id	sid	tid	token	lemma	pos
fbem-1048	1	1	frontiers	frontier	NOUN
fbem-1048	1	2	in	in	ADP
fbem-1048	1	3	business	business	NOUN
fbem-1048	1	4	,	,	PUNCT
fbem-1048	1	5	economics	economic	NOUN
fbem-1048	1	6	and	and	CCONJ
fbem-1048	1	7	management	management	NOUN
fbem-1048	1	8	issn	issn	PROPN
fbem-1048	1	9	:	:	PUNCT
fbem-1048	1	10	2766	2766	NUM
fbem-1048	1	11	-	-	PUNCT
fbem-1048	1	12	824x	824x	NUM
fbem-1048	1	13	|	|	ADJ
fbem-1048	1	14	vol	vol	NOUN
fbem-1048	1	15	.	.	PROPN
fbem-1048	2	1	4	4	NUM
fbem-1048	2	2	,	,	PUNCT
fbem-1048	2	3	no	no	INTJ
fbem-1048	2	4	.	.	NOUN
fbem-1048	2	5	3	3	NUM
fbem-1048	2	6	,	,	PUNCT
fbem-1048	2	7	2022	2022	NUM
fbem-1048	2	8	25	25	NUM
fbem-1048	2	9	sentiment	sentiment	NOUN
fbem-1048	2	10	analysis	analysis	NOUN
fbem-1048	2	11	and	and	CCONJ
fbem-1048	2	12	personalized	personalized	ADJ
fbem-1048	2	13	recommendations	recommendation	NOUN
fbem-1048	2	14	based	base	VERB
fbem-1048	2	15	on	on	ADP
fbem-1048	2	16	jd.com	jd.com	PROPN
fbem-1048	2	17	reviews	review	NOUN
fbem-1048	2	18	yutong	yutong	PROPN
fbem-1048	2	19	zhang1	zhang1	PROPN
fbem-1048	2	20	,	,	PUNCT
fbem-1048	2	21	a	a	DET
fbem-1048	2	22	1school	1school	NUM
fbem-1048	2	23	of	of	ADP
fbem-1048	2	24	computer	computer	NOUN
fbem-1048	2	25	science	science	NOUN
fbem-1048	2	26	and	and	CCONJ
fbem-1048	2	27	information	information	NOUN
fbem-1048	2	28	engineering	engineering	PROPN
fbem-1048	2	29	,	,	PUNCT
fbem-1048	2	30	hubei	hubei	PROPN
fbem-1048	2	31	university	university	PROPN
fbem-1048	2	32	,	,	PUNCT
fbem-1048	2	33	wuhan	wuhan	PROPN
fbem-1048	2	34	430000	430000	NUM
fbem-1048	2	35	china	china	PROPN
fbem-1048	2	36	acorresponding	acorresponding	NOUN
fbem-1048	2	37	author	author	NOUN
fbem-1048	2	38	’s	’s	PART
fbem-1048	2	39	email	email	NOUN
fbem-1048	2	40	:	:	PUNCT
fbem-1048	2	41	201931119020041@stu.hubu.edu.cn	201931119020041@stu.hubu.edu.cn	NUM
fbem-1048	2	42	abstract	abstract	NOUN
fbem-1048	2	43	:	:	PUNCT
fbem-1048	2	44	the	the	DET
fbem-1048	2	45	general	general	ADJ
fbem-1048	2	46	big	big	ADJ
fbem-1048	2	47	data	datum	NOUN
fbem-1048	2	48	personalized	personalize	VERB
fbem-1048	2	49	recommendation	recommendation	NOUN
fbem-1048	2	50	is	be	AUX
fbem-1048	2	51	based	base	VERB
fbem-1048	2	52	on	on	ADP
fbem-1048	2	53	the	the	DET
fbem-1048	2	54	number	number	NOUN
fbem-1048	2	55	of	of	ADP
fbem-1048	2	56	times	time	NOUN
fbem-1048	2	57	or	or	CCONJ
fbem-1048	2	58	the	the	DET
fbem-1048	2	59	length	length	NOUN
fbem-1048	2	60	of	of	ADP
fbem-1048	2	61	time	time	NOUN
fbem-1048	2	62	users	user	NOUN
fbem-1048	2	63	click	click	VERB
fbem-1048	2	64	on	on	ADP
fbem-1048	2	65	a	a	DET
fbem-1048	2	66	related	related	ADJ
fbem-1048	2	67	content	content	NOUN
fbem-1048	2	68	,	,	PUNCT
fbem-1048	2	69	but	but	CCONJ
fbem-1048	2	70	in	in	ADP
fbem-1048	2	71	many	many	ADJ
fbem-1048	2	72	cases	case	NOUN
fbem-1048	2	73	,	,	PUNCT
fbem-1048	2	74	these	these	PRON
fbem-1048	2	75	can	can	AUX
fbem-1048	2	76	not	not	PART
fbem-1048	2	77	be	be	AUX
fbem-1048	2	78	the	the	DET
fbem-1048	2	79	most	most	ADV
fbem-1048	2	80	direct	direct	ADJ
fbem-1048	2	81	basis	basis	NOUN
fbem-1048	2	82	for	for	ADP
fbem-1048	2	83	accurate	accurate	ADJ
fbem-1048	2	84	recommendation	recommendation	NOUN
fbem-1048	2	85	,	,	PUNCT
fbem-1048	2	86	and	and	CCONJ
fbem-1048	2	87	there	there	PRON
fbem-1048	2	88	may	may	AUX
fbem-1048	2	89	be	be	AUX
fbem-1048	2	90	cases	case	NOUN
fbem-1048	2	91	such	such	ADJ
fbem-1048	2	92	as	as	ADP
fbem-1048	2	93	wrong	wrong	ADJ
fbem-1048	2	94	clicks	click	NOUN
fbem-1048	2	95	by	by	ADP
fbem-1048	2	96	users	user	NOUN
fbem-1048	2	97	.	.	PUNCT
fbem-1048	3	1	these	these	DET
fbem-1048	3	2	factors	factor	NOUN
fbem-1048	3	3	and	and	CCONJ
fbem-1048	3	4	a	a	DET
fbem-1048	3	5	large	large	ADJ
fbem-1048	3	6	number	number	NOUN
fbem-1048	3	7	of	of	ADP
fbem-1048	3	8	related	related	ADJ
fbem-1048	3	9	products	product	NOUN
fbem-1048	3	10	or	or	CCONJ
fbem-1048	3	11	articles	article	NOUN
fbem-1048	3	12	recommended	recommend	VERB
fbem-1048	3	13	to	to	ADP
fbem-1048	3	14	users	user	NOUN
fbem-1048	3	15	may	may	AUX
fbem-1048	3	16	cause	cause	VERB
fbem-1048	3	17	users	user	NOUN
fbem-1048	3	18	'	'	PART
fbem-1048	3	19	disgust	disgust	NOUN
fbem-1048	3	20	.	.	PUNCT
fbem-1048	4	1	this	this	DET
fbem-1048	4	2	article	article	NOUN
fbem-1048	4	3	conducts	conduct	VERB
fbem-1048	4	4	sentiment	sentiment	NOUN
fbem-1048	4	5	analysis	analysis	NOUN
fbem-1048	4	6	on	on	ADP
fbem-1048	4	7	jd.com	jd.com	PROPN
fbem-1048	4	8	reviews	review	NOUN
fbem-1048	4	9	as	as	ADP
fbem-1048	4	10	an	an	DET
fbem-1048	4	11	example	example	NOUN
fbem-1048	4	12	,	,	PUNCT
fbem-1048	4	13	obtains	obtain	VERB
fbem-1048	4	14	the	the	DET
fbem-1048	4	15	user	user	NOUN
fbem-1048	4	16	's	's	PART
fbem-1048	4	17	likes	like	NOUN
fbem-1048	4	18	and	and	CCONJ
fbem-1048	4	19	dislikes	dislike	NOUN
fbem-1048	4	20	,	,	PUNCT
fbem-1048	4	21	and	and	CCONJ
fbem-1048	4	22	then	then	ADV
fbem-1048	4	23	makes	make	VERB
fbem-1048	4	24	accurate	accurate	ADJ
fbem-1048	4	25	personalized	personalized	ADJ
fbem-1048	4	26	recommendations	recommendation	NOUN
fbem-1048	4	27	,	,	PUNCT
fbem-1048	4	28	so	so	SCONJ
fbem-1048	4	29	that	that	SCONJ
fbem-1048	4	30	a	a	DET
fbem-1048	4	31	greater	great	ADJ
fbem-1048	4	32	understanding	understanding	NOUN
fbem-1048	4	33	of	of	ADP
fbem-1048	4	34	preferences	preference	NOUN
fbem-1048	4	35	can	can	AUX
fbem-1048	4	36	improve	improve	VERB
fbem-1048	4	37	the	the	DET
fbem-1048	4	38	effect	effect	NOUN
fbem-1048	4	39	of	of	ADP
fbem-1048	4	40	recommendations	recommendation	NOUN
fbem-1048	4	41	,	,	PUNCT
fbem-1048	4	42	and	and	CCONJ
fbem-1048	4	43	more	more	ADV
fbem-1048	4	44	accurate	accurate	ADJ
fbem-1048	4	45	personalized	personalized	ADJ
fbem-1048	4	46	recommendations	recommendation	NOUN
fbem-1048	4	47	.	.	PUNCT
fbem-1048	5	1	keywords	keyword	NOUN
fbem-1048	5	2	:	:	PUNCT
fbem-1048	5	3	natural	natural	ADJ
fbem-1048	5	4	language	language	NOUN
fbem-1048	5	5	processing	processing	NOUN
fbem-1048	5	6	,	,	PUNCT
fbem-1048	5	7	sentiment	sentiment	NOUN
fbem-1048	5	8	analysis	analysis	NOUN
fbem-1048	5	9	,	,	PUNCT
fbem-1048	5	10	personalized	personalized	ADJ
fbem-1048	5	11	recommendation	recommendation	NOUN
fbem-1048	5	12	,	,	PUNCT
fbem-1048	5	13	jd.com	jd.com	PROPN
fbem-1048	5	14	review	review	NOUN
fbem-1048	5	15	.	.	PUNCT
fbem-1048	6	1	1	1	X
fbem-1048	6	2	.	.	X
fbem-1048	6	3	introduction	introduction	NOUN
fbem-1048	6	4	nowadays	nowadays	ADV
fbem-1048	6	5	,	,	PUNCT
fbem-1048	6	6	the	the	DET
fbem-1048	6	7	science	science	NOUN
fbem-1048	6	8	and	and	CCONJ
fbem-1048	6	9	technology	technology	NOUN
fbem-1048	6	10	are	be	AUX
fbem-1048	6	11	developed	develop	VERB
fbem-1048	6	12	rapidly	rapidly	ADV
fbem-1048	6	13	,	,	PUNCT
fbem-1048	6	14	people	people	NOUN
fbem-1048	6	15	are	be	AUX
fbem-1048	6	16	in	in	ADP
fbem-1048	6	17	an	an	DET
fbem-1048	6	18	environment	environment	NOUN
fbem-1048	6	19	where	where	SCONJ
fbem-1048	6	20	information	information	NOUN
fbem-1048	6	21	is	be	AUX
fbem-1048	6	22	almost	almost	ADV
fbem-1048	6	23	transparent	transparent	ADJ
fbem-1048	6	24	.	.	PUNCT
fbem-1048	7	1	big	big	ADJ
fbem-1048	7	2	data	datum	NOUN
fbem-1048	7	3	can	can	AUX
fbem-1048	7	4	track	track	VERB
fbem-1048	7	5	any	any	DET
fbem-1048	7	6	user	user	NOUN
fbem-1048	7	7	using	use	VERB
fbem-1048	7	8	the	the	DET
fbem-1048	7	9	internet	internet	NOUN
fbem-1048	7	10	and	and	CCONJ
fbem-1048	7	11	obtain	obtain	VERB
fbem-1048	7	12	all	all	DET
fbem-1048	7	13	their	their	PRON
fbem-1048	7	14	information	information	NOUN
fbem-1048	7	15	exposed	expose	VERB
fbem-1048	7	16	on	on	ADP
fbem-1048	7	17	the	the	DET
fbem-1048	7	18	internet	internet	NOUN
fbem-1048	7	19	.	.	PUNCT
fbem-1048	8	1	many	many	ADJ
fbem-1048	8	2	merchants	merchant	NOUN
fbem-1048	8	3	saw	see	VERB
fbem-1048	8	4	business	business	NOUN
fbem-1048	8	5	opportunities	opportunity	NOUN
fbem-1048	8	6	from	from	ADP
fbem-1048	8	7	it	it	PRON
fbem-1048	8	8	,	,	PUNCT
fbem-1048	8	9	obtained	obtain	VERB
fbem-1048	8	10	a	a	DET
fbem-1048	8	11	large	large	ADJ
fbem-1048	8	12	amount	amount	NOUN
fbem-1048	8	13	of	of	ADP
fbem-1048	8	14	information	information	NOUN
fbem-1048	8	15	through	through	ADP
fbem-1048	8	16	users	user	NOUN
fbem-1048	8	17	'	'	PART
fbem-1048	8	18	use	use	NOUN
fbem-1048	8	19	of	of	ADP
fbem-1048	8	20	their	their	PRON
fbem-1048	8	21	software	software	NOUN
fbem-1048	8	22	,	,	PUNCT
fbem-1048	8	23	and	and	CCONJ
fbem-1048	8	24	analyzed	analyze	VERB
fbem-1048	8	25	the	the	DET
fbem-1048	8	26	data	datum	NOUN
fbem-1048	8	27	through	through	ADP
fbem-1048	8	28	a	a	DET
fbem-1048	8	29	series	series	NOUN
fbem-1048	8	30	of	of	ADP
fbem-1048	8	31	methods	method	NOUN
fbem-1048	8	32	to	to	PART
fbem-1048	8	33	obtain	obtain	VERB
fbem-1048	8	34	each	each	DET
fbem-1048	8	35	user	user	NOUN
fbem-1048	8	36	's	's	PART
fbem-1048	8	37	preferences	preference	NOUN
fbem-1048	8	38	,	,	PUNCT
fbem-1048	8	39	and	and	CCONJ
fbem-1048	8	40	made	make	VERB
fbem-1048	8	41	personalized	personalized	ADJ
fbem-1048	8	42	recommendations	recommendation	NOUN
fbem-1048	8	43	,	,	PUNCT
fbem-1048	8	44	which	which	PRON
fbem-1048	8	45	greatly	greatly	ADV
fbem-1048	8	46	increased	increase	VERB
fbem-1048	8	47	traffic	traffic	NOUN
fbem-1048	8	48	or	or	CCONJ
fbem-1048	8	49	profits	profit	NOUN
fbem-1048	8	50	.	.	PUNCT
fbem-1048	9	1	profit	profit	NOUN
fbem-1048	9	2	target	target	NOUN
fbem-1048	9	3	.	.	PUNCT
fbem-1048	10	1	personalized	personalized	ADJ
fbem-1048	10	2	recommendation	recommendation	NOUN
fbem-1048	10	3	exists	exist	VERB
fbem-1048	10	4	in	in	ADP
fbem-1048	10	5	almost	almost	ADV
fbem-1048	10	6	every	every	PRON
fbem-1048	10	7	software	software	NOUN
fbem-1048	10	8	,	,	PUNCT
fbem-1048	10	9	but	but	CCONJ
fbem-1048	10	10	the	the	DET
fbem-1048	10	11	data	datum	NOUN
fbem-1048	10	12	of	of	ADP
fbem-1048	10	13	most	most	ADJ
fbem-1048	10	14	software	software	NOUN
fbem-1048	10	15	personalized	personalize	VERB
fbem-1048	10	16	recommendation	recommendation	NOUN
fbem-1048	10	17	basically	basically	ADV
fbem-1048	10	18	comes	come	VERB
fbem-1048	10	19	from	from	ADP
fbem-1048	10	20	the	the	DET
fbem-1048	10	21	length	length	NOUN
fbem-1048	10	22	of	of	ADP
fbem-1048	10	23	time	time	NOUN
fbem-1048	10	24	users	user	NOUN
fbem-1048	10	25	stay	stay	VERB
fbem-1048	10	26	in	in	ADP
fbem-1048	10	27	a	a	DET
fbem-1048	10	28	certain	certain	ADJ
fbem-1048	10	29	aspect	aspect	NOUN
fbem-1048	10	30	of	of	ADP
fbem-1048	10	31	content	content	NOUN
fbem-1048	10	32	or	or	CCONJ
fbem-1048	10	33	the	the	DET
fbem-1048	10	34	number	number	NOUN
fbem-1048	10	35	of	of	ADP
fbem-1048	10	36	clicks	click	NOUN
fbem-1048	10	37	,	,	PUNCT
fbem-1048	10	38	but	but	CCONJ
fbem-1048	10	39	in	in	ADP
fbem-1048	10	40	fact	fact	NOUN
fbem-1048	10	41	not	not	PART
fbem-1048	10	42	all	all	PRON
fbem-1048	10	43	of	of	ADP
fbem-1048	10	44	these	these	DET
fbem-1048	10	45	data	datum	NOUN
fbem-1048	10	46	are	be	AUX
fbem-1048	10	47	is	be	AUX
fbem-1048	10	48	the	the	DET
fbem-1048	10	49	user	user	NOUN
fbem-1048	10	50	's	's	PART
fbem-1048	10	51	real	real	ADJ
fbem-1048	10	52	preferences	preference	NOUN
fbem-1048	10	53	.	.	PUNCT
fbem-1048	11	1	in	in	ADP
fbem-1048	11	2	china	china	PROPN
fbem-1048	11	3	,	,	PUNCT
fbem-1048	11	4	jd.com	jd.com	PROPN
fbem-1048	11	5	which	which	PRON
fbem-1048	11	6	is	be	AUX
fbem-1048	11	7	one	one	NUM
fbem-1048	11	8	of	of	ADP
fbem-1048	11	9	the	the	DET
fbem-1048	11	10	biggest	big	ADJ
fbem-1048	11	11	e	e	NOUN
fbem-1048	11	12	-	-	ADJ
fbem-1048	11	13	commerce	commerce	NOUN
fbem-1048	11	14	companies	company	NOUN
fbem-1048	11	15	collects	collect	VERB
fbem-1048	11	16	a	a	DET
fbem-1048	11	17	large	large	ADJ
fbem-1048	11	18	mass	mass	NOUN
fbem-1048	11	19	of	of	ADP
fbem-1048	11	20	data	datum	NOUN
fbem-1048	11	21	on	on	ADP
fbem-1048	11	22	user	user	NOUN
fbem-1048	11	23	reviews	review	NOUN
fbem-1048	11	24	as	as	SCONJ
fbem-1048	11	25	many	many	ADJ
fbem-1048	11	26	users	user	NOUN
fbem-1048	11	27	place	place	VERB
fbem-1048	11	28	orders	order	NOUN
fbem-1048	11	29	for	for	ADP
fbem-1048	11	30	shopping	shopping	NOUN
fbem-1048	11	31	every	every	DET
fbem-1048	11	32	day	day	NOUN
fbem-1048	11	33	.	.	PUNCT
fbem-1048	12	1	it	it	PRON
fbem-1048	12	2	is	be	AUX
fbem-1048	12	3	not	not	PART
fbem-1048	12	4	difficult	difficult	ADJ
fbem-1048	12	5	to	to	PART
fbem-1048	12	6	find	find	VERB
fbem-1048	12	7	that	that	SCONJ
fbem-1048	12	8	there	there	PRON
fbem-1048	12	9	are	be	VERB
fbem-1048	12	10	many	many	ADJ
fbem-1048	12	11	emotional	emotional	ADJ
fbem-1048	12	12	factors	factor	NOUN
fbem-1048	12	13	in	in	ADP
fbem-1048	12	14	these	these	DET
fbem-1048	12	15	user	user	NOUN
fbem-1048	12	16	comments	comment	NOUN
fbem-1048	12	17	.	.	PUNCT
fbem-1048	13	1	for	for	ADP
fbem-1048	13	2	example	example	NOUN
fbem-1048	13	3	,	,	PUNCT
fbem-1048	13	4	users	user	NOUN
fbem-1048	13	5	who	who	PRON
fbem-1048	13	6	dislike	dislike	VERB
fbem-1048	13	7	will	will	AUX
fbem-1048	13	8	publish	publish	VERB
fbem-1048	13	9	some	some	DET
fbem-1048	13	10	negative	negative	ADJ
fbem-1048	13	11	words	word	NOUN
fbem-1048	13	12	,	,	PUNCT
fbem-1048	13	13	while	while	SCONJ
fbem-1048	13	14	users	user	NOUN
fbem-1048	13	15	who	who	PRON
fbem-1048	13	16	like	like	VERB
fbem-1048	13	17	them	they	PRON
fbem-1048	13	18	will	will	AUX
fbem-1048	13	19	have	have	VERB
fbem-1048	13	20	positive	positive	ADJ
fbem-1048	13	21	words	word	NOUN
fbem-1048	13	22	.	.	PUNCT
fbem-1048	14	1	according	accord	VERB
fbem-1048	14	2	to	to	ADP
fbem-1048	14	3	these	these	DET
fbem-1048	14	4	emotional	emotional	ADJ
fbem-1048	14	5	comments	comment	NOUN
fbem-1048	14	6	,	,	PUNCT
fbem-1048	14	7	the	the	DET
fbem-1048	14	8	software	software	NOUN
fbem-1048	14	9	background	background	NOUN
fbem-1048	14	10	can	can	AUX
fbem-1048	14	11	intuitively	intuitively	ADV
fbem-1048	14	12	see	see	VERB
fbem-1048	14	13	the	the	DET
fbem-1048	14	14	user	user	NOUN
fbem-1048	14	15	's	's	PART
fbem-1048	14	16	satisfaction	satisfaction	NOUN
fbem-1048	14	17	with	with	ADP
fbem-1048	14	18	the	the	DET
fbem-1048	14	19	products	product	NOUN
fbem-1048	14	20	and	and	CCONJ
fbem-1048	14	21	analyze	analyze	VERB
fbem-1048	14	22	whether	whether	SCONJ
fbem-1048	14	23	the	the	DET
fbem-1048	14	24	products	product	NOUN
fbem-1048	14	25	of	of	ADP
fbem-1048	14	26	the	the	DET
fbem-1048	14	27	same	same	ADJ
fbem-1048	14	28	category	category	NOUN
fbem-1048	14	29	or	or	CCONJ
fbem-1048	14	30	derived	derive	VERB
fbem-1048	14	31	supporting	support	VERB
fbem-1048	14	32	products	product	NOUN
fbem-1048	14	33	can	can	AUX
fbem-1048	14	34	attract	attract	VERB
fbem-1048	14	35	users	user	NOUN
fbem-1048	14	36	,	,	PUNCT
fbem-1048	14	37	so	so	SCONJ
fbem-1048	14	38	as	as	SCONJ
fbem-1048	14	39	to	to	PART
fbem-1048	14	40	infer	infer	VERB
fbem-1048	14	41	future	future	ADJ
fbem-1048	14	42	recommendations	recommendation	NOUN
fbem-1048	14	43	for	for	ADP
fbem-1048	14	44	users	user	NOUN
fbem-1048	14	45	.	.	PUNCT
fbem-1048	15	1	content	content	PROPN
fbem-1048	15	2	,	,	PUNCT
fbem-1048	15	3	the	the	DET
fbem-1048	15	4	platform	platform	NOUN
fbem-1048	15	5	adds	add	VERB
fbem-1048	15	6	this	this	DET
fbem-1048	15	7	method	method	NOUN
fbem-1048	15	8	to	to	PART
fbem-1048	15	9	greatly	greatly	ADV
fbem-1048	15	10	improve	improve	VERB
fbem-1048	15	11	the	the	DET
fbem-1048	15	12	effect	effect	NOUN
fbem-1048	15	13	of	of	ADP
fbem-1048	15	14	the	the	DET
fbem-1048	15	15	recommendation	recommendation	NOUN
fbem-1048	15	16	.	.	PUNCT
fbem-1048	16	1	therefore	therefore	ADV
fbem-1048	16	2	,	,	PUNCT
fbem-1048	16	3	sentiment	sentiment	NOUN
fbem-1048	16	4	analysis	analysis	NOUN
fbem-1048	16	5	based	base	VERB
fbem-1048	16	6	on	on	ADP
fbem-1048	16	7	comments	comment	NOUN
fbem-1048	16	8	must	must	AUX
fbem-1048	16	9	be	be	AUX
fbem-1048	16	10	considered	consider	VERB
fbem-1048	16	11	in	in	ADP
fbem-1048	16	12	personalized	personalized	ADJ
fbem-1048	16	13	analysis	analysis	NOUN
fbem-1048	16	14	.	.	PUNCT
fbem-1048	17	1	in	in	ADP
fbem-1048	17	2	the	the	DET
fbem-1048	17	3	field	field	NOUN
fbem-1048	17	4	of	of	ADP
fbem-1048	17	5	the	the	DET
fbem-1048	17	6	computer	computer	NOUN
fbem-1048	17	7	science	science	NOUN
fbem-1048	17	8	and	and	CCONJ
fbem-1048	17	9	artificial	artificial	ADJ
fbem-1048	17	10	intelligence	intelligence	NOUN
fbem-1048	17	11	,	,	PUNCT
fbem-1048	17	12	the	the	DET
fbem-1048	17	13	natural	natural	ADJ
fbem-1048	17	14	language	language	NOUN
fbem-1048	17	15	processing	processing	NOUN
fbem-1048	17	16	is	be	AUX
fbem-1048	17	17	a	a	DET
fbem-1048	17	18	very	very	ADV
fbem-1048	17	19	important	important	ADJ
fbem-1048	17	20	orientation	orientation	NOUN
fbem-1048	17	21	.	.	PUNCT
fbem-1048	18	1	it	it	PRON
fbem-1048	18	2	refers	refer	VERB
fbem-1048	18	3	to	to	ADP
fbem-1048	18	4	the	the	DET
fbem-1048	18	5	technology	technology	NOUN
fbem-1048	18	6	which	which	PRON
fbem-1048	18	7	human	human	NOUN
fbem-1048	18	8	could	could	AUX
fbem-1048	18	9	use	use	VERB
fbem-1048	18	10	the	the	DET
fbem-1048	18	11	natural	natural	ADJ
fbem-1048	18	12	language	language	NOUN
fbem-1048	18	13	to	to	PART
fbem-1048	18	14	communicate	communicate	VERB
fbem-1048	18	15	with	with	ADP
fbem-1048	18	16	the	the	DET
fbem-1048	18	17	machines	machine	NOUN
fbem-1048	18	18	.	.	PUNCT
fbem-1048	19	1	various	various	ADJ
fbem-1048	19	2	theories	theory	NOUN
fbem-1048	19	3	and	and	CCONJ
fbem-1048	19	4	methods	method	NOUN
fbem-1048	19	5	of	of	ADP
fbem-1048	19	6	communication	communication	NOUN
fbem-1048	19	7	.	.	PUNCT
fbem-1048	20	1	nasukawa	nasukawa	PROPN
fbem-1048	20	2	et	et	PROPN
fbem-1048	20	3	al	al	PROPN
fbem-1048	20	4	.	.	PUNCT
fbem-1048	21	1	[	[	X
fbem-1048	21	2	1	1	X
fbem-1048	21	3	]	]	PUNCT
fbem-1048	21	4	were	be	AUX
fbem-1048	21	5	the	the	DET
fbem-1048	21	6	first	first	ADJ
fbem-1048	21	7	group	group	NOUN
fbem-1048	21	8	of	of	ADP
fbem-1048	21	9	people	people	NOUN
fbem-1048	21	10	who	who	PRON
fbem-1048	21	11	originally	originally	ADV
fbem-1048	21	12	proposed	propose	VERB
fbem-1048	21	13	the	the	DET
fbem-1048	21	14	natural	natural	ADJ
fbem-1048	21	15	language	language	NOUN
fbem-1048	21	16	processing	processing	NOUN
fbem-1048	21	17	technology	technology	NOUN
fbem-1048	21	18	of	of	ADP
fbem-1048	21	19	the	the	DET
fbem-1048	21	20	sentiment	sentiment	NOUN
fbem-1048	21	21	analysis	analysis	NOUN
fbem-1048	21	22	,	,	PUNCT
fbem-1048	21	23	which	which	PRON
fbem-1048	21	24	refers	refer	VERB
fbem-1048	21	25	to	to	ADP
fbem-1048	21	26	extracting	extract	VERB
fbem-1048	21	27	people	people	NOUN
fbem-1048	21	28	's	's	PART
fbem-1048	21	29	emotions	emotion	NOUN
fbem-1048	21	30	or	or	CCONJ
fbem-1048	21	31	opinions	opinion	NOUN
fbem-1048	21	32	from	from	ADP
fbem-1048	21	33	written	write	VERB
fbem-1048	21	34	words	word	NOUN
fbem-1048	21	35	.	.	PUNCT
fbem-1048	22	1	sentiment	sentiment	NOUN
fbem-1048	22	2	analysis	analysis	NOUN
fbem-1048	22	3	is	be	AUX
fbem-1048	22	4	currently	currently	ADV
fbem-1048	22	5	a	a	DET
fbem-1048	22	6	popular	popular	ADJ
fbem-1048	22	7	research	research	NOUN
fbem-1048	22	8	direction	direction	NOUN
fbem-1048	22	9	at	at	ADP
fbem-1048	22	10	home	home	NOUN
fbem-1048	22	11	and	and	CCONJ
fbem-1048	22	12	abroad	abroad	ADV
fbem-1048	22	13	.	.	PUNCT
fbem-1048	23	1	compared	compare	VERB
fbem-1048	23	2	with	with	ADP
fbem-1048	23	3	sentiment	sentiment	NOUN
fbem-1048	23	4	analysis	analysis	NOUN
fbem-1048	23	5	in	in	ADP
fbem-1048	23	6	chinese	chinese	ADJ
fbem-1048	23	7	,	,	PUNCT
fbem-1048	23	8	sentiment	sentiment	NOUN
fbem-1048	23	9	analysis	analysis	NOUN
fbem-1048	23	10	in	in	ADP
fbem-1048	23	11	english	english	PROPN
fbem-1048	23	12	will	will	AUX
fbem-1048	23	13	be	be	AUX
fbem-1048	23	14	more	more	ADV
fbem-1048	23	15	accurate	accurate	ADJ
fbem-1048	23	16	and	and	CCONJ
fbem-1048	23	17	simple	simple	ADJ
fbem-1048	23	18	,	,	PUNCT
fbem-1048	23	19	because	because	SCONJ
fbem-1048	23	20	english	english	PROPN
fbem-1048	23	21	has	have	VERB
fbem-1048	23	22	no	no	DET
fbem-1048	23	23	major	major	ADJ
fbem-1048	23	24	semantic	semantic	ADJ
fbem-1048	23	25	and	and	CCONJ
fbem-1048	23	26	grammatical	grammatical	ADJ
fbem-1048	23	27	changes	change	NOUN
fbem-1048	23	28	,	,	PUNCT
fbem-1048	23	29	but	but	CCONJ
fbem-1048	23	30	chinese	chinese	ADJ
fbem-1048	23	31	sentiment	sentiment	NOUN
fbem-1048	23	32	analysis	analysis	NOUN
fbem-1048	23	33	will	will	AUX
fbem-1048	23	34	be	be	AUX
fbem-1048	23	35	more	more	ADV
fbem-1048	23	36	accurate	accurate	ADJ
fbem-1048	23	37	and	and	CCONJ
fbem-1048	23	38	simple	simple	ADJ
fbem-1048	23	39	.	.	PUNCT
fbem-1048	24	1	since	since	SCONJ
fbem-1048	24	2	it	it	PRON
fbem-1048	24	3	is	be	AUX
fbem-1048	24	4	everchanging	everchange	VERB
fbem-1048	24	5	,	,	PUNCT
fbem-1048	24	6	it	it	PRON
fbem-1048	24	7	is	be	AUX
fbem-1048	24	8	relatively	relatively	ADV
fbem-1048	24	9	difficult	difficult	ADJ
fbem-1048	24	10	to	to	PART
fbem-1048	24	11	process	process	VERB
fbem-1048	24	12	,	,	PUNCT
fbem-1048	24	13	and	and	CCONJ
fbem-1048	24	14	the	the	DET
fbem-1048	24	15	accuracy	accuracy	NOUN
fbem-1048	24	16	of	of	ADP
fbem-1048	24	17	the	the	DET
fbem-1048	24	18	processing	processing	NOUN
fbem-1048	24	19	results	result	NOUN
fbem-1048	24	20	will	will	AUX
fbem-1048	24	21	also	also	ADV
fbem-1048	24	22	decrease	decrease	VERB
fbem-1048	24	23	.	.	PUNCT
fbem-1048	25	1	sentiment	sentiment	NOUN
fbem-1048	25	2	analysis	analysis	NOUN
fbem-1048	25	3	and	and	CCONJ
fbem-1048	25	4	topic	topic	NOUN
fbem-1048	25	5	model	model	NOUN
fbem-1048	25	6	are	be	AUX
fbem-1048	25	7	two	two	NUM
fbem-1048	25	8	hot	hot	ADJ
fbem-1048	25	9	research	research	NOUN
fbem-1048	25	10	directions	direction	NOUN
fbem-1048	25	11	.	.	PUNCT
fbem-1048	26	1	according	accord	VERB
fbem-1048	26	2	to	to	ADP
fbem-1048	26	3	different	different	ADJ
fbem-1048	26	4	application	application	NOUN
fbem-1048	26	5	fields	field	NOUN
fbem-1048	26	6	,	,	PUNCT
fbem-1048	26	7	text	text	NOUN
fbem-1048	26	8	sentiment	sentiment	NOUN
fbem-1048	26	9	analysis	analysis	NOUN
fbem-1048	26	10	technology	technology	NOUN
fbem-1048	26	11	can	can	AUX
fbem-1048	26	12	be	be	AUX
fbem-1048	26	13	divided	divide	VERB
fbem-1048	26	14	:	:	PUNCT
fbem-1048	26	15	(	(	PUNCT
fbem-1048	26	16	1	1	X
fbem-1048	26	17	)	)	PUNCT
fbem-1048	26	18	text	text	NOUN
fbem-1048	26	19	sentiment	sentiment	NOUN
fbem-1048	26	20	analysis	analysis	NOUN
fbem-1048	26	21	of	of	ADP
fbem-1048	26	22	product	product	NOUN
fbem-1048	26	23	reviews	review	NOUN
fbem-1048	26	24	is	be	AUX
fbem-1048	26	25	generally	generally	ADV
fbem-1048	26	26	used	use	VERB
fbem-1048	26	27	to	to	PART
fbem-1048	26	28	assist	assist	VERB
fbem-1048	26	29	consumers	consumer	NOUN
fbem-1048	26	30	in	in	ADP
fbem-1048	26	31	decision	decision	NOUN
fbem-1048	26	32	-	-	PUNCT
fbem-1048	26	33	making	make	VERB
fbem-1048	26	34	and	and	CCONJ
fbem-1048	26	35	business	business	NOUN
fbem-1048	26	36	public	public	ADJ
fbem-1048	26	37	opinion	opinion	NOUN
fbem-1048	26	38	monitoring	monitor	VERB
fbem-1048	26	39	[	[	X
fbem-1048	26	40	2	2	NUM
fbem-1048	26	41	-	-	SYM
fbem-1048	26	42	3	3	NUM
fbem-1048	26	43	]	]	PUNCT
fbem-1048	26	44	;	;	PUNCT
fbem-1048	26	45	(	(	PUNCT
fbem-1048	26	46	2	2	X
fbem-1048	26	47	)	)	PUNCT
fbem-1048	26	48	text	text	NOUN
fbem-1048	26	49	sentiment	sentiment	NOUN
fbem-1048	26	50	analysis	analysis	NOUN
fbem-1048	26	51	for	for	ADP
fbem-1048	26	52	news	news	NOUN
fbem-1048	26	53	comments	comment	NOUN
fbem-1048	26	54	is	be	AUX
fbem-1048	26	55	used	use	VERB
fbem-1048	26	56	to	to	PART
fbem-1048	26	57	process	process	VERB
fbem-1048	26	58	text	text	NOUN
fbem-1048	26	59	comments	comment	NOUN
fbem-1048	26	60	published	publish	VERB
fbem-1048	26	61	by	by	ADP
fbem-1048	26	62	news	news	NOUN
fbem-1048	26	63	events	event	NOUN
fbem-1048	26	64	,	,	PUNCT
fbem-1048	26	65	and	and	CCONJ
fbem-1048	26	66	is	be	AUX
fbem-1048	26	67	generally	generally	ADV
fbem-1048	26	68	used	use	VERB
fbem-1048	26	69	to	to	PART
fbem-1048	26	70	guide	guide	VERB
fbem-1048	26	71	public	public	ADJ
fbem-1048	26	72	opinion	opinion	NOUN
fbem-1048	26	73	based	base	VERB
fbem-1048	26	74	on	on	ADP
fbem-1048	26	75	the	the	DET
fbem-1048	26	76	correct	correct	ADJ
fbem-1048	26	77	public	public	ADJ
fbem-1048	26	78	opinion	opinion	NOUN
fbem-1048	27	1	[	[	X
fbem-1048	27	2	4	4	NUM
fbem-1048	27	3	-	-	SYM
fbem-1048	27	4	5	5	NUM
fbem-1048	27	5	]	]	PUNCT
fbem-1048	27	6	.	.	PUNCT
fbem-1048	28	1	two	two	NUM
fbem-1048	28	2	main	main	ADJ
fbem-1048	28	3	research	research	NOUN
fbem-1048	28	4	methods	method	NOUN
fbem-1048	28	5	in	in	ADP
fbem-1048	28	6	the	the	DET
fbem-1048	28	7	field	field	NOUN
fbem-1048	28	8	of	of	ADP
fbem-1048	28	9	the	the	DET
fbem-1048	28	10	sentiment	sentiment	NOUN
fbem-1048	28	11	analysis	analysis	NOUN
fbem-1048	28	12	include	include	VERB
fbem-1048	28	13	the	the	DET
fbem-1048	28	14	unsupervised	unsupervised	ADJ
fbem-1048	28	15	one	one	NOUN
fbem-1048	28	16	and	and	CCONJ
fbem-1048	28	17	the	the	DET
fbem-1048	28	18	supervised	supervised	ADJ
fbem-1048	28	19	one	one	NUM
fbem-1048	28	20	.	.	PUNCT
fbem-1048	29	1	in	in	ADP
fbem-1048	29	2	the	the	DET
fbem-1048	29	3	early	early	ADJ
fbem-1048	29	4	period	period	NOUN
fbem-1048	29	5	,	,	PUNCT
fbem-1048	29	6	the	the	DET
fbem-1048	29	7	supervised	supervised	ADJ
fbem-1048	29	8	learning	learning	NOUN
fbem-1048	29	9	is	be	AUX
fbem-1048	29	10	about	about	ADP
fbem-1048	29	11	shallow	shallow	ADJ
fbem-1048	29	12	models	model	NOUN
fbem-1048	29	13	such	such	ADJ
fbem-1048	29	14	as	as	ADP
fbem-1048	29	15	support	support	NOUN
fbem-1048	29	16	vector	vector	NOUN
fbem-1048	29	17	machine	machine	NOUN
fbem-1048	29	18	(	(	PUNCT
fbem-1048	29	19	svm	svm	PROPN
fbem-1048	29	20	)	)	PUNCT
fbem-1048	29	21	,	,	PUNCT
fbem-1048	29	22	maximum	maximum	PROPN
fbem-1048	29	23	entropy	entropy	PROPN
fbem-1048	29	24	,	,	PUNCT
fbem-1048	29	25	naïve	naïve	ADJ
fbem-1048	29	26	bayes	bayes	NOUN
fbem-1048	29	27	,	,	PUNCT
fbem-1048	29	28	etc	etc	X
fbem-1048	29	29	.	.	X
fbem-1048	30	1	the	the	DET
fbem-1048	30	2	unsupervised	unsupervised	ADJ
fbem-1048	30	3	learning	learning	NOUN
fbem-1048	30	4	is	be	AUX
fbem-1048	30	5	based	base	VERB
fbem-1048	30	6	on	on	ADP
fbem-1048	30	7	semantic	semantic	ADJ
fbem-1048	30	8	analysis	analysis	NOUN
fbem-1048	30	9	,	,	PUNCT
fbem-1048	30	10	dictionary	dictionary	NOUN
fbem-1048	30	11	and	and	CCONJ
fbem-1048	30	12	the	the	DET
fbem-1048	30	13	other	other	ADJ
fbem-1048	30	14	methods	method	NOUN
fbem-1048	30	15	.	.	PUNCT
fbem-1048	31	1	because	because	SCONJ
fbem-1048	31	2	of	of	ADP
fbem-1048	31	3	the	the	DET
fbem-1048	31	4	appearance	appearance	NOUN
fbem-1048	31	5	of	of	ADP
fbem-1048	31	6	the	the	DET
fbem-1048	31	7	deep	deep	ADJ
fbem-1048	31	8	learning	learning	NOUN
fbem-1048	31	9	,	,	PUNCT
fbem-1048	31	10	regression	regression	NOUN
fbem-1048	31	11	tasks	task	NOUN
fbem-1048	31	12	and	and	CCONJ
fbem-1048	31	13	many	many	ADJ
fbem-1048	31	14	classifications	classification	NOUN
fbem-1048	31	15	have	have	AUX
fbem-1048	31	16	achieved	achieve	VERB
fbem-1048	31	17	best	good	ADJ
fbem-1048	31	18	results	result	NOUN
fbem-1048	31	19	.	.	PUNCT
fbem-1048	32	1	in	in	ADP
fbem-1048	32	2	recent	recent	ADJ
fbem-1048	32	3	years	year	NOUN
fbem-1048	32	4	,	,	PUNCT
fbem-1048	32	5	it	it	PRON
fbem-1048	32	6	has	have	AUX
fbem-1048	32	7	become	become	VERB
fbem-1048	32	8	a	a	DET
fbem-1048	32	9	hotspot	hotspot	NOUN
fbem-1048	32	10	of	of	ADP
fbem-1048	32	11	the	the	DET
fbem-1048	32	12	research	research	NOUN
fbem-1048	32	13	in	in	ADP
fbem-1048	32	14	the	the	DET
fbem-1048	32	15	field	field	NOUN
fbem-1048	32	16	of	of	ADP
fbem-1048	32	17	the	the	DET
fbem-1048	32	18	sentiment	sentiment	NOUN
fbem-1048	32	19	analysis	analysis	NOUN
fbem-1048	32	20	which	which	PRON
fbem-1048	32	21	is	be	AUX
fbem-1048	32	22	applicated	applicate	VERB
fbem-1048	32	23	with	with	ADP
fbem-1048	32	24	the	the	DET
fbem-1048	32	25	deep	deep	ADJ
fbem-1048	32	26	learning	learning	NOUN
fbem-1048	32	27	.	.	PUNCT
fbem-1048	33	1	in	in	ADP
fbem-1048	33	2	terms	term	NOUN
fbem-1048	33	3	of	of	ADP
fbem-1048	33	4	model	model	NOUN
fbem-1048	33	5	establishment	establishment	NOUN
fbem-1048	33	6	,	,	PUNCT
fbem-1048	33	7	the	the	DET
fbem-1048	33	8	lda	lda	PROPN
fbem-1048	33	9	model	model	NOUN
fbem-1048	33	10	is	be	AUX
fbem-1048	33	11	the	the	DET
fbem-1048	33	12	mainstream	mainstream	NOUN
fbem-1048	33	13	model	model	NOUN
fbem-1048	33	14	in	in	ADP
fbem-1048	33	15	the	the	DET
fbem-1048	33	16	current	current	ADJ
fbem-1048	33	17	topic	topic	NOUN
fbem-1048	33	18	model	model	NOUN
fbem-1048	33	19	,	,	PUNCT
fbem-1048	33	20	it	it	PRON
fbem-1048	33	21	’s	’	VERB
fbem-1048	33	22	about	about	ADP
fbem-1048	33	23	the	the	DET
fbem-1048	33	24	psla	psla	NOUN
fbem-1048	33	25	which	which	PRON
fbem-1048	33	26	is	be	AUX
fbem-1048	33	27	added	add	VERB
fbem-1048	33	28	with	with	ADP
fbem-1048	33	29	the	the	DET
fbem-1048	33	30	bayesian	bayesian	NOUN
fbem-1048	33	31	framework	framework	NOUN
fbem-1048	33	32	,	,	PUNCT
fbem-1048	33	33	meanwhile	meanwhile	ADV
fbem-1048	33	34	it	it	PRON
fbem-1048	33	35	’s	’	VERB
fbem-1048	33	36	also	also	ADV
fbem-1048	33	37	a	a	DET
fbem-1048	33	38	typical	typical	ADJ
fbem-1048	33	39	bag	bag	NOUN
fbem-1048	33	40	model	model	NOUN
fbem-1048	33	41	.	.	PUNCT
fbem-1048	34	1	it	it	PRON
fbem-1048	34	2	can	can	AUX
fbem-1048	34	3	give	give	VERB
fbem-1048	34	4	the	the	DET
fbem-1048	34	5	subject	subject	NOUN
fbem-1048	34	6	of	of	ADP
fbem-1048	34	7	each	each	DET
fbem-1048	34	8	document	document	NOUN
fbem-1048	34	9	in	in	ADP
fbem-1048	34	10	the	the	DET
fbem-1048	34	11	document	document	NOUN
fbem-1048	34	12	set	set	VERB
fbem-1048	34	13	in	in	ADP
fbem-1048	34	14	the	the	DET
fbem-1048	34	15	form	form	NOUN
fbem-1048	34	16	of	of	ADP
fbem-1048	34	17	a	a	DET
fbem-1048	34	18	probability	probability	NOUN
fbem-1048	34	19	distribution	distribution	NOUN
fbem-1048	34	20	,	,	PUNCT
fbem-1048	34	21	so	so	SCONJ
fbem-1048	34	22	as	as	SCONJ
fbem-1048	34	23	to	to	PART
fbem-1048	34	24	extract	extract	VERB
fbem-1048	34	25	their	their	PRON
fbem-1048	34	26	subject	subject	NOUN
fbem-1048	34	27	by	by	ADP
fbem-1048	34	28	analyzing	analyze	VERB
fbem-1048	34	29	the	the	DET
fbem-1048	34	30	documents	document	NOUN
fbem-1048	34	31	,	,	PUNCT
fbem-1048	34	32	and	and	CCONJ
fbem-1048	34	33	perform	perform	VERB
fbem-1048	34	34	clustering	cluster	VERB
fbem-1048	34	35	or	or	CCONJ
fbem-1048	34	36	text	text	NOUN
fbem-1048	34	37	classification	classification	NOUN
fbem-1048	34	38	on	on	ADP
fbem-1048	34	39	the	the	DET
fbem-1048	34	40	subject	subject	NOUN
fbem-1048	34	41	.	.	PUNCT
fbem-1048	35	1	2	2	X
fbem-1048	35	2	.	.	X
fbem-1048	35	3	technical	technical	ADJ
fbem-1048	35	4	implementation	implementation	NOUN
fbem-1048	35	5	method	method	NOUN
fbem-1048	35	6	2.1	2.1	NUM
fbem-1048	35	7	.	.	PUNCT
fbem-1048	36	1	technical	technical	ADJ
fbem-1048	36	2	assumptions	assumption	NOUN
fbem-1048	36	3	according	accord	VERB
fbem-1048	36	4	to	to	ADP
fbem-1048	36	5	the	the	DET
fbem-1048	36	6	articles	article	NOUN
fbem-1048	36	7	or	or	CCONJ
fbem-1048	36	8	some	some	DET
fbem-1048	36	9	paragraphs	paragraph	NOUN
fbem-1048	36	10	published	publish	VERB
fbem-1048	36	11	by	by	ADP
fbem-1048	36	12	26	26	NUM
fbem-1048	36	13	the	the	DET
fbem-1048	36	14	user	user	NOUN
fbem-1048	36	15	,	,	PUNCT
fbem-1048	36	16	the	the	DET
fbem-1048	36	17	sentiment	sentiment	NOUN
fbem-1048	36	18	analysis	analysis	NOUN
fbem-1048	36	19	is	be	AUX
fbem-1048	36	20	carried	carry	VERB
fbem-1048	36	21	out	out	ADP
fbem-1048	36	22	to	to	PART
fbem-1048	36	23	obtain	obtain	VERB
fbem-1048	36	24	the	the	DET
fbem-1048	36	25	content	content	NOUN
fbem-1048	36	26	that	that	PRON
fbem-1048	36	27	the	the	DET
fbem-1048	36	28	user	user	NOUN
fbem-1048	36	29	likes	like	VERB
fbem-1048	36	30	and	and	CCONJ
fbem-1048	36	31	dislikes	dislike	NOUN
fbem-1048	36	32	,	,	PUNCT
fbem-1048	36	33	and	and	CCONJ
fbem-1048	36	34	the	the	DET
fbem-1048	36	35	obtained	obtain	VERB
fbem-1048	36	36	results	result	NOUN
fbem-1048	36	37	are	be	AUX
fbem-1048	36	38	used	use	VERB
fbem-1048	36	39	to	to	PART
fbem-1048	36	40	make	make	VERB
fbem-1048	36	41	personalized	personalized	ADJ
fbem-1048	36	42	recommendations	recommendation	NOUN
fbem-1048	36	43	,	,	PUNCT
fbem-1048	36	44	so	so	SCONJ
fbem-1048	36	45	as	as	SCONJ
fbem-1048	36	46	to	to	PART
fbem-1048	36	47	obtain	obtain	VERB
fbem-1048	36	48	the	the	DET
fbem-1048	36	49	user	user	NOUN
fbem-1048	36	50	's	's	PART
fbem-1048	36	51	preferences	preference	NOUN
fbem-1048	36	52	more	more	ADV
fbem-1048	36	53	accurately	accurately	ADV
fbem-1048	36	54	.	.	PUNCT
fbem-1048	37	1	to	to	PART
fbem-1048	37	2	solve	solve	VERB
fbem-1048	37	3	the	the	DET
fbem-1048	37	4	key	key	ADJ
fbem-1048	37	5	problem	problem	NOUN
fbem-1048	37	6	,	,	PUNCT
fbem-1048	37	7	we	we	PRON
fbem-1048	37	8	need	need	VERB
fbem-1048	37	9	to	to	PART
fbem-1048	37	10	collect	collect	VERB
fbem-1048	37	11	the	the	DET
fbem-1048	37	12	data	datum	NOUN
fbem-1048	37	13	of	of	ADP
fbem-1048	37	14	these	these	DET
fbem-1048	37	15	segments	segment	NOUN
fbem-1048	37	16	,	,	PUNCT
fbem-1048	37	17	and	and	CCONJ
fbem-1048	37	18	use	use	VERB
fbem-1048	37	19	the	the	DET
fbem-1048	37	20	relevant	relevant	ADJ
fbem-1048	37	21	libraries	library	NOUN
fbem-1048	37	22	in	in	ADP
fbem-1048	37	23	python	python	NOUN
fbem-1048	37	24	to	to	PART
fbem-1048	37	25	complete	complete	VERB
fbem-1048	37	26	word	word	NOUN
fbem-1048	37	27	segmentation	segmentation	NOUN
fbem-1048	37	28	,	,	PUNCT
fbem-1048	37	29	model	model	NOUN
fbem-1048	37	30	establishment	establishment	NOUN
fbem-1048	37	31	,	,	PUNCT
fbem-1048	37	32	etc	etc	X
fbem-1048	37	33	.	.	X
fbem-1048	37	34	to	to	PART
fbem-1048	37	35	obtain	obtain	VERB
fbem-1048	37	36	the	the	DET
fbem-1048	37	37	results	result	NOUN
fbem-1048	37	38	.	.	PUNCT
fbem-1048	38	1	2.2	2.2	NUM
fbem-1048	38	2	.	.	PUNCT
fbem-1048	39	1	word	word	NOUN
fbem-1048	39	2	vector	vector	NOUN
fbem-1048	39	3	computers	computer	NOUN
fbem-1048	39	4	can	can	AUX
fbem-1048	39	5	not	not	PART
fbem-1048	39	6	directly	directly	ADV
fbem-1048	39	7	read	read	VERB
fbem-1048	39	8	human	human	ADJ
fbem-1048	39	9	vocabulary	vocabulary	NOUN
fbem-1048	39	10	,	,	PUNCT
fbem-1048	39	11	so	so	SCONJ
fbem-1048	39	12	they	they	PRON
fbem-1048	39	13	can	can	AUX
fbem-1048	39	14	not	not	PART
fbem-1048	39	15	directly	directly	ADV
fbem-1048	39	16	process	process	VERB
fbem-1048	39	17	text	text	NOUN
fbem-1048	39	18	.	.	PUNCT
fbem-1048	40	1	it	it	PRON
fbem-1048	40	2	needs	need	VERB
fbem-1048	40	3	to	to	PART
fbem-1048	40	4	be	be	AUX
fbem-1048	40	5	converted	convert	VERB
fbem-1048	40	6	into	into	ADP
fbem-1048	40	7	data	datum	NOUN
fbem-1048	40	8	that	that	PRON
fbem-1048	40	9	can	can	AUX
fbem-1048	40	10	be	be	AUX
fbem-1048	40	11	recognized	recognize	VERB
fbem-1048	40	12	by	by	ADP
fbem-1048	40	13	computers	computer	NOUN
fbem-1048	40	14	.	.	PUNCT
fbem-1048	41	1	therefore	therefore	ADV
fbem-1048	41	2	,	,	PUNCT
fbem-1048	41	3	first	first	ADV
fbem-1048	41	4	of	of	ADP
fbem-1048	41	5	all	all	PRON
fbem-1048	41	6	,	,	PUNCT
fbem-1048	41	7	the	the	DET
fbem-1048	41	8	acquired	acquire	VERB
fbem-1048	41	9	text	text	NOUN
fbem-1048	41	10	data	datum	NOUN
fbem-1048	41	11	must	must	AUX
fbem-1048	41	12	be	be	AUX
fbem-1048	41	13	converted	convert	VERB
fbem-1048	41	14	into	into	ADP
fbem-1048	41	15	word	word	NOUN
fbem-1048	41	16	vectors	vector	NOUN
fbem-1048	41	17	so	so	SCONJ
fbem-1048	41	18	that	that	SCONJ
fbem-1048	41	19	computers	computer	NOUN
fbem-1048	41	20	can	can	AUX
fbem-1048	41	21	recognize	recognize	VERB
fbem-1048	41	22	them	they	PRON
fbem-1048	41	23	.	.	PUNCT
fbem-1048	42	1	this	this	DET
fbem-1048	42	2	process	process	NOUN
fbem-1048	42	3	can	can	AUX
fbem-1048	42	4	also	also	ADV
fbem-1048	42	5	be	be	AUX
fbem-1048	42	6	called	call	VERB
fbem-1048	42	7	the	the	DET
fbem-1048	42	8	process	process	NOUN
fbem-1048	42	9	of	of	ADP
fbem-1048	42	10	word	word	NOUN
fbem-1048	42	11	embedding	embed	VERB
fbem-1048	42	12	,	,	PUNCT
fbem-1048	42	13	and	and	CCONJ
fbem-1048	42	14	there	there	PRON
fbem-1048	42	15	are	be	VERB
fbem-1048	42	16	mainly	mainly	ADV
fbem-1048	42	17	two	two	NUM
fbem-1048	42	18	ways	way	NOUN
fbem-1048	42	19	based	base	VERB
fbem-1048	42	20	on	on	ADP
fbem-1048	42	21	statistics	statistic	NOUN
fbem-1048	42	22	and	and	CCONJ
fbem-1048	42	23	based	base	VERB
fbem-1048	42	24	on	on	ADP
fbem-1048	42	25	language	language	NOUN
fbem-1048	42	26	models	model	NOUN
fbem-1048	42	27	.	.	PUNCT
fbem-1048	43	1	statistics	statistic	NOUN
fbem-1048	43	2	-	-	PUNCT
fbem-1048	43	3	based	base	VERB
fbem-1048	43	4	methods	method	NOUN
fbem-1048	43	5	include	include	VERB
fbem-1048	43	6	bag	bag	NOUN
fbem-1048	43	7	-	-	PUNCT
fbem-1048	43	8	of	of	ADP
fbem-1048	43	9	-	-	PUNCT
fbem-1048	43	10	words	word	NOUN
fbem-1048	43	11	model	model	NOUN
fbem-1048	43	12	,	,	PUNCT
fbem-1048	43	13	onehot	onehot	ADJ
fbem-1048	43	14	encoding	encoding	NOUN
fbem-1048	43	15	and	and	CCONJ
fbem-1048	43	16	term	term	NOUN
fbem-1048	43	17	frequency	frequency	NOUN
fbem-1048	43	18	-	-	PUNCT
fbem-1048	43	19	inverse	inverse	NOUN
fbem-1048	43	20	document	document	NOUN
fbem-1048	43	21	frequency	frequency	NOUN
fbem-1048	43	22	(	(	PUNCT
fbem-1048	43	23	tf	tf	PROPN
fbem-1048	43	24	-	-	PUNCT
fbem-1048	43	25	idf	idf	NOUN
fbem-1048	43	26	)	)	PUNCT
fbem-1048	44	1	[	[	X
fbem-1048	44	2	6	6	NUM
fbem-1048	44	3	]	]	PUNCT
fbem-1048	44	4	,	,	PUNCT
fbem-1048	44	5	etc	etc	X
fbem-1048	44	6	.	.	X
fbem-1048	44	7	;	;	PUNCT
fbem-1048	44	8	language	language	NOUN
fbem-1048	44	9	model	model	NOUN
fbem-1048	44	10	-	-	PUNCT
fbem-1048	44	11	based	base	VERB
fbem-1048	44	12	methods	method	NOUN
fbem-1048	44	13	include	include	VERB
fbem-1048	44	14	well	well	ADV
fbem-1048	44	15	-	-	PUNCT
fbem-1048	44	16	known	know	VERB
fbem-1048	44	17	word2vec	word2vec	X
fbem-1048	44	18	and	and	CCONJ
fbem-1048	44	19	bert	bert	PROPN
fbem-1048	44	20	.	.	PUNCT
fbem-1048	44	21	word2vec[7	word2vec[7	PROPN
fbem-1048	44	22	-	-	SYM
fbem-1048	44	23	8	8	NUM
fbem-1048	44	24	]	]	PUNCT
fbem-1048	44	25	is	be	AUX
fbem-1048	44	26	an	an	DET
fbem-1048	44	27	open	open	ADJ
fbem-1048	44	28	source	source	NOUN
fbem-1048	44	29	word	word	NOUN
fbem-1048	44	30	embedding	embed	VERB
fbem-1048	44	31	tool	tool	NOUN
fbem-1048	44	32	based	base	VERB
fbem-1048	44	33	on	on	ADP
fbem-1048	44	34	deep	deep	ADJ
fbem-1048	44	35	learning	learning	NOUN
fbem-1048	44	36	proposed	propose	VERB
fbem-1048	44	37	by	by	ADP
fbem-1048	44	38	google	google	PROPN
fbem-1048	44	39	,	,	PUNCT
fbem-1048	44	40	which	which	PRON
fbem-1048	44	41	mainly	mainly	ADV
fbem-1048	44	42	includes	include	VERB
fbem-1048	44	43	two	two	NUM
fbem-1048	44	44	models	model	NOUN
fbem-1048	44	45	,	,	PUNCT
fbem-1048	44	46	namely	namely	ADV
fbem-1048	44	47	skip	skip	VERB
fbem-1048	44	48	-	-	PUNCT
fbem-1048	44	49	gram	gram	NOUN
fbem-1048	44	50	and	and	CCONJ
fbem-1048	44	51	continuous	continuous	ADJ
fbem-1048	44	52	bag	bag	NOUN
fbem-1048	44	53	of	of	ADP
fbem-1048	44	54	words	word	NOUN
fbem-1048	44	55	(	(	PUNCT
fbem-1048	44	56	cbow	cbow	PROPN
fbem-1048	44	57	)	)	PUNCT
fbem-1048	44	58	.	.	PUNCT
fbem-1048	45	1	tf	tf	PROPN
fbem-1048	45	2	-	-	PUNCT
fbem-1048	45	3	idf	idf	PROPN
fbem-1048	45	4	,	,	PUNCT
fbem-1048	45	5	word2vec	word2vec	X
fbem-1048	45	6	and	and	CCONJ
fbem-1048	45	7	a	a	DET
fbem-1048	45	8	weighted	weight	VERB
fbem-1048	45	9	word2vec	word2vec	AUX
fbem-1048	45	10	make	make	VERB
fbem-1048	45	11	three	three	NUM
fbem-1048	45	12	methods	method	NOUN
fbem-1048	45	13	of	of	ADP
fbem-1048	45	14	extracting	extract	VERB
fbem-1048	45	15	word	word	NOUN
fbem-1048	45	16	vectors	vector	NOUN
fbem-1048	45	17	[	[	X
fbem-1048	45	18	9	9	NUM
fbem-1048	45	19	]	]	SYM
fbem-1048	45	20	2.3	2.3	NUM
fbem-1048	45	21	.	.	PUNCT
fbem-1048	46	1	stop	stop	VERB
fbem-1048	46	2	words	word	NOUN
fbem-1048	46	3	the	the	DET
fbem-1048	46	4	concept	concept	NOUN
fbem-1048	46	5	of	of	ADP
fbem-1048	46	6	the	the	DET
fbem-1048	46	7	stop	stop	NOUN
fbem-1048	46	8	words	word	NOUN
fbem-1048	46	9	is	be	AUX
fbem-1048	46	10	that	that	SCONJ
fbem-1048	46	11	in	in	ADP
fbem-1048	46	12	order	order	NOUN
fbem-1048	46	13	to	to	PART
fbem-1048	46	14	save	save	VERB
fbem-1048	46	15	the	the	DET
fbem-1048	46	16	storage	storage	NOUN
fbem-1048	46	17	space	space	NOUN
fbem-1048	46	18	and	and	CCONJ
fbem-1048	46	19	promote	promote	VERB
fbem-1048	46	20	the	the	DET
fbem-1048	46	21	efficiency	efficiency	NOUN
fbem-1048	46	22	of	of	ADP
fbem-1048	46	23	searching	search	VERB
fbem-1048	46	24	the	the	DET
fbem-1048	46	25	information	information	NOUN
fbem-1048	46	26	in	in	ADP
fbem-1048	46	27	the	the	DET
fbem-1048	46	28	information	information	NOUN
fbem-1048	46	29	retrieval	retrieval	NOUN
fbem-1048	46	30	,	,	PUNCT
fbem-1048	46	31	there	there	PRON
fbem-1048	46	32	will	will	AUX
fbem-1048	46	33	automatically	automatically	ADV
fbem-1048	46	34	filtered	filter	VERB
fbem-1048	46	35	out	out	ADP
fbem-1048	46	36	some	some	DET
fbem-1048	46	37	word	word	NOUN
fbem-1048	46	38	or	or	CCONJ
fbem-1048	46	39	words	word	NOUN
fbem-1048	46	40	before	before	ADP
fbem-1048	46	41	or	or	CCONJ
fbem-1048	46	42	after	after	ADP
fbem-1048	46	43	the	the	DET
fbem-1048	46	44	processing	processing	NOUN
fbem-1048	46	45	of	of	ADP
fbem-1048	46	46	the	the	DET
fbem-1048	46	47	natural	natural	ADJ
fbem-1048	46	48	language	language	NOUN
fbem-1048	46	49	data	datum	NOUN
fbem-1048	46	50	(	(	PUNCT
fbem-1048	46	51	or	or	CCONJ
fbem-1048	46	52	text	text	NOUN
fbem-1048	46	53	)	)	PUNCT
fbem-1048	46	54	.	.	PUNCT
fbem-1048	47	1	these	these	DET
fbem-1048	47	2	word	word	NOUN
fbem-1048	47	3	or	or	CCONJ
fbem-1048	47	4	words	word	NOUN
fbem-1048	47	5	are	be	AUX
fbem-1048	47	6	named	name	VERB
fbem-1048	47	7	stop	stop	VERB
fbem-1048	47	8	words	word	NOUN
fbem-1048	47	9	.	.	PUNCT
fbem-1048	48	1	those	those	PRON
fbem-1048	48	2	stop	stop	VERB
fbem-1048	48	3	words	word	NOUN
fbem-1048	48	4	are	be	AUX
fbem-1048	48	5	input	input	VERB
fbem-1048	48	6	by	by	ADP
fbem-1048	48	7	human	human	NOUN
fbem-1048	48	8	and	and	CCONJ
fbem-1048	48	9	generated	generate	VERB
fbem-1048	48	10	without	without	ADP
fbem-1048	48	11	automation	automation	NOUN
fbem-1048	48	12	.	.	PUNCT
fbem-1048	49	1	then	then	ADV
fbem-1048	49	2	,	,	PUNCT
fbem-1048	49	3	the	the	DET
fbem-1048	49	4	generated	generate	VERB
fbem-1048	49	5	stop	stop	NOUN
fbem-1048	49	6	words	word	NOUN
fbem-1048	49	7	would	would	AUX
fbem-1048	49	8	become	become	VERB
fbem-1048	49	9	a	a	DET
fbem-1048	49	10	stop	stop	NOUN
fbem-1048	49	11	word	word	NOUN
fbem-1048	49	12	list	list	NOUN
fbem-1048	49	13	.	.	PUNCT
fbem-1048	50	1	nevertheless	nevertheless	ADV
fbem-1048	50	2	,	,	PUNCT
fbem-1048	50	3	it	it	PRON
fbem-1048	50	4	’s	’	VERB
fbem-1048	50	5	no	no	DET
fbem-1048	50	6	certain	certain	ADJ
fbem-1048	50	7	list	list	NOUN
fbem-1048	50	8	of	of	ADP
fbem-1048	50	9	the	the	DET
fbem-1048	50	10	stop	stop	NOUN
fbem-1048	50	11	words	word	NOUN
fbem-1048	50	12	which	which	PRON
fbem-1048	50	13	could	could	AUX
fbem-1048	50	14	work	work	VERB
fbem-1048	50	15	for	for	ADP
fbem-1048	50	16	all	all	DET
fbem-1048	50	17	tools	tool	NOUN
fbem-1048	50	18	.	.	PUNCT
fbem-1048	51	1	stop	stop	VERB
fbem-1048	51	2	words	word	NOUN
fbem-1048	51	3	can	can	AUX
fbem-1048	51	4	be	be	AUX
fbem-1048	51	5	roughly	roughly	ADV
fbem-1048	51	6	divided	divide	VERB
fbem-1048	51	7	into	into	ADP
fbem-1048	51	8	two	two	NUM
fbem-1048	51	9	categories	category	NOUN
fbem-1048	51	10	:	:	PUNCT
fbem-1048	51	11	(	(	PUNCT
fbem-1048	51	12	1	1	X
fbem-1048	51	13	)	)	PUNCT
fbem-1048	51	14	it	it	PRON
fbem-1048	51	15	is	be	AUX
fbem-1048	51	16	widely	widely	ADV
fbem-1048	51	17	used	use	VERB
fbem-1048	51	18	and	and	CCONJ
fbem-1048	51	19	can	can	AUX
fbem-1048	51	20	be	be	AUX
fbem-1048	51	21	seen	see	VERB
fbem-1048	51	22	everywhere	everywhere	ADV
fbem-1048	51	23	on	on	ADP
fbem-1048	51	24	the	the	DET
fbem-1048	51	25	internet	internet	NOUN
fbem-1048	51	26	.	.	PUNCT
fbem-1048	52	1	for	for	ADP
fbem-1048	52	2	example	example	NOUN
fbem-1048	52	3	,	,	PUNCT
fbem-1048	52	4	the	the	DET
fbem-1048	52	5	word	word	NOUN
fbem-1048	52	6	"	"	PUNCT
fbem-1048	52	7	web	web	NOUN
fbem-1048	52	8	"	"	PUNCT
fbem-1048	52	9	appears	appear	VERB
fbem-1048	52	10	on	on	ADP
fbem-1048	52	11	almost	almost	ADV
fbem-1048	52	12	every	every	PRON
fbem-1048	52	13	website	website	NOUN
fbem-1048	52	14	.	.	PUNCT
fbem-1048	53	1	search	search	NOUN
fbem-1048	53	2	engines	engine	NOUN
fbem-1048	53	3	could	could	AUX
fbem-1048	53	4	n’t	not	PART
fbem-1048	53	5	make	make	VERB
fbem-1048	53	6	sure	sure	ADJ
fbem-1048	53	7	that	that	SCONJ
fbem-1048	53	8	those	those	DET
fbem-1048	53	9	words	word	NOUN
fbem-1048	53	10	could	could	AUX
fbem-1048	53	11	work	work	VERB
fbem-1048	53	12	out	out	ADP
fbem-1048	53	13	the	the	DET
fbem-1048	53	14	true	true	ADJ
fbem-1048	53	15	search	search	NOUN
fbem-1048	53	16	results	result	NOUN
fbem-1048	53	17	.	.	PUNCT
fbem-1048	54	1	also	also	ADV
fbem-1048	54	2	,	,	PUNCT
fbem-1048	54	3	it	it	PRON
fbem-1048	54	4	’s	’	VERB
fbem-1048	54	5	hard	hard	ADJ
fbem-1048	54	6	to	to	PART
fbem-1048	54	7	help	help	VERB
fbem-1048	54	8	narrow	narrow	VERB
fbem-1048	54	9	the	the	DET
fbem-1048	54	10	search	search	NOUN
fbem-1048	54	11	,	,	PUNCT
fbem-1048	54	12	meanwhile	meanwhile	ADV
fbem-1048	54	13	it	it	PRON
fbem-1048	54	14	also	also	ADV
fbem-1048	54	15	will	will	AUX
fbem-1048	54	16	reduce	reduce	VERB
fbem-1048	54	17	the	the	DET
fbem-1048	54	18	search	search	NOUN
fbem-1048	54	19	’s	’s	PART
fbem-1048	54	20	efficiency	efficiency	NOUN
fbem-1048	54	21	;	;	PUNCT
fbem-1048	54	22	(	(	PUNCT
fbem-1048	54	23	2	2	X
fbem-1048	54	24	)	)	PUNCT
fbem-1048	54	25	those	those	DET
fbem-1048	54	26	modal	modal	ADJ
fbem-1048	54	27	particles	particle	NOUN
fbem-1048	54	28	,	,	PUNCT
fbem-1048	54	29	adverbs	adverb	NOUN
fbem-1048	54	30	,	,	PUNCT
fbem-1048	54	31	prepositions	preposition	NOUN
fbem-1048	54	32	,	,	PUNCT
fbem-1048	54	33	conjunctions	conjunction	NOUN
fbem-1048	54	34	and	and	CCONJ
fbem-1048	54	35	so	so	ADV
fbem-1048	54	36	on	on	ADV
fbem-1048	54	37	usually	usually	ADV
fbem-1048	55	1	do	do	AUX
fbem-1048	55	2	n’t	not	PART
fbem-1048	55	3	have	have	VERB
fbem-1048	55	4	clear	clear	ADJ
fbem-1048	55	5	meanings	meaning	NOUN
fbem-1048	55	6	by	by	ADP
fbem-1048	55	7	their	their	PRON
fbem-1048	55	8	own	own	ADJ
fbem-1048	55	9	.	.	PUNCT
fbem-1048	56	1	they	they	PRON
fbem-1048	56	2	just	just	ADV
fbem-1048	56	3	have	have	VERB
fbem-1048	56	4	certain	certain	ADJ
fbem-1048	56	5	useful	useful	ADJ
fbem-1048	56	6	effects	effect	NOUN
fbem-1048	56	7	when	when	SCONJ
fbem-1048	56	8	they	they	PRON
fbem-1048	56	9	are	be	AUX
fbem-1048	56	10	added	add	VERB
fbem-1048	56	11	to	to	ADP
fbem-1048	56	12	complete	complete	ADJ
fbem-1048	56	13	sentences	sentence	NOUN
fbem-1048	56	14	.	.	PUNCT
fbem-1048	57	1	2.4	2.4	NUM
fbem-1048	57	2	.	.	PUNCT
fbem-1048	58	1	lda	lda	PROPN
fbem-1048	58	2	model	model	VERB
fbem-1048	58	3	the	the	DET
fbem-1048	58	4	lda	lda	PROPN
fbem-1048	58	5	model	model	NOUN
fbem-1048	58	6	is	be	AUX
fbem-1048	58	7	a	a	DET
fbem-1048	58	8	potential	potential	ADJ
fbem-1048	58	9	dirichlet	dirichlet	NOUN
fbem-1048	58	10	allocation	allocation	NOUN
fbem-1048	58	11	,	,	PUNCT
fbem-1048	58	12	and	and	CCONJ
fbem-1048	58	13	it	it	PRON
fbem-1048	58	14	’s	’	VERB
fbem-1048	58	15	a	a	DET
fbem-1048	58	16	very	very	ADV
fbem-1048	58	17	common	common	ADJ
fbem-1048	58	18	topic	topic	NOUN
fbem-1048	58	19	model	model	NOUN
fbem-1048	58	20	which	which	PRON
fbem-1048	58	21	often	often	ADV
fbem-1048	58	22	used	use	VERB
fbem-1048	58	23	in	in	ADP
fbem-1048	58	24	experiment	experiment	NOUN
fbem-1048	58	25	.	.	PUNCT
fbem-1048	59	1	this	this	DET
fbem-1048	59	2	kind	kind	NOUN
fbem-1048	59	3	of	of	ADP
fbem-1048	59	4	model	model	NOUN
fbem-1048	59	5	belongs	belong	VERB
fbem-1048	59	6	to	to	ADP
fbem-1048	59	7	unsupervised	unsupervised	ADJ
fbem-1048	59	8	learning	learning	NOUN
fbem-1048	59	9	model	model	NOUN
fbem-1048	59	10	which	which	PRON
fbem-1048	59	11	could	could	AUX
fbem-1048	59	12	be	be	AUX
fbem-1048	59	13	applied	apply	VERB
fbem-1048	59	14	in	in	ADP
fbem-1048	59	15	identifying	identify	VERB
fbem-1048	59	16	the	the	DET
fbem-1048	59	17	potential	potential	ADJ
fbem-1048	59	18	topic	topic	NOUN
fbem-1048	59	19	information	information	NOUN
fbem-1048	59	20	in	in	ADP
fbem-1048	59	21	the	the	DET
fbem-1048	59	22	large	large	ADJ
fbem-1048	59	23	-	-	PUNCT
fbem-1048	59	24	scale	scale	NOUN
fbem-1048	59	25	document	document	NOUN
fbem-1048	59	26	sets	set	NOUN
fbem-1048	59	27	or	or	CCONJ
fbem-1048	59	28	corpora	corpus	NOUN
fbem-1048	59	29	.	.	PUNCT
fbem-1048	60	1	the	the	DET
fbem-1048	60	2	lda	lda	PROPN
fbem-1048	60	3	model	model	NOUN
fbem-1048	60	4	also	also	ADV
fbem-1048	60	5	known	know	VERB
fbem-1048	60	6	as	as	ADP
fbem-1048	60	7	a	a	DET
fbem-1048	60	8	three	three	NUM
fbem-1048	60	9	-	-	PUNCT
fbem-1048	60	10	layer	layer	NOUN
fbem-1048	60	11	bayesian	bayesian	NOUN
fbem-1048	60	12	probability	probability	NOUN
fbem-1048	60	13	model	model	NOUN
fbem-1048	60	14	,	,	PUNCT
fbem-1048	60	15	includes	include	VERB
fbem-1048	60	16	three	three	NUM
fbem-1048	60	17	-	-	PUNCT
fbem-1048	60	18	layer	layer	NOUN
fbem-1048	60	19	structures	structure	NOUN
fbem-1048	60	20	:	:	PUNCT
fbem-1048	60	21	words	word	NOUN
fbem-1048	60	22	,	,	PUNCT
fbem-1048	60	23	topics	topic	NOUN
fbem-1048	60	24	and	and	CCONJ
fbem-1048	60	25	documents	document	NOUN
fbem-1048	60	26	.	.	PUNCT
fbem-1048	61	1	for	for	ADP
fbem-1048	61	2	each	each	DET
fbem-1048	61	3	document	document	NOUN
fbem-1048	61	4	in	in	ADP
fbem-1048	61	5	the	the	DET
fbem-1048	61	6	corpus	corpus	NOUN
fbem-1048	61	7	,	,	PUNCT
fbem-1048	61	8	the	the	DET
fbem-1048	61	9	generation	generation	NOUN
fbem-1048	61	10	process	process	NOUN
fbem-1048	61	11	is	be	AUX
fbem-1048	61	12	as	as	SCONJ
fbem-1048	61	13	follows	follow	VERB
fbem-1048	61	14	:	:	PUNCT
fbem-1048	61	15	(	(	PUNCT
fbem-1048	61	16	1	1	X
fbem-1048	61	17	)	)	PUNCT
fbem-1048	61	18	extracting	extract	VERB
fbem-1048	61	19	a	a	DET
fbem-1048	61	20	topic	topic	NOUN
fbem-1048	61	21	from	from	ADP
fbem-1048	61	22	the	the	DET
fbem-1048	61	23	topic	topic	NOUN
fbem-1048	61	24	distribution	distribution	NOUN
fbem-1048	61	25	in	in	ADP
fbem-1048	61	26	each	each	DET
fbem-1048	61	27	document	document	NOUN
fbem-1048	61	28	;	;	PUNCT
fbem-1048	61	29	(	(	PUNCT
fbem-1048	61	30	2	2	X
fbem-1048	61	31	)	)	PUNCT
fbem-1048	61	32	extracting	extract	VERB
fbem-1048	61	33	a	a	DET
fbem-1048	61	34	word	word	NOUN
fbem-1048	61	35	from	from	ADP
fbem-1048	61	36	the	the	DET
fbem-1048	61	37	word	word	NOUN
fbem-1048	61	38	distribution	distribution	NOUN
fbem-1048	61	39	corresponding	correspond	VERB
fbem-1048	61	40	to	to	ADP
fbem-1048	61	41	the	the	DET
fbem-1048	61	42	step	step	NOUN
fbem-1048	61	43	one	one	NUM
fbem-1048	61	44	’s	’s	PART
fbem-1048	61	45	topic	topic	NOUN
fbem-1048	61	46	;	;	PUNCT
fbem-1048	61	47	(	(	PUNCT
fbem-1048	61	48	3	3	X
fbem-1048	61	49	)	)	PUNCT
fbem-1048	61	50	repeating	repeat	VERB
fbem-1048	61	51	the	the	DET
fbem-1048	61	52	step	step	NOUN
fbem-1048	61	53	two	two	NUM
fbem-1048	61	54	’s	’s	PART
fbem-1048	61	55	process	process	NOUN
fbem-1048	61	56	until	until	SCONJ
fbem-1048	61	57	that	that	SCONJ
fbem-1048	61	58	all	all	DET
fbem-1048	61	59	the	the	DET
fbem-1048	61	60	words	word	NOUN
fbem-1048	61	61	in	in	ADP
fbem-1048	61	62	the	the	DET
fbem-1048	61	63	document	document	NOUN
fbem-1048	61	64	are	be	AUX
fbem-1048	61	65	traversed	traverse	VERB
fbem-1048	61	66	the	the	DET
fbem-1048	61	67	overall	overall	ADJ
fbem-1048	61	68	process	process	NOUN
fbem-1048	61	69	of	of	ADP
fbem-1048	61	70	lda	lda	PROPN
fbem-1048	61	71	is	be	AUX
fbem-1048	61	72	as	as	SCONJ
fbem-1048	61	73	follows	follow	VERB
fbem-1048	61	74	(	(	PUNCT
fbem-1048	61	75	define	define	VERB
fbem-1048	61	76	the	the	DET
fbem-1048	61	77	meaning	meaning	NOUN
fbem-1048	61	78	of	of	ADP
fbem-1048	61	79	some	some	DET
fbem-1048	61	80	letters	letter	NOUN
fbem-1048	61	81	:	:	PUNCT
fbem-1048	61	82	document	document	NOUN
fbem-1048	61	83	set	set	VERB
fbem-1048	61	84	d	d	NOUN
fbem-1048	61	85	,	,	PUNCT
fbem-1048	61	86	topic	topic	NOUN
fbem-1048	61	87	set	set	VERB
fbem-1048	61	88	t	t	PROPN
fbem-1048	61	89	):	):	PUNCT
fbem-1048	61	90	(	(	PUNCT
fbem-1048	61	91	1	1	X
fbem-1048	61	92	)	)	PUNCT
fbem-1048	61	93	each	each	DET
fbem-1048	61	94	document	document	NOUN
fbem-1048	62	1	d	d	NOUN
fbem-1048	62	2	in	in	ADP
fbem-1048	62	3	d	d	PROPN
fbem-1048	62	4	is	be	AUX
fbem-1048	62	5	regarded	regard	VERB
fbem-1048	62	6	as	as	ADP
fbem-1048	62	7	a	a	DET
fbem-1048	62	8	word	word	NOUN
fbem-1048	62	9	sequence	sequence	NOUN
fbem-1048	62	10	<	<	X
fbem-1048	62	11	w1,w2,	w1,w2,	PROPN
fbem-1048	62	12	…	…	SYM
fbem-1048	62	13	,wn	,wn	PROPN
fbem-1048	62	14	>	>	X
fbem-1048	62	15	,	,	PUNCT
fbem-1048	62	16	wi	wi	PROPN
fbem-1048	62	17	represents	represent	VERB
fbem-1048	62	18	the	the	DET
fbem-1048	62	19	i	i	PROPN
fbem-1048	62	20	-	-	PUNCT
fbem-1048	62	21	th	th	VERB
fbem-1048	62	22	word	word	NOUN
fbem-1048	62	23	,	,	PUNCT
fbem-1048	62	24	and	and	CCONJ
fbem-1048	62	25	d	d	NOUN
fbem-1048	62	26	has	have	VERB
fbem-1048	63	1	n	n	NUM
fbem-1048	63	2	words	word	NOUN
fbem-1048	63	3	.	.	PUNCT
fbem-1048	64	1	(	(	PUNCT
fbem-1048	64	2	it	it	PRON
fbem-1048	64	3	is	be	AUX
fbem-1048	64	4	called	call	VERB
fbem-1048	64	5	wordbag	wordbag	NOUN
fbem-1048	64	6	in	in	ADP
fbem-1048	64	7	lda	lda	PROPN
fbem-1048	64	8	,	,	PUNCT
fbem-1048	64	9	in	in	ADP
fbem-1048	64	10	fact	fact	NOUN
fbem-1048	64	11	,	,	PUNCT
fbem-1048	64	12	the	the	DET
fbem-1048	64	13	appearance	appearance	NOUN
fbem-1048	64	14	position	position	NOUN
fbem-1048	64	15	of	of	ADP
fbem-1048	64	16	each	each	DET
fbem-1048	64	17	word	word	NOUN
fbem-1048	64	18	has	have	VERB
fbem-1048	64	19	no	no	DET
fbem-1048	64	20	effect	effect	NOUN
fbem-1048	64	21	on	on	ADP
fbem-1048	64	22	the	the	DET
fbem-1048	64	23	lda	lda	PROPN
fbem-1048	64	24	algorithm	algorithm	NOUN
fbem-1048	64	25	)	)	PUNCT
fbem-1048	64	26	(	(	PUNCT
fbem-1048	64	27	2	2	X
fbem-1048	64	28	)	)	PUNCT
fbem-1048	64	29	all	all	DET
fbem-1048	64	30	the	the	DET
fbem-1048	64	31	different	different	ADJ
fbem-1048	64	32	words	word	NOUN
fbem-1048	64	33	involved	involve	VERB
fbem-1048	64	34	in	in	ADP
fbem-1048	64	35	d	d	PROPN
fbem-1048	64	36	form	form	VERB
fbem-1048	64	37	a	a	DET
fbem-1048	64	38	large	large	ADJ
fbem-1048	64	39	set	set	VERB
fbem-1048	64	40	vocabulary	vocabulary	NOUN
fbem-1048	64	41	(	(	PUNCT
fbem-1048	64	42	voc	voc	NOUN
fbem-1048	64	43	for	for	ADP
fbem-1048	64	44	short	short	ADJ
fbem-1048	64	45	)	)	PUNCT
fbem-1048	64	46	,	,	PUNCT
fbem-1048	64	47	lda	lda	PROPN
fbem-1048	64	48	takes	take	VERB
fbem-1048	64	49	the	the	DET
fbem-1048	64	50	document	document	NOUN
fbem-1048	64	51	set	set	VERB
fbem-1048	64	52	d	d	NOUN
fbem-1048	64	53	as	as	ADP
fbem-1048	64	54	input	input	NOUN
fbem-1048	64	55	,	,	PUNCT
fbem-1048	64	56	and	and	CCONJ
fbem-1048	64	57	hopes	hope	VERB
fbem-1048	64	58	to	to	PART
fbem-1048	64	59	train	train	VERB
fbem-1048	64	60	two	two	NUM
fbem-1048	64	61	result	result	NOUN
fbem-1048	64	62	vectors	vector	NOUN
fbem-1048	64	63	(	(	PUNCT
fbem-1048	64	64	set	set	VERB
fbem-1048	64	65	to	to	PART
fbem-1048	64	66	be	be	AUX
fbem-1048	64	67	clustered	cluster	VERB
fbem-1048	64	68	into	into	ADP
fbem-1048	64	69	k	k	PROPN
fbem-1048	64	70	topics	topic	NOUN
fbem-1048	64	71	,	,	PUNCT
fbem-1048	64	72	and	and	CCONJ
fbem-1048	64	73	voc	voc	NOUN
fbem-1048	64	74	contains	contain	VERB
fbem-1048	64	75	m	m	VERB
fbem-1048	64	76	words	word	NOUN
fbem-1048	64	77	in	in	ADP
fbem-1048	64	78	total	total	NOUN
fbem-1048	64	79	)	)	PUNCT
fbem-1048	64	80	)	)	PUNCT
fbem-1048	64	81	;	;	PUNCT
fbem-1048	64	82	(	(	PUNCT
fbem-1048	64	83	3	3	X
fbem-1048	64	84	)	)	PUNCT
fbem-1048	64	85	for	for	ADP
fbem-1048	64	86	each	each	DET
fbem-1048	64	87	document	document	NOUN
fbem-1048	64	88	d	d	NOUN
fbem-1048	64	89	in	in	ADP
fbem-1048	64	90	d	d	PROPN
fbem-1048	64	91	,	,	PUNCT
fbem-1048	64	92	the	the	DET
fbem-1048	64	93	probability	probability	NOUN
fbem-1048	64	94	θd	θd	VERB
fbem-1048	64	95	<	<	X
fbem-1048	64	96	pt1,	pt1,	X
fbem-1048	64	97	…	…	PUNCT
fbem-1048	64	98	,ptk	,ptk	PUNCT
fbem-1048	64	99	>	>	X
fbem-1048	64	100	corresponding	correspond	VERB
fbem-1048	64	101	to	to	ADP
fbem-1048	64	102	different	different	ADJ
fbem-1048	64	103	topics	topic	NOUN
fbem-1048	64	104	,	,	PUNCT
fbem-1048	64	105	where	where	SCONJ
fbem-1048	64	106	pti	pti	PROPN
fbem-1048	64	107	represents	represent	VERB
fbem-1048	64	108	the	the	DET
fbem-1048	64	109	probability	probability	NOUN
fbem-1048	64	110	that	that	SCONJ
fbem-1048	64	111	d	d	NOUN
fbem-1048	64	112	corresponds	correspond	VERB
fbem-1048	64	113	to	to	ADP
fbem-1048	64	114	the	the	DET
fbem-1048	64	115	ith	ith	PROPN
fbem-1048	64	116	topic	topic	NOUN
fbem-1048	64	117	in	in	ADP
fbem-1048	64	118	t.	t.	PROPN
fbem-1048	64	119	the	the	DET
fbem-1048	64	120	calculation	calculation	NOUN
fbem-1048	64	121	method	method	NOUN
fbem-1048	64	122	is	be	AUX
fbem-1048	64	123	intuitive	intuitive	ADJ
fbem-1048	64	124	,	,	PUNCT
fbem-1048	64	125	pti	pti	PROPN
fbem-1048	64	126	=	=	PROPN
fbem-1048	64	127	nti	nti	NOUN
fbem-1048	64	128	/	/	SYM
fbem-1048	64	129	n	n	CCONJ
fbem-1048	64	130	,	,	PUNCT
fbem-1048	64	131	where	where	SCONJ
fbem-1048	64	132	nti	nti	PROPN
fbem-1048	64	133	represents	represent	VERB
fbem-1048	64	134	the	the	DET
fbem-1048	64	135	number	number	NOUN
fbem-1048	64	136	of	of	ADP
fbem-1048	64	137	words	word	NOUN
fbem-1048	64	138	in	in	ADP
fbem-1048	64	139	d	d	PROPN
fbem-1048	64	140	corresponding	correspond	VERB
fbem-1048	64	141	to	to	ADP
fbem-1048	64	142	the	the	DET
fbem-1048	64	143	ith	ith	PROPN
fbem-1048	64	144	topic	topic	NOUN
fbem-1048	64	145	,	,	PUNCT
fbem-1048	64	146	and	and	CCONJ
fbem-1048	64	147	n	n	PRON
fbem-1048	64	148	is	be	AUX
fbem-1048	64	149	the	the	DET
fbem-1048	64	150	total	total	ADJ
fbem-1048	64	151	number	number	NOUN
fbem-1048	64	152	of	of	ADP
fbem-1048	64	153	all	all	DET
fbem-1048	64	154	words	word	NOUN
fbem-1048	64	155	in	in	ADP
fbem-1048	64	156	d.	d.	PROPN
fbem-1048	64	157	(	(	PUNCT
fbem-1048	64	158	4	4	NUM
fbem-1048	64	159	)	)	PUNCT
fbem-1048	64	160	for	for	ADP
fbem-1048	64	161	each	each	DET
fbem-1048	64	162	topic	topic	NOUN
fbem-1048	64	163	in	in	ADP
fbem-1048	64	164	t	t	PROPN
fbem-1048	64	165	,	,	PUNCT
fbem-1048	64	166	the	the	DET
fbem-1048	64	167	probability	probability	NOUN
fbem-1048	64	168	of	of	ADP
fbem-1048	64	169	generating	generate	VERB
fbem-1048	64	170	different	different	ADJ
fbem-1048	64	171	words	word	NOUN
fbem-1048	64	172	φt	φt	ADP
fbem-1048	64	173	<	<	X
fbem-1048	64	174	pw1,	pw1,	X
fbem-1048	64	175	…	…	SYM
fbem-1048	64	176	,pwm	,pwm	PUNCT
fbem-1048	64	177	>	>	X
fbem-1048	64	178	,	,	PUNCT
fbem-1048	64	179	where	where	SCONJ
fbem-1048	64	180	pwi	pwi	PROPN
fbem-1048	64	181	represents	represent	VERB
fbem-1048	64	182	the	the	DET
fbem-1048	64	183	probability	probability	NOUN
fbem-1048	64	184	that	that	SCONJ
fbem-1048	64	185	t	t	PROPN
fbem-1048	64	186	generates	generate	VERB
fbem-1048	64	187	the	the	DET
fbem-1048	64	188	i	i	PROPN
fbem-1048	64	189	-	-	PUNCT
fbem-1048	64	190	th	th	VERB
fbem-1048	64	191	word	word	NOUN
fbem-1048	64	192	in	in	ADP
fbem-1048	64	193	voc	voc	NOUN
fbem-1048	64	194	.	.	PUNCT
fbem-1048	65	1	the	the	DET
fbem-1048	65	2	calculation	calculation	NOUN
fbem-1048	65	3	method	method	NOUN
fbem-1048	65	4	is	be	AUX
fbem-1048	65	5	also	also	ADV
fbem-1048	65	6	very	very	ADV
fbem-1048	65	7	intuitive	intuitive	ADJ
fbem-1048	65	8	,	,	PUNCT
fbem-1048	65	9	pwi	pwi	PROPN
fbem-1048	65	10	=	=	PROPN
fbem-1048	65	11	nwi	nwi	PROPN
fbem-1048	65	12	/	/	SYM
fbem-1048	65	13	n	n	CCONJ
fbem-1048	65	14	,	,	PUNCT
fbem-1048	65	15	where	where	SCONJ
fbem-1048	65	16	nwi	nwi	NOUN
fbem-1048	65	17	represents	represent	VERB
fbem-1048	65	18	the	the	DET
fbem-1048	65	19	number	number	NOUN
fbem-1048	65	20	of	of	ADP
fbem-1048	65	21	the	the	DET
fbem-1048	65	22	i	i	PROPN
fbem-1048	65	23	-	-	PUNCT
fbem-1048	65	24	th	th	VERB
fbem-1048	65	25	word	word	NOUN
fbem-1048	65	26	in	in	ADP
fbem-1048	65	27	the	the	DET
fbem-1048	65	28	voc	voc	NOUN
fbem-1048	65	29	corresponding	correspond	VERB
fbem-1048	65	30	to	to	ADP
fbem-1048	65	31	topict	topict	NOUN
fbem-1048	65	32	,	,	PUNCT
fbem-1048	65	33	and	and	CCONJ
fbem-1048	65	34	n	n	CCONJ
fbem-1048	65	35	represents	represent	VERB
fbem-1048	65	36	the	the	DET
fbem-1048	65	37	total	total	ADJ
fbem-1048	65	38	number	number	NOUN
fbem-1048	65	39	of	of	ADP
fbem-1048	65	40	words	word	NOUN
fbem-1048	65	41	corresponding	correspond	VERB
fbem-1048	65	42	to	to	ADP
fbem-1048	65	43	topict	topict	NOUN
fbem-1048	65	44	.	.	PUNCT
fbem-1048	66	1	the	the	DET
fbem-1048	66	2	core	core	NOUN
fbem-1048	66	3	formula	formula	NOUN
fbem-1048	66	4	of	of	ADP
fbem-1048	66	5	lda	lda	NOUN
fbem-1048	66	6	:	:	PUNCT
fbem-1048	66	7	p(w|d)=p(w|t)*p(t|d	p(w|d)=p(w|t)*p(t|d	NUM
fbem-1048	66	8	)	)	PUNCT
fbem-1048	66	9	(	(	PUNCT
fbem-1048	66	10	1	1	NUM
fbem-1048	66	11	)	)	PUNCT
fbem-1048	66	12	3	3	NUM
fbem-1048	66	13	.	.	PUNCT
fbem-1048	67	1	jingdong	jingdong	ADP
fbem-1048	67	2	comments	comment	NOUN
fbem-1048	67	3	are	be	AUX
fbem-1048	67	4	expected	expect	VERB
fbem-1048	67	5	to	to	PART
fbem-1048	67	6	be	be	AUX
fbem-1048	67	7	obtained	obtain	VERB
fbem-1048	67	8	3.1	3.1	NUM
fbem-1048	67	9	.	.	PUNCT
fbem-1048	67	10	obtaining	obtain	VERB
fbem-1048	67	11	method	method	NOUN
fbem-1048	67	12	in	in	ADP
fbem-1048	67	13	order	order	NOUN
fbem-1048	67	14	to	to	PART
fbem-1048	67	15	obtain	obtain	VERB
fbem-1048	67	16	the	the	DET
fbem-1048	67	17	latest	late	ADJ
fbem-1048	67	18	comment	comment	NOUN
fbem-1048	67	19	data	datum	NOUN
fbem-1048	67	20	,	,	PUNCT
fbem-1048	67	21	it	it	PRON
fbem-1048	67	22	is	be	AUX
fbem-1048	67	23	implemented	implement	VERB
fbem-1048	67	24	through	through	ADP
fbem-1048	67	25	crawler	crawler	NOUN
fbem-1048	67	26	technology	technology	NOUN
fbem-1048	67	27	.	.	PUNCT
fbem-1048	68	1	the	the	DET
fbem-1048	68	2	crawler	crawler	NOUN
fbem-1048	68	3	technology	technology	NOUN
fbem-1048	68	4	can	can	AUX
fbem-1048	68	5	automatically	automatically	ADV
fbem-1048	68	6	crawl	crawl	VERB
fbem-1048	68	7	the	the	DET
fbem-1048	68	8	information	information	NOUN
fbem-1048	68	9	on	on	ADP
fbem-1048	68	10	the	the	DET
fbem-1048	68	11	webpage	webpage	NOUN
fbem-1048	68	12	by	by	ADP
fbem-1048	68	13	writing	write	VERB
fbem-1048	68	14	a	a	DET
fbem-1048	68	15	program	program	NOUN
fbem-1048	68	16	,	,	PUNCT
fbem-1048	68	17	and	and	CCONJ
fbem-1048	68	18	write	write	VERB
fbem-1048	68	19	the	the	DET
fbem-1048	68	20	code	code	NOUN
fbem-1048	68	21	according	accord	VERB
fbem-1048	68	22	to	to	ADP
fbem-1048	68	23	the	the	DET
fbem-1048	68	24	url	url	NOUN
fbem-1048	68	25	and	and	CCONJ
fbem-1048	68	26	user	user	NOUN
fbem-1048	68	27	-	-	PUNCT
fbem-1048	68	28	agent	agent	NOUN
fbem-1048	68	29	obtained	obtain	VERB
fbem-1048	68	30	in	in	ADP
fbem-1048	68	31	the	the	DET
fbem-1048	68	32	webpage	webpage	NOUN
fbem-1048	68	33	to	to	PART
fbem-1048	68	34	crawl	crawl	VERB
fbem-1048	68	35	the	the	DET
fbem-1048	68	36	information	information	NOUN
fbem-1048	68	37	of	of	ADP
fbem-1048	68	38	the	the	DET
fbem-1048	68	39	specified	specified	ADJ
fbem-1048	68	40	page	page	NOUN
fbem-1048	68	41	.	.	PUNCT
fbem-1048	69	1	the	the	DET
fbem-1048	69	2	crawler	crawler	NOUN
fbem-1048	69	3	technology	technology	NOUN
fbem-1048	69	4	is	be	AUX
fbem-1048	69	5	often	often	ADV
fbem-1048	69	6	limited	limit	VERB
fbem-1048	69	7	by	by	ADP
fbem-1048	69	8	various	various	ADJ
fbem-1048	69	9	websites	website	NOUN
fbem-1048	69	10	,	,	PUNCT
fbem-1048	69	11	and	and	CCONJ
fbem-1048	69	12	sometimes	sometimes	ADV
fbem-1048	69	13	it	it	PRON
fbem-1048	69	14	is	be	AUX
fbem-1048	69	15	frequently	frequently	ADV
fbem-1048	69	16	accessed	access	VERB
fbem-1048	69	17	.	.	PUNCT
fbem-1048	70	1	it	it	PRON
fbem-1048	70	2	will	will	AUX
fbem-1048	70	3	be	be	AUX
fbem-1048	70	4	monitored	monitor	VERB
fbem-1048	70	5	by	by	ADP
fbem-1048	70	6	the	the	DET
fbem-1048	70	7	platform	platform	NOUN
fbem-1048	70	8	in	in	ADP
fbem-1048	70	9	the	the	DET
fbem-1048	70	10	background	background	NOUN
fbem-1048	70	11	to	to	PART
fbem-1048	70	12	limit	limit	VERB
fbem-1048	70	13	the	the	DET
fbem-1048	70	14	number	number	NOUN
fbem-1048	70	15	of	of	ADP
fbem-1048	70	16	visits	visit	NOUN
fbem-1048	70	17	,	,	PUNCT
fbem-1048	70	18	which	which	PRON
fbem-1048	70	19	is	be	AUX
fbem-1048	70	20	not	not	PART
fbem-1048	70	21	applicable	applicable	ADJ
fbem-1048	70	22	when	when	SCONJ
fbem-1048	70	23	obtaining	obtain	VERB
fbem-1048	70	24	a	a	DET
fbem-1048	70	25	large	large	ADJ
fbem-1048	70	26	amount	amount	NOUN
fbem-1048	70	27	of	of	ADP
fbem-1048	70	28	data	datum	NOUN
fbem-1048	70	29	.	.	PUNCT
fbem-1048	71	1	the	the	DET
fbem-1048	71	2	amount	amount	NOUN
fbem-1048	71	3	of	of	ADP
fbem-1048	71	4	data	datum	NOUN
fbem-1048	71	5	obtained	obtain	VERB
fbem-1048	71	6	here	here	ADV
fbem-1048	71	7	is	be	AUX
fbem-1048	71	8	not	not	PART
fbem-1048	71	9	large	large	ADJ
fbem-1048	71	10	,	,	PUNCT
fbem-1048	71	11	so	so	SCONJ
fbem-1048	71	12	the	the	DET
fbem-1048	71	13	crawler	crawler	NOUN
fbem-1048	71	14	is	be	AUX
fbem-1048	71	15	directly	directly	ADV
fbem-1048	71	16	used	use	VERB
fbem-1048	71	17	to	to	PART
fbem-1048	71	18	obtain	obtain	VERB
fbem-1048	71	19	the	the	DET
fbem-1048	71	20	data	datum	NOUN
fbem-1048	71	21	.	.	PUNCT
fbem-1048	72	1	the	the	DET
fbem-1048	72	2	data	datum	NOUN
fbem-1048	72	3	crawled	crawl	VERB
fbem-1048	72	4	from	from	ADP
fbem-1048	72	5	jd.com	jd.com	PROPN
fbem-1048	72	6	comments	comment	NOUN
fbem-1048	72	7	in	in	ADP
fbem-1048	72	8	this	this	DET
fbem-1048	72	9	paper	paper	NOUN
fbem-1048	72	10	are	be	AUX
fbem-1048	72	11	divided	divide	VERB
fbem-1048	72	12	into	into	ADP
fbem-1048	72	13	the	the	DET
fbem-1048	72	14	two	two	NUM
fbem-1048	72	15	categories	category	NOUN
fbem-1048	72	16	,	,	PUNCT
fbem-1048	72	17	one	one	NUM
fbem-1048	72	18	is	be	AUX
fbem-1048	72	19	about	about	ADP
fbem-1048	72	20	negative	negative	ADJ
fbem-1048	72	21	comments	comment	NOUN
fbem-1048	72	22	and	and	CCONJ
fbem-1048	72	23	another	another	PRON
fbem-1048	72	24	is	be	AUX
fbem-1048	72	25	about	about	ADP
fbem-1048	72	26	the	the	DET
fbem-1048	72	27	positive	positive	ADJ
fbem-1048	72	28	comments	comment	NOUN
fbem-1048	72	29	.	.	PUNCT
fbem-1048	73	1	according	accord	VERB
fbem-1048	73	2	to	to	ADP
fbem-1048	73	3	the	the	DET
fbem-1048	73	4	corresponding	corresponding	ADJ
fbem-1048	73	5	parameters	parameter	NOUN
fbem-1048	73	6	in	in	ADP
fbem-1048	73	7	the	the	DET
fbem-1048	73	8	transformation	transformation	NOUN
fbem-1048	73	9	code	code	NOUN
fbem-1048	73	10	,	,	PUNCT
fbem-1048	73	11	the	the	DET
fbem-1048	73	12	two	two	NUM
fbem-1048	73	13	types	type	NOUN
fbem-1048	73	14	of	of	ADP
fbem-1048	73	15	data	datum	NOUN
fbem-1048	73	16	are	be	AUX
fbem-1048	73	17	saved	save	VERB
fbem-1048	73	18	in	in	ADP
fbem-1048	73	19	different	different	ADJ
fbem-1048	73	20	documents	document	NOUN
fbem-1048	73	21	for	for	ADP
fbem-1048	73	22	backup	backup	NOUN
fbem-1048	73	23	(	(	PUNCT
fbem-1048	73	24	the	the	DET
fbem-1048	73	25	data	data	NOUN
fbem-1048	73	26	content	content	NOUN
fbem-1048	73	27	is	be	AUX
fbem-1048	73	28	the	the	DET
fbem-1048	73	29	comments	comment	NOUN
fbem-1048	73	30	of	of	ADP
fbem-1048	73	31	a	a	DET
fbem-1048	73	32	certain	certain	ADJ
fbem-1048	73	33	brand	brand	NOUN
fbem-1048	73	34	of	of	ADP
fbem-1048	73	35	computer	computer	NOUN
fbem-1048	73	36	)	)	PUNCT
fbem-1048	73	37	.	.	PUNCT
fbem-1048	74	1	27	27	NUM
fbem-1048	74	2	figure	figure	NOUN
fbem-1048	74	3	1	1	NUM
fbem-1048	74	4	.	.	PUNCT
fbem-1048	74	5	extraction	extraction	NOUN
fbem-1048	74	6	of	of	ADP
fbem-1048	74	7	bad	bad	ADJ
fbem-1048	74	8	reviews	review	NOUN
fbem-1048	74	9	figure	figure	VERB
fbem-1048	74	10	2	2	NUM
fbem-1048	74	11	.	.	PUNCT
fbem-1048	74	12	extraction	extraction	NOUN
fbem-1048	74	13	of	of	ADP
fbem-1048	74	14	positive	positive	ADJ
fbem-1048	74	15	comments	comment	NOUN
fbem-1048	74	16	4	4	NUM
fbem-1048	74	17	.	.	PUNCT
fbem-1048	75	1	data	datum	NOUN
fbem-1048	75	2	preprocessing	preprocesse	VERB
fbem-1048	75	3	4.1	4.1	NUM
fbem-1048	75	4	.	.	PUNCT
fbem-1048	76	1	the	the	DET
fbem-1048	76	2	overall	overall	ADJ
fbem-1048	76	3	preprocessing	preprocessing	NOUN
fbem-1048	76	4	process	process	NOUN
fbem-1048	76	5	first	first	ADV
fbem-1048	76	6	,	,	PUNCT
fbem-1048	76	7	perform	perform	VERB
fbem-1048	76	8	word	word	NOUN
fbem-1048	76	9	segmentation	segmentation	NOUN
fbem-1048	76	10	on	on	ADP
fbem-1048	76	11	the	the	DET
fbem-1048	76	12	obtained	obtain	VERB
fbem-1048	76	13	jingdong	jingdong	ADP
fbem-1048	76	14	comment	comment	NOUN
fbem-1048	76	15	text	text	NOUN
fbem-1048	76	16	data	datum	NOUN
fbem-1048	76	17	,	,	PUNCT
fbem-1048	76	18	then	then	ADV
fbem-1048	76	19	perform	perform	VERB
fbem-1048	76	20	part	part	NOUN
fbem-1048	76	21	-	-	PUNCT
fbem-1048	76	22	of	of	ADP
fbem-1048	76	23	-	-	PUNCT
fbem-1048	76	24	speech	speech	NOUN
fbem-1048	76	25	tagging	tagging	NOUN
fbem-1048	76	26	,	,	PUNCT
fbem-1048	76	27	stop	stop	VERB
fbem-1048	76	28	word	word	NOUN
fbem-1048	76	29	deletion	deletion	NOUN
fbem-1048	76	30	,	,	PUNCT
fbem-1048	76	31	etc	etc	X
fbem-1048	76	32	.	.	X
fbem-1048	77	1	the	the	DET
fbem-1048	77	2	next	next	ADJ
fbem-1048	77	3	step	step	NOUN
fbem-1048	77	4	is	be	AUX
fbem-1048	77	5	to	to	PART
fbem-1048	77	6	perform	perform	VERB
fbem-1048	77	7	syntactic	syntactic	ADJ
fbem-1048	77	8	analysis	analysis	NOUN
fbem-1048	77	9	,	,	PUNCT
fbem-1048	77	10	and	and	CCONJ
fbem-1048	77	11	finally	finally	ADV
fbem-1048	77	12	generate	generate	VERB
fbem-1048	77	13	word	word	NOUN
fbem-1048	77	14	vectors	vector	NOUN
fbem-1048	77	15	as	as	SCONJ
fbem-1048	77	16	needed	need	VERB
fbem-1048	77	17	.	.	PUNCT
fbem-1048	78	1	it	it	PRON
fbem-1048	78	2	should	should	AUX
fbem-1048	78	3	be	be	AUX
fbem-1048	78	4	noted	note	VERB
fbem-1048	78	5	that	that	SCONJ
fbem-1048	78	6	when	when	SCONJ
fbem-1048	78	7	analyzing	analyze	VERB
fbem-1048	78	8	chinese	chinese	ADJ
fbem-1048	78	9	text	text	NOUN
fbem-1048	78	10	,	,	PUNCT
fbem-1048	78	11	a	a	DET
fbem-1048	78	12	special	special	ADJ
fbem-1048	78	13	word	word	NOUN
fbem-1048	78	14	segmentation	segmentation	NOUN
fbem-1048	78	15	algorithm	algorithm	NOUN
fbem-1048	78	16	is	be	AUX
fbem-1048	78	17	required	require	VERB
fbem-1048	78	18	to	to	PART
fbem-1048	78	19	assist	assist	VERB
fbem-1048	78	20	the	the	DET
fbem-1048	78	21	computer	computer	NOUN
fbem-1048	78	22	in	in	ADP
fbem-1048	78	23	identifying	identify	VERB
fbem-1048	78	24	words	word	NOUN
fbem-1048	78	25	,	,	PUNCT
fbem-1048	78	26	phrases	phrase	NOUN
fbem-1048	78	27	and	and	CCONJ
fbem-1048	78	28	fixed	fix	VERB
fbem-1048	78	29	collocations	collocation	NOUN
fbem-1048	78	30	in	in	ADP
fbem-1048	78	31	chinese	chinese	ADJ
fbem-1048	78	32	text	text	NOUN
fbem-1048	78	33	.	.	PUNCT
fbem-1048	79	1	there	there	PRON
fbem-1048	79	2	are	be	VERB
fbem-1048	79	3	roughly	roughly	ADV
fbem-1048	79	4	four	four	NUM
fbem-1048	79	5	kinds	kind	NOUN
fbem-1048	79	6	of	of	ADP
fbem-1048	79	7	common	common	ADJ
fbem-1048	79	8	stop	stop	NOUN
fbem-1048	79	9	word	word	NOUN
fbem-1048	79	10	lists	list	NOUN
fbem-1048	79	11	in	in	ADP
fbem-1048	79	12	chinese	chinese	PROPN
fbem-1048	79	13	:	:	PUNCT
fbem-1048	79	14	(	(	PUNCT
fbem-1048	79	15	1	1	X
fbem-1048	79	16	)	)	PUNCT
fbem-1048	79	17	the	the	DET
fbem-1048	79	18	harbin	harbin	PROPN
fbem-1048	79	19	institute	institute	PROPN
fbem-1048	79	20	of	of	ADP
fbem-1048	79	21	technology	technology	PROPN
fbem-1048	79	22	chinese	chinese	PROPN
fbem-1048	79	23	stop	stop	VERB
fbem-1048	79	24	word	word	NOUN
fbem-1048	79	25	list	list	NOUN
fbem-1048	79	26	;	;	PUNCT
fbem-1048	79	27	(	(	PUNCT
fbem-1048	79	28	2	2	X
fbem-1048	79	29	)	)	PUNCT
fbem-1048	79	30	the	the	DET
fbem-1048	79	31	chinese	chinese	PROPN
fbem-1048	79	32	stop	stop	VERB
fbem-1048	79	33	word	word	NOUN
fbem-1048	79	34	list	list	NOUN
fbem-1048	79	35	;	;	PUNCT
fbem-1048	79	36	(	(	PUNCT
fbem-1048	79	37	3	3	X
fbem-1048	79	38	)	)	PUNCT
fbem-1048	79	39	baidu	baidu	NOUN
fbem-1048	79	40	stop	stop	VERB
fbem-1048	79	41	word	word	NOUN
fbem-1048	79	42	list	list	NOUN
fbem-1048	79	43	;	;	PUNCT
fbem-1048	79	44	(	(	PUNCT
fbem-1048	79	45	4	4	X
fbem-1048	79	46	)	)	PUNCT
fbem-1048	79	47	sichuan	sichuan	PROPN
fbem-1048	79	48	university	university	PROPN
fbem-1048	79	49	machine	machine	NOUN
fbem-1048	79	50	smart	smart	ADJ
fbem-1048	79	51	lab	lab	NOUN
fbem-1048	79	52	stop	stop	NOUN
fbem-1048	79	53	-	-	PUNCT
fbem-1048	79	54	word	word	NOUN
fbem-1048	79	55	library	library	NOUN
fbem-1048	79	56	.	.	PUNCT
fbem-1048	80	1	in	in	ADP
fbem-1048	80	2	this	this	DET
fbem-1048	80	3	experiment	experiment	NOUN
fbem-1048	80	4	,	,	PUNCT
fbem-1048	80	5	the	the	DET
fbem-1048	80	6	using	use	VERB
fbem-1048	80	7	stop	stop	NOUN
fbem-1048	80	8	word	word	NOUN
fbem-1048	80	9	list	list	NOUN
fbem-1048	80	10	in	in	ADP
fbem-1048	80	11	chinese	chinese	PROPN
fbem-1048	80	12	is	be	AUX
fbem-1048	80	13	the	the	DET
fbem-1048	80	14	forth	forth	ADJ
fbem-1048	80	15	one	one	NUM
fbem-1048	80	16	above	above	ADV
fbem-1048	80	17	.	.	PUNCT
fbem-1048	81	1	5	5	X
fbem-1048	81	2	.	.	PUNCT
fbem-1048	81	3	model	model	NOUN
fbem-1048	81	4	establishment	establishment	NOUN
fbem-1048	81	5	5.1	5.1	NUM
fbem-1048	81	6	.	.	PUNCT
fbem-1048	82	1	lda	lda	PROPN
fbem-1048	82	2	model	model	PROPN
fbem-1048	82	3	establishment	establishment	NOUN
fbem-1048	82	4	first	first	ADV
fbem-1048	82	5	use	use	VERB
fbem-1048	82	6	gensim	gensim	NOUN
fbem-1048	82	7	for	for	ADP
fbem-1048	82	8	model	model	NOUN
fbem-1048	82	9	building	building	NOUN
fbem-1048	82	10	:	:	PUNCT
fbem-1048	82	11	the	the	DET
fbem-1048	82	12	corpora	corpora	PROPN
fbem-1048	82	13	.	.	PUNCT
fbem-1048	83	1	dictionary	dictionary	PROPN
fbem-1048	83	2	method	method	PROPN
fbem-1048	83	3	builds	build	VERB
fbem-1048	83	4	a	a	DET
fbem-1048	83	5	specific	specific	ADJ
fbem-1048	83	6	object	object	NOUN
fbem-1048	83	7	to	to	ADP
fbem-1048	83	8	store	store	NOUN
fbem-1048	83	9	corpus	corpus	NOUN
fbem-1048	83	10	data	datum	NOUN
fbem-1048	83	11	,	,	PUNCT
fbem-1048	83	12	and	and	CCONJ
fbem-1048	83	13	converts	convert	VERB
fbem-1048	83	14	the	the	DET
fbem-1048	83	15	cleaned	clean	VERB
fbem-1048	83	16	information	information	NOUN
fbem-1048	83	17	into	into	ADP
fbem-1048	83	18	a	a	DET
fbem-1048	83	19	gensimapproved	gensimapprove	VERB
fbem-1048	83	20	corpus	corpus	NOUN
fbem-1048	83	21	form	form	NOUN
fbem-1048	83	22	.	.	PUNCT
fbem-1048	84	1	the	the	DET
fbem-1048	84	2	second	second	ADJ
fbem-1048	84	3	step	step	NOUN
fbem-1048	84	4	uses	use	VERB
fbem-1048	84	5	models	model	NOUN
fbem-1048	84	6	.	.	PUNCT
fbem-1048	85	1	ldamodel	ldamodel	PROPN
fbem-1048	85	2	to	to	PART
fbem-1048	85	3	train	train	VERB
fbem-1048	85	4	the	the	DET
fbem-1048	85	5	model	model	NOUN
fbem-1048	85	6	on	on	ADP
fbem-1048	85	7	the	the	DET
fbem-1048	85	8	processed	process	VERB
fbem-1048	85	9	expectations	expectation	NOUN
fbem-1048	85	10	.	.	PUNCT
fbem-1048	86	1	6	6	X
fbem-1048	86	2	.	.	X
fbem-1048	86	3	result	result	NOUN
fbem-1048	86	4	after	after	ADP
fbem-1048	86	5	writing	write	VERB
fbem-1048	86	6	and	and	CCONJ
fbem-1048	86	7	running	run	VERB
fbem-1048	86	8	the	the	DET
fbem-1048	86	9	code	code	NOUN
fbem-1048	86	10	,	,	PUNCT
fbem-1048	86	11	the	the	DET
fbem-1048	86	12	topic	topic	NOUN
fbem-1048	86	13	extraction	extraction	NOUN
fbem-1048	86	14	situation	situation	NOUN
fbem-1048	86	15	after	after	SCONJ
fbem-1048	86	16	lda	lda	PROPN
fbem-1048	86	17	modeling	modeling	NOUN
fbem-1048	86	18	is	be	AUX
fbem-1048	86	19	obtained	obtain	VERB
fbem-1048	86	20	(	(	PUNCT
fbem-1048	86	21	the	the	DET
fbem-1048	86	22	experimental	experimental	ADJ
fbem-1048	86	23	results	result	NOUN
fbem-1048	86	24	of	of	ADP
fbem-1048	86	25	the	the	DET
fbem-1048	86	26	positive	positive	ADJ
fbem-1048	86	27	comments	comment	NOUN
fbem-1048	86	28	and	and	CCONJ
fbem-1048	86	29	the	the	DET
fbem-1048	86	30	negative	negative	ADJ
fbem-1048	86	31	comments	comment	NOUN
fbem-1048	86	32	are	be	AUX
fbem-1048	86	33	illustrated	illustrate	VERB
fbem-1048	86	34	in	in	ADP
fbem-1048	86	35	the	the	DET
fbem-1048	86	36	figure	figure	NOUN
fbem-1048	86	37	3	3	NUM
fbem-1048	86	38	and	and	CCONJ
fbem-1048	86	39	figure	figure	VERB
fbem-1048	86	40	4	4	NUM
fbem-1048	86	41	)	)	PUNCT
fbem-1048	86	42	,	,	PUNCT
fbem-1048	86	43	which	which	PRON
fbem-1048	86	44	are	be	AUX
fbem-1048	86	45	divided	divide	VERB
fbem-1048	86	46	into	into	ADP
fbem-1048	86	47	two	two	NUM
fbem-1048	86	48	categories	category	NOUN
fbem-1048	86	49	:	:	PUNCT
fbem-1048	86	50	(	(	PUNCT
fbem-1048	86	51	1	1	X
fbem-1048	86	52	)	)	PUNCT
fbem-1048	86	53	positive	positive	ADJ
fbem-1048	86	54	comment	comment	NOUN
fbem-1048	86	55	topics	topic	NOUN
fbem-1048	86	56	extracted	extract	VERB
fbem-1048	86	57	from	from	ADP
fbem-1048	86	58	positive	positive	ADJ
fbem-1048	86	59	comments	comment	NOUN
fbem-1048	86	60	;	;	PUNCT
fbem-1048	86	61	(	(	PUNCT
fbem-1048	86	62	2	2	X
fbem-1048	86	63	)	)	PUNCT
fbem-1048	86	64	negative	negative	ADJ
fbem-1048	86	65	comment	comment	NOUN
fbem-1048	86	66	topics	topic	NOUN
fbem-1048	86	67	extracted	extract	VERB
fbem-1048	86	68	from	from	ADP
fbem-1048	86	69	negative	negative	ADJ
fbem-1048	86	70	comments	comment	NOUN
fbem-1048	86	71	.	.	PUNCT
fbem-1048	87	1	through	through	ADP
fbem-1048	87	2	the	the	DET
fbem-1048	87	3	theme	theme	NOUN
fbem-1048	87	4	results	result	NOUN
fbem-1048	87	5	extracted	extract	VERB
fbem-1048	87	6	from	from	ADP
fbem-1048	87	7	the	the	DET
fbem-1048	87	8	experiment	experiment	NOUN
fbem-1048	87	9	,	,	PUNCT
fbem-1048	87	10	the	the	DET
fbem-1048	87	11	emotional	emotional	ADJ
fbem-1048	87	12	trend	trend	NOUN
fbem-1048	87	13	of	of	ADP
fbem-1048	87	14	each	each	DET
fbem-1048	87	15	user	user	NOUN
fbem-1048	87	16	for	for	ADP
fbem-1048	87	17	this	this	DET
fbem-1048	87	18	product	product	NOUN
fbem-1048	87	19	is	be	AUX
fbem-1048	87	20	analyzed	analyze	VERB
fbem-1048	87	21	,	,	PUNCT
fbem-1048	87	22	so	so	SCONJ
fbem-1048	87	23	as	as	SCONJ
fbem-1048	87	24	to	to	PART
fbem-1048	87	25	make	make	VERB
fbem-1048	87	26	a	a	DET
fbem-1048	87	27	more	more	ADV
fbem-1048	87	28	accurate	accurate	ADJ
fbem-1048	87	29	recommendation	recommendation	NOUN
fbem-1048	87	30	for	for	ADP
fbem-1048	87	31	the	the	DET
fbem-1048	87	32	next	next	ADJ
fbem-1048	87	33	step	step	NOUN
fbem-1048	87	34	according	accord	VERB
fbem-1048	87	35	to	to	ADP
fbem-1048	87	36	the	the	DET
fbem-1048	87	37	emotional	emotional	ADJ
fbem-1048	87	38	trend	trend	NOUN
fbem-1048	87	39	.	.	PUNCT
fbem-1048	88	1	table	table	NOUN
fbem-1048	88	2	1	1	NUM
fbem-1048	88	3	.	.	PUNCT
fbem-1048	89	1	experimental	experimental	ADJ
fbem-1048	89	2	results	result	NOUN
fbem-1048	89	3	of	of	ADP
fbem-1048	89	4	positive	positive	ADJ
fbem-1048	89	5	0	0	NUM
fbem-1048	89	6	1	1	NUM
fbem-1048	89	7	2	2	NUM
fbem-1048	89	8	3	3	NUM
fbem-1048	89	9	4	4	NUM
fbem-1048	89	10	5	5	NUM
fbem-1048	89	11	6	6	NUM
fbem-1048	89	12	7	7	NUM
fbem-1048	89	13	8	8	NUM
fbem-1048	89	14	9	9	NUM
fbem-1048	89	15	theme	theme	NOUN
fbem-1048	89	16	1	1	NUM
fbem-1048	89	17	very	very	ADV
fbem-1048	89	18	one	one	NUM
fbem-1048	89	19	machine	machine	NOUN
fbem-1048	89	20	run	run	NOUN
fbem-1048	89	21	computer	computer	NOUN
fbem-1048	89	22	totally	totally	ADV
fbem-1048	89	23	dell	dell	VERB
fbem-1048	89	24	very	very	ADV
fbem-1048	89	25	complete	complete	ADJ
fbem-1048	89	26	equipment	equipment	NOUN
fbem-1048	89	27	buy	buy	NOUN
fbem-1048	89	28	theme	theme	NOUN
fbem-1048	89	29	2	2	NUM
fbem-1048	89	30	very	very	ADV
fbem-1048	89	31	elegant	elegant	ADJ
fbem-1048	89	32	smooth	smooth	ADJ
fbem-1048	89	33	exterior	exterior	ADJ
fbem-1048	89	34	start	start	NOUN
fbem-1048	89	35	computer	computer	NOUN
fbem-1048	89	36	quick	quick	ADV
fbem-1048	89	37	worth	worth	ADJ
fbem-1048	89	38	one	one	NUM
fbem-1048	89	39	machine	machine	NOUN
fbem-1048	89	40	like	like	ADP
fbem-1048	89	41	theme	theme	NOUN
fbem-1048	89	42	3	3	NUM
fbem-1048	89	43	computer	computer	NOUN
fbem-1048	89	44	run	run	VERB
fbem-1048	89	45	special	special	ADJ
fbem-1048	89	46	very	very	ADV
fbem-1048	89	47	very	very	ADV
fbem-1048	89	48	speed	speed	NOUN
fbem-1048	89	49	screen	screen	NOUN
fbem-1048	89	50	very	very	ADV
fbem-1048	89	51	high	high	ADJ
fbem-1048	89	52	equipment	equipment	NOUN
fbem-1048	89	53	28	28	NUM
fbem-1048	89	54	table	table	NOUN
fbem-1048	89	55	2	2	NUM
fbem-1048	89	56	.	.	PUNCT
fbem-1048	89	57	experimental	experimental	ADJ
fbem-1048	89	58	results	result	NOUN
fbem-1048	89	59	of	of	ADP
fbem-1048	89	60	negative	negative	ADJ
fbem-1048	89	61	0	0	NUM
fbem-1048	89	62	1	1	NUM
fbem-1048	89	63	2	2	NUM
fbem-1048	89	64	3	3	NUM
fbem-1048	89	65	4	4	NUM
fbem-1048	89	66	5	5	NUM
fbem-1048	89	67	6	6	NUM
fbem-1048	89	68	7	7	NUM
fbem-1048	89	69	8	8	NUM
fbem-1048	89	70	9	9	NUM
fbem-1048	89	71	theme	theme	NOUN
fbem-1048	89	72	1	1	NUM
fbem-1048	89	73	twice	twice	ADV
fbem-1048	89	74	quality	quality	NOUN
fbem-1048	89	75	change	change	NOUN
fbem-1048	89	76	goods	good	NOUN
fbem-1048	89	77	problem	problem	NOUN
fbem-1048	89	78	a	a	DET
fbem-1048	89	79	little	little	ADJ
fbem-1048	89	80	carton	carton	NOUN
fbem-1048	89	81	shaking	shake	VERB
fbem-1048	89	82	tap	tap	NOUN
fbem-1048	89	83	bottom	bottom	ADJ
fbem-1048	89	84	theme	theme	NOUN
fbem-1048	89	85	2	2	NUM
fbem-1048	89	86	reaction	reaction	NOUN
fbem-1048	89	87	blue	blue	ADJ
fbem-1048	89	88	screen	screen	NOUN
fbem-1048	89	89	expensive	expensive	ADJ
fbem-1048	89	90	heavy	heavy	ADJ
fbem-1048	89	91	old	old	ADJ
fbem-1048	89	92	play	play	VERB
fbem-1048	89	93	exciting	exciting	ADJ
fbem-1048	89	94	splash	splash	NOUN
fbem-1048	89	95	screen	screen	NOUN
fbem-1048	89	96	battlefield	battlefield	NOUN
fbem-1048	89	97	none	none	NOUN
fbem-1048	89	98	theme	theme	NOUN
fbem-1048	89	99	3	3	NUM
fbem-1048	89	100	nothing	nothing	PRON
fbem-1048	89	101	after	after	ADP
fbem-1048	89	102	that	that	PRON
fbem-1048	89	103	do	do	AUX
fbem-1048	89	104	n’t	not	PART
fbem-1048	89	105	buy	buy	VERB
fbem-1048	89	106	all	all	DET
fbem-1048	89	107	not	not	PART
fbem-1048	89	108	good	good	ADJ
fbem-1048	89	109	knowledge	knowledge	NOUN
fbem-1048	89	110	computer	computer	NOUN
fbem-1048	89	111	game	game	NOUN
fbem-1048	89	112	shutdown	shutdown	NOUN
fbem-1048	89	113	high	high	ADJ
fbem-1048	89	114	useless	useless	ADJ
fbem-1048	89	115	for	for	ADP
fbem-1048	89	116	users	user	NOUN
fbem-1048	89	117	who	who	PRON
fbem-1048	89	118	are	be	AUX
fbem-1048	89	119	on	on	ADP
fbem-1048	89	120	the	the	DET
fbem-1048	89	121	positive	positive	ADJ
fbem-1048	89	122	side	side	NOUN
fbem-1048	89	123	,	,	PUNCT
fbem-1048	89	124	it	it	PRON
fbem-1048	89	125	can	can	AUX
fbem-1048	89	126	be	be	AUX
fbem-1048	89	127	found	find	VERB
fbem-1048	89	128	from	from	ADP
fbem-1048	89	129	the	the	DET
fbem-1048	89	130	keywords	keyword	NOUN
fbem-1048	89	131	that	that	SCONJ
fbem-1048	89	132	the	the	DET
fbem-1048	89	133	satisfaction	satisfaction	NOUN
fbem-1048	89	134	is	be	AUX
fbem-1048	89	135	very	very	ADV
fbem-1048	89	136	high	high	ADJ
fbem-1048	89	137	,	,	PUNCT
fbem-1048	89	138	and	and	CCONJ
fbem-1048	89	139	the	the	DET
fbem-1048	89	140	products	product	NOUN
fbem-1048	89	141	of	of	ADP
fbem-1048	89	142	the	the	DET
fbem-1048	89	143	same	same	ADJ
fbem-1048	89	144	category	category	NOUN
fbem-1048	89	145	can	can	AUX
fbem-1048	89	146	no	no	ADV
fbem-1048	89	147	longer	long	ADV
fbem-1048	89	148	be	be	AUX
fbem-1048	89	149	pushed	push	VERB
fbem-1048	89	150	,	,	PUNCT
fbem-1048	89	151	and	and	CCONJ
fbem-1048	89	152	can	can	AUX
fbem-1048	89	153	continue	continue	VERB
fbem-1048	89	154	to	to	PART
fbem-1048	89	155	recommend	recommend	VERB
fbem-1048	89	156	some	some	DET
fbem-1048	89	157	derivative	derivative	ADJ
fbem-1048	89	158	products	product	NOUN
fbem-1048	89	159	about	about	ADP
fbem-1048	89	160	the	the	DET
fbem-1048	89	161	product	product	NOUN
fbem-1048	89	162	,	,	PUNCT
fbem-1048	89	163	such	such	ADJ
fbem-1048	89	164	as	as	ADP
fbem-1048	89	165	peripherals	peripheral	NOUN
fbem-1048	89	166	,	,	PUNCT
fbem-1048	89	167	accessories	accessory	NOUN
fbem-1048	89	168	,	,	PUNCT
fbem-1048	89	169	etc	etc	X
fbem-1048	89	170	.	.	X
fbem-1048	89	171	for	for	ADP
fbem-1048	89	172	users	user	NOUN
fbem-1048	89	173	on	on	ADP
fbem-1048	89	174	the	the	DET
fbem-1048	89	175	negative	negative	ADJ
fbem-1048	89	176	side	side	NOUN
fbem-1048	89	177	,	,	PUNCT
fbem-1048	89	178	it	it	PRON
fbem-1048	89	179	can	can	AUX
fbem-1048	89	180	be	be	AUX
fbem-1048	89	181	found	find	VERB
fbem-1048	89	182	from	from	ADP
fbem-1048	89	183	the	the	DET
fbem-1048	89	184	subject	subject	ADJ
fbem-1048	89	185	words	word	NOUN
fbem-1048	89	186	that	that	PRON
fbem-1048	89	187	they	they	PRON
fbem-1048	89	188	are	be	AUX
fbem-1048	89	189	not	not	PART
fbem-1048	89	190	satisfied	satisfied	ADJ
fbem-1048	89	191	with	with	ADP
fbem-1048	89	192	the	the	DET
fbem-1048	89	193	product	product	NOUN
fbem-1048	89	194	.	.	PUNCT
fbem-1048	90	1	according	accord	VERB
fbem-1048	90	2	to	to	ADP
fbem-1048	90	3	the	the	DET
fbem-1048	90	4	main	main	ADJ
fbem-1048	90	5	dissatisfaction	dissatisfaction	NOUN
fbem-1048	90	6	problems	problem	NOUN
fbem-1048	90	7	raised	raise	VERB
fbem-1048	90	8	by	by	ADP
fbem-1048	90	9	the	the	DET
fbem-1048	90	10	negative	negative	ADJ
fbem-1048	90	11	comments	comment	NOUN
fbem-1048	90	12	,	,	PUNCT
fbem-1048	90	13	other	other	ADJ
fbem-1048	90	14	products	product	NOUN
fbem-1048	90	15	of	of	ADP
fbem-1048	90	16	the	the	DET
fbem-1048	90	17	same	same	ADJ
fbem-1048	90	18	category	category	NOUN
fbem-1048	90	19	and	and	CCONJ
fbem-1048	90	20	different	different	ADJ
fbem-1048	90	21	brands	brand	NOUN
fbem-1048	90	22	that	that	PRON
fbem-1048	90	23	meet	meet	VERB
fbem-1048	90	24	these	these	DET
fbem-1048	90	25	expectations	expectation	NOUN
fbem-1048	90	26	are	be	AUX
fbem-1048	90	27	recommended	recommend	VERB
fbem-1048	90	28	,	,	PUNCT
fbem-1048	90	29	such	such	ADJ
fbem-1048	90	30	as	as	ADP
fbem-1048	90	31	providing	provide	VERB
fbem-1048	90	32	higher	high	ADJ
fbem-1048	90	33	quality	quality	NOUN
fbem-1048	90	34	and	and	CCONJ
fbem-1048	90	35	more	more	ADV
fbem-1048	90	36	beautiful	beautiful	ADJ
fbem-1048	90	37	.	.	PUNCT
fbem-1048	91	1	product	product	NOUN
fbem-1048	91	2	push	push	VERB
fbem-1048	91	3	with	with	ADP
fbem-1048	91	4	better	well	ADJ
fbem-1048	91	5	performance	performance	NOUN
fbem-1048	91	6	.	.	PUNCT
fbem-1048	92	1	7	7	X
fbem-1048	92	2	.	.	X
fbem-1048	92	3	conclusion	conclusion	NOUN
fbem-1048	92	4	by	by	ADP
fbem-1048	92	5	using	use	VERB
fbem-1048	92	6	the	the	DET
fbem-1048	92	7	sentiment	sentiment	NOUN
fbem-1048	92	8	analysis	analysis	NOUN
fbem-1048	92	9	,	,	PUNCT
fbem-1048	92	10	the	the	DET
fbem-1048	92	11	efficiency	efficiency	NOUN
fbem-1048	92	12	and	and	CCONJ
fbem-1048	92	13	accuracy	accuracy	NOUN
fbem-1048	92	14	of	of	ADP
fbem-1048	92	15	the	the	DET
fbem-1048	92	16	recommendation	recommendation	NOUN
fbem-1048	92	17	can	can	AUX
fbem-1048	92	18	be	be	AUX
fbem-1048	92	19	greatly	greatly	ADV
fbem-1048	92	20	improved	improve	VERB
fbem-1048	92	21	,	,	PUNCT
fbem-1048	92	22	and	and	CCONJ
fbem-1048	92	23	a	a	DET
fbem-1048	92	24	large	large	ADJ
fbem-1048	92	25	number	number	NOUN
fbem-1048	92	26	of	of	ADP
fbem-1048	92	27	repeated	repeat	VERB
fbem-1048	92	28	products	product	NOUN
fbem-1048	92	29	can	can	AUX
fbem-1048	92	30	be	be	AUX
fbem-1048	92	31	avoided	avoid	VERB
fbem-1048	92	32	,	,	PUNCT
fbem-1048	92	33	thereby	thereby	ADV
fbem-1048	92	34	increasing	increase	VERB
fbem-1048	92	35	the	the	DET
fbem-1048	92	36	benefits	benefit	NOUN
fbem-1048	92	37	.	.	PUNCT
fbem-1048	93	1	references	reference	NOUN
fbem-1048	93	2	[	[	X
fbem-1048	93	3	1	1	X
fbem-1048	93	4	]	]	PUNCT
fbem-1048	93	5	nasukawa	nasukawa	PROPN
fbem-1048	93	6	t	t	PROPN
fbem-1048	93	7	&	&	CCONJ
fbem-1048	93	8	yi	yi	PROPN
fbem-1048	93	9	j.	j.	PROPN
fbem-1048	93	10	(	(	PUNCT
fbem-1048	93	11	2003	2003	NUM
fbem-1048	93	12	)	)	PUNCT
fbem-1048	93	13	sentiment	sentiment	NOUN
fbem-1048	93	14	analysis	analysis	NOUN
fbem-1048	93	15	:	:	PUNCT
fbem-1048	93	16	capturing	capture	VERB
fbem-1048	93	17	favorability	favorability	NOUN
fbem-1048	93	18	using	use	VERB
fbem-1048	93	19	natural	natural	ADJ
fbem-1048	93	20	language	language	NOUN
fbem-1048	93	21	processing	processing	NOUN
fbem-1048	93	22	.	.	PUNCT
fbem-1048	94	1	proc	proc	NOUN
fbem-1048	94	2	of	of	ADP
fbem-1048	94	3	the	the	DET
fbem-1048	94	4	2nd	2nd	ADJ
fbem-1048	94	5	international	international	ADJ
fbem-1048	94	6	conference	conference	NOUN
fbem-1048	94	7	on	on	ADP
fbem-1048	94	8	knowledge	knowledge	NOUN
fbem-1048	94	9	capture	capture	NOUN
fbem-1048	94	10	,	,	PUNCT
fbem-1048	94	11	70	70	NUM
fbem-1048	94	12	-	-	SYM
fbem-1048	94	13	77	77	NUM
fbem-1048	94	14	.	.	PUNCT
fbem-1048	95	1	[	[	X
fbem-1048	95	2	2	2	X
fbem-1048	95	3	]	]	PUNCT
fbem-1048	95	4	pang	pang	NOUN
fbem-1048	95	5	b	b	PROPN
fbem-1048	95	6	&	&	CCONJ
fbem-1048	95	7	lee	lee	PROPN
fbem-1048	95	8	l.	l.	PROPN
fbem-1048	95	9	(	(	PUNCT
fbem-1048	95	10	2008	2008	NUM
fbem-1048	95	11	)	)	PUNCT
fbem-1048	95	12	opinion	opinion	NOUN
fbem-1048	95	13	mining	mining	NOUN
fbem-1048	95	14	and	and	CCONJ
fbem-1048	95	15	sentiment	sentiment	NOUN
fbem-1048	95	16	analysis	analysis	NOUN
fbem-1048	95	17	.	.	PUNCT
fbem-1048	96	1	foundations	foundation	NOUN
fbem-1048	96	2	and	and	CCONJ
fbem-1048	96	3	trends	trend	NOUN
fbem-1048	96	4	in	in	ADP
fbem-1048	96	5	information	information	NOUN
fbem-1048	96	6	retrieval	retrieval	NOUN
fbem-1048	96	7	,	,	PUNCT
fbem-1048	96	8	2(1	2(1	ADV
fbem-1048	96	9	-	-	SYM
fbem-1048	96	10	2	2	NUM
fbem-1048	96	11	):	):	PUNCT
fbem-1048	96	12	1	1	NUM
fbem-1048	96	13	-	-	SYM
fbem-1048	96	14	135	135	NUM
fbem-1048	96	15	.	.	PUNCT
fbem-1048	97	1	[	[	X
fbem-1048	97	2	3	3	X
fbem-1048	97	3	]	]	X
fbem-1048	97	4	hu	hu	PROPN
fbem-1048	97	5	m	m	PROPN
fbem-1048	97	6	&	&	CCONJ
fbem-1048	97	7	liu	liu	PROPN
fbem-1048	97	8	b.	b.	PROPN
fbem-1048	97	9	(	(	PUNCT
fbem-1048	97	10	2004	2004	NUM
fbem-1048	97	11	)	)	PUNCT
fbem-1048	97	12	mining	mining	NOUN
fbem-1048	97	13	and	and	CCONJ
fbem-1048	97	14	summarizing	summarize	VERB
fbem-1048	97	15	customer	customer	NOUN
fbem-1048	97	16	reviews	review	NOUN
fbem-1048	97	17	.	.	PUNCT
fbem-1048	98	1	proceedings	proceeding	NOUN
fbem-1048	98	2	of	of	ADP
fbem-1048	98	3	the	the	DET
fbem-1048	98	4	tenth	tenth	ADJ
fbem-1048	98	5	acm	acm	NOUN
fbem-1048	98	6	sigkdd	sigkdd	NOUN
fbem-1048	98	7	international	international	ADJ
fbem-1048	98	8	conference	conference	NOUN
fbem-1048	98	9	on	on	ADP
fbem-1048	98	10	knowledge	knowledge	NOUN
fbem-1048	98	11	discovery	discovery	PROPN
fbem-1048	98	12	and	and	CCONJ
fbem-1048	98	13	data	datum	NOUN
fbem-1048	98	14	mining	mining	NOUN
fbem-1048	98	15	.	.	PUNCT
fbem-1048	99	1	acm	acm	PROPN
fbem-1048	99	2	,	,	PUNCT
fbem-1048	99	3	168	168	NUM
fbem-1048	99	4	-	-	SYM
fbem-1048	99	5	177	177	NUM
fbem-1048	99	6	.	.	PUNCT
fbem-1048	100	1	[	[	X
fbem-1048	100	2	4	4	X
fbem-1048	100	3	]	]	X
fbem-1048	100	4	ren	ren	PROPN
fbem-1048	100	5	f	f	PROPN
fbem-1048	100	6	&	&	CCONJ
fbem-1048	100	7	wu	wu	PROPN
fbem-1048	100	8	y.	y.	PROPN
fbem-1048	100	9	(	(	PUNCT
fbem-1048	100	10	2013	2013	NUM
fbem-1048	100	11	)	)	PUNCT
fbem-1048	100	12	predicting	predict	VERB
fbem-1048	100	13	user	user	NOUN
fbem-1048	100	14	-	-	PUNCT
fbem-1048	100	15	topic	topic	NOUN
fbem-1048	100	16	opinions	opinion	NOUN
fbem-1048	100	17	in	in	ADP
fbem-1048	100	18	twitter	twitter	NOUN
fbem-1048	100	19	with	with	ADP
fbem-1048	100	20	social	social	ADJ
fbem-1048	100	21	and	and	CCONJ
fbem-1048	100	22	topical	topical	ADJ
fbem-1048	100	23	context	context	NOUN
fbem-1048	100	24	.	.	PUNCT
fbem-1048	101	1	ieee	ieee	NOUN
fbem-1048	101	2	transactions	transaction	NOUN
fbem-1048	101	3	on	on	ADP
fbem-1048	101	4	affective	affective	ADJ
fbem-1048	101	5	computing	computing	NOUN
fbem-1048	101	6	,	,	PUNCT
fbem-1048	101	7	4(4	4(4	NUM
fbem-1048	101	8	):	):	PUNCT
fbem-1048	101	9	412	412	NUM
fbem-1048	101	10	-	-	SYM
fbem-1048	101	11	424	424	NUM
fbem-1048	101	12	.	.	PUNCT
fbem-1048	102	1	[	[	X
fbem-1048	102	2	5	5	X
fbem-1048	102	3	]	]	X
fbem-1048	102	4	hirschberg	hirschberg	PROPN
fbem-1048	102	5	j	j	PROPN
fbem-1048	102	6	&	&	CCONJ
fbem-1048	102	7	manning	manning	PROPN
fbem-1048	102	8	c	c	PROPN
fbem-1048	102	9	d.	d.	PROPN
fbem-1048	102	10	(	(	PUNCT
fbem-1048	102	11	2015	2015	NUM
fbem-1048	102	12	)	)	PUNCT
fbem-1048	102	13	advances	advance	NOUN
fbem-1048	102	14	in	in	ADP
fbem-1048	102	15	natural	natural	ADJ
fbem-1048	102	16	language	language	NOUN
fbem-1048	102	17	processing	processing	NOUN
fbem-1048	102	18	.	.	PUNCT
fbem-1048	103	1	science	science	NOUN
fbem-1048	103	2	,	,	PUNCT
fbem-1048	103	3	349(6245	349(6245	NUM
fbem-1048	103	4	):	):	PUNCT
fbem-1048	103	5	261	261	NUM
fbem-1048	103	6	-	-	SYM
fbem-1048	103	7	266	266	NUM
fbem-1048	103	8	.	.	PUNCT
fbem-1048	104	1	[	[	X
fbem-1048	104	2	6	6	X
fbem-1048	104	3	]	]	PUNCT
fbem-1048	104	4	aizawa	aizawa	PROPN
fbem-1048	104	5	a.	a.	PROPN
fbem-1048	104	6	(	(	PUNCT
fbem-1048	104	7	2003	2003	NUM
fbem-1048	104	8	)	)	PUNCT
fbem-1048	104	9	an	an	DET
fbem-1048	104	10	information	information	NOUN
fbem-1048	104	11	-	-	PUNCT
fbem-1048	104	12	theoretic	theoretic	NOUN
fbem-1048	104	13	perspective	perspective	NOUN
fbem-1048	104	14	of	of	ADP
fbem-1048	104	15	t	t	PROPN
fbem-1048	104	16	fidf	fidf	NOUN
fbem-1048	104	17	measures	measure	NOUN
fbem-1048	104	18	.	.	PUNCT
fbem-1048	105	1	information	information	NOUN
fbem-1048	105	2	processing	processing	NOUN
fbem-1048	105	3	&	&	CCONJ
fbem-1048	105	4	management	management	NOUN
fbem-1048	105	5	,	,	PUNCT
fbem-1048	105	6	39(1	39(1	NUM
fbem-1048	105	7	):	):	PUNCT
fbem-1048	105	8	45	45	NUM
fbem-1048	105	9	-	-	SYM
fbem-1048	105	10	65	65	NUM
fbem-1048	105	11	.	.	PUNCT
fbem-1048	106	1	[	[	X
fbem-1048	106	2	7	7	X
fbem-1048	106	3	]	]	X
fbem-1048	106	4	le	le	X
fbem-1048	106	5	q	q	PROPN
fbem-1048	106	6	&	&	CCONJ
fbem-1048	106	7	mikolov	mikolov	PROPN
fbem-1048	106	8	t.	t.	PROPN
fbem-1048	106	9	(	(	PUNCT
fbem-1048	106	10	2014	2014	NUM
fbem-1048	106	11	)	)	PUNCT
fbem-1048	106	12	distributed	distribute	VERB
fbem-1048	106	13	representations	representation	NOUN
fbem-1048	106	14	of	of	ADP
fbem-1048	106	15	sentences	sentence	NOUN
fbem-1048	106	16	and	and	CCONJ
fbem-1048	106	17	documents	document	NOUN
fbem-1048	106	18	.	.	PUNCT
fbem-1048	107	1	proc	proc	NOUN
fbem-1048	107	2	of	of	ADP
fbem-1048	107	3	the	the	DET
fbem-1048	107	4	31st	31st	ADJ
fbem-1048	107	5	international	international	ADJ
fbem-1048	107	6	conference	conference	NOUN
fbem-1048	107	7	on	on	ADP
fbem-1048	107	8	machine	machine	NOUN
fbem-1048	107	9	learning	learning	NOUN
fbem-1048	107	10	,	,	PUNCT
fbem-1048	107	11	1188	1188	NUM
fbem-1048	107	12	-	-	SYM
fbem-1048	107	13	1196	1196	NUM
fbem-1048	107	14	.	.	PUNCT
fbem-1048	108	1	[	[	X
fbem-1048	108	2	8	8	NUM
fbem-1048	108	3	]	]	X
fbem-1048	108	4	mikolov	mikolov	PROPN
fbem-1048	108	5	t	t	PROPN
fbem-1048	108	6	,	,	PUNCT
fbem-1048	108	7	chen	chen	PROPN
fbem-1048	108	8	k	k	PROPN
fbem-1048	108	9	,	,	PUNCT
fbem-1048	108	10	corrado	corrado	PROPN
fbem-1048	108	11	g	g	PROPN
fbem-1048	108	12	&	&	CCONJ
fbem-1048	108	13	et	et	PROPN
fbem-1048	108	14	al	al	PROPN
fbem-1048	108	15	.	.	PROPN
fbem-1048	109	1	(	(	PUNCT
fbem-1048	109	2	2014	2014	NUM
fbem-1048	109	3	)	)	PUNCT
fbem-1048	109	4	efficient	efficient	ADJ
fbem-1048	109	5	estimation	estimation	NOUN
fbem-1048	109	6	of	of	ADP
fbem-1048	109	7	word	word	NOUN
fbem-1048	109	8	representations	representation	NOUN
fbem-1048	109	9	in	in	ADP
fbem-1048	109	10	vector	vector	NOUN
fbem-1048	109	11	space	space	NOUN
fbem-1048	109	12	.	.	PUNCT
fbem-1048	110	1	arxiv	arxiv	NOUN
fbem-1048	110	2	:	:	PUNCT
fbem-1048	111	1	1301,3781	1301,3781	X
fbem-1048	111	2	.	.	PUNCT
fbem-1048	112	1	[	[	X
fbem-1048	112	2	9	9	NUM
fbem-1048	112	3	]	]	SYM
fbem-1048	112	4	li	li	PROPN
fbem-1048	112	5	rui	rui	PROPN
fbem-1048	112	6	,	,	PUNCT
fbem-1048	112	7	zhang	zhang	PROPN
fbem-1048	112	8	qian	qian	PROPN
fbem-1048	112	9	&	&	CCONJ
fbem-1048	112	10	liu	liu	PROPN
fbem-1048	112	11	jia	jia	PROPN
fbem-1048	112	12	-	-	PROPN
fbem-1048	112	13	yong	yong	PROPN
fbem-1048	112	14	.	.	PUNCT
fbem-1048	112	15	(	(	PUNCT
fbem-1048	112	16	2017	2017	NUM
fbem-1048	112	17	)	)	PUNCT
fbem-1048	112	18	microblog	microblog	NOUN
fbem-1048	112	19	sentiment	sentiment	NOUN
fbem-1048	112	20	analysis	analysis	NOUN
fbem-1048	112	21	based	base	VERB
fbem-1048	112	22	on	on	ADP
fbem-1048	112	23	weighted	weighted	ADJ
fbem-1048	112	24	word2vec	word2vec	PROPN
fbem-1048	112	25	.	.	PUNCT
fbem-1048	113	1	communications	communication	NOUN
fbem-1048	113	2	technology	technology	NOUN
fbem-1048	113	3	,	,	PUNCT
fbem-1048	113	4	50(3	50(3	NUM
fbem-1048	113	5	):	):	PUNCT
fbem-1048	113	6	502	502	NUM
fbem-1048	113	7	-	-	SYM
fbem-1048	113	8	506	506	NUM
fbem-1048	113	9	.	.	PUNCT