
fcis-32254	69	21	'	'	PUNCT
fcis-32254	69	22	(	(	PUNCT
fcis-32254	69	23	)	)	PUNCT
fcis-32254	69	24	0	0	NUM
fcis-32254	69	25	m	m	NOUN
fcis-32254	69	26	h	h	NOUN
fcis-32254	69	27			PRON
fcis-32254	69	28	exists	exist	VERB
fcis-32254	69	29	.	.	PUNCT
fcis-32254	70	1	through	through	ADP
fcis-32254	70	2	equation	equation	NOUN
fcis-32254	70	3	(	(	PUNCT
fcis-32254	70	4	9	9	NUM
fcis-32254	70	5	)	)	PUNCT
fcis-32254	70	6	,	,	PUNCT
fcis-32254	70	7	a	a	DET
fcis-32254	70	8	series	series	NOUN
fcis-32254	70	9	of	of	ADP
fcis-32254	70	10	sequences	sequence	NOUN
fcis-32254	70	11	that	that	PRON
fcis-32254	70	12	converge	converge	VERB
fcis-32254	70	13	to	to	ADP
fcis-32254	70	14	the	the	DET
fcis-32254	70	15	local	local	ADJ
fcis-32254	70	16	minimum	minimum	NOUN
fcis-32254	70	17	point	point	NOUN
fcis-32254	70	18	min	min	PROPN
fcis-32254	70	19	arg	arg	NOUN
fcis-32254	70	20	min	min	PROPN
fcis-32254	70	21	(	(	PUNCT
fcis-32254	70	22	)	)	PUNCT
fcis-32254	71	1	h	h	NOUN
fcis-32254	71	2	m	m	VERB
fcis-32254	71	3	m	m	VERB
fcis-32254	71	4	h	h	NOUN
fcis-32254	71	5	can	can	AUX
fcis-32254	71	6	be	be	AUX
fcis-32254	71	7	obtained	obtain	VERB
fcis-32254	71	8	:	:	PUNCT
fcis-32254	71	9	1	1	NUM
fcis-32254	71	10	min	min	NOUN
fcis-32254	71	11	2	2	NUM
fcis-32254	71	12	1	1	NUM
fcis-32254	71	13	0	0	NUM
fcis-32254	71	14	(	(	PUNCT
fcis-32254	71	15	)	)	PUNCT
fcis-32254	71	16	(	(	PUNCT
fcis-32254	71	17	)	)	PUNCT
fcis-32254	71	18	(	(	PUNCT
fcis-32254	71	19	)	)	PUNCT
fcis-32254	71	20	(	(	PUNCT
fcis-32254	71	21	)	)	PUNCT
fcis-32254	71	22	(	(	PUNCT
fcis-32254	71	23	)	)	PUNCT
fcis-32254	71	24	(	(	PUNCT
fcis-32254	71	25	)	)	PUNCT
fcis-32254	71	26	t	t	PROPN
fcis-32254	71	27	tm	tm	PROPN
fcis-32254	71	28	h	h	PROPN
fcis-32254	71	29	m	m	PROPN
fcis-32254	71	30	h	h	NOUN
fcis-32254	71	31	m	m	VERB
fcis-32254	71	32	h	h	NOUN
fcis-32254	71	33	m	m	VERB
fcis-32254	71	34	h	h	NOUN
fcis-32254	71	35	m	m	VERB
fcis-32254	71	36	h	h	NOUN
fcis-32254	71	37	m	m	VERB
fcis-32254	71	38	h	h	NOUN
fcis-32254	71	39			PROPN
fcis-32254	71	40			NOUN
fcis-32254	71	41			NOUN
fcis-32254	71	42			NOUN
fcis-32254	71	43			NOUN
fcis-32254	71	44			NOUN
fcis-32254	71	45			NOUN
fcis-32254	71	46			NOUN
fcis-32254	71	47			NOUN
fcis-32254	71	48	(	(	PUNCT
fcis-32254	71	49	11	11	NUM
fcis-32254	71	50	)	)	PUNCT
fcis-32254	71	51	for	for	ADP
fcis-32254	71	52	an	an	DET
fcis-32254	71	53	element	element	NOUN
fcis-32254	71	54	ijv	ijv	X
fcis-32254	71	55	in	in	ADP
fcis-32254	71	56	matrix	matrix	NOUN
fcis-32254	71	57	v	v	NOUN
fcis-32254	71	58	(	(	PUNCT
fcis-32254	71	59	from	from	ADP
fcis-32254	71	60	the	the	DET
fcis-32254	71	61	objective	objective	ADJ
fcis-32254	71	62	function	function	NOUN
fcis-32254	71	63	)	)	PUNCT
fcis-32254	71	64	,	,	PUNCT
fcis-32254	71	65	the	the	DET
fcis-32254	71	66	formulas	formula	NOUN
fcis-32254	71	67	for	for	ADP
fcis-32254	71	68	the	the	DET
fcis-32254	71	69	first	first	ADJ
fcis-32254	71	70	-	-	PUNCT
fcis-32254	71	71	order	order	NOUN
fcis-32254	71	72	partial	partial	ADJ
fcis-32254	71	73	derivative	derivative	NOUN
fcis-32254	71	74	'	'	PUNCT
fcis-32254	71	75	ijv	ijv	PROPN
fcis-32254	71	76	m	m	PROPN
fcis-32254	71	77	and	and	CCONJ
fcis-32254	71	78	the	the	DET
fcis-32254	71	79	second	second	ADJ
fcis-32254	71	80	-	-	PUNCT
fcis-32254	71	81	order	order	NOUN
fcis-32254	71	82	partial	partial	ADJ
fcis-32254	71	83	derivative	derivative	NOUN
fcis-32254	71	84	''	''	PUNCT
fcis-32254	71	85	ijv	ijv	PROPN
fcis-32254	71	86	m	m	PROPN
fcis-32254	71	87	of	of	ADP
fcis-32254	71	88	the	the	DET
fcis-32254	71	89	objective	objective	ADJ
fcis-32254	71	90	function	function	NOUN
fcis-32254	71	91	with	with	ADP
fcis-32254	71	92	respect	respect	NOUN
fcis-32254	71	93	to	to	ADP
fcis-32254	71	94	ijv	ijv	PROPN
fcis-32254	71	95	can	can	AUX
fcis-32254	71	96	be	be	AUX
fcis-32254	71	97	obtained	obtain	VERB
fcis-32254	71	98	as	as	SCONJ
fcis-32254	71	99	follows	follow	VERB
fcis-32254	71	100	:	:	PUNCT
fcis-32254	71	101	'	'	PUNCT
fcis-32254	71	102	(	(	PUNCT
fcis-32254	71	103	2	2	NUM
fcis-32254	71	104	2	2	NUM
fcis-32254	71	105	2	2	NUM
fcis-32254	71	106	2	2	NUM
fcis-32254	71	107	2	2	NUM
fcis-32254	71	108	)	)	PUNCT
fcis-32254	72	1	ij	ij	NOUN
fcis-32254	72	2	t	t	PROPN
fcis-32254	72	3	v	v	INTJ
fcis-32254	73	1	xu	xu	INTJ
fcis-32254	73	2	uu	uu	INTJ
fcis-32254	74	1	t	t	PROPN
fcis-32254	74	2	l	l	NOUN
fcis-32254	75	1	ij	ij	INTJ
fcis-32254	75	2	m	m	PROPN
fcis-32254	75	3	k	k	INTJ
fcis-32254	75	4	vk	vk	ADP
fcis-32254	75	5	q	q	PROPN
fcis-32254	75	6	a	a	DET
fcis-32254	75	7	v	v	NOUN
fcis-32254	75	8	a	a	DET
fcis-32254	75	9	a	a	DET
fcis-32254	75	10	lv	lv	PROPN
fcis-32254	75	11			NOUN
fcis-32254	75	12			X
fcis-32254	75	13			PROPN
fcis-32254	75	14			ADJ
fcis-32254	75	15			PROPN
fcis-32254	75	16			PROPN
fcis-32254	75	17			VERB
fcis-32254	75	18			PROPN
fcis-32254	75	19			ADV
fcis-32254	75	20			CCONJ
fcis-32254	75	21			X
fcis-32254	75	22	(	(	PUNCT
fcis-32254	75	23	12	12	NUM
fcis-32254	75	24	)	)	PUNCT
fcis-32254	75	25	''	''	PUNCT
fcis-32254	76	1	2	2	NUM
fcis-32254	76	2	(	(	PUNCT
fcis-32254	76	3	)	)	PUNCT
fcis-32254	76	4	2	2	NUM
fcis-32254	76	5	(	(	PUNCT
fcis-32254	76	6	)	)	PUNCT
fcis-32254	76	7	2	2	NUM
fcis-32254	76	8	,	,	PUNCT
fcis-32254	76	9	2	2	NUM
fcis-32254	76	10	(	(	PUNCT
fcis-32254	76	11	)	)	PUNCT
fcis-32254	76	12	2	2	NUM
fcis-32254	76	13	,	,	PUNCT
fcis-32254	76	14	ij	ij	INTJ
fcis-32254	76	15	t	t	PROPN
fcis-32254	76	16	uu	uu	INTJ
fcis-32254	76	17	jj	jj	PROPN
fcis-32254	76	18	jj	jj	PROPN
fcis-32254	76	19	ii	ii	PROPN
fcis-32254	76	20	i	i	PRON
fcis-32254	76	21	v	v	VERB
fcis-32254	76	22	uu	uu	INTJ
fcis-32254	76	23	j	j	PROPN
fcis-32254	76	24	j	j	PROPN
fcis-32254	76	25	ii	ii	PROPN
fcis-32254	77	1	k	k	PROPN
fcis-32254	77	2	a	a	DET
fcis-32254	77	3	a	a	PROPN
fcis-32254	77	4	l	l	NOUN
fcis-32254	77	5	if	if	SCONJ
fcis-32254	77	6	x	x	PRON
fcis-32254	77	7	m	m	NOUN
fcis-32254	77	8	has	have	VERB
fcis-32254	77	9	labels	label	NOUN
fcis-32254	77	10	othel	othel	PROPN
fcis-32254	77	11	rk	rk	PROPN
fcis-32254	77	12	s	s	PART
fcis-32254	77	13			NUM
fcis-32254	77	14			PROPN
fcis-32254	77	15			PROPN
fcis-32254	77	16			ADV
fcis-32254	77	17			PROPN
fcis-32254	77	18			PROPN
fcis-32254	77	19			ADV
fcis-32254	77	20			ADJ
fcis-32254	77	21			NUM
fcis-32254	77	22			NOUN
fcis-32254	77	23	(	(	PUNCT
fcis-32254	77	24	13	13	NUM
fcis-32254	77	25	)	)	PUNCT
fcis-32254	77	26	lemma	lemma	PROPN
fcis-32254	77	27	2	2	NUM
fcis-32254	77	28	:	:	PUNCT
fcis-32254	77	29	function	function	NOUN
fcis-32254	77	30	'	'	PUNCT
fcis-32254	77	31	2	2	NUM
fcis-32254	77	32	(	(	PUNCT
fcis-32254	77	33	,	,	PUNCT
fcis-32254	77	34	)	)	PUNCT
fcis-32254	77	35	(	(	PUNCT
fcis-32254	77	36	)	)	PUNCT
fcis-32254	77	37	(	(	PUNCT
fcis-32254	77	38	)	)	PUNCT
fcis-32254	77	39	(	(	PUNCT
fcis-32254	77	40	)	)	PUNCT
fcis-32254	77	41	(	(	PUNCT
fcis-32254	77	42	(	(	PUNCT
fcis-32254	77	43	)	)	PUNCT
fcis-32254	77	44	(	(	PUNCT
fcis-32254	77	45	)	)	PUNCT
fcis-32254	77	46	)	)	PUNCT
fcis-32254	78	1	(	(	PUNCT
fcis-32254	78	2	)	)	PUNCT
fcis-32254	78	3	ij	ij	INTJ
fcis-32254	79	1	ij	ij	INTJ
fcis-32254	80	1	t	t	X
fcis-32254	80	2	t	t	PROPN
fcis-32254	80	3	t	t	PROPN
fcis-32254	80	4	t	t	PROPN
fcis-32254	80	5	ij	ij	X
fcis-32254	80	6	v	v	NUM
fcis-32254	80	7	ij	ij	NUM
fcis-32254	80	8	v	v	NUM
fcis-32254	80	9	ij	ij	INTJ
fcis-32254	80	10	ij	ij	INTJ
fcis-32254	80	11	t	t	PROPN
fcis-32254	80	12	t	t	PROPN
fcis-32254	80	13	uu	uu	INTJ
fcis-32254	81	1	l	l	NOUN
fcis-32254	81	2	ij	ij	NOUN
fcis-32254	81	3	t	t	PROPN
fcis-32254	81	4	ijt	ijt	VERB
fcis-32254	81	5	ij	ij	INTJ
fcis-32254	81	6	g	g	PROPN
fcis-32254	81	7	v	v	NUM
fcis-32254	81	8	v	v	ADP
fcis-32254	81	9	m	m	NOUN
fcis-32254	81	10	v	v	NOUN
fcis-32254	81	11	m	m	NOUN
fcis-32254	81	12	v	v	ADP
fcis-32254	81	13	v	v	NUM
fcis-32254	81	14	v	v	NOUN
fcis-32254	81	15	vk	vk	ADP
fcis-32254	81	16	q	q	PROPN
fcis-32254	82	1	a	a	DET
fcis-32254	82	2	v	v	NOUN
fcis-32254	82	3	a	a	PRON
fcis-32254	82	4	a	a	DET
fcis-32254	82	5	dv	dv	PROPN
fcis-32254	82	6	v	v	NUM
fcis-32254	82	7	v	v	NUM
fcis-32254	82	8	v	v	NOUN
fcis-32254	82	9			NOUN
fcis-32254	82	10			NOUN
fcis-32254	82	11			NOUN
fcis-32254	82	12			NOUN
fcis-32254	82	13			PUNCT
fcis-32254	82	14			NUM
fcis-32254	82	15			ADJ
fcis-32254	82	16			PROPN
fcis-32254	82	17			VERB
fcis-32254	82	18			PROPN
fcis-32254	82	19			PROPN
fcis-32254	82	20			PUNCT
fcis-32254	82	21			PUNCT
fcis-32254	82	22			X
fcis-32254	82	23	(	(	PUNCT
fcis-32254	82	24	14	14	NUM
fcis-32254	82	25	)	)	PUNCT
fcis-32254	82	26	is	be	AUX
fcis-32254	82	27	the	the	DET
fcis-32254	82	28	auxiliary	auxiliary	ADJ
fcis-32254	82	29	function	function	NOUN
fcis-32254	82	30	of	of	ADP
fcis-32254	82	31	ijv	ijv	PROPN
fcis-32254	82	32	m	m	VERB
fcis-32254	82	33	.	.	PUNCT
fcis-32254	83	1	proof	proof	NOUN
fcis-32254	83	2	:	:	PUNCT
fcis-32254	83	3	as	as	SCONJ
fcis-32254	83	4	can	can	AUX
fcis-32254	83	5	be	be	AUX
fcis-32254	83	6	seen	see	VERB
fcis-32254	83	7	from	from	ADP
fcis-32254	83	8	definition	definition	NOUN
fcis-32254	83	9	1	1	NUM
fcis-32254	83	10	,	,	PUNCT
fcis-32254	83	11	to	to	PART
fcis-32254	83	12	prove	prove	VERB
fcis-32254	83	13	a	a	DET
fcis-32254	83	14	function	function	NOUN
fcis-32254	83	15	is	be	AUX
fcis-32254	83	16	the	the	DET
fcis-32254	83	17	auxiliary	auxiliary	ADJ
fcis-32254	83	18	function	function	NOUN
fcis-32254	83	19	(	(	PUNCT
fcis-32254	83	20	of	of	ADP
fcis-32254	83	21	another	another	DET
fcis-32254	83	22	function	function	NOUN
fcis-32254	83	23	)	)	PUNCT
fcis-32254	83	24	,	,	PUNCT
fcis-32254	83	25	it	it	PRON
fcis-32254	83	26	must	must	AUX
fcis-32254	83	27	satisfy	satisfy	VERB
fcis-32254	83	28	both	both	PRON
fcis-32254	83	29	(	(	PUNCT
fcis-32254	83	30	,	,	PUNCT
fcis-32254	83	31	)	)	PUNCT
fcis-32254	83	32	(	(	PUNCT
fcis-32254	83	33	)	)	PUNCT
fcis-32254	83	34	t	t	NOUN
fcis-32254	83	35	ij	ij	NUM
fcis-32254	83	36	ijg	ijg	NOUN
fcis-32254	83	37	v	v	PROPN
fcis-32254	83	38	v	v	PROPN
fcis-32254	83	39	m	m	NOUN
fcis-32254	83	40	v	v	NOUN
fcis-32254	83	41	and	and	CCONJ
fcis-32254	83	42	(	(	PUNCT
fcis-32254	83	43	,	,	PUNCT
fcis-32254	83	44	)	)	PUNCT
fcis-32254	83	45	(	(	PUNCT
fcis-32254	83	46	)	)	PUNCT
fcis-32254	83	47	ijg	ijg	NOUN
fcis-32254	83	48	v	v	NOUN
fcis-32254	83	49	v	v	ADP
fcis-32254	83	50	m	m	VERB
fcis-32254	83	51	v	v	ADV
fcis-32254	83	52	simultaneously	simultaneously	ADV
fcis-32254	83	53	.	.	PUNCT
fcis-32254	84	1	obviously	obviously	ADV
fcis-32254	84	2	,	,	PUNCT
fcis-32254	84	3	(	(	PUNCT
fcis-32254	84	4	,	,	PUNCT
fcis-32254	84	5	)	)	PUNCT
fcis-32254	84	6	(	(	PUNCT
fcis-32254	84	7	)	)	PUNCT
fcis-32254	84	8	ijg	ijg	NOUN
fcis-32254	84	9	v	v	NOUN
fcis-32254	84	10	v	v	ADP
fcis-32254	84	11	m	m	VERB
fcis-32254	84	12	v	v	ADV
fcis-32254	84	13	can	can	AUX
fcis-32254	84	14	be	be	AUX
fcis-32254	84	15	obtained	obtain	VERB
fcis-32254	84	16	directly	directly	ADV
fcis-32254	84	17	.	.	PUNCT
fcis-32254	85	1	next	next	ADV
fcis-32254	85	2	,	,	PUNCT
fcis-32254	85	3	we	we	PRON
fcis-32254	85	4	conduct	conduct	VERB
fcis-32254	85	5	a	a	DET
fcis-32254	85	6	detailed	detailed	ADJ
fcis-32254	85	7	proof	proof	NOUN
fcis-32254	85	8	for	for	ADP
fcis-32254	85	9	(	(	PUNCT
fcis-32254	85	10	,	,	PUNCT
fcis-32254	85	11	)	)	PUNCT
fcis-32254	85	12	(	(	PUNCT
fcis-32254	85	13	)	)	PUNCT
fcis-32254	85	14	t	t	NOUN
fcis-32254	85	15	ij	ij	NUM
fcis-32254	85	16	ijg	ijg	NOUN
fcis-32254	85	17	v	v	PROPN
fcis-32254	85	18	v	v	PROPN
fcis-32254	85	19	m	m	NOUN
fcis-32254	85	20	v	v	NOUN
fcis-32254	85	21	.	.	PUNCT
fcis-32254	86	1	first	first	ADV
fcis-32254	86	2	,	,	PUNCT
fcis-32254	86	3	perform	perform	VERB
fcis-32254	86	4	a	a	DET
fcis-32254	86	5	taylor	taylor	NOUN
fcis-32254	86	6	expansion	expansion	NOUN
fcis-32254	86	7	on	on	ADP
fcis-32254	86	8	(	(	PUNCT
fcis-32254	86	9	)	)	PUNCT
fcis-32254	86	10	ijm	ijm	PROPN
fcis-32254	86	11	v	v	NOUN
fcis-32254	86	12	;	;	PUNCT
fcis-32254	86	13	by	by	ADP
fcis-32254	86	14	comparing	compare	VERB
fcis-32254	86	15	it	it	PRON
fcis-32254	86	16	with	with	ADP
fcis-32254	86	17	equation	equation	NOUN
fcis-32254	86	18	(	(	PUNCT
fcis-32254	86	19	14	14	NUM
fcis-32254	86	20	)	)	PUNCT
fcis-32254	86	21	,	,	PUNCT
fcis-32254	86	22	to	to	PART
fcis-32254	86	23	prove	prove	VERB
fcis-32254	86	24	(	(	PUNCT
fcis-32254	86	25	,	,	PUNCT
fcis-32254	86	26	)	)	PUNCT
fcis-32254	86	27	(	(	PUNCT
fcis-32254	86	28	)	)	PUNCT
fcis-32254	86	29	t	t	NOUN
fcis-32254	86	30	ij	ij	NUM
fcis-32254	86	31	ijg	ijg	NOUN
fcis-32254	86	32	v	v	PROPN
fcis-32254	86	33	v	v	PROPN
fcis-32254	86	34	m	m	NOUN
fcis-32254	86	35	v	v	NOUN
fcis-32254	86	36	,	,	PUNCT
fcis-32254	86	37	it	it	PRON
fcis-32254	86	38	is	be	AUX
fcis-32254	86	39	equivalent	equivalent	ADJ
fcis-32254	86	40	to	to	ADP
fcis-32254	86	41	proving	prove	VERB
fcis-32254	86	42	that	that	PRON
fcis-32254	86	43	:	:	PUNCT
fcis-32254	86	44	''	''	PUNCT
fcis-32254	86	45	(	(	PUNCT
fcis-32254	86	46	(	(	PUNCT
fcis-32254	86	47	)	)	PUNCT
fcis-32254	86	48	(	(	PUNCT
fcis-32254	86	49	)	)	PUNCT
fcis-32254	86	50	)	)	PUNCT
fcis-32254	86	51	1	1	NUM
fcis-32254	87	1	(	(	PUNCT
fcis-32254	87	2	)	)	PUNCT
fcis-32254	87	3	2	2	NUM
fcis-32254	87	4	ij	ij	NOUN
fcis-32254	87	5	t	t	PROPN
fcis-32254	87	6	t	t	PROPN
fcis-32254	87	7	uu	uu	INTJ
fcis-32254	88	1	l	l	NOUN
fcis-32254	88	2	ij	ij	NOUN
fcis-32254	88	3	t	t	PROPN
fcis-32254	88	4	v	v	NOUN
fcis-32254	88	5	ijt	ijt	VERB
fcis-32254	88	6	ij	ij	INTJ
fcis-32254	88	7	vk	vk	ADP
fcis-32254	88	8	q	q	PROPN
fcis-32254	89	1	a	a	DET
fcis-32254	89	2	v	v	NOUN
fcis-32254	89	3	a	a	PRON
fcis-32254	89	4	a	a	DET
fcis-32254	89	5	dv	dv	PROPN
fcis-32254	89	6	m	m	PROPN
fcis-32254	89	7	v	v	ADP
fcis-32254	89	8	v	v	NUM
fcis-32254	89	9			NOUN
fcis-32254	89	10			NOUN
fcis-32254	89	11			NOUN
fcis-32254	89	12			NOUN
fcis-32254	89	13			PROPN
fcis-32254	89	14			NOUN
fcis-32254	89	15			PUNCT
fcis-32254	89	16			X
fcis-32254	89	17			NUM
fcis-32254	89	18	(	(	PUNCT
fcis-32254	89	19	15	15	NUM
fcis-32254	89	20	)	)	PUNCT
fcis-32254	89	21	substitute	substitute	NOUN
fcis-32254	89	22	equation	equation	NOUN
fcis-32254	89	23	(	(	PUNCT
fcis-32254	89	24	13	13	NUM
fcis-32254	89	25	)	)	PUNCT
fcis-32254	89	26	into	into	ADP
fcis-32254	89	27	equation	equation	NOUN
fcis-32254	89	28	(	(	PUNCT
fcis-32254	89	29	15	15	NUM
fcis-32254	89	30	)	)	PUNCT
fcis-32254	89	31	:	:	PUNCT
fcis-32254	89	32	(	(	PUNCT
fcis-32254	89	33	(	(	PUNCT
fcis-32254	89	34	)	)	PUNCT
fcis-32254	89	35	(	(	PUNCT
fcis-32254	89	36	)	)	PUNCT
fcis-32254	89	37	)	)	PUNCT
fcis-32254	90	1	[	[	X
fcis-32254	90	2	(	(	PUNCT
fcis-32254	90	3	)	)	PUNCT
fcis-32254	90	4	(	(	PUNCT
fcis-32254	90	5	)	)	PUNCT
fcis-32254	90	6	]	]	PUNCT
fcis-32254	90	7	(	(	PUNCT
fcis-32254	90	8	)	)	PUNCT
fcis-32254	90	9	[	[	X
fcis-32254	90	10	(	(	PUNCT
fcis-32254	90	11	)	)	PUNCT
fcis-32254	90	12	]	]	PUNCT
fcis-32254	91	1	t	t	PROPN
fcis-32254	91	2	t	t	PROPN
fcis-32254	91	3	uu	uu	INTJ
fcis-32254	91	4	l	l	NOUN
fcis-32254	91	5	ij	ij	NOUN
fcis-32254	91	6	t	t	X
fcis-32254	91	7	t	t	X
fcis-32254	91	8	ij	ij	INTJ
fcis-32254	91	9	uu	uu	INTJ
fcis-32254	91	10	jj	jj	PROPN
fcis-32254	91	11	jj	jj	PROPN
fcis-32254	91	12	ii	ii	PROPN
fcis-32254	92	1	i	i	PRON
fcis-32254	92	2	t	t	VERB
fcis-32254	92	3	uu	uu	INTJ
fcis-32254	93	1	ij	ij	INTJ
fcis-32254	93	2	ij	ij	INTJ
fcis-32254	93	3	uu	uu	INTJ
fcis-32254	93	4	jj	jj	PROPN
fcis-32254	93	5	ii	ii	PROPN
fcis-32254	93	6	has	have	AUX
fcis-32254	93	7	labels	label	VERB
fcis-32254	93	8	others	other	NOUN
fcis-32254	93	9	vk	vk	ADP
fcis-32254	93	10	q	q	PROPN
fcis-32254	93	11	a	a	DET
fcis-32254	93	12	v	v	NOUN
fcis-32254	93	13	a	a	DET
fcis-32254	93	14	a	a	DET
fcis-32254	93	15	dv	dv	PROPN
fcis-32254	93	16	v	v	X
fcis-32254	93	17	k	k	PROPN
fcis-32254	93	18	a	a	DET
fcis-32254	93	19	a	a	PROPN
fcis-32254	93	20	l	l	NOUN
fcis-32254	93	21	if	if	SCONJ
fcis-32254	93	22	x	x	PROPN
fcis-32254	93	23	vk	vk	VERB
fcis-32254	93	24	dv	dv	PROPN
fcis-32254	93	25	v	v	PROPN
fcis-32254	93	26	k	k	PROPN
fcis-32254	93	27	l	l	NOUN
fcis-32254	93	28			NOUN
fcis-32254	93	29			X
fcis-32254	93	30			PROPN
fcis-32254	93	31			ADJ
fcis-32254	93	32			NUM
fcis-32254	93	33			PROPN
fcis-32254	93	34			PROPN
fcis-32254	93	35			ADJ
fcis-32254	93	36			PROPN
fcis-32254	93	37			PROPN
fcis-32254	93	38			PROPN
fcis-32254	93	39			ADV
fcis-32254	93	40			PROPN
fcis-32254	93	41			PUNCT
fcis-32254	93	42			PUNCT
fcis-32254	93	43			PROPN
fcis-32254	93	44			NUM
fcis-32254	93	45			PUNCT
fcis-32254	93	46			VERB
fcis-32254	93	47			NUM
fcis-32254	93	48			X
fcis-32254	93	49			PUNCT
fcis-32254	93	50			PROPN
fcis-32254	94	1			PROPN
fcis-32254	94	2	(	(	PUNCT
fcis-32254	94	3	16	16	NUM
fcis-32254	94	4	)	)	PUNCT
fcis-32254	94	5	given	give	VERB
fcis-32254	94	6	(	(	PUNCT
fcis-32254	94	7	)	)	PUNCT
fcis-32254	94	8	(	(	PUNCT
fcis-32254	94	9	)	)	PUNCT
fcis-32254	94	10	t	t	NOUN
fcis-32254	94	11	uu	uu	INTJ
fcis-32254	94	12	ij	ij	X
fcis-32254	94	13	ik	ik	PROPN
fcis-32254	94	14	uu	uu	PROPN
fcis-32254	94	15	kj	kj	PROPN
fcis-32254	94	16	k	k	PROPN
fcis-32254	94	17	vk	vk	PROPN
fcis-32254	94	18	v	v	NUM
fcis-32254	94	19	k	k	ADJ
fcis-32254	94	20			X
fcis-32254	94	21	,	,	PUNCT
fcis-32254	94	22	thus	thus	ADV
fcis-32254	94	23	it	it	PRON
fcis-32254	94	24	is	be	AUX
fcis-32254	94	25	obvious	obvious	ADJ
fcis-32254	94	26	that	that	SCONJ
fcis-32254	94	27	:	:	PUNCT
fcis-32254	94	28	(	(	PUNCT
fcis-32254	94	29	)	)	PUNCT
fcis-32254	94	30	(	(	PUNCT
fcis-32254	94	31	)	)	PUNCT
fcis-32254	94	32	(	(	PUNCT
fcis-32254	94	33	)	)	PUNCT
fcis-32254	94	34	t	t	PROPN
fcis-32254	94	35	t	t	PROPN
fcis-32254	94	36	uu	uu	INTJ
fcis-32254	94	37	ij	ij	X
fcis-32254	94	38	ik	ik	PROPN
fcis-32254	94	39	uu	uu	PROPN
fcis-32254	94	40	kj	kj	PROPN
fcis-32254	95	1	ij	ij	INTJ
fcis-32254	95	2	uu	uu	INTJ
fcis-32254	95	3	jj	jj	PROPN
fcis-32254	95	4	k	k	PROPN
fcis-32254	95	5	vk	vk	PROPN
fcis-32254	95	6	v	v	PROPN
fcis-32254	95	7	k	k	PROPN
fcis-32254	95	8	v	v	X
fcis-32254	95	9	k	k	PROPN
fcis-32254	95	10			PROPN
fcis-32254	95	11	(	(	PUNCT
fcis-32254	95	12	17	17	NUM
fcis-32254	95	13	)	)	PUNCT
fcis-32254	95	14	similarly	similarly	ADV
fcis-32254	95	15	,	,	PUNCT
fcis-32254	95	16	it	it	PRON
fcis-32254	95	17	can	can	AUX
fcis-32254	95	18	be	be	AUX
fcis-32254	95	19	obtained	obtain	VERB
fcis-32254	95	20	that	that	PRON
fcis-32254	95	21	:	:	PUNCT
fcis-32254	95	22	(	(	PUNCT
fcis-32254	95	23	)	)	PUNCT
fcis-32254	95	24	(	(	PUNCT
fcis-32254	95	25	)	)	PUNCT
fcis-32254	95	26	t	t	NOUN
fcis-32254	95	27	t	t	PROPN
fcis-32254	95	28	ij	ij	X
fcis-32254	95	29	ik	ik	PROPN
fcis-32254	95	30	kj	kj	PROPN
fcis-32254	95	31	ii	ii	PROPN
fcis-32254	96	1	ij	ij	INTJ
fcis-32254	97	1	k	k	PROPN
fcis-32254	97	2	t	t	PROPN
fcis-32254	97	3	t	t	PROPN
fcis-32254	97	4	ii	ii	PROPN
fcis-32254	97	5	ii	ii	PROPN
fcis-32254	97	6	ij	ij	PROPN
fcis-32254	97	7	ii	ii	PROPN
fcis-32254	98	1	ij	ij	INTJ
fcis-32254	98	2	dv	dv	PROPN
fcis-32254	98	3	d	d	PROPN
fcis-32254	98	4	v	v	PROPN
fcis-32254	98	5	d	d	X
fcis-32254	98	6	v	v	PROPN
fcis-32254	98	7	d	d	PROPN
fcis-32254	98	8	w	w	PROPN
fcis-32254	98	9	v	v	PROPN
fcis-32254	98	10	l	l	NOUN
fcis-32254	98	11	v	v	ADP
fcis-32254	98	12			PROPN
fcis-32254	98	13			PROPN
fcis-32254	98	14			PROPN
fcis-32254	98	15			PROPN
fcis-32254	98	16			PROPN
fcis-32254	98	17			NUM
fcis-32254	98	18			NUM
fcis-32254	98	19			X
fcis-32254	98	20			PROPN
fcis-32254	98	21			PRON
fcis-32254	98	22			X
fcis-32254	98	23	(	(	PUNCT
fcis-32254	98	24	18	18	NUM
fcis-32254	98	25	)	)	PUNCT
fcis-32254	98	26	moreover	moreover	ADV
fcis-32254	98	27	,	,	PUNCT
fcis-32254	98	28	if	if	SCONJ
fcis-32254	98	29	ax	ax	NOUN
fcis-32254	98	30	has	have	VERB
fcis-32254	98	31	labels	label	NOUN
fcis-32254	98	32	,	,	PUNCT
fcis-32254	98	33	then	then	ADV
fcis-32254	98	34	both	both	DET
fcis-32254	98	35	(	(	PUNCT
fcis-32254	98	36	)	)	PUNCT
fcis-32254	98	37	l	l	NOUN
fcis-32254	98	38	ab	ab	PROPN
fcis-32254	98	39	abv	abv	PROPN
fcis-32254	98	40	v	v	ADV
fcis-32254	98	41	and	and	CCONJ
fcis-32254	98	42	(	(	PUNCT
fcis-32254	98	43	)	)	PUNCT
fcis-32254	98	44	0tq	0tq	NOUN
fcis-32254	98	45	a	a	DET
fcis-32254	98	46			PROPN
fcis-32254	98	47			NUM
fcis-32254	98	48	hold	hold	VERB
fcis-32254	98	49	;	;	PUNCT
fcis-32254	98	50	thus	thus	ADV
fcis-32254	98	51	it	it	PRON
fcis-32254	98	52	follows	follow	VERB
fcis-32254	98	53	that	that	SCONJ
fcis-32254	98	54	:	:	PUNCT
fcis-32254	98	55	[	[	PUNCT
fcis-32254	98	56	(	(	PUNCT
fcis-32254	98	57	)	)	PUNCT
fcis-32254	98	58	(	(	PUNCT
fcis-32254	98	59	)	)	PUNCT
fcis-32254	98	60	]	]	PUNCT
fcis-32254	99	1	[	[	PUNCT
fcis-32254	99	2	(	(	PUNCT
fcis-32254	99	3	)	)	PUNCT
fcis-32254	99	4	]	]	PUNCT
fcis-32254	99	5	[	[	PUNCT
fcis-32254	99	6	(	(	PUNCT
fcis-32254	99	7	)	)	PUNCT
fcis-32254	99	8	]	]	PUNCT
fcis-32254	99	9	(	(	PUNCT
fcis-32254	99	10	)	)	PUNCT
fcis-32254	99	11	(	(	PUNCT
fcis-32254	99	12	)	)	PUNCT
fcis-32254	99	13	t	t	NOUN
fcis-32254	99	14	t	t	PROPN
fcis-32254	99	15	t	t	PROPN
fcis-32254	99	16	l	l	PROPN
fcis-32254	99	17	ab	ab	PROPN
fcis-32254	99	18	l	l	PROPN
fcis-32254	99	19	ab	ab	PROPN
fcis-32254	99	20	t	t	PROPN
fcis-32254	99	21	t	t	PROPN
fcis-32254	99	22	t	t	PROPN
fcis-32254	99	23	t	t	PROPN
fcis-32254	99	24	t	t	PROPN
fcis-32254	99	25	l	l	PROPN
fcis-32254	99	26	ab	ab	PROPN
fcis-32254	99	27	ik	ik	PROPN
fcis-32254	99	28	kj	kj	PROPN
fcis-32254	99	29	ij	ij	INTJ
fcis-32254	100	1	jj	jj	PROPN
fcis-32254	100	2	k	k	PROPN
fcis-32254	100	3	q	q	PROPN
fcis-32254	100	4	a	a	DET
fcis-32254	100	5	v	v	NOUN
fcis-32254	100	6	a	a	DET
fcis-32254	100	7	a	a	DET
fcis-32254	100	8	v	v	NOUN
fcis-32254	100	9	a	a	DET
fcis-32254	100	10	a	a	DET
fcis-32254	100	11	v	v	NOUN
fcis-32254	100	12	a	a	DET
fcis-32254	100	13	a	a	DET
fcis-32254	100	14	v	v	NOUN
fcis-32254	100	15	a	a	DET
fcis-32254	100	16	a	a	DET
fcis-32254	100	17	v	v	NOUN
fcis-32254	100	18	a	a	DET
fcis-32254	100	19	a	a	DET
fcis-32254	100	20			NOUN
fcis-32254	100	21			X
fcis-32254	100	22			X
fcis-32254	100	23			X
fcis-32254	100	24			X
fcis-32254	100	25			X
fcis-32254	100	26			PROPN
fcis-32254	100	27			ADV
fcis-32254	100	28			PROPN
fcis-32254	100	29			PRON
fcis-32254	100	30			NUM
fcis-32254	100	31			PROPN
fcis-32254	100	32			PROPN
fcis-32254	100	33	(	(	PUNCT
fcis-32254	100	34	19	19	NUM
fcis-32254	100	35	)	)	PUNCT
fcis-32254	100	36	from	from	ADP
fcis-32254	100	37	equations	equation	NOUN
fcis-32254	100	38	(	(	PUNCT
fcis-32254	100	39	17	17	NUM
fcis-32254	100	40	)	)	PUNCT
fcis-32254	100	41	,	,	PUNCT
fcis-32254	100	42	(	(	PUNCT
fcis-32254	100	43	18	18	NUM
fcis-32254	100	44	)	)	PUNCT
fcis-32254	100	45	,	,	PUNCT
fcis-32254	100	46	and	and	CCONJ
fcis-32254	100	47	(	(	PUNCT
fcis-32254	100	48	19	19	NUM
fcis-32254	100	49	)	)	PUNCT
fcis-32254	101	1	,	,	PUNCT
fcis-32254	101	2	it	it	PRON
fcis-32254	101	3	can	can	AUX
fcis-32254	101	4	be	be	AUX
fcis-32254	101	5	proven	prove	VERB
fcis-32254	101	6	that	that	SCONJ
fcis-32254	101	7	equation	equation	NOUN
fcis-32254	101	8	(	(	PUNCT
fcis-32254	101	9	17	17	NUM
fcis-32254	101	10	)	)	PUNCT
fcis-32254	101	11	holds	hold	VERB
fcis-32254	101	12	.	.	PUNCT
fcis-32254	102	1	thus	thus	ADV
fcis-32254	102	2	,	,	PUNCT
fcis-32254	102	3	equation	equation	NOUN
fcis-32254	102	4	(	(	PUNCT
fcis-32254	102	5	14	14	NUM
fcis-32254	102	6	)	)	PUNCT
fcis-32254	102	7	is	be	AUX
fcis-32254	102	8	the	the	DET
fcis-32254	102	9	auxiliary	auxiliary	ADJ
fcis-32254	102	10	function	function	NOUN
fcis-32254	102	11	of	of	ADP
fcis-32254	102	12	ijv	ijv	PROPN
fcis-32254	102	13	m	m	PROPN
fcis-32254	102	14	,	,	PUNCT
fcis-32254	102	15	and	and	CCONJ
fcis-32254	102	16	theorem	theorem	VERB
fcis-32254	102	17	2	2	NUM
fcis-32254	102	18	is	be	AUX
fcis-32254	102	19	proven	prove	VERB
fcis-32254	102	20	.	.	PUNCT
fcis-32254	103	1	the	the	DET
fcis-32254	103	2	above	above	ADJ
fcis-32254	103	3	is	be	AUX
fcis-32254	103	4	the	the	DET
fcis-32254	103	5	convergence	convergence	NOUN
fcis-32254	103	6	proof	proof	NOUN
fcis-32254	103	7	for	for	ADP
fcis-32254	103	8	the	the	DET
fcis-32254	103	9	update	update	NOUN
fcis-32254	103	10	rule	rule	NOUN
fcis-32254	103	11	of	of	ADP
fcis-32254	103	12	matrix	matrix	NOUN
fcis-32254	103	13	v	v	NOUN
fcis-32254	103	14	,	,	PUNCT
fcis-32254	103	15	and	and	CCONJ
fcis-32254	103	16	the	the	DET
fcis-32254	103	17	update	update	NOUN
fcis-32254	103	18	rule	rule	NOUN
fcis-32254	103	19	of	of	ADP
fcis-32254	103	20	matrix	matrix	NOUN
fcis-32254	103	21	u	u	NOUN
fcis-32254	103	22	can	can	AUX
fcis-32254	103	23	be	be	AUX
fcis-32254	103	24	proven	prove	VERB
fcis-32254	103	25	similarly	similarly	ADV
fcis-32254	103	26	.	.	PUNCT
fcis-32254	104	1	since	since	SCONJ
fcis-32254	104	2	objective	objective	ADJ
fcis-32254	104	3	function	function	NOUN
fcis-32254	104	4	a	a	PRON
fcis-32254	104	5	(	(	PUNCT
fcis-32254	104	6	with	with	ADP
fcis-32254	104	7	respect	respect	NOUN
fcis-32254	104	8	to	to	ADP
fcis-32254	104	9	the	the	DET
fcis-32254	104	10	matrix	matrix	NOUN
fcis-32254	104	11	)	)	PUNCT
fcis-32254	104	12	in	in	ADP
fcis-32254	104	13	algorithm	algorithm	NOUN
fcis-32254	104	14	model	model	NOUN
fcis-32254	104	15	(	(	PUNCT
fcis-32254	104	16	1	1	X
fcis-32254	104	17	)	)	PUNCT
fcis-32254	104	18	is	be	AUX
fcis-32254	104	19	a	a	DET
fcis-32254	104	20	convex	convex	NOUN
fcis-32254	104	21	function	function	NOUN
fcis-32254	104	22	,	,	PUNCT
fcis-32254	104	23	the	the	DET
fcis-32254	104	24	update	update	NOUN
fcis-32254	104	25	rule	rule	NOUN
fcis-32254	104	26	can	can	AUX
fcis-32254	104	27	be	be	AUX
fcis-32254	104	28	obtained	obtain	VERB
fcis-32254	104	29	directly	directly	ADV
fcis-32254	104	30	by	by	ADP
fcis-32254	104	31	setting	set	VERB
fcis-32254	104	32	0	0	PUNCT
fcis-32254	104	33	a	a	DET
fcis-32254	104	34			ADJ
fcis-32254	104	35			PROPN
fcis-32254	104	36			ADJ
fcis-32254	104	37	l	l	NOUN
fcis-32254	104	38	and	and	CCONJ
fcis-32254	104	39	solving	solve	VERB
fcis-32254	104	40	it	it	PRON
fcis-32254	104	41	;	;	PUNCT
fcis-32254	104	42	thus	thus	ADV
fcis-32254	104	43	,	,	PUNCT
fcis-32254	104	44	the	the	DET
fcis-32254	104	45	update	update	NOUN
fcis-32254	104	46	formula	formula	NOUN
fcis-32254	104	47	of	of	ADP
fcis-32254	104	48	matrix	matrix	NOUN
fcis-32254	104	49	a	a	PRON
fcis-32254	104	50	is	be	AUX
fcis-32254	104	51	also	also	ADV
fcis-32254	104	52	convergent	convergent	ADJ
fcis-32254	104	53	.	.	PUNCT
fcis-32254	105	1	in	in	ADP
fcis-32254	105	2	conclusion	conclusion	NOUN
fcis-32254	105	3	,	,	PUNCT
fcis-32254	105	4	theorem	theorem	VERB
fcis-32254	105	5	1	1	NUM
fcis-32254	105	6	is	be	AUX
fcis-32254	105	7	proven	prove	VERB
fcis-32254	105	8	.	.	PUNCT
fcis-32254	106	1	3	3	X
fcis-32254	106	2	.	.	X
fcis-32254	106	3	experiments	experiment	NOUN
fcis-32254	106	4	and	and	CCONJ
fcis-32254	106	5	results	result	VERB
fcis-32254	106	6	analysis	analysis	NOUN
fcis-32254	106	7	this	this	DET
fcis-32254	106	8	paper	paper	NOUN
fcis-32254	106	9	tests	test	VERB
fcis-32254	106	10	the	the	DET
fcis-32254	106	11	clustering	clustering	ADJ
fcis-32254	106	12	performance	performance	NOUN
fcis-32254	106	13	of	of	ADP
fcis-32254	106	14	the	the	DET
fcis-32254	106	15	22	22	NUM
fcis-32254	106	16	skgdnmf	skgdnmf	NOUN
fcis-32254	106	17	algorithm	algorithm	NOUN
fcis-32254	106	18	on	on	ADP
fcis-32254	106	19	the	the	DET
fcis-32254	106	20	face	face	NOUN
fcis-32254	106	21	dataset	dataset	NOUN
fcis-32254	106	22	pie	pie	NOUN
fcis-32254	106	23	,	,	PUNCT
fcis-32254	106	24	the	the	DET
fcis-32254	106	25	object	object	NOUN
fcis-32254	106	26	dataset	dataset	NOUN
fcis-32254	106	27	coil20	coil20	NOUN
fcis-32254	106	28	,	,	PUNCT
fcis-32254	106	29	and	and	CCONJ
fcis-32254	106	30	the	the	DET
fcis-32254	106	31	handwritten	handwritten	ADJ
fcis-32254	106	32	digit	digit	NOUN
fcis-32254	106	33	dataset	dataset	NOUN
fcis-32254	106	34	mnist	mnist	NOUN
fcis-32254	106	35	,	,	PUNCT
fcis-32254	106	36	and	and	CCONJ
fcis-32254	106	37	compares	compare	VERB
fcis-32254	106	38	it	it	PRON
fcis-32254	106	39	with	with	ADP
fcis-32254	106	40	other	other	ADJ
fcis-32254	106	41	algorithms	algorithm	NOUN
fcis-32254	106	42	including	include	VERB
fcis-32254	106	43	k	k	NOUN
fcis-32254	106	44	-	-	PUNCT
fcis-32254	106	45	means	means	NOUN
fcis-32254	106	46	[	[	X
fcis-32254	106	47	5	5	NUM
fcis-32254	106	48	]	]	PUNCT
fcis-32254	106	49	,	,	PUNCT
fcis-32254	106	50	kpca	kpca	X
fcis-32254	106	51	[	[	X
fcis-32254	106	52	6	6	NUM
fcis-32254	106	53	]	]	PUNCT
fcis-32254	106	54	,	,	PUNCT
fcis-32254	106	55	nmf	nmf	PROPN
fcis-32254	107	1	[	[	X
fcis-32254	107	2	7	7	NUM
fcis-32254	107	3	]	]	PUNCT
fcis-32254	107	4	,	,	PUNCT
fcis-32254	107	5	gnmf	gnmf	NOUN
fcis-32254	107	6	,	,	PUNCT
fcis-32254	107	7	gdnmf	gdnmf	NOUN
fcis-32254	107	8	[	[	X
fcis-32254	107	9	8	8	NUM
fcis-32254	107	10	]	]	PUNCT
fcis-32254	107	11	,	,	PUNCT
fcis-32254	107	12	and	and	CCONJ
fcis-32254	107	13	kgdnmf	kgdnmf	NOUN
fcis-32254	108	1	[	[	X
fcis-32254	108	2	9	9	NUM
fcis-32254	108	3	]	]	PUNCT
fcis-32254	108	4	in	in	ADP
fcis-32254	108	5	terms	term	NOUN
fcis-32254	108	6	of	of	ADP
fcis-32254	108	7	clustering	clustering	ADJ
fcis-32254	108	8	performance	performance	NOUN
fcis-32254	108	9	.	.	PUNCT
fcis-32254	109	1	3.1	3.1	NUM
fcis-32254	109	2	.	.	PUNCT
fcis-32254	109	3	experimental	experimental	ADJ
fcis-32254	109	4	datasets	dataset	NOUN
fcis-32254	109	5	in	in	ADP
fcis-32254	109	6	the	the	DET
fcis-32254	109	7	coil20	coil20	NOUN
fcis-32254	109	8	dataset	dataset	NOUN
fcis-32254	109	9	,	,	PUNCT
fcis-32254	109	10	all	all	DET
fcis-32254	109	11	images	image	NOUN
fcis-32254	109	12	are	be	AUX
fcis-32254	109	13	uniformly	uniformly	ADV
fcis-32254	109	14	resized	resize	VERB
fcis-32254	109	15	to	to	ADP
fcis-32254	109	16	a	a	DET
fcis-32254	109	17	size	size	NOUN
fcis-32254	109	18	of	of	ADP
fcis-32254	109	19	32×32	32×32	NUM
fcis-32254	109	20	pixels	pixel	NOUN
fcis-32254	109	21	.	.	PUNCT
fcis-32254	110	1	in	in	ADP
fcis-32254	110	2	the	the	DET
fcis-32254	110	3	pie	pie	NOUN
fcis-32254	110	4	dataset	dataset	NOUN
fcis-32254	110	5	,	,	PUNCT
fcis-32254	110	6	68	68	NUM
fcis-32254	110	7	subjects	subject	NOUN
fcis-32254	110	8	are	be	AUX
fcis-32254	110	9	randomly	randomly	ADV
fcis-32254	110	10	selected	select	VERB
fcis-32254	110	11	,	,	PUNCT
fcis-32254	110	12	and	and	CCONJ
fcis-32254	110	13	42	42	NUM
fcis-32254	110	14	facial	facial	ADJ
fcis-32254	110	15	images	image	NOUN
fcis-32254	110	16	of	of	ADP
fcis-32254	110	17	each	each	DET
fcis-32254	110	18	subject	subject	NOUN
fcis-32254	110	19	under	under	ADP
fcis-32254	110	20	different	different	ADJ
fcis-32254	110	21	lighting	lighting	NOUN
fcis-32254	110	22	conditions	condition	NOUN
fcis-32254	110	23	are	be	AUX
fcis-32254	110	24	used	use	VERB
fcis-32254	110	25	;	;	PUNCT
fcis-32254	110	26	these	these	DET
fcis-32254	110	27	images	image	NOUN
fcis-32254	110	28	are	be	AUX
fcis-32254	110	29	uniformly	uniformly	ADV
fcis-32254	110	30	cropped	crop	VERB
fcis-32254	110	31	to	to	ADP
fcis-32254	110	32	32×32	32×32	NUM
fcis-32254	110	33	grayscale	grayscale	NOUN
fcis-32254	110	34	images	image	NOUN
fcis-32254	110	35	.	.	PUNCT
fcis-32254	111	1	for	for	SCONJ
fcis-32254	111	2	the	the	DET
fcis-32254	111	3	mnist	mnist	PROPN
fcis-32254	111	4	dataset	dataset	NOUN
fcis-32254	111	5	,	,	PUNCT
fcis-32254	111	6	200	200	NUM
fcis-32254	111	7	images	image	NOUN
fcis-32254	111	8	are	be	AUX
fcis-32254	111	9	randomly	randomly	ADV
fcis-32254	111	10	selected	select	VERB
fcis-32254	111	11	for	for	ADP
fcis-32254	111	12	each	each	DET
fcis-32254	111	13	digit	digit	NOUN
fcis-32254	111	14	(	(	PUNCT
fcis-32254	111	15	0	0	NUM
fcis-32254	111	16	-	-	SYM
fcis-32254	111	17	9	9	NUM
fcis-32254	111	18	)	)	PUNCT
fcis-32254	111	19	from	from	ADP
fcis-32254	111	20	the	the	DET
fcis-32254	111	21	training	training	NOUN
fcis-32254	111	22	set	set	NOUN
fcis-32254	111	23	,	,	PUNCT
fcis-32254	111	24	resulting	result	VERB
fcis-32254	111	25	in	in	ADP
fcis-32254	111	26	a	a	DET
fcis-32254	111	27	total	total	NOUN
fcis-32254	111	28	of	of	ADP
fcis-32254	111	29	2000	2000	NUM
fcis-32254	111	30	images	image	NOUN
fcis-32254	111	31	.	.	PUNCT
fcis-32254	112	1	3.2	3.2	NUM
fcis-32254	112	2	.	.	PUNCT
fcis-32254	112	3	experimental	experimental	ADJ
fcis-32254	112	4	process	process	NOUN
fcis-32254	112	5	at	at	ADP
fcis-32254	112	6	the	the	DET
fcis-32254	112	7	start	start	NOUN
fcis-32254	112	8	of	of	ADP
fcis-32254	112	9	the	the	DET
fcis-32254	112	10	experiment	experiment	NOUN
fcis-32254	112	11	,	,	PUNCT
fcis-32254	112	12	we	we	PRON
fcis-32254	112	13	first	first	ADV
fcis-32254	112	14	load	load	VERB
fcis-32254	112	15	and	and	CCONJ
fcis-32254	112	16	standardize	standardize	VERB
fcis-32254	112	17	the	the	DET
fcis-32254	112	18	three	three	NUM
fcis-32254	112	19	datasets	dataset	NOUN
fcis-32254	112	20	,	,	PUNCT
fcis-32254	112	21	then	then	ADV
fcis-32254	112	22	divide	divide	VERB
fcis-32254	112	23	them	they	PRON
fcis-32254	112	24	into	into	ADP
fcis-32254	112	25	a	a	DET
fcis-32254	112	26	preset	preset	ADJ
fcis-32254	112	27	number	number	NOUN
fcis-32254	112	28	of	of	ADP
fcis-32254	112	29	categories	category	NOUN
fcis-32254	112	30	.	.	PUNCT
fcis-32254	113	1	for	for	ADP
fcis-32254	113	2	semi	semi	ADJ
fcis-32254	113	3	-	-	ADJ
fcis-32254	113	4	supervised	supervised	ADJ
fcis-32254	113	5	clustering	clustering	ADJ
fcis-32254	113	6	algorithms	algorithm	NOUN
fcis-32254	113	7	,	,	PUNCT
fcis-32254	113	8	we	we	PRON
fcis-32254	113	9	randomly	randomly	VERB
fcis-32254	113	10	select	select	VERB
fcis-32254	113	11	k	k	PROPN
fcis-32254	113	12	categories	category	NOUN
fcis-32254	113	13	of	of	ADP
fcis-32254	113	14	data	datum	NOUN
fcis-32254	113	15	subsets	subset	NOUN
fcis-32254	113	16	each	each	DET
fcis-32254	113	17	time	time	NOUN
fcis-32254	113	18	,	,	PUNCT
fcis-32254	113	19	randomly	randomly	ADV
fcis-32254	113	20	extract	extract	VERB
fcis-32254	113	21	20	20	NUM
fcis-32254	113	22	%	%	NOUN
fcis-32254	113	23	of	of	ADP
fcis-32254	113	24	the	the	DET
fcis-32254	113	25	samples	sample	NOUN
fcis-32254	113	26	from	from	ADP
fcis-32254	113	27	each	each	DET
fcis-32254	113	28	category	category	NOUN
fcis-32254	113	29	to	to	PART
fcis-32254	113	30	construct	construct	VERB
fcis-32254	113	31	labeled	labeled	ADJ
fcis-32254	113	32	data	datum	NOUN
fcis-32254	113	33	,	,	PUNCT
fcis-32254	113	34	and	and	CCONJ
fcis-32254	113	35	then	then	ADV
fcis-32254	113	36	run	run	VERB
fcis-32254	113	37	the	the	DET
fcis-32254	113	38	algorithm	algorithm	NOUN
fcis-32254	113	39	model	model	NOUN
fcis-32254	113	40	for	for	ADP
fcis-32254	113	41	semi	semi	ADJ
fcis-32254	113	42	-	-	ADJ
fcis-32254	113	43	supervised	supervised	ADJ
fcis-32254	113	44	decomposition	decomposition	NOUN
fcis-32254	113	45	.	.	PUNCT
fcis-32254	114	1	for	for	ADP
fcis-32254	114	2	the	the	DET
fcis-32254	114	3	other	other	ADJ
fcis-32254	114	4	algorithms	algorithm	NOUN
fcis-32254	114	5	,	,	PUNCT
fcis-32254	114	6	we	we	PRON
fcis-32254	114	7	set	set	VERB
fcis-32254	114	8	the	the	DET
fcis-32254	114	9	dimensionality	dimensionality	NOUN
fcis-32254	114	10	reduction	reduction	NOUN
fcis-32254	114	11	dimension	dimension	NOUN
fcis-32254	114	12	to	to	ADP
fcis-32254	114	13	the	the	DET
fcis-32254	114	14	corresponding	corresponding	ADJ
fcis-32254	114	15	number	number	NOUN
fcis-32254	114	16	of	of	ADP
fcis-32254	114	17	clusters	cluster	NOUN
fcis-32254	114	18	k.	k.	PROPN
fcis-32254	115	1	in	in	ADP
fcis-32254	115	2	the	the	DET
fcis-32254	115	3	algorithm	algorithm	NOUN
fcis-32254	115	4	postprocessing	postprocessing	NOUN
fcis-32254	115	5	stage	stage	NOUN
fcis-32254	115	6	,	,	PUNCT
fcis-32254	115	7	clustering	cluster	VERB
fcis-32254	115	8	labels	label	NOUN
fcis-32254	115	9	are	be	AUX
fcis-32254	115	10	obtained	obtain	VERB
fcis-32254	115	11	through	through	ADP
fcis-32254	115	12	spectral	spectral	ADJ
fcis-32254	115	13	clustering	clustering	NOUN
fcis-32254	115	14	[	[	X
fcis-32254	115	15	10	10	NUM
fcis-32254	115	16	]	]	PUNCT
fcis-32254	115	17	.	.	PUNCT
fcis-32254	116	1	we	we	PRON
fcis-32254	116	2	calculate	calculate	VERB
fcis-32254	116	3	the	the	DET
fcis-32254	116	4	accuracy	accuracy	NOUN
fcis-32254	116	5	and	and	CCONJ
fcis-32254	116	6	mutual	mutual	ADJ
fcis-32254	116	7	information	information	NOUN
fcis-32254	116	8	(	(	PUNCT
fcis-32254	116	9	mi	mi	NOUN
fcis-32254	116	10	)	)	PUNCT
fcis-32254	116	11	metrics	metric	NOUN
fcis-32254	116	12	,	,	PUNCT
fcis-32254	116	13	repeat	repeat	VERB
fcis-32254	116	14	the	the	DET
fcis-32254	116	15	experiment	experiment	NOUN
fcis-32254	116	16	10	10	NUM
fcis-32254	116	17	times	time	NOUN
fcis-32254	116	18	for	for	SCONJ
fcis-32254	116	19	each	each	DET
fcis-32254	116	20	k	k	NOUN
fcis-32254	116	21	to	to	PART
fcis-32254	116	22	obtain	obtain	VERB
fcis-32254	116	23	the	the	DET
fcis-32254	116	24	mean	mean	ADJ
fcis-32254	116	25	and	and	CCONJ
fcis-32254	116	26	standard	standard	ADJ
fcis-32254	116	27	deviation	deviation	NOUN
fcis-32254	116	28	,	,	PUNCT
fcis-32254	116	29	and	and	CCONJ
fcis-32254	116	30	finally	finally	ADV
fcis-32254	116	31	output	output	VERB
fcis-32254	116	32	the	the	DET
fcis-32254	116	33	result	result	NOUN
fcis-32254	116	34	table	table	NOUN
fcis-32254	116	35	.	.	PUNCT
fcis-32254	117	1	the	the	DET
fcis-32254	117	2	algorithm	algorithm	NOUN
fcis-32254	117	3	proposed	propose	VERB
fcis-32254	117	4	in	in	ADP
fcis-32254	117	5	this	this	DET
fcis-32254	117	6	paper	paper	NOUN
fcis-32254	117	7	has	have	VERB
fcis-32254	117	8	a	a	DET
fcis-32254	117	9	total	total	NOUN
fcis-32254	117	10	of	of	ADP
fcis-32254	117	11	four	four	NUM
fcis-32254	117	12	parameters	parameter	NOUN
fcis-32254	117	13	,	,	PUNCT
fcis-32254	117	14	and	and	CCONJ
fcis-32254	117	15	different	different	ADJ
fcis-32254	117	16	parameters	parameter	NOUN
fcis-32254	117	17	exert	exert	VERB
fcis-32254	117	18	different	different	ADJ
fcis-32254	117	19	impacts	impact	NOUN
fcis-32254	117	20	on	on	ADP
fcis-32254	117	21	the	the	DET
fcis-32254	117	22	experimental	experimental	ADJ
fcis-32254	117	23	results	result	NOUN
fcis-32254	117	24	.	.	PUNCT
fcis-32254	118	1	therefore	therefore	ADV
fcis-32254	118	2	,	,	PUNCT
fcis-32254	118	3	this	this	DET
fcis-32254	118	4	paper	paper	NOUN
fcis-32254	118	5	adopts	adopt	VERB
fcis-32254	118	6	the	the	DET
fcis-32254	118	7	control	control	NOUN
fcis-32254	118	8	variable	variable	ADJ
fcis-32254	118	9	method	method	NOUN
fcis-32254	118	10	to	to	PART
fcis-32254	118	11	investigate	investigate	VERB
fcis-32254	118	12	the	the	DET
fcis-32254	118	13	selection	selection	NOUN
fcis-32254	118	14	of	of	ADP
fcis-32254	118	15	experimental	experimental	ADJ
fcis-32254	118	16	parameters	parameter	NOUN
fcis-32254	118	17	on	on	ADP
fcis-32254	118	18	the	the	DET
fcis-32254	118	19	three	three	NUM
fcis-32254	118	20	datasets	dataset	NOUN
fcis-32254	118	21	.	.	PUNCT
fcis-32254	119	1	with	with	ADP
fcis-32254	119	2	the	the	DET
fcis-32254	119	3	number	number	NOUN
fcis-32254	119	4	of	of	ADP
fcis-32254	119	5	neighbors	neighbor	NOUN
fcis-32254	119	6	fixed	fix	VERB
fcis-32254	119	7	,	,	PUNCT
fcis-32254	119	8	we	we	PRON
fcis-32254	119	9	conduct	conduct	VERB
fcis-32254	119	10	parameter	parameter	NOUN
fcis-32254	119	11	sensitivity	sensitivity	NOUN
fcis-32254	119	12	analysis	analysis	NOUN
fcis-32254	119	13	for	for	ADP
fcis-32254	119	14			NOUN
fcis-32254	119	15	,	,	PUNCT
fcis-32254	119	16			PROPN
fcis-32254	119	17	,	,	PUNCT
fcis-32254	119	18	and	and	CCONJ
fcis-32254	119	19			NUM
fcis-32254	119	20	respectively	respectively	ADV
fcis-32254	119	21	,	,	PUNCT
fcis-32254	119	22	repeat	repeat	VERB
fcis-32254	119	23	the	the	DET
fcis-32254	119	24	experiment	experiment	NOUN
fcis-32254	119	25	ten	ten	NUM
fcis-32254	119	26	times	time	NOUN
fcis-32254	119	27	to	to	PART
fcis-32254	119	28	calculate	calculate	VERB
fcis-32254	119	29	the	the	DET
fcis-32254	119	30	classification	classification	NOUN
fcis-32254	119	31	accuracy	accuracy	NOUN
fcis-32254	119	32	,	,	PUNCT
fcis-32254	119	33	and	and	CCONJ
fcis-32254	119	34	take	take	VERB
fcis-32254	119	35	the	the	DET
fcis-32254	119	36	average	average	ADJ
fcis-32254	119	37	value	value	NOUN
fcis-32254	119	38	.	.	PUNCT
fcis-32254	120	1	the	the	DET
fcis-32254	120	2	relevant	relevant	ADJ
fcis-32254	120	3	results	result	NOUN
fcis-32254	120	4	are	be	AUX
fcis-32254	120	5	shown	show	VERB
fcis-32254	120	6	in	in	ADP
fcis-32254	120	7	the	the	DET
fcis-32254	120	8	figure	figure	NOUN
fcis-32254	120	9	below	below	ADV
fcis-32254	120	10	:	:	PUNCT
fcis-32254	120	11	figure	figure	VERB
fcis-32254	120	12	1	1	NUM
fcis-32254	120	13	.	.	PUNCT
fcis-32254	121	1	parameter	parameter	NOUN
fcis-32254	121	2	sensitivity	sensitivity	NOUN
fcis-32254	121	3	analysis	analysis	NOUN
fcis-32254	121	4	from	from	ADP
fcis-32254	121	5	the	the	DET
fcis-32254	121	6	three	three	NUM
fcis-32254	121	7	groups	group	NOUN
fcis-32254	121	8	of	of	ADP
fcis-32254	121	9	experiments	experiment	NOUN
fcis-32254	121	10	,	,	PUNCT
fcis-32254	121	11	it	it	PRON
fcis-32254	121	12	can	can	AUX
fcis-32254	121	13	be	be	AUX
fcis-32254	121	14	observed	observe	VERB
fcis-32254	121	15	that	that	SCONJ
fcis-32254	121	16	the	the	DET
fcis-32254	121	17	proposed	propose	VERB
fcis-32254	121	18	algorithm	algorithm	NOUN
fcis-32254	121	19	exhibits	exhibit	VERB
fcis-32254	121	20	extremely	extremely	ADV
fcis-32254	121	21	high	high	ADJ
fcis-32254	121	22	robustness	robustness	NOUN
fcis-32254	121	23	to	to	ADP
fcis-32254	121	24	the	the	DET
fcis-32254	121	25	label	label	NOUN
fcis-32254	121	26	constraint	constraint	NOUN
fcis-32254	121	27	weight	weight	NOUN
fcis-32254	121	28			NOUN
fcis-32254	121	29	and	and	CCONJ
fcis-32254	121	30	the	the	DET
fcis-32254	121	31	warmstart	warmstart	ADJ
fcis-32254	121	32	weight	weight	PROPN
fcis-32254	121	33			NUM
fcis-32254	121	34	;	;	PUNCT
fcis-32254	121	35	parameter	parameter	NOUN
fcis-32254	121	36	variations	variation	NOUN
fcis-32254	121	37	hardly	hardly	ADV
fcis-32254	121	38	affect	affect	VERB
fcis-32254	121	39	the	the	DET
fcis-32254	121	40	clustering	clustering	ADJ
fcis-32254	121	41	performance	performance	NOUN
fcis-32254	121	42	.	.	PUNCT
fcis-32254	122	1	the	the	DET
fcis-32254	122	2	graph	graph	NOUN
fcis-32254	122	3	regularization	regularization	NOUN
fcis-32254	122	4	weight	weight	NOUN
fcis-32254	122	5			PROPN
fcis-32254	122	6	only	only	ADV
fcis-32254	122	7	has	have	VERB
fcis-32254	122	8	slight	slight	ADJ
fcis-32254	122	9	optimization	optimization	NOUN
fcis-32254	122	10	potential	potential	NOUN
fcis-32254	122	11	for	for	ADP
fcis-32254	122	12	specific	specific	ADJ
fcis-32254	122	13	datasets	dataset	NOUN
fcis-32254	122	14	,	,	PUNCT
fcis-32254	122	15	such	such	ADJ
fcis-32254	122	16	as	as	ADP
fcis-32254	122	17	the	the	DET
fcis-32254	122	18	coil20	coil20	NOUN
fcis-32254	122	19	dataset	dataset	VERB
fcis-32254	122	20	with	with	ADP
fcis-32254	122	21	high	high	ADJ
fcis-32254	122	22	graph	graph	NOUN
fcis-32254	122	23	structure	structure	NOUN
fcis-32254	122	24	dependence	dependence	NOUN
fcis-32254	122	25	.	.	PUNCT
fcis-32254	123	1	this	this	DET
fcis-32254	123	2	characteristic	characteristic	NOUN
fcis-32254	123	3	also	also	ADV
fcis-32254	123	4	indicates	indicate	VERB
fcis-32254	123	5	that	that	SCONJ
fcis-32254	123	6	most	most	ADJ
fcis-32254	123	7	parameters	parameter	NOUN
fcis-32254	123	8	of	of	ADP
fcis-32254	123	9	the	the	DET
fcis-32254	123	10	skgdnmf	skgdnmf	NOUN
fcis-32254	123	11	algorithm	algorithm	NOUN
fcis-32254	123	12	do	do	AUX
fcis-32254	123	13	not	not	PART
fcis-32254	123	14	require	require	VERB
fcis-32254	123	15	strict	strict	ADJ
fcis-32254	123	16	tuning	tuning	NOUN
fcis-32254	123	17	to	to	PART
fcis-32254	123	18	ensure	ensure	VERB
fcis-32254	123	19	model	model	NOUN
fcis-32254	123	20	stability	stability	NOUN
fcis-32254	123	21	,	,	PUNCT
fcis-32254	123	22	further	far	ADV
fcis-32254	123	23	enhancing	enhance	VERB
fcis-32254	123	24	the	the	DET
fcis-32254	123	25	algorithm	algorithm	NOUN
fcis-32254	123	26	's	's	PART
fcis-32254	123	27	practicality	practicality	NOUN
fcis-32254	123	28	.	.	PUNCT
fcis-32254	124	1	3.3	3.3	NUM
fcis-32254	124	2	.	.	PUNCT
fcis-32254	125	1	analysis	analysis	NOUN
fcis-32254	125	2	of	of	ADP
fcis-32254	125	3	experimental	experimental	ADJ
fcis-32254	125	4	results	result	NOUN
fcis-32254	125	5	table	table	NOUN
fcis-32254	125	6	1	1	NUM
fcis-32254	125	7	.	.	PUNCT
fcis-32254	126	1	three	three	NUM
fcis-32254	126	2	scheme	scheme	NOUN
fcis-32254	126	3	comparing	compare	VERB
fcis-32254	126	4	col20	col20	NOUN
fcis-32254	126	5	pie	pie	NOUN
fcis-32254	126	6	minst	minst	NOUN
fcis-32254	126	7	ac	ac	PROPN
fcis-32254	127	1	mih	mih	PROPN
fcis-32254	127	2	ac	ac	PROPN
fcis-32254	128	1	mih	mih	PROPN
fcis-32254	128	2	ac	ac	PROPN
fcis-32254	129	1	mih	mih	PROPN
fcis-32254	130	1	k	k	X
fcis-32254	130	2	-	-	PUNCT
fcis-32254	130	3	means	mean	VERB
fcis-32254	130	4	0.6920	0.6920	NUM
fcis-32254	130	5	0.7455	0.7455	NUM
fcis-32254	130	6	0.3140	0.3140	NUM
fcis-32254	130	7	0.3765	0.3765	NUM
fcis-32254	130	8	0.6400	0.6400	NUM
fcis-32254	130	9	0.5449	0.5449	NUM
fcis-32254	130	10	kpca	kpca	NOUN
fcis-32254	130	11	0.6979	0.6979	NUM
fcis-32254	130	12	0.7506	0.7506	NUM
fcis-32254	130	13	0.3162	0.3162	NUM
fcis-32254	130	14	0.3801	0.3801	NUM
fcis-32254	130	15	0.6461	0.6461	NUM
fcis-32254	130	16	0.5548	0.5548	NUM
fcis-32254	130	17	nmf	nmf	NOUN
fcis-32254	130	18	0.6840	0.6840	NUM
fcis-32254	130	19	0.7129	0.7129	NUM
fcis-32254	130	20	0.4904	0.4904	NUM
fcis-32254	130	21	0.5712	0.5712	NUM
fcis-32254	130	22	0.5992	0.5992	NUM
fcis-32254	130	23	0.4698	0.4698	NUM
fcis-32254	130	24	gnmf	gnmf	NOUN
fcis-32254	130	25	0.8356	0.8356	NUM
fcis-32254	130	26	0.8910	0.8910	NUM
fcis-32254	130	27	0.8332	0.8332	NUM
fcis-32254	130	28	0.8840	0.8840	NUM
fcis-32254	130	29	0.6830	0.6830	NUM
fcis-32254	130	30	0.6588	0.6588	NUM
fcis-32254	130	31	gdnmf	gdnmf	NOUN
fcis-32254	130	32	0.8813	0.8813	NUM
fcis-32254	130	33	0.9267	0.9267	NUM
fcis-32254	130	34	0.9389	0.9389	NUM
fcis-32254	130	35	0.9498	0.9498	NUM
fcis-32254	130	36	0.9527	0.9527	NUM
fcis-32254	130	37	0.8736	0.8736	NUM
fcis-32254	130	38	kgdnmf	kgdnmf	NOUN
fcis-32254	130	39	0.8756	0.8756	NUM
fcis-32254	130	40	0.9250	0.9250	NUM
fcis-32254	130	41	0.9344	0.9344	NUM
fcis-32254	130	42	0.9470	0.9470	NUM
fcis-32254	130	43	0.8914	0.8914	NUM
fcis-32254	130	44	0.8133	0.8133	NUM
fcis-32254	130	45	skgdnmf	skgdnmf	VERB
fcis-32254	130	46	0.9584	0.9584	NUM
fcis-32254	130	47	0.9534	0.9534	NUM
fcis-32254	130	48	0.9833	0.9833	NUM
fcis-32254	130	49	0.9724	0.9724	NUM
fcis-32254	130	50	0.9543	0.9543	NUM
fcis-32254	130	51	0.8769	0.8769	NUM
fcis-32254	130	52	table	table	NOUN
fcis-32254	130	53	1	1	NUM
fcis-32254	130	54	presents	present	VERB
fcis-32254	130	55	the	the	DET
fcis-32254	130	56	experimental	experimental	ADJ
fcis-32254	130	57	results	result	NOUN
fcis-32254	130	58	of	of	ADP
fcis-32254	130	59	the	the	DET
fcis-32254	130	60	six	six	NUM
fcis-32254	130	61	algorithms	algorithm	NOUN
fcis-32254	130	62	on	on	ADP
fcis-32254	130	63	the	the	DET
fcis-32254	130	64	three	three	NUM
fcis-32254	130	65	datasets	dataset	NOUN
fcis-32254	130	66	.	.	PUNCT
fcis-32254	131	1	among	among	ADP
fcis-32254	131	2	them	they	PRON
fcis-32254	131	3	,	,	PUNCT
fcis-32254	131	4	the	the	DET
fcis-32254	131	5	results	result	NOUN
fcis-32254	131	6	for	for	ADP
fcis-32254	131	7	the	the	DET
fcis-32254	131	8	coil20	coil20	NOUN
fcis-32254	131	9	and	and	CCONJ
fcis-32254	131	10	pie	pie	NOUN
fcis-32254	131	11	datasets	dataset	NOUN
fcis-32254	131	12	are	be	AUX
fcis-32254	131	13	the	the	DET
fcis-32254	131	14	average	average	ADJ
fcis-32254	131	15	values	value	NOUN
fcis-32254	131	16	when	when	SCONJ
fcis-32254	131	17	the	the	DET
fcis-32254	131	18	number	number	NOUN
fcis-32254	131	19	of	of	ADP
fcis-32254	131	20	clusters	cluster	NOUN
fcis-32254	131	21	ranges	range	VERB
fcis-32254	131	22	from	from	ADP
fcis-32254	131	23	2	2	NUM
fcis-32254	131	24	to	to	ADP
fcis-32254	131	25	20	20	NUM
fcis-32254	131	26	(	(	PUNCT
fcis-32254	131	27	with	with	ADP
fcis-32254	131	28	a	a	DET
fcis-32254	131	29	step	step	NOUN
fcis-32254	131	30	size	size	NOUN
fcis-32254	131	31	of	of	ADP
fcis-32254	131	32	2	2	NUM
fcis-32254	131	33	)	)	PUNCT
fcis-32254	131	34	;	;	PUNCT
fcis-32254	131	35	the	the	DET
fcis-32254	131	36	results	result	NOUN
fcis-32254	131	37	for	for	ADP
fcis-32254	131	38	the	the	DET
fcis-32254	131	39	mnist	mnist	NOUN
fcis-32254	131	40	dataset	dataset	PROPN
fcis-32254	131	41	are	be	AUX
fcis-32254	131	42	the	the	DET
fcis-32254	131	43	average	average	ADJ
fcis-32254	131	44	values	value	NOUN
fcis-32254	131	45	when	when	SCONJ
fcis-32254	131	46	the	the	DET
fcis-32254	131	47	number	number	NOUN
fcis-32254	131	48	of	of	ADP
fcis-32254	131	49	clusters	cluster	NOUN
fcis-32254	131	50	ranges	range	VERB
fcis-32254	131	51	from	from	ADP
fcis-32254	131	52	2	2	NUM
fcis-32254	131	53	to	to	ADP
fcis-32254	131	54	10	10	NUM
fcis-32254	131	55	.	.	PUNCT
fcis-32254	132	1	it	it	PRON
fcis-32254	132	2	can	can	AUX
fcis-32254	132	3	be	be	AUX
fcis-32254	132	4	seen	see	VERB
fcis-32254	132	5	that	that	SCONJ
fcis-32254	132	6	23	23	NUM
fcis-32254	132	7	the	the	DET
fcis-32254	132	8	proposed	propose	VERB
fcis-32254	132	9	algorithm	algorithm	NOUN
fcis-32254	132	10	exhibits	exhibit	VERB
fcis-32254	132	11	significant	significant	ADJ
fcis-32254	132	12	performance	performance	NOUN
fcis-32254	132	13	advantages	advantage	NOUN
fcis-32254	132	14	in	in	ADP
fcis-32254	132	15	the	the	DET
fcis-32254	132	16	clustering	clustering	ADJ
fcis-32254	132	17	tasks	task	NOUN
fcis-32254	132	18	of	of	ADP
fcis-32254	132	19	the	the	DET
fcis-32254	132	20	three	three	NUM
fcis-32254	132	21	datasets	dataset	NOUN
fcis-32254	132	22	,	,	PUNCT
fcis-32254	132	23	achieving	achieve	VERB
fcis-32254	132	24	higher	high	ADJ
fcis-32254	132	25	accuracy	accuracy	NOUN
fcis-32254	132	26	and	and	CCONJ
fcis-32254	132	27	mutual	mutual	ADJ
fcis-32254	132	28	information	information	NOUN
fcis-32254	132	29	(	(	PUNCT
fcis-32254	132	30	mi	mi	NOUN
fcis-32254	132	31	)	)	PUNCT
fcis-32254	132	32	.	.	PUNCT
fcis-32254	133	1	this	this	PRON
fcis-32254	133	2	indicates	indicate	VERB
fcis-32254	133	3	that	that	SCONJ
fcis-32254	133	4	the	the	DET
fcis-32254	133	5	algorithm	algorithm	NOUN
fcis-32254	133	6	can	can	AUX
fcis-32254	133	7	stabilize	stabilize	VERB
fcis-32254	133	8	the	the	DET
fcis-32254	133	9	discriminative	discriminative	NOUN
fcis-32254	133	10	structure	structure	NOUN
fcis-32254	133	11	of	of	ADP
fcis-32254	133	12	manifold	manifold	ADJ
fcis-32254	133	13	data	datum	NOUN
fcis-32254	133	14	across	across	ADP
fcis-32254	133	15	different	different	ADJ
fcis-32254	133	16	image	image	NOUN
fcis-32254	133	17	datasets	dataset	NOUN
fcis-32254	133	18	,	,	PUNCT
fcis-32254	133	19	proving	prove	VERB
fcis-32254	133	20	its	its	PRON
fcis-32254	133	21	effectiveness	effectiveness	NOUN
fcis-32254	133	22	in	in	ADP
fcis-32254	133	23	aspects	aspect	NOUN
fcis-32254	133	24	such	such	ADJ
fcis-32254	133	25	as	as	ADP
fcis-32254	133	26	preserving	preserve	VERB
fcis-32254	133	27	manifold	manifold	ADJ
fcis-32254	133	28	structures	structure	NOUN
fcis-32254	133	29	and	and	CCONJ
fcis-32254	133	30	suppressing	suppress	VERB
fcis-32254	133	31	noise	noise	NOUN
fcis-32254	133	32	interference	interference	NOUN
fcis-32254	133	33	.	.	PUNCT
fcis-32254	134	1	in	in	ADP
fcis-32254	134	2	conclusion	conclusion	NOUN
fcis-32254	134	3	,	,	PUNCT
fcis-32254	134	4	the	the	DET
fcis-32254	134	5	proposed	propose	VERB
fcis-32254	134	6	algorithm	algorithm	NOUN
fcis-32254	134	7	provides	provide	VERB
fcis-32254	134	8	a	a	DET
fcis-32254	134	9	new	new	ADJ
fcis-32254	134	10	solution	solution	NOUN
fcis-32254	134	11	for	for	ADP
fcis-32254	134	12	semi	semi	ADJ
fcis-32254	134	13	-	-	ADJ
fcis-32254	134	14	supervised	supervised	ADJ
fcis-32254	134	15	clustering	clustering	ADJ
fcis-32254	134	16	tasks	task	NOUN
fcis-32254	134	17	,	,	PUNCT
fcis-32254	134	18	is	be	AUX
fcis-32254	134	19	more	more	ADV
fcis-32254	134	20	suitable	suitable	ADJ
fcis-32254	134	21	for	for	ADP
fcis-32254	134	22	feature	feature	NOUN
fcis-32254	134	23	learning	learning	NOUN
fcis-32254	134	24	and	and	CCONJ
fcis-32254	134	25	structure	structure	NOUN
fcis-32254	134	26	mining	mining	NOUN
fcis-32254	134	27	of	of	ADP
fcis-32254	134	28	highdimensional	highdimensional	PROPN
fcis-32254	134	29	nonlinear	nonlinear	ADJ
fcis-32254	134	30	data	datum	NOUN
fcis-32254	134	31	,	,	PUNCT
fcis-32254	134	32	and	and	CCONJ
fcis-32254	134	33	has	have	VERB
fcis-32254	134	34	important	important	ADJ
fcis-32254	134	35	theoretical	theoretical	ADJ
fcis-32254	134	36	reference	reference	NOUN
fcis-32254	134	37	value	value	NOUN
fcis-32254	134	38	and	and	CCONJ
fcis-32254	134	39	practical	practical	ADJ
fcis-32254	134	40	application	application	NOUN
fcis-32254	134	41	prospects	prospect	NOUN
fcis-32254	134	42	.	.	PUNCT
fcis-32254	135	1	4	4	X
fcis-32254	135	2	.	.	X
fcis-32254	135	3	conclusion	conclusion	NOUN
fcis-32254	135	4	this	this	DET
fcis-32254	135	5	paper	paper	NOUN
fcis-32254	135	6	proposes	propose	VERB
fcis-32254	135	7	the	the	DET
fcis-32254	135	8	sparse	sparse	ADJ
fcis-32254	135	9	kernel	kernel	NOUN
fcis-32254	135	10	graph	graph	VERB
fcis-32254	135	11	discriminative	discriminative	ADJ
fcis-32254	135	12	non	non	ADJ
fcis-32254	135	13	-	-	ADJ
fcis-32254	135	14	negative	negative	ADJ
fcis-32254	135	15	matrix	matrix	NOUN
fcis-32254	135	16	factorization	factorization	NOUN
fcis-32254	135	17	(	(	PUNCT
fcis-32254	135	18	skgdnmf	skgdnmf	NOUN
fcis-32254	135	19	)	)	PUNCT
fcis-32254	135	20	algorithm	algorithm	NOUN
fcis-32254	135	21	,	,	PUNCT
fcis-32254	135	22	which	which	PRON
fcis-32254	135	23	integrates	integrate	VERB
fcis-32254	135	24	semi	semi	ADJ
fcis-32254	135	25	-	-	ADJ
fcis-32254	135	26	supervised	supervised	ADJ
fcis-32254	135	27	learning	learning	NOUN
fcis-32254	135	28	,	,	PUNCT
fcis-32254	135	29	sparse	sparse	ADJ
fcis-32254	135	30	constraints	constraint	NOUN
fcis-32254	135	31	,	,	PUNCT
fcis-32254	135	32	label	label	NOUN
fcis-32254	135	33	information	information	NOUN
fcis-32254	135	34	,	,	PUNCT
fcis-32254	135	35	capture	capture	NOUN
fcis-32254	135	36	of	of	ADP
fcis-32254	135	37	the	the	DET
fcis-32254	135	38	inherent	inherent	ADJ
fcis-32254	135	39	geometric	geometric	ADJ
fcis-32254	135	40	structure	structure	NOUN
fcis-32254	135	41	of	of	ADP
fcis-32254	135	42	data	datum	NOUN
fcis-32254	135	43	,	,	PUNCT
fcis-32254	135	44	and	and	CCONJ
fcis-32254	135	45	nonlinear	nonlinear	ADJ
fcis-32254	135	46	data	datum	NOUN
fcis-32254	135	47	processing	processing	NOUN
fcis-32254	135	48	capabilities	capability	NOUN
fcis-32254	135	49	into	into	ADP
fcis-32254	135	50	a	a	DET
fcis-32254	135	51	comprehensive	comprehensive	ADJ
fcis-32254	135	52	model	model	NOUN
fcis-32254	135	53	.	.	PUNCT
fcis-32254	136	1	experimental	experimental	ADJ
fcis-32254	136	2	results	result	NOUN
fcis-32254	136	3	show	show	VERB
fcis-32254	136	4	that	that	SCONJ
fcis-32254	136	5	compared	compare	VERB
fcis-32254	136	6	with	with	ADP
fcis-32254	136	7	existing	exist	VERB
fcis-32254	136	8	algorithms	algorithm	NOUN
fcis-32254	136	9	,	,	PUNCT
fcis-32254	136	10	the	the	DET
fcis-32254	136	11	proposed	propose	VERB
fcis-32254	136	12	algorithm	algorithm	NOUN
fcis-32254	136	13	exhibits	exhibit	VERB
fcis-32254	136	14	significant	significant	ADJ
fcis-32254	136	15	comprehensive	comprehensive	ADJ
fcis-32254	136	16	advantages	advantage	NOUN
fcis-32254	136	17	in	in	ADP
fcis-32254	136	18	image	image	NOUN
fcis-32254	136	19	clustering	clustering	NOUN
fcis-32254	136	20	tasks	task	NOUN
fcis-32254	136	21	,	,	PUNCT
fcis-32254	136	22	with	with	ADP
fcis-32254	136	23	improvements	improvement	NOUN
fcis-32254	136	24	in	in	ADP
fcis-32254	136	25	clustering	clustering	ADJ
fcis-32254	136	26	accuracy	accuracy	NOUN
fcis-32254	136	27	,	,	PUNCT
fcis-32254	136	28	robustness	robustness	NOUN
fcis-32254	136	29	,	,	PUNCT
fcis-32254	136	30	and	and	CCONJ
fcis-32254	136	31	sparsity	sparsity	NOUN
fcis-32254	136	32	.	.	PUNCT
fcis-32254	137	1	through	through	ADP
fcis-32254	137	2	its	its	PRON
fcis-32254	137	3	innovative	innovative	ADJ
fcis-32254	137	4	design	design	NOUN
fcis-32254	137	5	,	,	PUNCT
fcis-32254	137	6	the	the	DET
fcis-32254	137	7	algorithm	algorithm	NOUN
fcis-32254	137	8	effectively	effectively	ADV
fcis-32254	137	9	overcomes	overcome	VERB
fcis-32254	137	10	the	the	DET
fcis-32254	137	11	limitations	limitation	NOUN
fcis-32254	137	12	of	of	ADP
fcis-32254	137	13	traditional	traditional	ADJ
fcis-32254	137	14	methods	method	NOUN
fcis-32254	137	15	,	,	PUNCT
fcis-32254	137	16	provides	provide	VERB
fcis-32254	137	17	a	a	DET
fcis-32254	137	18	more	more	ADV
fcis-32254	137	19	effective	effective	ADJ
fcis-32254	137	20	solution	solution	NOUN
fcis-32254	137	21	for	for	ADP
fcis-32254	137	22	the	the	DET
fcis-32254	137	23	clustering	clustering	ADJ
fcis-32254	137	24	analysis	analysis	NOUN
fcis-32254	137	25	of	of	ADP
fcis-32254	137	26	complex	complex	ADJ
fcis-32254	137	27	image	image	NOUN
fcis-32254	137	28	data	datum	NOUN
fcis-32254	137	29	,	,	PUNCT
fcis-32254	137	30	and	and	CCONJ
fcis-32254	137	31	holds	hold	VERB
fcis-32254	137	32	important	important	ADJ
fcis-32254	137	33	theoretical	theoretical	ADJ
fcis-32254	137	34	significance	significance	NOUN
fcis-32254	137	35	and	and	CCONJ
fcis-32254	137	36	promising	promise	VERB
fcis-32254	137	37	application	application	NOUN
fcis-32254	137	38	prospects	prospect	NOUN
fcis-32254	137	39	.	.	PUNCT
fcis-32254	138	1	references	reference	NOUN
fcis-32254	138	2	[	[	X
fcis-32254	138	3	1	1	NUM
fcis-32254	138	4	]	]	X
fcis-32254	138	5	cai	cai	X
fcis-32254	139	1	d	d	X
fcis-32254	139	2	,	,	PUNCT
fcis-32254	139	3	he	he	PRON
fcis-32254	139	4	x	x	NOUN
fcis-32254	139	5	,	,	PUNCT
fcis-32254	139	6	wu	wu	PROPN
fcis-32254	139	7	x	x	PROPN
fcis-32254	139	8	,	,	PUNCT
fcis-32254	139	9	et	et	PROPN
fcis-32254	139	10	al	al	PROPN
fcis-32254	139	11	.	.	PUNCT
fcis-32254	139	12	non	non	ADJ
fcis-32254	139	13	-	-	ADJ
fcis-32254	139	14	negative	negative	ADJ
fcis-32254	139	15	matrix	matrix	NOUN
fcis-32254	139	16	factorization	factorization	NOUN
fcis-32254	139	17	on	on	ADP
fcis-32254	139	18	manifold[c]//2008	manifold[c]//2008	PROPN
fcis-32254	139	19	eighth	eighth	ADJ
fcis-32254	139	20	ieee	ieee	PROPN
fcis-32254	139	21	international	international	PROPN
fcis-32254	139	22	conference	conference	NOUN
fcis-32254	139	23	on	on	ADP
fcis-32254	139	24	data	datum	NOUN
fcis-32254	139	25	mining	mining	NOUN
fcis-32254	139	26	.	.	PUNCT
fcis-32254	140	1	ieee	ieee	PROPN
fcis-32254	140	2	,	,	PUNCT
fcis-32254	140	3	2008	2008	NUM
fcis-32254	140	4	:	:	PUNCT
fcis-32254	140	5	63	63	NUM
fcis-32254	140	6	-	-	SYM
fcis-32254	140	7	72	72	NUM
fcis-32254	140	8	.	.	PUNCT
fcis-32254	141	1	[	[	X
fcis-32254	141	2	2	2	NUM
fcis-32254	141	3	]	]	PUNCT
fcis-32254	141	4	babaee	babaee	PROPN
fcis-32254	141	5	m	m	PROPN
fcis-32254	141	6	,	,	PUNCT
fcis-32254	141	7	tsoukalas	tsoukalas	PROPN
fcis-32254	141	8	s	s	PART
fcis-32254	141	9	,	,	PUNCT
fcis-32254	141	10	babaee	babaee	PROPN
fcis-32254	141	11	m	m	PROPN
fcis-32254	141	12	,	,	PUNCT
fcis-32254	141	13	et	et	PROPN
fcis-32254	141	14	al	al	PROPN
fcis-32254	141	15	.	.	PROPN
fcis-32254	141	16	discriminative	discriminative	VERB
fcis-32254	141	17	nonnegative	nonnegative	ADJ
fcis-32254	141	18	matrix	matrix	NOUN
fcis-32254	141	19	factorization	factorization	NOUN
fcis-32254	141	20	for	for	ADP
fcis-32254	141	21	dimensionality	dimensionality	NOUN
fcis-32254	141	22	reduction	reduction	NOUN
fcis-32254	142	1	[	[	X
fcis-32254	142	2	j	j	X
fcis-32254	142	3	]	]	X
fcis-32254	142	4	.	.	PUNCT
fcis-32254	143	1	neurocomputing	neurocomputing	NOUN
fcis-32254	143	2	,	,	PUNCT
fcis-32254	143	3	2016	2016	NUM
fcis-32254	143	4	,	,	PUNCT
fcis-32254	143	5	173	173	NUM
fcis-32254	143	6	:	:	PUNCT
fcis-32254	143	7	212	212	NUM
fcis-32254	143	8	-	-	SYM
fcis-32254	143	9	223	223	NUM
fcis-32254	143	10	.	.	PUNCT
fcis-32254	144	1	[	[	X
fcis-32254	144	2	3	3	X
fcis-32254	144	3	]	]	X
fcis-32254	144	4	zhang	zhang	PROPN
fcis-32254	144	5	d	d	PROPN
fcis-32254	144	6	,	,	PUNCT
fcis-32254	144	7	zhou	zhou	PROPN
fcis-32254	144	8	z	z	PROPN
fcis-32254	144	9	h	h	PROPN
fcis-32254	144	10	,	,	PUNCT
fcis-32254	144	11	chen	chen	PROPN
fcis-32254	144	12	s.	s.	PROPN
fcis-32254	144	13	non	non	ADJ
fcis-32254	144	14	-	-	ADJ
fcis-32254	144	15	negative	negative	ADJ
fcis-32254	144	16	matrix	matrix	NOUN
fcis-32254	144	17	factorization	factorization	NOUN
fcis-32254	144	18	on	on	ADP
fcis-32254	144	19	kernels[c]//pricai	kernels[c]//pricai	PROPN
fcis-32254	144	20	2006	2006	NUM
fcis-32254	144	21	:	:	PUNCT
fcis-32254	144	22	trends	trend	NOUN
fcis-32254	144	23	in	in	ADP
fcis-32254	144	24	artificial	artificial	ADJ
fcis-32254	144	25	intelligence	intelligence	NOUN
fcis-32254	144	26	:	:	PUNCT
fcis-32254	144	27	9th	9th	ADJ
fcis-32254	144	28	pacific	pacific	PROPN
fcis-32254	144	29	rim	rim	PROPN
fcis-32254	144	30	international	international	ADJ
fcis-32254	144	31	conference	conference	NOUN
fcis-32254	144	32	on	on	ADP
fcis-32254	144	33	artificial	artificial	ADJ
fcis-32254	144	34	intelligence	intelligence	NOUN
fcis-32254	144	35	guilin	guilin	PROPN
fcis-32254	144	36	,	,	PUNCT
fcis-32254	144	37	china	china	PROPN
fcis-32254	144	38	,	,	PUNCT
fcis-32254	144	39	august	august	PROPN
fcis-32254	144	40	7	7	NUM
fcis-32254	144	41	-	-	SYM
fcis-32254	144	42	11	11	NUM
fcis-32254	144	43	,	,	PUNCT
fcis-32254	144	44	2006	2006	NUM
fcis-32254	144	45	proceedings	proceeding	NOUN
fcis-32254	144	46	9	9	NUM
fcis-32254	144	47	.	.	PUNCT
fcis-32254	144	48	springer	springer	PROPN
fcis-32254	144	49	berlin	berlin	PROPN
fcis-32254	144	50	heidelberg	heidelberg	PROPN
fcis-32254	144	51	,	,	PUNCT
fcis-32254	144	52	2006	2006	NUM
fcis-32254	144	53	:	:	PUNCT
fcis-32254	144	54	404	404	NUM
fcis-32254	144	55	-	-	SYM
fcis-32254	144	56	412	412	NUM
fcis-32254	144	57	.	.	PUNCT
fcis-32254	145	1	[	[	X
fcis-32254	145	2	4	4	X
fcis-32254	145	3	]	]	X
fcis-32254	145	4	sun	sun	PROPN
fcis-32254	145	5	l	l	PROPN
fcis-32254	145	6	,	,	PUNCT
fcis-32254	145	7	zhao	zhao	PROPN
fcis-32254	145	8	k	k	PROPN
fcis-32254	145	9	,	,	PUNCT
fcis-32254	145	10	han	han	PROPN
fcis-32254	145	11	c	c	PROPN
fcis-32254	145	12	y	y	PROPN
fcis-32254	145	13	,	,	PUNCT
fcis-32254	145	14	et	et	PROPN
fcis-32254	145	15	al	al	PROPN
fcis-32254	145	16	.	.	PUNCT
fcis-32254	145	17	enhancing	enhance	VERB
fcis-32254	145	18	hyperspectral	hyperspectral	ADJ
fcis-32254	145	19	unmixing	unmixing	NOUN
fcis-32254	145	20	with	with	ADP
fcis-32254	145	21	two	two	NUM
fcis-32254	145	22	-	-	PUNCT
fcis-32254	145	23	stage	stage	NOUN
fcis-32254	145	24	multiplicative	multiplicative	ADJ
fcis-32254	145	25	update	update	NOUN
fcis-32254	145	26	nonnegative	nonnegative	ADJ
fcis-32254	145	27	matrix	matrix	NOUN
fcis-32254	145	28	factorization[j	factorization[j	NOUN
fcis-32254	145	29	]	]	PUNCT
fcis-32254	145	30	.	.	PUNCT
fcis-32254	146	1	ieee	ieee	NOUN
fcis-32254	146	2	access	access	NOUN
fcis-32254	146	3	,	,	PUNCT
fcis-32254	146	4	2019,7:171023	2019,7:171023	NUM
fcis-32254	146	5	-	-	SYM
fcis-32254	146	6	171031	171031	NUM
fcis-32254	146	7	.	.	PUNCT
fcis-32254	147	1	[	[	X
fcis-32254	147	2	5	5	NUM
fcis-32254	147	3	]	]	PUNCT
fcis-32254	147	4	ashabi	ashabi	NOUN
fcis-32254	147	5	a	a	NOUN
fcis-32254	147	6	,	,	PUNCT
fcis-32254	147	7	sahibuddin	sahibuddin	PROPN
fcis-32254	147	8	s	s	PROPN
fcis-32254	147	9	b	b	PROPN
fcis-32254	147	10	,	,	PUNCT
fcis-32254	147	11	salkhordeh	salkhordeh	PROPN
fcis-32254	147	12	haghighi	haghighi	PROPN
fcis-32254	147	13	m.	m.	PROPN
fcis-32254	147	14	the	the	DET
fcis-32254	147	15	systematic	systematic	ADJ
fcis-32254	147	16	review	review	NOUN
fcis-32254	147	17	of	of	ADP
fcis-32254	147	18	k	k	NOUN
fcis-32254	147	19	-	-	PUNCT
fcis-32254	147	20	means	means	NOUN
fcis-32254	147	21	clustering	cluster	VERB
fcis-32254	147	22	algorithm[c]//	algorithm[c]//	PROPN
fcis-32254	147	23	proceedings	proceeding	NOUN
fcis-32254	147	24	of	of	ADP
fcis-32254	147	25	the	the	DET
fcis-32254	147	26	2020	2020	NUM
fcis-32254	147	27	9th	9th	ADJ
fcis-32254	147	28	international	international	ADJ
fcis-32254	147	29	conference	conference	NOUN
fcis-32254	147	30	on	on	ADP
fcis-32254	147	31	networks	network	NOUN
fcis-32254	147	32	,	,	PUNCT
fcis-32254	147	33	communication	communication	NOUN
fcis-32254	147	34	and	and	CCONJ
fcis-32254	147	35	computing	computing	NOUN
fcis-32254	147	36	.	.	PUNCT
fcis-32254	148	1	2020	2020	NUM
fcis-32254	148	2	:	:	PUNCT
fcis-32254	148	3	13	13	NUM
fcis-32254	148	4	-	-	SYM
fcis-32254	148	5	18	18	NUM
fcis-32254	148	6	.	.	PUNCT
fcis-32254	149	1	[	[	X
fcis-32254	149	2	6	6	NUM
fcis-32254	149	3	]	]	PUNCT
fcis-32254	149	4	schölkopf	schölkopf	ADJ
fcis-32254	149	5	b	b	PROPN
fcis-32254	149	6	,	,	PUNCT
fcis-32254	149	7	smola	smola	ADP
fcis-32254	149	8	a	a	DET
fcis-32254	149	9	,	,	PUNCT
fcis-32254	149	10	müller	müller	PROPN
fcis-32254	149	11	,	,	PUNCT
fcis-32254	149	12	k.nonlinear	k.nonlinear	ADJ
fcis-32254	149	13	component	component	NOUN
fcis-32254	149	14	analysis	analysis	NOUN
fcis-32254	149	15	as	as	ADP
fcis-32254	149	16	a	a	DET
fcis-32254	149	17	kernel	kernel	PROPN
fcis-32254	149	18	eigenvalue	eigenvalue	PROPN
fcis-32254	149	19	problem[j].neural	problem[j].neural	ADJ
fcis-32254	149	20	computation	computation	NOUN
fcis-32254	149	21	,	,	PUNCT
fcis-32254	149	22	1998	1998	NUM
fcis-32254	149	23	,	,	PUNCT
fcis-32254	149	24	10(5):1299	10(5):1299	NUM
fcis-32254	149	25	-	-	SYM
fcis-32254	149	26	1319	1319	NUM
fcis-32254	149	27	.	.	PUNCT
fcis-32254	150	1	[	[	X
fcis-32254	150	2	7	7	X
fcis-32254	150	3	]	]	X
fcis-32254	150	4	lee	lee	PROPN
fcis-32254	150	5	d	d	PROPN
fcis-32254	150	6	d	d	PROPN
fcis-32254	150	7	,	,	PUNCT
fcis-32254	150	8	seung	seung	PROPN
fcis-32254	150	9	h	h	PROPN
fcis-32254	150	10	s.	s.	PROPN
fcis-32254	150	11	learning	learn	VERB
fcis-32254	150	12	the	the	DET
fcis-32254	150	13	parts	part	NOUN
fcis-32254	150	14	of	of	ADP
fcis-32254	150	15	objects	object	NOUN
fcis-32254	150	16	by	by	ADP
fcis-32254	150	17	nonnegative	nonnegative	ADJ
fcis-32254	150	18	matrix	matrix	NOUN
fcis-32254	150	19	factorization[j	factorization[j	NOUN
fcis-32254	150	20	]	]	PUNCT
fcis-32254	150	21	.	.	PUNCT
fcis-32254	151	1	nature	nature	NOUN
fcis-32254	151	2	,	,	PUNCT
fcis-32254	151	3	1999	1999	NUM
fcis-32254	151	4	,	,	PUNCT
fcis-32254	151	5	401(6755):788791	401(6755):788791	NOUN
fcis-32254	151	6	.	.	PUNCT
fcis-32254	152	1	[	[	X
fcis-32254	152	2	8	8	NUM
fcis-32254	152	3	]	]	X
fcis-32254	152	4	li	li	PROPN
fcis-32254	152	5	h	h	PROPN
fcis-32254	152	6	,	,	PUNCT
fcis-32254	152	7	zhang	zhang	PROPN
fcis-32254	152	8	j	j	PROPN
fcis-32254	152	9	,	,	PUNCT
fcis-32254	152	10	shi	shi	PROPN
fcis-32254	152	11	g	g	PROPN
fcis-32254	152	12	,	,	PUNCT
fcis-32254	152	13	et	et	PROPN
fcis-32254	152	14	al	al	PROPN
fcis-32254	152	15	.	.	PROPN
fcis-32254	152	16	graph	graph	NOUN
fcis-32254	152	17	-	-	PUNCT
fcis-32254	152	18	based	base	VERB
fcis-32254	152	19	discriminative	discriminative	NOUN
fcis-32254	152	20	nonnegative	nonnegative	ADJ
fcis-32254	152	21	matrix	matrix	NOUN
fcis-32254	152	22	factorization	factorization	NOUN
fcis-32254	152	23	with	with	ADP
fcis-32254	152	24	label	label	NOUN
fcis-32254	152	25	information[j	information[j	PROPN
fcis-32254	152	26	]	]	PUNCT
fcis-32254	152	27	.	.	PUNCT
fcis-32254	153	1	neurocomputing	neurocomputing	NOUN
fcis-32254	153	2	,	,	PUNCT
fcis-32254	153	3	2017	2017	NUM
fcis-32254	153	4	,	,	PUNCT
fcis-32254	153	5	266	266	NUM
fcis-32254	153	6	:	:	SYM
fcis-32254	153	7	91	91	NUM
fcis-32254	153	8	-	-	SYM
fcis-32254	153	9	100	100	NUM
fcis-32254	153	10	.	.	PUNCT
fcis-32254	154	1	[	[	X
fcis-32254	154	2	9	9	NUM
fcis-32254	154	3	]	]	SYM
fcis-32254	154	4	li	li	NOUN
fcis-32254	154	5	x	x	SYM
fcis-32254	154	6	l	l	PROPN
fcis-32254	154	7	,	,	PUNCT
fcis-32254	154	8	zhang	zhang	PROPN
fcis-32254	154	9	y.	y.	PROPN
fcis-32254	154	10	kernel	kernel	PROPN
fcis-32254	154	11	-	-	PUNCT
fcis-32254	154	12	based	base	VERB
fcis-32254	154	13	discriminative	discriminative	NOUN
fcis-32254	154	14	graph	graph	NOUN
fcis-32254	154	15	regularized	regularize	VERB
fcis-32254	154	16	non	non	ADJ
fcis-32254	154	17	-	-	ADJ
fcis-32254	154	18	negative	negative	ADJ
fcis-32254	154	19	matrix	matrix	NOUN
fcis-32254	154	20	factorization	factorization	NOUN
fcis-32254	154	21	[	[	X
fcis-32254	154	22	in	in	ADP
fcis-32254	154	23	chinese][j	chinese][j	PROPN
fcis-32254	154	24	]	]	PUNCT
fcis-32254	154	25	.	.	PUNCT
fcis-32254	155	1	journal	journal	PROPN
fcis-32254	155	2	of	of	ADP
fcis-32254	155	3	computer	computer	NOUN
fcis-32254	155	4	science	science	NOUN
fcis-32254	155	5	and	and	CCONJ
fcis-32254	155	6	exploration	exploration	NOUN
fcis-32254	155	7	,	,	PUNCT
fcis-32254	155	8	2020,14	2020,14	NUM
fcis-32254	155	9	(	(	PUNCT
fcis-32254	155	10	11	11	NUM
fcis-32254	155	11	):	):	PUNCT
fcis-32254	155	12	1899	1899	NUM
fcis-32254	155	13	-	-	SYM
fcis-32254	155	14	1907	1907	NUM
fcis-32254	155	15	.	.	PUNCT
fcis-32254	156	1	[	[	X
fcis-32254	156	2	10	10	NUM
fcis-32254	156	3	]	]	X
fcis-32254	156	4	ng	ng	PROPN
fcis-32254	156	5	a	a	DET
fcis-32254	156	6	y	y	PROPN
fcis-32254	156	7	,	,	PUNCT
fcis-32254	156	8	jordan	jordan	PROPN
fcis-32254	156	9	m	m	PROPN
fcis-32254	156	10	i	i	PROPN
fcis-32254	156	11	,	,	PUNCT
fcis-32254	156	12	weiss	weiss	VERB
fcis-32254	156	13	y	y	PROPN
fcis-32254	156	14	.on	.on	PROPN
fcis-32254	156	15	spectral	spectral	ADJ
fcis-32254	156	16	clustering	clustering	NOUN
fcis-32254	156	17	:	:	PUNCT
fcis-32254	156	18	analysis	analysis	NOUN
fcis-32254	156	19	and	and	CCONJ
fcis-32254	156	20	an	an	DET
fcis-32254	156	21	algorithm[j].proc	algorithm[j].proc	NOUN
fcis-32254	156	22	nips	nip	NOUN
fcis-32254	156	23	,	,	PUNCT
fcis-32254	156	24	2002	2002	NUM
fcis-32254	156	25	.	.	PUNCT