id	sid	tid	token	lemma	pos
fcis-8581	1	1	frontiers	frontier	NOUN
fcis-8581	1	2	in	in	ADP
fcis-8581	1	3	computing	computing	NOUN
fcis-8581	1	4	and	and	CCONJ
fcis-8581	1	5	intelligent	intelligent	ADJ
fcis-8581	1	6	systems	system	NOUN
fcis-8581	1	7	issn	issn	VERB
fcis-8581	1	8	:	:	PUNCT
fcis-8581	1	9	2832	2832	NUM
fcis-8581	1	10	-	-	SYM
fcis-8581	1	11	6024	6024	NUM
fcis-8581	1	12	|	|	NOUN
fcis-8581	1	13	vol	vol	NOUN
fcis-8581	1	14	.	.	PROPN
fcis-8581	2	1	3	3	NUM
fcis-8581	2	2	,	,	PUNCT
fcis-8581	2	3	no	no	INTJ
fcis-8581	2	4	.	.	NOUN
fcis-8581	2	5	3	3	NUM
fcis-8581	2	6	,	,	PUNCT
fcis-8581	2	7	2023	2023	NUM
fcis-8581	2	8	113	113	NUM
fcis-8581	2	9	algorithm	algorithm	NOUN
fcis-8581	2	10	improvement	improvement	NOUN
fcis-8581	2	11	of	of	ADP
fcis-8581	2	12	movie	movie	NOUN
fcis-8581	2	13	recommendation	recommendation	NOUN
fcis-8581	2	14	system	system	NOUN
fcis-8581	2	15	based	base	VERB
fcis-8581	2	16	on	on	ADP
fcis-8581	2	17	hybrid	hybrid	ADJ
fcis-8581	2	18	recommendation	recommendation	NOUN
fcis-8581	2	19	algorithm	algorithm	NOUN
fcis-8581	2	20	ziyun	ziyun	NOUN
fcis-8581	2	21	liu	liu	PROPN
fcis-8581	2	22	1	1	NUM
fcis-8581	2	23	,	,	PUNCT
fcis-8581	2	24	a	a	PRON
fcis-8581	2	25	,	,	PUNCT
fcis-8581	2	26	*	*	ADJ
fcis-8581	2	27	,	,	PUNCT
fcis-8581	2	28	feiyu	feiyu	PROPN
fcis-8581	2	29	ren	ren	PROPN
fcis-8581	2	30	2	2	NUM
fcis-8581	2	31	,	,	PUNCT
fcis-8581	2	32	b	b	PROPN
fcis-8581	2	33	1	1	NUM
fcis-8581	2	34	xi'an	xi'an	PROPN
fcis-8581	2	35	polytechnic	polytechnic	PROPN
fcis-8581	2	36	university	university	PROPN
fcis-8581	2	37	,	,	PUNCT
fcis-8581	2	38	china	china	PROPN
fcis-8581	2	39	2	2	NUM
fcis-8581	2	40	xinji	xinji	X
fcis-8581	2	41	fani	fani	X
fcis-8581	2	42	leather	leather	PROPN
fcis-8581	2	43	industry	industry	PROPN
fcis-8581	2	44	co.	co.	PROPN
fcis-8581	2	45	,	,	PUNCT
fcis-8581	2	46	china	china	PROPN
fcis-8581	2	47	a	a	X
fcis-8581	2	48	,	,	PUNCT
fcis-8581	2	49	*	*	PUNCT
fcis-8581	2	50	18134282382@163.com	18134282382@163.com	NUM
fcis-8581	2	51	,	,	PUNCT
fcis-8581	2	52	b	b	NOUN
fcis-8581	2	53	1151333999@qq.com	1151333999@qq.com	NUM
fcis-8581	2	54	*	*	PUNCT
fcis-8581	2	55	corresponding	correspond	VERB
fcis-8581	2	56	author	author	NOUN
fcis-8581	2	57	:	:	PUNCT
fcis-8581	2	58	ziyun	ziyun	PROPN
fcis-8581	2	59	liu	liu	PROPN
fcis-8581	2	60	(	(	PUNCT
fcis-8581	2	61	email	email	NOUN
fcis-8581	2	62	:	:	PUNCT
fcis-8581	2	63	18134282382@163.com	18134282382@163.com	NUM
fcis-8581	2	64	)	)	PUNCT
fcis-8581	2	65	abstract	abstract	NOUN
fcis-8581	2	66	:	:	PUNCT
fcis-8581	2	67	in	in	ADP
fcis-8581	2	68	recent	recent	ADJ
fcis-8581	2	69	years	year	NOUN
fcis-8581	2	70	,	,	PUNCT
fcis-8581	2	71	the	the	DET
fcis-8581	2	72	internet	internet	NOUN
fcis-8581	2	73	has	have	AUX
fcis-8581	2	74	developed	develop	VERB
fcis-8581	2	75	rapidly	rapidly	ADV
fcis-8581	2	76	,	,	PUNCT
fcis-8581	2	77	and	and	CCONJ
fcis-8581	2	78	in	in	ADP
fcis-8581	2	79	the	the	DET
fcis-8581	2	80	face	face	NOUN
fcis-8581	2	81	of	of	ADP
fcis-8581	2	82	thousands	thousand	NOUN
fcis-8581	2	83	of	of	ADP
fcis-8581	2	84	data	datum	NOUN
fcis-8581	2	85	and	and	CCONJ
fcis-8581	2	86	information	information	NOUN
fcis-8581	2	87	,	,	PUNCT
fcis-8581	2	88	it	it	PRON
fcis-8581	2	89	has	have	AUX
fcis-8581	2	90	become	become	VERB
fcis-8581	2	91	very	very	ADV
fcis-8581	2	92	critical	critical	ADJ
fcis-8581	2	93	for	for	SCONJ
fcis-8581	2	94	users	user	NOUN
fcis-8581	2	95	to	to	PART
fcis-8581	2	96	find	find	VERB
fcis-8581	2	97	the	the	DET
fcis-8581	2	98	information	information	NOUN
fcis-8581	2	99	that	that	PRON
fcis-8581	2	100	is	be	AUX
fcis-8581	2	101	of	of	ADP
fcis-8581	2	102	high	high	ADJ
fcis-8581	2	103	value	value	NOUN
fcis-8581	2	104	to	to	ADP
fcis-8581	2	105	them	they	PRON
fcis-8581	2	106	in	in	ADP
fcis-8581	2	107	the	the	DET
fcis-8581	2	108	mass	mass	NOUN
fcis-8581	2	109	of	of	ADP
fcis-8581	2	110	information	information	NOUN
fcis-8581	2	111	,	,	PUNCT
fcis-8581	2	112	and	and	CCONJ
fcis-8581	2	113	the	the	DET
fcis-8581	2	114	recommendation	recommendation	NOUN
fcis-8581	2	115	system	system	NOUN
fcis-8581	2	116	is	be	AUX
fcis-8581	2	117	one	one	NUM
fcis-8581	2	118	of	of	ADP
fcis-8581	2	119	the	the	DET
fcis-8581	2	120	most	most	ADV
fcis-8581	2	121	effective	effective	ADJ
fcis-8581	2	122	ways	way	NOUN
fcis-8581	2	123	to	to	PART
fcis-8581	2	124	solve	solve	VERB
fcis-8581	2	125	this	this	DET
fcis-8581	2	126	information	information	NOUN
fcis-8581	2	127	overload	overload	NOUN
fcis-8581	2	128	phenomenon	phenomenon	NOUN
fcis-8581	2	129	.	.	PUNCT
fcis-8581	3	1	in	in	ADP
fcis-8581	3	2	this	this	DET
fcis-8581	3	3	paper	paper	NOUN
fcis-8581	3	4	,	,	PUNCT
fcis-8581	3	5	the	the	DET
fcis-8581	3	6	current	current	ADJ
fcis-8581	3	7	movie	movie	NOUN
fcis-8581	3	8	recommendation	recommendation	NOUN
fcis-8581	3	9	algorithm	algorithm	NOUN
fcis-8581	3	10	is	be	AUX
fcis-8581	3	11	improved	improve	VERB
fcis-8581	3	12	by	by	ADP
fcis-8581	3	13	using	use	VERB
fcis-8581	3	14	an	an	DET
fcis-8581	3	15	item	item	NOUN
fcis-8581	3	16	-	-	PUNCT
fcis-8581	3	17	based	base	VERB
fcis-8581	3	18	collaborative	collaborative	ADJ
fcis-8581	3	19	filtering	filtering	NOUN
fcis-8581	3	20	algorithm	algorithm	NOUN
fcis-8581	3	21	for	for	ADP
fcis-8581	3	22	the	the	DET
fcis-8581	3	23	similarity	similarity	NOUN
fcis-8581	3	24	measure	measure	NOUN
fcis-8581	3	25	of	of	ADP
fcis-8581	3	26	items	item	NOUN
fcis-8581	3	27	in	in	ADP
fcis-8581	3	28	the	the	DET
fcis-8581	3	29	item	item	NOUN
fcis-8581	3	30	-	-	PUNCT
fcis-8581	3	31	based	base	VERB
fcis-8581	3	32	recommendation	recommendation	NOUN
fcis-8581	3	33	process	process	NOUN
fcis-8581	3	34	;	;	PUNCT
fcis-8581	3	35	in	in	ADP
fcis-8581	3	36	the	the	DET
fcis-8581	3	37	recommendation	recommendation	NOUN
fcis-8581	3	38	process	process	NOUN
fcis-8581	3	39	,	,	PUNCT
fcis-8581	3	40	two	two	NUM
fcis-8581	3	41	more	more	ADV
fcis-8581	3	42	applicable	applicable	ADJ
fcis-8581	3	43	recommendation	recommendation	NOUN
fcis-8581	3	44	methods	method	NOUN
fcis-8581	3	45	are	be	AUX
fcis-8581	3	46	considered	consider	VERB
fcis-8581	3	47	:	:	PUNCT
fcis-8581	3	48	collaborative	collaborative	ADJ
fcis-8581	3	49	filtering	filtering	NOUN
fcis-8581	3	50	content	content	NOUN
fcis-8581	3	51	-	-	PUNCT
fcis-8581	3	52	based	base	VERB
fcis-8581	3	53	recommendation	recommendation	NOUN
fcis-8581	3	54	and	and	CCONJ
fcis-8581	3	55	matrix	matrix	NOUN
fcis-8581	3	56	decompositionbased	decompositionbase	VERB
fcis-8581	3	57	recommendation	recommendation	NOUN
fcis-8581	3	58	.	.	PUNCT
fcis-8581	4	1	it	it	PRON
fcis-8581	4	2	saves	save	VERB
fcis-8581	4	3	users	user	NOUN
fcis-8581	4	4	time	time	NOUN
fcis-8581	4	5	in	in	ADP
fcis-8581	4	6	searching	searching	NOUN
fcis-8581	4	7	,	,	PUNCT
fcis-8581	4	8	viewing	view	VERB
fcis-8581	4	9	and	and	CCONJ
fcis-8581	4	10	filtering	filtering	NOUN
fcis-8581	4	11	,	,	PUNCT
fcis-8581	4	12	while	while	SCONJ
fcis-8581	4	13	discovering	discover	VERB
fcis-8581	4	14	information	information	NOUN
fcis-8581	4	15	about	about	ADP
fcis-8581	4	16	their	their	PRON
fcis-8581	4	17	potential	potential	ADJ
fcis-8581	4	18	movie	movie	NOUN
fcis-8581	4	19	preferences	preference	NOUN
fcis-8581	4	20	.	.	PUNCT
fcis-8581	5	1	keywords	keyword	NOUN
fcis-8581	5	2	:	:	PUNCT
fcis-8581	5	3	recommended	recommend	VERB
fcis-8581	5	4	system	system	NOUN
fcis-8581	5	5	;	;	PUNCT
fcis-8581	5	6	similarity	similarity	NOUN
fcis-8581	5	7	algorithm	algorithm	NOUN
fcis-8581	5	8	;	;	PUNCT
fcis-8581	5	9	recommendation	recommendation	NOUN
fcis-8581	5	10	algorithm	algorithm	NOUN
fcis-8581	5	11	based	base	VERB
fcis-8581	5	12	on	on	ADP
fcis-8581	5	13	matrix	matrix	NOUN
fcis-8581	5	14	factorization	factorization	NOUN
fcis-8581	5	15	;	;	PUNCT
fcis-8581	5	16	alternating	alternate	VERB
fcis-8581	5	17	least	least	ADJ
fcis-8581	5	18	squares	square	NOUN
fcis-8581	5	19	.	.	PUNCT
fcis-8581	6	1	1	1	X
fcis-8581	6	2	.	.	X
fcis-8581	6	3	introduction	introduction	NOUN
fcis-8581	6	4	in	in	ADP
fcis-8581	6	5	the	the	DET
fcis-8581	6	6	past	past	ADJ
fcis-8581	6	7	two	two	NUM
fcis-8581	6	8	decades	decade	NOUN
fcis-8581	6	9	,	,	PUNCT
fcis-8581	6	10	internet	internet	NOUN
fcis-8581	6	11	technology	technology	NOUN
fcis-8581	6	12	has	have	AUX
fcis-8581	6	13	developed	develop	VERB
fcis-8581	6	14	rapidly	rapidly	ADV
fcis-8581	6	15	,	,	PUNCT
fcis-8581	6	16	and	and	CCONJ
fcis-8581	6	17	the	the	DET
fcis-8581	6	18	number	number	NOUN
fcis-8581	6	19	of	of	ADP
fcis-8581	6	20	active	active	ADJ
fcis-8581	6	21	users	user	NOUN
fcis-8581	6	22	on	on	ADP
fcis-8581	6	23	the	the	DET
fcis-8581	6	24	internet	internet	NOUN
fcis-8581	6	25	has	have	AUX
fcis-8581	6	26	grown	grow	VERB
fcis-8581	6	27	and	and	CCONJ
fcis-8581	6	28	the	the	DET
fcis-8581	6	29	amount	amount	NOUN
fcis-8581	6	30	of	of	ADP
fcis-8581	6	31	data	datum	NOUN
fcis-8581	6	32	generated	generate	VERB
fcis-8581	6	33	has	have	AUX
fcis-8581	6	34	increased	increase	VERB
fcis-8581	6	35	.	.	PUNCT
fcis-8581	7	1	how	how	SCONJ
fcis-8581	7	2	to	to	PART
fcis-8581	7	3	quickly	quickly	ADV
fcis-8581	7	4	find	find	VERB
fcis-8581	7	5	the	the	DET
fcis-8581	7	6	information	information	NOUN
fcis-8581	7	7	users	user	NOUN
fcis-8581	7	8	want	want	VERB
fcis-8581	7	9	from	from	ADP
fcis-8581	7	10	such	such	ADJ
fcis-8581	7	11	massive	massive	ADJ
fcis-8581	7	12	data	datum	NOUN
fcis-8581	7	13	becomes	become	VERB
fcis-8581	7	14	a	a	DET
fcis-8581	7	15	challenge	challenge	NOUN
fcis-8581	7	16	,	,	PUNCT
fcis-8581	7	17	i.e.	i.e.	X
fcis-8581	7	18	,	,	PUNCT
fcis-8581	7	19	the	the	DET
fcis-8581	7	20	information	information	NOUN
fcis-8581	7	21	overload	overload	NOUN
fcis-8581	7	22	problem	problem	NOUN
fcis-8581	8	1	[	[	X
fcis-8581	8	2	1	1	NUM
fcis-8581	8	3	]	]	PUNCT
fcis-8581	8	4	.	.	PUNCT
fcis-8581	9	1	for	for	ADP
fcis-8581	9	2	this	this	DET
fcis-8581	9	3	reason	reason	NOUN
fcis-8581	9	4	,	,	PUNCT
fcis-8581	9	5	search	search	NOUN
fcis-8581	9	6	engine	engine	NOUN
fcis-8581	9	7	and	and	CCONJ
fcis-8581	9	8	recommendation	recommendation	NOUN
fcis-8581	9	9	engine	engine	NOUN
fcis-8581	9	10	technologies	technology	NOUN
fcis-8581	9	11	were	be	AUX
fcis-8581	9	12	created	create	VERB
fcis-8581	9	13	.	.	PUNCT
fcis-8581	10	1	at	at	ADP
fcis-8581	10	2	present	present	ADJ
fcis-8581	10	3	,	,	PUNCT
fcis-8581	10	4	the	the	DET
fcis-8581	10	5	scale	scale	NOUN
fcis-8581	10	6	of	of	ADP
fcis-8581	10	7	china	china	PROPN
fcis-8581	10	8	's	's	PART
fcis-8581	10	9	internet	internet	NOUN
fcis-8581	10	10	search	search	NOUN
fcis-8581	10	11	engine	engine	NOUN
fcis-8581	10	12	users	user	NOUN
fcis-8581	10	13	reached	reach	VERB
fcis-8581	10	14	687	687	NUM
fcis-8581	10	15	million	million	NUM
fcis-8581	10	16	,	,	PUNCT
fcis-8581	10	17	and	and	CCONJ
fcis-8581	10	18	the	the	DET
fcis-8581	10	19	scale	scale	NOUN
fcis-8581	10	20	of	of	ADP
fcis-8581	10	21	cell	cell	NOUN
fcis-8581	10	22	phone	phone	NOUN
fcis-8581	10	23	search	search	NOUN
fcis-8581	10	24	applications	application	NOUN
fcis-8581	10	25	also	also	ADV
fcis-8581	10	26	has	have	VERB
fcis-8581	10	27	as	as	ADV
fcis-8581	10	28	many	many	ADJ
fcis-8581	10	29	as	as	ADP
fcis-8581	10	30	680	680	NUM
fcis-8581	10	31	million	million	NUM
fcis-8581	10	32	users	user	NOUN
fcis-8581	10	33	.	.	PUNCT
fcis-8581	11	1	search	search	NOUN
fcis-8581	11	2	engine	engine	NOUN
fcis-8581	11	3	technology	technology	NOUN
fcis-8581	11	4	has	have	AUX
fcis-8581	11	5	solved	solve	VERB
fcis-8581	11	6	the	the	DET
fcis-8581	11	7	problem	problem	NOUN
fcis-8581	11	8	of	of	ADP
fcis-8581	11	9	how	how	SCONJ
fcis-8581	11	10	to	to	PART
fcis-8581	11	11	filter	filter	VERB
fcis-8581	11	12	useful	useful	ADJ
fcis-8581	11	13	information	information	NOUN
fcis-8581	11	14	from	from	ADP
fcis-8581	11	15	large	large	ADJ
fcis-8581	11	16	-	-	PUNCT
fcis-8581	11	17	scale	scale	NOUN
fcis-8581	11	18	data	datum	NOUN
fcis-8581	11	19	for	for	ADP
fcis-8581	11	20	users	user	NOUN
fcis-8581	11	21	.	.	PUNCT
fcis-8581	12	1	but	but	CCONJ
fcis-8581	12	2	search	search	NOUN
fcis-8581	12	3	engine	engine	NOUN
fcis-8581	12	4	technology	technology	NOUN
fcis-8581	12	5	is	be	AUX
fcis-8581	12	6	for	for	ADP
fcis-8581	12	7	all	all	DET
fcis-8581	12	8	users	user	NOUN
fcis-8581	12	9	,	,	PUNCT
fcis-8581	12	10	and	and	CCONJ
fcis-8581	12	11	the	the	DET
fcis-8581	12	12	recommendation	recommendation	NOUN
fcis-8581	12	13	results	result	NOUN
fcis-8581	12	14	are	be	AUX
fcis-8581	12	15	universal	universal	ADJ
fcis-8581	12	16	and	and	CCONJ
fcis-8581	12	17	not	not	PART
fcis-8581	12	18	personalized	personalize	VERB
fcis-8581	12	19	enough	enough	ADV
fcis-8581	12	20	.	.	PUNCT
fcis-8581	13	1	movie	movie	NOUN
fcis-8581	13	2	search	search	NOUN
fcis-8581	13	3	also	also	ADV
fcis-8581	13	4	faces	face	VERB
fcis-8581	13	5	the	the	DET
fcis-8581	13	6	above	above	ADJ
fcis-8581	13	7	problem	problem	NOUN
fcis-8581	13	8	,	,	PUNCT
fcis-8581	13	9	and	and	CCONJ
fcis-8581	13	10	for	for	ADP
fcis-8581	13	11	this	this	DET
fcis-8581	13	12	reason	reason	NOUN
fcis-8581	13	13	,	,	PUNCT
fcis-8581	13	14	movie	movie	NOUN
fcis-8581	13	15	recommendation	recommendation	NOUN
fcis-8581	13	16	system	system	NOUN
fcis-8581	13	17	[	[	X
fcis-8581	13	18	2	2	NUM
fcis-8581	13	19	]	]	PUNCT
fcis-8581	13	20	was	be	AUX
fcis-8581	13	21	born	bear	VERB
fcis-8581	13	22	.	.	PUNCT
fcis-8581	14	1	by	by	ADP
fcis-8581	14	2	studying	study	VERB
fcis-8581	14	3	users	user	NOUN
fcis-8581	14	4	'	'	PART
fcis-8581	14	5	personal	personal	ADJ
fcis-8581	14	6	information	information	NOUN
fcis-8581	14	7	,	,	PUNCT
fcis-8581	14	8	item	item	NOUN
fcis-8581	14	9	information	information	NOUN
fcis-8581	14	10	and	and	CCONJ
fcis-8581	14	11	log	log	VERB
fcis-8581	14	12	records	record	NOUN
fcis-8581	14	13	to	to	PART
fcis-8581	14	14	filter	filter	VERB
fcis-8581	14	15	and	and	CCONJ
fcis-8581	14	16	sort	sort	VERB
fcis-8581	14	17	through	through	ADP
fcis-8581	14	18	the	the	DET
fcis-8581	14	19	clutter	clutter	NOUN
fcis-8581	14	20	of	of	ADP
fcis-8581	14	21	data	datum	NOUN
fcis-8581	14	22	,	,	PUNCT
fcis-8581	14	23	it	it	PRON
fcis-8581	14	24	emphasizes	emphasize	VERB
fcis-8581	14	25	providing	provide	VERB
fcis-8581	14	26	personalized	personalized	ADJ
fcis-8581	14	27	services	service	NOUN
fcis-8581	14	28	that	that	PRON
fcis-8581	14	29	help	help	VERB
fcis-8581	14	30	users	user	NOUN
fcis-8581	14	31	find	find	VERB
fcis-8581	14	32	the	the	DET
fcis-8581	14	33	information	information	NOUN
fcis-8581	14	34	they	they	PRON
fcis-8581	14	35	want	want	VERB
fcis-8581	14	36	better	well	ADJ
fcis-8581	14	37	and	and	CCONJ
fcis-8581	14	38	faster	fast	ADV
fcis-8581	14	39	.	.	PUNCT
fcis-8581	15	1	at	at	ADP
fcis-8581	15	2	the	the	DET
fcis-8581	15	3	same	same	ADJ
fcis-8581	15	4	time	time	NOUN
fcis-8581	15	5	,	,	PUNCT
fcis-8581	15	6	as	as	SCONJ
fcis-8581	15	7	he	he	PRON
fcis-8581	15	8	will	will	AUX
fcis-8581	15	9	track	track	VERB
fcis-8581	15	10	the	the	DET
fcis-8581	15	11	user	user	NOUN
fcis-8581	15	12	's	's	PART
fcis-8581	15	13	log	log	NOUN
fcis-8581	15	14	records	record	NOUN
fcis-8581	15	15	and	and	CCONJ
fcis-8581	15	16	constantly	constantly	ADV
fcis-8581	15	17	modify	modify	VERB
fcis-8581	15	18	the	the	DET
fcis-8581	15	19	recommendation	recommendation	NOUN
fcis-8581	15	20	model	model	NOUN
fcis-8581	15	21	according	accord	VERB
fcis-8581	15	22	to	to	ADP
fcis-8581	15	23	the	the	DET
fcis-8581	15	24	user	user	NOUN
fcis-8581	15	25	's	's	PART
fcis-8581	15	26	behavior	behavior	NOUN
fcis-8581	15	27	,	,	PUNCT
fcis-8581	15	28	thus	thus	ADV
fcis-8581	15	29	providing	provide	VERB
fcis-8581	15	30	more	more	ADJ
fcis-8581	15	31	real	real	ADJ
fcis-8581	15	32	-	-	PUNCT
fcis-8581	15	33	time	time	NOUN
fcis-8581	15	34	and	and	CCONJ
fcis-8581	15	35	accurate	accurate	ADJ
fcis-8581	15	36	recommendations	recommendation	NOUN
fcis-8581	15	37	.	.	PUNCT
fcis-8581	16	1	in	in	ADP
fcis-8581	16	2	this	this	DET
fcis-8581	16	3	paper	paper	NOUN
fcis-8581	16	4	,	,	PUNCT
fcis-8581	16	5	the	the	DET
fcis-8581	16	6	current	current	ADJ
fcis-8581	16	7	movie	movie	NOUN
fcis-8581	16	8	recommendation	recommendation	NOUN
fcis-8581	16	9	algorithm	algorithm	NOUN
fcis-8581	16	10	is	be	AUX
fcis-8581	16	11	improved	improve	VERB
fcis-8581	16	12	by	by	ADP
fcis-8581	16	13	first	first	ADV
fcis-8581	16	14	using	use	VERB
fcis-8581	16	15	an	an	DET
fcis-8581	16	16	itembased	itembased	ADJ
fcis-8581	16	17	collaborative	collaborative	ADJ
fcis-8581	16	18	filtering	filtering	NOUN
fcis-8581	16	19	algorithm	algorithm	NOUN
fcis-8581	16	20	for	for	ADP
fcis-8581	16	21	the	the	DET
fcis-8581	16	22	item	item	NOUN
fcis-8581	16	23	similarity	similarity	NOUN
fcis-8581	16	24	measure	measure	NOUN
fcis-8581	16	25	in	in	ADP
fcis-8581	16	26	the	the	DET
fcis-8581	16	27	item	item	NOUN
fcis-8581	16	28	-	-	PUNCT
fcis-8581	16	29	based	base	VERB
fcis-8581	16	30	recommendation	recommendation	NOUN
fcis-8581	16	31	process	process	NOUN
fcis-8581	16	32	.	.	PUNCT
fcis-8581	17	1	in	in	ADP
fcis-8581	17	2	the	the	DET
fcis-8581	17	3	recommendation	recommendation	NOUN
fcis-8581	17	4	process	process	NOUN
fcis-8581	17	5	,	,	PUNCT
fcis-8581	17	6	a	a	DET
fcis-8581	17	7	matrix	matrix	NOUN
fcis-8581	17	8	decomposition	decomposition	NOUN
fcis-8581	17	9	-	-	PUNCT
fcis-8581	17	10	based	base	VERB
fcis-8581	17	11	recommendation	recommendation	NOUN
fcis-8581	17	12	algorithm	algorithm	NOUN
fcis-8581	17	13	and	and	CCONJ
fcis-8581	17	14	a	a	DET
fcis-8581	17	15	regularized	regularize	VERB
fcis-8581	17	16	singular	singular	ADJ
fcis-8581	17	17	value	value	NOUN
fcis-8581	17	18	decomposition	decomposition	NOUN
fcis-8581	17	19	algorithm	algorithm	NOUN
fcis-8581	17	20	are	be	AUX
fcis-8581	17	21	used	use	VERB
fcis-8581	17	22	,	,	PUNCT
fcis-8581	17	23	and	and	CCONJ
fcis-8581	17	24	alternating	alternate	VERB
fcis-8581	17	25	least	least	ADJ
fcis-8581	17	26	squares	square	NOUN
fcis-8581	17	27	is	be	AUX
fcis-8581	17	28	used	use	VERB
fcis-8581	17	29	to	to	PART
fcis-8581	17	30	train	train	VERB
fcis-8581	17	31	the	the	DET
fcis-8581	17	32	data	datum	NOUN
fcis-8581	17	33	and	and	CCONJ
fcis-8581	17	34	find	find	VERB
fcis-8581	17	35	the	the	DET
fcis-8581	17	36	objective	objective	ADJ
fcis-8581	17	37	function	function	NOUN
fcis-8581	17	38	.	.	PUNCT
fcis-8581	18	1	2	2	X
fcis-8581	18	2	.	.	X
fcis-8581	18	3	collaborative	collaborative	ADJ
fcis-8581	18	4	filtering	filtering	NOUN
fcis-8581	18	5	recommendation	recommendation	NOUN
fcis-8581	18	6	algorithm	algorithm	NOUN
fcis-8581	18	7	2.1	2.1	NUM
fcis-8581	18	8	.	.	PUNCT
fcis-8581	19	1	user	user	NOUN
fcis-8581	19	2	-	-	PUNCT
fcis-8581	19	3	based	base	VERB
fcis-8581	19	4	recommendation	recommendation	NOUN
fcis-8581	19	5	algorithm	algorithm	NOUN
fcis-8581	19	6	the	the	DET
fcis-8581	19	7	main	main	ADJ
fcis-8581	19	8	idea	idea	NOUN
fcis-8581	19	9	of	of	ADP
fcis-8581	19	10	the	the	DET
fcis-8581	19	11	algorithm	algorithm	NOUN
fcis-8581	19	12	is	be	AUX
fcis-8581	19	13	to	to	PART
fcis-8581	19	14	calculate	calculate	VERB
fcis-8581	19	15	the	the	DET
fcis-8581	19	16	similarity	similarity	NOUN
fcis-8581	19	17	between	between	ADP
fcis-8581	19	18	two	two	NUM
fcis-8581	19	19	users	user	NOUN
fcis-8581	19	20	using	use	VERB
fcis-8581	19	21	some	some	DET
fcis-8581	19	22	similarity	similarity	NOUN
fcis-8581	19	23	measure	measure	NOUN
fcis-8581	19	24	based	base	VERB
fcis-8581	19	25	on	on	ADP
fcis-8581	19	26	user	user	NOUN
fcis-8581	19	27	-	-	PUNCT
fcis-8581	19	28	item	item	NOUN
fcis-8581	19	29	association	association	NOUN
fcis-8581	19	30	data	datum	NOUN
fcis-8581	19	31	,	,	PUNCT
fcis-8581	19	32	such	such	ADJ
fcis-8581	19	33	as	as	ADP
fcis-8581	19	34	the	the	DET
fcis-8581	19	35	user	user	NOUN
fcis-8581	19	36	's	's	PART
fcis-8581	19	37	evaluation	evaluation	NOUN
fcis-8581	19	38	information	information	NOUN
fcis-8581	19	39	about	about	ADP
fcis-8581	19	40	the	the	DET
fcis-8581	19	41	item	item	NOUN
fcis-8581	19	42	,	,	PUNCT
fcis-8581	19	43	and	and	CCONJ
fcis-8581	19	44	thus	thus	ADV
fcis-8581	19	45	obtain	obtain	VERB
fcis-8581	19	46	a	a	DET
fcis-8581	19	47	set	set	NOUN
fcis-8581	19	48	of	of	ADP
fcis-8581	19	49	users	user	NOUN
fcis-8581	19	50	containing	contain	VERB
fcis-8581	19	51	similarity	similarity	NOUN
fcis-8581	19	52	values	value	NOUN
fcis-8581	19	53	.	.	PUNCT
fcis-8581	20	1	its	its	PRON
fcis-8581	20	2	recommendation	recommendation	NOUN
fcis-8581	20	3	value	value	NOUN
fcis-8581	20	4	is	be	AUX
fcis-8581	20	5	related	relate	VERB
fcis-8581	20	6	to	to	ADP
fcis-8581	20	7	the	the	DET
fcis-8581	20	8	similarity	similarity	NOUN
fcis-8581	20	9	between	between	ADP
fcis-8581	20	10	users	user	NOUN
fcis-8581	20	11	and	and	CCONJ
fcis-8581	20	12	the	the	DET
fcis-8581	20	13	rating	rating	NOUN
fcis-8581	20	14	of	of	ADP
fcis-8581	20	15	recommended	recommend	VERB
fcis-8581	20	16	items	item	NOUN
fcis-8581	20	17	by	by	ADP
fcis-8581	20	18	users	user	NOUN
fcis-8581	20	19	'	'	PART
fcis-8581	20	20	neighbors	neighbor	NOUN
fcis-8581	20	21	,	,	PUNCT
fcis-8581	20	22	and	and	CCONJ
fcis-8581	20	23	the	the	DET
fcis-8581	20	24	recommendation	recommendation	NOUN
fcis-8581	20	25	results	result	NOUN
fcis-8581	20	26	are	be	AUX
fcis-8581	20	27	ranked	rank	VERB
fcis-8581	20	28	in	in	ADP
fcis-8581	20	29	descending	descend	VERB
fcis-8581	20	30	order	order	NOUN
fcis-8581	20	31	according	accord	VERB
fcis-8581	20	32	to	to	ADP
fcis-8581	20	33	the	the	DET
fcis-8581	20	34	recommendation	recommendation	NOUN
fcis-8581	20	35	value	value	NOUN
fcis-8581	20	36	.	.	PUNCT
fcis-8581	21	1	user	user	NOUN
fcis-8581	21	2	-	-	PUNCT
fcis-8581	21	3	based	base	VERB
fcis-8581	21	4	recommendations	recommendation	NOUN
fcis-8581	21	5	are	be	AUX
fcis-8581	21	6	highly	highly	ADV
fcis-8581	21	7	community	community	NOUN
fcis-8581	21	8	-	-	PUNCT
fcis-8581	21	9	based	base	VERB
fcis-8581	21	10	,	,	PUNCT
fcis-8581	21	11	timesensitive	timesensitive	ADJ
fcis-8581	21	12	and	and	CCONJ
fcis-8581	21	13	influenced	influence	VERB
fcis-8581	21	14	by	by	ADP
fcis-8581	21	15	user	user	NOUN
fcis-8581	21	16	size	size	NOUN
fcis-8581	21	17	.	.	PUNCT
fcis-8581	22	1	but	but	CCONJ
fcis-8581	22	2	also	also	ADV
fcis-8581	22	3	suffer	suffer	VERB
fcis-8581	22	4	from	from	ADP
fcis-8581	22	5	high	high	ADJ
fcis-8581	22	6	data	datum	NOUN
fcis-8581	22	7	sparsity	sparsity	NOUN
fcis-8581	22	8	[	[	X
fcis-8581	22	9	3	3	NUM
fcis-8581	22	10	]	]	PUNCT
fcis-8581	22	11	and	and	CCONJ
fcis-8581	22	12	poor	poor	ADJ
fcis-8581	22	13	scalability	scalability	NOUN
fcis-8581	22	14	.	.	PUNCT
fcis-8581	23	1	therefore	therefore	ADV
fcis-8581	23	2	,	,	PUNCT
fcis-8581	23	3	a	a	DET
fcis-8581	23	4	recommendation	recommendation	NOUN
fcis-8581	23	5	system	system	NOUN
fcis-8581	23	6	based	base	VERB
fcis-8581	23	7	on	on	ADP
fcis-8581	23	8	the	the	DET
fcis-8581	23	9	nearest	near	ADJ
fcis-8581	23	10	neighbor	neighbor	NOUN
fcis-8581	23	11	algorithm	algorithm	NOUN
fcis-8581	23	12	may	may	AUX
fcis-8581	23	13	not	not	PART
fcis-8581	23	14	be	be	AUX
fcis-8581	23	15	able	able	ADJ
fcis-8581	23	16	to	to	PART
fcis-8581	23	17	make	make	VERB
fcis-8581	23	18	effective	effective	ADJ
fcis-8581	23	19	recommendations	recommendation	NOUN
fcis-8581	23	20	for	for	ADP
fcis-8581	23	21	users	user	NOUN
fcis-8581	23	22	whose	whose	DET
fcis-8581	23	23	relevant	relevant	ADJ
fcis-8581	23	24	data	datum	NOUN
fcis-8581	23	25	are	be	AUX
fcis-8581	23	26	too	too	ADV
fcis-8581	23	27	sparse	sparse	ADJ
fcis-8581	23	28	.	.	PUNCT
fcis-8581	24	1	2.2	2.2	NUM
fcis-8581	24	2	.	.	PUNCT
fcis-8581	24	3	item	item	NOUN
fcis-8581	24	4	-	-	PUNCT
fcis-8581	24	5	based	base	VERB
fcis-8581	24	6	recommendation	recommendation	NOUN
fcis-8581	24	7	algorithm	algorithm	NOUN
fcis-8581	24	8	the	the	DET
fcis-8581	24	9	algorithm	algorithm	NOUN
fcis-8581	24	10	is	be	AUX
fcis-8581	24	11	based	base	VERB
fcis-8581	24	12	on	on	ADP
fcis-8581	24	13	the	the	DET
fcis-8581	24	14	assumption	assumption	NOUN
fcis-8581	24	15	that	that	SCONJ
fcis-8581	24	16	most	most	ADJ
fcis-8581	24	17	users	user	NOUN
fcis-8581	24	18	'	'	PART
fcis-8581	24	19	preferences	preference	NOUN
fcis-8581	24	20	are	be	AUX
fcis-8581	24	21	focused	focused	ADJ
fcis-8581	24	22	,	,	PUNCT
fcis-8581	24	23	that	that	SCONJ
fcis-8581	24	24	two	two	NUM
fcis-8581	24	25	items	item	NOUN
fcis-8581	24	26	are	be	AUX
fcis-8581	24	27	similar	similar	ADJ
fcis-8581	24	28	if	if	SCONJ
fcis-8581	24	29	they	they	PRON
fcis-8581	24	30	have	have	AUX
fcis-8581	24	31	been	be	AUX
fcis-8581	24	32	expressed	express	VERB
fcis-8581	24	33	as	as	ADP
fcis-8581	24	34	preferences	preference	NOUN
fcis-8581	24	35	by	by	ADP
fcis-8581	24	36	more	more	ADJ
fcis-8581	24	37	than	than	ADP
fcis-8581	24	38	one	one	NUM
fcis-8581	24	39	user	user	NOUN
fcis-8581	24	40	at	at	ADP
fcis-8581	24	41	the	the	DET
fcis-8581	24	42	same	same	ADJ
fcis-8581	24	43	time	time	NOUN
fcis-8581	24	44	,	,	PUNCT
fcis-8581	24	45	and	and	CCONJ
fcis-8581	24	46	that	that	SCONJ
fcis-8581	24	47	the	the	PRON
fcis-8581	24	48	closer	close	ADJ
fcis-8581	24	49	the	the	DET
fcis-8581	24	50	ratings	rating	NOUN
fcis-8581	24	51	given	give	VERB
fcis-8581	24	52	by	by	ADP
fcis-8581	24	53	the	the	DET
fcis-8581	24	54	users	user	NOUN
fcis-8581	24	55	who	who	PRON
fcis-8581	24	56	expressed	express	VERB
fcis-8581	24	57	their	their	PRON
fcis-8581	24	58	preferences	preference	NOUN
fcis-8581	24	59	for	for	ADP
fcis-8581	24	60	both	both	DET
fcis-8581	24	61	items	item	NOUN
fcis-8581	24	62	together	together	ADV
fcis-8581	24	63	,	,	PUNCT
fcis-8581	24	64	the	the	PRON
fcis-8581	24	65	more	more	ADV
fcis-8581	24	66	similar	similar	ADJ
fcis-8581	24	67	the	the	DET
fcis-8581	24	68	two	two	NUM
fcis-8581	24	69	items	item	NOUN
fcis-8581	24	70	are	be	AUX
fcis-8581	24	71	.	.	PUNCT
fcis-8581	25	1	collaborative	collaborative	ADJ
fcis-8581	25	2	item	item	NOUN
fcis-8581	25	3	-	-	PUNCT
fcis-8581	25	4	based	base	VERB
fcis-8581	25	5	filtering	filtering	NOUN
fcis-8581	25	6	has	have	VERB
fcis-8581	25	7	the	the	DET
fcis-8581	25	8	following	follow	VERB
fcis-8581	25	9	three	three	NUM
fcis-8581	25	10	features	feature	NOUN
fcis-8581	25	11	:	:	PUNCT
fcis-8581	25	12	first	first	ADV
fcis-8581	25	13	,	,	PUNCT
fcis-8581	25	14	it	it	PRON
fcis-8581	25	15	does	do	AUX
fcis-8581	25	16	not	not	PART
fcis-8581	25	17	need	need	VERB
fcis-8581	25	18	to	to	PART
fcis-8581	25	19	calculate	calculate	VERB
fcis-8581	25	20	similar	similar	ADJ
fcis-8581	25	21	users	user	NOUN
fcis-8581	25	22	and	and	CCONJ
fcis-8581	25	23	directly	directly	ADV
fcis-8581	25	24	calculates	calculate	VERB
fcis-8581	25	25	similar	similar	ADJ
fcis-8581	25	26	items	item	NOUN
fcis-8581	25	27	.	.	PUNCT
fcis-8581	26	1	second	second	ADJ
fcis-8581	26	2	,	,	PUNCT
fcis-8581	26	3	the	the	DET
fcis-8581	26	4	recommendation	recommendation	NOUN
fcis-8581	26	5	results	result	NOUN
fcis-8581	26	6	are	be	AUX
fcis-8581	26	7	more	more	ADV
fcis-8581	26	8	personalized	personalized	ADJ
fcis-8581	26	9	,	,	PUNCT
fcis-8581	26	10	rich	rich	ADJ
fcis-8581	26	11	in	in	ADP
fcis-8581	26	12	long	long	ADJ
fcis-8581	26	13	-	-	PUNCT
fcis-8581	26	14	tail	tail	NOUN
fcis-8581	26	15	items	item	NOUN
fcis-8581	26	16	,	,	PUNCT
fcis-8581	26	17	and	and	CCONJ
fcis-8581	26	18	their	their	PRON
fcis-8581	26	19	recommendation	recommendation	NOUN
fcis-8581	26	20	results	result	NOUN
fcis-8581	26	21	are	be	AUX
fcis-8581	26	22	more	more	ADV
fcis-8581	26	23	targeted	targeted	ADJ
fcis-8581	26	24	.	.	PUNCT
fcis-8581	27	1	third	third	ADJ
fcis-8581	27	2	,	,	PUNCT
fcis-8581	27	3	collaborative	collaborative	ADJ
fcis-8581	27	4	filtering	filtering	NOUN
fcis-8581	27	5	based	base	VERB
fcis-8581	27	6	on	on	ADP
fcis-8581	27	7	items	item	NOUN
fcis-8581	27	8	needs	need	VERB
fcis-8581	27	9	to	to	PART
fcis-8581	27	10	calculate	calculate	VERB
fcis-8581	27	11	the	the	DET
fcis-8581	27	12	set	set	NOUN
fcis-8581	27	13	of	of	ADP
fcis-8581	27	14	items	item	NOUN
fcis-8581	27	15	similar	similar	ADJ
fcis-8581	27	16	to	to	ADP
fcis-8581	27	17	the	the	DET
fcis-8581	27	18	user	user	NOUN
fcis-8581	27	19	's	's	PART
fcis-8581	27	20	favorite	favorite	ADJ
fcis-8581	27	21	items	item	NOUN
fcis-8581	27	22	,	,	PUNCT
fcis-8581	27	23	and	and	CCONJ
fcis-8581	27	24	its	its	PRON
fcis-8581	27	25	calculation	calculation	NOUN
fcis-8581	27	26	is	be	AUX
fcis-8581	27	27	proportional	proportional	ADJ
fcis-8581	27	28	to	to	ADP
fcis-8581	27	29	the	the	DET
fcis-8581	27	30	square	square	NOUN
fcis-8581	27	31	of	of	ADP
fcis-8581	27	32	the	the	DET
fcis-8581	27	33	number	number	NOUN
fcis-8581	27	34	of	of	ADP
fcis-8581	27	35	items	item	NOUN
fcis-8581	27	36	,	,	PUNCT
fcis-8581	27	37	which	which	PRON
fcis-8581	27	38	is	be	AUX
fcis-8581	27	39	suitable	suitable	ADJ
fcis-8581	27	40	for	for	ADP
fcis-8581	27	41	business	business	NOUN
fcis-8581	27	42	scenarios	scenario	NOUN
fcis-8581	27	43	with	with	ADP
fcis-8581	27	44	small	small	ADJ
fcis-8581	27	45	114	114	NUM
fcis-8581	27	46	number	number	NOUN
fcis-8581	27	47	of	of	ADP
fcis-8581	27	48	items	item	NOUN
fcis-8581	27	49	and	and	CCONJ
fcis-8581	27	50	large	large	ADJ
fcis-8581	27	51	user	user	NOUN
fcis-8581	27	52	size	size	NOUN
fcis-8581	27	53	.	.	PUNCT
fcis-8581	28	1	the	the	DET
fcis-8581	28	2	core	core	NOUN
fcis-8581	28	3	algorithm	algorithm	NOUN
fcis-8581	28	4	in	in	ADP
fcis-8581	28	5	the	the	DET
fcis-8581	28	6	item	item	NOUN
fcis-8581	28	7	-	-	PUNCT
fcis-8581	28	8	based	base	VERB
fcis-8581	28	9	collaborative	collaborative	ADJ
fcis-8581	28	10	filtering	filtering	NOUN
fcis-8581	28	11	recommendation	recommendation	NOUN
fcis-8581	28	12	algorithm	algorithm	NOUN
fcis-8581	28	13	is	be	AUX
fcis-8581	28	14	to	to	PART
fcis-8581	28	15	calculate	calculate	VERB
fcis-8581	28	16	the	the	DET
fcis-8581	28	17	similarity	similarity	NOUN
fcis-8581	28	18	between	between	ADP
fcis-8581	28	19	items	item	NOUN
fcis-8581	28	20	and	and	CCONJ
fcis-8581	28	21	then	then	ADV
fcis-8581	28	22	use	use	VERB
fcis-8581	28	23	the	the	DET
fcis-8581	28	24	similarity	similarity	NOUN
fcis-8581	28	25	between	between	ADP
fcis-8581	28	26	items	item	NOUN
fcis-8581	28	27	to	to	PART
fcis-8581	28	28	predict	predict	VERB
fcis-8581	28	29	the	the	DET
fcis-8581	28	30	user	user	NOUN
fcis-8581	28	31	's	's	PART
fcis-8581	28	32	rating	rating	NOUN
fcis-8581	28	33	system	system	NOUN
fcis-8581	28	34	for	for	ADP
fcis-8581	28	35	candidate	candidate	NOUN
fcis-8581	28	36	items	item	NOUN
fcis-8581	28	37	.	.	PUNCT
fcis-8581	29	1	therefore	therefore	ADV
fcis-8581	29	2	,	,	PUNCT
fcis-8581	29	3	the	the	DET
fcis-8581	29	4	choice	choice	NOUN
fcis-8581	29	5	of	of	ADP
fcis-8581	29	6	the	the	DET
fcis-8581	29	7	similarity	similarity	NOUN
fcis-8581	29	8	measure	measure	NOUN
fcis-8581	29	9	plays	play	VERB
fcis-8581	29	10	a	a	DET
fcis-8581	29	11	decisive	decisive	ADJ
fcis-8581	29	12	role	role	NOUN
fcis-8581	29	13	in	in	ADP
fcis-8581	29	14	the	the	DET
fcis-8581	29	15	final	final	ADJ
fcis-8581	29	16	result	result	NOUN
fcis-8581	29	17	.	.	PUNCT
fcis-8581	30	1	three	three	NUM
fcis-8581	30	2	main	main	ADJ
fcis-8581	30	3	methods	method	NOUN
fcis-8581	30	4	are	be	AUX
fcis-8581	30	5	included	include	VERB
fcis-8581	30	6	as	as	SCONJ
fcis-8581	30	7	follows	follow	VERB
fcis-8581	30	8	:	:	PUNCT
fcis-8581	30	9	(	(	PUNCT
fcis-8581	30	10	1	1	X
fcis-8581	30	11	)	)	PUNCT
fcis-8581	30	12	euclidean	euclidean	ADJ
fcis-8581	30	13	distance	distance	NOUN
fcis-8581	30	14	calculation	calculation	NOUN
fcis-8581	30	15	formula	formula	NOUN
fcis-8581	30	16	in	in	ADP
fcis-8581	30	17	calculating	calculate	VERB
fcis-8581	30	18	user	user	NOUN
fcis-8581	30	19	similarity	similarity	NOUN
fcis-8581	30	20	,	,	PUNCT
fcis-8581	30	21	the	the	DET
fcis-8581	30	22	euclidean	euclidean	ADJ
fcis-8581	30	23	distance	distance	NOUN
fcis-8581	30	24	[	[	X
fcis-8581	30	25	4	4	X
fcis-8581	30	26	]	]	PUNCT
fcis-8581	30	27	is	be	AUX
fcis-8581	30	28	a	a	DET
fcis-8581	30	29	more	more	ADV
fcis-8581	30	30	intuitive	intuitive	ADJ
fcis-8581	30	31	similarity	similarity	NOUN
fcis-8581	30	32	algorithm	algorithm	NOUN
fcis-8581	30	33	.	.	PUNCT
fcis-8581	31	1	according	accord	VERB
fcis-8581	31	2	to	to	ADP
fcis-8581	31	3	the	the	DET
fcis-8581	31	4	common	common	ADJ
fcis-8581	31	5	evaluation	evaluation	NOUN
fcis-8581	31	6	of	of	ADP
fcis-8581	31	7	item	item	NOUN
fcis-8581	31	8	as	as	ADP
fcis-8581	31	9	a	a	DET
fcis-8581	31	10	dimension	dimension	NOUN
fcis-8581	31	11	between	between	ADP
fcis-8581	31	12	users	user	NOUN
fcis-8581	31	13	,	,	PUNCT
fcis-8581	31	14	a	a	DET
fcis-8581	31	15	multidimensional	multidimensional	ADJ
fcis-8581	31	16	space	space	NOUN
fcis-8581	31	17	is	be	AUX
fcis-8581	31	18	established	establish	VERB
fcis-8581	31	19	,	,	PUNCT
fcis-8581	31	20	and	and	CCONJ
fcis-8581	31	21	the	the	DET
fcis-8581	31	22	location	location	NOUN
fcis-8581	31	23	of	of	ADP
fcis-8581	31	24	this	this	DET
fcis-8581	31	25	user	user	NOUN
fcis-8581	31	26	in	in	ADP
fcis-8581	31	27	this	this	DET
fcis-8581	31	28	multidimensional	multidimensional	ADJ
fcis-8581	31	29	space	space	NOUN
fcis-8581	31	30	can	can	AUX
fcis-8581	31	31	be	be	AUX
fcis-8581	31	32	located	locate	VERB
fcis-8581	31	33	by	by	ADP
fcis-8581	31	34	the	the	DET
fcis-8581	31	35	coordinate	coordinate	NOUN
fcis-8581	31	36	system	system	NOUN
fcis-8581	31	37	x	x	INTJ
fcis-8581	31	38	(	(	PUNCT
fcis-8581	31	39	s1	s1	NOUN
fcis-8581	31	40	,	,	PUNCT
fcis-8581	31	41	s2	s2	PROPN
fcis-8581	31	42	,	,	PUNCT
fcis-8581	31	43	......	......	PUNCT
fcis-8581	31	44	,	,	PUNCT
fcis-8581	31	45	si	si	X
fcis-8581	31	46	)	)	PUNCT
fcis-8581	31	47	composed	compose	VERB
fcis-8581	31	48	of	of	ADP
fcis-8581	31	49	the	the	DET
fcis-8581	31	50	user	user	NOUN
fcis-8581	31	51	's	's	PART
fcis-8581	31	52	evaluation	evaluation	NOUN
fcis-8581	31	53	score	score	NOUN
fcis-8581	31	54	on	on	ADP
fcis-8581	31	55	a	a	DET
fcis-8581	31	56	single	single	ADJ
fcis-8581	31	57	dimension	dimension	NOUN
fcis-8581	31	58	,	,	PUNCT
fcis-8581	31	59	then	then	ADV
fcis-8581	31	60	any	any	DET
fcis-8581	31	61	two	two	NUM
fcis-8581	31	62	locations	location	NOUN
fcis-8581	31	63	distance	distance	NOUN
fcis-8581	31	64	(	(	PUNCT
fcis-8581	31	65	x	x	NOUN
fcis-8581	31	66	,	,	PUNCT
fcis-8581	31	67	y	y	NOUN
fcis-8581	31	68	)	)	PUNCT
fcis-8581	31	69	can	can	AUX
fcis-8581	31	70	reflect	reflect	VERB
fcis-8581	31	71	the	the	DET
fcis-8581	31	72	degree	degree	NOUN
fcis-8581	31	73	of	of	ADP
fcis-8581	31	74	similarity	similarity	NOUN
fcis-8581	31	75	between	between	ADP
fcis-8581	31	76	two	two	NUM
fcis-8581	31	77	users	user	NOUN
fcis-8581	31	78	or	or	CCONJ
fcis-8581	31	79	items	item	NOUN
fcis-8581	31	80	.	.	PUNCT
fcis-8581	32	1	the	the	DET
fcis-8581	32	2	calculation	calculation	NOUN
fcis-8581	32	3	formula	formula	NOUN
fcis-8581	32	4	is	be	AUX
fcis-8581	32	5	shown	show	VERB
fcis-8581	32	6	in	in	ADP
fcis-8581	32	7	equation	equation	NOUN
fcis-8581	32	8	(	(	PUNCT
fcis-8581	32	9	1	1	NUM
fcis-8581	32	10	):	):	PUNCT
fcis-8581	32	11	dist	dist	NOUN
fcis-8581	32	12	(	(	PUNCT
fcis-8581	32	13	x	x	NOUN
fcis-8581	32	14	,	,	PUNCT
fcis-8581	32	15	y	y	NOUN
fcis-8581	32	16	)	)	PUNCT
fcis-8581	32	17	=	=	PUNCT
fcis-8581	32	18	∑	∑	PUNCT
fcis-8581	32	19	(	(	PUNCT
fcis-8581	32	20	1	1	NUM
fcis-8581	32	21	)	)	PUNCT
fcis-8581	32	22	(	(	PUNCT
fcis-8581	32	23	2	2	X
fcis-8581	32	24	)	)	PUNCT
fcis-8581	32	25	cosine	cosine	NOUN
fcis-8581	32	26	similarity	similarity	NOUN
fcis-8581	32	27	cosine	cosine	NOUN
fcis-8581	32	28	similarity	similarity	NOUN
fcis-8581	33	1	[	[	X
fcis-8581	33	2	5	5	NUM
fcis-8581	33	3	]	]	PUNCT
fcis-8581	33	4	is	be	AUX
fcis-8581	33	5	a	a	DET
fcis-8581	33	6	measure	measure	NOUN
fcis-8581	33	7	of	of	ADP
fcis-8581	33	8	the	the	DET
fcis-8581	33	9	similarity	similarity	NOUN
fcis-8581	33	10	between	between	ADP
fcis-8581	33	11	two	two	NUM
fcis-8581	33	12	texts	text	NOUN
fcis-8581	33	13	using	use	VERB
fcis-8581	33	14	the	the	DET
fcis-8581	33	15	cosine	cosine	NOUN
fcis-8581	33	16	of	of	ADP
fcis-8581	33	17	the	the	DET
fcis-8581	33	18	angle	angle	NOUN
fcis-8581	33	19	between	between	ADP
fcis-8581	33	20	two	two	NUM
fcis-8581	33	21	vectors	vector	NOUN
fcis-8581	33	22	in	in	ADP
fcis-8581	33	23	the	the	DET
fcis-8581	33	24	vector	vector	NOUN
fcis-8581	33	25	space	space	NOUN
fcis-8581	33	26	,	,	PUNCT
fcis-8581	33	27	which	which	PRON
fcis-8581	33	28	focuses	focus	VERB
fcis-8581	33	29	more	more	ADJ
fcis-8581	33	30	on	on	ADP
fcis-8581	33	31	the	the	DET
fcis-8581	33	32	difference	difference	NOUN
fcis-8581	33	33	in	in	ADP
fcis-8581	33	34	direction	direction	NOUN
fcis-8581	33	35	between	between	ADP
fcis-8581	33	36	two	two	NUM
fcis-8581	33	37	vectors	vector	NOUN
fcis-8581	33	38	than	than	ADP
fcis-8581	33	39	the	the	DET
fcis-8581	33	40	distance	distance	NOUN
fcis-8581	33	41	measure	measure	NOUN
fcis-8581	33	42	.	.	PUNCT
fcis-8581	34	1	in	in	ADP
fcis-8581	34	2	general	general	ADJ
fcis-8581	34	3	,	,	PUNCT
fcis-8581	34	4	after	after	ADP
fcis-8581	34	5	obtaining	obtain	VERB
fcis-8581	34	6	the	the	DET
fcis-8581	34	7	vector	vector	NOUN
fcis-8581	34	8	representation	representation	NOUN
fcis-8581	34	9	of	of	ADP
fcis-8581	34	10	two	two	NUM
fcis-8581	34	11	users	user	NOUN
fcis-8581	34	12	or	or	CCONJ
fcis-8581	34	13	items	item	NOUN
fcis-8581	34	14	in	in	ADP
fcis-8581	34	15	the	the	DET
fcis-8581	34	16	generated	generate	VERB
fcis-8581	34	17	collaborative	collaborative	ADJ
fcis-8581	34	18	filtering	filtering	NOUN
fcis-8581	34	19	matrix	matrix	NOUN
fcis-8581	34	20	,	,	PUNCT
fcis-8581	34	21	the	the	DET
fcis-8581	34	22	similarity	similarity	NOUN
fcis-8581	34	23	between	between	ADP
fcis-8581	34	24	the	the	DET
fcis-8581	34	25	two	two	NUM
fcis-8581	34	26	texts	text	NOUN
fcis-8581	34	27	can	can	AUX
fcis-8581	34	28	be	be	AUX
fcis-8581	34	29	calculated	calculate	VERB
fcis-8581	34	30	using	use	VERB
fcis-8581	34	31	the	the	DET
fcis-8581	34	32	cosine	cosine	NOUN
fcis-8581	34	33	similarity	similarity	NOUN
fcis-8581	34	34	,	,	PUNCT
fcis-8581	34	35	where	where	SCONJ
fcis-8581	34	36	the	the	DET
fcis-8581	34	37	result	result	NOUN
fcis-8581	34	38	is	be	AUX
fcis-8581	34	39	independent	independent	ADJ
fcis-8581	34	40	of	of	ADP
fcis-8581	34	41	the	the	DET
fcis-8581	34	42	length	length	NOUN
fcis-8581	34	43	of	of	ADP
fcis-8581	34	44	the	the	DET
fcis-8581	34	45	vector	vector	NOUN
fcis-8581	34	46	and	and	CCONJ
fcis-8581	34	47	is	be	AUX
fcis-8581	34	48	only	only	ADV
fcis-8581	34	49	related	relate	VERB
fcis-8581	34	50	to	to	ADP
fcis-8581	34	51	the	the	DET
fcis-8581	34	52	direction	direction	NOUN
fcis-8581	34	53	in	in	ADP
fcis-8581	34	54	which	which	PRON
fcis-8581	34	55	the	the	DET
fcis-8581	34	56	vector	vector	NOUN
fcis-8581	34	57	is	be	AUX
fcis-8581	34	58	pointing	point	VERB
fcis-8581	34	59	.	.	PUNCT
fcis-8581	35	1	the	the	DET
fcis-8581	35	2	cosine	cosine	NOUN
fcis-8581	35	3	between	between	ADP
fcis-8581	35	4	the	the	DET
fcis-8581	35	5	vectors	vector	NOUN
fcis-8581	35	6	can	can	AUX
fcis-8581	35	7	be	be	AUX
fcis-8581	35	8	found	find	VERB
fcis-8581	35	9	by	by	ADP
fcis-8581	35	10	using	use	VERB
fcis-8581	35	11	the	the	DET
fcis-8581	35	12	euclidean	euclidean	ADJ
fcis-8581	35	13	dot	dot	NOUN
fcis-8581	35	14	product	product	NOUN
fcis-8581	35	15	formula	formula	NOUN
fcis-8581	35	16	:	:	PUNCT
fcis-8581	35	17	a	a	DET
fcis-8581	35	18	∙	∙	PROPN
fcis-8581	35	19	b	b	PROPN
fcis-8581	35	20	|a|	|a|	PROPN
fcis-8581	35	21	|b|	|b|	PROPN
fcis-8581	35	22	cos	cos	PROPN
fcis-8581	35	23	θ	θ	PROPN
fcis-8581	35	24	(	(	PUNCT
fcis-8581	35	25	2	2	NUM
fcis-8581	35	26	)	)	PUNCT
fcis-8581	35	27	(	(	PUNCT
fcis-8581	35	28	3	3	X
fcis-8581	35	29	)	)	PUNCT
fcis-8581	35	30	pearson	pearson	NOUN
fcis-8581	35	31	correlation	correlation	NOUN
fcis-8581	35	32	coefficient	coefficient	PROPN
fcis-8581	35	33	pearson	pearson	PROPN
fcis-8581	35	34	correlation	correlation	PROPN
fcis-8581	35	35	coefficient	coefficient	NOUN
fcis-8581	35	36	,	,	PUNCT
fcis-8581	35	37	also	also	ADV
fcis-8581	35	38	known	know	VERB
fcis-8581	35	39	as	as	ADP
fcis-8581	35	40	pearson	pearson	PROPN
fcis-8581	35	41	product	product	NOUN
fcis-8581	35	42	-	-	PUNCT
fcis-8581	35	43	moment	moment	NOUN
fcis-8581	35	44	correlation	correlation	NOUN
fcis-8581	35	45	coefficient	coefficient	NOUN
fcis-8581	36	1	[	[	X
fcis-8581	36	2	6	6	NUM
fcis-8581	36	3	]	]	PUNCT
fcis-8581	36	4	,	,	PUNCT
fcis-8581	36	5	is	be	AUX
fcis-8581	36	6	a	a	DET
fcis-8581	36	7	measure	measure	NOUN
fcis-8581	36	8	of	of	ADP
fcis-8581	36	9	the	the	DET
fcis-8581	36	10	linear	linear	ADJ
fcis-8581	36	11	correlation	correlation	NOUN
fcis-8581	36	12	between	between	ADP
fcis-8581	36	13	two	two	NUM
fcis-8581	36	14	variables	variable	NOUN
fcis-8581	36	15	x	x	PUNCT
fcis-8581	36	16	and	and	CCONJ
fcis-8581	36	17	y	y	PROPN
fcis-8581	36	18	,	,	PUNCT
fcis-8581	36	19	with	with	ADP
fcis-8581	36	20	a	a	DET
fcis-8581	36	21	value	value	NOUN
fcis-8581	36	22	between	between	ADP
fcis-8581	36	23	-1	-1	PUNCT
fcis-8581	36	24	and	and	CCONJ
fcis-8581	36	25	1	1	X
fcis-8581	36	26	.	.	X
fcis-8581	37	1	the	the	DET
fcis-8581	37	2	pearson	pearson	PROPN
fcis-8581	37	3	correlation	correlation	NOUN
fcis-8581	37	4	coefficient	coefficient	NOUN
fcis-8581	37	5	between	between	ADP
fcis-8581	37	6	two	two	NUM
fcis-8581	37	7	variables	variable	NOUN
fcis-8581	37	8	is	be	AUX
fcis-8581	37	9	defined	define	VERB
fcis-8581	37	10	as	as	ADP
fcis-8581	37	11	the	the	DET
fcis-8581	37	12	quotient	quotient	NOUN
fcis-8581	37	13	of	of	ADP
fcis-8581	37	14	the	the	DET
fcis-8581	37	15	covariance	covariance	NOUN
fcis-8581	37	16	and	and	CCONJ
fcis-8581	37	17	standard	standard	ADJ
fcis-8581	37	18	deviation	deviation	NOUN
fcis-8581	37	19	between	between	ADP
fcis-8581	37	20	the	the	DET
fcis-8581	37	21	two	two	NUM
fcis-8581	37	22	variables	variable	NOUN
fcis-8581	37	23	and	and	CCONJ
fcis-8581	37	24	is	be	AUX
fcis-8581	37	25	calculated	calculate	VERB
fcis-8581	37	26	as	as	SCONJ
fcis-8581	37	27	shown	show	VERB
fcis-8581	37	28	in	in	ADP
fcis-8581	37	29	equation	equation	NOUN
fcis-8581	37	30	(	(	PUNCT
fcis-8581	37	31	3	3	NUM
fcis-8581	37	32	):	):	PUNCT
fcis-8581	37	33	ρ	ρ	PROPN
fcis-8581	37	34	x	x	PROPN
fcis-8581	37	35	,	,	PUNCT
fcis-8581	37	36	y	y	PROPN
fcis-8581	37	37	,	,	PUNCT
fcis-8581	37	38	(	(	PUNCT
fcis-8581	37	39	3	3	X
fcis-8581	37	40	)	)	PUNCT
fcis-8581	37	41	the	the	DET
fcis-8581	37	42	above	above	ADJ
fcis-8581	37	43	equation	equation	NOUN
fcis-8581	37	44	defines	define	VERB
fcis-8581	37	45	the	the	DET
fcis-8581	37	46	overall	overall	ADJ
fcis-8581	37	47	correlation	correlation	NOUN
fcis-8581	37	48	coefficient	coefficient	NOUN
fcis-8581	37	49	,	,	PUNCT
fcis-8581	37	50	which	which	PRON
fcis-8581	37	51	is	be	AUX
fcis-8581	37	52	usually	usually	ADV
fcis-8581	37	53	expressed	express	VERB
fcis-8581	37	54	using	use	VERB
fcis-8581	37	55	lowercase	lowercase	NOUN
fcis-8581	37	56	letters	letter	NOUN
fcis-8581	37	57	ρ	ρ	PROPN
fcis-8581	37	58	.	.	PUNCT
fcis-8581	38	1	estimating	estimate	VERB
fcis-8581	38	2	the	the	DET
fcis-8581	38	3	covariance	covariance	NOUN
fcis-8581	38	4	and	and	CCONJ
fcis-8581	38	5	standard	standard	ADJ
fcis-8581	38	6	deviation	deviation	NOUN
fcis-8581	38	7	of	of	ADP
fcis-8581	38	8	the	the	DET
fcis-8581	38	9	samples	sample	NOUN
fcis-8581	38	10	gives	give	VERB
fcis-8581	38	11	the	the	DET
fcis-8581	38	12	pearson	pearson	PROPN
fcis-8581	38	13	correlation	correlation	NOUN
fcis-8581	38	14	coefficient	coefficient	NOUN
fcis-8581	38	15	,	,	PUNCT
fcis-8581	38	16	which	which	PRON
fcis-8581	38	17	is	be	AUX
fcis-8581	38	18	often	often	ADV
fcis-8581	38	19	expressed	express	VERB
fcis-8581	38	20	in	in	ADP
fcis-8581	38	21	lowercase	lowercase	PROPN
fcis-8581	38	22	letters	letter	NOUN
fcis-8581	38	23	γ	γ	PROPN
fcis-8581	38	24	and	and	CCONJ
fcis-8581	38	25	calculated	calculate	VERB
fcis-8581	38	26	as	as	SCONJ
fcis-8581	38	27	shown	show	VERB
fcis-8581	38	28	in	in	ADP
fcis-8581	38	29	equation	equation	NOUN
fcis-8581	38	30	(	(	PUNCT
fcis-8581	38	31	4	4	NUM
fcis-8581	38	32	):	):	PUNCT
fcis-8581	38	33	γ	γ	X
fcis-8581	38	34	∑	∑	PROPN
fcis-8581	38	35	∑	∑	PROPN
fcis-8581	38	36	∑	∑	PUNCT
fcis-8581	38	37	(	(	PUNCT
fcis-8581	38	38	4	4	NUM
fcis-8581	38	39	)	)	PUNCT
fcis-8581	38	40	2.3	2.3	NUM
fcis-8581	38	41	.	.	PUNCT
fcis-8581	39	1	matrix	matrix	NOUN
fcis-8581	39	2	decomposition	decomposition	NOUN
fcis-8581	39	3	-	-	PUNCT
fcis-8581	39	4	based	base	VERB
fcis-8581	39	5	recommendation	recommendation	NOUN
fcis-8581	39	6	algorithm	algorithm	NOUN
fcis-8581	39	7	the	the	DET
fcis-8581	39	8	recommendation	recommendation	NOUN
fcis-8581	39	9	algorithm	algorithm	NOUN
fcis-8581	39	10	based	base	VERB
fcis-8581	39	11	on	on	ADP
fcis-8581	39	12	singular	singular	ADJ
fcis-8581	39	13	value	value	NOUN
fcis-8581	39	14	decomposition	decomposition	NOUN
fcis-8581	40	1	[	[	X
fcis-8581	40	2	7	7	X
fcis-8581	40	3	]	]	X
fcis-8581	40	4	(	(	PUNCT
fcis-8581	40	5	svd	svd	PROPN
fcis-8581	40	6	)	)	PUNCT
fcis-8581	40	7	is	be	AUX
fcis-8581	40	8	a	a	DET
fcis-8581	40	9	relatively	relatively	ADV
fcis-8581	40	10	novel	novel	ADJ
fcis-8581	40	11	algorithm	algorithm	NOUN
fcis-8581	40	12	in	in	ADP
fcis-8581	40	13	collaborative	collaborative	ADJ
fcis-8581	40	14	filtering	filtering	NOUN
fcis-8581	40	15	recommendation	recommendation	NOUN
fcis-8581	40	16	at	at	ADP
fcis-8581	40	17	present	present	ADJ
fcis-8581	40	18	.	.	PUNCT
fcis-8581	41	1	it	it	PRON
fcis-8581	41	2	is	be	AUX
fcis-8581	41	3	an	an	DET
fcis-8581	41	4	important	important	ADJ
fcis-8581	41	5	technique	technique	NOUN
fcis-8581	41	6	introduced	introduce	VERB
fcis-8581	41	7	from	from	ADP
fcis-8581	41	8	linear	linear	PROPN
fcis-8581	41	9	algebra	algebra	NOUN
fcis-8581	41	10	into	into	ADP
fcis-8581	41	11	machine	machine	NOUN
fcis-8581	41	12	learning	learning	NOUN
fcis-8581	41	13	.	.	PUNCT
fcis-8581	42	1	the	the	DET
fcis-8581	42	2	svd	svd	PROPN
fcis-8581	42	3	technique	technique	NOUN
fcis-8581	42	4	is	be	AUX
fcis-8581	42	5	used	use	VERB
fcis-8581	42	6	to	to	PART
fcis-8581	42	7	complement	complement	VERB
fcis-8581	42	8	the	the	DET
fcis-8581	42	9	user	user	NOUN
fcis-8581	42	10	-	-	PUNCT
fcis-8581	42	11	item	item	NOUN
fcis-8581	42	12	rating	rating	NOUN
fcis-8581	42	13	matrix	matrix	NOUN
fcis-8581	42	14	to	to	PART
fcis-8581	42	15	predict	predict	VERB
fcis-8581	42	16	user	user	NOUN
fcis-8581	42	17	ratings	rating	NOUN
fcis-8581	42	18	of	of	ADP
fcis-8581	42	19	items	item	NOUN
fcis-8581	42	20	,	,	PUNCT
fcis-8581	42	21	which	which	PRON
fcis-8581	42	22	requires	require	VERB
fcis-8581	42	23	us	we	PRON
fcis-8581	42	24	to	to	PART
fcis-8581	42	25	find	find	VERB
fcis-8581	42	26	a	a	DET
fcis-8581	42	27	complementary	complementary	ADJ
fcis-8581	42	28	method	method	NOUN
fcis-8581	42	29	that	that	PRON
fcis-8581	42	30	minimizes	minimize	VERB
fcis-8581	42	31	the	the	DET
fcis-8581	42	32	perturbation	perturbation	NOUN
fcis-8581	42	33	of	of	ADP
fcis-8581	42	34	the	the	DET
fcis-8581	42	35	original	original	ADJ
fcis-8581	42	36	matrix	matrix	NOUN
fcis-8581	42	37	.	.	PUNCT
fcis-8581	43	1	we	we	PRON
fcis-8581	43	2	generally	generally	ADV
fcis-8581	43	3	use	use	VERB
fcis-8581	43	4	the	the	DET
fcis-8581	43	5	difference	difference	NOUN
fcis-8581	43	6	in	in	ADP
fcis-8581	43	7	matrix	matrix	NOUN
fcis-8581	43	8	eigenvalues	eigenvalue	VERB
fcis-8581	43	9	to	to	PART
fcis-8581	43	10	represent	represent	VERB
fcis-8581	43	11	the	the	DET
fcis-8581	43	12	difference	difference	NOUN
fcis-8581	43	13	between	between	ADP
fcis-8581	43	14	the	the	DET
fcis-8581	43	15	complementary	complementary	ADJ
fcis-8581	43	16	matrix	matrix	NOUN
fcis-8581	43	17	and	and	CCONJ
fcis-8581	43	18	the	the	DET
fcis-8581	43	19	original	original	ADJ
fcis-8581	43	20	matrix	matrix	NOUN
fcis-8581	43	21	.	.	PUNCT
fcis-8581	44	1	3	3	X
fcis-8581	44	2	.	.	X
fcis-8581	44	3	algorithm	algorithm	NOUN
fcis-8581	44	4	improvement	improvement	NOUN
fcis-8581	44	5	3.1	3.1	NUM
fcis-8581	44	6	.	.	PUNCT
fcis-8581	45	1	algorithmic	algorithmic	ADJ
fcis-8581	45	2	ideas	idea	NOUN
fcis-8581	45	3	the	the	DET
fcis-8581	45	4	goal	goal	NOUN
fcis-8581	45	5	of	of	ADP
fcis-8581	45	6	this	this	DET
fcis-8581	45	7	paper	paper	NOUN
fcis-8581	45	8	is	be	AUX
fcis-8581	45	9	to	to	PART
fcis-8581	45	10	implement	implement	VERB
fcis-8581	45	11	a	a	DET
fcis-8581	45	12	personalized	personalize	VERB
fcis-8581	45	13	movie	movie	NOUN
fcis-8581	45	14	recommendation	recommendation	NOUN
fcis-8581	45	15	algorithm	algorithm	NOUN
fcis-8581	45	16	based	base	VERB
fcis-8581	45	17	on	on	ADP
fcis-8581	45	18	the	the	DET
fcis-8581	45	19	movie	movie	NOUN
fcis-8581	45	20	dataset	dataset	NOUN
fcis-8581	45	21	of	of	ADP
fcis-8581	45	22	movielens	movielen	NOUN
fcis-8581	45	23	website	website	NOUN
fcis-8581	45	24	.	.	PUNCT
fcis-8581	46	1	according	accord	VERB
fcis-8581	46	2	to	to	ADP
fcis-8581	46	3	the	the	DET
fcis-8581	46	4	characteristics	characteristic	NOUN
fcis-8581	46	5	of	of	ADP
fcis-8581	46	6	movie	movie	NOUN
fcis-8581	46	7	recommendation	recommendation	NOUN
fcis-8581	46	8	,	,	PUNCT
fcis-8581	46	9	the	the	DET
fcis-8581	46	10	similarity	similarity	NOUN
fcis-8581	46	11	measure	measure	NOUN
fcis-8581	46	12	used	use	VERB
fcis-8581	46	13	in	in	ADP
fcis-8581	46	14	the	the	DET
fcis-8581	46	15	collaborative	collaborative	ADJ
fcis-8581	46	16	filtering	filtering	NOUN
fcis-8581	46	17	model	model	NOUN
fcis-8581	46	18	is	be	AUX
fcis-8581	46	19	gradually	gradually	ADV
fcis-8581	46	20	improved	improve	VERB
fcis-8581	46	21	,	,	PUNCT
fcis-8581	46	22	and	and	CCONJ
fcis-8581	46	23	the	the	DET
fcis-8581	46	24	number	number	NOUN
fcis-8581	46	25	of	of	ADP
fcis-8581	46	26	users	user	NOUN
fcis-8581	46	27	who	who	PRON
fcis-8581	46	28	jointly	jointly	ADV
fcis-8581	46	29	evaluate	evaluate	VERB
fcis-8581	46	30	two	two	NUM
fcis-8581	46	31	movies	movie	NOUN
fcis-8581	46	32	based	base	VERB
fcis-8581	46	33	on	on	ADP
fcis-8581	46	34	their	their	PRON
fcis-8581	46	35	ratings	rating	NOUN
fcis-8581	46	36	will	will	AUX
fcis-8581	46	37	be	be	AUX
fcis-8581	46	38	used	use	VERB
fcis-8581	46	39	to	to	PART
fcis-8581	46	40	determine	determine	VERB
fcis-8581	46	41	the	the	DET
fcis-8581	46	42	degree	degree	NOUN
fcis-8581	46	43	of	of	ADP
fcis-8581	46	44	similarity	similarity	NOUN
fcis-8581	46	45	between	between	ADP
fcis-8581	46	46	two	two	NUM
fcis-8581	46	47	movies	movie	NOUN
fcis-8581	46	48	more	more	ADV
fcis-8581	46	49	precisely	precisely	ADV
fcis-8581	46	50	.	.	PUNCT
fcis-8581	47	1	in	in	ADP
fcis-8581	47	2	a	a	DET
fcis-8581	47	3	movie	movie	NOUN
fcis-8581	47	4	recommendation	recommendation	NOUN
fcis-8581	47	5	website	website	NOUN
fcis-8581	47	6	,	,	PUNCT
fcis-8581	47	7	each	each	DET
fcis-8581	47	8	user	user	NOUN
fcis-8581	47	9	generally	generally	ADV
fcis-8581	47	10	sees	see	VERB
fcis-8581	47	11	only	only	ADV
fcis-8581	47	12	a	a	DET
fcis-8581	47	13	small	small	ADJ
fcis-8581	47	14	fraction	fraction	NOUN
fcis-8581	47	15	of	of	ADP
fcis-8581	47	16	all	all	DET
fcis-8581	47	17	movies	movie	NOUN
fcis-8581	47	18	in	in	ADP
fcis-8581	47	19	the	the	DET
fcis-8581	47	20	dataset	dataset	NOUN
fcis-8581	47	21	,	,	PUNCT
fcis-8581	47	22	and	and	CCONJ
fcis-8581	47	23	the	the	DET
fcis-8581	47	24	movielens	movielen	NOUN
fcis-8581	47	25	dataset	dataset	VERB
fcis-8581	47	26	used	use	VERB
fcis-8581	47	27	in	in	ADP
fcis-8581	47	28	this	this	DET
fcis-8581	47	29	paper	paper	NOUN
fcis-8581	47	30	has	have	AUX
fcis-8581	47	31	filtered	filter	VERB
fcis-8581	47	32	out	out	ADP
fcis-8581	47	33	users	user	NOUN
fcis-8581	47	34	with	with	ADP
fcis-8581	47	35	less	less	ADJ
fcis-8581	47	36	than	than	ADP
fcis-8581	47	37	20	20	NUM
fcis-8581	47	38	reviews	review	NOUN
fcis-8581	47	39	.	.	PUNCT
fcis-8581	48	1	the	the	DET
fcis-8581	48	2	statistics	statistic	NOUN
fcis-8581	48	3	reveal	reveal	VERB
fcis-8581	48	4	that	that	SCONJ
fcis-8581	48	5	each	each	DET
fcis-8581	48	6	user	user	NOUN
fcis-8581	48	7	has	have	AUX
fcis-8581	48	8	rated	rate	VERB
fcis-8581	48	9	about	about	ADV
fcis-8581	48	10	100	100	NUM
fcis-8581	48	11	movies	movie	NOUN
fcis-8581	48	12	on	on	ADP
fcis-8581	48	13	average	average	ADJ
fcis-8581	48	14	,	,	PUNCT
fcis-8581	48	15	and	and	CCONJ
fcis-8581	48	16	the	the	DET
fcis-8581	48	17	sparsity	sparsity	NOUN
fcis-8581	48	18	of	of	ADP
fcis-8581	48	19	the	the	DET
fcis-8581	48	20	data	datum	NOUN
fcis-8581	48	21	can	can	AUX
fcis-8581	48	22	be	be	AUX
fcis-8581	48	23	expressed	express	VERB
fcis-8581	48	24	using	use	VERB
fcis-8581	48	25	equation	equation	NOUN
fcis-8581	48	26	(	(	PUNCT
fcis-8581	48	27	5	5	NUM
fcis-8581	48	28	)	)	PUNCT
fcis-8581	48	29	.	.	PUNCT
fcis-8581	49	1	the	the	DET
fcis-8581	49	2	sparsity	sparsity	NOUN
fcis-8581	49	3	in	in	ADP
fcis-8581	49	4	the	the	DET
fcis-8581	49	5	movielens	movielen	NOUN
fcis-8581	49	6	dataset	dataset	VERB
fcis-8581	49	7	is	be	AUX
fcis-8581	49	8	about	about	ADV
fcis-8581	49	9	99.83	99.83	NUM
fcis-8581	49	10	%	%	NOUN
fcis-8581	49	11	.	.	PUNCT
fcis-8581	50	1	thus	thus	ADV
fcis-8581	50	2	,	,	PUNCT
fcis-8581	50	3	the	the	DET
fcis-8581	50	4	rating	rating	NOUN
fcis-8581	50	5	matrix	matrix	NOUN
fcis-8581	50	6	of	of	ADP
fcis-8581	50	7	user	user	NOUN
fcis-8581	50	8	movies	movie	NOUN
fcis-8581	50	9	is	be	AUX
fcis-8581	50	10	a	a	DET
fcis-8581	50	11	very	very	ADV
fcis-8581	50	12	sparse	sparse	ADJ
fcis-8581	50	13	matrix	matrix	NOUN
fcis-8581	50	14	,	,	PUNCT
fcis-8581	50	15	and	and	CCONJ
fcis-8581	50	16	it	it	PRON
fcis-8581	50	17	is	be	AUX
fcis-8581	50	18	possible	possible	ADJ
fcis-8581	50	19	that	that	SCONJ
fcis-8581	50	20	only	only	ADV
fcis-8581	50	21	few	few	ADJ
fcis-8581	50	22	users	user	NOUN
fcis-8581	50	23	rate	rate	VERB
fcis-8581	50	24	a	a	DET
fcis-8581	50	25	movie	movie	NOUN
fcis-8581	50	26	.	.	PUNCT
fcis-8581	51	1	also	also	ADV
fcis-8581	51	2	,	,	PUNCT
fcis-8581	51	3	when	when	SCONJ
fcis-8581	51	4	a	a	DET
fcis-8581	51	5	new	new	ADJ
fcis-8581	51	6	movie	movie	NOUN
fcis-8581	51	7	is	be	AUX
fcis-8581	51	8	added	add	VERB
fcis-8581	51	9	to	to	ADP
fcis-8581	51	10	a	a	DET
fcis-8581	51	11	movie	movie	NOUN
fcis-8581	51	12	website	website	NOUN
fcis-8581	51	13	,	,	PUNCT
fcis-8581	51	14	the	the	DET
fcis-8581	51	15	sparsity	sparsity	NOUN
fcis-8581	51	16	of	of	ADP
fcis-8581	51	17	data	datum	NOUN
fcis-8581	51	18	is	be	AUX
fcis-8581	51	19	further	far	ADV
fcis-8581	51	20	increased	increase	VERB
fcis-8581	51	21	because	because	SCONJ
fcis-8581	51	22	there	there	PRON
fcis-8581	51	23	is	be	VERB
fcis-8581	51	24	no	no	DET
fcis-8581	51	25	rating	rating	NOUN
fcis-8581	51	26	record	record	NOUN
fcis-8581	51	27	about	about	ADP
fcis-8581	51	28	this	this	DET
fcis-8581	51	29	movie	movie	NOUN
fcis-8581	51	30	,	,	PUNCT
fcis-8581	51	31	also	also	ADV
fcis-8581	51	32	known	know	VERB
fcis-8581	51	33	as	as	ADP
fcis-8581	51	34	the	the	DET
fcis-8581	51	35	cold	cold	ADJ
fcis-8581	51	36	start	start	NOUN
fcis-8581	51	37	problem	problem	NOUN
fcis-8581	51	38	(	(	PUNCT
fcis-8581	51	39	lack	lack	NOUN
fcis-8581	51	40	of	of	ADP
fcis-8581	51	41	data	datum	NOUN
fcis-8581	51	42	information	information	NOUN
fcis-8581	51	43	to	to	PART
fcis-8581	51	44	carry	carry	VERB
fcis-8581	51	45	out	out	ADP
fcis-8581	51	46	personalized	personalized	ADJ
fcis-8581	51	47	recommendation	recommendation	NOUN
fcis-8581	51	48	service	service	NOUN
fcis-8581	51	49	)	)	PUNCT
fcis-8581	51	50	.	.	PUNCT
fcis-8581	52	1	1	1	NUM
fcis-8581	52	2	(	(	PUNCT
fcis-8581	52	3	5	5	NUM
fcis-8581	52	4	)	)	PUNCT
fcis-8581	52	5	to	to	PART
fcis-8581	52	6	solve	solve	VERB
fcis-8581	52	7	the	the	DET
fcis-8581	52	8	problem	problem	NOUN
fcis-8581	52	9	of	of	ADP
fcis-8581	52	10	sparse	sparse	ADJ
fcis-8581	52	11	data	datum	NOUN
fcis-8581	52	12	,	,	PUNCT
fcis-8581	52	13	content	content	NOUN
fcis-8581	52	14	-	-	PUNCT
fcis-8581	52	15	based	base	VERB
fcis-8581	52	16	recommendation	recommendation	NOUN
fcis-8581	52	17	is	be	AUX
fcis-8581	52	18	introduced	introduce	VERB
fcis-8581	52	19	to	to	PART
fcis-8581	52	20	describe	describe	VERB
fcis-8581	52	21	a	a	DET
fcis-8581	52	22	movie	movie	NOUN
fcis-8581	52	23	using	use	VERB
fcis-8581	52	24	its	its	PRON
fcis-8581	52	25	attributes	attribute	NOUN
fcis-8581	52	26	,	,	PUNCT
fcis-8581	52	27	which	which	PRON
fcis-8581	52	28	include	include	VERB
fcis-8581	52	29	movie	movie	NOUN
fcis-8581	52	30	title	title	NOUN
fcis-8581	52	31	,	,	PUNCT
fcis-8581	52	32	genre	genre	NOUN
fcis-8581	52	33	,	,	PUNCT
fcis-8581	52	34	and	and	CCONJ
fcis-8581	52	35	other	other	ADJ
fcis-8581	52	36	movie	movie	NOUN
fcis-8581	52	37	information	information	NOUN
fcis-8581	52	38	.	.	PUNCT
fcis-8581	53	1	finally	finally	ADV
fcis-8581	53	2	,	,	PUNCT
fcis-8581	53	3	a	a	DET
fcis-8581	53	4	combination	combination	NOUN
fcis-8581	53	5	of	of	ADP
fcis-8581	53	6	content	content	NOUN
fcis-8581	53	7	-	-	PUNCT
fcis-8581	53	8	based	base	VERB
fcis-8581	53	9	recommendations	recommendation	NOUN
fcis-8581	53	10	and	and	CCONJ
fcis-8581	53	11	matrix	matrix	NOUN
fcis-8581	53	12	decomposition	decomposition	NOUN
fcis-8581	53	13	-	-	PUNCT
fcis-8581	53	14	based	base	VERB
fcis-8581	53	15	recommendations	recommendation	NOUN
fcis-8581	53	16	are	be	AUX
fcis-8581	53	17	used	use	VERB
fcis-8581	53	18	to	to	PART
fcis-8581	53	19	make	make	VERB
fcis-8581	53	20	hybrid	hybrid	ADJ
fcis-8581	53	21	recommendations	recommendation	NOUN
fcis-8581	53	22	.	.	PUNCT
fcis-8581	54	1	3.2	3.2	NUM
fcis-8581	54	2	.	.	PUNCT
fcis-8581	55	1	similarity	similarity	NOUN
fcis-8581	55	2	algorithm	algorithm	NOUN
fcis-8581	55	3	improvement	improvement	NOUN
fcis-8581	55	4	the	the	DET
fcis-8581	55	5	similarity	similarity	NOUN
fcis-8581	55	6	algorithm	algorithm	NOUN
fcis-8581	55	7	of	of	ADP
fcis-8581	55	8	this	this	DET
fcis-8581	55	9	system	system	NOUN
fcis-8581	55	10	uses	use	VERB
fcis-8581	55	11	the	the	DET
fcis-8581	55	12	cosine	cosine	NOUN
fcis-8581	55	13	similarity	similarity	NOUN
fcis-8581	55	14	algorithm	algorithm	NOUN
fcis-8581	55	15	,	,	PUNCT
fcis-8581	55	16	considering	consider	VERB
fcis-8581	55	17	that	that	SCONJ
fcis-8581	55	18	each	each	DET
fcis-8581	55	19	data	data	NOUN
fcis-8581	55	20	record	record	NOUN
fcis-8581	55	21	in	in	ADP
fcis-8581	55	22	the	the	DET
fcis-8581	55	23	data	data	NOUN
fcis-8581	55	24	set	set	VERB
fcis-8581	55	25	keeps	keep	VERB
fcis-8581	55	26	a	a	DET
fcis-8581	55	27	time	time	NOUN
fcis-8581	55	28	stamp	stamp	NOUN
fcis-8581	55	29	,	,	PUNCT
fcis-8581	55	30	and	and	CCONJ
fcis-8581	55	31	according	accord	VERB
fcis-8581	55	32	to	to	ADP
fcis-8581	55	33	the	the	DET
fcis-8581	55	34	characteristics	characteristic	NOUN
fcis-8581	55	35	of	of	ADP
fcis-8581	55	36	human	human	ADJ
fcis-8581	55	37	memory	memory	NOUN
fcis-8581	55	38	,	,	PUNCT
fcis-8581	55	39	the	the	DET
fcis-8581	55	40	interests	interest	NOUN
fcis-8581	55	41	of	of	ADP
fcis-8581	55	42	each	each	DET
fcis-8581	55	43	user	user	NOUN
fcis-8581	55	44	will	will	AUX
fcis-8581	55	45	change	change	VERB
fcis-8581	55	46	accordingly	accordingly	ADV
fcis-8581	55	47	over	over	ADP
fcis-8581	55	48	time	time	NOUN
fcis-8581	55	49	,	,	PUNCT
fcis-8581	55	50	and	and	CCONJ
fcis-8581	55	51	the	the	PRON
fcis-8581	55	52	longer	long	ADJ
fcis-8581	55	53	the	the	DET
fcis-8581	55	54	interval	interval	NOUN
fcis-8581	55	55	of	of	ADP
fcis-8581	55	56	data	data	NOUN
fcis-8581	55	57	is	be	AUX
fcis-8581	55	58	,	,	PUNCT
fcis-8581	55	59	the	the	PRON
fcis-8581	55	60	more	more	ADV
fcis-8581	55	61	unreliable	unreliable	ADJ
fcis-8581	55	62	the	the	DET
fcis-8581	55	63	recommendation	recommendation	NOUN
fcis-8581	55	64	of	of	ADP
fcis-8581	55	65	this	this	DET
fcis-8581	55	66	system	system	NOUN
fcis-8581	55	67	is	be	AUX
fcis-8581	55	68	,	,	PUNCT
fcis-8581	55	69	therefore	therefore	ADV
fcis-8581	55	70	,	,	PUNCT
fcis-8581	55	71	the	the	DET
fcis-8581	55	72	time	time	NOUN
fcis-8581	55	73	factor	factor	NOUN
fcis-8581	55	74	of	of	ADP
fcis-8581	55	75	the	the	DET
fcis-8581	55	76	rating	rating	NOUN
fcis-8581	55	77	is	be	AUX
fcis-8581	55	78	taken	take	VERB
fcis-8581	55	79	into	into	ADP
fcis-8581	55	80	account	account	NOUN
fcis-8581	55	81	to	to	PART
fcis-8581	55	82	improve	improve	VERB
fcis-8581	55	83	the	the	DET
fcis-8581	55	84	cosine	cosine	NOUN
fcis-8581	55	85	similarity	similarity	NOUN
fcis-8581	55	86	calculation	calculation	NOUN
fcis-8581	55	87	method	method	NOUN
fcis-8581	55	88	to	to	PART
fcis-8581	55	89	achieve	achieve	VERB
fcis-8581	55	90	the	the	DET
fcis-8581	55	91	expected	expect	VERB
fcis-8581	55	92	purpose	purpose	NOUN
fcis-8581	55	93	.	.	PUNCT
fcis-8581	56	1	in	in	ADP
fcis-8581	56	2	this	this	DET
fcis-8581	56	3	paper	paper	NOUN
fcis-8581	56	4	,	,	PUNCT
fcis-8581	56	5	we	we	PRON
fcis-8581	56	6	use	use	VERB
fcis-8581	56	7	the	the	DET
fcis-8581	56	8	time	time	NOUN
fcis-8581	56	9	interval	interval	NOUN
fcis-8581	56	10	between	between	ADP
fcis-8581	56	11	users	user	NOUN
fcis-8581	56	12	'	'	PART
fcis-8581	56	13	ratings	rating	NOUN
fcis-8581	56	14	of	of	ADP
fcis-8581	56	15	items	item	NOUN
fcis-8581	56	16	to	to	PART
fcis-8581	56	17	optimize	optimize	VERB
fcis-8581	56	18	the	the	DET
fcis-8581	56	19	data	datum	NOUN
fcis-8581	56	20	.	.	PUNCT
fcis-8581	57	1	when	when	SCONJ
fcis-8581	57	2	the	the	DET
fcis-8581	57	3	time	time	NOUN
fcis-8581	57	4	interval	interval	NOUN
fcis-8581	57	5	between	between	ADP
fcis-8581	57	6	users	user	NOUN
fcis-8581	57	7	'	'	PART
fcis-8581	57	8	ratings	rating	NOUN
fcis-8581	57	9	of	of	ADP
fcis-8581	57	10	two	two	NUM
fcis-8581	57	11	items	item	NOUN
fcis-8581	57	12	is	be	AUX
fcis-8581	57	13	larger	large	ADJ
fcis-8581	57	14	,	,	PUNCT
fcis-8581	57	15	the	the	DET
fcis-8581	57	16	probability	probability	NOUN
fcis-8581	57	17	of	of	ADP
fcis-8581	57	18	users	user	NOUN
fcis-8581	57	19	'	'	PART
fcis-8581	57	20	interests	interest	NOUN
fcis-8581	57	21	changing	change	VERB
fcis-8581	57	22	is	be	AUX
fcis-8581	57	23	larger	large	ADJ
fcis-8581	57	24	,	,	PUNCT
fcis-8581	57	25	and	and	CCONJ
fcis-8581	57	26	the	the	DET
fcis-8581	57	27	error	error	NOUN
fcis-8581	57	28	after	after	ADP
fcis-8581	57	29	using	use	VERB
fcis-8581	57	30	the	the	DET
fcis-8581	57	31	original	original	ADJ
fcis-8581	57	32	rating	rating	NOUN
fcis-8581	57	33	data	datum	NOUN
fcis-8581	57	34	for	for	ADP
fcis-8581	57	35	similarity	similarity	NOUN
fcis-8581	57	36	calculation	calculation	NOUN
fcis-8581	57	37	recommendation	recommendation	NOUN
fcis-8581	57	38	is	be	AUX
fcis-8581	57	39	larger	large	ADJ
fcis-8581	57	40	.	.	PUNCT
fcis-8581	58	1	on	on	ADP
fcis-8581	58	2	the	the	DET
fcis-8581	58	3	contrary	contrary	NOUN
fcis-8581	58	4	,	,	PUNCT
fcis-8581	58	5	the	the	DET
fcis-8581	58	6	more	more	ADV
fcis-8581	58	7	recent	recent	ADJ
fcis-8581	58	8	the	the	DET
fcis-8581	58	9	user	user	NOUN
fcis-8581	58	10	rating	rating	NOUN
fcis-8581	58	11	data	datum	NOUN
fcis-8581	58	12	timestamp	timestamp	NOUN
fcis-8581	58	13	is	be	AUX
fcis-8581	58	14	from	from	ADP
fcis-8581	58	15	now	now	ADV
fcis-8581	58	16	,	,	PUNCT
fcis-8581	58	17	the	the	PRON
fcis-8581	58	18	more	more	ADV
fcis-8581	58	19	informative	informative	ADJ
fcis-8581	58	20	this	this	DET
fcis-8581	58	21	record	record	NOUN
fcis-8581	58	22	is	be	AUX
fcis-8581	58	23	and	and	CCONJ
fcis-8581	58	24	the	the	DET
fcis-8581	58	25	more	more	ADV
fcis-8581	58	26	reliable	reliable	ADJ
fcis-8581	58	27	the	the	DET
fcis-8581	58	28	recommended	recommend	VERB
fcis-8581	58	29	results	result	NOUN
fcis-8581	58	30	are	be	AUX
fcis-8581	58	31	.	.	PUNCT
fcis-8581	59	1	based	base	VERB
fcis-8581	59	2	on	on	ADP
fcis-8581	59	3	the	the	DET
fcis-8581	59	4	above	above	ADJ
fcis-8581	59	5	description	description	NOUN
fcis-8581	59	6	,	,	PUNCT
fcis-8581	59	7	this	this	DET
fcis-8581	59	8	paper	paper	NOUN
fcis-8581	59	9	introduces	introduce	VERB
fcis-8581	59	10	a	a	DET
fcis-8581	59	11	temporal	temporal	ADJ
fcis-8581	59	12	factor	factor	NOUN
fcis-8581	59	13	[	[	NOUN
fcis-8581	59	14	4	4	NUM
fcis-8581	59	15	]	]	PUNCT
fcis-8581	59	16	to	to	PART
fcis-8581	59	17	reflect	reflect	VERB
fcis-8581	59	18	the	the	DET
fcis-8581	59	19	changes	change	NOUN
fcis-8581	59	20	of	of	ADP
fcis-8581	59	21	users	user	NOUN
fcis-8581	59	22	'	'	PART
fcis-8581	59	23	interests	interest	NOUN
fcis-8581	59	24	over	over	ADP
fcis-8581	59	25	time	time	NOUN
fcis-8581	59	26	.	.	PUNCT
fcis-8581	60	1	assuming	assume	VERB
fcis-8581	60	2	that	that	SCONJ
fcis-8581	60	3	the	the	DET
fcis-8581	60	4	user	user	NOUN
fcis-8581	60	5	's	's	PART
fcis-8581	60	6	rating	rating	NOUN
fcis-8581	60	7	time	time	NOUN
fcis-8581	60	8	for	for	ADP
fcis-8581	60	9	movie	movie	NOUN
fcis-8581	60	10	x	x	PUNCT
fcis-8581	60	11	is	be	AUX
fcis-8581	60	12	tx	tx	PROPN
fcis-8581	60	13	and	and	CCONJ
fcis-8581	60	14	the	the	DET
fcis-8581	60	15	user	user	NOUN
fcis-8581	60	16	's	's	PART
fcis-8581	60	17	rating	rating	NOUN
fcis-8581	60	18	time	time	NOUN
fcis-8581	60	19	for	for	ADP
fcis-8581	60	20	movie	movie	NOUN
fcis-8581	60	21	y	y	PROPN
fcis-8581	60	22	is	be	AUX
fcis-8581	60	23	ty	ty	INTJ
fcis-8581	60	24	,	,	PUNCT
fcis-8581	60	25	the	the	DET
fcis-8581	60	26	formula	formula	NOUN
fcis-8581	60	27	for	for	ADP
fcis-8581	60	28	the	the	DET
fcis-8581	60	29	time	time	NOUN
fcis-8581	60	30	factor	factor	NOUN
fcis-8581	60	31	in	in	ADP
fcis-8581	60	32	this	this	DET
fcis-8581	60	33	paper	paper	NOUN
fcis-8581	60	34	is	be	AUX
fcis-8581	60	35	shown	show	VERB
fcis-8581	60	36	in	in	ADP
fcis-8581	60	37	equation	equation	NOUN
fcis-8581	60	38	(	(	PUNCT
fcis-8581	60	39	6	6	NUM
fcis-8581	60	40	)	)	PUNCT
fcis-8581	60	41	.	.	PUNCT
fcis-8581	61	1	the	the	PRON
fcis-8581	61	2	smaller	small	ADJ
fcis-8581	61	3	the	the	DET
fcis-8581	61	4	time	time	NOUN
fcis-8581	61	5	interval	interval	NOUN
fcis-8581	61	6	,	,	PUNCT
fcis-8581	61	7	the	the	PRON
fcis-8581	61	8	smaller	small	ADJ
fcis-8581	61	9	the	the	DET
fcis-8581	61	10	time	time	NOUN
fcis-8581	61	11	factor	factor	NOUN
fcis-8581	61	12	,	,	PUNCT
fcis-8581	61	13	and	and	CCONJ
fcis-8581	61	14	the	the	DET
fcis-8581	61	15	minimum	minimum	ADJ
fcis-8581	61	16	case	case	NOUN
fcis-8581	61	17	is	be	AUX
fcis-8581	61	18	the	the	DET
fcis-8581	61	19	time	time	NOUN
fcis-8581	61	20	factor	factor	NOUN
fcis-8581	61	21	result	result	NOUN
fcis-8581	61	22	of	of	ADP
fcis-8581	61	23	1	1	NUM
fcis-8581	61	24	when	when	SCONJ
fcis-8581	61	25	the	the	DET
fcis-8581	61	26	interval	interval	NOUN
fcis-8581	61	27	time	time	NOUN
fcis-8581	61	28	is	be	AUX
fcis-8581	61	29	0	0	NUM
fcis-8581	61	30	.	.	X
fcis-8581	62	1	α=	α=	NOUN
fcis-8581	62	2	in	in	ADP
fcis-8581	62	3	e+	e+	ADJ
fcis-8581	62	4	|tx	|tx	NOUN
fcis-8581	62	5	-	-	ADJ
fcis-8581	62	6	ty|	ty|	ADJ
fcis-8581	62	7	(	(	PUNCT
fcis-8581	62	8	6	6	NUM
fcis-8581	62	9	)	)	PUNCT
fcis-8581	62	10	the	the	DET
fcis-8581	62	11	formula	formula	NOUN
fcis-8581	62	12	for	for	ADP
fcis-8581	62	13	calculating	calculate	VERB
fcis-8581	62	14	the	the	DET
fcis-8581	62	15	cosine	cosine	NOUN
fcis-8581	62	16	similarity	similarity	NOUN
fcis-8581	62	17	after	after	ADP
fcis-8581	62	18	adding	add	VERB
fcis-8581	62	19	the	the	DET
fcis-8581	62	20	time	time	NOUN
fcis-8581	62	21	factor	factor	NOUN
fcis-8581	62	22	is	be	AUX
fcis-8581	62	23	shown	show	VERB
fcis-8581	62	24	in	in	ADP
fcis-8581	62	25	equation	equation	NOUN
fcis-8581	62	26	(	(	PUNCT
fcis-8581	62	27	7	7	NUM
fcis-8581	62	28	)	)	PUNCT
fcis-8581	62	29	.	.	PUNCT
fcis-8581	63	1	115	115	NUM
fcis-8581	64	1	cos	cos	ADP
fcis-8581	64	2	∙	∙	PROPN
fcis-8581	65	1	|	|	ADV
fcis-8581	66	1	|	|	ADV
fcis-8581	66	2	|	|	ADV
fcis-8581	66	3	|	|	ADV
fcis-8581	66	4	∑	∑	ADV
fcis-8581	66	5	∑	∑	PROPN
fcis-8581	66	6	∑	∑	PUNCT
fcis-8581	66	7	(	(	PUNCT
fcis-8581	66	8	7	7	NUM
fcis-8581	66	9	)	)	PUNCT
fcis-8581	66	10	3.3	3.3	NUM
fcis-8581	66	11	.	.	PUNCT
fcis-8581	67	1	recommendation	recommendation	NOUN
fcis-8581	67	2	algorithm	algorithm	NOUN
fcis-8581	67	3	improvement	improvement	NOUN
fcis-8581	67	4	3.3.1	3.3.1	NUM
fcis-8581	67	5	.	.	PUNCT
fcis-8581	68	1	user	user	NOUN
fcis-8581	68	2	-	-	PUNCT
fcis-8581	68	3	based	base	VERB
fcis-8581	68	4	recommendation	recommendation	NOUN
fcis-8581	68	5	algorithm	algorithm	NOUN
fcis-8581	68	6	it	it	PRON
fcis-8581	68	7	is	be	AUX
fcis-8581	68	8	assumed	assume	VERB
fcis-8581	68	9	that	that	SCONJ
fcis-8581	68	10	users	user	NOUN
fcis-8581	68	11	with	with	ADP
fcis-8581	68	12	similar	similar	ADJ
fcis-8581	68	13	interest	interest	NOUN
fcis-8581	68	14	preferences	preference	NOUN
fcis-8581	68	15	may	may	AUX
fcis-8581	68	16	show	show	VERB
fcis-8581	68	17	some	some	DET
fcis-8581	68	18	degree	degree	NOUN
fcis-8581	68	19	of	of	ADP
fcis-8581	68	20	interest	interest	NOUN
fcis-8581	68	21	preference	preference	NOUN
fcis-8581	68	22	for	for	ADP
fcis-8581	68	23	the	the	DET
fcis-8581	68	24	same	same	ADJ
fcis-8581	68	25	items	item	NOUN
fcis-8581	68	26	among	among	ADP
fcis-8581	68	27	themselves	themselves	PRON
fcis-8581	68	28	.	.	PUNCT
fcis-8581	69	1	the	the	DET
fcis-8581	69	2	basic	basic	ADJ
fcis-8581	69	3	principle	principle	NOUN
fcis-8581	69	4	of	of	ADP
fcis-8581	69	5	user	user	NOUN
fcis-8581	69	6	-	-	PUNCT
fcis-8581	69	7	based	base	VERB
fcis-8581	69	8	recommendation	recommendation	NOUN
fcis-8581	69	9	algorithm	algorithm	NOUN
fcis-8581	69	10	is	be	AUX
fcis-8581	69	11	to	to	PART
fcis-8581	69	12	find	find	VERB
fcis-8581	69	13	users	user	NOUN
fcis-8581	69	14	with	with	ADP
fcis-8581	69	15	similar	similar	ADJ
fcis-8581	69	16	preferences	preference	NOUN
fcis-8581	69	17	for	for	ADP
fcis-8581	69	18	the	the	DET
fcis-8581	69	19	target	target	NOUN
fcis-8581	69	20	object	object	NOUN
fcis-8581	69	21	based	base	VERB
fcis-8581	69	22	on	on	ADP
fcis-8581	69	23	the	the	DET
fcis-8581	69	24	evaluation	evaluation	NOUN
fcis-8581	69	25	data	datum	NOUN
fcis-8581	69	26	recorded	record	VERB
fcis-8581	69	27	in	in	ADP
fcis-8581	69	28	the	the	DET
fcis-8581	69	29	system	system	NOUN
fcis-8581	69	30	,	,	PUNCT
fcis-8581	69	31	and	and	CCONJ
fcis-8581	69	32	then	then	ADV
fcis-8581	69	33	generate	generate	VERB
fcis-8581	69	34	recommendation	recommendation	NOUN
fcis-8581	69	35	results	result	NOUN
fcis-8581	69	36	for	for	ADP
fcis-8581	69	37	the	the	DET
fcis-8581	69	38	target	target	NOUN
fcis-8581	69	39	object	object	NOUN
fcis-8581	69	40	based	base	VERB
fcis-8581	69	41	on	on	ADP
fcis-8581	69	42	the	the	DET
fcis-8581	69	43	preference	preference	NOUN
fcis-8581	69	44	information	information	NOUN
fcis-8581	69	45	possessed	possess	VERB
fcis-8581	69	46	by	by	ADP
fcis-8581	69	47	the	the	DET
fcis-8581	69	48	near	near	ADJ
fcis-8581	69	49	-	-	PUNCT
fcis-8581	69	50	neighbor	neighbor	NOUN
fcis-8581	69	51	users	user	NOUN
fcis-8581	69	52	.	.	PUNCT
fcis-8581	70	1	the	the	DET
fcis-8581	70	2	main	main	ADJ
fcis-8581	70	3	steps	step	NOUN
fcis-8581	70	4	are	be	AUX
fcis-8581	70	5	as	as	SCONJ
fcis-8581	70	6	follows	follow	VERB
fcis-8581	70	7	:	:	PUNCT
fcis-8581	70	8	(	(	PUNCT
fcis-8581	70	9	1	1	X
fcis-8581	70	10	)	)	PUNCT
fcis-8581	70	11	establishing	establish	VERB
fcis-8581	70	12	a	a	DET
fcis-8581	70	13	scoring	scoring	NOUN
fcis-8581	70	14	model	model	NOUN
fcis-8581	70	15	this	this	DET
fcis-8581	70	16	step	step	NOUN
fcis-8581	70	17	is	be	AUX
fcis-8581	70	18	the	the	DET
fcis-8581	70	19	basic	basic	ADJ
fcis-8581	70	20	step	step	NOUN
fcis-8581	70	21	of	of	ADP
fcis-8581	70	22	the	the	DET
fcis-8581	70	23	algorithm	algorithm	NOUN
fcis-8581	70	24	,	,	PUNCT
fcis-8581	70	25	which	which	PRON
fcis-8581	70	26	lays	lay	VERB
fcis-8581	70	27	the	the	DET
fcis-8581	70	28	foundation	foundation	NOUN
fcis-8581	70	29	for	for	ADP
fcis-8581	70	30	the	the	DET
fcis-8581	70	31	subsequent	subsequent	ADJ
fcis-8581	70	32	steps	step	NOUN
fcis-8581	70	33	such	such	ADJ
fcis-8581	70	34	as	as	ADP
fcis-8581	70	35	similarity	similarity	NOUN
fcis-8581	70	36	calculation	calculation	NOUN
fcis-8581	70	37	.	.	PUNCT
fcis-8581	71	1	a	a	DET
fcis-8581	71	2	two	two	NUM
fcis-8581	71	3	-	-	PUNCT
fcis-8581	71	4	dimensional	dimensional	ADJ
fcis-8581	71	5	matrix	matrix	NOUN
fcis-8581	71	6	is	be	AUX
fcis-8581	71	7	used	use	VERB
fcis-8581	71	8	in	in	ADP
fcis-8581	71	9	the	the	DET
fcis-8581	71	10	algorithm	algorithm	NOUN
fcis-8581	71	11	to	to	PART
fcis-8581	71	12	represent	represent	VERB
fcis-8581	71	13	the	the	DET
fcis-8581	71	14	scoring	scoring	NOUN
fcis-8581	71	15	model	model	NOUN
fcis-8581	71	16	,	,	PUNCT
fcis-8581	71	17	so	so	SCONJ
fcis-8581	71	18	the	the	DET
fcis-8581	71	19	similarity	similarity	NOUN
fcis-8581	71	20	calculation	calculation	NOUN
fcis-8581	71	21	of	of	ADP
fcis-8581	71	22	different	different	ADJ
fcis-8581	71	23	objects	object	NOUN
fcis-8581	71	24	can	can	AUX
fcis-8581	71	25	be	be	AUX
fcis-8581	71	26	transformed	transform	VERB
fcis-8581	71	27	into	into	ADP
fcis-8581	71	28	the	the	DET
fcis-8581	71	29	similarity	similarity	NOUN
fcis-8581	71	30	calculation	calculation	NOUN
fcis-8581	71	31	between	between	ADP
fcis-8581	71	32	different	different	ADJ
fcis-8581	71	33	vectors	vector	NOUN
fcis-8581	71	34	in	in	ADP
fcis-8581	71	35	this	this	DET
fcis-8581	71	36	matrix	matrix	NOUN
fcis-8581	71	37	.	.	PUNCT
fcis-8581	72	1	a	a	DET
fcis-8581	72	2	value	value	NOUN
fcis-8581	72	3	between	between	ADP
fcis-8581	72	4	0	0	NUM
fcis-8581	72	5	and	and	CCONJ
fcis-8581	72	6	5	5	NUM
fcis-8581	72	7	is	be	AUX
fcis-8581	72	8	used	use	VERB
fcis-8581	72	9	in	in	ADP
fcis-8581	72	10	this	this	DET
fcis-8581	72	11	system	system	NOUN
fcis-8581	72	12	to	to	PART
fcis-8581	72	13	indicate	indicate	VERB
fcis-8581	72	14	the	the	DET
fcis-8581	72	15	user	user	NOUN
fcis-8581	72	16	's	's	PART
fcis-8581	72	17	liking	liking	NOUN
fcis-8581	72	18	of	of	ADP
fcis-8581	72	19	the	the	DET
fcis-8581	72	20	movie	movie	NOUN
fcis-8581	72	21	,	,	PUNCT
fcis-8581	72	22	each	each	DET
fcis-8581	72	23	score	score	NOUN
fcis-8581	72	24	represents	represent	VERB
fcis-8581	72	25	a	a	DET
fcis-8581	72	26	different	different	ADJ
fcis-8581	72	27	rating	rating	NOUN
fcis-8581	72	28	or	or	CCONJ
fcis-8581	72	29	liking	liking	NOUN
fcis-8581	72	30	level	level	NOUN
fcis-8581	72	31	,	,	PUNCT
fcis-8581	72	32	and	and	CCONJ
fcis-8581	72	33	a	a	DET
fcis-8581	72	34	higher	high	ADJ
fcis-8581	72	35	score	score	NOUN
fcis-8581	72	36	means	mean	VERB
fcis-8581	72	37	a	a	DET
fcis-8581	72	38	higher	high	ADJ
fcis-8581	72	39	liking	liking	NOUN
fcis-8581	72	40	.	.	PUNCT
fcis-8581	73	1	(	(	PUNCT
fcis-8581	73	2	2	2	X
fcis-8581	73	3	)	)	PUNCT
fcis-8581	73	4	calculate	calculate	VERB
fcis-8581	73	5	the	the	DET
fcis-8581	73	6	similarity	similarity	NOUN
fcis-8581	73	7	between	between	ADP
fcis-8581	73	8	users	user	NOUN
fcis-8581	73	9	calculating	calculate	VERB
fcis-8581	73	10	the	the	DET
fcis-8581	73	11	similarity	similarity	NOUN
fcis-8581	73	12	between	between	ADP
fcis-8581	73	13	users	user	NOUN
fcis-8581	73	14	is	be	AUX
fcis-8581	73	15	the	the	DET
fcis-8581	73	16	basic	basic	ADJ
fcis-8581	73	17	step	step	NOUN
fcis-8581	73	18	of	of	ADP
fcis-8581	73	19	establishing	establish	VERB
fcis-8581	73	20	the	the	DET
fcis-8581	73	21	set	set	NOUN
fcis-8581	73	22	of	of	ADP
fcis-8581	73	23	nearest	near	ADJ
fcis-8581	73	24	neighbors	neighbor	NOUN
fcis-8581	73	25	,	,	PUNCT
fcis-8581	73	26	and	and	CCONJ
fcis-8581	73	27	the	the	DET
fcis-8581	73	28	accuracy	accuracy	NOUN
fcis-8581	73	29	of	of	ADP
fcis-8581	73	30	similarity	similarity	NOUN
fcis-8581	73	31	calculation	calculation	NOUN
fcis-8581	73	32	will	will	AUX
fcis-8581	73	33	directly	directly	ADV
fcis-8581	73	34	affect	affect	VERB
fcis-8581	73	35	the	the	DET
fcis-8581	73	36	selection	selection	NOUN
fcis-8581	73	37	of	of	ADP
fcis-8581	73	38	nearest	near	ADJ
fcis-8581	73	39	neighbor	neighbor	NOUN
fcis-8581	73	40	users	user	NOUN
fcis-8581	73	41	and	and	CCONJ
fcis-8581	73	42	finally	finally	ADV
fcis-8581	73	43	affect	affect	VERB
fcis-8581	73	44	the	the	DET
fcis-8581	73	45	accuracy	accuracy	NOUN
fcis-8581	73	46	of	of	ADP
fcis-8581	73	47	prediction	prediction	NOUN
fcis-8581	73	48	results	result	NOUN
fcis-8581	73	49	.	.	PUNCT
fcis-8581	74	1	the	the	DET
fcis-8581	74	2	similarity	similarity	NOUN
fcis-8581	74	3	algorithm	algorithm	NOUN
fcis-8581	74	4	used	use	VERB
fcis-8581	74	5	in	in	ADP
fcis-8581	74	6	this	this	DET
fcis-8581	74	7	system	system	NOUN
fcis-8581	74	8	is	be	AUX
fcis-8581	74	9	the	the	DET
fcis-8581	74	10	cosine	cosine	NOUN
fcis-8581	74	11	similarity	similarity	NOUN
fcis-8581	74	12	algorithm	algorithm	NOUN
fcis-8581	74	13	,	,	PUNCT
fcis-8581	74	14	which	which	PRON
fcis-8581	74	15	is	be	AUX
fcis-8581	74	16	a	a	DET
fcis-8581	74	17	common	common	ADJ
fcis-8581	74	18	method	method	NOUN
fcis-8581	74	19	to	to	PART
fcis-8581	74	20	calculate	calculate	VERB
fcis-8581	74	21	the	the	DET
fcis-8581	74	22	similarity	similarity	NOUN
fcis-8581	74	23	of	of	ADP
fcis-8581	74	24	user	user	NOUN
fcis-8581	74	25	rating	rating	NOUN
fcis-8581	74	26	vectors	vector	NOUN
fcis-8581	74	27	.	.	PUNCT
fcis-8581	75	1	(	(	PUNCT
fcis-8581	75	2	3	3	X
fcis-8581	75	3	)	)	PUNCT
fcis-8581	75	4	establishing	establish	VERB
fcis-8581	75	5	the	the	DET
fcis-8581	75	6	set	set	NOUN
fcis-8581	75	7	of	of	ADP
fcis-8581	75	8	similar	similar	ADJ
fcis-8581	75	9	nearest	near	ADJ
fcis-8581	75	10	neighbors	neighbor	NOUN
fcis-8581	75	11	there	there	PRON
fcis-8581	75	12	are	be	VERB
fcis-8581	75	13	two	two	NUM
fcis-8581	75	14	main	main	ADJ
fcis-8581	75	15	methods	method	NOUN
fcis-8581	75	16	commonly	commonly	ADV
fcis-8581	75	17	used	use	VERB
fcis-8581	75	18	to	to	PART
fcis-8581	75	19	select	select	VERB
fcis-8581	75	20	nearest	near	ADJ
fcis-8581	75	21	neighbor	neighbor	NOUN
fcis-8581	75	22	users	user	NOUN
fcis-8581	75	23	:	:	PUNCT
fcis-8581	75	24	1	1	X
fcis-8581	75	25	)	)	PUNCT
fcis-8581	75	26	set	set	VERB
fcis-8581	75	27	threshold	threshold	NOUN
fcis-8581	75	28	method	method	NOUN
fcis-8581	75	29	,	,	PUNCT
fcis-8581	75	30	i.e.	i.e.	X
fcis-8581	75	31	,	,	PUNCT
fcis-8581	75	32	a	a	DET
fcis-8581	75	33	threshold	threshold	NOUN
fcis-8581	75	34	value	value	NOUN
fcis-8581	75	35	δ	δ	PROPN
fcis-8581	75	36	is	be	AUX
fcis-8581	75	37	given	give	VERB
fcis-8581	75	38	in	in	ADP
fcis-8581	75	39	the	the	DET
fcis-8581	75	40	system	system	NOUN
fcis-8581	75	41	,	,	PUNCT
fcis-8581	75	42	and	and	CCONJ
fcis-8581	75	43	the	the	DET
fcis-8581	75	44	similarity	similarity	NOUN
fcis-8581	75	45	calculated	calculate	VERB
fcis-8581	75	46	between	between	ADP
fcis-8581	75	47	two	two	NUM
fcis-8581	75	48	users	user	NOUN
fcis-8581	75	49	can	can	AUX
fcis-8581	75	50	be	be	AUX
fcis-8581	75	51	selected	select	VERB
fcis-8581	75	52	as	as	ADP
fcis-8581	75	53	nearest	near	ADJ
fcis-8581	75	54	neighbor	neighbor	NOUN
fcis-8581	75	55	if	if	SCONJ
fcis-8581	75	56	it	it	PRON
fcis-8581	75	57	exceeds	exceed	VERB
fcis-8581	75	58	the	the	DET
fcis-8581	75	59	threshold	threshold	NOUN
fcis-8581	75	60	value	value	NOUN
fcis-8581	75	61	δ	δ	PROPN
fcis-8581	75	62	.	.	PUNCT
fcis-8581	76	1	the	the	DET
fcis-8581	76	2	number	number	NOUN
fcis-8581	76	3	of	of	ADP
fcis-8581	76	4	nearest	near	ADJ
fcis-8581	76	5	neighbor	neighbor	NOUN
fcis-8581	76	6	users	user	NOUN
fcis-8581	76	7	obtained	obtain	VERB
fcis-8581	76	8	by	by	ADP
fcis-8581	76	9	the	the	DET
fcis-8581	76	10	set	set	NOUN
fcis-8581	76	11	threshold	threshold	NOUN
fcis-8581	76	12	method	method	NOUN
fcis-8581	76	13	is	be	AUX
fcis-8581	76	14	predetermined	predetermine	VERB
fcis-8581	76	15	,	,	PUNCT
fcis-8581	76	16	but	but	CCONJ
fcis-8581	76	17	the	the	DET
fcis-8581	76	18	similarity	similarity	NOUN
fcis-8581	76	19	between	between	ADP
fcis-8581	76	20	the	the	DET
fcis-8581	76	21	obtained	obtain	VERB
fcis-8581	76	22	nearest	near	ADJ
fcis-8581	76	23	neighbor	neighbor	NOUN
fcis-8581	76	24	users	user	NOUN
fcis-8581	76	25	and	and	CCONJ
fcis-8581	76	26	the	the	DET
fcis-8581	76	27	target	target	NOUN
fcis-8581	76	28	users	user	NOUN
fcis-8581	76	29	generally	generally	ADV
fcis-8581	76	30	does	do	AUX
fcis-8581	76	31	not	not	PART
fcis-8581	76	32	show	show	VERB
fcis-8581	76	33	a	a	DET
fcis-8581	76	34	large	large	ADJ
fcis-8581	76	35	deviation	deviation	NOUN
fcis-8581	76	36	.	.	PUNCT
fcis-8581	77	1	2	2	X
fcis-8581	77	2	)	)	PUNCT
fcis-8581	77	3	k	k	PROPN
fcis-8581	77	4	nearest	near	ADJ
fcis-8581	77	5	neighbor	neighbor	NOUN
fcis-8581	77	6	method	method	NOUN
fcis-8581	77	7	,	,	PUNCT
fcis-8581	77	8	firstly	firstly	ADV
fcis-8581	77	9	,	,	PUNCT
fcis-8581	77	10	the	the	DET
fcis-8581	77	11	similarity	similarity	NOUN
fcis-8581	77	12	between	between	ADP
fcis-8581	77	13	the	the	DET
fcis-8581	77	14	target	target	NOUN
fcis-8581	77	15	object	object	NOUN
fcis-8581	77	16	and	and	CCONJ
fcis-8581	77	17	other	other	ADJ
fcis-8581	77	18	users	user	NOUN
fcis-8581	77	19	is	be	AUX
fcis-8581	77	20	arranged	arrange	VERB
fcis-8581	77	21	in	in	ADP
fcis-8581	77	22	the	the	DET
fcis-8581	77	23	order	order	NOUN
fcis-8581	77	24	from	from	ADP
fcis-8581	77	25	largest	large	ADJ
fcis-8581	77	26	to	to	AUX
fcis-8581	77	27	smallest	small	ADJ
fcis-8581	77	28	,	,	PUNCT
fcis-8581	77	29	and	and	CCONJ
fcis-8581	77	30	the	the	DET
fcis-8581	77	31	top	top	ADJ
fcis-8581	77	32	k	k	PROPN
fcis-8581	77	33	users	user	NOUN
fcis-8581	77	34	are	be	AUX
fcis-8581	77	35	taken	take	VERB
fcis-8581	77	36	to	to	PART
fcis-8581	77	37	form	form	VERB
fcis-8581	77	38	the	the	DET
fcis-8581	77	39	set	set	NOUN
fcis-8581	77	40	of	of	ADP
fcis-8581	77	41	nearest	near	ADJ
fcis-8581	77	42	neighbors	neighbor	NOUN
fcis-8581	77	43	.	.	PUNCT
fcis-8581	78	1	in	in	ADP
fcis-8581	78	2	the	the	DET
fcis-8581	78	3	set	set	NOUN
fcis-8581	78	4	threshold	threshold	NOUN
fcis-8581	78	5	method	method	NOUN
fcis-8581	78	6	,	,	PUNCT
fcis-8581	78	7	if	if	SCONJ
fcis-8581	78	8	the	the	DET
fcis-8581	78	9	threshold	threshold	NOUN
fcis-8581	78	10	δ	δ	PROPN
fcis-8581	78	11	is	be	AUX
fcis-8581	78	12	set	set	VERB
fcis-8581	78	13	high	high	ADV
fcis-8581	78	14	,	,	PUNCT
fcis-8581	78	15	it	it	PRON
fcis-8581	78	16	will	will	AUX
fcis-8581	78	17	select	select	VERB
fcis-8581	78	18	users	user	NOUN
fcis-8581	78	19	with	with	ADP
fcis-8581	78	20	high	high	ADJ
fcis-8581	78	21	similarity	similarity	NOUN
fcis-8581	78	22	to	to	ADP
fcis-8581	78	23	the	the	DET
fcis-8581	78	24	target	target	NOUN
fcis-8581	78	25	user	user	NOUN
fcis-8581	78	26	's	's	PART
fcis-8581	78	27	interest	interest	NOUN
fcis-8581	78	28	preferences	preference	NOUN
fcis-8581	78	29	,	,	PUNCT
fcis-8581	78	30	but	but	CCONJ
fcis-8581	78	31	at	at	ADP
fcis-8581	78	32	the	the	DET
fcis-8581	78	33	same	same	ADJ
fcis-8581	78	34	time	time	NOUN
fcis-8581	78	35	,	,	PUNCT
fcis-8581	78	36	it	it	PRON
fcis-8581	78	37	may	may	AUX
fcis-8581	78	38	result	result	VERB
fcis-8581	78	39	in	in	ADP
fcis-8581	78	40	fewer	few	ADJ
fcis-8581	78	41	or	or	CCONJ
fcis-8581	78	42	even	even	ADV
fcis-8581	78	43	no	no	DET
fcis-8581	78	44	users	user	NOUN
fcis-8581	78	45	satisfying	satisfy	VERB
fcis-8581	78	46	the	the	DET
fcis-8581	78	47	threshold	threshold	NOUN
fcis-8581	78	48	,	,	PUNCT
fcis-8581	78	49	so	so	SCONJ
fcis-8581	78	50	that	that	SCONJ
fcis-8581	78	51	it	it	PRON
fcis-8581	78	52	is	be	AUX
fcis-8581	78	53	impossible	impossible	ADJ
fcis-8581	78	54	to	to	PART
fcis-8581	78	55	find	find	VERB
fcis-8581	78	56	the	the	DET
fcis-8581	78	57	set	set	NOUN
fcis-8581	78	58	of	of	ADP
fcis-8581	78	59	the	the	DET
fcis-8581	78	60	target	target	NOUN
fcis-8581	78	61	user	user	NOUN
fcis-8581	78	62	's	's	PART
fcis-8581	78	63	near	near	ADJ
fcis-8581	78	64	-	-	PUNCT
fcis-8581	78	65	neighbor	neighbor	NOUN
fcis-8581	78	66	users	user	NOUN
fcis-8581	78	67	.	.	PUNCT
fcis-8581	79	1	however	however	ADV
fcis-8581	79	2	,	,	PUNCT
fcis-8581	79	3	the	the	DET
fcis-8581	79	4	threshold	threshold	NOUN
fcis-8581	79	5	δ	δ	PROPN
fcis-8581	79	6	,	,	PUNCT
fcis-8581	79	7	if	if	SCONJ
fcis-8581	79	8	set	set	VERB
fcis-8581	79	9	lower	lower	ADV
fcis-8581	79	10	,	,	PUNCT
fcis-8581	79	11	will	will	AUX
fcis-8581	79	12	lead	lead	VERB
fcis-8581	79	13	to	to	ADP
fcis-8581	79	14	the	the	DET
fcis-8581	79	15	existence	existence	NOUN
fcis-8581	79	16	of	of	ADP
fcis-8581	79	17	users	user	NOUN
fcis-8581	79	18	with	with	ADP
fcis-8581	79	19	lower	low	ADJ
fcis-8581	79	20	similarity	similarity	NOUN
fcis-8581	79	21	among	among	ADP
fcis-8581	79	22	the	the	DET
fcis-8581	79	23	similar	similar	ADJ
fcis-8581	79	24	nearest	near	ADJ
fcis-8581	79	25	neighbors	neighbor	NOUN
fcis-8581	79	26	satisfying	satisfy	VERB
fcis-8581	79	27	the	the	DET
fcis-8581	79	28	threshold	threshold	NOUN
fcis-8581	79	29	.	.	PUNCT
fcis-8581	80	1	although	although	SCONJ
fcis-8581	80	2	the	the	DET
fcis-8581	80	3	number	number	NOUN
fcis-8581	80	4	of	of	ADP
fcis-8581	80	5	similar	similar	ADJ
fcis-8581	80	6	nearest	near	ADJ
fcis-8581	80	7	neighbor	neighbor	NOUN
fcis-8581	80	8	users	user	NOUN
fcis-8581	80	9	satisfying	satisfy	VERB
fcis-8581	80	10	the	the	DET
fcis-8581	80	11	lower	low	ADJ
fcis-8581	80	12	threshold	threshold	NOUN
fcis-8581	80	13	is	be	AUX
fcis-8581	80	14	higher	high	ADJ
fcis-8581	80	15	,	,	PUNCT
fcis-8581	80	16	the	the	DET
fcis-8581	80	17	similarity	similarity	NOUN
fcis-8581	80	18	between	between	ADP
fcis-8581	80	19	the	the	DET
fcis-8581	80	20	set	set	NOUN
fcis-8581	80	21	of	of	ADP
fcis-8581	80	22	such	such	ADJ
fcis-8581	80	23	nearest	near	ADJ
fcis-8581	80	24	neighbor	neighbor	NOUN
fcis-8581	80	25	users	user	NOUN
fcis-8581	80	26	and	and	CCONJ
fcis-8581	80	27	the	the	DET
fcis-8581	80	28	target	target	NOUN
fcis-8581	80	29	users	user	NOUN
fcis-8581	80	30	can	can	AUX
fcis-8581	80	31	not	not	PART
fcis-8581	80	32	be	be	AUX
fcis-8581	80	33	guaranteed	guarantee	VERB
fcis-8581	80	34	,	,	PUNCT
fcis-8581	80	35	and	and	CCONJ
fcis-8581	80	36	it	it	PRON
fcis-8581	80	37	increases	increase	VERB
fcis-8581	80	38	the	the	DET
fcis-8581	80	39	computation	computation	NOUN
fcis-8581	80	40	and	and	CCONJ
fcis-8581	80	41	error	error	NOUN
fcis-8581	80	42	for	for	ADP
fcis-8581	80	43	the	the	DET
fcis-8581	80	44	subsequent	subsequent	ADJ
fcis-8581	80	45	calculation	calculation	NOUN
fcis-8581	80	46	.	.	PUNCT
fcis-8581	81	1	similarly	similarly	ADV
fcis-8581	81	2	,	,	PUNCT
fcis-8581	81	3	the	the	DET
fcis-8581	81	4	setting	setting	NOUN
fcis-8581	81	5	of	of	ADP
fcis-8581	81	6	the	the	DET
fcis-8581	81	7	k	k	PROPN
fcis-8581	81	8	value	value	NOUN
fcis-8581	81	9	in	in	ADP
fcis-8581	81	10	the	the	DET
fcis-8581	81	11	k	k	NOUN
fcis-8581	81	12	-	-	PUNCT
fcis-8581	81	13	nearest	near	ADJ
fcis-8581	81	14	neighbor	neighbor	NOUN
fcis-8581	81	15	method	method	NOUN
fcis-8581	81	16	affects	affect	VERB
fcis-8581	81	17	the	the	DET
fcis-8581	81	18	quality	quality	NOUN
fcis-8581	81	19	of	of	ADP
fcis-8581	81	20	the	the	DET
fcis-8581	81	21	nearest	near	ADJ
fcis-8581	81	22	neighbor	neighbor	NOUN
fcis-8581	81	23	user	user	NOUN
fcis-8581	81	24	selection	selection	NOUN
fcis-8581	81	25	and	and	CCONJ
fcis-8581	81	26	thus	thus	ADV
fcis-8581	81	27	the	the	DET
fcis-8581	81	28	subsequent	subsequent	ADJ
fcis-8581	81	29	calculation	calculation	NOUN
fcis-8581	81	30	.	.	PUNCT
fcis-8581	82	1	the	the	DET
fcis-8581	82	2	effect	effect	NOUN
fcis-8581	82	3	of	of	ADP
fcis-8581	82	4	a	a	DET
fcis-8581	82	5	larger	large	ADJ
fcis-8581	82	6	k	k	NOUN
fcis-8581	82	7	value	value	NOUN
fcis-8581	82	8	is	be	AUX
fcis-8581	82	9	similar	similar	ADJ
fcis-8581	82	10	to	to	ADP
fcis-8581	82	11	the	the	DET
fcis-8581	82	12	effect	effect	NOUN
fcis-8581	82	13	of	of	ADP
fcis-8581	82	14	a	a	DET
fcis-8581	82	15	smaller	small	ADJ
fcis-8581	82	16	threshold	threshold	NOUN
fcis-8581	82	17	in	in	ADP
fcis-8581	82	18	the	the	DET
fcis-8581	82	19	set	set	NOUN
fcis-8581	82	20	threshold	threshold	NOUN
fcis-8581	82	21	method	method	NOUN
fcis-8581	82	22	,	,	PUNCT
fcis-8581	82	23	but	but	CCONJ
fcis-8581	82	24	since	since	SCONJ
fcis-8581	82	25	the	the	DET
fcis-8581	82	26	k	k	PROPN
fcis-8581	82	27	value	value	NOUN
fcis-8581	82	28	is	be	AUX
fcis-8581	82	29	fixed	fix	VERB
fcis-8581	82	30	in	in	ADP
fcis-8581	82	31	the	the	DET
fcis-8581	82	32	k	k	PROPN
fcis-8581	82	33	nearest	near	ADJ
fcis-8581	82	34	neighbor	neighbor	NOUN
fcis-8581	82	35	method	method	NOUN
fcis-8581	82	36	,	,	PUNCT
fcis-8581	82	37	there	there	PRON
fcis-8581	82	38	is	be	VERB
fcis-8581	82	39	no	no	DET
fcis-8581	82	40	uncontrolled	uncontrolled	ADJ
fcis-8581	82	41	introduction	introduction	NOUN
fcis-8581	82	42	of	of	ADP
fcis-8581	82	43	a	a	DET
fcis-8581	82	44	large	large	ADJ
fcis-8581	82	45	number	number	NOUN
fcis-8581	82	46	of	of	ADP
fcis-8581	82	47	nearest	near	ADJ
fcis-8581	82	48	neighbor	neighbor	NOUN
fcis-8581	82	49	users	user	NOUN
fcis-8581	82	50	with	with	ADP
fcis-8581	82	51	low	low	ADJ
fcis-8581	82	52	similarity	similarity	NOUN
fcis-8581	82	53	as	as	ADP
fcis-8581	82	54	in	in	ADP
fcis-8581	82	55	the	the	DET
fcis-8581	82	56	set	set	NOUN
fcis-8581	82	57	threshold	threshold	NOUN
fcis-8581	82	58	method	method	NOUN
fcis-8581	82	59	.	.	PUNCT
fcis-8581	83	1	the	the	DET
fcis-8581	83	2	effect	effect	NOUN
fcis-8581	83	3	of	of	ADP
fcis-8581	83	4	setting	set	VERB
fcis-8581	83	5	a	a	DET
fcis-8581	83	6	small	small	ADJ
fcis-8581	83	7	k	k	NOUN
fcis-8581	83	8	value	value	NOUN
fcis-8581	83	9	is	be	AUX
fcis-8581	83	10	similar	similar	ADJ
fcis-8581	83	11	to	to	ADP
fcis-8581	83	12	the	the	DET
fcis-8581	83	13	effect	effect	NOUN
fcis-8581	83	14	of	of	ADP
fcis-8581	83	15	setting	set	VERB
fcis-8581	83	16	a	a	DET
fcis-8581	83	17	large	large	ADJ
fcis-8581	83	18	threshold	threshold	NOUN
fcis-8581	83	19	in	in	ADP
fcis-8581	83	20	the	the	DET
fcis-8581	83	21	threshold	threshold	NOUN
fcis-8581	83	22	method	method	NOUN
fcis-8581	83	23	,	,	PUNCT
fcis-8581	83	24	but	but	CCONJ
fcis-8581	83	25	because	because	SCONJ
fcis-8581	83	26	the	the	DET
fcis-8581	83	27	k	k	PROPN
fcis-8581	83	28	value	value	NOUN
fcis-8581	83	29	is	be	AUX
fcis-8581	83	30	fixed	fix	VERB
fcis-8581	83	31	in	in	ADP
fcis-8581	83	32	the	the	DET
fcis-8581	83	33	k	k	PROPN
fcis-8581	83	34	nearest	near	ADJ
fcis-8581	83	35	neighbor	neighbor	NOUN
fcis-8581	83	36	method	method	NOUN
fcis-8581	83	37	,	,	PUNCT
fcis-8581	83	38	there	there	PRON
fcis-8581	83	39	will	will	AUX
fcis-8581	83	40	not	not	PART
fcis-8581	83	41	be	be	AUX
fcis-8581	83	42	too	too	ADV
fcis-8581	83	43	few	few	ADJ
fcis-8581	83	44	or	or	CCONJ
fcis-8581	83	45	even	even	ADV
fcis-8581	83	46	no	no	DET
fcis-8581	83	47	nearest	near	ADJ
fcis-8581	83	48	neighbor	neighbor	NOUN
fcis-8581	83	49	users	user	NOUN
fcis-8581	83	50	as	as	ADP
fcis-8581	83	51	in	in	ADP
fcis-8581	83	52	the	the	DET
fcis-8581	83	53	threshold	threshold	NOUN
fcis-8581	83	54	method	method	NOUN
fcis-8581	83	55	.	.	PUNCT
fcis-8581	84	1	this	this	DET
fcis-8581	84	2	system	system	NOUN
fcis-8581	84	3	calculates	calculate	VERB
fcis-8581	84	4	a	a	DET
fcis-8581	84	5	unique	unique	ADJ
fcis-8581	84	6	one	one	NUM
fcis-8581	84	7	-	-	PUNCT
fcis-8581	84	8	dimensional	dimensional	ADJ
fcis-8581	84	9	value	value	NOUN
fcis-8581	84	10	based	base	VERB
fcis-8581	84	11	on	on	ADP
fcis-8581	84	12	the	the	DET
fcis-8581	84	13	cosine	cosine	NOUN
fcis-8581	84	14	similarity	similarity	NOUN
fcis-8581	84	15	,	,	PUNCT
fcis-8581	84	16	and	and	CCONJ
fcis-8581	84	17	it	it	PRON
fcis-8581	84	18	is	be	AUX
fcis-8581	84	19	more	more	ADV
fcis-8581	84	20	appropriate	appropriate	ADJ
fcis-8581	84	21	to	to	PART
fcis-8581	84	22	use	use	VERB
fcis-8581	84	23	the	the	DET
fcis-8581	84	24	threshold	threshold	NOUN
fcis-8581	84	25	method	method	NOUN
fcis-8581	84	26	.	.	PUNCT
fcis-8581	85	1	based	base	VERB
fcis-8581	85	2	on	on	ADP
fcis-8581	85	3	certain	certain	ADJ
fcis-8581	85	4	experiments	experiment	NOUN
fcis-8581	85	5	,	,	PUNCT
fcis-8581	85	6	it	it	PRON
fcis-8581	85	7	is	be	AUX
fcis-8581	85	8	judged	judge	VERB
fcis-8581	85	9	that	that	SCONJ
fcis-8581	85	10	when	when	SCONJ
fcis-8581	85	11	the	the	DET
fcis-8581	85	12	threshold	threshold	NOUN
fcis-8581	85	13	value	value	NOUN
fcis-8581	85	14	is	be	AUX
fcis-8581	85	15	0.8	0.8	NUM
fcis-8581	85	16	,	,	PUNCT
fcis-8581	85	17	there	there	PRON
fcis-8581	85	18	are	be	VERB
fcis-8581	85	19	both	both	DET
fcis-8581	85	20	new	new	ADJ
fcis-8581	85	21	recommendations	recommendation	NOUN
fcis-8581	85	22	for	for	ADP
fcis-8581	85	23	users	user	NOUN
fcis-8581	85	24	with	with	ADP
fcis-8581	85	25	few	few	ADJ
fcis-8581	85	26	ratings	rating	NOUN
fcis-8581	85	27	and	and	CCONJ
fcis-8581	85	28	the	the	DET
fcis-8581	85	29	recommended	recommend	VERB
fcis-8581	85	30	movies	movie	NOUN
fcis-8581	85	31	meet	meet	VERB
fcis-8581	85	32	the	the	DET
fcis-8581	85	33	needs	need	NOUN
fcis-8581	85	34	of	of	ADP
fcis-8581	85	35	users	user	NOUN
fcis-8581	85	36	,	,	PUNCT
fcis-8581	85	37	so	so	SCONJ
fcis-8581	85	38	this	this	DET
fcis-8581	85	39	department	department	NOUN
fcis-8581	85	40	adopts	adopt	VERB
fcis-8581	85	41	the	the	DET
fcis-8581	85	42	set	set	NOUN
fcis-8581	85	43	threshold	threshold	NOUN
fcis-8581	85	44	method	method	NOUN
fcis-8581	85	45	with	with	ADP
fcis-8581	85	46	a	a	DET
fcis-8581	85	47	threshold	threshold	NOUN
fcis-8581	85	48	value	value	NOUN
fcis-8581	85	49	of	of	ADP
fcis-8581	85	50	0.8	0.8	NUM
fcis-8581	85	51	.	.	PUNCT
fcis-8581	86	1	(	(	PUNCT
fcis-8581	86	2	4	4	X
fcis-8581	86	3	)	)	PUNCT
fcis-8581	86	4	predictive	predictive	ADJ
fcis-8581	86	5	scoring	scoring	NOUN
fcis-8581	86	6	there	there	PRON
fcis-8581	86	7	are	be	VERB
fcis-8581	86	8	two	two	NUM
fcis-8581	86	9	ways	way	NOUN
fcis-8581	86	10	to	to	PART
fcis-8581	86	11	generate	generate	VERB
fcis-8581	86	12	recommendation	recommendation	NOUN
fcis-8581	86	13	results	result	NOUN
fcis-8581	86	14	for	for	ADP
fcis-8581	86	15	the	the	DET
fcis-8581	86	16	target	target	NOUN
fcis-8581	86	17	user	user	NOUN
fcis-8581	86	18	based	base	VERB
fcis-8581	86	19	on	on	ADP
fcis-8581	86	20	the	the	DET
fcis-8581	86	21	set	set	NOUN
fcis-8581	86	22	of	of	ADP
fcis-8581	86	23	similar	similar	ADJ
fcis-8581	86	24	nearest	near	ADJ
fcis-8581	86	25	neighbors	neighbor	NOUN
fcis-8581	86	26	obtained	obtain	VERB
fcis-8581	86	27	in	in	ADP
fcis-8581	86	28	the	the	DET
fcis-8581	86	29	previous	previous	ADJ
fcis-8581	86	30	step	step	NOUN
fcis-8581	86	31	,	,	PUNCT
fcis-8581	86	32	one	one	NUM
fcis-8581	86	33	is	be	AUX
fcis-8581	86	34	the	the	DET
fcis-8581	86	35	prediction	prediction	NOUN
fcis-8581	86	36	of	of	ADP
fcis-8581	86	37	the	the	DET
fcis-8581	86	38	target	target	NOUN
fcis-8581	86	39	user	user	NOUN
fcis-8581	86	40	's	's	PART
fcis-8581	86	41	rating	rating	NOUN
fcis-8581	86	42	for	for	ADP
fcis-8581	86	43	any	any	DET
fcis-8581	86	44	item	item	NOUN
fcis-8581	86	45	,	,	PUNCT
fcis-8581	86	46	and	and	CCONJ
fcis-8581	86	47	the	the	DET
fcis-8581	86	48	other	other	ADJ
fcis-8581	86	49	is	be	AUX
fcis-8581	86	50	to	to	PART
fcis-8581	86	51	generate	generate	VERB
fcis-8581	86	52	a	a	DET
fcis-8581	86	53	recommendation	recommendation	NOUN
fcis-8581	86	54	list	list	NOUN
fcis-8581	86	55	.	.	PUNCT
fcis-8581	87	1	both	both	DET
fcis-8581	87	2	ways	way	NOUN
fcis-8581	87	3	of	of	ADP
fcis-8581	87	4	generating	generate	VERB
fcis-8581	87	5	recommendation	recommendation	NOUN
fcis-8581	87	6	results	result	NOUN
fcis-8581	87	7	require	require	VERB
fcis-8581	87	8	the	the	DET
fcis-8581	87	9	prediction	prediction	NOUN
fcis-8581	87	10	of	of	ADP
fcis-8581	87	11	the	the	DET
fcis-8581	87	12	target	target	NOUN
fcis-8581	87	13	user	user	NOUN
fcis-8581	87	14	's	's	PART
fcis-8581	87	15	rating	rating	NOUN
fcis-8581	87	16	for	for	ADP
fcis-8581	87	17	an	an	DET
fcis-8581	87	18	item	item	NOUN
fcis-8581	87	19	first	first	ADV
fcis-8581	87	20	.	.	PUNCT
fcis-8581	88	1	the	the	DET
fcis-8581	88	2	common	common	ADJ
fcis-8581	88	3	calculation	calculation	NOUN
fcis-8581	88	4	method	method	NOUN
fcis-8581	88	5	used	use	VERB
fcis-8581	88	6	to	to	PART
fcis-8581	88	7	perform	perform	VERB
fcis-8581	88	8	predictive	predictive	ADJ
fcis-8581	88	9	scoring	scoring	NOUN
fcis-8581	88	10	is	be	AUX
fcis-8581	88	11	the	the	DET
fcis-8581	88	12	ordinary	ordinary	ADJ
fcis-8581	88	13	weighted	weight	VERB
fcis-8581	88	14	average	average	ADJ
fcis-8581	88	15	,	,	PUNCT
fcis-8581	88	16	where	where	SCONJ
fcis-8581	88	17	similarity	similarity	NOUN
fcis-8581	88	18	is	be	AUX
fcis-8581	88	19	used	use	VERB
fcis-8581	88	20	as	as	ADP
fcis-8581	88	21	a	a	DET
fcis-8581	88	22	weight	weight	NOUN
fcis-8581	88	23	for	for	ADP
fcis-8581	88	24	weighted	weighted	ADJ
fcis-8581	88	25	averaging	averaging	NOUN
fcis-8581	88	26	.	.	PUNCT
fcis-8581	89	1	under	under	ADP
fcis-8581	89	2	this	this	DET
fcis-8581	89	3	calculation	calculation	NOUN
fcis-8581	89	4	,	,	PUNCT
fcis-8581	89	5	the	the	DET
fcis-8581	89	6	more	more	ADV
fcis-8581	89	7	similar	similar	ADJ
fcis-8581	89	8	nearest	near	ADJ
fcis-8581	89	9	neighbor	neighbor	NOUN
fcis-8581	89	10	users	user	NOUN
fcis-8581	89	11	to	to	ADP
fcis-8581	89	12	the	the	DET
fcis-8581	89	13	target	target	NOUN
fcis-8581	89	14	user	user	NOUN
fcis-8581	89	15	contribute	contribute	VERB
fcis-8581	89	16	to	to	ADP
fcis-8581	89	17	the	the	DET
fcis-8581	89	18	final	final	ADJ
fcis-8581	89	19	predicted	predict	VERB
fcis-8581	89	20	rating	rating	NOUN
fcis-8581	89	21	,	,	PUNCT
fcis-8581	89	22	the	the	PRON
fcis-8581	89	23	more	more	ADV
fcis-8581	89	24	the	the	DET
fcis-8581	89	25	predicted	predict	VERB
fcis-8581	89	26	rating	rating	NOUN
fcis-8581	89	27	is	be	AUX
fcis-8581	89	28	eventually	eventually	ADV
fcis-8581	89	29	dominated	dominate	VERB
fcis-8581	89	30	by	by	ADP
fcis-8581	89	31	some	some	DET
fcis-8581	89	32	nearest	near	ADJ
fcis-8581	89	33	neighbor	neighbor	NOUN
fcis-8581	89	34	users	user	NOUN
fcis-8581	89	35	with	with	ADP
fcis-8581	89	36	high	high	ADJ
fcis-8581	89	37	similarity	similarity	NOUN
fcis-8581	89	38	,	,	PUNCT
fcis-8581	89	39	and	and	CCONJ
fcis-8581	89	40	due	due	ADP
fcis-8581	89	41	to	to	ADP
fcis-8581	89	42	this	this	DET
fcis-8581	89	43	feature	feature	NOUN
fcis-8581	89	44	,	,	PUNCT
fcis-8581	89	45	the	the	DET
fcis-8581	89	46	system	system	NOUN
fcis-8581	89	47	can	can	AUX
fcis-8581	89	48	abstract	abstract	VERB
fcis-8581	89	49	a	a	DET
fcis-8581	89	50	specific	specific	ADJ
fcis-8581	89	51	number	number	NOUN
fcis-8581	89	52	of	of	ADP
fcis-8581	89	53	users	user	NOUN
fcis-8581	89	54	to	to	PART
fcis-8581	89	55	make	make	VERB
fcis-8581	89	56	their	their	PRON
fcis-8581	89	57	ratings	rating	NOUN
fcis-8581	89	58	for	for	ADP
fcis-8581	89	59	movies	movie	NOUN
fcis-8581	89	60	representative	representative	NOUN
fcis-8581	89	61	in	in	ADP
fcis-8581	89	62	order	order	NOUN
fcis-8581	89	63	to	to	PART
fcis-8581	89	64	reduce	reduce	VERB
fcis-8581	89	65	the	the	DET
fcis-8581	89	66	huge	huge	ADJ
fcis-8581	89	67	amount	amount	NOUN
fcis-8581	89	68	of	of	ADP
fcis-8581	89	69	computation	computation	NOUN
fcis-8581	89	70	due	due	ADP
fcis-8581	89	71	to	to	ADP
fcis-8581	89	72	the	the	DET
fcis-8581	89	73	large	large	ADJ
fcis-8581	89	74	number	number	NOUN
fcis-8581	89	75	of	of	ADP
fcis-8581	89	76	movies	movie	NOUN
fcis-8581	89	77	and	and	CCONJ
fcis-8581	89	78	users	user	NOUN
fcis-8581	89	79	.	.	PUNCT
fcis-8581	90	1	3.3.2	3.3.2	X
fcis-8581	90	2	.	.	X
fcis-8581	90	3	matrix	matrix	NOUN
fcis-8581	90	4	decomposition	decomposition	NOUN
fcis-8581	90	5	-	-	PUNCT
fcis-8581	90	6	based	base	VERB
fcis-8581	90	7	recommendation	recommendation	NOUN
fcis-8581	90	8	algorithm	algorithm	NOUN
fcis-8581	90	9	matrix	matrix	NOUN
fcis-8581	90	10	decomposition	decomposition	NOUN
fcis-8581	90	11	is	be	AUX
fcis-8581	90	12	the	the	DET
fcis-8581	90	13	decomposition	decomposition	NOUN
fcis-8581	90	14	of	of	ADP
fcis-8581	90	15	a	a	DET
fcis-8581	90	16	highdimensional	highdimensional	ADJ
fcis-8581	90	17	matrix	matrix	NOUN
fcis-8581	90	18	into	into	ADP
fcis-8581	90	19	a	a	DET
fcis-8581	90	20	representation	representation	NOUN
fcis-8581	90	21	of	of	ADP
fcis-8581	90	22	a	a	DET
fcis-8581	90	23	lowdimensional	lowdimensional	ADJ
fcis-8581	90	24	matrix	matrix	NOUN
fcis-8581	90	25	.	.	PUNCT
fcis-8581	91	1	this	this	DET
fcis-8581	91	2	approach	approach	NOUN
fcis-8581	91	3	reduces	reduce	VERB
fcis-8581	91	4	the	the	DET
fcis-8581	91	5	dimensionality	dimensionality	NOUN
fcis-8581	91	6	of	of	ADP
fcis-8581	91	7	the	the	DET
fcis-8581	91	8	target	target	NOUN
fcis-8581	91	9	matrix	matrix	NOUN
fcis-8581	91	10	and	and	CCONJ
fcis-8581	91	11	allows	allow	VERB
fcis-8581	91	12	the	the	DET
fcis-8581	91	13	program	program	NOUN
fcis-8581	91	14	to	to	PART
fcis-8581	91	15	run	run	VERB
fcis-8581	91	16	at	at	ADP
fcis-8581	91	17	a	a	DET
fcis-8581	91	18	faster	fast	ADJ
fcis-8581	91	19	speed	speed	NOUN
fcis-8581	91	20	.	.	PUNCT
fcis-8581	92	1	the	the	DET
fcis-8581	92	2	most	most	ADV
fcis-8581	92	3	used	used	ADJ
fcis-8581	92	4	decomposition	decomposition	NOUN
fcis-8581	92	5	is	be	AUX
fcis-8581	92	6	the	the	DET
fcis-8581	92	7	singular	singular	ADJ
fcis-8581	92	8	value	value	NOUN
fcis-8581	92	9	[	[	X
fcis-8581	92	10	6	6	NUM
fcis-8581	92	11	]	]	PUNCT
fcis-8581	92	12	(	(	PUNCT
fcis-8581	92	13	svd	svd	PROPN
fcis-8581	92	14	)	)	PUNCT
fcis-8581	92	15	decomposition	decomposition	NOUN
fcis-8581	92	16	.	.	PUNCT
fcis-8581	93	1	when	when	SCONJ
fcis-8581	93	2	the	the	DET
fcis-8581	93	3	number	number	NOUN
fcis-8581	93	4	of	of	ADP
fcis-8581	93	5	user	user	NOUN
fcis-8581	93	6	and	and	CCONJ
fcis-8581	93	7	item	item	NOUN
fcis-8581	93	8	ratings	rating	NOUN
fcis-8581	93	9	is	be	AUX
fcis-8581	93	10	very	very	ADV
fcis-8581	93	11	large	large	ADJ
fcis-8581	93	12	,	,	PUNCT
fcis-8581	93	13	the	the	DET
fcis-8581	93	14	target	target	NOUN
fcis-8581	93	15	matrix	matrix	NOUN
fcis-8581	93	16	is	be	AUX
fcis-8581	93	17	also	also	ADV
fcis-8581	93	18	very	very	ADV
fcis-8581	93	19	large	large	ADJ
fcis-8581	93	20	and	and	CCONJ
fcis-8581	93	21	a	a	DET
fcis-8581	93	22	sparse	sparse	ADJ
fcis-8581	93	23	matrix	matrix	NOUN
fcis-8581	93	24	,	,	PUNCT
fcis-8581	93	25	which	which	PRON
fcis-8581	93	26	is	be	AUX
fcis-8581	93	27	very	very	ADV
fcis-8581	93	28	unfriendly	unfriendly	ADJ
fcis-8581	93	29	for	for	ADP
fcis-8581	93	30	the	the	DET
fcis-8581	93	31	operation	operation	NOUN
fcis-8581	93	32	of	of	ADP
fcis-8581	93	33	the	the	DET
fcis-8581	93	34	algorithm	algorithm	NOUN
fcis-8581	93	35	,	,	PUNCT
fcis-8581	93	36	and	and	CCONJ
fcis-8581	93	37	after	after	ADP
fcis-8581	93	38	reducing	reduce	VERB
fcis-8581	93	39	the	the	DET
fcis-8581	93	40	dimensionality	dimensionality	NOUN
fcis-8581	93	41	of	of	ADP
fcis-8581	93	42	the	the	DET
fcis-8581	93	43	target	target	NOUN
fcis-8581	93	44	matrix	matrix	NOUN
fcis-8581	93	45	using	use	VERB
fcis-8581	93	46	svd	svd	PROPN
fcis-8581	93	47	technique	technique	NOUN
fcis-8581	93	48	,	,	PUNCT
fcis-8581	93	49	the	the	DET
fcis-8581	93	50	potential	potential	ADJ
fcis-8581	93	51	properties	property	NOUN
fcis-8581	93	52	in	in	ADP
fcis-8581	93	53	the	the	DET
fcis-8581	93	54	original	original	ADJ
fcis-8581	93	55	matrix	matrix	NOUN
fcis-8581	93	56	can	can	AUX
fcis-8581	93	57	be	be	AUX
fcis-8581	93	58	inferred	infer	VERB
fcis-8581	93	59	and	and	CCONJ
fcis-8581	93	60	the	the	DET
fcis-8581	93	61	potential	potential	ADJ
fcis-8581	93	62	correlations	correlation	NOUN
fcis-8581	93	63	between	between	ADP
fcis-8581	93	64	users	user	NOUN
fcis-8581	93	65	and	and	CCONJ
fcis-8581	93	66	items	item	NOUN
fcis-8581	93	67	can	can	AUX
fcis-8581	93	68	be	be	AUX
fcis-8581	93	69	found	find	VERB
fcis-8581	93	70	.	.	PUNCT
fcis-8581	94	1	assuming	assume	VERB
fcis-8581	94	2	that	that	SCONJ
fcis-8581	94	3	the	the	DET
fcis-8581	94	4	system	system	NOUN
fcis-8581	94	5	has	have	VERB
fcis-8581	94	6	m	m	NOUN
fcis-8581	94	7	users	user	NOUN
fcis-8581	94	8	and	and	CCONJ
fcis-8581	94	9	n	n	PRON
fcis-8581	94	10	items	item	NOUN
fcis-8581	94	11	,	,	PUNCT
fcis-8581	94	12	the	the	DET
fcis-8581	94	13	collaborative	collaborative	ADJ
fcis-8581	94	14	filtering	filtering	NOUN
fcis-8581	94	15	matrix	matrix	NOUN
fcis-8581	94	16	r	r	NOUN
fcis-8581	94	17	is	be	AUX
fcis-8581	94	18	constructed	construct	VERB
fcis-8581	94	19	by	by	ADP
fcis-8581	94	20	the	the	DET
fcis-8581	94	21	rating	rating	NOUN
fcis-8581	94	22	relationship	relationship	NOUN
fcis-8581	94	23	between	between	ADP
fcis-8581	94	24	items	item	NOUN
fcis-8581	94	25	and	and	CCONJ
fcis-8581	94	26	items	item	NOUN
fcis-8581	94	27	,	,	PUNCT
fcis-8581	94	28	and	and	CCONJ
fcis-8581	94	29	then	then	ADV
fcis-8581	94	30	the	the	DET
fcis-8581	94	31	complete	complete	ADJ
fcis-8581	94	32	rating	rating	NOUN
fcis-8581	94	33	matrix	matrix	NOUN
fcis-8581	94	34	r	r	NOUN
fcis-8581	94	35	needs	need	VERB
fcis-8581	94	36	to	to	PART
fcis-8581	94	37	be	be	AUX
fcis-8581	94	38	constructed	construct	VERB
fcis-8581	94	39	when	when	SCONJ
fcis-8581	94	40	recommending	recommend	VERB
fcis-8581	94	41	information	information	NOUN
fcis-8581	94	42	to	to	ADP
fcis-8581	94	43	users	user	NOUN
fcis-8581	94	44	,	,	PUNCT
fcis-8581	94	45	so	so	ADV
fcis-8581	94	46	the	the	DET
fcis-8581	94	47	corresponding	correspond	VERB
fcis-8581	94	48	prediction	prediction	NOUN
fcis-8581	94	49	filling	filling	NOUN
fcis-8581	94	50	is	be	AUX
fcis-8581	94	51	needed	need	VERB
fcis-8581	94	52	for	for	ADP
fcis-8581	94	53	the	the	DET
fcis-8581	94	54	true	true	ADJ
fcis-8581	94	55	information	information	NOUN
fcis-8581	94	56	of	of	ADP
fcis-8581	94	57	the	the	DET
fcis-8581	94	58	current	current	ADJ
fcis-8581	94	59	matrix	matrix	NOUN
fcis-8581	94	60	.	.	PUNCT
fcis-8581	95	1	initially	initially	ADV
fcis-8581	95	2	,	,	PUNCT
fcis-8581	95	3	the	the	DET
fcis-8581	95	4	mean	mean	ADJ
fcis-8581	95	5	values	value	NOUN
fcis-8581	95	6	of	of	ADP
fcis-8581	95	7	the	the	DET
fcis-8581	95	8	corresponding	correspond	VERB
fcis-8581	95	9	user	user	NOUN
fcis-8581	95	10	ratings	rating	NOUN
fcis-8581	95	11	can	can	AUX
fcis-8581	95	12	be	be	AUX
fcis-8581	95	13	filled	fill	VERB
fcis-8581	95	14	in	in	ADP
fcis-8581	95	15	uniformly	uniformly	ADV
fcis-8581	95	16	where	where	SCONJ
fcis-8581	95	17	the	the	DET
fcis-8581	95	18	data	datum	NOUN
fcis-8581	95	19	are	be	AUX
fcis-8581	95	20	missing	miss	VERB
fcis-8581	95	21	.	.	PUNCT
fcis-8581	96	1	in	in	ADP
fcis-8581	96	2	addition	addition	NOUN
fcis-8581	96	3	,	,	PUNCT
fcis-8581	96	4	the	the	DET
fcis-8581	96	5	data	datum	NOUN
fcis-8581	96	6	can	can	AUX
fcis-8581	96	7	also	also	ADV
fcis-8581	96	8	be	be	AUX
fcis-8581	96	9	normalized	normalize	VERB
fcis-8581	96	10	,	,	PUNCT
fcis-8581	96	11	i.e.	i.e.	X
fcis-8581	96	12	,	,	PUNCT
fcis-8581	96	13	the	the	DET
fcis-8581	96	14	entire	entire	ADJ
fcis-8581	96	15	data	datum	NOUN
fcis-8581	96	16	is	be	AUX
fcis-8581	96	17	subtracted	subtract	VERB
fcis-8581	96	18	from	from	ADP
fcis-8581	96	19	the	the	DET
fcis-8581	96	20	mean	mean	NOUN
fcis-8581	96	21	,	,	PUNCT
fcis-8581	96	22	which	which	PRON
fcis-8581	96	23	in	in	ADP
fcis-8581	96	24	a	a	DET
fcis-8581	96	25	geometric	geometric	ADJ
fcis-8581	96	26	sense	sense	NOUN
fcis-8581	96	27	behaves	behave	VERB
fcis-8581	96	28	as	as	ADP
fcis-8581	96	29	data	datum	NOUN
fcis-8581	96	30	centered	center	VERB
fcis-8581	96	31	at	at	ADP
fcis-8581	96	32	the	the	DET
fcis-8581	96	33	origin	origin	NOUN
fcis-8581	96	34	,	,	PUNCT
fcis-8581	96	35	and	and	CCONJ
fcis-8581	96	36	thus	thus	ADV
fcis-8581	96	37	yields	yield	VERB
fcis-8581	96	38	the	the	DET
fcis-8581	96	39	final	final	ADJ
fcis-8581	96	40	rating	rating	NOUN
fcis-8581	96	41	matrix	matrix	NOUN
fcis-8581	96	42	r.	r.	NOUN
fcis-8581	96	43	the	the	DET
fcis-8581	96	44	svd	svd	PROPN
fcis-8581	96	45	decomposition	decomposition	NOUN
fcis-8581	96	46	of	of	ADP
fcis-8581	96	47	the	the	DET
fcis-8581	96	48	matrix	matrix	NOUN
fcis-8581	96	49	r	r	NOUN
fcis-8581	96	50	yields	yield	NOUN
fcis-8581	96	51	(	(	PUNCT
fcis-8581	96	52	8)	8)	NUM
fcis-8581	96	53	:	:	PUNCT
fcis-8581	96	54	∙	∙	PROPN
fcis-8581	96	55	∙	∙	X
fcis-8581	96	56	(	(	PUNCT
fcis-8581	96	57	8)	8)	NUM
fcis-8581	96	58	where	where	SCONJ
fcis-8581	96	59	u	u	NOUN
fcis-8581	96	60	and	and	CCONJ
fcis-8581	96	61	v	v	NOUN
fcis-8581	96	62	denote	denote	VERB
fcis-8581	96	63	two	two	NUM
fcis-8581	96	64	orthogonal	orthogonal	ADJ
fcis-8581	96	65	matrices	matrix	NOUN
fcis-8581	96	66	of	of	ADP
fcis-8581	96	67	116	116	NUM
fcis-8581	96	68	dimension	dimension	NOUN
fcis-8581	96	69	m×r	m×r	PROPN
fcis-8581	96	70	and	and	CCONJ
fcis-8581	96	71	r×n	r×n	PROPN
fcis-8581	96	72	,	,	PUNCT
fcis-8581	96	73	respectively	respectively	ADV
fcis-8581	96	74	,	,	PUNCT
fcis-8581	96	75	used	use	VERB
fcis-8581	96	76	to	to	PART
fcis-8581	96	77	represent	represent	VERB
fcis-8581	96	78	the	the	DET
fcis-8581	96	79	potential	potential	ADJ
fcis-8581	96	80	factors	factor	NOUN
fcis-8581	96	81	of	of	ADP
fcis-8581	96	82	users	user	NOUN
fcis-8581	96	83	and	and	CCONJ
fcis-8581	96	84	items	item	NOUN
fcis-8581	96	85	,	,	PUNCT
fcis-8581	96	86	respectively	respectively	ADV
fcis-8581	96	87	.	.	PUNCT
fcis-8581	97	1	s	s	PART
fcis-8581	97	2	is	be	AUX
fcis-8581	97	3	a	a	DET
fcis-8581	97	4	diagonal	diagonal	ADJ
fcis-8581	97	5	matrix	matrix	NOUN
fcis-8581	97	6	of	of	ADP
fcis-8581	97	7	order	order	NOUN
fcis-8581	97	8	r×r	r×r	PROPN
fcis-8581	97	9	,	,	PUNCT
fcis-8581	97	10	consisting	consist	VERB
fcis-8581	97	11	of	of	ADP
fcis-8581	97	12	the	the	DET
fcis-8581	97	13	singular	singular	ADJ
fcis-8581	97	14	values	value	NOUN
fcis-8581	97	15	of	of	ADP
fcis-8581	97	16	the	the	DET
fcis-8581	97	17	original	original	ADJ
fcis-8581	97	18	matrix	matrix	NOUN
fcis-8581	97	19	,	,	PUNCT
fcis-8581	97	20	similar	similar	ADJ
fcis-8581	97	21	to	to	ADP
fcis-8581	97	22	the	the	DET
fcis-8581	97	23	orthogonalized	orthogonalized	ADJ
fcis-8581	97	24	decomposition	decomposition	NOUN
fcis-8581	97	25	of	of	ADP
fcis-8581	97	26	a	a	DET
fcis-8581	97	27	matrix	matrix	NOUN
fcis-8581	97	28	.	.	PUNCT
fcis-8581	98	1	after	after	ADP
fcis-8581	98	2	getting	get	VERB
fcis-8581	98	3	the	the	DET
fcis-8581	98	4	r	r	NOUN
fcis-8581	98	5	singular	singular	ADJ
fcis-8581	98	6	values	value	NOUN
fcis-8581	98	7	,	,	PUNCT
fcis-8581	98	8	the	the	DET
fcis-8581	98	9	selection	selection	NOUN
fcis-8581	98	10	of	of	ADP
fcis-8581	98	11	singular	singular	ADJ
fcis-8581	98	12	values	value	NOUN
fcis-8581	98	13	can	can	AUX
fcis-8581	98	14	be	be	AUX
fcis-8581	98	15	carried	carry	VERB
fcis-8581	98	16	out	out	ADP
fcis-8581	98	17	,	,	PUNCT
fcis-8581	98	18	and	and	CCONJ
fcis-8581	98	19	the	the	DET
fcis-8581	98	20	first	first	ADJ
fcis-8581	98	21	five	five	NUM
fcis-8581	98	22	larger	large	ADJ
fcis-8581	98	23	singular	singular	ADJ
fcis-8581	98	24	values	value	NOUN
fcis-8581	98	25	are	be	AUX
fcis-8581	98	26	selected	select	VERB
fcis-8581	98	27	in	in	ADP
fcis-8581	98	28	this	this	DET
fcis-8581	98	29	paper	paper	NOUN
fcis-8581	98	30	,	,	PUNCT
fcis-8581	98	31	at	at	ADP
fcis-8581	98	32	which	which	DET
fcis-8581	98	33	time	time	NOUN
fcis-8581	98	34	the	the	DET
fcis-8581	98	35	diagonal	diagonal	ADJ
fcis-8581	98	36	matrix	matrix	NOUN
fcis-8581	98	37	s	s	NOUN
fcis-8581	98	38	becomes	become	VERB
fcis-8581	98	39	sk	sk	ADP
fcis-8581	98	40	,	,	PUNCT
fcis-8581	98	41	and	and	CCONJ
fcis-8581	98	42	u	u	NOUN
fcis-8581	98	43	and	and	CCONJ
fcis-8581	98	44	v	v	ADP
fcis-8581	98	45	become	become	VERB
fcis-8581	98	46	k	k	ADV
fcis-8581	98	47	-	-	ADJ
fcis-8581	98	48	dimensional	dimensional	ADJ
fcis-8581	98	49	accordingly	accordingly	ADV
fcis-8581	98	50	,	,	PUNCT
fcis-8581	98	51	and	and	CCONJ
fcis-8581	98	52	the	the	DET
fcis-8581	98	53	sum	sum	NOUN
fcis-8581	98	54	is	be	AUX
fcis-8581	98	55	obtained	obtain	VERB
fcis-8581	98	56	,	,	PUNCT
fcis-8581	98	57	which	which	PRON
fcis-8581	98	58	in	in	ADP
fcis-8581	98	59	turn	turn	NOUN
fcis-8581	98	60	achieves	achieve	VERB
fcis-8581	98	61	the	the	DET
fcis-8581	98	62	dimensionality	dimensionality	NOUN
fcis-8581	98	63	reduction	reduction	NOUN
fcis-8581	98	64	needed	need	VERB
fcis-8581	98	65	in	in	ADP
fcis-8581	98	66	this	this	DET
fcis-8581	98	67	paper	paper	NOUN
fcis-8581	98	68	.	.	PUNCT
fcis-8581	99	1	next	next	ADV
fcis-8581	99	2	,	,	PUNCT
fcis-8581	99	3	the	the	DET
fcis-8581	99	4	three	three	NUM
fcis-8581	99	5	matrices	matrix	NOUN
fcis-8581	99	6	are	be	AUX
fcis-8581	99	7	dotted	dot	VERB
fcis-8581	99	8	and	and	CCONJ
fcis-8581	99	9	multiplied	multiply	VERB
fcis-8581	99	10	to	to	PART
fcis-8581	99	11	obtain	obtain	VERB
fcis-8581	99	12	the	the	DET
fcis-8581	99	13	desired	desire	VERB
fcis-8581	99	14	reconstruction	reconstruction	NOUN
fcis-8581	99	15	matrix	matrix	NOUN
fcis-8581	99	16	(	(	PUNCT
fcis-8581	99	17	9	9	NUM
fcis-8581	99	18	):	):	PUNCT
fcis-8581	99	19	∙	∙	PROPN
fcis-8581	99	20	∙	∙	PROPN
fcis-8581	99	21	(	(	PUNCT
fcis-8581	99	22	9	9	X
fcis-8581	99	23	)	)	PUNCT
fcis-8581	99	24	using	use	VERB
fcis-8581	99	25	the	the	DET
fcis-8581	99	26	decomposed	decompose	VERB
fcis-8581	99	27	matrix	matrix	NOUN
fcis-8581	99	28	,	,	PUNCT
fcis-8581	99	29	the	the	DET
fcis-8581	99	30	corresponding	correspond	VERB
fcis-8581	99	31	prediction	prediction	NOUN
fcis-8581	99	32	can	can	AUX
fcis-8581	99	33	be	be	AUX
fcis-8581	99	34	made	make	VERB
fcis-8581	99	35	for	for	ADP
fcis-8581	99	36	the	the	DET
fcis-8581	99	37	user	user	NOUN
fcis-8581	99	38	's	's	PART
fcis-8581	99	39	unrated	unrated	ADJ
fcis-8581	99	40	items	item	NOUN
fcis-8581	99	41	by	by	ADP
fcis-8581	99	42	dotting	dot	VERB
fcis-8581	99	43	the	the	DET
fcis-8581	99	44	i	i	PROPN
fcis-8581	99	45	-	-	PUNCT
fcis-8581	99	46	th	th	X
fcis-8581	99	47	row	row	NOUN
fcis-8581	99	48	and	and	CCONJ
fcis-8581	99	49	the	the	DET
fcis-8581	99	50	j	j	PROPN
fcis-8581	99	51	-	-	PUNCT
fcis-8581	99	52	th	th	X
fcis-8581	99	53	column	column	NOUN
fcis-8581	99	54	,	,	PUNCT
fcis-8581	99	55	and	and	CCONJ
fcis-8581	99	56	adding	add	VERB
fcis-8581	99	57	the	the	DET
fcis-8581	99	58	mean	mean	ADJ
fcis-8581	99	59	value	value	NOUN
fcis-8581	99	60	of	of	ADP
fcis-8581	99	61	the	the	DET
fcis-8581	99	62	ratings	rating	NOUN
fcis-8581	99	63	of	of	ADP
fcis-8581	99	64	the	the	DET
fcis-8581	99	65	previous	previous	ADJ
fcis-8581	99	66	user	user	NOUN
fcis-8581	99	67	rating	rating	NOUN
fcis-8581	99	68	data	datum	NOUN
fcis-8581	99	69	,	,	PUNCT
fcis-8581	99	70	i.e.	i.e.	X
fcis-8581	99	71	,	,	PUNCT
fcis-8581	99	72	after	after	ADP
fcis-8581	99	73	decentering	decentere	VERB
fcis-8581	99	74	,	,	PUNCT
fcis-8581	99	75	to	to	PART
fcis-8581	99	76	obtain	obtain	VERB
fcis-8581	99	77	the	the	DET
fcis-8581	99	78	predicted	predict	VERB
fcis-8581	99	79	rating	rating	NOUN
fcis-8581	99	80	value	value	NOUN
fcis-8581	99	81	of	of	ADP
fcis-8581	99	82	user	user	NOUN
fcis-8581	99	83	i	i	PRON
fcis-8581	99	84	for	for	ADP
fcis-8581	99	85	item	item	PROPN
fcis-8581	99	86	j	j	PROPN
fcis-8581	99	87	,	,	PUNCT
fcis-8581	99	88	as	as	SCONJ
fcis-8581	99	89	shown	show	VERB
fcis-8581	99	90	in	in	ADP
fcis-8581	99	91	(	(	PUNCT
fcis-8581	99	92	10	10	NUM
fcis-8581	99	93	):	):	PUNCT
fcis-8581	99	94	,	,	PUNCT
fcis-8581	100	1	∙	∙	PROPN
fcis-8581	100	2	∙	∙	PUNCT
fcis-8581	100	3	∙	∙	PROPN
fcis-8581	100	4	(	(	PUNCT
fcis-8581	100	5	10	10	NUM
fcis-8581	100	6	)	)	PUNCT
fcis-8581	100	7	the	the	DET
fcis-8581	100	8	svd	svd	PROPN
fcis-8581	100	9	matrix	matrix	NOUN
fcis-8581	100	10	decomposition	decomposition	NOUN
fcis-8581	100	11	algorithm	algorithm	NOUN
fcis-8581	100	12	has	have	VERB
fcis-8581	100	13	great	great	ADJ
fcis-8581	100	14	applications	application	NOUN
fcis-8581	100	15	in	in	ADP
fcis-8581	100	16	recommendation	recommendation	NOUN
fcis-8581	100	17	systems	system	NOUN
fcis-8581	100	18	,	,	PUNCT
fcis-8581	100	19	improving	improve	VERB
fcis-8581	100	20	the	the	DET
fcis-8581	100	21	performance	performance	NOUN
fcis-8581	100	22	of	of	ADP
fcis-8581	100	23	the	the	DET
fcis-8581	100	24	recommendation	recommendation	NOUN
fcis-8581	100	25	system	system	NOUN
fcis-8581	100	26	and	and	CCONJ
fcis-8581	100	27	realizing	realize	VERB
fcis-8581	100	28	real	real	ADJ
fcis-8581	100	29	-	-	PUNCT
fcis-8581	100	30	time	time	NOUN
fcis-8581	100	31	recommendations	recommendation	NOUN
fcis-8581	100	32	even	even	ADV
fcis-8581	100	33	in	in	ADP
fcis-8581	100	34	the	the	DET
fcis-8581	100	35	case	case	NOUN
fcis-8581	100	36	of	of	ADP
fcis-8581	100	37	extremely	extremely	ADV
fcis-8581	100	38	sparse	sparse	ADJ
fcis-8581	100	39	matrices	matrix	NOUN
fcis-8581	100	40	,	,	PUNCT
fcis-8581	100	41	laying	lay	VERB
fcis-8581	100	42	a	a	DET
fcis-8581	100	43	good	good	ADJ
fcis-8581	100	44	foundation	foundation	NOUN
fcis-8581	100	45	for	for	ADP
fcis-8581	100	46	more	more	ADJ
fcis-8581	100	47	matrix	matrix	NOUN
fcis-8581	100	48	decomposition	decomposition	NOUN
fcis-8581	100	49	recommendation	recommendation	NOUN
fcis-8581	100	50	algorithms	algorithm	NOUN
fcis-8581	100	51	in	in	ADP
fcis-8581	100	52	the	the	DET
fcis-8581	100	53	future	future	NOUN
fcis-8581	100	54	.	.	PUNCT
fcis-8581	101	1	3.3.3	3.3.3	NUM
fcis-8581	101	2	.	.	PUNCT
fcis-8581	101	3	regularized	regularize	VERB
fcis-8581	101	4	singular	singular	ADJ
fcis-8581	101	5	value	value	NOUN
fcis-8581	101	6	decomposition	decomposition	NOUN
fcis-8581	101	7	algorithm	algorithm	NOUN
fcis-8581	101	8	the	the	DET
fcis-8581	101	9	svd	svd	PROPN
fcis-8581	101	10	algorithm	algorithm	PROPN
fcis-8581	101	11	introduced	introduce	VERB
fcis-8581	101	12	in	in	ADP
fcis-8581	101	13	section	section	NOUN
fcis-8581	101	14	3.3.2	3.3.2	NUM
fcis-8581	101	15	needs	need	VERB
fcis-8581	101	16	to	to	PART
fcis-8581	101	17	calculate	calculate	VERB
fcis-8581	101	18	the	the	DET
fcis-8581	101	19	eigenvectors	eigenvector	NOUN
fcis-8581	101	20	and	and	CCONJ
fcis-8581	101	21	eigenvalues	eigenvalue	NOUN
fcis-8581	101	22	of	of	ADP
fcis-8581	101	23	the	the	DET
fcis-8581	101	24	matrix	matrix	NOUN
fcis-8581	101	25	during	during	ADP
fcis-8581	101	26	the	the	DET
fcis-8581	101	27	computation	computation	NOUN
fcis-8581	101	28	,	,	PUNCT
fcis-8581	101	29	and	and	CCONJ
fcis-8581	101	30	the	the	DET
fcis-8581	101	31	matrix	matrix	NOUN
fcis-8581	101	32	can	can	AUX
fcis-8581	101	33	not	not	PART
fcis-8581	101	34	have	have	VERB
fcis-8581	101	35	null	null	ADJ
fcis-8581	101	36	values	value	NOUN
fcis-8581	101	37	during	during	ADP
fcis-8581	101	38	the	the	DET
fcis-8581	101	39	computation	computation	NOUN
fcis-8581	101	40	,	,	PUNCT
fcis-8581	101	41	which	which	PRON
fcis-8581	101	42	is	be	AUX
fcis-8581	101	43	not	not	PART
fcis-8581	101	44	friendly	friendly	ADJ
fcis-8581	101	45	to	to	ADP
fcis-8581	101	46	the	the	DET
fcis-8581	101	47	program	program	NOUN
fcis-8581	101	48	and	and	CCONJ
fcis-8581	101	49	the	the	DET
fcis-8581	101	50	computation	computation	NOUN
fcis-8581	101	51	process	process	NOUN
fcis-8581	101	52	is	be	AUX
fcis-8581	101	53	complicated	complicate	VERB
fcis-8581	101	54	.	.	PUNCT
fcis-8581	102	1	the	the	DET
fcis-8581	102	2	regularized	regularize	VERB
fcis-8581	102	3	singular	singular	ADJ
fcis-8581	102	4	value	value	NOUN
fcis-8581	102	5	decomposition	decomposition	NOUN
fcis-8581	102	6	(	(	PUNCT
fcis-8581	102	7	rsvd	rsvd	NOUN
fcis-8581	102	8	)	)	PUNCT
fcis-8581	102	9	algorithm	algorithm	NOUN
fcis-8581	102	10	is	be	AUX
fcis-8581	102	11	a	a	DET
fcis-8581	102	12	further	further	ADJ
fcis-8581	102	13	improvement	improvement	NOUN
fcis-8581	102	14	of	of	ADP
fcis-8581	102	15	the	the	DET
fcis-8581	102	16	svd	svd	PROPN
fcis-8581	102	17	algorithm	algorithm	PROPN
fcis-8581	102	18	,	,	PUNCT
fcis-8581	102	19	which	which	PRON
fcis-8581	102	20	is	be	AUX
fcis-8581	102	21	a	a	DET
fcis-8581	102	22	very	very	ADV
fcis-8581	102	23	popular	popular	ADJ
fcis-8581	102	24	recommendation	recommendation	NOUN
fcis-8581	102	25	system	system	NOUN
fcis-8581	102	26	algorithm	algorithm	NOUN
fcis-8581	102	27	nowadays	nowadays	ADV
fcis-8581	102	28	.	.	PUNCT
fcis-8581	103	1	in	in	ADP
fcis-8581	103	2	the	the	DET
fcis-8581	103	3	rsvd	rsvd	NOUN
fcis-8581	103	4	model	model	NOUN
fcis-8581	103	5	,	,	PUNCT
fcis-8581	103	6	the	the	DET
fcis-8581	103	7	initial	initial	ADJ
fcis-8581	103	8	collaborative	collaborative	ADJ
fcis-8581	103	9	filtering	filtering	NOUN
fcis-8581	103	10	matrix	matrix	NOUN
fcis-8581	103	11	r	r	NOUN
fcis-8581	103	12	is	be	AUX
fcis-8581	103	13	decomposed	decompose	VERB
fcis-8581	103	14	into	into	ADP
fcis-8581	103	15	two	two	NUM
fcis-8581	103	16	low	low	ADJ
fcis-8581	103	17	-	-	PUNCT
fcis-8581	103	18	rank	rank	NOUN
fcis-8581	103	19	matrices	matrix	NOUN
fcis-8581	103	20	u	u	NOUN
fcis-8581	103	21	and	and	CCONJ
fcis-8581	103	22	v	v	ADP
fcis-8581	103	23	with	with	ADP
fcis-8581	103	24	matrix	matrix	NOUN
fcis-8581	103	25	dimensions	dimension	NOUN
fcis-8581	103	26	m×f	m×f	PROPN
fcis-8581	103	27	and	and	CCONJ
fcis-8581	103	28	n×f	n×f	PROPN
fcis-8581	103	29	,	,	PUNCT
fcis-8581	103	30	respectively	respectively	ADV
fcis-8581	103	31	,	,	PUNCT
fcis-8581	103	32	where	where	SCONJ
fcis-8581	103	33	f<<min(m	f<<min(m	NOUN
fcis-8581	103	34	,	,	PUNCT
fcis-8581	103	35	n	n	CCONJ
fcis-8581	103	36	)	)	PUNCT
fcis-8581	103	37	,	,	PUNCT
fcis-8581	103	38	the	the	DET
fcis-8581	103	39	mathematical	mathematical	ADJ
fcis-8581	103	40	representation	representation	NOUN
fcis-8581	103	41	of	of	ADP
fcis-8581	103	42	this	this	DET
fcis-8581	103	43	algorithm	algorithm	NOUN
fcis-8581	103	44	is	be	AUX
fcis-8581	103	45	given	give	VERB
fcis-8581	103	46	in	in	ADP
fcis-8581	103	47	(	(	PUNCT
fcis-8581	103	48	11	11	NUM
fcis-8581	103	49	):	):	PUNCT
fcis-8581	104	1	∙	∙	PROPN
fcis-8581	104	2	(	(	PUNCT
fcis-8581	104	3	11	11	NUM
fcis-8581	104	4	)	)	PUNCT
fcis-8581	104	5	where	where	SCONJ
fcis-8581	104	6	,	,	PUNCT
fcis-8581	104	7	matrix	matrix	NOUN
fcis-8581	104	8	u	u	NOUN
fcis-8581	104	9	is	be	AUX
fcis-8581	104	10	the	the	DET
fcis-8581	104	11	user	user	NOUN
fcis-8581	104	12	potential	potential	ADJ
fcis-8581	104	13	factor	factor	NOUN
fcis-8581	104	14	matrix	matrix	NOUN
fcis-8581	104	15	,	,	PUNCT
fcis-8581	104	16	which	which	PRON
fcis-8581	104	17	indicates	indicate	VERB
fcis-8581	104	18	the	the	DET
fcis-8581	104	19	liking	liking	NOUN
fcis-8581	104	20	of	of	ADP
fcis-8581	104	21	each	each	DET
fcis-8581	104	22	user	user	NOUN
fcis-8581	104	23	for	for	ADP
fcis-8581	104	24	the	the	DET
fcis-8581	104	25	item	item	NOUN
fcis-8581	104	26	feature	feature	NOUN
fcis-8581	104	27	factors	factor	NOUN
fcis-8581	104	28	,	,	PUNCT
fcis-8581	104	29	and	and	CCONJ
fcis-8581	104	30	matrix	matrix	NOUN
fcis-8581	104	31	v	v	NOUN
fcis-8581	104	32	is	be	AUX
fcis-8581	104	33	the	the	DET
fcis-8581	104	34	item	item	NOUN
fcis-8581	104	35	potential	potential	ADJ
fcis-8581	104	36	factor	factor	NOUN
fcis-8581	104	37	matrix	matrix	NOUN
fcis-8581	104	38	,	,	PUNCT
fcis-8581	104	39	which	which	PRON
fcis-8581	104	40	indicates	indicate	VERB
fcis-8581	104	41	the	the	DET
fcis-8581	104	42	feature	feature	NOUN
fcis-8581	104	43	factor	factor	NOUN
fcis-8581	104	44	composition	composition	NOUN
fcis-8581	104	45	of	of	ADP
fcis-8581	104	46	each	each	DET
fcis-8581	104	47	item	item	NOUN
fcis-8581	104	48	.	.	PUNCT
fcis-8581	105	1	where	where	SCONJ
fcis-8581	105	2	each	each	DET
fcis-8581	105	3	row	row	NOUN
fcis-8581	105	4	of	of	ADP
fcis-8581	105	5	the	the	DET
fcis-8581	105	6	u	u	NOUN
fcis-8581	105	7	matrix	matrix	NOUN
fcis-8581	105	8	indicates	indicate	VERB
fcis-8581	105	9	the	the	DET
fcis-8581	105	10	popularity	popularity	NOUN
fcis-8581	105	11	of	of	ADP
fcis-8581	105	12	each	each	DET
fcis-8581	105	13	user	user	NOUN
fcis-8581	105	14	for	for	ADP
fcis-8581	105	15	different	different	ADJ
fcis-8581	105	16	item	item	NOUN
fcis-8581	105	17	feature	feature	NOUN
fcis-8581	105	18	attributes	attribute	VERB
fcis-8581	105	19	,	,	PUNCT
fcis-8581	105	20	and	and	CCONJ
fcis-8581	105	21	each	each	DET
fcis-8581	105	22	row	row	NOUN
fcis-8581	105	23	of	of	ADP
fcis-8581	105	24	the	the	DET
fcis-8581	105	25	v	v	NOUN
fcis-8581	105	26	matrix	matrix	NOUN
fcis-8581	105	27	indicates	indicate	VERB
fcis-8581	105	28	the	the	DET
fcis-8581	105	29	weight	weight	NOUN
fcis-8581	105	30	of	of	ADP
fcis-8581	105	31	each	each	DET
fcis-8581	105	32	item	item	NOUN
fcis-8581	105	33	's	's	PART
fcis-8581	105	34	feature	feature	NOUN
fcis-8581	105	35	attributes	attribute	VERB
fcis-8581	105	36	.	.	PUNCT
fcis-8581	106	1	each	each	DET
fcis-8581	106	2	column	column	NOUN
fcis-8581	106	3	of	of	ADP
fcis-8581	106	4	the	the	DET
fcis-8581	106	5	potential	potential	ADJ
fcis-8581	106	6	factor	factor	NOUN
fcis-8581	106	7	matrix	matrix	NOUN
fcis-8581	106	8	represents	represent	VERB
fcis-8581	106	9	a	a	DET
fcis-8581	106	10	potential	potential	ADJ
fcis-8581	106	11	factor	factor	NOUN
fcis-8581	106	12	,	,	PUNCT
fcis-8581	106	13	and	and	CCONJ
fcis-8581	106	14	these	these	DET
fcis-8581	106	15	potential	potential	ADJ
fcis-8581	106	16	factors	factor	NOUN
fcis-8581	106	17	form	form	VERB
fcis-8581	106	18	a	a	DET
fcis-8581	106	19	vector	vector	NOUN
fcis-8581	106	20	space	space	NOUN
fcis-8581	106	21	,	,	PUNCT
fcis-8581	106	22	which	which	PRON
fcis-8581	106	23	in	in	ADP
fcis-8581	106	24	turn	turn	NOUN
fcis-8581	106	25	represents	represent	VERB
fcis-8581	106	26	the	the	DET
fcis-8581	106	27	characteristics	characteristic	NOUN
fcis-8581	106	28	possessed	possess	VERB
fcis-8581	106	29	by	by	ADP
fcis-8581	106	30	a	a	DET
fcis-8581	106	31	user	user	NOUN
fcis-8581	106	32	or	or	CCONJ
fcis-8581	106	33	an	an	DET
fcis-8581	106	34	item	item	NOUN
fcis-8581	106	35	.	.	PUNCT
fcis-8581	107	1	in	in	ADP
fcis-8581	107	2	other	other	ADJ
fcis-8581	107	3	words	word	NOUN
fcis-8581	107	4	,	,	PUNCT
fcis-8581	107	5	the	the	DET
fcis-8581	107	6	matrix	matrix	NOUN
fcis-8581	107	7	decomposition	decomposition	NOUN
fcis-8581	107	8	model	model	NOUN
fcis-8581	107	9	extracts	extract	NOUN
fcis-8581	107	10	f	f	NOUN
fcis-8581	107	11	potential	potential	ADJ
fcis-8581	107	12	factors	factor	NOUN
fcis-8581	107	13	from	from	ADP
fcis-8581	107	14	the	the	DET
fcis-8581	107	15	original	original	ADJ
fcis-8581	107	16	matrix	matrix	NOUN
fcis-8581	107	17	,	,	PUNCT
fcis-8581	107	18	which	which	PRON
fcis-8581	107	19	in	in	ADP
fcis-8581	107	20	turn	turn	NOUN
fcis-8581	107	21	indicates	indicate	VERB
fcis-8581	107	22	the	the	DET
fcis-8581	107	23	degree	degree	NOUN
fcis-8581	107	24	of	of	ADP
fcis-8581	107	25	preference	preference	NOUN
fcis-8581	107	26	of	of	ADP
fcis-8581	107	27	this	this	DET
fcis-8581	107	28	user	user	NOUN
fcis-8581	107	29	for	for	ADP
fcis-8581	107	30	the	the	DET
fcis-8581	107	31	item	item	NOUN
fcis-8581	107	32	.	.	PUNCT
fcis-8581	108	1	the	the	DET
fcis-8581	108	2	preference	preference	NOUN
fcis-8581	108	3	r	r	NOUN
fcis-8581	108	4	of	of	ADP
fcis-8581	108	5	user	user	NOUN
fcis-8581	108	6	i	i	PRON
fcis-8581	108	7	for	for	ADP
fcis-8581	108	8	item	item	NOUN
fcis-8581	108	9	j	j	PROPN
fcis-8581	108	10	can	can	AUX
fcis-8581	108	11	be	be	AUX
fcis-8581	108	12	obtained	obtain	VERB
fcis-8581	108	13	by	by	ADP
fcis-8581	108	14	equation	equation	NOUN
fcis-8581	108	15	(	(	PUNCT
fcis-8581	108	16	12	12	NUM
fcis-8581	108	17	):	):	PUNCT
fcis-8581	108	18	,	,	PUNCT
fcis-8581	108	19	∙	∙	PROPN
fcis-8581	108	20	(	(	PUNCT
fcis-8581	108	21	12	12	NUM
fcis-8581	108	22	)	)	PUNCT
fcis-8581	108	23	where	where	SCONJ
fcis-8581	108	24	and	and	CCONJ
fcis-8581	108	25	denote	denote	VERB
fcis-8581	108	26	the	the	DET
fcis-8581	108	27	i	i	PROPN
fcis-8581	108	28	-	-	PUNCT
fcis-8581	108	29	th	th	X
fcis-8581	108	30	row	row	NOUN
fcis-8581	108	31	of	of	ADP
fcis-8581	108	32	the	the	DET
fcis-8581	108	33	potential	potential	ADJ
fcis-8581	108	34	factor	factor	NOUN
fcis-8581	108	35	prototype	prototype	NOUN
fcis-8581	108	36	matrix	matrix	NOUN
fcis-8581	108	37	u	u	NOUN
fcis-8581	108	38	and	and	CCONJ
fcis-8581	108	39	the	the	DET
fcis-8581	108	40	j	j	PROPN
fcis-8581	108	41	-	-	PUNCT
fcis-8581	108	42	th	th	VERB
fcis-8581	108	43	row	row	NOUN
fcis-8581	108	44	of	of	ADP
fcis-8581	108	45	v	v	NOUN
fcis-8581	108	46	,	,	PUNCT
fcis-8581	108	47	respectively	respectively	ADV
fcis-8581	108	48	,	,	PUNCT
fcis-8581	108	49	and	and	CCONJ
fcis-8581	108	50	the	the	DET
fcis-8581	108	51	objective	objective	ADJ
fcis-8581	108	52	function	function	NOUN
fcis-8581	108	53	l	l	NOUN
fcis-8581	108	54	can	can	AUX
fcis-8581	108	55	be	be	AUX
fcis-8581	108	56	determined	determine	VERB
fcis-8581	108	57	to	to	PART
fcis-8581	108	58	calculate	calculate	VERB
fcis-8581	108	59	the	the	DET
fcis-8581	108	60	matrices	matrix	NOUN
fcis-8581	108	61	u	u	NOUN
fcis-8581	108	62	and	and	CCONJ
fcis-8581	108	63	v	v	NOUN
fcis-8581	108	64	,	,	PUNCT
fcis-8581	108	65	i.e.	i.e.	X
fcis-8581	108	66	,	,	PUNCT
fcis-8581	108	67	equation	equation	NOUN
fcis-8581	108	68	(	(	PUNCT
fcis-8581	108	69	13	13	NUM
fcis-8581	108	70	):	):	PUNCT
fcis-8581	108	71	∑	∑	ADV
fcis-8581	108	72	,	,	PUNCT
fcis-8581	108	73	,	,	PUNCT
fcis-8581	108	74	,	,	PUNCT
fcis-8581	108	75	⋋	⋋	NUM
fcis-8581	108	76	∑	∑	ADP
fcis-8581	108	77	⋋	⋋	NUM
fcis-8581	108	78	∑	∑	PUNCT
fcis-8581	108	79	(	(	PUNCT
fcis-8581	108	80	13	13	NUM
fcis-8581	108	81	)	)	PUNCT
fcis-8581	108	82	where	where	SCONJ
fcis-8581	108	83	,	,	PUNCT
fcis-8581	108	84	denotes	denote	VERB
fcis-8581	108	85	the	the	DET
fcis-8581	108	86	original	original	ADJ
fcis-8581	108	87	rating	rating	NOUN
fcis-8581	108	88	of	of	ADP
fcis-8581	108	89	the	the	DET
fcis-8581	108	90	user	user	NOUN
fcis-8581	108	91	,	,	PUNCT
fcis-8581	108	92	̂	̂	X
fcis-8581	108	93	,	,	PUNCT
fcis-8581	108	94	denotes	denote	VERB
fcis-8581	108	95	the	the	DET
fcis-8581	108	96	predicted	predict	VERB
fcis-8581	108	97	value	value	NOUN
fcis-8581	108	98	of	of	ADP
fcis-8581	108	99	the	the	DET
fcis-8581	108	100	user	user	NOUN
fcis-8581	108	101	's	's	PART
fcis-8581	108	102	rating	rating	NOUN
fcis-8581	108	103	using	use	VERB
fcis-8581	108	104	this	this	DET
fcis-8581	108	105	model	model	NOUN
fcis-8581	108	106	,	,	PUNCT
fcis-8581	108	107	and	and	CCONJ
fcis-8581	108	108	λ	λ	PROPN
fcis-8581	108	109	is	be	AUX
fcis-8581	108	110	a	a	DET
fcis-8581	108	111	regularization	regularization	NOUN
fcis-8581	108	112	parameter	parameter	NOUN
fcis-8581	108	113	that	that	PRON
fcis-8581	108	114	varies	vary	VERB
fcis-8581	108	115	continuously	continuously	ADV
fcis-8581	108	116	with	with	ADP
fcis-8581	108	117	the	the	DET
fcis-8581	108	118	operation	operation	NOUN
fcis-8581	108	119	.	.	PUNCT
fcis-8581	109	1	the	the	DET
fcis-8581	109	2	regularization	regularization	NOUN
fcis-8581	109	3	terms	term	NOUN
fcis-8581	109	4	of	of	ADP
fcis-8581	109	5	u	u	NOUN
fcis-8581	109	6	and	and	CCONJ
fcis-8581	109	7	v	v	NOUN
fcis-8581	109	8	are	be	AUX
fcis-8581	109	9	added	add	VERB
fcis-8581	109	10	to	to	ADP
fcis-8581	109	11	the	the	DET
fcis-8581	109	12	objective	objective	ADJ
fcis-8581	109	13	function	function	NOUN
fcis-8581	109	14	in	in	ADP
fcis-8581	109	15	eq	eq	ADP
fcis-8581	109	16	.	.	PUNCT
fcis-8581	110	1	(	(	PUNCT
fcis-8581	110	2	13	13	NUM
fcis-8581	110	3	)	)	PUNCT
fcis-8581	110	4	,	,	PUNCT
fcis-8581	110	5	which	which	PRON
fcis-8581	110	6	in	in	ADP
fcis-8581	110	7	turn	turn	NOUN
fcis-8581	110	8	controls	control	VERB
fcis-8581	110	9	the	the	DET
fcis-8581	110	10	complexity	complexity	NOUN
fcis-8581	110	11	of	of	ADP
fcis-8581	110	12	the	the	DET
fcis-8581	110	13	project	project	NOUN
fcis-8581	110	14	.	.	PUNCT
fcis-8581	111	1	λ	λ	NOUN
fcis-8581	111	2	indicates	indicate	VERB
fcis-8581	111	3	the	the	DET
fcis-8581	111	4	ability	ability	NOUN
fcis-8581	111	5	of	of	ADP
fcis-8581	111	6	this	this	DET
fcis-8581	111	7	regularization	regularization	NOUN
fcis-8581	111	8	term	term	NOUN
fcis-8581	111	9	to	to	PART
fcis-8581	111	10	constrain	constrain	VERB
fcis-8581	111	11	the	the	DET
fcis-8581	111	12	model	model	NOUN
fcis-8581	111	13	.	.	PUNCT
fcis-8581	112	1	if	if	SCONJ
fcis-8581	112	2	λ	λ	PROPN
fcis-8581	112	3	is	be	AUX
fcis-8581	112	4	too	too	ADV
fcis-8581	112	5	small	small	ADJ
fcis-8581	112	6	,	,	PUNCT
fcis-8581	112	7	this	this	DET
fcis-8581	112	8	constraint	constraint	NOUN
fcis-8581	112	9	is	be	AUX
fcis-8581	112	10	not	not	PART
fcis-8581	112	11	sufficient	sufficient	ADJ
fcis-8581	112	12	to	to	PART
fcis-8581	112	13	reduce	reduce	VERB
fcis-8581	112	14	the	the	DET
fcis-8581	112	15	complexity	complexity	NOUN
fcis-8581	112	16	,	,	PUNCT
fcis-8581	112	17	and	and	CCONJ
fcis-8581	112	18	if	if	SCONJ
fcis-8581	112	19	λ	λ	NOUN
fcis-8581	112	20	is	be	AUX
fcis-8581	112	21	too	too	ADV
fcis-8581	112	22	large	large	ADJ
fcis-8581	112	23	,	,	PUNCT
fcis-8581	112	24	it	it	PRON
fcis-8581	112	25	may	may	AUX
fcis-8581	112	26	indicate	indicate	VERB
fcis-8581	112	27	the	the	DET
fcis-8581	112	28	loss	loss	NOUN
fcis-8581	112	29	of	of	ADP
fcis-8581	112	30	some	some	DET
fcis-8581	112	31	important	important	ADJ
fcis-8581	112	32	parameters	parameter	NOUN
fcis-8581	112	33	,	,	PUNCT
fcis-8581	112	34	which	which	PRON
fcis-8581	112	35	in	in	ADP
fcis-8581	112	36	turn	turn	NOUN
fcis-8581	112	37	leads	lead	VERB
fcis-8581	112	38	to	to	ADP
fcis-8581	112	39	a	a	DET
fcis-8581	112	40	decrease	decrease	NOUN
fcis-8581	112	41	in	in	ADP
fcis-8581	112	42	model	model	NOUN
fcis-8581	112	43	accuracy	accuracy	NOUN
fcis-8581	112	44	.	.	PUNCT
fcis-8581	113	1	in	in	ADP
fcis-8581	113	2	the	the	DET
fcis-8581	113	3	actual	actual	ADJ
fcis-8581	113	4	operation	operation	NOUN
fcis-8581	113	5	process	process	NOUN
fcis-8581	113	6	,	,	PUNCT
fcis-8581	113	7	it	it	PRON
fcis-8581	113	8	is	be	AUX
fcis-8581	113	9	necessary	necessary	ADJ
fcis-8581	113	10	to	to	PART
fcis-8581	113	11	balance	balance	VERB
fcis-8581	113	12	the	the	DET
fcis-8581	113	13	degree	degree	NOUN
fcis-8581	113	14	of	of	ADP
fcis-8581	113	15	fit	fit	NOUN
fcis-8581	113	16	and	and	CCONJ
fcis-8581	113	17	accuracy	accuracy	NOUN
fcis-8581	113	18	of	of	ADP
fcis-8581	113	19	the	the	DET
fcis-8581	113	20	model	model	NOUN
fcis-8581	113	21	by	by	ADP
fcis-8581	113	22	constantly	constantly	ADV
fcis-8581	113	23	grid	grid	VERB
fcis-8581	113	24	operations	operation	NOUN
fcis-8581	113	25	to	to	PART
fcis-8581	113	26	find	find	VERB
fcis-8581	113	27	the	the	DET
fcis-8581	113	28	appropriate	appropriate	ADJ
fcis-8581	113	29	λ	λ	NOUN
fcis-8581	113	30	.	.	PROPN
fcis-8581	113	31	3.4	3.4	NUM
fcis-8581	113	32	.	.	PUNCT
fcis-8581	114	1	alternating	alternate	VERB
fcis-8581	114	2	least	least	ADJ
fcis-8581	114	3	squares	square	NOUN
fcis-8581	114	4	method	method	NOUN
fcis-8581	114	5	in	in	ADP
fcis-8581	114	6	the	the	DET
fcis-8581	114	7	recommendation	recommendation	NOUN
fcis-8581	114	8	system	system	NOUN
fcis-8581	114	9	of	of	ADP
fcis-8581	114	10	matrix	matrix	NOUN
fcis-8581	114	11	decomposition	decomposition	NOUN
fcis-8581	114	12	used	use	VERB
fcis-8581	114	13	,	,	PUNCT
fcis-8581	114	14	there	there	PRON
fcis-8581	114	15	are	be	VERB
fcis-8581	114	16	usually	usually	ADV
fcis-8581	114	17	stochastic	stochastic	ADJ
fcis-8581	114	18	gradient	gradient	ADJ
fcis-8581	114	19	descent	descent	NOUN
fcis-8581	114	20	(	(	PUNCT
fcis-8581	114	21	sgd	sgd	PROPN
fcis-8581	114	22	)	)	PUNCT
fcis-8581	115	1	[	[	X
fcis-8581	115	2	4	4	X
fcis-8581	115	3	]	]	PUNCT
fcis-8581	115	4	and	and	CCONJ
fcis-8581	115	5	alternating	alternate	VERB
fcis-8581	115	6	least	least	ADJ
fcis-8581	115	7	squares	square	NOUN
fcis-8581	115	8	(	(	PUNCT
fcis-8581	115	9	als	als	NOUN
fcis-8581	115	10	)	)	PUNCT
fcis-8581	116	1	[	[	X
fcis-8581	116	2	6	6	NUM
fcis-8581	116	3	]	]	PUNCT
fcis-8581	116	4	for	for	ADP
fcis-8581	116	5	the	the	DET
fcis-8581	116	6	solution	solution	NOUN
fcis-8581	116	7	of	of	ADP
fcis-8581	116	8	the	the	DET
fcis-8581	116	9	loss	loss	NOUN
fcis-8581	116	10	function	function	NOUN
fcis-8581	116	11	(	(	PUNCT
fcis-8581	116	12	e.g.	e.g.	ADV
fcis-8581	116	13	,	,	PUNCT
fcis-8581	116	14	equation	equation	NOUN
fcis-8581	116	15	(	(	PUNCT
fcis-8581	116	16	13	13	NUM
fcis-8581	116	17	)	)	PUNCT
fcis-8581	116	18	)	)	PUNCT
fcis-8581	116	19	,	,	PUNCT
fcis-8581	116	20	and	and	CCONJ
fcis-8581	116	21	the	the	DET
fcis-8581	116	22	alternating	alternate	VERB
fcis-8581	116	23	least	least	ADJ
fcis-8581	116	24	squares	square	NOUN
fcis-8581	116	25	(	(	PUNCT
fcis-8581	116	26	als	als	NOUN
fcis-8581	116	27	)	)	PUNCT
fcis-8581	116	28	method	method	NOUN
fcis-8581	116	29	is	be	AUX
fcis-8581	116	30	used	use	VERB
fcis-8581	116	31	in	in	ADP
fcis-8581	116	32	this	this	DET
fcis-8581	116	33	paper	paper	NOUN
fcis-8581	116	34	to	to	PART
fcis-8581	116	35	optimize	optimize	VERB
fcis-8581	116	36	the	the	DET
fcis-8581	116	37	loss	loss	NOUN
fcis-8581	116	38	function	function	NOUN
fcis-8581	116	39	,	,	PUNCT
fcis-8581	116	40	and	and	CCONJ
fcis-8581	116	41	the	the	DET
fcis-8581	116	42	solution	solution	NOUN
fcis-8581	116	43	process	process	NOUN
fcis-8581	116	44	of	of	ADP
fcis-8581	116	45	the	the	DET
fcis-8581	116	46	alternating	alternate	VERB
fcis-8581	116	47	least	least	ADJ
fcis-8581	116	48	squares	square	NOUN
fcis-8581	116	49	method	method	NOUN
fcis-8581	116	50	is	be	AUX
fcis-8581	116	51	described	describe	VERB
fcis-8581	116	52	below	below	ADV
fcis-8581	116	53	.	.	PUNCT
fcis-8581	117	1	the	the	DET
fcis-8581	117	2	alternating	alternate	VERB
fcis-8581	117	3	least	least	ADJ
fcis-8581	117	4	squares	square	NOUN
fcis-8581	117	5	method	method	NOUN
fcis-8581	117	6	uses	use	VERB
fcis-8581	117	7	the	the	DET
fcis-8581	117	8	idea	idea	NOUN
fcis-8581	117	9	of	of	ADP
fcis-8581	117	10	constantly	constantly	ADV
fcis-8581	117	11	performing	perform	VERB
fcis-8581	117	12	further	further	ADJ
fcis-8581	117	13	updates	update	NOUN
fcis-8581	117	14	to	to	PART
fcis-8581	117	15	solve	solve	VERB
fcis-8581	117	16	the	the	DET
fcis-8581	117	17	objective	objective	ADJ
fcis-8581	117	18	function	function	NOUN
fcis-8581	117	19	.	.	PUNCT
fcis-8581	118	1	when	when	SCONJ
fcis-8581	118	2	updating	update	VERB
fcis-8581	118	3	a	a	DET
fcis-8581	118	4	parameter	parameter	NOUN
fcis-8581	118	5	to	to	PART
fcis-8581	118	6	be	be	AUX
fcis-8581	118	7	updated	update	VERB
fcis-8581	118	8	,	,	PUNCT
fcis-8581	118	9	the	the	DET
fcis-8581	118	10	control	control	NOUN
fcis-8581	118	11	variables	variable	NOUN
fcis-8581	118	12	method	method	NOUN
fcis-8581	118	13	is	be	AUX
fcis-8581	118	14	used	use	VERB
fcis-8581	118	15	to	to	PART
fcis-8581	118	16	treat	treat	VERB
fcis-8581	118	17	the	the	DET
fcis-8581	118	18	other	other	ADJ
fcis-8581	118	19	parameters	parameter	NOUN
fcis-8581	118	20	as	as	ADP
fcis-8581	118	21	fixed	fix	VERB
fcis-8581	118	22	values	value	NOUN
fcis-8581	118	23	,	,	PUNCT
fcis-8581	118	24	so	so	SCONJ
fcis-8581	118	25	that	that	SCONJ
fcis-8581	118	26	the	the	DET
fcis-8581	118	27	objective	objective	ADJ
fcis-8581	118	28	function	function	NOUN
fcis-8581	118	29	is	be	AUX
fcis-8581	118	30	reduced	reduce	VERB
fcis-8581	118	31	to	to	ADP
fcis-8581	118	32	a	a	DET
fcis-8581	118	33	quadratic	quadratic	ADJ
fcis-8581	118	34	equation	equation	NOUN
fcis-8581	118	35	,	,	PUNCT
fcis-8581	118	36	which	which	PRON
fcis-8581	118	37	in	in	ADP
fcis-8581	118	38	turn	turn	NOUN
fcis-8581	118	39	leads	lead	VERB
fcis-8581	118	40	to	to	ADP
fcis-8581	118	41	an	an	DET
fcis-8581	118	42	optimal	optimal	ADJ
fcis-8581	118	43	solution	solution	NOUN
fcis-8581	118	44	.	.	PUNCT
fcis-8581	119	1	taking	take	VERB
fcis-8581	119	2	equation	equation	NOUN
fcis-8581	119	3	(	(	PUNCT
fcis-8581	119	4	13	13	NUM
fcis-8581	119	5	)	)	PUNCT
fcis-8581	119	6	as	as	ADP
fcis-8581	119	7	an	an	DET
fcis-8581	119	8	example	example	NOUN
fcis-8581	119	9	,	,	PUNCT
fcis-8581	119	10	the	the	DET
fcis-8581	119	11	potential	potential	ADJ
fcis-8581	119	12	factor	factor	NOUN
fcis-8581	119	13	matrix	matrix	NOUN
fcis-8581	119	14	v	v	NOUN
fcis-8581	119	15	is	be	AUX
fcis-8581	119	16	first	first	ADV
fcis-8581	119	17	randomly	randomly	ADV
fcis-8581	119	18	generated	generate	VERB
fcis-8581	119	19	,	,	PUNCT
fcis-8581	119	20	and	and	CCONJ
fcis-8581	119	21	then	then	ADV
fcis-8581	119	22	the	the	DET
fcis-8581	119	23	bias	bias	NOUN
fcis-8581	119	24	derivative	derivative	NOUN
fcis-8581	119	25	is	be	AUX
fcis-8581	119	26	found	find	VERB
fcis-8581	119	27	for	for	ADP
fcis-8581	119	28	the	the	DET
fcis-8581	119	29	user	user	NOUN
fcis-8581	119	30	potential	potential	ADJ
fcis-8581	119	31	factor	factor	NOUN
fcis-8581	119	32	matrix	matrix	NOUN
fcis-8581	119	33	,	,	PUNCT
fcis-8581	119	34	and	and	CCONJ
fcis-8581	119	35	the	the	DET
fcis-8581	119	36	result	result	NOUN
fcis-8581	119	37	is	be	AUX
fcis-8581	119	38	as	as	SCONJ
fcis-8581	119	39	follows	follow	VERB
fcis-8581	119	40	equation	equation	NOUN
fcis-8581	119	41	(	(	PUNCT
fcis-8581	119	42	14	14	NUM
fcis-8581	119	43	):	):	SYM
fcis-8581	119	44	2	2	NUM
fcis-8581	119	45	2	2	NUM
fcis-8581	119	46	⋋	⋋	NUM
fcis-8581	119	47	(	(	PUNCT
fcis-8581	119	48	14	14	NUM
fcis-8581	119	49	)	)	PUNCT
fcis-8581	119	50	according	accord	VERB
fcis-8581	119	51	to	to	ADP
fcis-8581	119	52	the	the	DET
fcis-8581	119	53	symmetry	symmetry	NOUN
fcis-8581	119	54	,	,	PUNCT
fcis-8581	119	55	the	the	DET
fcis-8581	119	56	potential	potential	ADJ
fcis-8581	119	57	factor	factor	NOUN
fcis-8581	119	58	matrix	matrix	NOUN
fcis-8581	119	59	u	u	NOUN
fcis-8581	119	60	is	be	AUX
fcis-8581	119	61	fixed	fix	VERB
fcis-8581	119	62	and	and	CCONJ
fcis-8581	119	63	the	the	DET
fcis-8581	119	64	partial	partial	ADJ
fcis-8581	119	65	derivatives	derivative	NOUN
fcis-8581	119	66	are	be	AUX
fcis-8581	119	67	obtained	obtain	VERB
fcis-8581	119	68	for	for	ADP
fcis-8581	119	69	the	the	DET
fcis-8581	119	70	potential	potential	ADJ
fcis-8581	119	71	factor	factor	NOUN
fcis-8581	119	72	vectors	vector	NOUN
fcis-8581	119	73	of	of	ADP
fcis-8581	119	74	the	the	DET
fcis-8581	119	75	users	user	NOUN
fcis-8581	119	76	as	as	ADP
fcis-8581	119	77	in	in	ADP
fcis-8581	119	78	equation	equation	NOUN
fcis-8581	119	79	(	(	PUNCT
fcis-8581	119	80	15	15	NUM
fcis-8581	119	81	):	):	SYM
fcis-8581	119	82	2	2	NUM
fcis-8581	119	83	2	2	NUM
fcis-8581	119	84	⋋	⋋	NUM
fcis-8581	119	85	(	(	PUNCT
fcis-8581	119	86	15	15	NUM
fcis-8581	119	87	)	)	PUNCT
fcis-8581	119	88	let	let	VERB
fcis-8581	119	89	both	both	DET
fcis-8581	119	90	equation	equation	NOUN
fcis-8581	119	91	(	(	PUNCT
fcis-8581	119	92	14	14	NUM
fcis-8581	119	93	)	)	PUNCT
fcis-8581	119	94	and	and	CCONJ
fcis-8581	119	95	equation	equation	NOUN
fcis-8581	119	96	(	(	PUNCT
fcis-8581	119	97	15	15	NUM
fcis-8581	119	98	)	)	PUNCT
fcis-8581	119	99	be	be	VERB
fcis-8581	119	100	0	0	NUM
fcis-8581	119	101	.	.	PUNCT
fcis-8581	120	1	iterative	iterative	ADJ
fcis-8581	120	2	computation	computation	NOUN
fcis-8581	120	3	is	be	AUX
fcis-8581	120	4	performed	perform	VERB
fcis-8581	120	5	by	by	ADP
fcis-8581	120	6	calculating	calculate	VERB
fcis-8581	120	7	the	the	DET
fcis-8581	120	8	extreme	extreme	ADJ
fcis-8581	120	9	values	value	NOUN
fcis-8581	120	10	of	of	ADP
fcis-8581	120	11	the	the	DET
fcis-8581	120	12	binary	binary	ADJ
fcis-8581	120	13	function	function	NOUN
fcis-8581	120	14	to	to	PART
fcis-8581	120	15	find	find	VERB
fcis-8581	120	16	the	the	DET
fcis-8581	120	17	suitable	suitable	ADJ
fcis-8581	120	18	and	and	CCONJ
fcis-8581	120	19	thus	thus	ADV
fcis-8581	120	20	the	the	DET
fcis-8581	120	21	desired	desire	VERB
fcis-8581	120	22	objective	objective	ADJ
fcis-8581	120	23	function	function	NOUN
fcis-8581	120	24	.	.	PUNCT
fcis-8581	121	1	each	each	DET
fcis-8581	121	2	complete	complete	ADJ
fcis-8581	121	3	traversal	traversal	NOUN
fcis-8581	121	4	of	of	ADP
fcis-8581	121	5	the	the	DET
fcis-8581	121	6	entire	entire	ADJ
fcis-8581	121	7	data	datum	NOUN
fcis-8581	121	8	set	set	VERB
fcis-8581	121	9	is	be	AUX
fcis-8581	121	10	called	call	VERB
fcis-8581	121	11	an	an	DET
fcis-8581	121	12	iteration	iteration	NOUN
fcis-8581	121	13	.	.	PUNCT
fcis-8581	122	1	there	there	PRON
fcis-8581	122	2	are	be	VERB
fcis-8581	122	3	generally	generally	ADV
fcis-8581	122	4	three	three	NUM
fcis-8581	122	5	ways	way	NOUN
fcis-8581	122	6	to	to	PART
fcis-8581	122	7	end	end	VERB
fcis-8581	122	8	an	an	DET
fcis-8581	122	9	iteration	iteration	NOUN
fcis-8581	122	10	:	:	PUNCT
fcis-8581	122	11	the	the	DET
fcis-8581	122	12	first	first	ADJ
fcis-8581	122	13	is	be	AUX
fcis-8581	122	14	to	to	PART
fcis-8581	122	15	set	set	VERB
fcis-8581	122	16	a	a	DET
fcis-8581	122	17	threshold	threshold	NOUN
fcis-8581	122	18	value	value	NOUN
fcis-8581	122	19	and	and	CCONJ
fcis-8581	122	20	end	end	VERB
fcis-8581	122	21	the	the	DET
fcis-8581	122	22	training	training	NOUN
fcis-8581	122	23	when	when	SCONJ
fcis-8581	122	24	the	the	DET
fcis-8581	122	25	desired	desire	VERB
fcis-8581	122	26	objective	objective	ADJ
fcis-8581	122	27	function	function	NOUN
fcis-8581	122	28	value	value	NOUN
fcis-8581	122	29	is	be	AUX
fcis-8581	122	30	below	below	ADP
fcis-8581	122	31	the	the	DET
fcis-8581	122	32	threshold	threshold	NOUN
fcis-8581	122	33	,	,	PUNCT
fcis-8581	122	34	which	which	PRON
fcis-8581	122	35	is	be	AUX
fcis-8581	122	36	very	very	ADV
fcis-8581	122	37	simple	simple	ADJ
fcis-8581	122	38	and	and	CCONJ
fcis-8581	122	39	different	different	ADJ
fcis-8581	122	40	from	from	ADP
fcis-8581	122	41	the	the	DET
fcis-8581	122	42	operation	operation	NOUN
fcis-8581	122	43	,	,	PUNCT
fcis-8581	122	44	but	but	CCONJ
fcis-8581	122	45	is	be	AUX
fcis-8581	122	46	usually	usually	ADV
fcis-8581	122	47	used	use	VERB
fcis-8581	122	48	rarely	rarely	ADV
fcis-8581	122	49	;	;	PUNCT
fcis-8581	122	50	the	the	DET
fcis-8581	122	51	second	second	ADJ
fcis-8581	122	52	method	method	NOUN
fcis-8581	122	53	is	be	AUX
fcis-8581	122	54	:	:	PUNCT
fcis-8581	122	55	set	set	VERB
fcis-8581	122	56	a	a	DET
fcis-8581	122	57	threshold	threshold	NOUN
fcis-8581	122	58	value	value	NOUN
fcis-8581	122	59	,	,	PUNCT
fcis-8581	122	60	calculate	calculate	VERB
fcis-8581	122	61	the	the	DET
fcis-8581	122	62	absolute	absolute	ADJ
fcis-8581	122	63	value	value	NOUN
fcis-8581	122	64	of	of	ADP
fcis-8581	122	65	the	the	DET
fcis-8581	122	66	difference	difference	NOUN
fcis-8581	122	67	between	between	ADP
fcis-8581	122	68	the	the	DET
fcis-8581	122	69	objective	objective	ADJ
fcis-8581	122	70	function	function	NOUN
fcis-8581	122	71	value	value	NOUN
fcis-8581	122	72	of	of	ADP
fcis-8581	122	73	this	this	DET
fcis-8581	122	74	iteration	iteration	NOUN
fcis-8581	122	75	and	and	CCONJ
fcis-8581	122	76	the	the	DET
fcis-8581	122	77	objective	objective	ADJ
fcis-8581	122	78	function	function	NOUN
fcis-8581	122	79	value	value	NOUN
fcis-8581	122	80	of	of	ADP
fcis-8581	122	81	the	the	DET
fcis-8581	122	82	next	next	ADJ
fcis-8581	122	83	iteration	iteration	NOUN
fcis-8581	122	84	,	,	PUNCT
fcis-8581	122	85	and	and	CCONJ
fcis-8581	122	86	terminate	terminate	VERB
fcis-8581	122	87	the	the	DET
fcis-8581	122	88	iteration	iteration	NOUN
fcis-8581	122	89	if	if	SCONJ
fcis-8581	122	90	it	it	PRON
fcis-8581	122	91	is	be	AUX
fcis-8581	122	92	less	less	ADJ
fcis-8581	122	93	than	than	ADP
fcis-8581	122	94	the	the	DET
fcis-8581	122	95	threshold	threshold	NOUN
fcis-8581	122	96	value	value	NOUN
fcis-8581	122	97	;	;	PUNCT
fcis-8581	122	98	the	the	DET
fcis-8581	122	99	third	third	ADJ
fcis-8581	122	100	method	method	NOUN
fcis-8581	122	101	is	be	AUX
fcis-8581	122	102	:	:	PUNCT
fcis-8581	122	103	fixing	fix	VERB
fcis-8581	122	104	the	the	DET
fcis-8581	122	105	number	number	NOUN
fcis-8581	122	106	of	of	ADP
fcis-8581	122	107	iterations	iteration	NOUN
fcis-8581	122	108	.	.	PUNCT
fcis-8581	123	1	in	in	ADP
fcis-8581	123	2	the	the	DET
fcis-8581	123	3	usual	usual	ADJ
fcis-8581	123	4	experimental	experimental	ADJ
fcis-8581	123	5	procedure	procedure	NOUN
fcis-8581	123	6	,	,	PUNCT
fcis-8581	123	7	the	the	DET
fcis-8581	123	8	iterations	iteration	NOUN
fcis-8581	123	9	are	be	AUX
fcis-8581	123	10	usually	usually	ADV
fcis-8581	123	11	performed	perform	VERB
fcis-8581	123	12	by	by	ADP
fcis-8581	123	13	combining	combine	VERB
fcis-8581	123	14	the	the	DET
fcis-8581	123	15	second	second	ADJ
fcis-8581	123	16	and	and	CCONJ
fcis-8581	123	17	third	third	ADJ
fcis-8581	123	18	methods	method	NOUN
fcis-8581	123	19	.	.	PUNCT
fcis-8581	124	1	4	4	X
fcis-8581	124	2	.	.	X
fcis-8581	124	3	summary	summary	NOUN
fcis-8581	124	4	and	and	CCONJ
fcis-8581	124	5	outlook	outlook	NOUN
fcis-8581	124	6	in	in	ADP
fcis-8581	124	7	this	this	DET
fcis-8581	124	8	paper	paper	NOUN
fcis-8581	124	9	,	,	PUNCT
fcis-8581	124	10	two	two	NUM
fcis-8581	124	11	recommendation	recommendation	NOUN
fcis-8581	124	12	techniques	technique	NOUN
fcis-8581	124	13	based	base	VERB
fcis-8581	124	14	on	on	ADP
fcis-8581	124	15	content	content	NOUN
fcis-8581	124	16	and	and	CCONJ
fcis-8581	124	17	matrix	matrix	NOUN
fcis-8581	124	18	-	-	PUNCT
fcis-8581	124	19	based	base	VERB
fcis-8581	124	20	decomposition	decomposition	NOUN
fcis-8581	124	21	are	be	AUX
fcis-8581	124	22	studied	study	VERB
fcis-8581	124	23	and	and	CCONJ
fcis-8581	124	24	analyzed	analyze	VERB
fcis-8581	124	25	for	for	ADP
fcis-8581	124	26	the	the	DET
fcis-8581	124	27	personalized	personalize	VERB
fcis-8581	124	28	movie	movie	NOUN
fcis-8581	124	29	recommendation	recommendation	NOUN
fcis-8581	124	30	problem	problem	NOUN
fcis-8581	124	31	,	,	PUNCT
fcis-8581	124	32	and	and	CCONJ
fcis-8581	124	33	a	a	DET
fcis-8581	124	34	hybrid	hybrid	ADJ
fcis-8581	124	35	recommendation	recommendation	NOUN
fcis-8581	124	36	scheme	scheme	NOUN
fcis-8581	124	37	introducing	introduce	VERB
fcis-8581	124	38	content	content	NOUN
fcis-8581	124	39	-	-	PUNCT
fcis-8581	124	40	based	base	VERB
fcis-8581	124	41	recommendations	recommendation	NOUN
fcis-8581	124	42	is	be	AUX
fcis-8581	124	43	proposed	propose	VERB
fcis-8581	124	44	for	for	ADP
fcis-8581	124	45	the	the	DET
fcis-8581	124	46	classical	classical	ADJ
fcis-8581	124	47	collaborative	collaborative	ADJ
fcis-8581	124	48	filtering	filtering	NOUN
fcis-8581	124	49	algorithm	algorithm	NOUN
fcis-8581	124	50	data	datum	NOUN
fcis-8581	124	51	sparsity	sparsity	NOUN
fcis-8581	124	52	and	and	CCONJ
fcis-8581	124	53	item	item	NOUN
fcis-8581	124	54	cold	cold	ADJ
fcis-8581	124	55	117	117	NUM
fcis-8581	124	56	start	start	NOUN
fcis-8581	124	57	problems	problem	NOUN
fcis-8581	124	58	.	.	PUNCT
fcis-8581	125	1	since	since	SCONJ
fcis-8581	125	2	the	the	DET
fcis-8581	125	3	number	number	NOUN
fcis-8581	125	4	of	of	ADP
fcis-8581	125	5	users	user	NOUN
fcis-8581	125	6	in	in	ADP
fcis-8581	125	7	movie	movie	NOUN
fcis-8581	125	8	recommendation	recommendation	NOUN
fcis-8581	125	9	is	be	AUX
fcis-8581	125	10	generally	generally	ADV
fcis-8581	125	11	much	much	ADV
fcis-8581	125	12	larger	large	ADJ
fcis-8581	125	13	than	than	ADP
fcis-8581	125	14	the	the	DET
fcis-8581	125	15	number	number	NOUN
fcis-8581	125	16	of	of	ADP
fcis-8581	125	17	movies	movie	NOUN
fcis-8581	125	18	,	,	PUNCT
fcis-8581	125	19	the	the	DET
fcis-8581	125	20	item	item	NOUN
fcis-8581	125	21	-	-	PUNCT
fcis-8581	125	22	based	base	VERB
fcis-8581	125	23	collaborative	collaborative	ADJ
fcis-8581	125	24	filtering	filtering	NOUN
fcis-8581	125	25	recommendation	recommendation	NOUN
fcis-8581	125	26	method	method	NOUN
fcis-8581	125	27	is	be	AUX
fcis-8581	125	28	chosen	choose	VERB
fcis-8581	125	29	.	.	PUNCT
fcis-8581	126	1	also	also	ADV
fcis-8581	126	2	considering	consider	VERB
fcis-8581	126	3	some	some	DET
fcis-8581	126	4	characteristics	characteristic	NOUN
fcis-8581	126	5	of	of	ADP
fcis-8581	126	6	movie	movie	NOUN
fcis-8581	126	7	recommendation	recommendation	NOUN
fcis-8581	126	8	,	,	PUNCT
fcis-8581	126	9	several	several	ADJ
fcis-8581	126	10	similarity	similarity	NOUN
fcis-8581	126	11	measures	measure	NOUN
fcis-8581	126	12	used	use	VERB
fcis-8581	126	13	by	by	ADP
fcis-8581	126	14	the	the	DET
fcis-8581	126	15	item	item	NOUN
fcis-8581	126	16	-	-	PUNCT
fcis-8581	126	17	based	base	VERB
fcis-8581	126	18	collaborative	collaborative	ADJ
fcis-8581	126	19	filtering	filtering	NOUN
fcis-8581	126	20	algorithm	algorithm	NOUN
fcis-8581	126	21	are	be	AUX
fcis-8581	126	22	improved	improve	VERB
fcis-8581	126	23	in	in	ADP
fcis-8581	126	24	three	three	NUM
fcis-8581	126	25	aspects	aspect	NOUN
fcis-8581	126	26	,	,	PUNCT
fcis-8581	126	27	namely	namely	ADV
fcis-8581	126	28	,	,	PUNCT
fcis-8581	126	29	user	user	NOUN
fcis-8581	126	30	rating	rating	NOUN
fcis-8581	126	31	characteristics	characteristic	NOUN
fcis-8581	126	32	,	,	PUNCT
fcis-8581	126	33	the	the	DET
fcis-8581	126	34	number	number	NOUN
fcis-8581	126	35	of	of	ADP
fcis-8581	126	36	users	user	NOUN
fcis-8581	126	37	who	who	PRON
fcis-8581	126	38	jointly	jointly	ADV
fcis-8581	126	39	rate	rate	VERB
fcis-8581	126	40	two	two	NUM
fcis-8581	126	41	movies	movie	NOUN
fcis-8581	126	42	and	and	CCONJ
fcis-8581	126	43	the	the	DET
fcis-8581	126	44	time	time	NOUN
fcis-8581	126	45	factor	factor	NOUN
fcis-8581	126	46	of	of	ADP
fcis-8581	126	47	user	user	NOUN
fcis-8581	126	48	rating	rating	NOUN
fcis-8581	126	49	,	,	PUNCT
fcis-8581	126	50	in	in	ADP
fcis-8581	126	51	turn	turn	NOUN
fcis-8581	126	52	.	.	PUNCT
fcis-8581	127	1	the	the	DET
fcis-8581	127	2	improvement	improvement	NOUN
fcis-8581	127	3	of	of	ADP
fcis-8581	127	4	the	the	DET
fcis-8581	127	5	recommendation	recommendation	NOUN
fcis-8581	127	6	algorithm	algorithm	NOUN
fcis-8581	127	7	in	in	ADP
fcis-8581	127	8	this	this	DET
fcis-8581	127	9	paper	paper	NOUN
fcis-8581	127	10	is	be	AUX
fcis-8581	127	11	specific	specific	ADJ
fcis-8581	127	12	to	to	ADP
fcis-8581	127	13	the	the	DET
fcis-8581	127	14	movielens	movielen	NOUN
fcis-8581	127	15	dataset	dataset	VERB
fcis-8581	127	16	,	,	PUNCT
fcis-8581	127	17	but	but	CCONJ
fcis-8581	127	18	the	the	DET
fcis-8581	127	19	idea	idea	NOUN
fcis-8581	127	20	of	of	ADP
fcis-8581	127	21	the	the	DET
fcis-8581	127	22	improvement	improvement	NOUN
fcis-8581	127	23	is	be	AUX
fcis-8581	127	24	somewhat	somewhat	ADV
fcis-8581	127	25	general	general	ADJ
fcis-8581	127	26	.	.	PUNCT
fcis-8581	128	1	when	when	SCONJ
fcis-8581	128	2	designing	design	VERB
fcis-8581	128	3	personalized	personalized	ADJ
fcis-8581	128	4	recommendation	recommendation	NOUN
fcis-8581	128	5	systems	system	NOUN
fcis-8581	128	6	for	for	ADP
fcis-8581	128	7	other	other	ADJ
fcis-8581	128	8	items	item	NOUN
fcis-8581	128	9	,	,	PUNCT
fcis-8581	128	10	we	we	PRON
fcis-8581	128	11	can	can	AUX
fcis-8581	128	12	also	also	ADV
fcis-8581	128	13	improve	improve	VERB
fcis-8581	128	14	the	the	DET
fcis-8581	128	15	similarity	similarity	NOUN
fcis-8581	128	16	measure	measure	NOUN
fcis-8581	128	17	used	use	VERB
fcis-8581	128	18	by	by	ADP
fcis-8581	128	19	the	the	DET
fcis-8581	128	20	collaborative	collaborative	ADJ
fcis-8581	128	21	filtering	filtering	NOUN
fcis-8581	128	22	algorithm	algorithm	NOUN
fcis-8581	128	23	according	accord	VERB
fcis-8581	128	24	to	to	ADP
fcis-8581	128	25	the	the	DET
fcis-8581	128	26	characteristics	characteristic	NOUN
fcis-8581	128	27	of	of	ADP
fcis-8581	128	28	the	the	DET
fcis-8581	128	29	domain	domain	NOUN
fcis-8581	128	30	.	.	PUNCT
fcis-8581	129	1	in	in	ADP
fcis-8581	129	2	this	this	DET
fcis-8581	129	3	paper	paper	NOUN
fcis-8581	129	4	,	,	PUNCT
fcis-8581	129	5	we	we	PRON
fcis-8581	129	6	propose	propose	VERB
fcis-8581	129	7	a	a	DET
fcis-8581	129	8	scheme	scheme	NOUN
fcis-8581	129	9	to	to	PART
fcis-8581	129	10	improve	improve	VERB
fcis-8581	129	11	the	the	DET
fcis-8581	129	12	item	item	NOUN
fcis-8581	129	13	similarity	similarity	NOUN
fcis-8581	129	14	measure	measure	NOUN
fcis-8581	129	15	according	accord	VERB
fcis-8581	129	16	to	to	ADP
fcis-8581	129	17	the	the	DET
fcis-8581	129	18	characteristics	characteristic	NOUN
fcis-8581	129	19	of	of	ADP
fcis-8581	129	20	the	the	DET
fcis-8581	129	21	domain	domain	NOUN
fcis-8581	129	22	to	to	PART
fcis-8581	129	23	which	which	PRON
fcis-8581	129	24	the	the	DET
fcis-8581	129	25	recommended	recommend	VERB
fcis-8581	129	26	items	item	NOUN
fcis-8581	129	27	belong	belong	VERB
fcis-8581	129	28	in	in	ADP
fcis-8581	129	29	the	the	DET
fcis-8581	129	30	study	study	NOUN
fcis-8581	129	31	of	of	ADP
fcis-8581	129	32	recommendation	recommendation	NOUN
fcis-8581	129	33	systems	system	NOUN
fcis-8581	129	34	,	,	PUNCT
fcis-8581	129	35	and	and	CCONJ
fcis-8581	129	36	we	we	PRON
fcis-8581	129	37	also	also	ADV
fcis-8581	129	38	propose	propose	VERB
fcis-8581	129	39	the	the	DET
fcis-8581	129	40	use	use	NOUN
fcis-8581	129	41	of	of	ADP
fcis-8581	129	42	item	item	NOUN
fcis-8581	129	43	attributes	attribute	NOUN
fcis-8581	129	44	to	to	PART
fcis-8581	129	45	solve	solve	VERB
fcis-8581	129	46	the	the	DET
fcis-8581	129	47	problems	problem	NOUN
fcis-8581	129	48	of	of	ADP
fcis-8581	129	49	data	datum	NOUN
fcis-8581	129	50	sparsity	sparsity	NOUN
fcis-8581	129	51	and	and	CCONJ
fcis-8581	129	52	item	item	NOUN
fcis-8581	129	53	cold	cold	ADJ
fcis-8581	129	54	start	start	NOUN
fcis-8581	129	55	in	in	ADP
fcis-8581	129	56	collaborative	collaborative	ADJ
fcis-8581	129	57	item	item	NOUN
fcis-8581	129	58	filtering	filtering	NOUN
fcis-8581	129	59	.	.	PUNCT
fcis-8581	130	1	however	however	ADV
fcis-8581	130	2	,	,	PUNCT
fcis-8581	130	3	the	the	DET
fcis-8581	130	4	optimization	optimization	NOUN
fcis-8581	130	5	of	of	ADP
fcis-8581	130	6	recommendation	recommendation	NOUN
fcis-8581	130	7	methods	method	NOUN
fcis-8581	130	8	is	be	AUX
fcis-8581	130	9	endless	endless	ADJ
fcis-8581	130	10	,	,	PUNCT
fcis-8581	130	11	we	we	PRON
fcis-8581	130	12	can	can	AUX
fcis-8581	130	13	build	build	VERB
fcis-8581	130	14	user	user	NOUN
fcis-8581	130	15	models	model	NOUN
fcis-8581	130	16	and	and	CCONJ
fcis-8581	130	17	design	design	NOUN
fcis-8581	130	18	user	user	NOUN
fcis-8581	130	19	feedback	feedback	NOUN
fcis-8581	130	20	mechanisms	mechanism	NOUN
fcis-8581	130	21	to	to	PART
fcis-8581	130	22	discover	discover	VERB
fcis-8581	130	23	users	user	NOUN
fcis-8581	130	24	'	'	PART
fcis-8581	130	25	interest	interest	NOUN
fcis-8581	130	26	preferences	preference	NOUN
fcis-8581	130	27	,	,	PUNCT
fcis-8581	130	28	or	or	CCONJ
fcis-8581	130	29	we	we	PRON
fcis-8581	130	30	can	can	AUX
fcis-8581	130	31	further	far	ADV
fcis-8581	130	32	use	use	VERB
fcis-8581	130	33	deep	deep	ADJ
fcis-8581	130	34	learning	learning	NOUN
fcis-8581	130	35	methods	method	NOUN
fcis-8581	130	36	to	to	PART
fcis-8581	130	37	train	train	VERB
fcis-8581	130	38	our	our	PRON
fcis-8581	130	39	recommendation	recommendation	NOUN
fcis-8581	130	40	models	model	NOUN
fcis-8581	130	41	.	.	PUNCT
fcis-8581	131	1	references	reference	NOUN
fcis-8581	131	2	[	[	X
fcis-8581	131	3	1	1	NUM
fcis-8581	131	4	]	]	X
fcis-8581	131	5	hua	hua	PROPN
fcis-8581	131	6	-	-	PUNCT
fcis-8581	131	7	fu	fu	PROPN
fcis-8581	131	8	duan	duan	PROPN
fcis-8581	131	9	.	.	PUNCT
fcis-8581	131	10	application	application	NOUN
fcis-8581	131	11	of	of	ADP
fcis-8581	131	12	matrix	matrix	NOUN
fcis-8581	131	13	decomposition	decomposition	NOUN
fcis-8581	131	14	techniques	technique	NOUN
fcis-8581	131	15	considering	consider	VERB
fcis-8581	131	16	time	time	NOUN
fcis-8581	131	17	effects	effect	NOUN
fcis-8581	131	18	in	in	ADP
fcis-8581	131	19	recommender	recommender	NOUN
fcis-8581	131	20	systems[j	systems[j	PROPN
fcis-8581	131	21	]	]	PUNCT
fcis-8581	131	22	.	.	PUNCT
fcis-8581	132	1	microcomputer	microcomputer	NOUN
fcis-8581	132	2	applications	application	NOUN
fcis-8581	132	3	,	,	PUNCT
fcis-8581	132	4	2013	2013	NUM
fcis-8581	132	5	,	,	PUNCT
fcis-8581	132	6	29(03	29(03	NUM
fcis-8581	132	7	):	):	PUNCT
fcis-8581	132	8	53	53	NUM
fcis-8581	132	9	-	-	SYM
fcis-8581	132	10	55	55	NUM
fcis-8581	132	11	+	+	NOUN
fcis-8581	132	12	64	64	NUM
fcis-8581	132	13	.	.	PUNCT
fcis-8581	133	1	[	[	X
fcis-8581	133	2	2	2	NUM
fcis-8581	133	3	]	]	X
fcis-8581	133	4	liang	liang	PROPN
fcis-8581	133	5	-	-	PUNCT
fcis-8581	133	6	jun	jun	PROPN
fcis-8581	133	7	liu	liu	PROPN
fcis-8581	133	8	,	,	PUNCT
fcis-8581	133	9	liu	liu	PROPN
fcis-8581	133	10	yang	yang	PROPN
fcis-8581	133	11	.	.	PUNCT
fcis-8581	134	1	design	design	NOUN
fcis-8581	134	2	and	and	CCONJ
fcis-8581	134	3	implementation	implementation	NOUN
fcis-8581	134	4	of	of	ADP
fcis-8581	134	5	movie	movie	NOUN
fcis-8581	134	6	recommendation	recommendation	NOUN
fcis-8581	134	7	system[j	system[j	NOUN
fcis-8581	134	8	]	]	PUNCT
fcis-8581	134	9	.	.	PUNCT
fcis-8581	135	1	internet	internet	NOUN
fcis-8581	135	2	of	of	ADP
fcis-8581	135	3	things	thing	NOUN
fcis-8581	135	4	technology	technology	NOUN
fcis-8581	135	5	,	,	PUNCT
fcis-8581	135	6	2021	2021	NUM
fcis-8581	135	7	,	,	PUNCT
fcis-8581	135	8	11(03	11(03	NUM
fcis-8581	135	9	):	):	PUNCT
fcis-8581	135	10	86	86	NUM
fcis-8581	135	11	-	-	SYM
fcis-8581	135	12	88	88	NUM
fcis-8581	135	13	+	+	NOUN
fcis-8581	135	14	92	92	NUM
fcis-8581	135	15	.	.	PUNCT
fcis-8581	136	1	[	[	X
fcis-8581	136	2	3	3	NUM
fcis-8581	136	3	]	]	X
fcis-8581	136	4	zong	zong	PROPN
fcis-8581	136	5	-	-	PUNCT
fcis-8581	136	6	yan	yan	PROPN
fcis-8581	136	7	chen	chen	PROPN
fcis-8581	136	8	.	.	PUNCT
fcis-8581	137	1	collaborative	collaborative	ADJ
fcis-8581	137	2	filtering	filtering	NOUN
fcis-8581	137	3	recommendation	recommendation	NOUN
fcis-8581	137	4	algorithm	algorithm	NOUN
fcis-8581	137	5	for	for	ADP
fcis-8581	137	6	sparse	sparse	ADJ
fcis-8581	137	7	data	datum	NOUN
fcis-8581	137	8	optimization[d	optimization[d	PROPN
fcis-8581	137	9	]	]	PUNCT
fcis-8581	137	10	.	.	PUNCT
fcis-8581	138	1	nanjing	nanjing	PROPN
fcis-8581	138	2	university	university	PROPN
fcis-8581	138	3	of	of	ADP
fcis-8581	138	4	posts	post	NOUN
fcis-8581	138	5	and	and	CCONJ
fcis-8581	138	6	telecommunications	telecommunication	NOUN
fcis-8581	138	7	,	,	PUNCT
fcis-8581	138	8	2017	2017	NUM
fcis-8581	138	9	.	.	PUNCT
fcis-8581	139	1	[	[	X
fcis-8581	139	2	4	4	X
fcis-8581	139	3	]	]	PUNCT
fcis-8581	139	4	deuk	deuk	NOUN
fcis-8581	139	5	hee	hee	PROPN
fcis-8581	139	6	park	park	NOUN
fcis-8581	139	7	,	,	PUNCT
fcis-8581	139	8	hyea	hyea	PROPN
fcis-8581	139	9	kyeong	kyeong	PROPN
fcis-8581	139	10	kim	kim	PROPN
fcis-8581	139	11	,	,	PUNCT
fcis-8581	139	12	il	il	PROPN
fcis-8581	139	13	young	young	ADJ
fcis-8581	139	14	choi	choi	NOUN
fcis-8581	139	15	,	,	PUNCT
fcis-8581	139	16	jae	jae	PROPN
fcis-8581	139	17	kyeong	kyeong	PROPN
fcis-8581	139	18	kim	kim	PROPN
fcis-8581	139	19	.	.	PUNCT
fcis-8581	140	1	a	a	DET
fcis-8581	140	2	literature	literature	NOUN
fcis-8581	140	3	review	review	NOUN
fcis-8581	140	4	and	and	CCONJ
fcis-8581	140	5	classification	classification	NOUN
fcis-8581	140	6	of	of	ADP
fcis-8581	140	7	recommender	recommender	NOUN
fcis-8581	140	8	systems	system	NOUN
fcis-8581	140	9	research[j	research[j	NOUN
fcis-8581	140	10	]	]	PUNCT
fcis-8581	140	11	.	.	PUNCT
fcis-8581	141	1	expert	expert	NOUN
fcis-8581	141	2	systems	system	NOUN
fcis-8581	141	3	with	with	ADP
fcis-8581	141	4	applications	application	NOUN
fcis-8581	141	5	,	,	PUNCT
fcis-8581	141	6	2012	2012	NUM
fcis-8581	141	7	,	,	PUNCT
fcis-8581	141	8	39(11	39(11	NUM
fcis-8581	141	9	)	)	PUNCT
fcis-8581	141	10	.	.	PUNCT
fcis-8581	142	1	[	[	X
fcis-8581	142	2	5	5	NUM
fcis-8581	142	3	]	]	X
fcis-8581	142	4	yao	yao	NOUN
fcis-8581	142	5	-	-	PUNCT
fcis-8581	142	6	ning	ne	VERB
fcis-8581	142	7	fang	fang	PROPN
fcis-8581	142	8	,	,	PUNCT
fcis-8581	142	9	yun	yun	PROPN
fcis-8581	142	10	-	-	PUNCT
fcis-8581	142	11	fei	fei	PROPN
fcis-8581	142	12	guo	guo	PROPN
fcis-8581	142	13	,	,	PUNCT
fcis-8581	142	14	xue	xue	PROPN
fcis-8581	142	15	-	-	PUNCT
fcis-8581	142	16	tao	tao	PROPN
fcis-8581	142	17	ding	ding	NOUN
fcis-8581	142	18	,	,	PUNCT
fcis-8581	142	19	ju	ju	NOUN
fcis-8581	142	20	-	-	PUNCT
fcis-8581	142	21	long	long	ADJ
fcis-8581	142	22	lan	lan	PROPN
fcis-8581	142	23	.	.	PUNCT
fcis-8581	143	1	an	an	DET
fcis-8581	143	2	improved	improve	VERB
fcis-8581	143	3	recommendation	recommendation	NOUN
fcis-8581	143	4	algorithm	algorithm	NOUN
fcis-8581	143	5	for	for	ADP
fcis-8581	143	6	singular	singular	ADJ
fcis-8581	143	7	value	value	NOUN
fcis-8581	143	8	decomposition	decomposition	NOUN
fcis-8581	143	9	based	base	VERB
fcis-8581	143	10	on	on	ADP
fcis-8581	143	11	local	local	ADJ
fcis-8581	143	12	structure[j	structure[j	NOUN
fcis-8581	143	13	]	]	PUNCT
fcis-8581	143	14	.	.	PUNCT
fcis-8581	144	1	journal	journal	PROPN
fcis-8581	144	2	of	of	ADP
fcis-8581	144	3	electronics	electronic	NOUN
fcis-8581	144	4	and	and	CCONJ
fcis-8581	144	5	information	information	NOUN
fcis-8581	144	6	,	,	PUNCT
fcis-8581	144	7	2013	2013	NUM
fcis-8581	144	8	,	,	PUNCT
fcis-8581	144	9	35(06	35(06	NUM
fcis-8581	144	10	):	):	PUNCT
fcis-8581	144	11	1284	1284	NUM
fcis-8581	144	12	-	-	SYM
fcis-8581	144	13	1289	1289	NUM
fcis-8581	144	14	.	.	PUNCT
fcis-8581	145	1	[	[	X
fcis-8581	145	2	6	6	NUM
fcis-8581	145	3	]	]	X
fcis-8581	145	4	shi	shi	PROPN
fcis-8581	145	5	-	-	PUNCT
fcis-8581	145	6	chao	chao	PROPN
fcis-8581	145	7	dai	dai	PROPN
fcis-8581	145	8	.	.	PROPN
fcis-8581	145	9	research	research	PROPN
fcis-8581	145	10	on	on	ADP
fcis-8581	145	11	parallelization	parallelization	NOUN
fcis-8581	145	12	of	of	ADP
fcis-8581	145	13	matrix	matrix	NOUN
fcis-8581	145	14	decomposition	decomposition	NOUN
fcis-8581	145	15	based	base	VERB
fcis-8581	145	16	on	on	ADP
fcis-8581	145	17	graph	graph	NOUN
fcis-8581	145	18	computation	computation	NOUN
fcis-8581	145	19	model	model	NOUN
fcis-8581	146	1	[	[	X
fcis-8581	146	2	d	d	X
fcis-8581	146	3	]	]	X
fcis-8581	146	4	.	.	PUNCT
fcis-8581	147	1	zhejiang	zhejiang	PROPN
fcis-8581	147	2	sci	sci	PROPN
fcis-8581	147	3	-	-	PUNCT
fcis-8581	147	4	tech	tech	PROPN
fcis-8581	147	5	university	university	NOUN
fcis-8581	147	6	,	,	PUNCT
fcis-8581	147	7	2016	2016	NUM
fcis-8581	147	8	.	.	PUNCT
fcis-8581	148	1	[	[	X
fcis-8581	148	2	7	7	X
fcis-8581	148	3	]	]	X
fcis-8581	148	4	li	li	PROPN
fcis-8581	148	5	h.	h.	PROPN
fcis-8581	148	6	,	,	PUNCT
fcis-8581	148	7	liu	liu	PROPN
fcis-8581	148	8	y.	y.	PROPN
fcis-8581	148	9	,	,	PUNCT
fcis-8581	148	10	qian	qian	PROPN
fcis-8581	148	11	y.	y.	PROPN
fcis-8581	148	12	et	et	PROPN
fcis-8581	148	13	al	al	PROPN
fcis-8581	148	14	.	.	PROPN
fcis-8581	148	15	hhmf	hhmf	NOUN
fcis-8581	148	16	:	:	PUNCT
fcis-8581	148	17	hidden	hidden	ADJ
fcis-8581	148	18	hierarchical	hierarchical	ADJ
fcis-8581	148	19	matrix	matrix	NOUN
fcis-8581	148	20	factorization	factorization	NOUN
fcis-8581	148	21	for	for	ADP
fcis-8581	148	22	recommender	recommender	NOUN
fcis-8581	148	23	systems	system	NOUN
fcis-8581	148	24	.	.	PUNCT
fcis-8581	149	1	data	data	PROPN
fcis-8581	149	2	min	min	PROPN
fcis-8581	149	3	knowl	knowl	PROPN
fcis-8581	149	4	disc	disc	VERB
fcis-8581	149	5	33	33	NUM
fcis-8581	149	6	,	,	PUNCT
fcis-8581	149	7	1548–1582	1548–1582	NUM
fcis-8581	149	8	(	(	PUNCT
fcis-8581	149	9	2019	2019	NUM
fcis-8581	149	10	)	)	PUNCT
fcis-8581	149	11	.	.	PUNCT
fcis-8581	150	1	[	[	X
fcis-8581	150	2	8	8	NUM
fcis-8581	150	3	]	]	X
fcis-8581	150	4	dong	dong	PROPN
fcis-8581	150	5	-	-	PUNCT
fcis-8581	150	6	ting	ting	PROPN
fcis-8581	150	7	sun	sun	PROPN
fcis-8581	150	8	.	.	PUNCT
fcis-8581	150	9	research	research	NOUN
fcis-8581	150	10	on	on	ADP
fcis-8581	150	11	cold	cold	ADJ
fcis-8581	150	12	start	start	NOUN
fcis-8581	150	13	problem	problem	NOUN
fcis-8581	150	14	in	in	ADP
fcis-8581	150	15	collaborative	collaborative	ADJ
fcis-8581	150	16	filtering	filtering	NOUN
fcis-8581	150	17	recommender	recommender	PROPN
fcis-8581	150	18	system[d	system[d	PROPN
fcis-8581	150	19	]	]	PUNCT
fcis-8581	150	20	.	.	PUNCT
fcis-8581	151	1	national	national	PROPN
fcis-8581	151	2	university	university	PROPN
fcis-8581	151	3	of	of	ADP
fcis-8581	151	4	defense	defense	PROPN
fcis-8581	151	5	technology	technology	NOUN
fcis-8581	151	6	,	,	PUNCT
fcis-8581	151	7	2011	2011	NUM
fcis-8581	151	8	.	.	PUNCT