id	sid	tid	token	lemma	pos
ajst-24984	1	1	academic	academic	ADJ
ajst-24984	1	2	journal	journal	NOUN
ajst-24984	1	3	of	of	ADP
ajst-24984	1	4	science	science	NOUN
ajst-24984	1	5	and	and	CCONJ
ajst-24984	1	6	technology	technology	NOUN
ajst-24984	1	7	issn	issn	NOUN
ajst-24984	1	8	:	:	PUNCT
ajst-24984	1	9	2771	2771	NUM
ajst-24984	1	10	-	-	SYM
ajst-24984	1	11	3032	3032	NUM
ajst-24984	1	12	|	|	NOUN
ajst-24984	1	13	vol	vol	NOUN
ajst-24984	1	14	.	.	PROPN
ajst-24984	2	1	12	12	NUM
ajst-24984	2	2	,	,	PUNCT
ajst-24984	2	3	no	no	INTJ
ajst-24984	2	4	.	.	NOUN
ajst-24984	2	5	1	1	NUM
ajst-24984	2	6	,	,	PUNCT
ajst-24984	2	7	2024	2024	NUM
ajst-24984	2	8	192	192	NUM
ajst-24984	2	9	aspect	aspect	NOUN
ajst-24984	2	10	-	-	PUNCT
ajst-24984	2	11	based	base	VERB
ajst-24984	2	12	sentiment	sentiment	NOUN
ajst-24984	2	13	analysis	analysis	NOUN
ajst-24984	2	14	based	base	VERB
ajst-24984	2	15	on	on	ADP
ajst-24984	2	16	feature	feature	NOUN
ajst-24984	2	17	extraction	extraction	NOUN
ajst-24984	2	18	and	and	CCONJ
ajst-24984	2	19	attention	attention	NOUN
ajst-24984	2	20	bo	bo	VERB
ajst-24984	2	21	he	he	PRON
ajst-24984	2	22	a	a	PRON
ajst-24984	2	23	,	,	PUNCT
ajst-24984	2	24	ruoyu	ruoyu	ADJ
ajst-24984	2	25	zhao	zhao	PROPN
ajst-24984	2	26	*	*	PUNCT
ajst-24984	2	27	,	,	PUNCT
ajst-24984	2	28	lei	lei	PROPN
ajst-24984	2	29	wang	wang	PROPN
ajst-24984	2	30	b	b	PROPN
ajst-24984	2	31	,	,	PUNCT
ajst-24984	2	32	yongfen	yongfen	PROPN
ajst-24984	2	33	yang	yang	PROPN
ajst-24984	2	34	c	c	PROPN
ajst-24984	2	35	college	college	PROPN
ajst-24984	2	36	of	of	ADP
ajst-24984	2	37	computer	computer	NOUN
ajst-24984	2	38	science	science	NOUN
ajst-24984	2	39	and	and	CCONJ
ajst-24984	2	40	engineering	engineering	NOUN
ajst-24984	2	41	,	,	PUNCT
ajst-24984	2	42	chongqing	chongqing	PROPN
ajst-24984	2	43	university	university	PROPN
ajst-24984	2	44	of	of	ADP
ajst-24984	2	45	technology	technology	NOUN
ajst-24984	2	46	,	,	PUNCT
ajst-24984	2	47	chongqing	chongqe	VERB
ajst-24984	2	48	40054	40054	NUM
ajst-24984	2	49	,	,	PUNCT
ajst-24984	2	50	china	china	PROPN
ajst-24984	2	51	*	*	PUNCT
ajst-24984	2	52	corresponding	correspond	VERB
ajst-24984	2	53	author	author	NOUN
ajst-24984	2	54	:	:	PUNCT
ajst-24984	2	55	ruoyu	ruoyu	ADJ
ajst-24984	2	56	zhao	zhao	NOUN
ajst-24984	2	57	(	(	PUNCT
ajst-24984	2	58	email	email	NOUN
ajst-24984	2	59	:	:	PUNCT
ajst-24984	2	60	qiaqiaeric@outlook.com	qiaqiaeric@outlook.com	X
ajst-24984	2	61	)	)	PUNCT
ajst-24984	2	62	,	,	PUNCT
ajst-24984	2	63	a	a	DET
ajst-24984	2	64	hebo@cqut.edu.cn	hebo@cqut.edu.cn	PROPN
ajst-24984	2	65	,	,	PUNCT
ajst-24984	2	66	b	b	X
ajst-24984	2	67	wanglei6731@163.com	wanglei6731@163.com	X
ajst-24984	2	68	,	,	PUNCT
ajst-24984	2	69	c	c	PROPN
ajst-24984	2	70	yangff0518@163.com	yangff0518@163.com	X
ajst-24984	2	71	abstract	abstract	NOUN
ajst-24984	2	72	:	:	PUNCT
ajst-24984	2	73	aspect	aspect	NOUN
ajst-24984	2	74	-	-	PUNCT
ajst-24984	2	75	based	base	VERB
ajst-24984	2	76	sentiment	sentiment	NOUN
ajst-24984	2	77	analysis	analysis	NOUN
ajst-24984	2	78	(	(	PUNCT
ajst-24984	2	79	absa	absa	X
ajst-24984	2	80	)	)	PUNCT
ajst-24984	2	81	is	be	AUX
ajst-24984	2	82	a	a	DET
ajst-24984	2	83	fine	fine	ADV
ajst-24984	2	84	-	-	PUNCT
ajst-24984	2	85	grained	grain	VERB
ajst-24984	2	86	sentiment	sentiment	NOUN
ajst-24984	2	87	analysis	analysis	NOUN
ajst-24984	2	88	task	task	NOUN
ajst-24984	2	89	whose	whose	DET
ajst-24984	2	90	goal	goal	NOUN
ajst-24984	2	91	is	be	AUX
ajst-24984	2	92	to	to	PART
ajst-24984	2	93	analyze	analyze	VERB
ajst-24984	2	94	the	the	DET
ajst-24984	2	95	sentiment	sentiment	NOUN
ajst-24984	2	96	polarity	polarity	NOUN
ajst-24984	2	97	of	of	ADP
ajst-24984	2	98	specific	specific	ADJ
ajst-24984	2	99	aspects	aspect	NOUN
ajst-24984	2	100	of	of	ADP
ajst-24984	2	101	a	a	DET
ajst-24984	2	102	text	text	NOUN
ajst-24984	2	103	.	.	PUNCT
ajst-24984	3	1	currently	currently	ADV
ajst-24984	3	2	,	,	PUNCT
ajst-24984	3	3	aspect	aspect	NOUN
ajst-24984	3	4	-	-	PUNCT
ajst-24984	3	5	based	base	VERB
ajst-24984	3	6	sentiment	sentiment	NOUN
ajst-24984	3	7	analysis	analysis	NOUN
ajst-24984	3	8	is	be	AUX
ajst-24984	3	9	widely	widely	ADV
ajst-24984	3	10	used	use	VERB
ajst-24984	3	11	in	in	ADP
ajst-24984	3	12	opinion	opinion	NOUN
ajst-24984	3	13	analysis	analysis	NOUN
ajst-24984	3	14	and	and	CCONJ
ajst-24984	3	15	content	content	NOUN
ajst-24984	3	16	recommendation	recommendation	NOUN
ajst-24984	3	17	.	.	PUNCT
ajst-24984	4	1	in	in	ADP
ajst-24984	4	2	this	this	DET
ajst-24984	4	3	paper	paper	NOUN
ajst-24984	4	4	,	,	PUNCT
ajst-24984	4	5	we	we	PRON
ajst-24984	4	6	provide	provide	VERB
ajst-24984	4	7	an	an	DET
ajst-24984	4	8	overview	overview	NOUN
ajst-24984	4	9	of	of	ADP
ajst-24984	4	10	aspect	aspect	NOUN
ajst-24984	4	11	-	-	PUNCT
ajst-24984	4	12	based	base	VERB
ajst-24984	4	13	sentiment	sentiment	NOUN
ajst-24984	4	14	analysis	analysis	NOUN
ajst-24984	4	15	methods	method	NOUN
ajst-24984	4	16	such	such	ADJ
ajst-24984	4	17	as	as	ADP
ajst-24984	4	18	sentiment	sentiment	NOUN
ajst-24984	4	19	dictionary	dictionary	NOUN
ajst-24984	4	20	,	,	PUNCT
ajst-24984	4	21	machine	machine	NOUN
ajst-24984	4	22	learning	learning	NOUN
ajst-24984	4	23	,	,	PUNCT
ajst-24984	4	24	and	and	CCONJ
ajst-24984	4	25	deep	deep	ADJ
ajst-24984	4	26	learning	learning	NOUN
ajst-24984	4	27	,	,	PUNCT
ajst-24984	4	28	introduce	introduce	VERB
ajst-24984	4	29	their	their	PRON
ajst-24984	4	30	latest	late	ADJ
ajst-24984	4	31	research	research	NOUN
ajst-24984	4	32	results	result	NOUN
ajst-24984	4	33	,	,	PUNCT
ajst-24984	4	34	and	and	CCONJ
ajst-24984	4	35	analyze	analyze	VERB
ajst-24984	4	36	the	the	DET
ajst-24984	4	37	advantages	advantage	NOUN
ajst-24984	4	38	and	and	CCONJ
ajst-24984	4	39	disadvantages	disadvantage	NOUN
ajst-24984	4	40	of	of	ADP
ajst-24984	4	41	using	use	VERB
ajst-24984	4	42	different	different	ADJ
ajst-24984	4	43	methods	method	NOUN
ajst-24984	4	44	.	.	PUNCT
ajst-24984	5	1	the	the	DET
ajst-24984	5	2	current	current	ADJ
ajst-24984	5	3	status	status	NOUN
ajst-24984	5	4	of	of	ADP
ajst-24984	5	5	research	research	NOUN
ajst-24984	5	6	on	on	ADP
ajst-24984	5	7	aspect	aspect	NOUN
ajst-24984	5	8	-	-	PUNCT
ajst-24984	5	9	based	base	VERB
ajst-24984	5	10	sentiment	sentiment	NOUN
ajst-24984	5	11	analysis	analysis	NOUN
ajst-24984	5	12	at	at	ADP
ajst-24984	5	13	home	home	NOUN
ajst-24984	5	14	and	and	CCONJ
ajst-24984	5	15	abroad	abroad	ADV
ajst-24984	5	16	is	be	AUX
ajst-24984	5	17	analyzed	analyze	VERB
ajst-24984	5	18	in	in	ADP
ajst-24984	5	19	depth	depth	NOUN
ajst-24984	5	20	by	by	ADP
ajst-24984	5	21	combing	comb	VERB
ajst-24984	5	22	this	this	DET
ajst-24984	5	23	class	class	NOUN
ajst-24984	5	24	of	of	ADP
ajst-24984	5	25	methods	method	NOUN
ajst-24984	5	26	,	,	PUNCT
ajst-24984	5	27	and	and	CCONJ
ajst-24984	5	28	the	the	DET
ajst-24984	5	29	future	future	ADJ
ajst-24984	5	30	development	development	NOUN
ajst-24984	5	31	and	and	CCONJ
ajst-24984	5	32	trends	trend	NOUN
ajst-24984	5	33	of	of	ADP
ajst-24984	5	34	textual	textual	ADJ
ajst-24984	5	35	aspect	aspect	NOUN
ajst-24984	5	36	-	-	PUNCT
ajst-24984	5	37	based	base	VERB
ajst-24984	5	38	sentiment	sentiment	NOUN
ajst-24984	5	39	analysis	analysis	NOUN
ajst-24984	5	40	are	be	AUX
ajst-24984	5	41	proposed	propose	VERB
ajst-24984	5	42	.	.	PUNCT
ajst-24984	6	1	keywords	keyword	NOUN
ajst-24984	6	2	:	:	PUNCT
ajst-24984	6	3	aspect	aspect	NOUN
ajst-24984	6	4	-	-	PUNCT
ajst-24984	6	5	based	base	VERB
ajst-24984	6	6	sentiment	sentiment	NOUN
ajst-24984	6	7	analysis	analysis	NOUN
ajst-24984	6	8	;	;	PUNCT
ajst-24984	6	9	natural	natural	ADJ
ajst-24984	6	10	language	language	NOUN
ajst-24984	6	11	processing	processing	NOUN
ajst-24984	6	12	;	;	PUNCT
ajst-24984	6	13	sentiment	sentiment	NOUN
ajst-24984	6	14	dictionary	dictionary	NOUN
ajst-24984	6	15	;	;	PUNCT
ajst-24984	6	16	machine	machine	NOUN
ajst-24984	6	17	learning	learning	NOUN
ajst-24984	6	18	;	;	PUNCT
ajst-24984	6	19	deep	deep	ADJ
ajst-24984	6	20	learning	learning	NOUN
ajst-24984	6	21	.	.	PUNCT
ajst-24984	7	1	1	1	X
ajst-24984	7	2	.	.	X
ajst-24984	7	3	introduction	introduction	NOUN
ajst-24984	7	4	the	the	DET
ajst-24984	7	5	internet	internet	NOUN
ajst-24984	7	6	is	be	AUX
ajst-24984	7	7	growing	grow	VERB
ajst-24984	7	8	by	by	ADP
ajst-24984	7	9	leaps	leap	NOUN
ajst-24984	7	10	and	and	CCONJ
ajst-24984	7	11	bounds	bound	NOUN
ajst-24984	7	12	in	in	ADP
ajst-24984	7	13	the	the	DET
ajst-24984	7	14	digital	digital	ADJ
ajst-24984	7	15	age	age	NOUN
ajst-24984	7	16	and	and	CCONJ
ajst-24984	7	17	internet	internet	NOUN
ajst-24984	7	18	users	user	NOUN
ajst-24984	7	19	are	be	AUX
ajst-24984	7	20	increasing	increase	VERB
ajst-24984	7	21	at	at	ADP
ajst-24984	7	22	a	a	DET
ajst-24984	7	23	breakneck	breakneck	NOUN
ajst-24984	7	24	pace	pace	NOUN
ajst-24984	7	25	,	,	PUNCT
ajst-24984	7	26	with	with	ADP
ajst-24984	7	27	around	around	ADP
ajst-24984	7	28	900,000	900,000	NUM
ajst-24984	7	29	people	people	NOUN
ajst-24984	7	30	going	go	VERB
ajst-24984	7	31	online	online	ADV
ajst-24984	7	32	for	for	ADP
ajst-24984	7	33	the	the	DET
ajst-24984	7	34	first	first	ADJ
ajst-24984	7	35	time	time	NOUN
ajst-24984	7	36	every	every	DET
ajst-24984	7	37	day	day	NOUN
ajst-24984	7	38	since	since	SCONJ
ajst-24984	7	39	2018	2018	NUM
ajst-24984	7	40	.	.	PUNCT
ajst-24984	8	1	currently	currently	ADV
ajst-24984	8	2	,	,	PUNCT
ajst-24984	8	3	the	the	DET
ajst-24984	8	4	internet	internet	NOUN
ajst-24984	8	5	has	have	AUX
ajst-24984	8	6	long	long	ADV
ajst-24984	8	7	been	be	AUX
ajst-24984	8	8	used	use	VERB
ajst-24984	8	9	by	by	ADP
ajst-24984	8	10	more	more	ADJ
ajst-24984	8	11	than	than	ADP
ajst-24984	8	12	5.3	5.3	NUM
ajst-24984	8	13	billion	billion	NUM
ajst-24984	8	14	people	people	NOUN
ajst-24984	8	15	worldwide	worldwide	ADV
ajst-24984	8	16	,	,	PUNCT
ajst-24984	8	17	equivalent	equivalent	ADJ
ajst-24984	8	18	to	to	ADP
ajst-24984	8	19	66	66	NUM
ajst-24984	8	20	%	%	NOUN
ajst-24984	8	21	of	of	ADP
ajst-24984	8	22	the	the	DET
ajst-24984	8	23	world	world	NOUN
ajst-24984	8	24	's	's	PART
ajst-24984	8	25	population	population	NOUN
ajst-24984	8	26	,	,	PUNCT
ajst-24984	8	27	and	and	CCONJ
ajst-24984	8	28	this	this	DET
ajst-24984	8	29	number	number	NOUN
ajst-24984	8	30	is	be	AUX
ajst-24984	8	31	expected	expect	VERB
ajst-24984	8	32	to	to	PART
ajst-24984	8	33	top	top	VERB
ajst-24984	8	34	6.54	6.54	NUM
ajst-24984	8	35	billion	billion	NUM
ajst-24984	8	36	in	in	ADP
ajst-24984	8	37	2025	2025	NUM
ajst-24984	8	38	.	.	PUNCT
ajst-24984	9	1	along	along	ADP
ajst-24984	9	2	with	with	ADP
ajst-24984	9	3	the	the	DET
ajst-24984	9	4	increase	increase	NOUN
ajst-24984	9	5	in	in	ADP
ajst-24984	9	6	internet	internet	NOUN
ajst-24984	9	7	usage	usage	NOUN
ajst-24984	9	8	comes	come	VERB
ajst-24984	9	9	data	datum	NOUN
ajst-24984	9	10	:	:	PUNCT
ajst-24984	9	11	idc	idc	PROPN
ajst-24984	9	12	expects	expect	VERB
ajst-24984	9	13	to	to	PART
ajst-24984	9	14	generate	generate	VERB
ajst-24984	9	15	159.2	159.2	NUM
ajst-24984	9	16	zettabytes	zettabyte	NOUN
ajst-24984	9	17	of	of	ADP
ajst-24984	9	18	data	datum	NOUN
ajst-24984	9	19	in	in	ADP
ajst-24984	9	20	2024	2024	NUM
ajst-24984	9	21	and	and	CCONJ
ajst-24984	9	22	384.6	384.6	NUM
ajst-24984	9	23	zettabytes	zettabyte	NOUN
ajst-24984	9	24	in	in	ADP
ajst-24984	9	25	one	one	NUM
ajst-24984	9	26	year	year	NOUN
ajst-24984	9	27	by	by	ADP
ajst-24984	9	28	2028	2028	NUM
ajst-24984	9	29	.	.	PUNCT
ajst-24984	10	1	this	this	DET
ajst-24984	10	2	huge	huge	ADJ
ajst-24984	10	3	amount	amount	NOUN
ajst-24984	10	4	of	of	ADP
ajst-24984	10	5	data	datum	NOUN
ajst-24984	10	6	has	have	AUX
ajst-24984	10	7	greatly	greatly	ADV
ajst-24984	10	8	accelerated	accelerate	VERB
ajst-24984	10	9	the	the	DET
ajst-24984	10	10	development	development	NOUN
ajst-24984	10	11	of	of	ADP
ajst-24984	10	12	fields	field	NOUN
ajst-24984	10	13	such	such	ADJ
ajst-24984	10	14	as	as	ADP
ajst-24984	10	15	natural	natural	ADJ
ajst-24984	10	16	language	language	NOUN
ajst-24984	10	17	processing	processing	NOUN
ajst-24984	10	18	(	(	PUNCT
ajst-24984	10	19	nlp	nlp	NOUN
ajst-24984	10	20	)	)	PUNCT
ajst-24984	10	21	and	and	CCONJ
ajst-24984	10	22	computer	computer	NOUN
ajst-24984	10	23	vision	vision	NOUN
ajst-24984	10	24	(	(	PUNCT
ajst-24984	10	25	cv	cv	PROPN
ajst-24984	10	26	)	)	PUNCT
ajst-24984	10	27	.	.	PUNCT
ajst-24984	11	1	natural	natural	ADJ
ajst-24984	11	2	language	language	NOUN
ajst-24984	11	3	processing	processing	NOUN
ajst-24984	11	4	(	(	PUNCT
ajst-24984	11	5	nlp	nlp	NOUN
ajst-24984	11	6	)	)	PUNCT
ajst-24984	11	7	is	be	AUX
ajst-24984	11	8	a	a	DET
ajst-24984	11	9	branch	branch	NOUN
ajst-24984	11	10	of	of	ADP
ajst-24984	11	11	artificial	artificial	ADJ
ajst-24984	11	12	intelligence	intelligence	NOUN
ajst-24984	11	13	that	that	PRON
ajst-24984	11	14	deals	deal	VERB
ajst-24984	11	15	with	with	ADP
ajst-24984	11	16	the	the	DET
ajst-24984	11	17	design	design	NOUN
ajst-24984	11	18	and	and	CCONJ
ajst-24984	11	19	implementation	implementation	NOUN
ajst-24984	11	20	of	of	ADP
ajst-24984	11	21	systems	system	NOUN
ajst-24984	11	22	and	and	CCONJ
ajst-24984	11	23	algorithms	algorithm	NOUN
ajst-24984	11	24	capable	capable	ADJ
ajst-24984	11	25	of	of	ADP
ajst-24984	11	26	interacting	interact	VERB
ajst-24984	11	27	through	through	ADP
ajst-24984	11	28	human	human	ADJ
ajst-24984	11	29	language	language	NOUN
ajst-24984	11	30	[	[	X
ajst-24984	11	31	1	1	NUM
ajst-24984	11	32	]	]	PUNCT
ajst-24984	11	33	.	.	PUNCT
ajst-24984	12	1	in	in	ADP
ajst-24984	12	2	the	the	DET
ajst-24984	12	3	field	field	NOUN
ajst-24984	12	4	of	of	ADP
ajst-24984	12	5	nlp	nlp	NOUN
ajst-24984	12	6	,	,	PUNCT
ajst-24984	12	7	there	there	PRON
ajst-24984	12	8	has	have	AUX
ajst-24984	12	9	been	be	AUX
ajst-24984	12	10	continuous	continuous	ADJ
ajst-24984	12	11	research	research	NOUN
ajst-24984	12	12	on	on	ADP
ajst-24984	12	13	sentimen	sentiman	NOUN
ajst-24984	12	14	analysis	analysis	NOUN
ajst-24984	12	15	.	.	PUNCT
ajst-24984	13	1	sentiment	sentiment	NOUN
ajst-24984	13	2	analysis	analysis	NOUN
ajst-24984	13	3	(	(	PUNCT
ajst-24984	13	4	sa	sa	NOUN
ajst-24984	13	5	)	)	PUNCT
ajst-24984	13	6	is	be	AUX
ajst-24984	13	7	mainly	mainly	ADV
ajst-24984	13	8	categorized	categorize	VERB
ajst-24984	13	9	into	into	ADP
ajst-24984	13	10	chapterlevel	chapterlevel	ADJ
ajst-24984	13	11	sentiment	sentiment	NOUN
ajst-24984	13	12	analysis	analysis	NOUN
ajst-24984	13	13	,	,	PUNCT
ajst-24984	13	14	sentence	sentence	NOUN
ajst-24984	13	15	-	-	PUNCT
ajst-24984	13	16	level	level	NOUN
ajst-24984	13	17	sentiment	sentiment	NOUN
ajst-24984	13	18	analysis	analysis	NOUN
ajst-24984	13	19	and	and	CCONJ
ajst-24984	13	20	aspect	aspect	NOUN
ajst-24984	13	21	-	-	PUNCT
ajst-24984	13	22	based	base	VERB
ajst-24984	13	23	sentiment	sentiment	NOUN
ajst-24984	13	24	analysis	analysis	NOUN
ajst-24984	13	25	(	(	PUNCT
ajst-24984	13	26	absa	absa	X
ajst-24984	13	27	)	)	PUNCT
ajst-24984	14	1	[	[	X
ajst-24984	14	2	2	2	NUM
ajst-24984	14	3	]	]	PUNCT
ajst-24984	14	4	.	.	PUNCT
ajst-24984	15	1	chapterlevel	chapterlevel	PROPN
ajst-24984	15	2	sentiment	sentiment	NOUN
ajst-24984	15	3	analysis	analysis	NOUN
ajst-24984	15	4	analyzes	analyze	VERB
ajst-24984	15	5	the	the	DET
ajst-24984	15	6	whole	whole	ADJ
ajst-24984	15	7	text	text	NOUN
ajst-24984	15	8	in	in	ADP
ajst-24984	15	9	general	general	ADJ
ajst-24984	15	10	,	,	PUNCT
ajst-24984	15	11	while	while	SCONJ
ajst-24984	15	12	sentence	sentence	NOUN
ajst-24984	15	13	-	-	PUNCT
ajst-24984	15	14	level	level	NOUN
ajst-24984	15	15	sentiment	sentiment	NOUN
ajst-24984	15	16	analysis	analysis	NOUN
ajst-24984	15	17	analyzes	analyze	VERB
ajst-24984	15	18	the	the	DET
ajst-24984	15	19	whole	whole	ADJ
ajst-24984	15	20	sentence	sentence	NOUN
ajst-24984	15	21	as	as	ADP
ajst-24984	15	22	a	a	DET
ajst-24984	15	23	basic	basic	ADJ
ajst-24984	15	24	unit	unit	NOUN
ajst-24984	15	25	,	,	PUNCT
ajst-24984	15	26	and	and	CCONJ
ajst-24984	15	27	both	both	PRON
ajst-24984	15	28	of	of	ADP
ajst-24984	15	29	them	they	PRON
ajst-24984	15	30	can	can	AUX
ajst-24984	15	31	only	only	ADV
ajst-24984	15	32	analyze	analyze	VERB
ajst-24984	15	33	the	the	DET
ajst-24984	15	34	overall	overall	ADJ
ajst-24984	15	35	sentiment	sentiment	NOUN
ajst-24984	15	36	of	of	ADP
ajst-24984	15	37	a	a	DET
ajst-24984	15	38	text	text	NOUN
ajst-24984	15	39	or	or	CCONJ
ajst-24984	15	40	sentence	sentence	NOUN
ajst-24984	15	41	,	,	PUNCT
ajst-24984	15	42	not	not	PART
ajst-24984	15	43	the	the	DET
ajst-24984	15	44	sentiment	sentiment	NOUN
ajst-24984	15	45	of	of	ADP
ajst-24984	15	46	multiple	multiple	ADJ
ajst-24984	15	47	aspects	aspect	NOUN
ajst-24984	15	48	,	,	PUNCT
ajst-24984	15	49	so	so	SCONJ
ajst-24984	15	50	they	they	PRON
ajst-24984	15	51	are	be	AUX
ajst-24984	15	52	classified	classify	VERB
ajst-24984	15	53	as	as	ADP
ajst-24984	15	54	coarse	coarse	NOUN
ajst-24984	15	55	-	-	PUNCT
ajst-24984	15	56	grained	grain	VERB
ajst-24984	15	57	sentiment	sentiment	NOUN
ajst-24984	15	58	analysis	analysis	NOUN
ajst-24984	15	59	.	.	PUNCT
ajst-24984	16	1	aspect	aspect	NOUN
ajst-24984	16	2	-	-	PUNCT
ajst-24984	16	3	based	base	VERB
ajst-24984	16	4	sentiment	sentiment	NOUN
ajst-24984	16	5	analysis	analysis	NOUN
ajst-24984	16	6	is	be	AUX
ajst-24984	16	7	a	a	DET
ajst-24984	16	8	fine	fine	ADV
ajst-24984	16	9	-	-	PUNCT
ajst-24984	16	10	grained	grain	VERB
ajst-24984	16	11	sentiment	sentiment	NOUN
ajst-24984	16	12	analysis	analysis	NOUN
ajst-24984	16	13	[	[	X
ajst-24984	16	14	3	3	X
ajst-24984	16	15	]	]	PUNCT
ajst-24984	16	16	that	that	PRON
ajst-24984	16	17	analyzes	analyze	VERB
ajst-24984	16	18	the	the	DET
ajst-24984	16	19	sentiment	sentiment	NOUN
ajst-24984	16	20	of	of	ADP
ajst-24984	16	21	specific	specific	ADJ
ajst-24984	16	22	aspects	aspect	NOUN
ajst-24984	16	23	of	of	ADP
ajst-24984	16	24	a	a	DET
ajst-24984	16	25	sentence	sentence	NOUN
ajst-24984	16	26	rather	rather	ADV
ajst-24984	16	27	than	than	ADP
ajst-24984	16	28	the	the	DET
ajst-24984	16	29	overall	overall	ADJ
ajst-24984	16	30	sentiment	sentiment	NOUN
ajst-24984	16	31	.	.	PUNCT
ajst-24984	17	1	for	for	ADP
ajst-24984	17	2	example	example	NOUN
ajst-24984	17	3	,	,	PUNCT
ajst-24984	17	4	"	"	PUNCT
ajst-24984	17	5	this	this	DET
ajst-24984	17	6	video	video	NOUN
ajst-24984	17	7	card	card	NOUN
ajst-24984	17	8	works	work	VERB
ajst-24984	17	9	great	great	ADJ
ajst-24984	17	10	,	,	PUNCT
ajst-24984	17	11	but	but	CCONJ
ajst-24984	17	12	the	the	DET
ajst-24984	17	13	video	video	NOUN
ajst-24984	17	14	memory	memory	NOUN
ajst-24984	17	15	is	be	AUX
ajst-24984	17	16	too	too	ADV
ajst-24984	17	17	small	small	ADJ
ajst-24984	17	18	.	.	PUNCT
ajst-24984	17	19	"	"	PUNCT
ajst-24984	18	1	(	(	PUNCT
ajst-24984	18	2	fig.1	fig.1	PROPN
ajst-24984	18	3	)	)	PUNCT
ajst-24984	18	4	this	this	DET
ajst-24984	18	5	sentence	sentence	NOUN
ajst-24984	18	6	has	have	VERB
ajst-24984	18	7	two	two	NUM
ajst-24984	18	8	opposite	opposite	ADJ
ajst-24984	18	9	sentiment	sentiment	NOUN
ajst-24984	18	10	polarities	polarity	NOUN
ajst-24984	18	11	,	,	PUNCT
ajst-24984	18	12	positive	positive	ADJ
ajst-24984	18	13	and	and	CCONJ
ajst-24984	18	14	negative	negative	ADJ
ajst-24984	18	15	,	,	PUNCT
ajst-24984	18	16	with	with	ADP
ajst-24984	18	17	the	the	DET
ajst-24984	18	18	sentiment	sentiment	NOUN
ajst-24984	18	19	polarity	polarity	NOUN
ajst-24984	18	20	of	of	ADP
ajst-24984	18	21	"	"	PUNCT
ajst-24984	18	22	graphics	graphic	NOUN
ajst-24984	18	23	card	card	NOUN
ajst-24984	18	24	"	"	PUNCT
ajst-24984	18	25	being	be	AUX
ajst-24984	18	26	positive	positive	ADJ
ajst-24984	18	27	and	and	CCONJ
ajst-24984	18	28	the	the	DET
ajst-24984	18	29	sentiment	sentiment	NOUN
ajst-24984	18	30	polarity	polarity	NOUN
ajst-24984	18	31	of	of	ADP
ajst-24984	18	32	"	"	PUNCT
ajst-24984	18	33	video	video	NOUN
ajst-24984	18	34	memory	memory	NOUN
ajst-24984	18	35	"	"	PUNCT
ajst-24984	18	36	being	be	AUX
ajst-24984	18	37	negative	negative	ADJ
ajst-24984	18	38	.	.	PUNCT
ajst-24984	19	1	aspect	aspect	NOUN
ajst-24984	19	2	-	-	PUNCT
ajst-24984	19	3	based	base	VERB
ajst-24984	19	4	sentiment	sentiment	NOUN
ajst-24984	19	5	analysis	analysis	NOUN
ajst-24984	19	6	and	and	CCONJ
ajst-24984	19	7	sentence	sentence	NOUN
ajst-24984	19	8	-	-	PUNCT
ajst-24984	19	9	level	level	NOUN
ajst-24984	19	10	sentiment	sentiment	NOUN
ajst-24984	19	11	analysis	analysis	NOUN
ajst-24984	19	12	are	be	AUX
ajst-24984	19	13	difficult	difficult	ADJ
ajst-24984	19	14	to	to	PART
ajst-24984	19	15	accurately	accurately	ADV
ajst-24984	19	16	analyze	analyze	VERB
ajst-24984	19	17	the	the	DET
ajst-24984	19	18	sentiments	sentiment	NOUN
ajst-24984	19	19	of	of	ADP
ajst-24984	19	20	the	the	DET
ajst-24984	19	21	text	text	NOUN
ajst-24984	19	22	.	.	PUNCT
ajst-24984	20	1	therefore	therefore	ADV
ajst-24984	20	2	,	,	PUNCT
ajst-24984	20	3	traditional	traditional	ADJ
ajst-24984	20	4	coarse	coarse	NOUN
ajst-24984	20	5	-	-	PUNCT
ajst-24984	20	6	grained	grain	VERB
ajst-24984	20	7	sentiment	sentiment	NOUN
ajst-24984	20	8	analysis	analysis	NOUN
ajst-24984	20	9	methods	method	NOUN
ajst-24984	20	10	can	can	AUX
ajst-24984	20	11	no	no	ADV
ajst-24984	20	12	longer	long	ADV
ajst-24984	20	13	meet	meet	VERB
ajst-24984	20	14	the	the	DET
ajst-24984	20	15	needs	need	NOUN
ajst-24984	20	16	of	of	ADP
ajst-24984	20	17	sentiment	sentiment	NOUN
ajst-24984	20	18	analysis	analysis	NOUN
ajst-24984	20	19	in	in	ADP
ajst-24984	20	20	today	today	NOUN
ajst-24984	20	21	's	's	PART
ajst-24984	20	22	complex	complex	ADJ
ajst-24984	20	23	semantic	semantic	ADJ
ajst-24984	20	24	environment	environment	NOUN
ajst-24984	20	25	.	.	PUNCT
ajst-24984	21	1	finer	fine	ADJ
ajst-24984	21	2	-	-	PUNCT
ajst-24984	21	3	grained	grain	VERB
ajst-24984	21	4	sentiment	sentiment	NOUN
ajst-24984	21	5	analysis	analysis	NOUN
ajst-24984	21	6	methods	method	NOUN
ajst-24984	21	7	have	have	AUX
ajst-24984	21	8	become	become	VERB
ajst-24984	21	9	a	a	DET
ajst-24984	21	10	key	key	ADJ
ajst-24984	21	11	research	research	NOUN
ajst-24984	21	12	issue	issue	NOUN
ajst-24984	21	13	in	in	ADP
ajst-24984	21	14	natural	natural	ADJ
ajst-24984	21	15	language	language	NOUN
ajst-24984	21	16	processing	processing	NOUN
ajst-24984	21	17	[	[	X
ajst-24984	21	18	4	4	NUM
ajst-24984	21	19	]	]	PUNCT
ajst-24984	21	20	.	.	PUNCT
ajst-24984	22	1	fig	fig	PROPN
ajst-24984	22	2	1	1	NUM
ajst-24984	22	3	.	.	PUNCT
ajst-24984	23	1	example	example	NOUN
ajst-24984	23	2	of	of	ADP
ajst-24984	23	3	a	a	DET
ajst-24984	23	4	multifaceted	multifaceted	ADJ
ajst-24984	23	5	,	,	PUNCT
ajst-24984	23	6	multilevel	multilevel	ADJ
ajst-24984	23	7	sentence	sentence	NOUN
ajst-24984	23	8	2	2	NUM
ajst-24984	23	9	.	.	PUNCT
ajst-24984	24	1	development	development	NOUN
ajst-24984	24	2	of	of	ADP
ajst-24984	24	3	aspect	aspect	NOUN
ajst-24984	24	4	-	-	PUNCT
ajst-24984	24	5	based	base	VERB
ajst-24984	24	6	sentiment	sentiment	NOUN
ajst-24984	24	7	analysis	analysis	NOUN
ajst-24984	24	8	aspect	aspect	NOUN
ajst-24984	24	9	-	-	PUNCT
ajst-24984	24	10	based	base	VERB
ajst-24984	24	11	sentiment	sentiment	NOUN
ajst-24984	24	12	analysis	analysis	NOUN
ajst-24984	24	13	(	(	PUNCT
ajst-24984	24	14	absa	absa	X
ajst-24984	24	15	)	)	PUNCT
ajst-24984	24	16	is	be	AUX
ajst-24984	24	17	a	a	DET
ajst-24984	24	18	fine	fine	ADV
ajst-24984	24	19	-	-	PUNCT
ajst-24984	24	20	grained	grain	VERB
ajst-24984	24	21	sentiment	sentiment	NOUN
ajst-24984	24	22	analysis	analysis	NOUN
ajst-24984	24	23	task	task	NOUN
ajst-24984	24	24	that	that	PRON
ajst-24984	24	25	can	can	AUX
ajst-24984	24	26	deal	deal	VERB
ajst-24984	24	27	with	with	ADP
ajst-24984	24	28	texts	text	NOUN
ajst-24984	24	29	that	that	PRON
ajst-24984	24	30	contain	contain	VERB
ajst-24984	24	31	many	many	ADJ
ajst-24984	24	32	sentiment	sentiment	NOUN
ajst-24984	24	33	polarities	polarity	NOUN
ajst-24984	24	34	and	and	CCONJ
ajst-24984	24	35	are	be	AUX
ajst-24984	24	36	complex	complex	ADJ
ajst-24984	24	37	,	,	PUNCT
ajst-24984	24	38	and	and	CCONJ
ajst-24984	24	39	it	it	PRON
ajst-24984	24	40	contains	contain	VERB
ajst-24984	24	41	several	several	ADJ
ajst-24984	24	42	subtasks	subtask	NOUN
ajst-24984	24	43	such	such	ADJ
ajst-24984	24	44	as	as	ADP
ajst-24984	24	45	aspect	aspect	NOUN
ajst-24984	24	46	sentiment	sentiment	NOUN
ajst-24984	24	47	classification	classification	NOUN
ajst-24984	24	48	and	and	CCONJ
ajst-24984	24	49	aspect	aspect	NOUN
ajst-24984	24	50	extraction	extraction	NOUN
ajst-24984	24	51	[	[	X
ajst-24984	24	52	5][6][7	5][6][7	NUM
ajst-24984	24	53	]	]	X
ajst-24984	24	54	.	.	PUNCT
ajst-24984	25	1	the	the	DET
ajst-24984	25	2	aspect	aspect	NOUN
ajst-24984	25	3	sentiment	sentiment	NOUN
ajst-24984	25	4	classification	classification	NOUN
ajst-24984	25	5	task	task	NOUN
ajst-24984	25	6	refers	refer	VERB
ajst-24984	25	7	to	to	ADP
ajst-24984	25	8	classifying	classify	VERB
ajst-24984	25	9	the	the	DET
ajst-24984	25	10	sentiment	sentiment	NOUN
ajst-24984	25	11	polarity	polarity	NOUN
ajst-24984	25	12	of	of	ADP
ajst-24984	25	13	target	target	NOUN
ajst-24984	25	14	aspect	aspect	NOUN
ajst-24984	25	15	words	word	NOUN
ajst-24984	25	16	in	in	ADP
ajst-24984	25	17	the	the	DET
ajst-24984	25	18	text	text	NOUN
ajst-24984	25	19	.	.	PUNCT
ajst-24984	26	1	for	for	ADP
ajst-24984	26	2	example	example	NOUN
ajst-24984	26	3	,	,	PUNCT
ajst-24984	26	4	"	"	PUNCT
ajst-24984	26	5	graphics	graphics	NOUN
ajst-24984	26	6	card	card	NOUN
ajst-24984	26	7	"	"	PUNCT
ajst-24984	26	8	and	and	CCONJ
ajst-24984	26	9	"	"	PUNCT
ajst-24984	26	10	memory	memory	NOUN
ajst-24984	26	11	"	"	PUNCT
ajst-24984	26	12	in	in	ADP
ajst-24984	26	13	fig.1	fig.1	PROPN
ajst-24984	26	14	are	be	AUX
ajst-24984	26	15	the	the	DET
ajst-24984	26	16	aspect	aspect	NOUN
ajst-24984	26	17	words	word	NOUN
ajst-24984	26	18	in	in	ADP
ajst-24984	26	19	this	this	DET
ajst-24984	26	20	line	line	NOUN
ajst-24984	26	21	of	of	ADP
ajst-24984	26	22	text	text	NOUN
ajst-24984	26	23	,	,	PUNCT
ajst-24984	26	24	and	and	CCONJ
ajst-24984	26	25	in	in	ADP
ajst-24984	26	26	this	this	DET
ajst-24984	26	27	sentence	sentence	NOUN
ajst-24984	26	28	,	,	PUNCT
ajst-24984	26	29	the	the	DET
ajst-24984	26	30	polarity	polarity	NOUN
ajst-24984	26	31	of	of	ADP
ajst-24984	26	32	these	these	DET
ajst-24984	26	33	two	two	NUM
ajst-24984	26	34	words	word	NOUN
ajst-24984	26	35	is	be	AUX
ajst-24984	26	36	judged	judge	VERB
ajst-24984	26	37	as	as	ADP
ajst-24984	26	38	positive	positive	ADJ
ajst-24984	26	39	and	and	CCONJ
ajst-24984	26	40	negative	negative	ADJ
ajst-24984	26	41	respectively	respectively	ADV
ajst-24984	26	42	,	,	PUNCT
ajst-24984	26	43	while	while	SCONJ
ajst-24984	26	44	the	the	DET
ajst-24984	26	45	aspect	aspect	NOUN
ajst-24984	26	46	extraction	extraction	NOUN
ajst-24984	26	47	task	task	NOUN
ajst-24984	26	48	is	be	AUX
ajst-24984	26	49	a	a	DET
ajst-24984	26	50	part	part	NOUN
ajst-24984	26	51	of	of	ADP
ajst-24984	26	52	the	the	DET
ajst-24984	26	53	aspect	aspect	NOUN
ajst-24984	26	54	sentiment	sentiment	NOUN
ajst-24984	26	55	classification	classification	NOUN
ajst-24984	26	56	task	task	NOUN
ajst-24984	26	57	.	.	PUNCT
ajst-24984	27	1	task	task	NOUN
ajst-24984	27	2	is	be	AUX
ajst-24984	27	3	the	the	DET
ajst-24984	27	4	basis	basis	NOUN
ajst-24984	27	5	of	of	ADP
ajst-24984	27	6	the	the	DET
ajst-24984	27	7	aspect	aspect	NOUN
ajst-24984	27	8	sentiment	sentiment	NOUN
ajst-24984	27	9	categorization	categorization	NOUN
ajst-24984	27	10	task	task	NOUN
ajst-24984	27	11	,	,	PUNCT
ajst-24984	27	12	and	and	CCONJ
ajst-24984	27	13	all	all	PRON
ajst-24984	27	14	it	it	PRON
ajst-24984	27	15	has	have	VERB
ajst-24984	27	16	to	to	PART
ajst-24984	27	17	do	do	VERB
ajst-24984	27	18	is	be	AUX
ajst-24984	27	19	to	to	PART
ajst-24984	27	20	extract	extract	VERB
ajst-24984	27	21	all	all	DET
ajst-24984	27	22	the	the	DET
ajst-24984	27	23	aspect	aspect	NOUN
ajst-24984	27	24	words	word	NOUN
ajst-24984	27	25	in	in	ADP
ajst-24984	27	26	the	the	DET
ajst-24984	27	27	given	give	VERB
ajst-24984	27	28	text	text	NOUN
ajst-24984	27	29	.	.	PUNCT
ajst-24984	28	1	aspect	aspect	NOUN
ajst-24984	28	2	-	-	PUNCT
ajst-24984	28	3	based	base	VERB
ajst-24984	28	4	sentiment	sentiment	NOUN
ajst-24984	28	5	analysis	analysis	NOUN
ajst-24984	28	6	subtasks	subtask	NOUN
ajst-24984	28	7	and	and	CCONJ
ajst-24984	28	8	examples	example	NOUN
ajst-24984	28	9	are	be	AUX
ajst-24984	28	10	shown	show	VERB
ajst-24984	28	11	in	in	ADP
ajst-24984	28	12	fig.2	fig.2	PROPN
ajst-24984	28	13	.	.	PUNCT
ajst-24984	29	1	this	this	DET
ajst-24984	29	2	paper	paper	NOUN
ajst-24984	29	3	focuses	focus	VERB
ajst-24984	29	4	on	on	ADP
ajst-24984	29	5	the	the	DET
ajst-24984	29	6	task	task	NOUN
ajst-24984	29	7	of	of	ADP
ajst-24984	29	8	aspectual	aspectual	ADJ
ajst-24984	29	9	sentiment	sentiment	NOUN
ajst-24984	29	10	categorization	categorization	NOUN
ajst-24984	29	11	.	.	PUNCT
ajst-24984	30	1	early	early	ADJ
ajst-24984	30	2	research	research	NOUN
ajst-24984	30	3	on	on	ADP
ajst-24984	30	4	aspect	aspect	NOUN
ajst-24984	30	5	-	-	PUNCT
ajst-24984	30	6	based	base	VERB
ajst-24984	30	7	sentiment	sentiment	NOUN
ajst-24984	30	8	analysis	analysis	NOUN
ajst-24984	30	9	of	of	ADP
ajst-24984	30	10	text	text	NOUN
ajst-24984	30	11	mainly	mainly	ADV
ajst-24984	30	12	used	use	VERB
ajst-24984	30	13	methods	method	NOUN
ajst-24984	30	14	such	such	ADJ
ajst-24984	30	15	as	as	ADP
ajst-24984	30	16	machine	machine	NOUN
ajst-24984	30	17	learning	learning	NOUN
ajst-24984	30	18	-	-	PUNCT
ajst-24984	30	19	based	base	VERB
ajst-24984	30	20	and	and	CCONJ
ajst-24984	30	21	sentiment	sentiment	NOUN
ajst-24984	30	22	dictionary	dictionary	NOUN
ajst-24984	30	23	-	-	PUNCT
ajst-24984	30	24	based	base	VERB
ajst-24984	30	25	because	because	SCONJ
ajst-24984	30	26	of	of	ADP
ajst-24984	30	27	their	their	PRON
ajst-24984	30	28	ability	ability	NOUN
ajst-24984	30	29	to	to	PART
ajst-24984	30	30	capture	capture	VERB
ajst-24984	30	31	feature	feature	NOUN
ajst-24984	30	32	representations	representation	NOUN
ajst-24984	30	33	with	with	ADP
ajst-24984	30	34	relative	relative	ADJ
ajst-24984	30	35	ease	ease	NOUN
ajst-24984	30	36	[	[	X
ajst-24984	30	37	8][9	8][9	NOUN
ajst-24984	30	38	]	]	X
ajst-24984	30	39	.	.	PUNCT
ajst-24984	31	1	however	however	ADV
ajst-24984	31	2	,	,	PUNCT
ajst-24984	31	3	the	the	DET
ajst-24984	31	4	input	input	NOUN
ajst-24984	31	5	text	text	NOUN
ajst-24984	31	6	must	must	AUX
ajst-24984	31	7	undergo	undergo	VERB
ajst-24984	31	8	extensive	extensive	ADJ
ajst-24984	31	9	preprocessing	preprocessing	NOUN
ajst-24984	31	10	as	as	ADV
ajst-24984	31	11	well	well	ADV
ajst-24984	31	12	as	as	ADP
ajst-24984	31	13	complex	complex	ADJ
ajst-24984	31	14	feature	feature	NOUN
ajst-24984	31	15	engineering	engineering	NOUN
ajst-24984	31	16	before	before	ADP
ajst-24984	31	17	using	use	VERB
ajst-24984	31	18	these	these	DET
ajst-24984	31	19	methods	method	NOUN
ajst-24984	31	20	,	,	PUNCT
ajst-24984	31	21	and	and	CCONJ
ajst-24984	31	22	the	the	DET
ajst-24984	31	23	effectiveness	effectiveness	NOUN
ajst-24984	31	24	of	of	ADP
ajst-24984	31	25	the	the	DET
ajst-24984	31	26	results	result	NOUN
ajst-24984	31	27	they	they	PRON
ajst-24984	31	28	obtain	obtain	VERB
ajst-24984	31	29	often	often	ADV
ajst-24984	31	30	depends	depend	VERB
ajst-24984	31	31	on	on	ADP
ajst-24984	31	32	the	the	DET
ajst-24984	31	33	quality	quality	NOUN
ajst-24984	31	34	of	of	ADP
ajst-24984	31	35	the	the	DET
ajst-24984	31	36	manually	manually	ADV
ajst-24984	31	37	labeled	label	VERB
ajst-24984	31	38	data	datum	NOUN
ajst-24984	31	39	,	,	PUNCT
ajst-24984	31	40	which	which	PRON
ajst-24984	31	41	requires	require	VERB
ajst-24984	31	42	a	a	DET
ajst-24984	31	43	large	large	ADJ
ajst-24984	31	44	investment	investment	NOUN
ajst-24984	31	45	in	in	ADP
ajst-24984	31	46	labor	labor	NOUN
ajst-24984	31	47	costs	cost	NOUN
ajst-24984	31	48	when	when	SCONJ
ajst-24984	31	49	dealing	deal	VERB
ajst-24984	31	50	with	with	ADP
ajst-24984	31	51	the	the	DET
ajst-24984	31	52	massive	massive	ADJ
ajst-24984	31	53	193	193	NUM
ajst-24984	31	54	amount	amount	NOUN
ajst-24984	31	55	of	of	ADP
ajst-24984	31	56	text	text	NOUN
ajst-24984	31	57	data	datum	NOUN
ajst-24984	31	58	nowadays	nowadays	ADV
ajst-24984	31	59	.	.	PUNCT
ajst-24984	32	1	in	in	ADP
ajst-24984	32	2	addition	addition	NOUN
ajst-24984	32	3	,	,	PUNCT
ajst-24984	32	4	these	these	DET
ajst-24984	32	5	models	model	NOUN
ajst-24984	32	6	do	do	VERB
ajst-24984	32	7	not	not	PART
ajst-24984	32	8	or	or	CCONJ
ajst-24984	32	9	less	less	ADV
ajst-24984	32	10	take	take	VERB
ajst-24984	32	11	into	into	ADP
ajst-24984	32	12	account	account	NOUN
ajst-24984	32	13	the	the	DET
ajst-24984	32	14	semantic	semantic	ADJ
ajst-24984	32	15	relations	relation	NOUN
ajst-24984	32	16	within	within	ADP
ajst-24984	32	17	the	the	DET
ajst-24984	32	18	utterance	utterance	ADJ
ajst-24984	32	19	,	,	PUNCT
ajst-24984	32	20	contextual	contextual	ADJ
ajst-24984	32	21	relations	relation	NOUN
ajst-24984	32	22	and	and	CCONJ
ajst-24984	32	23	other	other	ADJ
ajst-24984	32	24	elements	element	NOUN
ajst-24984	32	25	.	.	PUNCT
ajst-24984	33	1	fig	fig	NOUN
ajst-24984	33	2	2	2	NUM
ajst-24984	33	3	.	.	PUNCT
ajst-24984	33	4	aspect	aspect	NOUN
ajst-24984	33	5	extraction	extraction	NOUN
ajst-24984	33	6	and	and	CCONJ
ajst-24984	33	7	aspect	aspect	NOUN
ajst-24984	33	8	sentiment	sentiment	NOUN
ajst-24984	33	9	classification	classification	NOUN
ajst-24984	33	10	with	with	ADP
ajst-24984	33	11	the	the	DET
ajst-24984	33	12	development	development	NOUN
ajst-24984	33	13	of	of	ADP
ajst-24984	33	14	deep	deep	ADJ
ajst-24984	33	15	learning	learning	NOUN
ajst-24984	33	16	,	,	PUNCT
ajst-24984	33	17	neural	neural	ADJ
ajst-24984	33	18	networks	network	NOUN
ajst-24984	33	19	are	be	AUX
ajst-24984	33	20	increasingly	increasingly	ADV
ajst-24984	33	21	used	use	VERB
ajst-24984	33	22	in	in	ADP
ajst-24984	33	23	natural	natural	ADJ
ajst-24984	33	24	language	language	NOUN
ajst-24984	33	25	processing	processing	NOUN
ajst-24984	33	26	(	(	PUNCT
ajst-24984	33	27	nlp	nlp	NOUN
ajst-24984	33	28	)	)	PUNCT
ajst-24984	33	29	tasks	task	NOUN
ajst-24984	33	30	[	[	X
ajst-24984	33	31	10	10	NUM
ajst-24984	33	32	]	]	PUNCT
ajst-24984	33	33	.	.	PUNCT
ajst-24984	34	1	deep	deep	ADJ
ajst-24984	34	2	learning	learning	NOUN
ajst-24984	34	3	based	base	VERB
ajst-24984	34	4	neural	neural	ADJ
ajst-24984	34	5	networks	network	NOUN
ajst-24984	34	6	are	be	AUX
ajst-24984	34	7	able	able	ADJ
ajst-24984	34	8	to	to	PART
ajst-24984	34	9	convert	convert	VERB
ajst-24984	34	10	text	text	NOUN
ajst-24984	34	11	into	into	ADP
ajst-24984	34	12	entity	entity	NOUN
ajst-24984	34	13	vectors	vector	NOUN
ajst-24984	34	14	using	use	VERB
ajst-24984	34	15	word	word	NOUN
ajst-24984	34	16	embeddings	embedding	NOUN
ajst-24984	34	17	thus	thus	ADV
ajst-24984	34	18	overcoming	overcome	VERB
ajst-24984	34	19	the	the	DET
ajst-24984	34	20	problem	problem	NOUN
ajst-24984	34	21	of	of	ADP
ajst-24984	34	22	traditional	traditional	ADJ
ajst-24984	34	23	machine	machine	NOUN
ajst-24984	34	24	learning	learning	NOUN
ajst-24984	34	25	that	that	PRON
ajst-24984	34	26	relies	rely	VERB
ajst-24984	34	27	on	on	ADP
ajst-24984	34	28	manual	manual	ADJ
ajst-24984	34	29	extraction	extraction	NOUN
ajst-24984	34	30	of	of	ADP
ajst-24984	34	31	text	text	NOUN
ajst-24984	34	32	features	feature	NOUN
ajst-24984	34	33	.	.	PUNCT
ajst-24984	35	1	recurrent	recurrent	ADJ
ajst-24984	35	2	neural	neural	ADJ
ajst-24984	35	3	networks	network	NOUN
ajst-24984	35	4	(	(	PUNCT
ajst-24984	35	5	rnn	rnn	PROPN
ajst-24984	35	6	)	)	PUNCT
ajst-24984	35	7	and	and	CCONJ
ajst-24984	35	8	attention	attention	NOUN
ajst-24984	35	9	mechanisms	mechanism	NOUN
ajst-24984	35	10	have	have	AUX
ajst-24984	35	11	been	be	AUX
ajst-24984	35	12	widely	widely	ADV
ajst-24984	35	13	introduced	introduce	VERB
ajst-24984	35	14	for	for	ADP
ajst-24984	35	15	aspect	aspect	NOUN
ajst-24984	35	16	-	-	PUNCT
ajst-24984	35	17	based	base	VERB
ajst-24984	35	18	sentiment	sentiment	NOUN
ajst-24984	35	19	analysis	analysis	NOUN
ajst-24984	35	20	,	,	PUNCT
ajst-24984	35	21	these	these	DET
ajst-24984	35	22	methods	method	NOUN
ajst-24984	35	23	have	have	VERB
ajst-24984	35	24	the	the	DET
ajst-24984	35	25	advantage	advantage	NOUN
ajst-24984	35	26	of	of	ADP
ajst-24984	35	27	being	be	AUX
ajst-24984	35	28	able	able	ADJ
ajst-24984	35	29	to	to	PART
ajst-24984	35	30	capture	capture	VERB
ajst-24984	35	31	sequential	sequential	ADJ
ajst-24984	35	32	and	and	CCONJ
ajst-24984	35	33	contextual	contextual	ADJ
ajst-24984	35	34	semantic	semantic	ADJ
ajst-24984	35	35	information	information	NOUN
ajst-24984	35	36	of	of	ADP
ajst-24984	35	37	the	the	DET
ajst-24984	35	38	text	text	NOUN
ajst-24984	36	1	[	[	X
ajst-24984	36	2	11	11	NUM
ajst-24984	36	3	]	]	PUNCT
ajst-24984	36	4	and	and	CCONJ
ajst-24984	36	5	are	be	AUX
ajst-24984	36	6	therefore	therefore	ADV
ajst-24984	36	7	suitable	suitable	ADJ
ajst-24984	36	8	for	for	ADP
ajst-24984	36	9	sentiment	sentiment	NOUN
ajst-24984	36	10	analysis	analysis	NOUN
ajst-24984	36	11	.	.	PUNCT
ajst-24984	37	1	however	however	ADV
ajst-24984	37	2	,	,	PUNCT
ajst-24984	37	3	when	when	SCONJ
ajst-24984	37	4	multiple	multiple	ADJ
ajst-24984	37	5	sentiment	sentiment	NOUN
ajst-24984	37	6	information	information	NOUN
ajst-24984	37	7	is	be	AUX
ajst-24984	37	8	in	in	ADP
ajst-24984	37	9	the	the	DET
ajst-24984	37	10	same	same	ADJ
ajst-24984	37	11	text	text	NOUN
ajst-24984	37	12	,	,	PUNCT
ajst-24984	37	13	aspect	aspect	NOUN
ajst-24984	37	14	-	-	PUNCT
ajst-24984	37	15	based	base	VERB
ajst-24984	37	16	sentiment	sentiment	NOUN
ajst-24984	37	17	analysis	analysis	NOUN
ajst-24984	37	18	requires	require	VERB
ajst-24984	37	19	deeper	deep	ADJ
ajst-24984	37	20	and	and	CCONJ
ajst-24984	37	21	more	more	ADV
ajst-24984	37	22	subtle	subtle	ADJ
ajst-24984	37	23	sentiment	sentiment	NOUN
ajst-24984	37	24	information	information	NOUN
ajst-24984	37	25	than	than	ADP
ajst-24984	37	26	sentence	sentence	NOUN
ajst-24984	37	27	-	-	PUNCT
ajst-24984	37	28	level	level	NOUN
ajst-24984	37	29	or	or	CCONJ
ajst-24984	37	30	chapterlevel	chapterlevel	ADJ
ajst-24984	37	31	sentiment	sentiment	NOUN
ajst-24984	37	32	analysis	analysis	NOUN
ajst-24984	37	33	[	[	X
ajst-24984	37	34	12	12	NUM
ajst-24984	37	35	]	]	PUNCT
ajst-24984	37	36	.	.	PUNCT
ajst-24984	38	1	in	in	ADP
ajst-24984	38	2	this	this	DET
ajst-24984	38	3	case	case	NOUN
ajst-24984	38	4	,	,	PUNCT
ajst-24984	38	5	such	such	ADJ
ajst-24984	38	6	models	model	NOUN
ajst-24984	38	7	,	,	PUNCT
ajst-24984	38	8	which	which	PRON
ajst-24984	38	9	are	be	AUX
ajst-24984	38	10	difficult	difficult	ADJ
ajst-24984	38	11	to	to	PART
ajst-24984	38	12	fully	fully	ADV
ajst-24984	38	13	utilize	utilize	VERB
ajst-24984	38	14	contextual	contextual	ADJ
ajst-24984	38	15	and	and	CCONJ
ajst-24984	38	16	inter	inter	ADJ
ajst-24984	38	17	-	-	ADJ
ajst-24984	38	18	word	word	ADJ
ajst-24984	38	19	syntactic	syntactic	ADJ
ajst-24984	38	20	dependencies	dependency	NOUN
ajst-24984	38	21	in	in	ADP
ajst-24984	38	22	sentences	sentence	NOUN
ajst-24984	38	23	,	,	PUNCT
ajst-24984	38	24	are	be	AUX
ajst-24984	38	25	less	less	ADV
ajst-24984	38	26	adequate	adequate	ADJ
ajst-24984	38	27	.	.	PUNCT
ajst-24984	39	1	in	in	ADP
ajst-24984	39	2	this	this	DET
ajst-24984	39	3	context	context	NOUN
ajst-24984	39	4	,	,	PUNCT
ajst-24984	39	5	using	use	VERB
ajst-24984	39	6	pre	pre	ADJ
ajst-24984	39	7	-	-	ADJ
ajst-24984	39	8	training	training	ADJ
ajst-24984	39	9	models	model	NOUN
ajst-24984	39	10	to	to	PART
ajst-24984	39	11	enhance	enhance	VERB
ajst-24984	39	12	the	the	DET
ajst-24984	39	13	semantic	semantic	ADJ
ajst-24984	39	14	representation	representation	NOUN
ajst-24984	39	15	and	and	CCONJ
ajst-24984	39	16	thus	thus	ADV
ajst-24984	39	17	deepen	deepen	VERB
ajst-24984	39	18	the	the	DET
ajst-24984	39	19	model	model	NOUN
ajst-24984	39	20	's	's	PART
ajst-24984	39	21	learning	learning	NOUN
ajst-24984	39	22	of	of	ADP
ajst-24984	39	23	the	the	DET
ajst-24984	39	24	syntactic	syntactic	ADJ
ajst-24984	39	25	dependencies	dependency	NOUN
ajst-24984	39	26	of	of	ADP
ajst-24984	39	27	the	the	DET
ajst-24984	39	28	text	text	NOUN
ajst-24984	39	29	is	be	AUX
ajst-24984	39	30	a	a	DET
ajst-24984	39	31	good	good	ADJ
ajst-24984	39	32	approach	approach	NOUN
ajst-24984	39	33	.	.	PUNCT
ajst-24984	40	1	the	the	DET
ajst-24984	40	2	"	"	PUNCT
ajst-24984	40	3	embeddings	embedding	NOUN
ajst-24984	40	4	from	from	ADP
ajst-24984	40	5	language	language	NOUN
ajst-24984	40	6	models	model	NOUN
ajst-24984	40	7	(	(	PUNCT
ajst-24984	40	8	elmo	elmo	PROPN
ajst-24984	40	9	)	)	PUNCT
ajst-24984	40	10	"	"	PUNCT
ajst-24984	41	1	[	[	X
ajst-24984	41	2	13	13	NUM
ajst-24984	41	3	]	]	PUNCT
ajst-24984	41	4	and	and	CCONJ
ajst-24984	41	5	"	"	PUNCT
ajst-24984	41	6	generative	generative	ADJ
ajst-24984	41	7	pre	pre	NOUN
ajst-24984	41	8	-	-	NOUN
ajst-24984	41	9	training	training	ADJ
ajst-24984	41	10	(	(	PUNCT
ajst-24984	41	11	gpt	gpt	NOUN
ajst-24984	41	12	)	)	PUNCT
ajst-24984	41	13	"	"	PUNCT
ajst-24984	42	1	[	[	X
ajst-24984	42	2	14	14	NUM
ajst-24984	42	3	]	]	ADJ
ajst-24984	42	4	models	model	NOUN
ajst-24984	42	5	,	,	PUNCT
ajst-24984	42	6	which	which	PRON
ajst-24984	42	7	exhibit	exhibit	VERB
ajst-24984	42	8	very	very	ADV
ajst-24984	42	9	good	good	ADJ
ajst-24984	42	10	performance	performance	NOUN
ajst-24984	42	11	,	,	PUNCT
ajst-24984	42	12	and	and	CCONJ
ajst-24984	42	13	the	the	DET
ajst-24984	42	14	subsequent	subsequent	ADJ
ajst-24984	42	15	emergence	emergence	NOUN
ajst-24984	42	16	of	of	ADP
ajst-24984	42	17	the	the	DET
ajst-24984	42	18	bert	bert	PROPN
ajst-24984	42	19	model	model	NOUN
ajst-24984	42	20	[	[	X
ajst-24984	42	21	15	15	NUM
ajst-24984	42	22	]	]	PUNCT
ajst-24984	42	23	also	also	ADV
ajst-24984	42	24	performs	perform	VERB
ajst-24984	42	25	well	well	ADV
ajst-24984	42	26	,	,	PUNCT
ajst-24984	42	27	having	having	AUX
ajst-24984	42	28	topped	top	VERB
ajst-24984	42	29	the	the	DET
ajst-24984	42	30	list	list	NOUN
ajst-24984	42	31	on	on	ADP
ajst-24984	42	32	various	various	ADJ
ajst-24984	42	33	natural	natural	ADJ
ajst-24984	42	34	language	language	NOUN
ajst-24984	42	35	processing	processing	NOUN
ajst-24984	42	36	tasks	task	NOUN
ajst-24984	42	37	.	.	PUNCT
ajst-24984	43	1	karimi	karimi	PROPN
ajst-24984	43	2	et	et	PROPN
ajst-24984	43	3	al	al	PROPN
ajst-24984	43	4	.	.	PUNCT
ajst-24984	44	1	[	[	X
ajst-24984	44	2	16	16	NUM
ajst-24984	44	3	]	]	PUNCT
ajst-24984	44	4	used	use	VERB
ajst-24984	44	5	the	the	DET
ajst-24984	44	6	standard	standard	ADJ
ajst-24984	44	7	bert	bert	PROPN
ajst-24984	44	8	model	model	NOUN
ajst-24984	44	9	to	to	PART
ajst-24984	44	10	complete	complete	VERB
ajst-24984	44	11	the	the	DET
ajst-24984	44	12	absa	absa	PROPN
ajst-24984	44	13	processing	processing	NOUN
ajst-24984	44	14	task	task	NOUN
ajst-24984	44	15	and	and	CCONJ
ajst-24984	44	16	found	find	VERB
ajst-24984	44	17	that	that	SCONJ
ajst-24984	44	18	the	the	DET
ajst-24984	44	19	performance	performance	NOUN
ajst-24984	44	20	of	of	ADP
ajst-24984	44	21	the	the	DET
ajst-24984	44	22	bert	bert	PROPN
ajst-24984	44	23	model	model	NOUN
ajst-24984	44	24	can	can	AUX
ajst-24984	44	25	significantly	significantly	ADV
ajst-24984	44	26	outperform	outperform	VERB
ajst-24984	44	27	the	the	DET
ajst-24984	44	28	general	general	ADJ
ajst-24984	44	29	neural	neural	ADJ
ajst-24984	44	30	network	network	NOUN
ajst-24984	44	31	model	model	NOUN
ajst-24984	44	32	,	,	PUNCT
ajst-24984	44	33	which	which	PRON
ajst-24984	44	34	can	can	AUX
ajst-24984	44	35	show	show	VERB
ajst-24984	44	36	that	that	SCONJ
ajst-24984	44	37	bert	bert	PROPN
ajst-24984	44	38	is	be	AUX
ajst-24984	44	39	capable	capable	ADJ
ajst-24984	44	40	of	of	ADP
ajst-24984	44	41	performing	perform	VERB
ajst-24984	44	42	the	the	DET
ajst-24984	44	43	absa	absa	PROPN
ajst-24984	44	44	task	task	NOUN
ajst-24984	44	45	to	to	ADP
ajst-24984	44	46	a	a	DET
ajst-24984	44	47	certain	certain	ADJ
ajst-24984	44	48	extent	extent	NOUN
ajst-24984	44	49	.	.	PUNCT
ajst-24984	45	1	although	although	SCONJ
ajst-24984	45	2	the	the	DET
ajst-24984	45	3	bert	bert	PROPN
ajst-24984	45	4	model	model	NOUN
ajst-24984	45	5	has	have	VERB
ajst-24984	45	6	excellent	excellent	ADJ
ajst-24984	45	7	performance	performance	NOUN
ajst-24984	45	8	,	,	PUNCT
ajst-24984	45	9	it	it	PRON
ajst-24984	45	10	has	have	VERB
ajst-24984	45	11	too	too	ADV
ajst-24984	45	12	many	many	ADJ
ajst-24984	45	13	parameters	parameter	NOUN
ajst-24984	45	14	,	,	PUNCT
ajst-24984	45	15	which	which	PRON
ajst-24984	45	16	leads	lead	VERB
ajst-24984	45	17	to	to	ADP
ajst-24984	45	18	slow	slow	ADJ
ajst-24984	45	19	convergence	convergence	NOUN
ajst-24984	45	20	of	of	ADP
ajst-24984	45	21	the	the	DET
ajst-24984	45	22	model	model	NOUN
ajst-24984	45	23	during	during	ADP
ajst-24984	45	24	training	training	NOUN
ajst-24984	45	25	,	,	PUNCT
ajst-24984	45	26	and	and	CCONJ
ajst-24984	45	27	it	it	PRON
ajst-24984	45	28	is	be	AUX
ajst-24984	45	29	very	very	ADV
ajst-24984	45	30	easy	easy	ADJ
ajst-24984	45	31	to	to	ADP
ajst-24984	45	32	overfitting	overfitte	VERB
ajst-24984	45	33	when	when	SCONJ
ajst-24984	45	34	the	the	DET
ajst-24984	45	35	data	datum	NOUN
ajst-24984	45	36	set	set	VERB
ajst-24984	45	37	is	be	AUX
ajst-24984	45	38	not	not	PART
ajst-24984	45	39	large	large	ADJ
ajst-24984	45	40	enough	enough	ADV
ajst-24984	45	41	.	.	PUNCT
ajst-24984	46	1	for	for	ADP
ajst-24984	46	2	this	this	DET
ajst-24984	46	3	kind	kind	NOUN
ajst-24984	46	4	of	of	ADP
ajst-24984	46	5	problem	problem	NOUN
ajst-24984	46	6	,	,	PUNCT
ajst-24984	46	7	some	some	DET
ajst-24984	46	8	researchers	researcher	NOUN
ajst-24984	46	9	start	start	VERB
ajst-24984	46	10	at	at	ADP
ajst-24984	46	11	the	the	DET
ajst-24984	46	12	pre	pre	ADJ
ajst-24984	46	13	-	-	ADJ
ajst-24984	46	14	training	training	ADJ
ajst-24984	46	15	stage	stage	NOUN
ajst-24984	46	16	,	,	PUNCT
ajst-24984	46	17	so	so	SCONJ
ajst-24984	46	18	that	that	SCONJ
ajst-24984	46	19	the	the	DET
ajst-24984	46	20	model	model	NOUN
ajst-24984	46	21	is	be	AUX
ajst-24984	46	22	first	first	ADJ
ajst-24984	46	23	pre	pre	ADJ
ajst-24984	46	24	-	-	VERB
ajst-24984	46	25	trained	train	VERB
ajst-24984	46	26	on	on	ADP
ajst-24984	46	27	a	a	DET
ajst-24984	46	28	larger	large	ADJ
ajst-24984	46	29	sentence	sentence	NOUN
ajst-24984	46	30	set	set	VERB
ajst-24984	46	31	sentiment	sentiment	NOUN
ajst-24984	46	32	analysis	analysis	NOUN
ajst-24984	46	33	dataset	dataset	NOUN
ajst-24984	46	34	,	,	PUNCT
ajst-24984	46	35	and	and	CCONJ
ajst-24984	46	36	then	then	ADV
ajst-24984	46	37	fine	fine	ADV
ajst-24984	46	38	-	-	PUNCT
ajst-24984	46	39	tuned	tune	VERB
ajst-24984	46	40	on	on	ADP
ajst-24984	46	41	the	the	DET
ajst-24984	46	42	dataset	dataset	NOUN
ajst-24984	46	43	of	of	ADP
ajst-24984	46	44	the	the	DET
ajst-24984	46	45	absa	absa	PROPN
ajst-24984	46	46	task	task	NOUN
ajst-24984	46	47	,	,	PUNCT
ajst-24984	46	48	so	so	SCONJ
ajst-24984	46	49	that	that	SCONJ
ajst-24984	46	50	the	the	DET
ajst-24984	46	51	model	model	NOUN
ajst-24984	46	52	performance	performance	NOUN
ajst-24984	46	53	can	can	AUX
ajst-24984	46	54	be	be	AUX
ajst-24984	46	55	improved	improve	VERB
ajst-24984	47	1	[	[	X
ajst-24984	47	2	17][18	17][18	NUM
ajst-24984	47	3	]	]	PUNCT
ajst-24984	47	4	.	.	PUNCT
ajst-24984	48	1	3	3	X
ajst-24984	48	2	.	.	X
ajst-24984	48	3	aspect	aspect	NOUN
ajst-24984	48	4	-	-	PUNCT
ajst-24984	48	5	based	base	VERB
ajst-24984	48	6	sentiment	sentiment	NOUN
ajst-24984	48	7	analysis	analysis	NOUN
ajst-24984	48	8	approaches	approach	VERB
ajst-24984	48	9	sentiment	sentiment	NOUN
ajst-24984	48	10	analysis	analysis	NOUN
ajst-24984	48	11	is	be	AUX
ajst-24984	48	12	a	a	DET
ajst-24984	48	13	technical	technical	ADJ
ajst-24984	48	14	means	mean	NOUN
ajst-24984	48	15	to	to	PART
ajst-24984	48	16	probe	probe	VERB
ajst-24984	48	17	deeply	deeply	ADV
ajst-24984	48	18	into	into	ADP
ajst-24984	48	19	the	the	DET
ajst-24984	48	20	sentiment	sentiment	NOUN
ajst-24984	48	21	connotation	connotation	NOUN
ajst-24984	48	22	of	of	ADP
ajst-24984	48	23	text	text	NOUN
ajst-24984	48	24	,	,	PUNCT
ajst-24984	48	25	aiming	aim	VERB
ajst-24984	48	26	to	to	PART
ajst-24984	48	27	reveal	reveal	VERB
ajst-24984	48	28	the	the	DET
ajst-24984	48	29	hidden	hidden	ADJ
ajst-24984	48	30	emotions	emotion	NOUN
ajst-24984	48	31	or	or	CCONJ
ajst-24984	48	32	attitudinal	attitudinal	ADJ
ajst-24984	48	33	tendencies	tendency	NOUN
ajst-24984	48	34	behind	behind	ADP
ajst-24984	48	35	the	the	DET
ajst-24984	48	36	text	text	NOUN
ajst-24984	48	37	.	.	PUNCT
ajst-24984	49	1	the	the	DET
ajst-24984	49	2	technique	technique	NOUN
ajst-24984	49	3	is	be	AUX
ajst-24984	49	4	widely	widely	ADV
ajst-24984	49	5	used	use	VERB
ajst-24984	49	6	in	in	ADP
ajst-24984	49	7	several	several	ADJ
ajst-24984	49	8	fields	field	NOUN
ajst-24984	49	9	,	,	PUNCT
ajst-24984	49	10	such	such	ADJ
ajst-24984	49	11	as	as	ADP
ajst-24984	49	12	public	public	ADJ
ajst-24984	49	13	opinion	opinion	NOUN
ajst-24984	49	14	monitoring	monitoring	NOUN
ajst-24984	49	15	,	,	PUNCT
ajst-24984	49	16	product	product	NOUN
ajst-24984	49	17	review	review	VERB
ajst-24984	49	18	analysis	analysis	NOUN
ajst-24984	49	19	,	,	PUNCT
ajst-24984	49	20	event	event	NOUN
ajst-24984	49	21	monitoring	monitoring	NOUN
ajst-24984	49	22	,	,	PUNCT
ajst-24984	49	23	etc	etc	X
ajst-24984	49	24	.	.	X
ajst-24984	49	25	aspect	aspect	NOUN
ajst-24984	49	26	-	-	PUNCT
ajst-24984	49	27	based	base	VERB
ajst-24984	49	28	sentiment	sentiment	NOUN
ajst-24984	49	29	analysis	analysis	NOUN
ajst-24984	49	30	can	can	AUX
ajst-24984	49	31	be	be	AUX
ajst-24984	49	32	categorized	categorize	VERB
ajst-24984	49	33	into	into	ADP
ajst-24984	49	34	document	document	NOUN
ajst-24984	49	35	-	-	PUNCT
ajst-24984	49	36	level	level	NOUN
ajst-24984	49	37	,	,	PUNCT
ajst-24984	49	38	sentence	sentence	NOUN
ajst-24984	49	39	-	-	PUNCT
ajst-24984	49	40	level	level	NOUN
ajst-24984	49	41	and	and	CCONJ
ajst-24984	49	42	aspectbased	aspectbase	VERB
ajst-24984	49	43	sentiment	sentiment	NOUN
ajst-24984	49	44	analysis	analysis	NOUN
ajst-24984	49	45	according	accord	VERB
ajst-24984	49	46	to	to	ADP
ajst-24984	49	47	the	the	DET
ajst-24984	49	48	different	different	ADJ
ajst-24984	49	49	granularity	granularity	NOUN
ajst-24984	49	50	of	of	ADP
ajst-24984	49	51	analysis	analysis	NOUN
ajst-24984	49	52	.	.	PUNCT
ajst-24984	50	1	among	among	ADP
ajst-24984	50	2	them	they	PRON
ajst-24984	50	3	,	,	PUNCT
ajst-24984	50	4	document	document	NOUN
ajst-24984	50	5	-	-	PUNCT
ajst-24984	50	6	level	level	NOUN
ajst-24984	50	7	and	and	CCONJ
ajst-24984	50	8	sentence	sentence	NOUN
ajst-24984	50	9	-	-	PUNCT
ajst-24984	50	10	level	level	NOUN
ajst-24984	50	11	belong	belong	VERB
ajst-24984	50	12	to	to	ADP
ajst-24984	50	13	coarse	coarse	NOUN
ajst-24984	50	14	-	-	PUNCT
ajst-24984	50	15	grained	grain	VERB
ajst-24984	50	16	sentiment	sentiment	NOUN
ajst-24984	50	17	analysis	analysis	NOUN
ajst-24984	50	18	,	,	PUNCT
ajst-24984	50	19	while	while	SCONJ
ajst-24984	50	20	aspect	aspect	NOUN
ajst-24984	50	21	-	-	PUNCT
ajst-24984	50	22	based	base	VERB
ajst-24984	50	23	sentiment	sentiment	NOUN
ajst-24984	50	24	analysis	analysis	NOUN
ajst-24984	50	25	belongs	belong	VERB
ajst-24984	50	26	to	to	ADP
ajst-24984	50	27	finegrained	finegraine	VERB
ajst-24984	50	28	sentiment	sentiment	NOUN
ajst-24984	50	29	analysis	analysis	NOUN
ajst-24984	50	30	.	.	PUNCT
ajst-24984	51	1	document	document	NOUN
ajst-24984	51	2	-	-	PUNCT
ajst-24984	51	3	level	level	NOUN
ajst-24984	51	4	sentiment	sentiment	NOUN
ajst-24984	51	5	analysis	analysis	NOUN
ajst-24984	51	6	is	be	AUX
ajst-24984	51	7	the	the	DET
ajst-24984	51	8	earliest	early	ADJ
ajst-24984	51	9	form	form	NOUN
ajst-24984	51	10	of	of	ADP
ajst-24984	51	11	sentiment	sentiment	NOUN
ajst-24984	51	12	analysis	analysis	NOUN
ajst-24984	51	13	,	,	PUNCT
ajst-24984	51	14	which	which	PRON
ajst-24984	51	15	targets	target	VERB
ajst-24984	51	16	the	the	DET
ajst-24984	51	17	whole	whole	ADJ
ajst-24984	51	18	document	document	NOUN
ajst-24984	51	19	,	,	PUNCT
ajst-24984	51	20	extracts	extract	VERB
ajst-24984	51	21	all	all	DET
ajst-24984	51	22	the	the	DET
ajst-24984	51	23	sentiment	sentiment	NOUN
ajst-24984	51	24	words	word	NOUN
ajst-24984	51	25	and	and	CCONJ
ajst-24984	51	26	performs	perform	VERB
ajst-24984	51	27	sentiment	sentiment	NOUN
ajst-24984	51	28	classification	classification	NOUN
ajst-24984	51	29	statistics	statistic	NOUN
ajst-24984	51	30	to	to	PART
ajst-24984	51	31	determine	determine	VERB
ajst-24984	51	32	the	the	DET
ajst-24984	51	33	sentiment	sentiment	NOUN
ajst-24984	51	34	tendency	tendency	NOUN
ajst-24984	51	35	of	of	ADP
ajst-24984	51	36	the	the	DET
ajst-24984	51	37	whole	whole	ADJ
ajst-24984	51	38	document	document	NOUN
ajst-24984	51	39	.	.	PUNCT
ajst-24984	52	1	sentencelevel	sentencelevel	NOUN
ajst-24984	52	2	sentiment	sentiment	NOUN
ajst-24984	52	3	analysis	analysis	NOUN
ajst-24984	52	4	is	be	AUX
ajst-24984	52	5	similar	similar	ADJ
ajst-24984	52	6	to	to	ADP
ajst-24984	52	7	the	the	DET
ajst-24984	52	8	methodology	methodology	NOUN
ajst-24984	52	9	of	of	ADP
ajst-24984	52	10	document	document	NOUN
ajst-24984	52	11	-	-	PUNCT
ajst-24984	52	12	level	level	NOUN
ajst-24984	52	13	sentiment	sentiment	NOUN
ajst-24984	52	14	analysis	analysis	NOUN
ajst-24984	52	15	,	,	PUNCT
ajst-24984	52	16	but	but	CCONJ
ajst-24984	52	17	the	the	DET
ajst-24984	52	18	object	object	NOUN
ajst-24984	52	19	of	of	ADP
ajst-24984	52	20	analysis	analysis	NOUN
ajst-24984	52	21	is	be	AUX
ajst-24984	52	22	narrowed	narrow	VERB
ajst-24984	52	23	down	down	ADP
ajst-24984	52	24	from	from	ADP
ajst-24984	52	25	the	the	DET
ajst-24984	52	26	whole	whole	ADJ
ajst-24984	52	27	document	document	NOUN
ajst-24984	52	28	to	to	ADP
ajst-24984	52	29	a	a	DET
ajst-24984	52	30	single	single	ADJ
ajst-24984	52	31	sentence	sentence	NOUN
ajst-24984	52	32	.	.	PUNCT
ajst-24984	53	1	aspect	aspect	NOUN
ajst-24984	53	2	-	-	PUNCT
ajst-24984	53	3	based	base	VERB
ajst-24984	53	4	sentiment	sentiment	NOUN
ajst-24984	53	5	analysis	analysis	NOUN
ajst-24984	53	6	is	be	AUX
ajst-24984	53	7	a	a	DET
ajst-24984	53	8	sub	sub	NOUN
ajst-24984	53	9	-	-	NOUN
ajst-24984	53	10	domain	domain	NOUN
ajst-24984	53	11	of	of	ADP
ajst-24984	53	12	sentiment	sentiment	NOUN
ajst-24984	53	13	analysis	analysis	NOUN
ajst-24984	53	14	whose	whose	DET
ajst-24984	53	15	task	task	NOUN
ajst-24984	53	16	is	be	AUX
ajst-24984	53	17	to	to	PART
ajst-24984	53	18	mine	mine	VERB
ajst-24984	53	19	multiple	multiple	ADJ
ajst-24984	53	20	aspects	aspect	NOUN
ajst-24984	53	21	of	of	ADP
ajst-24984	53	22	a	a	DET
ajst-24984	53	23	given	give	VERB
ajst-24984	53	24	entity	entity	NOUN
ajst-24984	53	25	in	in	ADP
ajst-24984	53	26	a	a	DET
ajst-24984	53	27	given	give	VERB
ajst-24984	53	28	text	text	NOUN
ajst-24984	53	29	sentence	sentence	NOUN
ajst-24984	53	30	.	.	PUNCT
ajst-24984	54	1	by	by	ADP
ajst-24984	54	2	extracting	extract	VERB
ajst-24984	54	3	the	the	DET
ajst-24984	54	4	sentiment	sentiment	NOUN
ajst-24984	54	5	words	word	NOUN
ajst-24984	54	6	contained	contain	VERB
ajst-24984	54	7	in	in	ADP
ajst-24984	54	8	these	these	DET
ajst-24984	54	9	aspects	aspect	NOUN
ajst-24984	54	10	,	,	PUNCT
ajst-24984	54	11	aspect	aspect	NOUN
ajst-24984	54	12	-	-	PUNCT
ajst-24984	54	13	based	base	VERB
ajst-24984	54	14	sentiment	sentiment	NOUN
ajst-24984	54	15	analysis	analysis	NOUN
ajst-24984	54	16	is	be	AUX
ajst-24984	54	17	able	able	ADJ
ajst-24984	54	18	to	to	PART
ajst-24984	54	19	determine	determine	VERB
ajst-24984	54	20	the	the	DET
ajst-24984	54	21	sentiment	sentiment	NOUN
ajst-24984	54	22	polarity	polarity	NOUN
ajst-24984	54	23	of	of	ADP
ajst-24984	54	24	that	that	DET
ajst-24984	54	25	entity	entity	NOUN
ajst-24984	54	26	.	.	PUNCT
ajst-24984	55	1	the	the	DET
ajst-24984	55	2	techniques	technique	NOUN
ajst-24984	55	3	used	use	VERB
ajst-24984	55	4	for	for	ADP
ajst-24984	55	5	different	different	ADJ
ajst-24984	55	6	levels	level	NOUN
ajst-24984	55	7	of	of	ADP
ajst-24984	55	8	sentiment	sentiment	NOUN
ajst-24984	55	9	analysis	analysis	NOUN
ajst-24984	55	10	vary	vary	VERB
ajst-24984	55	11	due	due	ADP
ajst-24984	55	12	to	to	ADP
ajst-24984	55	13	the	the	DET
ajst-24984	55	14	granularity	granularity	NOUN
ajst-24984	55	15	of	of	ADP
ajst-24984	55	16	the	the	DET
ajst-24984	55	17	analysis	analysis	NOUN
ajst-24984	55	18	.	.	PUNCT
ajst-24984	56	1	over	over	ADP
ajst-24984	56	2	the	the	DET
ajst-24984	56	3	past	past	ADJ
ajst-24984	56	4	years	year	NOUN
ajst-24984	56	5	,	,	PUNCT
ajst-24984	56	6	the	the	DET
ajst-24984	56	7	field	field	NOUN
ajst-24984	56	8	of	of	ADP
ajst-24984	56	9	sentiment	sentiment	NOUN
ajst-24984	56	10	analysis	analysis	NOUN
ajst-24984	56	11	has	have	AUX
ajst-24984	56	12	witnessed	witness	VERB
ajst-24984	56	13	the	the	DET
ajst-24984	56	14	emergence	emergence	NOUN
ajst-24984	56	15	of	of	ADP
ajst-24984	56	16	multiple	multiple	ADJ
ajst-24984	56	17	solution	solution	NOUN
ajst-24984	56	18	approaches	approach	NOUN
ajst-24984	56	19	.	.	PUNCT
ajst-24984	57	1	in	in	ADP
ajst-24984	57	2	the	the	DET
ajst-24984	57	3	early	early	ADJ
ajst-24984	57	4	stages	stage	NOUN
ajst-24984	57	5	of	of	ADP
ajst-24984	57	6	artificial	artificial	ADJ
ajst-24984	57	7	intelligence	intelligence	NOUN
ajst-24984	57	8	,	,	PUNCT
ajst-24984	57	9	researchers	researcher	NOUN
ajst-24984	57	10	commonly	commonly	ADV
ajst-24984	57	11	used	use	VERB
ajst-24984	57	12	sentiment	sentiment	NOUN
ajst-24984	57	13	dictionaries	dictionary	NOUN
ajst-24984	57	14	or	or	CCONJ
ajst-24984	57	15	traditional	traditional	ADJ
ajst-24984	57	16	machine	machine	NOUN
ajst-24984	57	17	learning	learning	NOUN
ajst-24984	57	18	methods	method	NOUN
ajst-24984	57	19	for	for	ADP
ajst-24984	57	20	sentiment	sentiment	NOUN
ajst-24984	57	21	analysis	analysis	NOUN
ajst-24984	57	22	.	.	PUNCT
ajst-24984	58	1	however	however	ADV
ajst-24984	58	2	,	,	PUNCT
ajst-24984	58	3	with	with	ADP
ajst-24984	58	4	the	the	DET
ajst-24984	58	5	advancement	advancement	NOUN
ajst-24984	58	6	of	of	ADP
ajst-24984	58	7	ai	ai	PROPN
ajst-24984	58	8	technology	technology	NOUN
ajst-24984	58	9	,	,	PUNCT
ajst-24984	58	10	a	a	DET
ajst-24984	58	11	single	single	ADJ
ajst-24984	58	12	solution	solution	NOUN
ajst-24984	58	13	strategy	strategy	NOUN
ajst-24984	58	14	can	can	AUX
ajst-24984	58	15	no	no	ADV
ajst-24984	58	16	longer	long	ADV
ajst-24984	58	17	fully	fully	ADV
ajst-24984	58	18	satisfy	satisfy	VERB
ajst-24984	58	19	the	the	DET
ajst-24984	58	20	needs	need	NOUN
ajst-24984	58	21	of	of	ADP
ajst-24984	58	22	sentiment	sentiment	NOUN
ajst-24984	58	23	analysis	analysis	NOUN
ajst-24984	58	24	tasks	task	NOUN
ajst-24984	58	25	.	.	PUNCT
ajst-24984	59	1	as	as	ADP
ajst-24984	59	2	a	a	DET
ajst-24984	59	3	result	result	NOUN
ajst-24984	59	4	,	,	PUNCT
ajst-24984	59	5	researchers	researcher	NOUN
ajst-24984	59	6	have	have	AUX
ajst-24984	59	7	begun	begin	VERB
ajst-24984	59	8	to	to	PART
ajst-24984	59	9	explore	explore	VERB
ajst-24984	59	10	ways	way	NOUN
ajst-24984	59	11	to	to	PART
ajst-24984	59	12	combine	combine	VERB
ajst-24984	59	13	sentiment	sentiment	NOUN
ajst-24984	59	14	dictionaries	dictionary	NOUN
ajst-24984	59	15	with	with	ADP
ajst-24984	59	16	machine	machine	NOUN
ajst-24984	59	17	learning	learn	VERB
ajst-24984	59	18	methods	method	NOUN
ajst-24984	59	19	to	to	PART
ajst-24984	59	20	optimize	optimize	VERB
ajst-24984	59	21	model	model	NOUN
ajst-24984	59	22	performance	performance	NOUN
ajst-24984	59	23	by	by	ADP
ajst-24984	59	24	introducing	introduce	VERB
ajst-24984	59	25	specific	specific	ADJ
ajst-24984	59	26	constraint	constraint	NOUN
ajst-24984	59	27	rules	rule	NOUN
ajst-24984	59	28	,	,	PUNCT
ajst-24984	59	29	with	with	ADP
ajst-24984	59	30	a	a	DET
ajst-24984	59	31	view	view	NOUN
ajst-24984	59	32	to	to	ADP
ajst-24984	59	33	obtaining	obtain	VERB
ajst-24984	59	34	more	more	ADV
ajst-24984	59	35	accurate	accurate	ADJ
ajst-24984	59	36	sentiment	sentiment	NOUN
ajst-24984	59	37	analysis	analysis	NOUN
ajst-24984	59	38	results	result	NOUN
ajst-24984	59	39	.	.	PUNCT
ajst-24984	60	1	in	in	ADP
ajst-24984	60	2	recent	recent	ADJ
ajst-24984	60	3	years	year	NOUN
ajst-24984	60	4	,	,	PUNCT
ajst-24984	60	5	the	the	DET
ajst-24984	60	6	rise	rise	NOUN
ajst-24984	60	7	of	of	ADP
ajst-24984	60	8	deep	deep	ADJ
ajst-24984	60	9	learning	learning	NOUN
ajst-24984	60	10	has	have	AUX
ajst-24984	60	11	gradually	gradually	ADV
ajst-24984	60	12	replaced	replace	VERB
ajst-24984	60	13	the	the	DET
ajst-24984	60	14	traditional	traditional	ADJ
ajst-24984	60	15	manual	manual	ADJ
ajst-24984	60	16	feature	feature	NOUN
ajst-24984	60	17	design	design	NOUN
ajst-24984	60	18	and	and	CCONJ
ajst-24984	60	19	selection	selection	NOUN
ajst-24984	60	20	process	process	NOUN
ajst-24984	60	21	as	as	ADP
ajst-24984	60	22	a	a	DET
ajst-24984	60	23	mainstream	mainstream	NOUN
ajst-24984	60	24	technique	technique	NOUN
ajst-24984	60	25	in	in	ADP
ajst-24984	60	26	sentiment	sentiment	NOUN
ajst-24984	60	27	analysis	analysis	NOUN
ajst-24984	60	28	tasks	task	NOUN
ajst-24984	60	29	.	.	PUNCT
ajst-24984	61	1	currently	currently	ADV
ajst-24984	61	2	,	,	PUNCT
ajst-24984	61	3	there	there	PRON
ajst-24984	61	4	are	be	VERB
ajst-24984	61	5	diverse	diverse	ADJ
ajst-24984	61	6	implementation	implementation	NOUN
ajst-24984	61	7	strategies	strategy	NOUN
ajst-24984	61	8	for	for	ADP
ajst-24984	61	9	sentiment	sentiment	NOUN
ajst-24984	61	10	analysis	analysis	NOUN
ajst-24984	61	11	,	,	PUNCT
ajst-24984	61	12	including	include	VERB
ajst-24984	61	13	but	but	CCONJ
ajst-24984	61	14	not	not	PART
ajst-24984	61	15	limited	limit	VERB
ajst-24984	61	16	to	to	PART
ajst-24984	61	17	:	:	PUNCT
ajst-24984	61	18	sentiment	sentiment	NOUN
ajst-24984	61	19	lexicon	lexicon	NOUN
ajst-24984	61	20	-	-	PUNCT
ajst-24984	61	21	based	base	VERB
ajst-24984	61	22	approaches	approach	NOUN
ajst-24984	61	23	,	,	PUNCT
ajst-24984	61	24	traditional	traditional	ADJ
ajst-24984	61	25	machine	machine	NOUN
ajst-24984	61	26	learning	learn	VERB
ajst-24984	61	27	techniques	technique	NOUN
ajst-24984	61	28	,	,	PUNCT
ajst-24984	61	29	and	and	CCONJ
ajst-24984	61	30	deep	deep	ADJ
ajst-24984	61	31	learning	learning	NOUN
ajst-24984	61	32	techniques	technique	NOUN
ajst-24984	61	33	.	.	PUNCT
ajst-24984	62	1	each	each	PRON
ajst-24984	62	2	of	of	ADP
ajst-24984	62	3	these	these	DET
ajst-24984	62	4	approaches	approach	NOUN
ajst-24984	62	5	has	have	VERB
ajst-24984	62	6	its	its	PRON
ajst-24984	62	7	own	own	ADJ
ajst-24984	62	8	advantages	advantage	NOUN
ajst-24984	62	9	,	,	PUNCT
ajst-24984	62	10	and	and	CCONJ
ajst-24984	62	11	together	together	ADV
ajst-24984	62	12	they	they	PRON
ajst-24984	62	13	have	have	AUX
ajst-24984	62	14	contributed	contribute	VERB
ajst-24984	62	15	to	to	ADP
ajst-24984	62	16	the	the	DET
ajst-24984	62	17	advancement	advancement	NOUN
ajst-24984	62	18	of	of	ADP
ajst-24984	62	19	sentiment	sentiment	NOUN
ajst-24984	62	20	analysis	analysis	NOUN
ajst-24984	62	21	technology	technology	NOUN
ajst-24984	62	22	and	and	CCONJ
ajst-24984	62	23	the	the	DET
ajst-24984	62	24	expansion	expansion	NOUN
ajst-24984	62	25	of	of	ADP
ajst-24984	62	26	its	its	PRON
ajst-24984	62	27	application	application	NOUN
ajst-24984	62	28	.	.	PUNCT
ajst-24984	63	1	3.1	3.1	NUM
ajst-24984	63	2	.	.	PUNCT
ajst-24984	63	3	sentiment	sentiment	PROPN
ajst-24984	63	4	dictionary	dictionary	NOUN
ajst-24984	63	5	based	base	VERB
ajst-24984	63	6	approaches	approach	NOUN
ajst-24984	63	7	in	in	ADP
ajst-24984	63	8	the	the	DET
ajst-24984	63	9	early	early	ADJ
ajst-24984	63	10	stages	stage	NOUN
ajst-24984	63	11	of	of	ADP
ajst-24984	63	12	sentiment	sentiment	NOUN
ajst-24984	63	13	analysis	analysis	NOUN
ajst-24984	63	14	,	,	PUNCT
ajst-24984	63	15	researchers	researcher	NOUN
ajst-24984	63	16	relied	rely	VERB
ajst-24984	63	17	heavily	heavily	ADV
ajst-24984	63	18	on	on	ADP
ajst-24984	63	19	rules	rule	NOUN
ajst-24984	63	20	to	to	PART
ajst-24984	63	21	construct	construct	VERB
ajst-24984	63	22	sentiment	sentiment	NOUN
ajst-24984	63	23	dictionaries	dictionary	NOUN
ajst-24984	63	24	,	,	PUNCT
ajst-24984	63	25	which	which	PRON
ajst-24984	63	26	infer	infer	VERB
ajst-24984	63	27	the	the	DET
ajst-24984	63	28	sentiment	sentiment	NOUN
ajst-24984	63	29	tendency	tendency	NOUN
ajst-24984	63	30	of	of	ADP
ajst-24984	63	31	a	a	DET
ajst-24984	63	32	text	text	NOUN
ajst-24984	63	33	by	by	ADP
ajst-24984	63	34	counting	count	VERB
ajst-24984	63	35	the	the	DET
ajst-24984	63	36	frequency	frequency	NOUN
ajst-24984	63	37	of	of	ADP
ajst-24984	63	38	sentiment	sentiment	NOUN
ajst-24984	63	39	words	word	NOUN
ajst-24984	63	40	or	or	CCONJ
ajst-24984	63	41	phrases	phrase	NOUN
ajst-24984	63	42	in	in	ADP
ajst-24984	63	43	the	the	DET
ajst-24984	63	44	text	text	NOUN
ajst-24984	63	45	.	.	PUNCT
ajst-24984	64	1	however	however	ADV
ajst-24984	64	2	,	,	PUNCT
ajst-24984	64	3	the	the	DET
ajst-24984	64	4	effectiveness	effectiveness	NOUN
ajst-24984	64	5	of	of	ADP
ajst-24984	64	6	this	this	DET
ajst-24984	64	7	approach	approach	NOUN
ajst-24984	64	8	largely	largely	ADV
ajst-24984	64	9	depends	depend	VERB
ajst-24984	64	10	on	on	ADP
ajst-24984	64	11	the	the	DET
ajst-24984	64	12	completeness	completeness	NOUN
ajst-24984	64	13	of	of	ADP
ajst-24984	64	14	the	the	DET
ajst-24984	64	15	sentiment	sentiment	NOUN
ajst-24984	64	16	dictionary	dictionary	NOUN
ajst-24984	64	17	.	.	PUNCT
ajst-24984	65	1	as	as	ADP
ajst-24984	65	2	an	an	DET
ajst-24984	65	3	important	important	ADJ
ajst-24984	65	4	resource	resource	NOUN
ajst-24984	65	5	reflecting	reflect	VERB
ajst-24984	65	6	the	the	DET
ajst-24984	65	7	unstructured	unstructured	ADJ
ajst-24984	65	8	features	feature	NOUN
ajst-24984	65	9	of	of	ADP
ajst-24984	65	10	texts	text	NOUN
ajst-24984	65	11	,	,	PUNCT
ajst-24984	65	12	sentiment	sentiment	NOUN
ajst-24984	65	13	dictionaries	dictionary	NOUN
ajst-24984	65	14	play	play	VERB
ajst-24984	65	15	a	a	DET
ajst-24984	65	16	key	key	ADJ
ajst-24984	65	17	role	role	NOUN
ajst-24984	65	18	in	in	ADP
ajst-24984	65	19	sentiment	sentiment	NOUN
ajst-24984	65	20	analysis	analysis	NOUN
ajst-24984	65	21	of	of	ADP
ajst-24984	65	22	texts	text	NOUN
ajst-24984	65	23	.	.	PUNCT
ajst-24984	66	1	with	with	ADP
ajst-24984	66	2	the	the	DET
ajst-24984	66	3	in	in	ADP
ajst-24984	66	4	-	-	PUNCT
ajst-24984	66	5	depth	depth	NOUN
ajst-24984	66	6	study	study	NOUN
ajst-24984	66	7	of	of	ADP
ajst-24984	66	8	sentiment	sentiment	NOUN
ajst-24984	66	9	dictionaries	dictionary	NOUN
ajst-24984	66	10	by	by	ADP
ajst-24984	66	11	194	194	NUM
ajst-24984	66	12	scholars	scholar	NOUN
ajst-24984	66	13	at	at	ADP
ajst-24984	66	14	home	home	NOUN
ajst-24984	66	15	and	and	CCONJ
ajst-24984	66	16	abroad	abroad	ADV
ajst-24984	66	17	,	,	PUNCT
ajst-24984	66	18	many	many	ADJ
ajst-24984	66	19	sentiment	sentiment	NOUN
ajst-24984	66	20	dictionaries	dictionary	NOUN
ajst-24984	66	21	have	have	AUX
ajst-24984	66	22	been	be	AUX
ajst-24984	66	23	constructed	construct	VERB
ajst-24984	66	24	artificially	artificially	ADV
ajst-24984	66	25	.	.	PUNCT
ajst-24984	67	1	internationally	internationally	ADV
ajst-24984	67	2	,	,	PUNCT
ajst-24984	67	3	one	one	NUM
ajst-24984	67	4	of	of	ADP
ajst-24984	67	5	the	the	DET
ajst-24984	67	6	earliest	early	ADJ
ajst-24984	67	7	english	english	ADJ
ajst-24984	67	8	sentiment	sentiment	NOUN
ajst-24984	67	9	dictionaries	dictionary	NOUN
ajst-24984	67	10	is	be	AUX
ajst-24984	67	11	sentiwordnet	sentiwordnet	NOUN
ajst-24984	67	12	.	.	PUNCT
ajst-24984	68	1	in	in	ADP
ajst-24984	68	2	addition	addition	NOUN
ajst-24984	68	3	,	,	PUNCT
ajst-24984	68	4	there	there	PRON
ajst-24984	68	5	are	be	VERB
ajst-24984	68	6	some	some	DET
ajst-24984	68	7	other	other	ADJ
ajst-24984	68	8	well	well	ADV
ajst-24984	68	9	-	-	PUNCT
ajst-24984	68	10	known	know	VERB
ajst-24984	68	11	sentiment	sentiment	NOUN
ajst-24984	68	12	dictionaries	dictionary	NOUN
ajst-24984	68	13	,	,	PUNCT
ajst-24984	68	14	such	such	ADJ
ajst-24984	68	15	as	as	ADP
ajst-24984	68	16	the	the	DET
ajst-24984	68	17	general	general	ADJ
ajst-24984	68	18	inquirer	inquirer	NOUN
ajst-24984	68	19	(	(	PUNCT
ajst-24984	68	20	gi	gi	NOUN
ajst-24984	68	21	)	)	PUNCT
ajst-24984	68	22	,	,	PUNCT
ajst-24984	68	23	linguistic	linguistic	ADJ
ajst-24984	68	24	inquiry	inquiry	NOUN
ajst-24984	68	25	and	and	CCONJ
ajst-24984	68	26	word	word	NOUN
ajst-24984	68	27	count	count	NOUN
ajst-24984	68	28	(	(	PUNCT
ajst-24984	68	29	liwc	liwc	PROPN
ajst-24984	68	30	)	)	PUNCT
ajst-24984	68	31	,	,	PUNCT
ajst-24984	68	32	mpqa	mpqa	NOUN
ajst-24984	68	33	subjectivity	subjectivity	NOUN
ajst-24984	68	34	cues	cue	NOUN
ajst-24984	68	35	lexicon	lexicon	X
ajst-24984	68	36	,	,	PUNCT
ajst-24984	68	37	etc	etc	X
ajst-24984	68	38	.	.	X
ajst-24984	69	1	in	in	ADP
ajst-24984	69	2	china	china	PROPN
ajst-24984	69	3	,	,	PUNCT
ajst-24984	69	4	the	the	DET
ajst-24984	69	5	main	main	ADJ
ajst-24984	69	6	representatives	representative	NOUN
ajst-24984	69	7	of	of	ADP
ajst-24984	69	8	chinese	chinese	ADJ
ajst-24984	69	9	sentiment	sentiment	NOUN
ajst-24984	69	10	dictionaries	dictionary	NOUN
ajst-24984	69	11	include	include	VERB
ajst-24984	69	12	ntusd	ntusd	PROPN
ajst-24984	69	13	dictionary	dictionary	PROPN
ajst-24984	69	14	,	,	PUNCT
ajst-24984	69	15	zhiwang	zhiwang	PROPN
ajst-24984	69	16	dictionary	dictionary	PROPN
ajst-24984	69	17	hownet	hownet	PROPN
ajst-24984	69	18	and	and	CCONJ
ajst-24984	69	19	sentiment	sentiment	NOUN
ajst-24984	69	20	vocabulary	vocabulary	NOUN
ajst-24984	69	21	ontology	ontology	NOUN
ajst-24984	69	22	library	library	NOUN
ajst-24984	69	23	.	.	PUNCT
ajst-24984	70	1	the	the	DET
ajst-24984	70	2	construction	construction	NOUN
ajst-24984	70	3	and	and	CCONJ
ajst-24984	70	4	application	application	NOUN
ajst-24984	70	5	of	of	ADP
ajst-24984	70	6	these	these	DET
ajst-24984	70	7	dictionaries	dictionary	NOUN
ajst-24984	70	8	have	have	AUX
ajst-24984	70	9	greatly	greatly	ADV
ajst-24984	70	10	promoted	promote	VERB
ajst-24984	70	11	the	the	DET
ajst-24984	70	12	development	development	NOUN
ajst-24984	70	13	of	of	ADP
ajst-24984	70	14	the	the	DET
ajst-24984	70	15	field	field	NOUN
ajst-24984	70	16	of	of	ADP
ajst-24984	70	17	sentiment	sentiment	NOUN
ajst-24984	70	18	analysis	analysis	NOUN
ajst-24984	70	19	.	.	PUNCT
ajst-24984	71	1	with	with	ADP
ajst-24984	71	2	the	the	DET
ajst-24984	71	3	booming	booming	ADJ
ajst-24984	71	4	development	development	NOUN
ajst-24984	71	5	of	of	ADP
ajst-24984	71	6	the	the	DET
ajst-24984	71	7	social	social	ADJ
ajst-24984	71	8	media	medium	NOUN
ajst-24984	71	9	industry	industry	NOUN
ajst-24984	71	10	,	,	PUNCT
ajst-24984	71	11	more	more	ADJ
ajst-24984	71	12	and	and	CCONJ
ajst-24984	71	13	more	more	ADJ
ajst-24984	71	14	scholars	scholar	NOUN
ajst-24984	71	15	began	begin	VERB
ajst-24984	71	16	to	to	PART
ajst-24984	71	17	pay	pay	VERB
ajst-24984	71	18	attention	attention	NOUN
ajst-24984	71	19	to	to	ADP
ajst-24984	71	20	the	the	DET
ajst-24984	71	21	social	social	ADJ
ajst-24984	71	22	sentiment	sentiment	NOUN
ajst-24984	71	23	in	in	ADP
ajst-24984	71	24	social	social	ADJ
ajst-24984	71	25	media	medium	NOUN
ajst-24984	71	26	events	event	NOUN
ajst-24984	71	27	,	,	PUNCT
ajst-24984	71	28	and	and	CCONJ
ajst-24984	71	29	this	this	PRON
ajst-24984	71	30	has	have	AUX
ajst-24984	71	31	given	give	VERB
ajst-24984	71	32	rise	rise	NOUN
ajst-24984	71	33	to	to	ADP
ajst-24984	71	34	numerous	numerous	ADJ
ajst-24984	71	35	novel	novel	ADJ
ajst-24984	71	36	analysis	analysis	NOUN
ajst-24984	71	37	methods	method	NOUN
ajst-24984	71	38	based	base	VERB
ajst-24984	71	39	on	on	ADP
ajst-24984	71	40	sentiment	sentiment	NOUN
ajst-24984	71	41	dictionaries	dictionary	NOUN
ajst-24984	71	42	.	.	PUNCT
ajst-24984	72	1	jia	jia	PROPN
ajst-24984	72	2	k	k	PROPN
ajst-24984	72	3	et	et	PROPN
ajst-24984	72	4	al	al	PROPN
ajst-24984	73	1	[	[	X
ajst-24984	73	2	19	19	NUM
ajst-24984	73	3	]	]	PUNCT
ajst-24984	73	4	combined	combined	ADJ
ajst-24984	73	5	sentiment	sentiment	NOUN
ajst-24984	73	6	dictionaries	dictionary	NOUN
ajst-24984	73	7	and	and	CCONJ
ajst-24984	73	8	semantic	semantic	ADJ
ajst-24984	73	9	rules	rule	NOUN
ajst-24984	73	10	to	to	PART
ajst-24984	73	11	improve	improve	VERB
ajst-24984	73	12	the	the	DET
ajst-24984	73	13	method	method	NOUN
ajst-24984	73	14	of	of	ADP
ajst-24984	73	15	calculating	calculate	VERB
ajst-24984	73	16	the	the	DET
ajst-24984	73	17	sentiment	sentiment	NOUN
ajst-24984	73	18	intensity	intensity	NOUN
ajst-24984	73	19	of	of	ADP
ajst-24984	73	20	sentiment	sentiment	NOUN
ajst-24984	73	21	words	word	NOUN
ajst-24984	73	22	with	with	ADP
ajst-24984	73	23	different	different	ADJ
ajst-24984	73	24	sentiment	sentiment	NOUN
ajst-24984	73	25	categories	category	NOUN
ajst-24984	73	26	,	,	PUNCT
ajst-24984	73	27	and	and	CCONJ
ajst-24984	73	28	enhanced	enhance	VERB
ajst-24984	73	29	the	the	DET
ajst-24984	73	30	effectiveness	effectiveness	NOUN
ajst-24984	73	31	of	of	ADP
ajst-24984	73	32	sentiment	sentiment	NOUN
ajst-24984	73	33	classification	classification	NOUN
ajst-24984	73	34	.	.	PUNCT
ajst-24984	74	1	alharthi	alharthi	PROPN
ajst-24984	75	1	k	k	X
ajst-24984	75	2	a	a	DET
ajst-24984	75	3	et	et	NOUN
ajst-24984	75	4	al	al	PROPN
ajst-24984	75	5	[	[	PUNCT
ajst-24984	75	6	20	20	NUM
ajst-24984	75	7	]	]	PUNCT
ajst-24984	75	8	efficiently	efficiently	ADV
ajst-24984	75	9	constructed	construct	VERB
ajst-24984	75	10	a	a	DET
ajst-24984	75	11	reusable	reusable	ADJ
ajst-24984	75	12	sentiment	sentiment	NOUN
ajst-24984	75	13	dictionary	dictionary	NOUN
ajst-24984	75	14	on	on	ADP
ajst-24984	75	15	a	a	DET
ajst-24984	75	16	large	large	ADJ
ajst-24984	75	17	-	-	PUNCT
ajst-24984	75	18	scale	scale	NOUN
ajst-24984	75	19	system	system	NOUN
ajst-24984	75	20	log	log	NOUN
ajst-24984	75	21	and	and	CCONJ
ajst-24984	75	22	utilized	utilize	VERB
ajst-24984	75	23	the	the	DET
ajst-24984	75	24	dictionary	dictionary	NOUN
ajst-24984	75	25	to	to	PART
ajst-24984	75	26	detect	detect	VERB
ajst-24984	75	27	errors	error	NOUN
ajst-24984	75	28	and	and	CCONJ
ajst-24984	75	29	identify	identify	VERB
ajst-24984	75	30	faulty	faulty	ADJ
ajst-24984	75	31	components	component	NOUN
ajst-24984	75	32	.	.	PUNCT
ajst-24984	76	1	in	in	ADP
ajst-24984	76	2	view	view	NOUN
ajst-24984	76	3	of	of	ADP
ajst-24984	76	4	the	the	DET
ajst-24984	76	5	colloquial	colloquial	ADJ
ajst-24984	76	6	,	,	PUNCT
ajst-24984	76	7	irregular	irregular	ADJ
ajst-24984	76	8	and	and	CCONJ
ajst-24984	76	9	diverse	diverse	ADJ
ajst-24984	76	10	nature	nature	NOUN
ajst-24984	76	11	of	of	ADP
ajst-24984	76	12	english	english	ADJ
ajst-24984	76	13	social	social	ADJ
ajst-24984	76	14	media	medium	NOUN
ajst-24984	76	15	texts	text	NOUN
ajst-24984	76	16	,	,	PUNCT
ajst-24984	76	17	rao	rao	PROPN
ajst-24984	76	18	l	l	NOUN
ajst-24984	76	19	et	et	NOUN
ajst-24984	76	20	al	al	PROPN
ajst-24984	77	1	[	[	X
ajst-24984	77	2	21	21	NUM
ajst-24984	77	3	]	]	PUNCT
ajst-24984	77	4	proposed	propose	VERB
ajst-24984	77	5	an	an	DET
ajst-24984	77	6	innovative	innovative	ADJ
ajst-24984	77	7	approach	approach	NOUN
ajst-24984	77	8	to	to	ADP
ajst-24984	77	9	text	text	NOUN
ajst-24984	77	10	sentiment	sentiment	NOUN
ajst-24984	77	11	analysis	analysis	NOUN
ajst-24984	77	12	that	that	PRON
ajst-24984	77	13	incorporates	incorporate	VERB
ajst-24984	77	14	dictionary	dictionary	ADJ
ajst-24984	77	15	sentiment	sentiment	NOUN
ajst-24984	77	16	strength	strength	NOUN
ajst-24984	77	17	,	,	PUNCT
ajst-24984	77	18	expression	expression	NOUN
ajst-24984	77	19	patterns	pattern	NOUN
ajst-24984	77	20	,	,	PUNCT
ajst-24984	77	21	and	and	CCONJ
ajst-24984	77	22	optimized	optimize	VERB
ajst-24984	77	23	semantic	semantic	ADJ
ajst-24984	77	24	contextual	contextual	ADJ
ajst-24984	77	25	features	feature	NOUN
ajst-24984	77	26	and	and	CCONJ
ajst-24984	77	27	constructs	construct	VERB
ajst-24984	77	28	a	a	DET
ajst-24984	77	29	sentiment	sentiment	NOUN
ajst-24984	77	30	analysis	analysis	NOUN
ajst-24984	77	31	model	model	NOUN
ajst-24984	77	32	based	base	VERB
ajst-24984	77	33	on	on	ADP
ajst-24984	77	34	these	these	DET
ajst-24984	77	35	features	feature	NOUN
ajst-24984	77	36	.	.	PUNCT
ajst-24984	78	1	based	base	VERB
ajst-24984	78	2	on	on	ADP
ajst-24984	78	3	word	word	NOUN
ajst-24984	78	4	embedding	embed	VERB
ajst-24984	78	5	,	,	PUNCT
ajst-24984	78	6	sentiment	sentiment	NOUN
ajst-24984	78	7	dictionary	dictionary	NOUN
ajst-24984	78	8	and	and	CCONJ
ajst-24984	78	9	fuzzy	fuzzy	ADJ
ajst-24984	78	10	inference	inference	NOUN
ajst-24984	78	11	,	,	PUNCT
ajst-24984	78	12	yan	yan	PROPN
ajst-24984	78	13	r	r	NOUN
ajst-24984	78	14	et	et	NOUN
ajst-24984	78	15	al	al	PROPN
ajst-24984	79	1	[	[	X
ajst-24984	79	2	22	22	NUM
ajst-24984	79	3	]	]	PUNCT
ajst-24984	79	4	proposed	propose	VERB
ajst-24984	79	5	a	a	DET
ajst-24984	79	6	low	low	ADV
ajst-24984	79	7	-	-	PUNCT
ajst-24984	79	8	dimensional	dimensional	ADJ
ajst-24984	79	9	hybrid	hybrid	ADJ
ajst-24984	79	10	feature	feature	NOUN
ajst-24984	79	11	model	model	NOUN
ajst-24984	79	12	and	and	CCONJ
ajst-24984	79	13	a	a	DET
ajst-24984	79	14	sentimentenhanced	sentimentenhanced	ADJ
ajst-24984	79	15	inference	inference	NOUN
ajst-24984	79	16	model	model	NOUN
ajst-24984	79	17	.	.	PUNCT
ajst-24984	80	1	although	although	SCONJ
ajst-24984	80	2	the	the	DET
ajst-24984	80	3	sentiment	sentiment	NOUN
ajst-24984	80	4	dictionary	dictionary	ADJ
ajst-24984	80	5	approach	approach	NOUN
ajst-24984	80	6	is	be	AUX
ajst-24984	80	7	widely	widely	ADV
ajst-24984	80	8	used	use	VERB
ajst-24984	80	9	in	in	ADP
ajst-24984	80	10	sentiment	sentiment	NOUN
ajst-24984	80	11	analysis	analysis	NOUN
ajst-24984	80	12	,	,	PUNCT
ajst-24984	80	13	its	its	PRON
ajst-24984	80	14	effectiveness	effectiveness	NOUN
ajst-24984	80	15	depends	depend	VERB
ajst-24984	80	16	largely	largely	ADV
ajst-24984	80	17	on	on	ADP
ajst-24984	80	18	the	the	DET
ajst-24984	80	19	quality	quality	NOUN
ajst-24984	80	20	of	of	ADP
ajst-24984	80	21	the	the	DET
ajst-24984	80	22	sentiment	sentiment	NOUN
ajst-24984	80	23	dictionary	dictionary	NOUN
ajst-24984	80	24	.	.	PUNCT
ajst-24984	81	1	since	since	SCONJ
ajst-24984	81	2	the	the	DET
ajst-24984	81	3	construction	construction	NOUN
ajst-24984	81	4	of	of	ADP
ajst-24984	81	5	sentiment	sentiment	NOUN
ajst-24984	81	6	dictionaries	dictionary	NOUN
ajst-24984	81	7	often	often	ADV
ajst-24984	81	8	requires	require	VERB
ajst-24984	81	9	significant	significant	ADJ
ajst-24984	81	10	human	human	ADJ
ajst-24984	81	11	input	input	NOUN
ajst-24984	81	12	and	and	CCONJ
ajst-24984	81	13	specialized	specialized	ADJ
ajst-24984	81	14	knowledge	knowledge	NOUN
ajst-24984	81	15	,	,	PUNCT
ajst-24984	81	16	there	there	PRON
ajst-24984	81	17	are	be	VERB
ajst-24984	81	18	limitations	limitation	NOUN
ajst-24984	81	19	to	to	ADP
ajst-24984	81	20	the	the	DET
ajst-24984	81	21	process	process	NOUN
ajst-24984	81	22	of	of	ADP
ajst-24984	81	23	their	their	PRON
ajst-24984	81	24	creation	creation	NOUN
ajst-24984	81	25	.	.	PUNCT
ajst-24984	82	1	in	in	ADP
ajst-24984	82	2	addition	addition	NOUN
ajst-24984	82	3	,	,	PUNCT
ajst-24984	82	4	with	with	ADP
ajst-24984	82	5	the	the	DET
ajst-24984	82	6	development	development	NOUN
ajst-24984	82	7	of	of	ADP
ajst-24984	82	8	society	society	NOUN
ajst-24984	82	9	and	and	CCONJ
ajst-24984	82	10	the	the	DET
ajst-24984	82	11	evolution	evolution	NOUN
ajst-24984	82	12	of	of	ADP
ajst-24984	82	13	language	language	NOUN
ajst-24984	82	14	,	,	PUNCT
ajst-24984	82	15	the	the	DET
ajst-24984	82	16	existing	exist	VERB
ajst-24984	82	17	sentiment	sentiment	NOUN
ajst-24984	82	18	dictionaries	dictionary	NOUN
ajst-24984	82	19	may	may	AUX
ajst-24984	82	20	not	not	PART
ajst-24984	82	21	be	be	AUX
ajst-24984	82	22	able	able	ADJ
ajst-24984	82	23	to	to	PART
ajst-24984	82	24	cover	cover	VERB
ajst-24984	82	25	specific	specific	ADJ
ajst-24984	82	26	sentiment	sentiment	NOUN
ajst-24984	82	27	expressions	expression	NOUN
ajst-24984	82	28	in	in	ADP
ajst-24984	82	29	emerging	emerge	VERB
ajst-24984	82	30	domains	domain	NOUN
ajst-24984	82	31	,	,	PUNCT
ajst-24984	82	32	and	and	CCONJ
ajst-24984	82	33	there	there	PRON
ajst-24984	82	34	are	be	VERB
ajst-24984	82	35	domain	domain	NOUN
ajst-24984	82	36	adaptation	adaptation	NOUN
ajst-24984	82	37	limitations	limitation	NOUN
ajst-24984	82	38	.	.	PUNCT
ajst-24984	83	1	the	the	DET
ajst-24984	83	2	rapid	rapid	ADJ
ajst-24984	83	3	development	development	NOUN
ajst-24984	83	4	of	of	ADP
ajst-24984	83	5	the	the	DET
ajst-24984	83	6	internet	internet	NOUN
ajst-24984	83	7	has	have	AUX
ajst-24984	83	8	led	lead	VERB
ajst-24984	83	9	to	to	ADP
ajst-24984	83	10	the	the	DET
ajst-24984	83	11	emergence	emergence	NOUN
ajst-24984	83	12	of	of	ADP
ajst-24984	83	13	new	new	ADJ
ajst-24984	83	14	words	word	NOUN
ajst-24984	83	15	,	,	PUNCT
ajst-24984	83	16	and	and	CCONJ
ajst-24984	83	17	the	the	DET
ajst-24984	83	18	sentiment	sentiment	NOUN
ajst-24984	83	19	tendencies	tendency	NOUN
ajst-24984	83	20	of	of	ADP
ajst-24984	83	21	these	these	DET
ajst-24984	83	22	new	new	ADJ
ajst-24984	83	23	words	word	NOUN
ajst-24984	83	24	may	may	AUX
ajst-24984	83	25	not	not	PART
ajst-24984	83	26	be	be	AUX
ajst-24984	83	27	reflected	reflect	VERB
ajst-24984	83	28	in	in	ADP
ajst-24984	83	29	sentiment	sentiment	NOUN
ajst-24984	83	30	dictionaries	dictionary	NOUN
ajst-24984	83	31	in	in	ADP
ajst-24984	83	32	a	a	DET
ajst-24984	83	33	timely	timely	ADJ
ajst-24984	83	34	manner	manner	NOUN
ajst-24984	83	35	,	,	PUNCT
ajst-24984	83	36	thus	thus	ADV
ajst-24984	83	37	affecting	affect	VERB
ajst-24984	83	38	the	the	DET
ajst-24984	83	39	accuracy	accuracy	NOUN
ajst-24984	83	40	of	of	ADP
ajst-24984	83	41	sentiment	sentiment	NOUN
ajst-24984	83	42	analysis	analysis	NOUN
ajst-24984	83	43	.	.	PUNCT
ajst-24984	84	1	the	the	DET
ajst-24984	84	2	updating	update	VERB
ajst-24984	84	3	speed	speed	NOUN
ajst-24984	84	4	of	of	ADP
ajst-24984	84	5	the	the	DET
ajst-24984	84	6	sentiment	sentiment	NOUN
ajst-24984	84	7	dictionary	dictionary	NOUN
ajst-24984	84	8	can	can	AUX
ajst-24984	84	9	not	not	PART
ajst-24984	84	10	keep	keep	VERB
ajst-24984	84	11	up	up	ADP
ajst-24984	84	12	with	with	ADP
ajst-24984	84	13	the	the	DET
ajst-24984	84	14	development	development	NOUN
ajst-24984	84	15	of	of	ADP
ajst-24984	84	16	language	language	NOUN
ajst-24984	84	17	,	,	PUNCT
ajst-24984	84	18	resulting	result	VERB
ajst-24984	84	19	in	in	ADP
ajst-24984	84	20	some	some	DET
ajst-24984	84	21	new	new	ADJ
ajst-24984	84	22	sentiment	sentiment	NOUN
ajst-24984	84	23	expressions	expression	NOUN
ajst-24984	84	24	not	not	PART
ajst-24984	84	25	finding	find	VERB
ajst-24984	84	26	corresponding	corresponding	ADJ
ajst-24984	84	27	entries	entry	NOUN
ajst-24984	84	28	in	in	ADP
ajst-24984	84	29	the	the	DET
ajst-24984	84	30	dictionary	dictionary	NOUN
ajst-24984	84	31	,	,	PUNCT
ajst-24984	84	32	which	which	PRON
ajst-24984	84	33	makes	make	VERB
ajst-24984	84	34	these	these	DET
ajst-24984	84	35	expressions	expression	NOUN
ajst-24984	84	36	neglected	neglect	VERB
ajst-24984	84	37	or	or	CCONJ
ajst-24984	84	38	misjudged	misjudge	VERB
ajst-24984	84	39	in	in	ADP
ajst-24984	84	40	the	the	DET
ajst-24984	84	41	process	process	NOUN
ajst-24984	84	42	of	of	ADP
ajst-24984	84	43	sentiment	sentiment	NOUN
ajst-24984	84	44	analysis	analysis	NOUN
ajst-24984	84	45	.	.	PUNCT
ajst-24984	85	1	another	another	DET
ajst-24984	85	2	problem	problem	NOUN
ajst-24984	85	3	is	be	AUX
ajst-24984	85	4	that	that	DET
ajst-24984	85	5	sentiment	sentiment	NOUN
ajst-24984	85	6	dictionary	dictionary	NOUN
ajst-24984	85	7	-	-	PUNCT
ajst-24984	85	8	based	base	VERB
ajst-24984	85	9	approaches	approach	NOUN
ajst-24984	85	10	usually	usually	ADV
ajst-24984	85	11	do	do	AUX
ajst-24984	85	12	not	not	PART
ajst-24984	85	13	consider	consider	VERB
ajst-24984	85	14	the	the	DET
ajst-24984	85	15	contextual	contextual	ADJ
ajst-24984	85	16	connection	connection	NOUN
ajst-24984	85	17	of	of	ADP
ajst-24984	85	18	words	word	NOUN
ajst-24984	85	19	,	,	PUNCT
ajst-24984	85	20	resulting	result	VERB
ajst-24984	85	21	in	in	ADP
ajst-24984	85	22	the	the	DET
ajst-24984	85	23	sentiment	sentiment	NOUN
ajst-24984	85	24	value	value	NOUN
ajst-24984	85	25	of	of	ADP
ajst-24984	85	26	each	each	DET
ajst-24984	85	27	word	word	NOUN
ajst-24984	85	28	remaining	remain	VERB
ajst-24984	85	29	static	static	NOUN
ajst-24984	85	30	in	in	ADP
ajst-24984	85	31	different	different	ADJ
ajst-24984	85	32	articles	article	NOUN
ajst-24984	85	33	or	or	CCONJ
ajst-24984	85	34	sentences	sentence	NOUN
ajst-24984	85	35	,	,	PUNCT
ajst-24984	85	36	and	and	CCONJ
ajst-24984	85	37	not	not	PART
ajst-24984	85	38	being	be	AUX
ajst-24984	85	39	able	able	ADJ
ajst-24984	85	40	to	to	PART
ajst-24984	85	41	dynamically	dynamically	ADV
ajst-24984	85	42	adapt	adapt	VERB
ajst-24984	85	43	to	to	ADP
ajst-24984	85	44	contextual	contextual	ADJ
ajst-24984	85	45	changes	change	NOUN
ajst-24984	85	46	.	.	PUNCT
ajst-24984	86	1	this	this	DET
ajst-24984	86	2	approach	approach	NOUN
ajst-24984	86	3	has	have	VERB
ajst-24984	86	4	weak	weak	ADJ
ajst-24984	86	5	generalization	generalization	NOUN
ajst-24984	86	6	ability	ability	NOUN
ajst-24984	86	7	and	and	CCONJ
ajst-24984	86	8	poor	poor	ADJ
ajst-24984	86	9	system	system	NOUN
ajst-24984	86	10	migration	migration	NOUN
ajst-24984	86	11	,	,	PUNCT
ajst-24984	86	12	making	make	VERB
ajst-24984	86	13	it	it	PRON
ajst-24984	86	14	difficult	difficult	ADJ
ajst-24984	86	15	to	to	PART
ajst-24984	86	16	adapt	adapt	VERB
ajst-24984	86	17	to	to	ADP
ajst-24984	86	18	the	the	DET
ajst-24984	86	19	needs	need	NOUN
ajst-24984	86	20	of	of	ADP
ajst-24984	86	21	sentiment	sentiment	NOUN
ajst-24984	86	22	analysis	analysis	NOUN
ajst-24984	86	23	in	in	ADP
ajst-24984	86	24	different	different	ADJ
ajst-24984	86	25	domains	domain	NOUN
ajst-24984	86	26	and	and	CCONJ
ajst-24984	86	27	contexts	context	NOUN
ajst-24984	86	28	.	.	PUNCT
ajst-24984	87	1	in	in	ADP
ajst-24984	87	2	view	view	NOUN
ajst-24984	87	3	of	of	ADP
ajst-24984	87	4	the	the	DET
ajst-24984	87	5	above	above	ADJ
ajst-24984	87	6	limitations	limitation	NOUN
ajst-24984	87	7	,	,	PUNCT
ajst-24984	87	8	more	more	ADJ
ajst-24984	87	9	and	and	CCONJ
ajst-24984	87	10	more	more	ADJ
ajst-24984	87	11	scholars	scholar	NOUN
ajst-24984	87	12	are	be	AUX
ajst-24984	87	13	turning	turn	VERB
ajst-24984	87	14	to	to	ADP
ajst-24984	87	15	machine	machine	NOUN
ajst-24984	87	16	learning	learning	NOUN
ajst-24984	87	17	-	-	PUNCT
ajst-24984	87	18	based	base	VERB
ajst-24984	87	19	sentiment	sentiment	NOUN
ajst-24984	87	20	analysis	analysis	NOUN
ajst-24984	87	21	methods	method	NOUN
ajst-24984	87	22	.	.	PUNCT
ajst-24984	88	1	machine	machine	NOUN
ajst-24984	88	2	learning	learn	VERB
ajst-24984	88	3	methods	method	NOUN
ajst-24984	88	4	are	be	AUX
ajst-24984	88	5	able	able	ADJ
ajst-24984	88	6	to	to	PART
ajst-24984	88	7	automatically	automatically	ADV
ajst-24984	88	8	learn	learn	VERB
ajst-24984	88	9	sentiment	sentiment	NOUN
ajst-24984	88	10	features	feature	VERB
ajst-24984	88	11	in	in	ADP
ajst-24984	88	12	text	text	NOUN
ajst-24984	88	13	by	by	ADP
ajst-24984	88	14	training	training	NOUN
ajst-24984	88	15	on	on	ADP
ajst-24984	88	16	a	a	DET
ajst-24984	88	17	large	large	ADJ
ajst-24984	88	18	amount	amount	NOUN
ajst-24984	88	19	of	of	ADP
ajst-24984	88	20	labeled	label	VERB
ajst-24984	88	21	data	datum	NOUN
ajst-24984	88	22	,	,	PUNCT
ajst-24984	88	23	thus	thus	ADV
ajst-24984	88	24	improving	improve	VERB
ajst-24984	88	25	the	the	DET
ajst-24984	88	26	accuracy	accuracy	NOUN
ajst-24984	88	27	and	and	CCONJ
ajst-24984	88	28	adaptability	adaptability	NOUN
ajst-24984	88	29	of	of	ADP
ajst-24984	88	30	sentiment	sentiment	NOUN
ajst-24984	88	31	analysis	analysis	NOUN
ajst-24984	88	32	.	.	PUNCT
ajst-24984	89	1	these	these	DET
ajst-24984	89	2	methods	method	NOUN
ajst-24984	89	3	are	be	AUX
ajst-24984	89	4	usually	usually	ADV
ajst-24984	89	5	able	able	ADJ
ajst-24984	89	6	to	to	PART
ajst-24984	89	7	better	well	ADV
ajst-24984	89	8	handle	handle	VERB
ajst-24984	89	9	contextual	contextual	ADJ
ajst-24984	89	10	information	information	NOUN
ajst-24984	89	11	,	,	PUNCT
ajst-24984	89	12	dynamically	dynamically	ADV
ajst-24984	89	13	adjust	adjust	VERB
ajst-24984	89	14	the	the	DET
ajst-24984	89	15	sentiment	sentiment	NOUN
ajst-24984	89	16	values	value	NOUN
ajst-24984	89	17	of	of	ADP
ajst-24984	89	18	words	word	NOUN
ajst-24984	89	19	,	,	PUNCT
ajst-24984	89	20	and	and	CCONJ
ajst-24984	89	21	improve	improve	VERB
ajst-24984	89	22	the	the	DET
ajst-24984	89	23	generalization	generalization	NOUN
ajst-24984	89	24	ability	ability	NOUN
ajst-24984	89	25	and	and	CCONJ
ajst-24984	89	26	system	system	NOUN
ajst-24984	89	27	migration	migration	NOUN
ajst-24984	89	28	of	of	ADP
ajst-24984	89	29	sentiment	sentiment	NOUN
ajst-24984	89	30	analysis	analysis	NOUN
ajst-24984	89	31	.	.	PUNCT
ajst-24984	90	1	3.2	3.2	NUM
ajst-24984	90	2	.	.	PUNCT
ajst-24984	90	3	machine	machine	NOUN
ajst-24984	90	4	learning	learning	NOUN
ajst-24984	90	5	based	base	VERB
ajst-24984	90	6	approaches	approach	VERB
ajst-24984	90	7	the	the	DET
ajst-24984	90	8	process	process	NOUN
ajst-24984	90	9	of	of	ADP
ajst-24984	90	10	sentiment	sentiment	NOUN
ajst-24984	90	11	analysis	analysis	NOUN
ajst-24984	90	12	based	base	VERB
ajst-24984	90	13	on	on	ADP
ajst-24984	90	14	traditional	traditional	ADJ
ajst-24984	90	15	machine	machine	NOUN
ajst-24984	90	16	learning	learning	NOUN
ajst-24984	90	17	methods	method	NOUN
ajst-24984	90	18	usually	usually	ADV
ajst-24984	90	19	involves	involve	VERB
ajst-24984	90	20	the	the	DET
ajst-24984	90	21	following	follow	VERB
ajst-24984	90	22	steps	step	NOUN
ajst-24984	90	23	:	:	PUNCT
ajst-24984	90	24	first	first	ADV
ajst-24984	90	25	,	,	PUNCT
ajst-24984	90	26	sentiment	sentiment	NOUN
ajst-24984	90	27	words	word	NOUN
ajst-24984	90	28	are	be	AUX
ajst-24984	90	29	identified	identify	VERB
ajst-24984	90	30	and	and	CCONJ
ajst-24984	90	31	selected	select	VERB
ajst-24984	90	32	as	as	ADP
ajst-24984	90	33	features	feature	NOUN
ajst-24984	90	34	from	from	ADP
ajst-24984	90	35	the	the	DET
ajst-24984	90	36	text	text	NOUN
ajst-24984	90	37	,	,	PUNCT
ajst-24984	90	38	which	which	PRON
ajst-24984	90	39	are	be	AUX
ajst-24984	90	40	the	the	DET
ajst-24984	90	41	basic	basic	ADJ
ajst-24984	90	42	units	unit	NOUN
ajst-24984	90	43	for	for	ADP
ajst-24984	90	44	expressing	express	VERB
ajst-24984	90	45	sentiment	sentiment	NOUN
ajst-24984	90	46	and	and	CCONJ
ajst-24984	90	47	are	be	AUX
ajst-24984	90	48	crucial	crucial	ADJ
ajst-24984	90	49	for	for	ADP
ajst-24984	90	50	sentiment	sentiment	NOUN
ajst-24984	90	51	recognition	recognition	NOUN
ajst-24984	90	52	and	and	CCONJ
ajst-24984	90	53	classification	classification	NOUN
ajst-24984	90	54	.	.	PUNCT
ajst-24984	91	1	next	next	ADJ
ajst-24984	91	2	,	,	PUNCT
ajst-24984	91	3	classical	classical	ADJ
ajst-24984	91	4	machine	machine	NOUN
ajst-24984	91	5	learning	learn	VERB
ajst-24984	91	6	algorithms	algorithm	NOUN
ajst-24984	91	7	such	such	ADJ
ajst-24984	91	8	as	as	ADP
ajst-24984	91	9	logistic	logistic	ADJ
ajst-24984	91	10	regression	regression	NOUN
ajst-24984	91	11	,	,	PUNCT
ajst-24984	91	12	plain	plain	ADJ
ajst-24984	91	13	bayes	bayes	NOUN
ajst-24984	91	14	,	,	PUNCT
ajst-24984	91	15	and	and	CCONJ
ajst-24984	91	16	support	support	VERB
ajst-24984	91	17	vector	vector	NOUN
ajst-24984	91	18	machines	machine	NOUN
ajst-24984	91	19	are	be	AUX
ajst-24984	91	20	used	use	VERB
ajst-24984	91	21	to	to	PART
ajst-24984	91	22	train	train	VERB
ajst-24984	91	23	on	on	ADP
ajst-24984	91	24	the	the	DET
ajst-24984	91	25	already	already	ADV
ajst-24984	91	26	labeled	label	VERB
ajst-24984	91	27	feature	feature	NOUN
ajst-24984	91	28	dataset	dataset	NOUN
ajst-24984	91	29	.	.	PUNCT
ajst-24984	92	1	these	these	DET
ajst-24984	92	2	algorithms	algorithm	NOUN
ajst-24984	92	3	are	be	AUX
ajst-24984	92	4	able	able	ADJ
ajst-24984	92	5	to	to	PART
ajst-24984	92	6	learn	learn	VERB
ajst-24984	92	7	the	the	DET
ajst-24984	92	8	complex	complex	ADJ
ajst-24984	92	9	relationship	relationship	NOUN
ajst-24984	92	10	between	between	ADP
ajst-24984	92	11	sentiment	sentiment	NOUN
ajst-24984	92	12	words	word	NOUN
ajst-24984	92	13	and	and	CCONJ
ajst-24984	92	14	sentiment	sentiment	NOUN
ajst-24984	92	15	tendencies	tendency	NOUN
ajst-24984	92	16	and	and	CCONJ
ajst-24984	92	17	construct	construct	VERB
ajst-24984	92	18	a	a	DET
ajst-24984	92	19	sentiment	sentiment	NOUN
ajst-24984	92	20	classification	classification	NOUN
ajst-24984	92	21	model	model	NOUN
ajst-24984	92	22	.	.	PUNCT
ajst-24984	93	1	finally	finally	ADV
ajst-24984	93	2	,	,	PUNCT
ajst-24984	93	3	the	the	DET
ajst-24984	93	4	trained	train	VERB
ajst-24984	93	5	models	model	NOUN
ajst-24984	93	6	are	be	AUX
ajst-24984	93	7	utilized	utilize	VERB
ajst-24984	93	8	to	to	PART
ajst-24984	93	9	predict	predict	VERB
ajst-24984	93	10	sentiment	sentiment	NOUN
ajst-24984	93	11	tendencies	tendency	NOUN
ajst-24984	93	12	on	on	ADP
ajst-24984	93	13	the	the	DET
ajst-24984	93	14	text	text	NOUN
ajst-24984	93	15	data	datum	NOUN
ajst-24984	93	16	within	within	ADP
ajst-24984	93	17	the	the	DET
ajst-24984	93	18	test	test	NOUN
ajst-24984	93	19	set	set	NOUN
ajst-24984	93	20	.	.	PUNCT
ajst-24984	94	1	this	this	DET
ajst-24984	94	2	prediction	prediction	NOUN
ajst-24984	94	3	process	process	NOUN
ajst-24984	94	4	relies	rely	VERB
ajst-24984	94	5	on	on	ADP
ajst-24984	94	6	the	the	DET
ajst-24984	94	7	quality	quality	NOUN
ajst-24984	94	8	of	of	ADP
ajst-24984	94	9	the	the	DET
ajst-24984	94	10	selection	selection	NOUN
ajst-24984	94	11	of	of	ADP
ajst-24984	94	12	the	the	DET
ajst-24984	94	13	text	text	NOUN
ajst-24984	94	14	data	datum	NOUN
ajst-24984	94	15	used	use	VERB
ajst-24984	94	16	in	in	ADP
ajst-24984	94	17	the	the	DET
ajst-24984	94	18	training	training	NOUN
ajst-24984	94	19	phase	phase	NOUN
ajst-24984	94	20	as	as	ADV
ajst-24984	94	21	well	well	ADV
ajst-24984	94	22	as	as	ADP
ajst-24984	94	23	the	the	DET
ajst-24984	94	24	accuracy	accuracy	NOUN
ajst-24984	94	25	of	of	ADP
ajst-24984	94	26	the	the	DET
ajst-24984	94	27	sentiment	sentiment	NOUN
ajst-24984	94	28	labeling	labeling	NOUN
ajst-24984	94	29	.	.	PUNCT
ajst-24984	95	1	therefore	therefore	ADV
ajst-24984	95	2	,	,	PUNCT
ajst-24984	95	3	in	in	ADP
ajst-24984	95	4	order	order	NOUN
ajst-24984	95	5	to	to	PART
ajst-24984	95	6	obtain	obtain	VERB
ajst-24984	95	7	the	the	DET
ajst-24984	95	8	desired	desire	VERB
ajst-24984	95	9	classification	classification	NOUN
ajst-24984	95	10	results	result	NOUN
ajst-24984	95	11	,	,	PUNCT
ajst-24984	95	12	not	not	PART
ajst-24984	95	13	only	only	ADV
ajst-24984	95	14	do	do	AUX
ajst-24984	95	15	we	we	PRON
ajst-24984	95	16	need	need	VERB
ajst-24984	95	17	to	to	PART
ajst-24984	95	18	carefully	carefully	ADV
ajst-24984	95	19	select	select	VERB
ajst-24984	95	20	the	the	DET
ajst-24984	95	21	training	training	NOUN
ajst-24984	95	22	texts	text	NOUN
ajst-24984	95	23	to	to	PART
ajst-24984	95	24	ensure	ensure	VERB
ajst-24984	95	25	their	their	PRON
ajst-24984	95	26	representativeness	representativeness	NOUN
ajst-24984	95	27	,	,	PUNCT
ajst-24984	95	28	but	but	CCONJ
ajst-24984	95	29	we	we	PRON
ajst-24984	95	30	also	also	ADV
ajst-24984	95	31	need	need	VERB
ajst-24984	95	32	to	to	PART
ajst-24984	95	33	ensure	ensure	VERB
ajst-24984	95	34	the	the	DET
ajst-24984	95	35	accuracy	accuracy	NOUN
ajst-24984	95	36	of	of	ADP
ajst-24984	95	37	the	the	DET
ajst-24984	95	38	sentiment	sentiment	NOUN
ajst-24984	95	39	labeling	labeling	NOUN
ajst-24984	95	40	.	.	PUNCT
ajst-24984	96	1	these	these	DET
ajst-24984	96	2	two	two	NUM
ajst-24984	96	3	factors	factor	NOUN
ajst-24984	96	4	together	together	ADV
ajst-24984	96	5	determine	determine	VERB
ajst-24984	96	6	the	the	DET
ajst-24984	96	7	performance	performance	NOUN
ajst-24984	96	8	of	of	ADP
ajst-24984	96	9	the	the	DET
ajst-24984	96	10	final	final	ADJ
ajst-24984	96	11	sentiment	sentiment	NOUN
ajst-24984	96	12	classification	classification	NOUN
ajst-24984	96	13	model	model	NOUN
ajst-24984	96	14	.	.	PUNCT
ajst-24984	97	1	pang	pang	PROPN
ajst-24984	97	2	b	b	PROPN
ajst-24984	97	3	et	et	PROPN
ajst-24984	97	4	al	al	PROPN
ajst-24984	98	1	[	[	X
ajst-24984	98	2	23	23	NUM
ajst-24984	98	3	]	]	PUNCT
ajst-24984	98	4	pioneered	pioneer	VERB
ajst-24984	98	5	the	the	DET
ajst-24984	98	6	application	application	NOUN
ajst-24984	98	7	of	of	ADP
ajst-24984	98	8	machine	machine	NOUN
ajst-24984	98	9	learning	learning	NOUN
ajst-24984	98	10	in	in	ADP
ajst-24984	98	11	the	the	DET
ajst-24984	98	12	field	field	NOUN
ajst-24984	98	13	of	of	ADP
ajst-24984	98	14	sentiment	sentiment	NOUN
ajst-24984	98	15	analysis	analysis	NOUN
ajst-24984	98	16	by	by	ADP
ajst-24984	98	17	using	use	VERB
ajst-24984	98	18	machine	machine	NOUN
ajst-24984	98	19	learning	learning	NOUN
ajst-24984	98	20	methods	method	NOUN
ajst-24984	98	21	to	to	PART
ajst-24984	98	22	classify	classify	VERB
ajst-24984	98	23	the	the	DET
ajst-24984	98	24	sentiment	sentiment	NOUN
ajst-24984	98	25	of	of	ADP
ajst-24984	98	26	movie	movie	NOUN
ajst-24984	98	27	reviews	review	NOUN
ajst-24984	98	28	.	.	PUNCT
ajst-24984	99	1	aiming	aim	VERB
ajst-24984	99	2	at	at	ADP
ajst-24984	99	3	the	the	DET
ajst-24984	99	4	problem	problem	NOUN
ajst-24984	99	5	of	of	ADP
ajst-24984	99	6	over	over	ADP
ajst-24984	99	7	-	-	PUNCT
ajst-24984	99	8	reliance	reliance	NOUN
ajst-24984	99	9	on	on	ADP
ajst-24984	99	10	sentiment	sentiment	NOUN
ajst-24984	99	11	word	word	NOUN
ajst-24984	99	12	statistics	statistic	NOUN
ajst-24984	99	13	for	for	ADP
ajst-24984	99	14	text	text	NOUN
ajst-24984	99	15	sentiment	sentiment	NOUN
ajst-24984	99	16	analysis	analysis	NOUN
ajst-24984	99	17	,	,	PUNCT
ajst-24984	99	18	han	han	PROPN
ajst-24984	99	19	k	k	PROPN
ajst-24984	99	20	x	x	PUNCT
ajst-24984	99	21	et	et	PROPN
ajst-24984	99	22	al	al	PROPN
ajst-24984	100	1	[	[	X
ajst-24984	100	2	24	24	NUM
ajst-24984	100	3	]	]	PUNCT
ajst-24984	100	4	proposed	propose	VERB
ajst-24984	100	5	an	an	DET
ajst-24984	100	6	improved	improved	ADJ
ajst-24984	100	7	support	support	NOUN
ajst-24984	100	8	vector	vector	NOUN
ajst-24984	100	9	machine	machine	NOUN
ajst-24984	100	10	(	(	PUNCT
ajst-24984	100	11	fk	fk	INTJ
ajst-24984	100	12	-	-	PUNCT
ajst-24984	100	13	svm	svm	NOUN
ajst-24984	100	14	)	)	PUNCT
ajst-24984	100	15	method	method	NOUN
ajst-24984	100	16	based	base	VERB
ajst-24984	100	17	on	on	ADP
ajst-24984	100	18	fisher	fisher	PROPN
ajst-24984	100	19	kernel	kernel	PROPN
ajst-24984	100	20	function	function	VERB
ajst-24984	100	21	to	to	PART
ajst-24984	100	22	fully	fully	ADV
ajst-24984	100	23	consider	consider	VERB
ajst-24984	100	24	lexical	lexical	ADJ
ajst-24984	100	25	and	and	CCONJ
ajst-24984	100	26	potential	potential	ADJ
ajst-24984	100	27	semantic	semantic	ADJ
ajst-24984	100	28	information	information	NOUN
ajst-24984	100	29	in	in	ADP
ajst-24984	100	30	text	text	NOUN
ajst-24984	100	31	.	.	PUNCT
ajst-24984	101	1	without	without	ADP
ajst-24984	101	2	relying	rely	VERB
ajst-24984	101	3	on	on	ADP
ajst-24984	101	4	manual	manual	ADJ
ajst-24984	101	5	labeling	labeling	NOUN
ajst-24984	101	6	,	,	PUNCT
ajst-24984	101	7	wang	wang	PROPN
ajst-24984	101	8	y	y	PROPN
ajst-24984	101	9	et	et	PROPN
ajst-24984	101	10	al	al	PROPN
ajst-24984	102	1	[	[	X
ajst-24984	102	2	25	25	NUM
ajst-24984	102	3	]	]	PUNCT
ajst-24984	102	4	proposed	propose	VERB
ajst-24984	102	5	a	a	DET
ajst-24984	102	6	new	new	ADJ
ajst-24984	102	7	aspect	aspect	NOUN
ajst-24984	102	8	-	-	PUNCT
ajst-24984	102	9	based	base	VERB
ajst-24984	102	10	sentiment	sentiment	NOUN
ajst-24984	102	11	analysis	analysis	NOUN
ajst-24984	102	12	method	method	NOUN
ajst-24984	102	13	based	base	VERB
ajst-24984	102	14	on	on	ADP
ajst-24984	102	15	an	an	DET
ajst-24984	102	16	incremental	incremental	ADJ
ajst-24984	102	17	machine	machine	NOUN
ajst-24984	102	18	learning	learning	NOUN
ajst-24984	102	19	paradigm	paradigm	NOUN
ajst-24984	102	20	to	to	PART
ajst-24984	102	21	achieve	achieve	VERB
ajst-24984	102	22	accurate	accurate	ADJ
ajst-24984	102	23	machine	machine	NOUN
ajst-24984	102	24	labeling	labeling	NOUN
ajst-24984	102	25	.	.	PUNCT
ajst-24984	103	1	kurniawan	kurniawan	PROPN
ajst-24984	103	2	c	c	NOUN
ajst-24984	103	3	et	et	PROPN
ajst-24984	103	4	al	al	PROPN
ajst-24984	104	1	[	[	X
ajst-24984	104	2	26	26	NUM
ajst-24984	104	3	]	]	PUNCT
ajst-24984	104	4	,	,	PUNCT
ajst-24984	104	5	on	on	ADP
ajst-24984	104	6	the	the	DET
ajst-24984	104	7	other	other	ADJ
ajst-24984	104	8	hand	hand	NOUN
ajst-24984	104	9	,	,	PUNCT
ajst-24984	104	10	used	use	VERB
ajst-24984	104	11	a	a	DET
ajst-24984	104	12	support	support	NOUN
ajst-24984	104	13	vector	vector	NOUN
ajst-24984	104	14	machine	machine	NOUN
ajst-24984	104	15	approach	approach	NOUN
ajst-24984	104	16	for	for	ADP
ajst-24984	104	17	sentiment	sentiment	NOUN
ajst-24984	104	18	analysis	analysis	NOUN
ajst-24984	104	19	,	,	PUNCT
ajst-24984	104	20	which	which	PRON
ajst-24984	104	21	was	be	AUX
ajst-24984	104	22	able	able	ADJ
ajst-24984	104	23	to	to	PART
ajst-24984	104	24	categorize	categorize	VERB
ajst-24984	104	25	student	student	NOUN
ajst-24984	104	26	responses	response	NOUN
ajst-24984	104	27	based	base	VERB
ajst-24984	104	28	on	on	ADP
ajst-24984	104	29	positive	positive	ADJ
ajst-24984	104	30	and	and	CCONJ
ajst-24984	104	31	negative	negative	ADJ
ajst-24984	104	32	polarities	polarity	NOUN
ajst-24984	104	33	.	.	PUNCT
ajst-24984	105	1	these	these	DET
ajst-24984	105	2	studies	study	NOUN
ajst-24984	105	3	demonstrate	demonstrate	VERB
ajst-24984	105	4	the	the	DET
ajst-24984	105	5	versatility	versatility	NOUN
ajst-24984	105	6	and	and	CCONJ
ajst-24984	105	7	effectiveness	effectiveness	NOUN
ajst-24984	105	8	of	of	ADP
ajst-24984	105	9	machine	machine	NOUN
ajst-24984	105	10	learning	learn	VERB
ajst-24984	105	11	methods	method	NOUN
ajst-24984	105	12	in	in	ADP
ajst-24984	105	13	the	the	DET
ajst-24984	105	14	field	field	NOUN
ajst-24984	105	15	of	of	ADP
ajst-24984	105	16	sentiment	sentiment	NOUN
ajst-24984	105	17	analysis	analysis	NOUN
ajst-24984	105	18	.	.	PUNCT
ajst-24984	106	1	the	the	DET
ajst-24984	106	2	effectiveness	effectiveness	NOUN
ajst-24984	106	3	of	of	ADP
ajst-24984	106	4	machine	machine	NOUN
ajst-24984	106	5	learning	learn	VERB
ajst-24984	106	6	in	in	ADP
ajst-24984	106	7	sentiment	sentiment	NOUN
ajst-24984	106	8	analysis	analysis	NOUN
ajst-24984	106	9	is	be	AUX
ajst-24984	106	10	highly	highly	ADV
ajst-24984	106	11	dependent	dependent	ADJ
ajst-24984	106	12	on	on	ADP
ajst-24984	106	13	the	the	DET
ajst-24984	106	14	quality	quality	NOUN
ajst-24984	106	15	of	of	ADP
ajst-24984	106	16	feature	feature	NOUN
ajst-24984	106	17	representation	representation	NOUN
ajst-24984	106	18	,	,	PUNCT
ajst-24984	106	19	and	and	CCONJ
ajst-24984	106	20	the	the	DET
ajst-24984	106	21	complexity	complexity	NOUN
ajst-24984	106	22	of	of	ADP
ajst-24984	106	23	feature	feature	NOUN
ajst-24984	106	24	engineering	engineer	VERB
ajst-24984	106	25	itself	itself	PRON
ajst-24984	106	26	often	often	ADV
ajst-24984	106	27	leads	lead	VERB
ajst-24984	106	28	to	to	ADP
ajst-24984	106	29	difficulties	difficulty	NOUN
ajst-24984	106	30	in	in	ADP
ajst-24984	106	31	obtaining	obtain	VERB
ajst-24984	106	32	ideal	ideal	ADJ
ajst-24984	106	33	classification	classification	NOUN
ajst-24984	106	34	results	result	NOUN
ajst-24984	106	35	.	.	PUNCT
ajst-24984	107	1	in	in	ADP
ajst-24984	107	2	order	order	NOUN
ajst-24984	107	3	to	to	PART
ajst-24984	107	4	overcome	overcome	VERB
ajst-24984	107	5	the	the	DET
ajst-24984	107	6	limitations	limitation	NOUN
ajst-24984	107	7	of	of	ADP
ajst-24984	107	8	a	a	DET
ajst-24984	107	9	single	single	ADJ
ajst-24984	107	10	machine	machine	NOUN
ajst-24984	107	11	learning	learning	NOUN
ajst-24984	107	12	approach	approach	NOUN
ajst-24984	107	13	,	,	PUNCT
ajst-24984	107	14	many	many	ADJ
ajst-24984	107	15	researchers	researcher	NOUN
ajst-24984	107	16	and	and	CCONJ
ajst-24984	107	17	scholars	scholar	NOUN
ajst-24984	107	18	have	have	AUX
ajst-24984	107	19	begun	begin	VERB
ajst-24984	107	20	to	to	PART
ajst-24984	107	21	explore	explore	VERB
ajst-24984	107	22	the	the	DET
ajst-24984	107	23	combination	combination	NOUN
ajst-24984	107	24	of	of	ADP
ajst-24984	107	25	sentiment	sentiment	NOUN
ajst-24984	107	26	dictionaries	dictionary	NOUN
ajst-24984	107	27	and	and	CCONJ
ajst-24984	107	28	machine	machine	NOUN
ajst-24984	107	29	learning	learn	VERB
ajst-24984	107	30	in	in	ADP
ajst-24984	107	31	order	order	NOUN
ajst-24984	107	32	to	to	PART
ajst-24984	107	33	construct	construct	VERB
ajst-24984	107	34	novel	novel	ADJ
ajst-24984	107	35	sentiment	sentiment	NOUN
ajst-24984	107	36	analysis	analysis	NOUN
ajst-24984	107	37	models	model	NOUN
ajst-24984	107	38	.	.	PUNCT
ajst-24984	108	1	li	li	PROPN
ajst-24984	108	2	z	z	PROPN
ajst-24984	108	3	et	et	PROPN
ajst-24984	108	4	al	al	PROPN
ajst-24984	109	1	[	[	X
ajst-24984	109	2	27	27	NUM
ajst-24984	109	3	]	]	PUNCT
ajst-24984	109	4	constructed	construct	VERB
ajst-24984	109	5	a	a	DET
ajst-24984	109	6	sentiment	sentiment	NOUN
ajst-24984	109	7	dictionary	dictionary	NOUN
ajst-24984	109	8	for	for	ADP
ajst-24984	109	9	popup	popup	NOUN
ajst-24984	109	10	comments	comment	NOUN
ajst-24984	109	11	and	and	CCONJ
ajst-24984	109	12	proposed	propose	VERB
ajst-24984	109	13	a	a	DET
ajst-24984	109	14	new	new	ADJ
ajst-24984	109	15	sentiment	sentiment	NOUN
ajst-24984	109	16	analysis	analysis	NOUN
ajst-24984	109	17	method	method	NOUN
ajst-24984	109	18	that	that	PRON
ajst-24984	109	19	combines	combine	VERB
ajst-24984	109	20	a	a	DET
ajst-24984	109	21	sentiment	sentiment	NOUN
ajst-24984	109	22	dictionary	dictionary	NOUN
ajst-24984	109	23	and	and	CCONJ
ajst-24984	109	24	a	a	DET
ajst-24984	109	25	plain	plain	ADJ
ajst-24984	109	26	bayesian	bayesian	NOUN
ajst-24984	109	27	model	model	NOUN
ajst-24984	109	28	to	to	PART
ajst-24984	109	29	analyze	analyze	VERB
ajst-24984	109	30	pop	pop	VERB
ajst-24984	109	31	-	-	PUNCT
ajst-24984	109	32	up	up	ADP
ajst-24984	109	33	comments	comment	NOUN
ajst-24984	109	34	.	.	PUNCT
ajst-24984	110	1	li	li	PROPN
ajst-24984	110	2	h	h	PROPN
ajst-24984	110	3	et	et	PROPN
ajst-24984	110	4	al	al	PROPN
ajst-24984	111	1	[	[	X
ajst-24984	111	2	28	28	NUM
ajst-24984	111	3	]	]	PUNCT
ajst-24984	111	4	proposed	propose	VERB
ajst-24984	111	5	a	a	DET
ajst-24984	111	6	new	new	ADJ
ajst-24984	111	7	method	method	NOUN
ajst-24984	111	8	that	that	PRON
ajst-24984	111	9	combines	combine	VERB
ajst-24984	111	10	machine	machine	NOUN
ajst-24984	111	11	learning	learning	NOUN
ajst-24984	111	12	and	and	CCONJ
ajst-24984	111	13	dictionary	dictionary	ADJ
ajst-24984	111	14	techniques	technique	NOUN
ajst-24984	111	15	to	to	PART
ajst-24984	111	16	construct	construct	VERB
ajst-24984	111	17	a	a	DET
ajst-24984	111	18	new	new	ADJ
ajst-24984	111	19	dictionary	dictionary	NOUN
ajst-24984	111	20	by	by	ADP
ajst-24984	111	21	learning	learn	VERB
ajst-24984	111	22	the	the	DET
ajst-24984	111	23	labels	label	NOUN
ajst-24984	111	24	of	of	ADP
ajst-24984	111	25	the	the	DET
ajst-24984	111	26	classifiers	classifier	NOUN
ajst-24984	111	27	of	of	ADP
ajst-24984	111	28	the	the	DET
ajst-24984	111	29	training	training	NOUN
ajst-24984	111	30	data	datum	NOUN
ajst-24984	111	31	and	and	CCONJ
ajst-24984	111	32	uses	use	VERB
ajst-24984	111	33	the	the	DET
ajst-24984	111	34	new	new	ADJ
ajst-24984	111	35	dictionary	dictionary	NOUN
ajst-24984	111	36	to	to	PART
ajst-24984	111	37	correct	correct	VERB
ajst-24984	111	38	the	the	DET
ajst-24984	111	39	classifier	classifier	NOUN
ajst-24984	111	40	's	's	PART
ajst-24984	111	41	's	's	PART
ajst-24984	111	42	erroneous	erroneous	ADJ
ajst-24984	111	43	prediction	prediction	NOUN
ajst-24984	111	44	results	result	NOUN
ajst-24984	111	45	.	.	PUNCT
ajst-24984	112	1	compared	compare	VERB
ajst-24984	112	2	to	to	ADP
ajst-24984	112	3	the	the	DET
ajst-24984	112	4	traditional	traditional	ADJ
ajst-24984	112	5	sentiment	sentiment	NOUN
ajst-24984	112	6	dictionary	dictionary	ADJ
ajst-24984	112	7	analysis	analysis	NOUN
ajst-24984	112	8	195	195	NUM
ajst-24984	112	9	methods	method	NOUN
ajst-24984	112	10	,	,	PUNCT
ajst-24984	112	11	machine	machine	NOUN
ajst-24984	112	12	learning	learning	NOUN
ajst-24984	112	13	has	have	AUX
ajst-24984	112	14	demonstrated	demonstrate	VERB
ajst-24984	112	15	higher	high	ADJ
ajst-24984	112	16	convenience	convenience	NOUN
ajst-24984	112	17	and	and	CCONJ
ajst-24984	112	18	accuracy	accuracy	NOUN
ajst-24984	112	19	in	in	ADP
ajst-24984	112	20	the	the	DET
ajst-24984	112	21	field	field	NOUN
ajst-24984	112	22	of	of	ADP
ajst-24984	112	23	sentiment	sentiment	NOUN
ajst-24984	112	24	analysis	analysis	NOUN
ajst-24984	112	25	.	.	PUNCT
ajst-24984	113	1	it	it	PRON
ajst-24984	113	2	not	not	PART
ajst-24984	113	3	only	only	ADV
ajst-24984	113	4	has	have	VERB
ajst-24984	113	5	excellent	excellent	ADJ
ajst-24984	113	6	scalability	scalability	NOUN
ajst-24984	113	7	and	and	CCONJ
ajst-24984	113	8	repeatability	repeatability	NOUN
ajst-24984	113	9	,	,	PUNCT
ajst-24984	113	10	but	but	CCONJ
ajst-24984	113	11	also	also	ADV
ajst-24984	113	12	achieves	achieve	VERB
ajst-24984	113	13	more	more	ADV
ajst-24984	113	14	efficient	efficient	ADJ
ajst-24984	113	15	classification	classification	NOUN
ajst-24984	113	16	performance	performance	NOUN
ajst-24984	113	17	.	.	PUNCT
ajst-24984	114	1	however	however	ADV
ajst-24984	114	2	,	,	PUNCT
ajst-24984	114	3	the	the	DET
ajst-24984	114	4	performance	performance	NOUN
ajst-24984	114	5	of	of	ADP
ajst-24984	114	6	machine	machine	NOUN
ajst-24984	114	7	learning	learn	VERB
ajst-24984	114	8	algorithms	algorithm	NOUN
ajst-24984	114	9	is	be	AUX
ajst-24984	114	10	highly	highly	ADV
ajst-24984	114	11	dependent	dependent	ADJ
ajst-24984	114	12	on	on	ADP
ajst-24984	114	13	well	well	ADV
ajst-24984	114	14	-	-	PUNCT
ajst-24984	114	15	labeled	label	VERB
ajst-24984	114	16	training	training	NOUN
ajst-24984	114	17	datasets	dataset	NOUN
ajst-24984	114	18	.	.	PUNCT
ajst-24984	115	1	obtaining	obtain	VERB
ajst-24984	115	2	largescale	largescale	NOUN
ajst-24984	115	3	and	and	CCONJ
ajst-24984	115	4	high	high	ADJ
ajst-24984	115	5	-	-	PUNCT
ajst-24984	115	6	quality	quality	NOUN
ajst-24984	115	7	training	training	NOUN
ajst-24984	115	8	data	datum	NOUN
ajst-24984	115	9	usually	usually	ADV
ajst-24984	115	10	requires	require	VERB
ajst-24984	115	11	expensive	expensive	ADJ
ajst-24984	115	12	labor	labor	NOUN
ajst-24984	115	13	costs	cost	NOUN
ajst-24984	115	14	,	,	PUNCT
ajst-24984	115	15	and	and	CCONJ
ajst-24984	115	16	the	the	DET
ajst-24984	115	17	subjectivity	subjectivity	NOUN
ajst-24984	115	18	of	of	ADP
ajst-24984	115	19	human	human	ADJ
ajst-24984	115	20	labeling	labeling	NOUN
ajst-24984	115	21	may	may	AUX
ajst-24984	115	22	have	have	VERB
ajst-24984	115	23	some	some	DET
ajst-24984	115	24	impact	impact	NOUN
ajst-24984	115	25	on	on	ADP
ajst-24984	115	26	the	the	DET
ajst-24984	115	27	classification	classification	NOUN
ajst-24984	115	28	results	result	NOUN
ajst-24984	115	29	.	.	PUNCT
ajst-24984	116	1	although	although	SCONJ
ajst-24984	116	2	combining	combine	VERB
ajst-24984	116	3	machine	machine	NOUN
ajst-24984	116	4	learning	learn	VERB
ajst-24984	116	5	with	with	ADP
ajst-24984	116	6	sentiment	sentiment	NOUN
ajst-24984	116	7	dictionaries	dictionary	NOUN
ajst-24984	116	8	is	be	AUX
ajst-24984	116	9	an	an	DET
ajst-24984	116	10	improvement	improvement	NOUN
ajst-24984	116	11	over	over	ADP
ajst-24984	116	12	a	a	DET
ajst-24984	116	13	single	single	ADJ
ajst-24984	116	14	approach	approach	NOUN
ajst-24984	116	15	,	,	PUNCT
ajst-24984	116	16	this	this	DET
ajst-24984	116	17	approach	approach	NOUN
ajst-24984	116	18	is	be	AUX
ajst-24984	116	19	still	still	ADV
ajst-24984	116	20	susceptible	susceptible	ADJ
ajst-24984	116	21	to	to	ADP
ajst-24984	116	22	the	the	DET
ajst-24984	116	23	quality	quality	NOUN
ajst-24984	116	24	of	of	ADP
ajst-24984	116	25	sentiment	sentiment	NOUN
ajst-24984	116	26	dictionaries	dictionary	NOUN
ajst-24984	116	27	and	and	CCONJ
ajst-24984	116	28	the	the	DET
ajst-24984	116	29	labeling	labeling	NOUN
ajst-24984	116	30	of	of	ADP
ajst-24984	116	31	text	text	NOUN
ajst-24984	116	32	features	feature	NOUN
ajst-24984	116	33	.	.	PUNCT
ajst-24984	117	1	in	in	ADP
ajst-24984	117	2	addition	addition	NOUN
ajst-24984	117	3	,	,	PUNCT
ajst-24984	117	4	machine	machine	NOUN
ajst-24984	117	5	learning	learning	NOUN
ajst-24984	117	6	may	may	AUX
ajst-24984	117	7	fall	fall	VERB
ajst-24984	117	8	into	into	ADP
ajst-24984	117	9	inefficient	inefficient	ADJ
ajst-24984	117	10	work	work	NOUN
ajst-24984	117	11	when	when	SCONJ
ajst-24984	117	12	dealing	deal	VERB
ajst-24984	117	13	with	with	ADP
ajst-24984	117	14	large	large	ADJ
ajst-24984	117	15	-	-	PUNCT
ajst-24984	117	16	scale	scale	NOUN
ajst-24984	117	17	data	datum	NOUN
ajst-24984	117	18	,	,	PUNCT
ajst-24984	117	19	resulting	result	VERB
ajst-24984	117	20	in	in	ADP
ajst-24984	117	21	a	a	DET
ajst-24984	117	22	sluggish	sluggish	ADJ
ajst-24984	117	23	model	model	NOUN
ajst-24984	117	24	response	response	NOUN
ajst-24984	117	25	.	.	PUNCT
ajst-24984	118	1	therefore	therefore	ADV
ajst-24984	118	2	,	,	PUNCT
ajst-24984	118	3	in	in	ADP
ajst-24984	118	4	practical	practical	ADJ
ajst-24984	118	5	applications	application	NOUN
ajst-24984	118	6	,	,	PUNCT
ajst-24984	118	7	more	more	ADJ
ajst-24984	118	8	and	and	CCONJ
ajst-24984	118	9	more	more	ADJ
ajst-24984	118	10	scholars	scholar	NOUN
ajst-24984	118	11	have	have	AUX
ajst-24984	118	12	shifted	shift	VERB
ajst-24984	118	13	the	the	DET
ajst-24984	118	14	research	research	NOUN
ajst-24984	118	15	focus	focus	NOUN
ajst-24984	118	16	of	of	ADP
ajst-24984	118	17	sentiment	sentiment	NOUN
ajst-24984	118	18	analysis	analysis	NOUN
ajst-24984	118	19	to	to	ADP
ajst-24984	118	20	deep	deep	ADJ
ajst-24984	118	21	learning	learning	NOUN
ajst-24984	118	22	.	.	PUNCT
ajst-24984	119	1	deep	deep	ADJ
ajst-24984	119	2	learning	learning	NOUN
ajst-24984	119	3	is	be	AUX
ajst-24984	119	4	able	able	ADJ
ajst-24984	119	5	to	to	PART
ajst-24984	119	6	reduce	reduce	VERB
ajst-24984	119	7	the	the	DET
ajst-24984	119	8	reliance	reliance	NOUN
ajst-24984	119	9	on	on	ADP
ajst-24984	119	10	manually	manually	ADV
ajst-24984	119	11	labeled	label	VERB
ajst-24984	119	12	data	datum	NOUN
ajst-24984	119	13	by	by	ADP
ajst-24984	119	14	automatically	automatically	ADV
ajst-24984	119	15	extracting	extract	VERB
ajst-24984	119	16	features	feature	NOUN
ajst-24984	119	17	and	and	CCONJ
ajst-24984	119	18	end	end	NOUN
ajst-24984	119	19	-	-	PUNCT
ajst-24984	119	20	to	to	ADP
ajst-24984	119	21	-	-	PUNCT
ajst-24984	119	22	end	end	NOUN
ajst-24984	119	23	model	model	NOUN
ajst-24984	119	24	training	training	NOUN
ajst-24984	119	25	,	,	PUNCT
ajst-24984	119	26	and	and	CCONJ
ajst-24984	119	27	usually	usually	ADV
ajst-24984	119	28	provides	provide	VERB
ajst-24984	119	29	more	more	ADV
ajst-24984	119	30	efficient	efficient	ADJ
ajst-24984	119	31	performance	performance	NOUN
ajst-24984	119	32	when	when	SCONJ
ajst-24984	119	33	dealing	deal	VERB
ajst-24984	119	34	with	with	ADP
ajst-24984	119	35	large	large	ADJ
ajst-24984	119	36	-	-	PUNCT
ajst-24984	119	37	scale	scale	NOUN
ajst-24984	119	38	data	datum	NOUN
ajst-24984	119	39	.	.	PUNCT
ajst-24984	120	1	3.3	3.3	NUM
ajst-24984	120	2	.	.	PUNCT
ajst-24984	121	1	deep	deep	ADJ
ajst-24984	121	2	learning	learning	NOUN
ajst-24984	121	3	based	base	VERB
ajst-24984	121	4	approaches	approach	VERB
ajst-24984	121	5	traditional	traditional	ADJ
ajst-24984	121	6	sentiment	sentiment	NOUN
ajst-24984	121	7	analysis	analysis	NOUN
ajst-24984	121	8	methods	method	NOUN
ajst-24984	121	9	are	be	AUX
ajst-24984	121	10	no	no	ADV
ajst-24984	121	11	longer	long	ADV
ajst-24984	121	12	able	able	ADJ
ajst-24984	121	13	to	to	PART
ajst-24984	121	14	meet	meet	VERB
ajst-24984	121	15	the	the	DET
ajst-24984	121	16	demand	demand	NOUN
ajst-24984	121	17	for	for	ADP
ajst-24984	121	18	efficient	efficient	ADJ
ajst-24984	121	19	and	and	CCONJ
ajst-24984	121	20	high	high	ADJ
ajst-24984	121	21	-	-	PUNCT
ajst-24984	121	22	quality	quality	NOUN
ajst-24984	121	23	sentiment	sentiment	NOUN
ajst-24984	121	24	classification	classification	NOUN
ajst-24984	121	25	tasks	task	NOUN
ajst-24984	121	26	in	in	ADP
ajst-24984	121	27	the	the	DET
ajst-24984	121	28	era	era	NOUN
ajst-24984	121	29	of	of	ADP
ajst-24984	121	30	big	big	ADJ
ajst-24984	121	31	data	datum	NOUN
ajst-24984	121	32	in	in	ADP
ajst-24984	121	33	terms	term	NOUN
ajst-24984	121	34	of	of	ADP
ajst-24984	121	35	labor	labor	NOUN
ajst-24984	121	36	cost	cost	NOUN
ajst-24984	121	37	and	and	CCONJ
ajst-24984	121	38	time	time	NOUN
ajst-24984	121	39	efficiency	efficiency	NOUN
ajst-24984	121	40	.	.	PUNCT
ajst-24984	122	1	thanks	thank	NOUN
ajst-24984	122	2	to	to	ADP
ajst-24984	122	3	the	the	DET
ajst-24984	122	4	rise	rise	NOUN
ajst-24984	122	5	of	of	ADP
ajst-24984	122	6	neural	neural	ADJ
ajst-24984	122	7	network	network	NOUN
ajst-24984	122	8	technology	technology	NOUN
ajst-24984	122	9	,	,	PUNCT
ajst-24984	122	10	the	the	DET
ajst-24984	122	11	field	field	NOUN
ajst-24984	122	12	of	of	ADP
ajst-24984	122	13	sentiment	sentiment	NOUN
ajst-24984	122	14	analysis	analysis	NOUN
ajst-24984	122	15	has	have	AUX
ajst-24984	122	16	been	be	AUX
ajst-24984	122	17	able	able	ADJ
ajst-24984	122	18	to	to	PART
ajst-24984	122	19	convert	convert	VERB
ajst-24984	122	20	text	text	NOUN
ajst-24984	122	21	sequences	sequence	NOUN
ajst-24984	122	22	into	into	ADP
ajst-24984	122	23	word	word	NOUN
ajst-24984	122	24	vectors	vector	NOUN
ajst-24984	122	25	through	through	ADP
ajst-24984	122	26	word	word	NOUN
ajst-24984	122	27	embedding	embed	VERB
ajst-24984	122	28	models	model	NOUN
ajst-24984	122	29	,	,	PUNCT
ajst-24984	122	30	and	and	CCONJ
ajst-24984	122	31	then	then	ADV
ajst-24984	122	32	feed	feed	VERB
ajst-24984	122	33	the	the	DET
ajst-24984	122	34	vector	vector	NOUN
ajst-24984	122	35	sequences	sequence	NOUN
ajst-24984	122	36	as	as	ADP
ajst-24984	122	37	inputs	input	NOUN
ajst-24984	122	38	into	into	ADP
ajst-24984	122	39	neural	neural	ADJ
ajst-24984	122	40	network	network	NOUN
ajst-24984	122	41	models	model	NOUN
ajst-24984	122	42	,	,	PUNCT
ajst-24984	122	43	thus	thus	ADV
ajst-24984	122	44	making	make	VERB
ajst-24984	122	45	the	the	DET
ajst-24984	122	46	features	feature	NOUN
ajst-24984	122	47	rich	rich	ADJ
ajst-24984	122	48	in	in	ADP
ajst-24984	122	49	semantic	semantic	ADJ
ajst-24984	122	50	information	information	NOUN
ajst-24984	122	51	.	.	PUNCT
ajst-24984	123	1	with	with	ADP
ajst-24984	123	2	the	the	DET
ajst-24984	123	3	successful	successful	ADJ
ajst-24984	123	4	application	application	NOUN
ajst-24984	123	5	of	of	ADP
ajst-24984	123	6	attention	attention	NOUN
ajst-24984	123	7	and	and	CCONJ
ajst-24984	123	8	memory	memory	NOUN
ajst-24984	123	9	mechanisms	mechanism	NOUN
ajst-24984	123	10	in	in	ADP
ajst-24984	123	11	natural	natural	ADJ
ajst-24984	123	12	language	language	NOUN
ajst-24984	123	13	processing	processing	NOUN
ajst-24984	123	14	tasks	task	NOUN
ajst-24984	123	15	,	,	PUNCT
ajst-24984	123	16	deep	deep	ADJ
ajst-24984	123	17	learning	learning	NOUN
ajst-24984	123	18	-	-	PUNCT
ajst-24984	123	19	based	base	VERB
ajst-24984	123	20	sentiment	sentiment	NOUN
ajst-24984	123	21	analysis	analysis	NOUN
ajst-24984	123	22	methods	method	NOUN
ajst-24984	123	23	are	be	AUX
ajst-24984	123	24	able	able	ADJ
ajst-24984	123	25	to	to	PART
ajst-24984	123	26	achieve	achieve	VERB
ajst-24984	123	27	even	even	ADV
ajst-24984	123	28	better	well	ADJ
ajst-24984	123	29	performance	performance	NOUN
ajst-24984	123	30	.	.	PUNCT
ajst-24984	124	1	the	the	DET
ajst-24984	124	2	main	main	ADJ
ajst-24984	124	3	advantage	advantage	NOUN
ajst-24984	124	4	of	of	ADP
ajst-24984	124	5	this	this	DET
ajst-24984	124	6	approach	approach	NOUN
ajst-24984	124	7	is	be	AUX
ajst-24984	124	8	that	that	SCONJ
ajst-24984	124	9	it	it	PRON
ajst-24984	124	10	does	do	AUX
ajst-24984	124	11	not	not	PART
ajst-24984	124	12	need	need	VERB
ajst-24984	124	13	to	to	PART
ajst-24984	124	14	define	define	VERB
ajst-24984	124	15	the	the	DET
ajst-24984	124	16	features	feature	NOUN
ajst-24984	124	17	manually	manually	ADV
ajst-24984	124	18	,	,	PUNCT
ajst-24984	124	19	but	but	CCONJ
ajst-24984	124	20	simulates	simulate	VERB
ajst-24984	124	21	the	the	DET
ajst-24984	124	22	working	work	VERB
ajst-24984	124	23	mechanism	mechanism	NOUN
ajst-24984	124	24	of	of	ADP
ajst-24984	124	25	the	the	DET
ajst-24984	124	26	human	human	ADJ
ajst-24984	124	27	brain	brain	NOUN
ajst-24984	124	28	through	through	ADP
ajst-24984	124	29	deep	deep	ADJ
ajst-24984	124	30	learning	learning	NOUN
ajst-24984	124	31	,	,	PUNCT
ajst-24984	124	32	parses	parse	VERB
ajst-24984	124	33	the	the	DET
ajst-24984	124	34	text	text	NOUN
ajst-24984	124	35	and	and	CCONJ
ajst-24984	124	36	extracts	extract	VERB
ajst-24984	124	37	key	key	ADJ
ajst-24984	124	38	features	feature	NOUN
ajst-24984	124	39	in	in	ADP
ajst-24984	124	40	a	a	DET
ajst-24984	124	41	hierarchical	hierarchical	ADJ
ajst-24984	124	42	way	way	NOUN
ajst-24984	124	43	,	,	PUNCT
ajst-24984	124	44	and	and	CCONJ
ajst-24984	124	45	automatically	automatically	ADV
ajst-24984	124	46	adjusts	adjust	VERB
ajst-24984	124	47	the	the	DET
ajst-24984	124	48	model	model	NOUN
ajst-24984	124	49	parameters	parameter	NOUN
ajst-24984	124	50	to	to	PART
ajst-24984	124	51	optimize	optimize	VERB
ajst-24984	124	52	the	the	DET
ajst-24984	124	53	output	output	NOUN
ajst-24984	124	54	results	result	NOUN
ajst-24984	124	55	.	.	PUNCT
ajst-24984	125	1	in	in	ADP
ajst-24984	125	2	addition	addition	NOUN
ajst-24984	125	3	,	,	PUNCT
ajst-24984	125	4	the	the	DET
ajst-24984	125	5	neural	neural	ADJ
ajst-24984	125	6	network	network	NOUN
ajst-24984	125	7	is	be	AUX
ajst-24984	125	8	able	able	ADJ
ajst-24984	125	9	to	to	PART
ajst-24984	125	10	learn	learn	VERB
ajst-24984	125	11	the	the	DET
ajst-24984	125	12	sentiment	sentiment	NOUN
ajst-24984	125	13	information	information	NOUN
ajst-24984	125	14	embedded	embed	VERB
ajst-24984	125	15	in	in	ADP
ajst-24984	125	16	the	the	DET
ajst-24984	125	17	text	text	NOUN
ajst-24984	125	18	on	on	ADP
ajst-24984	125	19	its	its	PRON
ajst-24984	125	20	own	own	ADJ
ajst-24984	125	21	,	,	PUNCT
ajst-24984	125	22	significantly	significantly	ADV
ajst-24984	125	23	improving	improve	VERB
ajst-24984	125	24	the	the	DET
ajst-24984	125	25	overall	overall	ADJ
ajst-24984	125	26	accuracy	accuracy	NOUN
ajst-24984	125	27	of	of	ADP
ajst-24984	125	28	sentiment	sentiment	NOUN
ajst-24984	125	29	analysis	analysis	NOUN
ajst-24984	125	30	.	.	PUNCT
ajst-24984	126	1	in	in	ADP
ajst-24984	126	2	recent	recent	ADJ
ajst-24984	126	3	years	year	NOUN
ajst-24984	126	4	,	,	PUNCT
ajst-24984	126	5	recurrent	recurrent	ADJ
ajst-24984	126	6	neural	neural	ADJ
ajst-24984	126	7	networks	network	NOUN
ajst-24984	126	8	(	(	PUNCT
ajst-24984	126	9	rnn	rnn	PROPN
ajst-24984	126	10	)	)	PUNCT
ajst-24984	126	11	represent	represent	VERB
ajst-24984	126	12	an	an	DET
ajst-24984	126	13	end	end	NOUN
ajst-24984	126	14	-	-	PUNCT
ajst-24984	126	15	to	to	ADP
ajst-24984	126	16	-	-	PUNCT
ajst-24984	126	17	end	end	NOUN
ajst-24984	126	18	neural	neural	ADJ
ajst-24984	126	19	network	network	NOUN
ajst-24984	126	20	model	model	NOUN
ajst-24984	126	21	that	that	PRON
ajst-24984	126	22	has	have	AUX
ajst-24984	126	23	attracted	attract	VERB
ajst-24984	126	24	attention	attention	NOUN
ajst-24984	126	25	in	in	ADP
ajst-24984	126	26	the	the	DET
ajst-24984	126	27	field	field	NOUN
ajst-24984	126	28	of	of	ADP
ajst-24984	126	29	aspect	aspect	NOUN
ajst-24984	126	30	-	-	PUNCT
ajst-24984	126	31	based	base	VERB
ajst-24984	126	32	sentiment	sentiment	NOUN
ajst-24984	126	33	analysis	analysis	NOUN
ajst-24984	126	34	with	with	ADP
ajst-24984	126	35	its	its	PRON
ajst-24984	126	36	excellent	excellent	ADJ
ajst-24984	126	37	automatic	automatic	ADJ
ajst-24984	126	38	feature	feature	NOUN
ajst-24984	126	39	extraction	extraction	NOUN
ajst-24984	126	40	capability	capability	NOUN
ajst-24984	126	41	and	and	CCONJ
ajst-24984	126	42	good	good	ADJ
ajst-24984	126	43	classification	classification	NOUN
ajst-24984	126	44	results	result	NOUN
ajst-24984	126	45	[	[	X
ajst-24984	126	46	29	29	NUM
ajst-24984	126	47	]	]	PUNCT
ajst-24984	126	48	.	.	PUNCT
ajst-24984	127	1	however	however	ADV
ajst-24984	127	2	,	,	PUNCT
ajst-24984	127	3	rnn	rnn	VERB
ajst-24984	127	4	faces	face	VERB
ajst-24984	127	5	the	the	DET
ajst-24984	127	6	problem	problem	NOUN
ajst-24984	127	7	of	of	ADP
ajst-24984	127	8	vanishing	vanish	VERB
ajst-24984	127	9	or	or	CCONJ
ajst-24984	127	10	exploding	explode	VERB
ajst-24984	127	11	gradients	gradient	NOUN
ajst-24984	127	12	when	when	SCONJ
ajst-24984	127	13	dealing	deal	VERB
ajst-24984	127	14	with	with	ADP
ajst-24984	127	15	long	long	ADJ
ajst-24984	127	16	sequence	sequence	NOUN
ajst-24984	127	17	data	datum	NOUN
ajst-24984	127	18	,	,	PUNCT
ajst-24984	127	19	which	which	PRON
ajst-24984	127	20	limits	limit	VERB
ajst-24984	127	21	its	its	PRON
ajst-24984	127	22	ability	ability	NOUN
ajst-24984	127	23	to	to	PART
ajst-24984	127	24	capture	capture	VERB
ajst-24984	127	25	long	long	ADJ
ajst-24984	127	26	-	-	PUNCT
ajst-24984	127	27	term	term	NOUN
ajst-24984	127	28	dependencies	dependency	NOUN
ajst-24984	127	29	in	in	ADP
ajst-24984	127	30	sequences	sequence	NOUN
ajst-24984	127	31	.	.	PUNCT
ajst-24984	128	1	in	in	ADP
ajst-24984	128	2	addition	addition	NOUN
ajst-24984	128	3	,	,	PUNCT
ajst-24984	128	4	rnn	rnn	PROPN
ajst-24984	128	5	's	's	PART
ajst-24984	128	6	limitations	limitation	NOUN
ajst-24984	128	7	in	in	ADP
ajst-24984	128	8	memory	memory	NOUN
ajst-24984	128	9	affect	affect	VERB
ajst-24984	128	10	its	its	PRON
ajst-24984	128	11	performance	performance	NOUN
ajst-24984	128	12	in	in	ADP
ajst-24984	128	13	processing	processing	NOUN
ajst-24984	128	14	complex	complex	ADJ
ajst-24984	128	15	sequence	sequence	NOUN
ajst-24984	128	16	data	datum	NOUN
ajst-24984	128	17	,	,	PUNCT
ajst-24984	128	18	as	as	SCONJ
ajst-24984	128	19	it	it	PRON
ajst-24984	128	20	struggles	struggle	VERB
ajst-24984	128	21	to	to	PART
ajst-24984	128	22	retain	retain	VERB
ajst-24984	128	23	and	and	CCONJ
ajst-24984	128	24	utilize	utilize	VERB
ajst-24984	128	25	early	early	ADJ
ajst-24984	128	26	history	history	NOUN
ajst-24984	128	27	information	information	NOUN
ajst-24984	128	28	.	.	PUNCT
ajst-24984	129	1	to	to	PART
ajst-24984	129	2	overcome	overcome	VERB
ajst-24984	129	3	these	these	DET
ajst-24984	129	4	challenges	challenge	NOUN
ajst-24984	129	5	,	,	PUNCT
ajst-24984	129	6	researchers	researcher	NOUN
ajst-24984	129	7	introduced	introduce	VERB
ajst-24984	129	8	gating	gate	VERB
ajst-24984	129	9	mechanisms	mechanism	NOUN
ajst-24984	129	10	and	and	CCONJ
ajst-24984	129	11	proposed	propose	VERB
ajst-24984	129	12	the	the	DET
ajst-24984	129	13	long	long	ADJ
ajst-24984	129	14	short	short	ADJ
ajst-24984	129	15	-	-	PUNCT
ajst-24984	129	16	term	term	NOUN
ajst-24984	129	17	memory	memory	NOUN
ajst-24984	129	18	network	network	NOUN
ajst-24984	129	19	(	(	PUNCT
ajst-24984	129	20	lstm	lstm	NOUN
ajst-24984	129	21	)	)	PUNCT
ajst-24984	129	22	model	model	NOUN
ajst-24984	129	23	[	[	X
ajst-24984	129	24	30].the	30].the	NUM
ajst-24984	129	25	lstm	lstm	ADJ
ajst-24984	129	26	model	model	NOUN
ajst-24984	129	27	is	be	AUX
ajst-24984	129	28	able	able	ADJ
ajst-24984	129	29	to	to	PART
ajst-24984	129	30	accurately	accurately	ADV
ajst-24984	129	31	filter	filter	VERB
ajst-24984	129	32	and	and	CCONJ
ajst-24984	129	33	retain	retain	VERB
ajst-24984	129	34	critical	critical	ADJ
ajst-24984	129	35	historical	historical	ADJ
ajst-24984	129	36	information	information	NOUN
ajst-24984	129	37	through	through	ADP
ajst-24984	129	38	the	the	DET
ajst-24984	129	39	synergistic	synergistic	ADJ
ajst-24984	129	40	action	action	NOUN
ajst-24984	129	41	of	of	ADP
ajst-24984	129	42	forgetting	forget	VERB
ajst-24984	129	43	gates	gate	NOUN
ajst-24984	129	44	,	,	PUNCT
ajst-24984	129	45	input	input	NOUN
ajst-24984	129	46	gates	gate	NOUN
ajst-24984	129	47	,	,	PUNCT
ajst-24984	129	48	and	and	CCONJ
ajst-24984	129	49	output	output	NOUN
ajst-24984	129	50	gates	gate	NOUN
ajst-24984	129	51	,	,	PUNCT
ajst-24984	129	52	effectively	effectively	ADV
ajst-24984	129	53	solving	solve	VERB
ajst-24984	129	54	the	the	DET
ajst-24984	129	55	problems	problem	NOUN
ajst-24984	129	56	of	of	ADP
ajst-24984	129	57	gradient	gradient	ADJ
ajst-24984	129	58	vanishing	vanishing	NOUN
ajst-24984	129	59	and	and	CCONJ
ajst-24984	129	60	long	long	ADJ
ajst-24984	129	61	-	-	PUNCT
ajst-24984	129	62	term	term	NOUN
ajst-24984	129	63	dependencies	dependency	NOUN
ajst-24984	129	64	,	,	PUNCT
ajst-24984	129	65	and	and	CCONJ
ajst-24984	129	66	thus	thus	ADV
ajst-24984	129	67	demonstrating	demonstrate	VERB
ajst-24984	129	68	excellent	excellent	ADJ
ajst-24984	129	69	performance	performance	NOUN
ajst-24984	129	70	when	when	SCONJ
ajst-24984	129	71	processing	process	VERB
ajst-24984	129	72	long	long	ADJ
ajst-24984	129	73	sequence	sequence	NOUN
ajst-24984	129	74	data	datum	NOUN
ajst-24984	129	75	.	.	PUNCT
ajst-24984	130	1	therefore	therefore	ADV
ajst-24984	130	2	,	,	PUNCT
ajst-24984	130	3	lstm	lstm	NOUN
ajst-24984	130	4	models	model	NOUN
ajst-24984	130	5	are	be	AUX
ajst-24984	130	6	widely	widely	ADV
ajst-24984	130	7	used	use	VERB
ajst-24984	130	8	in	in	ADP
ajst-24984	130	9	aspect	aspect	NOUN
ajst-24984	130	10	-	-	PUNCT
ajst-24984	130	11	based	base	VERB
ajst-24984	130	12	sentiment	sentiment	NOUN
ajst-24984	130	13	analysis	analysis	NOUN
ajst-24984	130	14	tasks	task	NOUN
ajst-24984	130	15	to	to	PART
ajst-24984	130	16	improve	improve	VERB
ajst-24984	130	17	the	the	DET
ajst-24984	130	18	correctness	correctness	NOUN
ajst-24984	130	19	and	and	CCONJ
ajst-24984	130	20	efficiency	efficiency	NOUN
ajst-24984	130	21	of	of	ADP
ajst-24984	130	22	sentiment	sentiment	NOUN
ajst-24984	130	23	analysis	analysis	NOUN
ajst-24984	130	24	.	.	PUNCT
ajst-24984	131	1	based	base	VERB
ajst-24984	131	2	on	on	ADP
ajst-24984	131	3	the	the	DET
ajst-24984	131	4	lstm	lstm	PROPN
ajst-24984	131	5	model	model	PROPN
ajst-24984	131	6	li	li	PROPN
ajst-24984	131	7	et	et	PROPN
ajst-24984	131	8	al	al	PROPN
ajst-24984	132	1	[	[	X
ajst-24984	132	2	31	31	NUM
ajst-24984	132	3	]	]	PUNCT
ajst-24984	132	4	developed	develop	VERB
ajst-24984	132	5	a	a	DET
ajst-24984	132	6	"	"	PUNCT
ajst-24984	132	7	semisupervised	semisupervised	ADJ
ajst-24984	132	8	and	and	CCONJ
ajst-24984	132	9	multi	multi	ADJ
ajst-24984	132	10	-	-	ADJ
ajst-24984	132	11	task	task	ADJ
ajst-24984	132	12	learning	learning	NOUN
ajst-24984	132	13	"	"	PUNCT
ajst-24984	132	14	(	(	PUNCT
ajst-24984	132	15	seml	seml	NOUN
ajst-24984	132	16	)	)	PUNCT
ajst-24984	132	17	framework	framework	NOUN
ajst-24984	132	18	that	that	PRON
ajst-24984	132	19	uses	use	VERB
ajst-24984	132	20	embedding	embed	VERB
ajst-24984	132	21	and	and	CCONJ
ajst-24984	132	22	lstm	lstm	ADJ
ajst-24984	132	23	layers	layer	NOUN
ajst-24984	132	24	to	to	PART
ajst-24984	132	25	transfer	transfer	VERB
ajst-24984	132	26	sentiment	sentiment	NOUN
ajst-24984	132	27	information	information	NOUN
ajst-24984	132	28	to	to	ADP
ajst-24984	132	29	document	document	NOUN
ajst-24984	132	30	-	-	PUNCT
ajst-24984	132	31	level	level	NOUN
ajst-24984	132	32	data	datum	NOUN
ajst-24984	132	33	in	in	ADP
ajst-24984	132	34	an	an	DET
ajst-24984	132	35	absa	absa	PROPN
ajst-24984	132	36	model	model	NOUN
ajst-24984	132	37	.	.	PUNCT
ajst-24984	133	1	chen	chen	PROPN
ajst-24984	133	2	and	and	CCONJ
ajst-24984	133	3	qian	qian	PROPN
ajst-24984	134	1	[	[	X
ajst-24984	134	2	32	32	NUM
ajst-24984	134	3	]	]	PUNCT
ajst-24984	134	4	developed	develop	VERB
ajst-24984	134	5	a	a	DET
ajst-24984	134	6	"	"	PUNCT
ajst-24984	134	7	transmission	transmission	NOUN
ajst-24984	134	8	capsule	capsule	NOUN
ajst-24984	134	9	network	network	NOUN
ajst-24984	134	10	model	model	NOUN
ajst-24984	134	11	(	(	PUNCT
ajst-24984	134	12	transcap	transcap	NOUN
ajst-24984	134	13	)	)	PUNCT
ajst-24984	134	14	"	"	PUNCT
ajst-24984	134	15	for	for	ADP
ajst-24984	134	16	sharing	share	VERB
ajst-24984	134	17	prior	prior	ADJ
ajst-24984	134	18	knowledge	knowledge	NOUN
ajst-24984	134	19	at	at	ADP
ajst-24984	134	20	the	the	DET
ajst-24984	134	21	document	document	NOUN
ajst-24984	134	22	level	level	NOUN
ajst-24984	134	23	to	to	ADP
ajst-24984	134	24	an	an	DET
ajst-24984	134	25	aspect	aspect	NOUN
ajst-24984	134	26	-	-	PUNCT
ajst-24984	134	27	based	base	VERB
ajst-24984	134	28	sentiment	sentiment	NOUN
ajst-24984	134	29	analysis	analysis	NOUN
ajst-24984	134	30	task	task	NOUN
ajst-24984	134	31	.	.	PUNCT
ajst-24984	135	1	"	"	PUNCT
ajst-24984	135	2	transmission	transmission	NOUN
ajst-24984	135	3	capsule	capsule	NOUN
ajst-24984	135	4	network	network	NOUN
ajst-24984	135	5	model	model	NOUN
ajst-24984	135	6	(	(	PUNCT
ajst-24984	135	7	transcap	transcap	NOUN
ajst-24984	135	8	)	)	PUNCT
ajst-24984	135	9	"	"	PUNCT
ajst-24984	135	10	for	for	ADP
ajst-24984	135	11	sharing	share	VERB
ajst-24984	135	12	document	document	NOUN
ajst-24984	135	13	-	-	PUNCT
ajst-24984	135	14	level	level	NOUN
ajst-24984	135	15	prior	prior	ADJ
ajst-24984	135	16	knowledge	knowledge	NOUN
ajst-24984	135	17	to	to	PART
ajst-24984	135	18	aspect	aspect	NOUN
ajst-24984	135	19	-	-	PUNCT
ajst-24984	135	20	based	base	VERB
ajst-24984	135	21	sentiment	sentiment	NOUN
ajst-24984	135	22	analysis	analysis	NOUN
ajst-24984	135	23	tasks	task	NOUN
ajst-24984	135	24	.	.	PUNCT
ajst-24984	136	1	su	su	PROPN
ajst-24984	136	2	et	et	PROPN
ajst-24984	136	3	al	al	PROPN
ajst-24984	137	1	[	[	X
ajst-24984	137	2	33	33	NUM
ajst-24984	137	3	]	]	PUNCT
ajst-24984	137	4	discussed	discuss	VERB
ajst-24984	137	5	an	an	DET
ajst-24984	137	6	absa	absa	PROPN
ajst-24984	137	7	model	model	NOUN
ajst-24984	137	8	based	base	VERB
ajst-24984	137	9	on	on	ADP
ajst-24984	137	10	capsule	capsule	NOUN
ajst-24984	137	11	networks	network	NOUN
ajst-24984	137	12	and	and	CCONJ
ajst-24984	137	13	xlnet	xlnet	PROPN
ajst-24984	137	14	,	,	PUNCT
ajst-24984	137	15	which	which	PRON
ajst-24984	137	16	first	first	ADV
ajst-24984	137	17	obtains	obtain	VERB
ajst-24984	137	18	the	the	DET
ajst-24984	137	19	relationship	relationship	NOUN
ajst-24984	137	20	between	between	ADP
ajst-24984	137	21	sequences	sequence	NOUN
ajst-24984	137	22	and	and	CCONJ
ajst-24984	137	23	aspects	aspect	NOUN
ajst-24984	137	24	,	,	PUNCT
ajst-24984	137	25	and	and	CCONJ
ajst-24984	137	26	then	then	ADV
ajst-24984	137	27	generates	generate	VERB
ajst-24984	137	28	aspect	aspect	NOUN
ajst-24984	137	29	-	-	PUNCT
ajst-24984	137	30	specific	specific	ADJ
ajst-24984	137	31	representations	representation	NOUN
ajst-24984	137	32	the	the	DET
ajst-24984	137	33	global	global	ADJ
ajst-24984	137	34	relationship	relationship	NOUN
ajst-24984	137	35	between	between	ADP
ajst-24984	137	36	aspects	aspect	NOUN
ajst-24984	137	37	facilitates	facilitates	AUX
ajst-24984	137	38	aspect	aspect	VERB
ajst-24984	137	39	awareness	awareness	NOUN
ajst-24984	137	40	and	and	CCONJ
ajst-24984	137	41	ensures	ensure	VERB
ajst-24984	137	42	that	that	SCONJ
ajst-24984	137	43	pre	pre	NOUN
ajst-24984	137	44	-	-	NOUN
ajst-24984	137	45	training	training	NOUN
ajst-24984	137	46	of	of	ADP
ajst-24984	137	47	xlnet	xlnet	NOUN
ajst-24984	137	48	ultimately	ultimately	ADV
ajst-24984	137	49	improves	improve	VERB
ajst-24984	137	50	task	task	NOUN
ajst-24984	137	51	uncertainty	uncertainty	NOUN
ajst-24984	137	52	.	.	PUNCT
ajst-24984	138	1	dong	dong	PROPN
ajst-24984	138	2	et	et	PROPN
ajst-24984	138	3	al	al	PROPN
ajst-24984	138	4	.	.	PUNCT
ajst-24984	139	1	[	[	X
ajst-24984	139	2	34	34	NUM
ajst-24984	139	3	]	]	PUNCT
ajst-24984	139	4	describe	describe	VERB
ajst-24984	139	5	an	an	DET
ajst-24984	139	6	adaptive	adaptive	ADJ
ajst-24984	139	7	recurrent	recurrent	ADJ
ajst-24984	139	8	neural	neural	ADJ
ajst-24984	139	9	network	network	NOUN
ajst-24984	139	10	(	(	PUNCT
ajst-24984	139	11	adarnn	adarnn	PROPN
ajst-24984	139	12	)	)	PUNCT
ajst-24984	139	13	model	model	NOUN
ajst-24984	139	14	for	for	ADP
ajst-24984	139	15	categorizing	categorize	VERB
ajst-24984	139	16	sentiment	sentiment	NOUN
ajst-24984	139	17	in	in	ADP
ajst-24984	139	18	twitter	twitter	NOUN
ajst-24984	139	19	comments	comment	NOUN
ajst-24984	139	20	.	.	PUNCT
ajst-24984	140	1	however	however	ADV
ajst-24984	140	2	,	,	PUNCT
ajst-24984	140	3	all	all	DET
ajst-24984	140	4	these	these	DET
ajst-24984	140	5	models	model	NOUN
ajst-24984	140	6	lose	lose	VERB
ajst-24984	140	7	more	more	ADJ
ajst-24984	140	8	or	or	CCONJ
ajst-24984	140	9	less	less	ADV
ajst-24984	140	10	implicit	implicit	ADJ
ajst-24984	140	11	features	feature	NOUN
ajst-24984	140	12	computed	compute	VERB
ajst-24984	140	13	in	in	ADP
ajst-24984	140	14	the	the	DET
ajst-24984	140	15	middle	middle	ADJ
ajst-24984	140	16	part	part	NOUN
ajst-24984	140	17	when	when	SCONJ
ajst-24984	140	18	using	use	VERB
ajst-24984	140	19	recurrent	recurrent	ADJ
ajst-24984	140	20	neural	neural	ADJ
ajst-24984	140	21	networks	network	NOUN
ajst-24984	140	22	,	,	PUNCT
ajst-24984	140	23	while	while	SCONJ
ajst-24984	140	24	the	the	DET
ajst-24984	140	25	structure	structure	NOUN
ajst-24984	140	26	of	of	ADP
ajst-24984	140	27	elmo	elmo	PROPN
ajst-24984	140	28	's	's	PART
ajst-24984	140	29	[	[	NOUN
ajst-24984	140	30	13	13	NUM
ajst-24984	140	31	]	]	PUNCT
ajst-24984	140	32	model	model	NOUN
ajst-24984	140	33	can	can	AUX
ajst-24984	140	34	effectively	effectively	ADV
ajst-24984	140	35	retain	retain	VERB
ajst-24984	140	36	the	the	DET
ajst-24984	140	37	implicit	implicit	ADJ
ajst-24984	140	38	features	feature	NOUN
ajst-24984	140	39	computed	compute	VERB
ajst-24984	140	40	by	by	ADP
ajst-24984	140	41	each	each	DET
ajst-24984	140	42	layer	layer	NOUN
ajst-24984	140	43	of	of	ADP
ajst-24984	140	44	the	the	DET
ajst-24984	140	45	recurrent	recurrent	ADJ
ajst-24984	140	46	neural	neural	ADJ
ajst-24984	140	47	network	network	NOUN
ajst-24984	140	48	and	and	CCONJ
ajst-24984	140	49	enable	enable	VERB
ajst-24984	140	50	the	the	DET
ajst-24984	140	51	model	model	NOUN
ajst-24984	140	52	to	to	PART
ajst-24984	140	53	learn	learn	VERB
ajst-24984	140	54	deep	deep	ADJ
ajst-24984	140	55	contextualized	contextualized	ADJ
ajst-24984	140	56	word	word	NOUN
ajst-24984	140	57	representations	representation	NOUN
ajst-24984	140	58	.	.	PUNCT
ajst-24984	141	1	however	however	ADV
ajst-24984	141	2	,	,	PUNCT
ajst-24984	141	3	the	the	DET
ajst-24984	141	4	elmo	elmo	PROPN
ajst-24984	141	5	[	[	X
ajst-24984	141	6	13	13	NUM
ajst-24984	141	7	]	]	PUNCT
ajst-24984	141	8	model	model	NOUN
ajst-24984	141	9	merges	merge	VERB
ajst-24984	141	10	the	the	DET
ajst-24984	141	11	implicit	implicit	ADJ
ajst-24984	141	12	features	feature	NOUN
ajst-24984	141	13	of	of	ADP
ajst-24984	141	14	each	each	DET
ajst-24984	141	15	layer	layer	NOUN
ajst-24984	141	16	in	in	ADP
ajst-24984	141	17	such	such	DET
ajst-24984	141	18	a	a	DET
ajst-24984	141	19	way	way	NOUN
ajst-24984	141	20	that	that	PRON
ajst-24984	141	21	the	the	DET
ajst-24984	141	22	key	key	ADJ
ajst-24984	141	23	features	feature	NOUN
ajst-24984	141	24	of	of	ADP
ajst-24984	141	25	some	some	DET
ajst-24984	141	26	layers	layer	NOUN
ajst-24984	141	27	do	do	AUX
ajst-24984	141	28	not	not	PART
ajst-24984	141	29	receive	receive	VERB
ajst-24984	141	30	enough	enough	ADJ
ajst-24984	141	31	attention	attention	NOUN
ajst-24984	141	32	.	.	PUNCT
ajst-24984	142	1	therefore	therefore	ADV
ajst-24984	142	2	,	,	PUNCT
ajst-24984	142	3	we	we	PRON
ajst-24984	142	4	improve	improve	VERB
ajst-24984	142	5	the	the	DET
ajst-24984	142	6	feature	feature	NOUN
ajst-24984	142	7	merging	merge	VERB
ajst-24984	142	8	approach	approach	NOUN
ajst-24984	142	9	of	of	ADP
ajst-24984	142	10	the	the	DET
ajst-24984	142	11	elmo	elmo	PROPN
ajst-24984	142	12	[	[	X
ajst-24984	142	13	13	13	NUM
ajst-24984	142	14	]	]	PUNCT
ajst-24984	142	15	model	model	NOUN
ajst-24984	142	16	merging	merging	NOUN
ajst-24984	142	17	by	by	ADP
ajst-24984	142	18	first	first	ADV
ajst-24984	142	19	highlighting	highlight	VERB
ajst-24984	142	20	the	the	DET
ajst-24984	142	21	focus	focus	NOUN
ajst-24984	142	22	of	of	ADP
ajst-24984	142	23	each	each	DET
ajst-24984	142	24	layer	layer	NOUN
ajst-24984	142	25	using	use	VERB
ajst-24984	142	26	the	the	DET
ajst-24984	142	27	multi	multi	ADJ
ajst-24984	142	28	-	-	ADJ
ajst-24984	142	29	head	head	ADJ
ajst-24984	142	30	selfattention	selfattention	NOUN
ajst-24984	142	31	mechanism	mechanism	NOUN
ajst-24984	142	32	,	,	PUNCT
ajst-24984	142	33	and	and	CCONJ
ajst-24984	142	34	then	then	ADV
ajst-24984	142	35	merging	merge	VERB
ajst-24984	142	36	the	the	DET
ajst-24984	142	37	data	datum	NOUN
ajst-24984	142	38	of	of	ADP
ajst-24984	142	39	each	each	DET
ajst-24984	142	40	layer	layer	NOUN
ajst-24984	142	41	using	use	VERB
ajst-24984	142	42	convolution	convolution	NOUN
ajst-24984	142	43	,	,	PUNCT
ajst-24984	142	44	which	which	PRON
ajst-24984	142	45	allows	allow	VERB
ajst-24984	142	46	the	the	DET
ajst-24984	142	47	model	model	NOUN
ajst-24984	142	48	to	to	PART
ajst-24984	142	49	retain	retain	VERB
ajst-24984	142	50	the	the	DET
ajst-24984	142	51	important	important	ADJ
ajst-24984	142	52	implicit	implicit	ADJ
ajst-24984	142	53	features	feature	NOUN
ajst-24984	142	54	and	and	CCONJ
ajst-24984	142	55	reduces	reduce	VERB
ajst-24984	142	56	the	the	DET
ajst-24984	142	57	influence	influence	NOUN
ajst-24984	142	58	of	of	ADP
ajst-24984	142	59	noise	noise	NOUN
ajst-24984	142	60	,	,	PUNCT
ajst-24984	142	61	and	and	CCONJ
ajst-24984	142	62	deepens	deepen	VERB
ajst-24984	142	63	the	the	DET
ajst-24984	142	64	model	model	NOUN
ajst-24984	142	65	's	's	PART
ajst-24984	142	66	understanding	understanding	NOUN
ajst-24984	142	67	of	of	ADP
ajst-24984	142	68	the	the	DET
ajst-24984	142	69	input	input	NOUN
ajst-24984	142	70	text	text	NOUN
ajst-24984	142	71	.	.	PUNCT
ajst-24984	143	1	in	in	ADP
ajst-24984	143	2	the	the	DET
ajst-24984	143	3	field	field	NOUN
ajst-24984	143	4	of	of	ADP
ajst-24984	143	5	text	text	NOUN
ajst-24984	143	6	data	datum	NOUN
ajst-24984	143	7	processing	processing	NOUN
ajst-24984	143	8	,	,	PUNCT
ajst-24984	143	9	the	the	DET
ajst-24984	143	10	attention	attention	NOUN
ajst-24984	143	11	mechanism	mechanism	NOUN
ajst-24984	143	12	is	be	AUX
ajst-24984	143	13	favored	favor	VERB
ajst-24984	143	14	for	for	ADP
ajst-24984	143	15	its	its	PRON
ajst-24984	143	16	significant	significant	ADJ
ajst-24984	143	17	advantages	advantage	NOUN
ajst-24984	143	18	,	,	PUNCT
ajst-24984	143	19	which	which	PRON
ajst-24984	143	20	can	can	AUX
ajst-24984	143	21	effectively	effectively	ADV
ajst-24984	143	22	obtain	obtain	VERB
ajst-24984	143	23	the	the	DET
ajst-24984	143	24	importance	importance	NOUN
ajst-24984	143	25	of	of	ADP
ajst-24984	143	26	individual	individual	ADJ
ajst-24984	143	27	words	word	NOUN
ajst-24984	143	28	in	in	ADP
ajst-24984	143	29	a	a	DET
ajst-24984	143	30	sequence	sequence	NOUN
ajst-24984	143	31	,	,	PUNCT
ajst-24984	143	32	so	so	ADV
ajst-24984	143	33	often	often	ADV
ajst-24984	143	34	researchers	researcher	NOUN
ajst-24984	143	35	combine	combine	VERB
ajst-24984	143	36	the	the	DET
ajst-24984	143	37	attention	attention	NOUN
ajst-24984	143	38	mechanism	mechanism	NOUN
ajst-24984	143	39	with	with	ADP
ajst-24984	143	40	recurrent	recurrent	ADJ
ajst-24984	143	41	neural	neural	ADJ
ajst-24984	143	42	networks	network	NOUN
ajst-24984	143	43	to	to	PART
ajst-24984	143	44	solve	solve	VERB
ajst-24984	143	45	the	the	DET
ajst-24984	143	46	problem	problem	NOUN
ajst-24984	143	47	.	.	PUNCT
ajst-24984	144	1	aiming	aim	VERB
ajst-24984	144	2	at	at	ADP
ajst-24984	144	3	the	the	DET
ajst-24984	144	4	problem	problem	NOUN
ajst-24984	144	5	of	of	ADP
ajst-24984	144	6	small	small	ADJ
ajst-24984	144	7	dataset	dataset	NOUN
ajst-24984	144	8	size	size	NOUN
ajst-24984	144	9	for	for	ADP
ajst-24984	144	10	aspect	aspect	NOUN
ajst-24984	144	11	-	-	PUNCT
ajst-24984	144	12	based	base	VERB
ajst-24984	144	13	sentiment	sentiment	NOUN
ajst-24984	144	14	analysis	analysis	NOUN
ajst-24984	144	15	,	,	PUNCT
ajst-24984	144	16	xu	xu	PROPN
ajst-24984	144	17	g	g	PROPN
ajst-24984	144	18	et	et	PROPN
ajst-24984	144	19	al	al	PROPN
ajst-24984	145	1	[	[	X
ajst-24984	145	2	35	35	NUM
ajst-24984	145	3	]	]	PUNCT
ajst-24984	145	4	adopted	adopt	VERB
ajst-24984	145	5	the	the	DET
ajst-24984	145	6	structure	structure	NOUN
ajst-24984	145	7	of	of	ADP
ajst-24984	145	8	bidirectional	bidirectional	ADJ
ajst-24984	145	9	long	long	ADJ
ajst-24984	145	10	and	and	CCONJ
ajst-24984	145	11	short	short	ADJ
ajst-24984	145	12	-	-	PUNCT
ajst-24984	145	13	term	term	NOUN
ajst-24984	145	14	memory	memory	NOUN
ajst-24984	145	15	network	network	NOUN
ajst-24984	145	16	(	(	PUNCT
ajst-24984	145	17	bi	bi	NOUN
ajst-24984	145	18	-	-	ADJ
ajst-24984	145	19	lstm	lstm	NOUN
ajst-24984	145	20	)	)	PUNCT
ajst-24984	145	21	that	that	PRON
ajst-24984	145	22	combines	combine	VERB
ajst-24984	145	23	the	the	DET
ajst-24984	145	24	attention	attention	NOUN
ajst-24984	145	25	mechanism	mechanism	NOUN
ajst-24984	145	26	and	and	CCONJ
ajst-24984	145	27	introduced	introduce	VERB
ajst-24984	145	28	the	the	DET
ajst-24984	145	29	transfer	transfer	NOUN
ajst-24984	145	30	learning	learning	NOUN
ajst-24984	145	31	technique	technique	NOUN
ajst-24984	145	32	in	in	ADP
ajst-24984	145	33	order	order	NOUN
ajst-24984	145	34	to	to	PART
ajst-24984	145	35	realize	realize	VERB
ajst-24984	145	36	more	more	ADV
ajst-24984	145	37	effective	effective	ADJ
ajst-24984	145	38	aspectbased	aspectbase	VERB
ajst-24984	145	39	sentiment	sentiment	NOUN
ajst-24984	145	40	analysis	analysis	NOUN
ajst-24984	145	41	.	.	PUNCT
ajst-24984	146	1	although	although	SCONJ
ajst-24984	146	2	the	the	DET
ajst-24984	146	3	attention	attention	NOUN
ajst-24984	146	4	mechanism	mechanism	NOUN
ajst-24984	146	5	has	have	AUX
ajst-24984	146	6	made	make	VERB
ajst-24984	146	7	significant	significant	ADJ
ajst-24984	146	8	progress	progress	NOUN
ajst-24984	146	9	in	in	ADP
ajst-24984	146	10	aspect	aspect	NOUN
ajst-24984	146	11	-	-	PUNCT
ajst-24984	146	12	level	level	NOUN
ajst-24984	146	13	sentiment	sentiment	NOUN
ajst-24984	146	14	categorization	categorization	NOUN
ajst-24984	146	15	,	,	PUNCT
ajst-24984	146	16	the	the	DET
ajst-24984	146	17	problem	problem	NOUN
ajst-24984	146	18	of	of	ADP
ajst-24984	146	19	information	information	NOUN
ajst-24984	146	20	loss	loss	NOUN
ajst-24984	146	21	may	may	AUX
ajst-24984	146	22	still	still	ADV
ajst-24984	146	23	exist	exist	VERB
ajst-24984	146	24	when	when	SCONJ
ajst-24984	146	25	dealing	deal	VERB
ajst-24984	146	26	with	with	ADP
ajst-24984	146	27	longer	long	ADJ
ajst-24984	146	28	or	or	CCONJ
ajst-24984	146	29	contextual	contextual	ADJ
ajst-24984	146	30	sentences	sentence	NOUN
ajst-24984	146	31	containing	contain	VERB
ajst-24984	146	32	multiple	multiple	ADJ
ajst-24984	146	33	sentiments	sentiment	NOUN
ajst-24984	146	34	.	.	PUNCT
ajst-24984	147	1	for	for	ADP
ajst-24984	147	2	this	this	DET
ajst-24984	147	3	reason	reason	NOUN
ajst-24984	147	4	,	,	PUNCT
ajst-24984	147	5	liao	liao	PROPN
ajst-24984	147	6	w	w	PROPN
ajst-24984	147	7	et	et	PROPN
ajst-24984	147	8	al	al	PROPN
ajst-24984	148	1	[	[	X
ajst-24984	148	2	36	36	NUM
ajst-24984	148	3	]	]	PUNCT
ajst-24984	148	4	proposed	propose	VERB
ajst-24984	148	5	a	a	DET
ajst-24984	148	6	fine	fine	ADV
ajst-24984	148	7	-	-	PUNCT
ajst-24984	148	8	grained	grain	VERB
ajst-24984	148	9	attention	attention	NOUN
ajst-24984	148	10	-	-	PUNCT
ajst-24984	148	11	based	base	VERB
ajst-24984	148	12	phrase	phrase	NOUN
ajst-24984	148	13	-	-	PUNCT
ajst-24984	148	14	aware	aware	ADJ
ajst-24984	148	15	neural	neural	ADJ
ajst-24984	148	16	network	network	NOUN
ajst-24984	148	17	to	to	PART
ajst-24984	148	18	address	address	VERB
ajst-24984	148	19	the	the	DET
ajst-24984	148	20	problem	problem	NOUN
ajst-24984	148	21	of	of	ADP
ajst-24984	148	22	ignoring	ignore	VERB
ajst-24984	148	23	the	the	DET
ajst-24984	148	24	importance	importance	NOUN
ajst-24984	148	25	of	of	ADP
ajst-24984	148	26	key	key	ADJ
ajst-24984	148	27	phrases	phrase	NOUN
ajst-24984	148	28	in	in	ADP
ajst-24984	148	29	a	a	DET
ajst-24984	148	30	sentence	sentence	NOUN
ajst-24984	148	31	.	.	PUNCT
ajst-24984	149	1	these	these	DET
ajst-24984	149	2	studies	study	NOUN
ajst-24984	149	3	show	show	VERB
ajst-24984	149	4	that	that	SCONJ
ajst-24984	149	5	deep	deep	ADJ
ajst-24984	149	6	learning	learning	NOUN
ajst-24984	149	7	models	model	NOUN
ajst-24984	149	8	incorporating	incorporate	VERB
ajst-24984	149	9	attention	attention	NOUN
ajst-24984	149	10	mechanisms	mechanism	NOUN
ajst-24984	149	11	have	have	VERB
ajst-24984	149	12	strong	strong	ADJ
ajst-24984	149	13	potential	potential	ADJ
ajst-24984	149	14	and	and	CCONJ
ajst-24984	149	15	application	application	NOUN
ajst-24984	149	16	prospects	prospect	NOUN
ajst-24984	149	17	in	in	ADP
ajst-24984	149	18	sentiment	sentiment	NOUN
ajst-24984	149	19	analysis	analysis	NOUN
ajst-24984	149	20	tasks	task	NOUN
ajst-24984	149	21	.	.	PUNCT
ajst-24984	150	1	graph	graph	NOUN
ajst-24984	150	2	convolution	convolution	NOUN
ajst-24984	150	3	network	network	NOUN
ajst-24984	150	4	(	(	PUNCT
ajst-24984	150	5	gcn	gcn	NOUN
ajst-24984	150	6	)	)	PUNCT
ajst-24984	150	7	has	have	AUX
ajst-24984	150	8	attracted	attract	VERB
ajst-24984	150	9	a	a	DET
ajst-24984	150	10	lot	lot	NOUN
ajst-24984	150	11	of	of	ADP
ajst-24984	150	12	attention	attention	NOUN
ajst-24984	150	13	in	in	ADP
ajst-24984	150	14	the	the	DET
ajst-24984	150	15	research	research	NOUN
ajst-24984	150	16	field	field	NOUN
ajst-24984	150	17	due	due	ADP
ajst-24984	150	18	to	to	ADP
ajst-24984	150	19	its	its	PRON
ajst-24984	150	20	remarkable	remarkable	ADJ
ajst-24984	150	21	ability	ability	NOUN
ajst-24984	150	22	to	to	PART
ajst-24984	150	23	capture	capture	VERB
ajst-24984	150	24	the	the	DET
ajst-24984	150	25	structure	structure	NOUN
ajst-24984	150	26	and	and	CCONJ
ajst-24984	150	27	relationships	relationship	NOUN
ajst-24984	150	28	between	between	ADP
ajst-24984	150	29	graph	graph	NOUN
ajst-24984	150	30	nodes	node	NOUN
ajst-24984	150	31	as	as	ADV
ajst-24984	150	32	well	well	ADV
ajst-24984	150	33	as	as	ADP
ajst-24984	150	34	syntactic	syntactic	ADJ
ajst-24984	150	35	features	feature	NOUN
ajst-24984	150	36	between	between	ADP
ajst-24984	150	37	words	word	NOUN
ajst-24984	150	38	.	.	PUNCT
ajst-24984	151	1	numerous	numerous	ADJ
ajst-24984	151	2	researchers	researcher	NOUN
ajst-24984	151	3	have	have	AUX
ajst-24984	151	4	applied	apply	VERB
ajst-24984	151	5	it	it	PRON
ajst-24984	151	6	to	to	ADP
ajst-24984	151	7	dependency	dependency	NOUN
ajst-24984	151	8	tree	tree	NOUN
ajst-24984	151	9	analysis	analysis	NOUN
ajst-24984	151	10	to	to	PART
ajst-24984	151	11	account	account	VERB
ajst-24984	151	12	for	for	ADP
ajst-24984	151	13	syntactic	syntactic	ADJ
ajst-24984	151	14	constraints	constraint	NOUN
ajst-24984	151	15	and	and	CCONJ
ajst-24984	151	16	long	long	ADJ
ajst-24984	151	17	-	-	PUNCT
ajst-24984	151	18	term	term	NOUN
ajst-24984	151	19	lexical	lexical	ADJ
ajst-24984	151	20	dependencies	dependency	NOUN
ajst-24984	151	21	.	.	PUNCT
ajst-24984	152	1	zhang	zhang	PROPN
ajst-24984	152	2	c	c	PROPN
ajst-24984	152	3	et	et	PROPN
ajst-24984	152	4	al	al	PROPN
ajst-24984	153	1	[	[	X
ajst-24984	153	2	37	37	NUM
ajst-24984	153	3	]	]	PUNCT
ajst-24984	153	4	proposed	propose	VERB
ajst-24984	153	5	to	to	PART
ajst-24984	153	6	construct	construct	VERB
ajst-24984	153	7	a	a	DET
ajst-24984	153	8	graph	graph	NOUN
ajst-24984	153	9	convolution	convolution	NOUN
ajst-24984	153	10	network	network	NOUN
ajst-24984	153	11	on	on	ADP
ajst-24984	153	12	the	the	DET
ajst-24984	153	13	dependency	dependency	NOUN
ajst-24984	153	14	tree	tree	NOUN
ajst-24984	153	15	of	of	ADP
ajst-24984	153	16	a	a	DET
ajst-24984	153	17	196	196	NUM
ajst-24984	153	18	sentence	sentence	NOUN
ajst-24984	153	19	to	to	ADP
ajst-24984	153	20	mine	mine	VERB
ajst-24984	153	21	syntactic	syntactic	ADJ
ajst-24984	153	22	information	information	NOUN
ajst-24984	153	23	and	and	CCONJ
ajst-24984	153	24	lexical	lexical	ADJ
ajst-24984	153	25	dependencies	dependency	NOUN
ajst-24984	153	26	.	.	PUNCT
ajst-24984	154	1	in	in	ADP
ajst-24984	154	2	order	order	NOUN
ajst-24984	154	3	to	to	PART
ajst-24984	154	4	enhance	enhance	VERB
ajst-24984	154	5	the	the	DET
ajst-24984	154	6	model	model	NOUN
ajst-24984	154	7	's	's	PART
ajst-24984	154	8	focus	focus	NOUN
ajst-24984	154	9	on	on	ADP
ajst-24984	154	10	local	local	ADJ
ajst-24984	154	11	syntactic	syntactic	ADJ
ajst-24984	154	12	context	context	NOUN
ajst-24984	154	13	,	,	PUNCT
ajst-24984	154	14	xiao	xiao	PROPN
ajst-24984	154	15	l	l	PROPN
ajst-24984	154	16	et	et	NOUN
ajst-24984	154	17	al	al	PROPN
ajst-24984	154	18	[	[	X
ajst-24984	154	19	38	38	NUM
ajst-24984	154	20	]	]	PUNCT
ajst-24984	154	21	constructed	construct	VERB
ajst-24984	154	22	a	a	DET
ajst-24984	154	23	context	context	NOUN
ajst-24984	154	24	dynamic	dynamic	ADJ
ajst-24984	154	25	weighting	weighting	NOUN
ajst-24984	154	26	mechanism	mechanism	NOUN
ajst-24984	154	27	incorporating	incorporate	VERB
ajst-24984	154	28	syntactic	syntactic	ADJ
ajst-24984	154	29	awareness	awareness	NOUN
ajst-24984	154	30	.	.	PUNCT
ajst-24984	155	1	meng	meng	PROPN
ajst-24984	155	2	f	f	PROPN
ajst-24984	155	3	et	et	PROPN
ajst-24984	155	4	al	al	PROPN
ajst-24984	156	1	[	[	X
ajst-24984	156	2	39	39	NUM
ajst-24984	156	3	]	]	PUNCT
ajst-24984	156	4	designed	design	VERB
ajst-24984	156	5	a	a	DET
ajst-24984	156	6	weighted	weighted	ADJ
ajst-24984	156	7	graph	graph	NOUN
ajst-24984	156	8	convolutional	convolutional	ADJ
ajst-24984	156	9	network	network	NOUN
ajst-24984	156	10	(	(	PUNCT
ajst-24984	156	11	wgcn	wgcn	PROPN
ajst-24984	156	12	)	)	PUNCT
ajst-24984	156	13	based	base	VERB
ajst-24984	156	14	on	on	ADP
ajst-24984	156	15	feature	feature	NOUN
ajst-24984	156	16	combinations	combination	NOUN
ajst-24984	156	17	to	to	PART
ajst-24984	156	18	mine	mine	VERB
ajst-24984	156	19	rich	rich	ADJ
ajst-24984	156	20	syntactic	syntactic	ADJ
ajst-24984	156	21	information	information	NOUN
ajst-24984	156	22	,	,	PUNCT
ajst-24984	156	23	which	which	PRON
ajst-24984	156	24	solved	solve	VERB
ajst-24984	156	25	the	the	DET
ajst-24984	156	26	problem	problem	NOUN
ajst-24984	156	27	of	of	ADP
ajst-24984	156	28	failing	fail	VERB
ajst-24984	156	29	to	to	PART
ajst-24984	156	30	make	make	VERB
ajst-24984	156	31	full	full	ADJ
ajst-24984	156	32	use	use	NOUN
ajst-24984	156	33	of	of	ADP
ajst-24984	156	34	the	the	DET
ajst-24984	156	35	dependency	dependency	NOUN
ajst-24984	156	36	analysis	analysis	NOUN
ajst-24984	156	37	gained	gain	VERB
ajst-24984	156	38	from	from	ADP
ajst-24984	156	39	dependency	dependency	NOUN
ajst-24984	156	40	analysis	analysis	NOUN
ajst-24984	156	41	when	when	SCONJ
ajst-24984	156	42	using	use	VERB
ajst-24984	156	43	a	a	DET
ajst-24984	156	44	graph	graph	NOUN
ajst-24984	156	45	convolutional	convolutional	ADJ
ajst-24984	156	46	network	network	NOUN
ajst-24984	156	47	on	on	ADP
ajst-24984	156	48	a	a	DET
ajst-24984	156	49	dependency	dependency	NOUN
ajst-24984	156	50	tree	tree	NOUN
ajst-24984	156	51	of	of	ADP
ajst-24984	156	52	a	a	DET
ajst-24984	156	53	sentence	sentence	NOUN
ajst-24984	156	54	.	.	PUNCT
ajst-24984	157	1	of	of	ADP
ajst-24984	157	2	syntactic	syntactic	ADJ
ajst-24984	157	3	information	information	NOUN
ajst-24984	157	4	.	.	PUNCT
ajst-24984	158	1	du	du	PROPN
ajst-24984	158	2	j	j	PROPN
ajst-24984	158	3	et	et	PROPN
ajst-24984	158	4	al	al	PROPN
ajst-24984	159	1	[	[	X
ajst-24984	159	2	40	40	NUM
ajst-24984	159	3	]	]	PUNCT
ajst-24984	159	4	designed	design	VERB
ajst-24984	159	5	an	an	DET
ajst-24984	159	6	enhancement	enhancement	NOUN
ajst-24984	159	7	module	module	NOUN
ajst-24984	159	8	that	that	PRON
ajst-24984	159	9	dynamically	dynamically	ADV
ajst-24984	159	10	updates	update	VERB
ajst-24984	159	11	dependency	dependency	NOUN
ajst-24984	159	12	edges	edge	NOUN
ajst-24984	159	13	based	base	VERB
ajst-24984	159	14	on	on	ADP
ajst-24984	159	15	dependency	dependency	NOUN
ajst-24984	159	16	and	and	CCONJ
ajst-24984	159	17	contextual	contextual	ADJ
ajst-24984	159	18	information	information	NOUN
ajst-24984	159	19	.	.	PUNCT
ajst-24984	160	1	in	in	ADP
ajst-24984	160	2	order	order	NOUN
ajst-24984	160	3	to	to	PART
ajst-24984	160	4	fully	fully	ADV
ajst-24984	160	5	analyze	analyze	VERB
ajst-24984	160	6	implicit	implicit	ADJ
ajst-24984	160	7	knowledge	knowledge	NOUN
ajst-24984	160	8	and	and	CCONJ
ajst-24984	160	9	learn	learn	VERB
ajst-24984	160	10	critical	critical	ADJ
ajst-24984	160	11	fine	fine	ADV
ajst-24984	160	12	-	-	PUNCT
ajst-24984	160	13	grained	grain	VERB
ajst-24984	160	14	implicit	implicit	ADJ
ajst-24984	160	15	distance	distance	NOUN
ajst-24984	160	16	information	information	NOUN
ajst-24984	160	17	through	through	ADP
ajst-24984	160	18	soft	soft	ADJ
ajst-24984	160	19	pruning	pruning	NOUN
ajst-24984	160	20	strategies	strategy	NOUN
ajst-24984	160	21	,	,	PUNCT
ajst-24984	160	22	xiao	xiao	PROPN
ajst-24984	160	23	l	l	PROPN
ajst-24984	160	24	et	et	NOUN
ajst-24984	160	25	al	al	PROPN
ajst-24984	161	1	[	[	X
ajst-24984	161	2	41	41	NUM
ajst-24984	161	3	]	]	PUNCT
ajst-24984	161	4	designed	design	VERB
ajst-24984	161	5	an	an	DET
ajst-24984	161	6	innovative	innovative	ADJ
ajst-24984	161	7	aspect	aspect	NOUN
ajst-24984	161	8	-	-	PUNCT
ajst-24984	161	9	based	base	VERB
ajst-24984	161	10	sentiment	sentiment	NOUN
ajst-24984	161	11	analysis	analysis	NOUN
ajst-24984	161	12	framework	framework	NOUN
ajst-24984	161	13	(	(	PUNCT
ajst-24984	161	14	sdgcn	sdgcn	PROPN
ajst-24984	161	15	)	)	PUNCT
ajst-24984	161	16	that	that	PRON
ajst-24984	161	17	incorporates	incorporate	VERB
ajst-24984	161	18	graph	graph	VERB
ajst-24984	161	19	convolutional	convolutional	ADJ
ajst-24984	161	20	networks	network	NOUN
ajst-24984	161	21	and	and	CCONJ
ajst-24984	161	22	enhances	enhance	VERB
ajst-24984	161	23	the	the	DET
ajst-24984	161	24	model	model	NOUN
ajst-24984	161	25	performance	performance	NOUN
ajst-24984	161	26	through	through	ADP
ajst-24984	161	27	distanceattention	distanceattention	NOUN
ajst-24984	161	28	guidance	guidance	NOUN
ajst-24984	161	29	.	.	PUNCT
ajst-24984	162	1	with	with	ADP
ajst-24984	162	2	the	the	DET
ajst-24984	162	3	emergence	emergence	NOUN
ajst-24984	162	4	of	of	ADP
ajst-24984	162	5	the	the	DET
ajst-24984	162	6	dependency	dependency	NOUN
ajst-24984	162	7	-	-	PUNCT
ajst-24984	162	8	based	base	VERB
ajst-24984	162	9	dual	dual	ADJ
ajst-24984	162	10	graph	graph	NOUN
ajst-24984	162	11	(	(	PUNCT
ajst-24984	162	12	ld2	ld2	PROPN
ajst-24984	162	13	g	g	NOUN
ajst-24984	162	14	)	)	PUNCT
ajst-24984	162	15	model	model	NOUN
ajst-24984	162	16	[	[	X
ajst-24984	162	17	42	42	NUM
ajst-24984	162	18	]	]	PUNCT
ajst-24984	162	19	,	,	PUNCT
ajst-24984	162	20	the	the	DET
ajst-24984	162	21	interactive	interactive	ADJ
ajst-24984	162	22	dual	dual	ADJ
ajst-24984	162	23	graph	graph	NOUN
ajst-24984	162	24	convolutional	convolutional	ADJ
ajst-24984	162	25	network	network	NOUN
ajst-24984	162	26	(	(	PUNCT
ajst-24984	162	27	inter	inter	ADJ
ajst-24984	162	28	-	-	ADJ
ajst-24984	162	29	dgcn	dgcn	ADJ
ajst-24984	162	30	)	)	PUNCT
ajst-24984	162	31	model	model	NOUN
ajst-24984	163	1	[	[	X
ajst-24984	163	2	43	43	NUM
ajst-24984	163	3	]	]	PUNCT
ajst-24984	163	4	,	,	PUNCT
ajst-24984	163	5	the	the	DET
ajst-24984	163	6	syntactically	syntactically	ADV
ajst-24984	163	7	and	and	CCONJ
ajst-24984	163	8	semantically	semantically	ADV
ajst-24984	163	9	enhanced	enhance	VERB
ajst-24984	163	10	graph	graph	NOUN
ajst-24984	163	11	convolutional	convolutional	ADJ
ajst-24984	163	12	network	network	NOUN
ajst-24984	163	13	(	(	PUNCT
ajst-24984	163	14	ssegcn	ssegcn	PROPN
ajst-24984	163	15	)	)	PUNCT
ajst-24984	163	16	model	model	NOUN
ajst-24984	164	1	[	[	X
ajst-24984	164	2	44	44	NUM
ajst-24984	164	3	]	]	PUNCT
ajst-24984	164	4	,	,	PUNCT
ajst-24984	164	5	and	and	CCONJ
ajst-24984	164	6	the	the	DET
ajst-24984	164	7	combination	combination	NOUN
ajst-24984	164	8	of	of	ADP
ajst-24984	164	9	local	local	ADJ
ajst-24984	164	10	global	global	ADJ
ajst-24984	164	11	contextual	contextual	ADJ
ajst-24984	164	12	guided	guide	VERB
ajst-24984	164	13	networks	network	NOUN
ajst-24984	164	14	(	(	PUNCT
ajst-24984	164	15	lgcg	lgcg	PROPN
ajst-24984	164	16	)	)	PUNCT
ajst-24984	164	17	model	model	NOUN
ajst-24984	164	18	[	[	X
ajst-24984	164	19	45	45	NUM
ajst-24984	164	20	]	]	PUNCT
ajst-24984	164	21	,	,	PUNCT
ajst-24984	164	22	the	the	DET
ajst-24984	164	23	research	research	NOUN
ajst-24984	164	24	community	community	NOUN
ajst-24984	164	25	has	have	AUX
ajst-24984	164	26	mitigated	mitigate	VERB
ajst-24984	164	27	the	the	DET
ajst-24984	164	28	noisy	noisy	ADJ
ajst-24984	164	29	information	information	NOUN
ajst-24984	164	30	introduced	introduce	VERB
ajst-24984	164	31	during	during	ADP
ajst-24984	164	32	the	the	DET
ajst-24984	164	33	use	use	NOUN
ajst-24984	164	34	of	of	ADP
ajst-24984	164	35	dependency	dependency	NOUN
ajst-24984	164	36	trees	tree	NOUN
ajst-24984	164	37	problem	problem	NOUN
ajst-24984	164	38	and	and	CCONJ
ajst-24984	164	39	provides	provide	VERB
ajst-24984	164	40	rich	rich	ADJ
ajst-24984	164	41	relational	relational	ADJ
ajst-24984	164	42	information	information	NOUN
ajst-24984	164	43	between	between	ADP
ajst-24984	164	44	words	word	NOUN
ajst-24984	164	45	.	.	PUNCT
ajst-24984	165	1	the	the	DET
ajst-24984	165	2	emergence	emergence	NOUN
ajst-24984	165	3	of	of	ADP
ajst-24984	165	4	these	these	DET
ajst-24984	165	5	models	model	NOUN
ajst-24984	165	6	has	have	AUX
ajst-24984	165	7	further	far	ADV
ajst-24984	165	8	advanced	advance	VERB
ajst-24984	165	9	the	the	DET
ajst-24984	165	10	field	field	NOUN
ajst-24984	165	11	of	of	ADP
ajst-24984	165	12	sentiment	sentiment	NOUN
ajst-24984	165	13	analysis	analysis	NOUN
ajst-24984	165	14	and	and	CCONJ
ajst-24984	165	15	provided	provide	VERB
ajst-24984	165	16	new	new	ADJ
ajst-24984	165	17	perspectives	perspective	NOUN
ajst-24984	165	18	for	for	ADP
ajst-24984	165	19	understanding	understand	VERB
ajst-24984	165	20	complex	complex	ADJ
ajst-24984	165	21	textual	textual	ADJ
ajst-24984	165	22	data	datum	NOUN
ajst-24984	165	23	.	.	PUNCT
ajst-24984	166	1	and	and	CCONJ
ajst-24984	166	2	with	with	ADP
ajst-24984	166	3	the	the	DET
ajst-24984	166	4	advent	advent	NOUN
ajst-24984	166	5	of	of	ADP
ajst-24984	166	6	the	the	DET
ajst-24984	166	7	transformer	transformer	NOUN
ajst-24984	166	8	[	[	X
ajst-24984	166	9	46	46	NUM
ajst-24984	166	10	]	]	PUNCT
ajst-24984	166	11	,	,	PUNCT
ajst-24984	166	12	the	the	DET
ajst-24984	166	13	number	number	NOUN
ajst-24984	166	14	of	of	ADP
ajst-24984	166	15	models	model	NOUN
ajst-24984	166	16	that	that	PRON
ajst-24984	166	17	use	use	VERB
ajst-24984	166	18	attention	attention	NOUN
ajst-24984	166	19	mechanisms	mechanism	NOUN
ajst-24984	166	20	as	as	ADP
ajst-24984	166	21	their	their	PRON
ajst-24984	166	22	primary	primary	ADJ
ajst-24984	166	23	structure	structure	NOUN
ajst-24984	166	24	has	have	AUX
ajst-24984	166	25	increased	increase	VERB
ajst-24984	166	26	dramatically	dramatically	ADV
ajst-24984	166	27	.	.	PUNCT
ajst-24984	167	1	zhao	zhao	PROPN
ajst-24984	167	2	et	et	PROPN
ajst-24984	167	3	al	al	PROPN
ajst-24984	167	4	.	.	PUNCT
ajst-24984	168	1	[	[	X
ajst-24984	168	2	47	47	NUM
ajst-24984	168	3	]	]	PUNCT
ajst-24984	168	4	describe	describe	VERB
ajst-24984	168	5	an	an	DET
ajst-24984	168	6	"	"	PUNCT
ajst-24984	168	7	attention	attention	NOUN
ajst-24984	168	8	transfer	transfer	NOUN
ajst-24984	168	9	network	network	NOUN
ajst-24984	168	10	(	(	PUNCT
ajst-24984	168	11	atn	atn	PROPN
ajst-24984	168	12	)	)	PUNCT
ajst-24984	168	13	"	"	PUNCT
ajst-24984	168	14	model	model	NOUN
ajst-24984	168	15	that	that	PRON
ajst-24984	168	16	incorporates	incorporate	VERB
ajst-24984	168	17	both	both	CCONJ
ajst-24984	168	18	attention	attention	NOUN
ajst-24984	168	19	-	-	PUNCT
ajst-24984	168	20	guided	guide	VERB
ajst-24984	168	21	and	and	CCONJ
ajst-24984	168	22	fused	fused	ADJ
ajst-24984	168	23	models	model	NOUN
ajst-24984	168	24	.	.	PUNCT
ajst-24984	169	1	the	the	DET
ajst-24984	169	2	former	former	ADJ
ajst-24984	169	3	uses	use	VERB
ajst-24984	169	4	the	the	DET
ajst-24984	169	5	attention	attention	NOUN
ajst-24984	169	6	weights	weight	NOUN
ajst-24984	169	7	of	of	ADP
ajst-24984	169	8	a	a	DET
ajst-24984	169	9	document	document	NOUN
ajst-24984	169	10	-	-	PUNCT
ajst-24984	169	11	level	level	NOUN
ajst-24984	169	12	sentiment	sentiment	NOUN
ajst-24984	169	13	analysis	analysis	NOUN
ajst-24984	169	14	model	model	NOUN
ajst-24984	169	15	as	as	ADP
ajst-24984	169	16	learning	learn	VERB
ajst-24984	169	17	signals	signal	NOUN
ajst-24984	169	18	for	for	ADP
ajst-24984	169	19	guiding	guide	VERB
ajst-24984	169	20	the	the	DET
ajst-24984	169	21	absa	absa	PROPN
ajst-24984	169	22	model	model	NOUN
ajst-24984	169	23	during	during	ADP
ajst-24984	169	24	the	the	DET
ajst-24984	169	25	training	training	NOUN
ajst-24984	169	26	phase	phase	NOUN
ajst-24984	169	27	.	.	PUNCT
ajst-24984	170	1	besides	besides	SCONJ
ajst-24984	170	2	,	,	PUNCT
ajst-24984	170	3	some	some	DET
ajst-24984	170	4	pretrained	pretraine	VERB
ajst-24984	170	5	language	language	NOUN
ajst-24984	170	6	models	model	NOUN
ajst-24984	170	7	such	such	ADJ
ajst-24984	170	8	as	as	ADP
ajst-24984	170	9	gpt	gpt	NOUN
ajst-24984	170	10	[	[	X
ajst-24984	170	11	14	14	NUM
ajst-24984	170	12	]	]	PUNCT
ajst-24984	170	13	and	and	CCONJ
ajst-24984	170	14	bert	bert	X
ajst-24984	170	15	[	[	X
ajst-24984	170	16	15	15	NUM
ajst-24984	170	17	]	]	PUNCT
ajst-24984	170	18	shine	shine	NOUN
ajst-24984	170	19	on	on	ADP
ajst-24984	170	20	various	various	ADJ
ajst-24984	170	21	text	text	NOUN
ajst-24984	170	22	processing	processing	NOUN
ajst-24984	170	23	tasks	task	NOUN
ajst-24984	170	24	and	and	CCONJ
ajst-24984	170	25	show	show	VERB
ajst-24984	170	26	extremely	extremely	ADV
ajst-24984	170	27	good	good	ADJ
ajst-24984	170	28	performance	performance	NOUN
ajst-24984	170	29	.	.	PUNCT
ajst-24984	171	1	currently	currently	ADV
ajst-24984	171	2	most	most	ADJ
ajst-24984	171	3	researchers	researcher	NOUN
ajst-24984	171	4	tend	tend	VERB
ajst-24984	171	5	to	to	PART
ajst-24984	171	6	optimize	optimize	VERB
ajst-24984	171	7	or	or	CCONJ
ajst-24984	171	8	improve	improve	VERB
ajst-24984	171	9	auto	auto	NOUN
ajst-24984	171	10	-	-	PUNCT
ajst-24984	171	11	coding	code	VERB
ajst-24984	171	12	based	base	VERB
ajst-24984	171	13	bert	bert	NOUN
ajst-24984	171	14	models	model	NOUN
ajst-24984	171	15	for	for	ADP
ajst-24984	171	16	absa	absa	NOUN
ajst-24984	171	17	tasks	task	NOUN
ajst-24984	171	18	.	.	PUNCT
ajst-24984	172	1	zhao	zhao	PROPN
ajst-24984	172	2	et	et	PROPN
ajst-24984	172	3	al	al	PROPN
ajst-24984	173	1	[	[	X
ajst-24984	173	2	48	48	NUM
ajst-24984	173	3	]	]	PUNCT
ajst-24984	173	4	argued	argue	VERB
ajst-24984	173	5	that	that	SCONJ
ajst-24984	173	6	the	the	DET
ajst-24984	173	7	two	two	NUM
ajst-24984	173	8	subtasks	subtask	NOUN
ajst-24984	173	9	of	of	ADP
ajst-24984	173	10	aspect	aspect	NOUN
ajst-24984	173	11	extraction	extraction	NOUN
ajst-24984	173	12	and	and	CCONJ
ajst-24984	173	13	aspect	aspect	NOUN
ajst-24984	173	14	sentiment	sentiment	NOUN
ajst-24984	173	15	classification	classification	NOUN
ajst-24984	173	16	in	in	ADP
ajst-24984	173	17	absa	absa	PROPN
ajst-24984	173	18	are	be	AUX
ajst-24984	173	19	intrinsically	intrinsically	ADV
ajst-24984	173	20	linked	link	VERB
ajst-24984	173	21	,	,	PUNCT
ajst-24984	173	22	and	and	CCONJ
ajst-24984	173	23	also	also	ADV
ajst-24984	173	24	they	they	PRON
ajst-24984	173	25	found	find	VERB
ajst-24984	173	26	that	that	SCONJ
ajst-24984	173	27	the	the	DET
ajst-24984	173	28	performance	performance	NOUN
ajst-24984	173	29	of	of	ADP
ajst-24984	173	30	aspect	aspect	NOUN
ajst-24984	173	31	sentiment	sentiment	NOUN
ajst-24984	173	32	classification	classification	NOUN
ajst-24984	173	33	task	task	NOUN
ajst-24984	173	34	can	can	AUX
ajst-24984	173	35	be	be	AUX
ajst-24984	173	36	further	far	ADV
ajst-24984	173	37	improved	improve	VERB
ajst-24984	173	38	by	by	ADP
ajst-24984	173	39	utilizing	utilize	VERB
ajst-24984	173	40	the	the	DET
ajst-24984	173	41	dependency	dependency	NOUN
ajst-24984	173	42	syntax	syntax	NOUN
ajst-24984	173	43	information	information	NOUN
ajst-24984	173	44	of	of	ADP
ajst-24984	173	45	graph	graph	NOUN
ajst-24984	173	46	neural	neural	ADJ
ajst-24984	173	47	networks	network	NOUN
ajst-24984	173	48	,	,	PUNCT
ajst-24984	173	49	so	so	SCONJ
ajst-24984	173	50	this	this	DET
ajst-24984	173	51	research	research	NOUN
ajst-24984	173	52	team	team	NOUN
ajst-24984	173	53	utilized	utilize	VERB
ajst-24984	173	54	the	the	DET
ajst-24984	173	55	multi	multi	ADJ
ajst-24984	173	56	-	-	ADJ
ajst-24984	173	57	head	head	ADJ
ajst-24984	173	58	attention	attention	NOUN
ajst-24984	173	59	to	to	ADP
ajst-24984	173	60	associate	associate	VERB
ajst-24984	173	61	dependency	dependency	NOUN
ajst-24984	173	62	trees	tree	NOUN
ajst-24984	173	63	with	with	ADP
ajst-24984	173	64	aspect	aspect	NOUN
ajst-24984	173	65	extraction	extraction	NOUN
ajst-24984	173	66	and	and	CCONJ
ajst-24984	173	67	use	use	VERB
ajst-24984	173	68	bert	bert	NOUN
ajst-24984	173	69	to	to	PART
ajst-24984	173	70	process	process	VERB
ajst-24984	173	71	the	the	DET
ajst-24984	173	72	original	original	ADJ
ajst-24984	173	73	text	text	NOUN
ajst-24984	173	74	and	and	CCONJ
ajst-24984	173	75	aspect	aspect	NOUN
ajst-24984	173	76	features	feature	NOUN
ajst-24984	173	77	proposed	propose	VERB
ajst-24984	173	78	a	a	DET
ajst-24984	173	79	multi	multi	ADJ
ajst-24984	173	80	-	-	ADJ
ajst-24984	173	81	task	task	ADJ
ajst-24984	173	82	learning	learning	NOUN
ajst-24984	173	83	model	model	NOUN
ajst-24984	173	84	which	which	PRON
ajst-24984	173	85	improved	improve	VERB
ajst-24984	173	86	the	the	DET
ajst-24984	173	87	overall	overall	ADJ
ajst-24984	173	88	performance	performance	NOUN
ajst-24984	173	89	of	of	ADP
ajst-24984	173	90	the	the	DET
ajst-24984	173	91	model	model	NOUN
ajst-24984	173	92	on	on	ADP
ajst-24984	173	93	the	the	DET
ajst-24984	173	94	absa	absa	PROPN
ajst-24984	173	95	task	task	NOUN
ajst-24984	173	96	.	.	PUNCT
ajst-24984	174	1	liu	liu	PROPN
ajst-24984	174	2	et	et	PROPN
ajst-24984	174	3	al	al	PROPN
ajst-24984	175	1	[	[	X
ajst-24984	175	2	49	49	NUM
ajst-24984	175	3	]	]	PUNCT
ajst-24984	175	4	found	find	VERB
ajst-24984	175	5	that	that	SCONJ
ajst-24984	175	6	severe	severe	ADJ
ajst-24984	175	7	domain	domain	NOUN
ajst-24984	175	8	shifting	shifting	NOUN
ajst-24984	175	9	leads	lead	VERB
ajst-24984	175	10	to	to	ADP
ajst-24984	175	11	poor	poor	ADJ
ajst-24984	175	12	results	result	NOUN
ajst-24984	175	13	when	when	SCONJ
ajst-24984	175	14	using	use	VERB
ajst-24984	175	15	self	self	NOUN
ajst-24984	175	16	-	-	PUNCT
ajst-24984	175	17	coding	code	VERB
ajst-24984	175	18	pre	pre	ADJ
ajst-24984	175	19	-	-	ADJ
ajst-24984	175	20	training	training	ADJ
ajst-24984	175	21	models	model	NOUN
ajst-24984	175	22	such	such	ADJ
ajst-24984	175	23	as	as	ADP
ajst-24984	175	24	bert	bert	NOUN
ajst-24984	175	25	to	to	PART
ajst-24984	175	26	process	process	VERB
ajst-24984	175	27	downstream	downstream	ADJ
ajst-24984	175	28	tasks	task	NOUN
ajst-24984	175	29	of	of	ADP
ajst-24984	175	30	absa	absa	PROPN
ajst-24984	175	31	.	.	PUNCT
ajst-24984	176	1	to	to	PART
ajst-24984	176	2	address	address	VERB
ajst-24984	176	3	this	this	DET
ajst-24984	176	4	problem	problem	NOUN
ajst-24984	176	5	liu	liu	PROPN
ajst-24984	176	6	et	et	PROPN
ajst-24984	176	7	al	al	PROPN
ajst-24984	176	8	.	.	PROPN
ajst-24984	176	9	used	use	VERB
ajst-24984	176	10	a	a	DET
ajst-24984	176	11	stepwise	stepwise	ADJ
ajst-24984	176	12	training	training	NOUN
ajst-24984	176	13	approach	approach	NOUN
ajst-24984	176	14	.	.	PUNCT
ajst-24984	177	1	first	first	ADV
ajst-24984	177	2	,	,	PUNCT
ajst-24984	177	3	the	the	DET
ajst-24984	177	4	team	team	NOUN
ajst-24984	177	5	selected	select	VERB
ajst-24984	177	6	instances	instance	NOUN
ajst-24984	177	7	related	relate	VERB
ajst-24984	177	8	to	to	ADP
ajst-24984	177	9	the	the	DET
ajst-24984	177	10	target	target	NOUN
ajst-24984	177	11	domain	domain	NOUN
ajst-24984	177	12	from	from	ADP
ajst-24984	177	13	a	a	DET
ajst-24984	177	14	large	large	ADJ
ajst-24984	177	15	-	-	PUNCT
ajst-24984	177	16	scale	scale	NOUN
ajst-24984	177	17	pre	pre	ADJ
ajst-24984	177	18	-	-	ADJ
ajst-24984	177	19	training	training	ADJ
ajst-24984	177	20	dataset	dataset	NOUN
ajst-24984	177	21	to	to	PART
ajst-24984	177	22	align	align	VERB
ajst-24984	177	23	the	the	DET
ajst-24984	177	24	instances	instance	NOUN
ajst-24984	177	25	between	between	ADP
ajst-24984	177	26	the	the	DET
ajst-24984	177	27	pre	pre	NOUN
ajst-24984	177	28	-	-	NOUN
ajst-24984	177	29	training	training	ADJ
ajst-24984	177	30	and	and	CCONJ
ajst-24984	177	31	target	target	NOUN
ajst-24984	177	32	domains	domain	NOUN
ajst-24984	177	33	,	,	PUNCT
ajst-24984	177	34	and	and	CCONJ
ajst-24984	177	35	used	use	VERB
ajst-24984	177	36	these	these	DET
ajst-24984	177	37	data	datum	NOUN
ajst-24984	177	38	for	for	ADP
ajst-24984	177	39	the	the	DET
ajst-24984	177	40	first	first	ADJ
ajst-24984	177	41	stage	stage	NOUN
ajst-24984	177	42	of	of	ADP
ajst-24984	177	43	training	training	NOUN
ajst-24984	177	44	.	.	PUNCT
ajst-24984	178	1	then	then	ADV
ajst-24984	178	2	,	,	PUNCT
ajst-24984	178	3	a	a	DET
ajst-24984	178	4	knowledge	knowledge	NOUN
ajst-24984	178	5	-	-	PUNCT
ajst-24984	178	6	based	base	VERB
ajst-24984	178	7	guidance	guidance	NOUN
ajst-24984	178	8	strategy	strategy	NOUN
ajst-24984	178	9	was	be	AUX
ajst-24984	178	10	introduced	introduce	VERB
ajst-24984	178	11	in	in	ADP
ajst-24984	178	12	the	the	DET
ajst-24984	178	13	second	second	ADJ
ajst-24984	178	14	phase	phase	NOUN
ajst-24984	178	15	of	of	ADP
ajst-24984	178	16	training	training	NOUN
ajst-24984	178	17	to	to	PART
ajst-24984	178	18	further	far	ADV
ajst-24984	178	19	bridge	bridge	VERB
ajst-24984	178	20	the	the	DET
ajst-24984	178	21	domain	domain	NOUN
ajst-24984	178	22	gap	gap	NOUN
ajst-24984	178	23	at	at	ADP
ajst-24984	178	24	the	the	DET
ajst-24984	178	25	knowledge	knowledge	NOUN
ajst-24984	178	26	level	level	NOUN
ajst-24984	178	27	.	.	PUNCT
ajst-24984	179	1	finally	finally	ADV
ajst-24984	179	2	,	,	PUNCT
ajst-24984	179	3	in	in	ADP
ajst-24984	179	4	the	the	DET
ajst-24984	179	5	third	third	ADJ
ajst-24984	179	6	phase	phase	NOUN
ajst-24984	179	7	,	,	PUNCT
ajst-24984	179	8	the	the	DET
ajst-24984	179	9	learner	learner	NOUN
ajst-24984	179	10	model	model	NOUN
ajst-24984	179	11	is	be	AUX
ajst-24984	179	12	fine	fine	ADV
ajst-24984	179	13	-	-	PUNCT
ajst-24984	179	14	tuned	tune	VERB
ajst-24984	179	15	to	to	PART
ajst-24984	179	16	better	well	ADV
ajst-24984	179	17	adapt	adapt	VERB
ajst-24984	179	18	its	its	PRON
ajst-24984	179	19	learned	learn	VERB
ajst-24984	179	20	knowledge	knowledge	NOUN
ajst-24984	179	21	to	to	ADP
ajst-24984	179	22	the	the	DET
ajst-24984	179	23	target	target	NOUN
ajst-24984	179	24	dataset	dataset	VERB
ajst-24984	179	25	.	.	PUNCT
ajst-24984	180	1	4	4	X
ajst-24984	180	2	.	.	X
ajst-24984	180	3	conclusion	conclusion	NOUN
ajst-24984	180	4	aspect	aspect	NOUN
ajst-24984	180	5	-	-	PUNCT
ajst-24984	180	6	based	base	VERB
ajst-24984	180	7	sentiment	sentiment	NOUN
ajst-24984	180	8	analysis	analysis	NOUN
ajst-24984	180	9	is	be	AUX
ajst-24984	180	10	an	an	DET
ajst-24984	180	11	important	important	ADJ
ajst-24984	180	12	research	research	NOUN
ajst-24984	180	13	area	area	NOUN
ajst-24984	180	14	in	in	ADP
ajst-24984	180	15	the	the	DET
ajst-24984	180	16	field	field	NOUN
ajst-24984	180	17	of	of	ADP
ajst-24984	180	18	natural	natural	ADJ
ajst-24984	180	19	language	language	NOUN
ajst-24984	180	20	processing	processing	NOUN
ajst-24984	180	21	,	,	PUNCT
ajst-24984	180	22	which	which	PRON
ajst-24984	180	23	provides	provide	VERB
ajst-24984	180	24	data	datum	NOUN
ajst-24984	180	25	support	support	NOUN
ajst-24984	180	26	for	for	ADP
ajst-24984	180	27	decision	decision	NOUN
ajst-24984	180	28	making	making	NOUN
ajst-24984	180	29	by	by	ADP
ajst-24984	180	30	identifying	identify	VERB
ajst-24984	180	31	and	and	CCONJ
ajst-24984	180	32	evaluating	evaluate	VERB
ajst-24984	180	33	the	the	DET
ajst-24984	180	34	sentiment	sentiment	NOUN
ajst-24984	180	35	tendencies	tendency	NOUN
ajst-24984	180	36	of	of	ADP
ajst-24984	180	37	specific	specific	ADJ
ajst-24984	180	38	aspects	aspect	NOUN
ajst-24984	180	39	of	of	ADP
ajst-24984	180	40	a	a	DET
ajst-24984	180	41	text	text	NOUN
ajst-24984	180	42	.	.	PUNCT
ajst-24984	181	1	research	research	NOUN
ajst-24984	181	2	methods	method	NOUN
ajst-24984	181	3	in	in	ADP
ajst-24984	181	4	this	this	DET
ajst-24984	181	5	area	area	NOUN
ajst-24984	181	6	have	have	AUX
ajst-24984	181	7	evolved	evolve	VERB
ajst-24984	181	8	from	from	ADP
ajst-24984	181	9	traditional	traditional	ADJ
ajst-24984	181	10	approaches	approach	NOUN
ajst-24984	181	11	based	base	VERB
ajst-24984	181	12	on	on	ADP
ajst-24984	181	13	sentiment	sentiment	NOUN
ajst-24984	181	14	dictionaries	dictionary	NOUN
ajst-24984	181	15	,	,	PUNCT
ajst-24984	181	16	to	to	ADP
ajst-24984	181	17	machine	machine	NOUN
ajst-24984	181	18	learning	learning	NOUN
ajst-24984	181	19	algorithms	algorithm	NOUN
ajst-24984	181	20	,	,	PUNCT
ajst-24984	181	21	to	to	ADP
ajst-24984	181	22	deep	deep	ADJ
ajst-24984	181	23	learning	learning	NOUN
ajst-24984	181	24	models	model	NOUN
ajst-24984	181	25	.	.	PUNCT
ajst-24984	182	1	each	each	PRON
ajst-24984	182	2	of	of	ADP
ajst-24984	182	3	these	these	DET
ajst-24984	182	4	approaches	approach	NOUN
ajst-24984	182	5	has	have	VERB
ajst-24984	182	6	its	its	PRON
ajst-24984	182	7	own	own	ADJ
ajst-24984	182	8	characteristics	characteristic	NOUN
ajst-24984	182	9	,	,	PUNCT
ajst-24984	182	10	with	with	ADP
ajst-24984	182	11	sentiment	sentiment	NOUN
ajst-24984	182	12	dictionary	dictionary	ADJ
ajst-24984	182	13	approaches	approach	NOUN
ajst-24984	182	14	relying	rely	VERB
ajst-24984	182	15	on	on	ADP
ajst-24984	182	16	pre	pre	ADJ
ajst-24984	182	17	-	-	ADJ
ajst-24984	182	18	constructed	constructed	ADJ
ajst-24984	182	19	sentiment	sentiment	NOUN
ajst-24984	182	20	vocabulary	vocabulary	NOUN
ajst-24984	182	21	resources	resource	NOUN
ajst-24984	182	22	,	,	PUNCT
ajst-24984	182	23	machine	machine	NOUN
ajst-24984	182	24	learning	learning	NOUN
ajst-24984	182	25	approaches	approach	NOUN
ajst-24984	182	26	learning	learn	VERB
ajst-24984	182	27	features	feature	VERB
ajst-24984	182	28	automatically	automatically	ADV
ajst-24984	182	29	through	through	ADP
ajst-24984	182	30	algorithms	algorithm	NOUN
ajst-24984	182	31	,	,	PUNCT
ajst-24984	182	32	and	and	CCONJ
ajst-24984	182	33	deep	deep	ADJ
ajst-24984	182	34	learning	learning	NOUN
ajst-24984	182	35	approaches	approach	VERB
ajst-24984	182	36	utilizing	utilize	VERB
ajst-24984	182	37	neural	neural	ADJ
ajst-24984	182	38	networks	network	NOUN
ajst-24984	182	39	to	to	PART
ajst-24984	182	40	capture	capture	VERB
ajst-24984	182	41	complex	complex	ADJ
ajst-24984	182	42	text	text	NOUN
ajst-24984	182	43	features	feature	NOUN
ajst-24984	182	44	.	.	PUNCT
ajst-24984	183	1	with	with	ADP
ajst-24984	183	2	the	the	DET
ajst-24984	183	3	increase	increase	NOUN
ajst-24984	183	4	in	in	ADP
ajst-24984	183	5	computational	computational	ADJ
ajst-24984	183	6	power	power	NOUN
ajst-24984	183	7	and	and	CCONJ
ajst-24984	183	8	data	datum	NOUN
ajst-24984	183	9	volume	volume	NOUN
ajst-24984	183	10	,	,	PUNCT
ajst-24984	183	11	large	large	ADJ
ajst-24984	183	12	models	model	NOUN
ajst-24984	183	13	,	,	PUNCT
ajst-24984	183	14	various	various	ADJ
ajst-24984	183	15	transformers	transformer	NOUN
ajst-24984	183	16	and	and	CCONJ
ajst-24984	183	17	their	their	PRON
ajst-24984	183	18	variants	variant	NOUN
ajst-24984	183	19	are	be	AUX
ajst-24984	183	20	increasingly	increasingly	ADV
ajst-24984	183	21	used	use	VERB
ajst-24984	183	22	in	in	ADP
ajst-24984	183	23	aspect	aspect	NOUN
ajst-24984	183	24	-	-	PUNCT
ajst-24984	183	25	based	base	VERB
ajst-24984	183	26	sentiment	sentiment	NOUN
ajst-24984	183	27	analysis	analysis	NOUN
ajst-24984	183	28	,	,	PUNCT
ajst-24984	183	29	which	which	PRON
ajst-24984	183	30	are	be	AUX
ajst-24984	183	31	capable	capable	ADJ
ajst-24984	183	32	of	of	ADP
ajst-24984	183	33	handling	handle	VERB
ajst-24984	183	34	more	more	ADV
ajst-24984	183	35	complex	complex	ADJ
ajst-24984	183	36	tasks	task	NOUN
ajst-24984	183	37	and	and	CCONJ
ajst-24984	183	38	providing	provide	VERB
ajst-24984	183	39	more	more	ADV
ajst-24984	183	40	accurate	accurate	ADJ
ajst-24984	183	41	results	result	NOUN
ajst-24984	183	42	.	.	PUNCT
ajst-24984	184	1	in	in	ADP
ajst-24984	184	2	addition	addition	NOUN
ajst-24984	184	3	,	,	PUNCT
ajst-24984	184	4	multimodal	multimodal	ADJ
ajst-24984	184	5	aspect	aspect	NOUN
ajst-24984	184	6	-	-	PUNCT
ajst-24984	184	7	based	base	VERB
ajst-24984	184	8	sentiment	sentiment	NOUN
ajst-24984	184	9	analysis	analysis	NOUN
ajst-24984	184	10	combines	combine	VERB
ajst-24984	184	11	visual	visual	ADJ
ajst-24984	184	12	and	and	CCONJ
ajst-24984	184	13	auditory	auditory	ADJ
ajst-24984	184	14	information	information	NOUN
ajst-24984	184	15	other	other	ADJ
ajst-24984	184	16	than	than	ADP
ajst-24984	184	17	text	text	NOUN
ajst-24984	184	18	,	,	PUNCT
ajst-24984	184	19	such	such	ADJ
ajst-24984	184	20	as	as	ADP
ajst-24984	184	21	videos	video	NOUN
ajst-24984	184	22	,	,	PUNCT
ajst-24984	184	23	images	image	NOUN
ajst-24984	184	24	,	,	PUNCT
ajst-24984	184	25	and	and	CCONJ
ajst-24984	184	26	speech	speech	NOUN
ajst-24984	184	27	,	,	PUNCT
ajst-24984	184	28	to	to	PART
ajst-24984	184	29	provide	provide	VERB
ajst-24984	184	30	a	a	DET
ajst-24984	184	31	richer	rich	ADJ
ajst-24984	184	32	context	context	NOUN
ajst-24984	184	33	for	for	ADP
ajst-24984	184	34	sentiment	sentiment	NOUN
ajst-24984	184	35	analysis	analysis	NOUN
ajst-24984	184	36	and	and	CCONJ
ajst-24984	184	37	enhance	enhance	VERB
ajst-24984	184	38	the	the	DET
ajst-24984	184	39	accuracy	accuracy	NOUN
ajst-24984	184	40	and	and	CCONJ
ajst-24984	184	41	usefulness	usefulness	NOUN
ajst-24984	184	42	of	of	ADP
ajst-24984	184	43	the	the	DET
ajst-24984	184	44	analysis	analysis	NOUN
ajst-24984	184	45	.	.	PUNCT
ajst-24984	185	1	acknowledgments	acknowledgment	NOUN
ajst-24984	185	2	this	this	DET
ajst-24984	185	3	work	work	NOUN
ajst-24984	185	4	was	be	AUX
ajst-24984	185	5	supported	support	VERB
ajst-24984	185	6	by	by	ADP
ajst-24984	185	7	the	the	DET
ajst-24984	185	8	2023	2023	NUM
ajst-24984	185	9	high	high	ADJ
ajst-24984	185	10	-	-	PUNCT
ajst-24984	185	11	quality	quality	NOUN
ajst-24984	185	12	innovation	innovation	NOUN
ajst-24984	185	13	project	project	NOUN
ajst-24984	185	14	for	for	ADP
ajst-24984	185	15	postgraduates	postgraduate	NOUN
ajst-24984	185	16	of	of	ADP
ajst-24984	185	17	chongqing	chongqing	PROPN
ajst-24984	185	18	university	university	PROPN
ajst-24984	185	19	of	of	ADP
ajst-24984	185	20	technology	technology	NOUN
ajst-24984	185	21	(	(	PUNCT
ajst-24984	185	22	gzlcx20233249	gzlcx20233249	PROPN
ajst-24984	185	23	)	)	PUNCT
ajst-24984	185	24	.	.	PUNCT
ajst-24984	186	1	references	reference	NOUN
ajst-24984	187	1	[	[	X
ajst-24984	187	2	1	1	NUM
ajst-24984	187	3	]	]	X
ajst-24984	187	4	lauriola	lauriola	PROPN
ajst-24984	187	5	i	i	PRON
ajst-24984	187	6	,	,	PUNCT
ajst-24984	187	7	lavelli	lavelli	PROPN
ajst-24984	187	8	a	a	PROPN
ajst-24984	187	9	,	,	PUNCT
ajst-24984	187	10	aiolli	aiolli	PROPN
ajst-24984	187	11	f.	f.	PROPN
ajst-24984	188	1	an	an	DET
ajst-24984	188	2	introduction	introduction	NOUN
ajst-24984	188	3	to	to	ADP
ajst-24984	188	4	deep	deep	ADJ
ajst-24984	188	5	learning	learning	NOUN
ajst-24984	188	6	in	in	ADP
ajst-24984	188	7	natural	natural	ADJ
ajst-24984	188	8	language	language	NOUN
ajst-24984	188	9	processing	processing	NOUN
ajst-24984	188	10	:	:	PUNCT
ajst-24984	188	11	models	model	NOUN
ajst-24984	188	12	,	,	PUNCT
ajst-24984	188	13	techniques	technique	NOUN
ajst-24984	188	14	,	,	PUNCT
ajst-24984	188	15	and	and	CCONJ
ajst-24984	188	16	tools[j	tools[j	PROPN
ajst-24984	188	17	]	]	PUNCT
ajst-24984	188	18	.	.	PUNCT
ajst-24984	189	1	neurocomputing	neurocomputing	NOUN
ajst-24984	189	2	,	,	PUNCT
ajst-24984	189	3	2022	2022	NUM
ajst-24984	189	4	,	,	PUNCT
ajst-24984	189	5	470	470	NUM
ajst-24984	189	6	:	:	SYM
ajst-24984	189	7	443	443	NUM
ajst-24984	189	8	-	-	NUM
ajst-24984	189	9	456	456	NUM
ajst-24984	189	10	.	.	PUNCT
ajst-24984	190	1	[	[	X
ajst-24984	190	2	2	2	NUM
ajst-24984	190	3	]	]	X
ajst-24984	190	4	zhang	zhang	PROPN
ajst-24984	190	5	y	y	PROPN
ajst-24984	190	6	,	,	PUNCT
ajst-24984	190	7	li	li	PROPN
ajst-24984	190	8	t.	t.	PROPN
ajst-24984	190	9	review	review	PROPN
ajst-24984	190	10	of	of	ADP
ajst-24984	190	11	comment	comment	NOUN
ajst-24984	190	12	-	-	PUNCT
ajst-24984	190	13	oriented	orient	VERB
ajst-24984	190	14	aspect	aspect	NOUN
ajst-24984	190	15	-	-	PUNCT
ajst-24984	190	16	based	base	VERB
ajst-24984	190	17	sentiment	sentiment	NOUN
ajst-24984	190	18	analysis[j	analysis[j	PROPN
ajst-24984	190	19	]	]	PUNCT
ajst-24984	190	20	.	.	PUNCT
ajst-24984	191	1	comput	comput	NOUN
ajst-24984	191	2	.	.	PUNCT
ajst-24984	192	1	sci	sci	PROPN
ajst-24984	192	2	,	,	PUNCT
ajst-24984	192	3	2020	2020	NUM
ajst-24984	192	4	,	,	PUNCT
ajst-24984	192	5	47(7	47(7	NOUN
ajst-24984	192	6	)	)	PUNCT
ajst-24984	192	7	.	.	PUNCT
ajst-24984	193	1	[	[	X
ajst-24984	193	2	3	3	NUM
ajst-24984	193	3	]	]	X
ajst-24984	193	4	n.	n.	NOUN
ajst-24984	193	5	majumder	majumder	PROPN
ajst-24984	193	6	,	,	PUNCT
ajst-24984	193	7	s.	s.	PROPN
ajst-24984	193	8	poria	poria	PROPN
ajst-24984	193	9	,	,	PUNCT
ajst-24984	193	10	h.	h.	PROPN
ajst-24984	193	11	peng	peng	PROPN
ajst-24984	193	12	,	,	PUNCT
ajst-24984	193	13	n.	n.	PROPN
ajst-24984	193	14	chhaya	chhaya	PROPN
ajst-24984	193	15	,	,	PUNCT
ajst-24984	193	16	e.	e.	PROPN
ajst-24984	193	17	cambria	cambria	PROPN
ajst-24984	193	18	and	and	CCONJ
ajst-24984	193	19	a.	a.	NOUN
ajst-24984	193	20	gelbukh	gelbukh	PROPN
ajst-24984	193	21	,	,	PUNCT
ajst-24984	193	22	"	"	PUNCT
ajst-24984	193	23	sentiment	sentiment	NOUN
ajst-24984	193	24	and	and	CCONJ
ajst-24984	193	25	sarcasm	sarcasm	NOUN
ajst-24984	193	26	classification	classification	NOUN
ajst-24984	193	27	with	with	ADP
ajst-24984	193	28	multitask	multitask	ADJ
ajst-24984	193	29	learning	learning	NOUN
ajst-24984	193	30	,	,	PUNCT
ajst-24984	193	31	"	"	PUNCT
ajst-24984	193	32	in	in	ADP
ajst-24984	193	33	ieee	ieee	NOUN
ajst-24984	193	34	intelligent	intelligent	ADJ
ajst-24984	193	35	systems	system	NOUN
ajst-24984	193	36	,	,	PUNCT
ajst-24984	193	37	vol	vol	NOUN
ajst-24984	193	38	.	.	PROPN
ajst-24984	193	39	34	34	NUM
ajst-24984	193	40	,	,	PUNCT
ajst-24984	193	41	no	no	INTJ
ajst-24984	193	42	.	.	NOUN
ajst-24984	193	43	3	3	NUM
ajst-24984	193	44	,	,	PUNCT
ajst-24984	193	45	pp	pp	ADJ
ajst-24984	193	46	.	.	PUNCT
ajst-24984	194	1	38	38	NUM
ajst-24984	194	2	-	-	SYM
ajst-24984	194	3	43	43	NUM
ajst-24984	194	4	,	,	PUNCT
ajst-24984	194	5	1	1	NUM
ajst-24984	194	6	may	may	PROPN
ajst-24984	194	7	-	-	PUNCT
ajst-24984	194	8	june	june	PROPN
ajst-24984	194	9	2019	2019	NUM
ajst-24984	194	10	,	,	PUNCT
ajst-24984	194	11	doi	doi	NOUN
ajst-24984	194	12	:	:	PUNCT
ajst-24984	194	13	10.1109/	10.1109/	NUM
ajst-24984	194	14	mis	mis	X
ajst-24984	194	15	.	.	PROPN
ajst-24984	194	16	2019	2019	NUM
ajst-24984	194	17	.	.	PUNCT
ajst-24984	195	1	2904691	2904691	NUM
ajst-24984	195	2	.	.	PUNCT
ajst-24984	196	1	[	[	X
ajst-24984	196	2	4	4	NUM
ajst-24984	196	3	]	]	X
ajst-24984	196	4	tang	tang	X
ajst-24984	196	5	x	x	SYM
ajst-24984	196	6	b	b	PROPN
ajst-24984	196	7	,	,	PUNCT
ajst-24984	196	8	liu	liu	PROPN
ajst-24984	196	9	g	g	PROPN
ajst-24984	196	10	c.	c.	PROPN
ajst-24984	196	11	research	research	PROPN
ajst-24984	196	12	review	review	NOUN
ajst-24984	196	13	on	on	ADP
ajst-24984	196	14	fine	fine	ADV
ajst-24984	196	15	-	-	PUNCT
ajst-24984	196	16	grained	grain	VERB
ajst-24984	196	17	sentiment	sentiment	NOUN
ajst-24984	196	18	analysis[j	analysis[j	PROPN
ajst-24984	196	19	]	]	PUNCT
ajst-24984	196	20	.	.	PUNCT
ajst-24984	197	1	library	library	NOUN
ajst-24984	197	2	and	and	CCONJ
ajst-24984	197	3	information	information	NOUN
ajst-24984	197	4	service	service	NOUN
ajst-24984	197	5	,	,	PUNCT
ajst-24984	197	6	2017	2017	NUM
ajst-24984	197	7	,	,	PUNCT
ajst-24984	197	8	61(5	61(5	NUM
ajst-24984	197	9	):	):	PUNCT
ajst-24984	197	10	132140	132140	NUM
ajst-24984	197	11	.	.	PUNCT
ajst-24984	198	1	[	[	X
ajst-24984	198	2	5	5	X
ajst-24984	198	3	]	]	PUNCT
ajst-24984	198	4	b.	b.	PROPN
ajst-24984	198	5	liu	liu	PROPN
ajst-24984	198	6	and	and	CCONJ
ajst-24984	198	7	l.	l.	PROPN
ajst-24984	198	8	zhang	zhang	PROPN
ajst-24984	198	9	,	,	PUNCT
ajst-24984	198	10	“	"	PUNCT
ajst-24984	198	11	a	a	DET
ajst-24984	198	12	survey	survey	NOUN
ajst-24984	198	13	of	of	ADP
ajst-24984	198	14	opinion	opinion	NOUN
ajst-24984	198	15	mining	mining	NOUN
ajst-24984	198	16	and	and	CCONJ
ajst-24984	198	17	sentiment	sentiment	NOUN
ajst-24984	198	18	analysis	analysis	NOUN
ajst-24984	198	19	,	,	PUNCT
ajst-24984	198	20	”	"	PUNCT
ajst-24984	198	21	in	in	ADP
ajst-24984	198	22	mining	mining	NOUN
ajst-24984	198	23	text	text	NOUN
ajst-24984	198	24	data	datum	NOUN
ajst-24984	198	25	.	.	PUNCT
ajst-24984	199	1	springer	springer	NOUN
ajst-24984	199	2	,	,	PUNCT
ajst-24984	199	3	2012	2012	NUM
ajst-24984	199	4	,	,	PUNCT
ajst-24984	199	5	pp	pp	ADV
ajst-24984	199	6	.	.	PUNCT
ajst-24984	200	1	415–463	415–463	NUM
ajst-24984	200	2	.	.	PUNCT
ajst-24984	201	1	[	[	X
ajst-24984	201	2	6	6	NUM
ajst-24984	201	3	]	]	PUNCT
ajst-24984	201	4	m.	m.	NOUN
ajst-24984	201	5	pontiki	pontiki	PROPN
ajst-24984	201	6	,	,	PUNCT
ajst-24984	201	7	d.	d.	PROPN
ajst-24984	201	8	galanis	galanis	PROPN
ajst-24984	201	9	,	,	PUNCT
ajst-24984	201	10	j.	j.	PROPN
ajst-24984	201	11	pavlopoulos	pavlopoulos	PROPN
ajst-24984	201	12	,	,	PUNCT
ajst-24984	201	13	h.	h.	PROPN
ajst-24984	201	14	papageorgiou	papageorgiou	PROPN
ajst-24984	201	15	,	,	PUNCT
ajst-24984	201	16	i.	i.	PROPN
ajst-24984	201	17	androut	androut	PROPN
ajst-24984	201	18	-	-	PUNCT
ajst-24984	201	19	sopoulos	sopoulo	NOUN
ajst-24984	201	20	,	,	PUNCT
ajst-24984	201	21	and	and	CCONJ
ajst-24984	201	22	s.	s.	PROPN
ajst-24984	201	23	manandhar	manandhar	PROPN
ajst-24984	201	24	,	,	PUNCT
ajst-24984	201	25	“	"	PUNCT
ajst-24984	201	26	semeval-2014	semeval-2014	NOUN
ajst-24984	201	27	task	task	NOUN
ajst-24984	201	28	4	4	NUM
ajst-24984	201	29	:	:	PUNCT
ajst-24984	201	30	aspect	aspect	NOUN
ajst-24984	201	31	based	base	VERB
ajst-24984	201	32	sentiment	sentiment	NOUN
ajst-24984	201	33	analysis	analysis	NOUN
ajst-24984	201	34	,	,	PUNCT
ajst-24984	201	35	”	"	PUNCT
ajst-24984	201	36	in	in	ADP
ajst-24984	201	37	semeval	semeval	NOUN
ajst-24984	201	38	2014	2014	NUM
ajst-24984	201	39	,	,	PUNCT
ajst-24984	201	40	2014	2014	NUM
ajst-24984	201	41	,	,	PUNCT
ajst-24984	201	42	pp	pp	ADV
ajst-24984	201	43	.	.	PUNCT
ajst-24984	202	1	27–35	27–35	NUM
ajst-24984	202	2	.	.	PUNCT
ajst-24984	203	1	[	[	X
ajst-24984	203	2	7	7	X
ajst-24984	203	3	]	]	X
ajst-24984	203	4	d.	d.	PROPN
ajst-24984	203	5	tang	tang	PROPN
ajst-24984	203	6	,	,	PUNCT
ajst-24984	203	7	b.	b.	PROPN
ajst-24984	203	8	qin	qin	PROPN
ajst-24984	203	9	,	,	PUNCT
ajst-24984	203	10	and	and	CCONJ
ajst-24984	203	11	t.	t.	PROPN
ajst-24984	203	12	liu	liu	PROPN
ajst-24984	203	13	,	,	PUNCT
ajst-24984	203	14	“	"	PUNCT
ajst-24984	203	15	aspect	aspect	NOUN
ajst-24984	203	16	level	level	NOUN
ajst-24984	203	17	sentiment	sentiment	NOUN
ajst-24984	203	18	classification	classification	NOUN
ajst-24984	203	19	with	with	ADP
ajst-24984	203	20	deep	deep	ADJ
ajst-24984	203	21	memory	memory	NOUN
ajst-24984	203	22	network	network	NOUN
ajst-24984	203	23	,	,	PUNCT
ajst-24984	203	24	”	"	PUNCT
ajst-24984	203	25	in	in	ADP
ajst-24984	203	26	proceedings	proceeding	NOUN
ajst-24984	203	27	of	of	ADP
ajst-24984	203	28	the	the	DET
ajst-24984	203	29	2016	2016	NUM
ajst-24984	203	30	conference	conference	NOUN
ajst-24984	203	31	on	on	ADP
ajst-24984	203	32	empirical	empirical	ADJ
ajst-24984	203	33	methods	method	NOUN
ajst-24984	203	34	in	in	ADP
ajst-24984	203	35	natural	natural	ADJ
ajst-24984	203	36	language	language	NOUN
ajst-24984	203	37	processing	processing	NOUN
ajst-24984	203	38	,	,	PUNCT
ajst-24984	203	39	2016	2016	NUM
ajst-24984	203	40	,	,	PUNCT
ajst-24984	203	41	pp	pp	ADV
ajst-24984	203	42	.	.	PUNCT
ajst-24984	204	1	214–224	214–224	NUM
ajst-24984	204	2	.	.	PUNCT
ajst-24984	205	1	[	[	X
ajst-24984	205	2	8	8	NUM
ajst-24984	205	3	]	]	PUNCT
ajst-24984	205	4	m.	m.	NOUN
ajst-24984	205	5	al	al	PROPN
ajst-24984	205	6	-	-	PUNCT
ajst-24984	205	7	smadi	smadi	NOUN
ajst-24984	205	8	,	,	PUNCT
ajst-24984	205	9	o.	o.	NOUN
ajst-24984	205	10	qawasmeh	qawasmeh	NOUN
ajst-24984	205	11	,	,	PUNCT
ajst-24984	205	12	m.	m.	NOUN
ajst-24984	205	13	al	al	PROPN
ajst-24984	205	14	-	-	PUNCT
ajst-24984	205	15	ayyoub	ayyoub	PROPN
ajst-24984	205	16	,	,	PUNCT
ajst-24984	205	17	y.	y.	PROPN
ajst-24984	205	18	jararweh	jararweh	PROPN
ajst-24984	205	19	,	,	PUNCT
ajst-24984	205	20	b.	b.	PROPN
ajst-24984	205	21	gupta	gupta	PROPN
ajst-24984	205	22	,	,	PUNCT
ajst-24984	205	23	deep	deep	ADJ
ajst-24984	205	24	recurrent	recurrent	ADJ
ajst-24984	205	25	neural	neural	ADJ
ajst-24984	205	26	network	network	NOUN
ajst-24984	205	27	vs.	vs.	ADP
ajst-24984	205	28	support	support	NOUN
ajst-24984	205	29	vector	vector	NOUN
ajst-24984	205	30	machine	machine	NOUN
ajst-24984	205	31	for	for	ADP
ajst-24984	205	32	aspect	aspect	NOUN
ajst-24984	205	33	-	-	PUNCT
ajst-24984	205	34	based	base	VERB
ajst-24984	205	35	sentiment	sentiment	NOUN
ajst-24984	205	36	analysis	analysis	NOUN
ajst-24984	205	37	of	of	ADP
ajst-24984	205	38	arabic	arabic	ADJ
ajst-24984	205	39	hotels	hotel	NOUN
ajst-24984	205	40	’	'	PUNCT
ajst-24984	205	41	reviews	review	NOUN
ajst-24984	205	42	,	,	PUNCT
ajst-24984	205	43	j.	j.	PROPN
ajst-24984	205	44	comput	comput	PROPN
ajst-24984	205	45	.	.	PUNCT
ajst-24984	206	1	sci	sci	PROPN
ajst-24984	206	2	.	.	PROPN
ajst-24984	206	3	27	27	NUM
ajst-24984	206	4	(	(	PUNCT
ajst-24984	206	5	2018	2018	NUM
ajst-24984	206	6	)	)	PUNCT
ajst-24984	206	7	386–393	386–393	NUM
ajst-24984	206	8	.	.	PUNCT
ajst-24984	206	9	197	197	NUM
ajst-24984	207	1	[	[	SYM
ajst-24984	207	2	9	9	NUM
ajst-24984	207	3	]	]	PUNCT
ajst-24984	207	4	t.	t.	NOUN
ajst-24984	207	5	brychcín	brychcín	PROPN
ajst-24984	207	6	,	,	PUNCT
ajst-24984	207	7	m.	m.	NOUN
ajst-24984	207	8	konkol	konkol	PROPN
ajst-24984	207	9	,	,	PUNCT
ajst-24984	207	10	j.	j.	PROPN
ajst-24984	207	11	steinberger	steinberger	PROPN
ajst-24984	207	12	,	,	PUNCT
ajst-24984	207	13	uwb	uwb	PROPN
ajst-24984	207	14	:	:	PUNCT
ajst-24984	207	15	machine	machine	NOUN
ajst-24984	207	16	learning	learn	VERB
ajst-24984	207	17	approach	approach	NOUN
ajst-24984	207	18	to	to	ADP
ajst-24984	207	19	aspect	aspect	NOUN
ajst-24984	207	20	-	-	PUNCT
ajst-24984	207	21	based	base	VERB
ajst-24984	207	22	sentiment	sentiment	NOUN
ajst-24984	207	23	analysis	analysis	NOUN
ajst-24984	207	24	,	,	PUNCT
ajst-24984	207	25	in	in	ADP
ajst-24984	207	26	:	:	PUNCT
ajst-24984	207	27	proceedings	proceeding	NOUN
ajst-24984	207	28	of	of	ADP
ajst-24984	207	29	the	the	DET
ajst-24984	207	30	8th	8th	ADJ
ajst-24984	207	31	international	international	ADJ
ajst-24984	207	32	workshop	workshop	NOUN
ajst-24984	207	33	on	on	ADP
ajst-24984	207	34	semantic	semantic	ADJ
ajst-24984	207	35	evaluation	evaluation	NOUN
ajst-24984	207	36	,	,	PUNCT
ajst-24984	207	37	semeval	semeval	NOUN
ajst-24984	207	38	2014	2014	NUM
ajst-24984	207	39	,	,	PUNCT
ajst-24984	207	40	2014	2014	NUM
ajst-24984	207	41	,	,	PUNCT
ajst-24984	207	42	pp	pp	ADV
ajst-24984	207	43	.	.	PUNCT
ajst-24984	208	1	817–822	817–822	NUM
ajst-24984	208	2	.	.	PUNCT
ajst-24984	209	1	[	[	X
ajst-24984	209	2	10	10	NUM
ajst-24984	209	3	]	]	X
ajst-24984	209	4	w.	w.	PROPN
ajst-24984	209	5	wang	wang	PROPN
ajst-24984	209	6	,	,	PUNCT
ajst-24984	209	7	n.	n.	PROPN
ajst-24984	209	8	yang	yang	PROPN
ajst-24984	209	9	,	,	PUNCT
ajst-24984	209	10	f.	f.	PROPN
ajst-24984	209	11	wei	wei	PROPN
ajst-24984	209	12	,	,	PUNCT
ajst-24984	209	13	b.	b.	PROPN
ajst-24984	209	14	chang	chang	PROPN
ajst-24984	209	15	,	,	PUNCT
ajst-24984	209	16	m.	m.	PROPN
ajst-24984	209	17	zhou	zhou	PROPN
ajst-24984	209	18	,	,	PUNCT
ajst-24984	209	19	gated	gate	VERB
ajst-24984	209	20	selfmatching	selfmatche	VERB
ajst-24984	209	21	networks	network	NOUN
ajst-24984	209	22	for	for	ADP
ajst-24984	209	23	reading	read	VERB
ajst-24984	209	24	comprehension	comprehension	NOUN
ajst-24984	209	25	and	and	CCONJ
ajst-24984	209	26	question	question	NOUN
ajst-24984	209	27	answering	answering	NOUN
ajst-24984	209	28	,	,	PUNCT
ajst-24984	209	29	in	in	ADP
ajst-24984	209	30	:	:	PUNCT
ajst-24984	209	31	pro	pro	ADJ
ajst-24984	209	32	-	-	NOUN
ajst-24984	209	33	ceedings	ceeding	NOUN
ajst-24984	209	34	of	of	ADP
ajst-24984	209	35	the	the	DET
ajst-24984	209	36	55th	55th	ADJ
ajst-24984	209	37	annual	annual	ADJ
ajst-24984	209	38	meeting	meeting	NOUN
ajst-24984	209	39	of	of	ADP
ajst-24984	209	40	the	the	DET
ajst-24984	209	41	association	association	NOUN
ajst-24984	209	42	for	for	ADP
ajst-24984	209	43	computational	computational	ADJ
ajst-24984	209	44	linguistics	linguistic	NOUN
ajst-24984	209	45	(	(	PUNCT
ajst-24984	209	46	volume	volume	NOUN
ajst-24984	209	47	1	1	NUM
ajst-24984	209	48	:	:	PUNCT
ajst-24984	209	49	long	long	ADJ
ajst-24984	209	50	papers	paper	NOUN
ajst-24984	209	51	)	)	PUNCT
ajst-24984	209	52	,	,	PUNCT
ajst-24984	209	53	2017	2017	NUM
ajst-24984	209	54	,	,	PUNCT
ajst-24984	209	55	pp	pp	ADV
ajst-24984	209	56	.	.	PUNCT
ajst-24984	210	1	189–198	189–198	NUM
ajst-24984	210	2	.	.	PUNCT
ajst-24984	211	1	[	[	X
ajst-24984	211	2	11	11	NUM
ajst-24984	211	3	]	]	PUNCT
ajst-24984	211	4	m.	m.	NOUN
ajst-24984	211	5	al	al	PROPN
ajst-24984	211	6	-	-	PUNCT
ajst-24984	211	7	smadi	smadi	NOUN
ajst-24984	211	8	,	,	PUNCT
ajst-24984	211	9	o.	o.	NOUN
ajst-24984	211	10	qawasmeh	qawasmeh	NOUN
ajst-24984	211	11	,	,	PUNCT
ajst-24984	211	12	m.	m.	NOUN
ajst-24984	211	13	al	al	PROPN
ajst-24984	211	14	-	-	PUNCT
ajst-24984	211	15	ayyoub	ayyoub	PROPN
ajst-24984	211	16	,	,	PUNCT
ajst-24984	211	17	y.	y.	PROPN
ajst-24984	211	18	jararweh	jararweh	PROPN
ajst-24984	211	19	,	,	PUNCT
ajst-24984	211	20	b.	b.	PROPN
ajst-24984	211	21	gupta	gupta	PROPN
ajst-24984	211	22	,	,	PUNCT
ajst-24984	211	23	deep	deep	ADJ
ajst-24984	211	24	recurrent	recurrent	ADJ
ajst-24984	211	25	neural	neural	ADJ
ajst-24984	211	26	network	network	NOUN
ajst-24984	211	27	vs.	vs.	ADP
ajst-24984	211	28	support	support	NOUN
ajst-24984	211	29	vector	vector	NOUN
ajst-24984	211	30	machine	machine	NOUN
ajst-24984	211	31	for	for	ADP
ajst-24984	211	32	aspect	aspect	NOUN
ajst-24984	211	33	-	-	PUNCT
ajst-24984	211	34	based	base	VERB
ajst-24984	211	35	sentiment	sentiment	NOUN
ajst-24984	211	36	analysis	analysis	NOUN
ajst-24984	211	37	of	of	ADP
ajst-24984	211	38	arabic	arabic	ADJ
ajst-24984	211	39	hotels	hotel	NOUN
ajst-24984	211	40	’	'	PUNCT
ajst-24984	211	41	reviews	review	NOUN
ajst-24984	211	42	,	,	PUNCT
ajst-24984	211	43	j.	j.	PROPN
ajst-24984	211	44	comput	comput	PROPN
ajst-24984	211	45	.	.	PUNCT
ajst-24984	212	1	sci	sci	PROPN
ajst-24984	212	2	.	.	PROPN
ajst-24984	212	3	27	27	NUM
ajst-24984	212	4	(	(	PUNCT
ajst-24984	212	5	2018	2018	NUM
ajst-24984	212	6	)	)	PUNCT
ajst-24984	212	7	386–393	386–393	NUM
ajst-24984	212	8	.	.	PUNCT
ajst-24984	213	1	[	[	X
ajst-24984	213	2	12	12	NUM
ajst-24984	213	3	]	]	X
ajst-24984	213	4	yadav	yadav	PROPN
ajst-24984	213	5	rk	rk	PROPN
ajst-24984	213	6	,	,	PUNCT
ajst-24984	213	7	jiao	jiao	PROPN
ajst-24984	213	8	l	l	PROPN
ajst-24984	213	9	,	,	PUNCT
ajst-24984	213	10	goodwin	goodwin	PROPN
ajst-24984	213	11	m	m	PROPN
ajst-24984	213	12	,	,	PUNCT
ajst-24984	213	13	granmo	granmo	PROPN
ajst-24984	213	14	o	o	PROPN
ajst-24984	213	15	-	-	PROPN
ajst-24984	213	16	c	c	PROPN
ajst-24984	213	17	(	(	PUNCT
ajst-24984	213	18	2021	2021	NUM
ajst-24984	213	19	)	)	PUNCT
ajst-24984	213	20	positionless	positionless	ADJ
ajst-24984	213	21	aspect	aspect	NOUN
ajst-24984	213	22	based	base	VERB
ajst-24984	213	23	sentiment	sentiment	NOUN
ajst-24984	213	24	analysis	analysis	NOUN
ajst-24984	213	25	using	use	VERB
ajst-24984	213	26	attention	attention	NOUN
ajst-24984	213	27	mechanism	mechanism	NOUN
ajst-24984	213	28	.	.	PUNCT
ajst-24984	214	1	know	know	NOUN
ajst-24984	214	2	-	-	PUNCT
ajst-24984	214	3	based	base	VERB
ajst-24984	214	4	syst	syst	NOUN
ajst-24984	214	5	226:107136	226:107136	NOUN
ajst-24984	214	6	.	.	PUNCT
ajst-24984	215	1	[	[	X
ajst-24984	215	2	13	13	NUM
ajst-24984	215	3	]	]	PUNCT
ajst-24984	215	4	peters	peters	PROPN
ajst-24984	215	5	me	i	PRON
ajst-24984	215	6	,	,	PUNCT
ajst-24984	215	7	neumann	neumann	PROPN
ajst-24984	215	8	m	m	PROPN
ajst-24984	215	9	,	,	PUNCT
ajst-24984	215	10	iyyer	iyyer	PROPN
ajst-24984	215	11	m	m	PROPN
ajst-24984	215	12	,	,	PUNCT
ajst-24984	215	13	gardner	gardner	PROPN
ajst-24984	215	14	m	m	PROPN
ajst-24984	215	15	,	,	PUNCT
ajst-24984	215	16	clark	clark	PROPN
ajst-24984	215	17	c	c	PROPN
ajst-24984	215	18	,	,	PUNCT
ajst-24984	215	19	lee	lee	PROPN
ajst-24984	215	20	k	k	PROPN
ajst-24984	215	21	,	,	PUNCT
ajst-24984	215	22	zettlemoyer	zettlemoyer	NOUN
ajst-24984	215	23	l	l	PROPN
ajst-24984	215	24	(	(	PUNCT
ajst-24984	215	25	2018	2018	NUM
ajst-24984	215	26	)	)	PUNCT
ajst-24984	215	27	deep	deep	ADJ
ajst-24984	215	28	con	con	NOUN
ajst-24984	215	29	-	-	PUNCT
ajst-24984	215	30	textualized	textualize	VERB
ajst-24984	215	31	word	word	NOUN
ajst-24984	215	32	representations	representation	NOUN
ajst-24984	215	33	.	.	PUNCT
ajst-24984	216	1	corr	corr	NOUN
ajst-24984	216	2	,	,	PUNCT
ajst-24984	216	3	abs/1802.05365	abs/1802.05365	VERB
ajst-24984	216	4	.	.	PUNCT
ajst-24984	217	1	[	[	X
ajst-24984	217	2	14	14	NUM
ajst-24984	217	3	]	]	X
ajst-24984	217	4	alec	alec	PROPN
ajst-24984	217	5	r	r	PROPN
ajst-24984	217	6	,	,	PUNCT
ajst-24984	217	7	karthik	karthik	PROPN
ajst-24984	217	8	n	n	CCONJ
ajst-24984	217	9	,	,	PUNCT
ajst-24984	217	10	tim	tim	PROPN
ajst-24984	217	11	s	s	PROPN
ajst-24984	217	12	,	,	PUNCT
ajst-24984	217	13	ilya	ilya	PROPN
ajst-24984	217	14	s	s	PROPN
ajst-24984	217	15	(	(	PUNCT
ajst-24984	217	16	2018	2018	NUM
ajst-24984	217	17	)	)	PUNCT
ajst-24984	217	18	improving	improve	VERB
ajst-24984	217	19	language	language	NOUN
ajst-24984	217	20	understanding	understanding	NOUN
ajst-24984	217	21	by	by	ADP
ajst-24984	217	22	generative	generative	ADJ
ajst-24984	217	23	pre	pre	NOUN
ajst-24984	217	24	-	-	NOUN
ajst-24984	217	25	training	training	NOUN
ajst-24984	217	26	.	.	PUNCT
ajst-24984	218	1	the	the	DET
ajst-24984	218	2	university	university	PROPN
ajst-24984	218	3	of	of	ADP
ajst-24984	218	4	british	british	PROPN
ajst-24984	218	5	columbia	columbia	PROPN
ajst-24984	218	6	vancouver	vancouver	PROPN
ajst-24984	218	7	campus	campus	PROPN
ajst-24984	218	8	,	,	PUNCT
ajst-24984	218	9	vancouver	vancouver	PROPN
ajst-24984	218	10	.	.	PUNCT
ajst-24984	219	1	[	[	X
ajst-24984	219	2	15	15	NUM
ajst-24984	219	3	]	]	X
ajst-24984	219	4	devlin	devlin	PROPN
ajst-24984	219	5	j	j	PROPN
ajst-24984	219	6	,	,	PUNCT
ajst-24984	219	7	chang	chang	PROPN
ajst-24984	219	8	m	m	PROPN
ajst-24984	219	9	w	w	PROPN
ajst-24984	219	10	,	,	PUNCT
ajst-24984	219	11	lee	lee	PROPN
ajst-24984	219	12	k	k	PROPN
ajst-24984	219	13	,	,	PUNCT
ajst-24984	219	14	et	et	PROPN
ajst-24984	219	15	al	al	PROPN
ajst-24984	219	16	.	.	PUNCT
ajst-24984	219	17	bert	bert	PROPN
ajst-24984	219	18	:	:	PUNCT
ajst-24984	219	19	pre	pre	ADJ
ajst-24984	219	20	-	-	NOUN
ajst-24984	219	21	training	training	NOUN
ajst-24984	219	22	of	of	ADP
ajst-24984	219	23	deep	deep	ADJ
ajst-24984	219	24	bidirectional	bidirectional	ADJ
ajst-24984	219	25	transformers	transformer	NOUN
ajst-24984	219	26	for	for	ADP
ajst-24984	219	27	language	language	NOUN
ajst-24984	219	28	understanding[j	understanding[j	NOUN
ajst-24984	219	29	]	]	PUNCT
ajst-24984	219	30	.	.	PUNCT
ajst-24984	220	1	arxiv	arxiv	PROPN
ajst-24984	220	2	preprint	preprint	PROPN
ajst-24984	220	3	arxiv:1810.04805	arxiv:1810.04805	PROPN
ajst-24984	220	4	,	,	PUNCT
ajst-24984	220	5	2018	2018	NUM
ajst-24984	220	6	.	.	PUNCT
ajst-24984	221	1	[	[	X
ajst-24984	221	2	16	16	NUM
ajst-24984	221	3	]	]	X
ajst-24984	221	4	karimi	karimi	PROPN
ajst-24984	221	5	a	a	PROPN
ajst-24984	221	6	,	,	PUNCT
ajst-24984	221	7	rossi	rossi	ADJ
ajst-24984	221	8	l	l	PROPN
ajst-24984	221	9	,	,	PUNCT
ajst-24984	221	10	prati	prati	PROPN
ajst-24984	221	11	a.	a.	PROPN
ajst-24984	221	12	adversarial	adversarial	PROPN
ajst-24984	221	13	training	training	NOUN
ajst-24984	221	14	for	for	ADP
ajst-24984	221	15	aspectbased	aspectbase	VERB
ajst-24984	221	16	sentiment	sentiment	NOUN
ajst-24984	221	17	analysis	analysis	NOUN
ajst-24984	221	18	with	with	ADP
ajst-24984	221	19	bert[c]//2020	bert[c]//2020	ADP
ajst-24984	221	20	25th	25th	ADJ
ajst-24984	221	21	international	international	ADJ
ajst-24984	221	22	conference	conference	NOUN
ajst-24984	221	23	on	on	ADP
ajst-24984	221	24	pattern	pattern	NOUN
ajst-24984	221	25	recognition	recognition	NOUN
ajst-24984	221	26	(	(	PUNCT
ajst-24984	221	27	icpr	icpr	PROPN
ajst-24984	221	28	)	)	PUNCT
ajst-24984	221	29	.	.	PUNCT
ajst-24984	222	1	ieee	ieee	NOUN
ajst-24984	222	2	,	,	PUNCT
ajst-24984	222	3	2021	2021	NUM
ajst-24984	222	4	:	:	PUNCT
ajst-24984	222	5	87978803	87978803	NUM
ajst-24984	222	6	.	.	PUNCT
ajst-24984	223	1	[	[	X
ajst-24984	223	2	17	17	NUM
ajst-24984	223	3	]	]	PUNCT
ajst-24984	223	4	r.	r.	PROPN
ajst-24984	223	5	he	he	PRON
ajst-24984	223	6	,	,	PUNCT
ajst-24984	223	7	w.	w.	PROPN
ajst-24984	223	8	s.	s.	PROPN
ajst-24984	223	9	lee	lee	PROPN
ajst-24984	223	10	,	,	PUNCT
ajst-24984	223	11	h.	h.	PROPN
ajst-24984	223	12	t.	t.	PROPN
ajst-24984	223	13	ng	ng	PROPN
ajst-24984	223	14	,	,	PUNCT
ajst-24984	223	15	and	and	CCONJ
ajst-24984	223	16	d.	d.	PROPN
ajst-24984	223	17	dahlmeier	dahlmeier	PROPN
ajst-24984	223	18	,	,	PUNCT
ajst-24984	223	19	“	"	PUNCT
ajst-24984	223	20	exploiting	exploit	VERB
ajst-24984	223	21	document	document	NOUN
ajst-24984	223	22	knowledge	knowledge	NOUN
ajst-24984	223	23	for	for	ADP
ajst-24984	223	24	aspect	aspect	NOUN
ajst-24984	223	25	-	-	PUNCT
ajst-24984	223	26	level	level	NOUN
ajst-24984	223	27	sentiment	sentiment	NOUN
ajst-24984	223	28	classification	classification	NOUN
ajst-24984	223	29	,	,	PUNCT
ajst-24984	223	30	”	"	PUNCT
ajst-24984	223	31	in	in	ADP
ajst-24984	223	32	proceedings	proceeding	NOUN
ajst-24984	223	33	of	of	ADP
ajst-24984	223	34	the	the	DET
ajst-24984	223	35	56th	56th	ADJ
ajst-24984	223	36	annual	annual	ADJ
ajst-24984	223	37	meeting	meeting	NOUN
ajst-24984	223	38	of	of	ADP
ajst-24984	223	39	the	the	DET
ajst-24984	223	40	association	association	NOUN
ajst-24984	223	41	for	for	ADP
ajst-24984	223	42	computational	computational	ADJ
ajst-24984	223	43	linguistics	linguistic	NOUN
ajst-24984	223	44	(	(	PUNCT
ajst-24984	223	45	volume	volume	NOUN
ajst-24984	223	46	2	2	NUM
ajst-24984	223	47	:	:	PUNCT
ajst-24984	223	48	short	short	ADJ
ajst-24984	223	49	papers	paper	NOUN
ajst-24984	223	50	)	)	PUNCT
ajst-24984	223	51	,	,	PUNCT
ajst-24984	223	52	2018	2018	NUM
ajst-24984	223	53	,	,	PUNCT
ajst-24984	223	54	pp	pp	ADJ
ajst-24984	223	55	.	.	PUNCT
ajst-24984	224	1	579–585	579–585	NUM
ajst-24984	224	2	.	.	PUNCT
ajst-24984	225	1	[	[	X
ajst-24984	225	2	18	18	NUM
ajst-24984	225	3	]	]	PUNCT
ajst-24984	225	4	z.	z.	PROPN
ajst-24984	225	5	chen	chen	PROPN
ajst-24984	225	6	and	and	CCONJ
ajst-24984	225	7	t.	t.	PROPN
ajst-24984	225	8	qian	qian	PROPN
ajst-24984	225	9	,	,	PUNCT
ajst-24984	225	10	“	"	PUNCT
ajst-24984	225	11	transfer	transfer	NOUN
ajst-24984	225	12	capsule	capsule	NOUN
ajst-24984	225	13	network	network	NOUN
ajst-24984	225	14	for	for	ADP
ajst-24984	225	15	aspect	aspect	NOUN
ajst-24984	225	16	level	level	NOUN
ajst-24984	225	17	sentiment	sentiment	NOUN
ajst-24984	225	18	classification	classification	NOUN
ajst-24984	225	19	,	,	PUNCT
ajst-24984	225	20	”	"	PUNCT
ajst-24984	225	21	in	in	ADP
ajst-24984	225	22	proceedings	proceeding	NOUN
ajst-24984	225	23	of	of	ADP
ajst-24984	225	24	the	the	DET
ajst-24984	225	25	57th	57th	ADJ
ajst-24984	225	26	annual	annual	ADJ
ajst-24984	225	27	meeting	meeting	NOUN
ajst-24984	225	28	of	of	ADP
ajst-24984	225	29	the	the	DET
ajst-24984	225	30	association	association	NOUN
ajst-24984	225	31	for	for	ADP
ajst-24984	225	32	computational	computational	ADJ
ajst-24984	225	33	linguistics	linguistic	NOUN
ajst-24984	225	34	,	,	PUNCT
ajst-24984	225	35	2019	2019	NUM
ajst-24984	225	36	,	,	PUNCT
ajst-24984	225	37	pp	pp	ADJ
ajst-24984	225	38	.	.	PUNCT
ajst-24984	226	1	547–556	547–556	NUM
ajst-24984	226	2	.	.	PUNCT
ajst-24984	227	1	[	[	X
ajst-24984	227	2	19	19	NUM
ajst-24984	227	3	]	]	X
ajst-24984	227	4	jia	jia	PROPN
ajst-24984	227	5	k	k	PROPN
ajst-24984	227	6	,	,	PUNCT
ajst-24984	227	7	li	li	PROPN
ajst-24984	227	8	z.	z.	PROPN
ajst-24984	227	9	chinese	chinese	PROPN
ajst-24984	227	10	micro	micro	PROPN
ajst-24984	227	11	-	-	ADJ
ajst-24984	227	12	blog	blog	ADJ
ajst-24984	227	13	sentiment	sentiment	NOUN
ajst-24984	227	14	classification	classification	NOUN
ajst-24984	227	15	based	base	VERB
ajst-24984	227	16	on	on	ADP
ajst-24984	227	17	emotion	emotion	NOUN
ajst-24984	227	18	dictionary	dictionary	NOUN
ajst-24984	227	19	and	and	CCONJ
ajst-24984	227	20	semantic	semantic	ADJ
ajst-24984	227	21	rules[c]//2020	rules[c]//2020	VERB
ajst-24984	227	22	international	international	ADJ
ajst-24984	227	23	conference	conference	NOUN
ajst-24984	227	24	on	on	ADP
ajst-24984	227	25	computer	computer	NOUN
ajst-24984	227	26	information	information	NOUN
ajst-24984	227	27	and	and	CCONJ
ajst-24984	227	28	big	big	ADJ
ajst-24984	227	29	data	datum	NOUN
ajst-24984	227	30	applications	application	NOUN
ajst-24984	227	31	(	(	PUNCT
ajst-24984	227	32	cibda	cibda	PROPN
ajst-24984	227	33	)	)	PUNCT
ajst-24984	227	34	.	.	PUNCT
ajst-24984	228	1	ieee	ieee	PROPN
ajst-24984	228	2	,	,	PUNCT
ajst-24984	228	3	2020	2020	NUM
ajst-24984	228	4	:	:	PUNCT
ajst-24984	228	5	309	309	NUM
ajst-24984	228	6	-	-	SYM
ajst-24984	228	7	312	312	NUM
ajst-24984	228	8	.	.	PUNCT
ajst-24984	229	1	[	[	X
ajst-24984	229	2	20	20	NUM
ajst-24984	229	3	]	]	X
ajst-24984	229	4	alharthi	alharthi	PROPN
ajst-24984	230	1	k	k	PROPN
ajst-24984	230	2	a	a	X
ajst-24984	230	3	,	,	PUNCT
ajst-24984	230	4	jhumka	jhumka	NOUN
ajst-24984	230	5	a	a	NOUN
ajst-24984	230	6	,	,	PUNCT
ajst-24984	230	7	di	di	X
ajst-24984	230	8	s	s	PROPN
ajst-24984	230	9	,	,	PUNCT
ajst-24984	230	10	et	et	PROPN
ajst-24984	230	11	al	al	PROPN
ajst-24984	230	12	.	.	PROPN
ajst-24984	230	13	sentiment	sentiment	PROPN
ajst-24984	230	14	analysis	analysis	NOUN
ajst-24984	230	15	based	base	VERB
ajst-24984	230	16	error	error	NOUN
ajst-24984	230	17	detection	detection	NOUN
ajst-24984	230	18	for	for	ADP
ajst-24984	230	19	large	large	ADJ
ajst-24984	230	20	-	-	PUNCT
ajst-24984	230	21	scale	scale	NOUN
ajst-24984	230	22	systems[c]//2021	systems[c]//2021	NUM
ajst-24984	230	23	51st	51st	ADJ
ajst-24984	230	24	annual	annual	ADJ
ajst-24984	230	25	ieee	ieee	NOUN
ajst-24984	230	26	/	/	SYM
ajst-24984	230	27	ifip	ifip	PROPN
ajst-24984	230	28	international	international	ADJ
ajst-24984	230	29	conference	conference	NOUN
ajst-24984	230	30	on	on	ADP
ajst-24984	230	31	dependable	dependable	ADJ
ajst-24984	230	32	systems	system	NOUN
ajst-24984	230	33	and	and	CCONJ
ajst-24984	230	34	networks	network	NOUN
ajst-24984	230	35	(	(	PUNCT
ajst-24984	230	36	dsn	dsn	PROPN
ajst-24984	230	37	)	)	PUNCT
ajst-24984	230	38	.	.	PUNCT
ajst-24984	231	1	ieee	ieee	NOUN
ajst-24984	231	2	,	,	PUNCT
ajst-24984	231	3	2021	2021	NUM
ajst-24984	231	4	:	:	PUNCT
ajst-24984	231	5	237	237	NUM
ajst-24984	231	6	-	-	SYM
ajst-24984	231	7	249	249	NUM
ajst-24984	231	8	.	.	PUNCT
ajst-24984	232	1	[	[	X
ajst-24984	232	2	21	21	NUM
ajst-24984	232	3	]	]	X
ajst-24984	232	4	rao	rao	PROPN
ajst-24984	232	5	l.	l.	PROPN
ajst-24984	232	6	sentiment	sentiment	PROPN
ajst-24984	232	7	analysis	analysis	NOUN
ajst-24984	232	8	of	of	ADP
ajst-24984	232	9	english	english	ADJ
ajst-24984	232	10	text	text	NOUN
ajst-24984	232	11	with	with	ADP
ajst-24984	232	12	multilevel	multilevel	ADJ
ajst-24984	232	13	features[j	features[j	NOUN
ajst-24984	232	14	]	]	PUNCT
ajst-24984	232	15	.	.	PUNCT
ajst-24984	233	1	scientific	scientific	ADJ
ajst-24984	233	2	programming	programming	NOUN
ajst-24984	233	3	,	,	PUNCT
ajst-24984	233	4	2022	2022	NUM
ajst-24984	233	5	,	,	PUNCT
ajst-24984	233	6	2022	2022	NUM
ajst-24984	233	7	:	:	PUNCT
ajst-24984	233	8	1	1	NUM
ajst-24984	233	9	-	-	SYM
ajst-24984	233	10	10	10	NUM
ajst-24984	233	11	.	.	PUNCT
ajst-24984	234	1	[	[	X
ajst-24984	234	2	22	22	NUM
ajst-24984	234	3	]	]	X
ajst-24984	234	4	yan	yan	PROPN
ajst-24984	234	5	r	r	PROPN
ajst-24984	234	6	,	,	PUNCT
ajst-24984	234	7	yu	yu	PROPN
ajst-24984	234	8	y	y	PROPN
ajst-24984	234	9	,	,	PUNCT
ajst-24984	234	10	qiu	qiu	PROPN
ajst-24984	234	11	d.	d.	PROPN
ajst-24984	234	12	emotion	emotion	NOUN
ajst-24984	234	13	-	-	PUNCT
ajst-24984	234	14	enhanced	enhance	VERB
ajst-24984	234	15	classification	classification	NOUN
ajst-24984	234	16	based	base	VERB
ajst-24984	234	17	on	on	ADP
ajst-24984	234	18	fuzzy	fuzzy	ADJ
ajst-24984	234	19	reasoning[j	reasoning[j	PROPN
ajst-24984	234	20	]	]	PUNCT
ajst-24984	234	21	.	.	PUNCT
ajst-24984	235	1	international	international	ADJ
ajst-24984	235	2	journal	journal	PROPN
ajst-24984	235	3	of	of	ADP
ajst-24984	235	4	machine	machine	NOUN
ajst-24984	235	5	learning	learning	NOUN
ajst-24984	235	6	and	and	CCONJ
ajst-24984	235	7	cybernetics	cybernetic	NOUN
ajst-24984	235	8	,	,	PUNCT
ajst-24984	235	9	2021,13(3):1	2021,13(3):1	NOUN
ajst-24984	235	10	-	-	SYM
ajst-24984	235	11	12	12	NUM
ajst-24984	235	12	.	.	PUNCT
ajst-24984	236	1	[	[	X
ajst-24984	236	2	23	23	NUM
ajst-24984	236	3	]	]	PUNCT
ajst-24984	236	4	pang	pang	NOUN
ajst-24984	236	5	b	b	PROPN
ajst-24984	236	6	,	,	PUNCT
ajst-24984	236	7	lee	lee	PROPN
ajst-24984	236	8	l	l	PROPN
ajst-24984	236	9	,	,	PUNCT
ajst-24984	236	10	vaithyanathan	vaithyanathan	PROPN
ajst-24984	236	11	s.	s.	PROPN
ajst-24984	236	12	thumbs	thumb	VERB
ajst-24984	236	13	up	up	ADP
ajst-24984	236	14	?	?	PUNCT
ajst-24984	237	1	sentiment	sentiment	NOUN
ajst-24984	237	2	classification	classification	NOUN
ajst-24984	237	3	using	use	VERB
ajst-24984	237	4	machine	machine	NOUN
ajst-24984	237	5	learning	learn	VERB
ajst-24984	237	6	techniques[j	techniques[j	NOUN
ajst-24984	237	7	]	]	PUNCT
ajst-24984	237	8	.	.	PUNCT
ajst-24984	238	1	arxiv	arxiv	PROPN
ajst-24984	238	2	preprint	preprint	NOUN
ajst-24984	238	3	cs/0205070	cs/0205070	NOUN
ajst-24984	238	4	,	,	PUNCT
ajst-24984	238	5	2002	2002	NUM
ajst-24984	238	6	.	.	PUNCT
ajst-24984	239	1	[	[	X
ajst-24984	239	2	24	24	NUM
ajst-24984	239	3	]	]	X
ajst-24984	239	4	han	han	PROPN
ajst-24984	239	5	k	k	PROPN
ajst-24984	239	6	x	x	PROPN
ajst-24984	239	7	,	,	PUNCT
ajst-24984	239	8	chien	chien	PROPN
ajst-24984	239	9	w	w	PROPN
ajst-24984	239	10	,	,	PUNCT
ajst-24984	239	11	chiu	chiu	PROPN
ajst-24984	239	12	c	c	PROPN
ajst-24984	239	13	c	c	NOUN
ajst-24984	239	14	,	,	PUNCT
ajst-24984	239	15	et	et	PROPN
ajst-24984	239	16	al	al	PROPN
ajst-24984	239	17	.	.	PUNCT
ajst-24984	239	18	application	application	NOUN
ajst-24984	239	19	of	of	ADP
ajst-24984	239	20	support	support	NOUN
ajst-24984	239	21	vector	vector	NOUN
ajst-24984	239	22	machine	machine	NOUN
ajst-24984	239	23	(	(	PUNCT
ajst-24984	239	24	svm	svm	PROPN
ajst-24984	239	25	)	)	PUNCT
ajst-24984	239	26	in	in	ADP
ajst-24984	239	27	the	the	DET
ajst-24984	239	28	sentiment	sentiment	NOUN
ajst-24984	239	29	analysis	analysis	NOUN
ajst-24984	239	30	of	of	ADP
ajst-24984	239	31	twitter	twitter	NOUN
ajst-24984	239	32	dataset[j	dataset[j	PROPN
ajst-24984	239	33	]	]	PUNCT
ajst-24984	239	34	.	.	PUNCT
ajst-24984	240	1	applied	apply	VERB
ajst-24984	240	2	sciences	science	NOUN
ajst-24984	240	3	,	,	PUNCT
ajst-24984	240	4	2020,10(3	2020,10(3	NOUN
ajst-24984	240	5	)	)	PUNCT
ajst-24984	240	6	.	.	PUNCT
ajst-24984	241	1	[	[	X
ajst-24984	241	2	25	25	NUM
ajst-24984	241	3	]	]	X
ajst-24984	241	4	wang	wang	PROPN
ajst-24984	241	5	y	y	PROPN
ajst-24984	241	6	,	,	PUNCT
ajst-24984	241	7	chen	chen	PROPN
ajst-24984	241	8	q	q	PROPN
ajst-24984	241	9	,	,	PUNCT
ajst-24984	241	10	shen	shen	PROPN
ajst-24984	241	11	j	j	PROPN
ajst-24984	241	12	,	,	PUNCT
ajst-24984	241	13	et	et	PROPN
ajst-24984	241	14	al	al	PROPN
ajst-24984	241	15	.	.	PROPN
ajst-24984	241	16	aspect	aspect	NOUN
ajst-24984	241	17	-	-	PUNCT
ajst-24984	241	18	level	level	NOUN
ajst-24984	241	19	sentiment	sentiment	NOUN
ajst-24984	241	20	analysis	analysis	NOUN
ajst-24984	241	21	based	base	VERB
ajst-24984	241	22	on	on	ADP
ajst-24984	241	23	gradual	gradual	ADJ
ajst-24984	241	24	machine	machine	NOUN
ajst-24984	241	25	learning[j	learning[j	NOUN
ajst-24984	241	26	]	]	PUNCT
ajst-24984	241	27	.	.	PUNCT
ajst-24984	242	1	knowledge	knowledge	NOUN
ajst-24984	242	2	-	-	PUNCT
ajst-24984	242	3	based	base	VERB
ajst-24984	242	4	systems	system	NOUN
ajst-24984	242	5	,	,	PUNCT
ajst-24984	242	6	2021	2021	NUM
ajst-24984	242	7	,	,	PUNCT
ajst-24984	242	8	212	212	NUM
ajst-24984	242	9	:	:	PUNCT
ajst-24984	242	10	106509	106509	NUM
ajst-24984	242	11	.	.	PUNCT
ajst-24984	243	1	[	[	X
ajst-24984	243	2	26	26	NUM
ajst-24984	243	3	]	]	X
ajst-24984	243	4	kurniawan	kurniawan	X
ajst-24984	243	5	c	c	PROPN
ajst-24984	243	6	,	,	PUNCT
ajst-24984	243	7	wahyuni	wahyuni	PROPN
ajst-24984	243	8	f.	f.	PROPN
ajst-24984	243	9	sentiment	sentiment	PROPN
ajst-24984	243	10	analysis	analysis	NOUN
ajst-24984	243	11	of	of	ADP
ajst-24984	243	12	online	online	ADJ
ajst-24984	243	13	learning	learn	VERB
ajst-24984	243	14	students	student	NOUN
ajst-24984	243	15	feedback	feedback	VERB
ajst-24984	243	16	for	for	ADP
ajst-24984	243	17	facing	face	VERB
ajst-24984	243	18	new	new	ADJ
ajst-24984	243	19	semester	semester	NOUN
ajst-24984	243	20	:	:	PUNCT
ajst-24984	243	21	a	a	DET
ajst-24984	243	22	support	support	NOUN
ajst-24984	243	23	vector	vector	NOUN
ajst-24984	243	24	machine	machine	NOUN
ajst-24984	243	25	approach[c]//2021	approach[c]//2021	PROPN
ajst-24984	243	26	7th	7th	ADJ
ajst-24984	243	27	international	international	ADJ
ajst-24984	243	28	conference	conference	NOUN
ajst-24984	243	29	on	on	ADP
ajst-24984	243	30	education	education	NOUN
ajst-24984	243	31	and	and	CCONJ
ajst-24984	243	32	technology	technology	NOUN
ajst-24984	243	33	(	(	PUNCT
ajst-24984	243	34	icet	icet	PROPN
ajst-24984	243	35	)	)	PUNCT
ajst-24984	243	36	.	.	PUNCT
ajst-24984	244	1	ieee	ieee	NOUN
ajst-24984	244	2	,	,	PUNCT
ajst-24984	244	3	2021	2021	NUM
ajst-24984	244	4	:	:	PUNCT
ajst-24984	244	5	1	1	NUM
ajst-24984	244	6	-	-	SYM
ajst-24984	244	7	6	6	NUM
ajst-24984	244	8	.	.	PUNCT
ajst-24984	245	1	[	[	X
ajst-24984	245	2	27	27	NUM
ajst-24984	245	3	]	]	X
ajst-24984	245	4	li	li	PROPN
ajst-24984	245	5	z	z	PROPN
ajst-24984	245	6	,	,	PUNCT
ajst-24984	245	7	li	li	PROPN
ajst-24984	245	8	r	r	PROPN
ajst-24984	245	9	,	,	PUNCT
ajst-24984	245	10	jin	jin	PROPN
ajst-24984	245	11	g.	g.	PROPN
ajst-24984	245	12	sentiment	sentiment	PROPN
ajst-24984	245	13	analysis	analysis	NOUN
ajst-24984	245	14	of	of	ADP
ajst-24984	245	15	danmaku	danmaku	NOUN
ajst-24984	245	16	videos	video	NOUN
ajst-24984	245	17	based	base	VERB
ajst-24984	245	18	on	on	ADP
ajst-24984	245	19	naïve	naïve	ADJ
ajst-24984	245	20	bayes	baye	NOUN
ajst-24984	245	21	and	and	CCONJ
ajst-24984	245	22	sentiment	sentiment	NOUN
ajst-24984	245	23	dictionary[j	dictionary[j	NOUN
ajst-24984	245	24	]	]	X
ajst-24984	245	25	.	.	PUNCT
ajst-24984	246	1	ieee	ieee	NOUN
ajst-24984	246	2	access	access	NOUN
ajst-24984	246	3	,	,	PUNCT
ajst-24984	246	4	2020	2020	NUM
ajst-24984	246	5	,	,	PUNCT
ajst-24984	246	6	8	8	NUM
ajst-24984	246	7	:	:	SYM
ajst-24984	246	8	75073	75073	NUM
ajst-24984	246	9	-	-	SYM
ajst-24984	246	10	75084	75084	NUM
ajst-24984	246	11	.	.	PUNCT
ajst-24984	247	1	[	[	X
ajst-24984	247	2	28	28	NUM
ajst-24984	247	3	]	]	X
ajst-24984	247	4	li	li	PROPN
ajst-24984	247	5	h	h	PROPN
ajst-24984	247	6	,	,	PUNCT
ajst-24984	247	7	chen	chen	PROPN
ajst-24984	247	8	q	q	PROPN
ajst-24984	247	9	,	,	PUNCT
ajst-24984	247	10	zhong	zhong	PROPN
ajst-24984	247	11	z	z	X
ajst-24984	247	12	,	,	PUNCT
ajst-24984	247	13	et	et	PROPN
ajst-24984	247	14	al	al	PROPN
ajst-24984	247	15	.	.	PUNCT
ajst-24984	248	1	e	e	X
ajst-24984	248	2	-	-	NOUN
ajst-24984	248	3	word	word	NOUN
ajst-24984	248	4	of	of	ADP
ajst-24984	248	5	mouth	mouth	NOUN
ajst-24984	248	6	sentiment	sentiment	NOUN
ajst-24984	248	7	analysis	analysis	NOUN
ajst-24984	248	8	for	for	ADP
ajst-24984	248	9	user	user	NOUN
ajst-24984	248	10	behavior	behavior	PROPN
ajst-24984	248	11	studies[j	studies[j	PROPN
ajst-24984	248	12	]	]	PUNCT
ajst-24984	248	13	.	.	PUNCT
ajst-24984	249	1	information	information	NOUN
ajst-24984	249	2	processing	processing	NOUN
ajst-24984	249	3	&	&	CCONJ
ajst-24984	249	4	management	management	NOUN
ajst-24984	249	5	,	,	PUNCT
ajst-24984	249	6	2022,59(1	2022,59(1	NUM
ajst-24984	249	7	)	)	PUNCT
ajst-24984	249	8	.	.	PUNCT
ajst-24984	250	1	[	[	X
ajst-24984	250	2	29	29	NUM
ajst-24984	250	3	]	]	X
ajst-24984	250	4	zhou	zhou	PROPN
ajst-24984	250	5	j	j	PROPN
ajst-24984	250	6	,	,	PUNCT
ajst-24984	250	7	huang	huang	PROPN
ajst-24984	250	8	jx	jx	PROPN
ajst-24984	250	9	,	,	PUNCT
ajst-24984	250	10	chen	chen	PROPN
ajst-24984	250	11	q	q	PROPN
ajst-24984	250	12	,	,	PUNCT
ajst-24984	250	13	qinmin	qinmin	NOUN
ajst-24984	250	14	vivian	vivian	PROPN
ajst-24984	250	15	h	h	PROPN
ajst-24984	250	16	,	,	PUNCT
ajst-24984	250	17	wang	wang	PROPN
ajst-24984	250	18	t	t	PROPN
ajst-24984	250	19	,	,	PUNCT
ajst-24984	250	20	he	he	PRON
ajst-24984	250	21	l	l	NOUN
ajst-24984	250	22	(	(	PUNCT
ajst-24984	250	23	2019	2019	NUM
ajst-24984	250	24	)	)	PUNCT
ajst-24984	250	25	deep	deep	ADJ
ajst-24984	250	26	learning	learning	NOUN
ajst-24984	250	27	for	for	ADP
ajst-24984	250	28	aspect	aspect	NOUN
ajst-24984	250	29	-	-	PUNCT
ajst-24984	250	30	level	level	NOUN
ajst-24984	250	31	sentiment	sentiment	NOUN
ajst-24984	250	32	classification	classification	NOUN
ajst-24984	250	33	:	:	PUNCT
ajst-24984	250	34	survey	survey	NOUN
ajst-24984	250	35	,	,	PUNCT
ajst-24984	250	36	vision	vision	NOUN
ajst-24984	250	37	,	,	PUNCT
ajst-24984	250	38	and	and	CCONJ
ajst-24984	250	39	challenges	challenge	NOUN
ajst-24984	250	40	.	.	PUNCT
ajst-24984	251	1	ieee	ieee	NOUN
ajst-24984	251	2	access	access	NOUN
ajst-24984	251	3	7:78454–78483	7:78454–78483	PROPN
ajst-24984	251	4	.	.	PUNCT
ajst-24984	252	1	[	[	X
ajst-24984	252	2	30	30	NUM
ajst-24984	252	3	]	]	X
ajst-24984	252	4	s.	s.	PROPN
ajst-24984	252	5	hochreiter	hochreiter	PROPN
ajst-24984	252	6	and	and	CCONJ
ajst-24984	252	7	j.	j.	PROPN
ajst-24984	252	8	schmidhuber	schmidhuber	PROPN
ajst-24984	252	9	,	,	PUNCT
ajst-24984	252	10	"	"	PUNCT
ajst-24984	252	11	long	long	ADJ
ajst-24984	252	12	short	short	ADJ
ajst-24984	252	13	-	-	PUNCT
ajst-24984	252	14	term	term	NOUN
ajst-24984	252	15	memory	memory	NOUN
ajst-24984	252	16	,	,	PUNCT
ajst-24984	252	17	"	"	PUNCT
ajst-24984	252	18	in	in	ADP
ajst-24984	252	19	neural	neural	ADJ
ajst-24984	252	20	computation	computation	NOUN
ajst-24984	252	21	,	,	PUNCT
ajst-24984	252	22	vol	vol	NOUN
ajst-24984	252	23	.	.	PROPN
ajst-24984	252	24	9	9	NUM
ajst-24984	252	25	,	,	PUNCT
ajst-24984	252	26	no	no	INTJ
ajst-24984	252	27	.	.	NOUN
ajst-24984	252	28	8	8	NUM
ajst-24984	252	29	,	,	PUNCT
ajst-24984	252	30	pp	pp	ADJ
ajst-24984	252	31	.	.	PUNCT
ajst-24984	253	1	1735	1735	NUM
ajst-24984	253	2	-	-	SYM
ajst-24984	253	3	1780	1780	NUM
ajst-24984	253	4	,	,	PUNCT
ajst-24984	253	5	15	15	NUM
ajst-24984	253	6	nov	nov	NOUN
ajst-24984	253	7	.	.	PROPN
ajst-24984	253	8	1997	1997	NUM
ajst-24984	253	9	,	,	PUNCT
ajst-24984	253	10	doi	doi	NOUN
ajst-24984	253	11	:	:	PUNCT
ajst-24984	253	12	10.1162	10.1162	NUM
ajst-24984	253	13	/	/	SYM
ajst-24984	253	14	neco.1997.9.8.1735	neco.1997.9.8.1735	NOUN
ajst-24984	253	15	.	.	PUNCT
ajst-24984	254	1	[	[	X
ajst-24984	254	2	31	31	NUM
ajst-24984	254	3	]	]	X
ajst-24984	254	4	li	li	PROPN
ajst-24984	254	5	n	n	PROPN
ajst-24984	254	6	,	,	PUNCT
ajst-24984	254	7	chow	chow	PROPN
ajst-24984	254	8	cy	cy	PROPN
ajst-24984	254	9	,	,	PUNCT
ajst-24984	254	10	zhang	zhang	PROPN
ajst-24984	254	11	jd	jd	PROPN
ajst-24984	254	12	(	(	PUNCT
ajst-24984	254	13	2020	2020	NUM
ajst-24984	254	14	)	)	PUNCT
ajst-24984	254	15	seml	seml	NOUN
ajst-24984	254	16	:	:	PUNCT
ajst-24984	254	17	a	a	DET
ajst-24984	254	18	semi	semi	ADJ
ajst-24984	254	19	-	-	ADJ
ajst-24984	254	20	supervised	supervised	ADJ
ajst-24984	254	21	multi	multi	ADJ
ajst-24984	254	22	-	-	ADJ
ajst-24984	254	23	task	task	ADJ
ajst-24984	254	24	learning	learn	VERB
ajst-24984	254	25	framework	framework	NOUN
ajst-24984	254	26	for	for	ADP
ajst-24984	254	27	aspect	aspect	NOUN
ajst-24984	254	28	-	-	PUNCT
ajst-24984	254	29	based	base	VERB
ajst-24984	254	30	sentiment	sentiment	NOUN
ajst-24984	254	31	analysis	analysis	NOUN
ajst-24984	254	32	.	.	PUNCT
ajst-24984	255	1	ieee	ieee	NOUN
ajst-24984	255	2	access	access	NOUN
ajst-24984	255	3	8:189287–189297	8:189287–189297	NUM
ajst-24984	256	1	.	.	PUNCT
ajst-24984	257	1	[	[	X
ajst-24984	257	2	32	32	NUM
ajst-24984	257	3	]	]	X
ajst-24984	257	4	chen	chen	PROPN
ajst-24984	257	5	z	z	PROPN
ajst-24984	257	6	,	,	PUNCT
ajst-24984	257	7	qian	qian	PROPN
ajst-24984	257	8	t(2019	t(2019	PROPN
ajst-24984	257	9	)	)	PUNCT
ajst-24984	257	10	transfer	transfer	NOUN
ajst-24984	257	11	capsule	capsule	NOUN
ajst-24984	257	12	network	network	NOUN
ajst-24984	257	13	for	for	ADP
ajst-24984	257	14	aspect	aspect	NOUN
ajst-24984	257	15	level	level	NOUN
ajst-24984	257	16	sentiment	sentiment	NOUN
ajst-24984	257	17	classification	classification	NOUN
ajst-24984	257	18	.	.	PUNCT
ajst-24984	258	1	in	in	ADP
ajst-24984	258	2	:	:	PUNCT
ajst-24984	258	3	proceedings	proceeding	NOUN
ajst-24984	258	4	of	of	ADP
ajst-24984	258	5	the	the	DET
ajst-24984	258	6	57th	57th	ADJ
ajst-24984	258	7	annual	annual	ADJ
ajst-24984	258	8	meeting	meeting	NOUN
ajst-24984	258	9	of	of	ADP
ajst-24984	258	10	the	the	DET
ajst-24984	258	11	association	association	NOUN
ajst-24984	258	12	for	for	ADP
ajst-24984	258	13	computational	computational	ADJ
ajst-24984	258	14	linguistics	linguistic	NOUN
ajst-24984	258	15	,	,	PUNCT
ajst-24984	258	16	pp	pp	ADP
ajst-24984	258	17	547–556	547–556	NUM
ajst-24984	258	18	.	.	PUNCT
ajst-24984	259	1	[	[	X
ajst-24984	259	2	33	33	NUM
ajst-24984	259	3	]	]	X
ajst-24984	259	4	jindian	jindian	PROPN
ajst-24984	259	5	s	s	PROPN
ajst-24984	259	6	,	,	PUNCT
ajst-24984	259	7	shanshan	shanshan	PROPN
ajst-24984	259	8	yu	yu	PROPN
ajst-24984	259	9	,	,	PUNCT
ajst-24984	259	10	luo	luo	PROPN
ajst-24984	259	11	d	d	PROPN
ajst-24984	259	12	(	(	PUNCT
ajst-24984	259	13	2020	2020	NUM
ajst-24984	259	14	)	)	PUNCT
ajst-24984	259	15	enhancing	enhance	VERB
ajst-24984	259	16	aspect	aspect	NOUN
ajst-24984	259	17	-	-	PUNCT
ajst-24984	259	18	based	base	VERB
ajst-24984	259	19	sentiment	sentiment	NOUN
ajst-24984	259	20	analysis	analysis	NOUN
ajst-24984	259	21	with	with	ADP
ajst-24984	259	22	capsule	capsule	ADJ
ajst-24984	259	23	network	network	NOUN
ajst-24984	259	24	.	.	PUNCT
ajst-24984	260	1	ieee	ieee	NOUN
ajst-24984	260	2	access	access	NOUN
ajst-24984	260	3	8:100551–100561	8:100551–100561	NUM
ajst-24984	260	4	.	.	PUNCT
ajst-24984	261	1	[	[	X
ajst-24984	261	2	34	34	NUM
ajst-24984	261	3	]	]	X
ajst-24984	261	4	dong	dong	PROPN
ajst-24984	261	5	l	l	PROPN
ajst-24984	261	6	,	,	PUNCT
ajst-24984	261	7	wei	wei	PROPN
ajst-24984	261	8	f	f	PROPN
ajst-24984	261	9	,	,	PUNCT
ajst-24984	261	10	tan	tan	PROPN
ajst-24984	261	11	c	c	NOUN
ajst-24984	261	12	,	,	PUNCT
ajst-24984	261	13	tang	tang	X
ajst-24984	261	14	d	d	NOUN
ajst-24984	261	15	,	,	PUNCT
ajst-24984	261	16	zhou	zhou	PROPN
ajst-24984	261	17	m	m	PROPN
ajst-24984	261	18	,	,	PUNCT
ajst-24984	261	19	xu	xu	PROPN
ajst-24984	261	20	k(2014	k(2014	PROPN
ajst-24984	261	21	)	)	PUNCT
ajst-24984	261	22	adaptive	adaptive	ADJ
ajst-24984	261	23	recursive	recursive	ADJ
ajst-24984	261	24	neural	neural	ADJ
ajst-24984	261	25	network	network	NOUN
ajst-24984	261	26	for	for	ADP
ajst-24984	261	27	target	target	NOUN
ajst-24984	261	28	-	-	PUNCT
ajst-24984	261	29	dependent	dependent	ADJ
ajst-24984	261	30	twitter	twitter	NOUN
ajst-24984	261	31	sentiment	sentiment	NOUN
ajst-24984	261	32	classification	classification	NOUN
ajst-24984	261	33	.	.	PUNCT
ajst-24984	262	1	in	in	ADP
ajst-24984	262	2	:	:	PUNCT
ajst-24984	262	3	proceedings	proceeding	NOUN
ajst-24984	262	4	of	of	ADP
ajst-24984	262	5	the	the	DET
ajst-24984	262	6	52nd	52nd	ADJ
ajst-24984	262	7	annual	annual	ADJ
ajst-24984	262	8	meeting	meeting	NOUN
ajst-24984	262	9	of	of	ADP
ajst-24984	262	10	the	the	DET
ajst-24984	262	11	association	association	NOUN
ajst-24984	262	12	for	for	ADP
ajst-24984	262	13	computational	computational	ADJ
ajst-24984	262	14	linguistics	linguistic	NOUN
ajst-24984	262	15	(	(	PUNCT
ajst-24984	262	16	volume	volume	NOUN
ajst-24984	262	17	2	2	NUM
ajst-24984	262	18	:	:	PUNCT
ajst-24984	262	19	short	short	ADJ
ajst-24984	262	20	papers	paper	NOUN
ajst-24984	262	21	)	)	PUNCT
ajst-24984	262	22	,	,	PUNCT
ajst-24984	262	23	pp	pp	ADP
ajst-24984	262	24	49–54	49–54	NUM
ajst-24984	262	25	.	.	PUNCT
ajst-24984	263	1	[	[	X
ajst-24984	263	2	35	35	NUM
ajst-24984	263	3	]	]	PUNCT
ajst-24984	263	4	xu	xu	PROPN
ajst-24984	264	1	g	g	PROPN
ajst-24984	264	2	,	,	PUNCT
ajst-24984	264	3	zhang	zhang	PROPN
ajst-24984	264	4	z	z	PROPN
ajst-24984	264	5	,	,	PUNCT
ajst-24984	264	6	zhang	zhang	PROPN
ajst-24984	264	7	t	t	PROPN
ajst-24984	264	8	,	,	PUNCT
ajst-24984	264	9	et	et	PROPN
ajst-24984	264	10	al	al	PROPN
ajst-24984	264	11	.	.	PROPN
ajst-24984	264	12	aspect	aspect	NOUN
ajst-24984	264	13	-	-	PUNCT
ajst-24984	264	14	level	level	NOUN
ajst-24984	264	15	sentiment	sentiment	NOUN
ajst-24984	264	16	classification	classification	NOUN
ajst-24984	264	17	based	base	VERB
ajst-24984	264	18	on	on	ADP
ajst-24984	264	19	attention	attention	NOUN
ajst-24984	264	20	-	-	PUNCT
ajst-24984	264	21	bilstm	bilstm	NOUN
ajst-24984	264	22	model	model	NOUN
ajst-24984	264	23	and	and	CCONJ
ajst-24984	264	24	transfer	transfer	NOUN
ajst-24984	264	25	learning[j	learning[j	NOUN
ajst-24984	264	26	]	]	PUNCT
ajst-24984	264	27	.	.	PUNCT
ajst-24984	265	1	knowledge	knowledge	NOUN
ajst-24984	265	2	-	-	PUNCT
ajst-24984	265	3	based	base	VERB
ajst-24984	265	4	systems	system	NOUN
ajst-24984	265	5	,	,	PUNCT
ajst-24984	265	6	2022,245	2022,245	NUM
ajst-24984	265	7	.	.	PUNCT
ajst-24984	266	1	[	[	X
ajst-24984	266	2	36	36	NUM
ajst-24984	266	3	]	]	X
ajst-24984	266	4	liao	liao	PROPN
ajst-24984	266	5	w	w	PROPN
ajst-24984	266	6	,	,	PUNCT
ajst-24984	266	7	zhou	zhou	PROPN
ajst-24984	266	8	j	j	PROPN
ajst-24984	266	9	,	,	PUNCT
ajst-24984	266	10	wang	wang	PROPN
ajst-24984	266	11	y	y	PROPN
ajst-24984	266	12	,	,	PUNCT
ajst-24984	266	13	et	et	PROPN
ajst-24984	266	14	al	al	PROPN
ajst-24984	266	15	.	.	PUNCT
ajst-24984	266	16	fine	fine	ADV
ajst-24984	266	17	-	-	PUNCT
ajst-24984	266	18	grained	grain	VERB
ajst-24984	266	19	attention	attention	NOUN
ajst-24984	266	20	-	-	PUNCT
ajst-24984	266	21	based	base	VERB
ajst-24984	266	22	phrase	phrase	NOUN
ajst-24984	266	23	-	-	PUNCT
ajst-24984	266	24	aware	aware	ADJ
ajst-24984	266	25	network	network	NOUN
ajst-24984	266	26	for	for	ADP
ajst-24984	266	27	aspect	aspect	NOUN
ajst-24984	266	28	-	-	PUNCT
ajst-24984	266	29	level	level	NOUN
ajst-24984	266	30	sentiment	sentiment	NOUN
ajst-24984	266	31	analysis[j	analysis[j	PROPN
ajst-24984	266	32	]	]	PUNCT
ajst-24984	266	33	.	.	PUNCT
ajst-24984	267	1	artificial	artificial	ADJ
ajst-24984	267	2	intelligence	intelligence	NOUN
ajst-24984	267	3	review	review	NOUN
ajst-24984	267	4	,	,	PUNCT
ajst-24984	267	5	2021,55(5):1	2021,55(5):1	PROPN
ajst-24984	267	6	-	-	SYM
ajst-24984	267	7	20	20	NUM
ajst-24984	267	8	.	.	PUNCT
ajst-24984	268	1	[	[	X
ajst-24984	268	2	37	37	NUM
ajst-24984	268	3	]	]	X
ajst-24984	268	4	zhang	zhang	PROPN
ajst-24984	268	5	c	c	PROPN
ajst-24984	268	6	,	,	PUNCT
ajst-24984	268	7	li	li	PROPN
ajst-24984	268	8	q	q	PROPN
ajst-24984	268	9	,	,	PUNCT
ajst-24984	268	10	song	song	NOUN
ajst-24984	268	11	d.	d.	PROPN
ajst-24984	268	12	aspect	aspect	PROPN
ajst-24984	268	13	-	-	PUNCT
ajst-24984	268	14	based	base	VERB
ajst-24984	268	15	sentiment	sentiment	NOUN
ajst-24984	268	16	classification	classification	NOUN
ajst-24984	268	17	with	with	ADP
ajst-24984	268	18	aspect	aspect	NOUN
ajst-24984	268	19	-	-	PUNCT
ajst-24984	268	20	specific	specific	ADJ
ajst-24984	268	21	graph	graph	NOUN
ajst-24984	268	22	convolutional	convolutional	ADJ
ajst-24984	268	23	networks[j	networks[j	PROPN
ajst-24984	268	24	]	]	X
ajst-24984	268	25	.	.	PUNCT
ajst-24984	269	1	arxiv	arxiv	PROPN
ajst-24984	269	2	preprint	preprint	PROPN
ajst-24984	269	3	arxiv:1909.03477	arxiv:1909.03477	NOUN
ajst-24984	269	4	,	,	PUNCT
ajst-24984	269	5	2019	2019	NUM
ajst-24984	269	6	.	.	PUNCT
ajst-24984	270	1	[	[	X
ajst-24984	270	2	38	38	NUM
ajst-24984	270	3	]	]	PUNCT
ajst-24984	270	4	xiao	xiao	PROPN
ajst-24984	270	5	l	l	PROPN
ajst-24984	270	6	,	,	PUNCT
ajst-24984	270	7	hu	hu	PROPN
ajst-24984	270	8	x	x	PROPN
ajst-24984	270	9	,	,	PUNCT
ajst-24984	270	10	chen	chen	PROPN
ajst-24984	270	11	y	y	PROPN
ajst-24984	270	12	,	,	PUNCT
ajst-24984	270	13	et	et	PROPN
ajst-24984	270	14	al	al	PROPN
ajst-24984	270	15	.	.	PUNCT
ajst-24984	271	1	multi	multi	ADJ
ajst-24984	271	2	-	-	ADJ
ajst-24984	271	3	head	head	ADJ
ajst-24984	271	4	self	self	NOUN
ajst-24984	271	5	-	-	PUNCT
ajst-24984	271	6	attention	attention	NOUN
ajst-24984	271	7	based	base	VERB
ajst-24984	271	8	gated	gate	VERB
ajst-24984	271	9	graph	graph	NOUN
ajst-24984	271	10	convolutional	convolutional	ADJ
ajst-24984	271	11	networks	network	NOUN
ajst-24984	271	12	for	for	ADP
ajst-24984	271	13	aspect	aspect	NOUN
ajst-24984	271	14	-	-	PUNCT
ajst-24984	271	15	based	base	VERB
ajst-24984	271	16	sentiment	sentiment	NOUN
ajst-24984	271	17	classification[j	classification[j	PROPN
ajst-24984	271	18	]	]	PUNCT
ajst-24984	271	19	.	.	PUNCT
ajst-24984	272	1	multimedia	multimedia	NOUN
ajst-24984	272	2	tools	tool	NOUN
ajst-24984	272	3	and	and	CCONJ
ajst-24984	272	4	applications	application	NOUN
ajst-24984	272	5	,	,	PUNCT
ajst-24984	272	6	2020,81	2020,81	NUM
ajst-24984	272	7	(	(	PUNCT
ajst-24984	272	8	14	14	NUM
ajst-24984	272	9	):	):	PUNCT
ajst-24984	272	10	19051	19051	NUM
ajst-24984	272	11	-	-	SYM
ajst-24984	272	12	19070	19070	NUM
ajst-24984	272	13	.	.	PUNCT
ajst-24984	273	1	[	[	X
ajst-24984	273	2	39	39	NUM
ajst-24984	273	3	]	]	PUNCT
ajst-24984	273	4	meng	meng	PROPN
ajst-24984	273	5	f	f	PROPN
ajst-24984	273	6	,	,	PUNCT
ajst-24984	273	7	feng	feng	PROPN
ajst-24984	273	8	j	j	PROPN
ajst-24984	273	9	,	,	PUNCT
ajst-24984	273	10	yin	yin	PROPN
ajst-24984	273	11	d	d	PROPN
ajst-24984	273	12	,	,	PUNCT
ajst-24984	273	13	et	et	PROPN
ajst-24984	273	14	al	al	PROPN
ajst-24984	273	15	.	.	PROPN
ajst-24984	273	16	sentiment	sentiment	PROPN
ajst-24984	273	17	analysis	analysis	NOUN
ajst-24984	273	18	with	with	ADP
ajst-24984	273	19	weighted	weight	VERB
ajst-24984	273	20	graph	graph	NOUN
ajst-24984	273	21	convolutional	convolutional	ADJ
ajst-24984	273	22	networks[c]//emnlp	networks[c]//emnlp	PROPN
ajst-24984	273	23	(	(	PUNCT
ajst-24984	273	24	findings	finding	NOUN
ajst-24984	273	25	)	)	PUNCT
ajst-24984	273	26	.	.	PUNCT
ajst-24984	274	1	2020	2020	NUM
ajst-24984	274	2	:	:	PUNCT
ajst-24984	275	1	586	586	NUM
ajst-24984	275	2	-	-	SYM
ajst-24984	275	3	595	595	NUM
ajst-24984	275	4	.	.	PUNCT
ajst-24984	276	1	[	[	X
ajst-24984	276	2	40	40	NUM
ajst-24984	276	3	]	]	X
ajst-24984	276	4	du	du	PROPN
ajst-24984	276	5	j	j	PROPN
ajst-24984	276	6	,	,	PUNCT
ajst-24984	276	7	zhang	zhang	PROPN
ajst-24984	276	8	y	y	PROPN
ajst-24984	276	9	,	,	PUNCT
ajst-24984	276	10	yue	yue	PROPN
ajst-24984	276	11	b	b	PROPN
ajst-24984	276	12	,	,	PUNCT
ajst-24984	276	13	et	et	PROPN
ajst-24984	276	14	al	al	PROPN
ajst-24984	276	15	.	.	PUNCT
ajst-24984	276	16	bidirectional	bidirectional	ADJ
ajst-24984	276	17	edge	edge	NOUN
ajst-24984	276	18	-	-	PUNCT
ajst-24984	276	19	enhanced	enhance	VERB
ajst-24984	276	20	graph	graph	NOUN
ajst-24984	276	21	convolutional	convolutional	ADJ
ajst-24984	276	22	networks	network	NOUN
ajst-24984	276	23	for	for	ADP
ajst-24984	276	24	aspect	aspect	NOUN
ajst-24984	276	25	-	-	PUNCT
ajst-24984	276	26	based	base	VERB
ajst-24984	276	27	sentiment	sentiment	NOUN
ajst-24984	276	28	classification[c]//2021	classification[c]//2021	PRON
ajst-24984	276	29	ieee	ieee	NOUN
ajst-24984	276	30	45th	45th	ADJ
ajst-24984	276	31	annual	annual	ADJ
ajst-24984	276	32	computers	computer	NOUN
ajst-24984	276	33	,	,	PUNCT
ajst-24984	276	34	software	software	NOUN
ajst-24984	276	35	,	,	PUNCT
ajst-24984	276	36	and	and	CCONJ
ajst-24984	276	37	applications	application	NOUN
ajst-24984	276	38	conference	conference	NOUN
ajst-24984	276	39	(	(	PUNCT
ajst-24984	276	40	compsac	compsac	NOUN
ajst-24984	276	41	)	)	PUNCT
ajst-24984	276	42	.	.	PUNCT
ajst-24984	277	1	ieee	ieee	NOUN
ajst-24984	277	2	,	,	PUNCT
ajst-24984	277	3	2021	2021	NUM
ajst-24984	277	4	:	:	PUNCT
ajst-24984	277	5	1599	1599	NUM
ajst-24984	277	6	-	-	SYM
ajst-24984	277	7	1604	1604	NUM
ajst-24984	277	8	.	.	PUNCT
ajst-24984	278	1	[	[	X
ajst-24984	278	2	41	41	NUM
ajst-24984	278	3	]	]	PUNCT
ajst-24984	278	4	xiao	xiao	PROPN
ajst-24984	278	5	l	l	PROPN
ajst-24984	278	6	,	,	PUNCT
ajst-24984	278	7	gu	gu	PROPN
ajst-24984	278	8	d	d	PROPN
ajst-24984	278	9	,	,	PUNCT
ajst-24984	278	10	xue	xue	PROPN
ajst-24984	278	11	y	y	PROPN
ajst-24984	278	12	,	,	PUNCT
ajst-24984	278	13	et	et	PROPN
ajst-24984	278	14	al	al	PROPN
ajst-24984	278	15	.	.	PROPN
ajst-24984	278	16	sintactical	sintactical	ADJ
ajst-24984	278	17	distance	distance	NOUN
ajst-24984	278	18	attention	attention	NOUN
ajst-24984	278	19	guided	guide	VERB
ajst-24984	278	20	graph	graph	NOUN
ajst-24984	278	21	convolutional	convolutional	ADJ
ajst-24984	278	22	network	network	NOUN
ajst-24984	278	23	for	for	ADP
ajst-24984	278	24	aspect	aspect	NOUN
ajst-24984	278	25	-	-	PUNCT
ajst-24984	278	26	based	base	VERB
ajst-24984	278	27	sentiment	sentiment	NOUN
ajst-24984	278	28	analisis[c]//2021	analisis[c]//2021	PROPN
ajst-24984	278	29	international	international	ADJ
ajst-24984	278	30	joint	joint	ADJ
ajst-24984	278	31	conference	conference	NOUN
ajst-24984	278	32	on	on	ADP
ajst-24984	278	33	neural	neural	ADJ
ajst-24984	278	34	networks	network	NOUN
ajst-24984	278	35	(	(	PUNCT
ajst-24984	278	36	ijcnn	ijcnn	PROPN
ajst-24984	278	37	)	)	PUNCT
ajst-24984	278	38	.	.	PUNCT
ajst-24984	279	1	ieee	ieee	PROPN
ajst-24984	279	2	,	,	PUNCT
ajst-24984	279	3	2021	2021	NUM
ajst-24984	279	4	:	:	PUNCT
ajst-24984	279	5	1	1	NUM
ajst-24984	279	6	-	-	SYM
ajst-24984	279	7	7	7	NUM
ajst-24984	279	8	.	.	PUNCT
ajst-24984	280	1	[	[	X
ajst-24984	280	2	42	42	NUM
ajst-24984	280	3	]	]	X
ajst-24984	280	4	ma	ma	PROPN
ajst-24984	280	5	y	y	PROPN
ajst-24984	280	6	,	,	PUNCT
ajst-24984	280	7	pang	pang	NOUN
ajst-24984	280	8	y.	y.	PROPN
ajst-24984	280	9	learnable	learnable	ADJ
ajst-24984	280	10	dependency	dependency	NOUN
ajst-24984	280	11	-	-	PUNCT
ajst-24984	280	12	based	base	VERB
ajst-24984	280	13	double	double	ADJ
ajst-24984	280	14	graph	graph	NOUN
ajst-24984	280	15	structure	structure	NOUN
ajst-24984	280	16	for	for	ADP
ajst-24984	280	17	aspect	aspect	NOUN
ajst-24984	280	18	-	-	PUNCT
ajst-24984	280	19	based	base	VERB
ajst-24984	280	20	sentiment	sentiment	NOUN
ajst-24984	280	21	analysis[c]//	analysis[c]//	PROPN
ajst-24984	280	22	proceedings	proceeding	NOUN
ajst-24984	280	23	of	of	ADP
ajst-24984	280	24	the	the	DET
ajst-24984	280	25	29th	29th	ADJ
ajst-24984	280	26	international	international	ADJ
ajst-24984	280	27	conference	conference	NOUN
ajst-24984	280	28	on	on	ADP
ajst-24984	280	29	computational	computational	ADJ
ajst-24984	280	30	linguistics	linguistic	NOUN
ajst-24984	280	31	.	.	PUNCT
ajst-24984	281	1	2022	2022	NUM
ajst-24984	281	2	:	:	PUNCT
ajst-24984	281	3	7086	7086	NUM
ajst-24984	281	4	-	-	SYM
ajst-24984	281	5	7092	7092	NUM
ajst-24984	281	6	.	.	PUNCT
ajst-24984	282	1	[	[	X
ajst-24984	282	2	43	43	NUM
ajst-24984	282	3	]	]	X
ajst-24984	282	4	wang	wang	PROPN
ajst-24984	282	5	x	x	PROPN
ajst-24984	282	6	,	,	PUNCT
ajst-24984	282	7	liu	liu	PROPN
ajst-24984	282	8	p	p	PROPN
ajst-24984	282	9	,	,	PUNCT
ajst-24984	282	10	zhu	zhu	PROPN
ajst-24984	282	11	z	z	X
ajst-24984	282	12	,	,	PUNCT
ajst-24984	282	13	et	et	PROPN
ajst-24984	282	14	al	al	PROPN
ajst-24984	282	15	.	.	PROPN
ajst-24984	282	16	interactive	interactive	ADJ
ajst-24984	282	17	double	double	ADJ
ajst-24984	282	18	graph	graph	NOUN
ajst-24984	282	19	convolutional	convolutional	ADJ
ajst-24984	282	20	networks	network	NOUN
ajst-24984	282	21	for	for	ADP
ajst-24984	282	22	aspect	aspect	NOUN
ajst-24984	282	23	-	-	PUNCT
ajst-24984	282	24	based	base	VERB
ajst-24984	282	25	sentiment	sentiment	NOUN
ajst-24984	282	26	analysis[c]//2022	analysis[c]//2022	PROPN
ajst-24984	282	27	international	international	ADJ
ajst-24984	282	28	joint	joint	ADJ
ajst-24984	282	29	conference	conference	NOUN
ajst-24984	282	30	on	on	ADP
ajst-24984	282	31	neural	neural	ADJ
ajst-24984	282	32	networks	network	NOUN
ajst-24984	282	33	(	(	PUNCT
ajst-24984	282	34	ijcnn	ijcnn	PROPN
ajst-24984	282	35	)	)	PUNCT
ajst-24984	282	36	.	.	PUNCT
ajst-24984	283	1	ieee	ieee	PROPN
ajst-24984	283	2	,	,	PUNCT
ajst-24984	283	3	2022	2022	NUM
ajst-24984	283	4	:	:	PUNCT
ajst-24984	283	5	1	1	NUM
ajst-24984	283	6	-	-	SYM
ajst-24984	283	7	7	7	NUM
ajst-24984	283	8	.	.	NOUN
ajst-24984	283	9	198	198	NUM
ajst-24984	284	1	[	[	SYM
ajst-24984	284	2	44	44	NUM
ajst-24984	284	3	]	]	X
ajst-24984	285	1	zhang	zhang	PROPN
ajst-24984	285	2	z	z	PROPN
ajst-24984	285	3	,	,	PUNCT
ajst-24984	285	4	zhou	zhou	PROPN
ajst-24984	285	5	z	z	PROPN
ajst-24984	285	6	,	,	PUNCT
ajst-24984	285	7	wang	wang	PROPN
ajst-24984	285	8	y.	y.	PROPN
ajst-24984	285	9	ssegcn	ssegcn	PROPN
ajst-24984	285	10	:	:	PUNCT
ajst-24984	285	11	syntactic	syntactic	ADJ
ajst-24984	285	12	and	and	CCONJ
ajst-24984	285	13	semantic	semantic	ADJ
ajst-24984	285	14	enhanced	enhance	VERB
ajst-24984	285	15	graph	graph	NOUN
ajst-24984	285	16	convolutional	convolutional	ADJ
ajst-24984	285	17	network	network	NOUN
ajst-24984	285	18	for	for	ADP
ajst-24984	285	19	aspect	aspect	NOUN
ajst-24984	285	20	-	-	PUNCT
ajst-24984	285	21	based	base	VERB
ajst-24984	285	22	sentiment	sentiment	NOUN
ajst-24984	285	23	analysis[c]//proceedings	analysis[c]//proceedings	PROPN
ajst-24984	285	24	of	of	ADP
ajst-24984	285	25	the	the	DET
ajst-24984	285	26	2022	2022	NUM
ajst-24984	285	27	conference	conference	NOUN
ajst-24984	285	28	of	of	ADP
ajst-24984	285	29	the	the	DET
ajst-24984	285	30	north	north	ADJ
ajst-24984	285	31	american	american	ADJ
ajst-24984	285	32	chapter	chapter	NOUN
ajst-24984	285	33	of	of	ADP
ajst-24984	285	34	the	the	DET
ajst-24984	285	35	association	association	NOUN
ajst-24984	285	36	for	for	ADP
ajst-24984	285	37	computational	computational	ADJ
ajst-24984	285	38	linguistics	linguistic	NOUN
ajst-24984	285	39	:	:	PUNCT
ajst-24984	285	40	human	human	ADJ
ajst-24984	285	41	language	language	NOUN
ajst-24984	285	42	technologies	technology	NOUN
ajst-24984	285	43	.	.	PUNCT
ajst-24984	286	1	2022	2022	NUM
ajst-24984	286	2	:	:	PUNCT
ajst-24984	286	3	4916	4916	NUM
ajst-24984	286	4	-	-	SYM
ajst-24984	286	5	4925	4925	NUM
ajst-24984	286	6	.	.	PUNCT
ajst-24984	287	1	[	[	X
ajst-24984	287	2	45	45	NUM
ajst-24984	287	3	]	]	X
ajst-24984	287	4	meirong	meirong	NOUN
ajst-24984	287	5	d	d	NOUN
ajst-24984	287	6	,	,	PUNCT
ajst-24984	287	7	jinqian	jinqian	PROPN
ajst-24984	287	8	l	l	NOUN
ajst-24984	287	9	,	,	PUNCT
ajst-24984	287	10	biqing	biqe	VERB
ajst-24984	287	11	z	z	PROPN
ajst-24984	287	12	,	,	PUNCT
ajst-24984	287	13	et	et	PROPN
ajst-24984	287	14	al	al	PROPN
ajst-24984	287	15	.	.	PROPN
ajst-24984	287	16	aspect	aspect	PROPN
ajst-24984	287	17	sentiment	sentiment	NOUN
ajst-24984	287	18	analysis	analysis	NOUN
ajst-24984	287	19	based	base	VERB
ajst-24984	287	20	on	on	ADP
ajst-24984	287	21	local	local	ADJ
ajst-24984	287	22	-	-	PUNCT
ajst-24984	287	23	global	global	ADJ
ajst-24984	287	24	context	context	NOUN
ajst-24984	287	25	guidance[j	guidance[j	PROPN
ajst-24984	287	26	]	]	PUNCT
ajst-24984	287	27	.	.	PUNCT
ajst-24984	288	1	softwareguide	softwareguide	ADJ
ajst-24984	288	2	,	,	PUNCT
ajst-24984	288	3	2024	2024	NUM
ajst-24984	288	4	,	,	PUNCT
ajst-24984	288	5	23	23	NUM
ajst-24984	288	6	(	(	PUNCT
ajst-24984	288	7	01	01	NUM
ajst-24984	288	8	):	):	PUNCT
ajst-24984	288	9	190	190	NUM
ajst-24984	288	10	-	-	SYM
ajst-24984	288	11	196	196	NUM
ajst-24984	288	12	.	.	PUNCT
ajst-24984	289	1	[	[	X
ajst-24984	289	2	46	46	NUM
ajst-24984	289	3	]	]	X
ajst-24984	289	4	vaswani	vaswani	NOUN
ajst-24984	289	5	a	a	PRON
ajst-24984	289	6	,	,	PUNCT
ajst-24984	289	7	shazeer	shazeer	NOUN
ajst-24984	289	8	n	n	NOUN
ajst-24984	289	9	,	,	PUNCT
ajst-24984	289	10	parmar	parmar	PROPN
ajst-24984	289	11	n	n	PROPN
ajst-24984	289	12	,	,	PUNCT
ajst-24984	289	13	et	et	NOUN
ajst-24984	289	14	al.attention	al.attention	NOUN
ajst-24984	289	15	is	be	AUX
ajst-24984	289	16	all	all	DET
ajst-24984	289	17	you	you	PRON
ajst-24984	289	18	need[j].arxiv	need[j].arxiv	ADJ
ajst-24984	289	19	,	,	PUNCT
ajst-24984	289	20	2017	2017	NUM
ajst-24984	289	21	.	.	PUNCT
ajst-24984	290	1	[	[	X
ajst-24984	290	2	47	47	NUM
ajst-24984	290	3	]	]	X
ajst-24984	290	4	zhao	zhao	PROPN
ajst-24984	290	5	f	f	PROPN
ajst-24984	290	6	,	,	PUNCT
ajst-24984	290	7	wu	wu	PROPN
ajst-24984	290	8	z	z	PROPN
ajst-24984	290	9	,	,	PUNCT
ajst-24984	290	10	dai	dai	PROPN
ajst-24984	290	11	x(2020b	x(2020b	PROPN
ajst-24984	290	12	)	)	PUNCT
ajst-24984	290	13	attention	attention	NOUN
ajst-24984	290	14	transfer	transfer	NOUN
ajst-24984	290	15	network	network	NOUN
ajst-24984	290	16	for	for	ADP
ajst-24984	290	17	aspect	aspect	NOUN
ajst-24984	290	18	-	-	PUNCT
ajst-24984	290	19	level	level	NOUN
ajst-24984	290	20	sentiment	sentiment	NOUN
ajst-24984	290	21	classification	classification	NOUN
ajst-24984	290	22	.	.	PUNCT
ajst-24984	291	1	in	in	ADP
ajst-24984	291	2	:	:	PUNCT
ajst-24984	291	3	proceedings	proceeding	NOUN
ajst-24984	291	4	of	of	ADP
ajst-24984	291	5	the	the	DET
ajst-24984	291	6	28th	28th	ADJ
ajst-24984	291	7	international	international	ADJ
ajst-24984	291	8	conference	conference	NOUN
ajst-24984	291	9	on	on	ADP
ajst-24984	291	10	computational	computational	ADJ
ajst-24984	291	11	linguistics	linguistic	NOUN
ajst-24984	291	12	,	,	PUNCT
ajst-24984	291	13	pp	pp	ADP
ajst-24984	291	14	811–821	811–821	NUM
ajst-24984	291	15	.	.	PUNCT
ajst-24984	292	1	[	[	X
ajst-24984	292	2	48	48	NUM
ajst-24984	292	3	]	]	X
ajst-24984	292	4	zhao	zhao	PROPN
ajst-24984	292	5	,	,	PUNCT
ajst-24984	292	6	g.	g.	PROPN
ajst-24984	292	7	,	,	PUNCT
ajst-24984	292	8	luo	luo	PROPN
ajst-24984	292	9	,	,	PUNCT
ajst-24984	292	10	y.	y.	PROPN
ajst-24984	292	11	,	,	PUNCT
ajst-24984	292	12	chen	chen	PROPN
ajst-24984	292	13	,	,	PUNCT
ajst-24984	292	14	q.	q.	PROPN
ajst-24984	292	15	,	,	PUNCT
ajst-24984	292	16	&	&	CCONJ
ajst-24984	292	17	qian	qian	PROPN
ajst-24984	292	18	,	,	PUNCT
ajst-24984	292	19	x.	x.	NOUN
ajst-24984	292	20	(	(	PUNCT
ajst-24984	292	21	2023	2023	NUM
ajst-24984	292	22	)	)	PUNCT
ajst-24984	292	23	.	.	PUNCT
ajst-24984	292	24	aspect	aspect	NOUN
ajst-24984	292	25	-	-	PUNCT
ajst-24984	292	26	based	base	VERB
ajst-24984	292	27	sentiment	sentiment	NOUN
ajst-24984	292	28	analysis	analysis	NOUN
ajst-24984	292	29	via	via	ADP
ajst-24984	292	30	multitask	multitask	ADJ
ajst-24984	292	31	learning	learning	NOUN
ajst-24984	292	32	for	for	ADP
ajst-24984	292	33	online	online	ADJ
ajst-24984	292	34	reviews	review	NOUN
ajst-24984	292	35	.	.	PUNCT
ajst-24984	293	1	knowledge	knowledge	NOUN
ajst-24984	293	2	-	-	PUNCT
ajst-24984	293	3	based	base	VERB
ajst-24984	293	4	systems	system	NOUN
ajst-24984	293	5	,	,	PUNCT
ajst-24984	293	6	264	264	NUM
ajst-24984	293	7	,	,	PUNCT
ajst-24984	293	8	110326	110326	NUM
ajst-24984	293	9	.	.	PUNCT
ajst-24984	294	1	[	[	X
ajst-24984	294	2	49	49	NUM
ajst-24984	294	3	]	]	X
ajst-24984	294	4	liu	liu	PROPN
ajst-24984	294	5	j	j	PROPN
ajst-24984	294	6	,	,	PUNCT
ajst-24984	294	7	zhong	zhong	PROPN
ajst-24984	294	8	q	q	X
ajst-24984	294	9	,	,	PUNCT
ajst-24984	294	10	ding	ding	NOUN
ajst-24984	294	11	l	l	NOUN
ajst-24984	294	12	,	,	PUNCT
ajst-24984	294	13	et	et	PROPN
ajst-24984	294	14	al	al	PROPN
ajst-24984	294	15	.	.	PROPN
ajst-24984	294	16	unified	unified	ADJ
ajst-24984	294	17	instance	instance	NOUN
ajst-24984	294	18	and	and	CCONJ
ajst-24984	294	19	knowledge	knowledge	NOUN
ajst-24984	294	20	alignment	alignment	NOUN
ajst-24984	294	21	pretraining	pretraine	VERB
ajst-24984	294	22	for	for	ADP
ajst-24984	294	23	aspect	aspect	NOUN
ajst-24984	294	24	-	-	PUNCT
ajst-24984	294	25	based	base	VERB
ajst-24984	294	26	sentiment	sentiment	NOUN
ajst-24984	294	27	analysis[j	analysis[j	PROPN
ajst-24984	294	28	]	]	PUNCT
ajst-24984	294	29	.	.	PUNCT
ajst-24984	295	1	ieee	ieee	PROPN
ajst-24984	295	2	/	/	SYM
ajst-24984	295	3	acm	acm	PROPN
ajst-24984	295	4	transactions	transaction	NOUN
ajst-24984	295	5	on	on	ADP
ajst-24984	295	6	audio	audio	NOUN
ajst-24984	295	7	,	,	PUNCT
ajst-24984	295	8	speech	speech	NOUN
ajst-24984	295	9	,	,	PUNCT
ajst-24984	295	10	and	and	CCONJ
ajst-24984	295	11	language	language	NOUN
ajst-24984	295	12	processing	processing	NOUN
ajst-24984	295	13	,	,	PUNCT
ajst-24984	295	14	2023	2023	NUM
ajst-24984	295	15	,	,	PUNCT
ajst-24984	295	16	31	31	NUM
ajst-24984	295	17	:	:	PUNCT
ajst-24984	295	18	2629	2629	NUM
ajst-24984	295	19	-	-	SYM
ajst-24984	295	20	2642	2642	NUM
ajst-24984	295	21	.	.	PUNCT