id	sid	tid	token	lemma	pos
ajst-12262	1	1	academic	academic	ADJ
ajst-12262	1	2	journal	journal	NOUN
ajst-12262	1	3	of	of	ADP
ajst-12262	1	4	science	science	NOUN
ajst-12262	1	5	and	and	CCONJ
ajst-12262	1	6	technology	technology	NOUN
ajst-12262	1	7	issn	issn	NOUN
ajst-12262	1	8	:	:	PUNCT
ajst-12262	1	9	2771	2771	NUM
ajst-12262	1	10	-	-	SYM
ajst-12262	1	11	3032	3032	NUM
ajst-12262	1	12	|	|	NOUN
ajst-12262	1	13	vol	vol	NOUN
ajst-12262	1	14	.	.	PROPN
ajst-12262	2	1	7	7	NUM
ajst-12262	2	2	,	,	PUNCT
ajst-12262	2	3	no	no	INTJ
ajst-12262	2	4	.	.	NOUN
ajst-12262	2	5	2	2	NUM
ajst-12262	2	6	,	,	PUNCT
ajst-12262	2	7	2023	2023	NUM
ajst-12262	2	8	170	170	NUM
ajst-12262	2	9	research	research	NOUN
ajst-12262	2	10	on	on	ADP
ajst-12262	2	11	emotion	emotion	NOUN
ajst-12262	2	12	recognition	recognition	NOUN
ajst-12262	2	13	model	model	NOUN
ajst-12262	2	14	of	of	ADP
ajst-12262	2	15	takeaway	takeaway	NOUN
ajst-12262	2	16	evaluation	evaluation	NOUN
ajst-12262	2	17	text	text	NOUN
ajst-12262	2	18	based	base	VERB
ajst-12262	2	19	on	on	ADP
ajst-12262	2	20	lstm‐cnn	lstm‐cnn	PROPN
ajst-12262	2	21	ruiqing	ruiqe	VERB
ajst-12262	2	22	gao	gao	PROPN
ajst-12262	2	23	university	university	PROPN
ajst-12262	2	24	of	of	ADP
ajst-12262	2	25	technology	technology	PROPN
ajst-12262	2	26	sydney	sydney	PROPN
ajst-12262	2	27	,	,	PUNCT
ajst-12262	2	28	sydney	sydney	PROPN
ajst-12262	2	29	,	,	PUNCT
ajst-12262	2	30	australia	australia	PROPN
ajst-12262	2	31	abstract	abstract	NOUN
ajst-12262	2	32	:	:	PUNCT
ajst-12262	2	33	takeaway	takeaway	PROPN
ajst-12262	2	34	evaluation	evaluation	NOUN
ajst-12262	2	35	is	be	AUX
ajst-12262	2	36	a	a	DET
ajst-12262	2	37	literal	literal	ADJ
ajst-12262	2	38	evaluation	evaluation	NOUN
ajst-12262	2	39	of	of	ADP
ajst-12262	2	40	the	the	DET
ajst-12262	2	41	products	product	NOUN
ajst-12262	2	42	and	and	CCONJ
ajst-12262	2	43	services	service	NOUN
ajst-12262	2	44	experienced	experience	VERB
ajst-12262	2	45	by	by	ADP
ajst-12262	2	46	consumers	consumer	NOUN
ajst-12262	2	47	,	,	PUNCT
ajst-12262	2	48	which	which	PRON
ajst-12262	2	49	not	not	PART
ajst-12262	2	50	only	only	ADV
ajst-12262	2	51	provides	provide	VERB
ajst-12262	2	52	a	a	DET
ajst-12262	2	53	realistic	realistic	ADJ
ajst-12262	2	54	basis	basis	NOUN
ajst-12262	2	55	for	for	ADP
ajst-12262	2	56	the	the	DET
ajst-12262	2	57	improvement	improvement	NOUN
ajst-12262	2	58	of	of	ADP
ajst-12262	2	59	products	product	NOUN
ajst-12262	2	60	and	and	CCONJ
ajst-12262	2	61	services	service	NOUN
ajst-12262	2	62	of	of	ADP
ajst-12262	2	63	merchants	merchant	NOUN
ajst-12262	2	64	,	,	PUNCT
ajst-12262	2	65	but	but	CCONJ
ajst-12262	2	66	also	also	ADV
ajst-12262	2	67	affects	affect	VERB
ajst-12262	2	68	the	the	DET
ajst-12262	2	69	purchase	purchase	NOUN
ajst-12262	2	70	decision	decision	NOUN
ajst-12262	2	71	of	of	ADP
ajst-12262	2	72	consumers	consumer	NOUN
ajst-12262	2	73	in	in	ADP
ajst-12262	2	74	the	the	DET
ajst-12262	2	75	future	future	NOUN
ajst-12262	2	76	.	.	PUNCT
ajst-12262	3	1	in	in	ADP
ajst-12262	3	2	this	this	DET
ajst-12262	3	3	paper	paper	NOUN
ajst-12262	3	4	,	,	PUNCT
ajst-12262	3	5	nlp(natural	nlp(natural	ADJ
ajst-12262	3	6	language	language	NOUN
ajst-12262	3	7	processing	processing	NOUN
ajst-12262	3	8	)	)	PUNCT
ajst-12262	3	9	is	be	AUX
ajst-12262	3	10	used	use	VERB
ajst-12262	3	11	for	for	ADP
ajst-12262	3	12	preprocessing	preprocesse	VERB
ajst-12262	3	13	in	in	ADP
ajst-12262	3	14	tensorflow	tensorflow	NOUN
ajst-12262	3	15	environment	environment	NOUN
ajst-12262	3	16	,	,	PUNCT
ajst-12262	3	17	and	and	CCONJ
ajst-12262	3	18	an	an	DET
ajst-12262	3	19	emotion	emotion	NOUN
ajst-12262	3	20	recognition	recognition	NOUN
ajst-12262	3	21	model	model	NOUN
ajst-12262	3	22	of	of	ADP
ajst-12262	3	23	takeaway	takeaway	NOUN
ajst-12262	3	24	evaluation	evaluation	NOUN
ajst-12262	3	25	text	text	NOUN
ajst-12262	3	26	based	base	VERB
ajst-12262	3	27	on	on	ADP
ajst-12262	3	28	lstm	lstm	PROPN
ajst-12262	3	29	-	-	PUNCT
ajst-12262	3	30	cnn	cnn	PROPN
ajst-12262	3	31	is	be	AUX
ajst-12262	3	32	established	establish	VERB
ajst-12262	3	33	.	.	PUNCT
ajst-12262	4	1	the	the	DET
ajst-12262	4	2	model	model	NOUN
ajst-12262	4	3	is	be	AUX
ajst-12262	4	4	based	base	VERB
ajst-12262	4	5	on	on	ADP
ajst-12262	4	6	bilstm	bilstm	NOUN
ajst-12262	4	7	and	and	CCONJ
ajst-12262	4	8	attention	attention	NOUN
ajst-12262	4	9	mechanism	mechanism	NOUN
ajst-12262	4	10	,	,	PUNCT
ajst-12262	4	11	and	and	CCONJ
ajst-12262	4	12	uses	use	VERB
ajst-12262	4	13	word2vec	word2vec	PROPN
ajst-12262	4	14	method	method	VERB
ajst-12262	4	15	to	to	PART
ajst-12262	4	16	vectorize	vectorize	VERB
ajst-12262	4	17	text	text	NOUN
ajst-12262	4	18	vocabulary	vocabulary	NOUN
ajst-12262	4	19	.	.	PUNCT
ajst-12262	5	1	then	then	ADV
ajst-12262	5	2	,	,	PUNCT
ajst-12262	5	3	the	the	DET
ajst-12262	5	4	lstm	lstm	PROPN
ajst-12262	5	5	-	-	PUNCT
ajst-12262	5	6	cnn	cnn	PROPN
ajst-12262	5	7	serial	serial	ADJ
ajst-12262	5	8	hybrid	hybrid	ADJ
ajst-12262	5	9	model	model	NOUN
ajst-12262	5	10	is	be	AUX
ajst-12262	5	11	used	use	VERB
ajst-12262	5	12	to	to	PART
ajst-12262	5	13	extract	extract	VERB
ajst-12262	5	14	the	the	DET
ajst-12262	5	15	context	context	NOUN
ajst-12262	5	16	and	and	CCONJ
ajst-12262	5	17	local	local	ADJ
ajst-12262	5	18	semantics	semantic	NOUN
ajst-12262	5	19	of	of	ADP
ajst-12262	5	20	the	the	DET
ajst-12262	5	21	text	text	NOUN
ajst-12262	5	22	,	,	PUNCT
ajst-12262	5	23	and	and	CCONJ
ajst-12262	5	24	the	the	DET
ajst-12262	5	25	effectiveness	effectiveness	NOUN
ajst-12262	5	26	of	of	ADP
ajst-12262	5	27	the	the	DET
ajst-12262	5	28	algorithm	algorithm	NOUN
ajst-12262	5	29	is	be	AUX
ajst-12262	5	30	verified	verify	VERB
ajst-12262	5	31	by	by	ADP
ajst-12262	5	32	an	an	DET
ajst-12262	5	33	example	example	NOUN
ajst-12262	5	34	.	.	PUNCT
ajst-12262	6	1	the	the	DET
ajst-12262	6	2	precision	precision	NOUN
ajst-12262	6	3	reaches	reach	VERB
ajst-12262	6	4	0.937	0.937	NUM
ajst-12262	6	5	,	,	PUNCT
ajst-12262	6	6	the	the	DET
ajst-12262	6	7	recall	recall	NOUN
ajst-12262	6	8	rate	rate	NOUN
ajst-12262	6	9	reaches	reach	VERB
ajst-12262	6	10	0.896,f1	0.896,f1	NUM
ajst-12262	6	11	and	and	CCONJ
ajst-12262	6	12	the	the	DET
ajst-12262	6	13	f1	f1	NOUN
ajst-12262	6	14	value	value	NOUN
ajst-12262	6	15	reaches	reach	VERB
ajst-12262	6	16	0.906	0.906	NUM
ajst-12262	6	17	.	.	PUNCT
ajst-12262	7	1	through	through	ADP
ajst-12262	7	2	horizontal	horizontal	ADJ
ajst-12262	7	3	comparison	comparison	NOUN
ajst-12262	7	4	,	,	PUNCT
ajst-12262	7	5	it	it	PRON
ajst-12262	7	6	can	can	AUX
ajst-12262	7	7	be	be	AUX
ajst-12262	7	8	found	find	VERB
ajst-12262	7	9	that	that	SCONJ
ajst-12262	7	10	the	the	DET
ajst-12262	7	11	model	model	NOUN
ajst-12262	7	12	in	in	ADP
ajst-12262	7	13	this	this	DET
ajst-12262	7	14	paper	paper	NOUN
ajst-12262	7	15	is	be	AUX
ajst-12262	7	16	outstanding	outstanding	ADJ
ajst-12262	7	17	in	in	ADP
ajst-12262	7	18	this	this	DET
ajst-12262	7	19	task	task	NOUN
ajst-12262	7	20	,	,	PUNCT
ajst-12262	7	21	and	and	CCONJ
ajst-12262	7	22	the	the	DET
ajst-12262	7	23	emotion	emotion	NOUN
ajst-12262	7	24	enhancement	enhancement	NOUN
ajst-12262	7	25	model	model	NOUN
ajst-12262	7	26	also	also	ADV
ajst-12262	7	27	improves	improve	VERB
ajst-12262	7	28	the	the	DET
ajst-12262	7	29	classification	classification	NOUN
ajst-12262	7	30	accuracy	accuracy	NOUN
ajst-12262	7	31	to	to	ADP
ajst-12262	7	32	some	some	DET
ajst-12262	7	33	extent	extent	NOUN
ajst-12262	7	34	.	.	PUNCT
ajst-12262	8	1	keywords	keyword	NOUN
ajst-12262	8	2	:	:	PUNCT
ajst-12262	8	3	lstm	lstm	PROPN
ajst-12262	8	4	-	-	PUNCT
ajst-12262	8	5	cnn	cnn	PROPN
ajst-12262	8	6	,	,	PUNCT
ajst-12262	8	7	takeaway	takeaway	PROPN
ajst-12262	8	8	,	,	PUNCT
ajst-12262	8	9	evaluation	evaluation	NOUN
ajst-12262	8	10	text	text	NOUN
ajst-12262	8	11	,	,	PUNCT
ajst-12262	8	12	emotion	emotion	NOUN
ajst-12262	8	13	recognition	recognition	NOUN
ajst-12262	8	14	.	.	PUNCT
ajst-12262	9	1	1	1	X
ajst-12262	9	2	.	.	X
ajst-12262	9	3	introduction	introduction	NOUN
ajst-12262	9	4	in	in	ADP
ajst-12262	9	5	the	the	DET
ajst-12262	9	6	fast	fast	ADV
ajst-12262	9	7	-	-	PUNCT
ajst-12262	9	8	paced	pace	VERB
ajst-12262	9	9	internet	internet	NOUN
ajst-12262	9	10	era	era	NOUN
ajst-12262	9	11	,	,	PUNCT
ajst-12262	9	12	more	more	ADJ
ajst-12262	9	13	and	and	CCONJ
ajst-12262	9	14	more	more	ADJ
ajst-12262	9	15	people	people	NOUN
ajst-12262	9	16	use	use	VERB
ajst-12262	9	17	the	the	DET
ajst-12262	9	18	take	take	VERB
ajst-12262	9	19	-	-	PUNCT
ajst-12262	9	20	away	away	NOUN
ajst-12262	9	21	platform	platform	NOUN
ajst-12262	9	22	to	to	PART
ajst-12262	9	23	order	order	VERB
ajst-12262	9	24	food	food	NOUN
ajst-12262	9	25	,	,	PUNCT
ajst-12262	9	26	resulting	result	VERB
ajst-12262	9	27	in	in	ADP
ajst-12262	9	28	a	a	DET
ajst-12262	9	29	large	large	ADJ
ajst-12262	9	30	number	number	NOUN
ajst-12262	9	31	of	of	ADP
ajst-12262	9	32	take	take	NOUN
ajst-12262	9	33	-	-	PUNCT
ajst-12262	9	34	away	away	ADP
ajst-12262	9	35	comment	comment	NOUN
ajst-12262	9	36	data	datum	NOUN
ajst-12262	9	37	.	.	PUNCT
ajst-12262	10	1	customers	customer	NOUN
ajst-12262	10	2	can	can	AUX
ajst-12262	10	3	choose	choose	VERB
ajst-12262	10	4	goods	good	NOUN
ajst-12262	10	5	according	accord	VERB
ajst-12262	10	6	to	to	ADP
ajst-12262	10	7	the	the	DET
ajst-12262	10	8	number	number	NOUN
ajst-12262	10	9	and	and	CCONJ
ajst-12262	10	10	content	content	NOUN
ajst-12262	10	11	of	of	ADP
ajst-12262	10	12	comments	comment	NOUN
ajst-12262	10	13	,	,	PUNCT
ajst-12262	10	14	and	and	CCONJ
ajst-12262	10	15	businesses	business	NOUN
ajst-12262	10	16	can	can	AUX
ajst-12262	10	17	also	also	ADV
ajst-12262	10	18	mine	mine	VERB
ajst-12262	10	19	valuable	valuable	ADJ
ajst-12262	10	20	information	information	NOUN
ajst-12262	10	21	from	from	ADP
ajst-12262	10	22	positive	positive	ADJ
ajst-12262	10	23	and	and	CCONJ
ajst-12262	10	24	negative	negative	ADJ
ajst-12262	10	25	comments	comment	NOUN
ajst-12262	10	26	to	to	PART
ajst-12262	10	27	adjust	adjust	VERB
ajst-12262	10	28	food	food	NOUN
ajst-12262	10	29	,	,	PUNCT
ajst-12262	10	30	sales	sale	NOUN
ajst-12262	10	31	and	and	CCONJ
ajst-12262	10	32	industrial	industrial	ADJ
ajst-12262	10	33	structure	structure	NOUN
ajst-12262	11	1	[	[	X
ajst-12262	11	2	1	1	NUM
ajst-12262	11	3	]	]	PUNCT
ajst-12262	11	4	.	.	PUNCT
ajst-12262	12	1	takeaway	takeaway	PROPN
ajst-12262	12	2	evaluation	evaluation	NOUN
ajst-12262	12	3	is	be	AUX
ajst-12262	12	4	a	a	DET
ajst-12262	12	5	literal	literal	ADJ
ajst-12262	12	6	evaluation	evaluation	NOUN
ajst-12262	12	7	of	of	ADP
ajst-12262	12	8	the	the	DET
ajst-12262	12	9	products	product	NOUN
ajst-12262	12	10	and	and	CCONJ
ajst-12262	12	11	services	service	NOUN
ajst-12262	12	12	experienced	experience	VERB
ajst-12262	12	13	by	by	ADP
ajst-12262	12	14	consumers	consumer	NOUN
ajst-12262	12	15	,	,	PUNCT
ajst-12262	12	16	which	which	PRON
ajst-12262	12	17	not	not	PART
ajst-12262	12	18	only	only	ADV
ajst-12262	12	19	provides	provide	VERB
ajst-12262	12	20	a	a	DET
ajst-12262	12	21	realistic	realistic	ADJ
ajst-12262	12	22	basis	basis	NOUN
ajst-12262	12	23	for	for	ADP
ajst-12262	12	24	the	the	DET
ajst-12262	12	25	improvement	improvement	NOUN
ajst-12262	12	26	of	of	ADP
ajst-12262	12	27	products	product	NOUN
ajst-12262	12	28	and	and	CCONJ
ajst-12262	12	29	services	service	NOUN
ajst-12262	12	30	of	of	ADP
ajst-12262	12	31	merchants	merchant	NOUN
ajst-12262	12	32	,	,	PUNCT
ajst-12262	12	33	but	but	CCONJ
ajst-12262	12	34	also	also	ADV
ajst-12262	12	35	affects	affect	VERB
ajst-12262	12	36	the	the	DET
ajst-12262	12	37	purchase	purchase	NOUN
ajst-12262	12	38	decision	decision	NOUN
ajst-12262	12	39	of	of	ADP
ajst-12262	12	40	consumers	consumer	NOUN
ajst-12262	12	41	in	in	ADP
ajst-12262	12	42	the	the	DET
ajst-12262	12	43	future	future	NOUN
ajst-12262	12	44	.	.	PUNCT
ajst-12262	13	1	takeaway	takeaway	PROPN
ajst-12262	13	2	merchant	merchant	NOUN
ajst-12262	13	3	rating	rating	NOUN
ajst-12262	13	4	refers	refer	VERB
ajst-12262	13	5	to	to	ADP
ajst-12262	13	6	the	the	DET
ajst-12262	13	7	quantitative	quantitative	ADJ
ajst-12262	13	8	evaluation	evaluation	NOUN
ajst-12262	13	9	method	method	NOUN
ajst-12262	13	10	of	of	ADP
ajst-12262	13	11	products	product	NOUN
ajst-12262	13	12	or	or	CCONJ
ajst-12262	13	13	services	service	NOUN
ajst-12262	13	14	provided	provide	VERB
ajst-12262	13	15	by	by	ADP
ajst-12262	13	16	foreign	foreign	ADJ
ajst-12262	13	17	sellers	seller	NOUN
ajst-12262	13	18	.	.	PUNCT
ajst-12262	14	1	accurate	accurate	ADJ
ajst-12262	14	2	take	take	VERB
ajst-12262	14	3	-	-	PUNCT
ajst-12262	14	4	away	away	NOUN
ajst-12262	14	5	merchant	merchant	NOUN
ajst-12262	14	6	rating	rating	NOUN
ajst-12262	14	7	can	can	AUX
ajst-12262	14	8	not	not	PART
ajst-12262	14	9	only	only	ADV
ajst-12262	14	10	help	help	VERB
ajst-12262	14	11	the	the	DET
ajst-12262	14	12	platform	platform	NOUN
ajst-12262	14	13	realize	realize	VERB
ajst-12262	14	14	the	the	DET
ajst-12262	14	15	management	management	NOUN
ajst-12262	14	16	of	of	ADP
ajst-12262	14	17	survival	survival	NOUN
ajst-12262	14	18	of	of	ADP
ajst-12262	14	19	the	the	DET
ajst-12262	14	20	fittest	fit	ADJ
ajst-12262	14	21	,	,	PUNCT
ajst-12262	14	22	but	but	CCONJ
ajst-12262	14	23	also	also	ADV
ajst-12262	14	24	reduce	reduce	VERB
ajst-12262	14	25	the	the	DET
ajst-12262	14	26	cost	cost	NOUN
ajst-12262	14	27	of	of	ADP
ajst-12262	14	28	consumers	consumer	NOUN
ajst-12262	14	29	'	'	PART
ajst-12262	14	30	choice	choice	NOUN
ajst-12262	14	31	among	among	ADP
ajst-12262	14	32	a	a	DET
ajst-12262	14	33	large	large	ADJ
ajst-12262	14	34	number	number	NOUN
ajst-12262	14	35	of	of	ADP
ajst-12262	14	36	similar	similar	ADJ
ajst-12262	14	37	merchants	merchant	NOUN
ajst-12262	14	38	,	,	PUNCT
ajst-12262	14	39	which	which	PRON
ajst-12262	14	40	has	have	VERB
ajst-12262	14	41	important	important	ADJ
ajst-12262	14	42	practical	practical	ADJ
ajst-12262	14	43	guiding	guiding	NOUN
ajst-12262	14	44	significance	significance	NOUN
ajst-12262	14	45	for	for	ADP
ajst-12262	14	46	platform	platform	NOUN
ajst-12262	14	47	managers	manager	NOUN
ajst-12262	14	48	and	and	CCONJ
ajst-12262	14	49	consumers	consumer	NOUN
ajst-12262	14	50	.	.	PUNCT
ajst-12262	15	1	massive	massive	ADJ
ajst-12262	15	2	texts	text	NOUN
ajst-12262	15	3	come	come	VERB
ajst-12262	15	4	from	from	ADP
ajst-12262	15	5	many	many	ADJ
ajst-12262	15	6	users	user	NOUN
ajst-12262	15	7	of	of	ADP
ajst-12262	15	8	the	the	DET
ajst-12262	15	9	internet	internet	NOUN
ajst-12262	15	10	,	,	PUNCT
ajst-12262	15	11	with	with	ADP
ajst-12262	15	12	various	various	ADJ
ajst-12262	15	13	forms	form	NOUN
ajst-12262	15	14	and	and	CCONJ
ajst-12262	15	15	no	no	DET
ajst-12262	15	16	fixed	fixed	ADJ
ajst-12262	15	17	format	format	NOUN
ajst-12262	15	18	,	,	PUNCT
ajst-12262	15	19	so	so	SCONJ
ajst-12262	15	20	it	it	PRON
ajst-12262	15	21	is	be	AUX
ajst-12262	15	22	difficult	difficult	ADJ
ajst-12262	15	23	to	to	PART
ajst-12262	15	24	process	process	VERB
ajst-12262	15	25	them	they	PRON
ajst-12262	15	26	by	by	ADP
ajst-12262	15	27	simple	simple	ADJ
ajst-12262	15	28	automatic	automatic	ADJ
ajst-12262	15	29	means	mean	NOUN
ajst-12262	15	30	.	.	PUNCT
ajst-12262	16	1	if	if	SCONJ
ajst-12262	16	2	you	you	PRON
ajst-12262	16	3	rely	rely	VERB
ajst-12262	16	4	on	on	ADP
ajst-12262	16	5	manual	manual	ADJ
ajst-12262	16	6	processing	processing	NOUN
ajst-12262	16	7	,	,	PUNCT
ajst-12262	16	8	there	there	PRON
ajst-12262	16	9	are	be	VERB
ajst-12262	16	10	problems	problem	NOUN
ajst-12262	16	11	such	such	ADJ
ajst-12262	16	12	as	as	ADP
ajst-12262	16	13	excessive	excessive	ADJ
ajst-12262	16	14	workload	workload	NOUN
ajst-12262	16	15	and	and	CCONJ
ajst-12262	16	16	poor	poor	ADJ
ajst-12262	16	17	real	real	ADJ
ajst-12262	16	18	-	-	PUNCT
ajst-12262	16	19	time	time	NOUN
ajst-12262	16	20	performance	performance	NOUN
ajst-12262	16	21	[	[	X
ajst-12262	16	22	2	2	NUM
ajst-12262	16	23	]	]	PUNCT
ajst-12262	16	24	.	.	PUNCT
ajst-12262	17	1	in	in	ADP
ajst-12262	17	2	order	order	NOUN
ajst-12262	17	3	to	to	PART
ajst-12262	17	4	realize	realize	VERB
ajst-12262	17	5	the	the	DET
ajst-12262	17	6	emotional	emotional	ADJ
ajst-12262	17	7	analysis	analysis	NOUN
ajst-12262	17	8	of	of	ADP
ajst-12262	17	9	massive	massive	ADJ
ajst-12262	17	10	texts	text	NOUN
ajst-12262	17	11	,	,	PUNCT
ajst-12262	17	12	scholars	scholar	NOUN
ajst-12262	17	13	analyze	analyze	VERB
ajst-12262	17	14	the	the	DET
ajst-12262	17	15	emotional	emotional	ADJ
ajst-12262	17	16	tendency	tendency	NOUN
ajst-12262	17	17	of	of	ADP
ajst-12262	17	18	texts	text	NOUN
ajst-12262	17	19	through	through	ADP
ajst-12262	17	20	statistics	statistic	NOUN
ajst-12262	17	21	,	,	PUNCT
ajst-12262	17	22	machine	machine	NOUN
ajst-12262	17	23	learning	learning	NOUN
ajst-12262	17	24	and	and	CCONJ
ajst-12262	17	25	neural	neural	ADJ
ajst-12262	17	26	networks	network	NOUN
ajst-12262	18	1	[	[	X
ajst-12262	18	2	3	3	NUM
ajst-12262	18	3	-	-	SYM
ajst-12262	18	4	5	5	NUM
ajst-12262	18	5	]	]	PUNCT
ajst-12262	18	6	.	.	PUNCT
ajst-12262	19	1	in	in	ADP
ajst-12262	19	2	this	this	DET
ajst-12262	19	3	paper	paper	NOUN
ajst-12262	19	4	,	,	PUNCT
ajst-12262	19	5	nlp(natural	nlp(natural	ADJ
ajst-12262	19	6	language	language	NOUN
ajst-12262	19	7	processing	processing	NOUN
ajst-12262	19	8	)	)	PUNCT
ajst-12262	19	9	method	method	NOUN
ajst-12262	19	10	is	be	AUX
ajst-12262	19	11	used	use	VERB
ajst-12262	19	12	for	for	ADP
ajst-12262	19	13	preprocessing	preprocesse	VERB
ajst-12262	19	14	in	in	ADP
ajst-12262	19	15	tensorflow	tensorflow	NOUN
ajst-12262	19	16	environment	environment	NOUN
ajst-12262	19	17	,	,	PUNCT
ajst-12262	19	18	and	and	CCONJ
ajst-12262	19	19	an	an	DET
ajst-12262	19	20	emotion	emotion	NOUN
ajst-12262	19	21	recognition	recognition	NOUN
ajst-12262	19	22	model	model	NOUN
ajst-12262	19	23	of	of	ADP
ajst-12262	19	24	takeaway	takeaway	NOUN
ajst-12262	19	25	evaluation	evaluation	NOUN
ajst-12262	19	26	text	text	NOUN
ajst-12262	19	27	based	base	VERB
ajst-12262	19	28	on	on	ADP
ajst-12262	19	29	lstm	lstm	PROPN
ajst-12262	19	30	-	-	PUNCT
ajst-12262	19	31	cnn	cnn	PROPN
ajst-12262	19	32	is	be	AUX
ajst-12262	19	33	established	establish	VERB
ajst-12262	19	34	to	to	PART
ajst-12262	19	35	carry	carry	VERB
ajst-12262	19	36	out	out	ADP
ajst-12262	19	37	experiments	experiment	NOUN
ajst-12262	19	38	on	on	ADP
ajst-12262	19	39	the	the	DET
ajst-12262	19	40	takeaway	takeaway	PROPN
ajst-12262	19	41	data	datum	NOUN
ajst-12262	19	42	set	set	VERB
ajst-12262	19	43	.	.	PUNCT
ajst-12262	20	1	the	the	DET
ajst-12262	20	2	experimental	experimental	ADJ
ajst-12262	20	3	results	result	NOUN
ajst-12262	20	4	show	show	VERB
ajst-12262	20	5	that	that	SCONJ
ajst-12262	20	6	the	the	DET
ajst-12262	20	7	lstm	lstm	PROPN
ajst-12262	20	8	-	-	PUNCT
ajst-12262	20	9	cnn	cnn	PROPN
ajst-12262	20	10	model	model	NOUN
ajst-12262	20	11	has	have	VERB
ajst-12262	20	12	the	the	DET
ajst-12262	20	13	highest	high	ADJ
ajst-12262	20	14	accuracy	accuracy	NOUN
ajst-12262	20	15	.	.	PUNCT
ajst-12262	21	1	2	2	X
ajst-12262	21	2	.	.	X
ajst-12262	21	3	brief	brief	ADJ
ajst-12262	21	4	introduction	introduction	NOUN
ajst-12262	21	5	of	of	ADP
ajst-12262	21	6	text	text	NOUN
ajst-12262	21	7	sentiment	sentiment	NOUN
ajst-12262	21	8	analysis	analysis	NOUN
ajst-12262	21	9	method	method	ADJ
ajst-12262	21	10	emotional	emotional	ADJ
ajst-12262	21	11	analysis	analysis	NOUN
ajst-12262	21	12	of	of	ADP
ajst-12262	21	13	text	text	NOUN
ajst-12262	21	14	,	,	PUNCT
ajst-12262	21	15	also	also	ADV
ajst-12262	21	16	known	know	VERB
ajst-12262	21	17	as	as	ADP
ajst-12262	21	18	subjective	subjective	ADJ
ajst-12262	21	19	analysis	analysis	NOUN
ajst-12262	21	20	,	,	PUNCT
ajst-12262	21	21	refers	refer	VERB
ajst-12262	21	22	to	to	ADP
ajst-12262	21	23	identifying	identify	VERB
ajst-12262	21	24	and	and	CCONJ
ajst-12262	21	25	mining	mining	NOUN
ajst-12262	21	26	subjective	subjective	ADJ
ajst-12262	21	27	information	information	NOUN
ajst-12262	21	28	in	in	ADP
ajst-12262	21	29	the	the	DET
ajst-12262	21	30	original	original	ADJ
ajst-12262	21	31	text	text	NOUN
ajst-12262	21	32	by	by	ADP
ajst-12262	21	33	npl	npl	PROPN
ajst-12262	21	34	,	,	PUNCT
ajst-12262	21	35	text	text	NOUN
ajst-12262	21	36	mining	mining	NOUN
ajst-12262	21	37	and	and	CCONJ
ajst-12262	21	38	computer	computer	NOUN
ajst-12262	21	39	linguistics	linguistic	NOUN
ajst-12262	21	40	,	,	PUNCT
ajst-12262	21	41	so	so	SCONJ
ajst-12262	21	42	as	as	SCONJ
ajst-12262	21	43	to	to	PART
ajst-12262	21	44	get	get	VERB
ajst-12262	21	45	the	the	DET
ajst-12262	21	46	attitude	attitude	NOUN
ajst-12262	21	47	and	and	CCONJ
ajst-12262	21	48	opinions	opinion	NOUN
ajst-12262	21	49	of	of	ADP
ajst-12262	21	50	critics	critic	NOUN
ajst-12262	21	51	on	on	ADP
ajst-12262	21	52	the	the	DET
ajst-12262	21	53	research	research	NOUN
ajst-12262	21	54	object	object	NOUN
ajst-12262	21	55	[	[	X
ajst-12262	21	56	6	6	NUM
ajst-12262	21	57	]	]	PUNCT
ajst-12262	21	58	.	.	PUNCT
ajst-12262	22	1	the	the	DET
ajst-12262	22	2	analysis	analysis	NOUN
ajst-12262	22	3	method	method	NOUN
ajst-12262	22	4	based	base	VERB
ajst-12262	22	5	on	on	ADP
ajst-12262	22	6	emotion	emotion	NOUN
ajst-12262	22	7	dictionary	dictionary	NOUN
ajst-12262	22	8	comes	come	VERB
ajst-12262	22	9	from	from	ADP
ajst-12262	22	10	the	the	DET
ajst-12262	22	11	basic	basic	ADJ
ajst-12262	22	12	grammar	grammar	NOUN
ajst-12262	22	13	rule	rule	NOUN
ajst-12262	22	14	text	text	NOUN
ajst-12262	22	15	analysis	analysis	NOUN
ajst-12262	22	16	,	,	PUNCT
ajst-12262	22	17	and	and	CCONJ
ajst-12262	22	18	the	the	DET
ajst-12262	22	19	method	method	NOUN
ajst-12262	22	20	is	be	AUX
ajst-12262	22	21	relatively	relatively	ADV
ajst-12262	22	22	simple	simple	ADJ
ajst-12262	22	23	.	.	PUNCT
ajst-12262	23	1	based	base	VERB
ajst-12262	23	2	on	on	ADP
ajst-12262	23	3	the	the	DET
ajst-12262	23	4	method	method	NOUN
ajst-12262	23	5	of	of	ADP
ajst-12262	23	6	machine	machine	NOUN
ajst-12262	23	7	learning	learning	NOUN
ajst-12262	23	8	,	,	PUNCT
ajst-12262	23	9	it	it	PRON
ajst-12262	23	10	is	be	AUX
ajst-12262	23	11	necessary	necessary	ADJ
ajst-12262	23	12	to	to	PART
ajst-12262	23	13	manually	manually	ADV
ajst-12262	23	14	label	label	VERB
ajst-12262	23	15	a	a	DET
ajst-12262	23	16	large	large	ADJ
ajst-12262	23	17	number	number	NOUN
ajst-12262	23	18	of	of	ADP
ajst-12262	23	19	corpus	corpus	NOUN
ajst-12262	23	20	before	before	SCONJ
ajst-12262	23	21	it	it	PRON
ajst-12262	23	22	can	can	AUX
ajst-12262	23	23	be	be	AUX
ajst-12262	23	24	used	use	VERB
ajst-12262	23	25	as	as	ADP
ajst-12262	23	26	a	a	DET
ajst-12262	23	27	training	training	NOUN
ajst-12262	23	28	set	set	NOUN
ajst-12262	23	29	,	,	PUNCT
ajst-12262	23	30	and	and	CCONJ
ajst-12262	23	31	then	then	ADV
ajst-12262	23	32	extract	extract	VERB
ajst-12262	23	33	text	text	NOUN
ajst-12262	23	34	features	feature	NOUN
ajst-12262	23	35	through	through	ADP
ajst-12262	23	36	svm	svm	PROPN
ajst-12262	23	37	(	(	PUNCT
ajst-12262	23	38	support	support	NOUN
ajst-12262	23	39	vector	vector	NOUN
ajst-12262	23	40	machine	machine	NOUN
ajst-12262	23	41	)	)	PUNCT
ajst-12262	23	42	,	,	PUNCT
ajst-12262	23	43	random	random	ADJ
ajst-12262	23	44	forest	forest	NOUN
ajst-12262	23	45	and	and	CCONJ
ajst-12262	23	46	naive	naive	ADJ
ajst-12262	23	47	bayes	bayes	NOUN
ajst-12262	23	48	algorithms	algorithm	NOUN
ajst-12262	23	49	,	,	PUNCT
ajst-12262	23	50	and	and	CCONJ
ajst-12262	23	51	finally	finally	ADV
ajst-12262	23	52	build	build	VERB
ajst-12262	23	53	a	a	DET
ajst-12262	23	54	classifier	classifier	NOUN
ajst-12262	23	55	model	model	NOUN
ajst-12262	23	56	,	,	PUNCT
ajst-12262	23	57	and	and	CCONJ
ajst-12262	23	58	then	then	ADV
ajst-12262	23	59	use	use	VERB
ajst-12262	23	60	the	the	DET
ajst-12262	23	61	model	model	NOUN
ajst-12262	23	62	to	to	PART
ajst-12262	23	63	identify	identify	VERB
ajst-12262	23	64	the	the	DET
ajst-12262	23	65	emotional	emotional	ADJ
ajst-12262	23	66	tendency	tendency	NOUN
ajst-12262	23	67	of	of	ADP
ajst-12262	23	68	new	new	ADJ
ajst-12262	23	69	texts	text	NOUN
ajst-12262	23	70	after	after	ADP
ajst-12262	23	71	analyzing	analyze	VERB
ajst-12262	23	72	and	and	CCONJ
ajst-12262	23	73	verifying	verify	VERB
ajst-12262	23	74	the	the	DET
ajst-12262	23	75	model	model	NOUN
ajst-12262	23	76	[	[	X
ajst-12262	23	77	7	7	NUM
ajst-12262	23	78	-	-	SYM
ajst-12262	23	79	8	8	NUM
ajst-12262	23	80	]	]	PUNCT
ajst-12262	23	81	.	.	PUNCT
ajst-12262	24	1	lstm	lstm	NOUN
ajst-12262	24	2	(	(	PUNCT
ajst-12262	24	3	long	long	ADJ
ajst-12262	24	4	short	short	ADJ
ajst-12262	24	5	term	term	NOUN
ajst-12262	24	6	memory	memory	NOUN
ajst-12262	24	7	)	)	PUNCT
ajst-12262	24	8	is	be	AUX
ajst-12262	24	9	a	a	DET
ajst-12262	24	10	complex	complex	ADJ
ajst-12262	24	11	circulatory	circulatory	ADJ
ajst-12262	24	12	neural	neural	ADJ
ajst-12262	24	13	network	network	NOUN
ajst-12262	24	14	with	with	ADP
ajst-12262	24	15	long	long	ADJ
ajst-12262	24	16	-	-	PUNCT
ajst-12262	24	17	term	term	NOUN
ajst-12262	24	18	memory	memory	NOUN
ajst-12262	24	19	.	.	PUNCT
ajst-12262	25	1	lstm	lstm	PROPN
ajst-12262	25	2	uses	use	VERB
ajst-12262	25	3	three	three	NUM
ajst-12262	25	4	gates	gate	NOUN
ajst-12262	25	5	:	:	PUNCT
ajst-12262	25	6	input	input	NOUN
ajst-12262	25	7	gate	gate	NOUN
ajst-12262	25	8	,	,	PUNCT
ajst-12262	25	9	forgetting	forget	VERB
ajst-12262	25	10	gate	gate	NOUN
ajst-12262	25	11	and	and	CCONJ
ajst-12262	25	12	output	output	NOUN
ajst-12262	25	13	gate	gate	NOUN
ajst-12262	25	14	to	to	PART
ajst-12262	25	15	control	control	VERB
ajst-12262	25	16	information	information	NOUN
ajst-12262	25	17	.	.	PUNCT
ajst-12262	26	1	in	in	ADP
ajst-12262	26	2	most	most	ADJ
ajst-12262	26	3	nlp	nlp	ADJ
ajst-12262	26	4	problems	problem	NOUN
ajst-12262	26	5	,	,	PUNCT
ajst-12262	26	6	in	in	ADP
ajst-12262	26	7	addition	addition	NOUN
ajst-12262	26	8	to	to	ADP
ajst-12262	26	9	the	the	DET
ajst-12262	26	10	data	datum	NOUN
ajst-12262	26	11	that	that	PRON
ajst-12262	26	12	needs	need	VERB
ajst-12262	26	13	to	to	PART
ajst-12262	26	14	be	be	AUX
ajst-12262	26	15	input	input	NOUN
ajst-12262	26	16	first	first	ADV
ajst-12262	26	17	,	,	PUNCT
ajst-12262	26	18	the	the	DET
ajst-12262	26	19	subsequent	subsequent	ADJ
ajst-12262	26	20	data	data	NOUN
ajst-12262	26	21	is	be	AUX
ajst-12262	26	22	also	also	ADV
ajst-12262	26	23	needed	need	VERB
ajst-12262	26	24	[	[	PUNCT
ajst-12262	26	25	9	9	NUM
ajst-12262	26	26	]	]	PUNCT
ajst-12262	26	27	.	.	PUNCT
ajst-12262	27	1	for	for	ADP
ajst-12262	27	2	example	example	NOUN
ajst-12262	27	3	,	,	PUNCT
ajst-12262	27	4	in	in	ADP
ajst-12262	27	5	the	the	DET
ajst-12262	27	6	task	task	NOUN
ajst-12262	27	7	of	of	ADP
ajst-12262	27	8	aspect	aspect	NOUN
ajst-12262	27	9	-	-	PUNCT
ajst-12262	27	10	level	level	NOUN
ajst-12262	27	11	emotion	emotion	NOUN
ajst-12262	27	12	analysis	analysis	NOUN
ajst-12262	27	13	,	,	PUNCT
ajst-12262	27	14	the	the	DET
ajst-12262	27	15	emotional	emotional	ADJ
ajst-12262	27	16	tendency	tendency	NOUN
ajst-12262	27	17	of	of	ADP
ajst-12262	27	18	a	a	DET
ajst-12262	27	19	target	target	NOUN
ajst-12262	27	20	word	word	NOUN
ajst-12262	27	21	is	be	AUX
ajst-12262	27	22	analyzed	analyze	VERB
ajst-12262	27	23	,	,	PUNCT
ajst-12262	27	24	and	and	CCONJ
ajst-12262	27	25	the	the	DET
ajst-12262	27	26	emotional	emotional	ADJ
ajst-12262	27	27	words	word	NOUN
ajst-12262	27	28	expressing	express	VERB
ajst-12262	27	29	this	this	DET
ajst-12262	27	30	aspect	aspect	NOUN
ajst-12262	27	31	may	may	AUX
ajst-12262	27	32	appear	appear	VERB
ajst-12262	27	33	in	in	ADP
ajst-12262	27	34	front	front	NOUN
ajst-12262	27	35	of	of	ADP
ajst-12262	27	36	or	or	CCONJ
ajst-12262	27	37	behind	behind	ADP
ajst-12262	27	38	the	the	DET
ajst-12262	27	39	target	target	NOUN
ajst-12262	27	40	word	word	NOUN
ajst-12262	27	41	.	.	PUNCT
ajst-12262	28	1	because	because	SCONJ
ajst-12262	28	2	of	of	ADP
ajst-12262	28	3	the	the	DET
ajst-12262	28	4	excellent	excellent	ADJ
ajst-12262	28	5	performance	performance	NOUN
ajst-12262	28	6	of	of	ADP
ajst-12262	28	7	deep	deep	ADJ
ajst-12262	28	8	learning	learning	NOUN
ajst-12262	28	9	algorithm	algorithm	NOUN
ajst-12262	28	10	,	,	PUNCT
ajst-12262	28	11	many	many	ADJ
ajst-12262	28	12	current	current	ADJ
ajst-12262	28	13	researches	research	NOUN
ajst-12262	28	14	on	on	ADP
ajst-12262	28	15	emotion	emotion	NOUN
ajst-12262	28	16	recognition	recognition	NOUN
ajst-12262	28	17	use	use	VERB
ajst-12262	28	18	deep	deep	ADJ
ajst-12262	28	19	learning	learning	NOUN
ajst-12262	28	20	related	relate	VERB
ajst-12262	28	21	algorithms	algorithm	NOUN
ajst-12262	28	22	to	to	PART
ajst-12262	28	23	build	build	VERB
ajst-12262	28	24	models	model	NOUN
ajst-12262	28	25	.	.	PUNCT
ajst-12262	29	1	cnn	cnn	PROPN
ajst-12262	29	2	(	(	PUNCT
ajst-12262	29	3	convolutional	convolutional	ADJ
ajst-12262	29	4	neural	neural	ADJ
ajst-12262	29	5	network	network	NOUN
ajst-12262	29	6	)	)	PUNCT
ajst-12262	29	7	can	can	AUX
ajst-12262	29	8	reduce	reduce	VERB
ajst-12262	29	9	the	the	DET
ajst-12262	29	10	difference	difference	NOUN
ajst-12262	29	11	of	of	ADP
ajst-12262	29	12	input	input	NOUN
ajst-12262	29	13	frequency	frequency	NOUN
ajst-12262	29	14	and	and	CCONJ
ajst-12262	29	15	capture	capture	VERB
ajst-12262	29	16	local	local	ADJ
ajst-12262	29	17	information	information	NOUN
ajst-12262	29	18	,	,	PUNCT
ajst-12262	29	19	but	but	CCONJ
ajst-12262	29	20	it	it	PRON
ajst-12262	29	21	does	do	AUX
ajst-12262	29	22	not	not	PART
ajst-12262	29	23	consider	consider	VERB
ajst-12262	29	24	the	the	DET
ajst-12262	29	25	global	global	ADJ
ajst-12262	29	26	characteristics	characteristic	NOUN
ajst-12262	29	27	and	and	CCONJ
ajst-12262	29	28	background	background	NOUN
ajst-12262	29	29	.	.	PUNCT
ajst-12262	30	1	in	in	ADP
ajst-12262	30	2	short	short	ADJ
ajst-12262	30	3	,	,	PUNCT
ajst-12262	30	4	the	the	DET
ajst-12262	30	5	modeling	modeling	NOUN
ajst-12262	30	6	capabilities	capability	NOUN
ajst-12262	30	7	of	of	ADP
ajst-12262	30	8	cnn	cnn	PROPN
ajst-12262	30	9	and	and	CCONJ
ajst-12262	30	10	lstm	lstm	NOUN
ajst-12262	30	11	are	be	AUX
ajst-12262	30	12	limited	limit	VERB
ajst-12262	30	13	[	[	PUNCT
ajst-12262	30	14	1011	1011	NUM
ajst-12262	30	15	]	]	PUNCT
ajst-12262	30	16	.	.	PUNCT
ajst-12262	31	1	on	on	ADP
ajst-12262	31	2	this	this	DET
ajst-12262	31	3	basis	basis	NOUN
ajst-12262	31	4	,	,	PUNCT
ajst-12262	31	5	combining	combine	VERB
ajst-12262	31	6	cnn	cnn	PROPN
ajst-12262	31	7	and	and	CCONJ
ajst-12262	31	8	lstm	lstm	NOUN
ajst-12262	31	9	to	to	PART
ajst-12262	31	10	build	build	VERB
ajst-12262	31	11	a	a	DET
ajst-12262	31	12	network	network	NOUN
ajst-12262	31	13	for	for	ADP
ajst-12262	31	14	speech	speech	NOUN
ajst-12262	31	15	emotion	emotion	NOUN
ajst-12262	31	16	recognition	recognition	NOUN
ajst-12262	31	17	,	,	PUNCT
ajst-12262	31	18	so	so	SCONJ
ajst-12262	31	19	as	as	SCONJ
ajst-12262	31	20	to	to	PART
ajst-12262	31	21	learn	learn	VERB
ajst-12262	31	22	the	the	DET
ajst-12262	31	23	best	good	ADJ
ajst-12262	31	24	description	description	NOUN
ajst-12262	31	25	of	of	ADP
ajst-12262	31	26	information	information	NOUN
ajst-12262	31	27	.	.	PUNCT
ajst-12262	32	1	3	3	X
ajst-12262	32	2	.	.	X
ajst-12262	32	3	emotion	emotion	NOUN
ajst-12262	32	4	recognition	recognition	NOUN
ajst-12262	32	5	model	model	NOUN
ajst-12262	32	6	of	of	ADP
ajst-12262	32	7	takeaway	takeaway	NOUN
ajst-12262	32	8	evaluation	evaluation	NOUN
ajst-12262	32	9	text	text	NOUN
ajst-12262	32	10	based	base	VERB
ajst-12262	32	11	on	on	ADP
ajst-12262	32	12	lstm	lstm	PROPN
ajst-12262	32	13	-	-	PUNCT
ajst-12262	32	14	cnn	cnn	PROPN
ajst-12262	32	15	the	the	DET
ajst-12262	32	16	existing	exist	VERB
ajst-12262	32	17	take	take	VERB
ajst-12262	32	18	-	-	PUNCT
ajst-12262	32	19	away	away	NOUN
ajst-12262	32	20	data	datum	NOUN
ajst-12262	32	21	on	on	ADP
ajst-12262	32	22	the	the	DET
ajst-12262	32	23	internet	internet	NOUN
ajst-12262	32	24	are	be	AUX
ajst-12262	32	25	uneven	uneven	ADJ
ajst-12262	32	26	,	,	PUNCT
ajst-12262	32	27	with	with	ADP
ajst-12262	32	28	an	an	DET
ajst-12262	32	29	average	average	ADJ
ajst-12262	32	30	length	length	NOUN
ajst-12262	32	31	of	of	ADP
ajst-12262	32	32	20	20	NUM
ajst-12262	32	33	-	-	SYM
ajst-12262	32	34	50	50	NUM
ajst-12262	32	35	.	.	PUNCT
ajst-12262	33	1	most	most	ADJ
ajst-12262	33	2	take	take	VERB
ajst-12262	33	3	-	-	PUNCT
ajst-12262	33	4	away	away	ADP
ajst-12262	33	5	comments	comment	NOUN
ajst-12262	33	6	are	be	AUX
ajst-12262	33	7	short	short	ADJ
ajst-12262	33	8	in	in	ADP
ajst-12262	33	9	text	text	NOUN
ajst-12262	33	10	length	length	NOUN
ajst-12262	33	11	,	,	PUNCT
ajst-12262	33	12	but	but	CCONJ
ajst-12262	33	13	there	there	PRON
ajst-12262	33	14	are	be	VERB
ajst-12262	33	15	still	still	ADV
ajst-12262	33	16	a	a	DET
ajst-12262	33	17	few	few	ADJ
ajst-12262	33	18	take	take	VERB
ajst-12262	33	19	-	-	PUNCT
ajst-12262	33	20	away	away	NOUN
ajst-12262	33	21	comments	comment	NOUN
ajst-12262	33	22	with	with	ADP
ajst-12262	33	23	a	a	DET
ajst-12262	33	24	long	long	ADJ
ajst-12262	33	25	text	text	NOUN
ajst-12262	33	26	length	length	NOUN
ajst-12262	33	27	ranging	range	VERB
ajst-12262	33	28	from	from	ADP
ajst-12262	33	29	100	100	NUM
ajst-12262	33	30	to	to	ADP
ajst-12262	33	31	several	several	ADJ
ajst-12262	33	32	171	171	NUM
ajst-12262	33	33	hundred	hundred	NUM
ajst-12262	33	34	words	word	NOUN
ajst-12262	33	35	.	.	PUNCT
ajst-12262	34	1	in	in	ADP
ajst-12262	34	2	addition	addition	NOUN
ajst-12262	34	3	to	to	ADP
ajst-12262	34	4	the	the	DET
ajst-12262	34	5	dishes	dish	NOUN
ajst-12262	34	6	will	will	AUX
ajst-12262	34	7	affect	affect	VERB
ajst-12262	34	8	the	the	DET
ajst-12262	34	9	emotional	emotional	ADJ
ajst-12262	34	10	tendency	tendency	NOUN
ajst-12262	34	11	of	of	ADP
ajst-12262	34	12	user	user	NOUN
ajst-12262	34	13	comments	comment	NOUN
ajst-12262	34	14	,	,	PUNCT
ajst-12262	34	15	whether	whether	SCONJ
ajst-12262	34	16	the	the	DET
ajst-12262	34	17	user	user	NOUN
ajst-12262	34	18	's	's	PART
ajst-12262	34	19	order	order	NOUN
ajst-12262	34	20	will	will	AUX
ajst-12262	34	21	affect	affect	VERB
ajst-12262	34	22	the	the	DET
ajst-12262	34	23	emotional	emotional	ADJ
ajst-12262	34	24	tendency	tendency	NOUN
ajst-12262	34	25	.	.	PUNCT
ajst-12262	35	1	therefore	therefore	ADV
ajst-12262	35	2	,	,	PUNCT
ajst-12262	35	3	we	we	PRON
ajst-12262	35	4	can	can	AUX
ajst-12262	35	5	continue	continue	VERB
ajst-12262	35	6	to	to	PART
ajst-12262	35	7	use	use	VERB
ajst-12262	35	8	the	the	DET
ajst-12262	35	9	captured	capture	VERB
ajst-12262	35	10	information	information	NOUN
ajst-12262	35	11	,	,	PUNCT
ajst-12262	35	12	that	that	ADV
ajst-12262	35	13	is	is	ADV
ajst-12262	35	14	,	,	PUNCT
ajst-12262	35	15	the	the	DET
ajst-12262	35	16	different	different	ADJ
ajst-12262	35	17	delivery	delivery	NOUN
ajst-12262	35	18	time	time	NOUN
ajst-12262	35	19	and	and	CCONJ
ajst-12262	35	20	ordering	order	VERB
ajst-12262	35	21	time	time	NOUN
ajst-12262	35	22	of	of	ADP
ajst-12262	35	23	the	the	DET
ajst-12262	35	24	user	user	NOUN
ajst-12262	35	25	's	's	PART
ajst-12262	35	26	order	order	NOUN
ajst-12262	35	27	,	,	PUNCT
ajst-12262	35	28	to	to	PART
ajst-12262	35	29	study	study	VERB
ajst-12262	35	30	whether	whether	SCONJ
ajst-12262	35	31	it	it	PRON
ajst-12262	35	32	will	will	AUX
ajst-12262	35	33	affect	affect	VERB
ajst-12262	35	34	the	the	DET
ajst-12262	35	35	user	user	NOUN
ajst-12262	35	36	's	's	PART
ajst-12262	35	37	emotional	emotional	ADJ
ajst-12262	35	38	tendency	tendency	NOUN
ajst-12262	35	39	.	.	PUNCT
ajst-12262	36	1	through	through	ADP
ajst-12262	36	2	the	the	DET
ajst-12262	36	3	comparative	comparative	ADJ
ajst-12262	36	4	analysis	analysis	NOUN
ajst-12262	36	5	of	of	ADP
ajst-12262	36	6	the	the	DET
ajst-12262	36	7	user	user	NOUN
ajst-12262	36	8	's	's	PART
ajst-12262	36	9	order	order	NOUN
ajst-12262	36	10	emotion	emotion	NOUN
ajst-12262	36	11	,	,	PUNCT
ajst-12262	36	12	we	we	PRON
ajst-12262	36	13	can	can	AUX
ajst-12262	36	14	provide	provide	VERB
ajst-12262	36	15	reference	reference	NOUN
ajst-12262	36	16	for	for	ADP
ajst-12262	36	17	merchants	merchant	NOUN
ajst-12262	36	18	and	and	CCONJ
ajst-12262	36	19	take	take	VERB
ajst-12262	36	20	-	-	PUNCT
ajst-12262	36	21	away	away	NOUN
ajst-12262	36	22	enterprises	enterprise	NOUN
ajst-12262	36	23	to	to	PART
ajst-12262	36	24	make	make	VERB
ajst-12262	36	25	decisions	decision	NOUN
ajst-12262	36	26	in	in	ADP
ajst-12262	36	27	many	many	ADJ
ajst-12262	36	28	ways	way	NOUN
ajst-12262	36	29	.	.	PUNCT
ajst-12262	37	1	this	this	DET
ajst-12262	37	2	project	project	NOUN
ajst-12262	37	3	intends	intend	VERB
ajst-12262	37	4	to	to	PART
ajst-12262	37	5	study	study	VERB
ajst-12262	37	6	an	an	DET
ajst-12262	37	7	emotion	emotion	NOUN
ajst-12262	37	8	recognition	recognition	NOUN
ajst-12262	37	9	method	method	NOUN
ajst-12262	37	10	of	of	ADP
ajst-12262	37	11	take	take	VERB
ajst-12262	37	12	-	-	PUNCT
ajst-12262	37	13	away	away	ADP
ajst-12262	37	14	comment	comment	NOUN
ajst-12262	37	15	text	text	NOUN
ajst-12262	37	16	based	base	VERB
ajst-12262	37	17	on	on	ADP
ajst-12262	37	18	lstm	lstm	PROPN
ajst-12262	37	19	-	-	PUNCT
ajst-12262	37	20	cnn	cnn	PROPN
ajst-12262	37	21	,	,	PUNCT
ajst-12262	37	22	in	in	ADP
ajst-12262	37	23	order	order	NOUN
ajst-12262	37	24	to	to	PART
ajst-12262	37	25	solve	solve	VERB
ajst-12262	37	26	the	the	DET
ajst-12262	37	27	problems	problem	NOUN
ajst-12262	37	28	of	of	ADP
ajst-12262	37	29	poor	poor	ADJ
ajst-12262	37	30	real	real	ADJ
ajst-12262	37	31	-	-	PUNCT
ajst-12262	37	32	time	time	NOUN
ajst-12262	37	33	performance	performance	NOUN
ajst-12262	37	34	,	,	PUNCT
ajst-12262	37	35	difficulty	difficulty	NOUN
ajst-12262	37	36	in	in	ADP
ajst-12262	37	37	applying	apply	VERB
ajst-12262	37	38	to	to	ADP
ajst-12262	37	39	large	large	ADJ
ajst-12262	37	40	-	-	PUNCT
ajst-12262	37	41	scale	scale	NOUN
ajst-12262	37	42	text	text	NOUN
ajst-12262	37	43	,	,	PUNCT
ajst-12262	37	44	and	and	CCONJ
ajst-12262	37	45	inability	inability	NOUN
ajst-12262	37	46	to	to	PART
ajst-12262	37	47	extract	extract	VERB
ajst-12262	37	48	text	text	NOUN
ajst-12262	37	49	context	context	NOUN
ajst-12262	37	50	and	and	CCONJ
ajst-12262	37	51	local	local	ADJ
ajst-12262	37	52	semantic	semantic	ADJ
ajst-12262	37	53	features	feature	NOUN
ajst-12262	37	54	at	at	ADP
ajst-12262	37	55	the	the	DET
ajst-12262	37	56	same	same	ADJ
ajst-12262	37	57	time	time	NOUN
ajst-12262	37	58	.	.	PUNCT
ajst-12262	38	1	this	this	DET
ajst-12262	38	2	structure	structure	NOUN
ajst-12262	38	3	is	be	AUX
ajst-12262	38	4	shown	show	VERB
ajst-12262	38	5	in	in	ADP
ajst-12262	38	6	figure	figure	NOUN
ajst-12262	38	7	1	1	NUM
ajst-12262	38	8	.	.	PUNCT
ajst-12262	38	9	figure	figure	NOUN
ajst-12262	38	10	1	1	NUM
ajst-12262	38	11	.	.	PUNCT
ajst-12262	39	1	emotion	emotion	NOUN
ajst-12262	39	2	recognition	recognition	NOUN
ajst-12262	39	3	model	model	NOUN
ajst-12262	39	4	of	of	ADP
ajst-12262	39	5	takeaway	takeaway	NOUN
ajst-12262	39	6	evaluation	evaluation	NOUN
ajst-12262	39	7	text	text	NOUN
ajst-12262	39	8	based	base	VERB
ajst-12262	39	9	on	on	ADP
ajst-12262	39	10	lstm	lstm	PROPN
ajst-12262	39	11	-	-	PROPN
ajst-12262	39	12	cnn	cnn	PROPN
ajst-12262	39	13	a	a	DET
ajst-12262	39	14	attention	attention	NOUN
ajst-12262	39	15	mechanism	mechanism	NOUN
ajst-12262	39	16	based	base	VERB
ajst-12262	39	17	on	on	ADP
ajst-12262	39	18	bi	bi	ADJ
ajst-12262	39	19	-	-	ADJ
ajst-12262	39	20	lstm	lstm	ADJ
ajst-12262	39	21	is	be	AUX
ajst-12262	39	22	used	use	VERB
ajst-12262	39	23	,	,	PUNCT
ajst-12262	39	24	and	and	CCONJ
ajst-12262	39	25	the	the	DET
ajst-12262	39	26	vectorization	vectorization	NOUN
ajst-12262	39	27	of	of	ADP
ajst-12262	39	28	text	text	NOUN
ajst-12262	39	29	vocabulary	vocabulary	NOUN
ajst-12262	39	30	is	be	AUX
ajst-12262	39	31	realized	realize	VERB
ajst-12262	39	32	by	by	ADP
ajst-12262	39	33	word2vec	word2vec	PROPN
ajst-12262	39	34	.	.	PUNCT
ajst-12262	40	1	on	on	ADP
ajst-12262	40	2	this	this	DET
ajst-12262	40	3	basis	basis	NOUN
ajst-12262	40	4	,	,	PUNCT
ajst-12262	40	5	using	use	VERB
ajst-12262	40	6	the	the	DET
ajst-12262	40	7	lstm	lstm	PROPN
ajst-12262	40	8	-	-	PUNCT
ajst-12262	40	9	cnn	cnn	PROPN
ajst-12262	40	10	serial	serial	ADJ
ajst-12262	40	11	hybrid	hybrid	ADJ
ajst-12262	40	12	model	model	NOUN
ajst-12262	40	13	,	,	PUNCT
ajst-12262	40	14	the	the	DET
ajst-12262	40	15	semantic	semantic	ADJ
ajst-12262	40	16	information	information	NOUN
ajst-12262	40	17	of	of	ADP
ajst-12262	40	18	the	the	DET
ajst-12262	40	19	text	text	NOUN
ajst-12262	40	20	is	be	AUX
ajst-12262	40	21	extracted	extract	VERB
ajst-12262	40	22	from	from	ADP
ajst-12262	40	23	the	the	DET
ajst-12262	40	24	context	context	NOUN
ajst-12262	40	25	and	and	CCONJ
ajst-12262	40	26	local	local	ADJ
ajst-12262	40	27	semantic	semantic	ADJ
ajst-12262	40	28	information	information	NOUN
ajst-12262	40	29	respectively	respectively	ADV
ajst-12262	40	30	.	.	PUNCT
ajst-12262	41	1	finally	finally	ADV
ajst-12262	41	2	,	,	PUNCT
ajst-12262	41	3	the	the	DET
ajst-12262	41	4	text	text	NOUN
ajst-12262	41	5	is	be	AUX
ajst-12262	41	6	classified	classify	VERB
ajst-12262	41	7	by	by	ADP
ajst-12262	41	8	softmax	softmax	NOUN
ajst-12262	41	9	classifier	classifier	NOUN
ajst-12262	41	10	.	.	PUNCT
ajst-12262	42	1	a	a	DET
ajst-12262	42	2	large	large	ADJ
ajst-12262	42	3	number	number	NOUN
ajst-12262	42	4	of	of	ADP
ajst-12262	42	5	comments	comment	NOUN
ajst-12262	42	6	are	be	AUX
ajst-12262	42	7	freely	freely	ADV
ajst-12262	42	8	written	write	VERB
ajst-12262	42	9	by	by	ADP
ajst-12262	42	10	different	different	ADJ
ajst-12262	42	11	users	user	NOUN
ajst-12262	42	12	,	,	PUNCT
ajst-12262	42	13	and	and	CCONJ
ajst-12262	42	14	there	there	PRON
ajst-12262	42	15	is	be	VERB
ajst-12262	42	16	no	no	DET
ajst-12262	42	17	structured	structured	ADJ
ajst-12262	42	18	or	or	CCONJ
ajst-12262	42	19	standardized	standardized	ADJ
ajst-12262	42	20	grammar	grammar	NOUN
ajst-12262	42	21	and	and	CCONJ
ajst-12262	42	22	pattern	pattern	NOUN
ajst-12262	42	23	,	,	PUNCT
ajst-12262	42	24	which	which	PRON
ajst-12262	42	25	is	be	AUX
ajst-12262	42	26	highly	highly	ADV
ajst-12262	42	27	unstructured	unstructured	ADJ
ajst-12262	42	28	.	.	PUNCT
ajst-12262	43	1	the	the	DET
ajst-12262	43	2	open	open	ADJ
ajst-12262	43	3	source	source	NOUN
ajst-12262	43	4	word	word	NOUN
ajst-12262	43	5	vector	vector	NOUN
ajst-12262	43	6	tool	tool	NOUN
ajst-12262	43	7	word2vec	word2vec	PRON
ajst-12262	43	8	can	can	AUX
ajst-12262	43	9	transform	transform	VERB
ajst-12262	43	10	text	text	NOUN
ajst-12262	43	11	words	word	NOUN
ajst-12262	43	12	into	into	ADP
ajst-12262	43	13	highdimensional	highdimensional	ADJ
ajst-12262	43	14	real	real	ADJ
ajst-12262	43	15	number	number	NOUN
ajst-12262	43	16	vectors	vector	NOUN
ajst-12262	43	17	with	with	ADP
ajst-12262	43	18	certain	certain	ADJ
ajst-12262	43	19	semantic	semantic	ADJ
ajst-12262	43	20	information	information	NOUN
ajst-12262	43	21	by	by	ADP
ajst-12262	43	22	using	use	VERB
ajst-12262	43	23	cbow(continueus	cbow(continueus	ADJ
ajst-12262	43	24	bag	bag	NOUN
ajst-12262	43	25	of	of	ADP
ajst-12262	43	26	words	word	NOUN
ajst-12262	43	27	)	)	PUNCT
ajst-12262	43	28	or	or	CCONJ
ajst-12262	43	29	skip	skip	VERB
ajst-12262	43	30	-	-	PUNCT
ajst-12262	43	31	gram	gram	NOUN
ajst-12262	44	1	[	[	X
ajst-12262	44	2	12	12	NUM
ajst-12262	44	3	-	-	SYM
ajst-12262	44	4	13	13	NUM
ajst-12262	44	5	]	]	PUNCT
ajst-12262	44	6	.	.	PUNCT
ajst-12262	45	1	in	in	ADP
ajst-12262	45	2	this	this	DET
ajst-12262	45	3	paper	paper	NOUN
ajst-12262	45	4	,	,	PUNCT
ajst-12262	45	5	the	the	DET
ajst-12262	45	6	skip	skip	NOUN
ajst-12262	45	7	-	-	PUNCT
ajst-12262	45	8	gram	gram	NOUN
ajst-12262	45	9	model	model	NOUN
ajst-12262	45	10	of	of	ADP
ajst-12262	45	11	word2vec	word2vec	PRON
ajst-12262	45	12	is	be	AUX
ajst-12262	45	13	selected	select	VERB
ajst-12262	45	14	to	to	PART
ajst-12262	45	15	vectorize	vectorize	VERB
ajst-12262	45	16	the	the	DET
ajst-12262	45	17	text	text	NOUN
ajst-12262	45	18	vocabulary	vocabulary	NOUN
ajst-12262	45	19	.	.	PUNCT
ajst-12262	46	1	bi	bi	ADJ
ajst-12262	46	2	-	-	ADJ
ajst-12262	46	3	lstm	lstm	ADJ
ajst-12262	46	4	model	model	NOUN
ajst-12262	46	5	based	base	VERB
ajst-12262	46	6	on	on	ADP
ajst-12262	46	7	attention	attention	NOUN
ajst-12262	46	8	mechanism	mechanism	NOUN
ajst-12262	46	9	is	be	AUX
ajst-12262	46	10	a	a	DET
ajst-12262	46	11	baseline	baseline	ADJ
ajst-12262	46	12	model	model	NOUN
ajst-12262	46	13	in	in	ADP
ajst-12262	46	14	this	this	DET
ajst-12262	46	15	chapter	chapter	NOUN
ajst-12262	46	16	,	,	PUNCT
ajst-12262	46	17	which	which	PRON
ajst-12262	46	18	is	be	AUX
ajst-12262	46	19	developed	develop	VERB
ajst-12262	46	20	by	by	ADP
ajst-12262	46	21	subsequent	subsequent	ADJ
ajst-12262	46	22	improvements	improvement	NOUN
ajst-12262	46	23	.	.	PUNCT
ajst-12262	47	1	the	the	DET
ajst-12262	47	2	input	input	NOUN
ajst-12262	47	3	layer	layer	NOUN
ajst-12262	47	4	is	be	AUX
ajst-12262	47	5	responsible	responsible	ADJ
ajst-12262	47	6	for	for	ADP
ajst-12262	47	7	converting	convert	VERB
ajst-12262	47	8	the	the	DET
ajst-12262	47	9	input	input	NOUN
ajst-12262	47	10	sentence	sentence	NOUN
ajst-12262	47	11	and	and	CCONJ
ajst-12262	47	12	given	give	VERB
ajst-12262	47	13	aspect	aspect	NOUN
ajst-12262	47	14	into	into	ADP
ajst-12262	47	15	vector	vector	NOUN
ajst-12262	47	16	representation	representation	NOUN
ajst-12262	47	17	,	,	PUNCT
ajst-12262	47	18	and	and	CCONJ
ajst-12262	47	19	then	then	ADV
ajst-12262	47	20	the	the	DET
ajst-12262	47	21	bi	bi	ADJ
ajst-12262	47	22	lstm	lstm	NOUN
ajst-12262	47	23	layer	layer	NOUN
ajst-12262	47	24	constructs	construct	VERB
ajst-12262	47	25	aspect	aspect	VERB
ajst-12262	47	26	related	relate	VERB
ajst-12262	47	27	memory	memory	NOUN
ajst-12262	47	28	matrices	matrix	NOUN
ajst-12262	47	29	from	from	ADP
ajst-12262	47	30	the	the	DET
ajst-12262	47	31	word	word	NOUN
ajst-12262	47	32	vector	vector	NOUN
ajst-12262	47	33	sequence	sequence	NOUN
ajst-12262	47	34	of	of	ADP
ajst-12262	47	35	the	the	DET
ajst-12262	47	36	sentence	sentence	NOUN
ajst-12262	47	37	.	.	PUNCT
ajst-12262	48	1	the	the	DET
ajst-12262	48	2	attention	attention	NOUN
ajst-12262	48	3	layer	layer	NOUN
ajst-12262	48	4	then	then	ADV
ajst-12262	48	5	captures	capture	VERB
ajst-12262	48	6	the	the	DET
ajst-12262	48	7	information	information	NOUN
ajst-12262	48	8	that	that	PRON
ajst-12262	48	9	needs	need	VERB
ajst-12262	48	10	to	to	PART
ajst-12262	48	11	be	be	AUX
ajst-12262	48	12	focused	focus	VERB
ajst-12262	48	13	on	on	ADP
ajst-12262	48	14	the	the	DET
ajst-12262	48	15	given	give	VERB
ajst-12262	48	16	aspect	aspect	NOUN
ajst-12262	48	17	from	from	ADP
ajst-12262	48	18	the	the	DET
ajst-12262	48	19	memory	memory	NOUN
ajst-12262	48	20	matrix	matrix	NOUN
ajst-12262	48	21	output	output	NOUN
ajst-12262	48	22	by	by	ADP
ajst-12262	48	23	the	the	DET
ajst-12262	48	24	bi	bi	ADJ
ajst-12262	48	25	lstm	lstm	PROPN
ajst-12262	48	26	layer	layer	NOUN
ajst-12262	48	27	.	.	PUNCT
ajst-12262	49	1	the	the	DET
ajst-12262	49	2	classification	classification	NOUN
ajst-12262	49	3	layer	layer	NOUN
ajst-12262	49	4	consists	consist	VERB
ajst-12262	49	5	of	of	ADP
ajst-12262	49	6	a	a	DET
ajst-12262	49	7	multi	multi	ADJ
ajst-12262	49	8	-	-	ADJ
ajst-12262	49	9	layer	layer	ADJ
ajst-12262	49	10	perceptron	perceptron	NOUN
ajst-12262	49	11	and	and	CCONJ
ajst-12262	49	12	a	a	DET
ajst-12262	49	13	soflmax	soflmax	NOUN
ajst-12262	49	14	classifier	classifier	NOUN
ajst-12262	49	15	.	.	PUNCT
ajst-12262	50	1	attention	attention	NOUN
ajst-12262	50	2	calculates	calculate	VERB
ajst-12262	50	3			PROPN
ajst-12262	50	4	npppp	npppp	NOUN
ajst-12262	50	5	,	,	PUNCT
ajst-12262	50	6	,	,	PUNCT
ajst-12262	50	7	,	,	PUNCT
ajst-12262	50	8	21	21	NUM
ajst-12262	50	9	�	�	NOUN
ajst-12262	50	10			NUM
ajst-12262	50	11	for	for	ADP
ajst-12262	50	12	each	each	DET
ajst-12262	50	13	hidden	hide	VERB
ajst-12262	50	14	layer	layer	NOUN
ajst-12262	50	15	vector	vector	NOUN
ajst-12262	50	16	,	,	PUNCT
ajst-12262	50	17	and	and	CCONJ
ajst-12262	50	18	the	the	DET
ajst-12262	50	19	attention	attention	NOUN
ajst-12262	50	20	weight	weight	NOUN
ajst-12262	50	21	can	can	AUX
ajst-12262	50	22	be	be	AUX
ajst-12262	50	23	interpreted	interpret	VERB
ajst-12262	50	24	as	as	ADP
ajst-12262	50	25	the	the	DET
ajst-12262	50	26	probability	probability	NOUN
ajst-12262	50	27	that	that	SCONJ
ajst-12262	50	28	the	the	DET
ajst-12262	50	29	correct	correct	ADJ
ajst-12262	50	30	word	word	NOUN
ajst-12262	50	31	should	should	AUX
ajst-12262	50	32	be	be	AUX
ajst-12262	50	33	paid	pay	VERB
ajst-12262	50	34	attention	attention	NOUN
ajst-12262	50	35	to	to	ADP
ajst-12262	50	36	when	when	SCONJ
ajst-12262	50	37	judging	judge	VERB
ajst-12262	50	38	the	the	DET
ajst-12262	50	39	emotional	emotional	ADJ
ajst-12262	50	40	polarity	polarity	NOUN
ajst-12262	50	41	.	.	PUNCT
ajst-12262	51	1	the	the	DET
ajst-12262	51	2	calculation	calculation	NOUN
ajst-12262	51	3	process	process	NOUN
ajst-12262	51	4	of	of	ADP
ajst-12262	51	5	attention	attention	NOUN
ajst-12262	51	6	can	can	AUX
ajst-12262	51	7	be	be	AUX
ajst-12262	51	8	expressed	express	VERB
ajst-12262	51	9	as	as	ADP
ajst-12262	51	10	the	the	DET
ajst-12262	51	11	following	follow	VERB
ajst-12262	51	12	formula	formula	NOUN
ajst-12262	51	13	:	:	PUNCT
ajst-12262	51	14			PROPN
ajst-12262	51	15	st	st	PROPN
ajst-12262	51	16	ii	ii	PROPN
ajst-12262	51	17	twhd	twhd	PROPN
ajst-12262	51	18	,	,	PUNCT
ajst-12262	51	19	,	,	PUNCT
ajst-12262	51	20	tanh	tanh	PROPN
ajst-12262	51	21	(	(	PUNCT
ajst-12262	51	22	1	1	X
ajst-12262	51	23	)	)	PUNCT
ajst-12262	51	24			NOUN
ajst-12262	51	25			SYM
ajst-12262	51	26			NOUN
ajst-12262	51	27			PROPN
ajst-12262	51	28			PRON
ajst-12262	51	29			PROPN
ajst-12262	51	30	n	n	PRON
ajst-12262	51	31	j	j	PROPN
ajst-12262	52	1	j	j	PROPN
ajst-12262	52	2	j	j	PROPN
ajst-12262	53	1	i	i	INTJ
ajst-12262	53	2	d	d	PROPN
ajst-12262	54	1	d	d	X
ajst-12262	54	2	p	p	NOUN
ajst-12262	54	3	1	1	NUM
ajst-12262	54	4	exp	exp	NOUN
ajst-12262	54	5	exp	exp	NOUN
ajst-12262	54	6	(	(	PUNCT
ajst-12262	54	7	2	2	NUM
ajst-12262	54	8	)	)	PUNCT
ajst-12262	54	9	where	where	SCONJ
ajst-12262	54	10	st	st	PROPN
ajst-12262	54	11	is	be	AUX
ajst-12262	54	12	that	that	PRON
ajst-12262	54	13	vector	vector	NOUN
ajst-12262	54	14	representation	representation	NOUN
ajst-12262	54	15	of	of	ADP
ajst-12262	54	16	the	the	DET
ajst-12262	54	17	target	target	NOUN
ajst-12262	54	18	word	word	NOUN
ajst-12262	54	19	and	and	CCONJ
ajst-12262	54	20	w	w	NOUN
ajst-12262	54	21	is	be	AUX
ajst-12262	54	22	the	the	DET
ajst-12262	54	23	train	train	NOUN
ajst-12262	54	24	parameter	parameter	NOUN
ajst-12262	54	25	matrix	matrix	NOUN
ajst-12262	54	26	.	.	PUNCT
ajst-12262	55	1	weighting	weight	VERB
ajst-12262	55	2	the	the	DET
ajst-12262	55	3	output	output	NOUN
ajst-12262	55	4	of	of	ADP
ajst-12262	55	5	hidden	hide	VERB
ajst-12262	55	6	layer	layer	NOUN
ajst-12262	55	7	with	with	ADP
ajst-12262	55	8	attention	attention	NOUN
ajst-12262	55	9	weight	weight	NOUN
ajst-12262	55	10	:	:	PUNCT
ajst-12262	55	11			X
ajst-12262	55	12			NOUN
ajst-12262	56	1			NOUN
ajst-12262	56	2	n	n	INTJ
ajst-12262	57	1	i	i	PRON
ajst-12262	57	2	iis	iis	PROPN
ajst-12262	57	3	hpz	hpz	NOUN
ajst-12262	57	4	1	1	NUM
ajst-12262	57	5	(	(	PUNCT
ajst-12262	57	6	3	3	NUM
ajst-12262	57	7	)	)	PUNCT
ajst-12262	57	8	finally	finally	ADV
ajst-12262	57	9	,	,	PUNCT
ajst-12262	57	10	the	the	DET
ajst-12262	57	11	output	output	NOUN
ajst-12262	57	12	of	of	ADP
ajst-12262	57	13	the	the	DET
ajst-12262	57	14	attention	attention	NOUN
ajst-12262	57	15	layer	layer	NOUN
ajst-12262	57	16	passes	pass	VERB
ajst-12262	57	17	through	through	ADP
ajst-12262	57	18	the	the	DET
ajst-12262	57	19	full	full	ADJ
ajst-12262	57	20	connection	connection	NOUN
ajst-12262	57	21	layer	layer	NOUN
ajst-12262	57	22	to	to	PART
ajst-12262	57	23	get	get	VERB
ajst-12262	57	24	the	the	DET
ajst-12262	57	25	final	final	ADJ
ajst-12262	57	26	emotion	emotion	NOUN
ajst-12262	57	27	classification	classification	NOUN
ajst-12262	57	28	result	result	NOUN
ajst-12262	57	29	.	.	PUNCT
ajst-12262	58	1	in	in	ADP
ajst-12262	58	2	the	the	DET
ajst-12262	58	3	task	task	NOUN
ajst-12262	58	4	of	of	ADP
ajst-12262	58	5	text	text	NOUN
ajst-12262	58	6	classification	classification	NOUN
ajst-12262	58	7	of	of	ADP
ajst-12262	58	8	takeaway	takeaway	NOUN
ajst-12262	58	9	evaluation	evaluation	NOUN
ajst-12262	58	10	,	,	PUNCT
ajst-12262	58	11	because	because	SCONJ
ajst-12262	58	12	takeaway	takeaway	NOUN
ajst-12262	58	13	evaluation	evaluation	NOUN
ajst-12262	58	14	text	text	NOUN
ajst-12262	58	15	has	have	VERB
ajst-12262	58	16	the	the	DET
ajst-12262	58	17	characteristics	characteristic	NOUN
ajst-12262	58	18	of	of	ADP
ajst-12262	58	19	short	short	ADJ
ajst-12262	58	20	length	length	NOUN
ajst-12262	58	21	,	,	PUNCT
ajst-12262	58	22	compact	compact	ADJ
ajst-12262	58	23	structure	structure	NOUN
ajst-12262	58	24	and	and	CCONJ
ajst-12262	58	25	clear	clear	ADJ
ajst-12262	58	26	meaning	meaning	NOUN
ajst-12262	58	27	,	,	PUNCT
ajst-12262	58	28	cnn	cnn	PROPN
ajst-12262	58	29	can	can	AUX
ajst-12262	58	30	be	be	AUX
ajst-12262	58	31	used	use	VERB
ajst-12262	58	32	to	to	PART
ajst-12262	58	33	classify	classify	VERB
ajst-12262	58	34	emotions	emotion	NOUN
ajst-12262	58	35	.	.	PUNCT
ajst-12262	59	1	through	through	ADP
ajst-12262	59	2	the	the	DET
ajst-12262	59	3	calculation	calculation	NOUN
ajst-12262	59	4	of	of	ADP
ajst-12262	59	5	different	different	ADJ
ajst-12262	59	6	kernel	kernel	NOUN
ajst-12262	59	7	functions	function	NOUN
ajst-12262	59	8	in	in	ADP
ajst-12262	59	9	convolution	convolution	NOUN
ajst-12262	59	10	layer	layer	NOUN
ajst-12262	59	11	,	,	PUNCT
ajst-12262	59	12	multiple	multiple	ADJ
ajst-12262	59	13	feature	feature	NOUN
ajst-12262	59	14	maps	map	NOUN
ajst-12262	59	15	can	can	AUX
ajst-12262	59	16	be	be	AUX
ajst-12262	59	17	obtained	obtain	VERB
ajst-12262	59	18	,	,	PUNCT
ajst-12262	59	19	and	and	CCONJ
ajst-12262	59	20	then	then	ADV
ajst-12262	59	21	different	different	ADJ
ajst-12262	59	22	features	feature	NOUN
ajst-12262	59	23	in	in	ADP
ajst-12262	59	24	the	the	DET
ajst-12262	59	25	text	text	NOUN
ajst-12262	59	26	can	can	AUX
ajst-12262	59	27	be	be	AUX
ajst-12262	59	28	mined	mine	VERB
ajst-12262	59	29	.	.	PUNCT
ajst-12262	60	1	the	the	DET
ajst-12262	60	2	basic	basic	ADJ
ajst-12262	60	3	unit	unit	NOUN
ajst-12262	60	4	of	of	ADP
ajst-12262	60	5	the	the	DET
ajst-12262	60	6	convolution	convolution	NOUN
ajst-12262	60	7	layer	layer	NOUN
ajst-12262	60	8	is	be	AUX
ajst-12262	60	9	n	n	ADV
ajst-12262	60	10	-	-	PUNCT
ajst-12262	60	11	gram	gram	NOUN
ajst-12262	60	12	,	,	PUNCT
ajst-12262	60	13	and	and	CCONJ
ajst-12262	60	14	if	if	SCONJ
ajst-12262	60	15	the	the	DET
ajst-12262	60	16	window	window	NOUN
ajst-12262	60	17	size	size	NOUN
ajst-12262	60	18	is	be	AUX
ajst-12262	60	19	d	d	NOUN
ajst-12262	60	20	,	,	PUNCT
ajst-12262	60	21	the	the	DET
ajst-12262	60	22	word	word	NOUN
ajst-12262	60	23	vectors	vector	NOUN
ajst-12262	60	24	of	of	ADP
ajst-12262	60	25	consecutive	consecutive	ADJ
ajst-12262	60	26	d	d	NOUN
ajst-12262	60	27	words	word	NOUN
ajst-12262	60	28	can	can	AUX
ajst-12262	60	29	form	form	VERB
ajst-12262	60	30	an	an	DET
ajst-12262	60	31	n	n	CCONJ
ajst-12262	60	32	-	-	PUNCT
ajst-12262	60	33	gram	gram	NOUN
ajst-12262	60	34	vector	vector	NOUN
ajst-12262	60	35	ic	ic	PROPN
ajst-12262	60	36	,	,	PUNCT
ajst-12262	60	37	as	as	SCONJ
ajst-12262	60	38	shown	show	VERB
ajst-12262	60	39	in	in	ADP
ajst-12262	60	40	formula	formula	NOUN
ajst-12262	60	41	(	(	PUNCT
ajst-12262	60	42	4	4	NUM
ajst-12262	60	43	)	)	PUNCT
ajst-12262	60	44	.	.	PUNCT
ajst-12262	61	1	11	11	NUM
ajst-12262	61	2			NOUN
ajst-12262	61	3			NOUN
ajst-12262	61	4	diiii	diiii	NOUN
ajst-12262	61	5	vvvc	vvvc	X
ajst-12262	61	6	�	�	PROPN
ajst-12262	61	7	(	(	PUNCT
ajst-12262	61	8	4	4	NUM
ajst-12262	61	9	)	)	PUNCT
ajst-12262	61	10	where	where	SCONJ
ajst-12262	61	11			PROPN
ajst-12262	61	12	is	be	AUX
ajst-12262	61	13	the	the	DET
ajst-12262	61	14	join	join	NOUN
ajst-12262	61	15	operation	operation	NOUN
ajst-12262	61	16	,	,	PUNCT
ajst-12262	61	17	which	which	PRON
ajst-12262	61	18	is	be	AUX
ajst-12262	61	19	used	use	VERB
ajst-12262	61	20	to	to	PART
ajst-12262	61	21	join	join	VERB
ajst-12262	61	22	the	the	DET
ajst-12262	61	23	word	word	NOUN
ajst-12262	61	24	vectors	vector	NOUN
ajst-12262	61	25	from	from	ADP
ajst-12262	61	26	head	head	NOUN
ajst-12262	61	27	to	to	ADP
ajst-12262	61	28	tail	tail	NOUN
ajst-12262	61	29	.	.	PUNCT
ajst-12262	62	1	the	the	DET
ajst-12262	62	2	purpose	purpose	NOUN
ajst-12262	62	3	of	of	ADP
ajst-12262	62	4	convolution	convolution	NOUN
ajst-12262	62	5	layer	layer	NOUN
ajst-12262	62	6	is	be	AUX
ajst-12262	62	7	to	to	PART
ajst-12262	62	8	extract	extract	VERB
ajst-12262	62	9	the	the	DET
ajst-12262	62	10	semantic	semantic	ADJ
ajst-12262	62	11	features	feature	NOUN
ajst-12262	62	12	of	of	ADP
ajst-12262	62	13	sentences	sentence	NOUN
ajst-12262	62	14	,	,	PUNCT
ajst-12262	62	15	and	and	CCONJ
ajst-12262	62	16	each	each	DET
ajst-12262	62	17	convolution	convolution	NOUN
ajst-12262	62	18	kernel	kernel	NOUN
ajst-12262	62	19	should	should	AUX
ajst-12262	62	20	extract	extract	VERB
ajst-12262	62	21	some	some	DET
ajst-12262	62	22	features	feature	NOUN
ajst-12262	62	23	.	.	PUNCT
ajst-12262	63	1	in	in	ADP
ajst-12262	63	2	this	this	DET
ajst-12262	63	3	paper	paper	NOUN
ajst-12262	63	4	,	,	PUNCT
ajst-12262	63	5	the	the	DET
ajst-12262	63	6	number	number	NOUN
ajst-12262	63	7	of	of	ADP
ajst-12262	63	8	convolution	convolution	NOUN
ajst-12262	63	9	kernels	kernel	NOUN
ajst-12262	63	10	is	be	AUX
ajst-12262	63	11	set	set	VERB
ajst-12262	63	12	to	to	ADP
ajst-12262	63	13	128	128	NUM
ajst-12262	63	14	.	.	PUNCT
ajst-12262	64	1	for	for	ADP
ajst-12262	64	2	each	each	DET
ajst-12262	64	3	sentence	sentence	NOUN
ajst-12262	64	4	matrix	matrix	NOUN
ajst-12262	64	5	z	z	NOUN
ajst-12262	64	6	172	172	NUM
ajst-12262	64	7	output	output	NOUN
ajst-12262	64	8	by	by	ADP
ajst-12262	64	9	embedding	embed	VERB
ajst-12262	64	10	layer	layer	NOUN
ajst-12262	64	11	,	,	PUNCT
ajst-12262	64	12	convolution	convolution	NOUN
ajst-12262	64	13	operation	operation	NOUN
ajst-12262	64	14	is	be	AUX
ajst-12262	64	15	performed	perform	VERB
ajst-12262	64	16	:	:	PUNCT
ajst-12262	65	1			NOUN
ajst-12262	65	2	bwzfs	bwzfs	ADJ
ajst-12262	65	3			ADJ
ajst-12262	65	4	(	(	PUNCT
ajst-12262	65	5	5	5	NUM
ajst-12262	65	6	)	)	PUNCT
ajst-12262	65	7	where	where	SCONJ
ajst-12262	65	8	:	:	PUNCT
ajst-12262	65	9	s	s	X
ajst-12262	65	10	represents	represent	VERB
ajst-12262	65	11	the	the	DET
ajst-12262	65	12	feature	feature	NOUN
ajst-12262	65	13	matrix	matrix	NOUN
ajst-12262	65	14	extracted	extract	VERB
ajst-12262	65	15	by	by	ADP
ajst-12262	65	16	convolution	convolution	NOUN
ajst-12262	65	17	operation	operation	NOUN
ajst-12262	65	18	;	;	PUNCT
ajst-12262	65	19	weight	weight	NOUN
ajst-12262	65	20	matrixw	matrixw	VERB
ajst-12262	65	21	and	and	CCONJ
ajst-12262	65	22	bias	bias	NOUN
ajst-12262	65	23	vectorb	vectorb	NOUN
ajst-12262	65	24	are	be	AUX
ajst-12262	65	25	the	the	DET
ajst-12262	65	26	learning	learn	VERB
ajst-12262	65	27	parameters	parameter	NOUN
ajst-12262	65	28	of	of	ADP
ajst-12262	65	29	the	the	DET
ajst-12262	65	30	network	network	NOUN
ajst-12262	65	31	.	.	PUNCT
ajst-12262	66	1	then	then	ADV
ajst-12262	66	2	,	,	PUNCT
ajst-12262	66	3	the	the	DET
ajst-12262	66	4	extracted	extract	VERB
ajst-12262	66	5	features	feature	NOUN
ajst-12262	66	6	are	be	AUX
ajst-12262	66	7	transferred	transfer	VERB
ajst-12262	66	8	to	to	ADP
ajst-12262	66	9	the	the	DET
ajst-12262	66	10	library	library	NOUN
ajst-12262	66	11	,	,	PUNCT
ajst-12262	66	12	and	and	CCONJ
ajst-12262	66	13	all	all	DET
ajst-12262	66	14	the	the	DET
ajst-12262	66	15	features	feature	NOUN
ajst-12262	66	16	are	be	AUX
ajst-12262	66	17	merged	merge	VERB
ajst-12262	66	18	in	in	ADP
ajst-12262	66	19	the	the	DET
ajst-12262	66	20	library	library	NOUN
ajst-12262	66	21	,	,	PUNCT
ajst-12262	66	22	thus	thus	ADV
ajst-12262	66	23	simplifying	simplify	VERB
ajst-12262	66	24	the	the	DET
ajst-12262	66	25	representation	representation	NOUN
ajst-12262	66	26	of	of	ADP
ajst-12262	66	27	features	feature	NOUN
ajst-12262	66	28	.	.	PUNCT
ajst-12262	67	1	on	on	ADP
ajst-12262	67	2	this	this	DET
ajst-12262	67	3	basis	basis	NOUN
ajst-12262	67	4	,	,	PUNCT
ajst-12262	67	5	a	a	DET
ajst-12262	67	6	clustering	cluster	VERB
ajst-12262	67	7	algorithm	algorithm	NOUN
ajst-12262	67	8	based	base	VERB
ajst-12262	67	9	on	on	ADP
ajst-12262	67	10	k	k	PROPN
ajst-12262	67	11	-	-	PUNCT
ajst-12262	67	12	max	max	PROPN
ajst-12262	67	13	function	function	NOUN
ajst-12262	67	14	is	be	AUX
ajst-12262	67	15	proposed	propose	VERB
ajst-12262	67	16	,	,	PUNCT
ajst-12262	67	17	and	and	CCONJ
ajst-12262	67	18	the	the	DET
ajst-12262	67	19	semantic	semantic	ADJ
ajst-12262	67	20	information	information	NOUN
ajst-12262	67	21	represented	represent	VERB
ajst-12262	67	22	by	by	ADP
ajst-12262	67	23	each	each	DET
ajst-12262	67	24	filter	filter	NOUN
ajst-12262	67	25	is	be	AUX
ajst-12262	67	26	expressed	express	VERB
ajst-12262	67	27	by	by	ADP
ajst-12262	67	28	selecting	select	VERB
ajst-12262	67	29	the	the	DET
ajst-12262	67	30	first	first	ADJ
ajst-12262	67	31	k	k	PROPN
ajst-12262	67	32	maxima	maxima	NOUN
ajst-12262	67	33	of	of	ADP
ajst-12262	67	34	each	each	DET
ajst-12262	67	35	filter	filter	NOUN
ajst-12262	67	36	.	.	PUNCT
ajst-12262	68	1	in	in	ADP
ajst-12262	68	2	the	the	DET
ajst-12262	68	3	convolution	convolution	NOUN
ajst-12262	68	4	layer	layer	NOUN
ajst-12262	68	5	,	,	PUNCT
ajst-12262	68	6	the	the	DET
ajst-12262	68	7	phrase	phrase	NOUN
ajst-12262	68	8	is	be	AUX
ajst-12262	68	9	convolved	convolve	VERB
ajst-12262	68	10	,	,	PUNCT
ajst-12262	68	11	pooled	pool	VERB
ajst-12262	68	12	and	and	CCONJ
ajst-12262	68	13	feature	feature	NOUN
ajst-12262	68	14	extracted	extract	VERB
ajst-12262	68	15	to	to	PART
ajst-12262	68	16	obtain	obtain	VERB
ajst-12262	68	17	generalized	generalized	ADJ
ajst-12262	68	18	binary	binary	ADJ
ajst-12262	68	19	and	and	CCONJ
ajst-12262	68	20	ternary	ternary	ADJ
ajst-12262	68	21	feature	feature	NOUN
ajst-12262	68	22	vectors	vector	NOUN
ajst-12262	68	23	.	.	PUNCT
ajst-12262	69	1	on	on	ADP
ajst-12262	69	2	this	this	DET
ajst-12262	69	3	basis	basis	NOUN
ajst-12262	69	4	,	,	PUNCT
ajst-12262	69	5	the	the	DET
ajst-12262	69	6	two	two	NUM
ajst-12262	69	7	types	type	NOUN
ajst-12262	69	8	of	of	ADP
ajst-12262	69	9	feature	feature	NOUN
ajst-12262	69	10	vectors	vector	NOUN
ajst-12262	69	11	are	be	AUX
ajst-12262	69	12	fused	fuse	VERB
ajst-12262	69	13	to	to	PART
ajst-12262	69	14	form	form	VERB
ajst-12262	69	15	a	a	DET
ajst-12262	69	16	least	least	ADJ
ajst-12262	69	17	squares	square	NOUN
ajst-12262	69	18	model	model	NOUN
ajst-12262	69	19	.	.	PUNCT
ajst-12262	70	1	finally	finally	ADV
ajst-12262	70	2	,	,	PUNCT
ajst-12262	70	3	input	input	NOUN
ajst-12262	70	4	to	to	ADP
ajst-12262	70	5	the	the	DET
ajst-12262	70	6	full	full	ADJ
ajst-12262	70	7	connection	connection	NOUN
ajst-12262	70	8	layer	layer	NOUN
ajst-12262	70	9	for	for	ADP
ajst-12262	70	10	emotional	emotional	ADJ
ajst-12262	70	11	analysis	analysis	NOUN
ajst-12262	70	12	:	:	PUNCT
ajst-12262	70	13			NOUN
ajst-12262	70	14			SYM
ajst-12262	70	15			NOUN
ajst-12262	71	1	ff	ff	PROPN
ajst-12262	71	2	t	t	PROPN
ajst-12262	71	3	bqwsoftwwyp	bqwsoftwwyp	VERB
ajst-12262	71	4			ADJ
ajst-12262	71	5	max	max	PROPN
ajst-12262	71	6	,	,	PUNCT
ajst-12262	71	7	(	(	PUNCT
ajst-12262	71	8	6	6	NUM
ajst-12262	71	9	)	)	PUNCT
ajst-12262	71	10	where	where	SCONJ
ajst-12262	71	11	ff	ff	NOUN
ajst-12262	71	12	bw	bw	PROPN
ajst-12262	71	13	,	,	PUNCT
ajst-12262	71	14	is	be	AUX
ajst-12262	71	15	a	a	DET
ajst-12262	71	16	parameter	parameter	NOUN
ajst-12262	71	17	that	that	PRON
ajst-12262	71	18	can	can	AUX
ajst-12262	71	19	be	be	AUX
ajst-12262	71	20	learned	learn	VERB
ajst-12262	71	21	.	.	PUNCT
ajst-12262	72	1	4	4	X
ajst-12262	72	2	.	.	X
ajst-12262	72	3	experimental	experimental	ADJ
ajst-12262	72	4	analysis	analysis	NOUN
ajst-12262	72	5	the	the	DET
ajst-12262	72	6	experimental	experimental	ADJ
ajst-12262	72	7	take	take	VERB
ajst-12262	72	8	-	-	PUNCT
ajst-12262	72	9	away	away	NOUN
ajst-12262	72	10	data	datum	NOUN
ajst-12262	72	11	set	set	NOUN
ajst-12262	72	12	comes	come	VERB
ajst-12262	72	13	from	from	ADP
ajst-12262	72	14	the	the	DET
ajst-12262	72	15	network	network	NOUN
ajst-12262	72	16	,	,	PUNCT
ajst-12262	72	17	and	and	CCONJ
ajst-12262	72	18	it	it	PRON
ajst-12262	72	19	is	be	AUX
ajst-12262	72	20	the	the	DET
ajst-12262	72	21	take	take	VERB
ajst-12262	72	22	-	-	PUNCT
ajst-12262	72	23	away	away	NOUN
ajst-12262	72	24	review	review	NOUN
ajst-12262	72	25	data	datum	NOUN
ajst-12262	72	26	of	of	ADP
ajst-12262	72	27	a	a	DET
ajst-12262	72	28	take	take	VERB
ajst-12262	72	29	-	-	PUNCT
ajst-12262	72	30	away	away	NOUN
ajst-12262	72	31	platform	platform	NOUN
ajst-12262	72	32	.	.	PUNCT
ajst-12262	73	1	from	from	ADP
ajst-12262	73	2	it	it	PRON
ajst-12262	73	3	,	,	PUNCT
ajst-12262	73	4	7	7	NUM
ajst-12262	73	5	000	000	NUM
ajst-12262	73	6	pieces	piece	NOUN
ajst-12262	73	7	are	be	AUX
ajst-12262	73	8	randomly	randomly	ADV
ajst-12262	73	9	selected	select	VERB
ajst-12262	73	10	as	as	ADP
ajst-12262	73	11	training	training	NOUN
ajst-12262	73	12	set	set	NOUN
ajst-12262	73	13	,	,	PUNCT
ajst-12262	73	14	1500	1500	NUM
ajst-12262	73	15	pieces	piece	NOUN
ajst-12262	73	16	as	as	ADP
ajst-12262	73	17	verification	verification	NOUN
ajst-12262	73	18	set	set	VERB
ajst-12262	73	19	and	and	CCONJ
ajst-12262	73	20	1	1	NUM
ajst-12262	73	21	500	500	NUM
ajst-12262	73	22	pieces	piece	NOUN
ajst-12262	73	23	as	as	ADP
ajst-12262	73	24	test	test	NOUN
ajst-12262	73	25	set	set	NOUN
ajst-12262	73	26	,	,	PUNCT
ajst-12262	73	27	and	and	CCONJ
ajst-12262	73	28	the	the	DET
ajst-12262	73	29	training	training	NOUN
ajst-12262	73	30	set	set	NOUN
ajst-12262	73	31	,	,	PUNCT
ajst-12262	73	32	test	test	NOUN
ajst-12262	73	33	set	set	NOUN
ajst-12262	73	34	and	and	CCONJ
ajst-12262	73	35	verification	verification	NOUN
ajst-12262	73	36	set	set	NOUN
ajst-12262	73	37	are	be	AUX
ajst-12262	73	38	not	not	PART
ajst-12262	73	39	intersected	intersect	VERB
ajst-12262	73	40	.	.	PUNCT
ajst-12262	74	1	in	in	ADP
ajst-12262	74	2	order	order	NOUN
ajst-12262	74	3	to	to	PART
ajst-12262	74	4	avoid	avoid	VERB
ajst-12262	74	5	over	over	ADP
ajst-12262	74	6	-	-	PUNCT
ajst-12262	74	7	matching	matching	NOUN
ajst-12262	74	8	,	,	PUNCT
ajst-12262	74	9	a	a	DET
ajst-12262	74	10	leakage	leakage	NOUN
ajst-12262	74	11	layer	layer	NOUN
ajst-12262	74	12	is	be	AUX
ajst-12262	74	13	added	add	VERB
ajst-12262	74	14	to	to	ADP
ajst-12262	74	15	the	the	DET
ajst-12262	74	16	output	output	NOUN
ajst-12262	74	17	layer	layer	NOUN
ajst-12262	74	18	of	of	ADP
ajst-12262	74	19	lstm	lstm	NOUN
ajst-12262	74	20	,	,	PUNCT
ajst-12262	74	21	and	and	CCONJ
ajst-12262	74	22	the	the	DET
ajst-12262	74	23	leakage	leakage	NOUN
ajst-12262	74	24	probability	probability	NOUN
ajst-12262	74	25	is	be	AUX
ajst-12262	74	26	0.5	0.5	NUM
ajst-12262	74	27	.	.	PUNCT
ajst-12262	75	1	on	on	ADP
ajst-12262	75	2	this	this	DET
ajst-12262	75	3	basis	basis	NOUN
ajst-12262	75	4	,	,	PUNCT
ajst-12262	75	5	l1	l1	PROPN
ajst-12262	75	6	regularization	regularization	NOUN
ajst-12262	75	7	and	and	CCONJ
ajst-12262	75	8	l2	l2	NOUN
ajst-12262	75	9	regularization	regularization	NOUN
ajst-12262	75	10	are	be	AUX
ajst-12262	75	11	adopted	adopt	VERB
ajst-12262	75	12	,	,	PUNCT
ajst-12262	75	13	and	and	CCONJ
ajst-12262	75	14	the	the	DET
ajst-12262	75	15	parameter	parameter	NOUN
ajst-12262	75	16	is	be	AUX
ajst-12262	75	17	set	set	VERB
ajst-12262	75	18	to	to	ADP
ajst-12262	75	19	1	1	NUM
ajst-12262	75	20	.	.	PUNCT
ajst-12262	76	1	the	the	DET
ajst-12262	76	2	initial	initial	ADJ
ajst-12262	76	3	value	value	NOUN
ajst-12262	76	4	of	of	ADP
ajst-12262	76	5	the	the	DET
ajst-12262	76	6	learning	learning	NOUN
ajst-12262	76	7	rate	rate	NOUN
ajst-12262	76	8	is	be	AUX
ajst-12262	76	9	set	set	VERB
ajst-12262	76	10	at	at	ADP
ajst-12262	76	11	0.001	0.001	NUM
ajst-12262	76	12	and	and	CCONJ
ajst-12262	76	13	0.9	0.9	NUM
ajst-12262	76	14	.	.	PUNCT
ajst-12262	77	1	in	in	ADP
ajst-12262	77	2	the	the	DET
ajst-12262	77	3	learning	learning	NOUN
ajst-12262	77	4	stage	stage	NOUN
ajst-12262	77	5	,	,	PUNCT
ajst-12262	77	6	an	an	DET
ajst-12262	77	7	"	"	PUNCT
ajst-12262	77	8	early	early	ADJ
ajst-12262	77	9	termination	termination	NOUN
ajst-12262	77	10	"	"	PUNCT
ajst-12262	77	11	mechanism	mechanism	NOUN
ajst-12262	77	12	is	be	AUX
ajst-12262	77	13	introduced	introduce	VERB
ajst-12262	77	14	,	,	PUNCT
ajst-12262	77	15	that	that	ADV
ajst-12262	77	16	is	is	ADV
ajst-12262	77	17	,	,	PUNCT
ajst-12262	77	18	once	once	ADV
ajst-12262	77	19	the	the	DET
ajst-12262	77	20	performance	performance	NOUN
ajst-12262	77	21	of	of	ADP
ajst-12262	77	22	the	the	DET
ajst-12262	77	23	test	test	NOUN
ajst-12262	77	24	sample	sample	NOUN
ajst-12262	77	25	can	can	AUX
ajst-12262	77	26	not	not	PART
ajst-12262	77	27	be	be	AUX
ajst-12262	77	28	further	far	ADV
ajst-12262	77	29	improved	improve	VERB
ajst-12262	77	30	,	,	PUNCT
ajst-12262	77	31	the	the	DET
ajst-12262	77	32	learning	learning	NOUN
ajst-12262	77	33	is	be	AUX
ajst-12262	77	34	terminated	terminate	VERB
ajst-12262	77	35	.	.	PUNCT
ajst-12262	78	1	this	this	DET
ajst-12262	78	2	experiment	experiment	NOUN
ajst-12262	78	3	is	be	AUX
ajst-12262	78	4	written	write	VERB
ajst-12262	78	5	in	in	ADP
ajst-12262	78	6	python	python	PROPN
ajst-12262	78	7	,	,	PUNCT
ajst-12262	78	8	and	and	CCONJ
ajst-12262	78	9	the	the	DET
ajst-12262	78	10	data	datum	NOUN
ajst-12262	78	11	of	of	ADP
ajst-12262	78	12	the	the	DET
ajst-12262	78	13	input	input	NOUN
ajst-12262	78	14	model	model	NOUN
ajst-12262	78	15	is	be	AUX
ajst-12262	78	16	represented	represent	VERB
ajst-12262	78	17	by	by	ADP
ajst-12262	78	18	a	a	DET
ajst-12262	78	19	400	400	NUM
ajst-12262	78	20	-	-	PUNCT
ajst-12262	78	21	dimensional	dimensional	ADJ
ajst-12262	78	22	word	word	NOUN
ajst-12262	78	23	vector	vector	NOUN
ajst-12262	78	24	trained	train	VERB
ajst-12262	78	25	by	by	ADP
ajst-12262	78	26	word2vec	word2vec	PROPN
ajst-12262	78	27	.	.	PUNCT
ajst-12262	79	1	the	the	DET
ajst-12262	79	2	threshold	threshold	NOUN
ajst-12262	79	3	for	for	ADP
ajst-12262	79	4	judging	judge	VERB
ajst-12262	79	5	a	a	DET
ajst-12262	79	6	long	long	ADJ
ajst-12262	79	7	text	text	NOUN
ajst-12262	79	8	is	be	AUX
ajst-12262	79	9	set	set	VERB
ajst-12262	79	10	to	to	ADP
ajst-12262	79	11	100	100	NUM
ajst-12262	79	12	,	,	PUNCT
ajst-12262	79	13	and	and	CCONJ
ajst-12262	79	14	the	the	DET
ajst-12262	79	15	number	number	NOUN
ajst-12262	79	16	of	of	ADP
ajst-12262	79	17	keyword	keyword	NOUN
ajst-12262	79	18	extraction	extraction	NOUN
ajst-12262	79	19	is	be	AUX
ajst-12262	79	20	20	20	NUM
ajst-12262	79	21	.	.	PUNCT
ajst-12262	80	1	the	the	DET
ajst-12262	80	2	accuracy	accuracy	NOUN
ajst-12262	80	3	and	and	CCONJ
ajst-12262	80	4	loss	loss	NOUN
ajst-12262	80	5	changes	change	NOUN
ajst-12262	80	6	of	of	ADP
ajst-12262	80	7	the	the	DET
ajst-12262	80	8	training	training	NOUN
ajst-12262	80	9	set	set	VERB
ajst-12262	80	10	during	during	ADP
ajst-12262	80	11	the	the	DET
ajst-12262	80	12	experiment	experiment	NOUN
ajst-12262	80	13	are	be	AUX
ajst-12262	80	14	shown	show	VERB
ajst-12262	80	15	in	in	ADP
ajst-12262	80	16	figure	figure	NOUN
ajst-12262	80	17	2	2	NUM
ajst-12262	80	18	.	.	PUNCT
ajst-12262	80	19	figure	figure	NOUN
ajst-12262	80	20	2	2	NUM
ajst-12262	80	21	.	.	NOUN
ajst-12262	80	22	accuracy	accuracy	NOUN
ajst-12262	80	23	and	and	CCONJ
ajst-12262	80	24	loss	loss	NOUN
ajst-12262	80	25	change	change	NOUN
ajst-12262	80	26	of	of	ADP
ajst-12262	80	27	training	training	NOUN
ajst-12262	80	28	set	set	VERB
ajst-12262	80	29	in	in	ADP
ajst-12262	80	30	order	order	NOUN
ajst-12262	80	31	to	to	PART
ajst-12262	80	32	verify	verify	VERB
ajst-12262	80	33	the	the	DET
ajst-12262	80	34	validity	validity	NOUN
ajst-12262	80	35	of	of	ADP
ajst-12262	80	36	this	this	DET
ajst-12262	80	37	model	model	NOUN
ajst-12262	80	38	,	,	PUNCT
ajst-12262	80	39	precision	precision	NOUN
ajst-12262	80	40	,	,	PUNCT
ajst-12262	80	41	recall	recall	NOUN
ajst-12262	80	42	rate	rate	NOUN
ajst-12262	80	43	and	and	CCONJ
ajst-12262	80	44	f1	f1	NOUN
ajst-12262	80	45	value	value	NOUN
ajst-12262	80	46	are	be	AUX
ajst-12262	80	47	used	use	VERB
ajst-12262	80	48	for	for	ADP
ajst-12262	80	49	evaluation	evaluation	NOUN
ajst-12262	80	50	,	,	PUNCT
ajst-12262	80	51	and	and	CCONJ
ajst-12262	80	52	the	the	DET
ajst-12262	80	53	experimental	experimental	ADJ
ajst-12262	80	54	results	result	NOUN
ajst-12262	80	55	are	be	AUX
ajst-12262	80	56	compared	compare	VERB
ajst-12262	80	57	with	with	ADP
ajst-12262	80	58	the	the	DET
ajst-12262	80	59	evaluation	evaluation	NOUN
ajst-12262	80	60	results	result	NOUN
ajst-12262	80	61	of	of	ADP
ajst-12262	80	62	different	different	ADJ
ajst-12262	80	63	models	model	NOUN
ajst-12262	80	64	,	,	PUNCT
ajst-12262	80	65	as	as	SCONJ
ajst-12262	80	66	shown	show	VERB
ajst-12262	80	67	in	in	ADP
ajst-12262	80	68	table	table	NOUN
ajst-12262	81	1	1	1	NUM
ajst-12262	81	2	.	.	PUNCT
ajst-12262	81	3	table	table	NOUN
ajst-12262	81	4	1	1	NUM
ajst-12262	81	5	.	.	PUNCT
ajst-12262	82	1	performance	performance	NOUN
ajst-12262	82	2	comparison	comparison	NOUN
ajst-12262	82	3	between	between	ADP
ajst-12262	82	4	the	the	DET
ajst-12262	82	5	proposed	propose	VERB
ajst-12262	82	6	model	model	NOUN
ajst-12262	82	7	and	and	CCONJ
ajst-12262	82	8	the	the	DET
ajst-12262	82	9	hybrid	hybrid	ADJ
ajst-12262	82	10	model	model	NOUN
ajst-12262	82	11	model	model	PROPN
ajst-12262	82	12	precision	precision	PROPN
ajst-12262	82	13	recall	recall	PROPN
ajst-12262	82	14	f1	f1	PROPN
ajst-12262	82	15	lstm	lstm	PROPN
ajst-12262	82	16	0.774	0.774	NUM
ajst-12262	82	17	0.78	0.78	NUM
ajst-12262	82	18	0.714	0.714	NUM
ajst-12262	82	19	cnn	cnn	NOUN
ajst-12262	82	20	0.789	0.789	NUM
ajst-12262	82	21	0.769	0.769	NUM
ajst-12262	82	22	0.725	0.725	NUM
ajst-12262	82	23	bi	bi	NOUN
ajst-12262	82	24	-	-	ADJ
ajst-12262	82	25	lstm	lstm	ADJ
ajst-12262	82	26	0.812	0.812	NUM
ajst-12262	82	27	0.797	0.797	NUM
ajst-12262	82	28	0.81	0.81	NUM
ajst-12262	82	29	lstm+	lstm+	DET
ajst-12262	82	30	attention	attention	NOUN
ajst-12262	82	31	mechanism	mechanism	NOUN
ajst-12262	82	32	0.835	0.835	NOUN
ajst-12262	82	33	0.805	0.805	NUM
ajst-12262	82	34	0.826	0.826	NUM
ajst-12262	82	35	the	the	DET
ajst-12262	82	36	proposed	propose	VERB
ajst-12262	82	37	model	model	NOUN
ajst-12262	82	38	0.937	0.937	NUM
ajst-12262	82	39	0.896	0.896	NUM
ajst-12262	82	40	0.906	0.906	NUM
ajst-12262	82	41	it	it	PRON
ajst-12262	82	42	can	can	AUX
ajst-12262	82	43	be	be	AUX
ajst-12262	82	44	seen	see	VERB
ajst-12262	82	45	that	that	SCONJ
ajst-12262	82	46	the	the	DET
ajst-12262	82	47	accuracy	accuracy	NOUN
ajst-12262	82	48	of	of	ADP
ajst-12262	82	49	bi	bi	ADJ
ajst-12262	82	50	-	-	ADJ
ajst-12262	82	51	lstm	lstm	ADJ
ajst-12262	82	52	model	model	NOUN
ajst-12262	82	53	is	be	AUX
ajst-12262	82	54	higher	high	ADJ
ajst-12262	82	55	than	than	ADP
ajst-12262	82	56	that	that	PRON
ajst-12262	82	57	of	of	ADP
ajst-12262	82	58	simple	simple	ADJ
ajst-12262	82	59	lstm	lstm	NOUN
ajst-12262	82	60	and	and	CCONJ
ajst-12262	82	61	cnn	cnn	PROPN
ajst-12262	82	62	model	model	NOUN
ajst-12262	82	63	,	,	PUNCT
ajst-12262	82	64	while	while	SCONJ
ajst-12262	82	65	the	the	DET
ajst-12262	82	66	accuracy	accuracy	NOUN
ajst-12262	82	67	of	of	ADP
ajst-12262	82	68	lstm+	lstm+	DET
ajst-12262	82	69	attention	attention	NOUN
ajst-12262	82	70	mechanism	mechanism	NOUN
ajst-12262	82	71	model	model	NOUN
ajst-12262	82	72	is	be	AUX
ajst-12262	82	73	higher	high	ADJ
ajst-12262	82	74	than	than	ADP
ajst-12262	82	75	that	that	PRON
ajst-12262	82	76	of	of	ADP
ajst-12262	82	77	bi	bi	ADJ
ajst-12262	82	78	-	-	ADJ
ajst-12262	82	79	lstm	lstm	ADJ
ajst-12262	82	80	model	model	NOUN
ajst-12262	82	81	.	.	PUNCT
ajst-12262	83	1	the	the	DET
ajst-12262	83	2	accuracy	accuracy	NOUN
ajst-12262	83	3	of	of	ADP
ajst-12262	83	4	this	this	DET
ajst-12262	83	5	model	model	NOUN
ajst-12262	83	6	is	be	AUX
ajst-12262	83	7	the	the	DET
ajst-12262	83	8	highest	high	ADJ
ajst-12262	83	9	,	,	PUNCT
ajst-12262	83	10	with	with	ADP
ajst-12262	83	11	precision	precision	NOUN
ajst-12262	83	12	,	,	PUNCT
ajst-12262	83	13	recall	recall	NOUN
ajst-12262	83	14	and	and	CCONJ
ajst-12262	83	15	f1	f1	NOUN
ajst-12262	83	16	being	be	AUX
ajst-12262	83	17	0.937	0.937	NUM
ajst-12262	83	18	,	,	PUNCT
ajst-12262	83	19	0.896	0.896	NUM
ajst-12262	83	20	and	and	CCONJ
ajst-12262	83	21	0.906	0.906	NUM
ajst-12262	83	22	respectively	respectively	ADV
ajst-12262	83	23	.	.	PUNCT
ajst-12262	84	1	through	through	ADP
ajst-12262	84	2	horizontal	horizontal	ADJ
ajst-12262	84	3	comparison	comparison	NOUN
ajst-12262	84	4	,	,	PUNCT
ajst-12262	84	5	it	it	PRON
ajst-12262	84	6	can	can	AUX
ajst-12262	84	7	be	be	AUX
ajst-12262	84	8	found	find	VERB
ajst-12262	84	9	that	that	SCONJ
ajst-12262	84	10	the	the	DET
ajst-12262	84	11	model	model	NOUN
ajst-12262	84	12	in	in	ADP
ajst-12262	84	13	this	this	DET
ajst-12262	84	14	paper	paper	NOUN
ajst-12262	84	15	is	be	AUX
ajst-12262	84	16	outstanding	outstanding	ADJ
ajst-12262	84	17	in	in	ADP
ajst-12262	84	18	this	this	DET
ajst-12262	84	19	task	task	NOUN
ajst-12262	84	20	,	,	PUNCT
ajst-12262	84	21	and	and	CCONJ
ajst-12262	84	22	the	the	DET
ajst-12262	84	23	emotion	emotion	NOUN
ajst-12262	84	24	enhancement	enhancement	NOUN
ajst-12262	84	25	model	model	NOUN
ajst-12262	84	26	also	also	ADV
ajst-12262	84	27	improves	improve	VERB
ajst-12262	84	28	the	the	DET
ajst-12262	84	29	classification	classification	NOUN
ajst-12262	84	30	accuracy	accuracy	NOUN
ajst-12262	84	31	to	to	ADP
ajst-12262	84	32	some	some	DET
ajst-12262	84	33	extent	extent	NOUN
ajst-12262	84	34	.	.	PUNCT
ajst-12262	85	1	173	173	NUM
ajst-12262	85	2	5	5	NUM
ajst-12262	85	3	.	.	PUNCT
ajst-12262	86	1	conclusions	conclusion	NOUN
ajst-12262	86	2	massive	massive	ADJ
ajst-12262	86	3	texts	text	NOUN
ajst-12262	86	4	come	come	VERB
ajst-12262	86	5	from	from	ADP
ajst-12262	86	6	many	many	ADJ
ajst-12262	86	7	users	user	NOUN
ajst-12262	86	8	of	of	ADP
ajst-12262	86	9	the	the	DET
ajst-12262	86	10	internet	internet	NOUN
ajst-12262	86	11	,	,	PUNCT
ajst-12262	86	12	with	with	ADP
ajst-12262	86	13	various	various	ADJ
ajst-12262	86	14	forms	form	NOUN
ajst-12262	86	15	and	and	CCONJ
ajst-12262	86	16	no	no	DET
ajst-12262	86	17	fixed	fixed	ADJ
ajst-12262	86	18	format	format	NOUN
ajst-12262	86	19	,	,	PUNCT
ajst-12262	86	20	so	so	SCONJ
ajst-12262	86	21	it	it	PRON
ajst-12262	86	22	is	be	AUX
ajst-12262	86	23	difficult	difficult	ADJ
ajst-12262	86	24	to	to	PART
ajst-12262	86	25	process	process	VERB
ajst-12262	86	26	them	they	PRON
ajst-12262	86	27	by	by	ADP
ajst-12262	86	28	simple	simple	ADJ
ajst-12262	86	29	automatic	automatic	ADJ
ajst-12262	86	30	means	mean	NOUN
ajst-12262	86	31	.	.	PUNCT
ajst-12262	87	1	if	if	SCONJ
ajst-12262	87	2	you	you	PRON
ajst-12262	87	3	rely	rely	VERB
ajst-12262	87	4	on	on	ADP
ajst-12262	87	5	manual	manual	ADJ
ajst-12262	87	6	processing	processing	NOUN
ajst-12262	87	7	,	,	PUNCT
ajst-12262	87	8	there	there	PRON
ajst-12262	87	9	are	be	VERB
ajst-12262	87	10	problems	problem	NOUN
ajst-12262	87	11	such	such	ADJ
ajst-12262	87	12	as	as	ADP
ajst-12262	87	13	excessive	excessive	ADJ
ajst-12262	87	14	workload	workload	NOUN
ajst-12262	87	15	and	and	CCONJ
ajst-12262	87	16	poor	poor	ADJ
ajst-12262	87	17	real	real	ADJ
ajst-12262	87	18	-	-	PUNCT
ajst-12262	87	19	time	time	NOUN
ajst-12262	87	20	performance	performance	NOUN
ajst-12262	87	21	.	.	PUNCT
ajst-12262	88	1	in	in	ADP
ajst-12262	88	2	this	this	DET
ajst-12262	88	3	paper	paper	NOUN
ajst-12262	88	4	,	,	PUNCT
ajst-12262	88	5	the	the	DET
ajst-12262	88	6	word	word	NOUN
ajst-12262	88	7	embedding	embed	VERB
ajst-12262	88	8	nlp	nlp	NOUN
ajst-12262	88	9	is	be	AUX
ajst-12262	88	10	used	use	VERB
ajst-12262	88	11	for	for	ADP
ajst-12262	88	12	preprocessing	preprocesse	VERB
ajst-12262	88	13	in	in	ADP
ajst-12262	88	14	tensorflow	tensorflow	NOUN
ajst-12262	88	15	environment	environment	NOUN
ajst-12262	88	16	,	,	PUNCT
ajst-12262	88	17	and	and	CCONJ
ajst-12262	88	18	an	an	DET
ajst-12262	88	19	emotion	emotion	NOUN
ajst-12262	88	20	recognition	recognition	NOUN
ajst-12262	88	21	model	model	NOUN
ajst-12262	88	22	of	of	ADP
ajst-12262	88	23	takeaway	takeaway	NOUN
ajst-12262	88	24	evaluation	evaluation	NOUN
ajst-12262	88	25	text	text	NOUN
ajst-12262	88	26	based	base	VERB
ajst-12262	88	27	on	on	ADP
ajst-12262	88	28	lstm	lstm	PROPN
ajst-12262	88	29	-	-	PUNCT
ajst-12262	88	30	cnn	cnn	PROPN
ajst-12262	88	31	is	be	AUX
ajst-12262	88	32	established	establish	VERB
ajst-12262	88	33	.	.	PUNCT
ajst-12262	89	1	the	the	DET
ajst-12262	89	2	experimental	experimental	ADJ
ajst-12262	89	3	results	result	NOUN
ajst-12262	89	4	show	show	VERB
ajst-12262	89	5	that	that	SCONJ
ajst-12262	89	6	the	the	DET
ajst-12262	89	7	accuracy	accuracy	NOUN
ajst-12262	89	8	of	of	ADP
ajst-12262	89	9	this	this	DET
ajst-12262	89	10	model	model	NOUN
ajst-12262	89	11	is	be	AUX
ajst-12262	89	12	the	the	DET
ajst-12262	89	13	highest	high	ADJ
ajst-12262	89	14	,	,	PUNCT
ajst-12262	89	15	and	and	CCONJ
ajst-12262	89	16	the	the	DET
ajst-12262	89	17	precision	precision	NOUN
ajst-12262	89	18	,	,	PUNCT
ajst-12262	89	19	recall	recall	NOUN
ajst-12262	89	20	and	and	CCONJ
ajst-12262	89	21	f1	f1	NOUN
ajst-12262	89	22	are	be	AUX
ajst-12262	89	23	0.937	0.937	NUM
ajst-12262	89	24	,	,	PUNCT
ajst-12262	89	25	0.896	0.896	NUM
ajst-12262	89	26	and	and	CCONJ
ajst-12262	89	27	0.906	0.906	NUM
ajst-12262	89	28	respectively	respectively	ADV
ajst-12262	89	29	.	.	PUNCT
ajst-12262	90	1	through	through	ADP
ajst-12262	90	2	horizontal	horizontal	ADJ
ajst-12262	90	3	comparison	comparison	NOUN
ajst-12262	90	4	,	,	PUNCT
ajst-12262	90	5	it	it	PRON
ajst-12262	90	6	can	can	AUX
ajst-12262	90	7	be	be	AUX
ajst-12262	90	8	found	find	VERB
ajst-12262	90	9	that	that	SCONJ
ajst-12262	90	10	the	the	DET
ajst-12262	90	11	model	model	NOUN
ajst-12262	90	12	in	in	ADP
ajst-12262	90	13	this	this	DET
ajst-12262	90	14	paper	paper	NOUN
ajst-12262	90	15	is	be	AUX
ajst-12262	90	16	outstanding	outstanding	ADJ
ajst-12262	90	17	in	in	ADP
ajst-12262	90	18	this	this	DET
ajst-12262	90	19	task	task	NOUN
ajst-12262	90	20	,	,	PUNCT
ajst-12262	90	21	and	and	CCONJ
ajst-12262	90	22	the	the	DET
ajst-12262	90	23	emotion	emotion	NOUN
ajst-12262	90	24	enhancement	enhancement	NOUN
ajst-12262	90	25	model	model	NOUN
ajst-12262	90	26	also	also	ADV
ajst-12262	90	27	improves	improve	VERB
ajst-12262	90	28	the	the	DET
ajst-12262	90	29	classification	classification	NOUN
ajst-12262	90	30	accuracy	accuracy	NOUN
ajst-12262	90	31	to	to	ADP
ajst-12262	90	32	some	some	DET
ajst-12262	90	33	extent	extent	NOUN
ajst-12262	90	34	.	.	PUNCT
ajst-12262	91	1	references	reference	NOUN
ajst-12262	91	2	[	[	X
ajst-12262	91	3	1	1	NUM
ajst-12262	91	4	]	]	PUNCT
ajst-12262	91	5	bernhard	bernhard	PROPN
ajst-12262	91	6	,	,	PUNCT
ajst-12262	91	7	k.	k.	PROPN
ajst-12262	91	8	,	,	PUNCT
ajst-12262	91	9	suzana	suzana	PROPN
ajst-12262	91	10	,	,	PUNCT
ajst-12262	91	11	i.	i.	PROPN
ajst-12262	91	12	,	,	PUNCT
ajst-12262	91	13	mathias	mathias	PROPN
ajst-12262	91	14	,	,	PUNCT
ajst-12262	91	15	k.	k.	PROPN
ajst-12262	91	16	,	,	PUNCT
ajst-12262	91	17	stefan	stefan	PROPN
ajst-12262	91	18	,	,	PUNCT
ajst-12262	91	19	f.	f.	PROPN
ajst-12262	91	20	,	,	PUNCT
ajst-12262	91	21	&	&	CCONJ
ajst-12262	91	22	helmut	helmut	PROPN
ajst-12262	91	23	,	,	PUNCT
ajst-12262	91	24	p.	p.	NOUN
ajst-12262	91	25	(	(	PUNCT
ajst-12262	91	26	2018	2018	NUM
ajst-12262	91	27	)	)	PUNCT
ajst-12262	91	28	.	.	PUNCT
ajst-12262	92	1	deep	deep	ADJ
ajst-12262	92	2	learning	learning	NOUN
ajst-12262	92	3	for	for	ADP
ajst-12262	92	4	affective	affective	ADJ
ajst-12262	92	5	computing	computing	NOUN
ajst-12262	92	6	:	:	PUNCT
ajst-12262	92	7	text	text	NOUN
ajst-12262	92	8	-	-	PUNCT
ajst-12262	92	9	based	base	VERB
ajst-12262	92	10	emotion	emotion	NOUN
ajst-12262	92	11	recognition	recognition	NOUN
ajst-12262	92	12	in	in	ADP
ajst-12262	92	13	decision	decision	NOUN
ajst-12262	92	14	support	support	NOUN
ajst-12262	92	15	.	.	PUNCT
ajst-12262	93	1	decision	decision	NOUN
ajst-12262	93	2	support	support	NOUN
ajst-12262	93	3	systems	system	NOUN
ajst-12262	93	4	,	,	PUNCT
ajst-12262	93	5	115(10	115(10	NUM
ajst-12262	93	6	)	)	PUNCT
ajst-12262	93	7	,	,	PUNCT
ajst-12262	93	8	24	24	NUM
ajst-12262	93	9	-	-	SYM
ajst-12262	93	10	35	35	NUM
ajst-12262	93	11	.	.	PUNCT
ajst-12262	94	1	[	[	X
ajst-12262	94	2	2	2	NUM
ajst-12262	94	3	]	]	X
ajst-12262	94	4	zhang	zhang	PROPN
ajst-12262	94	5	,	,	PUNCT
ajst-12262	94	6	g.	g.	PROPN
ajst-12262	94	7	,	,	PUNCT
ajst-12262	94	8	ananiadou	ananiadou	PROPN
ajst-12262	94	9	,	,	PUNCT
ajst-12262	94	10	s.	s.	PROPN
ajst-12262	94	11	,	,	PUNCT
ajst-12262	94	12	&	&	CCONJ
ajst-12262	94	13	odbal	odbal	PROPN
ajst-12262	94	14	,	,	PUNCT
ajst-12262	94	15	s.	s.	PROPN
ajst-12262	94	16	(	(	PUNCT
ajst-12262	94	17	2022	2022	NUM
ajst-12262	94	18	)	)	PUNCT
ajst-12262	94	19	.	.	PUNCT
ajst-12262	95	1	examining	examine	VERB
ajst-12262	95	2	and	and	CCONJ
ajst-12262	95	3	mitigating	mitigate	VERB
ajst-12262	95	4	gender	gender	NOUN
ajst-12262	95	5	bias	bias	NOUN
ajst-12262	95	6	in	in	ADP
ajst-12262	95	7	text	text	NOUN
ajst-12262	95	8	emotion	emotion	NOUN
ajst-12262	95	9	detection	detection	NOUN
ajst-12262	95	10	task	task	NOUN
ajst-12262	95	11	.	.	PUNCT
ajst-12262	96	1	neurocomputing(7	neurocomputing(7	PROPN
ajst-12262	96	2	)	)	PUNCT
ajst-12262	96	3	,	,	PUNCT
ajst-12262	96	4	493	493	NUM
ajst-12262	96	5	.	.	PUNCT
ajst-12262	97	1	[	[	X
ajst-12262	97	2	3	3	NUM
ajst-12262	97	3	]	]	X
ajst-12262	97	4	ian	ian	PROPN
ajst-12262	97	5	,	,	PUNCT
ajst-12262	97	6	w.	w.	PROPN
ajst-12262	97	7	,	,	PUNCT
ajst-12262	97	8	john	john	PROPN
ajst-12262	97	9	,	,	PUNCT
ajst-12262	97	10	m.	m.	PROPN
ajst-12262	97	11	c.	c.	PROPN
ajst-12262	97	12	,	,	PUNCT
ajst-12262	97	13	vladimir	vladimir	PROPN
ajst-12262	97	14	,	,	PUNCT
ajst-12262	97	15	a.	a.	PROPN
ajst-12262	97	16	,	,	PUNCT
ajst-12262	97	17	&	&	CCONJ
ajst-12262	97	18	paul	paul	PROPN
ajst-12262	97	19	,	,	PUNCT
ajst-12262	97	20	b.	b.	PROPN
ajst-12262	97	21	(	(	PUNCT
ajst-12262	97	22	2018	2018	NUM
ajst-12262	97	23	)	)	PUNCT
ajst-12262	97	24	.	.	PUNCT
ajst-12262	98	1	a	a	DET
ajst-12262	98	2	comparison	comparison	NOUN
ajst-12262	98	3	of	of	ADP
ajst-12262	98	4	emotion	emotion	NOUN
ajst-12262	98	5	annotation	annotation	NOUN
ajst-12262	98	6	approaches	approach	NOUN
ajst-12262	98	7	for	for	ADP
ajst-12262	98	8	text	text	NOUN
ajst-12262	98	9	.	.	PUNCT
ajst-12262	99	1	information	information	NOUN
ajst-12262	99	2	,	,	PUNCT
ajst-12262	99	3	9(5	9(5	NUM
ajst-12262	99	4	)	)	PUNCT
ajst-12262	99	5	,	,	PUNCT
ajst-12262	99	6	117	117	NUM
ajst-12262	99	7	.	.	PUNCT
ajst-12262	100	1	[	[	X
ajst-12262	100	2	4	4	NUM
ajst-12262	100	3	]	]	SYM
ajst-12262	100	4	xu	xu	PROPN
ajst-12262	100	5	,	,	PUNCT
ajst-12262	100	6	q.	q.	PROPN
ajst-12262	100	7	,	,	PUNCT
ajst-12262	100	8	zhang	zhang	PROPN
ajst-12262	100	9	,	,	PUNCT
ajst-12262	100	10	c.	c.	PROPN
ajst-12262	100	11	,	,	PUNCT
ajst-12262	100	12	&	&	CCONJ
ajst-12262	100	13	sun	sun	PROPN
ajst-12262	100	14	,	,	PUNCT
ajst-12262	100	15	b.	b.	PROPN
ajst-12262	100	16	(	(	PUNCT
ajst-12262	100	17	2020	2020	NUM
ajst-12262	100	18	)	)	PUNCT
ajst-12262	100	19	.	.	PUNCT
ajst-12262	101	1	emotion	emotion	NOUN
ajst-12262	101	2	recognition	recognition	NOUN
ajst-12262	101	3	model	model	NOUN
ajst-12262	101	4	based	base	VERB
ajst-12262	101	5	on	on	ADP
ajst-12262	101	6	the	the	DET
ajst-12262	101	7	dempster	dempster	PROPN
ajst-12262	101	8	–	–	PUNCT
ajst-12262	101	9	shafer	shafer	NOUN
ajst-12262	101	10	evidence	evidence	NOUN
ajst-12262	101	11	theory	theory	NOUN
ajst-12262	101	12	.	.	PUNCT
ajst-12262	102	1	journal	journal	PROPN
ajst-12262	102	2	of	of	ADP
ajst-12262	102	3	electronic	electronic	ADJ
ajst-12262	102	4	imaging	imaging	NOUN
ajst-12262	102	5	,	,	PUNCT
ajst-12262	102	6	29(2	29(2	NUM
ajst-12262	102	7	)	)	PUNCT
ajst-12262	102	8	,	,	PUNCT
ajst-12262	102	9	1	1	X
ajst-12262	102	10	.	.	PUNCT
ajst-12262	103	1	[	[	X
ajst-12262	103	2	5	5	NUM
ajst-12262	103	3	]	]	X
ajst-12262	103	4	sourina	sourina	ADJ
ajst-12262	103	5	,	,	PUNCT
ajst-12262	103	6	o.	o.	PROPN
ajst-12262	103	7	,	,	PUNCT
ajst-12262	103	8	li	li	PROPN
ajst-12262	103	9	,	,	PUNCT
ajst-12262	103	10	l.	l.	PROPN
ajst-12262	103	11	,	,	PUNCT
ajst-12262	103	12	&	&	CCONJ
ajst-12262	103	13	pan	pan	PROPN
ajst-12262	103	14	,	,	PUNCT
ajst-12262	103	15	z.	z.	PROPN
ajst-12262	103	16	(	(	PUNCT
ajst-12262	103	17	2012	2012	NUM
ajst-12262	103	18	)	)	PUNCT
ajst-12262	103	19	.	.	PUNCT
ajst-12262	104	1	emotion	emotion	NOUN
ajst-12262	104	2	-	-	PUNCT
ajst-12262	104	3	based	base	VERB
ajst-12262	104	4	interaction	interaction	NOUN
ajst-12262	104	5	.	.	PUNCT
ajst-12262	105	1	journal	journal	PROPN
ajst-12262	105	2	on	on	ADP
ajst-12262	105	3	multimodal	multimodal	NOUN
ajst-12262	105	4	user	user	NOUN
ajst-12262	105	5	interfaces	interface	NOUN
ajst-12262	105	6	,	,	PUNCT
ajst-12262	105	7	5(1	5(1	NUM
ajst-12262	105	8	-	-	SYM
ajst-12262	105	9	2	2	NUM
ajst-12262	105	10	)	)	PUNCT
ajst-12262	105	11	,	,	PUNCT
ajst-12262	105	12	p.1	p.1	NOUN
ajst-12262	105	13	.	.	PUNCT
ajst-12262	106	1	[	[	X
ajst-12262	106	2	6	6	NUM
ajst-12262	106	3	]	]	X
ajst-12262	106	4	lee	lee	PROPN
ajst-12262	106	5	,	,	PUNCT
ajst-12262	106	6	s.	s.	PROPN
ajst-12262	106	7	y.	y.	PROPN
ajst-12262	106	8	m.	m.	PROPN
ajst-12262	106	9	,	,	PUNCT
ajst-12262	106	10	chen	chen	PROPN
ajst-12262	106	11	,	,	PUNCT
ajst-12262	106	12	y.	y.	PROPN
ajst-12262	106	13	,	,	PUNCT
ajst-12262	106	14	huang	huang	PROPN
ajst-12262	106	15	,	,	PUNCT
ajst-12262	106	16	c.	c.	PROPN
ajst-12262	106	17	r.	r.	PROPN
ajst-12262	106	18	,	,	PUNCT
ajst-12262	106	19	&	&	CCONJ
ajst-12262	106	20	li	li	PROPN
ajst-12262	106	21	,	,	PUNCT
ajst-12262	106	22	s.	s.	PROPN
ajst-12262	106	23	(	(	PUNCT
ajst-12262	106	24	2013	2013	NUM
ajst-12262	106	25	)	)	PUNCT
ajst-12262	106	26	.	.	PUNCT
ajst-12262	107	1	detecting	detect	VERB
ajst-12262	107	2	emotion	emotion	NOUN
ajst-12262	107	3	causes	cause	VERB
ajst-12262	107	4	with	with	ADP
ajst-12262	107	5	a	a	DET
ajst-12262	107	6	linguistic	linguistic	ADJ
ajst-12262	107	7	rule	rule	NOUN
ajst-12262	107	8	-	-	PUNCT
ajst-12262	107	9	based	base	VERB
ajst-12262	107	10	approach	approach	NOUN
ajst-12262	107	11	.	.	PUNCT
ajst-12262	108	1	computational	computational	ADJ
ajst-12262	108	2	intelligence	intelligence	NOUN
ajst-12262	108	3	,	,	PUNCT
ajst-12262	108	4	29(3	29(3	NUM
ajst-12262	108	5	)	)	PUNCT
ajst-12262	108	6	,	,	PUNCT
ajst-12262	108	7	390	390	NUM
ajst-12262	108	8	-	-	SYM
ajst-12262	108	9	416	416	NUM
ajst-12262	108	10	.	.	PUNCT
ajst-12262	109	1	[	[	X
ajst-12262	109	2	7	7	X
ajst-12262	109	3	]	]	SYM
ajst-12262	109	4	xing	xing	PROPN
ajst-12262	109	5	,	,	PUNCT
ajst-12262	109	6	w.	w.	PROPN
ajst-12262	109	7	u.	u.	PROPN
ajst-12262	109	8	,	,	PUNCT
ajst-12262	109	9	hai	hai	PROPN
ajst-12262	109	10	-	-	PUNCT
ajst-12262	109	11	tao	tao	PROPN
ajst-12262	109	12	,	,	PUNCT
ajst-12262	109	13	l.	l.	PROPN
ajst-12262	109	14	,	,	PUNCT
ajst-12262	109	15	&	&	CCONJ
ajst-12262	109	16	shao	shao	PROPN
ajst-12262	109	17	-	-	PUNCT
ajst-12262	109	18	jian	jian	PROPN
ajst-12262	109	19	,	,	PUNCT
ajst-12262	109	20	z.	z.	PROPN
ajst-12262	109	21	(	(	PUNCT
ajst-12262	109	22	2015	2015	NUM
ajst-12262	109	23	)	)	PUNCT
ajst-12262	109	24	.	.	PUNCT
ajst-12262	109	25	sentiment	sentiment	NOUN
ajst-12262	109	26	analysis	analysis	NOUN
ajst-12262	109	27	for	for	ADP
ajst-12262	109	28	chinese	chinese	ADJ
ajst-12262	109	29	text	text	NOUN
ajst-12262	109	30	based	base	VERB
ajst-12262	109	31	on	on	ADP
ajst-12262	109	32	emotion	emotion	NOUN
ajst-12262	109	33	degree	degree	NOUN
ajst-12262	109	34	lexicon	lexicon	NOUN
ajst-12262	109	35	and	and	CCONJ
ajst-12262	109	36	cognitive	cognitive	ADJ
ajst-12262	109	37	theories	theory	NOUN
ajst-12262	109	38	.	.	PUNCT
ajst-12262	110	1	journal	journal	PROPN
ajst-12262	110	2	of	of	ADP
ajst-12262	110	3	shanghai	shanghai	PROPN
ajst-12262	110	4	jiaotong	jiaotong	PROPN
ajst-12262	110	5	university	university	PROPN
ajst-12262	110	6	,	,	PUNCT
ajst-12262	110	7	20(1	20(1	NUM
ajst-12262	110	8	)	)	PUNCT
ajst-12262	110	9	,	,	PUNCT
ajst-12262	110	10	6	6	NUM
ajst-12262	110	11	.	.	PUNCT
ajst-12262	111	1	[	[	X
ajst-12262	111	2	8	8	NUM
ajst-12262	111	3	]	]	PUNCT
ajst-12262	111	4	bandhakavi	bandhakavi	NOUN
ajst-12262	111	5	,	,	PUNCT
ajst-12262	111	6	a.	a.	NOUN
ajst-12262	111	7	,	,	PUNCT
ajst-12262	111	8	wiratunga	wiratunga	NOUN
ajst-12262	111	9	,	,	PUNCT
ajst-12262	111	10	n.	n.	NOUN
ajst-12262	111	11	,	,	PUNCT
ajst-12262	111	12	massie	massie	PROPN
ajst-12262	111	13	,	,	PUNCT
ajst-12262	111	14	s.	s.	PROPN
ajst-12262	111	15	,	,	PUNCT
ajst-12262	111	16	&	&	CCONJ
ajst-12262	111	17	padmanabhan	padmanabhan	PROPN
ajst-12262	111	18	,	,	PUNCT
ajst-12262	111	19	d.	d.	PROPN
ajst-12262	111	20	(	(	PUNCT
ajst-12262	111	21	2017	2017	NUM
ajst-12262	111	22	)	)	PUNCT
ajst-12262	111	23	.	.	PUNCT
ajst-12262	112	1	lexicon	lexicon	ADJ
ajst-12262	112	2	generation	generation	NOUN
ajst-12262	112	3	for	for	ADP
ajst-12262	112	4	emotion	emotion	NOUN
ajst-12262	112	5	detection	detection	NOUN
ajst-12262	112	6	from	from	ADP
ajst-12262	112	7	text	text	NOUN
ajst-12262	112	8	.	.	PUNCT
ajst-12262	113	1	intelligent	intelligent	ADJ
ajst-12262	113	2	systems	system	NOUN
ajst-12262	113	3	,	,	PUNCT
ajst-12262	113	4	ieee	ieee	NOUN
ajst-12262	113	5	,	,	PUNCT
ajst-12262	113	6	32(1	32(1	NUM
ajst-12262	113	7	)	)	PUNCT
ajst-12262	113	8	,	,	PUNCT
ajst-12262	113	9	102	102	NUM
ajst-12262	113	10	-	-	SYM
ajst-12262	113	11	108	108	NUM
ajst-12262	113	12	.	.	PUNCT
ajst-12262	114	1	[	[	X
ajst-12262	114	2	9	9	NUM
ajst-12262	114	3	]	]	X
ajst-12262	114	4	yamamoto	yamamoto	NOUN
ajst-12262	114	5	,	,	PUNCT
ajst-12262	114	6	y.	y.	PROPN
ajst-12262	114	7	,	,	PUNCT
ajst-12262	114	8	niitsuma	niitsuma	PROPN
ajst-12262	114	9	,	,	PUNCT
ajst-12262	114	10	m.	m.	NOUN
ajst-12262	114	11	,	,	PUNCT
ajst-12262	114	12	&	&	CCONJ
ajst-12262	114	13	yamashita	yamashita	PROPN
ajst-12262	114	14	,	,	PUNCT
ajst-12262	114	15	y.	y.	PROPN
ajst-12262	114	16	(	(	PUNCT
ajst-12262	114	17	2016	2016	NUM
ajst-12262	114	18	)	)	PUNCT
ajst-12262	114	19	.	.	PUNCT
ajst-12262	115	1	automatic	automatic	ADJ
ajst-12262	115	2	recognition	recognition	NOUN
ajst-12262	115	3	of	of	ADP
ajst-12262	115	4	negative	negative	ADJ
ajst-12262	115	5	emotion	emotion	NOUN
ajst-12262	115	6	in	in	ADP
ajst-12262	115	7	speech	speech	NOUN
ajst-12262	115	8	using	use	VERB
ajst-12262	115	9	support	support	NOUN
ajst-12262	115	10	vector	vector	NOUN
ajst-12262	115	11	machine	machine	NOUN
ajst-12262	115	12	.	.	PUNCT
ajst-12262	116	1	the	the	DET
ajst-12262	116	2	journal	journal	NOUN
ajst-12262	116	3	of	of	ADP
ajst-12262	116	4	the	the	DET
ajst-12262	116	5	acoustical	acoustical	ADJ
ajst-12262	116	6	society	society	NOUN
ajst-12262	116	7	of	of	ADP
ajst-12262	116	8	america	america	PROPN
ajst-12262	116	9	,	,	PUNCT
ajst-12262	116	10	140(4	140(4	NUM
ajst-12262	116	11	)	)	PUNCT
ajst-12262	116	12	,	,	PUNCT
ajst-12262	116	13	3400	3400	NUM
ajst-12262	116	14	-	-	SYM
ajst-12262	116	15	3400	3400	NUM
ajst-12262	116	16	.	.	PUNCT
ajst-12262	117	1	[	[	X
ajst-12262	117	2	10	10	NUM
ajst-12262	117	3	]	]	X
ajst-12262	117	4	pais	pais	PROPN
ajst-12262	117	5	,	,	PUNCT
ajst-12262	117	6	a.	a.	PROPN
ajst-12262	117	7	l.	l.	PROPN
ajst-12262	117	8	,	,	PUNCT
ajst-12262	117	9	moga	moga	PROPN
ajst-12262	117	10	,	,	PUNCT
ajst-12262	117	11	a.	a.	PROPN
ajst-12262	117	12	s.	s.	PROPN
ajst-12262	117	13	,	,	PUNCT
ajst-12262	117	14	&	&	CCONJ
ajst-12262	117	15	buiu	buiu	PROPN
ajst-12262	117	16	,	,	PUNCT
ajst-12262	117	17	c.	c.	PROPN
ajst-12262	117	18	(	(	PUNCT
ajst-12262	117	19	2010	2010	NUM
ajst-12262	117	20	)	)	PUNCT
ajst-12262	117	21	.	.	PUNCT
ajst-12262	118	1	an	an	DET
ajst-12262	118	2	integrated	integrate	VERB
ajst-12262	118	3	framework	framework	NOUN
ajst-12262	118	4	for	for	ADP
ajst-12262	118	5	emotion	emotion	NOUN
ajst-12262	118	6	recognition	recognition	NOUN
ajst-12262	118	7	and	and	CCONJ
ajst-12262	118	8	expression	expression	NOUN
ajst-12262	118	9	using	use	VERB
ajst-12262	118	10	robot	robot	NOUN
ajst-12262	118	11	artists	artist	NOUN
ajst-12262	118	12	.	.	PUNCT
ajst-12262	119	1	icic	icic	PROPN
ajst-12262	119	2	express	express	PROPN
ajst-12262	119	3	letters	letter	NOUN
ajst-12262	119	4	,	,	PUNCT
ajst-12262	119	5	1(2	1(2	NUM
ajst-12262	119	6	)	)	PUNCT
ajst-12262	119	7	,	,	PUNCT
ajst-12262	119	8	169	169	NUM
ajst-12262	119	9	-	-	SYM
ajst-12262	119	10	174	174	NUM
ajst-12262	119	11	.	.	PUNCT
ajst-12262	120	1	[	[	X
ajst-12262	120	2	11	11	NUM
ajst-12262	120	3	]	]	X
ajst-12262	120	4	cai	cai	X
ajst-12262	120	5	,	,	PUNCT
ajst-12262	120	6	l.	l.	PROPN
ajst-12262	120	7	,	,	PUNCT
ajst-12262	120	8	hu	hu	PROPN
ajst-12262	120	9	,	,	PUNCT
ajst-12262	120	10	y.	y.	PROPN
ajst-12262	120	11	,	,	PUNCT
ajst-12262	120	12	dong	dong	PROPN
ajst-12262	120	13	,	,	PUNCT
ajst-12262	120	14	j.	j.	PROPN
ajst-12262	120	15	,	,	PUNCT
ajst-12262	120	16	&	&	CCONJ
ajst-12262	120	17	zhou	zhou	PROPN
ajst-12262	120	18	,	,	PUNCT
ajst-12262	120	19	s.	s.	PROPN
ajst-12262	120	20	(	(	PUNCT
ajst-12262	120	21	2019	2019	NUM
ajst-12262	120	22	)	)	PUNCT
ajst-12262	120	23	.	.	PUNCT
ajst-12262	121	1	audio	audio	ADJ
ajst-12262	121	2	-	-	ADJ
ajst-12262	121	3	textual	textual	ADJ
ajst-12262	121	4	emotion	emotion	NOUN
ajst-12262	121	5	recognition	recognition	NOUN
ajst-12262	121	6	based	base	VERB
ajst-12262	121	7	on	on	ADP
ajst-12262	121	8	improved	improve	VERB
ajst-12262	121	9	neural	neural	ADJ
ajst-12262	121	10	networks	network	NOUN
ajst-12262	121	11	.	.	PUNCT
ajst-12262	122	1	mathematical	mathematical	ADJ
ajst-12262	122	2	problems	problem	NOUN
ajst-12262	122	3	in	in	ADP
ajst-12262	122	4	engineering	engineering	NOUN
ajst-12262	122	5	,	,	PUNCT
ajst-12262	122	6	2019(6	2019(6	NUM
ajst-12262	122	7	)	)	PUNCT
ajst-12262	122	8	,	,	PUNCT
ajst-12262	122	9	1	1	NUM
ajst-12262	122	10	-	-	SYM
ajst-12262	122	11	9	9	NUM
ajst-12262	122	12	.	.	PUNCT
ajst-12262	123	1	[	[	X
ajst-12262	123	2	12	12	NUM
ajst-12262	123	3	]	]	X
ajst-12262	123	4	huang	huang	PROPN
ajst-12262	123	5	,	,	PUNCT
ajst-12262	123	6	h.	h.	PROPN
ajst-12262	123	7	,	,	PUNCT
ajst-12262	123	8	liu	liu	PROPN
ajst-12262	123	9	,	,	PUNCT
ajst-12262	123	10	q.	q.	PROPN
ajst-12262	123	11	,	,	PUNCT
ajst-12262	123	12	&	&	CCONJ
ajst-12262	123	13	wu	wu	PROPN
ajst-12262	123	14	,	,	PUNCT
ajst-12262	123	15	l.	l.	PROPN
ajst-12262	123	16	(	(	PUNCT
ajst-12262	123	17	2014	2014	NUM
ajst-12262	123	18	)	)	PUNCT
ajst-12262	123	19	.	.	PUNCT
ajst-12262	124	1	improving	improve	VERB
ajst-12262	124	2	emotion	emotion	NOUN
ajst-12262	124	3	recognition	recognition	NOUN
ajst-12262	124	4	from	from	ADP
ajst-12262	124	5	imbalanced	imbalanced	ADJ
ajst-12262	124	6	corpus	corpus	NOUN
ajst-12262	124	7	with	with	ADP
ajst-12262	124	8	a	a	DET
ajst-12262	124	9	novel	novel	ADJ
ajst-12262	124	10	re	re	NOUN
ajst-12262	124	11	-	-	ADJ
ajst-12262	124	12	sampling	sample	VERB
ajst-12262	124	13	method	method	NOUN
ajst-12262	124	14	.	.	PUNCT
ajst-12262	125	1	icic	icic	PROPN
ajst-12262	125	2	express	express	PROPN
ajst-12262	125	3	letters	letter	NOUN
ajst-12262	125	4	,	,	PUNCT
ajst-12262	125	5	8(7	8(7	NUM
ajst-12262	125	6	)	)	PUNCT
ajst-12262	125	7	,	,	PUNCT
ajst-12262	125	8	1845	1845	NUM
ajst-12262	125	9	-	-	SYM
ajst-12262	125	10	1850	1850	NUM
ajst-12262	125	11	.	.	PUNCT
ajst-12262	126	1	[	[	X
ajst-12262	126	2	13	13	NUM
ajst-12262	126	3	]	]	SYM
ajst-12262	126	4	xie	xie	PROPN
ajst-12262	126	5	,	,	PUNCT
ajst-12262	126	6	b.	b.	PROPN
ajst-12262	126	7	,	,	PUNCT
ajst-12262	126	8	sidulova	sidulova	VERB
ajst-12262	126	9	,	,	PUNCT
ajst-12262	126	10	m.	m.	NOUN
ajst-12262	126	11	,	,	PUNCT
ajst-12262	126	12	&	&	CCONJ
ajst-12262	126	13	park	park	PROPN
ajst-12262	126	14	,	,	PUNCT
ajst-12262	126	15	c.	c.	PROPN
ajst-12262	126	16	h.	h.	PROPN
ajst-12262	126	17	(	(	PUNCT
ajst-12262	126	18	2021	2021	NUM
ajst-12262	126	19	)	)	PUNCT
ajst-12262	126	20	.	.	PUNCT
ajst-12262	127	1	robust	robust	ADJ
ajst-12262	127	2	multimodal	multimodal	NOUN
ajst-12262	127	3	emotion	emotion	NOUN
ajst-12262	127	4	recognition	recognition	NOUN
ajst-12262	127	5	from	from	ADP
ajst-12262	127	6	conversation	conversation	NOUN
ajst-12262	127	7	with	with	ADP
ajst-12262	127	8	transformer	transformer	NOUN
ajst-12262	127	9	-	-	PUNCT
ajst-12262	127	10	based	base	VERB
ajst-12262	127	11	crossmodality	crossmodality	NOUN
ajst-12262	127	12	fusion	fusion	NOUN
ajst-12262	127	13	.	.	PUNCT
ajst-12262	128	1	sensors	sensor	NOUN
ajst-12262	128	2	,	,	PUNCT
ajst-12262	128	3	21(14	21(14	NUM
ajst-12262	128	4	)	)	PUNCT
ajst-12262	128	5	,	,	PUNCT
ajst-12262	128	6	4913	4913	NUM
ajst-12262	128	7	.	.	PUNCT