id	sid	tid	token	lemma	pos
ajst-9439	1	1	academic	academic	ADJ
ajst-9439	1	2	journal	journal	NOUN
ajst-9439	1	3	of	of	ADP
ajst-9439	1	4	science	science	NOUN
ajst-9439	1	5	and	and	CCONJ
ajst-9439	1	6	technology	technology	NOUN
ajst-9439	1	7	issn	issn	NOUN
ajst-9439	1	8	:	:	PUNCT
ajst-9439	1	9	2771	2771	NUM
ajst-9439	1	10	-	-	SYM
ajst-9439	1	11	3032	3032	NUM
ajst-9439	1	12	|	|	NOUN
ajst-9439	1	13	vol	vol	NOUN
ajst-9439	1	14	.	.	PROPN
ajst-9439	2	1	6	6	NUM
ajst-9439	2	2	,	,	PUNCT
ajst-9439	2	3	no	no	INTJ
ajst-9439	2	4	.	.	NOUN
ajst-9439	2	5	2	2	NUM
ajst-9439	2	6	,	,	PUNCT
ajst-9439	2	7	2023	2023	NUM
ajst-9439	2	8	26	26	NUM
ajst-9439	2	9	practical	practical	ADJ
ajst-9439	2	10	application	application	NOUN
ajst-9439	2	11	of	of	ADP
ajst-9439	2	12	platform	platform	NOUN
ajst-9439	2	13	comment	comment	NOUN
ajst-9439	2	14	sentiment	sentiment	NOUN
ajst-9439	2	15	binary	binary	ADJ
ajst-9439	2	16	classification	classification	NOUN
ajst-9439	2	17	based	base	VERB
ajst-9439	2	18	on	on	ADP
ajst-9439	2	19	deep	deep	ADJ
ajst-9439	2	20	learning	learning	NOUN
ajst-9439	2	21	jihan	jihan	PROPN
ajst-9439	2	22	huang1	huang1	PROPN
ajst-9439	2	23	,	,	PUNCT
ajst-9439	2	24	a	a	DET
ajst-9439	2	25	1zhuhai	1zhuhai	NUM
ajst-9439	2	26	college	college	NOUN
ajst-9439	2	27	,	,	PUNCT
ajst-9439	2	28	jilin	jilin	PROPN
ajst-9439	2	29	university	university	PROPN
ajst-9439	2	30	,	,	PUNCT
ajst-9439	2	31	zhuhai	zhuhai	PROPN
ajst-9439	2	32	,	,	PUNCT
ajst-9439	2	33	guangdong	guangdong	PROPN
ajst-9439	2	34	,	,	PUNCT
ajst-9439	2	35	519041	519041	NUM
ajst-9439	2	36	,	,	PUNCT
ajst-9439	2	37	china	china	PROPN
ajst-9439	2	38	ahuangjihan1@gmail.com	ahuangjihan1@gmail.com	X
ajst-9439	2	39	abstract	abstract	ADJ
ajst-9439	2	40	:	:	PUNCT
ajst-9439	2	41	the	the	DET
ajst-9439	2	42	role	role	NOUN
ajst-9439	2	43	of	of	ADP
ajst-9439	2	44	product	product	NOUN
ajst-9439	2	45	reviews	review	NOUN
ajst-9439	2	46	is	be	AUX
ajst-9439	2	47	to	to	PART
ajst-9439	2	48	display	display	VERB
ajst-9439	2	49	real	real	ADJ
ajst-9439	2	50	product	product	NOUN
ajst-9439	2	51	information	information	NOUN
ajst-9439	2	52	,	,	PUNCT
ajst-9439	2	53	help	help	VERB
ajst-9439	2	54	users	user	NOUN
ajst-9439	2	55	judge	judge	VERB
ajst-9439	2	56	whether	whether	SCONJ
ajst-9439	2	57	the	the	DET
ajst-9439	2	58	real	real	ADJ
ajst-9439	2	59	situation	situation	NOUN
ajst-9439	2	60	of	of	ADP
ajst-9439	2	61	the	the	DET
ajst-9439	2	62	product	product	NOUN
ajst-9439	2	63	is	be	AUX
ajst-9439	2	64	consistent	consistent	ADJ
ajst-9439	2	65	with	with	ADP
ajst-9439	2	66	the	the	DET
ajst-9439	2	67	merchant	merchant	NOUN
ajst-9439	2	68	's	's	PART
ajst-9439	2	69	description	description	NOUN
ajst-9439	2	70	,	,	PUNCT
ajst-9439	2	71	and	and	CCONJ
ajst-9439	2	72	indirectly	indirectly	ADV
ajst-9439	2	73	help	help	VERB
ajst-9439	2	74	users	user	NOUN
ajst-9439	2	75	judge	judge	VERB
ajst-9439	2	76	the	the	DET
ajst-9439	2	77	price	price	NOUN
ajst-9439	2	78	range	range	NOUN
ajst-9439	2	79	of	of	ADP
ajst-9439	2	80	the	the	DET
ajst-9439	2	81	product	product	NOUN
ajst-9439	2	82	.	.	PUNCT
ajst-9439	3	1	at	at	ADP
ajst-9439	3	2	present	present	ADJ
ajst-9439	3	3	,	,	PUNCT
ajst-9439	3	4	most	most	ADJ
ajst-9439	3	5	commodity	commodity	NOUN
ajst-9439	3	6	platforms	platform	NOUN
ajst-9439	3	7	and	and	CCONJ
ajst-9439	3	8	accommodation	accommodation	NOUN
ajst-9439	3	9	intermediary	intermediary	NOUN
ajst-9439	3	10	platforms	platform	NOUN
ajst-9439	3	11	adopt	adopt	VERB
ajst-9439	3	12	the	the	DET
ajst-9439	3	13	method	method	NOUN
ajst-9439	3	14	of	of	ADP
ajst-9439	3	15	user	user	NOUN
ajst-9439	3	16	star	star	NOUN
ajst-9439	3	17	rating	rating	NOUN
ajst-9439	3	18	,	,	PUNCT
ajst-9439	3	19	which	which	PRON
ajst-9439	3	20	leads	lead	VERB
ajst-9439	3	21	to	to	ADP
ajst-9439	3	22	many	many	ADJ
ajst-9439	3	23	invalid	invalid	ADJ
ajst-9439	3	24	positive	positive	ADJ
ajst-9439	3	25	reviews	review	NOUN
ajst-9439	3	26	and	and	CCONJ
ajst-9439	3	27	malicious	malicious	ADJ
ajst-9439	3	28	negative	negative	ADJ
ajst-9439	3	29	reviews	review	NOUN
ajst-9439	3	30	,	,	PUNCT
ajst-9439	3	31	thus	thus	ADV
ajst-9439	3	32	bringing	bring	VERB
ajst-9439	3	33	a	a	DET
ajst-9439	3	34	lot	lot	NOUN
ajst-9439	3	35	of	of	ADP
ajst-9439	3	36	disturbing	disturb	VERB
ajst-9439	3	37	information	information	NOUN
ajst-9439	3	38	to	to	ADP
ajst-9439	3	39	consumers	consumer	NOUN
ajst-9439	3	40	.	.	PUNCT
ajst-9439	4	1	therefore	therefore	ADV
ajst-9439	4	2	,	,	PUNCT
ajst-9439	4	3	this	this	DET
ajst-9439	4	4	paper	paper	NOUN
ajst-9439	4	5	introduces	introduce	VERB
ajst-9439	4	6	a	a	DET
ajst-9439	4	7	mechanism	mechanism	NOUN
ajst-9439	4	8	based	base	VERB
ajst-9439	4	9	on	on	ADP
ajst-9439	4	10	artificial	artificial	ADJ
ajst-9439	4	11	intelligence	intelligence	NOUN
ajst-9439	4	12	to	to	PART
ajst-9439	4	13	judge	judge	VERB
ajst-9439	4	14	the	the	DET
ajst-9439	4	15	emotion	emotion	NOUN
ajst-9439	4	16	expressed	express	VERB
ajst-9439	4	17	by	by	ADP
ajst-9439	4	18	a	a	DET
ajst-9439	4	19	machine	machine	NOUN
ajst-9439	4	20	,	,	PUNCT
ajst-9439	4	21	and	and	CCONJ
ajst-9439	4	22	further	further	ADJ
ajst-9439	4	23	verifies	verifie	NOUN
ajst-9439	4	24	the	the	DET
ajst-9439	4	25	validity	validity	NOUN
ajst-9439	4	26	and	and	CCONJ
ajst-9439	4	27	accuracy	accuracy	NOUN
ajst-9439	4	28	of	of	ADP
ajst-9439	4	29	the	the	DET
ajst-9439	4	30	model	model	NOUN
ajst-9439	4	31	on	on	ADP
ajst-9439	4	32	the	the	DET
ajst-9439	4	33	bert	bert	PROPN
ajst-9439	4	34	hugging	hug	VERB
ajst-9439	4	35	face	face	NOUN
ajst-9439	4	36	dataset	dataset	NOUN
ajst-9439	4	37	.	.	PUNCT
ajst-9439	5	1	keywords	keyword	NOUN
ajst-9439	5	2	:	:	PUNCT
ajst-9439	5	3	sentiment	sentiment	NOUN
ajst-9439	5	4	classification	classification	NOUN
ajst-9439	5	5	,	,	PUNCT
ajst-9439	5	6	deep	deep	ADJ
ajst-9439	5	7	learning	learning	NOUN
ajst-9439	5	8	,	,	PUNCT
ajst-9439	5	9	natural	natural	ADJ
ajst-9439	5	10	language	language	NOUN
ajst-9439	5	11	processing	processing	NOUN
ajst-9439	5	12	.	.	PUNCT
ajst-9439	6	1	1	1	X
ajst-9439	6	2	.	.	X
ajst-9439	6	3	introduction	introduction	NOUN
ajst-9439	6	4	at	at	ADP
ajst-9439	6	5	present	present	ADJ
ajst-9439	6	6	,	,	PUNCT
ajst-9439	6	7	mainstream	mainstream	ADJ
ajst-9439	6	8	commodity	commodity	NOUN
ajst-9439	6	9	platforms	platform	NOUN
ajst-9439	6	10	and	and	CCONJ
ajst-9439	6	11	accommodation	accommodation	NOUN
ajst-9439	6	12	intermediary	intermediary	NOUN
ajst-9439	6	13	platforms	platform	NOUN
ajst-9439	6	14	,	,	PUNCT
ajst-9439	6	15	such	such	ADJ
ajst-9439	6	16	as	as	ADP
ajst-9439	6	17	taobao[1	taobao[1	X
ajst-9439	6	18	]	]	PUNCT
ajst-9439	6	19	and	and	CCONJ
ajst-9439	6	20	ctrip[2	ctrip[2	PROPN
ajst-9439	6	21	]	]	PUNCT
ajst-9439	6	22	,	,	PUNCT
ajst-9439	6	23	all	all	PRON
ajst-9439	6	24	use	use	VERB
ajst-9439	6	25	star	star	NOUN
ajst-9439	6	26	ratings	rating	NOUN
ajst-9439	6	27	to	to	PART
ajst-9439	6	28	assist	assist	VERB
ajst-9439	6	29	consumers	consumer	NOUN
ajst-9439	6	30	in	in	ADP
ajst-9439	6	31	making	make	VERB
ajst-9439	6	32	purchase	purchase	NOUN
ajst-9439	6	33	decisions	decision	NOUN
ajst-9439	6	34	.	.	PUNCT
ajst-9439	7	1	it	it	PRON
ajst-9439	7	2	is	be	AUX
ajst-9439	7	3	no	no	ADV
ajst-9439	7	4	longer	long	ADV
ajst-9439	7	5	a	a	DET
ajst-9439	7	6	good	good	ADJ
ajst-9439	7	7	solution	solution	NOUN
ajst-9439	7	8	to	to	ADP
ajst-9439	7	9	this	this	DET
ajst-9439	7	10	problem	problem	NOUN
ajst-9439	7	11	to	to	PART
ajst-9439	7	12	distinguish	distinguish	VERB
ajst-9439	7	13	whether	whether	SCONJ
ajst-9439	7	14	the	the	DET
ajst-9439	7	15	goods	good	NOUN
ajst-9439	7	16	are	be	AUX
ajst-9439	7	17	good	good	ADJ
ajst-9439	7	18	or	or	CCONJ
ajst-9439	7	19	bad	bad	ADJ
ajst-9439	7	20	.	.	PUNCT
ajst-9439	8	1	because	because	SCONJ
ajst-9439	8	2	there	there	PRON
ajst-9439	8	3	are	be	VERB
ajst-9439	8	4	a	a	DET
ajst-9439	8	5	lot	lot	NOUN
ajst-9439	8	6	of	of	ADP
ajst-9439	8	7	full	full	ADJ
ajst-9439	8	8	-	-	PUNCT
ajst-9439	8	9	star	star	NOUN
ajst-9439	8	10	reviews	review	NOUN
ajst-9439	8	11	,	,	PUNCT
ajst-9439	8	12	the	the	DET
ajst-9439	8	13	content	content	NOUN
ajst-9439	8	14	is	be	AUX
ajst-9439	8	15	perfunctory	perfunctory	ADJ
ajst-9439	8	16	,	,	PUNCT
ajst-9439	8	17	because	because	SCONJ
ajst-9439	8	18	many	many	ADJ
ajst-9439	8	19	people	people	NOUN
ajst-9439	8	20	are	be	AUX
ajst-9439	8	21	habitually	habitually	ADV
ajst-9439	8	22	full	full	ADJ
ajst-9439	8	23	-	-	PUNCT
ajst-9439	8	24	star	star	NOUN
ajst-9439	8	25	.	.	PUNCT
ajst-9439	9	1	furthermore	furthermore	ADV
ajst-9439	9	2	,	,	PUNCT
ajst-9439	9	3	when	when	SCONJ
ajst-9439	9	4	someone	someone	PRON
ajst-9439	9	5	gives	give	VERB
ajst-9439	9	6	a	a	DET
ajst-9439	9	7	full	full	ADJ
ajst-9439	9	8	-	-	PUNCT
ajst-9439	9	9	star	star	NOUN
ajst-9439	9	10	/	/	SYM
ajst-9439	9	11	one	one	NUM
ajst-9439	9	12	-	-	PUNCT
ajst-9439	9	13	star	star	NOUN
ajst-9439	9	14	review	review	NOUN
ajst-9439	9	15	without	without	ADP
ajst-9439	9	16	leaving	leave	VERB
ajst-9439	9	17	detailed	detailed	ADJ
ajst-9439	9	18	feedback	feedback	NOUN
ajst-9439	9	19	,	,	PUNCT
ajst-9439	9	20	how	how	SCONJ
ajst-9439	9	21	do	do	AUX
ajst-9439	9	22	you	you	PRON
ajst-9439	9	23	understand	understand	VERB
ajst-9439	9	24	this	this	DET
ajst-9439	9	25	full	full	ADJ
ajst-9439	9	26	-	-	PUNCT
ajst-9439	9	27	star	star	NOUN
ajst-9439	9	28	/	/	SYM
ajst-9439	9	29	one	one	NUM
ajst-9439	9	30	-	-	PUNCT
ajst-9439	9	31	star	star	NOUN
ajst-9439	9	32	review	review	NOUN
ajst-9439	9	33	?	?	PUNCT
ajst-9439	10	1	ratings	rating	NOUN
ajst-9439	10	2	given	give	VERB
ajst-9439	10	3	by	by	ADP
ajst-9439	10	4	reviews	review	NOUN
ajst-9439	10	5	are	be	AUX
ajst-9439	10	6	great	great	ADJ
ajst-9439	10	7	when	when	SCONJ
ajst-9439	10	8	there	there	PRON
ajst-9439	10	9	is	be	VERB
ajst-9439	10	10	no	no	DET
ajst-9439	10	11	explanation	explanation	NOUN
ajst-9439	10	12	or	or	CCONJ
ajst-9439	10	13	clarification	clarification	NOUN
ajst-9439	10	14	,	,	PUNCT
ajst-9439	10	15	but	but	CCONJ
ajst-9439	10	16	they	they	PRON
ajst-9439	10	17	are	be	AUX
ajst-9439	10	18	very	very	ADV
ajst-9439	10	19	bad	bad	ADJ
ajst-9439	10	20	for	for	SCONJ
ajst-9439	10	21	the	the	DET
ajst-9439	10	22	person	person	NOUN
ajst-9439	10	23	reading	read	VERB
ajst-9439	10	24	these	these	DET
ajst-9439	10	25	reviews	review	NOUN
ajst-9439	10	26	,	,	PUNCT
ajst-9439	10	27	because	because	SCONJ
ajst-9439	10	28	the	the	DET
ajst-9439	10	29	real	real	ADJ
ajst-9439	10	30	value	value	NOUN
ajst-9439	10	31	of	of	ADP
ajst-9439	10	32	the	the	DET
ajst-9439	10	33	product	product	NOUN
ajst-9439	10	34	can	can	AUX
ajst-9439	10	35	not	not	PART
ajst-9439	10	36	be	be	AUX
ajst-9439	10	37	understood	understand	VERB
ajst-9439	10	38	from	from	ADP
ajst-9439	10	39	these	these	DET
ajst-9439	10	40	scores	score	NOUN
ajst-9439	10	41	.	.	PUNCT
ajst-9439	11	1	as	as	ADP
ajst-9439	11	2	consumers	consumer	NOUN
ajst-9439	11	3	,	,	PUNCT
ajst-9439	11	4	what	what	PRON
ajst-9439	11	5	we	we	PRON
ajst-9439	11	6	need	need	VERB
ajst-9439	11	7	are	be	AUX
ajst-9439	11	8	more	more	ADV
ajst-9439	11	9	real	real	ADJ
ajst-9439	11	10	and	and	CCONJ
ajst-9439	11	11	contentbased	contentbased	ADJ
ajst-9439	11	12	evaluations	evaluation	NOUN
ajst-9439	11	13	.	.	PUNCT
ajst-9439	12	1	therefore	therefore	ADV
ajst-9439	12	2	,	,	PUNCT
ajst-9439	12	3	we	we	PRON
ajst-9439	12	4	have	have	AUX
ajst-9439	12	5	introduced	introduce	VERB
ajst-9439	12	6	an	an	DET
ajst-9439	12	7	evaluation	evaluation	NOUN
ajst-9439	12	8	method	method	NOUN
ajst-9439	12	9	that	that	PRON
ajst-9439	12	10	uses	use	VERB
ajst-9439	12	11	content	content	NOUN
ajst-9439	12	12	as	as	ADP
ajst-9439	12	13	the	the	DET
ajst-9439	12	14	judgment	judgment	NOUN
ajst-9439	12	15	material	material	NOUN
ajst-9439	12	16	,	,	PUNCT
ajst-9439	12	17	and	and	CCONJ
ajst-9439	12	18	directly	directly	ADV
ajst-9439	12	19	extracts	extract	VERB
ajst-9439	12	20	the	the	DET
ajst-9439	12	21	characteristics	characteristic	NOUN
ajst-9439	12	22	of	of	ADP
ajst-9439	12	23	the	the	DET
ajst-9439	12	24	content	content	NOUN
ajst-9439	12	25	to	to	PART
ajst-9439	12	26	judge	judge	VERB
ajst-9439	12	27	whether	whether	SCONJ
ajst-9439	12	28	it	it	PRON
ajst-9439	12	29	is	be	AUX
ajst-9439	12	30	a	a	DET
ajst-9439	12	31	good	good	ADJ
ajst-9439	12	32	or	or	CCONJ
ajst-9439	12	33	bad	bad	ADJ
ajst-9439	12	34	review	review	NOUN
ajst-9439	12	35	,	,	PUNCT
ajst-9439	12	36	without	without	ADP
ajst-9439	12	37	the	the	DET
ajst-9439	12	38	need	need	NOUN
ajst-9439	12	39	for	for	SCONJ
ajst-9439	12	40	people	people	NOUN
ajst-9439	12	41	to	to	PART
ajst-9439	12	42	manually	manually	ADV
ajst-9439	12	43	score	score	VERB
ajst-9439	12	44	.	.	PUNCT
ajst-9439	13	1	and	and	CCONJ
ajst-9439	13	2	if	if	SCONJ
ajst-9439	13	3	the	the	DET
ajst-9439	13	4	content	content	NOUN
ajst-9439	13	5	is	be	AUX
ajst-9439	13	6	too	too	ADV
ajst-9439	13	7	little	little	ADJ
ajst-9439	13	8	,	,	PUNCT
ajst-9439	13	9	no	no	DET
ajst-9439	13	10	judgment	judgment	NOUN
ajst-9439	13	11	will	will	AUX
ajst-9439	13	12	be	be	AUX
ajst-9439	13	13	made	make	VERB
ajst-9439	13	14	,	,	PUNCT
ajst-9439	13	15	because	because	SCONJ
ajst-9439	13	16	it	it	PRON
ajst-9439	13	17	is	be	AUX
ajst-9439	13	18	an	an	DET
ajst-9439	13	19	invalid	invalid	ADJ
ajst-9439	13	20	review	review	NOUN
ajst-9439	13	21	for	for	ADP
ajst-9439	13	22	consumers	consumer	NOUN
ajst-9439	13	23	.	.	PUNCT
ajst-9439	14	1	figure	figure	VERB
ajst-9439	14	2	1	1	NUM
ajst-9439	14	3	.	.	PUNCT
ajst-9439	15	1	scale	scale	NOUN
ajst-9439	15	2	of	of	ADP
ajst-9439	15	3	online	online	ADJ
ajst-9439	15	4	shopping	shopping	NOUN
ajst-9439	15	5	users	user	NOUN
ajst-9439	15	6	and	and	CCONJ
ajst-9439	15	7	their	their	PRON
ajst-9439	15	8	usage(2017	usage(2017	NOUN
ajst-9439	15	9	-	-	PUNCT
ajst-9439	15	10	2021	2021	NUM
ajst-9439	15	11	)	)	PUNCT
ajst-9439	16	1	it	it	PRON
ajst-9439	16	2	can	can	AUX
ajst-9439	16	3	be	be	AUX
ajst-9439	16	4	seen	see	VERB
ajst-9439	16	5	from	from	ADP
ajst-9439	16	6	table	table	NOUN
ajst-9439	16	7	1	1	NUM
ajst-9439	16	8	that	that	SCONJ
ajst-9439	16	9	the	the	DET
ajst-9439	16	10	scale	scale	NOUN
ajst-9439	16	11	of	of	ADP
ajst-9439	16	12	shopping	shopping	NOUN
ajst-9439	16	13	by	by	ADP
ajst-9439	16	14	online	online	ADJ
ajst-9439	16	15	users	user	NOUN
ajst-9439	16	16	is	be	AUX
ajst-9439	16	17	increasing	increase	VERB
ajst-9439	16	18	every	every	DET
ajst-9439	16	19	year	year	NOUN
ajst-9439	16	20	,	,	PUNCT
ajst-9439	16	21	and	and	CCONJ
ajst-9439	16	22	the	the	DET
ajst-9439	16	23	overall	overall	ADJ
ajst-9439	16	24	utilization	utilization	NOUN
ajst-9439	16	25	rate	rate	NOUN
ajst-9439	16	26	has	have	AUX
ajst-9439	16	27	reached	reach	VERB
ajst-9439	16	28	more	more	ADJ
ajst-9439	16	29	than	than	ADP
ajst-9439	16	30	80	80	NUM
ajst-9439	16	31	%	%	NOUN
ajst-9439	16	32	by	by	ADP
ajst-9439	16	33	december	december	PROPN
ajst-9439	16	34	21	21	NUM
ajst-9439	16	35	,	,	PUNCT
ajst-9439	16	36	which	which	PRON
ajst-9439	16	37	reflects	reflect	VERB
ajst-9439	16	38	people	people	NOUN
ajst-9439	16	39	's	's	PART
ajst-9439	16	40	dependence	dependence	NOUN
ajst-9439	16	41	on	on	ADP
ajst-9439	16	42	online	online	ADJ
ajst-9439	16	43	shopping	shopping	NOUN
ajst-9439	16	44	platforms	platform	NOUN
ajst-9439	16	45	.	.	PUNCT
ajst-9439	17	1	due	due	ADP
ajst-9439	17	2	to	to	ADP
ajst-9439	17	3	the	the	DET
ajst-9439	17	4	particularity	particularity	NOUN
ajst-9439	17	5	of	of	ADP
ajst-9439	17	6	online	online	ADJ
ajst-9439	17	7	shopping	shopping	NOUN
ajst-9439	17	8	,	,	PUNCT
ajst-9439	17	9	people	people	NOUN
ajst-9439	17	10	will	will	AUX
ajst-9439	17	11	pay	pay	VERB
ajst-9439	17	12	more	more	ADJ
ajst-9439	17	13	attention	attention	NOUN
ajst-9439	17	14	to	to	ADP
ajst-9439	17	15	the	the	DET
ajst-9439	17	16	reviews	review	NOUN
ajst-9439	17	17	of	of	ADP
ajst-9439	17	18	products	product	NOUN
ajst-9439	17	19	by	by	ADP
ajst-9439	17	20	buyers	buyer	NOUN
ajst-9439	17	21	,	,	PUNCT
ajst-9439	17	22	so	so	SCONJ
ajst-9439	17	23	the	the	DET
ajst-9439	17	24	sentiment	sentiment	NOUN
ajst-9439	17	25	classification	classification	NOUN
ajst-9439	17	26	for	for	ADP
ajst-9439	17	27	reviews	review	NOUN
ajst-9439	17	28	can	can	AUX
ajst-9439	17	29	be	be	AUX
ajst-9439	17	30	said	say	VERB
ajst-9439	17	31	to	to	PART
ajst-9439	17	32	be	be	AUX
ajst-9439	17	33	very	very	ADV
ajst-9439	17	34	important	important	ADJ
ajst-9439	17	35	.	.	PUNCT
ajst-9439	18	1	62.00	62.00	NUM
ajst-9439	18	2	%	%	NOUN
ajst-9439	18	3	64.00	64.00	NUM
ajst-9439	18	4	%	%	NOUN
ajst-9439	18	5	66.00	66.00	NUM
ajst-9439	18	6	%	%	NOUN
ajst-9439	18	7	68.00	68.00	NUM
ajst-9439	18	8	%	%	NOUN
ajst-9439	18	9	70.00	70.00	NUM
ajst-9439	18	10	%	%	NOUN
ajst-9439	18	11	72.00	72.00	NUM
ajst-9439	18	12	%	%	NOUN
ajst-9439	18	13	74.00	74.00	NUM
ajst-9439	18	14	%	%	NOUN
ajst-9439	18	15	76.00	76.00	NUM
ajst-9439	18	16	%	%	NOUN
ajst-9439	18	17	78.00	78.00	NUM
ajst-9439	18	18	%	%	NOUN
ajst-9439	18	19	80.00	80.00	NUM
ajst-9439	18	20	%	%	NOUN
ajst-9439	18	21	82.00	82.00	NUM
ajst-9439	18	22	%	%	NOUN
ajst-9439	18	23	84.00	84.00	NUM
ajst-9439	18	24	%	%	NOUN
ajst-9439	18	25	0	0	NUM
ajst-9439	18	26	10000	10000	NUM
ajst-9439	18	27	20000	20000	NUM
ajst-9439	18	28	30000	30000	NUM
ajst-9439	18	29	40000	40000	NUM
ajst-9439	18	30	50000	50000	NUM
ajst-9439	18	31	60000	60000	NUM
ajst-9439	18	32	70000	70000	NUM
ajst-9439	18	33	80000	80000	NUM
ajst-9439	18	34	90000	90000	NUM
ajst-9439	18	35	2017.12	2017.12	NUM
ajst-9439	18	36	2018.12	2018.12	NUM
ajst-9439	18	37	2020.3	2020.3	NUM
ajst-9439	18	38	2020.12	2020.12	NUM
ajst-9439	18	39	2021.12	2021.12	NUM
ajst-9439	18	40	use	use	NOUN
ajst-9439	18	41	scale	scale	NOUN
ajst-9439	18	42	：	：	PUNCT
ajst-9439	18	43	thousand	thousand	NUM
ajst-9439	18	44	utilization	utilization	NOUN
ajst-9439	18	45	rate	rate	NOUN
ajst-9439	18	46	：	：	PUNCT
ajst-9439	18	47	%	%	NOUN
ajst-9439	18	48	27	27	NUM
ajst-9439	18	49	figure	figure	NOUN
ajst-9439	18	50	2	2	NUM
ajst-9439	18	51	.	.	NUM
ajst-9439	18	52	double	double	ADJ
ajst-9439	18	53	11	11	NUM
ajst-9439	18	54	turnover	turnover	NOUN
ajst-9439	18	55	of	of	ADP
ajst-9439	18	56	online	online	ADJ
ajst-9439	18	57	shoppers(2017	shoppers(2017	NOUN
ajst-9439	18	58	-	-	PUNCT
ajst-9439	18	59	2021	2021	NUM
ajst-9439	18	60	)	)	PUNCT
ajst-9439	18	61	it	it	PRON
ajst-9439	18	62	can	can	AUX
ajst-9439	18	63	be	be	AUX
ajst-9439	18	64	seen	see	VERB
ajst-9439	18	65	from	from	ADP
ajst-9439	18	66	table	table	NOUN
ajst-9439	18	67	2	2	NUM
ajst-9439	18	68	that	that	SCONJ
ajst-9439	18	69	the	the	DET
ajst-9439	18	70	transaction	transaction	NOUN
ajst-9439	18	71	volume	volume	NOUN
ajst-9439	18	72	in	in	ADP
ajst-9439	18	73	recent	recent	ADJ
ajst-9439	18	74	years	year	NOUN
ajst-9439	18	75	has	have	AUX
ajst-9439	18	76	tended	tend	VERB
ajst-9439	18	77	to	to	PART
ajst-9439	18	78	be	be	AUX
ajst-9439	18	79	stable	stable	ADJ
ajst-9439	18	80	,	,	PUNCT
ajst-9439	18	81	and	and	CCONJ
ajst-9439	18	82	the	the	DET
ajst-9439	18	83	e	e	PROPN
ajst-9439	18	84	-	-	NOUN
ajst-9439	18	85	commerce	commerce	NOUN
ajst-9439	18	86	platform	platform	NOUN
ajst-9439	18	87	has	have	AUX
ajst-9439	18	88	entered	enter	VERB
ajst-9439	18	89	a	a	DET
ajst-9439	18	90	stable	stable	ADJ
ajst-9439	18	91	period	period	NOUN
ajst-9439	18	92	,	,	PUNCT
ajst-9439	18	93	with	with	ADP
ajst-9439	18	94	a	a	DET
ajst-9439	18	95	huge	huge	ADJ
ajst-9439	18	96	transaction	transaction	NOUN
ajst-9439	18	97	volume	volume	NOUN
ajst-9439	18	98	,	,	PUNCT
ajst-9439	18	99	which	which	PRON
ajst-9439	18	100	further	far	ADV
ajst-9439	18	101	shows	show	VERB
ajst-9439	18	102	the	the	DET
ajst-9439	18	103	importance	importance	NOUN
ajst-9439	18	104	of	of	ADP
ajst-9439	18	105	reviewassisted	reviewassiste	VERB
ajst-9439	18	106	purchases	purchase	NOUN
ajst-9439	18	107	.	.	PUNCT
ajst-9439	19	1	we	we	PRON
ajst-9439	19	2	conduct	conduct	VERB
ajst-9439	19	3	experiments	experiment	NOUN
ajst-9439	19	4	in	in	ADP
ajst-9439	19	5	the	the	DET
ajst-9439	19	6	context	context	NOUN
ajst-9439	19	7	of	of	ADP
ajst-9439	19	8	artificial	artificial	ADJ
ajst-9439	19	9	intelligence	intelligence	NOUN
ajst-9439	19	10	,	,	PUNCT
ajst-9439	19	11	and	and	CCONJ
ajst-9439	19	12	with	with	ADP
ajst-9439	19	13	the	the	DET
ajst-9439	19	14	help	help	NOUN
ajst-9439	19	15	of	of	ADP
ajst-9439	19	16	artificial	artificial	ADJ
ajst-9439	19	17	intelligence	intelligence	NOUN
ajst-9439	19	18	technology	technology	NOUN
ajst-9439	19	19	,	,	PUNCT
ajst-9439	19	20	we	we	PRON
ajst-9439	19	21	can	can	AUX
ajst-9439	19	22	achieve	achieve	VERB
ajst-9439	19	23	various	various	ADJ
ajst-9439	19	24	fake	fake	ADJ
ajst-9439	19	25	reviews	review	NOUN
ajst-9439	19	26	.	.	PUNCT
ajst-9439	20	1	ecommerce	ecommerce	NOUN
ajst-9439	20	2	platforms	platform	NOUN
ajst-9439	20	3	can	can	AUX
ajst-9439	20	4	use	use	VERB
ajst-9439	20	5	artificial	artificial	ADJ
ajst-9439	20	6	intelligence	intelligence	NOUN
ajst-9439	20	7	to	to	PART
ajst-9439	20	8	strengthen	strengthen	VERB
ajst-9439	20	9	verification	verification	NOUN
ajst-9439	20	10	and	and	CCONJ
ajst-9439	20	11	filter	filter	NOUN
ajst-9439	20	12	useful	useful	ADJ
ajst-9439	20	13	reviews	review	NOUN
ajst-9439	20	14	to	to	PART
ajst-9439	20	15	resist	resist	VERB
ajst-9439	20	16	malicious	malicious	ADJ
ajst-9439	20	17	marketing	marketing	NOUN
ajst-9439	20	18	.	.	PUNCT
ajst-9439	21	1	in	in	ADP
ajst-9439	21	2	addition	addition	NOUN
ajst-9439	21	3	,	,	PUNCT
ajst-9439	21	4	it	it	PRON
ajst-9439	21	5	can	can	AUX
ajst-9439	21	6	also	also	ADV
ajst-9439	21	7	help	help	VERB
ajst-9439	21	8	e	e	NOUN
ajst-9439	21	9	-	-	NOUN
ajst-9439	21	10	commerce	commerce	NOUN
ajst-9439	21	11	companies	company	NOUN
ajst-9439	21	12	realize	realize	VERB
ajst-9439	21	13	functions	function	NOUN
ajst-9439	21	14	such	such	ADJ
ajst-9439	21	15	as	as	ADP
ajst-9439	21	16	intelligent	intelligent	ADJ
ajst-9439	21	17	artificial	artificial	ADJ
ajst-9439	21	18	customer	customer	NOUN
ajst-9439	21	19	service	service	NOUN
ajst-9439	21	20	.	.	PUNCT
ajst-9439	22	1	in	in	ADP
ajst-9439	22	2	this	this	DET
ajst-9439	22	3	experiment	experiment	NOUN
ajst-9439	22	4	,	,	PUNCT
ajst-9439	22	5	we	we	PRON
ajst-9439	22	6	also	also	ADV
ajst-9439	22	7	need	need	VERB
ajst-9439	22	8	to	to	PART
ajst-9439	22	9	use	use	VERB
ajst-9439	22	10	artificial	artificial	ADJ
ajst-9439	22	11	intelligence	intelligence	NOUN
ajst-9439	22	12	technology	technology	NOUN
ajst-9439	22	13	to	to	PART
ajst-9439	22	14	carry	carry	VERB
ajst-9439	22	15	out	out	ADP
ajst-9439	22	16	the	the	DET
ajst-9439	22	17	sentimental	sentimental	ADJ
ajst-9439	22	18	classification	classification	NOUN
ajst-9439	22	19	of	of	ADP
ajst-9439	22	20	comments	comment	NOUN
ajst-9439	22	21	.	.	PUNCT
ajst-9439	23	1	in	in	ADP
ajst-9439	23	2	this	this	DET
ajst-9439	23	3	paper	paper	NOUN
ajst-9439	23	4	,	,	PUNCT
ajst-9439	23	5	we	we	PRON
ajst-9439	23	6	conducted	conduct	VERB
ajst-9439	23	7	experiments	experiment	NOUN
ajst-9439	23	8	to	to	PART
ajst-9439	23	9	test	test	VERB
ajst-9439	23	10	the	the	DET
ajst-9439	23	11	accuracy	accuracy	NOUN
ajst-9439	23	12	of	of	ADP
ajst-9439	23	13	the	the	DET
ajst-9439	23	14	model	model	NOUN
ajst-9439	23	15	in	in	ADP
ajst-9439	23	16	the	the	DET
ajst-9439	23	17	bert	bert	PROPN
ajst-9439	23	18	hugging	hug	VERB
ajst-9439	23	19	face	face	NOUN
ajst-9439	23	20	dataset	dataset	NOUN
ajst-9439	23	21	,	,	PUNCT
ajst-9439	23	22	and	and	CCONJ
ajst-9439	23	23	the	the	DET
ajst-9439	23	24	results	result	NOUN
ajst-9439	23	25	show	show	VERB
ajst-9439	23	26	that	that	SCONJ
ajst-9439	23	27	the	the	DET
ajst-9439	23	28	accuracy	accuracy	NOUN
ajst-9439	23	29	has	have	AUX
ajst-9439	23	30	reached	reach	VERB
ajst-9439	23	31	the	the	DET
ajst-9439	23	32	standard	standard	NOUN
ajst-9439	23	33	for	for	ADP
ajst-9439	23	34	industry	industry	NOUN
ajst-9439	23	35	applications	application	NOUN
ajst-9439	23	36	.	.	PUNCT
ajst-9439	24	1	2	2	X
ajst-9439	24	2	.	.	X
ajst-9439	24	3	related	relate	VERB
ajst-9439	24	4	word	word	NOUN
ajst-9439	24	5	2.1	2.1	NUM
ajst-9439	24	6	.	.	PUNCT
ajst-9439	25	1	convolutional	convolutional	ADJ
ajst-9439	25	2	neural	neural	ADJ
ajst-9439	25	3	network	network	NOUN
ajst-9439	25	4	the	the	DET
ajst-9439	25	5	widespread	widespread	ADJ
ajst-9439	25	6	application	application	NOUN
ajst-9439	25	7	of	of	ADP
ajst-9439	25	8	artificial	artificial	ADJ
ajst-9439	25	9	intelligence	intelligence	NOUN
ajst-9439	25	10	models	model	NOUN
ajst-9439	25	11	comes	come	VERB
ajst-9439	25	12	from	from	ADP
ajst-9439	25	13	the	the	DET
ajst-9439	25	14	rapid	rapid	ADJ
ajst-9439	25	15	development	development	NOUN
ajst-9439	25	16	of	of	ADP
ajst-9439	25	17	convolutional	convolutional	ADJ
ajst-9439	25	18	neural	neural	ADJ
ajst-9439	25	19	networks	network	NOUN
ajst-9439	25	20	.	.	PUNCT
ajst-9439	26	1	convolutional	convolutional	ADJ
ajst-9439	26	2	neural	neural	ADJ
ajst-9439	26	3	network	network	NOUN
ajst-9439	26	4	is	be	AUX
ajst-9439	26	5	a	a	DET
ajst-9439	26	6	kind	kind	NOUN
ajst-9439	26	7	of	of	ADP
ajst-9439	26	8	feed	feed	NOUN
ajst-9439	26	9	-	-	PUNCT
ajst-9439	26	10	forward	forward	ADV
ajst-9439	26	11	neural	neural	ADJ
ajst-9439	26	12	network	network	NOUN
ajst-9439	26	13	with	with	ADP
ajst-9439	26	14	convolution	convolution	NOUN
ajst-9439	26	15	calculation	calculation	NOUN
ajst-9439	26	16	and	and	CCONJ
ajst-9439	26	17	deep	deep	ADJ
ajst-9439	26	18	structure	structure	NOUN
ajst-9439	26	19	.	.	PUNCT
ajst-9439	27	1	it	it	PRON
ajst-9439	27	2	is	be	AUX
ajst-9439	27	3	a	a	DET
ajst-9439	27	4	very	very	ADV
ajst-9439	27	5	representative	representative	ADJ
ajst-9439	27	6	algorithm	algorithm	NOUN
ajst-9439	27	7	of	of	ADP
ajst-9439	27	8	deep	deep	ADJ
ajst-9439	27	9	learning	learning	NOUN
ajst-9439	27	10	.	.	PUNCT
ajst-9439	28	1	zhou	zhou	PROPN
ajst-9439	28	2	feiyan[3]and	feiyan[3]and	PROPN
ajst-9439	28	3	others	other	NOUN
ajst-9439	28	4	conducted	conduct	VERB
ajst-9439	28	5	a	a	DET
ajst-9439	28	6	review	review	NOUN
ajst-9439	28	7	of	of	ADP
ajst-9439	28	8	their	their	PRON
ajst-9439	28	9	convolutional	convolutional	ADJ
ajst-9439	28	10	neural	neural	ADJ
ajst-9439	28	11	networks	network	NOUN
ajst-9439	28	12	,	,	PUNCT
ajst-9439	28	13	and	and	CCONJ
ajst-9439	28	14	generally	generally	ADV
ajst-9439	28	15	introduced	introduce	VERB
ajst-9439	28	16	the	the	DET
ajst-9439	28	17	current	current	ADJ
ajst-9439	28	18	status	status	NOUN
ajst-9439	28	19	of	of	ADP
ajst-9439	28	20	convolutional	convolutional	ADJ
ajst-9439	28	21	neural	neural	ADJ
ajst-9439	28	22	networks	network	NOUN
ajst-9439	28	23	.	.	PUNCT
ajst-9439	29	1	zhang	zhang	PROPN
ajst-9439	29	2	shun[4]and	shun[4]and	PROPN
ajst-9439	29	3	others	other	NOUN
ajst-9439	29	4	introduced	introduce	VERB
ajst-9439	29	5	the	the	DET
ajst-9439	29	6	specific	specific	ADJ
ajst-9439	29	7	application	application	NOUN
ajst-9439	29	8	of	of	ADP
ajst-9439	29	9	convolutional	convolutional	ADJ
ajst-9439	29	10	neural	neural	ADJ
ajst-9439	29	11	network	network	NOUN
ajst-9439	29	12	.	.	PUNCT
ajst-9439	30	1	2.2	2.2	NUM
ajst-9439	30	2	.	.	PUNCT
ajst-9439	30	3	natural	natural	ADJ
ajst-9439	30	4	language	language	NOUN
ajst-9439	30	5	processing	processing	NOUN
ajst-9439	30	6	natural	natural	ADJ
ajst-9439	30	7	language	language	NOUN
ajst-9439	30	8	processing	processing	NOUN
ajst-9439	30	9	is	be	AUX
ajst-9439	30	10	a	a	DET
ajst-9439	30	11	specific	specific	ADJ
ajst-9439	30	12	application	application	NOUN
ajst-9439	30	13	of	of	ADP
ajst-9439	30	14	neural	neural	ADJ
ajst-9439	30	15	networks	network	NOUN
ajst-9439	30	16	,	,	PUNCT
ajst-9439	30	17	specifically	specifically	ADV
ajst-9439	30	18	studying	study	VERB
ajst-9439	30	19	various	various	ADJ
ajst-9439	30	20	theories	theory	NOUN
ajst-9439	30	21	and	and	CCONJ
ajst-9439	30	22	methods	method	NOUN
ajst-9439	30	23	for	for	ADP
ajst-9439	30	24	effective	effective	ADJ
ajst-9439	30	25	communication	communication	NOUN
ajst-9439	30	26	between	between	ADP
ajst-9439	30	27	humans	human	NOUN
ajst-9439	30	28	and	and	CCONJ
ajst-9439	30	29	computers	computer	NOUN
ajst-9439	30	30	using	use	VERB
ajst-9439	30	31	natural	natural	ADJ
ajst-9439	30	32	language	language	NOUN
ajst-9439	30	33	.	.	PUNCT
ajst-9439	31	1	wang	wang	PROPN
ajst-9439	31	2	canhui[5]and	canhui[5]and	PROPN
ajst-9439	31	3	others	other	NOUN
ajst-9439	31	4	summarized	summarize	VERB
ajst-9439	31	5	how	how	SCONJ
ajst-9439	31	6	to	to	PART
ajst-9439	31	7	use	use	VERB
ajst-9439	31	8	natural	natural	ADJ
ajst-9439	31	9	processing	processing	NOUN
ajst-9439	31	10	in	in	ADP
ajst-9439	31	11	information	information	NOUN
ajst-9439	31	12	retrieval	retrieval	NOUN
ajst-9439	31	13	,	,	PUNCT
ajst-9439	31	14	and	and	CCONJ
ajst-9439	31	15	natural	natural	ADJ
ajst-9439	31	16	language	language	NOUN
ajst-9439	31	17	processing	processing	NOUN
ajst-9439	31	18	needs	need	NOUN
ajst-9439	31	19	to	to	PART
ajst-9439	31	20	be	be	AUX
ajst-9439	31	21	optimized	optimize	VERB
ajst-9439	31	22	for	for	ADP
ajst-9439	31	23	information	information	NOUN
ajst-9439	31	24	retrieval	retrieval	NOUN
ajst-9439	31	25	to	to	PART
ajst-9439	31	26	play	play	VERB
ajst-9439	31	27	a	a	DET
ajst-9439	31	28	better	well	ADJ
ajst-9439	31	29	role	role	NOUN
ajst-9439	31	30	,	,	PUNCT
ajst-9439	31	31	while	while	SCONJ
ajst-9439	31	32	xi	xi	PROPN
ajst-9439	31	33	xuefeng[6]et	xuefeng[6]et	PROPN
ajst-9439	31	34	al	al	PROPN
ajst-9439	31	35	.	.	PUNCT
ajst-9439	32	1	the	the	DET
ajst-9439	32	2	research	research	NOUN
ajst-9439	32	3	was	be	AUX
ajst-9439	32	4	carried	carry	VERB
ajst-9439	32	5	out	out	ADP
ajst-9439	32	6	,	,	PUNCT
ajst-9439	32	7	focusing	focus	VERB
ajst-9439	32	8	on	on	ADP
ajst-9439	32	9	the	the	DET
ajst-9439	32	10	progress	progress	NOUN
ajst-9439	32	11	of	of	ADP
ajst-9439	32	12	natural	natural	ADJ
ajst-9439	32	13	language	language	NOUN
ajst-9439	32	14	processing	processing	NOUN
ajst-9439	32	15	in	in	ADP
ajst-9439	32	16	deep	deep	ADJ
ajst-9439	32	17	learning	learning	NOUN
ajst-9439	32	18	and	and	CCONJ
ajst-9439	32	19	its	its	PRON
ajst-9439	32	20	specific	specific	ADJ
ajst-9439	32	21	applications	application	NOUN
ajst-9439	32	22	.	.	PUNCT
ajst-9439	33	1	for	for	ADP
ajst-9439	33	2	natural	natural	ADJ
ajst-9439	33	3	language	language	NOUN
ajst-9439	33	4	processing	processing	NOUN
ajst-9439	33	5	,	,	PUNCT
ajst-9439	33	6	the	the	DET
ajst-9439	33	7	pre	pre	NOUN
ajst-9439	33	8	-	-	NOUN
ajst-9439	33	9	training	training	NOUN
ajst-9439	33	10	of	of	ADP
ajst-9439	33	11	the	the	DET
ajst-9439	33	12	model	model	NOUN
ajst-9439	33	13	is	be	AUX
ajst-9439	33	14	also	also	ADV
ajst-9439	33	15	necessary	necessary	ADJ
ajst-9439	33	16	.	.	PUNCT
ajst-9439	34	1	li	li	PROPN
ajst-9439	34	2	zhoujun[7]and	zhoujun[7]and	PROPN
ajst-9439	34	3	others	other	NOUN
ajst-9439	34	4	made	make	VERB
ajst-9439	34	5	a	a	DET
ajst-9439	34	6	good	good	ADJ
ajst-9439	34	7	summary	summary	NOUN
ajst-9439	34	8	of	of	ADP
ajst-9439	34	9	the	the	DET
ajst-9439	34	10	pretraining	pretraine	VERB
ajst-9439	34	11	technology	technology	NOUN
ajst-9439	34	12	of	of	ADP
ajst-9439	34	13	the	the	DET
ajst-9439	34	14	model	model	NOUN
ajst-9439	34	15	,	,	PUNCT
ajst-9439	34	16	and	and	CCONJ
ajst-9439	34	17	sorted	sort	VERB
ajst-9439	34	18	out	out	ADP
ajst-9439	34	19	a	a	DET
ajst-9439	34	20	series	series	NOUN
ajst-9439	34	21	of	of	ADP
ajst-9439	34	22	instructive	instructive	ADJ
ajst-9439	34	23	technologies	technology	NOUN
ajst-9439	34	24	,	,	PUNCT
ajst-9439	34	25	including	include	VERB
ajst-9439	34	26	the	the	DET
ajst-9439	34	27	improved	improved	ADJ
ajst-9439	34	28	model	model	NOUN
ajst-9439	34	29	of	of	ADP
ajst-9439	34	30	bert	bert	PROPN
ajst-9439	34	31	used	use	VERB
ajst-9439	34	32	today	today	NOUN
ajst-9439	34	33	.	.	PUNCT
ajst-9439	35	1	2.3	2.3	NUM
ajst-9439	35	2	.	.	PUNCT
ajst-9439	36	1	sentiment	sentiment	NOUN
ajst-9439	36	2	classification	classification	NOUN
ajst-9439	36	3	in	in	ADP
ajst-9439	36	4	the	the	DET
ajst-9439	36	5	early	early	ADJ
ajst-9439	36	6	stage	stage	NOUN
ajst-9439	36	7	of	of	ADP
ajst-9439	36	8	the	the	DET
ajst-9439	36	9	sentiment	sentiment	NOUN
ajst-9439	36	10	analysis	analysis	NOUN
ajst-9439	36	11	of	of	ADP
ajst-9439	36	12	internet	internet	NOUN
ajst-9439	36	13	comments	comment	NOUN
ajst-9439	36	14	,	,	PUNCT
ajst-9439	36	15	zhang	zhang	PROPN
ajst-9439	36	16	ziqiong	ziqiong	PROPN
ajst-9439	36	17	et	et	PROPN
ajst-9439	36	18	al.[8	al.[8	PROPN
ajst-9439	36	19	]	]	PUNCT
ajst-9439	36	20	conducted	conduct	VERB
ajst-9439	36	21	a	a	DET
ajst-9439	36	22	research	research	NOUN
ajst-9439	36	23	review	review	NOUN
ajst-9439	36	24	on	on	ADP
ajst-9439	36	25	it	it	PRON
ajst-9439	36	26	,	,	PUNCT
ajst-9439	36	27	and	and	CCONJ
ajst-9439	36	28	proposed	propose	VERB
ajst-9439	36	29	to	to	PART
ajst-9439	36	30	distinguish	distinguish	VERB
ajst-9439	36	31	between	between	ADP
ajst-9439	36	32	subjective	subjective	ADJ
ajst-9439	36	33	and	and	CCONJ
ajst-9439	36	34	objective	objective	ADJ
ajst-9439	36	35	text	text	NOUN
ajst-9439	36	36	information	information	NOUN
ajst-9439	36	37	,	,	PUNCT
ajst-9439	36	38	and	and	CCONJ
ajst-9439	36	39	to	to	PART
ajst-9439	36	40	carry	carry	VERB
ajst-9439	36	41	out	out	ADP
ajst-9439	36	42	emotional	emotional	ADJ
ajst-9439	36	43	binary	binary	ADJ
ajst-9439	36	44	classification	classification	NOUN
ajst-9439	36	45	.	.	PUNCT
ajst-9439	37	1	now	now	ADV
ajst-9439	37	2	text	text	NOUN
ajst-9439	37	3	sentiment	sentiment	NOUN
ajst-9439	37	4	analysis	analysis	NOUN
ajst-9439	37	5	is	be	AUX
ajst-9439	37	6	generally	generally	ADV
ajst-9439	37	7	summarized	summarize	VERB
ajst-9439	37	8	into	into	ADP
ajst-9439	37	9	three	three	NUM
ajst-9439	37	10	main	main	ADJ
ajst-9439	37	11	tasks	task	NOUN
ajst-9439	37	12	,	,	PUNCT
ajst-9439	37	13	namely	namely	ADV
ajst-9439	37	14	,	,	PUNCT
ajst-9439	37	15	emotional	emotional	ADJ
ajst-9439	37	16	information	information	NOUN
ajst-9439	37	17	extraction	extraction	NOUN
ajst-9439	37	18	,	,	PUNCT
ajst-9439	37	19	emotional	emotional	ADJ
ajst-9439	37	20	information	information	NOUN
ajst-9439	37	21	classification	classification	NOUN
ajst-9439	37	22	,	,	PUNCT
ajst-9439	37	23	and	and	CCONJ
ajst-9439	37	24	emotional	emotional	ADJ
ajst-9439	37	25	information	information	NOUN
ajst-9439	37	26	retrieval	retrieval	NOUN
ajst-9439	37	27	and	and	CCONJ
ajst-9439	37	28	induction	induction	NOUN
ajst-9439	37	29	[	[	X
ajst-9439	37	30	9	9	NUM
ajst-9439	37	31	]	]	PUNCT
ajst-9439	37	32	.	.	PUNCT
ajst-9439	38	1	in	in	ADP
ajst-9439	38	2	the	the	DET
ajst-9439	38	3	past	past	ADJ
ajst-9439	38	4	few	few	ADJ
ajst-9439	38	5	years	year	NOUN
ajst-9439	38	6	,	,	PUNCT
ajst-9439	38	7	there	there	PRON
ajst-9439	38	8	has	have	AUX
ajst-9439	38	9	been	be	AUX
ajst-9439	38	10	a	a	DET
ajst-9439	38	11	tendency	tendency	NOUN
ajst-9439	38	12	to	to	PART
ajst-9439	38	13	use	use	VERB
ajst-9439	38	14	deep	deep	ADJ
ajst-9439	38	15	learning	learning	NOUN
ajst-9439	38	16	technology	technology	NOUN
ajst-9439	38	17	to	to	PART
ajst-9439	38	18	solve	solve	VERB
ajst-9439	38	19	text	text	NOUN
ajst-9439	38	20	classification	classification	NOUN
ajst-9439	38	21	problems	problem	NOUN
ajst-9439	38	22	.	.	PUNCT
ajst-9439	39	1	chen	chen	PROPN
ajst-9439	39	2	long	long	PROPN
ajst-9439	39	3	et	et	PROPN
ajst-9439	39	4	al[10	al[10	PROPN
ajst-9439	39	5	]	]	PUNCT
ajst-9439	39	6	summarized	summarize	VERB
ajst-9439	39	7	traditional	traditional	ADJ
ajst-9439	39	8	sentiment	sentiment	NOUN
ajst-9439	39	9	classification	classification	NOUN
ajst-9439	39	10	methods	method	NOUN
ajst-9439	39	11	,	,	PUNCT
ajst-9439	39	12	including	include	VERB
ajst-9439	39	13	but	but	CCONJ
ajst-9439	39	14	not	not	PART
ajst-9439	39	15	limited	limit	VERB
ajst-9439	39	16	to	to	ADP
ajst-9439	39	17	methods	method	NOUN
ajst-9439	39	18	based	base	VERB
ajst-9439	39	19	on	on	ADP
ajst-9439	39	20	sentiment	sentiment	NOUN
ajst-9439	39	21	dictionaries	dictionary	NOUN
ajst-9439	39	22	,	,	PUNCT
ajst-9439	39	23	machine	machine	NOUN
ajst-9439	39	24	learning	learning	NOUN
ajst-9439	39	25	algorithms	algorithm	NOUN
ajst-9439	39	26	,	,	PUNCT
ajst-9439	39	27	or	or	CCONJ
ajst-9439	39	28	both	both	PRON
ajst-9439	39	29	.	.	PUNCT
ajst-9439	40	1	some	some	DET
ajst-9439	40	2	methods	method	NOUN
ajst-9439	40	3	,	,	PUNCT
ajst-9439	40	4	and	and	CCONJ
ajst-9439	40	5	methods	method	NOUN
ajst-9439	40	6	based	base	VERB
ajst-9439	40	7	on	on	ADP
ajst-9439	40	8	deep	deep	ADJ
ajst-9439	40	9	learning	learning	NOUN
ajst-9439	40	10	.	.	PUNCT
ajst-9439	41	1	whereas	whereas	SCONJ
ajst-9439	41	2	in	in	ADP
ajst-9439	41	3	this	this	DET
ajst-9439	41	4	paper	paper	NOUN
ajst-9439	41	5	,	,	PUNCT
ajst-9439	41	6	we	we	PRON
ajst-9439	41	7	have	have	AUX
ajst-9439	41	8	used	use	VERB
ajst-9439	41	9	a	a	DET
ajst-9439	41	10	deep	deep	ADJ
ajst-9439	41	11	learning	learning	NOUN
ajst-9439	41	12	based	base	VERB
ajst-9439	41	13	approach	approach	NOUN
ajst-9439	41	14	using	use	VERB
ajst-9439	41	15	bert[11	bert[11	X
ajst-9439	41	16	]	]	PUNCT
ajst-9439	41	17	for	for	ADP
ajst-9439	41	18	the	the	DET
ajst-9439	41	19	application	application	NOUN
ajst-9439	41	20	of	of	ADP
ajst-9439	41	21	sentiment	sentiment	NOUN
ajst-9439	41	22	classification	classification	NOUN
ajst-9439	41	23	.	.	PUNCT
ajst-9439	42	1	3	3	X
ajst-9439	42	2	.	.	X
ajst-9439	42	3	methods	method	NOUN
ajst-9439	42	4	we	we	PRON
ajst-9439	42	5	need	need	VERB
ajst-9439	42	6	to	to	PART
ajst-9439	42	7	know	know	VERB
ajst-9439	42	8	the	the	DET
ajst-9439	42	9	concept	concept	NOUN
ajst-9439	42	10	of	of	ADP
ajst-9439	42	11	feature	feature	NOUN
ajst-9439	42	12	first	first	ADV
ajst-9439	42	13	,	,	PUNCT
ajst-9439	42	14	which	which	PRON
ajst-9439	42	15	is	be	AUX
ajst-9439	42	16	an	an	DET
ajst-9439	42	17	independently	independently	ADV
ajst-9439	42	18	observable	observable	ADJ
ajst-9439	42	19	attribute	attribute	NOUN
ajst-9439	42	20	or	or	CCONJ
ajst-9439	42	21	characteristic	characteristic	NOUN
ajst-9439	42	22	of	of	ADP
ajst-9439	42	23	the	the	DET
ajst-9439	42	24	observed	observe	VERB
ajst-9439	42	25	object	object	NOUN
ajst-9439	42	26	.	.	PUNCT
ajst-9439	43	1	for	for	ADP
ajst-9439	43	2	example	example	NOUN
ajst-9439	43	3	,	,	PUNCT
ajst-9439	43	4	we	we	PRON
ajst-9439	43	5	need	need	VERB
ajst-9439	43	6	to	to	PART
ajst-9439	43	7	identify	identify	VERB
ajst-9439	43	8	the	the	DET
ajst-9439	43	9	category	category	NOUN
ajst-9439	43	10	of	of	ADP
ajst-9439	43	11	wine	wine	NOUN
ajst-9439	43	12	,	,	PUNCT
ajst-9439	43	13	and	and	CCONJ
ajst-9439	43	14	the	the	DET
ajst-9439	43	15	characteristics	characteristic	NOUN
ajst-9439	43	16	we	we	PRON
ajst-9439	43	17	need	need	VERB
ajst-9439	43	18	to	to	PART
ajst-9439	43	19	obtain	obtain	VERB
ajst-9439	43	20	are	be	AUX
ajst-9439	43	21	:	:	PUNCT
ajst-9439	43	22	brand	brand	NOUN
ajst-9439	43	23	,	,	PUNCT
ajst-9439	43	24	color	color	NOUN
ajst-9439	43	25	,	,	PUNCT
ajst-9439	43	26	taste	taste	NOUN
ajst-9439	43	27	,	,	PUNCT
ajst-9439	43	28	etc	etc	X
ajst-9439	43	29	.	.	X
ajst-9439	44	1	this	this	DET
ajst-9439	44	2	article	article	NOUN
ajst-9439	44	3	uses	use	VERB
ajst-9439	44	4	a	a	DET
ajst-9439	44	5	fully	fully	ADV
ajst-9439	44	6	connected	connect	VERB
ajst-9439	44	7	neural	neural	ADJ
ajst-9439	44	8	network	network	NOUN
ajst-9439	44	9	for	for	ADP
ajst-9439	44	10	calculations	calculation	NOUN
ajst-9439	44	11	,	,	PUNCT
ajst-9439	44	12	which	which	PRON
ajst-9439	44	13	is	be	AUX
ajst-9439	44	14	one	one	NUM
ajst-9439	44	15	of	of	ADP
ajst-9439	44	16	the	the	DET
ajst-9439	44	17	most	most	ADV
ajst-9439	44	18	basic	basic	ADJ
ajst-9439	44	19	network	network	NOUN
ajst-9439	44	20	types	type	NOUN
ajst-9439	44	21	in	in	ADP
ajst-9439	44	22	deep	deep	ADJ
ajst-9439	44	23	learning	learning	NOUN
ajst-9439	44	24	.	.	PUNCT
ajst-9439	45	1	the	the	DET
ajst-9439	45	2	fully	fully	ADV
ajst-9439	45	3	connected	connected	ADJ
ajst-9439	45	4	neural	neural	ADJ
ajst-9439	45	5	network	network	NOUN
ajst-9439	45	6	has	have	VERB
ajst-9439	45	7	an	an	DET
ajst-9439	45	8	input	input	NOUN
ajst-9439	45	9	layer	layer	NOUN
ajst-9439	45	10	,	,	PUNCT
ajst-9439	45	11	a	a	DET
ajst-9439	45	12	hidden	hidden	ADJ
ajst-9439	45	13	layer	layer	NOUN
ajst-9439	45	14	,	,	PUNCT
ajst-9439	45	15	and	and	CCONJ
ajst-9439	45	16	an	an	DET
ajst-9439	45	17	output	output	NOUN
ajst-9439	45	18	layer	layer	NOUN
ajst-9439	45	19	,	,	PUNCT
ajst-9439	45	20	which	which	PRON
ajst-9439	45	21	perform	perform	VERB
ajst-9439	45	22	input	input	NOUN
ajst-9439	45	23	,	,	PUNCT
ajst-9439	45	24	calculation	calculation	NOUN
ajst-9439	45	25	,	,	PUNCT
ajst-9439	45	26	and	and	CCONJ
ajst-9439	45	27	output	output	NOUN
ajst-9439	45	28	respectively	respectively	ADV
ajst-9439	45	29	.	.	PUNCT
ajst-9439	46	1	figure	figure	VERB
ajst-9439	46	2	3	3	NUM
ajst-9439	46	3	.	.	PUNCT
ajst-9439	46	4	simple	simple	ADJ
ajst-9439	46	5	operation	operation	NOUN
ajst-9439	46	6	diagram	diagram	NOUN
ajst-9439	46	7	fully	fully	ADV
ajst-9439	46	8	connected	connect	VERB
ajst-9439	46	9	neural	neural	ADJ
ajst-9439	46	10	network	network	NOUN
ajst-9439	46	11	(	(	PUNCT
ajst-9439	46	12	as	as	SCONJ
ajst-9439	46	13	shown	show	VERB
ajst-9439	46	14	in	in	ADP
ajst-9439	46	15	figure	figure	NOUN
ajst-9439	46	16	2	2	NUM
ajst-9439	46	17	):	):	PUNCT
ajst-9439	46	18	input	input	NOUN
ajst-9439	46	19	on	on	ADP
ajst-9439	46	20	the	the	DET
ajst-9439	46	21	left	left	ADJ
ajst-9439	46	22	,	,	PUNCT
ajst-9439	46	23	calculation	calculation	NOUN
ajst-9439	46	24	in	in	ADP
ajst-9439	46	25	the	the	DET
ajst-9439	46	26	middle	middle	ADJ
ajst-9439	46	27	,	,	PUNCT
ajst-9439	46	28	output	output	NOUN
ajst-9439	46	29	on	on	ADP
ajst-9439	46	30	the	the	DET
ajst-9439	46	31	right	right	NOUN
ajst-9439	46	32	.	.	PUNCT
ajst-9439	47	1	-30.00	-30.00	NOUN
ajst-9439	47	2	%	%	NOUN
ajst-9439	47	3	-20.00	-20.00	NOUN
ajst-9439	47	4	%	%	NOUN
ajst-9439	47	5	-10.00	-10.00	NUM
ajst-9439	47	6	%	%	NOUN
ajst-9439	47	7	0.00	0.00	NUM
ajst-9439	47	8	%	%	NOUN
ajst-9439	47	9	10.00	10.00	NUM
ajst-9439	47	10	%	%	NOUN
ajst-9439	47	11	20.00	20.00	NUM
ajst-9439	47	12	%	%	NOUN
ajst-9439	47	13	30.00	30.00	NUM
ajst-9439	47	14	%	%	NOUN
ajst-9439	47	15	40.00	40.00	NUM
ajst-9439	47	16	%	%	NOUN
ajst-9439	47	17	50.00	50.00	NUM
ajst-9439	47	18	%	%	NOUN
ajst-9439	47	19	0	0	NUM
ajst-9439	47	20	500	500	NUM
ajst-9439	47	21	1000	1000	NUM
ajst-9439	47	22	1500	1500	NUM
ajst-9439	47	23	2000	2000	NUM
ajst-9439	47	24	2500	2500	NUM
ajst-9439	47	25	3000	3000	NUM
ajst-9439	47	26	3500	3500	NUM
ajst-9439	47	27	4000	4000	NUM
ajst-9439	47	28	4500	4500	NUM
ajst-9439	47	29	2017	2017	NUM
ajst-9439	47	30	2018	2018	NUM
ajst-9439	47	31	2019	2019	NUM
ajst-9439	47	32	2020	2020	NUM
ajst-9439	47	33	2021	2021	NUM
ajst-9439	47	34	turnover(million	turnover(million	NOUN
ajst-9439	47	35	)	)	PUNCT
ajst-9439	47	36	year	year	NOUN
ajst-9439	47	37	-	-	PUNCT
ajst-9439	47	38	on	on	ADP
ajst-9439	47	39	-	-	PUNCT
ajst-9439	47	40	year	year	NOUN
ajst-9439	47	41	growth	growth	NOUN
ajst-9439	47	42	28	28	NUM
ajst-9439	47	43	figure	figure	NOUN
ajst-9439	47	44	4	4	NUM
ajst-9439	47	45	.	.	PUNCT
ajst-9439	47	46	schematic	schematic	ADJ
ajst-9439	47	47	diagram	diagram	NOUN
ajst-9439	47	48	of	of	ADP
ajst-9439	47	49	a	a	DET
ajst-9439	47	50	fully	fully	ADV
ajst-9439	47	51	connected	connected	ADJ
ajst-9439	47	52	neural	neural	ADJ
ajst-9439	47	53	network	network	NOUN
ajst-9439	47	54	in	in	ADP
ajst-9439	47	55	order	order	NOUN
ajst-9439	47	56	to	to	PART
ajst-9439	47	57	solve	solve	VERB
ajst-9439	47	58	the	the	DET
ajst-9439	47	59	problem	problem	NOUN
ajst-9439	47	60	of	of	ADP
ajst-9439	47	61	sentiment	sentiment	NOUN
ajst-9439	47	62	classification	classification	NOUN
ajst-9439	47	63	,	,	PUNCT
ajst-9439	47	64	we	we	PRON
ajst-9439	47	65	need	need	VERB
ajst-9439	47	66	to	to	PART
ajst-9439	47	67	extract	extract	VERB
ajst-9439	47	68	the	the	DET
ajst-9439	47	69	features	feature	NOUN
ajst-9439	47	70	of	of	ADP
ajst-9439	47	71	each	each	DET
ajst-9439	47	72	piece	piece	NOUN
ajst-9439	47	73	of	of	ADP
ajst-9439	47	74	data	datum	NOUN
ajst-9439	47	75	and	and	CCONJ
ajst-9439	47	76	classify	classify	VERB
ajst-9439	47	77	them	they	PRON
ajst-9439	47	78	,	,	PUNCT
ajst-9439	47	79	so	so	ADV
ajst-9439	47	80	we	we	PRON
ajst-9439	47	81	first	first	ADV
ajst-9439	47	82	use	use	VERB
ajst-9439	47	83	the	the	DET
ajst-9439	47	84	pre	pre	ADJ
ajst-9439	47	85	-	-	ADJ
ajst-9439	47	86	trained	train	VERB
ajst-9439	47	87	model	model	NOUN
ajst-9439	47	88	to	to	PART
ajst-9439	47	89	do	do	VERB
ajst-9439	47	90	a	a	DET
ajst-9439	47	91	calculation	calculation	NOUN
ajst-9439	47	92	,	,	PUNCT
ajst-9439	47	93	extract	extract	VERB
ajst-9439	47	94	the	the	DET
ajst-9439	47	95	features	feature	NOUN
ajst-9439	47	96	in	in	ADP
ajst-9439	47	97	the	the	DET
ajst-9439	47	98	data	datum	NOUN
ajst-9439	47	99	,	,	PUNCT
ajst-9439	47	100	and	and	CCONJ
ajst-9439	47	101	put	put	VERB
ajst-9439	47	102	the	the	DET
ajst-9439	47	103	extracted	extract	VERB
ajst-9439	47	104	features	feature	NOUN
ajst-9439	47	105	into	into	ADP
ajst-9439	47	106	the	the	DET
ajst-9439	47	107	fully	fully	ADV
ajst-9439	47	108	connected	connected	ADJ
ajst-9439	47	109	network	network	NOUN
ajst-9439	47	110	go	go	VERB
ajst-9439	47	111	to	to	ADP
ajst-9439	47	112	the	the	DET
ajst-9439	47	113	calculation	calculation	NOUN
ajst-9439	47	114	,	,	PUNCT
ajst-9439	47	115	and	and	CCONJ
ajst-9439	47	116	the	the	DET
ajst-9439	47	117	result	result	NOUN
ajst-9439	47	118	after	after	SCONJ
ajst-9439	47	119	the	the	DET
ajst-9439	47	120	calculation	calculation	NOUN
ajst-9439	47	121	will	will	AUX
ajst-9439	47	122	form	form	VERB
ajst-9439	47	123	a	a	DET
ajst-9439	47	124	2category	2category	NUM
ajst-9439	47	125	result	result	NOUN
ajst-9439	47	126	of	of	ADP
ajst-9439	47	127	16	16	NUM
ajst-9439	47	128	sentences	sentence	NOUN
ajst-9439	47	129	.	.	PUNCT
ajst-9439	48	1	the	the	DET
ajst-9439	48	2	optimizer	optimizer	NOUN
ajst-9439	48	3	can	can	AUX
ajst-9439	48	4	adjust	adjust	VERB
ajst-9439	48	5	the	the	DET
ajst-9439	48	6	parameters	parameter	NOUN
ajst-9439	48	7	of	of	ADP
ajst-9439	48	8	each	each	DET
ajst-9439	48	9	hidden	hide	VERB
ajst-9439	48	10	layer	layer	NOUN
ajst-9439	48	11	of	of	ADP
ajst-9439	48	12	the	the	DET
ajst-9439	48	13	loss	loss	NOUN
ajst-9439	48	14	function	function	NOUN
ajst-9439	48	15	in	in	ADP
ajst-9439	48	16	the	the	DET
ajst-9439	48	17	backpropagation	backpropagation	NOUN
ajst-9439	48	18	link	link	NOUN
ajst-9439	48	19	of	of	ADP
ajst-9439	48	20	deep	deep	ADJ
ajst-9439	48	21	learning	learning	NOUN
ajst-9439	48	22	,	,	PUNCT
ajst-9439	48	23	and	and	CCONJ
ajst-9439	48	24	the	the	DET
ajst-9439	48	25	changed	change	VERB
ajst-9439	48	26	parameters	parameter	NOUN
ajst-9439	48	27	will	will	AUX
ajst-9439	48	28	cause	cause	VERB
ajst-9439	48	29	the	the	DET
ajst-9439	48	30	value	value	NOUN
ajst-9439	48	31	of	of	ADP
ajst-9439	48	32	the	the	DET
ajst-9439	48	33	loss	loss	NOUN
ajst-9439	48	34	function	function	NOUN
ajst-9439	48	35	to	to	PART
ajst-9439	48	36	continuously	continuously	ADV
ajst-9439	48	37	approach	approach	VERB
ajst-9439	48	38	the	the	DET
ajst-9439	48	39	global	global	ADJ
ajst-9439	48	40	minimum	minimum	NOUN
ajst-9439	48	41	.	.	PUNCT
ajst-9439	49	1	this	this	PRON
ajst-9439	49	2	is	be	AUX
ajst-9439	49	3	also	also	ADV
ajst-9439	49	4	essential	essential	ADJ
ajst-9439	49	5	for	for	ADP
ajst-9439	49	6	deep	deep	ADJ
ajst-9439	49	7	learning	learning	NOUN
ajst-9439	49	8	.	.	PUNCT
ajst-9439	50	1	for	for	ADP
ajst-9439	50	2	the	the	DET
ajst-9439	50	3	optimizer	optimizer	NOUN
ajst-9439	50	4	,	,	PUNCT
ajst-9439	50	5	we	we	PRON
ajst-9439	50	6	chose	choose	VERB
ajst-9439	50	7	the	the	DET
ajst-9439	50	8	improved	improved	ADJ
ajst-9439	50	9	version	version	NOUN
ajst-9439	50	10	of	of	ADP
ajst-9439	50	11	the	the	DET
ajst-9439	50	12	adam	adam	PROPN
ajst-9439	50	13	optimizer	optimizer	NOUN
ajst-9439	50	14	,	,	PUNCT
ajst-9439	50	15	adamw	adamw	NOUN
ajst-9439	50	16	optimizer	optimizer	NOUN
ajst-9439	50	17	,	,	PUNCT
ajst-9439	50	18	which	which	PRON
ajst-9439	50	19	has	have	VERB
ajst-9439	50	20	the	the	DET
ajst-9439	50	21	advantages	advantage	NOUN
ajst-9439	50	22	of	of	ADP
ajst-9439	50	23	the	the	DET
ajst-9439	50	24	oadagrad	oadagrad	NOUN
ajst-9439	50	25	algorithm	algorithm	NOUN
ajst-9439	50	26	and	and	CCONJ
ajst-9439	50	27	the	the	DET
ajst-9439	50	28	rmsprop	rmsprop	NOUN
ajst-9439	50	29	algorithm	algorithm	NOUN
ajst-9439	50	30	.	.	PUNCT
ajst-9439	51	1	figure	figure	NOUN
ajst-9439	51	2	5	5	NUM
ajst-9439	51	3	.	.	PUNCT
ajst-9439	52	1	adam	adam	NOUN
ajst-9439	52	2	with	with	ADP
ajst-9439	52	3	l2	l2	NOUN
ajst-9439	52	4	regularization	regularization	NOUN
ajst-9439	52	5	and	and	CCONJ
ajst-9439	52	6	adam	adam	NOUN
ajst-9439	52	7	with	with	ADP
ajst-9439	52	8	decoupled	decouple	VERB
ajst-9439	52	9	weight	weight	NOUN
ajst-9439	52	10	decay(adamw	decay(adamw	PROPN
ajst-9439	52	11	)	)	PUNCT
ajst-9439	52	12	4	4	NUM
ajst-9439	52	13	.	.	X
ajst-9439	52	14	experiment	experiment	NOUN
ajst-9439	52	15	4.1	4.1	NUM
ajst-9439	52	16	.	.	PUNCT
ajst-9439	53	1	experimental	experimental	ADJ
ajst-9439	53	2	details	detail	NOUN
ajst-9439	53	3	and	and	CCONJ
ajst-9439	53	4	design	design	NOUN
ajst-9439	53	5	for	for	ADP
ajst-9439	53	6	deep	deep	ADJ
ajst-9439	53	7	learning	learning	NOUN
ajst-9439	53	8	,	,	PUNCT
ajst-9439	53	9	model	model	NOUN
ajst-9439	53	10	pre	pre	NOUN
ajst-9439	53	11	-	-	NOUN
ajst-9439	53	12	training	training	NOUN
ajst-9439	53	13	is	be	AUX
ajst-9439	53	14	essential	essential	ADJ
ajst-9439	53	15	.	.	PUNCT
ajst-9439	54	1	broadly	broadly	ADV
ajst-9439	54	2	speaking	speak	VERB
ajst-9439	54	3	,	,	PUNCT
ajst-9439	54	4	pre	pre	ADJ
ajst-9439	54	5	-	-	ADJ
ajst-9439	54	6	training	training	NOUN
ajst-9439	54	7	refers	refer	VERB
ajst-9439	54	8	to	to	ADP
ajst-9439	54	9	using	use	VERB
ajst-9439	54	10	a	a	DET
ajst-9439	54	11	large	large	ADJ
ajst-9439	54	12	amount	amount	NOUN
ajst-9439	54	13	of	of	ADP
ajst-9439	54	14	training	training	NOUN
ajst-9439	54	15	data	datum	NOUN
ajst-9439	54	16	to	to	PART
ajst-9439	54	17	train	train	VERB
ajst-9439	54	18	and	and	CCONJ
ajst-9439	54	19	adjust	adjust	VERB
ajst-9439	54	20	the	the	DET
ajst-9439	54	21	model	model	NOUN
ajst-9439	54	22	to	to	PART
ajst-9439	54	23	be	be	AUX
ajst-9439	54	24	trained	train	VERB
ajst-9439	54	25	,	,	PUNCT
ajst-9439	54	26	extracting	extract	VERB
ajst-9439	54	27	a	a	DET
ajst-9439	54	28	large	large	ADJ
ajst-9439	54	29	number	number	NOUN
ajst-9439	54	30	of	of	ADP
ajst-9439	54	31	common	common	ADJ
ajst-9439	54	32	features	feature	NOUN
ajst-9439	54	33	from	from	ADP
ajst-9439	54	34	it	it	PRON
ajst-9439	54	35	,	,	PUNCT
ajst-9439	54	36	so	so	SCONJ
ajst-9439	54	37	that	that	SCONJ
ajst-9439	54	38	the	the	DET
ajst-9439	54	39	learning	learning	NOUN
ajst-9439	54	40	burden	burden	NOUN
ajst-9439	54	41	of	of	ADP
ajst-9439	54	42	the	the	DET
ajst-9439	54	43	used	used	ADJ
ajst-9439	54	44	model	model	NOUN
ajst-9439	54	45	on	on	ADP
ajst-9439	54	46	specific	specific	ADJ
ajst-9439	54	47	tasks	task	NOUN
ajst-9439	54	48	is	be	AUX
ajst-9439	54	49	reduced	reduce	VERB
ajst-9439	54	50	,	,	PUNCT
ajst-9439	54	51	and	and	CCONJ
ajst-9439	54	52	finally	finally	ADV
ajst-9439	54	53	in	in	ADP
ajst-9439	54	54	the	the	DET
ajst-9439	54	55	downstream	downstream	ADJ
ajst-9439	54	56	tasks	task	NOUN
ajst-9439	54	57	fine	fine	ADV
ajst-9439	54	58	-	-	PUNCT
ajst-9439	54	59	tuning	tuning	NOUN
ajst-9439	54	60	of	of	ADP
ajst-9439	54	61	the	the	DET
ajst-9439	54	62	data	datum	NOUN
ajst-9439	54	63	.	.	PUNCT
ajst-9439	55	1	in	in	ADP
ajst-9439	55	2	this	this	DET
ajst-9439	55	3	article	article	NOUN
ajst-9439	55	4	,	,	PUNCT
ajst-9439	55	5	we	we	PRON
ajst-9439	55	6	only	only	ADV
ajst-9439	55	7	get	get	VERB
ajst-9439	55	8	text	text	NOUN
ajst-9439	55	9	and	and	CCONJ
ajst-9439	55	10	labels	label	NOUN
ajst-9439	55	11	for	for	ADP
ajst-9439	55	12	the	the	DET
ajst-9439	55	13	data	datum	NOUN
ajst-9439	55	14	,	,	PUNCT
ajst-9439	55	15	and	and	CCONJ
ajst-9439	55	16	use	use	VERB
ajst-9439	55	17	the	the	DET
ajst-9439	55	18	pre	pre	ADJ
ajst-9439	55	19	-	-	ADJ
ajst-9439	55	20	training	training	ADJ
ajst-9439	55	21	model	model	NOUN
ajst-9439	55	22	directly	directly	ADV
ajst-9439	55	23	.	.	PUNCT
ajst-9439	56	1	the	the	DET
ajst-9439	56	2	pre	pre	ADJ
ajst-9439	56	3	-	-	ADJ
ajst-9439	56	4	training	training	ADJ
ajst-9439	56	5	model	model	NOUN
ajst-9439	56	6	chooses	choose	VERB
ajst-9439	56	7	the	the	DET
ajst-9439	56	8	classic	classic	ADJ
ajst-9439	56	9	bert	bert	PROPN
ajst-9439	56	10	model	model	NOUN
ajst-9439	56	11	.	.	PUNCT
ajst-9439	57	1	for	for	ADP
ajst-9439	57	2	deep	deep	ADJ
ajst-9439	57	3	learning	learning	NOUN
ajst-9439	57	4	,	,	PUNCT
ajst-9439	57	5	model	model	NOUN
ajst-9439	57	6	pretraining	pretraine	VERB
ajst-9439	57	7	is	be	AUX
ajst-9439	57	8	essential	essential	ADJ
ajst-9439	57	9	.	.	PUNCT
ajst-9439	58	1	broadly	broadly	ADV
ajst-9439	58	2	speaking	speak	VERB
ajst-9439	58	3	,	,	PUNCT
ajst-9439	58	4	pre	pre	ADJ
ajst-9439	58	5	-	-	ADJ
ajst-9439	58	6	training	training	NOUN
ajst-9439	58	7	refers	refer	VERB
ajst-9439	58	8	to	to	ADP
ajst-9439	58	9	using	use	VERB
ajst-9439	58	10	a	a	DET
ajst-9439	58	11	large	large	ADJ
ajst-9439	58	12	amount	amount	NOUN
ajst-9439	58	13	of	of	ADP
ajst-9439	58	14	training	training	NOUN
ajst-9439	58	15	data	datum	NOUN
ajst-9439	58	16	to	to	PART
ajst-9439	58	17	train	train	VERB
ajst-9439	58	18	and	and	CCONJ
ajst-9439	58	19	adjust	adjust	VERB
ajst-9439	58	20	the	the	DET
ajst-9439	58	21	model	model	NOUN
ajst-9439	58	22	to	to	PART
ajst-9439	58	23	be	be	AUX
ajst-9439	58	24	trained	train	VERB
ajst-9439	58	25	,	,	PUNCT
ajst-9439	58	26	extracting	extract	VERB
ajst-9439	58	27	a	a	DET
ajst-9439	58	28	large	large	ADJ
ajst-9439	58	29	number	number	NOUN
ajst-9439	58	30	of	of	ADP
ajst-9439	58	31	common	common	ADJ
ajst-9439	58	32	features	feature	NOUN
ajst-9439	58	33	from	from	ADP
ajst-9439	58	34	it	it	PRON
ajst-9439	58	35	,	,	PUNCT
ajst-9439	58	36	so	so	SCONJ
ajst-9439	58	37	that	that	SCONJ
ajst-9439	58	38	the	the	DET
ajst-9439	58	39	learning	learning	NOUN
ajst-9439	58	40	burden	burden	NOUN
ajst-9439	58	41	of	of	ADP
ajst-9439	58	42	the	the	DET
ajst-9439	58	43	used	used	ADJ
ajst-9439	58	44	model	model	NOUN
ajst-9439	58	45	on	on	ADP
ajst-9439	58	46	specific	specific	ADJ
ajst-9439	58	47	tasks	task	NOUN
ajst-9439	58	48	is	be	AUX
ajst-9439	58	49	reduced	reduce	VERB
ajst-9439	58	50	,	,	PUNCT
ajst-9439	58	51	and	and	CCONJ
ajst-9439	58	52	finally	finally	ADV
ajst-9439	58	53	in	in	ADP
ajst-9439	58	54	the	the	DET
ajst-9439	58	55	downstream	downstream	ADJ
ajst-9439	58	56	tasks	task	NOUN
ajst-9439	58	57	fine	fine	ADV
ajst-9439	58	58	-	-	PUNCT
ajst-9439	58	59	tuning	tuning	NOUN
ajst-9439	58	60	of	of	ADP
ajst-9439	58	61	the	the	DET
ajst-9439	58	62	data	datum	NOUN
ajst-9439	58	63	.	.	PUNCT
ajst-9439	59	1	with	with	ADP
ajst-9439	59	2	the	the	DET
ajst-9439	59	3	increase	increase	NOUN
ajst-9439	59	4	of	of	ADP
ajst-9439	59	5	the	the	DET
ajst-9439	59	6	deep	deep	ADJ
ajst-9439	59	7	learning	learning	NOUN
ajst-9439	59	8	ability	ability	NOUN
ajst-9439	59	9	of	of	ADP
ajst-9439	59	10	the	the	DET
ajst-9439	59	11	model	model	NOUN
ajst-9439	59	12	,	,	PUNCT
ajst-9439	59	13	the	the	DET
ajst-9439	59	14	feature	feature	NOUN
ajst-9439	59	15	extraction	extraction	NOUN
ajst-9439	59	16	ability	ability	NOUN
ajst-9439	59	17	is	be	AUX
ajst-9439	59	18	improved	improve	VERB
ajst-9439	59	19	,	,	PUNCT
ajst-9439	59	20	and	and	CCONJ
ajst-9439	59	21	the	the	DET
ajst-9439	59	22	extraction	extraction	NOUN
ajst-9439	59	23	ability	ability	NOUN
ajst-9439	59	24	of	of	ADP
ajst-9439	59	25	specific	specific	ADJ
ajst-9439	59	26	features	feature	NOUN
ajst-9439	59	27	is	be	AUX
ajst-9439	59	28	increased	increase	VERB
ajst-9439	59	29	,	,	PUNCT
ajst-9439	59	30	so	so	SCONJ
ajst-9439	59	31	that	that	SCONJ
ajst-9439	59	32	the	the	DET
ajst-9439	59	33	model	model	NOUN
ajst-9439	59	34	can	can	AUX
ajst-9439	59	35	be	be	AUX
ajst-9439	59	36	generalized	generalize	VERB
ajst-9439	59	37	and	and	CCONJ
ajst-9439	59	38	widely	widely	ADV
ajst-9439	59	39	used	use	VERB
ajst-9439	59	40	.	.	PUNCT
ajst-9439	60	1	regarding	regard	VERB
ajst-9439	60	2	the	the	DET
ajst-9439	60	3	choice	choice	NOUN
ajst-9439	60	4	of	of	ADP
ajst-9439	60	5	optimizer	optimizer	NOUN
ajst-9439	60	6	,	,	PUNCT
ajst-9439	60	7	we	we	PRON
ajst-9439	60	8	use	use	VERB
ajst-9439	60	9	the	the	DET
ajst-9439	60	10	adamw	adamw	NOUN
ajst-9439	60	11	optimizer	optimizer	NOUN
ajst-9439	60	12	with	with	ADP
ajst-9439	60	13	an	an	DET
ajst-9439	60	14	initial	initial	ADJ
ajst-9439	60	15	learning	learning	NOUN
ajst-9439	60	16	rate	rate	NOUN
ajst-9439	60	17	of	of	ADP
ajst-9439	60	18	5e-4	5e-4	PROPN
ajst-9439	60	19	,	,	PUNCT
ajst-9439	60	20	and	and	CCONJ
ajst-9439	60	21	in	in	ADP
ajst-9439	60	22	order	order	NOUN
ajst-9439	60	23	to	to	PART
ajst-9439	60	24	make	make	VERB
ajst-9439	60	25	full	full	ADJ
ajst-9439	60	26	use	use	NOUN
ajst-9439	60	27	of	of	ADP
ajst-9439	60	28	our	our	PRON
ajst-9439	60	29	computing	computing	NOUN
ajst-9439	60	30	resources	resource	NOUN
ajst-9439	60	31	,	,	PUNCT
ajst-9439	60	32	we	we	PRON
ajst-9439	60	33	set	set	VERB
ajst-9439	60	34	the	the	DET
ajst-9439	60	35	batch	batch	NOUN
ajst-9439	60	36	size	size	NOUN
ajst-9439	60	37	to	to	ADP
ajst-9439	60	38	32	32	NUM
ajst-9439	60	39	.	.	PUNCT
ajst-9439	61	1	regarding	regard	VERB
ajst-9439	61	2	the	the	DET
ajst-9439	61	3	loss	loss	NOUN
ajst-9439	61	4	function	function	NOUN
ajst-9439	61	5	,	,	PUNCT
ajst-9439	61	6	we	we	PRON
ajst-9439	61	7	use	use	VERB
ajst-9439	61	8	the	the	DET
ajst-9439	61	9	loss	loss	NOUN
ajst-9439	61	10	function	function	NOUN
ajst-9439	61	11	crossentropyloss	crossentropyloss	NOUN
ajst-9439	61	12	commonly	commonly	ADV
ajst-9439	61	13	used	use	VERB
ajst-9439	61	14	in	in	ADP
ajst-9439	61	15	classification	classification	NOUN
ajst-9439	61	16	problems	problem	NOUN
ajst-9439	61	17	.	.	PUNCT
ajst-9439	62	1	all	all	DET
ajst-9439	62	2	experiments	experiment	NOUN
ajst-9439	62	3	are	be	AUX
ajst-9439	62	4	implemented	implement	VERB
ajst-9439	62	5	using	use	VERB
ajst-9439	62	6	kaggle	kaggle	NOUN
ajst-9439	62	7	on	on	ADP
ajst-9439	62	8	nvidia	nvidia	PROPN
ajst-9439	62	9	gtx1080ti	gtx1080ti	PROPN
ajst-9439	62	10	gpu	gpu	PROPN
ajst-9439	62	11	.	.	PROPN
ajst-9439	62	12	4.2	4.2	NUM
ajst-9439	62	13	.	.	PUNCT
ajst-9439	63	1	datasets	dataset	NOUN
ajst-9439	63	2	as	as	SCONJ
ajst-9439	63	3	shown	show	VERB
ajst-9439	63	4	in	in	ADP
ajst-9439	63	5	figure	figure	NOUN
ajst-9439	63	6	4	4	NUM
ajst-9439	63	7	,	,	PUNCT
ajst-9439	63	8	the	the	DET
ajst-9439	63	9	data	datum	NOUN
ajst-9439	63	10	set	set	VERB
ajst-9439	63	11	name	name	NOUN
ajst-9439	63	12	:	:	PUNCT
ajst-9439	63	13	bert	bert	PROPN
ajst-9439	63	14	hugging	hug	VERB
ajst-9439	63	15	face	face	NOUN
ajst-9439	63	16	dataset	dataset	NOUN
ajst-9439	63	17	,	,	PUNCT
ajst-9439	63	18	each	each	DET
ajst-9439	63	19	sentence	sentence	NOUN
ajst-9439	63	20	is	be	AUX
ajst-9439	63	21	an	an	DET
ajst-9439	63	22	evaluation	evaluation	NOUN
ajst-9439	63	23	of	of	ADP
ajst-9439	63	24	something	something	PRON
ajst-9439	63	25	,	,	PUNCT
ajst-9439	63	26	the	the	DET
ajst-9439	63	27	training	training	NOUN
ajst-9439	63	28	set	set	NOUN
ajst-9439	63	29	has	have	VERB
ajst-9439	63	30	a	a	DET
ajst-9439	63	31	total	total	NOUN
ajst-9439	63	32	of	of	ADP
ajst-9439	63	33	9600	9600	NUM
ajst-9439	63	34	evaluations	evaluation	NOUN
ajst-9439	63	35	,	,	PUNCT
ajst-9439	63	36	the	the	DET
ajst-9439	63	37	test	test	NOUN
ajst-9439	63	38	set	set	NOUN
ajst-9439	63	39	has	have	VERB
ajst-9439	63	40	a	a	DET
ajst-9439	63	41	total	total	NOUN
ajst-9439	63	42	of	of	ADP
ajst-9439	63	43	1200	1200	NUM
ajst-9439	63	44	evaluations	evaluation	NOUN
ajst-9439	63	45	,	,	PUNCT
ajst-9439	63	46	and	and	CCONJ
ajst-9439	63	47	the	the	DET
ajst-9439	63	48	verification	verification	NOUN
ajst-9439	63	49	set	set	NOUN
ajst-9439	63	50	has	have	VERB
ajst-9439	63	51	a	a	DET
ajst-9439	63	52	total	total	NOUN
ajst-9439	63	53	of	of	ADP
ajst-9439	63	54	1200	1200	NUM
ajst-9439	63	55	evaluations	evaluation	NOUN
ajst-9439	63	56	.	.	PUNCT
ajst-9439	64	1	a	a	DET
ajst-9439	64	2	total	total	NOUN
ajst-9439	64	3	of	of	ADP
ajst-9439	64	4	12,000	12,000	NUM
ajst-9439	64	5	entries	entry	NOUN
ajst-9439	64	6	.	.	PUNCT
ajst-9439	65	1	the	the	DET
ajst-9439	65	2	data	datum	NOUN
ajst-9439	65	3	in	in	ADP
ajst-9439	65	4	the	the	DET
ajst-9439	65	5	data	datum	NOUN
ajst-9439	65	6	set	set	VERB
ajst-9439	65	7	are	be	AUX
ajst-9439	65	8	all	all	PRON
ajst-9439	65	9	evaluations	evaluation	NOUN
ajst-9439	65	10	of	of	ADP
ajst-9439	65	11	things	thing	NOUN
ajst-9439	65	12	,	,	PUNCT
ajst-9439	65	13	such	such	ADJ
ajst-9439	65	14	as	as	ADP
ajst-9439	65	15	:	:	PUNCT
ajst-9439	65	16	(	(	PUNCT
ajst-9439	65	17	'	'	PUNCT
ajst-9439	65	18	original	original	ADJ
ajst-9439	65	19	goods	good	NOUN
ajst-9439	65	20	,	,	PUNCT
ajst-9439	65	21	when	when	SCONJ
ajst-9439	65	22	picking	pick	VERB
ajst-9439	65	23	up	up	ADP
ajst-9439	65	24	the	the	DET
ajst-9439	65	25	goods	good	NOUN
ajst-9439	65	26	,	,	PUNCT
ajst-9439	65	27	the	the	DET
ajst-9439	65	28	seals	seal	NOUN
ajst-9439	65	29	have	have	AUX
ajst-9439	65	30	not	not	PART
ajst-9439	65	31	been	be	AUX
ajst-9439	65	32	opened	open	VERB
ajst-9439	65	33	.	.	PUNCT
ajst-9439	66	1	because	because	SCONJ
ajst-9439	66	2	it	it	PRON
ajst-9439	66	3	is	be	AUX
ajst-9439	66	4	the	the	DET
ajst-9439	66	5	first	first	ADJ
ajst-9439	66	6	time	time	NOUN
ajst-9439	66	7	to	to	PART
ajst-9439	66	8	shop	shop	VERB
ajst-9439	66	9	in	in	ADP
ajst-9439	66	10	jingdong	jingdong	NOUN
ajst-9439	66	11	,	,	PUNCT
ajst-9439	66	12	pick	pick	VERB
ajst-9439	66	13	up	up	ADP
ajst-9439	66	14	the	the	DET
ajst-9439	66	15	goods	good	NOUN
ajst-9439	66	16	at	at	ADP
ajst-9439	66	17	the	the	DET
ajst-9439	66	18	door	door	NOUN
ajst-9439	66	19	,	,	PUNCT
ajst-9439	66	20	after	after	ADP
ajst-9439	66	21	seeing	see	VERB
ajst-9439	66	22	the	the	DET
ajst-9439	66	23	goods	good	NOUN
ajst-9439	66	24	,	,	PUNCT
ajst-9439	66	25	finally	finally	ADV
ajst-9439	66	26	do	do	AUX
ajst-9439	66	27	n't	not	PART
ajst-9439	66	28	worry	worry	VERB
ajst-9439	66	29	about	about	ADP
ajst-9439	66	30	the	the	DET
ajst-9439	66	31	quality	quality	NOUN
ajst-9439	66	32	.	.	PUNCT
ajst-9439	67	1	it	it	PRON
ajst-9439	67	2	's	be	AUX
ajst-9439	67	3	definitely	definitely	ADV
ajst-9439	67	4	a	a	DET
ajst-9439	67	5	genuine	genuine	ADJ
ajst-9439	67	6	product	product	NOUN
ajst-9439	67	7	.	.	PUNCT
ajst-9439	68	1	this	this	DET
ajst-9439	68	2	transaction	transaction	NOUN
ajst-9439	68	3	is	be	AUX
ajst-9439	68	4	very	very	ADV
ajst-9439	68	5	funny	funny	ADJ
ajst-9439	68	6	.	.	PUNCT
ajst-9439	69	1	i	i	PRON
ajst-9439	69	2	placed	place	VERB
ajst-9439	69	3	an	an	DET
ajst-9439	69	4	order	order	NOUN
ajst-9439	69	5	on	on	ADP
ajst-9439	69	6	the	the	DET
ajst-9439	69	7	evening	evening	NOUN
ajst-9439	69	8	of	of	ADP
ajst-9439	69	9	the	the	DET
ajst-9439	69	10	25th	25th	NOUN
ajst-9439	69	11	,	,	PUNCT
ajst-9439	69	12	and	and	CCONJ
ajst-9439	69	13	i	i	PRON
ajst-9439	69	14	will	will	AUX
ajst-9439	69	15	check	check	VERB
ajst-9439	69	16	it	it	PRON
ajst-9439	69	17	again	again	ADV
ajst-9439	69	18	in	in	ADP
ajst-9439	69	19	the	the	DET
ajst-9439	69	20	morning	morning	NOUN
ajst-9439	69	21	of	of	ADP
ajst-9439	69	22	the	the	DET
ajst-9439	69	23	26th	26th	NOUN
ajst-9439	69	24	.	.	PUNCT
ajst-9439	70	1	since	since	SCONJ
ajst-9439	70	2	i	i	PRON
ajst-9439	70	3	found	find	VERB
ajst-9439	70	4	out	out	ADP
ajst-9439	70	5	that	that	SCONJ
ajst-9439	70	6	there	there	PRON
ajst-9439	70	7	is	be	VERB
ajst-9439	70	8	a	a	DET
ajst-9439	70	9	new	new	ADJ
ajst-9439	70	10	promotion	promotion	NOUN
ajst-9439	70	11	with	with	ADP
ajst-9439	70	12	a	a	DET
ajst-9439	70	13	cashback	cashback	NOUN
ajst-9439	70	14	of	of	ADP
ajst-9439	70	15	100	100	NUM
ajst-9439	70	16	yuan	yuan	NOUN
ajst-9439	70	17	,	,	PUNCT
ajst-9439	70	18	fortunately	fortunately	ADV
ajst-9439	70	19	,	,	PUNCT
ajst-9439	70	20	i	i	PRON
ajst-9439	70	21	came	come	VERB
ajst-9439	70	22	to	to	PART
ajst-9439	70	23	pick	pick	VERB
ajst-9439	70	24	up	up	ADP
ajst-9439	70	25	the	the	DET
ajst-9439	70	26	goods	good	NOUN
ajst-9439	70	27	.	.	PUNCT
ajst-9439	71	1	immediately	immediately	ADV
ajst-9439	71	2	canceled	cancel	VERB
ajst-9439	71	3	the	the	DET
ajst-9439	71	4	order	order	NOUN
ajst-9439	71	5	on	on	ADP
ajst-9439	71	6	the	the	DET
ajst-9439	71	7	25th	25th	NOUN
ajst-9439	71	8	and	and	CCONJ
ajst-9439	71	9	placed	place	VERB
ajst-9439	71	10	another	another	DET
ajst-9439	71	11	order	order	NOUN
ajst-9439	71	12	,	,	PUNCT
ajst-9439	71	13	saving	save	VERB
ajst-9439	71	14	100	100	NUM
ajst-9439	71	15	yuan	yuan	NOUN
ajst-9439	71	16	.	.	PUNCT
ajst-9439	72	1	hehe	hehe	INTJ
ajst-9439	72	2	.	.	PUNCT
ajst-9439	72	3	'	'	PUNCT
ajst-9439	72	4	)	)	PUNCT
ajst-9439	72	5	29	29	NUM
ajst-9439	72	6	figure	figure	NOUN
ajst-9439	72	7	6	6	NUM
ajst-9439	72	8	.	.	PUNCT
ajst-9439	72	9	dataset	dataset	VERB
ajst-9439	72	10	4.3	4.3	NUM
ajst-9439	72	11	.	.	PUNCT
ajst-9439	72	12	machine	machine	NOUN
ajst-9439	72	13	learning	learn	VERB
ajst-9439	72	14	algorithms	algorithm	NOUN
ajst-9439	72	15	this	this	DET
ajst-9439	72	16	article	article	NOUN
ajst-9439	72	17	uses	use	VERB
ajst-9439	72	18	the	the	DET
ajst-9439	72	19	bert	bert	PROPN
ajst-9439	72	20	algorithm	algorithm	NOUN
ajst-9439	72	21	.	.	PUNCT
ajst-9439	73	1	the	the	DET
ajst-9439	73	2	bert	bert	PROPN
ajst-9439	73	3	algorithm	algorithm	PROPN
ajst-9439	73	4	is	be	AUX
ajst-9439	73	5	a	a	DET
ajst-9439	73	6	two	two	NUM
ajst-9439	73	7	-	-	PUNCT
ajst-9439	73	8	way	way	NOUN
ajst-9439	73	9	encoding	encoding	NOUN
ajst-9439	73	10	representation	representation	NOUN
ajst-9439	73	11	algorithm	algorithm	NOUN
ajst-9439	73	12	based	base	VERB
ajst-9439	73	13	on	on	ADP
ajst-9439	73	14	the	the	DET
ajst-9439	73	15	transformer	transformer	NOUN
ajst-9439	73	16	algorithm	algorithm	NOUN
ajst-9439	73	17	.	.	PUNCT
ajst-9439	74	1	the	the	DET
ajst-9439	74	2	transformer	transformer	ADJ
ajst-9439	74	3	algorithm	algorithm	NOUN
ajst-9439	74	4	is	be	AUX
ajst-9439	74	5	based	base	VERB
ajst-9439	74	6	on	on	ADP
ajst-9439	74	7	a	a	DET
ajst-9439	74	8	multi	multi	ADJ
ajst-9439	74	9	-	-	ADJ
ajst-9439	74	10	head	head	ADJ
ajst-9439	74	11	attention	attention	NOUN
ajst-9439	74	12	mechanism	mechanism	NOUN
ajst-9439	74	13	,	,	PUNCT
ajst-9439	74	14	and	and	CCONJ
ajst-9439	74	15	bert	bert	PROPN
ajst-9439	74	16	stacks	stack	VERB
ajst-9439	74	17	multiple	multiple	ADJ
ajst-9439	74	18	transformer	transformer	NOUN
ajst-9439	74	19	models	model	NOUN
ajst-9439	74	20	,	,	PUNCT
ajst-9439	74	21	and	and	CCONJ
ajst-9439	74	22	adjusts	adjust	VERB
ajst-9439	74	23	the	the	DET
ajst-9439	74	24	bidirectional	bidirectional	ADJ
ajst-9439	74	25	transformer	transformer	NOUN
ajst-9439	74	26	in	in	ADP
ajst-9439	74	27	all	all	DET
ajst-9439	74	28	layers	layer	NOUN
ajst-9439	74	29	.	.	PUNCT
ajst-9439	75	1	to	to	AUX
ajst-9439	75	2	pre	pre	VERB
ajst-9439	75	3	-	-	VERB
ajst-9439	75	4	train	train	VERB
ajst-9439	75	5	the	the	DET
ajst-9439	75	6	bidirectional	bidirectional	ADJ
ajst-9439	75	7	deep	deep	ADJ
ajst-9439	75	8	representation	representation	NOUN
ajst-9439	75	9	,	,	PUNCT
ajst-9439	75	10	and	and	CCONJ
ajst-9439	75	11	the	the	DET
ajst-9439	75	12	pre	pre	ADJ
ajst-9439	75	13	-	-	ADJ
ajst-9439	75	14	trained	train	VERB
ajst-9439	75	15	bert	bert	NOUN
ajst-9439	75	16	model	model	NOUN
ajst-9439	75	17	can	can	AUX
ajst-9439	75	18	be	be	AUX
ajst-9439	75	19	finetuned	finetune	VERB
ajst-9439	75	20	through	through	ADP
ajst-9439	75	21	an	an	DET
ajst-9439	75	22	additional	additional	ADJ
ajst-9439	75	23	output	output	NOUN
ajst-9439	75	24	layer	layer	NOUN
ajst-9439	75	25	,	,	PUNCT
ajst-9439	75	26	which	which	PRON
ajst-9439	75	27	has	have	VERB
ajst-9439	75	28	wider	wide	ADJ
ajst-9439	75	29	applicability	applicability	NOUN
ajst-9439	75	30	and	and	CCONJ
ajst-9439	75	31	does	do	AUX
ajst-9439	75	32	not	not	PART
ajst-9439	75	33	require	require	VERB
ajst-9439	75	34	more	more	ADV
ajst-9439	75	35	repetitive	repetitive	ADJ
ajst-9439	75	36	model	model	NOUN
ajst-9439	75	37	training	training	NOUN
ajst-9439	75	38	work	work	NOUN
ajst-9439	75	39	.	.	PUNCT
ajst-9439	76	1	figure	figure	VERB
ajst-9439	76	2	7	7	NUM
ajst-9439	76	3	.	.	PUNCT
ajst-9439	76	4	schematic	schematic	ADJ
ajst-9439	76	5	diagram	diagram	NOUN
ajst-9439	76	6	of	of	ADP
ajst-9439	76	7	bert	bert	PROPN
ajst-9439	76	8	algorithm	algorithm	PROPN
ajst-9439	76	9	4.4	4.4	NUM
ajst-9439	76	10	.	.	PUNCT
ajst-9439	77	1	neural	neural	ADJ
ajst-9439	77	2	network	network	NOUN
ajst-9439	77	3	model	model	NOUN
ajst-9439	77	4	fully	fully	ADV
ajst-9439	77	5	connected	connected	ADJ
ajst-9439	77	6	neural	neural	ADJ
ajst-9439	77	7	network	network	NOUN
ajst-9439	77	8	or	or	CCONJ
ajst-9439	77	9	multi	multi	ADJ
ajst-9439	77	10	-	-	ADJ
ajst-9439	77	11	layer	layer	ADJ
ajst-9439	77	12	perception	perception	NOUN
ajst-9439	77	13	is	be	AUX
ajst-9439	77	14	an	an	DET
ajst-9439	77	15	artificial	artificial	ADJ
ajst-9439	77	16	neural	neural	ADJ
ajst-9439	77	17	network	network	NOUN
ajst-9439	77	18	structure	structure	NOUN
ajst-9439	77	19	with	with	ADP
ajst-9439	77	20	a	a	DET
ajst-9439	77	21	relatively	relatively	ADV
ajst-9439	77	22	simple	simple	ADJ
ajst-9439	77	23	connection	connection	NOUN
ajst-9439	77	24	method	method	NOUN
ajst-9439	77	25	,	,	PUNCT
ajst-9439	77	26	which	which	PRON
ajst-9439	77	27	belongs	belong	VERB
ajst-9439	77	28	to	to	ADP
ajst-9439	77	29	a	a	DET
ajst-9439	77	30	kind	kind	NOUN
ajst-9439	77	31	of	of	ADP
ajst-9439	77	32	feedforward	feedforward	ADJ
ajst-9439	77	33	neural	neural	ADJ
ajst-9439	77	34	network	network	NOUN
ajst-9439	77	35	.	.	PUNCT
ajst-9439	78	1	it	it	PRON
ajst-9439	78	2	mainly	mainly	ADV
ajst-9439	78	3	consists	consist	VERB
ajst-9439	78	4	of	of	ADP
ajst-9439	78	5	input	input	NOUN
ajst-9439	78	6	layer	layer	NOUN
ajst-9439	78	7	,	,	PUNCT
ajst-9439	78	8	hidden	hide	VERB
ajst-9439	78	9	layer	layer	NOUN
ajst-9439	78	10	and	and	CCONJ
ajst-9439	78	11	output	output	NOUN
ajst-9439	78	12	layer	layer	NOUN
ajst-9439	78	13	,	,	PUNCT
ajst-9439	78	14	and	and	CCONJ
ajst-9439	78	15	there	there	PRON
ajst-9439	78	16	can	can	AUX
ajst-9439	78	17	be	be	AUX
ajst-9439	78	18	multiple	multiple	ADJ
ajst-9439	78	19	neurons	neuron	NOUN
ajst-9439	78	20	in	in	ADP
ajst-9439	78	21	each	each	DET
ajst-9439	78	22	hidden	hide	VERB
ajst-9439	78	23	layer	layer	NOUN
ajst-9439	78	24	.	.	PUNCT
ajst-9439	79	1	the	the	DET
ajst-9439	79	2	input	input	NOUN
ajst-9439	79	3	layer	layer	NOUN
ajst-9439	79	4	only	only	ADV
ajst-9439	79	5	receives	receive	VERB
ajst-9439	79	6	external	external	ADJ
ajst-9439	79	7	input	input	NOUN
ajst-9439	79	8	and	and	CCONJ
ajst-9439	79	9	does	do	AUX
ajst-9439	79	10	not	not	PART
ajst-9439	79	11	perform	perform	VERB
ajst-9439	79	12	any	any	DET
ajst-9439	79	13	function	function	NOUN
ajst-9439	79	14	processing	processing	NOUN
ajst-9439	79	15	,	,	PUNCT
ajst-9439	79	16	so	so	CCONJ
ajst-9439	79	17	the	the	DET
ajst-9439	79	18	number	number	NOUN
ajst-9439	79	19	of	of	ADP
ajst-9439	79	20	neurons	neuron	NOUN
ajst-9439	79	21	in	in	ADP
ajst-9439	79	22	the	the	DET
ajst-9439	79	23	input	input	NOUN
ajst-9439	79	24	layer	layer	NOUN
ajst-9439	79	25	is	be	AUX
ajst-9439	79	26	often	often	ADV
ajst-9439	79	27	the	the	DET
ajst-9439	79	28	same	same	ADJ
ajst-9439	79	29	as	as	ADP
ajst-9439	79	30	the	the	DET
ajst-9439	79	31	number	number	NOUN
ajst-9439	79	32	of	of	ADP
ajst-9439	79	33	input	input	NOUN
ajst-9439	79	34	features	feature	NOUN
ajst-9439	79	35	.	.	PUNCT
ajst-9439	80	1	the	the	DET
ajst-9439	80	2	neurons	neuron	NOUN
ajst-9439	80	3	in	in	ADP
ajst-9439	80	4	the	the	DET
ajst-9439	80	5	hidden	hide	VERB
ajst-9439	80	6	layer	layer	NOUN
ajst-9439	80	7	and	and	CCONJ
ajst-9439	80	8	the	the	DET
ajst-9439	80	9	output	output	NOUN
ajst-9439	80	10	layer	layer	NOUN
ajst-9439	80	11	process	process	NOUN
ajst-9439	80	12	the	the	DET
ajst-9439	80	13	signal	signal	NOUN
ajst-9439	80	14	,	,	PUNCT
ajst-9439	80	15	and	and	CCONJ
ajst-9439	80	16	the	the	DET
ajst-9439	80	17	final	final	ADJ
ajst-9439	80	18	result	result	NOUN
ajst-9439	80	19	is	be	AUX
ajst-9439	80	20	output	output	VERB
ajst-9439	80	21	by	by	ADP
ajst-9439	80	22	the	the	DET
ajst-9439	80	23	neurons	neuron	NOUN
ajst-9439	80	24	in	in	ADP
ajst-9439	80	25	the	the	DET
ajst-9439	80	26	output	output	NOUN
ajst-9439	80	27	layer	layer	NOUN
ajst-9439	80	28	.	.	PUNCT
ajst-9439	81	1	according	accord	VERB
ajst-9439	81	2	to	to	ADP
ajst-9439	81	3	the	the	DET
ajst-9439	81	4	number	number	NOUN
ajst-9439	81	5	of	of	ADP
ajst-9439	81	6	hidden	hidden	ADJ
ajst-9439	81	7	layers	layer	NOUN
ajst-9439	81	8	,	,	PUNCT
ajst-9439	81	9	it	it	PRON
ajst-9439	81	10	can	can	AUX
ajst-9439	81	11	be	be	AUX
ajst-9439	81	12	divided	divide	VERB
ajst-9439	81	13	into	into	ADP
ajst-9439	81	14	single	single	ADJ
ajst-9439	81	15	hidden	hide	VERB
ajst-9439	81	16	layer	layer	NOUN
ajst-9439	81	17	mlp	mlp	NOUN
ajst-9439	81	18	and	and	CCONJ
ajst-9439	81	19	multi	multi	ADJ
ajst-9439	81	20	hidden	hide	VERB
ajst-9439	81	21	layer	layer	NOUN
ajst-9439	81	22	mlp	mlp	NOUN
ajst-9439	81	23	.	.	PUNCT
ajst-9439	82	1	the	the	DET
ajst-9439	82	2	number	number	NOUN
ajst-9439	82	3	of	of	ADP
ajst-9439	82	4	neurons	neuron	NOUN
ajst-9439	82	5	in	in	ADP
ajst-9439	82	6	each	each	DET
ajst-9439	82	7	hidden	hide	VERB
ajst-9439	82	8	layer	layer	NOUN
ajst-9439	82	9	can	can	AUX
ajst-9439	82	10	be	be	AUX
ajst-9439	82	11	changed	change	VERB
ajst-9439	82	12	,	,	PUNCT
ajst-9439	82	13	and	and	CCONJ
ajst-9439	82	14	there	there	PRON
ajst-9439	82	15	is	be	VERB
ajst-9439	82	16	usually	usually	ADV
ajst-9439	82	17	no	no	DET
ajst-9439	82	18	good	good	ADJ
ajst-9439	82	19	standard	standard	NOUN
ajst-9439	82	20	for	for	ADP
ajst-9439	82	21	determining	determine	VERB
ajst-9439	82	22	the	the	DET
ajst-9439	82	23	number	number	NOUN
ajst-9439	82	24	of	of	ADP
ajst-9439	82	25	neurons	neuron	NOUN
ajst-9439	82	26	in	in	ADP
ajst-9439	82	27	each	each	DET
ajst-9439	82	28	layer	layer	NOUN
ajst-9439	82	29	and	and	CCONJ
ajst-9439	82	30	the	the	DET
ajst-9439	82	31	number	number	NOUN
ajst-9439	82	32	of	of	ADP
ajst-9439	82	33	hidden	hidden	ADJ
ajst-9439	82	34	layers	layer	NOUN
ajst-9439	82	35	,	,	PUNCT
ajst-9439	82	36	which	which	PRON
ajst-9439	82	37	has	have	AUX
ajst-9439	82	38	been	be	AUX
ajst-9439	82	39	tested	test	VERB
ajst-9439	82	40	and	and	CCONJ
ajst-9439	82	41	explored	explore	VERB
ajst-9439	82	42	so	so	ADV
ajst-9439	82	43	far	far	ADV
ajst-9439	82	44	.	.	PUNCT
ajst-9439	83	1	figure	figure	VERB
ajst-9439	83	2	8	8	NUM
ajst-9439	83	3	.	.	PUNCT
ajst-9439	84	1	schematic	schematic	ADJ
ajst-9439	84	2	diagram	diagram	NOUN
ajst-9439	84	3	of	of	ADP
ajst-9439	84	4	a	a	DET
ajst-9439	84	5	fully	fully	ADV
ajst-9439	84	6	connected	connected	ADJ
ajst-9439	84	7	neural	neural	ADJ
ajst-9439	84	8	network	network	NOUN
ajst-9439	84	9	4.5	4.5	NUM
ajst-9439	84	10	.	.	PUNCT
ajst-9439	85	1	experimental	experimental	ADJ
ajst-9439	85	2	results	result	NOUN
ajst-9439	85	3	the	the	DET
ajst-9439	85	4	experimental	experimental	ADJ
ajst-9439	85	5	results	result	NOUN
ajst-9439	85	6	of	of	ADP
ajst-9439	85	7	the	the	DET
ajst-9439	85	8	training	training	NOUN
ajst-9439	85	9	set	set	NOUN
ajst-9439	85	10	are	be	AUX
ajst-9439	85	11	shown	show	VERB
ajst-9439	85	12	in	in	ADP
ajst-9439	85	13	table	table	NOUN
ajst-9439	85	14	3	3	NUM
ajst-9439	85	15	.	.	PUNCT
ajst-9439	85	16	(	(	PUNCT
ajst-9439	85	17	16	16	NUM
ajst-9439	85	18	sentences	sentence	NOUN
ajst-9439	85	19	per	per	ADP
ajst-9439	85	20	group	group	NOUN
ajst-9439	85	21	)	)	PUNCT
ajst-9439	85	22	table	table	NOUN
ajst-9439	85	23	1	1	NUM
ajst-9439	85	24	.	.	PUNCT
ajst-9439	86	1	training	training	NOUN
ajst-9439	86	2	set	set	VERB
ajst-9439	86	3	experimental	experimental	ADJ
ajst-9439	86	4	results	result	NOUN
ajst-9439	86	5	number	number	NOUN
ajst-9439	86	6	of	of	ADP
ajst-9439	86	7	test	test	NOUN
ajst-9439	86	8	groups	group	NOUN
ajst-9439	86	9	loss	loss	VERB
ajst-9439	86	10	function	function	NOUN
ajst-9439	86	11	accuracy	accuracy	NOUN
ajst-9439	86	12	1	1	NUM
ajst-9439	86	13	0.6831	0.6831	NUM
ajst-9439	86	14	0.6875	0.6875	NUM
ajst-9439	86	15	20	20	NUM
ajst-9439	86	16	0.6375	0.6375	NUM
ajst-9439	86	17	0.7535	0.7535	NUM
ajst-9439	86	18	30	30	NUM
ajst-9439	86	19	0.6150	0.6150	NUM
ajst-9439	86	20	0.6875	0.6875	NUM
ajst-9439	86	21	50	50	NUM
ajst-9439	86	22	0.4962	0.4962	NUM
ajst-9439	86	23	0.9375	0.9375	NUM
ajst-9439	86	24	100	100	NUM
ajst-9439	86	25	0.4558	0.4558	NUM
ajst-9439	86	26	0.8755	0.8755	NUM
ajst-9439	86	27	200	200	NUM
ajst-9439	86	28	0.4121	0.4121	NUM
ajst-9439	86	29	0.9375	0.9375	NUM
ajst-9439	86	30	it	it	PRON
ajst-9439	86	31	can	can	AUX
ajst-9439	86	32	be	be	AUX
ajst-9439	86	33	seen	see	VERB
ajst-9439	86	34	that	that	SCONJ
ajst-9439	86	35	the	the	DET
ajst-9439	86	36	loss	loss	NOUN
ajst-9439	86	37	will	will	AUX
ajst-9439	86	38	decrease	decrease	VERB
ajst-9439	86	39	and	and	CCONJ
ajst-9439	86	40	the	the	DET
ajst-9439	86	41	accuracy	accuracy	NOUN
ajst-9439	86	42	rate	rate	NOUN
ajst-9439	86	43	will	will	AUX
ajst-9439	86	44	also	also	ADV
ajst-9439	86	45	decrease	decrease	VERB
ajst-9439	86	46	in	in	ADP
ajst-9439	86	47	the	the	DET
ajst-9439	86	48	training	training	NOUN
ajst-9439	86	49	set	set	NOUN
ajst-9439	86	50	.	.	PUNCT
ajst-9439	87	1	in	in	ADP
ajst-9439	87	2	the	the	DET
ajst-9439	87	3	classification	classification	NOUN
ajst-9439	87	4	problem	problem	NOUN
ajst-9439	87	5	,	,	PUNCT
ajst-9439	87	6	the	the	DET
ajst-9439	87	7	previous	previous	ADJ
ajst-9439	87	8	model	model	NOUN
ajst-9439	87	9	has	have	AUX
ajst-9439	87	10	not	not	PART
ajst-9439	87	11	been	be	AUX
ajst-9439	87	12	fully	fully	ADV
ajst-9439	87	13	trained	train	VERB
ajst-9439	87	14	.	.	PUNCT
ajst-9439	88	1	it	it	PRON
ajst-9439	88	2	is	be	AUX
ajst-9439	88	3	normal	normal	ADJ
ajst-9439	88	4	for	for	SCONJ
ajst-9439	88	5	this	this	PRON
ajst-9439	88	6	to	to	PART
ajst-9439	88	7	happen	happen	VERB
ajst-9439	88	8	occasionally	occasionally	ADV
ajst-9439	88	9	.	.	PUNCT
ajst-9439	89	1	when	when	SCONJ
ajst-9439	89	2	the	the	DET
ajst-9439	89	3	number	number	NOUN
ajst-9439	89	4	of	of	ADP
ajst-9439	89	5	test	test	NOUN
ajst-9439	89	6	groups	group	NOUN
ajst-9439	89	7	is	be	AUX
ajst-9439	89	8	50	50	NUM
ajst-9439	89	9	to	to	PART
ajst-9439	89	10	100	100	NUM
ajst-9439	89	11	,	,	PUNCT
ajst-9439	89	12	loss	loss	NOUN
ajst-9439	89	13	it	it	PRON
ajst-9439	89	14	has	have	AUX
ajst-9439	89	15	been	be	AUX
ajst-9439	89	16	maintained	maintain	VERB
ajst-9439	89	17	between	between	ADP
ajst-9439	89	18	0.4	0.4	NUM
ajst-9439	89	19	and	and	CCONJ
ajst-9439	89	20	0.5	0.5	NUM
ajst-9439	89	21	,	,	PUNCT
ajst-9439	89	22	and	and	CCONJ
ajst-9439	89	23	the	the	DET
ajst-9439	89	24	accuracy	accuracy	NOUN
ajst-9439	89	25	rate	rate	NOUN
ajst-9439	89	26	is	be	AUX
ajst-9439	89	27	above	above	ADP
ajst-9439	89	28	80	80	NUM
ajst-9439	89	29	%	%	NOUN
ajst-9439	89	30	.	.	PUNCT
ajst-9439	90	1	it	it	PRON
ajst-9439	90	2	is	be	AUX
ajst-9439	90	3	normal	normal	ADJ
ajst-9439	90	4	for	for	SCONJ
ajst-9439	90	5	the	the	DET
ajst-9439	90	6	accuracy	accuracy	NOUN
ajst-9439	90	7	rate	rate	NOUN
ajst-9439	90	8	to	to	PART
ajst-9439	90	9	fluctuate	fluctuate	VERB
ajst-9439	90	10	.	.	PUNCT
ajst-9439	91	1	the	the	DET
ajst-9439	91	2	experimental	experimental	ADJ
ajst-9439	91	3	results	result	NOUN
ajst-9439	91	4	of	of	ADP
ajst-9439	91	5	the	the	DET
ajst-9439	91	6	test	test	NOUN
ajst-9439	91	7	set	set	NOUN
ajst-9439	91	8	are	be	AUX
ajst-9439	91	9	shown	show	VERB
ajst-9439	91	10	in	in	ADP
ajst-9439	91	11	table	table	NOUN
ajst-9439	91	12	4	4	NUM
ajst-9439	91	13	.	.	SYM
ajst-9439	91	14	30	30	NUM
ajst-9439	91	15	table	table	NOUN
ajst-9439	91	16	2	2	NUM
ajst-9439	91	17	.	.	PUNCT
ajst-9439	91	18	test	test	NOUN
ajst-9439	91	19	set	set	VERB
ajst-9439	91	20	experimental	experimental	ADJ
ajst-9439	91	21	results	result	NOUN
ajst-9439	91	22	number	number	NOUN
ajst-9439	91	23	of	of	ADP
ajst-9439	91	24	test	test	NOUN
ajst-9439	91	25	groups	group	NOUN
ajst-9439	91	26	accuracy	accuracy	VERB
ajst-9439	91	27	1	1	NUM
ajst-9439	91	28	0.8125	0.8125	NUM
ajst-9439	91	29	20	20	NUM
ajst-9439	91	30	0.8645	0.8645	NUM
ajst-9439	91	31	30	30	NUM
ajst-9439	91	32	0.8677	0.8677	NUM
ajst-9439	91	33	50	50	NUM
ajst-9439	91	34	0.8660	0.8660	NUM
ajst-9439	91	35	100	100	NUM
ajst-9439	91	36	0.8657	0.8657	NUM
ajst-9439	91	37	it	it	PRON
ajst-9439	91	38	can	can	AUX
ajst-9439	91	39	be	be	AUX
ajst-9439	91	40	seen	see	VERB
ajst-9439	91	41	that	that	SCONJ
ajst-9439	91	42	in	in	ADP
ajst-9439	91	43	the	the	DET
ajst-9439	91	44	training	training	NOUN
ajst-9439	91	45	results	result	NOUN
ajst-9439	91	46	of	of	ADP
ajst-9439	91	47	the	the	DET
ajst-9439	91	48	test	test	NOUN
ajst-9439	91	49	set	set	NOUN
ajst-9439	91	50	,	,	PUNCT
ajst-9439	91	51	the	the	DET
ajst-9439	91	52	accuracy	accuracy	NOUN
ajst-9439	91	53	rate	rate	NOUN
ajst-9439	91	54	has	have	AUX
ajst-9439	91	55	been	be	AUX
ajst-9439	91	56	maintained	maintain	VERB
ajst-9439	91	57	at	at	ADP
ajst-9439	91	58	about	about	ADP
ajst-9439	91	59	86	86	NUM
ajst-9439	91	60	%	%	NOUN
ajst-9439	91	61	,	,	PUNCT
ajst-9439	91	62	which	which	PRON
ajst-9439	91	63	belongs	belong	VERB
ajst-9439	91	64	to	to	ADP
ajst-9439	91	65	a	a	DET
ajst-9439	91	66	relatively	relatively	ADV
ajst-9439	91	67	high	high	ADJ
ajst-9439	91	68	level	level	NOUN
ajst-9439	91	69	.	.	PUNCT
ajst-9439	92	1	the	the	DET
ajst-9439	92	2	accuracy	accuracy	NOUN
ajst-9439	92	3	of	of	ADP
ajst-9439	92	4	the	the	DET
ajst-9439	92	5	final	final	ADJ
ajst-9439	92	6	test	test	NOUN
ajst-9439	92	7	set	set	NOUN
ajst-9439	92	8	is	be	AUX
ajst-9439	92	9	shown	show	VERB
ajst-9439	92	10	in	in	ADP
ajst-9439	92	11	figure	figure	NOUN
ajst-9439	92	12	9	9	NUM
ajst-9439	92	13	.	.	PUNCT
ajst-9439	93	1	figure	figure	NOUN
ajst-9439	93	2	9	9	NUM
ajst-9439	93	3	.	.	PUNCT
ajst-9439	93	4	test	test	NOUN
ajst-9439	93	5	results	result	NOUN
ajst-9439	93	6	5	5	NUM
ajst-9439	93	7	.	.	PUNCT
ajst-9439	94	1	conclusion	conclusion	VERB
ajst-9439	94	2	the	the	DET
ajst-9439	94	3	disadvantage	disadvantage	NOUN
ajst-9439	94	4	of	of	ADP
ajst-9439	94	5	the	the	DET
ajst-9439	94	6	traditional	traditional	ADJ
ajst-9439	94	7	simple	simple	ADJ
ajst-9439	94	8	star	star	NOUN
ajst-9439	94	9	rating	rating	NOUN
ajst-9439	94	10	is	be	AUX
ajst-9439	94	11	that	that	SCONJ
ajst-9439	94	12	there	there	PRON
ajst-9439	94	13	are	be	VERB
ajst-9439	94	14	many	many	ADJ
ajst-9439	94	15	meaningless	meaningless	ADJ
ajst-9439	94	16	product	product	NOUN
ajst-9439	94	17	reviews	review	NOUN
ajst-9439	94	18	.	.	PUNCT
ajst-9439	95	1	for	for	ADP
ajst-9439	95	2	example	example	NOUN
ajst-9439	95	3	,	,	PUNCT
ajst-9439	95	4	the	the	DET
ajst-9439	95	5	number	number	NOUN
ajst-9439	95	6	of	of	ADP
ajst-9439	95	7	words	word	NOUN
ajst-9439	95	8	in	in	ADP
ajst-9439	95	9	the	the	DET
ajst-9439	95	10	review	review	NOUN
ajst-9439	95	11	is	be	AUX
ajst-9439	95	12	too	too	ADV
ajst-9439	95	13	small	small	ADJ
ajst-9439	95	14	to	to	PART
ajst-9439	95	15	provide	provide	VERB
ajst-9439	95	16	reference	reference	NOUN
ajst-9439	95	17	value	value	NOUN
ajst-9439	95	18	,	,	PUNCT
ajst-9439	95	19	or	or	CCONJ
ajst-9439	95	20	the	the	DET
ajst-9439	95	21	content	content	NOUN
ajst-9439	95	22	of	of	ADP
ajst-9439	95	23	the	the	DET
ajst-9439	95	24	full	full	ADJ
ajst-9439	95	25	-	-	PUNCT
ajst-9439	95	26	star	star	NOUN
ajst-9439	95	27	review	review	NOUN
ajst-9439	95	28	expresses	express	VERB
ajst-9439	95	29	dissatisfaction	dissatisfaction	NOUN
ajst-9439	95	30	with	with	ADP
ajst-9439	95	31	the	the	DET
ajst-9439	95	32	product	product	NOUN
ajst-9439	95	33	.	.	PUNCT
ajst-9439	96	1	this	this	PRON
ajst-9439	96	2	can	can	AUX
ajst-9439	96	3	be	be	AUX
ajst-9439	96	4	said	say	VERB
ajst-9439	96	5	to	to	PART
ajst-9439	96	6	be	be	AUX
ajst-9439	96	7	typical	typical	ADJ
ajst-9439	96	8	content	content	NOUN
ajst-9439	96	9	does	do	AUX
ajst-9439	96	10	not	not	PART
ajst-9439	96	11	match	match	VERB
ajst-9439	96	12	the	the	DET
ajst-9439	96	13	title	title	NOUN
ajst-9439	96	14	,	,	PUNCT
ajst-9439	96	15	which	which	PRON
ajst-9439	96	16	makes	make	VERB
ajst-9439	96	17	customers	customer	NOUN
ajst-9439	96	18	less	less	ADV
ajst-9439	96	19	trustworthy	trustworthy	ADJ
ajst-9439	96	20	in	in	ADP
ajst-9439	96	21	reviews	review	NOUN
ajst-9439	96	22	and	and	CCONJ
ajst-9439	96	23	affects	affect	VERB
ajst-9439	96	24	people	people	NOUN
ajst-9439	96	25	's	's	PART
ajst-9439	96	26	willingness	willingness	NOUN
ajst-9439	96	27	to	to	PART
ajst-9439	96	28	buy	buy	VERB
ajst-9439	96	29	to	to	ADP
ajst-9439	96	30	a	a	DET
ajst-9439	96	31	certain	certain	ADJ
ajst-9439	96	32	extent	extent	NOUN
ajst-9439	96	33	.	.	PUNCT
ajst-9439	97	1	this	this	DET
ajst-9439	97	2	situation	situation	NOUN
ajst-9439	97	3	is	be	AUX
ajst-9439	97	4	not	not	PART
ajst-9439	97	5	conducive	conducive	ADJ
ajst-9439	97	6	to	to	ADP
ajst-9439	97	7	the	the	DET
ajst-9439	97	8	merchant	merchant	NOUN
ajst-9439	97	9	's	's	PART
ajst-9439	97	10	commodity	commodity	NOUN
ajst-9439	97	11	purchase	purchase	NOUN
ajst-9439	97	12	and	and	CCONJ
ajst-9439	97	13	is	be	AUX
ajst-9439	97	14	not	not	PART
ajst-9439	97	15	conducive	conducive	ADJ
ajst-9439	97	16	to	to	ADP
ajst-9439	97	17	the	the	DET
ajst-9439	97	18	customer	customer	NOUN
ajst-9439	97	19	's	's	PART
ajst-9439	97	20	valuation	valuation	NOUN
ajst-9439	97	21	of	of	ADP
ajst-9439	97	22	the	the	DET
ajst-9439	97	23	commodity	commodity	NOUN
ajst-9439	97	24	.	.	PUNCT
ajst-9439	98	1	in	in	ADP
ajst-9439	98	2	the	the	DET
ajst-9439	98	3	context	context	NOUN
ajst-9439	98	4	of	of	ADP
ajst-9439	98	5	deep	deep	ADJ
ajst-9439	98	6	learning	learning	NOUN
ajst-9439	98	7	,	,	PUNCT
ajst-9439	98	8	we	we	PRON
ajst-9439	98	9	can	can	AUX
ajst-9439	98	10	quickly	quickly	ADV
ajst-9439	98	11	conduct	conduct	VERB
ajst-9439	98	12	sentiment	sentiment	NOUN
ajst-9439	98	13	analysis	analysis	NOUN
ajst-9439	98	14	of	of	ADP
ajst-9439	98	15	reviews	review	NOUN
ajst-9439	98	16	.	.	PUNCT
ajst-9439	99	1	instead	instead	ADV
ajst-9439	99	2	of	of	ADP
ajst-9439	99	3	relying	rely	VERB
ajst-9439	99	4	on	on	ADP
ajst-9439	99	5	traditional	traditional	ADJ
ajst-9439	99	6	star	star	NOUN
ajst-9439	99	7	ratings	rating	NOUN
ajst-9439	99	8	,	,	PUNCT
ajst-9439	99	9	we	we	PRON
ajst-9439	99	10	can	can	AUX
ajst-9439	99	11	choose	choose	VERB
ajst-9439	99	12	to	to	PART
ajst-9439	99	13	use	use	VERB
ajst-9439	99	14	machine	machine	NOUN
ajst-9439	99	15	learning	learn	VERB
ajst-9439	99	16	technology	technology	NOUN
ajst-9439	99	17	to	to	PART
ajst-9439	99	18	distinguish	distinguish	VERB
ajst-9439	99	19	the	the	DET
ajst-9439	99	20	emotions	emotion	NOUN
ajst-9439	99	21	of	of	ADP
ajst-9439	99	22	product	product	NOUN
ajst-9439	99	23	reviews	review	NOUN
ajst-9439	99	24	instead	instead	ADV
ajst-9439	99	25	of	of	ADP
ajst-9439	99	26	letting	let	VERB
ajst-9439	99	27	customers	customer	NOUN
ajst-9439	99	28	decide	decide	VERB
ajst-9439	99	29	,	,	PUNCT
ajst-9439	99	30	and	and	CCONJ
ajst-9439	99	31	we	we	PRON
ajst-9439	99	32	can	can	AUX
ajst-9439	99	33	also	also	ADV
ajst-9439	99	34	prevent	prevent	VERB
ajst-9439	99	35	reviews	review	NOUN
ajst-9439	99	36	with	with	ADP
ajst-9439	99	37	too	too	ADV
ajst-9439	99	38	little	little	ADJ
ajst-9439	99	39	content	content	NOUN
ajst-9439	99	40	from	from	ADP
ajst-9439	99	41	appearing	appear	VERB
ajst-9439	99	42	in	in	ADP
ajst-9439	99	43	the	the	DET
ajst-9439	99	44	reviews	review	NOUN
ajst-9439	99	45	that	that	PRON
ajst-9439	99	46	customers	customer	NOUN
ajst-9439	99	47	can	can	AUX
ajst-9439	99	48	see	see	VERB
ajst-9439	99	49	,	,	PUNCT
ajst-9439	99	50	so	so	SCONJ
ajst-9439	99	51	that	that	SCONJ
ajst-9439	99	52	we	we	PRON
ajst-9439	99	53	can	can	AUX
ajst-9439	99	54	avoid	avoid	VERB
ajst-9439	99	55	comments	comment	NOUN
ajst-9439	99	56	without	without	ADP
ajst-9439	99	57	content	content	NOUN
ajst-9439	99	58	,	,	PUNCT
ajst-9439	99	59	and	and	CCONJ
ajst-9439	99	60	you	you	PRON
ajst-9439	99	61	can	can	AUX
ajst-9439	99	62	also	also	ADV
ajst-9439	99	63	avoid	avoid	VERB
ajst-9439	99	64	comments	comment	NOUN
ajst-9439	99	65	that	that	PRON
ajst-9439	99	66	are	be	AUX
ajst-9439	99	67	not	not	PART
ajst-9439	99	68	on	on	ADP
ajst-9439	99	69	the	the	DET
ajst-9439	99	70	topic	topic	NOUN
ajst-9439	99	71	.	.	PUNCT
ajst-9439	100	1	we	we	PRON
ajst-9439	100	2	conducted	conduct	VERB
ajst-9439	100	3	a	a	DET
ajst-9439	100	4	preliminary	preliminary	ADJ
ajst-9439	100	5	application	application	NOUN
ajst-9439	100	6	of	of	ADP
ajst-9439	100	7	sentiment	sentiment	NOUN
ajst-9439	100	8	classification	classification	NOUN
ajst-9439	100	9	for	for	ADP
ajst-9439	100	10	comments	comment	NOUN
ajst-9439	100	11	,	,	PUNCT
ajst-9439	100	12	and	and	CCONJ
ajst-9439	100	13	proved	prove	VERB
ajst-9439	100	14	that	that	SCONJ
ajst-9439	100	15	our	our	PRON
ajst-9439	100	16	method	method	NOUN
ajst-9439	100	17	for	for	ADP
ajst-9439	100	18	sentiment	sentiment	NOUN
ajst-9439	100	19	classification	classification	NOUN
ajst-9439	100	20	of	of	ADP
ajst-9439	100	21	comments	comment	NOUN
ajst-9439	100	22	has	have	VERB
ajst-9439	100	23	a	a	DET
ajst-9439	100	24	high	high	ADJ
ajst-9439	100	25	success	success	NOUN
ajst-9439	100	26	rate	rate	NOUN
ajst-9439	100	27	through	through	ADP
ajst-9439	100	28	sentiment	sentiment	NOUN
ajst-9439	100	29	classification	classification	NOUN
ajst-9439	100	30	experiments	experiment	NOUN
ajst-9439	100	31	on	on	ADP
ajst-9439	100	32	the	the	DET
ajst-9439	100	33	bert	bert	PROPN
ajst-9439	100	34	hugging	hug	VERB
ajst-9439	100	35	face	face	NOUN
ajst-9439	100	36	dataset	dataset	NOUN
ajst-9439	100	37	.	.	PUNCT
ajst-9439	101	1	this	this	PRON
ajst-9439	101	2	is	be	AUX
ajst-9439	101	3	a	a	DET
ajst-9439	101	4	preliminary	preliminary	ADJ
ajst-9439	101	5	application	application	NOUN
ajst-9439	101	6	of	of	ADP
ajst-9439	101	7	machine	machine	NOUN
ajst-9439	101	8	learning	learn	VERB
ajst-9439	101	9	algorithms	algorithm	NOUN
ajst-9439	101	10	in	in	ADP
ajst-9439	101	11	comment	comment	NOUN
ajst-9439	101	12	sentiment	sentiment	NOUN
ajst-9439	101	13	classification	classification	NOUN
ajst-9439	101	14	,	,	PUNCT
ajst-9439	101	15	and	and	CCONJ
ajst-9439	101	16	more	more	ADJ
ajst-9439	101	17	in	in	ADP
ajst-9439	101	18	-	-	PUNCT
ajst-9439	101	19	depth	depth	NOUN
ajst-9439	101	20	applications	application	NOUN
ajst-9439	101	21	need	need	VERB
ajst-9439	101	22	to	to	PART
ajst-9439	101	23	be	be	AUX
ajst-9439	101	24	further	far	ADV
ajst-9439	101	25	carried	carry	VERB
ajst-9439	101	26	out	out	ADP
ajst-9439	101	27	.	.	PUNCT
ajst-9439	102	1	references	reference	NOUN
ajst-9439	102	2	[	[	X
ajst-9439	102	3	1	1	NUM
ajst-9439	102	4	]	]	X
ajst-9439	102	5	xue	xue	PROPN
ajst-9439	102	6	,	,	PUNCT
ajst-9439	102	7	y.z	y.z	PROPN
ajst-9439	102	8	.	.	PROPN
ajst-9439	102	9	,	,	PUNCT
ajst-9439	102	10	guo	guo	PROPN
ajst-9439	102	11	,	,	PUNCT
ajst-9439	102	12	y.f	y.f	PROPN
ajst-9439	102	13	.	.	PUNCT
ajst-9439	103	1	(	(	PUNCT
ajst-9439	103	2	2012	2012	NUM
ajst-9439	103	3	)	)	PUNCT
ajst-9439	103	4	ngaqqwc2c	ngaqqwc2c	ADJ
ajst-9439	103	5	research	research	NOUN
ajst-9439	103	6	on	on	ADP
ajst-9439	103	7	the	the	DET
ajst-9439	103	8	competitive	competitive	ADJ
ajst-9439	103	9	strategy	strategy	NOUN
ajst-9439	103	10	of	of	ADP
ajst-9439	103	11	e	e	NOUN
ajst-9439	103	12	-	-	NOUN
ajst-9439	103	13	commerce	commerce	NOUN
ajst-9439	103	14	sellers	seller	NOUN
ajst-9439	103	15	:	:	PUNCT
ajst-9439	103	16	an	an	DET
ajst-9439	103	17	analysis	analysis	NOUN
ajst-9439	103	18	based	base	VERB
ajst-9439	103	19	on	on	ADP
ajst-9439	103	20	taobao	taobao	PROPN
ajst-9439	103	21	.	.	PUNCT
ajst-9439	104	1	nankai	nankai	PROPN
ajst-9439	104	2	management	management	PROPN
ajst-9439	104	3	review	review	PROPN
ajst-9439	104	4	,	,	PUNCT
ajst-9439	104	5	015(005),129140	015(005),129140	NUM
ajst-9439	104	6	.	.	PUNCT
ajst-9439	105	1	[	[	X
ajst-9439	105	2	2	2	NUM
ajst-9439	105	3	]	]	X
ajst-9439	105	4	zhou	zhou	PROPN
ajst-9439	105	5	,	,	PUNCT
ajst-9439	105	6	y.g	y.g	PROPN
ajst-9439	105	7	.	.	PROPN
ajst-9439	105	8	,	,	PUNCT
ajst-9439	105	9	ma	ma	PROPN
ajst-9439	105	10	,	,	PUNCT
ajst-9439	105	11	y.h	y.h	PROPN
ajst-9439	105	12	.	.	PROPN
ajst-9439	105	13	(	(	PUNCT
ajst-9439	105	14	2007)research	2007)research	NUM
ajst-9439	105	15	on	on	ADP
ajst-9439	105	16	tourist	tourist	NOUN
ajst-9439	105	17	satisfaction	satisfaction	NOUN
ajst-9439	105	18	evaluation	evaluation	NOUN
ajst-9439	105	19	and	and	CCONJ
ajst-9439	105	20	visitor	visitor	NOUN
ajst-9439	105	21	management	management	NOUN
ajst-9439	105	22	based	base	VERB
ajst-9439	105	23	on	on	ADP
ajst-9439	105	24	ctrip.com	ctrip.com	NOUN
ajst-9439	105	25	's	's	PART
ajst-9439	105	26	free	free	ADJ
ajst-9439	105	27	reviews	review	NOUN
ajst-9439	105	28	:	:	PUNCT
ajst-9439	105	29	a	a	DET
ajst-9439	105	30	case	case	NOUN
ajst-9439	105	31	study	study	NOUN
ajst-9439	105	32	of	of	ADP
ajst-9439	105	33	huangshan	huangshan	PROPN
ajst-9439	105	34	scenic	scenic	ADJ
ajst-9439	105	35	area	area	NOUN
ajst-9439	105	36	.	.	PUNCT
ajst-9439	106	1	geography	geography	NOUN
ajst-9439	106	2	and	and	CCONJ
ajst-9439	106	3	geo	geo	PROPN
ajst-9439	106	4	-	-	PUNCT
ajst-9439	106	5	information	information	NOUN
ajst-9439	106	6	science	science	NOUN
ajst-9439	106	7	,	,	PUNCT
ajst-9439	106	8	23(2	23(2	NOUN
ajst-9439	106	9	)	)	PUNCT
ajst-9439	106	10	,	,	PUNCT
ajst-9439	106	11	97	97	NUM
ajst-9439	106	12	-	-	SYM
ajst-9439	106	13	100	100	NUM
ajst-9439	106	14	.	.	PUNCT
ajst-9439	107	1	[	[	X
ajst-9439	107	2	3	3	NUM
ajst-9439	107	3	]	]	X
ajst-9439	107	4	zhou	zhou	X
ajst-9439	107	5	,	,	PUNCT
ajst-9439	107	6	f.y	f.y	PROPN
ajst-9439	107	7	.	.	PROPN
ajst-9439	107	8	,	,	PUNCT
ajst-9439	107	9	jin	jin	PROPN
ajst-9439	107	10	,	,	PUNCT
ajst-9439	107	11	l.p	l.p	PROPN
ajst-9439	107	12	.	.	PROPN
ajst-9439	107	13	,	,	PUNCT
ajst-9439	107	14	dong	dong	PROPN
ajst-9439	107	15	,	,	PUNCT
ajst-9439	107	16	j.	j.	PROPN
ajst-9439	107	17	(	(	PUNCT
ajst-9439	107	18	2017	2017	NUM
ajst-9439	107	19	)	)	PUNCT
ajst-9439	107	20	a	a	DET
ajst-9439	107	21	review	review	NOUN
ajst-9439	107	22	of	of	ADP
ajst-9439	107	23	convolutional	convolutional	ADJ
ajst-9439	107	24	neural	neural	ADJ
ajst-9439	107	25	network	network	NOUN
ajst-9439	107	26	research	research	NOUN
ajst-9439	107	27	.	.	PUNCT
ajst-9439	108	1	chinese	chinese	ADJ
ajst-9439	108	2	journal	journal	PROPN
ajst-9439	108	3	of	of	ADP
ajst-9439	108	4	computers	computer	NOUN
ajst-9439	108	5	,	,	PUNCT
ajst-9439	108	6	40(6	40(6	NOUN
ajst-9439	108	7	)	)	PUNCT
ajst-9439	108	8	,	,	PUNCT
ajst-9439	108	9	1229	1229	NUM
ajst-9439	108	10	-	-	SYM
ajst-9439	108	11	1251	1251	NUM
ajst-9439	108	12	.	.	PUNCT
ajst-9439	109	1	[	[	X
ajst-9439	109	2	4	4	NUM
ajst-9439	109	3	]	]	X
ajst-9439	109	4	zhang	zhang	PROPN
ajst-9439	109	5	,	,	PUNCT
ajst-9439	109	6	s.	s.	PROPN
ajst-9439	109	7	,	,	PUNCT
ajst-9439	109	8	gong	gong	PROPN
ajst-9439	109	9	,	,	PUNCT
ajst-9439	109	10	y.h	y.h	PROPN
ajst-9439	109	11	.	.	PROPN
ajst-9439	109	12	,	,	PUNCT
ajst-9439	109	13	wang	wang	PROPN
ajst-9439	109	14	,	,	PUNCT
ajst-9439	109	15	j.j	j.j	PROPN
ajst-9439	109	16	.	.	PROPN
ajst-9439	109	17	(	(	PUNCT
ajst-9439	109	18	2019	2019	NUM
ajst-9439	109	19	)	)	PUNCT
ajst-9439	109	20	the	the	DET
ajst-9439	109	21	development	development	NOUN
ajst-9439	109	22	of	of	ADP
ajst-9439	109	23	deep	deep	ADJ
ajst-9439	109	24	convolutional	convolutional	ADJ
ajst-9439	109	25	neural	neural	ADJ
ajst-9439	109	26	networks	network	NOUN
ajst-9439	109	27	and	and	CCONJ
ajst-9439	109	28	their	their	PRON
ajst-9439	109	29	applications	application	NOUN
ajst-9439	109	30	in	in	ADP
ajst-9439	109	31	the	the	DET
ajst-9439	109	32	field	field	NOUN
ajst-9439	109	33	of	of	ADP
ajst-9439	109	34	computer	computer	NOUN
ajst-9439	109	35	vision	vision	NOUN
ajst-9439	109	36	.	.	PUNCT
ajst-9439	110	1	chinese	chinese	ADJ
ajst-9439	110	2	journal	journal	PROPN
ajst-9439	110	3	of	of	ADP
ajst-9439	110	4	computers	computer	NOUN
ajst-9439	110	5	,	,	PUNCT
ajst-9439	110	6	042(003),453	042(003),453	NUM
ajst-9439	110	7	-	-	SYM
ajst-9439	110	8	482	482	NUM
ajst-9439	110	9	.	.	PUNCT
ajst-9439	111	1	[	[	X
ajst-9439	111	2	5	5	NUM
ajst-9439	111	3	]	]	X
ajst-9439	111	4	wang	wang	PROPN
ajst-9439	111	5	,	,	PUNCT
ajst-9439	111	6	c.h	c.h	PROPN
ajst-9439	111	7	.	.	PROPN
ajst-9439	111	8	,	,	PUNCT
ajst-9439	111	9	zhang	zhang	PROPN
ajst-9439	111	10	,	,	PUNCT
ajst-9439	111	11	m.	m.	NOUN
ajst-9439	111	12	,	,	PUNCT
ajst-9439	111	13	ma	ma	PROPN
ajst-9439	111	14	,	,	PUNCT
ajst-9439	111	15	s.p	s.p	PROPN
ajst-9439	111	16	.	.	PROPN
ajst-9439	111	17	(	(	PUNCT
ajst-9439	111	18	2007	2007	NUM
ajst-9439	111	19	)	)	PUNCT
ajst-9439	111	20	a	a	DET
ajst-9439	111	21	review	review	NOUN
ajst-9439	111	22	of	of	ADP
ajst-9439	111	23	the	the	DET
ajst-9439	111	24	application	application	NOUN
ajst-9439	111	25	of	of	ADP
ajst-9439	111	26	natural	natural	ADJ
ajst-9439	111	27	language	language	NOUN
ajst-9439	111	28	processing	processing	NOUN
ajst-9439	111	29	in	in	ADP
ajst-9439	111	30	information	information	NOUN
ajst-9439	111	31	retrieval	retrieval	NOUN
ajst-9439	111	32	.	.	PUNCT
ajst-9439	112	1	journal	journal	NOUN
ajst-9439	112	2	of	of	ADP
ajst-9439	112	3	chinese	chinese	ADJ
ajst-9439	112	4	information	information	NOUN
ajst-9439	112	5	technology,(2	technology,(2	NUM
ajst-9439	112	6	)	)	PUNCT
ajst-9439	112	7	,	,	PUNCT
ajst-9439	112	8	3545	3545	NUM
ajst-9439	112	9	.	.	PUNCT
ajst-9439	113	1	[	[	X
ajst-9439	113	2	6	6	NUM
ajst-9439	113	3	]	]	SYM
ajst-9439	113	4	xi	xi	PROPN
ajst-9439	113	5	,	,	PUNCT
ajst-9439	113	6	x.f	x.f	PROPN
ajst-9439	113	7	.	.	PROPN
ajst-9439	113	8	,	,	PUNCT
ajst-9439	113	9	zhou	zhou	PROPN
ajst-9439	113	10	,	,	PUNCT
ajst-9439	113	11	g.d	g.d	PROPN
ajst-9439	113	12	.	.	PROPN
ajst-9439	113	13	(	(	PUNCT
ajst-9439	113	14	2016	2016	NUM
ajst-9439	113	15	)	)	PUNCT
ajst-9439	113	16	deep	deep	ADJ
ajst-9439	113	17	learning	learn	VERB
ajst-9439	113	18	research	research	NOUN
ajst-9439	113	19	for	for	ADP
ajst-9439	113	20	natural	natural	ADJ
ajst-9439	113	21	language	language	NOUN
ajst-9439	113	22	processing	processing	NOUN
ajst-9439	113	23	.	.	PUNCT
ajst-9439	114	1	acta	acta	PROPN
ajst-9439	114	2	automatica	automatica	PROPN
ajst-9439	114	3	sinica	sinica	PROPN
ajst-9439	114	4	,	,	PUNCT
ajst-9439	114	5	42(10	42(10	PROPN
ajst-9439	114	6	)	)	PUNCT
ajst-9439	114	7	,	,	PUNCT
ajst-9439	114	8	14451465	14451465	NUM
ajst-9439	114	9	.	.	PUNCT
ajst-9439	115	1	[	[	X
ajst-9439	115	2	7	7	NUM
ajst-9439	115	3	]	]	SYM
ajst-9439	115	4	li	li	PROPN
ajst-9439	115	5	,	,	PUNCT
ajst-9439	115	6	z.j	z.j	PROPN
ajst-9439	115	7	.	.	PROPN
ajst-9439	115	8	,	,	PUNCT
ajst-9439	115	9	fan	fan	PROPN
ajst-9439	115	10	,	,	PUNCT
ajst-9439	115	11	y.	y.	PROPN
ajst-9439	115	12	,	,	PUNCT
ajst-9439	115	13	wu	wu	PROPN
ajst-9439	115	14	,	,	PUNCT
ajst-9439	115	15	x.j	x.j	PROPN
ajst-9439	115	16	.	.	PROPN
ajst-9439	115	17	(	(	PUNCT
ajst-9439	115	18	2020	2020	NUM
ajst-9439	115	19	)	)	PUNCT
ajst-9439	115	20	a	a	DET
ajst-9439	115	21	review	review	NOUN
ajst-9439	115	22	of	of	ADP
ajst-9439	115	23	pre	pre	ADJ
ajst-9439	115	24	-	-	ADJ
ajst-9439	115	25	training	training	ADJ
ajst-9439	115	26	techniques	technique	NOUN
ajst-9439	115	27	for	for	ADP
ajst-9439	115	28	natural	natural	ADJ
ajst-9439	115	29	language	language	NOUN
ajst-9439	115	30	processing	processing	NOUN
ajst-9439	115	31	.	.	PUNCT
ajst-9439	116	1	computer	computer	NOUN
ajst-9439	116	2	science	science	NOUN
ajst-9439	116	3	,	,	PUNCT
ajst-9439	116	4	47(3	47(3	NOUN
ajst-9439	116	5	)	)	PUNCT
ajst-9439	116	6	,	,	PUNCT
ajst-9439	116	7	162	162	NUM
ajst-9439	116	8	-	-	SYM
ajst-9439	116	9	173	173	NUM
ajst-9439	116	10	.	.	PUNCT
ajst-9439	117	1	[	[	X
ajst-9439	117	2	8	8	NUM
ajst-9439	117	3	]	]	X
ajst-9439	117	4	zhang	zhang	PROPN
ajst-9439	117	5	,	,	PUNCT
ajst-9439	117	6	z.q	z.q	PROPN
ajst-9439	117	7	.	.	PROPN
ajst-9439	117	8	,	,	PUNCT
ajst-9439	117	9	ye	ye	PROPN
ajst-9439	117	10	,	,	PUNCT
ajst-9439	117	11	q.	q.	PROPN
ajst-9439	117	12	,	,	PUNCT
ajst-9439	117	13	li	li	PROPN
ajst-9439	117	14	,	,	PUNCT
ajst-9439	117	15	y.j	y.j	PROPN
ajst-9439	117	16	.	.	PUNCT
ajst-9439	117	17	(	(	PUNCT
ajst-9439	117	18	2010	2010	NUM
ajst-9439	117	19	)	)	PUNCT
ajst-9439	117	20	a	a	DET
ajst-9439	117	21	review	review	NOUN
ajst-9439	117	22	of	of	ADP
ajst-9439	117	23	sentiment	sentiment	NOUN
ajst-9439	117	24	analysis	analysis	NOUN
ajst-9439	117	25	of	of	ADP
ajst-9439	117	26	internet	internet	NOUN
ajst-9439	117	27	product	product	NOUN
ajst-9439	117	28	reviews	review	NOUN
ajst-9439	117	29	.	.	PUNCT
ajst-9439	118	1	journal	journal	NOUN
ajst-9439	118	2	of	of	ADP
ajst-9439	118	3	management	management	NOUN
ajst-9439	118	4	science	science	NOUN
ajst-9439	118	5	,	,	PUNCT
ajst-9439	118	6	13(6	13(6	PROPN
ajst-9439	118	7	)	)	PUNCT
ajst-9439	118	8	,	,	PUNCT
ajst-9439	118	9	84	84	NUM
ajst-9439	118	10	-	-	SYM
ajst-9439	118	11	96	96	NUM
ajst-9439	118	12	.	.	PUNCT
ajst-9439	119	1	[	[	X
ajst-9439	119	2	9	9	NUM
ajst-9439	119	3	]	]	X
ajst-9439	119	4	zhao	zhao	PROPN
ajst-9439	119	5	,	,	PUNCT
ajst-9439	119	6	y.y	y.y	PROPN
ajst-9439	119	7	.	.	PROPN
ajst-9439	119	8	,	,	PUNCT
ajst-9439	119	9	qin	qin	PROPN
ajst-9439	119	10	,	,	PUNCT
ajst-9439	119	11	b.	b.	PROPN
ajst-9439	119	12	,	,	PUNCT
ajst-9439	119	13	liu	liu	PROPN
ajst-9439	119	14	,	,	PUNCT
ajst-9439	119	15	t.	t.	PROPN
ajst-9439	119	16	(	(	PUNCT
ajst-9439	119	17	2010)sentiment	2010)sentiment	NUM
ajst-9439	119	18	analysis	analysis	NOUN
ajst-9439	119	19	.	.	PUNCT
ajst-9439	119	20	journal	journal	PROPN
ajst-9439	119	21	of	of	ADP
ajst-9439	119	22	software	software	NOUN
ajst-9439	119	23	,	,	PUNCT
ajst-9439	119	24	21(8),1834−1848	21(8),1834−1848	NUM
ajst-9439	119	25	.	.	PUNCT
ajst-9439	120	1	[	[	X
ajst-9439	120	2	10	10	NUM
ajst-9439	120	3	]	]	X
ajst-9439	120	4	chen	chen	PROPN
ajst-9439	120	5	,	,	PUNCT
ajst-9439	120	6	l.	l.	PROPN
ajst-9439	120	7	,	,	PUNCT
ajst-9439	120	8	guan	guan	PROPN
ajst-9439	120	9	,	,	PUNCT
ajst-9439	120	10	z.y	z.y	PROPN
ajst-9439	120	11	.	.	PROPN
ajst-9439	120	12	,	,	PUNCT
ajst-9439	120	13	he	he	PRON
ajst-9439	120	14	,	,	PUNCT
ajst-9439	120	15	j.h	j.h	PROPN
ajst-9439	120	16	.	.	PROPN
ajst-9439	120	17	,	,	PUNCT
ajst-9439	120	18	peng	peng	PROPN
ajst-9439	120	19	,	,	PUNCT
ajst-9439	120	20	j.y.(2017	j.y.(2017	PROPN
ajst-9439	120	21	)	)	PUNCT
ajst-9439	120	22	research	research	NOUN
ajst-9439	120	23	progress	progress	NOUN
ajst-9439	120	24	on	on	ADP
ajst-9439	120	25	sentiment	sentiment	NOUN
ajst-9439	120	26	classification	classification	NOUN
ajst-9439	120	27	.	.	PUNCT
ajst-9439	121	1	computer	computer	NOUN
ajst-9439	121	2	research	research	NOUN
ajst-9439	121	3	and	and	CCONJ
ajst-9439	121	4	development	development	NOUN
ajst-9439	121	5	,	,	PUNCT
ajst-9439	121	6	54(6	54(6	NOUN
ajst-9439	121	7	)	)	PUNCT
ajst-9439	121	8	,	,	PUNCT
ajst-9439	121	9	1150	1150	NUM
ajst-9439	121	10	-	-	SYM
ajst-9439	121	11	1170	1170	NUM
ajst-9439	121	12	.	.	PUNCT
ajst-9439	122	1	[	[	X
ajst-9439	122	2	11	11	NUM
ajst-9439	122	3	]	]	X
ajst-9439	122	4	hoang	hoang	PROPN
ajst-9439	122	5	,	,	PUNCT
ajst-9439	122	6	m.	m.	NOUN
ajst-9439	122	7	,	,	PUNCT
ajst-9439	122	8	bihorac	bihorac	PROPN
ajst-9439	122	9	,	,	PUNCT
ajst-9439	122	10	o.a	o.a	PROPN
ajst-9439	122	11	.	.	PROPN
ajst-9439	122	12	,	,	PUNCT
ajst-9439	122	13	rouces	rouce	NOUN
ajst-9439	122	14	,	,	PUNCT
ajst-9439	122	15	j.(2019	j.(2019	PROPN
ajst-9439	122	16	)	)	PUNCT
ajst-9439	122	17	aspect	aspect	NOUN
ajst-9439	122	18	-	-	PUNCT
ajst-9439	122	19	based	base	VERB
ajst-9439	122	20	sentiment	sentiment	NOUN
ajst-9439	122	21	analysis	analysis	NOUN
ajst-9439	122	22	use	use	NOUN
ajst-9439	122	23	bert//	bert//	NOUN
ajst-9439	122	24	proceedings	proceeding	NOUN
ajst-9439	122	25	of	of	ADP
ajst-9439	122	26	the	the	DET
ajst-9439	122	27	22nd	22nd	ADJ
ajst-9439	122	28	nordic	nordic	ADJ
ajst-9439	122	29	conference	conference	NOUN
ajst-9439	122	30	on	on	ADP
ajst-9439	122	31	computational	computational	ADJ
ajst-9439	122	32	linguistics	linguistic	NOUN
ajst-9439	122	33	.	.	PUNCT
ajst-9439	122	34	:	:	PUNCT
ajst-9439	122	35	187	187	NUM
ajst-9439	122	36	-	-	SYM
ajst-9439	122	37	196	196	NUM
ajst-9439	122	38	.	.	PUNCT
ajst-9439	122	39	0.80	0.80	NUM
ajst-9439	122	40	%	%	NOUN
ajst-9439	122	41	0.81	0.81	NUM
ajst-9439	122	42	%	%	NOUN
ajst-9439	122	43	0.82	0.82	NUM
ajst-9439	122	44	%	%	NOUN
ajst-9439	122	45	0.83	0.83	NUM
ajst-9439	122	46	%	%	NOUN
ajst-9439	122	47	0.84	0.84	NUM
ajst-9439	122	48	%	%	NOUN
ajst-9439	122	49	0.85	0.85	NUM
ajst-9439	122	50	%	%	NOUN
ajst-9439	122	51	0.86	0.86	NUM
ajst-9439	122	52	%	%	NOUN
ajst-9439	122	53	0.87	0.87	NUM
ajst-9439	122	54	%	%	NOUN
ajst-9439	122	55	0.88	0.88	NUM
ajst-9439	122	56	%	%	NOUN
ajst-9439	122	57	0.89	0.89	NUM
ajst-9439	122	58	%	%	NOUN
ajst-9439	122	59	0	0	NUM
ajst-9439	122	60	5	5	NUM
ajst-9439	122	61	10	10	NUM
ajst-9439	122	62	15	15	NUM
ajst-9439	122	63	20	20	NUM
ajst-9439	122	64	25	25	NUM
ajst-9439	122	65	30	30	NUM
ajst-9439	122	66	35	35	NUM
ajst-9439	122	67	abscissa	abscissa	ADJ
ajst-9439	122	68	:	:	PUNCT
ajst-9439	122	69	group	group	NOUN
ajst-9439	122	70	ordinate	ordinate	NOUN
ajst-9439	122	71	:	:	PUNCT
ajst-9439	122	72	accuracy	accuracy	NOUN