id	sid	tid	token	lemma	pos
fcis-29191	1	1	frontiers	frontier	NOUN
fcis-29191	1	2	in	in	ADP
fcis-29191	1	3	computing	computing	NOUN
fcis-29191	1	4	and	and	CCONJ
fcis-29191	1	5	intelligent	intelligent	ADJ
fcis-29191	1	6	systems	system	NOUN
fcis-29191	1	7	issn	issn	VERB
fcis-29191	1	8	:	:	PUNCT
fcis-29191	1	9	2832	2832	NUM
fcis-29191	1	10	-	-	SYM
fcis-29191	1	11	6024	6024	NUM
fcis-29191	1	12	|	|	NOUN
fcis-29191	1	13	vol	vol	NOUN
fcis-29191	1	14	.	.	PROPN
fcis-29191	2	1	11	11	NUM
fcis-29191	2	2	,	,	PUNCT
fcis-29191	2	3	no	no	INTJ
fcis-29191	2	4	.	.	NOUN
fcis-29191	2	5	1	1	NUM
fcis-29191	2	6	,	,	PUNCT
fcis-29191	2	7	2025	2025	NUM
fcis-29191	2	8	29	29	NUM
fcis-29191	2	9	cyber	cyber	NOUN
fcis-29191	2	10	violence	violence	NOUN
fcis-29191	2	11	text	text	NOUN
fcis-29191	2	12	classification	classification	NOUN
fcis-29191	2	13	model	model	NOUN
fcis-29191	2	14	based	base	VERB
fcis-29191	2	15	on	on	ADP
fcis-29191	2	16	graph	graph	NOUN
fcis-29191	2	17	convolutional	convolutional	ADJ
fcis-29191	2	18	networks	network	NOUN
fcis-29191	2	19	and	and	CCONJ
fcis-29191	2	20	syntactic	syntactic	ADJ
fcis-29191	2	21	parsing	parsing	NOUN
fcis-29191	2	22	binting	binte	VERB
fcis-29191	2	23	qi	qi	NOUN
fcis-29191	2	24	,	,	PUNCT
fcis-29191	2	25	hanming	hanme	VERB
fcis-29191	2	26	zhai	zhai	PROPN
fcis-29191	2	27	,	,	PUNCT
fcis-29191	2	28	yunyang	yunyang	PROPN
fcis-29191	2	29	bu	bu	PROPN
fcis-29191	2	30	,	,	PUNCT
fcis-29191	2	31	zhuxuan	zhuxuan	PROPN
fcis-29191	2	32	han	han	PROPN
fcis-29191	2	33	,	,	PUNCT
fcis-29191	2	34	fanliang	fanliang	PROPN
fcis-29191	2	35	bu	bu	PROPN
fcis-29191	2	36	*	*	PROPN
fcis-29191	2	37	department	department	PROPN
fcis-29191	2	38	of	of	ADP
fcis-29191	2	39	information	information	NOUN
fcis-29191	2	40	network	network	NOUN
fcis-29191	2	41	security	security	NOUN
fcis-29191	2	42	,	,	PUNCT
fcis-29191	2	43	people	people	NOUN
fcis-29191	2	44	's	's	PART
fcis-29191	2	45	public	public	ADJ
fcis-29191	2	46	security	security	NOUN
fcis-29191	2	47	university	university	PROPN
fcis-29191	2	48	of	of	ADP
fcis-29191	2	49	china	china	PROPN
fcis-29191	2	50	,	,	PUNCT
fcis-29191	2	51	beijing	beijing	PROPN
fcis-29191	2	52	,	,	PUNCT
fcis-29191	2	53	china	china	PROPN
fcis-29191	2	54	*	*	PUNCT
fcis-29191	2	55	corresponding	correspond	VERB
fcis-29191	2	56	author	author	NOUN
fcis-29191	2	57	:	:	PUNCT
fcis-29191	2	58	fanliang	fanliang	PROPN
fcis-29191	2	59	bu	bu	PROPN
fcis-29191	3	1	(	(	PUNCT
fcis-29191	3	2	email	email	NOUN
fcis-29191	3	3	:	:	PUNCT
fcis-29191	3	4	bufanliang@sina.com	bufanliang@sina.com	X
fcis-29191	3	5	)	)	PUNCT
fcis-29191	3	6	abstract	abstract	NOUN
fcis-29191	3	7	:	:	PUNCT
fcis-29191	3	8	there	there	PRON
fcis-29191	3	9	are	be	VERB
fcis-29191	3	10	problems	problem	NOUN
fcis-29191	3	11	such	such	ADJ
fcis-29191	3	12	as	as	ADP
fcis-29191	3	13	semantic	semantic	ADJ
fcis-29191	3	14	sparsity	sparsity	NOUN
fcis-29191	3	15	and	and	CCONJ
fcis-29191	3	16	incomplete	incomplete	ADJ
fcis-29191	3	17	context	context	NOUN
fcis-29191	3	18	in	in	ADP
fcis-29191	3	19	cyber	cyber	ADJ
fcis-29191	3	20	violence	violence	NOUN
fcis-29191	3	21	texts	text	NOUN
fcis-29191	3	22	in	in	ADP
fcis-29191	3	23	social	social	ADJ
fcis-29191	3	24	media	medium	NOUN
fcis-29191	3	25	,	,	PUNCT
fcis-29191	3	26	and	and	CCONJ
fcis-29191	3	27	the	the	DET
fcis-29191	3	28	current	current	ADJ
fcis-29191	3	29	research	research	NOUN
fcis-29191	3	30	on	on	ADP
fcis-29191	3	31	the	the	DET
fcis-29191	3	32	refinement	refinement	NOUN
fcis-29191	3	33	and	and	CCONJ
fcis-29191	3	34	classification	classification	NOUN
fcis-29191	3	35	of	of	ADP
fcis-29191	3	36	cyber	cyber	NOUN
fcis-29191	3	37	violence	violence	NOUN
fcis-29191	3	38	texts	text	NOUN
fcis-29191	3	39	is	be	AUX
fcis-29191	3	40	insufficient	insufficient	ADJ
fcis-29191	3	41	.	.	PUNCT
fcis-29191	4	1	this	this	DET
fcis-29191	4	2	paper	paper	NOUN
fcis-29191	4	3	constructs	construct	VERB
fcis-29191	4	4	a	a	DET
fcis-29191	4	5	text	text	NOUN
fcis-29191	4	6	classification	classification	NOUN
fcis-29191	4	7	model	model	NOUN
fcis-29191	4	8	based	base	VERB
fcis-29191	4	9	on	on	ADP
fcis-29191	4	10	graph	graph	NOUN
fcis-29191	4	11	neural	neural	ADJ
fcis-29191	4	12	networks	network	NOUN
fcis-29191	4	13	to	to	PART
fcis-29191	4	14	improve	improve	VERB
fcis-29191	4	15	the	the	DET
fcis-29191	4	16	fine	fine	ADV
fcis-29191	4	17	-	-	PUNCT
fcis-29191	4	18	grained	grain	VERB
fcis-29191	4	19	classification	classification	NOUN
fcis-29191	4	20	effect	effect	NOUN
fcis-29191	4	21	of	of	ADP
fcis-29191	4	22	multi	multi	ADJ
fcis-29191	4	23	category	category	NOUN
fcis-29191	4	24	cyber	cyber	NOUN
fcis-29191	4	25	violence	violence	NOUN
fcis-29191	4	26	texts	text	NOUN
fcis-29191	4	27	in	in	ADP
fcis-29191	4	28	social	social	ADJ
fcis-29191	4	29	media	medium	NOUN
fcis-29191	4	30	.	.	PUNCT
fcis-29191	5	1	introducing	introduce	VERB
fcis-29191	5	2	dependency	dependency	NOUN
fcis-29191	5	3	relationships	relationship	NOUN
fcis-29191	5	4	into	into	ADP
fcis-29191	5	5	bert	bert	PROPN
fcis-29191	5	6	through	through	ADP
fcis-29191	5	7	syntactic	syntactic	ADJ
fcis-29191	5	8	analysis	analysis	NOUN
fcis-29191	5	9	for	for	ADP
fcis-29191	5	10	deep	deep	ADJ
fcis-29191	5	11	semantic	semantic	ADJ
fcis-29191	5	12	representation	representation	NOUN
fcis-29191	5	13	learning	learning	NOUN
fcis-29191	5	14	enhances	enhance	VERB
fcis-29191	5	15	the	the	DET
fcis-29191	5	16	model	model	NOUN
fcis-29191	5	17	's	's	PART
fcis-29191	5	18	contextual	contextual	ADJ
fcis-29191	5	19	understanding	understanding	NOUN
fcis-29191	5	20	ability	ability	NOUN
fcis-29191	5	21	and	and	CCONJ
fcis-29191	5	22	better	well	ADJ
fcis-29191	5	23	captures	capture	NOUN
fcis-29191	5	24	fine	fine	ADV
fcis-29191	5	25	-	-	PUNCT
fcis-29191	5	26	grained	grain	VERB
fcis-29191	5	27	features	feature	NOUN
fcis-29191	5	28	in	in	ADP
fcis-29191	5	29	sentences	sentence	NOUN
fcis-29191	5	30	.	.	PUNCT
fcis-29191	6	1	build	build	VERB
fcis-29191	6	2	a	a	DET
fcis-29191	6	3	graph	graph	NOUN
fcis-29191	6	4	structure	structure	NOUN
fcis-29191	6	5	containing	contain	VERB
fcis-29191	6	6	two	two	NUM
fcis-29191	6	7	different	different	ADJ
fcis-29191	6	8	node	node	ADJ
fcis-29191	6	9	types	type	NOUN
fcis-29191	6	10	,	,	PUNCT
fcis-29191	6	11	words	word	NOUN
fcis-29191	6	12	and	and	CCONJ
fcis-29191	6	13	documents	document	NOUN
fcis-29191	6	14	,	,	PUNCT
fcis-29191	6	15	to	to	PART
fcis-29191	6	16	propagate	propagate	VERB
fcis-29191	6	17	semantic	semantic	ADJ
fcis-29191	6	18	information	information	NOUN
fcis-29191	6	19	through	through	ADP
fcis-29191	6	20	gcn	gcn	NOUN
fcis-29191	6	21	and	and	CCONJ
fcis-29191	6	22	strengthen	strengthen	VERB
fcis-29191	6	23	the	the	DET
fcis-29191	6	24	relationships	relationship	NOUN
fcis-29191	6	25	between	between	ADP
fcis-29191	6	26	nodes	node	NOUN
fcis-29191	6	27	.	.	PUNCT
fcis-29191	7	1	by	by	ADP
fcis-29191	7	2	combining	combine	VERB
fcis-29191	7	3	large	large	ADJ
fcis-29191	7	4	-	-	PUNCT
fcis-29191	7	5	scale	scale	NOUN
fcis-29191	7	6	pre	pre	ADJ
fcis-29191	7	7	-	-	ADJ
fcis-29191	7	8	trained	train	VERB
fcis-29191	7	9	models	model	NOUN
fcis-29191	7	10	with	with	ADP
fcis-29191	7	11	gcn	gcn	ADJ
fcis-29191	7	12	,	,	PUNCT
fcis-29191	7	13	complementary	complementary	ADJ
fcis-29191	7	14	advantages	advantage	NOUN
fcis-29191	7	15	can	can	AUX
fcis-29191	7	16	be	be	AUX
fcis-29191	7	17	achieved	achieve	VERB
fcis-29191	7	18	to	to	PART
fcis-29191	7	19	improve	improve	VERB
fcis-29191	7	20	the	the	DET
fcis-29191	7	21	overall	overall	ADJ
fcis-29191	7	22	performance	performance	NOUN
fcis-29191	7	23	of	of	ADP
fcis-29191	7	24	the	the	DET
fcis-29191	7	25	model	model	NOUN
fcis-29191	7	26	.	.	PUNCT
fcis-29191	8	1	the	the	DET
fcis-29191	8	2	experimental	experimental	ADJ
fcis-29191	8	3	results	result	NOUN
fcis-29191	8	4	show	show	VERB
fcis-29191	8	5	that	that	SCONJ
fcis-29191	8	6	the	the	DET
fcis-29191	8	7	method	method	NOUN
fcis-29191	8	8	achieves	achieve	VERB
fcis-29191	8	9	a	a	DET
fcis-29191	8	10	classification	classification	NOUN
fcis-29191	8	11	accuracy	accuracy	NOUN
fcis-29191	8	12	of	of	ADP
fcis-29191	8	13	84.77	84.77	NUM
fcis-29191	8	14	%	%	NOUN
fcis-29191	8	15	on	on	ADP
fcis-29191	8	16	publicly	publicly	ADV
fcis-29191	8	17	available	available	ADJ
fcis-29191	8	18	cyber	cyber	NOUN
fcis-29191	8	19	violence	violence	NOUN
fcis-29191	8	20	text	text	NOUN
fcis-29191	8	21	datasets	dataset	NOUN
fcis-29191	8	22	,	,	PUNCT
fcis-29191	8	23	which	which	PRON
fcis-29191	8	24	is	be	AUX
fcis-29191	8	25	2.36	2.36	NUM
fcis-29191	8	26	%	%	NOUN
fcis-29191	8	27	higher	high	ADJ
fcis-29191	8	28	than	than	ADP
fcis-29191	8	29	the	the	DET
fcis-29191	8	30	baseline	baseline	NOUN
fcis-29191	8	31	model	model	NOUN
fcis-29191	8	32	.	.	PUNCT
fcis-29191	9	1	it	it	PRON
fcis-29191	9	2	can	can	AUX
fcis-29191	9	3	effectively	effectively	ADV
fcis-29191	9	4	and	and	CCONJ
fcis-29191	9	5	accurately	accurately	ADV
fcis-29191	9	6	identify	identify	VERB
fcis-29191	9	7	multiple	multiple	ADJ
fcis-29191	9	8	types	type	NOUN
fcis-29191	9	9	of	of	ADP
fcis-29191	9	10	cyber	cyber	NOUN
fcis-29191	9	11	violence	violence	NOUN
fcis-29191	9	12	texts	text	NOUN
fcis-29191	9	13	.	.	PUNCT
fcis-29191	10	1	keywords	keyword	NOUN
fcis-29191	10	2	:	:	PUNCT
fcis-29191	10	3	graph	graph	VERB
fcis-29191	10	4	convolutional	convolutional	ADJ
fcis-29191	10	5	network	network	NOUN
fcis-29191	10	6	;	;	PUNCT
fcis-29191	10	7	cyber	cyber	ADJ
fcis-29191	10	8	violence	violence	NOUN
fcis-29191	10	9	;	;	PUNCT
fcis-29191	10	10	text	text	NOUN
fcis-29191	10	11	classification	classification	NOUN
fcis-29191	10	12	;	;	PUNCT
fcis-29191	10	13	bert	bert	NOUN
fcis-29191	10	14	;	;	PUNCT
fcis-29191	10	15	syntactic	syntactic	ADJ
fcis-29191	10	16	parsing	parsing	NOUN
fcis-29191	10	17	.	.	PUNCT
fcis-29191	11	1	1	1	X
fcis-29191	11	2	.	.	X
fcis-29191	11	3	introduction	introduction	NOUN
fcis-29191	11	4	with	with	ADP
fcis-29191	11	5	the	the	DET
fcis-29191	11	6	rapid	rapid	ADJ
fcis-29191	11	7	development	development	NOUN
fcis-29191	11	8	of	of	ADP
fcis-29191	11	9	the	the	DET
fcis-29191	11	10	internet	internet	NOUN
fcis-29191	11	11	and	and	CCONJ
fcis-29191	11	12	social	social	ADJ
fcis-29191	11	13	media	medium	NOUN
fcis-29191	11	14	,	,	PUNCT
fcis-29191	11	15	communication	communication	NOUN
fcis-29191	11	16	on	on	ADP
fcis-29191	11	17	social	social	ADJ
fcis-29191	11	18	media	medium	NOUN
fcis-29191	11	19	has	have	AUX
fcis-29191	11	20	become	become	VERB
fcis-29191	11	21	a	a	DET
fcis-29191	11	22	norm	norm	NOUN
fcis-29191	11	23	.	.	PUNCT
fcis-29191	12	1	users	user	NOUN
fcis-29191	12	2	can	can	AUX
fcis-29191	12	3	freely	freely	ADV
fcis-29191	12	4	express	express	VERB
fcis-29191	12	5	their	their	PRON
fcis-29191	12	6	opinions	opinion	NOUN
fcis-29191	12	7	on	on	ADP
fcis-29191	12	8	the	the	DET
fcis-29191	12	9	internet	internet	NOUN
fcis-29191	12	10	.	.	PUNCT
fcis-29191	13	1	this	this	DET
fcis-29191	13	2	freedom	freedom	NOUN
fcis-29191	13	3	also	also	ADV
fcis-29191	13	4	makes	make	VERB
fcis-29191	13	5	cyber	cyber	ADJ
fcis-29191	13	6	violence	violence	NOUN
fcis-29191	13	7	more	more	ADV
fcis-29191	13	8	frequent	frequent	ADJ
fcis-29191	13	9	.	.	PUNCT
fcis-29191	14	1	nowadays	nowadays	ADV
fcis-29191	14	2	,	,	PUNCT
fcis-29191	14	3	cyber	cyber	ADJ
fcis-29191	14	4	violence	violence	NOUN
fcis-29191	14	5	has	have	AUX
fcis-29191	14	6	attracted	attract	VERB
fcis-29191	14	7	high	high	ADJ
fcis-29191	14	8	attention	attention	NOUN
fcis-29191	14	9	worldwide	worldwide	ADV
fcis-29191	14	10	and	and	CCONJ
fcis-29191	14	11	has	have	AUX
fcis-29191	14	12	become	become	VERB
fcis-29191	14	13	a	a	DET
fcis-29191	14	14	major	major	ADJ
fcis-29191	14	15	hot	hot	ADJ
fcis-29191	14	16	issue	issue	NOUN
fcis-29191	14	17	that	that	PRON
fcis-29191	14	18	urgently	urgently	ADV
fcis-29191	14	19	needs	need	VERB
fcis-29191	14	20	to	to	PART
fcis-29191	14	21	be	be	AUX
fcis-29191	14	22	addressed	address	VERB
fcis-29191	14	23	.	.	PUNCT
fcis-29191	15	1	cyber	cyber	PROPN
fcis-29191	15	2	violence	violence	NOUN
fcis-29191	15	3	information	information	NOUN
fcis-29191	15	4	refers	refer	VERB
fcis-29191	15	5	to	to	ADP
fcis-29191	15	6	illegal	illegal	ADJ
fcis-29191	15	7	and	and	CCONJ
fcis-29191	15	8	harmful	harmful	ADJ
fcis-29191	15	9	content	content	NOUN
fcis-29191	15	10	that	that	PRON
fcis-29191	15	11	is	be	AUX
fcis-29191	15	12	concentrated	concentrate	VERB
fcis-29191	15	13	on	on	ADP
fcis-29191	15	14	individuals	individual	NOUN
fcis-29191	15	15	through	through	ADP
fcis-29191	15	16	the	the	DET
fcis-29191	15	17	internet	internet	NOUN
fcis-29191	15	18	in	in	ADP
fcis-29191	15	19	the	the	DET
fcis-29191	15	20	form	form	NOUN
fcis-29191	15	21	of	of	ADP
fcis-29191	15	22	text	text	NOUN
fcis-29191	15	23	,	,	PUNCT
fcis-29191	15	24	images	image	NOUN
fcis-29191	15	25	,	,	PUNCT
fcis-29191	15	26	audio	audio	NOUN
fcis-29191	15	27	,	,	PUNCT
fcis-29191	15	28	video	video	NOUN
fcis-29191	15	29	,	,	PUNCT
fcis-29191	15	30	etc	etc	X
fcis-29191	15	31	.	.	X
fcis-29191	15	32	,	,	PUNCT
fcis-29191	15	33	containing	contain	VERB
fcis-29191	15	34	insults	insult	NOUN
fcis-29191	15	35	,	,	PUNCT
fcis-29191	15	36	slander	slander	NOUN
fcis-29191	15	37	,	,	PUNCT
fcis-29191	15	38	rumors	rumor	NOUN
fcis-29191	15	39	,	,	PUNCT
fcis-29191	15	40	incitement	incitement	VERB
fcis-29191	15	41	to	to	PART
fcis-29191	15	42	hatred	hatred	VERB
fcis-29191	15	43	,	,	PUNCT
fcis-29191	15	44	coercion	coercion	NOUN
fcis-29191	15	45	,	,	PUNCT
fcis-29191	15	46	invasion	invasion	NOUN
fcis-29191	15	47	of	of	ADP
fcis-29191	15	48	privacy	privacy	NOUN
fcis-29191	15	49	,	,	PUNCT
fcis-29191	15	50	as	as	ADV
fcis-29191	15	51	well	well	ADV
fcis-29191	15	52	as	as	ADP
fcis-29191	15	53	accusations	accusation	NOUN
fcis-29191	15	54	,	,	PUNCT
fcis-29191	15	55	ridicule	ridicule	NOUN
fcis-29191	15	56	,	,	PUNCT
fcis-29191	15	57	belittling	belittle	VERB
fcis-29191	15	58	and	and	CCONJ
fcis-29191	15	59	discrimination	discrimination	NOUN
fcis-29191	15	60	that	that	PRON
fcis-29191	15	61	affect	affect	VERB
fcis-29191	15	62	physical	physical	ADJ
fcis-29191	15	63	and	and	CCONJ
fcis-29191	15	64	mental	mental	ADJ
fcis-29191	15	65	health	health	NOUN
fcis-29191	15	66	.	.	PUNCT
fcis-29191	16	1	despite	despite	SCONJ
fcis-29191	16	2	the	the	DET
fcis-29191	16	3	continuous	continuous	ADJ
fcis-29191	16	4	efforts	effort	NOUN
fcis-29191	16	5	of	of	ADP
fcis-29191	16	6	major	major	ADJ
fcis-29191	16	7	platforms	platform	NOUN
fcis-29191	16	8	to	to	PART
fcis-29191	16	9	develop	develop	VERB
fcis-29191	16	10	measures	measure	NOUN
fcis-29191	16	11	to	to	PART
fcis-29191	16	12	combat	combat	VERB
fcis-29191	16	13	cyber	cyber	NOUN
fcis-29191	16	14	violence	violence	NOUN
fcis-29191	16	15	,	,	PUNCT
fcis-29191	16	16	its	its	PRON
fcis-29191	16	17	scale	scale	NOUN
fcis-29191	16	18	and	and	CCONJ
fcis-29191	16	19	complexity	complexity	NOUN
fcis-29191	16	20	still	still	ADV
fcis-29191	16	21	make	make	VERB
fcis-29191	16	22	it	it	PRON
fcis-29191	16	23	difficult	difficult	ADJ
fcis-29191	16	24	to	to	PART
fcis-29191	16	25	completely	completely	ADV
fcis-29191	16	26	solve	solve	VERB
fcis-29191	16	27	the	the	DET
fcis-29191	16	28	problem	problem	NOUN
fcis-29191	16	29	.	.	PUNCT
fcis-29191	17	1	this	this	DET
fcis-29191	17	2	study	study	NOUN
fcis-29191	17	3	mainly	mainly	ADV
fcis-29191	17	4	explores	explore	VERB
fcis-29191	17	5	the	the	DET
fcis-29191	17	6	classification	classification	NOUN
fcis-29191	17	7	of	of	ADP
fcis-29191	17	8	cyber	cyber	ADJ
fcis-29191	17	9	violence	violence	NOUN
fcis-29191	17	10	texts	text	NOUN
fcis-29191	17	11	in	in	ADP
fcis-29191	17	12	social	social	ADJ
fcis-29191	17	13	media	medium	NOUN
fcis-29191	17	14	.	.	PUNCT
fcis-29191	18	1	text	text	NOUN
fcis-29191	18	2	classification	classification	NOUN
fcis-29191	18	3	is	be	AUX
fcis-29191	18	4	a	a	DET
fcis-29191	18	5	key	key	ADJ
fcis-29191	18	6	task	task	NOUN
fcis-29191	18	7	in	in	ADP
fcis-29191	18	8	natural	natural	ADJ
fcis-29191	18	9	language	language	NOUN
fcis-29191	18	10	processing	processing	NOUN
fcis-29191	18	11	,	,	PUNCT
fcis-29191	18	12	aimed	aim	VERB
fcis-29191	18	13	at	at	ADP
fcis-29191	18	14	assigning	assign	VERB
fcis-29191	18	15	text	text	NOUN
fcis-29191	18	16	content	content	NOUN
fcis-29191	18	17	to	to	ADP
fcis-29191	18	18	defined	define	VERB
fcis-29191	18	19	categories	category	NOUN
fcis-29191	18	20	.	.	PUNCT
fcis-29191	19	1	the	the	DET
fcis-29191	19	2	cyber	cyber	PROPN
fcis-29191	19	3	violence	violence	NOUN
fcis-29191	19	4	text	text	NOUN
fcis-29191	19	5	on	on	ADP
fcis-29191	19	6	social	social	ADJ
fcis-29191	19	7	platforms	platform	NOUN
fcis-29191	19	8	is	be	AUX
fcis-29191	19	9	usually	usually	ADV
fcis-29191	19	10	dominated	dominate	VERB
fcis-29191	19	11	by	by	ADP
fcis-29191	19	12	short	short	ADJ
fcis-29191	19	13	texts	text	NOUN
fcis-29191	19	14	,	,	PUNCT
fcis-29191	19	15	which	which	PRON
fcis-29191	19	16	have	have	VERB
fcis-29191	19	17	sparse	sparse	ADJ
fcis-29191	19	18	semantics	semantic	NOUN
fcis-29191	19	19	,	,	PUNCT
fcis-29191	19	20	vague	vague	ADJ
fcis-29191	19	21	information	information	NOUN
fcis-29191	19	22	,	,	PUNCT
fcis-29191	19	23	lack	lack	NOUN
fcis-29191	19	24	context	context	NOUN
fcis-29191	19	25	,	,	PUNCT
fcis-29191	19	26	and	and	CCONJ
fcis-29191	19	27	often	often	ADV
fcis-29191	19	28	contain	contain	VERB
fcis-29191	19	29	insulting	insulting	ADJ
fcis-29191	19	30	,	,	PUNCT
fcis-29191	19	31	defamatory	defamatory	ADJ
fcis-29191	19	32	,	,	PUNCT
fcis-29191	19	33	threatening	threaten	VERB
fcis-29191	19	34	words	word	NOUN
fcis-29191	19	35	as	as	ADV
fcis-29191	19	36	well	well	ADV
fcis-29191	19	37	as	as	ADP
fcis-29191	19	38	satirical	satirical	ADJ
fcis-29191	19	39	,	,	PUNCT
fcis-29191	19	40	mocking	mock	VERB
fcis-29191	19	41	tone	tone	NOUN
fcis-29191	19	42	features	feature	NOUN
fcis-29191	19	43	.	.	PUNCT
fcis-29191	20	1	therefore	therefore	ADV
fcis-29191	20	2	,	,	PUNCT
fcis-29191	20	3	existing	exist	VERB
fcis-29191	20	4	classification	classification	NOUN
fcis-29191	20	5	techniques	technique	NOUN
fcis-29191	20	6	can	can	AUX
fcis-29191	20	7	not	not	PART
fcis-29191	20	8	fully	fully	ADV
fcis-29191	20	9	adapt	adapt	VERB
fcis-29191	20	10	to	to	ADP
fcis-29191	20	11	such	such	ADJ
fcis-29191	20	12	texts	text	NOUN
fcis-29191	20	13	,	,	PUNCT
fcis-29191	20	14	making	make	VERB
fcis-29191	20	15	it	it	PRON
fcis-29191	20	16	difficult	difficult	ADJ
fcis-29191	20	17	to	to	PART
fcis-29191	20	18	fully	fully	ADV
fcis-29191	20	19	explore	explore	VERB
fcis-29191	20	20	the	the	DET
fcis-29191	20	21	characteristics	characteristic	NOUN
fcis-29191	20	22	of	of	ADP
fcis-29191	20	23	violent	violent	ADJ
fcis-29191	20	24	information	information	NOUN
fcis-29191	20	25	.	.	PUNCT
fcis-29191	21	1	this	this	DET
fcis-29191	21	2	article	article	NOUN
fcis-29191	21	3	designs	design	VERB
fcis-29191	21	4	a	a	DET
fcis-29191	21	5	cyber	cyber	NOUN
fcis-29191	21	6	violence	violence	NOUN
fcis-29191	21	7	text	text	NOUN
fcis-29191	21	8	classification	classification	NOUN
fcis-29191	21	9	model	model	NOUN
fcis-29191	21	10	based	base	VERB
fcis-29191	21	11	on	on	ADP
fcis-29191	21	12	graph	graph	NOUN
fcis-29191	21	13	convolutional	convolutional	ADJ
fcis-29191	21	14	networks	network	NOUN
fcis-29191	21	15	and	and	CCONJ
fcis-29191	21	16	syntactic	syntactic	ADJ
fcis-29191	21	17	analysis	analysis	NOUN
fcis-29191	21	18	,	,	PUNCT
fcis-29191	21	19	which	which	PRON
fcis-29191	21	20	subdivides	subdivide	VERB
fcis-29191	21	21	various	various	ADJ
fcis-29191	21	22	categories	category	NOUN
fcis-29191	21	23	of	of	ADP
fcis-29191	21	24	cyber	cyber	NOUN
fcis-29191	21	25	violence	violence	NOUN
fcis-29191	21	26	texts	text	NOUN
fcis-29191	21	27	and	and	CCONJ
fcis-29191	21	28	classifies	classify	VERB
fcis-29191	21	29	them	they	PRON
fcis-29191	21	30	according	accord	VERB
fcis-29191	21	31	to	to	ADP
fcis-29191	21	32	the	the	DET
fcis-29191	21	33	age	age	NOUN
fcis-29191	21	34	,	,	PUNCT
fcis-29191	21	35	race	race	NOUN
fcis-29191	21	36	,	,	PUNCT
fcis-29191	21	37	gender	gender	NOUN
fcis-29191	21	38	,	,	PUNCT
fcis-29191	21	39	religion	religion	NOUN
fcis-29191	21	40	,	,	PUNCT
fcis-29191	21	41	or	or	CCONJ
fcis-29191	21	42	other	other	ADJ
fcis-29191	21	43	attributes	attribute	NOUN
fcis-29191	21	44	of	of	ADP
fcis-29191	21	45	the	the	DET
fcis-29191	21	46	victims	victim	NOUN
fcis-29191	21	47	.	.	PUNCT
fcis-29191	22	1	this	this	DET
fcis-29191	22	2	model	model	NOUN
fcis-29191	22	3	introduces	introduce	VERB
fcis-29191	22	4	syntactic	syntactic	ADJ
fcis-29191	22	5	relationship	relationship	NOUN
fcis-29191	22	6	information	information	NOUN
fcis-29191	22	7	into	into	ADP
fcis-29191	22	8	the	the	DET
fcis-29191	22	9	bert	bert	PROPN
fcis-29191	22	10	model	model	NOUN
fcis-29191	22	11	by	by	ADP
fcis-29191	22	12	performing	perform	VERB
fcis-29191	22	13	syntactic	syntactic	ADJ
fcis-29191	22	14	analysis	analysis	NOUN
fcis-29191	22	15	on	on	ADP
fcis-29191	22	16	the	the	DET
fcis-29191	22	17	text	text	NOUN
fcis-29191	22	18	,	,	PUNCT
fcis-29191	22	19	helping	help	VERB
fcis-29191	22	20	the	the	DET
fcis-29191	22	21	model	model	NOUN
fcis-29191	22	22	better	well	ADV
fcis-29191	22	23	understand	understand	VERB
fcis-29191	22	24	the	the	DET
fcis-29191	22	25	grammar	grammar	NOUN
fcis-29191	22	26	and	and	CCONJ
fcis-29191	22	27	semantic	semantic	ADJ
fcis-29191	22	28	relationships	relationship	NOUN
fcis-29191	22	29	within	within	ADP
fcis-29191	22	30	sentences	sentence	NOUN
fcis-29191	22	31	and	and	CCONJ
fcis-29191	22	32	better	well	ADV
fcis-29191	22	33	represent	represent	VERB
fcis-29191	22	34	the	the	DET
fcis-29191	22	35	features	feature	NOUN
fcis-29191	22	36	of	of	ADP
fcis-29191	22	37	cyberbullying	cyberbullye	VERB
fcis-29191	22	38	texts	text	NOUN
fcis-29191	22	39	.	.	PUNCT
fcis-29191	23	1	graph	graph	VERB
fcis-29191	23	2	convolutional	convolutional	ADJ
fcis-29191	23	3	network	network	NOUN
fcis-29191	23	4	is	be	AUX
fcis-29191	23	5	a	a	DET
fcis-29191	23	6	deep	deep	ADJ
fcis-29191	23	7	learning	learning	NOUN
fcis-29191	23	8	method	method	NOUN
fcis-29191	23	9	suitable	suitable	ADJ
fcis-29191	23	10	for	for	ADP
fcis-29191	23	11	processing	processing	NOUN
fcis-29191	23	12	graph	graph	NOUN
fcis-29191	23	13	structured	structure	VERB
fcis-29191	23	14	data	datum	NOUN
fcis-29191	23	15	,	,	PUNCT
fcis-29191	23	16	which	which	PRON
fcis-29191	23	17	can	can	AUX
fcis-29191	23	18	deeply	deeply	ADV
fcis-29191	23	19	explore	explore	VERB
fcis-29191	23	20	the	the	DET
fcis-29191	23	21	semantic	semantic	ADJ
fcis-29191	23	22	and	and	CCONJ
fcis-29191	23	23	internal	internal	ADJ
fcis-29191	23	24	relationships	relationship	NOUN
fcis-29191	23	25	between	between	ADP
fcis-29191	23	26	texts	text	NOUN
fcis-29191	23	27	by	by	ADP
fcis-29191	23	28	mapping	map	VERB
fcis-29191	23	29	them	they	PRON
fcis-29191	23	30	as	as	ADP
fcis-29191	23	31	graph	graph	NOUN
fcis-29191	23	32	structures	structure	NOUN
fcis-29191	23	33	.	.	PUNCT
fcis-29191	24	1	the	the	DET
fcis-29191	24	2	joint	joint	ADJ
fcis-29191	24	3	training	training	NOUN
fcis-29191	24	4	of	of	ADP
fcis-29191	24	5	bert	bert	PROPN
fcis-29191	24	6	and	and	CCONJ
fcis-29191	24	7	gcn	gcn	NOUN
fcis-29191	24	8	has	have	AUX
fcis-29191	24	9	improved	improve	VERB
fcis-29191	24	10	the	the	DET
fcis-29191	24	11	effectiveness	effectiveness	NOUN
fcis-29191	24	12	of	of	ADP
fcis-29191	24	13	cyber	cyber	ADJ
fcis-29191	24	14	violence	violence	NOUN
fcis-29191	24	15	text	text	NOUN
fcis-29191	24	16	classification	classification	NOUN
fcis-29191	24	17	,	,	PUNCT
fcis-29191	24	18	demonstrating	demonstrate	VERB
fcis-29191	24	19	that	that	SCONJ
fcis-29191	24	20	through	through	ADP
fcis-29191	24	21	large	large	ADJ
fcis-29191	24	22	-	-	PUNCT
fcis-29191	24	23	scale	scale	NOUN
fcis-29191	24	24	pre	pre	NOUN
fcis-29191	24	25	training	training	NOUN
fcis-29191	24	26	,	,	PUNCT
fcis-29191	24	27	gcn	gcn	NOUN
fcis-29191	24	28	can	can	AUX
fcis-29191	24	29	better	well	ADV
fcis-29191	24	30	learn	learn	VERB
fcis-29191	24	31	the	the	DET
fcis-29191	24	32	complex	complex	ADJ
fcis-29191	24	33	relationships	relationship	NOUN
fcis-29191	24	34	between	between	ADP
fcis-29191	24	35	nodes	node	NOUN
fcis-29191	24	36	and	and	CCONJ
fcis-29191	24	37	graph	graph	NOUN
fcis-29191	24	38	structures	structure	NOUN
fcis-29191	24	39	,	,	PUNCT
fcis-29191	24	40	enhance	enhance	VERB
fcis-29191	24	41	its	its	PRON
fcis-29191	24	42	representation	representation	NOUN
fcis-29191	24	43	of	of	ADP
fcis-29191	24	44	node	node	ADJ
fcis-29191	24	45	features	feature	NOUN
fcis-29191	24	46	,	,	PUNCT
fcis-29191	24	47	and	and	CCONJ
fcis-29191	24	48	significantly	significantly	ADV
fcis-29191	24	49	optimize	optimize	VERB
fcis-29191	24	50	the	the	DET
fcis-29191	24	51	model	model	NOUN
fcis-29191	24	52	's	's	PART
fcis-29191	24	53	generalization	generalization	NOUN
fcis-29191	24	54	ability	ability	NOUN
fcis-29191	24	55	.	.	PUNCT
fcis-29191	25	1	the	the	DET
fcis-29191	25	2	main	main	ADJ
fcis-29191	25	3	contributions	contribution	NOUN
fcis-29191	25	4	and	and	CCONJ
fcis-29191	25	5	innovations	innovation	NOUN
fcis-29191	25	6	of	of	ADP
fcis-29191	25	7	this	this	DET
fcis-29191	25	8	article	article	NOUN
fcis-29191	25	9	are	be	AUX
fcis-29191	25	10	summarized	summarize	VERB
fcis-29191	25	11	as	as	SCONJ
fcis-29191	25	12	follows	follow	VERB
fcis-29191	25	13	:	:	PUNCT
fcis-29191	25	14	(	(	PUNCT
fcis-29191	25	15	1	1	X
fcis-29191	25	16	)	)	PUNCT
fcis-29191	25	17	this	this	DET
fcis-29191	25	18	article	article	NOUN
fcis-29191	25	19	conducts	conduct	VERB
fcis-29191	25	20	fine	fine	ADV
fcis-29191	25	21	-	-	PUNCT
fcis-29191	25	22	grained	grain	VERB
fcis-29191	25	23	classification	classification	NOUN
fcis-29191	25	24	of	of	ADP
fcis-29191	25	25	cyber	cyber	ADJ
fcis-29191	25	26	violence	violence	NOUN
fcis-29191	25	27	texts	text	NOUN
fcis-29191	25	28	into	into	ADP
fcis-29191	25	29	multiple	multiple	ADJ
fcis-29191	25	30	categories	category	NOUN
fcis-29191	25	31	.	.	PUNCT
fcis-29191	26	1	while	while	SCONJ
fcis-29191	26	2	determining	determine	VERB
fcis-29191	26	3	whether	whether	SCONJ
fcis-29191	26	4	they	they	PRON
fcis-29191	26	5	belong	belong	VERB
fcis-29191	26	6	to	to	ADP
fcis-29191	26	7	cyber	cyber	NOUN
fcis-29191	26	8	violence	violence	NOUN
fcis-29191	26	9	texts	text	NOUN
fcis-29191	26	10	,	,	PUNCT
fcis-29191	26	11	it	it	PRON
fcis-29191	26	12	is	be	AUX
fcis-29191	26	13	no	no	ADV
fcis-29191	26	14	longer	long	ADV
fcis-29191	26	15	limited	limit	VERB
fcis-29191	26	16	to	to	ADP
fcis-29191	26	17	simple	simple	ADJ
fcis-29191	26	18	classification	classification	NOUN
fcis-29191	26	19	based	base	VERB
fcis-29191	26	20	on	on	ADP
fcis-29191	26	21	gender	gender	NOUN
fcis-29191	26	22	discrimination	discrimination	NOUN
fcis-29191	26	23	or	or	CCONJ
fcis-29191	26	24	racial	racial	ADJ
fcis-29191	26	25	discrimination	discrimination	NOUN
fcis-29191	26	26	.	.	PUNCT
fcis-29191	27	1	based	base	VERB
fcis-29191	27	2	on	on	ADP
fcis-29191	27	3	different	different	ADJ
fcis-29191	27	4	characteristics	characteristic	NOUN
fcis-29191	27	5	of	of	ADP
fcis-29191	27	6	cyber	cyber	PROPN
fcis-29191	27	7	violence	violence	NOUN
fcis-29191	27	8	,	,	PUNCT
fcis-29191	27	9	a	a	DET
fcis-29191	27	10	comprehensive	comprehensive	ADJ
fcis-29191	27	11	classification	classification	NOUN
fcis-29191	27	12	system	system	NOUN
fcis-29191	27	13	is	be	AUX
fcis-29191	27	14	constructed	construct	VERB
fcis-29191	27	15	.	.	PUNCT
fcis-29191	28	1	(	(	PUNCT
fcis-29191	28	2	2	2	X
fcis-29191	28	3	)	)	PUNCT
fcis-29191	28	4	by	by	ADP
fcis-29191	28	5	performing	perform	VERB
fcis-29191	28	6	syntactic	syntactic	ADJ
fcis-29191	28	7	analysis	analysis	NOUN
fcis-29191	28	8	on	on	ADP
fcis-29191	28	9	the	the	DET
fcis-29191	28	10	text	text	NOUN
fcis-29191	28	11	and	and	CCONJ
fcis-29191	28	12	introducing	introduce	VERB
fcis-29191	28	13	syntactic	syntactic	ADJ
fcis-29191	28	14	relationship	relationship	NOUN
fcis-29191	28	15	information	information	NOUN
fcis-29191	28	16	into	into	ADP
fcis-29191	28	17	the	the	DET
fcis-29191	28	18	bert	bert	PROPN
fcis-29191	28	19	model	model	NOUN
fcis-29191	28	20	,	,	PUNCT
fcis-29191	28	21	the	the	DET
fcis-29191	28	22	model	model	NOUN
fcis-29191	28	23	can	can	AUX
fcis-29191	28	24	more	more	ADV
fcis-29191	28	25	accurately	accurately	ADV
fcis-29191	28	26	capture	capture	VERB
fcis-29191	28	27	the	the	DET
fcis-29191	28	28	dependencies	dependency	NOUN
fcis-29191	28	29	and	and	CCONJ
fcis-29191	28	30	semantic	semantic	ADJ
fcis-29191	28	31	connections	connection	NOUN
fcis-29191	28	32	within	within	ADP
fcis-29191	28	33	the	the	DET
fcis-29191	28	34	sentence	sentence	NOUN
fcis-29191	28	35	.	.	PUNCT
fcis-29191	29	1	joint	joint	ADJ
fcis-29191	29	2	training	training	NOUN
fcis-29191	29	3	of	of	ADP
fcis-29191	29	4	bert	bert	PROPN
fcis-29191	29	5	model	model	NOUN
fcis-29191	29	6	and	and	CCONJ
fcis-29191	29	7	gcn	gcn	NOUN
fcis-29191	29	8	can	can	AUX
fcis-29191	29	9	better	well	ADV
fcis-29191	29	10	learn	learn	VERB
fcis-29191	29	11	features	feature	NOUN
fcis-29191	29	12	in	in	ADP
fcis-29191	29	13	cyber	cyber	ADJ
fcis-29191	29	14	violence	violence	NOUN
fcis-29191	29	15	texts	text	NOUN
fcis-29191	29	16	,	,	PUNCT
fcis-29191	29	17	which	which	PRON
fcis-29191	29	18	can	can	AUX
fcis-29191	29	19	help	help	VERB
fcis-29191	29	20	improve	improve	VERB
fcis-29191	29	21	text	text	NOUN
fcis-29191	29	22	classification	classification	NOUN
fcis-29191	29	23	performance	performance	NOUN
fcis-29191	29	24	.	.	PUNCT
fcis-29191	30	1	(	(	PUNCT
fcis-29191	30	2	3	3	X
fcis-29191	30	3	)	)	PUNCT
fcis-29191	30	4	the	the	DET
fcis-29191	30	5	model	model	NOUN
fcis-29191	30	6	proposed	propose	VERB
fcis-29191	30	7	in	in	ADP
fcis-29191	30	8	this	this	DET
fcis-29191	30	9	article	article	NOUN
fcis-29191	30	10	was	be	AUX
fcis-29191	30	11	tested	test	VERB
fcis-29191	30	12	on	on	ADP
fcis-29191	30	13	a	a	DET
fcis-29191	30	14	publicly	publicly	ADV
fcis-29191	30	15	available	available	ADJ
fcis-29191	30	16	dataset	dataset	NOUN
fcis-29191	30	17	of	of	ADP
fcis-29191	30	18	cyber	cyber	NOUN
fcis-29191	30	19	violence	violence	NOUN
fcis-29191	30	20	,	,	PUNCT
fcis-29191	30	21	and	and	CCONJ
fcis-29191	30	22	its	its	PRON
fcis-29191	30	23	classification	classification	NOUN
fcis-29191	30	24	performance	performance	NOUN
fcis-29191	30	25	was	be	AUX
fcis-29191	30	26	better	well	ADJ
fcis-29191	30	27	than	than	ADP
fcis-29191	30	28	the	the	DET
fcis-29191	30	29	baseline	baseline	NOUN
fcis-29191	30	30	model	model	NOUN
fcis-29191	30	31	,	,	PUNCT
fcis-29191	30	32	improving	improve	VERB
fcis-29191	30	33	the	the	DET
fcis-29191	30	34	classification	classification	NOUN
fcis-29191	30	35	performance	performance	NOUN
fcis-29191	30	36	of	of	ADP
fcis-29191	30	37	cyber	cyber	ADJ
fcis-29191	30	38	violence	violence	NOUN
fcis-29191	30	39	text	text	NOUN
fcis-29191	30	40	in	in	ADP
fcis-29191	30	41	social	social	ADJ
fcis-29191	30	42	media	medium	NOUN
fcis-29191	30	43	.	.	PUNCT
fcis-29191	31	1	2	2	X
fcis-29191	31	2	.	.	X
fcis-29191	31	3	related	relate	VERB
fcis-29191	31	4	works	work	NOUN
fcis-29191	31	5	2.1	2.1	NUM
fcis-29191	31	6	.	.	PUNCT
fcis-29191	32	1	text	text	NOUN
fcis-29191	32	2	classification	classification	NOUN
fcis-29191	32	3	the	the	DET
fcis-29191	32	4	traditional	traditional	ADJ
fcis-29191	32	5	methods	method	NOUN
fcis-29191	32	6	for	for	ADP
fcis-29191	32	7	text	text	NOUN
fcis-29191	32	8	classification	classification	NOUN
fcis-29191	32	9	include	include	VERB
fcis-29191	32	10	support	support	NOUN
fcis-29191	32	11	vector	vector	NOUN
fcis-29191	32	12	machine	machine	NOUN
fcis-29191	32	13	(	(	PUNCT
fcis-29191	32	14	svm)[1	svm)[1	PROPN
fcis-29191	32	15	]	]	PUNCT
fcis-29191	32	16	,	,	PUNCT
fcis-29191	32	17	naive	naive	ADJ
fcis-29191	32	18	bayes	bayes	NOUN
fcis-29191	32	19	(	(	PUNCT
fcis-29191	32	20	nb)[2	nb)[2	PROPN
fcis-29191	32	21	]	]	X
fcis-29191	32	22	,	,	PUNCT
fcis-29191	32	23	k	k	PROPN
fcis-29191	32	24	30	30	NUM
fcis-29191	32	25	nearest	near	ADJ
fcis-29191	32	26	neighbor	neighbor	NOUN
fcis-29191	32	27	algorithm	algorithm	NOUN
fcis-29191	32	28	(	(	PUNCT
fcis-29191	32	29	knn)[3	knn)[3	PROPN
fcis-29191	32	30	]	]	PUNCT
fcis-29191	32	31	,	,	PUNCT
fcis-29191	32	32	decision	decision	NOUN
fcis-29191	32	33	tree	tree	NOUN
fcis-29191	32	34	algorithm[4	algorithm[4	PROPN
fcis-29191	32	35	]	]	PUNCT
fcis-29191	32	36	,	,	PUNCT
fcis-29191	32	37	etc	etc	X
fcis-29191	32	38	.	.	X
fcis-29191	33	1	these	these	DET
fcis-29191	33	2	traditional	traditional	ADJ
fcis-29191	33	3	classification	classification	NOUN
fcis-29191	33	4	algorithms	algorithm	NOUN
fcis-29191	33	5	have	have	VERB
fcis-29191	33	6	high	high	ADJ
fcis-29191	33	7	requirements	requirement	NOUN
fcis-29191	33	8	for	for	ADP
fcis-29191	33	9	feature	feature	NOUN
fcis-29191	33	10	engineering	engineering	NOUN
fcis-29191	33	11	,	,	PUNCT
fcis-29191	33	12	but	but	CCONJ
fcis-29191	33	13	in	in	ADP
fcis-29191	33	14	social	social	ADJ
fcis-29191	33	15	media	medium	NOUN
fcis-29191	33	16	,	,	PUNCT
fcis-29191	33	17	the	the	DET
fcis-29191	33	18	proportion	proportion	NOUN
fcis-29191	33	19	of	of	ADP
fcis-29191	33	20	short	short	ADJ
fcis-29191	33	21	texts	text	NOUN
fcis-29191	33	22	is	be	AUX
fcis-29191	33	23	large	large	ADJ
fcis-29191	33	24	,	,	PUNCT
fcis-29191	33	25	and	and	CCONJ
fcis-29191	33	26	the	the	DET
fcis-29191	33	27	characteristics	characteristic	NOUN
fcis-29191	33	28	of	of	ADP
fcis-29191	33	29	short	short	ADJ
fcis-29191	33	30	texts	text	NOUN
fcis-29191	33	31	themselves	themselves	PRON
fcis-29191	33	32	are	be	AUX
fcis-29191	33	33	insufficient	insufficient	ADJ
fcis-29191	33	34	,	,	PUNCT
fcis-29191	33	35	which	which	PRON
fcis-29191	33	36	limits	limit	VERB
fcis-29191	33	37	the	the	DET
fcis-29191	33	38	performance	performance	NOUN
fcis-29191	33	39	of	of	ADP
fcis-29191	33	40	these	these	DET
fcis-29191	33	41	methods	method	NOUN
fcis-29191	33	42	in	in	ADP
fcis-29191	33	43	processing	process	VERB
fcis-29191	33	44	such	such	ADJ
fcis-29191	33	45	data	datum	NOUN
fcis-29191	33	46	.	.	PUNCT
fcis-29191	34	1	with	with	ADP
fcis-29191	34	2	the	the	DET
fcis-29191	34	3	continuous	continuous	ADJ
fcis-29191	34	4	advancement	advancement	NOUN
fcis-29191	34	5	of	of	ADP
fcis-29191	34	6	neural	neural	ADJ
fcis-29191	34	7	network	network	NOUN
fcis-29191	34	8	technology	technology	NOUN
fcis-29191	34	9	,	,	PUNCT
fcis-29191	34	10	more	more	ADJ
fcis-29191	34	11	and	and	CCONJ
fcis-29191	34	12	more	more	ADV
fcis-29191	34	13	deep	deep	ADJ
fcis-29191	34	14	learning	learning	NOUN
fcis-29191	34	15	models	model	NOUN
fcis-29191	34	16	have	have	AUX
fcis-29191	34	17	been	be	AUX
fcis-29191	34	18	introduced	introduce	VERB
fcis-29191	34	19	into	into	ADP
fcis-29191	34	20	text	text	NOUN
fcis-29191	34	21	classification	classification	NOUN
fcis-29191	34	22	tasks	task	NOUN
fcis-29191	34	23	,	,	PUNCT
fcis-29191	34	24	including	include	VERB
fcis-29191	34	25	convolutional	convolutional	ADJ
fcis-29191	34	26	neural	neural	ADJ
fcis-29191	34	27	networks	network	NOUN
fcis-29191	34	28	(	(	PUNCT
fcis-29191	34	29	cnn)[5	cnn)[5	PROPN
fcis-29191	34	30	]	]	PUNCT
fcis-29191	34	31	,	,	PUNCT
fcis-29191	34	32	recurrent	recurrent	ADJ
fcis-29191	34	33	neural	neural	ADJ
fcis-29191	34	34	networks	network	NOUN
fcis-29191	34	35	(	(	PUNCT
fcis-29191	34	36	rnn)[6	rnn)[6	NOUN
fcis-29191	34	37	]	]	PUNCT
fcis-29191	34	38	,	,	PUNCT
fcis-29191	34	39	and	and	CCONJ
fcis-29191	34	40	attention	attention	NOUN
fcis-29191	34	41	based	base	VERB
fcis-29191	34	42	models[7	models[7	PROPN
fcis-29191	34	43	]	]	X
fcis-29191	34	44	.	.	PUNCT
fcis-29191	35	1	the	the	DET
fcis-29191	35	2	emergence	emergence	NOUN
fcis-29191	35	3	of	of	ADP
fcis-29191	35	4	bert[8	bert[8	NOUN
fcis-29191	35	5	]	]	PUNCT
fcis-29191	35	6	has	have	AUX
fcis-29191	35	7	brought	bring	VERB
fcis-29191	35	8	revolutionary	revolutionary	ADJ
fcis-29191	35	9	changes	change	NOUN
fcis-29191	35	10	to	to	ADP
fcis-29191	35	11	natural	natural	ADJ
fcis-29191	35	12	language	language	NOUN
fcis-29191	35	13	processing	processing	NOUN
fcis-29191	35	14	technology	technology	NOUN
fcis-29191	35	15	,	,	PUNCT
fcis-29191	35	16	as	as	SCONJ
fcis-29191	35	17	it	it	PRON
fcis-29191	35	18	can	can	AUX
fcis-29191	35	19	generate	generate	VERB
fcis-29191	35	20	word	word	NOUN
fcis-29191	35	21	vectors	vector	NOUN
fcis-29191	35	22	containing	contain	VERB
fcis-29191	35	23	contextual	contextual	ADJ
fcis-29191	35	24	information	information	NOUN
fcis-29191	35	25	,	,	PUNCT
fcis-29191	35	26	significantly	significantly	ADV
fcis-29191	35	27	improving	improve	VERB
fcis-29191	35	28	the	the	DET
fcis-29191	35	29	semantic	semantic	ADJ
fcis-29191	35	30	understanding	understanding	NOUN
fcis-29191	35	31	ability	ability	NOUN
fcis-29191	35	32	of	of	ADP
fcis-29191	35	33	models	model	NOUN
fcis-29191	35	34	.	.	PUNCT
fcis-29191	36	1	many	many	ADJ
fcis-29191	36	2	researchers	researcher	NOUN
fcis-29191	36	3	have	have	AUX
fcis-29191	36	4	studied	study	VERB
fcis-29191	36	5	bert	bert	NOUN
fcis-29191	36	6	based	base	VERB
fcis-29191	36	7	text	text	NOUN
fcis-29191	36	8	classification	classification	NOUN
fcis-29191	36	9	models	model	NOUN
fcis-29191	36	10	,	,	PUNCT
fcis-29191	36	11	which	which	PRON
fcis-29191	36	12	have	have	AUX
fcis-29191	36	13	achieved	achieve	VERB
fcis-29191	36	14	better	well	ADJ
fcis-29191	36	15	performance	performance	NOUN
fcis-29191	36	16	than	than	ADP
fcis-29191	36	17	the	the	DET
fcis-29191	36	18	above	above	ADJ
fcis-29191	36	19	models	model	NOUN
fcis-29191	36	20	in	in	ADP
fcis-29191	36	21	text	text	NOUN
fcis-29191	36	22	classification	classification	NOUN
fcis-29191	36	23	and	and	CCONJ
fcis-29191	36	24	multiple	multiple	ADJ
fcis-29191	36	25	nlp	nlp	NOUN
fcis-29191	36	26	tasks	task	NOUN
fcis-29191	36	27	.	.	PUNCT
fcis-29191	37	1	by	by	ADP
fcis-29191	37	2	conducting	conduct	VERB
fcis-29191	37	3	unsupervised	unsupervised	ADJ
fcis-29191	37	4	training	training	NOUN
fcis-29191	37	5	on	on	ADP
fcis-29191	37	6	large	large	ADJ
fcis-29191	37	7	-	-	PUNCT
fcis-29191	37	8	scale	scale	NOUN
fcis-29191	37	9	unlabeled	unlabeled	NOUN
fcis-29191	37	10	corpora	corpora	PROPN
fcis-29191	37	11	,	,	PUNCT
fcis-29191	37	12	pre	pre	X
fcis-29191	37	13	trained	train	VERB
fcis-29191	37	14	models	model	NOUN
fcis-29191	37	15	can	can	AUX
fcis-29191	37	16	deeply	deeply	ADV
fcis-29191	37	17	explore	explore	VERB
fcis-29191	37	18	potential	potential	ADJ
fcis-29191	37	19	semantic	semantic	ADJ
fcis-29191	37	20	information	information	NOUN
fcis-29191	37	21	in	in	ADP
fcis-29191	37	22	language	language	NOUN
fcis-29191	37	23	and	and	CCONJ
fcis-29191	37	24	demonstrate	demonstrate	VERB
fcis-29191	37	25	superior	superior	ADJ
fcis-29191	37	26	performance	performance	NOUN
fcis-29191	37	27	in	in	ADP
fcis-29191	37	28	specific	specific	ADJ
fcis-29191	37	29	tasks	task	NOUN
fcis-29191	37	30	through	through	ADP
fcis-29191	37	31	fine	fine	ADV
fcis-29191	37	32	-	-	PUNCT
fcis-29191	37	33	tuning	tuning	NOUN
fcis-29191	37	34	.	.	PUNCT
fcis-29191	38	1	with	with	ADP
fcis-29191	38	2	its	its	PRON
fcis-29191	38	3	powerful	powerful	ADJ
fcis-29191	38	4	ability	ability	NOUN
fcis-29191	38	5	to	to	PART
fcis-29191	38	6	express	express	VERB
fcis-29191	38	7	non	non	ADJ
fcis-29191	38	8	euclidean	euclidean	ADJ
fcis-29191	38	9	geometric	geometric	ADJ
fcis-29191	38	10	relationships	relationship	NOUN
fcis-29191	38	11	,	,	PUNCT
fcis-29191	38	12	graph	graph	NOUN
fcis-29191	38	13	neural	neural	ADJ
fcis-29191	38	14	networks	network	NOUN
fcis-29191	38	15	(	(	PUNCT
fcis-29191	38	16	gnns)[9	gnns)[9	X
fcis-29191	38	17	]	]	PUNCT
fcis-29191	38	18	have	have	AUX
fcis-29191	38	19	been	be	AUX
fcis-29191	38	20	applied	apply	VERB
fcis-29191	38	21	in	in	ADP
fcis-29191	38	22	multiple	multiple	ADJ
fcis-29191	38	23	practical	practical	ADJ
fcis-29191	38	24	scenarios	scenario	NOUN
fcis-29191	38	25	.	.	PUNCT
fcis-29191	39	1	graph	graph	VERB
fcis-29191	39	2	convolutional	convolutional	ADJ
fcis-29191	39	3	network	network	NOUN
fcis-29191	39	4	(	(	PUNCT
fcis-29191	39	5	gcn	gcn	NOUN
fcis-29191	39	6	)	)	PUNCT
fcis-29191	40	1	[	[	X
fcis-29191	40	2	10	10	NUM
fcis-29191	40	3	]	]	PUNCT
fcis-29191	40	4	is	be	AUX
fcis-29191	40	5	a	a	DET
fcis-29191	40	6	variant	variant	NOUN
fcis-29191	40	7	of	of	ADP
fcis-29191	40	8	gnn	gnn	PROPN
fcis-29191	40	9	that	that	PRON
fcis-29191	40	10	can	can	AUX
fcis-29191	40	11	capture	capture	VERB
fcis-29191	40	12	information	information	NOUN
fcis-29191	40	13	of	of	ADP
fcis-29191	40	14	high	high	ADJ
fcis-29191	40	15	-	-	PUNCT
fcis-29191	40	16	order	order	NOUN
fcis-29191	40	17	neighboring	neighboring	NOUN
fcis-29191	40	18	nodes	node	NOUN
fcis-29191	40	19	.	.	PUNCT
fcis-29191	41	1	gcn	gcn	NOUN
fcis-29191	41	2	updates	update	VERB
fcis-29191	41	3	node	node	NOUN
fcis-29191	41	4	features	feature	NOUN
fcis-29191	41	5	by	by	ADP
fcis-29191	41	6	multiplying	multiply	VERB
fcis-29191	41	7	the	the	DET
fcis-29191	41	8	adjacency	adjacency	NOUN
fcis-29191	41	9	matrix	matrix	NOUN
fcis-29191	41	10	and	and	CCONJ
fcis-29191	41	11	node	node	ADJ
fcis-29191	41	12	feature	feature	NOUN
fcis-29191	41	13	matrix	matrix	NOUN
fcis-29191	41	14	,	,	PUNCT
fcis-29191	41	15	and	and	CCONJ
fcis-29191	41	16	is	be	AUX
fcis-29191	41	17	used	use	VERB
fcis-29191	41	18	to	to	PART
fcis-29191	41	19	perform	perform	VERB
fcis-29191	41	20	various	various	ADJ
fcis-29191	41	21	tasks	task	NOUN
fcis-29191	41	22	such	such	ADJ
fcis-29191	41	23	as	as	ADP
fcis-29191	41	24	node	node	ADJ
fcis-29191	41	25	classification	classification	NOUN
fcis-29191	41	26	and	and	CCONJ
fcis-29191	41	27	graph	graph	NOUN
fcis-29191	41	28	classification	classification	NOUN
fcis-29191	41	29	.	.	PUNCT
fcis-29191	42	1	yao[11	yao[11	PROPN
fcis-29191	42	2	]	]	PUNCT
fcis-29191	42	3	et	et	PROPN
fcis-29191	42	4	al	al	PROPN
fcis-29191	42	5	.	.	PROPN
fcis-29191	42	6	proposed	propose	VERB
fcis-29191	42	7	a	a	DET
fcis-29191	42	8	text	text	NOUN
fcis-29191	42	9	graph	graph	NOUN
fcis-29191	42	10	convolutional	convolutional	ADJ
fcis-29191	42	11	network	network	NOUN
fcis-29191	42	12	(	(	PUNCT
fcis-29191	42	13	textgcn	textgcn	ADJ
fcis-29191	42	14	)	)	PUNCT
fcis-29191	42	15	model	model	NOUN
fcis-29191	42	16	,	,	PUNCT
fcis-29191	42	17	which	which	PRON
fcis-29191	42	18	constructs	construct	VERB
fcis-29191	42	19	the	the	DET
fcis-29191	42	20	entire	entire	ADJ
fcis-29191	42	21	dataset	dataset	NOUN
fcis-29191	42	22	into	into	ADP
fcis-29191	42	23	a	a	DET
fcis-29191	42	24	heterogeneous	heterogeneous	ADJ
fcis-29191	42	25	text	text	NOUN
fcis-29191	42	26	graph	graph	NOUN
fcis-29191	42	27	containing	contain	VERB
fcis-29191	42	28	word	word	NOUN
fcis-29191	42	29	nodes	node	NOUN
fcis-29191	42	30	and	and	CCONJ
fcis-29191	42	31	document	document	NOUN
fcis-29191	42	32	nodes	node	NOUN
fcis-29191	42	33	,	,	PUNCT
fcis-29191	42	34	and	and	CCONJ
fcis-29191	42	35	effectively	effectively	ADV
fcis-29191	42	36	extracts	extract	VERB
fcis-29191	42	37	global	global	ADJ
fcis-29191	42	38	word	word	NOUN
fcis-29191	42	39	cooccurrence	cooccurrence	NOUN
fcis-29191	42	40	features	feature	VERB
fcis-29191	42	41	.	.	PUNCT
fcis-29191	43	1	graph	graph	NOUN
fcis-29191	43	2	attention	attention	NOUN
fcis-29191	43	3	network	network	NOUN
fcis-29191	43	4	(	(	PUNCT
fcis-29191	43	5	gat)[12	gat)[12	PROPN
fcis-29191	43	6	]	]	X
fcis-29191	43	7	utilizes	utilize	VERB
fcis-29191	43	8	masked	mask	VERB
fcis-29191	43	9	self	self	NOUN
fcis-29191	43	10	-	-	PUNCT
fcis-29191	43	11	attention	attention	NOUN
fcis-29191	43	12	layers	layer	NOUN
fcis-29191	43	13	to	to	PART
fcis-29191	43	14	manipulate	manipulate	VERB
fcis-29191	43	15	graph	graph	NOUN
fcis-29191	43	16	structured	structure	VERB
fcis-29191	43	17	data	datum	NOUN
fcis-29191	43	18	,	,	PUNCT
fcis-29191	43	19	and	and	CCONJ
fcis-29191	43	20	evaluates	evaluate	VERB
fcis-29191	43	21	the	the	DET
fcis-29191	43	22	importance	importance	NOUN
fcis-29191	43	23	of	of	ADP
fcis-29191	43	24	neighboring	neighboring	NOUN
fcis-29191	43	25	nodes	node	NOUN
fcis-29191	43	26	to	to	ADP
fcis-29191	43	27	the	the	DET
fcis-29191	43	28	target	target	NOUN
fcis-29191	43	29	node	node	NOUN
fcis-29191	43	30	through	through	ADP
fcis-29191	43	31	attention	attention	NOUN
fcis-29191	43	32	mechanisms	mechanism	NOUN
fcis-29191	43	33	when	when	SCONJ
fcis-29191	43	34	aggregating	aggregate	VERB
fcis-29191	43	35	neighborhood	neighborhood	NOUN
fcis-29191	43	36	information	information	NOUN
fcis-29191	43	37	of	of	ADP
fcis-29191	43	38	graph	graph	NOUN
fcis-29191	43	39	nodes	node	NOUN
fcis-29191	43	40	.	.	PUNCT
fcis-29191	44	1	the	the	DET
fcis-29191	44	2	heterogeneous	heterogeneous	ADJ
fcis-29191	44	3	graph	graph	NOUN
fcis-29191	44	4	attention	attention	NOUN
fcis-29191	44	5	network	network	NOUN
fcis-29191	44	6	(	(	PUNCT
fcis-29191	44	7	hgat	hgat	NOUN
fcis-29191	44	8	)	)	PUNCT
fcis-29191	44	9	with	with	ADP
fcis-29191	44	10	dual	dual	ADJ
fcis-29191	44	11	layer	layer	NOUN
fcis-29191	44	12	attention	attention	NOUN
fcis-29191	44	13	mechanism[13	mechanism[13	NOUN
fcis-29191	44	14	]	]	PUNCT
fcis-29191	44	15	learns	learn	VERB
fcis-29191	44	16	the	the	DET
fcis-29191	44	17	importance	importance	NOUN
fcis-29191	44	18	of	of	ADP
fcis-29191	44	19	different	different	ADJ
fcis-29191	44	20	adjacent	adjacent	ADJ
fcis-29191	44	21	nodes	node	NOUN
fcis-29191	44	22	and	and	CCONJ
fcis-29191	44	23	node	node	ADJ
fcis-29191	44	24	types	type	NOUN
fcis-29191	44	25	in	in	ADP
fcis-29191	44	26	a	a	DET
fcis-29191	44	27	node	node	NOUN
fcis-29191	44	28	.	.	PUNCT
fcis-29191	45	1	this	this	DET
fcis-29191	45	2	model	model	NOUN
fcis-29191	45	3	solves	solve	VERB
fcis-29191	45	4	the	the	DET
fcis-29191	45	5	sparsity	sparsity	NOUN
fcis-29191	45	6	problem	problem	NOUN
fcis-29191	45	7	in	in	ADP
fcis-29191	45	8	semi	semi	ADV
fcis-29191	45	9	supervised	supervised	ADJ
fcis-29191	45	10	short	short	ADJ
fcis-29191	45	11	text	text	NOUN
fcis-29191	45	12	classification	classification	NOUN
fcis-29191	45	13	by	by	ADP
fcis-29191	45	14	propagating	propagate	VERB
fcis-29191	45	15	information	information	NOUN
fcis-29191	45	16	in	in	ADP
fcis-29191	45	17	the	the	DET
fcis-29191	45	18	graph	graph	NOUN
fcis-29191	45	19	structure	structure	NOUN
fcis-29191	45	20	and	and	CCONJ
fcis-29191	45	21	capturing	capture	VERB
fcis-29191	45	22	the	the	DET
fcis-29191	45	23	relationships	relationship	NOUN
fcis-29191	45	24	between	between	ADP
fcis-29191	45	25	nodes	node	NOUN
fcis-29191	45	26	.	.	PUNCT
fcis-29191	46	1	lin[14	lin[14	PROPN
fcis-29191	46	2	]	]	PUNCT
fcis-29191	46	3	combined	combine	VERB
fcis-29191	46	4	large	large	ADJ
fcis-29191	46	5	-	-	PUNCT
fcis-29191	46	6	scale	scale	NOUN
fcis-29191	46	7	pre	pre	ADJ
fcis-29191	46	8	-	-	ADJ
fcis-29191	46	9	trained	train	VERB
fcis-29191	46	10	models	model	NOUN
fcis-29191	46	11	with	with	ADP
fcis-29191	46	12	gcn	gcn	NOUN
fcis-29191	46	13	and	and	CCONJ
fcis-29191	46	14	used	use	VERB
fcis-29191	46	15	graphs	graph	NOUN
fcis-29191	46	16	to	to	PART
fcis-29191	46	17	model	model	VERB
fcis-29191	46	18	the	the	DET
fcis-29191	46	19	relationships	relationship	NOUN
fcis-29191	46	20	between	between	ADP
fcis-29191	46	21	different	different	ADJ
fcis-29191	46	22	samples	sample	NOUN
fcis-29191	46	23	in	in	ADP
fcis-29191	46	24	the	the	DET
fcis-29191	46	25	entire	entire	ADJ
fcis-29191	46	26	corpus	corpus	NOUN
fcis-29191	46	27	,	,	PUNCT
fcis-29191	46	28	achieving	achieve	VERB
fcis-29191	46	29	advantageous	advantageous	ADJ
fcis-29191	46	30	results	result	NOUN
fcis-29191	46	31	on	on	ADP
fcis-29191	46	32	public	public	ADJ
fcis-29191	46	33	datasets	dataset	NOUN
fcis-29191	46	34	.	.	PUNCT
fcis-29191	47	1	2.2	2.2	NUM
fcis-29191	47	2	.	.	PUNCT
fcis-29191	48	1	classification	classification	NOUN
fcis-29191	48	2	of	of	ADP
fcis-29191	48	3	cyber	cyber	ADJ
fcis-29191	48	4	violence	violence	NOUN
fcis-29191	48	5	text	text	NOUN
fcis-29191	48	6	due	due	ADP
fcis-29191	48	7	to	to	ADP
fcis-29191	48	8	the	the	DET
fcis-29191	48	9	enormous	enormous	ADJ
fcis-29191	48	10	threat	threat	NOUN
fcis-29191	48	11	posed	pose	VERB
fcis-29191	48	12	by	by	ADP
fcis-29191	48	13	online	online	ADJ
fcis-29191	48	14	violence	violence	NOUN
fcis-29191	48	15	to	to	ADP
fcis-29191	48	16	society	society	NOUN
fcis-29191	48	17	,	,	PUNCT
fcis-29191	48	18	online	online	ADJ
fcis-29191	48	19	violence	violence	NOUN
fcis-29191	48	20	detection	detection	NOUN
fcis-29191	48	21	has	have	AUX
fcis-29191	48	22	gradually	gradually	ADV
fcis-29191	48	23	become	become	VERB
fcis-29191	48	24	a	a	DET
fcis-29191	48	25	research	research	NOUN
fcis-29191	48	26	hotspot	hotspot	NOUN
fcis-29191	48	27	.	.	PUNCT
fcis-29191	49	1	dinakar[15	dinakar[15	NOUN
fcis-29191	49	2	]	]	X
fcis-29191	49	3	et	et	PROPN
fcis-29191	49	4	al	al	PROPN
fcis-29191	49	5	.	.	PROPN
fcis-29191	49	6	used	use	VERB
fcis-29191	49	7	naive	naive	ADJ
fcis-29191	49	8	bayes	bayes	NOUN
fcis-29191	49	9	and	and	CCONJ
fcis-29191	49	10	svm	svm	VERB
fcis-29191	49	11	to	to	PART
fcis-29191	49	12	construct	construct	VERB
fcis-29191	49	13	a	a	DET
fcis-29191	49	14	separate	separate	ADJ
fcis-29191	49	15	topic	topic	NOUN
fcis-29191	49	16	sensitive	sensitive	ADJ
fcis-29191	49	17	classifier	classifier	NOUN
fcis-29191	49	18	for	for	ADP
fcis-29191	49	19	detecting	detect	VERB
fcis-29191	49	20	cyberbullying	cyberbullye	VERB
fcis-29191	49	21	texts	text	NOUN
fcis-29191	49	22	.	.	PUNCT
fcis-29191	50	1	chavan	chavan	NOUN
fcis-29191	50	2	and	and	CCONJ
fcis-29191	50	3	shylaja[16	shylaja[16	PROPN
fcis-29191	50	4	]	]	PUNCT
fcis-29191	50	5	proposed	propose	VERB
fcis-29191	50	6	an	an	DET
fcis-29191	50	7	algorithm	algorithm	NOUN
fcis-29191	50	8	to	to	PART
fcis-29191	50	9	calculate	calculate	VERB
fcis-29191	50	10	a	a	DET
fcis-29191	50	11	score	score	NOUN
fcis-29191	50	12	representing	represent	VERB
fcis-29191	50	13	the	the	DET
fcis-29191	50	14	probability	probability	NOUN
fcis-29191	50	15	of	of	ADP
fcis-29191	50	16	comments	comment	NOUN
fcis-29191	50	17	offending	offend	VERB
fcis-29191	50	18	other	other	ADJ
fcis-29191	50	19	users	user	NOUN
fcis-29191	50	20	,	,	PUNCT
fcis-29191	50	21	utilizing	utilize	VERB
fcis-29191	50	22	a	a	DET
fcis-29191	50	23	series	series	NOUN
fcis-29191	50	24	of	of	ADP
fcis-29191	50	25	features	feature	NOUN
fcis-29191	50	26	such	such	ADJ
fcis-29191	50	27	as	as	ADP
fcis-29191	50	28	skip	skip	ADJ
fcis-29191	50	29	grams	gram	NOUN
fcis-29191	50	30	to	to	PART
fcis-29191	50	31	detect	detect	VERB
fcis-29191	50	32	aggressive	aggressive	ADJ
fcis-29191	50	33	comments	comment	NOUN
fcis-29191	50	34	targeting	target	VERB
fcis-29191	50	35	peers	peer	NOUN
fcis-29191	50	36	.	.	PUNCT
fcis-29191	51	1	squicciarini[17]et	squicciarini[17]et	PROPN
fcis-29191	51	2	al	al	PROPN
fcis-29191	51	3	.	.	PROPN
fcis-29191	51	4	applied	apply	VERB
fcis-29191	51	5	decision	decision	NOUN
fcis-29191	51	6	tree	tree	NOUN
fcis-29191	51	7	classifiers	classifier	NOUN
fcis-29191	51	8	to	to	ADP
fcis-29191	51	9	specific	specific	ADJ
fcis-29191	51	10	content	content	NOUN
fcis-29191	51	11	features	feature	NOUN
fcis-29191	51	12	to	to	PART
fcis-29191	51	13	detect	detect	VERB
fcis-29191	51	14	cyberbullying	cyberbullye	VERB
fcis-29191	51	15	behavior	behavior	NOUN
fcis-29191	51	16	in	in	ADP
fcis-29191	51	17	social	social	ADJ
fcis-29191	51	18	networks	network	NOUN
fcis-29191	51	19	such	such	ADJ
fcis-29191	51	20	as	as	ADP
fcis-29191	51	21	myspace	myspace	NOUN
fcis-29191	51	22	and	and	CCONJ
fcis-29191	51	23	spring.me	spring.me	PROPN
fcis-29191	51	24	.	.	PUNCT
fcis-29191	52	1	at	at	ADP
fcis-29191	52	2	present	present	ADJ
fcis-29191	52	3	,	,	PUNCT
fcis-29191	52	4	deep	deep	ADJ
fcis-29191	52	5	learning	learning	NOUN
fcis-29191	52	6	models	model	NOUN
fcis-29191	52	7	have	have	AUX
fcis-29191	52	8	also	also	ADV
fcis-29191	52	9	become	become	VERB
fcis-29191	52	10	the	the	DET
fcis-29191	52	11	main	main	ADJ
fcis-29191	52	12	method	method	NOUN
fcis-29191	52	13	for	for	ADP
fcis-29191	52	14	detecting	detect	VERB
fcis-29191	52	15	online	online	ADJ
fcis-29191	52	16	violent	violent	ADJ
fcis-29191	52	17	texts	text	NOUN
fcis-29191	52	18	.	.	PUNCT
fcis-29191	53	1	sweta	sweta	PROPN
fcis-29191	53	2	agrawal[18	agrawal[18	X
fcis-29191	53	3	]	]	PUNCT
fcis-29191	53	4	uses	use	VERB
fcis-29191	53	5	four	four	NUM
fcis-29191	53	6	models	model	NOUN
fcis-29191	53	7	,	,	PUNCT
fcis-29191	53	8	cnn	cnn	PROPN
fcis-29191	53	9	,	,	PUNCT
fcis-29191	53	10	lstm	lstm	PROPN
fcis-29191	53	11	,	,	PUNCT
fcis-29191	53	12	blstm	blstm	NOUN
fcis-29191	53	13	,	,	PUNCT
fcis-29191	53	14	and	and	CCONJ
fcis-29191	53	15	blstm	blstm	NOUN
fcis-29191	53	16	,	,	PUNCT
fcis-29191	53	17	to	to	PART
fcis-29191	53	18	analyze	analyze	VERB
fcis-29191	53	19	online	online	ADJ
fcis-29191	53	20	violent	violent	ADJ
fcis-29191	53	21	texts	text	NOUN
fcis-29191	53	22	on	on	ADP
fcis-29191	53	23	multiple	multiple	ADJ
fcis-29191	53	24	social	social	ADJ
fcis-29191	53	25	media	medium	NOUN
fcis-29191	53	26	platforms	platform	NOUN
fcis-29191	53	27	.	.	PUNCT
fcis-29191	54	1	daraghmi[19	daraghmi[19	VERB
fcis-29191	54	2	]	]	PUNCT
fcis-29191	54	3	et	et	PROPN
fcis-29191	54	4	al	al	PROPN
fcis-29191	54	5	.	.	PROPN
fcis-29191	54	6	compared	compare	VERB
fcis-29191	54	7	and	and	CCONJ
fcis-29191	54	8	analyzed	analyze	VERB
fcis-29191	54	9	various	various	ADJ
fcis-29191	54	10	machine	machine	NOUN
fcis-29191	54	11	learning	learning	NOUN
fcis-29191	54	12	and	and	CCONJ
fcis-29191	54	13	deep	deep	ADJ
fcis-29191	54	14	learning	learning	NOUN
fcis-29191	54	15	algorithms	algorithm	NOUN
fcis-29191	54	16	and	and	CCONJ
fcis-29191	54	17	proposed	propose	VERB
fcis-29191	54	18	a	a	DET
fcis-29191	54	19	cnn	cnn	PROPN
fcis-29191	54	20	bilstm	bilstm	PROPN
fcis-29191	54	21	gru	gru	PROPN
fcis-29191	54	22	hybrid	hybrid	PROPN
fcis-29191	54	23	model	model	NOUN
fcis-29191	54	24	for	for	ADP
fcis-29191	54	25	detecting	detect	VERB
fcis-29191	54	26	arabic	arabic	ADJ
fcis-29191	54	27	cyberbullying	cyberbullying	NOUN
fcis-29191	54	28	.	.	PUNCT
fcis-29191	55	1	the	the	DET
fcis-29191	55	2	model	model	NOUN
fcis-29191	55	3	has	have	AUX
fcis-29191	55	4	stacked	stack	VERB
fcis-29191	55	5	word	word	NOUN
fcis-29191	55	6	embedding	embed	VERB
fcis-29191	55	7	function	function	NOUN
fcis-29191	55	8	and	and	CCONJ
fcis-29191	55	9	can	can	AUX
fcis-29191	55	10	adapt	adapt	VERB
fcis-29191	55	11	to	to	ADP
fcis-29191	55	12	different	different	ADJ
fcis-29191	55	13	data	datum	NOUN
fcis-29191	55	14	changes	change	NOUN
fcis-29191	55	15	.	.	PUNCT
fcis-29191	56	1	the	the	DET
fcis-29191	56	2	research	research	NOUN
fcis-29191	56	3	on	on	ADP
fcis-29191	56	4	cyber	cyber	ADJ
fcis-29191	56	5	violence	violence	NOUN
fcis-29191	56	6	in	in	ADP
fcis-29191	56	7	china	china	PROPN
fcis-29191	56	8	is	be	AUX
fcis-29191	56	9	also	also	ADV
fcis-29191	56	10	increasing	increase	VERB
fcis-29191	56	11	.	.	PUNCT
fcis-29191	57	1	huang	huang	PROPN
fcis-29191	57	2	rui[20	rui[20	PROPN
fcis-29191	57	3	]	]	PUNCT
fcis-29191	57	4	proposed	propose	VERB
fcis-29191	57	5	a	a	DET
fcis-29191	57	6	dictionary	dictionary	ADJ
fcis-29191	57	7	and	and	CCONJ
fcis-29191	57	8	rule	rule	NOUN
fcis-29191	57	9	-	-	PUNCT
fcis-29191	57	10	based	base	VERB
fcis-29191	57	11	method	method	NOUN
fcis-29191	57	12	for	for	ADP
fcis-29191	57	13	detecting	detect	VERB
fcis-29191	57	14	cyberbullying	cyberbullye	VERB
fcis-29191	57	15	language	language	NOUN
fcis-29191	57	16	and	and	CCONJ
fcis-29191	57	17	established	establish	VERB
fcis-29191	57	18	a	a	DET
fcis-29191	57	19	corresponding	corresponding	ADJ
fcis-29191	57	20	detection	detection	NOUN
fcis-29191	57	21	system	system	NOUN
fcis-29191	57	22	.	.	PUNCT
fcis-29191	58	1	qiang	qiang	PROPN
fcis-29191	58	2	lan[21	lan[21	PROPN
fcis-29191	58	3	]	]	PUNCT
fcis-29191	58	4	constructed	construct	VERB
fcis-29191	58	5	a	a	DET
fcis-29191	58	6	chinese	chinese	ADJ
fcis-29191	58	7	cyber	cyber	NOUN
fcis-29191	58	8	violent	violent	ADJ
fcis-29191	58	9	text	text	NOUN
fcis-29191	58	10	language	language	NOUN
fcis-29191	58	11	database	database	NOUN
fcis-29191	58	12	and	and	CCONJ
fcis-29191	58	13	implemented	implement	VERB
fcis-29191	58	14	the	the	DET
fcis-29191	58	15	detection	detection	NOUN
fcis-29191	58	16	of	of	ADP
fcis-29191	58	17	cyber	cyber	ADJ
fcis-29191	58	18	violent	violent	ADJ
fcis-29191	58	19	text	text	NOUN
fcis-29191	58	20	using	use	VERB
fcis-29191	58	21	a	a	DET
fcis-29191	58	22	character	character	NOUN
fcis-29191	58	23	level	level	NOUN
fcis-29191	58	24	neural	neural	ADJ
fcis-29191	58	25	network	network	NOUN
fcis-29191	58	26	model	model	NOUN
fcis-29191	58	27	.	.	PUNCT
fcis-29191	59	1	wang	wang	PROPN
fcis-29191	59	2	runyu[22	runyu[22	PROPN
fcis-29191	59	3	]	]	PUNCT
fcis-29191	59	4	provided	provide	VERB
fcis-29191	59	5	a	a	DET
fcis-29191	59	6	comprehensive	comprehensive	ADJ
fcis-29191	59	7	summary	summary	NOUN
fcis-29191	59	8	and	and	CCONJ
fcis-29191	59	9	analysis	analysis	NOUN
fcis-29191	59	10	of	of	ADP
fcis-29191	59	11	research	research	NOUN
fcis-29191	59	12	results	result	NOUN
fcis-29191	59	13	on	on	ADP
fcis-29191	59	14	cyberbullying	cyberbullye	VERB
fcis-29191	59	15	,	,	PUNCT
fcis-29191	59	16	and	and	CCONJ
fcis-29191	59	17	put	put	VERB
fcis-29191	59	18	forward	forward	ADJ
fcis-29191	59	19	thoughts	thought	NOUN
fcis-29191	59	20	and	and	CCONJ
fcis-29191	59	21	suggestions	suggestion	NOUN
fcis-29191	59	22	on	on	ADP
fcis-29191	59	23	intelligent	intelligent	ADJ
fcis-29191	59	24	intervention	intervention	NOUN
fcis-29191	59	25	for	for	ADP
fcis-29191	59	26	cyberbullying	cyberbullye	VERB
fcis-29191	59	27	.	.	PUNCT
fcis-29191	60	1	3	3	X
fcis-29191	60	2	.	.	X
fcis-29191	60	3	cyber	cyber	NOUN
fcis-29191	60	4	violence	violence	NOUN
fcis-29191	60	5	text	text	NOUN
fcis-29191	60	6	classification	classification	NOUN
fcis-29191	60	7	model	model	NOUN
fcis-29191	60	8	fig	fig	NOUN
fcis-29191	60	9	1	1	NUM
fcis-29191	60	10	.	.	PUNCT
fcis-29191	61	1	r	r	X
fcis-29191	61	2	-	-	PUNCT
fcis-29191	61	3	bg	bg	PROPN
fcis-29191	61	4	model	model	NOUN
fcis-29191	61	5	structure	structure	NOUN
fcis-29191	61	6	31	31	NUM
fcis-29191	61	7	in	in	ADP
fcis-29191	61	8	the	the	DET
fcis-29191	61	9	cyber	cyber	NOUN
fcis-29191	61	10	violence	violence	NOUN
fcis-29191	61	11	text	text	NOUN
fcis-29191	61	12	classification	classification	NOUN
fcis-29191	61	13	model	model	NOUN
fcis-29191	61	14	based	base	VERB
fcis-29191	61	15	on	on	ADP
fcis-29191	61	16	graph	graph	NOUN
fcis-29191	61	17	convolutional	convolutional	ADJ
fcis-29191	61	18	networks	network	NOUN
fcis-29191	61	19	and	and	CCONJ
fcis-29191	61	20	syntactic	syntactic	ADJ
fcis-29191	61	21	relationship	relationship	NOUN
fcis-29191	61	22	analysis	analysis	NOUN
fcis-29191	61	23	proposed	propose	VERB
fcis-29191	61	24	in	in	ADP
fcis-29191	61	25	this	this	DET
fcis-29191	61	26	article	article	NOUN
fcis-29191	61	27	(	(	PUNCT
fcis-29191	61	28	relation	relation	NOUN
fcis-29191	61	29	bertgcn	bertgcn	PROPN
fcis-29191	61	30	,	,	PUNCT
fcis-29191	61	31	hereinafter	hereinafter	NOUN
fcis-29191	61	32	referred	refer	VERB
fcis-29191	61	33	to	to	ADP
fcis-29191	61	34	as	as	ADP
fcis-29191	61	35	r	r	NOUN
fcis-29191	61	36	-	-	PUNCT
fcis-29191	61	37	bg	bg	PROPN
fcis-29191	61	38	model	model	NOUN
fcis-29191	61	39	)	)	PUNCT
fcis-29191	61	40	,	,	PUNCT
fcis-29191	61	41	the	the	DET
fcis-29191	61	42	text	text	NOUN
fcis-29191	61	43	is	be	AUX
fcis-29191	61	44	first	first	ADV
fcis-29191	61	45	subjected	subject	VERB
fcis-29191	61	46	to	to	ADP
fcis-29191	61	47	syntactic	syntactic	ADJ
fcis-29191	61	48	analysis	analysis	NOUN
fcis-29191	61	49	,	,	PUNCT
fcis-29191	61	50	and	and	CCONJ
fcis-29191	61	51	the	the	DET
fcis-29191	61	52	extracted	extract	VERB
fcis-29191	61	53	dependency	dependency	NOUN
fcis-29191	61	54	relationship	relationship	NOUN
fcis-29191	61	55	features	feature	NOUN
fcis-29191	61	56	are	be	AUX
fcis-29191	61	57	input	input	VERB
fcis-29191	61	58	together	together	ADV
fcis-29191	61	59	with	with	ADP
fcis-29191	61	60	the	the	DET
fcis-29191	61	61	original	original	ADJ
fcis-29191	61	62	text	text	NOUN
fcis-29191	61	63	into	into	ADP
fcis-29191	61	64	bert	bert	NOUN
fcis-29191	61	65	related	relate	VERB
fcis-29191	61	66	models	model	NOUN
fcis-29191	61	67	(	(	PUNCT
fcis-29191	61	68	bert	bert	PROPN
fcis-29191	61	69	,	,	PUNCT
fcis-29191	61	70	roberta	roberta	PROPN
fcis-29191	61	71	)	)	PUNCT
fcis-29191	61	72	for	for	ADP
fcis-29191	61	73	training	training	NOUN
fcis-29191	61	74	.	.	PUNCT
fcis-29191	62	1	this	this	DET
fcis-29191	62	2	approach	approach	NOUN
fcis-29191	62	3	adds	add	VERB
fcis-29191	62	4	additional	additional	ADJ
fcis-29191	62	5	contextual	contextual	ADJ
fcis-29191	62	6	information	information	NOUN
fcis-29191	62	7	during	during	ADP
fcis-29191	62	8	model	model	NOUN
fcis-29191	62	9	training	training	NOUN
fcis-29191	62	10	,	,	PUNCT
fcis-29191	62	11	enabling	enable	VERB
fcis-29191	62	12	the	the	DET
fcis-29191	62	13	model	model	NOUN
fcis-29191	62	14	to	to	PART
fcis-29191	62	15	more	more	ADV
fcis-29191	62	16	accurately	accurately	ADV
fcis-29191	62	17	capture	capture	VERB
fcis-29191	62	18	the	the	DET
fcis-29191	62	19	dependency	dependency	NOUN
fcis-29191	62	20	relationships	relationship	NOUN
fcis-29191	62	21	and	and	CCONJ
fcis-29191	62	22	semantic	semantic	ADJ
fcis-29191	62	23	connections	connection	NOUN
fcis-29191	62	24	within	within	ADP
fcis-29191	62	25	sentences	sentence	NOUN
fcis-29191	62	26	.	.	PUNCT
fcis-29191	63	1	then	then	ADV
fcis-29191	63	2	use	use	VERB
fcis-29191	63	3	bert	bert	NOUN
fcis-29191	63	4	related	relate	VERB
fcis-29191	63	5	models	model	NOUN
fcis-29191	63	6	to	to	PART
fcis-29191	63	7	generate	generate	VERB
fcis-29191	63	8	initial	initial	ADJ
fcis-29191	63	9	representations	representation	NOUN
fcis-29191	63	10	for	for	ADP
fcis-29191	63	11	each	each	DET
fcis-29191	63	12	document	document	NOUN
fcis-29191	63	13	node	node	NOUN
fcis-29191	63	14	and	and	CCONJ
fcis-29191	63	15	input	input	VERB
fcis-29191	63	16	them	they	PRON
fcis-29191	63	17	into	into	ADP
fcis-29191	63	18	gcn	gcn	NOUN
fcis-29191	63	19	for	for	ADP
fcis-29191	63	20	processing	processing	NOUN
fcis-29191	63	21	.	.	PUNCT
fcis-29191	64	1	the	the	DET
fcis-29191	64	2	document	document	NOUN
fcis-29191	64	3	representation	representation	NOUN
fcis-29191	64	4	is	be	AUX
fcis-29191	64	5	iteratively	iteratively	ADV
fcis-29191	64	6	updated	update	VERB
fcis-29191	64	7	in	in	ADP
fcis-29191	64	8	the	the	DET
fcis-29191	64	9	graph	graph	NOUN
fcis-29191	64	10	structure	structure	NOUN
fcis-29191	64	11	through	through	ADP
fcis-29191	64	12	gcn	gcn	PROPN
fcis-29191	64	13	,	,	PUNCT
fcis-29191	64	14	and	and	CCONJ
fcis-29191	64	15	the	the	DET
fcis-29191	64	16	final	final	ADJ
fcis-29191	64	17	generated	generate	VERB
fcis-29191	64	18	document	document	NOUN
fcis-29191	64	19	node	node	NOUN
fcis-29191	64	20	features	feature	NOUN
fcis-29191	64	21	are	be	AUX
fcis-29191	64	22	fed	feed	VERB
fcis-29191	64	23	into	into	ADP
fcis-29191	64	24	a	a	DET
fcis-29191	64	25	softmax	softmax	NOUN
fcis-29191	64	26	classifier	classifier	NOUN
fcis-29191	64	27	for	for	ADP
fcis-29191	64	28	prediction	prediction	NOUN
fcis-29191	64	29	.	.	PUNCT
fcis-29191	65	1	the	the	DET
fcis-29191	65	2	structure	structure	NOUN
fcis-29191	65	3	of	of	ADP
fcis-29191	65	4	the	the	DET
fcis-29191	65	5	r	r	PROPN
fcis-29191	65	6	-	-	PUNCT
fcis-29191	65	7	bg	bg	PROPN
fcis-29191	65	8	model	model	NOUN
fcis-29191	65	9	is	be	AUX
fcis-29191	65	10	shown	show	VERB
fcis-29191	65	11	in	in	ADP
fcis-29191	65	12	figure	figure	NOUN
fcis-29191	65	13	1	1	NUM
fcis-29191	65	14	.	.	SYM
fcis-29191	65	15	3.1	3.1	NUM
fcis-29191	65	16	.	.	PUNCT
fcis-29191	66	1	syntactic	syntactic	ADJ
fcis-29191	66	2	parsing	parsing	NOUN
fcis-29191	66	3	firstly	firstly	ADV
fcis-29191	66	4	,	,	PUNCT
fcis-29191	66	5	label	label	VERB
fcis-29191	66	6	the	the	DET
fcis-29191	66	7	original	original	ADJ
fcis-29191	66	8	text	text	NOUN
fcis-29191	66	9	and	and	CCONJ
fcis-29191	66	10	group	group	NOUN
fcis-29191	66	11	the	the	DET
fcis-29191	66	12	labels	label	NOUN
fcis-29191	66	13	into	into	ADP
fcis-29191	66	14	sentences	sentence	NOUN
fcis-29191	66	15	.	.	PUNCT
fcis-29191	67	1	then	then	ADV
fcis-29191	67	2	,	,	PUNCT
fcis-29191	67	3	based	base	VERB
fcis-29191	67	4	on	on	ADP
fcis-29191	67	5	a	a	DET
fcis-29191	67	6	bidirectional	bidirectional	ADJ
fcis-29191	67	7	long	long	ADJ
fcis-29191	67	8	short	short	ADJ
fcis-29191	67	9	-	-	PUNCT
fcis-29191	67	10	term	term	NOUN
fcis-29191	67	11	memory	memory	NOUN
fcis-29191	67	12	network	network	NOUN
fcis-29191	67	13	,	,	PUNCT
fcis-29191	67	14	assign	assign	VERB
fcis-29191	67	15	a	a	DET
fcis-29191	67	16	part	part	NOUN
fcis-29191	67	17	of	of	ADP
fcis-29191	67	18	speech	speech	NOUN
fcis-29191	67	19	(	(	PUNCT
fcis-29191	67	20	pos	pos	NOUN
fcis-29191	67	21	)	)	PUNCT
fcis-29191	67	22	to	to	ADP
fcis-29191	67	23	each	each	DET
fcis-29191	67	24	word	word	NOUN
fcis-29191	67	25	in	in	ADP
fcis-29191	67	26	the	the	DET
fcis-29191	67	27	sentence	sentence	NOUN
fcis-29191	67	28	and	and	CCONJ
fcis-29191	67	29	analyze	analyze	VERB
fcis-29191	67	30	its	its	PRON
fcis-29191	67	31	universal	universal	ADJ
fcis-29191	67	32	morphological	morphological	ADJ
fcis-29191	67	33	features	feature	NOUN
fcis-29191	67	34	(	(	PUNCT
fcis-29191	67	35	ufeats	ufeat	NOUN
fcis-29191	67	36	)	)	PUNCT
fcis-29191	67	37	.	.	PUNCT
fcis-29191	68	1	in	in	ADP
fcis-29191	68	2	order	order	NOUN
fcis-29191	68	3	to	to	PART
fcis-29191	68	4	ensure	ensure	VERB
fcis-29191	68	5	consistency	consistency	NOUN
fcis-29191	68	6	between	between	ADP
fcis-29191	68	7	universal	universal	ADJ
fcis-29191	68	8	pos	pos	NOUN
fcis-29191	68	9	(	(	PUNCT
fcis-29191	68	10	upos	upos	NOUN
fcis-29191	68	11	)	)	PUNCT
fcis-29191	68	12	,	,	PUNCT
fcis-29191	68	13	tree	tree	NOUN
fcis-29191	68	14	library	library	NOUN
fcis-29191	68	15	specific	specific	ADJ
fcis-29191	68	16	pos	pos	NOUN
fcis-29191	68	17	(	(	PUNCT
fcis-29191	68	18	xpos	xpos	PROPN
fcis-29191	68	19	)	)	PUNCT
fcis-29191	68	20	,	,	PUNCT
fcis-29191	68	21	and	and	CCONJ
fcis-29191	68	22	ufeats	ufeat	NOUN
fcis-29191	68	23	,	,	PUNCT
fcis-29191	68	24	the	the	DET
fcis-29191	68	25	biafine	biafine	NOUN
fcis-29191	68	26	scoring	scoring	NOUN
fcis-29191	68	27	mechanism	mechanism	NOUN
fcis-29191	68	28	proposed	propose	VERB
fcis-29191	68	29	by	by	ADP
fcis-29191	68	30	dozat	dozat	NOUN
fcis-29191	68	31	and	and	CCONJ
fcis-29191	68	32	manning	manning	NOUN
fcis-29191	68	33	was	be	AUX
fcis-29191	68	34	used	use	VERB
fcis-29191	68	35	to	to	PART
fcis-29191	68	36	convert	convert	VERB
fcis-29191	68	37	xpos	xpos	PROPN
fcis-29191	68	38	and	and	CCONJ
fcis-29191	68	39	ufeats	ufeats	PROPN
fcis-29191	68	40	prediction	prediction	NOUN
fcis-29191	68	41	conditions	condition	NOUN
fcis-29191	68	42	into	into	ADP
fcis-29191	68	43	upos	upos	NOUN
fcis-29191	68	44	prediction	prediction	NOUN
fcis-29191	68	45	conditions	condition	NOUN
fcis-29191	68	46	.	.	PUNCT
fcis-29191	69	1	adjust	adjust	VERB
fcis-29191	69	2	the	the	DET
fcis-29191	69	3	predictions	prediction	NOUN
fcis-29191	69	4	of	of	ADP
fcis-29191	69	5	xpos	xpos	PROPN
fcis-29191	69	6	and	and	CCONJ
fcis-29191	69	7	ufeats	ufeat	NOUN
fcis-29191	69	8	based	base	VERB
fcis-29191	69	9	on	on	ADP
fcis-29191	69	10	upos	upos	NOUN
fcis-29191	69	11	predictions	prediction	NOUN
fcis-29191	69	12	.	.	PUNCT
fcis-29191	70	1	finally	finally	ADV
fcis-29191	70	2	,	,	PUNCT
fcis-29191	70	3	neural	neural	ADJ
fcis-29191	70	4	dependency	dependency	NOUN
fcis-29191	70	5	syntactic	syntactic	ADJ
fcis-29191	70	6	analysis	analysis	NOUN
fcis-29191	70	7	was	be	AUX
fcis-29191	70	8	implemented	implement	VERB
fcis-29191	70	9	based	base	VERB
fcis-29191	70	10	on	on	ADP
fcis-29191	70	11	bilstm	bilstm	NOUN
fcis-29191	70	12	's	's	PART
fcis-29191	70	13	deep	deep	ADJ
fcis-29191	70	14	double	double	ADJ
fcis-29191	70	15	affine	affine	NOUN
fcis-29191	70	16	approach	approach	NOUN
fcis-29191	70	17	,	,	PUNCT
fcis-29191	70	18	parsing	parse	VERB
fcis-29191	70	19	the	the	DET
fcis-29191	70	20	syntactic	syntactic	ADJ
fcis-29191	70	21	structure	structure	NOUN
fcis-29191	70	22	of	of	ADP
fcis-29191	70	23	each	each	DET
fcis-29191	70	24	sentence	sentence	NOUN
fcis-29191	70	25	.	.	PUNCT
fcis-29191	71	1	each	each	DET
fcis-29191	71	2	word	word	NOUN
fcis-29191	71	3	in	in	ADP
fcis-29191	71	4	the	the	DET
fcis-29191	71	5	sentence	sentence	NOUN
fcis-29191	71	6	was	be	AUX
fcis-29191	71	7	assigned	assign	VERB
fcis-29191	71	8	a	a	DET
fcis-29191	71	9	syntactic	syntactic	ADJ
fcis-29191	71	10	header	header	NOUN
fcis-29191	71	11	,	,	PUNCT
fcis-29191	71	12	which	which	PRON
fcis-29191	71	13	is	be	AUX
fcis-29191	71	14	another	another	DET
fcis-29191	71	15	word	word	NOUN
fcis-29191	71	16	or	or	CCONJ
fcis-29191	71	17	an	an	DET
fcis-29191	71	18	artificial	artificial	ADJ
fcis-29191	71	19	root	root	NOUN
fcis-29191	71	20	symbol	symbol	NOUN
fcis-29191	71	21	in	in	ADP
fcis-29191	71	22	the	the	DET
fcis-29191	71	23	sentence	sentence	NOUN
fcis-29191	71	24	.	.	PUNCT
fcis-29191	72	1	the	the	DET
fcis-29191	72	2	model	model	NOUN
fcis-29191	72	3	was	be	AUX
fcis-29191	72	4	further	far	ADV
fcis-29191	72	5	expanded	expand	VERB
fcis-29191	72	6	with	with	ADP
fcis-29191	72	7	the	the	DET
fcis-29191	72	8	linearization	linearization	NOUN
fcis-29191	72	9	order	order	NOUN
fcis-29191	72	10	of	of	ADP
fcis-29191	72	11	the	the	DET
fcis-29191	72	12	two	two	NUM
fcis-29191	72	13	words	word	NOUN
fcis-29191	72	14	and	and	CCONJ
fcis-29191	72	15	their	their	PRON
fcis-29191	72	16	typical	typical	ADJ
fcis-29191	72	17	linear	linear	ADJ
fcis-29191	72	18	distance	distance	NOUN
fcis-29191	72	19	to	to	PART
fcis-29191	72	20	improve	improve	VERB
fcis-29191	72	21	the	the	DET
fcis-29191	72	22	accuracy	accuracy	NOUN
fcis-29191	72	23	of	of	ADP
fcis-29191	72	24	syntactic	syntactic	ADJ
fcis-29191	72	25	analysis	analysis	NOUN
fcis-29191	72	26	.	.	PUNCT
fcis-29191	73	1	the	the	DET
fcis-29191	73	2	deep	deep	ADJ
fcis-29191	73	3	double	double	ADJ
fcis-29191	73	4	affine	affine	NOUN
fcis-29191	73	5	neural	neural	ADJ
fcis-29191	73	6	dependency	dependency	NOUN
fcis-29191	73	7	syntactic	syntactic	ADJ
fcis-29191	73	8	analysis	analysis	NOUN
fcis-29191	73	9	structure	structure	NOUN
fcis-29191	73	10	based	base	VERB
fcis-29191	73	11	on	on	ADP
fcis-29191	73	12	bilstm	bilstm	NOUN
fcis-29191	73	13	is	be	AUX
fcis-29191	73	14	shown	show	VERB
fcis-29191	73	15	in	in	ADP
fcis-29191	73	16	figure	figure	NOUN
fcis-29191	73	17	2	2	NUM
fcis-29191	73	18	.	.	PUNCT
fcis-29191	74	1	the	the	DET
fcis-29191	74	2	input	input	NOUN
fcis-29191	74	3	is	be	AUX
fcis-29191	74	4	a	a	DET
fcis-29191	74	5	vector	vector	NOUN
fcis-29191	74	6	representation	representation	NOUN
fcis-29191	74	7	obtained	obtain	VERB
fcis-29191	74	8	by	by	ADP
fcis-29191	74	9	concatenating	concatenate	VERB
fcis-29191	74	10	words	word	NOUN
fcis-29191	74	11	and	and	CCONJ
fcis-29191	74	12	part	part	NOUN
fcis-29191	74	13	of	of	ADP
fcis-29191	74	14	speech	speech	NOUN
fcis-29191	74	15	vectors	vector	NOUN
fcis-29191	74	16	.	.	PUNCT
fcis-29191	75	1	the	the	DET
fcis-29191	75	2	feature	feature	NOUN
fcis-29191	75	3	ri	ri	PROPN
fcis-29191	75	4	is	be	AUX
fcis-29191	75	5	extracted	extract	VERB
fcis-29191	75	6	using	use	VERB
fcis-29191	75	7	bilstm	bilstm	NOUN
fcis-29191	75	8	,	,	PUNCT
fcis-29191	75	9	and	and	CCONJ
fcis-29191	75	10	h(arc	h(arc	PROPN
fcis-29191	75	11	-	-	PUNCT
fcis-29191	75	12	dep	dep	NOUN
fcis-29191	75	13	)	)	PUNCT
fcis-29191	75	14	and	and	CCONJ
fcis-29191	75	15	h(arc	h(arc	PROPN
fcis-29191	75	16	-	-	PUNCT
fcis-29191	75	17	head	head	NOUN
fcis-29191	75	18	)	)	PUNCT
fcis-29191	75	19	are	be	AUX
fcis-29191	75	20	obtained	obtain	VERB
fcis-29191	75	21	through	through	ADP
fcis-29191	75	22	two	two	NUM
fcis-29191	75	23	different	different	ADJ
fcis-29191	75	24	mlps	mlp	NOUN
fcis-29191	75	25	.	.	PUNCT
fcis-29191	76	1	the	the	DET
fcis-29191	76	2	hstack	hstack	NOUN
fcis-29191	76	3	of	of	ADP
fcis-29191	76	4	all	all	DET
fcis-29191	76	5	words	word	NOUN
fcis-29191	76	6	yields	yield	VERB
fcis-29191	76	7	h(arc	h(arc	PROPN
fcis-29191	76	8	-	-	PUNCT
fcis-29191	76	9	dep	dep	NOUN
fcis-29191	76	10	)	)	PUNCT
fcis-29191	76	11	and	and	CCONJ
fcis-29191	76	12	h(arc	h(arc	PROPN
fcis-29191	76	13	-	-	PUNCT
fcis-29191	76	14	head	head	NOUN
fcis-29191	76	15	)	)	PUNCT
fcis-29191	76	16	,	,	PUNCT
fcis-29191	76	17	and	and	CCONJ
fcis-29191	76	18	an	an	DET
fcis-29191	76	19	additional	additional	ADJ
fcis-29191	76	20	unit	unit	NOUN
fcis-29191	76	21	vector	vector	NOUN
fcis-29191	76	22	is	be	AUX
fcis-29191	76	23	concatenated	concatenate	VERB
fcis-29191	76	24	for	for	ADP
fcis-29191	76	25	h(arc	h(arc	NOUN
fcis-29191	76	26	-	-	PUNCT
fcis-29191	76	27	head	head	NOUN
fcis-29191	76	28	)	)	PUNCT
fcis-29191	76	29	.	.	PUNCT
fcis-29191	77	1	using	use	VERB
fcis-29191	77	2	the	the	DET
fcis-29191	77	3	intermediate	intermediate	ADJ
fcis-29191	77	4	matrix	matrix	NOUN
fcis-29191	77	5	u(arc	u(arc	NOUN
fcis-29191	77	6	)	)	PUNCT
fcis-29191	77	7	for	for	ADP
fcis-29191	77	8	affine	affine	ADJ
fcis-29191	77	9	transformation	transformation	NOUN
fcis-29191	77	10	,	,	PUNCT
fcis-29191	77	11	dot	dot	NOUN
fcis-29191	77	12	product	product	NOUN
fcis-29191	77	13	each	each	DET
fcis-29191	77	14	word	word	NOUN
fcis-29191	77	15	as	as	ADP
fcis-29191	77	16	dep	dep	NOUN
fcis-29191	77	17	with	with	ADP
fcis-29191	77	18	each	each	DET
fcis-29191	77	19	word	word	NOUN
fcis-29191	77	20	as	as	ADP
fcis-29191	77	21	head	head	NOUN
fcis-29191	77	22	to	to	PART
fcis-29191	77	23	obtain	obtain	VERB
fcis-29191	77	24	the	the	DET
fcis-29191	77	25	score	score	NOUN
fcis-29191	77	26	matrix	matrix	NOUN
fcis-29191	77	27	s(arc	s(arc	NOUN
fcis-29191	77	28	)	)	PUNCT
fcis-29191	77	29	where	where	SCONJ
fcis-29191	77	30	arc	arc	NOUN
fcis-29191	77	31	holds	hold	VERB
fcis-29191	77	32	.	.	PUNCT
fcis-29191	78	1	fig	fig	NOUN
fcis-29191	78	2	2	2	NUM
fcis-29191	78	3	.	.	PUNCT
fcis-29191	78	4	syntactic	syntactic	ADJ
fcis-29191	78	5	parsing	parsing	NOUN
fcis-29191	78	6	structure	structure	NOUN
fcis-29191	78	7	3.2	3.2	NUM
fcis-29191	78	8	.	.	PUNCT
fcis-29191	79	1	graph	graph	VERB
fcis-29191	79	2	convolutional	convolutional	ADJ
fcis-29191	79	3	network	network	NOUN
fcis-29191	79	4	model	model	NOUN
fcis-29191	79	5	for	for	ADP
fcis-29191	79	6	cyber	cyber	NOUN
fcis-29191	79	7	violence	violence	NOUN
fcis-29191	79	8	text	text	NOUN
fcis-29191	79	9	a	a	DET
fcis-29191	79	10	heterogeneous	heterogeneous	ADJ
fcis-29191	79	11	graph	graph	NOUN
fcis-29191	79	12	(	(	PUNCT
fcis-29191	79	13	,	,	PUNCT
fcis-29191	79	14	)	)	PUNCT
fcis-29191	79	15	g	g	NOUN
fcis-29191	79	16	v	v	PROPN
fcis-29191	79	17	e	e	ADJ
fcis-29191	79	18	was	be	AUX
fcis-29191	79	19	constructed	construct	VERB
fcis-29191	79	20	based	base	VERB
fcis-29191	79	21	on	on	ADP
fcis-29191	79	22	textgcn	textgcn	PROPN
fcis-29191	80	1	[	[	X
fcis-29191	80	2	12	12	NUM
fcis-29191	80	3	]	]	PUNCT
fcis-29191	80	4	,	,	PUNCT
fcis-29191	80	5	which	which	PRON
fcis-29191	80	6	includes	include	VERB
fcis-29191	80	7	word	word	NOUN
fcis-29191	80	8	nodes	node	NOUN
fcis-29191	80	9	and	and	CCONJ
fcis-29191	80	10	document	document	NOUN
fcis-29191	80	11	nodes	node	NOUN
fcis-29191	80	12	.	.	PUNCT
fcis-29191	81	1	the	the	DET
fcis-29191	81	2	graph	graph	NOUN
fcis-29191	81	3	can	can	AUX
fcis-29191	81	4	visually	visually	ADV
fcis-29191	81	5	represent	represent	VERB
fcis-29191	81	6	the	the	DET
fcis-29191	81	7	co	co	NOUN
fcis-29191	81	8	-	-	NOUN
fcis-29191	81	9	occurrence	occurrence	ADJ
fcis-29191	81	10	relationship	relationship	NOUN
fcis-29191	81	11	of	of	ADP
fcis-29191	81	12	global	global	ADJ
fcis-29191	81	13	words	word	NOUN
fcis-29191	81	14	and	and	CCONJ
fcis-29191	81	15	adapt	adapt	VERB
fcis-29191	81	16	to	to	PART
fcis-29191	81	17	graph	graph	VERB
fcis-29191	81	18	convolutional	convolutional	ADJ
fcis-29191	81	19	networks	network	NOUN
fcis-29191	81	20	.	.	PUNCT
fcis-29191	82	1	it	it	PRON
fcis-29191	82	2	can	can	AUX
fcis-29191	82	3	explicitly	explicitly	ADV
fcis-29191	82	4	model	model	VERB
fcis-29191	82	5	global	global	ADJ
fcis-29191	82	6	word	word	NOUN
fcis-29191	82	7	co	co	NOUN
fcis-29191	82	8	-	-	NOUN
fcis-29191	82	9	occurrence	occurrence	NOUN
fcis-29191	82	10	and	and	CCONJ
fcis-29191	82	11	adapt	adapt	VERB
fcis-29191	82	12	well	well	ADV
fcis-29191	82	13	to	to	PART
fcis-29191	82	14	graph	graph	VERB
fcis-29191	82	15	convolutional	convolutional	ADJ
fcis-29191	82	16	networks	network	NOUN
fcis-29191	82	17	.	.	PUNCT
fcis-29191	83	1	v	v	NOUN
fcis-29191	83	2	represents	represent	VERB
fcis-29191	83	3	the	the	DET
fcis-29191	83	4	set	set	NOUN
fcis-29191	83	5	of	of	ADP
fcis-29191	83	6	nodes	node	NOUN
fcis-29191	83	7	in	in	ADP
fcis-29191	83	8	the	the	DET
fcis-29191	83	9	graph	graph	NOUN
fcis-29191	83	10	,	,	PUNCT
fcis-29191	83	11	ande	ande	PROPN
fcis-29191	83	12	represents	represent	VERB
fcis-29191	83	13	the	the	DET
fcis-29191	83	14	set	set	NOUN
fcis-29191	83	15	of	of	ADP
fcis-29191	83	16	edges	edge	NOUN
fcis-29191	83	17	in	in	ADP
fcis-29191	83	18	the	the	DET
fcis-29191	83	19	graph	graph	NOUN
fcis-29191	83	20	.	.	PUNCT
fcis-29191	84	1	by	by	ADP
fcis-29191	84	2	analyzing	analyze	VERB
fcis-29191	84	3	the	the	DET
fcis-29191	84	4	occurrence	occurrence	NOUN
fcis-29191	84	5	of	of	ADP
fcis-29191	84	6	words	word	NOUN
fcis-29191	84	7	in	in	ADP
fcis-29191	84	8	a	a	DET
fcis-29191	84	9	document	document	NOUN
fcis-29191	84	10	,	,	PUNCT
fcis-29191	84	11	edges	edge	NOUN
fcis-29191	84	12	between	between	ADP
fcis-29191	84	13	document	document	NOUN
fcis-29191	84	14	nodes	node	NOUN
fcis-29191	84	15	are	be	AUX
fcis-29191	84	16	established	establish	VERB
fcis-29191	84	17	,	,	PUNCT
fcis-29191	84	18	and	and	CCONJ
fcis-29191	84	19	their	their	PRON
fcis-29191	84	20	weights	weight	NOUN
fcis-29191	84	21	are	be	AUX
fcis-29191	84	22	determined	determine	VERB
fcis-29191	84	23	by	by	ADP
fcis-29191	84	24	the	the	DET
fcis-29191	84	25	word	word	NOUN
fcis-29191	84	26	frequency	frequency	NOUN
fcis-29191	84	27	inverse	inverse	NOUN
fcis-29191	84	28	document	document	NOUN
fcis-29191	84	29	frequency	frequency	NOUN
fcis-29191	84	30	(	(	PUNCT
fcis-29191	84	31	tf	tf	PROPN
fcis-29191	84	32	-	-	PUNCT
fcis-29191	84	33	idf	idf	NOUN
fcis-29191	84	34	)	)	PUNCT
fcis-29191	84	35	of	of	ADP
fcis-29191	84	36	the	the	DET
fcis-29191	84	37	words	word	NOUN
fcis-29191	84	38	.	.	PUNCT
fcis-29191	85	1	word	word	NOUN
fcis-29191	85	2	frequency	frequency	NOUN
fcis-29191	85	3	represents	represent	VERB
fcis-29191	85	4	the	the	DET
fcis-29191	85	5	number	number	NOUN
fcis-29191	85	6	of	of	ADP
fcis-29191	85	7	times	time	NOUN
fcis-29191	85	8	a	a	DET
fcis-29191	85	9	word	word	NOUN
fcis-29191	85	10	appears	appear	VERB
fcis-29191	85	11	in	in	ADP
fcis-29191	85	12	a	a	DET
fcis-29191	85	13	document	document	NOUN
fcis-29191	85	14	,	,	PUNCT
fcis-29191	85	15	while	while	SCONJ
fcis-29191	85	16	inverse	inverse	ADJ
fcis-29191	85	17	document	document	NOUN
fcis-29191	85	18	frequency	frequency	NOUN
fcis-29191	85	19	is	be	AUX
fcis-29191	85	20	the	the	DET
fcis-29191	85	21	reciprocal	reciprocal	NOUN
fcis-29191	85	22	of	of	ADP
fcis-29191	85	23	the	the	DET
fcis-29191	85	24	number	number	NOUN
fcis-29191	85	25	of	of	ADP
fcis-29191	85	26	documents	document	NOUN
fcis-29191	85	27	containing	contain	VERB
fcis-29191	85	28	that	that	DET
fcis-29191	85	29	word	word	NOUN
fcis-29191	85	30	.	.	PUNCT
fcis-29191	86	1	the	the	PRON
fcis-29191	86	2	larger	large	ADJ
fcis-29191	86	3	the	the	DET
fcis-29191	86	4	tf	tf	PROPN
fcis-29191	86	5	-	-	PUNCT
fcis-29191	86	6	idf	idf	PROPN
fcis-29191	86	7	value	value	NOUN
fcis-29191	86	8	of	of	ADP
fcis-29191	86	9	a	a	DET
fcis-29191	86	10	word	word	NOUN
fcis-29191	86	11	,	,	PUNCT
fcis-29191	86	12	the	the	PRON
fcis-29191	86	13	higher	high	ADJ
fcis-29191	86	14	its	its	PRON
fcis-29191	86	15	importance	importance	NOUN
fcis-29191	86	16	in	in	ADP
fcis-29191	86	17	the	the	DET
fcis-29191	86	18	document	document	NOUN
fcis-29191	86	19	.	.	PUNCT
fcis-29191	87	1	based	base	VERB
fcis-29191	87	2	on	on	ADP
fcis-29191	87	3	the	the	DET
fcis-29191	87	4	co	co	NOUN
fcis-29191	87	5	-	-	NOUN
fcis-29191	87	6	occurrence	occurrence	NOUN
fcis-29191	87	7	of	of	ADP
fcis-29191	87	8	words	word	NOUN
fcis-29191	87	9	in	in	ADP
fcis-29191	87	10	the	the	DET
fcis-29191	87	11	entire	entire	ADJ
fcis-29191	87	12	corpus	corpus	NOUN
fcis-29191	87	13	,	,	PUNCT
fcis-29191	87	14	edges	edge	NOUN
fcis-29191	87	15	are	be	AUX
fcis-29191	87	16	constructed	construct	VERB
fcis-29191	87	17	between	between	ADP
fcis-29191	87	18	word	word	NOUN
fcis-29191	87	19	nodes	node	NOUN
fcis-29191	87	20	.	.	PUNCT
fcis-29191	88	1	the	the	DET
fcis-29191	88	2	weights	weight	NOUN
fcis-29191	88	3	between	between	ADP
fcis-29191	88	4	two	two	NUM
fcis-29191	88	5	-	-	PUNCT
fcis-29191	88	6	word	word	NOUN
fcis-29191	88	7	nodes	node	NOUN
fcis-29191	88	8	are	be	AUX
fcis-29191	88	9	calculated	calculate	VERB
fcis-29191	88	10	using	use	VERB
fcis-29191	88	11	point	point	NOUN
fcis-29191	88	12	mutual	mutual	ADJ
fcis-29191	88	13	information	information	NOUN
fcis-29191	88	14	(	(	PUNCT
fcis-29191	88	15	pmi	pmi	PROPN
fcis-29191	88	16	)	)	PUNCT
fcis-29191	88	17	,	,	PUNCT
fcis-29191	88	18	which	which	PRON
fcis-29191	88	19	measures	measure	VERB
fcis-29191	88	20	word	word	NOUN
fcis-29191	88	21	associations	association	NOUN
fcis-29191	88	22	.	.	PUNCT
fcis-29191	89	1	based	base	VERB
fcis-29191	89	2	on	on	ADP
fcis-29191	89	3	the	the	DET
fcis-29191	89	4	co	co	NOUN
fcis-29191	89	5	-	-	NOUN
fcis-29191	89	6	occurrence	occurrence	ADJ
fcis-29191	89	7	pattern	pattern	NOUN
fcis-29191	89	8	of	of	ADP
fcis-29191	89	9	words	word	NOUN
fcis-29191	89	10	in	in	ADP
fcis-29191	89	11	the	the	DET
fcis-29191	89	12	corpus	corpus	NOUN
fcis-29191	89	13	,	,	PUNCT
fcis-29191	89	14	edges	edge	NOUN
fcis-29191	89	15	are	be	AUX
fcis-29191	89	16	generated	generate	VERB
fcis-29191	89	17	between	between	ADP
fcis-29191	89	18	word	word	NOUN
fcis-29191	89	19	nodes	node	NOUN
fcis-29191	89	20	,	,	PUNCT
fcis-29191	89	21	and	and	CCONJ
fcis-29191	89	22	their	their	PRON
fcis-29191	89	23	weights	weight	NOUN
fcis-29191	89	24	reflect	reflect	VERB
fcis-29191	89	25	the	the	DET
fcis-29191	89	26	strength	strength	NOUN
fcis-29191	89	27	of	of	ADP
fcis-29191	89	28	word	word	NOUN
fcis-29191	89	29	associations	association	NOUN
fcis-29191	89	30	by	by	ADP
fcis-29191	89	31	calculating	calculate	VERB
fcis-29191	89	32	pmi	pmi	NOUN
fcis-29191	89	33	.	.	PUNCT
fcis-29191	90	1	the	the	DET
fcis-29191	90	2	calculation	calculation	NOUN
fcis-29191	90	3	formula	formula	NOUN
fcis-29191	90	4	is	be	AUX
fcis-29191	90	5	as	as	SCONJ
fcis-29191	90	6	follows	follow	VERB
fcis-29191	90	7	:	:	PUNCT
fcis-29191	90	8	(	(	PUNCT
fcis-29191	90	9	,	,	PUNCT
fcis-29191	90	10	)	)	PUNCT
fcis-29191	90	11	(	(	PUNCT
fcis-29191	90	12	,	,	PUNCT
fcis-29191	90	13	)	)	PUNCT
fcis-29191	90	14	log	log	NOUN
fcis-29191	90	15	(	(	PUNCT
fcis-29191	90	16	)	)	PUNCT
fcis-29191	90	17	(	(	PUNCT
fcis-29191	90	18	)	)	PUNCT
fcis-29191	91	1	p	p	X
fcis-29191	92	1	i	i	PRON
fcis-29191	92	2	j	j	PROPN
fcis-29191	92	3	pmi	pmi	PROPN
fcis-29191	93	1	i	i	PRON
fcis-29191	93	2	j	j	PROPN
fcis-29191	94	1	p	p	X
fcis-29191	94	2	i	i	PRON
fcis-29191	94	3	p	p	PROPN
fcis-29191	94	4	j	j	PROPN
fcis-29191	94	5			NOUN
fcis-29191	94	6	(	(	PUNCT
fcis-29191	94	7	1	1	NUM
fcis-29191	94	8	)	)	PUNCT
fcis-29191	94	9	when	when	SCONJ
fcis-29191	94	10	the	the	DET
fcis-29191	94	11	pmi	pmi	NOUN
fcis-29191	94	12	value	value	NOUN
fcis-29191	94	13	is	be	AUX
fcis-29191	94	14	positive	positive	ADJ
fcis-29191	94	15	,	,	PUNCT
fcis-29191	94	16	it	it	PRON
fcis-29191	94	17	indicates	indicate	VERB
fcis-29191	94	18	that	that	SCONJ
fcis-29191	94	19	the	the	DET
fcis-29191	94	20	word	word	NOUN
fcis-29191	94	21	has	have	VERB
fcis-29191	94	22	high	high	ADJ
fcis-29191	94	23	semantic	semantic	ADJ
fcis-29191	94	24	relevance	relevance	NOUN
fcis-29191	94	25	in	in	ADP
fcis-29191	94	26	the	the	DET
fcis-29191	94	27	corpus	corpus	NOUN
fcis-29191	94	28	,	,	PUNCT
fcis-29191	94	29	while	while	SCONJ
fcis-29191	94	30	a	a	DET
fcis-29191	94	31	negative	negative	ADJ
fcis-29191	94	32	pmi	pmi	NOUN
fcis-29191	94	33	value	value	NOUN
fcis-29191	94	34	indicates	indicate	VERB
fcis-29191	94	35	low	low	ADJ
fcis-29191	94	36	or	or	CCONJ
fcis-29191	94	37	non	non	ADJ
fcis-29191	94	38	-	-	ADJ
fcis-29191	94	39	existent	existent	ADJ
fcis-29191	94	40	semantic	semantic	ADJ
fcis-29191	94	41	relevance	relevance	NOUN
fcis-29191	94	42	.	.	PUNCT
fcis-29191	95	1	therefore	therefore	ADV
fcis-29191	95	2	,	,	PUNCT
fcis-29191	95	3	only	only	ADV
fcis-29191	95	4	add	add	VERB
fcis-29191	95	5	edges	edge	NOUN
fcis-29191	95	6	between	between	ADP
fcis-29191	95	7	two	two	NUM
fcis-29191	95	8	-	-	PUNCT
fcis-29191	95	9	word	word	NOUN
fcis-29191	95	10	nodes	node	NOUN
fcis-29191	95	11	when	when	SCONJ
fcis-29191	95	12	the	the	DET
fcis-29191	95	13	pmi	pmi	NOUN
fcis-29191	95	14	value	value	NOUN
fcis-29191	95	15	is	be	AUX
fcis-29191	95	16	greater	great	ADJ
fcis-29191	95	17	than	than	ADP
fcis-29191	95	18	0	0	NUM
fcis-29191	95	19	.	.	PUNCT
fcis-29191	96	1	define	define	VERB
fcis-29191	96	2	the	the	DET
fcis-29191	96	3	weight	weight	NOUN
fcis-29191	96	4	of	of	ADP
fcis-29191	96	5	the	the	DET
fcis-29191	96	6	edge	edge	NOUN
fcis-29191	96	7	between	between	ADP
fcis-29191	96	8	nodes	node	NOUN
fcis-29191	96	9	i	i	PRON
fcis-29191	96	10	and	and	CCONJ
fcis-29191	96	11	j	j	PROPN
fcis-29191	96	12	using	use	VERB
fcis-29191	96	13	the	the	DET
fcis-29191	96	14	following	follow	VERB
fcis-29191	96	15	formula	formula	NOUN
fcis-29191	96	16	:	:	PUNCT
fcis-29191	96	17	32	32	NUM
fcis-29191	96	18	(	(	PUNCT
fcis-29191	96	19	,	,	PUNCT
fcis-29191	96	20	)	)	PUNCT
fcis-29191	96	21	,	,	PUNCT
fcis-29191	96	22	,	,	PUNCT
fcis-29191	96	23	,	,	PUNCT
fcis-29191	96	24	,	,	PUNCT
fcis-29191	96	25	1	1	NUM
fcis-29191	96	26	,	,	PUNCT
fcis-29191	96	27	0	0	NUM
fcis-29191	97	1	others	other	NOUN
fcis-29191	97	2	ij	ij	INTJ
fcis-29191	97	3	pmi	pmi	PROPN
fcis-29191	98	1	i	i	PRON
fcis-29191	98	2	j	j	PROPN
fcis-29191	99	1	i	i	PRON
fcis-29191	99	2	j	j	PROPN
fcis-29191	99	3	is	be	AUX
fcis-29191	99	4	words	word	NOUN
fcis-29191	100	1	tf	tf	INTJ
fcis-29191	100	2	idf	idf	PROPN
fcis-29191	101	1	i	i	PRON
fcis-29191	101	2	is	be	AUX
fcis-29191	101	3	document	document	NOUN
fcis-29191	101	4	j	j	PROPN
fcis-29191	101	5	is	be	AUX
fcis-29191	101	6	word	word	NOUN
fcis-29191	101	7	a	a	X
fcis-29191	102	1	i	i	NOUN
fcis-29191	102	2	j	j	PROPN
fcis-29191	102	3	is	be	AUX
fcis-29191	102	4	words	word	NOUN
fcis-29191	102	5			ADP
fcis-29191	103	1			NUM
fcis-29191	103	2			PROPN
fcis-29191	103	3			NUM
fcis-29191	103	4			NOUN
fcis-29191	103	5			PROPN
fcis-29191	103	6	，	，	PROPN
fcis-29191	103	7	(	(	PUNCT
fcis-29191	103	8	2	2	X
fcis-29191	103	9	)	)	PUNCT
fcis-29191	103	10	textgcn	textgcn	PROPN
fcis-29191	103	11	uses	use	VERB
fcis-29191	103	12	the	the	DET
fcis-29191	103	13	identity	identity	NOUN
fcis-29191	103	14	matrix	matrix	NOUN
fcis-29191	103	15	ndoc	ndoc	ADV
fcis-29191	103	16	wordnx	wordnx	VERB
fcis-29191	103	17	i	i	PRON
fcis-29191	103	18			ADV
fcis-29191	103	19	as	as	SCONJ
fcis-29191	103	20	the	the	DET
fcis-29191	103	21	node	node	ADJ
fcis-29191	103	22	feature	feature	NOUN
fcis-29191	103	23	initialization	initialization	NOUN
fcis-29191	103	24	,	,	PUNCT
fcis-29191	103	25	where	where	SCONJ
fcis-29191	103	26	docn	docn	NOUN
fcis-29191	103	27	and	and	CCONJ
fcis-29191	103	28	wordn	wordn	NOUN
fcis-29191	103	29	represent	represent	VERB
fcis-29191	103	30	the	the	DET
fcis-29191	103	31	number	number	NOUN
fcis-29191	103	32	of	of	ADP
fcis-29191	103	33	document	document	NOUN
fcis-29191	103	34	nodes	node	NOUN
fcis-29191	103	35	and	and	CCONJ
fcis-29191	103	36	word	word	NOUN
fcis-29191	103	37	nodes	node	NOUN
fcis-29191	103	38	,	,	PUNCT
fcis-29191	103	39	respectively	respectively	ADV
fcis-29191	103	40	.	.	PUNCT
fcis-29191	104	1	in	in	ADP
fcis-29191	104	2	the	the	DET
fcis-29191	104	3	r	r	PROPN
fcis-29191	104	4	-	-	PUNCT
fcis-29191	104	5	bg	bg	PROPN
fcis-29191	104	6	model	model	NOUN
fcis-29191	104	7	,	,	PUNCT
fcis-29191	104	8	document	document	NOUN
fcis-29191	104	9	embeddings	embedding	NOUN
fcis-29191	104	10	are	be	AUX
fcis-29191	104	11	generated	generate	VERB
fcis-29191	104	12	using	use	VERB
fcis-29191	104	13	models	model	NOUN
fcis-29191	104	14	such	such	ADJ
fcis-29191	104	15	as	as	ADP
fcis-29191	104	16	bert	bert	PROPN
fcis-29191	104	17	or	or	CCONJ
fcis-29191	104	18	roberta	roberta	PROPN
fcis-29191	104	19	,	,	PUNCT
fcis-29191	104	20	and	and	CCONJ
fcis-29191	104	21	b	b	X
fcis-29191	104	22	uses	use	VERB
fcis-29191	104	23	them	they	PRON
fcis-29191	104	24	as	as	ADP
fcis-29191	104	25	input	input	NOUN
fcis-29191	104	26	representations	representation	NOUN
fcis-29191	104	27	for	for	ADP
fcis-29191	104	28	document	document	NOUN
fcis-29191	104	29	nodes	node	NOUN
fcis-29191	104	30	,	,	PUNCT
fcis-29191	104	31	denoted	denote	VERB
fcis-29191	104	32	as	as	ADP
fcis-29191	104	33	docn	docn	NOUN
fcis-29191	104	34	d	d	NOUN
fcis-29191	104	35	docx	docx	NOUN
fcis-29191	104	36			PUNCT
fcis-29191	104	37	,	,	PUNCT
fcis-29191	104	38	with	with	ADP
fcis-29191	104	39	an	an	DET
fcis-29191	104	40	embedding	embed	VERB
fcis-29191	104	41	dimension	dimension	NOUN
fcis-29191	104	42	of	of	ADP
fcis-29191	104	43	d.	d.	PROPN
fcis-29191	104	44	the	the	DET
fcis-29191	104	45	specific	specific	ADJ
fcis-29191	104	46	form	form	NOUN
fcis-29191	104	47	of	of	ADP
fcis-29191	104	48	the	the	DET
fcis-29191	104	49	initial	initial	ADJ
fcis-29191	104	50	node	node	ADJ
fcis-29191	104	51	feature	feature	NOUN
fcis-29191	104	52	matrix	matrix	NOUN
fcis-29191	104	53	is	be	AUX
fcis-29191	104	54	as	as	SCONJ
fcis-29191	104	55	follows	follow	VERB
fcis-29191	104	56	:	:	PUNCT
fcis-29191	105	1			PROPN
fcis-29191	105	2	doc	doc	PROPN
fcis-29191	106	1	d	d	ADP
fcis-29191	106	2	0	0	NUM
fcis-29191	106	3	word	word	NOUN
fcis-29191	106	4	doc	doc	PROPN
fcis-29191	106	5	n	n	CCONJ
fcis-29191	106	6	n	n	NOUN
fcis-29191	106	7	x	x	PUNCT
fcis-29191	106	8	x	x	X
fcis-29191	106	9			PROPN
fcis-29191	106	10			PROPN
fcis-29191	106	11			NOUN
fcis-29191	106	12			PROPN
fcis-29191	107	1			PROPN
fcis-29191	108	1			PROPN
fcis-29191	108	2			PROPN
fcis-29191	109	1			PROPN
fcis-29191	109	2			NOUN
fcis-29191	109	3	(	(	PUNCT
fcis-29191	109	4	3	3	NUM
fcis-29191	109	5	)	)	PUNCT
fcis-29191	109	6	the	the	DET
fcis-29191	109	7	gcn	gcn	NOUN
fcis-29191	109	8	model	model	NOUN
fcis-29191	109	9	receives	receive	VERB
fcis-29191	109	10	the	the	DET
fcis-29191	109	11	initial	initial	ADJ
fcis-29191	109	12	feature	feature	NOUN
fcis-29191	109	13	matrix	matrix	NOUN
fcis-29191	109	14	x	x	PUNCT
fcis-29191	109	15	as	as	ADP
fcis-29191	109	16	input	input	NOUN
fcis-29191	109	17	and	and	CCONJ
fcis-29191	109	18	achieves	achieve	VERB
fcis-29191	109	19	information	information	NOUN
fcis-29191	109	20	propagation	propagation	NOUN
fcis-29191	109	21	through	through	ADP
fcis-29191	109	22	multiple	multiple	ADJ
fcis-29191	109	23	iterations	iteration	NOUN
fcis-29191	109	24	during	during	ADP
fcis-29191	109	25	training	training	NOUN
fcis-29191	109	26	and	and	CCONJ
fcis-29191	109	27	testing	testing	NOUN
fcis-29191	109	28	.	.	PUNCT
fcis-29191	110	1	for	for	ADP
fcis-29191	110	2	the	the	DET
fcis-29191	110	3	i	i	PROPN
fcis-29191	110	4	-	-	PUNCT
fcis-29191	110	5	th	th	VERB
fcis-29191	110	6	gcn	gcn	NOUN
fcis-29191	110	7	layer	layer	NOUN
fcis-29191	110	8	(	(	PUNCT
fcis-29191	110	9	denoted	denote	VERB
fcis-29191	110	10	as	as	ADP
fcis-29191	110	11	(	(	PUNCT
fcis-29191	110	12	)	)	PUNCT
fcis-29191	110	13	il	il	PROPN
fcis-29191	110	14	)	)	PUNCT
fcis-29191	110	15	,	,	PUNCT
fcis-29191	110	16	the	the	DET
fcis-29191	110	17	calculation	calculation	NOUN
fcis-29191	110	18	formula	formula	NOUN
fcis-29191	110	19	for	for	ADP
fcis-29191	110	20	its	its	PRON
fcis-29191	110	21	output	output	NOUN
fcis-29191	110	22	feature	feature	NOUN
fcis-29191	110	23	matrix	matrix	NOUN
fcis-29191	110	24	is	be	AUX
fcis-29191	110	25	:	:	PUNCT
fcis-29191	110	26			PROPN
fcis-29191	110	27			PROPN
fcis-29191	110	28	i	i	X
fcis-29191	110	29	(	(	PUNCT
fcis-29191	110	30	1	1	NUM
fcis-29191	110	31	)	)	PUNCT
fcis-29191	110	32	(	(	PUNCT
fcis-29191	110	33	)	)	PUNCT
fcis-29191	110	34	(	(	PUNCT
fcis-29191	110	35	)	)	PUNCT
fcis-29191	110	36	i	i	PRON
fcis-29191	110	37	il	il	PROPN
fcis-29191	110	38	al	al	PROPN
fcis-29191	110	39	w	w	PROPN
fcis-29191	110	40			PROPN
fcis-29191	110	41	(	(	PUNCT
fcis-29191	110	42	4	4	NUM
fcis-29191	110	43	)	)	PUNCT
fcis-29191	110	44	in	in	ADP
fcis-29191	110	45	the	the	DET
fcis-29191	110	46	formula	formula	NOUN
fcis-29191	110	47	,	,	PUNCT
fcis-29191	110	48			NOUN
fcis-29191	110	49	represents	represent	VERB
fcis-29191	110	50	the	the	DET
fcis-29191	110	51	activation	activation	NOUN
fcis-29191	110	52	function	function	NOUN
fcis-29191	110	53	,	,	PUNCT
fcis-29191	110	54			NOUN
fcis-29191	110	55	1	1	NUM
fcis-29191	110	56	1	1	NUM
fcis-29191	110	57	2	2	NUM
fcis-29191	110	58	2a	2a	NUM
fcis-29191	111	1	d	d	PRON
fcis-29191	111	2	ad	ad	NOUN
fcis-29191	111	3			PROPN
fcis-29191	111	4			PROPN
fcis-29191	111	5			NUM
fcis-29191	111	6	represents	represent	VERB
fcis-29191	111	7	the	the	DET
fcis-29191	111	8	normalized	normalize	VERB
fcis-29191	111	9	adjacency	adjacency	NOUN
fcis-29191	111	10	matrix	matrix	NOUN
fcis-29191	111	11	,	,	PUNCT
fcis-29191	111	12	and	and	CCONJ
fcis-29191	111	13	1	1	NUM
fcis-29191	111	14	(	(	PUNCT
fcis-29191	111	15	)	)	PUNCT
fcis-29191	111	16	i	i	PRON
fcis-29191	112	1	i	i	PROPN
fcis-29191	112	2	d	d	PROPN
fcis-29191	112	3	diw	diw	PROPN
fcis-29191	112	4			PROPN
fcis-29191	112	5			PUNCT
fcis-29191	112	6	is	be	AUX
fcis-29191	112	7	the	the	DET
fcis-29191	112	8	weight	weight	NOUN
fcis-29191	112	9	matrix	matrix	NOUN
fcis-29191	112	10	of	of	ADP
fcis-29191	112	11	that	that	DET
fcis-29191	112	12	layer	layer	NOUN
fcis-29191	112	13	,	,	PUNCT
fcis-29191	112	14	(	(	PUNCT
fcis-29191	112	15	0)l	0)l	NOUN
fcis-29191	112	16	x	x	PROPN
fcis-29191	112	17	is	be	AUX
fcis-29191	112	18	the	the	DET
fcis-29191	112	19	input	input	NOUN
fcis-29191	112	20	feature	feature	NOUN
fcis-29191	112	21	matrix	matrix	NOUN
fcis-29191	112	22	.	.	PUNCT
fcis-29191	113	1	generate	generate	VERB
fcis-29191	113	2	the	the	DET
fcis-29191	113	3	final	final	ADJ
fcis-29191	113	4	feature	feature	NOUN
fcis-29191	113	5	representation	representation	NOUN
fcis-29191	113	6	of	of	ADP
fcis-29191	113	7	the	the	DET
fcis-29191	113	8	document	document	NOUN
fcis-29191	113	9	through	through	ADP
fcis-29191	113	10	gcn	gcn	NOUN
fcis-29191	113	11	and	and	CCONJ
fcis-29191	113	12	input	input	VERB
fcis-29191	113	13	it	it	PRON
fcis-29191	113	14	into	into	ADP
fcis-29191	113	15	the	the	DET
fcis-29191	113	16	softmax	softmax	NOUN
fcis-29191	113	17	layer	layer	NOUN
fcis-29191	113	18	for	for	ADP
fcis-29191	113	19	category	category	NOUN
fcis-29191	113	20	prediction	prediction	NOUN
fcis-29191	113	21	:	:	PUNCT
fcis-29191	113	22	softmax	softmax	NOUN
fcis-29191	113	23	(	(	PUNCT
fcis-29191	113	24	(	(	PUNCT
fcis-29191	113	25	,	,	PUNCT
fcis-29191	113	26	)	)	PUNCT
fcis-29191	113	27	)	)	PUNCT
fcis-29191	114	1	gcnz	gcnz	ADV
fcis-29191	114	2	g	g	NOUN
fcis-29191	114	3	x	x	PUNCT
fcis-29191	114	4	a	a	NOUN
fcis-29191	114	5	(	(	PUNCT
fcis-29191	114	6	5	5	X
fcis-29191	114	7	)	)	PUNCT
fcis-29191	114	8	g	g	NOUN
fcis-29191	114	9	represents	represent	VERB
fcis-29191	114	10	the	the	DET
fcis-29191	114	11	gcn	gcn	ADJ
fcis-29191	114	12	model	model	NOUN
fcis-29191	114	13	.	.	PUNCT
fcis-29191	115	1	during	during	ADP
fcis-29191	115	2	the	the	DET
fcis-29191	115	3	training	training	NOUN
fcis-29191	115	4	process	process	NOUN
fcis-29191	115	5	,	,	PUNCT
fcis-29191	115	6	the	the	DET
fcis-29191	115	7	cross	cross	NOUN
fcis-29191	115	8	entropy	entropy	PROPN
fcis-29191	115	9	loss	loss	NOUN
fcis-29191	115	10	function	function	NOUN
fcis-29191	115	11	on	on	ADP
fcis-29191	115	12	the	the	DET
fcis-29191	115	13	labeled	label	VERB
fcis-29191	115	14	document	document	NOUN
fcis-29191	115	15	nodes	node	NOUN
fcis-29191	115	16	is	be	AUX
fcis-29191	115	17	used	use	VERB
fcis-29191	115	18	to	to	PART
fcis-29191	115	19	jointly	jointly	ADV
fcis-29191	115	20	optimize	optimize	VERB
fcis-29191	115	21	the	the	DET
fcis-29191	115	22	parameters	parameter	NOUN
fcis-29191	115	23	of	of	ADP
fcis-29191	115	24	bert	bert	PROPN
fcis-29191	115	25	and	and	CCONJ
fcis-29191	115	26	gcn	gcn	NOUN
fcis-29191	115	27	.	.	PUNCT
fcis-29191	116	1	the	the	DET
fcis-29191	116	2	calculation	calculation	NOUN
fcis-29191	116	3	formula	formula	NOUN
fcis-29191	116	4	is	be	AUX
fcis-29191	116	5	as	as	SCONJ
fcis-29191	116	6	follows	follow	VERB
fcis-29191	116	7	:	:	PUNCT
fcis-29191	116	8	1	1	NUM
fcis-29191	116	9	ln	ln	NOUN
fcis-29191	116	10	d	d	X
fcis-29191	116	11	f	f	X
fcis-29191	116	12	df	df	PROPN
fcis-29191	116	13	df	df	PROPN
fcis-29191	117	1	d	d	PROPN
fcis-29191	117	2	y	y	PROPN
fcis-29191	117	3	f	f	PROPN
fcis-29191	117	4	y	y	PROPN
fcis-29191	117	5	z	z	PROPN
fcis-29191	117	6			PROPN
fcis-29191	118	1			PROPN
fcis-29191	118	2			NUM
fcis-29191	119	1			X
fcis-29191	119	2			X
fcis-29191	119	3	(	(	PUNCT
fcis-29191	119	4	6	6	NUM
fcis-29191	119	5	)	)	PUNCT
fcis-29191	119	6	among	among	ADP
fcis-29191	119	7	them	they	PRON
fcis-29191	119	8	,	,	PUNCT
fcis-29191	119	9	yd	yd	PROPN
fcis-29191	119	10	represents	represent	VERB
fcis-29191	119	11	a	a	DET
fcis-29191	119	12	set	set	NOUN
fcis-29191	119	13	of	of	ADP
fcis-29191	119	14	labeled	label	VERB
fcis-29191	119	15	document	document	NOUN
fcis-29191	119	16	indexes	index	NOUN
fcis-29191	119	17	,	,	PUNCT
fcis-29191	119	18	f	f	PROPN
fcis-29191	119	19	is	be	AUX
fcis-29191	119	20	the	the	DET
fcis-29191	119	21	dimension	dimension	NOUN
fcis-29191	119	22	of	of	ADP
fcis-29191	119	23	output	output	NOUN
fcis-29191	119	24	features	feature	NOUN
fcis-29191	119	25	,	,	PUNCT
fcis-29191	119	26	and	and	CCONJ
fcis-29191	119	27	y	y	PROPN
fcis-29191	119	28	is	be	AUX
fcis-29191	119	29	the	the	DET
fcis-29191	119	30	label	label	NOUN
fcis-29191	119	31	index	index	NOUN
fcis-29191	119	32	matrix	matrix	NOUN
fcis-29191	119	33	used	use	VERB
fcis-29191	119	34	to	to	PART
fcis-29191	119	35	represent	represent	VERB
fcis-29191	119	36	the	the	DET
fcis-29191	119	37	structural	structural	ADJ
fcis-29191	119	38	format	format	NOUN
fcis-29191	119	39	of	of	ADP
fcis-29191	119	40	multi	multi	ADJ
fcis-29191	119	41	label	label	NOUN
fcis-29191	119	42	data	datum	NOUN
fcis-29191	119	43	.	.	PUNCT
fcis-29191	120	1	through	through	ADP
fcis-29191	120	2	research	research	NOUN
fcis-29191	120	3	,	,	PUNCT
fcis-29191	120	4	it	it	PRON
fcis-29191	120	5	has	have	AUX
fcis-29191	120	6	been	be	AUX
fcis-29191	120	7	found	find	VERB
fcis-29191	120	8	that	that	SCONJ
fcis-29191	120	9	interpolation	interpolation	NOUN
fcis-29191	120	10	prediction	prediction	NOUN
fcis-29191	120	11	can	can	AUX
fcis-29191	120	12	lead	lead	VERB
fcis-29191	120	13	to	to	ADP
fcis-29191	120	14	faster	fast	ADJ
fcis-29191	120	15	convergence	convergence	NOUN
fcis-29191	120	16	and	and	CCONJ
fcis-29191	120	17	better	well	ADJ
fcis-29191	120	18	performance	performance	NOUN
fcis-29191	120	19	of	of	ADP
fcis-29191	120	20	the	the	DET
fcis-29191	120	21	model	model	NOUN
fcis-29191	120	22	.	.	PUNCT
fcis-29191	121	1	therefore	therefore	ADV
fcis-29191	121	2	,	,	PUNCT
fcis-29191	121	3	embedding	embed	VERB
fcis-29191	121	4	the	the	DET
fcis-29191	121	5	document	document	NOUN
fcis-29191	121	6	directly	directly	ADV
fcis-29191	121	7	into	into	ADP
fcis-29191	121	8	a	a	DET
fcis-29191	121	9	fully	fully	ADV
fcis-29191	121	10	connected	connect	VERB
fcis-29191	121	11	layer	layer	NOUN
fcis-29191	121	12	containing	contain	VERB
fcis-29191	121	13	softmax	softmax	NOUN
fcis-29191	121	14	is	be	AUX
fcis-29191	121	15	used	use	VERB
fcis-29191	121	16	to	to	PART
fcis-29191	121	17	form	form	VERB
fcis-29191	121	18	an	an	DET
fcis-29191	121	19	auxiliary	auxiliary	ADJ
fcis-29191	121	20	classifier	classifier	NOUN
fcis-29191	121	21	:	:	PUNCT
fcis-29191	121	22	softmax	softmax	PROPN
fcis-29191	121	23	(	(	PUNCT
fcis-29191	121	24	)	)	PUNCT
fcis-29191	121	25	bertz	bertz	NOUN
fcis-29191	121	26	wx	wx	PROPN
fcis-29191	121	27	(	(	PUNCT
fcis-29191	121	28	7	7	X
fcis-29191	121	29	)	)	PUNCT
fcis-29191	121	30	the	the	DET
fcis-29191	121	31	ultimate	ultimate	ADJ
fcis-29191	121	32	training	training	NOUN
fcis-29191	121	33	objective	objective	NOUN
fcis-29191	121	34	is	be	AUX
fcis-29191	121	35	the	the	DET
fcis-29191	121	36	linear	linear	ADJ
fcis-29191	121	37	interpolation	interpolation	NOUN
fcis-29191	121	38	of	of	ADP
fcis-29191	121	39	predictions	prediction	NOUN
fcis-29191	121	40	from	from	ADP
fcis-29191	121	41	gcn	gcn	NOUN
fcis-29191	121	42	and	and	CCONJ
fcis-29191	121	43	bert	bert	PROPN
fcis-29191	121	44	,	,	PUNCT
fcis-29191	121	45	calculated	calculate	VERB
fcis-29191	121	46	using	use	VERB
fcis-29191	121	47	the	the	DET
fcis-29191	121	48	following	follow	VERB
fcis-29191	121	49	formula	formula	NOUN
fcis-29191	121	50	:	:	PUNCT
fcis-29191	121	51	(	(	PUNCT
fcis-29191	121	52	1	1	X
fcis-29191	121	53	)	)	PUNCT
fcis-29191	121	54	gcn	gcn	NOUN
fcis-29191	121	55	bertz	bertz	NOUN
fcis-29191	122	1	z	z	PROPN
fcis-29191	122	2	z	z	PROPN
fcis-29191	122	3			PROPN
fcis-29191	122	4			PUNCT
fcis-29191	122	5			X
fcis-29191	122	6	(	(	PUNCT
fcis-29191	122	7	8)	8)	NUM
fcis-29191	122	8	among	among	ADP
fcis-29191	122	9	them	they	PRON
fcis-29191	122	10	,	,	PUNCT
fcis-29191	122	11	λ	λ	PROPN
fcis-29191	122	12	controls	control	VERB
fcis-29191	122	13	the	the	DET
fcis-29191	122	14	weight	weight	NOUN
fcis-29191	122	15	of	of	ADP
fcis-29191	122	16	two	two	NUM
fcis-29191	122	17	objectives	objective	NOUN
fcis-29191	122	18	.	.	PUNCT
fcis-29191	123	1	λ=0	λ=0	PROPN
fcis-29191	123	2	indicates	indicate	VERB
fcis-29191	123	3	the	the	DET
fcis-29191	123	4	use	use	NOUN
fcis-29191	123	5	of	of	ADP
fcis-29191	123	6	only	only	ADV
fcis-29191	123	7	bert	bert	NOUN
fcis-29191	123	8	related	relate	VERB
fcis-29191	123	9	models	model	NOUN
fcis-29191	123	10	that	that	PRON
fcis-29191	123	11	introduce	introduce	VERB
fcis-29191	123	12	syntactic	syntactic	ADJ
fcis-29191	123	13	relationships	relationship	NOUN
fcis-29191	123	14	,	,	PUNCT
fcis-29191	123	15	while	while	SCONJ
fcis-29191	123	16	λ=1	λ=1	PRON
fcis-29191	123	17	indicates	indicate	VERB
fcis-29191	123	18	the	the	DET
fcis-29191	123	19	use	use	NOUN
fcis-29191	123	20	of	of	ADP
fcis-29191	123	21	only	only	ADJ
fcis-29191	123	22	gcn	gcn	NOUN
fcis-29191	123	23	models	model	NOUN
fcis-29191	123	24	.	.	PUNCT
fcis-29191	124	1	when	when	SCONJ
fcis-29191	124	2	λ∈(0,1	λ∈(0,1	NOUN
fcis-29191	124	3	)	)	PUNCT
fcis-29191	124	4	,	,	PUNCT
fcis-29191	124	5	the	the	DET
fcis-29191	124	6	predictions	prediction	NOUN
fcis-29191	124	7	of	of	ADP
fcis-29191	124	8	both	both	DET
fcis-29191	124	9	models	model	NOUN
fcis-29191	124	10	can	can	AUX
fcis-29191	124	11	be	be	AUX
fcis-29191	124	12	balanced	balance	VERB
fcis-29191	124	13	,	,	PUNCT
fcis-29191	124	14	and	and	CCONJ
fcis-29191	124	15	the	the	DET
fcis-29191	124	16	overall	overall	ADJ
fcis-29191	124	17	model	model	NOUN
fcis-29191	124	18	can	can	AUX
fcis-29191	124	19	be	be	AUX
fcis-29191	124	20	better	well	ADV
fcis-29191	124	21	optimized	optimize	VERB
fcis-29191	124	22	.	.	PUNCT
fcis-29191	125	1	directly	directly	ADV
fcis-29191	125	2	manipulate	manipulate	VERB
fcis-29191	125	3	the	the	DET
fcis-29191	125	4	input	input	NOUN
fcis-29191	125	5	of	of	ADP
fcis-29191	125	6	gcn	gcn	NOUN
fcis-29191	125	7	through	through	ADP
fcis-29191	125	8	zbert	zbert	NOUN
fcis-29191	125	9	to	to	PART
fcis-29191	125	10	ensure	ensure	VERB
fcis-29191	125	11	that	that	SCONJ
fcis-29191	125	12	the	the	DET
fcis-29191	125	13	optimization	optimization	NOUN
fcis-29191	125	14	direction	direction	NOUN
fcis-29191	125	15	of	of	ADP
fcis-29191	125	16	the	the	DET
fcis-29191	125	17	input	input	NOUN
fcis-29191	125	18	is	be	AUX
fcis-29191	125	19	consistent	consistent	ADJ
fcis-29191	125	20	with	with	ADP
fcis-29191	125	21	the	the	DET
fcis-29191	125	22	goal	goal	NOUN
fcis-29191	125	23	.	.	PUNCT
fcis-29191	126	1	this	this	DET
fcis-29191	126	2	method	method	NOUN
fcis-29191	126	3	effectively	effectively	ADV
fcis-29191	126	4	alleviates	alleviate	VERB
fcis-29191	126	5	the	the	DET
fcis-29191	126	6	common	common	ADJ
fcis-29191	126	7	problems	problem	NOUN
fcis-29191	126	8	of	of	ADP
fcis-29191	126	9	gradient	gradient	NOUN
fcis-29191	126	10	vanishing	vanishing	NOUN
fcis-29191	126	11	and	and	CCONJ
fcis-29191	126	12	excessive	excessive	ADJ
fcis-29191	126	13	smoothing	smoothing	NOUN
fcis-29191	126	14	in	in	ADP
fcis-29191	126	15	multi	multi	ADJ
fcis-29191	126	16	-	-	ADJ
fcis-29191	126	17	layer	layer	ADJ
fcis-29191	126	18	gcn	gcn	NOUN
fcis-29191	126	19	models	model	NOUN
fcis-29191	126	20	,	,	PUNCT
fcis-29191	126	21	thereby	thereby	ADV
fcis-29191	126	22	significantly	significantly	ADV
fcis-29191	126	23	enhancing	enhance	VERB
fcis-29191	126	24	the	the	DET
fcis-29191	126	25	performance	performance	NOUN
fcis-29191	126	26	of	of	ADP
fcis-29191	126	27	the	the	DET
fcis-29191	126	28	model	model	NOUN
fcis-29191	126	29	.	.	PUNCT
fcis-29191	127	1	4	4	X
fcis-29191	127	2	.	.	X
fcis-29191	127	3	experiment	experiment	NOUN
fcis-29191	127	4	4.1	4.1	NUM
fcis-29191	127	5	.	.	PUNCT
fcis-29191	128	1	dataset	dataset	VERB
fcis-29191	128	2	the	the	DET
fcis-29191	128	3	dataset	dataset	NOUN
fcis-29191	128	4	used	use	VERB
fcis-29191	128	5	in	in	ADP
fcis-29191	128	6	the	the	DET
fcis-29191	128	7	experiment	experiment	NOUN
fcis-29191	128	8	is	be	AUX
fcis-29191	128	9	the	the	DET
fcis-29191	128	10	cyberbullying	cyberbullye	VERB
fcis-29191	128	11	tweets	tweet	NOUN
fcis-29191	128	12	dataset	dataset	NOUN
fcis-29191	128	13	,	,	PUNCT
fcis-29191	128	14	which	which	PRON
fcis-29191	128	15	was	be	AUX
fcis-29191	128	16	constructed	construct	VERB
fcis-29191	128	17	by	by	ADP
fcis-29191	128	18	wang	wang	PROPN
fcis-29191	129	1	[	[	X
fcis-29191	129	2	23	23	NUM
fcis-29191	129	3	]	]	PUNCT
fcis-29191	129	4	et	et	PROPN
fcis-29191	129	5	al	al	PROPN
fcis-29191	129	6	.	.	PROPN
fcis-29191	130	1	in	in	ADP
fcis-29191	130	2	their	their	PRON
fcis-29191	130	3	study	study	NOUN
fcis-29191	130	4	of	of	ADP
fcis-29191	130	5	a	a	DET
fcis-29191	130	6	graph	graph	NOUN
fcis-29191	130	7	convolutional	convolutional	ADJ
fcis-29191	130	8	network	network	NOUN
fcis-29191	130	9	method	method	NOUN
fcis-29191	130	10	for	for	ADP
fcis-29191	130	11	finegrained	finegraine	VERB
fcis-29191	130	12	cyberbullying	cyberbullye	VERB
fcis-29191	130	13	detection	detection	NOUN
fcis-29191	130	14	.	.	PUNCT
fcis-29191	131	1	wang	wang	PROPN
fcis-29191	131	2	et	et	PROPN
fcis-29191	131	3	al	al	PROPN
fcis-29191	131	4	.	.	PROPN
fcis-29191	131	5	identified	identify	VERB
fcis-29191	131	6	and	and	CCONJ
fcis-29191	131	7	collected	collect	VERB
fcis-29191	131	8	16	16	NUM
fcis-29191	131	9	open	open	ADJ
fcis-29191	131	10	-	-	PUNCT
fcis-29191	131	11	source	source	NOUN
fcis-29191	131	12	datasets	dataset	NOUN
fcis-29191	131	13	on	on	ADP
fcis-29191	131	14	cybullying	cybullye	VERB
fcis-29191	131	15	through	through	ADP
fcis-29191	131	16	literature	literature	NOUN
fcis-29191	131	17	review	review	PROPN
fcis-29191	131	18	,	,	PUNCT
fcis-29191	131	19	and	and	CCONJ
fcis-29191	131	20	ultimately	ultimately	ADV
fcis-29191	131	21	selected	select	VERB
fcis-29191	131	22	six	six	NUM
fcis-29191	131	23	twitter	twitter	NOUN
fcis-29191	131	24	datasets	dataset	NOUN
fcis-29191	131	25	for	for	ADP
fcis-29191	131	26	further	further	ADJ
fcis-29191	131	27	analysis	analysis	NOUN
fcis-29191	131	28	.	.	PUNCT
fcis-29191	132	1	after	after	ADP
fcis-29191	132	2	using	use	VERB
fcis-29191	132	3	the	the	DET
fcis-29191	132	4	twitter	twitter	NOUN
fcis-29191	132	5	api	api	NOUN
fcis-29191	132	6	to	to	PART
fcis-29191	132	7	retrieve	retrieve	VERB
fcis-29191	132	8	tweet	tweet	NOUN
fcis-29191	132	9	content	content	NOUN
fcis-29191	132	10	,	,	PUNCT
fcis-29191	132	11	they	they	PRON
fcis-29191	132	12	manually	manually	ADV
fcis-29191	132	13	classified	classify	VERB
fcis-29191	132	14	tweets	tweet	NOUN
fcis-29191	132	15	and	and	CCONJ
fcis-29191	132	16	expanded	expand	VERB
fcis-29191	132	17	the	the	DET
fcis-29191	132	18	sample	sample	NOUN
fcis-29191	132	19	in	in	ADP
fcis-29191	132	20	a	a	DET
fcis-29191	132	21	semi	semi	ADV
fcis-29191	132	22	supervised	supervised	ADJ
fcis-29191	132	23	manner	manner	NOUN
fcis-29191	132	24	through	through	ADP
fcis-29191	132	25	dynamic	dynamic	ADJ
fcis-29191	132	26	query	query	NOUN
fcis-29191	132	27	expansion	expansion	NOUN
fcis-29191	132	28	methods	method	NOUN
fcis-29191	132	29	.	.	PUNCT
fcis-29191	133	1	finally	finally	ADV
fcis-29191	133	2	,	,	PUNCT
fcis-29191	133	3	they	they	PRON
fcis-29191	133	4	randomly	randomly	ADV
fcis-29191	133	5	selected	select	VERB
fcis-29191	133	6	not	not	PART
fcis-29191	133	7	-	-	PUNCT
fcis-29191	133	8	cyberbullying	cyberbullye	VERB
fcis-29191	133	9	texts	text	NOUN
fcis-29191	133	10	and	and	CCONJ
fcis-29191	133	11	cyberbullying	cyberbullye	VERB
fcis-29191	133	12	tweets	tweet	NOUN
fcis-29191	133	13	targeting	target	VERB
fcis-29191	133	14	age	age	NOUN
fcis-29191	133	15	,	,	PUNCT
fcis-29191	133	16	race	race	NOUN
fcis-29191	133	17	,	,	PUNCT
fcis-29191	133	18	gender	gender	NOUN
fcis-29191	133	19	,	,	PUNCT
fcis-29191	133	20	religion	religion	NOUN
fcis-29191	133	21	,	,	PUNCT
fcis-29191	133	22	and	and	CCONJ
fcis-29191	133	23	other	other	ADJ
fcis-29191	133	24	types	type	NOUN
fcis-29191	133	25	,	,	PUNCT
fcis-29191	133	26	forming	form	VERB
fcis-29191	133	27	a	a	DET
fcis-29191	133	28	balanced	balanced	ADJ
fcis-29191	133	29	dataset	dataset	NOUN
fcis-29191	133	30	of	of	ADP
fcis-29191	133	31	approximately	approximately	ADV
fcis-29191	133	32	48000	48000	NUM
fcis-29191	133	33	texts	text	NOUN
fcis-29191	133	34	.	.	PUNCT
fcis-29191	134	1	the	the	DET
fcis-29191	134	2	text	text	NOUN
fcis-29191	134	3	on	on	ADP
fcis-29191	134	4	age	age	NOUN
fcis-29191	134	5	discrimination	discrimination	NOUN
fcis-29191	134	6	and	and	CCONJ
fcis-29191	134	7	racism	racism	NOUN
fcis-29191	134	8	in	in	ADP
fcis-29191	134	9	the	the	DET
fcis-29191	134	10	dataset	dataset	NOUN
fcis-29191	134	11	may	may	AUX
fcis-29191	134	12	differ	differ	VERB
fcis-29191	134	13	from	from	ADP
fcis-29191	134	14	what	what	PRON
fcis-29191	134	15	is	be	AUX
fcis-29191	134	16	expressed	express	VERB
fcis-29191	134	17	in	in	ADP
fcis-29191	134	18	existing	exist	VERB
fcis-29191	134	19	cyberbullying	cyberbullye	VERB
fcis-29191	134	20	detection	detection	NOUN
fcis-29191	134	21	.	.	PUNCT
fcis-29191	135	1	the	the	DET
fcis-29191	135	2	data	data	NOUN
fcis-29191	135	3	labels	label	NOUN
fcis-29191	135	4	are	be	AUX
fcis-29191	135	5	divided	divide	VERB
fcis-29191	135	6	into	into	ADP
fcis-29191	135	7	six	six	NUM
fcis-29191	135	8	types	type	NOUN
fcis-29191	135	9	:	:	PUNCT
fcis-29191	135	10	not_cyberbullying	not_cyberbullying	NOUN
fcis-29191	135	11	,	,	PUNCT
fcis-29191	135	12	age	age	NOUN
fcis-29191	135	13	,	,	PUNCT
fcis-29191	135	14	ethnicity	ethnicity	NOUN
fcis-29191	135	15	,	,	PUNCT
fcis-29191	135	16	gender	gender	NOUN
fcis-29191	135	17	,	,	PUNCT
fcis-29191	135	18	religion	religion	NOUN
fcis-29191	135	19	,	,	PUNCT
fcis-29191	135	20	and	and	CCONJ
fcis-29191	135	21	other	other	ADJ
fcis-29191	135	22	_	_	PRON
fcis-29191	135	23	cyberbullying	cyberbullye	VERB
fcis-29191	135	24	.	.	PUNCT
fcis-29191	136	1	before	before	ADP
fcis-29191	136	2	validating	validate	VERB
fcis-29191	136	3	the	the	DET
fcis-29191	136	4	model	model	NOUN
fcis-29191	136	5	proposed	propose	VERB
fcis-29191	136	6	in	in	ADP
fcis-29191	136	7	this	this	DET
fcis-29191	136	8	article	article	NOUN
fcis-29191	136	9	on	on	ADP
fcis-29191	136	10	this	this	DET
fcis-29191	136	11	dataset	dataset	NOUN
fcis-29191	136	12	,	,	PUNCT
fcis-29191	136	13	a	a	DET
fcis-29191	136	14	simple	simple	ADJ
fcis-29191	136	15	preprocessing	preprocessing	NOUN
fcis-29191	136	16	was	be	AUX
fcis-29191	136	17	performed	perform	VERB
fcis-29191	136	18	to	to	PART
fcis-29191	136	19	remove	remove	VERB
fcis-29191	136	20	links	link	NOUN
fcis-29191	136	21	,	,	PUNCT
fcis-29191	136	22	mentions	mention	NOUN
fcis-29191	136	23	(	(	PUNCT
fcis-29191	136	24	@username	@username	PROPN
fcis-29191	136	25	)	)	PUNCT
fcis-29191	136	26	,	,	PUNCT
fcis-29191	136	27	forwarding	forward	VERB
fcis-29191	136	28	flags	flag	NOUN
fcis-29191	136	29	"	"	PUNCT
fcis-29191	136	30	rt	rt	NOUN
fcis-29191	136	31	"	"	PUNCT
fcis-29191	136	32	,	,	PUNCT
fcis-29191	136	33	and	and	CCONJ
fcis-29191	136	34	punctuation	punctuation	NOUN
fcis-29191	136	35	marks	mark	NOUN
fcis-29191	136	36	from	from	ADP
fcis-29191	136	37	tweets	tweet	NOUN
fcis-29191	136	38	.	.	PUNCT
fcis-29191	137	1	the	the	DET
fcis-29191	137	2	processed	process	VERB
fcis-29191	137	3	effective	effective	ADJ
fcis-29191	137	4	dataset	dataset	NOUN
fcis-29191	137	5	contains	contain	VERB
fcis-29191	137	6	a	a	DET
fcis-29191	137	7	total	total	NOUN
fcis-29191	137	8	of	of	ADP
fcis-29191	137	9	47242	47242	NUM
fcis-29191	137	10	texts	text	NOUN
fcis-29191	137	11	,	,	PUNCT
fcis-29191	137	12	with	with	ADP
fcis-29191	137	13	a	a	DET
fcis-29191	137	14	maximum	maximum	ADJ
fcis-29191	137	15	text	text	NOUN
fcis-29191	137	16	length	length	NOUN
fcis-29191	137	17	of	of	ADP
fcis-29191	137	18	322	322	NUM
fcis-29191	137	19	,	,	PUNCT
fcis-29191	137	20	a	a	DET
fcis-29191	137	21	minimum	minimum	NOUN
fcis-29191	137	22	text	text	NOUN
fcis-29191	137	23	length	length	NOUN
fcis-29191	137	24	of	of	ADP
fcis-29191	137	25	3	3	NUM
fcis-29191	137	26	,	,	PUNCT
fcis-29191	137	27	an	an	DET
fcis-29191	137	28	average	average	ADJ
fcis-29191	137	29	text	text	NOUN
fcis-29191	137	30	length	length	NOUN
fcis-29191	137	31	of	of	ADP
fcis-29191	137	32	25	25	NUM
fcis-29191	137	33	,	,	PUNCT
fcis-29191	137	34	and	and	CCONJ
fcis-29191	137	35	6	6	NUM
fcis-29191	137	36	types	type	NOUN
fcis-29191	137	37	of	of	ADP
fcis-29191	137	38	labels	label	NOUN
fcis-29191	137	39	.	.	PUNCT
fcis-29191	138	1	the	the	DET
fcis-29191	138	2	specific	specific	ADJ
fcis-29191	138	3	number	number	NOUN
fcis-29191	138	4	of	of	ADP
fcis-29191	138	5	texts	text	NOUN
fcis-29191	138	6	is	be	AUX
fcis-29191	138	7	shown	show	VERB
fcis-29191	138	8	in	in	ADP
fcis-29191	138	9	table	table	NOUN
fcis-29191	138	10	1	1	NUM
fcis-29191	138	11	.	.	PUNCT
fcis-29191	138	12	table	table	NOUN
fcis-29191	138	13	1	1	NUM
fcis-29191	138	14	.	.	PUNCT
fcis-29191	139	1	the	the	DET
fcis-29191	139	2	number	number	NOUN
fcis-29191	139	3	of	of	ADP
fcis-29191	139	4	texts	text	NOUN
fcis-29191	139	5	in	in	ADP
fcis-29191	139	6	each	each	DET
fcis-29191	139	7	category	category	NOUN
fcis-29191	139	8	of	of	ADP
fcis-29191	139	9	the	the	DET
fcis-29191	139	10	cyberbullying_tweets	cyberbullying_tweet	NOUN
fcis-29191	139	11	dataset	dataset	NOUN
fcis-29191	139	12	text	text	NOUN
fcis-29191	139	13	category	category	NOUN
fcis-29191	139	14	effective	effective	ADJ
fcis-29191	139	15	text	text	NOUN
fcis-29191	139	16	quantity	quantity	NOUN
fcis-29191	139	17	not_cyberbullying	not_cyberbullye	VERB
fcis-29191	139	18	7802	7802	NUM
fcis-29191	139	19	age	age	NOUN
fcis-29191	139	20	7990	7990	NUM
fcis-29191	139	21	ethnicity	ethnicity	NOUN
fcis-29191	139	22	7954	7954	NUM
fcis-29191	139	23	gender	gender	NOUN
fcis-29191	139	24	7918	7918	NUM
fcis-29191	139	25	religion	religion	NOUN
fcis-29191	139	26	7997	7997	NUM
fcis-29191	139	27	other_cyberbullying	other_cyberbullye	VERB
fcis-29191	139	28	7581	7581	NUM
fcis-29191	139	29	4.2	4.2	NUM
fcis-29191	139	30	.	.	PUNCT
fcis-29191	140	1	experimental	experimental	ADJ
fcis-29191	140	2	setup	setup	NOUN
fcis-29191	140	3	for	for	ADP
fcis-29191	140	4	bert	bert	PROPN
fcis-29191	140	5	and	and	CCONJ
fcis-29191	140	6	roberta	roberta	PROPN
fcis-29191	140	7	,	,	PUNCT
fcis-29191	140	8	use	use	VERB
fcis-29191	140	9	the	the	DET
fcis-29191	140	10	output	output	NOUN
fcis-29191	140	11	features	feature	NOUN
fcis-29191	140	12	of	of	ADP
fcis-29191	140	13	[	[	X
fcis-29191	140	14	cls	cls	X
fcis-29191	140	15	]	]	PUNCT
fcis-29191	140	16	tokens	token	NOUN
fcis-29191	140	17	as	as	ADP
fcis-29191	140	18	document	document	NOUN
fcis-29191	140	19	embeddings	embedding	NOUN
fcis-29191	140	20	,	,	PUNCT
fcis-29191	140	21	and	and	CCONJ
fcis-29191	140	22	then	then	ADV
fcis-29191	140	23	use	use	VERB
fcis-29191	140	24	a	a	DET
fcis-29191	140	25	feedforward	feedforward	ADJ
fcis-29191	140	26	layer	layer	NOUN
fcis-29191	140	27	to	to	PART
fcis-29191	140	28	output	output	VERB
fcis-29191	140	29	the	the	DET
fcis-29191	140	30	final	final	ADJ
fcis-29191	140	31	prediction	prediction	NOUN
fcis-29191	140	32	.	.	PUNCT
fcis-29191	141	1	use	use	VERB
fcis-29191	141	2	bert	bert	PROPN
fcis-29191	141	3	or	or	CCONJ
fcis-29191	141	4	roberta	roberta	PROPN
fcis-29191	141	5	and	and	CCONJ
fcis-29191	141	6	a	a	DET
fcis-29191	141	7	two	two	NUM
fcis-29191	141	8	-	-	PUNCT
fcis-29191	141	9	layer	layer	NOUN
fcis-29191	141	10	gcn	gcn	NOUN
fcis-29191	141	11	to	to	PART
fcis-29191	141	12	achieve	achieve	VERB
fcis-29191	141	13	joint	joint	ADJ
fcis-29191	141	14	training	training	NOUN
fcis-29191	141	15	.	.	PUNCT
fcis-29191	142	1	the	the	DET
fcis-29191	142	2	learning	learning	NOUN
fcis-29191	142	3	rate	rate	NOUN
fcis-29191	142	4	of	of	ADP
fcis-29191	142	5	gcn	gcn	NOUN
fcis-29191	142	6	module	module	NOUN
fcis-29191	142	7	is	be	AUX
fcis-29191	142	8	initialized	initialize	VERB
fcis-29191	142	9	to	to	ADP
fcis-29191	142	10	1e-3	1e-3	NUM
fcis-29191	142	11	,	,	PUNCT
fcis-29191	142	12	and	and	CCONJ
fcis-29191	142	13	for	for	ADP
fcis-29191	142	14	fine	fine	ADV
fcis-29191	142	15	-	-	PUNCT
fcis-29191	142	16	tuned	tune	VERB
fcis-29191	142	17	bert	bert	NOUN
fcis-29191	142	18	or	or	CCONJ
fcis-29191	142	19	roberta	roberta	PROPN
fcis-29191	142	20	modules	module	NOUN
fcis-29191	142	21	,	,	PUNCT
fcis-29191	142	22	the	the	DET
fcis-29191	142	23	learning	learning	NOUN
fcis-29191	142	24	rate	rate	NOUN
fcis-29191	142	25	is	be	AUX
fcis-29191	142	26	initialized	initialize	VERB
fcis-29191	142	27	to	to	ADP
fcis-29191	142	28	1e-5	1e-5	NUM
fcis-29191	142	29	.	.	PUNCT
fcis-29191	143	1	4.3	4.3	NUM
fcis-29191	143	2	.	.	PUNCT
fcis-29191	143	3	comparative	comparative	ADJ
fcis-29191	143	4	experimental	experimental	ADJ
fcis-29191	143	5	model	model	NOUN
fcis-29191	143	6	this	this	DET
fcis-29191	143	7	article	article	NOUN
fcis-29191	143	8	selects	select	VERB
fcis-29191	143	9	fasttext	fasttext	PROPN
fcis-29191	143	10	,	,	PUNCT
fcis-29191	143	11	textcnn	textcnn	PROPN
fcis-29191	143	12	,	,	PUNCT
fcis-29191	143	13	textrnn	textrnn	NOUN
fcis-29191	143	14	,	,	PUNCT
fcis-29191	143	15	bert	bert	PROPN
fcis-29191	143	16	,	,	PUNCT
fcis-29191	143	17	roberta	roberta	PROPN
fcis-29191	143	18	and	and	CCONJ
fcis-29191	143	19	other	other	ADJ
fcis-29191	143	20	benchmark	benchmark	NOUN
fcis-29191	143	21	models	model	NOUN
fcis-29191	143	22	for	for	ADP
fcis-29191	143	23	comparative	comparative	ADJ
fcis-29191	143	24	experiments	experiment	NOUN
fcis-29191	143	25	.	.	PUNCT
fcis-29191	144	1	specifically	specifically	ADV
fcis-29191	144	2	,	,	PUNCT
fcis-29191	144	3	as	as	SCONJ
fcis-29191	144	4	follows	follow	VERB
fcis-29191	144	5	:	:	PUNCT
fcis-29191	144	6	fasttext	fasttext	PROPN
fcis-29191	144	7	:	:	PUNCT
fcis-29191	144	8	a	a	DET
fcis-29191	144	9	simple	simple	ADJ
fcis-29191	144	10	and	and	CCONJ
fcis-29191	144	11	efficient	efficient	ADJ
fcis-29191	144	12	text	text	NOUN
fcis-29191	144	13	classification	classification	NOUN
fcis-29191	144	14	method	method	NOUN
fcis-29191	144	15	33	33	NUM
fcis-29191	144	16	that	that	PRON
fcis-29191	144	17	considers	consider	VERB
fcis-29191	144	18	the	the	DET
fcis-29191	144	19	average	average	ADJ
fcis-29191	144	20	value	value	NOUN
fcis-29191	144	21	of	of	ADP
fcis-29191	144	22	word	word	NOUN
fcis-29191	144	23	/	/	SYM
fcis-29191	144	24	n	n	CCONJ
fcis-29191	144	25	-	-	PUNCT
fcis-29191	144	26	gram	gram	NOUN
fcis-29191	144	27	embeddings	embedding	NOUN
fcis-29191	144	28	as	as	ADP
fcis-29191	144	29	document	document	NOUN
fcis-29191	144	30	embeddings	embedding	NOUN
fcis-29191	144	31	,	,	PUNCT
fcis-29191	144	32	and	and	CCONJ
fcis-29191	144	33	then	then	ADV
fcis-29191	144	34	inputs	input	VERB
fcis-29191	144	35	the	the	DET
fcis-29191	144	36	document	document	NOUN
fcis-29191	144	37	embeddings	embedding	NOUN
fcis-29191	144	38	into	into	ADP
fcis-29191	144	39	a	a	DET
fcis-29191	144	40	linear	linear	ADJ
fcis-29191	144	41	classifier	classifier	NOUN
fcis-29191	144	42	.	.	PUNCT
fcis-29191	145	1	textcnn	textcnn	PROPN
fcis-29191	145	2	:	:	PUNCT
fcis-29191	145	3	a	a	DET
fcis-29191	145	4	text	text	NOUN
fcis-29191	145	5	classification	classification	NOUN
fcis-29191	145	6	model	model	NOUN
fcis-29191	145	7	based	base	VERB
fcis-29191	145	8	on	on	ADP
fcis-29191	145	9	convolutional	convolutional	ADJ
fcis-29191	145	10	neural	neural	ADJ
fcis-29191	145	11	networks	network	NOUN
fcis-29191	145	12	.	.	PUNCT
fcis-29191	146	1	it	it	PRON
fcis-29191	146	2	extracts	extract	VERB
fcis-29191	146	3	text	text	NOUN
fcis-29191	146	4	features	feature	VERB
fcis-29191	146	5	through	through	ADP
fcis-29191	146	6	convolution	convolution	NOUN
fcis-29191	146	7	operations	operation	NOUN
fcis-29191	146	8	and	and	CCONJ
fcis-29191	146	9	then	then	ADV
fcis-29191	146	10	uses	use	VERB
fcis-29191	146	11	max	max	PROPN
fcis-29191	146	12	pooling	pool	VERB
fcis-29191	146	13	operations	operation	NOUN
fcis-29191	146	14	to	to	PART
fcis-29191	146	15	extract	extract	VERB
fcis-29191	146	16	the	the	DET
fcis-29191	146	17	most	most	ADV
fcis-29191	146	18	important	important	ADJ
fcis-29191	146	19	features	feature	NOUN
fcis-29191	146	20	.	.	PUNCT
fcis-29191	147	1	textrnn	textrnn	NOUN
fcis-29191	147	2	:	:	PUNCT
fcis-29191	147	3	use	use	NOUN
fcis-29191	147	4	rnn	rnn	NOUN
fcis-29191	147	5	to	to	PART
fcis-29191	147	6	calculate	calculate	VERB
fcis-29191	147	7	the	the	DET
fcis-29191	147	8	sequential	sequential	ADJ
fcis-29191	147	9	representation	representation	NOUN
fcis-29191	147	10	of	of	ADP
fcis-29191	147	11	text	text	NOUN
fcis-29191	147	12	.	.	PUNCT
fcis-29191	148	1	a	a	DET
fcis-29191	148	2	model	model	NOUN
fcis-29191	148	3	for	for	ADP
fcis-29191	148	4	text	text	NOUN
fcis-29191	148	5	classification	classification	NOUN
fcis-29191	148	6	tasks	task	NOUN
fcis-29191	148	7	using	use	VERB
fcis-29191	148	8	rnn	rnn	NOUN
fcis-29191	148	9	class	class	NOUN
fcis-29191	148	10	recurrent	recurrent	ADJ
fcis-29191	148	11	neural	neural	ADJ
fcis-29191	148	12	networks	network	NOUN
fcis-29191	148	13	.	.	PUNCT
fcis-29191	149	1	bert	bert	PROPN
fcis-29191	149	2	:	:	PUNCT
fcis-29191	149	3	a	a	DET
fcis-29191	149	4	powerful	powerful	ADJ
fcis-29191	149	5	pre	pre	ADJ
fcis-29191	149	6	-	-	ADJ
fcis-29191	149	7	trained	train	VERB
fcis-29191	149	8	model	model	NOUN
fcis-29191	149	9	that	that	PRON
fcis-29191	149	10	can	can	AUX
fcis-29191	149	11	utilize	utilize	VERB
fcis-29191	149	12	the	the	DET
fcis-29191	149	13	knowledge	knowledge	NOUN
fcis-29191	149	14	obtained	obtain	VERB
fcis-29191	149	15	from	from	ADP
fcis-29191	149	16	pre	pre	ADJ
fcis-29191	149	17	training	training	NOUN
fcis-29191	149	18	for	for	ADP
fcis-29191	149	19	classification	classification	NOUN
fcis-29191	149	20	.	.	PUNCT
fcis-29191	150	1	roberta	roberta	NOUN
fcis-29191	150	2	:	:	PUNCT
fcis-29191	150	3	it	it	PRON
fcis-29191	150	4	is	be	AUX
fcis-29191	150	5	a	a	DET
fcis-29191	150	6	further	further	ADJ
fcis-29191	150	7	improvement	improvement	NOUN
fcis-29191	150	8	of	of	ADP
fcis-29191	150	9	the	the	DET
fcis-29191	150	10	bert	bert	PROPN
fcis-29191	150	11	model	model	NOUN
fcis-29191	150	12	,	,	PUNCT
fcis-29191	150	13	with	with	ADP
fcis-29191	150	14	stronger	strong	ADJ
fcis-29191	150	15	understanding	understanding	NOUN
fcis-29191	150	16	and	and	CCONJ
fcis-29191	150	17	representation	representation	NOUN
fcis-29191	150	18	capabilities	capability	NOUN
fcis-29191	150	19	.	.	PUNCT
fcis-29191	151	1	4.4	4.4	NUM
fcis-29191	151	2	.	.	PUNCT
fcis-29191	152	1	evaluation	evaluation	NOUN
fcis-29191	152	2	indicators	indicator	NOUN
fcis-29191	152	3	text	text	VERB
fcis-29191	152	4	classification	classification	NOUN
fcis-29191	152	5	tasks	task	NOUN
fcis-29191	152	6	are	be	AUX
fcis-29191	152	7	usually	usually	ADV
fcis-29191	152	8	divided	divide	VERB
fcis-29191	152	9	into	into	ADP
fcis-29191	152	10	binary	binary	ADJ
fcis-29191	152	11	classification	classification	NOUN
fcis-29191	152	12	and	and	CCONJ
fcis-29191	152	13	multi	multi	ADJ
fcis-29191	152	14	classification	classification	NOUN
fcis-29191	152	15	tasks	task	NOUN
fcis-29191	152	16	,	,	PUNCT
fcis-29191	152	17	each	each	PRON
fcis-29191	152	18	with	with	ADP
fcis-29191	152	19	different	different	ADJ
fcis-29191	152	20	evaluation	evaluation	NOUN
fcis-29191	152	21	criteria	criterion	NOUN
fcis-29191	152	22	that	that	PRON
fcis-29191	152	23	help	help	AUX
fcis-29191	152	24	understand	understand	VERB
fcis-29191	152	25	the	the	DET
fcis-29191	152	26	accuracy	accuracy	NOUN
fcis-29191	152	27	,	,	PUNCT
fcis-29191	152	28	stability	stability	NOUN
fcis-29191	152	29	,	,	PUNCT
fcis-29191	152	30	and	and	CCONJ
fcis-29191	152	31	applicability	applicability	NOUN
fcis-29191	152	32	of	of	ADP
fcis-29191	152	33	the	the	DET
fcis-29191	152	34	model	model	NOUN
fcis-29191	152	35	in	in	ADP
fcis-29191	152	36	specific	specific	ADJ
fcis-29191	152	37	tasks	task	NOUN
fcis-29191	152	38	.	.	PUNCT
fcis-29191	153	1	common	common	ADJ
fcis-29191	153	2	evaluation	evaluation	NOUN
fcis-29191	153	3	metrics	metric	NOUN
fcis-29191	153	4	include	include	VERB
fcis-29191	153	5	accuracy	accuracy	NOUN
fcis-29191	153	6	,	,	PUNCT
fcis-29191	153	7	precision	precision	NOUN
fcis-29191	153	8	,	,	PUNCT
fcis-29191	153	9	recall	recall	NOUN
fcis-29191	153	10	,	,	PUNCT
fcis-29191	153	11	and	and	CCONJ
fcis-29191	153	12	f1	f1	NOUN
fcis-29191	153	13	score	score	NOUN
fcis-29191	153	14	.	.	PUNCT
fcis-29191	154	1	in	in	ADP
fcis-29191	154	2	the	the	DET
fcis-29191	154	3	binary	binary	PROPN
fcis-29191	154	4	classification	classification	NOUN
fcis-29191	154	5	task	task	NOUN
fcis-29191	154	6	,	,	PUNCT
fcis-29191	154	7	the	the	DET
fcis-29191	154	8	following	follow	VERB
fcis-29191	154	9	are	be	AUX
fcis-29191	154	10	the	the	DET
fcis-29191	154	11	calculation	calculation	NOUN
fcis-29191	154	12	formulas	formula	NOUN
fcis-29191	154	13	for	for	ADP
fcis-29191	154	14	each	each	DET
fcis-29191	154	15	indicator	indicator	NOUN
fcis-29191	154	16	:	:	PUNCT
fcis-29191	154	17	tp	tp	PART
fcis-29191	154	18	accuracy	accuracy	NOUN
fcis-29191	154	19	tp	tp	ADP
fcis-29191	154	20	fp	fp	PROPN
fcis-29191	154	21	fn	fn	PROPN
fcis-29191	155	1	fp	fp	PROPN
fcis-29191	156	1			PROPN
fcis-29191	157	1			PROPN
fcis-29191	158	1			PUNCT
fcis-29191	159	1			X
fcis-29191	159	2	(	(	PUNCT
fcis-29191	159	3	9	9	X
fcis-29191	159	4	)	)	PUNCT
fcis-29191	159	5	tp	tp	AUX
fcis-29191	159	6	precision	precision	NOUN
fcis-29191	159	7	tp	tp	ADP
fcis-29191	159	8	fp	fp	PROPN
fcis-29191	159	9			PROPN
fcis-29191	159	10			X
fcis-29191	159	11	(	(	PUNCT
fcis-29191	159	12	10	10	NUM
fcis-29191	159	13	)	)	PUNCT
fcis-29191	159	14	tp	tp	PART
fcis-29191	159	15	recall	recall	VERB
fcis-29191	159	16	tp	tp	ADP
fcis-29191	159	17	fn	fn	PROPN
fcis-29191	159	18			PROPN
fcis-29191	159	19			PUNCT
fcis-29191	159	20	(	(	PUNCT
fcis-29191	159	21	11	11	NUM
fcis-29191	159	22	)	)	SYM
fcis-29191	159	23	1	1	NUM
fcis-29191	159	24	2	2	NUM
fcis-29191	159	25	precision	precision	NOUN
fcis-29191	159	26	recall	recall	NOUN
fcis-29191	159	27	precision	precision	NOUN
fcis-29191	159	28	recall	recall	NOUN
fcis-29191	159	29	f	f	PROPN
fcis-29191	159	30			PROPN
fcis-29191	159	31			PROPN
fcis-29191	160	1			PROPN
fcis-29191	160	2			NOUN
fcis-29191	160	3	(	(	PUNCT
fcis-29191	160	4	12	12	NUM
fcis-29191	160	5	)	)	PUNCT
fcis-29191	160	6	tp	tp	NOUN
fcis-29191	160	7	(	(	PUNCT
fcis-29191	160	8	true	true	ADJ
fcis-29191	160	9	positive	positive	ADJ
fcis-29191	160	10	)	)	PUNCT
fcis-29191	160	11	,	,	PUNCT
fcis-29191	160	12	tn	tn	PROPN
fcis-29191	160	13	(	(	PUNCT
fcis-29191	160	14	true	true	ADJ
fcis-29191	160	15	negative	negative	NOUN
fcis-29191	160	16	)	)	PUNCT
fcis-29191	160	17	,	,	PUNCT
fcis-29191	160	18	fp	fp	INTJ
fcis-29191	160	19	(	(	PUNCT
fcis-29191	160	20	false	false	ADJ
fcis-29191	160	21	positive	positive	ADJ
fcis-29191	160	22	)	)	PUNCT
fcis-29191	160	23	,	,	PUNCT
fcis-29191	160	24	and	and	CCONJ
fcis-29191	160	25	fn	fn	INTJ
fcis-29191	160	26	(	(	PUNCT
fcis-29191	160	27	false	false	ADJ
fcis-29191	160	28	negative	negative	NOUN
fcis-29191	160	29	)	)	PUNCT
fcis-29191	160	30	represent	represent	VERB
fcis-29191	160	31	the	the	DET
fcis-29191	160	32	true	true	ADJ
fcis-29191	160	33	examples	example	NOUN
fcis-29191	160	34	,	,	PUNCT
fcis-29191	160	35	true	true	ADJ
fcis-29191	160	36	negative	negative	ADJ
fcis-29191	160	37	examples	example	NOUN
fcis-29191	160	38	,	,	PUNCT
fcis-29191	160	39	false	false	ADJ
fcis-29191	160	40	positive	positive	ADJ
fcis-29191	160	41	examples	example	NOUN
fcis-29191	160	42	,	,	PUNCT
fcis-29191	160	43	and	and	CCONJ
fcis-29191	160	44	false	false	ADJ
fcis-29191	160	45	negative	negative	ADJ
fcis-29191	160	46	examples	example	NOUN
fcis-29191	160	47	in	in	ADP
fcis-29191	160	48	the	the	DET
fcis-29191	160	49	confusion	confusion	NOUN
fcis-29191	160	50	matrix	matrix	NOUN
fcis-29191	160	51	.	.	PUNCT
fcis-29191	161	1	the	the	DET
fcis-29191	161	2	online	online	ADJ
fcis-29191	161	3	violence	violence	NOUN
fcis-29191	161	4	text	text	NOUN
fcis-29191	161	5	studied	study	VERB
fcis-29191	161	6	in	in	ADP
fcis-29191	161	7	this	this	DET
fcis-29191	161	8	article	article	NOUN
fcis-29191	161	9	belongs	belong	VERB
fcis-29191	161	10	to	to	ADP
fcis-29191	161	11	a	a	DET
fcis-29191	161	12	multi	multi	ADJ
fcis-29191	161	13	classification	classification	NOUN
fcis-29191	161	14	task	task	NOUN
fcis-29191	161	15	,	,	PUNCT
fcis-29191	161	16	therefore	therefore	ADV
fcis-29191	161	17	accuracy	accuracy	NOUN
fcis-29191	161	18	,	,	PUNCT
fcis-29191	161	19	weighted	weight	VERB
fcis-29191	161	20	average	average	ADJ
fcis-29191	161	21	accuracy	accuracy	NOUN
fcis-29191	161	22	,	,	PUNCT
fcis-29191	161	23	weighted	weight	VERB
fcis-29191	161	24	average	average	ADJ
fcis-29191	161	25	recall	recall	NOUN
fcis-29191	161	26	,	,	PUNCT
fcis-29191	161	27	and	and	CCONJ
fcis-29191	161	28	weighted	weight	VERB
fcis-29191	161	29	average	average	ADJ
fcis-29191	161	30	f1	f1	NOUN
fcis-29191	161	31	score	score	NOUN
fcis-29191	161	32	are	be	AUX
fcis-29191	161	33	used	use	VERB
fcis-29191	161	34	as	as	ADP
fcis-29191	161	35	indicators	indicator	NOUN
fcis-29191	161	36	for	for	ADP
fcis-29191	161	37	evaluating	evaluate	VERB
fcis-29191	161	38	the	the	DET
fcis-29191	161	39	model	model	NOUN
fcis-29191	161	40	.	.	PUNCT
fcis-29191	162	1	firstly	firstly	ADV
fcis-29191	162	2	,	,	PUNCT
fcis-29191	162	3	calculate	calculate	VERB
fcis-29191	162	4	the	the	DET
fcis-29191	162	5	indicators	indicator	NOUN
fcis-29191	162	6	for	for	ADP
fcis-29191	162	7	each	each	DET
fcis-29191	162	8	category	category	NOUN
fcis-29191	162	9	,	,	PUNCT
fcis-29191	162	10	then	then	ADV
fcis-29191	162	11	calculate	calculate	VERB
fcis-29191	162	12	the	the	DET
fcis-29191	162	13	weights	weight	NOUN
fcis-29191	162	14	of	of	ADP
fcis-29191	162	15	each	each	DET
fcis-29191	162	16	indicator	indicator	NOUN
fcis-29191	162	17	based	base	VERB
fcis-29191	162	18	on	on	ADP
fcis-29191	162	19	the	the	DET
fcis-29191	162	20	sample	sample	NOUN
fcis-29191	162	21	size	size	NOUN
fcis-29191	162	22	,	,	PUNCT
fcis-29191	162	23	and	and	CCONJ
fcis-29191	162	24	finally	finally	ADV
fcis-29191	162	25	calculate	calculate	VERB
fcis-29191	162	26	the	the	DET
fcis-29191	162	27	weighted	weighted	ADJ
fcis-29191	162	28	average	average	NOUN
fcis-29191	162	29	.	.	PUNCT
fcis-29191	163	1	the	the	DET
fcis-29191	163	2	specific	specific	ADJ
fcis-29191	163	3	calculation	calculation	NOUN
fcis-29191	163	4	formula	formula	NOUN
fcis-29191	163	5	is	be	AUX
fcis-29191	163	6	as	as	SCONJ
fcis-29191	163	7	follows	follow	VERB
fcis-29191	163	8	:	:	PUNCT
fcis-29191	163	9	1	1	NUM
fcis-29191	163	10	n	n	X
fcis-29191	163	11	ii	ii	NOUN
fcis-29191	163	12	tp	tp	NOUN
fcis-29191	163	13	accuracy	accuracy	NOUN
fcis-29191	163	14	sum	sum	NOUN
fcis-29191	163	15			NUM
fcis-29191	163	16			X
fcis-29191	163	17	(	(	PUNCT
fcis-29191	163	18	13	13	NUM
fcis-29191	163	19	)	)	SYM
fcis-29191	163	20	1	1	NUM
fcis-29191	163	21	n	n	DET
fcis-29191	163	22	weight	weight	NOUN
fcis-29191	164	1	i	i	PROPN
fcis-29191	164	2	ii	ii	PROPN
fcis-29191	164	3	precision	precision	VERB
fcis-29191	164	4	w	w	PROPN
fcis-29191	164	5	p	p	X
fcis-29191	164	6			ADJ
fcis-29191	164	7			PROPN
fcis-29191	164	8	(	(	PUNCT
fcis-29191	164	9	14	14	NUM
fcis-29191	164	10	)	)	SYM
fcis-29191	164	11	1	1	NUM
fcis-29191	164	12	n	n	DET
fcis-29191	164	13	weight	weight	NOUN
fcis-29191	164	14	i	i	PROPN
fcis-29191	164	15	ii	ii	PROPN
fcis-29191	164	16	recall	recall	VERB
fcis-29191	164	17	w	w	PROPN
fcis-29191	164	18	r	r	NOUN
fcis-29191	164	19			NUM
fcis-29191	164	20			PROPN
fcis-29191	164	21	(	(	PUNCT
fcis-29191	164	22	15	15	NUM
fcis-29191	164	23	)	)	PUNCT
fcis-29191	164	24	11	11	NUM
fcis-29191	164	25	1	1	NUM
fcis-29191	164	26	n	n	DET
fcis-29191	164	27	weight	weight	NOUN
fcis-29191	165	1	i	i	NOUN
fcis-29191	165	2	ii	ii	VERB
fcis-29191	166	1	f	f	PROPN
fcis-29191	166	2	w	w	PROPN
fcis-29191	166	3	f	f	PROPN
fcis-29191	166	4			NUM
fcis-29191	166	5			PROPN
fcis-29191	166	6	(	(	PUNCT
fcis-29191	166	7	16	16	NUM
fcis-29191	166	8	)	)	PUNCT
fcis-29191	166	9	among	among	ADP
fcis-29191	166	10	them	they	PRON
fcis-29191	166	11	,	,	PUNCT
fcis-29191	166	12	n	n	PRON
fcis-29191	166	13	represents	represent	VERB
fcis-29191	166	14	the	the	DET
fcis-29191	166	15	number	number	NOUN
fcis-29191	166	16	of	of	ADP
fcis-29191	166	17	categories	category	NOUN
fcis-29191	166	18	,	,	PUNCT
fcis-29191	166	19	and	and	CCONJ
fcis-29191	166	20	in	in	ADP
fcis-29191	166	21	this	this	DET
fcis-29191	166	22	study	study	NOUN
fcis-29191	166	23	,	,	PUNCT
fcis-29191	166	24	n=6	n=6	PROPN
fcis-29191	166	25	.	.	PUNCT
fcis-29191	167	1	tpi	tpi	PROPN
fcis-29191	167	2	refers	refer	VERB
fcis-29191	167	3	to	to	ADP
fcis-29191	167	4	the	the	DET
fcis-29191	167	5	number	number	NOUN
fcis-29191	167	6	of	of	ADP
fcis-29191	167	7	correctly	correctly	ADV
fcis-29191	167	8	predicted	predict	VERB
fcis-29191	167	9	samples	sample	NOUN
fcis-29191	167	10	in	in	ADP
fcis-29191	167	11	class	class	NOUN
fcis-29191	167	12	i	i	PROPN
fcis-29191	167	13	,	,	PUNCT
fcis-29191	167	14	and	and	CCONJ
fcis-29191	167	15	sum	sum	NOUN
fcis-29191	167	16	is	be	AUX
fcis-29191	167	17	the	the	DET
fcis-29191	167	18	total	total	ADJ
fcis-29191	167	19	number	number	NOUN
fcis-29191	167	20	of	of	ADP
fcis-29191	167	21	samples	sample	NOUN
fcis-29191	167	22	.	.	PUNCT
fcis-29191	168	1	wi	wi	PROPN
fcis-29191	168	2	,	,	PUNCT
fcis-29191	168	3	pi	pi	PROPN
fcis-29191	168	4	,	,	PUNCT
fcis-29191	168	5	ri	ri	PROPN
fcis-29191	168	6	,	,	PUNCT
fcis-29191	168	7	and	and	CCONJ
fcis-29191	168	8	f1i	f1i	PROPN
fcis-29191	168	9	respectively	respectively	ADV
fcis-29191	168	10	represent	represent	VERB
fcis-29191	168	11	the	the	DET
fcis-29191	168	12	weight	weight	NOUN
fcis-29191	168	13	,	,	PUNCT
fcis-29191	168	14	accuracy	accuracy	NOUN
fcis-29191	168	15	,	,	PUNCT
fcis-29191	168	16	recall	recall	NOUN
fcis-29191	168	17	,	,	PUNCT
fcis-29191	168	18	and	and	CCONJ
fcis-29191	168	19	f1	f1	NOUN
fcis-29191	168	20	value	value	NOUN
fcis-29191	168	21	of	of	ADP
fcis-29191	168	22	the	the	DET
fcis-29191	168	23	i	i	NOUN
fcis-29191	168	24	category	category	NOUN
fcis-29191	168	25	.	.	PUNCT
fcis-29191	169	1	4.5	4.5	NUM
fcis-29191	169	2	.	.	PUNCT
fcis-29191	169	3	result	result	VERB
fcis-29191	169	4	from	from	ADP
fcis-29191	169	5	table	table	NOUN
fcis-29191	169	6	2	2	NUM
fcis-29191	169	7	,	,	PUNCT
fcis-29191	169	8	it	it	PRON
fcis-29191	169	9	can	can	AUX
fcis-29191	169	10	be	be	AUX
fcis-29191	169	11	seen	see	VERB
fcis-29191	169	12	that	that	SCONJ
fcis-29191	169	13	among	among	ADP
fcis-29191	169	14	all	all	DET
fcis-29191	169	15	baseline	baseline	NOUN
fcis-29191	169	16	models	model	NOUN
fcis-29191	169	17	,	,	PUNCT
fcis-29191	169	18	bert	bert	PROPN
fcis-29191	169	19	and	and	CCONJ
fcis-29191	169	20	roberta	roberta	PROPN
fcis-29191	169	21	outperform	outperform	VERB
fcis-29191	169	22	other	other	ADJ
fcis-29191	169	23	baseline	baseline	NOUN
fcis-29191	169	24	models	model	NOUN
fcis-29191	169	25	due	due	ADP
fcis-29191	169	26	to	to	ADP
fcis-29191	169	27	their	their	PRON
fcis-29191	169	28	large	large	ADJ
fcis-29191	169	29	pre	pre	ADJ
fcis-29191	169	30	training	training	NOUN
fcis-29191	169	31	foundation	foundation	NOUN
fcis-29191	169	32	.	.	PUNCT
fcis-29191	170	1	roberta	roberta	PROPN
fcis-29191	170	2	uses	use	VERB
fcis-29191	170	3	larger	large	ADJ
fcis-29191	170	4	and	and	CCONJ
fcis-29191	170	5	richer	rich	ADJ
fcis-29191	170	6	text	text	NOUN
fcis-29191	170	7	data	datum	NOUN
fcis-29191	170	8	and	and	CCONJ
fcis-29191	170	9	longer	long	ADJ
fcis-29191	170	10	text	text	NOUN
fcis-29191	170	11	segments	segment	NOUN
fcis-29191	170	12	during	during	ADP
fcis-29191	170	13	the	the	DET
fcis-29191	170	14	training	training	NOUN
fcis-29191	170	15	process	process	NOUN
fcis-29191	170	16	,	,	PUNCT
fcis-29191	170	17	which	which	PRON
fcis-29191	170	18	enables	enable	VERB
fcis-29191	170	19	it	it	PRON
fcis-29191	170	20	to	to	PART
fcis-29191	170	21	more	more	ADV
fcis-29191	170	22	accurately	accurately	ADV
fcis-29191	170	23	understand	understand	VERB
fcis-29191	170	24	and	and	CCONJ
fcis-29191	170	25	capture	capture	VERB
fcis-29191	170	26	contextual	contextual	ADJ
fcis-29191	170	27	relationships	relationship	NOUN
fcis-29191	170	28	in	in	ADP
fcis-29191	170	29	text	text	NOUN
fcis-29191	170	30	.	.	PUNCT
fcis-29191	171	1	at	at	ADP
fcis-29191	171	2	the	the	DET
fcis-29191	171	3	same	same	ADJ
fcis-29191	171	4	time	time	NOUN
fcis-29191	171	5	,	,	PUNCT
fcis-29191	171	6	it	it	PRON
fcis-29191	171	7	adopts	adopt	VERB
fcis-29191	171	8	stricter	strict	ADJ
fcis-29191	171	9	masking	masking	NOUN
fcis-29191	171	10	strategies	strategy	NOUN
fcis-29191	171	11	and	and	CCONJ
fcis-29191	171	12	deeper	deep	ADJ
fcis-29191	171	13	network	network	NOUN
fcis-29191	171	14	structures	structure	NOUN
fcis-29191	171	15	,	,	PUNCT
fcis-29191	171	16	resulting	result	VERB
fcis-29191	171	17	in	in	ADP
fcis-29191	171	18	higher	high	ADJ
fcis-29191	171	19	classification	classification	NOUN
fcis-29191	171	20	performance	performance	NOUN
fcis-29191	171	21	than	than	ADP
fcis-29191	171	22	bert	bert	PROPN
fcis-29191	171	23	,	,	PUNCT
fcis-29191	171	24	achieving	achieve	VERB
fcis-29191	171	25	the	the	DET
fcis-29191	171	26	best	good	ADJ
fcis-29191	171	27	results	result	NOUN
fcis-29191	171	28	in	in	ADP
fcis-29191	171	29	the	the	DET
fcis-29191	171	30	baseline	baseline	NOUN
fcis-29191	171	31	model	model	NOUN
fcis-29191	171	32	with	with	ADP
fcis-29191	171	33	a	a	DET
fcis-29191	171	34	classification	classification	NOUN
fcis-29191	171	35	accuracy	accuracy	NOUN
fcis-29191	171	36	of	of	ADP
fcis-29191	171	37	82.41	82.41	NUM
fcis-29191	171	38	%	%	NOUN
fcis-29191	171	39	and	and	CCONJ
fcis-29191	171	40	a	a	DET
fcis-29191	171	41	weighted	weight	VERB
fcis-29191	171	42	average	average	ADJ
fcis-29191	171	43	f1	f1	NOUN
fcis-29191	171	44	value	value	NOUN
fcis-29191	171	45	of	of	ADP
fcis-29191	171	46	82.30	82.30	NUM
fcis-29191	171	47	%	%	NOUN
fcis-29191	171	48	.	.	PUNCT
fcis-29191	172	1	the	the	DET
fcis-29191	172	2	rrobertagcn	rrobertagcn	PROPN
fcis-29191	172	3	model	model	PROPN
fcis-29191	172	4	proposed	propose	VERB
fcis-29191	172	5	in	in	ADP
fcis-29191	172	6	this	this	DET
fcis-29191	172	7	article	article	NOUN
fcis-29191	172	8	outperforms	outperform	VERB
fcis-29191	172	9	the	the	DET
fcis-29191	172	10	baseline	baseline	NOUN
fcis-29191	172	11	model	model	NOUN
fcis-29191	172	12	in	in	ADP
fcis-29191	172	13	classifying	classify	VERB
fcis-29191	172	14	online	online	ADJ
fcis-29191	172	15	violent	violent	ADJ
fcis-29191	172	16	text	text	NOUN
fcis-29191	172	17	datasets	dataset	NOUN
fcis-29191	172	18	,	,	PUNCT
fcis-29191	172	19	with	with	ADP
fcis-29191	172	20	an	an	DET
fcis-29191	172	21	accuracy	accuracy	NOUN
fcis-29191	172	22	improvement	improvement	NOUN
fcis-29191	172	23	of	of	ADP
fcis-29191	172	24	2.36	2.36	NUM
fcis-29191	172	25	%	%	NOUN
fcis-29191	172	26	and	and	CCONJ
fcis-29191	172	27	an	an	DET
fcis-29191	172	28	f1	f1	NOUN
fcis-29191	172	29	score	score	NOUN
fcis-29191	172	30	improvement	improvement	NOUN
fcis-29191	172	31	of	of	ADP
fcis-29191	172	32	1.83	1.83	NUM
fcis-29191	172	33	%	%	NOUN
fcis-29191	172	34	compared	compare	VERB
fcis-29191	172	35	to	to	ADP
fcis-29191	172	36	the	the	DET
fcis-29191	172	37	best	well	ADV
fcis-29191	172	38	performing	perform	VERB
fcis-29191	172	39	baseline	baseline	NOUN
fcis-29191	172	40	model	model	NOUN
fcis-29191	172	41	.	.	PUNCT
fcis-29191	173	1	the	the	DET
fcis-29191	173	2	specific	specific	ADJ
fcis-29191	173	3	results	result	NOUN
fcis-29191	173	4	are	be	AUX
fcis-29191	173	5	shown	show	VERB
fcis-29191	173	6	in	in	ADP
fcis-29191	173	7	table	table	NOUN
fcis-29191	173	8	2	2	NUM
fcis-29191	173	9	.	.	PUNCT
fcis-29191	174	1	the	the	DET
fcis-29191	174	2	results	result	NOUN
fcis-29191	174	3	in	in	ADP
fcis-29191	174	4	table	table	NOUN
fcis-29191	174	5	2	2	NUM
fcis-29191	174	6	demonstrate	demonstrate	VERB
fcis-29191	174	7	the	the	DET
fcis-29191	174	8	effectiveness	effectiveness	NOUN
fcis-29191	174	9	of	of	ADP
fcis-29191	174	10	the	the	DET
fcis-29191	174	11	proposed	propose	VERB
fcis-29191	174	12	method	method	NOUN
fcis-29191	174	13	in	in	ADP
fcis-29191	174	14	fine	fine	ADV
fcis-29191	174	15	-	-	PUNCT
fcis-29191	174	16	grained	grain	VERB
fcis-29191	174	17	classification	classification	NOUN
fcis-29191	174	18	of	of	ADP
fcis-29191	174	19	online	online	ADJ
fcis-29191	174	20	violent	violent	ADJ
fcis-29191	174	21	texts	text	NOUN
fcis-29191	174	22	,	,	PUNCT
fcis-29191	174	23	mainly	mainly	ADV
fcis-29191	174	24	due	due	ADP
fcis-29191	174	25	to	to	ADP
fcis-29191	174	26	the	the	DET
fcis-29191	174	27	following	following	ADJ
fcis-29191	174	28	reasons	reason	NOUN
fcis-29191	174	29	:	:	PUNCT
fcis-29191	174	30	1	1	X
fcis-29191	174	31	.	.	X
fcis-29191	174	32	the	the	DET
fcis-29191	174	33	graph	graph	NOUN
fcis-29191	174	34	convolution	convolution	NOUN
fcis-29191	174	35	model	model	NOUN
fcis-29191	174	36	can	can	AUX
fcis-29191	174	37	capture	capture	VERB
fcis-29191	174	38	document	document	NOUN
fcis-29191	174	39	word	word	NOUN
fcis-29191	174	40	relationships	relationship	NOUN
fcis-29191	174	41	and	and	CCONJ
fcis-29191	174	42	global	global	ADJ
fcis-29191	174	43	word	word	NOUN
fcis-29191	174	44	relationships	relationship	NOUN
fcis-29191	174	45	,	,	PUNCT
fcis-29191	174	46	enabling	enable	VERB
fcis-29191	174	47	effective	effective	ADJ
fcis-29191	174	48	propagation	propagation	NOUN
fcis-29191	174	49	of	of	ADP
fcis-29191	174	50	document	document	NOUN
fcis-29191	174	51	label	label	NOUN
fcis-29191	174	52	information	information	NOUN
fcis-29191	174	53	through	through	ADP
fcis-29191	174	54	word	word	NOUN
fcis-29191	174	55	nodes	node	NOUN
fcis-29191	174	56	in	in	ADP
fcis-29191	174	57	the	the	DET
fcis-29191	174	58	graph	graph	NOUN
fcis-29191	174	59	.	.	PUNCT
fcis-29191	175	1	2	2	X
fcis-29191	175	2	.	.	X
fcis-29191	175	3	introducing	introduce	VERB
fcis-29191	175	4	syntactic	syntactic	ADJ
fcis-29191	175	5	relationships	relationship	NOUN
fcis-29191	175	6	into	into	ADP
fcis-29191	175	7	bert	bert	NOUN
fcis-29191	175	8	related	relate	VERB
fcis-29191	175	9	models	model	NOUN
fcis-29191	175	10	enables	enable	VERB
fcis-29191	175	11	the	the	DET
fcis-29191	175	12	model	model	NOUN
fcis-29191	175	13	to	to	PART
fcis-29191	175	14	more	more	ADV
fcis-29191	175	15	accurately	accurately	ADV
fcis-29191	175	16	capture	capture	VERB
fcis-29191	175	17	the	the	DET
fcis-29191	175	18	dependencies	dependency	NOUN
fcis-29191	175	19	and	and	CCONJ
fcis-29191	175	20	semantic	semantic	ADJ
fcis-29191	175	21	connections	connection	NOUN
fcis-29191	175	22	within	within	ADP
fcis-29191	175	23	sentences	sentence	NOUN
fcis-29191	175	24	,	,	PUNCT
fcis-29191	175	25	while	while	SCONJ
fcis-29191	175	26	significantly	significantly	ADV
fcis-29191	175	27	improving	improve	VERB
fcis-29191	175	28	classification	classification	NOUN
fcis-29191	175	29	performance	performance	NOUN
fcis-29191	175	30	by	by	ADP
fcis-29191	175	31	integrating	integrate	VERB
fcis-29191	175	32	the	the	DET
fcis-29191	175	33	advantages	advantage	NOUN
fcis-29191	175	34	of	of	ADP
fcis-29191	175	35	pre	pre	X
fcis-29191	175	36	trained	train	VERB
fcis-29191	175	37	models	model	NOUN
fcis-29191	175	38	and	and	CCONJ
fcis-29191	175	39	graph	graph	NOUN
fcis-29191	175	40	models	model	NOUN
fcis-29191	175	41	.	.	PUNCT
fcis-29191	176	1	table	table	NOUN
fcis-29191	176	2	2	2	NUM
fcis-29191	176	3	.	.	PUNCT
fcis-29191	176	4	comparative	comparative	ADJ
fcis-29191	176	5	experimental	experimental	ADJ
fcis-29191	176	6	results	result	NOUN
fcis-29191	176	7	model	model	NOUN
fcis-29191	176	8	accuracy	accuracy	NOUN
fcis-29191	176	9	(	(	PUNCT
fcis-29191	176	10	%	%	INTJ
fcis-29191	176	11	)	)	PUNCT
fcis-29191	176	12	weighted	weight	VERB
fcis-29191	176	13	average	average	ADJ
fcis-29191	176	14	precision	precision	NOUN
fcis-29191	176	15	(	(	PUNCT
fcis-29191	176	16	%	%	INTJ
fcis-29191	176	17	)	)	PUNCT
fcis-29191	176	18	weighted	weight	VERB
fcis-29191	176	19	average	average	ADJ
fcis-29191	176	20	recall	recall	NOUN
fcis-29191	176	21	(	(	PUNCT
fcis-29191	176	22	%	%	INTJ
fcis-29191	176	23	)	)	PUNCT
fcis-29191	176	24	weighted	weight	VERB
fcis-29191	176	25	f1	f1	NOUN
fcis-29191	176	26	(	(	PUNCT
fcis-29191	176	27	%	%	INTJ
fcis-29191	176	28	)	)	PUNCT
fcis-29191	177	1	fasttext	fasttext	NOUN
fcis-29191	177	2	76.52	76.52	NUM
fcis-29191	177	3	76.36	76.36	NUM
fcis-29191	177	4	76.31	76.31	NUM
fcis-29191	177	5	76.44	76.44	NUM
fcis-29191	177	6	textcnn	textcnn	NOUN
fcis-29191	177	7	78.12	78.12	NUM
fcis-29191	177	8	77.76	77.76	NUM
fcis-29191	177	9	77.82	77.82	NUM
fcis-29191	177	10	78.06	78.06	NUM
fcis-29191	177	11	textrnn	textrnn	NOUN
fcis-29191	177	12	80.52	80.52	NUM
fcis-29191	177	13	80.24	80.24	NUM
fcis-29191	177	14	80.28	80.28	NUM
fcis-29191	177	15	80.04	80.04	NUM
fcis-29191	177	16	bert	bert	PROPN
fcis-29191	177	17	81.51	81.51	NUM
fcis-29191	177	18	82.07	82.07	NUM
fcis-29191	177	19	81.01	81.01	NUM
fcis-29191	177	20	81.96	81.96	NUM
fcis-29191	177	21	roberta	roberta	PROPN
fcis-29191	177	22	82.41	82.41	NUM
fcis-29191	177	23	82.46	82.46	NUM
fcis-29191	177	24	82.33	82.33	NUM
fcis-29191	177	25	82.30	82.30	NUM
fcis-29191	177	26	r	r	NOUN
fcis-29191	177	27	-	-	PUNCT
fcis-29191	177	28	bertgcn	bertgcn	NOUN
fcis-29191	177	29	84.16	84.16	NUM
fcis-29191	177	30	84.23	84.23	NUM
fcis-29191	177	31	83.92	83.92	NUM
fcis-29191	177	32	84.02	84.02	NUM
fcis-29191	177	33	r	r	X
fcis-29191	177	34	-	-	PUNCT
fcis-29191	177	35	robertagcn	robertagcn	NOUN
fcis-29191	177	36	84.77	84.77	NUM
fcis-29191	177	37	84.65	84.65	NUM
fcis-29191	177	38	84.56	84.56	NUM
fcis-29191	177	39	84.13	84.13	NUM
fcis-29191	177	40	4.6	4.6	NUM
fcis-29191	177	41	.	.	PUNCT
fcis-29191	178	1	ablation	ablation	NOUN
fcis-29191	178	2	experiment	experiment	NOUN
fcis-29191	178	3	in	in	ADP
fcis-29191	178	4	order	order	NOUN
fcis-29191	178	5	to	to	PART
fcis-29191	178	6	investigate	investigate	VERB
fcis-29191	178	7	the	the	DET
fcis-29191	178	8	impact	impact	NOUN
fcis-29191	178	9	of	of	ADP
fcis-29191	178	10	each	each	DET
fcis-29191	178	11	module	module	NOUN
fcis-29191	178	12	on	on	ADP
fcis-29191	178	13	the	the	DET
fcis-29191	178	14	model	model	NOUN
fcis-29191	178	15	,	,	PUNCT
fcis-29191	178	16	ablation	ablation	NOUN
fcis-29191	178	17	experiments	experiment	NOUN
fcis-29191	178	18	were	be	AUX
fcis-29191	178	19	conducted	conduct	VERB
fcis-29191	178	20	.	.	PUNCT
fcis-29191	179	1	the	the	DET
fcis-29191	179	2	results	result	NOUN
fcis-29191	179	3	of	of	ADP
fcis-29191	179	4	the	the	DET
fcis-29191	179	5	ablation	ablation	NOUN
fcis-29191	179	6	experiment	experiment	NOUN
fcis-29191	179	7	are	be	AUX
fcis-29191	179	8	shown	show	VERB
fcis-29191	179	9	in	in	ADP
fcis-29191	179	10	table	table	NOUN
fcis-29191	179	11	3	3	NUM
fcis-29191	179	12	.	.	PUNCT
fcis-29191	179	13	from	from	ADP
fcis-29191	179	14	the	the	DET
fcis-29191	179	15	results	result	NOUN
fcis-29191	179	16	in	in	ADP
fcis-29191	179	17	table	table	NOUN
fcis-29191	179	18	3	3	NUM
fcis-29191	179	19	,	,	PUNCT
fcis-29191	179	20	it	it	PRON
fcis-29191	179	21	can	can	AUX
fcis-29191	179	22	be	be	AUX
fcis-29191	179	23	seen	see	VERB
fcis-29191	179	24	that	that	SCONJ
fcis-29191	179	25	when	when	SCONJ
fcis-29191	179	26	the	the	DET
fcis-29191	179	27	text	text	NOUN
fcis-29191	179	28	is	be	AUX
fcis-29191	179	29	not	not	PART
fcis-29191	179	30	subjected	subject	VERB
fcis-29191	179	31	to	to	ADP
fcis-29191	179	32	syntactic	syntactic	ADJ
fcis-29191	179	33	analysis	analysis	NOUN
fcis-29191	179	34	,	,	PUNCT
fcis-29191	179	35	directly	directly	ADV
fcis-29191	179	36	inputting	inputte	VERB
fcis-29191	179	37	the	the	DET
fcis-29191	179	38	text	text	NOUN
fcis-29191	179	39	into	into	ADP
fcis-29191	179	40	the	the	DET
fcis-29191	179	41	bert	bert	PROPN
fcis-29191	179	42	related	relate	VERB
fcis-29191	179	43	model	model	NOUN
fcis-29191	179	44	for	for	ADP
fcis-29191	179	45	text	text	NOUN
fcis-29191	179	46	representation	representation	NOUN
fcis-29191	179	47	and	and	CCONJ
fcis-29191	179	48	combining	combine	VERB
fcis-29191	179	49	it	it	PRON
fcis-29191	179	50	with	with	ADP
fcis-29191	179	51	gcn	gcn	NOUN
fcis-29191	179	52	training	training	NOUN
fcis-29191	179	53	results	result	NOUN
fcis-29191	179	54	in	in	ADP
fcis-29191	179	55	a	a	DET
fcis-29191	179	56	decrease	decrease	NOUN
fcis-29191	179	57	in	in	ADP
fcis-29191	179	58	the	the	DET
fcis-29191	179	59	34	34	NUM
fcis-29191	179	60	classification	classification	NOUN
fcis-29191	179	61	performance	performance	NOUN
fcis-29191	179	62	of	of	ADP
fcis-29191	179	63	online	online	ADJ
fcis-29191	179	64	violent	violent	ADJ
fcis-29191	179	65	text	text	NOUN
fcis-29191	179	66	compared	compare	VERB
fcis-29191	179	67	to	to	ADP
fcis-29191	179	68	the	the	DET
fcis-29191	179	69	complete	complete	ADJ
fcis-29191	179	70	model	model	NOUN
fcis-29191	179	71	.	.	PUNCT
fcis-29191	180	1	this	this	PRON
fcis-29191	180	2	is	be	AUX
fcis-29191	180	3	because	because	SCONJ
fcis-29191	180	4	compared	compare	VERB
fcis-29191	180	5	to	to	ADP
fcis-29191	180	6	directly	directly	ADV
fcis-29191	180	7	inputting	inputte	VERB
fcis-29191	180	8	the	the	DET
fcis-29191	180	9	text	text	NOUN
fcis-29191	180	10	into	into	ADP
fcis-29191	180	11	the	the	DET
fcis-29191	180	12	bert	bert	PROPN
fcis-29191	180	13	related	relate	VERB
fcis-29191	180	14	model	model	NOUN
fcis-29191	180	15	,	,	PUNCT
fcis-29191	180	16	sending	send	VERB
fcis-29191	180	17	the	the	DET
fcis-29191	180	18	syntactic	syntactic	ADJ
fcis-29191	180	19	analysis	analysis	NOUN
fcis-29191	180	20	results	result	VERB
fcis-29191	180	21	together	together	ADV
fcis-29191	180	22	with	with	ADP
fcis-29191	180	23	the	the	DET
fcis-29191	180	24	original	original	ADJ
fcis-29191	180	25	sentence	sentence	NOUN
fcis-29191	180	26	into	into	ADP
fcis-29191	180	27	the	the	DET
fcis-29191	180	28	bert	bert	PROPN
fcis-29191	180	29	related	relate	VERB
fcis-29191	180	30	model	model	NOUN
fcis-29191	180	31	can	can	AUX
fcis-29191	180	32	explicitly	explicitly	ADV
fcis-29191	180	33	express	express	VERB
fcis-29191	180	34	the	the	DET
fcis-29191	180	35	sentence	sentence	NOUN
fcis-29191	180	36	structure	structure	NOUN
fcis-29191	180	37	and	and	CCONJ
fcis-29191	180	38	dependency	dependency	NOUN
fcis-29191	180	39	relationships	relationship	NOUN
fcis-29191	180	40	,	,	PUNCT
fcis-29191	180	41	making	make	VERB
fcis-29191	180	42	it	it	PRON
fcis-29191	180	43	easier	easy	ADJ
fcis-29191	180	44	for	for	SCONJ
fcis-29191	180	45	the	the	DET
fcis-29191	180	46	model	model	NOUN
fcis-29191	180	47	to	to	PART
fcis-29191	180	48	understand	understand	VERB
fcis-29191	180	49	the	the	DET
fcis-29191	180	50	logical	logical	ADJ
fcis-29191	180	51	relationships	relationship	NOUN
fcis-29191	180	52	of	of	ADP
fcis-29191	180	53	the	the	DET
fcis-29191	180	54	sentence	sentence	NOUN
fcis-29191	180	55	.	.	PUNCT
fcis-29191	181	1	by	by	ADP
fcis-29191	181	2	explicitly	explicitly	ADV
fcis-29191	181	3	annotating	annotate	VERB
fcis-29191	181	4	dependency	dependency	NOUN
fcis-29191	181	5	relationships	relationship	NOUN
fcis-29191	181	6	,	,	PUNCT
fcis-29191	181	7	the	the	DET
fcis-29191	181	8	model	model	NOUN
fcis-29191	181	9	can	can	AUX
fcis-29191	181	10	better	well	ADV
fcis-29191	181	11	capture	capture	VERB
fcis-29191	181	12	fine	fine	ADV
fcis-29191	181	13	-	-	PUNCT
fcis-29191	181	14	grained	grain	VERB
fcis-29191	181	15	features	feature	NOUN
fcis-29191	181	16	in	in	ADP
fcis-29191	181	17	sentences	sentence	NOUN
fcis-29191	181	18	,	,	PUNCT
fcis-29191	181	19	thereby	thereby	ADV
fcis-29191	181	20	improving	improve	VERB
fcis-29191	181	21	classification	classification	NOUN
fcis-29191	181	22	performance	performance	NOUN
fcis-29191	181	23	.	.	PUNCT
fcis-29191	182	1	in	in	ADP
fcis-29191	182	2	addition	addition	NOUN
fcis-29191	182	3	,	,	PUNCT
fcis-29191	182	4	the	the	DET
fcis-29191	182	5	annotation	annotation	NOUN
fcis-29191	182	6	of	of	ADP
fcis-29191	182	7	syntactic	syntactic	ADJ
fcis-29191	182	8	dependency	dependency	NOUN
fcis-29191	182	9	relationships	relationship	NOUN
fcis-29191	182	10	can	can	AUX
fcis-29191	182	11	help	help	VERB
fcis-29191	182	12	the	the	DET
fcis-29191	182	13	model	model	NOUN
fcis-29191	182	14	eliminate	eliminate	VERB
fcis-29191	182	15	ambiguity	ambiguity	NOUN
fcis-29191	182	16	in	in	ADP
fcis-29191	182	17	sentences	sentence	NOUN
fcis-29191	182	18	,	,	PUNCT
fcis-29191	182	19	and	and	CCONJ
fcis-29191	182	20	the	the	DET
fcis-29191	182	21	model	model	NOUN
fcis-29191	182	22	can	can	AUX
fcis-29191	182	23	more	more	ADV
fcis-29191	182	24	clearly	clearly	ADV
fcis-29191	182	25	know	know	VERB
fcis-29191	182	26	which	which	DET
fcis-29191	182	27	words	word	NOUN
fcis-29191	182	28	are	be	AUX
fcis-29191	182	29	interdependent	interdependent	ADJ
fcis-29191	182	30	.	.	PUNCT
fcis-29191	183	1	in	in	ADP
fcis-29191	183	2	addition	addition	NOUN
fcis-29191	183	3	,	,	PUNCT
fcis-29191	183	4	it	it	PRON
fcis-29191	183	5	can	can	AUX
fcis-29191	183	6	be	be	AUX
fcis-29191	183	7	seen	see	VERB
fcis-29191	183	8	from	from	ADP
fcis-29191	183	9	the	the	DET
fcis-29191	183	10	table	table	NOUN
fcis-29191	183	11	that	that	SCONJ
fcis-29191	183	12	when	when	SCONJ
fcis-29191	183	13	the	the	DET
fcis-29191	183	14	robert	robert	PROPN
fcis-29191	183	15	module	module	PROPN
fcis-29191	183	16	or	or	CCONJ
fcis-29191	183	17	gcn	gcn	NOUN
fcis-29191	183	18	module	module	NOUN
fcis-29191	183	19	is	be	AUX
fcis-29191	183	20	removed	remove	VERB
fcis-29191	183	21	for	for	ADP
fcis-29191	183	22	training	training	NOUN
fcis-29191	183	23	,	,	PUNCT
fcis-29191	183	24	the	the	DET
fcis-29191	183	25	text	text	NOUN
fcis-29191	183	26	classification	classification	NOUN
fcis-29191	183	27	performance	performance	NOUN
fcis-29191	183	28	decreases	decrease	VERB
fcis-29191	183	29	.	.	PUNCT
fcis-29191	184	1	due	due	ADP
fcis-29191	184	2	to	to	ADP
fcis-29191	184	3	the	the	DET
fcis-29191	184	4	limited	limited	ADJ
fcis-29191	184	5	number	number	NOUN
fcis-29191	184	6	of	of	ADP
fcis-29191	184	7	nodes	node	NOUN
fcis-29191	184	8	and	and	CCONJ
fcis-29191	184	9	edges	edge	NOUN
fcis-29191	184	10	for	for	ADP
fcis-29191	184	11	shorter	short	ADJ
fcis-29191	184	12	texts	text	NOUN
fcis-29191	184	13	and	and	CCONJ
fcis-29191	184	14	the	the	DET
fcis-29191	184	15	limited	limited	ADJ
fcis-29191	184	16	ability	ability	NOUN
fcis-29191	184	17	of	of	ADP
fcis-29191	184	18	graphic	graphic	ADJ
fcis-29191	184	19	structures	structure	NOUN
fcis-29191	184	20	,	,	PUNCT
fcis-29191	184	21	the	the	DET
fcis-29191	184	22	impact	impact	NOUN
fcis-29191	184	23	of	of	ADP
fcis-29191	184	24	removing	remove	VERB
fcis-29191	184	25	the	the	DET
fcis-29191	184	26	gcn	gcn	NOUN
fcis-29191	184	27	module	module	NOUN
fcis-29191	184	28	is	be	AUX
fcis-29191	184	29	relatively	relatively	ADV
fcis-29191	184	30	small	small	ADJ
fcis-29191	184	31	.	.	PUNCT
fcis-29191	185	1	however	however	ADV
fcis-29191	185	2	,	,	PUNCT
fcis-29191	185	3	robert	robert	PROPN
fcis-29191	185	4	,	,	PUNCT
fcis-29191	185	5	with	with	ADP
fcis-29191	185	6	its	its	PRON
fcis-29191	185	7	large	large	ADJ
fcis-29191	185	8	pre	pre	NOUN
fcis-29191	185	9	training	training	NOUN
fcis-29191	185	10	features	feature	NOUN
fcis-29191	185	11	,	,	PUNCT
fcis-29191	185	12	obtains	obtain	VERB
fcis-29191	185	13	better	well	ADJ
fcis-29191	185	14	text	text	NOUN
fcis-29191	185	15	embeddings	embedding	NOUN
fcis-29191	185	16	and	and	CCONJ
fcis-29191	185	17	helps	help	VERB
fcis-29191	185	18	the	the	DET
fcis-29191	185	19	model	model	NOUN
fcis-29191	185	20	achieve	achieve	VERB
fcis-29191	185	21	better	well	ADJ
fcis-29191	185	22	classification	classification	NOUN
fcis-29191	185	23	performance	performance	NOUN
fcis-29191	185	24	.	.	PUNCT
fcis-29191	186	1	table	table	NOUN
fcis-29191	186	2	3	3	NUM
fcis-29191	186	3	.	.	PUNCT
fcis-29191	187	1	results	result	NOUN
fcis-29191	187	2	of	of	ADP
fcis-29191	187	3	ablation	ablation	NOUN
fcis-29191	187	4	experiment	experiment	NOUN
fcis-29191	187	5	model	model	NOUN
fcis-29191	187	6	accuracy	accuracy	NOUN
fcis-29191	187	7	(	(	PUNCT
fcis-29191	187	8	%	%	INTJ
fcis-29191	187	9	)	)	PUNCT
fcis-29191	187	10	weighted	weight	VERB
fcis-29191	187	11	f1(%	f1(%	NOUN
fcis-29191	187	12	)	)	PUNCT
fcis-29191	187	13	remove	remove	VERB
fcis-29191	187	14	robert	robert	PROPN
fcis-29191	187	15	79.64	79.64	NUM
fcis-29191	187	16	79.81	79.81	NUM
fcis-29191	187	17	remove	remove	ADJ
fcis-29191	187	18	gcn	gcn	NOUN
fcis-29191	187	19	82.95	82.95	NUM
fcis-29191	187	20	82.93	82.93	NUM
fcis-29191	187	21	remove	remove	VERB
fcis-29191	187	22	syntactic	syntactic	ADJ
fcis-29191	187	23	parsing	parse	VERB
fcis-29191	187	24	82.51	82.51	NUM
fcis-29191	187	25	82.34	82.34	NUM
fcis-29191	188	1	our	our	PRON
fcis-29191	188	2	model	model	NOUN
fcis-29191	188	3	84.77	84.77	NUM
fcis-29191	188	4	84.13	84.13	NUM
fcis-29191	188	5	5	5	NUM
fcis-29191	188	6	.	.	PUNCT
fcis-29191	189	1	summary	summary	NOUN
fcis-29191	189	2	this	this	DET
fcis-29191	189	3	article	article	NOUN
fcis-29191	189	4	proposes	propose	VERB
fcis-29191	189	5	a	a	DET
fcis-29191	189	6	graph	graph	NOUN
fcis-29191	189	7	convolutional	convolutional	ADJ
fcis-29191	189	8	network	network	NOUN
fcis-29191	189	9	classification	classification	NOUN
fcis-29191	189	10	model	model	NOUN
fcis-29191	189	11	r	r	PROPN
fcis-29191	189	12	-	-	PUNCT
fcis-29191	189	13	bg	bg	NOUN
fcis-29191	189	14	for	for	ADP
fcis-29191	189	15	cyber	cyber	ADJ
fcis-29191	189	16	violence	violence	NOUN
fcis-29191	189	17	text	text	NOUN
fcis-29191	189	18	.	.	PUNCT
fcis-29191	190	1	by	by	ADP
fcis-29191	190	2	introducing	introduce	VERB
fcis-29191	190	3	dependency	dependency	NOUN
fcis-29191	190	4	relationships	relationship	NOUN
fcis-29191	190	5	,	,	PUNCT
fcis-29191	190	6	bert	bert	X
fcis-29191	190	7	related	relate	VERB
fcis-29191	190	8	models	model	NOUN
fcis-29191	190	9	can	can	AUX
fcis-29191	190	10	better	well	ADV
fcis-29191	190	11	understand	understand	VERB
fcis-29191	190	12	the	the	DET
fcis-29191	190	13	grammar	grammar	NOUN
fcis-29191	190	14	and	and	CCONJ
fcis-29191	190	15	semantic	semantic	ADJ
fcis-29191	190	16	structure	structure	NOUN
fcis-29191	190	17	in	in	ADP
fcis-29191	190	18	sentences	sentence	NOUN
fcis-29191	190	19	,	,	PUNCT
fcis-29191	190	20	thereby	thereby	ADV
fcis-29191	190	21	more	more	ADV
fcis-29191	190	22	accurately	accurately	ADV
fcis-29191	190	23	representing	represent	VERB
fcis-29191	190	24	features	feature	NOUN
fcis-29191	190	25	in	in	ADP
fcis-29191	190	26	cyber	cyber	ADJ
fcis-29191	190	27	violence	violence	NOUN
fcis-29191	190	28	text	text	NOUN
fcis-29191	190	29	.	.	PUNCT
fcis-29191	191	1	the	the	DET
fcis-29191	191	2	proposed	propose	VERB
fcis-29191	191	3	model	model	NOUN
fcis-29191	191	4	combines	combine	VERB
fcis-29191	191	5	the	the	DET
fcis-29191	191	6	advantages	advantage	NOUN
fcis-29191	191	7	of	of	ADP
fcis-29191	191	8	bert	bert	NOUN
fcis-29191	191	9	related	relate	VERB
fcis-29191	191	10	models	model	NOUN
fcis-29191	191	11	and	and	CCONJ
fcis-29191	191	12	graph	graph	VERB
fcis-29191	191	13	convolutional	convolutional	ADJ
fcis-29191	191	14	networks	network	NOUN
fcis-29191	191	15	(	(	PUNCT
fcis-29191	191	16	gcn	gcn	NOUN
fcis-29191	191	17	)	)	PUNCT
fcis-29191	191	18	for	for	ADP
fcis-29191	191	19	joint	joint	ADJ
fcis-29191	191	20	training	training	NOUN
fcis-29191	191	21	,	,	PUNCT
fcis-29191	191	22	significantly	significantly	ADV
fcis-29191	191	23	improving	improve	VERB
fcis-29191	191	24	the	the	DET
fcis-29191	191	25	classification	classification	NOUN
fcis-29191	191	26	performance	performance	NOUN
fcis-29191	191	27	of	of	ADP
fcis-29191	191	28	cyber	cyber	ADJ
fcis-29191	191	29	violence	violence	NOUN
fcis-29191	191	30	text	text	NOUN
fcis-29191	191	31	.	.	PUNCT
fcis-29191	192	1	at	at	ADP
fcis-29191	192	2	the	the	DET
fcis-29191	192	3	same	same	ADJ
fcis-29191	192	4	time	time	NOUN
fcis-29191	192	5	,	,	PUNCT
fcis-29191	192	6	the	the	DET
fcis-29191	192	7	model	model	NOUN
fcis-29191	192	8	can	can	AUX
fcis-29191	192	9	effectively	effectively	ADV
fcis-29191	192	10	identify	identify	VERB
fcis-29191	192	11	and	and	CCONJ
fcis-29191	192	12	distinguish	distinguish	VERB
fcis-29191	192	13	different	different	ADJ
fcis-29191	192	14	categories	category	NOUN
fcis-29191	192	15	of	of	ADP
fcis-29191	192	16	cyber	cyber	NOUN
fcis-29191	192	17	violence	violence	NOUN
fcis-29191	192	18	text	text	NOUN
fcis-29191	192	19	,	,	PUNCT
fcis-29191	192	20	providing	provide	VERB
fcis-29191	192	21	scientific	scientific	ADJ
fcis-29191	192	22	basis	basis	NOUN
fcis-29191	192	23	for	for	ADP
fcis-29191	192	24	formulating	formulate	VERB
fcis-29191	192	25	response	response	NOUN
fcis-29191	192	26	strategies	strategy	NOUN
fcis-29191	192	27	and	and	CCONJ
fcis-29191	192	28	helping	help	VERB
fcis-29191	192	29	to	to	PART
fcis-29191	192	30	maintain	maintain	VERB
fcis-29191	192	31	a	a	DET
fcis-29191	192	32	healthy	healthy	ADJ
fcis-29191	192	33	online	online	ADJ
fcis-29191	192	34	environment	environment	NOUN
fcis-29191	192	35	.	.	PUNCT
fcis-29191	193	1	references	reference	NOUN
fcis-29191	193	2	[	[	X
fcis-29191	193	3	1	1	NUM
fcis-29191	193	4	]	]	PUNCT
fcis-29191	193	5	drucker	drucker	NOUN
fcis-29191	193	6	h	h	PROPN
fcis-29191	193	7	,	,	PUNCT
fcis-29191	193	8	wu	wu	PROPN
fcis-29191	193	9	d	d	PROPN
fcis-29191	193	10	,	,	PUNCT
fcis-29191	193	11	vapnik	vapnik	X
fcis-29191	193	12	v	v	NUM
fcis-29191	193	13	n.	n.	NOUN
fcis-29191	193	14	support	support	NOUN
fcis-29191	193	15	vector	vector	NOUN
fcis-29191	193	16	machines	machine	NOUN
fcis-29191	193	17	for	for	ADP
fcis-29191	193	18	spam	spam	NOUN
fcis-29191	193	19	categorization[j	categorization[j	PROPN
fcis-29191	193	20	]	]	PUNCT
fcis-29191	193	21	.	.	PUNCT
fcis-29191	194	1	ieee	ieee	NOUN
fcis-29191	194	2	transactions	transaction	NOUN
fcis-29191	194	3	on	on	ADP
fcis-29191	194	4	neural	neural	ADJ
fcis-29191	194	5	networks	network	NOUN
fcis-29191	194	6	,	,	PUNCT
fcis-29191	194	7	1999	1999	NUM
fcis-29191	194	8	,	,	PUNCT
fcis-29191	194	9	10(5	10(5	NUM
fcis-29191	194	10	):	):	PUNCT
fcis-29191	194	11	1048	1048	NUM
fcis-29191	194	12	-	-	SYM
fcis-29191	194	13	1054	1054	NUM
fcis-29191	194	14	.	.	PUNCT
fcis-29191	195	1	[	[	X
fcis-29191	195	2	2	2	NUM
fcis-29191	195	3	]	]	PUNCT
fcis-29191	195	4	mccallum	mccallum	PROPN
fcis-29191	195	5	a	a	PROPN
fcis-29191	195	6	,	,	PUNCT
fcis-29191	195	7	nigam	nigam	PROPN
fcis-29191	195	8	k.	k.	PROPN
fcis-29191	196	1	a	a	DET
fcis-29191	196	2	comparison	comparison	NOUN
fcis-29191	196	3	of	of	ADP
fcis-29191	196	4	event	event	NOUN
fcis-29191	196	5	models	model	NOUN
fcis-29191	196	6	for	for	ADP
fcis-29191	196	7	naive	naive	ADJ
fcis-29191	196	8	bayes	bayes	NOUN
fcis-29191	196	9	text	text	NOUN
fcis-29191	196	10	classification[c]//aaai-98	classification[c]//aaai-98	ADJ
fcis-29191	196	11	workshop	workshop	NOUN
fcis-29191	196	12	on	on	ADP
fcis-29191	196	13	learning	learn	VERB
fcis-29191	196	14	for	for	ADP
fcis-29191	196	15	text	text	NOUN
fcis-29191	196	16	categorization	categorization	NOUN
fcis-29191	196	17	.	.	PUNCT
fcis-29191	197	1	1998	1998	NUM
fcis-29191	197	2	,	,	PUNCT
fcis-29191	197	3	752(1	752(1	NUM
fcis-29191	197	4	):	):	PUNCT
fcis-29191	197	5	41	41	NUM
fcis-29191	197	6	-	-	SYM
fcis-29191	197	7	48	48	NUM
fcis-29191	197	8	.	.	PUNCT
fcis-29191	198	1	[	[	X
fcis-29191	198	2	3	3	X
fcis-29191	198	3	]	]	PUNCT
fcis-29191	198	4	cover	cover	NOUN
fcis-29191	198	5	t	t	PROPN
fcis-29191	198	6	,	,	PUNCT
fcis-29191	198	7	hart	hart	PROPN
fcis-29191	198	8	p.	p.	PROPN
fcis-29191	198	9	nearest	near	ADJ
fcis-29191	198	10	neighbor	neighbor	NOUN
fcis-29191	198	11	pattern	pattern	NOUN
fcis-29191	198	12	classification[j	classification[j	PROPN
fcis-29191	198	13	]	]	PUNCT
fcis-29191	198	14	.	.	PUNCT
fcis-29191	199	1	ieee	ieee	NOUN
fcis-29191	199	2	transactions	transaction	NOUN
fcis-29191	199	3	on	on	ADP
fcis-29191	199	4	information	information	NOUN
fcis-29191	199	5	theory	theory	NOUN
fcis-29191	199	6	,	,	PUNCT
fcis-29191	199	7	1967	1967	NUM
fcis-29191	199	8	,	,	PUNCT
fcis-29191	199	9	13(1	13(1	NUM
fcis-29191	199	10	):	):	PUNCT
fcis-29191	199	11	21	21	NUM
fcis-29191	199	12	-	-	SYM
fcis-29191	199	13	27	27	NUM
fcis-29191	199	14	.	.	PUNCT
fcis-29191	200	1	[	[	X
fcis-29191	200	2	4	4	NUM
fcis-29191	200	3	]	]	X
fcis-29191	200	4	kotsiantis	kotsiantis	PROPN
fcis-29191	200	5	s	s	PROPN
fcis-29191	200	6	b.	b.	PROPN
fcis-29191	200	7	decision	decision	NOUN
fcis-29191	200	8	trees	tree	NOUN
fcis-29191	200	9	:	:	PUNCT
fcis-29191	200	10	a	a	DET
fcis-29191	200	11	recent	recent	ADJ
fcis-29191	200	12	overview[j	overview[j	PROPN
fcis-29191	200	13	]	]	PUNCT
fcis-29191	200	14	.	.	PUNCT
fcis-29191	201	1	artificial	artificial	ADJ
fcis-29191	201	2	intelligence	intelligence	NOUN
fcis-29191	201	3	review	review	NOUN
fcis-29191	201	4	,	,	PUNCT
fcis-29191	201	5	2013	2013	NUM
fcis-29191	201	6	,	,	PUNCT
fcis-29191	201	7	39	39	NUM
fcis-29191	201	8	:	:	PUNCT
fcis-29191	201	9	261	261	NUM
fcis-29191	201	10	-	-	SYM
fcis-29191	201	11	283	283	NUM
fcis-29191	201	12	.	.	PUNCT
fcis-29191	202	1	[	[	X
fcis-29191	202	2	5	5	X
fcis-29191	202	3	]	]	PUNCT
fcis-29191	202	4	kim	kim	PROPN
fcis-29191	202	5	y.	y.	PROPN
fcis-29191	202	6	convolutional	convolutional	ADJ
fcis-29191	202	7	neural	neural	ADJ
fcis-29191	202	8	networks	network	NOUN
fcis-29191	202	9	for	for	ADP
fcis-29191	202	10	sentence	sentence	NOUN
fcis-29191	202	11	classification[c]//proceedings	classification[c]//proceeding	NOUN
fcis-29191	202	12	of	of	ADP
fcis-29191	202	13	conference	conference	NOUN
fcis-29191	202	14	on	on	ADP
fcis-29191	202	15	empirical	empirical	ADJ
fcis-29191	202	16	methods	method	NOUN
fcis-29191	202	17	in	in	ADP
fcis-29191	202	18	natural	natural	ADJ
fcis-29191	202	19	language	language	NOUN
fcis-29191	202	20	processing	processing	NOUN
fcis-29191	202	21	,	,	PUNCT
fcis-29191	202	22	2014	2014	NUM
fcis-29191	202	23	.	.	PUNCT
fcis-29191	203	1	[	[	X
fcis-29191	203	2	6	6	NUM
fcis-29191	203	3	]	]	PUNCT
fcis-29191	203	4	tang	tang	X
fcis-29191	203	5	d	d	NOUN
fcis-29191	203	6	,	,	PUNCT
fcis-29191	203	7	qin	qin	PROPN
fcis-29191	203	8	b	b	PROPN
fcis-29191	203	9	,	,	PUNCT
fcis-29191	203	10	liu	liu	PROPN
fcis-29191	203	11	t.	t.	PROPN
fcis-29191	203	12	document	document	NOUN
fcis-29191	203	13	modeling	modeling	NOUN
fcis-29191	203	14	with	with	ADP
fcis-29191	203	15	gated	gate	VERB
fcis-29191	203	16	recurrent	recurrent	ADJ
fcis-29191	203	17	neural	neural	ADJ
fcis-29191	203	18	network	network	NOUN
fcis-29191	203	19	for	for	ADP
fcis-29191	203	20	sentiment	sentiment	NOUN
fcis-29191	203	21	classification[c]//	classification[c]//	PROPN
fcis-29191	203	22	proceedings	proceeding	NOUN
fcis-29191	203	23	of	of	ADP
fcis-29191	203	24	the	the	DET
fcis-29191	203	25	2015	2015	NUM
fcis-29191	203	26	conference	conference	NOUN
fcis-29191	203	27	on	on	ADP
fcis-29191	203	28	empirical	empirical	ADJ
fcis-29191	203	29	methods	method	NOUN
fcis-29191	203	30	in	in	ADP
fcis-29191	203	31	natural	natural	ADJ
fcis-29191	203	32	language	language	NOUN
fcis-29191	203	33	processing	processing	NOUN
fcis-29191	203	34	.	.	PUNCT
fcis-29191	204	1	2015	2015	NUM
fcis-29191	204	2	:	:	PUNCT
fcis-29191	204	3	1422	1422	NUM
fcis-29191	204	4	-	-	SYM
fcis-29191	204	5	1432	1432	NUM
fcis-29191	204	6	.	.	PUNCT
fcis-29191	205	1	[	[	X
fcis-29191	205	2	7	7	X
fcis-29191	205	3	]	]	X
fcis-29191	205	4	vaswani	vaswani	ADJ
fcis-29191	205	5	a.	a.	NOUN
fcis-29191	205	6	attention	attention	NOUN
fcis-29191	205	7	is	be	AUX
fcis-29191	205	8	all	all	PRON
fcis-29191	205	9	you	you	PRON
fcis-29191	205	10	need[j	need[j	VERB
fcis-29191	205	11	]	]	PUNCT
fcis-29191	205	12	.	.	PUNCT
fcis-29191	206	1	advances	advance	NOUN
fcis-29191	206	2	in	in	ADP
fcis-29191	206	3	neural	neural	ADJ
fcis-29191	206	4	information	information	NOUN
fcis-29191	206	5	processing	processing	NOUN
fcis-29191	206	6	systems	system	NOUN
fcis-29191	206	7	,	,	PUNCT
fcis-29191	206	8	2017	2017	NUM
fcis-29191	206	9	.	.	PUNCT
fcis-29191	207	1	[	[	X
fcis-29191	207	2	8	8	NUM
fcis-29191	207	3	]	]	X
fcis-29191	207	4	devlin	devlin	PROPN
fcis-29191	207	5	j.	j.	PROPN
fcis-29191	207	6	bert	bert	PROPN
fcis-29191	207	7	:	:	PUNCT
fcis-29191	207	8	pre	pre	ADJ
fcis-29191	207	9	-	-	NOUN
fcis-29191	207	10	training	training	NOUN
fcis-29191	207	11	of	of	ADP
fcis-29191	207	12	deep	deep	ADJ
fcis-29191	207	13	bidirectional	bidirectional	ADJ
fcis-29191	207	14	transformers	transformer	NOUN
fcis-29191	207	15	for	for	ADP
fcis-29191	207	16	language	language	NOUN
fcis-29191	207	17	understanding[j	understanding[j	NOUN
fcis-29191	207	18	]	]	PUNCT
fcis-29191	207	19	.	.	PUNCT
fcis-29191	208	1	arxiv	arxiv	PROPN
fcis-29191	208	2	preprint	preprint	PROPN
fcis-29191	208	3	arxiv	arxiv	PROPN
fcis-29191	208	4	:	:	PUNCT
fcis-29191	208	5	1810	1810	NUM
fcis-29191	208	6	.	.	PUNCT
fcis-29191	208	7	04805	04805	NUM
fcis-29191	208	8	,	,	PUNCT
fcis-29191	208	9	2018	2018	NUM
fcis-29191	208	10	.	.	PUNCT
fcis-29191	209	1	[	[	X
fcis-29191	209	2	9	9	NUM
fcis-29191	209	3	]	]	X
fcis-29191	209	4	wu	wu	PROPN
fcis-29191	209	5	z	z	PROPN
fcis-29191	209	6	,	,	PUNCT
fcis-29191	209	7	pan	pan	PROPN
fcis-29191	209	8	s	s	PROPN
fcis-29191	209	9	,	,	PUNCT
fcis-29191	209	10	chen	chen	PROPN
fcis-29191	209	11	f	f	PROPN
fcis-29191	209	12	,	,	PUNCT
fcis-29191	209	13	et	et	PROPN
fcis-29191	209	14	al	al	PROPN
fcis-29191	209	15	.	.	PUNCT
fcis-29191	210	1	a	a	DET
fcis-29191	210	2	comprehensive	comprehensive	ADJ
fcis-29191	210	3	survey	survey	NOUN
fcis-29191	210	4	on	on	ADP
fcis-29191	210	5	graph	graph	NOUN
fcis-29191	210	6	neural	neural	ADJ
fcis-29191	210	7	networks[j	networks[j	PROPN
fcis-29191	210	8	]	]	X
fcis-29191	210	9	.	.	PUNCT
fcis-29191	211	1	ieee	ieee	NOUN
fcis-29191	211	2	transactions	transaction	NOUN
fcis-29191	211	3	on	on	ADP
fcis-29191	211	4	neural	neural	ADJ
fcis-29191	211	5	networks	network	NOUN
fcis-29191	211	6	and	and	CCONJ
fcis-29191	211	7	learning	learning	NOUN
fcis-29191	211	8	systems	system	NOUN
fcis-29191	211	9	,	,	PUNCT
fcis-29191	211	10	2020	2020	NUM
fcis-29191	211	11	,	,	PUNCT
fcis-29191	211	12	32(1	32(1	NUM
fcis-29191	211	13	):	):	PUNCT
fcis-29191	211	14	4	4	NUM
fcis-29191	211	15	-	-	SYM
fcis-29191	211	16	24	24	NUM
fcis-29191	211	17	.	.	PUNCT
fcis-29191	212	1	[	[	X
fcis-29191	212	2	10	10	NUM
fcis-29191	212	3	]	]	PUNCT
fcis-29191	212	4	kipf	kipf	PROPN
fcis-29191	212	5	t	t	PROPN
fcis-29191	212	6	n	n	CCONJ
fcis-29191	212	7	,	,	PUNCT
fcis-29191	212	8	welling	well	VERB
fcis-29191	212	9	m.	m.	NOUN
fcis-29191	212	10	semi	semi	ADJ
fcis-29191	212	11	-	-	ADJ
fcis-29191	212	12	supervised	supervised	ADJ
fcis-29191	212	13	classification	classification	NOUN
fcis-29191	212	14	with	with	ADP
fcis-29191	212	15	graph	graph	NOUN
fcis-29191	212	16	convolutional	convolutional	ADJ
fcis-29191	212	17	networks[j	networks[j	PROPN
fcis-29191	212	18	]	]	X
fcis-29191	212	19	.	.	PUNCT
fcis-29191	213	1	arxiv	arxiv	PROPN
fcis-29191	213	2	preprint	preprint	PROPN
fcis-29191	213	3	arxiv	arxiv	PROPN
fcis-29191	213	4	:	:	PUNCT
fcis-29191	213	5	1609	1609	NUM
fcis-29191	213	6	.	.	PUNCT
fcis-29191	213	7	02907	02907	NUM
fcis-29191	213	8	,	,	PUNCT
fcis-29191	213	9	2016	2016	NUM
fcis-29191	213	10	.	.	PUNCT
fcis-29191	214	1	[	[	X
fcis-29191	214	2	11	11	NUM
fcis-29191	214	3	]	]	X
fcis-29191	214	4	yao	yao	PROPN
fcis-29191	214	5	l	l	PROPN
fcis-29191	214	6	,	,	PUNCT
fcis-29191	214	7	mao	mao	X
fcis-29191	214	8	c	c	X
fcis-29191	214	9	,	,	PUNCT
fcis-29191	214	10	luo	luo	PROPN
fcis-29191	214	11	y.	y.	PROPN
fcis-29191	214	12	graph	graph	VERB
fcis-29191	214	13	convolutional	convolutional	ADJ
fcis-29191	214	14	networks	network	NOUN
fcis-29191	214	15	for	for	ADP
fcis-29191	214	16	text	text	NOUN
fcis-29191	214	17	classification[c]//proceedings	classification[c]//proceeding	NOUN
fcis-29191	214	18	of	of	ADP
fcis-29191	214	19	the	the	DET
fcis-29191	214	20	aaai	aaai	PROPN
fcis-29191	214	21	conference	conference	NOUN
fcis-29191	214	22	on	on	ADP
fcis-29191	214	23	artificial	artificial	ADJ
fcis-29191	214	24	intelligence	intelligence	NOUN
fcis-29191	214	25	.	.	PUNCT
fcis-29191	215	1	2019	2019	NUM
fcis-29191	215	2	,	,	PUNCT
fcis-29191	215	3	33(01	33(01	NUM
fcis-29191	215	4	):	):	PUNCT
fcis-29191	215	5	7370	7370	NUM
fcis-29191	215	6	-	-	SYM
fcis-29191	215	7	7377	7377	NUM
fcis-29191	215	8	.	.	PUNCT
fcis-29191	216	1	[	[	X
fcis-29191	216	2	12	12	NUM
fcis-29191	216	3	]	]	PUNCT
fcis-29191	216	4	velickovic	velickovic	NOUN
fcis-29191	216	5	p	p	X
fcis-29191	216	6	,	,	PUNCT
fcis-29191	216	7	cucurull	cucurull	NOUN
fcis-29191	216	8	g	g	NOUN
fcis-29191	216	9	,	,	PUNCT
fcis-29191	216	10	casanova	casanova	X
fcis-29191	216	11	a	a	X
fcis-29191	216	12	,	,	PUNCT
fcis-29191	216	13	et	et	PROPN
fcis-29191	216	14	al	al	PROPN
fcis-29191	216	15	.	.	PUNCT
fcis-29191	217	1	graph	graph	NOUN
fcis-29191	217	2	attention	attention	NOUN
fcis-29191	217	3	networks	network	NOUN
fcis-29191	218	1	[	[	X
fcis-29191	218	2	j	j	X
fcis-29191	218	3	]	]	X
fcis-29191	218	4	.	.	PUNCT
fcis-29191	219	1	stat	stat	PROPN
fcis-29191	219	2	,	,	PUNCT
fcis-29191	219	3	2017	2017	NUM
fcis-29191	219	4	,	,	PUNCT
fcis-29191	219	5	1050(20	1050(20	NUM
fcis-29191	219	6	):	):	PUNCT
fcis-29191	219	7	10	10	NUM
fcis-29191	219	8	-	-	SYM
fcis-29191	219	9	48550	48550	NUM
fcis-29191	219	10	.	.	PUNCT
fcis-29191	220	1	[	[	X
fcis-29191	220	2	13	13	NUM
fcis-29191	220	3	]	]	X
fcis-29191	220	4	linmei	linmei	PROPN
fcis-29191	220	5	h	h	NOUN
fcis-29191	220	6	,	,	PUNCT
fcis-29191	220	7	yang	yang	PROPN
fcis-29191	220	8	t	t	PROPN
fcis-29191	220	9	,	,	PUNCT
fcis-29191	220	10	shi	shi	PROPN
fcis-29191	220	11	c	c	PROPN
fcis-29191	220	12	,	,	PUNCT
fcis-29191	220	13	et	et	PROPN
fcis-29191	220	14	al	al	PROPN
fcis-29191	220	15	.	.	PROPN
fcis-29191	220	16	heterogeneous	heterogeneous	ADJ
fcis-29191	220	17	graph	graph	NOUN
fcis-29191	220	18	attention	attention	NOUN
fcis-29191	220	19	networks	network	NOUN
fcis-29191	220	20	for	for	ADP
fcis-29191	220	21	semi	semi	ADJ
fcis-29191	220	22	-	-	ADJ
fcis-29191	220	23	supervised	supervised	ADJ
fcis-29191	220	24	short	short	ADJ
fcis-29191	220	25	text	text	NOUN
fcis-29191	220	26	classification[c]//	classification[c]//	PROPN
fcis-29191	220	27	proceedings	proceeding	NOUN
fcis-29191	220	28	of	of	ADP
fcis-29191	220	29	the	the	DET
fcis-29191	220	30	2019	2019	NUM
fcis-29191	220	31	conference	conference	NOUN
fcis-29191	220	32	on	on	ADP
fcis-29191	220	33	empirical	empirical	ADJ
fcis-29191	220	34	methods	method	NOUN
fcis-29191	220	35	in	in	ADP
fcis-29191	220	36	natural	natural	ADJ
fcis-29191	220	37	language	language	NOUN
fcis-29191	220	38	processing	processing	NOUN
fcis-29191	220	39	and	and	CCONJ
fcis-29191	220	40	the	the	DET
fcis-29191	220	41	9th	9th	ADJ
fcis-29191	220	42	international	international	ADJ
fcis-29191	220	43	joint	joint	ADJ
fcis-29191	220	44	conference	conference	NOUN
fcis-29191	220	45	on	on	ADP
fcis-29191	220	46	natural	natural	ADJ
fcis-29191	220	47	language	language	NOUN
fcis-29191	220	48	processing	processing	NOUN
fcis-29191	220	49	(	(	PUNCT
fcis-29191	220	50	emnlp	emnlp	NOUN
fcis-29191	220	51	-	-	PUNCT
fcis-29191	220	52	ijcnlp	ijcnlp	NOUN
fcis-29191	220	53	)	)	PUNCT
fcis-29191	220	54	.	.	PUNCT
fcis-29191	221	1	2019	2019	NUM
fcis-29191	221	2	:	:	PUNCT
fcis-29191	221	3	4821	4821	NUM
fcis-29191	221	4	-	-	SYM
fcis-29191	221	5	4830	4830	NUM
fcis-29191	221	6	.	.	PUNCT
fcis-29191	222	1	[	[	X
fcis-29191	222	2	14	14	NUM
fcis-29191	222	3	]	]	X
fcis-29191	222	4	lin	lin	PROPN
fcis-29191	222	5	y	y	PROPN
fcis-29191	222	6	,	,	PUNCT
fcis-29191	222	7	meng	meng	PROPN
fcis-29191	222	8	y	y	PROPN
fcis-29191	222	9	,	,	PUNCT
fcis-29191	222	10	sun	sun	PROPN
fcis-29191	222	11	x	x	AUX
fcis-29191	222	12	,	,	PUNCT
fcis-29191	222	13	et	et	PROPN
fcis-29191	222	14	al	al	PROPN
fcis-29191	222	15	.	.	PROPN
fcis-29191	222	16	bertgcn	bertgcn	PROPN
fcis-29191	222	17	:	:	PUNCT
fcis-29191	222	18	transductive	transductive	ADJ
fcis-29191	222	19	text	text	NOUN
fcis-29191	222	20	classification	classification	NOUN
fcis-29191	222	21	by	by	ADP
fcis-29191	222	22	combining	combine	VERB
fcis-29191	222	23	gcn	gcn	NOUN
fcis-29191	222	24	and	and	CCONJ
fcis-29191	222	25	bert[j	bert[j	NOUN
fcis-29191	222	26	]	]	PUNCT
fcis-29191	222	27	.	.	PUNCT
fcis-29191	223	1	arxiv	arxiv	PROPN
fcis-29191	223	2	preprint	preprint	PROPN
fcis-29191	223	3	arxiv	arxiv	PROPN
fcis-29191	223	4	:	:	PUNCT
fcis-29191	223	5	2105.05727	2105.05727	NUM
fcis-29191	223	6	,	,	PUNCT
fcis-29191	223	7	2021	2021	NUM
fcis-29191	223	8	.	.	PUNCT
fcis-29191	224	1	[	[	X
fcis-29191	224	2	15	15	NUM
fcis-29191	224	3	]	]	PUNCT
fcis-29191	224	4	dinakar	dinakar	NOUN
fcis-29191	224	5	k	k	PROPN
fcis-29191	224	6	,	,	PUNCT
fcis-29191	224	7	reichart	reichart	PROPN
fcis-29191	224	8	r	r	NOUN
fcis-29191	224	9	,	,	PUNCT
fcis-29191	224	10	lieberman	lieberman	PROPN
fcis-29191	224	11	h.	h.	PROPN
fcis-29191	224	12	modeling	model	VERB
fcis-29191	224	13	the	the	DET
fcis-29191	224	14	detection	detection	NOUN
fcis-29191	224	15	of	of	ADP
fcis-29191	224	16	textual	textual	ADJ
fcis-29191	224	17	cyberbullying[c	cyberbullying[c	NOUN
fcis-29191	224	18	]	]	PUNCT
fcis-29191	224	19	//proceedings	//proceeding	NOUN
fcis-29191	224	20	of	of	ADP
fcis-29191	224	21	the	the	DET
fcis-29191	224	22	international	international	ADJ
fcis-29191	224	23	aaai	aaai	PROPN
fcis-29191	224	24	conference	conference	PROPN
fcis-29191	224	25	on	on	ADP
fcis-29191	224	26	web	web	NOUN
fcis-29191	224	27	and	and	CCONJ
fcis-29191	224	28	social	social	ADJ
fcis-29191	224	29	media	medium	NOUN
fcis-29191	224	30	.	.	PUNCT
fcis-29191	225	1	2011	2011	NUM
fcis-29191	225	2	,	,	PUNCT
fcis-29191	225	3	5(3	5(3	NUM
fcis-29191	225	4	):	):	PUNCT
fcis-29191	225	5	1117	1117	NUM
fcis-29191	225	6	.	.	PUNCT
fcis-29191	226	1	[	[	X
fcis-29191	226	2	16	16	NUM
fcis-29191	226	3	]	]	PUNCT
fcis-29191	226	4	chavan	chavan	NOUN
fcis-29191	226	5	v	v	ADP
fcis-29191	226	6	s	s	PROPN
fcis-29191	226	7	,	,	PUNCT
fcis-29191	226	8	shylaja	shylaja	PROPN
fcis-29191	226	9	s	s	PART
fcis-29191	226	10	s.	s.	PROPN
fcis-29191	226	11	machine	machine	PROPN
fcis-29191	226	12	learning	learn	VERB
fcis-29191	226	13	approach	approach	NOUN
fcis-29191	226	14	for	for	ADP
fcis-29191	226	15	detection	detection	NOUN
fcis-29191	226	16	of	of	ADP
fcis-29191	226	17	cyber	cyber	NOUN
fcis-29191	226	18	-	-	PUNCT
fcis-29191	226	19	aggressive	aggressive	ADJ
fcis-29191	226	20	comments	comment	NOUN
fcis-29191	226	21	by	by	ADP
fcis-29191	226	22	peers	peer	NOUN
fcis-29191	226	23	on	on	ADP
fcis-29191	226	24	social	social	ADJ
fcis-29191	226	25	media	medium	NOUN
fcis-29191	226	26	network[c]//2015	network[c]//2015	PART
fcis-29191	226	27	international	international	ADJ
fcis-29191	226	28	conference	conference	NOUN
fcis-29191	226	29	on	on	ADP
fcis-29191	226	30	advances	advance	NOUN
fcis-29191	226	31	in	in	ADP
fcis-29191	226	32	computing	computing	NOUN
fcis-29191	226	33	,	,	PUNCT
fcis-29191	226	34	communications	communication	NOUN
fcis-29191	226	35	and	and	CCONJ
fcis-29191	226	36	informatics	informatic	NOUN
fcis-29191	226	37	(	(	PUNCT
fcis-29191	226	38	icacci	icacci	PROPN
fcis-29191	226	39	)	)	PUNCT
fcis-29191	226	40	.	.	PUNCT
fcis-29191	227	1	ieee	ieee	PROPN
fcis-29191	227	2	,	,	PUNCT
fcis-29191	227	3	2015	2015	NUM
fcis-29191	227	4	:	:	PUNCT
fcis-29191	227	5	2354	2354	NUM
fcis-29191	227	6	-	-	SYM
fcis-29191	227	7	2358	2358	NUM
fcis-29191	227	8	.	.	PUNCT
fcis-29191	228	1	[	[	X
fcis-29191	228	2	17	17	NUM
fcis-29191	228	3	]	]	X
fcis-29191	228	4	squicciarini	squicciarini	NOUN
fcis-29191	228	5	a	a	PRON
fcis-29191	228	6	,	,	PUNCT
fcis-29191	228	7	rajtmajer	rajtmajer	NOUN
fcis-29191	228	8	s	s	PROPN
fcis-29191	228	9	,	,	PUNCT
fcis-29191	228	10	liu	liu	PROPN
fcis-29191	228	11	y	y	PROPN
fcis-29191	228	12	,	,	PUNCT
fcis-29191	228	13	et	et	PROPN
fcis-29191	228	14	al	al	PROPN
fcis-29191	228	15	.	.	PUNCT
fcis-29191	228	16	identification	identification	NOUN
fcis-29191	228	17	and	and	CCONJ
fcis-29191	228	18	characterization	characterization	NOUN
fcis-29191	228	19	of	of	ADP
fcis-29191	228	20	cyberbullying	cyberbullye	VERB
fcis-29191	228	21	dynamics	dynamic	NOUN
fcis-29191	228	22	in	in	ADP
fcis-29191	228	23	an	an	DET
fcis-29191	228	24	online	online	ADJ
fcis-29191	228	25	social	social	ADJ
fcis-29191	228	26	network	network	NOUN
fcis-29191	228	27	[	[	X
fcis-29191	228	28	c]//proceedings	c]//proceeding	NOUN
fcis-29191	228	29	of	of	ADP
fcis-29191	228	30	the	the	DET
fcis-29191	228	31	2015	2015	NUM
fcis-29191	228	32	ieee	ieee	NOUN
fcis-29191	228	33	/	/	SYM
fcis-29191	228	34	acm	acm	PROPN
fcis-29191	228	35	international	international	ADJ
fcis-29191	228	36	conference	conference	NOUN
fcis-29191	228	37	on	on	ADP
fcis-29191	228	38	advances	advance	NOUN
fcis-29191	228	39	in	in	ADP
fcis-29191	228	40	social	social	ADJ
fcis-29191	228	41	networks	network	NOUN
fcis-29191	228	42	analysis	analysis	NOUN
fcis-29191	228	43	and	and	CCONJ
fcis-29191	228	44	mining	mining	NOUN
fcis-29191	228	45	2015	2015	NUM
fcis-29191	228	46	.	.	PUNCT
fcis-29191	229	1	2015	2015	NUM
fcis-29191	229	2	:	:	PUNCT
fcis-29191	229	3	280	280	NUM
fcis-29191	229	4	-	-	SYM
fcis-29191	229	5	285	285	NUM
fcis-29191	229	6	.	.	PUNCT
fcis-29191	230	1	[	[	X
fcis-29191	230	2	18	18	NUM
fcis-29191	230	3	]	]	PUNCT
fcis-29191	230	4	agrawal	agrawal	PROPN
fcis-29191	230	5	s	s	PROPN
fcis-29191	230	6	,	,	PUNCT
fcis-29191	230	7	awekar	awekar	ADJ
fcis-29191	230	8	a.	a.	NOUN
fcis-29191	230	9	deep	deep	PROPN
fcis-29191	230	10	learning	learn	VERB
fcis-29191	230	11	for	for	ADP
fcis-29191	230	12	detecting	detect	VERB
fcis-29191	230	13	cyberbullying	cyberbullye	VERB
fcis-29191	230	14	across	across	ADP
fcis-29191	230	15	multiple	multiple	ADJ
fcis-29191	230	16	social	social	ADJ
fcis-29191	230	17	media	medium	NOUN
fcis-29191	230	18	platforms[c	platforms[c	PROPN
fcis-29191	230	19	]	]	PUNCT
fcis-29191	230	20	//	//	X
fcis-29191	230	21	european	european	PROPN
fcis-29191	230	22	conference	conference	PROPN
fcis-29191	230	23	on	on	ADP
fcis-29191	230	24	information	information	NOUN
fcis-29191	230	25	retrieval	retrieval	NOUN
fcis-29191	230	26	.	.	PUNCT
fcis-29191	231	1	cham	cham	PROPN
fcis-29191	231	2	:	:	PUNCT
fcis-29191	231	3	springer	springer	NOUN
fcis-29191	231	4	international	international	ADJ
fcis-29191	231	5	publishing	publishing	NOUN
fcis-29191	231	6	,	,	PUNCT
fcis-29191	231	7	2018	2018	NUM
fcis-29191	231	8	:	:	PUNCT
fcis-29191	231	9	141	141	NUM
fcis-29191	231	10	-	-	SYM
fcis-29191	231	11	153	153	NUM
fcis-29191	231	12	.	.	PUNCT
fcis-29191	232	1	[	[	X
fcis-29191	232	2	19	19	NUM
fcis-29191	232	3	]	]	PUNCT
fcis-29191	232	4	daraghmi	daraghmi	NOUN
fcis-29191	232	5	e	e	PROPN
fcis-29191	232	6	y	y	PROPN
fcis-29191	232	7	,	,	PUNCT
fcis-29191	232	8	qadan	qadan	PROPN
fcis-29191	232	9	s	s	PROPN
fcis-29191	232	10	,	,	PUNCT
fcis-29191	232	11	daraghmi	daraghmi	VERB
fcis-29191	232	12	y	y	PROPN
fcis-29191	232	13	,	,	PUNCT
fcis-29191	232	14	et	et	PROPN
fcis-29191	232	15	al	al	PROPN
fcis-29191	232	16	.	.	PROPN
fcis-29191	233	1	from	from	ADP
fcis-29191	233	2	text	text	NOUN
fcis-29191	233	3	to	to	ADP
fcis-29191	233	4	insight	insight	NOUN
fcis-29191	233	5	:	:	PUNCT
fcis-29191	233	6	an	an	DET
fcis-29191	233	7	integrated	integrated	ADJ
fcis-29191	233	8	cnn	cnn	PROPN
fcis-29191	233	9	-	-	PUNCT
fcis-29191	233	10	bilstm	bilstm	NOUN
fcis-29191	233	11	-	-	PUNCT
fcis-29191	233	12	gru	gru	NOUN
fcis-29191	233	13	model	model	NOUN
fcis-29191	233	14	for	for	ADP
fcis-29191	233	15	arabic	arabic	ADJ
fcis-29191	233	16	cyberbullying	cyberbullye	VERB
fcis-29191	233	17	detection[j	detection[j	PROPN
fcis-29191	233	18	]	]	PUNCT
fcis-29191	233	19	.	.	PUNCT
fcis-29191	234	1	ieee	ieee	NOUN
fcis-29191	234	2	access	access	NOUN
fcis-29191	234	3	,	,	PUNCT
fcis-29191	234	4	2024	2024	NUM
fcis-29191	234	5	.	.	PUNCT
fcis-29191	235	1	[	[	X
fcis-29191	235	2	20	20	NUM
fcis-29191	235	3	]	]	X
fcis-29191	235	4	huang	huang	PROPN
fcis-29191	235	5	r.	r.	PROPN
fcis-29191	235	6	implementation	implementation	NOUN
fcis-29191	235	7	of	of	ADP
fcis-29191	235	8	a	a	DET
fcis-29191	235	9	network	network	NOUN
fcis-29191	235	10	violent	violent	ADJ
fcis-29191	235	11	language	language	NOUN
fcis-29191	235	12	detection	detection	NOUN
fcis-29191	235	13	system[d	system[d	NOUN
fcis-29191	235	14	]	]	PUNCT
fcis-29191	235	15	.	.	PUNCT
fcis-29191	236	1	wuhan	wuhan	PROPN
fcis-29191	236	2	:	:	PUNCT
fcis-29191	236	3	school	school	NOUN
fcis-29191	236	4	of	of	ADP
fcis-29191	236	5	information	information	NOUN
fcis-29191	236	6	and	and	CCONJ
fcis-29191	236	7	communication	communication	NOUN
fcis-29191	236	8	engineering	engineering	NOUN
fcis-29191	236	9	,	,	PUNCT
fcis-29191	236	10	huazhong	huazhong	PROPN
fcis-29191	236	11	university	university	PROPN
fcis-29191	236	12	of	of	ADP
fcis-29191	236	13	science	science	NOUN
fcis-29191	236	14	and	and	CCONJ
fcis-29191	236	15	technology,2016	technology,2016	PRON
fcis-29191	236	16	.	.	PUNCT
fcis-29191	237	1	[	[	X
fcis-29191	237	2	21	21	NUM
fcis-29191	237	3	]	]	X
fcis-29191	237	4	qiang	qiang	PROPN
fcis-29191	237	5	l.	l.	PROPN
fcis-29191	237	6	research	research	PROPN
fcis-29191	237	7	on	on	ADP
fcis-29191	237	8	cyberbullying	cyberbullye	VERB
fcis-29191	237	9	language	language	NOUN
fcis-29191	237	10	based	base	VERB
fcis-29191	237	11	on	on	ADP
fcis-29191	237	12	social	social	ADJ
fcis-29191	237	13	networks[d	networks[d	NOUN
fcis-29191	237	14	]	]	PUNCT
fcis-29191	237	15	.	.	PUNCT
fcis-29191	238	1	taiyuan	taiyuan	NOUN
fcis-29191	238	2	:	:	PUNCT
fcis-29191	238	3	north	north	PROPN
fcis-29191	238	4	university	university	PROPN
fcis-29191	238	5	of	of	ADP
fcis-29191	238	6	china,2020	china,2020	PROPN
fcis-29191	239	1	[	[	X
fcis-29191	239	2	22	22	NUM
fcis-29191	239	3	]	]	X
fcis-29191	239	4	wang	wang	PROPN
fcis-29191	239	5	r	r	PROPN
fcis-29191	239	6	,	,	PUNCT
fcis-29191	239	7	zhang	zhang	PROPN
fcis-29191	239	8	p	p	X
fcis-29191	239	9	,	,	PUNCT
fcis-29191	239	10	gu	gu	NOUN
fcis-29191	239	11	n	n	CCONJ
fcis-29191	239	12	,	,	PUNCT
fcis-29191	239	13	et	et	PROPN
fcis-29191	239	14	al	al	PROPN
fcis-29191	239	15	.	.	PUNCT
fcis-29191	240	1	intelligent	intelligent	ADJ
fcis-29191	240	2	intervention	intervention	NOUN
fcis-29191	240	3	methods	method	NOUN
fcis-29191	240	4	for	for	ADP
fcis-29191	240	5	cyberbullying	cyberbullye	VERB
fcis-29191	240	6	in	in	ADP
fcis-29191	240	7	online	online	ADJ
fcis-29191	240	8	communities	community	NOUN
fcis-29191	240	9	:	:	PUNCT
fcis-29191	241	1	landscapes	landscape	NOUN
fcis-29191	241	2	,	,	PUNCT
fcis-29191	241	3	challenges	challenge	NOUN
fcis-29191	241	4	and	and	CCONJ
fcis-29191	241	5	prospects[j	prospects[j	NOUN
fcis-29191	241	6	]	]	PUNCT
fcis-29191	241	7	.	.	PUNCT
fcis-29191	242	1	journal	journal	PROPN
fcis-29191	242	2	of	of	ADP
fcis-29191	242	3	chinese	chinese	PROPN
fcis-29191	242	4	computer	computer	NOUN
fcis-29191	242	5	systems,2024,45(10):2312	systems,2024,45(10):2312	PROPN
fcis-29191	242	6	-	-	PUNCT
fcis-29191	242	7	2326	2326	NUM
fcis-29191	242	8	.	.	PUNCT
fcis-29191	243	1	[	[	X
fcis-29191	243	2	23	23	NUM
fcis-29191	243	3	]	]	X
fcis-29191	243	4	wang	wang	PROPN
fcis-29191	243	5	j	j	PROPN
fcis-29191	243	6	,	,	PUNCT
fcis-29191	243	7	fu	fu	PROPN
fcis-29191	243	8	k	k	PROPN
fcis-29191	243	9	,	,	PUNCT
fcis-29191	243	10	lu	lu	PROPN
fcis-29191	243	11	c	c	PROPN
fcis-29191	243	12	t.	t.	PROPN
fcis-29191	243	13	sosnet	sosnet	PROPN
fcis-29191	243	14	:	:	PUNCT
fcis-29191	243	15	a	a	DET
fcis-29191	243	16	graph	graph	NOUN
fcis-29191	243	17	convolutional	convolutional	ADJ
fcis-29191	243	18	network	network	NOUN
fcis-29191	243	19	approach	approach	NOUN
fcis-29191	243	20	to	to	ADP
fcis-29191	243	21	fine	fine	ADV
fcis-29191	243	22	-	-	PUNCT
fcis-29191	243	23	grained	grain	VERB
fcis-29191	243	24	cyberbullying	cyberbullying	NOUN
fcis-29191	243	25	detection[c	detection[c	X
fcis-29191	243	26	]	]	PUNCT
fcis-29191	243	27	//2020	//2020	PUNCT
fcis-29191	244	1	ieee	ieee	PROPN
fcis-29191	244	2	international	international	PROPN
fcis-29191	244	3	conference	conference	NOUN
fcis-29191	244	4	on	on	ADP
fcis-29191	244	5	big	big	ADJ
fcis-29191	244	6	data	datum	NOUN
fcis-29191	244	7	(	(	PUNCT
fcis-29191	244	8	big	big	ADJ
fcis-29191	244	9	data	datum	NOUN
fcis-29191	244	10	)	)	PUNCT
fcis-29191	244	11	.	.	PUNCT
fcis-29191	245	1	ieee	ieee	PROPN
fcis-29191	245	2	,	,	PUNCT
fcis-29191	245	3	2020	2020	NUM
fcis-29191	245	4	:	:	PUNCT
fcis-29191	245	5	1699	1699	NUM
fcis-29191	245	6	-	-	SYM
fcis-29191	245	7	1708	1708	NUM
fcis-29191	245	8	.	.	PUNCT