id	sid	tid	token	lemma	pos
futech-324	1	1	s.	s.	PROPN
futech-324	1	2	sushma	sushma	PROPN
futech-324	1	3	et	et	PROPN
futech-324	1	4	al	al	PROPN
futech-324	1	5	.	.	PROPN
futech-324	1	6	/future	/future	PUNCT
futech-324	1	7	technology	technology	PROPN
futech-324	1	8	august	august	PROPN
futech-324	1	9	2025|	2025|	NUM
futech-324	1	10	volume	volume	NOUN
futech-324	1	11	04	04	NUM
futech-324	2	1	|	|	NOUN
futech-324	2	2	issue	issue	NOUN
futech-324	2	3	03	03	NUM
futech-324	3	1	|	|	ADV
futech-324	3	2	pages	page	VERB
futech-324	3	3	76	76	NUM
futech-324	3	4	-	-	SYM
futech-324	3	5	84	84	NUM
futech-324	3	6	76	76	NUM
futech-324	3	7	article	article	NOUN
futech-324	3	8	enhanced	enhance	VERB
futech-324	3	9	toxic	toxic	ADJ
futech-324	3	10	comment	comment	NOUN
futech-324	3	11	detection	detection	NOUN
futech-324	3	12	model	model	NOUN
futech-324	3	13	through	through	ADP
futech-324	3	14	deep	deep	ADJ
futech-324	3	15	learning	learning	NOUN
futech-324	3	16	models	model	NOUN
futech-324	3	17	using	use	VERB
futech-324	3	18	word	word	NOUN
futech-324	3	19	embeddings	embedding	NOUN
futech-324	3	20	and	and	CCONJ
futech-324	3	21	transformer	transformer	NOUN
futech-324	3	22	architectures	architecture	NOUN
futech-324	3	23	s.	s.	PROPN
futech-324	3	24	sushma1,2	sushma1,2	PROPN
futech-324	3	25	*	*	PROPN
futech-324	3	26	,	,	PUNCT
futech-324	3	27	sasmita	sasmita	PROPN
futech-324	3	28	kumari	kumari	PROPN
futech-324	3	29	nayak1	nayak1	PROPN
futech-324	3	30	,	,	PUNCT
futech-324	3	31	m.	m.	NOUN
futech-324	3	32	vamsi	vamsi	PROPN
futech-324	3	33	krishna3	krishna3	PROPN
futech-324	3	34	1department	1department	NUM
futech-324	3	35	of	of	ADP
futech-324	3	36	cse	cse	PROPN
futech-324	3	37	,	,	PUNCT
futech-324	3	38	centurion	centurion	NOUN
futech-324	3	39	university	university	PROPN
futech-324	3	40	of	of	ADP
futech-324	3	41	technology	technology	NOUN
futech-324	3	42	and	and	CCONJ
futech-324	3	43	management	management	NOUN
futech-324	3	44	,	,	PUNCT
futech-324	3	45	bhubaneswar	bhubaneswar	NOUN
futech-324	3	46	,	,	PUNCT
futech-324	3	47	odisha	odisha	PROPN
futech-324	3	48	,	,	PUNCT
futech-324	3	49	india	india	PROPN
futech-324	3	50	2aditya	2aditya	PROPN
futech-324	3	51	university	university	NOUN
futech-324	3	52	,	,	PUNCT
futech-324	3	53	surampalem	surampalem	NOUN
futech-324	3	54	,	,	PUNCT
futech-324	3	55	india	india	PROPN
futech-324	3	56	3department	3department	PROPN
futech-324	3	57	of	of	ADP
futech-324	3	58	computer	computer	NOUN
futech-324	3	59	applications	application	NOUN
futech-324	3	60	,	,	PUNCT
futech-324	3	61	aditya	aditya	PROPN
futech-324	3	62	university	university	PROPN
futech-324	3	63	,	,	PUNCT
futech-324	3	64	surampalem	surampalem	NOUN
futech-324	3	65	,	,	PUNCT
futech-324	3	66	india	india	PROPN
futech-324	3	67	a	a	DET
futech-324	3	68	r	r	NOUN
futech-324	3	69	t	t	NOUN
futech-324	4	1	i	i	PRON
futech-324	4	2	c	c	NOUN
futech-324	4	3	l	l	NOUN
futech-324	4	4	e	e	NOUN
futech-324	5	1	i	i	PRON
futech-324	5	2	n	n	VERB
futech-324	5	3	f	f	NOUN
futech-324	5	4	o	o	NOUN
futech-324	5	5	article	article	NOUN
futech-324	5	6	history	history	NOUN
futech-324	5	7	:	:	PUNCT
futech-324	5	8	received	receive	VERB
futech-324	5	9	08	08	NUM
futech-324	5	10	april	april	PROPN
futech-324	5	11	2025	2025	NUM
futech-324	5	12	received	receive	VERB
futech-324	5	13	in	in	ADP
futech-324	5	14	revised	revise	VERB
futech-324	5	15	form	form	NOUN
futech-324	5	16	21	21	NUM
futech-324	5	17	may	may	AUX
futech-324	5	18	2025	2025	NUM
futech-324	5	19	accepted	accept	VERB
futech-324	5	20	31	31	NUM
futech-324	5	21	may	may	NOUN
futech-324	5	22	2025	2025	NUM
futech-324	5	23	keywords	keyword	NOUN
futech-324	5	24	:	:	PUNCT
futech-324	5	25	toxic	toxic	ADJ
futech-324	5	26	comment	comment	NOUN
futech-324	5	27	classification	classification	NOUN
futech-324	5	28	,	,	PUNCT
futech-324	5	29	word	word	NOUN
futech-324	5	30	embeddings	embedding	NOUN
futech-324	5	31	,	,	PUNCT
futech-324	5	32	ensemble	ensemble	ADJ
futech-324	5	33	modeling	modeling	NOUN
futech-324	5	34	*	*	PUNCT
futech-324	5	35	corresponding	correspond	VERB
futech-324	5	36	author	author	NOUN
futech-324	5	37	email	email	NOUN
futech-324	5	38	address	address	NOUN
futech-324	5	39	:	:	PUNCT
futech-324	5	40	sushma.cse2@gmail.com	sushma.cse2@gmail.com	X
futech-324	5	41	doi	doi	NOUN
futech-324	5	42	:	:	PUNCT
futech-324	5	43	10.55670	10.55670	NUM
futech-324	5	44	/	/	SYM
futech-324	5	45	fpll.futech.4.3.8	fpll.futech.4.3.8	NOUN
futech-324	5	46	a	a	DET
futech-324	5	47	b	b	PROPN
futech-324	5	48	s	s	ADP
futech-324	5	49	t	t	PROPN
futech-324	5	50	r	r	NOUN
futech-324	5	51	a	a	DET
futech-324	5	52	c	c	NOUN
futech-324	5	53	t	t	NOUN
futech-324	5	54	the	the	DET
futech-324	5	55	proliferation	proliferation	NOUN
futech-324	5	56	of	of	ADP
futech-324	5	57	harmful	harmful	ADJ
futech-324	5	58	and	and	CCONJ
futech-324	5	59	toxic	toxic	ADJ
futech-324	5	60	comments	comment	NOUN
futech-324	5	61	on	on	ADP
futech-324	5	62	social	social	ADJ
futech-324	5	63	media	medium	NOUN
futech-324	5	64	platforms	platform	NOUN
futech-324	5	65	necessitates	necessitate	VERB
futech-324	5	66	the	the	DET
futech-324	5	67	development	development	NOUN
futech-324	5	68	of	of	ADP
futech-324	5	69	robust	robust	ADJ
futech-324	5	70	methods	method	NOUN
futech-324	5	71	for	for	ADP
futech-324	5	72	automatically	automatically	ADV
futech-324	5	73	detecting	detect	VERB
futech-324	5	74	and	and	CCONJ
futech-324	5	75	classifying	classify	VERB
futech-324	5	76	such	such	ADJ
futech-324	5	77	content	content	NOUN
futech-324	5	78	.	.	PUNCT
futech-324	6	1	this	this	DET
futech-324	6	2	paper	paper	NOUN
futech-324	6	3	investigates	investigate	VERB
futech-324	6	4	the	the	DET
futech-324	6	5	application	application	NOUN
futech-324	6	6	of	of	ADP
futech-324	6	7	natural	natural	ADJ
futech-324	6	8	language	language	NOUN
futech-324	6	9	processing	processing	NOUN
futech-324	6	10	(	(	PUNCT
futech-324	6	11	nlp	nlp	NOUN
futech-324	6	12	)	)	PUNCT
futech-324	6	13	and	and	CCONJ
futech-324	6	14	ml	ml	ADP
futech-324	6	15	techniques	technique	NOUN
futech-324	6	16	for	for	ADP
futech-324	6	17	toxic	toxic	ADJ
futech-324	6	18	comment	comment	NOUN
futech-324	6	19	classification	classification	NOUN
futech-324	6	20	using	use	VERB
futech-324	6	21	the	the	DET
futech-324	6	22	jigsaw	jigsaw	NOUN
futech-324	6	23	toxic	toxic	ADJ
futech-324	6	24	comment	comment	NOUN
futech-324	6	25	dataset	dataset	NOUN
futech-324	6	26	.	.	PUNCT
futech-324	7	1	several	several	ADJ
futech-324	7	2	deep	deep	ADJ
futech-324	7	3	learning	learning	NOUN
futech-324	7	4	models	model	NOUN
futech-324	7	5	,	,	PUNCT
futech-324	7	6	including	include	VERB
futech-324	7	7	recurrent	recurrent	ADJ
futech-324	7	8	neural	neural	ADJ
futech-324	7	9	networks	network	NOUN
futech-324	7	10	(	(	PUNCT
futech-324	7	11	rnn	rnn	PROPN
futech-324	7	12	,	,	PUNCT
futech-324	7	13	lstm	lstm	NOUN
futech-324	7	14	,	,	PUNCT
futech-324	7	15	and	and	CCONJ
futech-324	7	16	gru	gru	PROPN
futech-324	7	17	)	)	PUNCT
futech-324	7	18	,	,	PUNCT
futech-324	7	19	are	be	AUX
futech-324	7	20	evaluated	evaluate	VERB
futech-324	7	21	in	in	ADP
futech-324	7	22	combination	combination	NOUN
futech-324	7	23	with	with	ADP
futech-324	7	24	feature	feature	NOUN
futech-324	7	25	extraction	extraction	NOUN
futech-324	7	26	methods	method	NOUN
futech-324	7	27	such	such	ADJ
futech-324	7	28	as	as	ADP
futech-324	7	29	tf	tf	PROPN
futech-324	7	30	-	-	PUNCT
futech-324	7	31	idf	idf	PROPN
futech-324	7	32	,	,	PUNCT
futech-324	7	33	word2vec	word2vec	X
futech-324	7	34	,	,	PUNCT
futech-324	7	35	and	and	CCONJ
futech-324	7	36	bert	bert	PROPN
futech-324	7	37	embeddings	embedding	NOUN
futech-324	7	38	.	.	PUNCT
futech-324	8	1	the	the	DET
futech-324	8	2	text	text	NOUN
futech-324	8	3	data	data	NOUN
futech-324	8	4	is	be	AUX
futech-324	8	5	pre	pre	ADJ
futech-324	8	6	-	-	VERB
futech-324	8	7	processed	processed	ADJ
futech-324	8	8	using	use	VERB
futech-324	8	9	both	both	PRON
futech-324	8	10	word2vec	word2vec	X
futech-324	8	11	and	and	CCONJ
futech-324	8	12	tfidf	tfidf	NOUN
futech-324	8	13	techniques	technique	NOUN
futech-324	8	14	for	for	ADP
futech-324	8	15	feature	feature	NOUN
futech-324	8	16	extraction	extraction	NOUN
futech-324	8	17	.	.	PUNCT
futech-324	9	1	rather	rather	ADV
futech-324	9	2	than	than	ADP
futech-324	9	3	implementing	implement	VERB
futech-324	9	4	a	a	DET
futech-324	9	5	combined	combine	VERB
futech-324	9	6	ensemble	ensemble	ADJ
futech-324	9	7	output	output	NOUN
futech-324	9	8	,	,	PUNCT
futech-324	9	9	the	the	DET
futech-324	9	10	study	study	NOUN
futech-324	9	11	conducts	conduct	VERB
futech-324	9	12	a	a	DET
futech-324	9	13	comparative	comparative	ADJ
futech-324	9	14	evaluation	evaluation	NOUN
futech-324	9	15	of	of	ADP
futech-324	9	16	modelembedding	modelembedde	VERB
futech-324	9	17	combinations	combination	NOUN
futech-324	9	18	to	to	PART
futech-324	9	19	determine	determine	VERB
futech-324	9	20	the	the	DET
futech-324	9	21	most	most	ADV
futech-324	9	22	effective	effective	ADJ
futech-324	9	23	pairings	pairing	NOUN
futech-324	9	24	.	.	PUNCT
futech-324	10	1	results	result	NOUN
futech-324	10	2	indicate	indicate	VERB
futech-324	10	3	that	that	SCONJ
futech-324	10	4	integrating	integrate	VERB
futech-324	10	5	bert	bert	NOUN
futech-324	10	6	with	with	ADP
futech-324	10	7	traditional	traditional	ADJ
futech-324	10	8	models	model	NOUN
futech-324	10	9	(	(	PUNCT
futech-324	10	10	rnn+bert	rnn+bert	NOUN
futech-324	10	11	,	,	PUNCT
futech-324	10	12	lstm+bert	lstm+bert	PROPN
futech-324	10	13	,	,	PUNCT
futech-324	10	14	gru+bert	gru+bert	PROPN
futech-324	10	15	)	)	PUNCT
futech-324	10	16	leads	lead	VERB
futech-324	10	17	to	to	ADP
futech-324	10	18	significant	significant	ADJ
futech-324	10	19	improvements	improvement	NOUN
futech-324	10	20	in	in	ADP
futech-324	10	21	classification	classification	NOUN
futech-324	10	22	accuracy	accuracy	NOUN
futech-324	10	23	,	,	PUNCT
futech-324	10	24	precision	precision	NOUN
futech-324	10	25	,	,	PUNCT
futech-324	10	26	recall	recall	NOUN
futech-324	10	27	,	,	PUNCT
futech-324	10	28	and	and	CCONJ
futech-324	10	29	f1	f1	NOUN
futech-324	10	30	-	-	PUNCT
futech-324	10	31	score	score	NOUN
futech-324	10	32	,	,	PUNCT
futech-324	10	33	demonstrating	demonstrate	VERB
futech-324	10	34	the	the	DET
futech-324	10	35	effectiveness	effectiveness	NOUN
futech-324	10	36	of	of	ADP
futech-324	10	37	bert	bert	PROPN
futech-324	10	38	embeddings	embedding	NOUN
futech-324	10	39	in	in	ADP
futech-324	10	40	capturing	capture	VERB
futech-324	10	41	nuanced	nuanced	ADJ
futech-324	10	42	text	text	NOUN
futech-324	10	43	features	feature	NOUN
futech-324	10	44	.	.	PUNCT
futech-324	11	1	among	among	ADP
futech-324	11	2	all	all	DET
futech-324	11	3	configurations	configuration	NOUN
futech-324	11	4	,	,	PUNCT
futech-324	11	5	lstm	lstm	NOUN
futech-324	11	6	combined	combine	VERB
futech-324	11	7	with	with	ADP
futech-324	11	8	word2vec	word2vec	PRON
futech-324	11	9	and	and	CCONJ
futech-324	11	10	lstm	lstm	NOUN
futech-324	11	11	with	with	ADP
futech-324	11	12	bert	bert	PROPN
futech-324	11	13	yielded	yield	VERB
futech-324	11	14	the	the	DET
futech-324	11	15	highest	high	ADJ
futech-324	11	16	performance	performance	NOUN
futech-324	11	17	.	.	PUNCT
futech-324	12	1	this	this	DET
futech-324	12	2	comparative	comparative	ADJ
futech-324	12	3	approach	approach	NOUN
futech-324	12	4	highlights	highlight	VERB
futech-324	12	5	the	the	DET
futech-324	12	6	potential	potential	NOUN
futech-324	12	7	of	of	ADP
futech-324	12	8	combining	combine	VERB
futech-324	12	9	classical	classical	ADJ
futech-324	12	10	recurrent	recurrent	NOUN
futech-324	12	11	models	model	NOUN
futech-324	12	12	with	with	ADP
futech-324	12	13	transformer	transformer	NOUN
futech-324	12	14	-	-	PUNCT
futech-324	12	15	based	base	VERB
futech-324	12	16	embeddings	embedding	NOUN
futech-324	12	17	as	as	ADP
futech-324	12	18	a	a	DET
futech-324	12	19	promising	promising	ADJ
futech-324	12	20	direction	direction	NOUN
futech-324	12	21	for	for	ADP
futech-324	12	22	detecting	detect	VERB
futech-324	12	23	toxic	toxic	ADJ
futech-324	12	24	comments	comment	NOUN
futech-324	12	25	.	.	PUNCT
futech-324	13	1	the	the	DET
futech-324	13	2	findings	finding	NOUN
futech-324	13	3	of	of	ADP
futech-324	13	4	this	this	DET
futech-324	13	5	work	work	NOUN
futech-324	13	6	provide	provide	VERB
futech-324	13	7	valuable	valuable	ADJ
futech-324	13	8	insights	insight	NOUN
futech-324	13	9	into	into	ADP
futech-324	13	10	leveraging	leverage	VERB
futech-324	13	11	deep	deep	ADJ
futech-324	13	12	learning	learning	NOUN
futech-324	13	13	techniques	technique	NOUN
futech-324	13	14	for	for	ADP
futech-324	13	15	toxic	toxic	ADJ
futech-324	13	16	comment	comment	NOUN
futech-324	13	17	detection	detection	NOUN
futech-324	13	18	,	,	PUNCT
futech-324	13	19	suggesting	suggest	VERB
futech-324	13	20	future	future	ADJ
futech-324	13	21	directions	direction	NOUN
futech-324	13	22	for	for	ADP
futech-324	13	23	refining	refine	VERB
futech-324	13	24	such	such	ADJ
futech-324	13	25	models	model	NOUN
futech-324	13	26	in	in	ADP
futech-324	13	27	real	real	ADJ
futech-324	13	28	-	-	PUNCT
futech-324	13	29	world	world	NOUN
futech-324	13	30	applications	application	NOUN
futech-324	13	31	.	.	PUNCT
futech-324	14	1	1	1	X
futech-324	14	2	.	.	X
futech-324	14	3	introduction	introduction	NOUN
futech-324	14	4	there	there	PRON
futech-324	14	5	is	be	VERB
futech-324	14	6	an	an	DET
futech-324	14	7	increasing	increase	VERB
futech-324	14	8	amount	amount	NOUN
futech-324	14	9	of	of	ADP
futech-324	14	10	harmful	harmful	ADJ
futech-324	14	11	and	and	CCONJ
futech-324	14	12	toxic	toxic	ADJ
futech-324	14	13	comments	comment	NOUN
futech-324	14	14	that	that	PRON
futech-324	14	15	may	may	AUX
futech-324	14	16	harm	harm	VERB
futech-324	14	17	users	user	NOUN
futech-324	14	18	'	'	PART
futech-324	14	19	experience	experience	NOUN
futech-324	14	20	,	,	PUNCT
futech-324	14	21	with	with	ADP
futech-324	14	22	the	the	DET
futech-324	14	23	exponentially	exponentially	ADV
futech-324	14	24	growing	grow	VERB
futech-324	14	25	user	user	NOUN
futech-324	14	26	-	-	PUNCT
futech-324	14	27	generated	generate	VERB
futech-324	14	28	content	content	NOUN
futech-324	14	29	available	available	ADJ
futech-324	14	30	on	on	ADP
futech-324	14	31	social	social	ADJ
futech-324	14	32	media	medium	NOUN
futech-324	14	33	platforms	platform	NOUN
futech-324	14	34	.	.	PUNCT
futech-324	15	1	hence	hence	ADV
futech-324	15	2	,	,	PUNCT
futech-324	15	3	the	the	DET
futech-324	15	4	need	need	NOUN
futech-324	15	5	for	for	ADP
futech-324	15	6	an	an	DET
futech-324	15	7	automated	automate	VERB
futech-324	15	8	system	system	NOUN
futech-324	15	9	that	that	PRON
futech-324	15	10	will	will	AUX
futech-324	15	11	detect	detect	VERB
futech-324	15	12	and	and	CCONJ
futech-324	15	13	filter	filter	VERB
futech-324	15	14	out	out	ADP
futech-324	15	15	such	such	ADJ
futech-324	15	16	rotten	rotten	ADJ
futech-324	15	17	content	content	NOUN
futech-324	15	18	has	have	AUX
futech-324	15	19	become	become	VERB
futech-324	15	20	crucial	crucial	ADJ
futech-324	15	21	to	to	ADP
futech-324	15	22	maintaining	maintain	VERB
futech-324	15	23	a	a	DET
futech-324	15	24	positive	positive	ADJ
futech-324	15	25	and	and	CCONJ
futech-324	15	26	healthy	healthy	ADJ
futech-324	15	27	environment	environment	NOUN
futech-324	15	28	on	on	ADP
futech-324	15	29	the	the	DET
futech-324	15	30	internet	internet	NOUN
futech-324	15	31	.	.	PUNCT
futech-324	16	1	this	this	PRON
futech-324	16	2	motivates	motivate	VERB
futech-324	16	3	the	the	DET
futech-324	16	4	adoption	adoption	NOUN
futech-324	16	5	of	of	ADP
futech-324	16	6	more	more	ADV
futech-324	16	7	advanced	advanced	ADJ
futech-324	16	8	ml	ml	NOUN
futech-324	16	9	and	and	CCONJ
futech-324	16	10	nlp	nlp	NOUN
futech-324	16	11	techniques	technique	NOUN
futech-324	16	12	,	,	PUNCT
futech-324	16	13	as	as	SCONJ
futech-324	16	14	traditional	traditional	ADJ
futech-324	16	15	content	content	NOUN
futech-324	16	16	moderation	moderation	NOUN
futech-324	16	17	systems	system	NOUN
futech-324	16	18	have	have	VERB
futech-324	16	19	high	high	ADJ
futech-324	16	20	false	false	ADJ
futech-324	16	21	-	-	PUNCT
futech-324	16	22	positive	positive	ADJ
futech-324	16	23	and	and	CCONJ
futech-324	16	24	falsenegative	falsenegative	ADJ
futech-324	16	25	rates	rate	NOUN
futech-324	16	26	and	and	CCONJ
futech-324	16	27	often	often	ADV
futech-324	16	28	fail	fail	VERB
futech-324	16	29	to	to	PART
futech-324	16	30	scale	scale	VERB
futech-324	16	31	well	well	ADV
futech-324	16	32	to	to	ADP
futech-324	16	33	larger	large	ADJ
futech-324	16	34	volumes	volume	NOUN
futech-324	16	35	of	of	ADP
futech-324	16	36	data	datum	NOUN
futech-324	16	37	.	.	PUNCT
futech-324	17	1	detecting	detect	VERB
futech-324	17	2	toxic	toxic	ADJ
futech-324	17	3	comments	comment	NOUN
futech-324	17	4	has	have	AUX
futech-324	17	5	been	be	AUX
futech-324	17	6	addressed	address	VERB
futech-324	17	7	with	with	ADP
futech-324	17	8	different	different	ADJ
futech-324	17	9	approaches	approach	NOUN
futech-324	17	10	in	in	ADP
futech-324	17	11	recent	recent	ADJ
futech-324	17	12	years	year	NOUN
futech-324	17	13	.	.	PUNCT
futech-324	18	1	now	now	ADV
futech-324	18	2	,	,	PUNCT
futech-324	18	3	there	there	PRON
futech-324	18	4	are	be	VERB
futech-324	18	5	two	two	NUM
futech-324	18	6	things	thing	NOUN
futech-324	18	7	we	we	PRON
futech-324	18	8	can	can	AUX
futech-324	18	9	guess	guess	VERB
futech-324	18	10	from	from	ADP
futech-324	18	11	the	the	DET
futech-324	18	12	name	name	NOUN
futech-324	18	13	of	of	ADP
futech-324	18	14	the	the	DET
futech-324	18	15	model	model	NOUN
futech-324	18	16	above	above	ADV
futech-324	18	17	:	:	PUNCT
futech-324	18	18	one	one	NUM
futech-324	18	19	is	be	AUX
futech-324	18	20	that	that	SCONJ
futech-324	18	21	it	it	PRON
futech-324	18	22	could	could	AUX
futech-324	18	23	be	be	AUX
futech-324	18	24	any	any	DET
futech-324	18	25	classical	classical	ADJ
futech-324	18	26	ml	ml	NOUN
futech-324	18	27	algorithm	algorithm	NOUN
futech-324	18	28	,	,	PUNCT
futech-324	18	29	and	and	CCONJ
futech-324	18	30	the	the	DET
futech-324	18	31	other	other	ADJ
futech-324	18	32	is	be	AUX
futech-324	18	33	that	that	SCONJ
futech-324	18	34	it	it	PRON
futech-324	18	35	is	be	AUX
futech-324	18	36	used	use	VERB
futech-324	18	37	for	for	ADP
futech-324	18	38	binary	binary	ADJ
futech-324	18	39	classification	classification	NOUN
futech-324	18	40	.	.	PUNCT
futech-324	19	1	these	these	DET
futech-324	19	2	techniques	technique	NOUN
futech-324	19	3	work	work	VERB
futech-324	19	4	reasonably	reasonably	ADV
futech-324	19	5	well	well	ADV
futech-324	19	6	most	most	ADJ
futech-324	19	7	of	of	ADP
futech-324	19	8	the	the	DET
futech-324	19	9	time	time	NOUN
futech-324	19	10	,	,	PUNCT
futech-324	19	11	but	but	CCONJ
futech-324	19	12	do	do	AUX
futech-324	19	13	not	not	PART
futech-324	19	14	necessarily	necessarily	ADV
futech-324	19	15	capture	capture	VERB
futech-324	19	16	all	all	DET
futech-324	19	17	the	the	DET
futech-324	19	18	subtleties	subtlety	NOUN
futech-324	19	19	of	of	ADP
futech-324	19	20	language	language	NOUN
futech-324	19	21	,	,	PUNCT
futech-324	19	22	particularly	particularly	ADV
futech-324	19	23	when	when	SCONJ
futech-324	19	24	processing	process	VERB
futech-324	19	25	unstructured	unstructured	ADJ
futech-324	19	26	content	content	NOUN
futech-324	19	27	like	like	ADP
futech-324	19	28	social	social	ADJ
futech-324	19	29	media	medium	NOUN
futech-324	19	30	comments	comment	NOUN
futech-324	19	31	.	.	PUNCT
futech-324	20	1	on	on	ADP
futech-324	20	2	this	this	DET
futech-324	20	3	training	training	NOUN
futech-324	20	4	,	,	PUNCT
futech-324	20	5	it	it	PRON
futech-324	20	6	's	be	AUX
futech-324	20	7	a	a	DET
futech-324	20	8	great	great	ADJ
futech-324	20	9	performance	performance	NOUN
futech-324	20	10	in	in	ADP
futech-324	20	11	sequence	sequence	NOUN
futech-324	20	12	data	datum	NOUN
futech-324	20	13	modeling	modeling	NOUN
futech-324	20	14	,	,	PUNCT
futech-324	20	15	especially	especially	ADV
futech-324	20	16	with	with	ADP
futech-324	20	17	dl	dl	PROPN
futech-324	20	18	models	model	NOUN
futech-324	20	19	like	like	ADP
futech-324	20	20	rnn	rnn	PROPN
futech-324	20	21	,	,	PUNCT
futech-324	20	22	lstm	lstm	NOUN
futech-324	20	23	,	,	PUNCT
futech-324	20	24	and	and	CCONJ
futech-324	20	25	gru	gru	NOUN
futech-324	20	26	.	.	PUNCT
futech-324	21	1	these	these	DET
futech-324	21	2	models	model	NOUN
futech-324	21	3	work	work	VERB
futech-324	21	4	proficiently	proficiently	ADV
futech-324	21	5	with	with	ADP
futech-324	21	6	text	text	NOUN
futech-324	21	7	classification	classification	NOUN
futech-324	21	8	.	.	PUNCT
futech-324	22	1	to	to	PART
futech-324	22	2	understand	understand	VERB
futech-324	22	3	contextuality	contextuality	NOUN
futech-324	22	4	and	and	CCONJ
futech-324	22	5	sentiment	sentiment	NOUN
futech-324	22	6	in	in	ADP
futech-324	22	7	textual	textual	ADJ
futech-324	22	8	data	datum	NOUN
futech-324	22	9	,	,	PUNCT
futech-324	22	10	these	these	DET
futech-324	22	11	models	model	NOUN
futech-324	22	12	have	have	VERB
futech-324	22	13	the	the	DET
futech-324	22	14	capability	capability	NOUN
futech-324	22	15	to	to	PART
futech-324	22	16	learn	learn	VERB
futech-324	22	17	temporal	temporal	ADJ
futech-324	22	18	dependencies	dependency	NOUN
futech-324	22	19	in	in	ADP
futech-324	22	20	sequential	sequential	ADJ
futech-324	22	21	data	datum	NOUN
futech-324	22	22	.	.	PUNCT
futech-324	23	1	furthermore	furthermore	ADV
futech-324	23	2	,	,	PUNCT
futech-324	23	3	transformerbased	transformerbase	VERB
futech-324	23	4	models	model	NOUN
futech-324	23	5	such	such	ADJ
futech-324	23	6	as	as	ADP
futech-324	23	7	bert	bert	PROPN
futech-324	23	8	have	have	AUX
futech-324	23	9	transformed	transform	VERB
futech-324	23	10	nlp	nlp	ADJ
futech-324	23	11	tasks	task	NOUN
futech-324	23	12	by	by	ADP
futech-324	23	13	capturing	capture	VERB
futech-324	23	14	extensive	extensive	ADJ
futech-324	23	15	contextual	contextual	ADJ
futech-324	23	16	information	information	NOUN
futech-324	23	17	and	and	CCONJ
futech-324	23	18	learning	learn	VERB
futech-324	23	19	contextual	contextual	ADJ
futech-324	23	20	word	word	NOUN
futech-324	23	21	embeddings	embedding	NOUN
futech-324	23	22	.	.	PUNCT
futech-324	24	1	despite	despite	SCONJ
futech-324	24	2	significant	significant	ADJ
futech-324	24	3	progress	progress	NOUN
futech-324	24	4	in	in	ADP
futech-324	24	5	automated	automate	VERB
futech-324	24	6	toxic	toxic	ADJ
futech-324	24	7	comment	comment	NOUN
futech-324	24	8	classification	classification	NOUN
futech-324	24	9	using	use	VERB
futech-324	24	10	ml	ml	X
futech-324	24	11	and	and	CCONJ
futech-324	24	12	dl	dl	PROPN
futech-324	24	13	techniques	technique	NOUN
futech-324	24	14	,	,	PUNCT
futech-324	24	15	existing	exist	VERB
futech-324	24	16	models	model	NOUN
futech-324	24	17	often	often	ADV
futech-324	24	18	struggle	struggle	VERB
futech-324	24	19	with	with	ADP
futech-324	24	20	accurately	accurately	ADV
futech-324	24	21	capturing	capture	VERB
futech-324	24	22	context	context	NOUN
futech-324	24	23	,	,	PUNCT
futech-324	24	24	handling	handle	VERB
futech-324	24	25	imbalanced	imbalanced	ADJ
futech-324	24	26	datasets	dataset	NOUN
futech-324	24	27	,	,	PUNCT
futech-324	24	28	and	and	CCONJ
futech-324	24	29	distinguishing	distinguish	VERB
futech-324	24	30	between	between	ADP
futech-324	24	31	subtle	subtle	ADJ
futech-324	24	32	toxic	toxic	ADJ
futech-324	24	33	and	and	CCONJ
futech-324	24	34	non	non	ADJ
futech-324	24	35	-	-	ADJ
futech-324	24	36	toxic	toxic	ADJ
futech-324	24	37	content	content	NOUN
futech-324	24	38	.	.	PUNCT
futech-324	25	1	additionally	additionally	ADV
futech-324	25	2	,	,	PUNCT
futech-324	25	3	many	many	ADJ
futech-324	25	4	previous	previous	ADJ
futech-324	25	5	works	work	NOUN
futech-324	25	6	either	either	CCONJ
futech-324	25	7	use	use	VERB
futech-324	25	8	only	only	ADV
futech-324	25	9	shallow	shallow	ADJ
futech-324	25	10	feature	feature	NOUN
futech-324	25	11	representations	representation	NOUN
futech-324	25	12	or	or	CCONJ
futech-324	25	13	do	do	AUX
futech-324	25	14	not	not	PART
futech-324	25	15	take	take	VERB
futech-324	25	16	full	full	ADJ
futech-324	25	17	advantage	advantage	NOUN
futech-324	25	18	of	of	ADP
futech-324	25	19	future	future	ADJ
futech-324	25	20	technology	technology	NOUN
futech-324	25	21	open	open	ADJ
futech-324	25	22	access	access	NOUN
futech-324	25	23	journal	journal	PROPN
futech-324	25	24	journal	journal	PROPN
futech-324	25	25	homepage	homepage	NOUN
futech-324	25	26	:	:	PUNCT
futech-324	25	27	https://fupubco.com/futech	https://fupubco.com/futech	PROPN
futech-324	25	28	issn	issn	VERB
futech-324	25	29	2832	2832	NUM
futech-324	25	30	-	-	SYM
futech-324	25	31	0379	0379	NUM
futech-324	25	32	august	august	PROPN
futech-324	25	33	2025|	2025|	NUM
futech-324	25	34	volume	volume	NOUN
futech-324	25	35	04	04	NUM
futech-324	26	1	|	|	NOUN
futech-324	26	2	issue	issue	NOUN
futech-324	26	3	03	03	NUM
futech-324	27	1	|	|	ADV
futech-324	27	2	pages	page	VERB
futech-324	27	3	76	76	NUM
futech-324	27	4	-	-	SYM
futech-324	27	5	84	84	NUM
futech-324	27	6	https://doi.org/10.55670/fpll.futech.4.3.8	https://doi.org/10.55670/fpll.futech.4.3.8	ADJ
futech-324	27	7	mailto:sushma.cse2@gmail.com	mailto:sushma.cse2@gmail.com	PROPN
futech-324	27	8	https://fupubco.com/futech	https://fupubco.com/futech	PROPN
futech-324	27	9	https://doi.org/10.55670/fpll.futech.4.3.8	https://doi.org/10.55670/fpll.futech.4.3.8	PROPN
futech-324	27	10	s.	s.	PROPN
futech-324	27	11	sushma	sushma	PROPN
futech-324	27	12	et	et	PROPN
futech-324	27	13	al	al	PROPN
futech-324	27	14	.	.	PROPN
futech-324	27	15	/future	/future	PUNCT
futech-324	27	16	technology	technology	PROPN
futech-324	27	17	august	august	PROPN
futech-324	27	18	2025|	2025|	NUM
futech-324	28	1	volume	volume	NOUN
futech-324	28	2	04	04	NUM
futech-324	29	1	|	|	NOUN
futech-324	29	2	issue	issue	NOUN
futech-324	29	3	03	03	NUM
futech-324	30	1	|	|	ADV
futech-324	30	2	pages	page	VERB
futech-324	30	3	76	76	NUM
futech-324	30	4	-	-	SYM
futech-324	30	5	84	84	NUM
futech-324	30	6	77	77	NUM
futech-324	30	7	classical	classical	ADJ
futech-324	30	8	and	and	CCONJ
futech-324	30	9	transformer	transformer	NOUN
futech-324	30	10	-	-	PUNCT
futech-324	30	11	based	base	VERB
futech-324	30	12	embeddings	embedding	NOUN
futech-324	30	13	.	.	PUNCT
futech-324	31	1	we	we	PRON
futech-324	31	2	aim	aim	VERB
futech-324	31	3	to	to	PART
futech-324	31	4	resolve	resolve	VERB
futech-324	31	5	these	these	DET
futech-324	31	6	issues	issue	NOUN
futech-324	31	7	in	in	ADP
futech-324	31	8	this	this	DET
futech-324	31	9	work	work	NOUN
futech-324	31	10	by	by	ADP
futech-324	31	11	conducting	conduct	VERB
futech-324	31	12	experiments	experiment	NOUN
futech-324	31	13	with	with	ADP
futech-324	31	14	hybrid	hybrid	ADJ
futech-324	31	15	dl	dl	PROPN
futech-324	31	16	models	model	NOUN
futech-324	31	17	that	that	PRON
futech-324	31	18	rely	rely	VERB
futech-324	31	19	on	on	ADP
futech-324	31	20	the	the	DET
futech-324	31	21	tf	tf	PROPN
futech-324	31	22	-	-	PUNCT
futech-324	31	23	idf	idf	PROPN
futech-324	31	24	,	,	PUNCT
futech-324	31	25	word2vec	word2vec	X
futech-324	31	26	,	,	PUNCT
futech-324	31	27	and	and	CCONJ
futech-324	31	28	bert	bert	NOUN
futech-324	31	29	embeddings	embedding	NOUN
futech-324	31	30	for	for	ADP
futech-324	31	31	better	well	ADJ
futech-324	31	32	toxic	toxic	ADJ
futech-324	31	33	comment	comment	NOUN
futech-324	31	34	detection	detection	NOUN
futech-324	31	35	.	.	PUNCT
futech-324	32	1	this	this	DET
futech-324	32	2	work	work	NOUN
futech-324	32	3	distinguishes	distinguish	VERB
futech-324	32	4	itself	itself	PRON
futech-324	32	5	from	from	ADP
futech-324	32	6	prior	prior	ADJ
futech-324	32	7	studies	study	NOUN
futech-324	32	8	by	by	ADP
futech-324	32	9	offering	offer	VERB
futech-324	32	10	a	a	DET
futech-324	32	11	comprehensive	comprehensive	ADJ
futech-324	32	12	and	and	CCONJ
futech-324	32	13	unified	unified	ADJ
futech-324	32	14	comparison	comparison	NOUN
futech-324	32	15	of	of	ADP
futech-324	32	16	various	various	ADJ
futech-324	32	17	word	word	NOUN
futech-324	32	18	embedding	embed	VERB
futech-324	32	19	strategies	strategy	NOUN
futech-324	32	20	(	(	PUNCT
futech-324	32	21	tf	tf	NOUN
futech-324	32	22	-	-	PUNCT
futech-324	32	23	idf	idf	PROPN
futech-324	32	24	,	,	PUNCT
futech-324	32	25	word2vec	word2vec	X
futech-324	32	26	,	,	PUNCT
futech-324	32	27	bert	bert	PROPN
futech-324	32	28	)	)	PUNCT
futech-324	32	29	in	in	ADP
futech-324	32	30	combination	combination	NOUN
futech-324	32	31	with	with	ADP
futech-324	32	32	sequential	sequential	ADJ
futech-324	32	33	models	model	NOUN
futech-324	32	34	(	(	PUNCT
futech-324	32	35	rnn	rnn	PROPN
futech-324	32	36	,	,	PUNCT
futech-324	32	37	lstm	lstm	PROPN
futech-324	32	38	,	,	PUNCT
futech-324	32	39	gru	gru	PROPN
futech-324	32	40	)	)	PUNCT
futech-324	32	41	,	,	PUNCT
futech-324	32	42	all	all	PRON
futech-324	32	43	evaluated	evaluate	VERB
futech-324	32	44	under	under	ADP
futech-324	32	45	consistent	consistent	ADJ
futech-324	32	46	preprocessing	preprocessing	NOUN
futech-324	32	47	,	,	PUNCT
futech-324	32	48	tokenization	tokenization	NOUN
futech-324	32	49	,	,	PUNCT
futech-324	32	50	and	and	CCONJ
futech-324	32	51	training	training	NOUN
futech-324	32	52	configurations	configuration	NOUN
futech-324	32	53	.	.	PUNCT
futech-324	33	1	unlike	unlike	ADP
futech-324	33	2	existing	exist	VERB
futech-324	33	3	research	research	NOUN
futech-324	33	4	that	that	PRON
futech-324	33	5	often	often	ADV
futech-324	33	6	benchmarks	benchmark	VERB
futech-324	33	7	one	one	NUM
futech-324	33	8	or	or	CCONJ
futech-324	33	9	two	two	NUM
futech-324	33	10	models	model	NOUN
futech-324	33	11	,	,	PUNCT
futech-324	33	12	this	this	DET
futech-324	33	13	study	study	NOUN
futech-324	33	14	explores	explore	VERB
futech-324	33	15	a	a	DET
futech-324	33	16	wide	wide	ADJ
futech-324	33	17	array	array	NOUN
futech-324	33	18	of	of	ADP
futech-324	33	19	hybrid	hybrid	ADJ
futech-324	33	20	architectures	architecture	NOUN
futech-324	33	21	(	(	PUNCT
futech-324	33	22	e.g.	e.g.	ADV
futech-324	33	23	,	,	PUNCT
futech-324	33	24	lstm+bert	lstm+bert	PROPN
futech-324	33	25	)	)	PUNCT
futech-324	33	26	to	to	PART
futech-324	33	27	identify	identify	VERB
futech-324	33	28	optimal	optimal	ADJ
futech-324	33	29	pairings	pairing	NOUN
futech-324	33	30	for	for	ADP
futech-324	33	31	toxic	toxic	ADJ
futech-324	33	32	comment	comment	NOUN
futech-324	33	33	classification	classification	NOUN
futech-324	33	34	.	.	PUNCT
futech-324	34	1	the	the	DET
futech-324	34	2	implementation	implementation	NOUN
futech-324	34	3	is	be	AUX
futech-324	34	4	designed	design	VERB
futech-324	34	5	to	to	PART
futech-324	34	6	reflect	reflect	VERB
futech-324	34	7	practical	practical	ADJ
futech-324	34	8	deployment	deployment	NOUN
futech-324	34	9	scenarios	scenario	NOUN
futech-324	34	10	using	use	VERB
futech-324	34	11	real	real	ADJ
futech-324	34	12	-	-	PUNCT
futech-324	34	13	world	world	NOUN
futech-324	34	14	metrics	metric	NOUN
futech-324	34	15	and	and	CCONJ
futech-324	34	16	balanced	balanced	ADJ
futech-324	34	17	experimental	experimental	ADJ
futech-324	34	18	design	design	NOUN
futech-324	34	19	.	.	PUNCT
futech-324	35	1	the	the	DET
futech-324	35	2	main	main	ADJ
futech-324	35	3	contributions	contribution	NOUN
futech-324	35	4	of	of	ADP
futech-324	35	5	this	this	DET
futech-324	35	6	work	work	NOUN
futech-324	35	7	are	be	AUX
futech-324	35	8	listed	list	VERB
futech-324	35	9	below	below	ADP
futech-324	35	10	:	:	PUNCT
futech-324	35	11	•	•	ADP
futech-324	35	12	to	to	PART
futech-324	35	13	design	design	VERB
futech-324	35	14	rnn	rnn	NOUN
futech-324	35	15	and	and	CCONJ
futech-324	35	16	other	other	ADJ
futech-324	35	17	deep	deep	ADJ
futech-324	35	18	learning	learning	NOUN
futech-324	35	19	models	model	NOUN
futech-324	35	20	to	to	PART
futech-324	35	21	compare	compare	VERB
futech-324	35	22	for	for	ADP
futech-324	35	23	this	this	DET
futech-324	35	24	toxic	toxic	ADJ
futech-324	35	25	comment	comment	NOUN
futech-324	35	26	classification	classification	NOUN
futech-324	35	27	.	.	PUNCT
futech-324	36	1	•	•	ADP
futech-324	36	2	to	to	PART
futech-324	36	3	explore	explore	VERB
futech-324	36	4	the	the	DET
futech-324	36	5	performance	performance	NOUN
futech-324	36	6	of	of	ADP
futech-324	36	7	different	different	ADJ
futech-324	36	8	embedding	embed	VERB
futech-324	36	9	approaches	approach	NOUN
futech-324	36	10	,	,	PUNCT
futech-324	36	11	such	such	ADJ
futech-324	36	12	as	as	ADP
futech-324	36	13	tf	tf	PROPN
futech-324	36	14	-	-	PUNCT
futech-324	36	15	idf	idf	PROPN
futech-324	36	16	,	,	PUNCT
futech-324	36	17	word2vec	word2vec	X
futech-324	36	18	,	,	PUNCT
futech-324	36	19	and	and	CCONJ
futech-324	36	20	bert	bert	PROPN
futech-324	36	21	,	,	PUNCT
futech-324	36	22	for	for	ADP
futech-324	36	23	semantics	semantic	NOUN
futech-324	36	24	and	and	CCONJ
futech-324	36	25	context	context	NOUN
futech-324	36	26	features	feature	NOUN
futech-324	36	27	.	.	PUNCT
futech-324	37	1	•	•	NUM
futech-324	37	2	investigate	investigate	VERB
futech-324	37	3	using	use	VERB
futech-324	37	4	hybrid	hybrid	ADJ
futech-324	37	5	architectures	architecture	NOUN
futech-324	37	6	that	that	PRON
futech-324	37	7	connect	connect	VERB
futech-324	37	8	traditional	traditional	ADJ
futech-324	37	9	deep	deep	ADJ
futech-324	37	10	learning	learning	NOUN
futech-324	37	11	with	with	ADP
futech-324	37	12	transformer	transformer	NOUN
futech-324	37	13	-	-	PUNCT
futech-324	37	14	related	relate	VERB
futech-324	37	15	embeddings	embedding	NOUN
futech-324	37	16	.	.	PUNCT
futech-324	38	1	•	•	ADP
futech-324	38	2	to	to	PART
futech-324	38	3	determine	determine	VERB
futech-324	38	4	the	the	DET
futech-324	38	5	classification	classification	NOUN
futech-324	38	6	accuracy	accuracy	NOUN
futech-324	38	7	,	,	PUNCT
futech-324	38	8	precision	precision	NOUN
futech-324	38	9	,	,	PUNCT
futech-324	38	10	recall	recall	NOUN
futech-324	38	11	,	,	PUNCT
futech-324	38	12	and	and	CCONJ
futech-324	38	13	f1	f1	NOUN
futech-324	38	14	-	-	PUNCT
futech-324	38	15	score	score	NOUN
futech-324	38	16	of	of	ADP
futech-324	38	17	all	all	DET
futech-324	38	18	models	model	NOUN
futech-324	38	19	on	on	ADP
futech-324	38	20	the	the	DET
futech-324	38	21	basis	basis	NOUN
futech-324	38	22	of	of	ADP
futech-324	38	23	realworld	realworld	PROPN
futech-324	38	24	toxic	toxic	PROPN
futech-324	38	25	comment	comment	NOUN
futech-324	38	26	data	datum	NOUN
futech-324	38	27	.	.	PUNCT
futech-324	39	1	2	2	X
futech-324	39	2	.	.	X
futech-324	39	3	literature	literature	NOUN
futech-324	39	4	survey	survey	PROPN
futech-324	39	5	a.	a.	NOUN
futech-324	39	6	albladi	albladi	PROPN
futech-324	39	7	et	et	PROPN
futech-324	39	8	al	al	PROPN
futech-324	39	9	.	.	PUNCT
futech-324	40	1	[	[	X
futech-324	40	2	1	1	NUM
futech-324	40	3	]	]	PUNCT
futech-324	40	4	investigated	investigate	VERB
futech-324	40	5	sentiment	sentiment	NOUN
futech-324	40	6	research	research	NOUN
futech-324	40	7	specifically	specifically	ADV
futech-324	40	8	on	on	ADP
futech-324	40	9	twitter	twitter	NOUN
futech-324	40	10	,	,	PUNCT
futech-324	40	11	which	which	PRON
futech-324	40	12	is	be	AUX
futech-324	40	13	extensive	extensive	ADJ
futech-324	40	14	in	in	ADP
futech-324	40	15	scale	scale	NOUN
futech-324	40	16	and	and	CCONJ
futech-324	40	17	a	a	DET
futech-324	40	18	valuable	valuable	ADJ
futech-324	40	19	data	datum	NOUN
futech-324	40	20	source	source	NOUN
futech-324	40	21	for	for	ADP
futech-324	40	22	understanding	understand	VERB
futech-324	40	23	public	public	ADJ
futech-324	40	24	opinion	opinion	NOUN
futech-324	40	25	.	.	PUNCT
futech-324	41	1	they	they	PRON
futech-324	41	2	reviewed	review	VERB
futech-324	41	3	desirable	desirable	ADJ
futech-324	41	4	and	and	CCONJ
futech-324	41	5	problematic	problematic	ADJ
futech-324	41	6	features	feature	NOUN
futech-324	41	7	and	and	CCONJ
futech-324	41	8	metrics	metric	NOUN
futech-324	41	9	of	of	ADP
futech-324	41	10	ml	ml	ADP
futech-324	41	11	,	,	PUNCT
futech-324	41	12	dl	dl	PROPN
futech-324	41	13	,	,	PUNCT
futech-324	41	14	and	and	CCONJ
futech-324	41	15	hybrid	hybrid	ADJ
futech-324	41	16	approaches	approach	NOUN
futech-324	41	17	.	.	PUNCT
futech-324	42	1	you	you	PRON
futech-324	42	2	focus	focus	VERB
futech-324	42	3	on	on	ADP
futech-324	42	4	the	the	DET
futech-324	42	5	bert	bert	PROPN
futech-324	42	6	and	and	CCONJ
futech-324	42	7	gpt	gpt	NOUN
futech-324	42	8	transformer	transformer	NOUN
futech-324	42	9	architectures	architecture	NOUN
futech-324	42	10	,	,	PUNCT
futech-324	42	11	although	although	SCONJ
futech-324	42	12	seismic	seismic	ADJ
futech-324	42	13	preprocessing	preprocessing	NOUN
futech-324	42	14	techniques	technique	NOUN
futech-324	42	15	,	,	PUNCT
futech-324	42	16	the	the	DET
futech-324	42	17	extraction	extraction	NOUN
futech-324	42	18	of	of	ADP
futech-324	42	19	features	feature	NOUN
futech-324	42	20	,	,	PUNCT
futech-324	42	21	and	and	CCONJ
futech-324	42	22	sentiment	sentiment	NOUN
futech-324	42	23	lexicons	lexicon	NOUN
futech-324	42	24	were	be	AUX
futech-324	42	25	also	also	ADV
futech-324	42	26	summarily	summarily	ADV
futech-324	42	27	considered	consider	VERB
futech-324	42	28	.	.	PUNCT
futech-324	43	1	this	this	DET
futech-324	43	2	research	research	NOUN
futech-324	43	3	aimed	aim	VERB
futech-324	43	4	to	to	PART
futech-324	43	5	give	give	VERB
futech-324	43	6	a	a	DET
futech-324	43	7	comprehensive	comprehensive	ADJ
futech-324	43	8	landscape	landscape	NOUN
futech-324	43	9	of	of	ADP
futech-324	43	10	use	use	NOUN
futech-324	43	11	cases	case	NOUN
futech-324	43	12	in	in	ADP
futech-324	43	13	twitter	twitter	NOUN
futech-324	43	14	sentiment	sentiment	NOUN
futech-324	43	15	analysis	analysis	NOUN
futech-324	43	16	,	,	PUNCT
futech-324	43	17	offering	offer	VERB
futech-324	43	18	useful	useful	ADJ
futech-324	43	19	insights	insight	NOUN
futech-324	43	20	for	for	ADP
futech-324	43	21	practitioners	practitioner	NOUN
futech-324	43	22	and	and	CCONJ
futech-324	43	23	researchers	researcher	NOUN
futech-324	43	24	in	in	ADP
futech-324	43	25	this	this	DET
futech-324	43	26	area	area	NOUN
futech-324	43	27	.	.	PUNCT
futech-324	44	1	z.	z.	PROPN
futech-324	44	2	hao	hao	PROPN
futech-324	44	3	et	et	PROPN
futech-324	44	4	al	al	PROPN
futech-324	44	5	.	.	PUNCT
futech-324	45	1	[	[	X
futech-324	45	2	2	2	X
futech-324	45	3	]	]	PUNCT
futech-324	45	4	discussed	discuss	VERB
futech-324	45	5	the	the	DET
futech-324	45	6	intricate	intricate	ADJ
futech-324	45	7	complexities	complexity	NOUN
futech-324	45	8	of	of	ADP
futech-324	45	9	gathering	gather	VERB
futech-324	45	10	public	public	ADJ
futech-324	45	11	views	view	NOUN
futech-324	45	12	,	,	PUNCT
futech-324	45	13	particularly	particularly	ADV
futech-324	45	14	regarding	regard	VERB
futech-324	45	15	social	social	ADJ
futech-324	45	16	media	medium	NOUN
futech-324	45	17	and	and	CCONJ
futech-324	45	18	its	its	PRON
futech-324	45	19	impacts	impact	NOUN
futech-324	45	20	on	on	ADP
futech-324	45	21	social	social	ADJ
futech-324	45	22	incidents	incident	NOUN
futech-324	45	23	.	.	PUNCT
futech-324	46	1	a	a	DET
futech-324	46	2	two	two	NUM
futech-324	46	3	-	-	PUNCT
futech-324	46	4	tier	tier	NOUN
futech-324	46	5	hierarchical	hierarchical	ADJ
futech-324	46	6	mechanism	mechanism	NOUN
futech-324	46	7	was	be	AUX
futech-324	46	8	proposed	propose	VERB
futech-324	46	9	for	for	ADP
futech-324	46	10	report	report	NOUN
futech-324	46	11	categorization	categorization	NOUN
futech-324	46	12	and	and	CCONJ
futech-324	46	13	review	review	NOUN
futech-324	46	14	-	-	PUNCT
futech-324	46	15	level	level	NOUN
futech-324	46	16	sentiment	sentiment	NOUN
futech-324	46	17	analysis	analysis	NOUN
futech-324	46	18	,	,	PUNCT
futech-324	46	19	aiming	aim	VERB
futech-324	46	20	to	to	PART
futech-324	46	21	effectively	effectively	ADV
futech-324	46	22	derive	derive	VERB
futech-324	46	23	sentiment	sentiment	NOUN
futech-324	46	24	through	through	ADP
futech-324	46	25	a	a	DET
futech-324	46	26	multi	multi	ADJ
futech-324	46	27	-	-	ADJ
futech-324	46	28	step	step	ADJ
futech-324	46	29	approach	approach	NOUN
futech-324	46	30	.	.	PUNCT
futech-324	47	1	this	this	DET
futech-324	47	2	method	method	NOUN
futech-324	47	3	was	be	AUX
futech-324	47	4	augmented	augment	VERB
futech-324	47	5	with	with	ADP
futech-324	47	6	advances	advance	NOUN
futech-324	47	7	in	in	ADP
futech-324	47	8	model	model	NOUN
futech-324	47	9	architecture	architecture	NOUN
futech-324	47	10	,	,	PUNCT
futech-324	47	11	including	include	VERB
futech-324	47	12	embedding	embed	VERB
futech-324	47	13	tables	table	NOUN
futech-324	47	14	and	and	CCONJ
futech-324	47	15	gating	gate	VERB
futech-324	47	16	mechanisms	mechanism	NOUN
futech-324	47	17	.	.	PUNCT
futech-324	48	1	moreover	moreover	ADV
futech-324	48	2	,	,	PUNCT
futech-324	48	3	they	they	PRON
futech-324	48	4	proposed	propose	VERB
futech-324	48	5	a	a	DET
futech-324	48	6	new	new	ADJ
futech-324	48	7	distributed	distribute	VERB
futech-324	48	8	dl	dl	PROPN
futech-324	48	9	model	model	NOUN
futech-324	48	10	which	which	PRON
futech-324	48	11	is	be	AUX
futech-324	48	12	based	base	VERB
futech-324	48	13	on	on	ADP
futech-324	48	14	blockchain	blockchain	PROPN
futech-324	48	15	isomerism	isomerism	PROPN
futech-324	48	16	learning	learn	VERB
futech-324	48	17	for	for	ADP
futech-324	48	18	the	the	DET
futech-324	48	19	security	security	NOUN
futech-324	48	20	risks	risk	NOUN
futech-324	48	21	and	and	CCONJ
futech-324	48	22	reducing	reduce	VERB
futech-324	48	23	the	the	DET
futech-324	48	24	model	model	NOUN
futech-324	48	25	silos	silo	NOUN
futech-324	48	26	.	.	PUNCT
futech-324	49	1	through	through	ADP
futech-324	49	2	a	a	DET
futech-324	49	3	series	series	NOUN
futech-324	49	4	of	of	ADP
futech-324	49	5	extensive	extensive	ADJ
futech-324	49	6	experiments	experiment	NOUN
futech-324	49	7	,	,	PUNCT
futech-324	49	8	they	they	PRON
futech-324	49	9	showed	show	VERB
futech-324	49	10	that	that	SCONJ
futech-324	49	11	,	,	PUNCT
futech-324	49	12	for	for	ADP
futech-324	49	13	performance	performance	NOUN
futech-324	49	14	and	and	CCONJ
futech-324	49	15	aggregation	aggregation	NOUN
futech-324	49	16	efficiency	efficiency	NOUN
futech-324	49	17	,	,	PUNCT
futech-324	49	18	their	their	PRON
futech-324	49	19	approach	approach	NOUN
futech-324	49	20	vastly	vastly	ADV
futech-324	49	21	outperforms	outperform	VERB
futech-324	49	22	existing	exist	VERB
futech-324	49	23	methods	method	NOUN
futech-324	49	24	.	.	PUNCT
futech-324	50	1	x.	x.	PROPN
futech-324	50	2	wang	wang	PROPN
futech-324	50	3	et	et	PROPN
futech-324	50	4	al	al	PROPN
futech-324	50	5	.	.	PUNCT
futech-324	51	1	[	[	X
futech-324	51	2	3	3	X
futech-324	51	3	]	]	PUNCT
futech-324	51	4	addressed	address	VERB
futech-324	51	5	the	the	DET
futech-324	51	6	challenges	challenge	NOUN
futech-324	51	7	in	in	ADP
futech-324	51	8	sentiment	sentiment	NOUN
futech-324	51	9	analysis	analysis	NOUN
futech-324	51	10	by	by	ADP
futech-324	51	11	introducing	introduce	VERB
futech-324	51	12	a	a	DET
futech-324	51	13	novel	novel	ADJ
futech-324	51	14	model	model	NOUN
futech-324	51	15	that	that	PRON
futech-324	51	16	integrates	integrate	VERB
futech-324	51	17	multimodal	multimodal	NOUN
futech-324	51	18	multiscale	multiscale	NOUN
futech-324	51	19	features	feature	NOUN
futech-324	51	20	based	base	VERB
futech-324	51	21	on	on	ADP
futech-324	51	22	a	a	DET
futech-324	51	23	fuzzy	fuzzy	ADJ
futech-324	51	24	-	-	PUNCT
futech-324	51	25	deep	deep	ADJ
futech-324	51	26	neural	neural	ADJ
futech-324	51	27	network	network	NOUN
futech-324	51	28	.	.	PUNCT
futech-324	52	1	it	it	PRON
futech-324	52	2	utilizes	utilize	VERB
futech-324	52	3	intrinsic	intrinsic	ADJ
futech-324	52	4	feature	feature	NOUN
futech-324	52	5	representations	representation	NOUN
futech-324	52	6	across	across	ADP
futech-324	52	7	text	text	NOUN
futech-324	52	8	,	,	PUNCT
futech-324	52	9	audio	audio	NOUN
futech-324	52	10	,	,	PUNCT
futech-324	52	11	and	and	CCONJ
futech-324	52	12	image	image	NOUN
futech-324	52	13	data	datum	NOUN
futech-324	52	14	.	.	PUNCT
futech-324	53	1	the	the	DET
futech-324	53	2	model	model	NOUN
futech-324	53	3	employs	employ	VERB
futech-324	53	4	fuzzy	fuzzy	ADJ
futech-324	53	5	logic	logic	NOUN
futech-324	53	6	rules	rule	NOUN
futech-324	53	7	to	to	PART
futech-324	53	8	increase	increase	VERB
futech-324	53	9	the	the	DET
futech-324	53	10	adaptability	adaptability	NOUN
futech-324	53	11	to	to	ADP
futech-324	53	12	the	the	DET
futech-324	53	13	ambiguity	ambiguity	NOUN
futech-324	53	14	present	present	ADJ
futech-324	53	15	in	in	ADP
futech-324	53	16	sentiment	sentiment	NOUN
futech-324	53	17	representations	representation	NOUN
futech-324	53	18	.	.	PUNCT
futech-324	54	1	additionally	additionally	ADV
futech-324	54	2	,	,	PUNCT
futech-324	54	3	they	they	PRON
futech-324	54	4	incorporated	incorporate	VERB
futech-324	54	5	a	a	DET
futech-324	54	6	dual	dual	ADJ
futech-324	54	7	attention	attention	NOUN
futech-324	54	8	mechanism	mechanism	NOUN
futech-324	54	9	to	to	PART
futech-324	54	10	flexibly	flexibly	ADV
futech-324	54	11	attend	attend	VERB
futech-324	54	12	to	to	ADP
futech-324	54	13	essential	essential	ADJ
futech-324	54	14	components	component	NOUN
futech-324	54	15	in	in	ADP
futech-324	54	16	the	the	DET
futech-324	54	17	multimodal	multimodal	NOUN
futech-324	54	18	data	datum	NOUN
futech-324	54	19	,	,	PUNCT
futech-324	54	20	thereby	thereby	ADV
futech-324	54	21	enhancing	enhance	VERB
futech-324	54	22	feature	feature	NOUN
futech-324	54	23	extraction	extraction	NOUN
futech-324	54	24	and	and	CCONJ
futech-324	54	25	improving	improve	VERB
futech-324	54	26	contextual	contextual	ADJ
futech-324	54	27	awareness	awareness	NOUN
futech-324	54	28	.	.	PUNCT
futech-324	55	1	we	we	PRON
futech-324	55	2	conducted	conduct	VERB
futech-324	55	3	extensive	extensive	ADJ
futech-324	55	4	validation	validation	NOUN
futech-324	55	5	of	of	ADP
futech-324	55	6	the	the	DET
futech-324	55	7	model	model	NOUN
futech-324	55	8	's	's	PART
futech-324	55	9	performance	performance	NOUN
futech-324	55	10	using	use	VERB
futech-324	55	11	several	several	ADJ
futech-324	55	12	datasets	dataset	NOUN
futech-324	55	13	and	and	CCONJ
futech-324	55	14	characterized	characterize	VERB
futech-324	55	15	its	its	PRON
futech-324	55	16	ability	ability	NOUN
futech-324	55	17	to	to	PART
futech-324	55	18	model	model	VERB
futech-324	55	19	the	the	DET
futech-324	55	20	complexities	complexity	NOUN
futech-324	55	21	of	of	ADP
futech-324	55	22	human	human	ADJ
futech-324	55	23	emotion	emotion	NOUN
futech-324	55	24	better	well	ADV
futech-324	55	25	than	than	ADP
futech-324	55	26	existing	exist	VERB
futech-324	55	27	methods	method	NOUN
futech-324	55	28	.	.	PUNCT
futech-324	56	1	h.	h.	PROPN
futech-324	56	2	t.	t.	PROPN
futech-324	56	3	phan	phan	PROPN
futech-324	56	4	et	et	PROPN
futech-324	56	5	al	al	PROPN
futech-324	56	6	.	.	PUNCT
futech-324	57	1	[	[	X
futech-324	57	2	4	4	X
futech-324	57	3	]	]	PUNCT
futech-324	57	4	devised	devise	VERB
futech-324	57	5	a	a	DET
futech-324	57	6	novel	novel	ADJ
futech-324	57	7	method	method	NOUN
futech-324	57	8	of	of	ADP
futech-324	57	9	aspect	aspect	NOUN
futech-324	57	10	-	-	PUNCT
futech-324	57	11	level	level	NOUN
futech-324	57	12	sentiment	sentiment	NOUN
futech-324	57	13	analysis	analysis	NOUN
futech-324	57	14	based	base	VERB
futech-324	57	15	on	on	ADP
futech-324	57	16	the	the	DET
futech-324	57	17	application	application	NOUN
futech-324	57	18	of	of	ADP
futech-324	57	19	three	three	NUM
futech-324	57	20	gcns	gcns	PROPN
futech-324	57	21	,	,	PUNCT
futech-324	57	22	which	which	PRON
futech-324	57	23	they	they	PRON
futech-324	57	24	coined	coin	VERB
futech-324	57	25	as	as	ADP
futech-324	57	26	multigraphic	multigraphic	ADJ
futech-324	57	27	convolutional	convolutional	ADJ
futech-324	57	28	network	network	NOUN
futech-324	57	29	(	(	PUNCT
futech-324	57	30	mulgcn	mulgcn	PROPN
futech-324	57	31	)	)	PUNCT
futech-324	57	32	.	.	PUNCT
futech-324	58	1	using	use	VERB
futech-324	58	2	the	the	DET
futech-324	58	3	dependency	dependency	NOUN
futech-324	58	4	parser	parser	NOUN
futech-324	58	5	tree	tree	NOUN
futech-324	58	6	,	,	PUNCT
futech-324	58	7	affective	affective	ADJ
futech-324	58	8	information	information	NOUN
futech-324	58	9	from	from	ADP
futech-324	58	10	senticnet	senticnet	NOUN
futech-324	58	11	,	,	PUNCT
futech-324	58	12	and	and	CCONJ
futech-324	58	13	inter	inter	ADJ
futech-324	58	14	-	-	ADJ
futech-324	58	15	aspectawareness	aspectawareness	ADJ
futech-324	58	16	,	,	PUNCT
futech-324	58	17	this	this	DET
futech-324	58	18	approach	approach	NOUN
futech-324	58	19	captures	capture	VERB
futech-324	58	20	both	both	DET
futech-324	58	21	syntax	syntax	NOUN
futech-324	58	22	and	and	CCONJ
futech-324	58	23	semantics	semantic	NOUN
futech-324	58	24	as	as	ADV
futech-324	58	25	well	well	ADV
futech-324	58	26	as	as	ADP
futech-324	58	27	context	context	NOUN
futech-324	58	28	.	.	PUNCT
futech-324	59	1	this	this	DET
futech-324	59	2	paper	paper	NOUN
futech-324	59	3	presents	present	VERB
futech-324	59	4	a	a	DET
futech-324	59	5	model	model	NOUN
futech-324	59	6	that	that	PRON
futech-324	59	7	introduces	introduce	VERB
futech-324	59	8	a	a	DET
futech-324	59	9	solution	solution	NOUN
futech-324	59	10	for	for	ADP
futech-324	59	11	aspect	aspect	NOUN
futech-324	59	12	-	-	PUNCT
futech-324	59	13	level	level	NOUN
futech-324	59	14	sentiment	sentiment	NOUN
futech-324	59	15	analysis	analysis	NOUN
futech-324	59	16	and	and	CCONJ
futech-324	59	17	improves	improve	VERB
futech-324	59	18	its	its	PRON
futech-324	59	19	performance	performance	NOUN
futech-324	59	20	by	by	ADP
futech-324	59	21	addressing	address	VERB
futech-324	59	22	the	the	DET
futech-324	59	23	difficulty	difficulty	NOUN
futech-324	59	24	of	of	ADP
futech-324	59	25	effectively	effectively	ADV
futech-324	59	26	leveraging	leverage	VERB
futech-324	59	27	relevant	relevant	ADJ
futech-324	59	28	features	feature	NOUN
futech-324	59	29	from	from	ADP
futech-324	59	30	different	different	ADJ
futech-324	59	31	knowledge	knowledge	NOUN
futech-324	59	32	sources	source	NOUN
futech-324	59	33	.	.	PUNCT
futech-324	60	1	experimental	experimental	ADJ
futech-324	60	2	evaluations	evaluation	NOUN
futech-324	60	3	over	over	ADP
futech-324	60	4	three	three	NUM
futech-324	60	5	benchmark	benchmark	NOUN
futech-324	60	6	datasets	dataset	NOUN
futech-324	60	7	demonstrate	demonstrate	VERB
futech-324	60	8	that	that	SCONJ
futech-324	60	9	the	the	DET
futech-324	60	10	mulgcn	mulgcn	PROPN
futech-324	60	11	model	model	PROPN
futech-324	60	12	beats	beat	VERB
futech-324	60	13	the	the	DET
futech-324	60	14	state	state	NOUN
futech-324	60	15	-	-	PUNCT
futech-324	60	16	of	of	ADP
futech-324	60	17	-	-	PUNCT
futech-324	60	18	the	the	DET
futech-324	60	19	-	-	PUNCT
futech-324	60	20	art	art	NOUN
futech-324	60	21	and	and	CCONJ
futech-324	60	22	improves	improve	VERB
futech-324	60	23	both	both	PRON
futech-324	60	24	accuracy	accuracy	NOUN
futech-324	60	25	and	and	CCONJ
futech-324	60	26	f1	f1	NOUN
futech-324	60	27	score	score	NOUN
futech-324	60	28	for	for	ADP
futech-324	60	29	aspect	aspect	NOUN
futech-324	60	30	-	-	PUNCT
futech-324	60	31	level	level	NOUN
futech-324	60	32	sentiment	sentiment	NOUN
futech-324	60	33	analysis	analysis	NOUN
futech-324	60	34	.	.	PUNCT
futech-324	61	1	s.	s.	PROPN
futech-324	61	2	ali	ali	PROPN
futech-324	61	3	et	et	PROPN
futech-324	61	4	al	al	PROPN
futech-324	61	5	.	.	PUNCT
futech-324	62	1	[	[	X
futech-324	62	2	5	5	NUM
futech-324	62	3	]	]	PUNCT
futech-324	62	4	highlighted	highlight	VERB
futech-324	62	5	the	the	DET
futech-324	62	6	issue	issue	NOUN
futech-324	62	7	of	of	ADP
futech-324	62	8	[	[	X
futech-324	62	9	event	event	NOUN
futech-324	62	10	classification	classification	NOUN
futech-324	62	11	]	]	PUNCT
futech-324	62	12	in	in	ADP
futech-324	62	13	low	low	ADJ
futech-324	62	14	-	-	PUNCT
futech-324	62	15	resource	resource	NOUN
futech-324	62	16	languages	language	NOUN
futech-324	62	17	,	,	PUNCT
futech-324	62	18	such	such	ADJ
futech-324	62	19	as	as	ADP
futech-324	62	20	urdu	urdu	PROPN
futech-324	62	21	,	,	PUNCT
futech-324	62	22	and	and	CCONJ
futech-324	62	23	underline	underline	VERB
futech-324	62	24	the	the	DET
futech-324	62	25	need	need	NOUN
futech-324	62	26	for	for	ADP
futech-324	62	27	wellformalised	wellformalise	VERB
futech-324	62	28	linguistic	linguistic	ADJ
futech-324	62	29	datasets	dataset	NOUN
futech-324	62	30	.	.	PUNCT
futech-324	63	1	the	the	DET
futech-324	63	2	dataset	dataset	NOUN
futech-324	63	3	contained	contain	VERB
futech-324	63	4	a	a	DET
futech-324	63	5	total	total	NOUN
futech-324	63	6	of	of	ADP
futech-324	63	7	103,771	103,771	NUM
futech-324	63	8	sentences	sentence	NOUN
futech-324	63	9	from	from	ADP
futech-324	63	10	five	five	NUM
futech-324	63	11	different	different	ADJ
futech-324	63	12	social	social	ADJ
futech-324	63	13	media	medium	NOUN
futech-324	63	14	platforms	platform	NOUN
futech-324	63	15	and	and	CCONJ
futech-324	63	16	was	be	AUX
futech-324	63	17	obtained	obtain	VERB
futech-324	63	18	for	for	ADP
futech-324	63	19	the	the	DET
futech-324	63	20	purpose	purpose	NOUN
futech-324	63	21	of	of	ADP
futech-324	63	22	classifying	classify	VERB
futech-324	63	23	text	text	NOUN
futech-324	63	24	in	in	ADP
futech-324	63	25	the	the	DET
futech-324	63	26	urdu	urdu	PROPN
futech-324	63	27	language	language	NOUN
futech-324	63	28	in	in	ADP
futech-324	63	29	a	a	DET
futech-324	63	30	multiclass	multiclass	ADJ
futech-324	63	31	classification	classification	NOUN
futech-324	63	32	approach	approach	NOUN
futech-324	63	33	.	.	PUNCT
futech-324	64	1	the	the	DET
futech-324	64	2	16	16	NUM
futech-324	64	3	event	event	NOUN
futech-324	64	4	categories	category	NOUN
futech-324	64	5	were	be	AUX
futech-324	64	6	used	use	VERB
futech-324	64	7	in	in	ADP
futech-324	64	8	the	the	DET
futech-324	64	9	classification	classification	NOUN
futech-324	64	10	task	task	NOUN
futech-324	64	11	.	.	PUNCT
futech-324	65	1	the	the	DET
futech-324	65	2	smfcnn	smfcnn	PROPN
futech-324	65	3	classifier	classifier	NOUN
futech-324	65	4	showed	show	VERB
futech-324	65	5	the	the	DET
futech-324	65	6	highest	high	ADJ
futech-324	65	7	accuracy	accuracy	NOUN
futech-324	65	8	(	(	PUNCT
futech-324	65	9	88.29	88.29	NUM
futech-324	65	10	%	%	NOUN
futech-324	65	11	)	)	PUNCT
futech-324	65	12	among	among	ADP
futech-324	65	13	the	the	DET
futech-324	65	14	methods	method	NOUN
futech-324	65	15	tested	test	VERB
futech-324	65	16	.	.	PUNCT
futech-324	66	1	furthermore	furthermore	ADV
futech-324	66	2	,	,	PUNCT
futech-324	66	3	on	on	ADP
futech-324	66	4	this	this	DET
futech-324	66	5	dataset	dataset	NOUN
futech-324	66	6	,	,	PUNCT
futech-324	66	7	xlm	xlm	PROPN
futech-324	66	8	-	-	PUNCT
futech-324	66	9	r+	r+	PUNCT
futech-324	66	10	(	(	PUNCT
futech-324	66	11	a	a	DET
futech-324	66	12	proposed	propose	VERB
futech-324	66	13	transformer	transformer	NOUN
futech-324	66	14	-	-	PUNCT
futech-324	66	15	based	base	VERB
futech-324	66	16	model	model	NOUN
futech-324	66	17	)	)	PUNCT
futech-324	66	18	outperformed	outperform	VERB
futech-324	66	19	them	they	PRON
futech-324	66	20	with	with	ADP
futech-324	66	21	an	an	DET
futech-324	66	22	accuracy	accuracy	NOUN
futech-324	66	23	of	of	ADP
futech-324	66	24	89.8	89.8	NUM
futech-324	66	25	%	%	NOUN
futech-324	66	26	as	as	ADV
futech-324	66	27	well	well	ADV
futech-324	66	28	.	.	PUNCT
futech-324	67	1	w.	w.	PROPN
futech-324	67	2	gong	gong	PROPN
futech-324	68	1	[	[	X
futech-324	68	2	6	6	NUM
futech-324	68	3	]	]	PUNCT
futech-324	68	4	developed	develop	VERB
futech-324	68	5	a	a	DET
futech-324	68	6	sentiment	sentiment	NOUN
futech-324	68	7	classification	classification	NOUN
futech-324	68	8	algorithm	algorithm	NOUN
futech-324	68	9	for	for	ADP
futech-324	68	10	textual	textual	ADJ
futech-324	68	11	data	datum	NOUN
futech-324	68	12	mining	mining	NOUN
futech-324	68	13	based	base	VERB
futech-324	68	14	on	on	ADP
futech-324	68	15	bidirectional	bidirectional	ADJ
futech-324	68	16	long	long	ADJ
futech-324	68	17	short	short	ADJ
futech-324	68	18	-	-	PUNCT
futech-324	68	19	term	term	NOUN
futech-324	68	20	memory	memory	NOUN
futech-324	68	21	network	network	NOUN
futech-324	68	22	(	(	PUNCT
futech-324	68	23	bilstm	bilstm	NOUN
futech-324	68	24	)	)	PUNCT
futech-324	68	25	,	,	PUNCT
futech-324	68	26	which	which	PRON
futech-324	68	27	demonstrated	demonstrate	VERB
futech-324	68	28	the	the	DET
futech-324	68	29	relevance	relevance	NOUN
futech-324	68	30	of	of	ADP
futech-324	68	31	emotion	emotion	NOUN
futech-324	68	32	manifestation	manifestation	NOUN
futech-324	68	33	in	in	ADP
futech-324	68	34	e	e	NOUN
futech-324	68	35	-	-	NOUN
futech-324	68	36	commerce	commerce	NOUN
futech-324	68	37	and	and	CCONJ
futech-324	68	38	social	social	ADJ
futech-324	68	39	media	medium	NOUN
futech-324	68	40	data	datum	NOUN
futech-324	68	41	with	with	ADP
futech-324	68	42	dl	dl	PROPN
futech-324	68	43	methods	method	NOUN
futech-324	68	44	.	.	PUNCT
futech-324	69	1	the	the	DET
futech-324	69	2	model	model	NOUN
futech-324	69	3	enhances	enhance	VERB
futech-324	69	4	the	the	DET
futech-324	69	5	accuracy	accuracy	NOUN
futech-324	69	6	in	in	ADP
futech-324	69	7	distinguishing	distinguish	VERB
futech-324	69	8	sentiment	sentiment	NOUN
futech-324	69	9	in	in	ADP
futech-324	69	10	such	such	ADJ
futech-324	69	11	cases	case	NOUN
futech-324	69	12	and	and	CCONJ
futech-324	69	13	outperforms	outperform	VERB
futech-324	69	14	conventional	conventional	ADJ
futech-324	69	15	sentiment	sentiment	NOUN
futech-324	69	16	classification	classification	NOUN
futech-324	69	17	techniques	technique	NOUN
futech-324	69	18	due	due	ADP
futech-324	69	19	to	to	ADP
futech-324	69	20	the	the	DET
futech-324	69	21	utilization	utilization	NOUN
futech-324	69	22	of	of	ADP
futech-324	69	23	bilstm	bilstm	NOUN
futech-324	69	24	,	,	PUNCT
futech-324	69	25	which	which	PRON
futech-324	69	26	adopts	adopt	VERB
futech-324	69	27	bidirectional	bidirectional	ADJ
futech-324	69	28	contexts	context	NOUN
futech-324	69	29	of	of	ADP
futech-324	69	30	text	text	NOUN
futech-324	69	31	segments	segment	NOUN
futech-324	69	32	[	[	X
futech-324	69	33	4,5	4,5	NUM
futech-324	69	34	]	]	PUNCT
futech-324	69	35	.	.	PUNCT
futech-324	70	1	this	this	DET
futech-324	70	2	study	study	NOUN
futech-324	70	3	demonstrates	demonstrate	VERB
futech-324	70	4	that	that	DET
futech-324	70	5	bilstm	bilstm	NOUN
futech-324	70	6	can	can	AUX
futech-324	70	7	also	also	ADV
futech-324	70	8	be	be	AUX
futech-324	70	9	a	a	DET
futech-324	70	10	viable	viable	ADJ
futech-324	70	11	option	option	NOUN
futech-324	70	12	for	for	ADP
futech-324	70	13	sentiment	sentiment	NOUN
futech-324	70	14	analysis	analysis	NOUN
futech-324	70	15	tasks	task	NOUN
futech-324	70	16	,	,	PUNCT
futech-324	70	17	such	such	ADJ
futech-324	70	18	as	as	ADP
futech-324	70	19	customer	customer	NOUN
futech-324	70	20	feedback	feedback	NOUN
futech-324	70	21	analysis	analysis	NOUN
futech-324	70	22	.	.	PUNCT
futech-324	71	1	s.	s.	PROPN
futech-324	71	2	a.	a.	PROPN
futech-324	71	3	mostafa	mostafa	PROPN
futech-324	71	4	et	et	PROPN
futech-324	71	5	al	al	PROPN
futech-324	71	6	.	.	PUNCT
futech-324	72	1	[	[	X
futech-324	72	2	7	7	X
futech-324	72	3	]	]	PUNCT
futech-324	72	4	employed	employ	VERB
futech-324	72	5	a	a	DET
futech-324	72	6	sentiment	sentiment	NOUN
futech-324	72	7	analysis	analysis	NOUN
futech-324	72	8	and	and	CCONJ
futech-324	72	9	classification	classification	NOUN
futech-324	72	10	method	method	NOUN
futech-324	72	11	for	for	ADP
futech-324	72	12	amazon	amazon	NOUN
futech-324	72	13	alexa	alexa	NOUN
futech-324	72	14	products	product	NOUN
futech-324	72	15	.	.	PUNCT
futech-324	73	1	the	the	DET
futech-324	73	2	dataset	dataset	NOUN
futech-324	73	3	contains	contain	VERB
futech-324	73	4	3150	3150	NUM
futech-324	73	5	reviews	review	NOUN
futech-324	73	6	and	and	CCONJ
futech-324	73	7	was	be	AUX
futech-324	73	8	labelled	label	VERB
futech-324	73	9	for	for	ADP
futech-324	73	10	positive	positive	ADJ
futech-324	73	11	or	or	CCONJ
futech-324	73	12	negative	negative	ADJ
futech-324	73	13	sentiment	sentiment	NOUN
futech-324	73	14	.	.	PUNCT
futech-324	74	1	they	they	PRON
futech-324	74	2	therefore	therefore	ADV
futech-324	74	3	trained	train	VERB
futech-324	74	4	four	four	NUM
futech-324	74	5	classifiers	classifier	NOUN
futech-324	74	6	and	and	CCONJ
futech-324	74	7	evaluated	evaluate	VERB
futech-324	74	8	the	the	DET
futech-324	74	9	performance	performance	NOUN
futech-324	74	10	metrics	metric	NOUN
futech-324	74	11	.	.	PUNCT
futech-324	75	1	analysis	analysis	NOUN
futech-324	75	2	of	of	ADP
futech-324	75	3	customer	customer	NOUN
futech-324	75	4	feedback	feedback	NOUN
futech-324	75	5	was	be	AUX
futech-324	75	6	found	find	VERB
futech-324	75	7	to	to	PART
futech-324	75	8	be	be	AUX
futech-324	75	9	highly	highly	ADV
futech-324	75	10	effective	effective	ADJ
futech-324	75	11	using	use	VERB
futech-324	75	12	rf	rf	NOUN
futech-324	75	13	,	,	PUNCT
futech-324	75	14	which	which	PRON
futech-324	75	15	was	be	AUX
futech-324	75	16	the	the	DET
futech-324	75	17	best	well	ADV
futech-324	75	18	-	-	PUNCT
futech-324	75	19	performing	perform	VERB
futech-324	75	20	classifier	classifier	NOUN
futech-324	75	21	.	.	PUNCT
futech-324	76	1	s.	s.	PROPN
futech-324	76	2	mehta	mehta	PROPN
futech-324	76	3	et	et	PROPN
futech-324	76	4	al	al	PROPN
futech-324	76	5	.	.	PUNCT
futech-324	77	1	[	[	X
futech-324	77	2	8	8	NUM
futech-324	77	3	]	]	PUNCT
futech-324	77	4	applied	apply	VERB
futech-324	77	5	cnn	cnn	PROPN
futech-324	77	6	as	as	ADP
futech-324	77	7	a	a	DET
futech-324	77	8	new	new	ADJ
futech-324	77	9	method	method	NOUN
futech-324	77	10	for	for	ADP
futech-324	77	11	sentiment	sentiment	NOUN
futech-324	77	12	analysis	analysis	NOUN
futech-324	77	13	with	with	ADP
futech-324	77	14	federated	federated	ADJ
futech-324	77	15	learning	learning	NOUN
futech-324	77	16	.	.	PUNCT
futech-324	78	1	it	it	PRON
futech-324	78	2	enables	enable	VERB
futech-324	78	3	model	model	NOUN
futech-324	78	4	-	-	PUNCT
futech-324	78	5	trained	train	VERB
futech-324	78	6	decentralization	decentralization	NOUN
futech-324	78	7	by	by	ADP
futech-324	78	8	processing	process	VERB
futech-324	78	9	multiple	multiple	ADJ
futech-324	78	10	devices	device	NOUN
futech-324	78	11	without	without	ADP
futech-324	78	12	transferring	transfer	VERB
futech-324	78	13	data	datum	NOUN
futech-324	78	14	to	to	ADP
futech-324	78	15	the	the	DET
futech-324	78	16	central	central	ADJ
futech-324	78	17	server	server	NOUN
futech-324	78	18	,	,	PUNCT
futech-324	78	19	thereby	thereby	ADV
futech-324	78	20	improving	improve	VERB
futech-324	78	21	privacy	privacy	NOUN
futech-324	78	22	while	while	SCONJ
futech-324	78	23	maintaining	maintain	VERB
futech-324	78	24	sufficient	sufficient	ADJ
futech-324	78	25	performance	performance	NOUN
futech-324	78	26	for	for	ADP
futech-324	78	27	the	the	DET
futech-324	78	28	sentiment	sentiment	NOUN
futech-324	78	29	classifier	classifier	NOUN
futech-324	78	30	.	.	PUNCT
futech-324	79	1	it	it	PRON
futech-324	79	2	is	be	AUX
futech-324	79	3	based	base	VERB
futech-324	79	4	on	on	ADP
futech-324	79	5	high	high	ADJ
futech-324	79	6	accuracy	accuracy	NOUN
futech-324	79	7	and	and	CCONJ
futech-324	79	8	a	a	DET
futech-324	79	9	good	good	ADJ
futech-324	79	10	roc	roc	PROPN
futech-324	79	11	auc	auc	NOUN
futech-324	79	12	score	score	NOUN
futech-324	79	13	,	,	PUNCT
futech-324	79	14	denoting	denote	VERB
futech-324	79	15	good	good	ADJ
futech-324	79	16	abbreviations	abbreviation	NOUN
futech-324	79	17	ann	ann	PROPN
futech-324	79	18	artificial	artificial	ADJ
futech-324	79	19	neural	neural	ADJ
futech-324	79	20	network	network	NOUN
futech-324	79	21	cnn	cnn	PROPN
futech-324	79	22	convolutional	convolutional	ADJ
futech-324	79	23	neural	neural	ADJ
futech-324	79	24	network	network	NOUN
futech-324	79	25	rnn	rnn	VERB
futech-324	79	26	recurrent	recurrent	ADJ
futech-324	79	27	neural	neural	ADJ
futech-324	79	28	network	network	NOUN
futech-324	79	29	lstm	lstm	PROPN
futech-324	79	30	long	long	ADJ
futech-324	79	31	short	short	ADJ
futech-324	79	32	-	-	PUNCT
futech-324	79	33	term	term	NOUN
futech-324	79	34	memory	memory	NOUN
futech-324	79	35	gru	gru	NOUN
futech-324	79	36	gated	gate	VERB
futech-324	79	37	recurrent	recurrent	ADJ
futech-324	79	38	unit	unit	NOUN
futech-324	79	39	nlp	nlp	PROPN
futech-324	79	40	natural	natural	ADJ
futech-324	79	41	language	language	NOUN
futech-324	79	42	processing	processing	NOUN
futech-324	79	43	bert	bert	PROPN
futech-324	79	44	bidirectional	bidirectional	ADJ
futech-324	79	45	encoder	encoder	NOUN
futech-324	79	46	representations	representation	VERB
futech-324	79	47	from	from	ADP
futech-324	79	48	transformers	transformer	NOUN
futech-324	79	49	tf	tf	PROPN
futech-324	79	50	-	-	PUNCT
futech-324	79	51	idf	idf	PROPN
futech-324	79	52	term	term	NOUN
futech-324	79	53	frequency	frequency	NOUN
futech-324	79	54	–	–	PUNCT
futech-324	79	55	inverse	inverse	ADJ
futech-324	79	56	document	document	NOUN
futech-324	79	57	frequency	frequency	NOUN
futech-324	79	58	word2vec	word2vec	AUX
futech-324	79	59	word	word	NOUN
futech-324	79	60	to	to	ADP
futech-324	79	61	vector	vector	NOUN
futech-324	79	62	s.	s.	PROPN
futech-324	79	63	sushma	sushma	PROPN
futech-324	79	64	et	et	PROPN
futech-324	79	65	al	al	PROPN
futech-324	79	66	.	.	PROPN
futech-324	79	67	/future	/future	PUNCT
futech-324	79	68	technology	technology	PROPN
futech-324	79	69	august	august	PROPN
futech-324	79	70	2025|	2025|	NUM
futech-324	80	1	volume	volume	NOUN
futech-324	80	2	04	04	NUM
futech-324	81	1	|	|	NOUN
futech-324	81	2	issue	issue	NOUN
futech-324	81	3	03	03	NUM
futech-324	82	1	|	|	ADV
futech-324	82	2	pages	page	VERB
futech-324	82	3	76	76	NUM
futech-324	82	4	-	-	SYM
futech-324	82	5	84	84	NUM
futech-324	82	6	78	78	NUM
futech-324	82	7	separation	separation	NOUN
futech-324	82	8	of	of	ADP
futech-324	82	9	classes	class	NOUN
futech-324	82	10	as	as	ADP
futech-324	82	11	different	different	ADJ
futech-324	82	12	sentiment	sentiment	NOUN
futech-324	82	13	groups	group	NOUN
futech-324	82	14	.	.	PUNCT
futech-324	83	1	as	as	ADP
futech-324	83	2	a	a	DET
futech-324	83	3	result	result	NOUN
futech-324	83	4	,	,	PUNCT
futech-324	83	5	it	it	PRON
futech-324	83	6	enhanced	enhance	VERB
futech-324	83	7	data	data	NOUN
futech-324	83	8	privacy	privacy	NOUN
futech-324	83	9	and	and	CCONJ
futech-324	83	10	efficiency	efficiency	NOUN
futech-324	83	11	due	due	ADP
futech-324	83	12	to	to	ADP
futech-324	83	13	the	the	DET
futech-324	83	14	use	use	NOUN
futech-324	83	15	of	of	ADP
futech-324	83	16	federated	federated	ADJ
futech-324	83	17	learning	learning	NOUN
futech-324	83	18	.	.	PUNCT
futech-324	84	1	x.	x.	NOUN
futech-324	85	1	he	he	PRON
futech-324	85	2	et	et	PROPN
futech-324	85	3	al	al	PROPN
futech-324	85	4	.	.	PUNCT
futech-324	86	1	[	[	X
futech-324	86	2	9	9	NUM
futech-324	86	3	]	]	PUNCT
futech-324	86	4	compared	compare	VERB
futech-324	86	5	svm	svm	PROPN
futech-324	86	6	for	for	ADP
futech-324	86	7	sentiment	sentiment	NOUN
futech-324	86	8	analysis	analysis	NOUN
futech-324	86	9	with	with	ADP
futech-324	86	10	other	other	ADJ
futech-324	86	11	traditional	traditional	ADJ
futech-324	86	12	classifiers	classifier	NOUN
futech-324	86	13	such	such	ADJ
futech-324	86	14	as	as	ADP
futech-324	86	15	lr	lr	PROPN
futech-324	86	16	,	,	PUNCT
futech-324	86	17	knn	knn	PROPN
futech-324	86	18	,	,	PUNCT
futech-324	86	19	nb	nb	PROPN
futech-324	86	20	,	,	PUNCT
futech-324	86	21	and	and	CCONJ
futech-324	86	22	xgboost	xgboost	X
futech-324	86	23	.	.	PUNCT
futech-324	87	1	this	this	DET
futech-324	87	2	study	study	NOUN
futech-324	87	3	applies	apply	VERB
futech-324	87	4	sentiment	sentiment	NOUN
futech-324	87	5	analysis	analysis	NOUN
futech-324	87	6	to	to	PART
futech-324	87	7	classify	classify	VERB
futech-324	87	8	comments	comment	NOUN
futech-324	87	9	about	about	ADP
futech-324	87	10	"	"	PUNCT
futech-324	87	11	huawei	huawei	PROPN
futech-324	87	12	mate60	mate60	NOUN
futech-324	87	13	"	"	PUNCT
futech-324	87	14	on	on	ADP
futech-324	87	15	the	the	DET
futech-324	87	16	little	little	ADJ
futech-324	87	17	red	red	ADJ
futech-324	87	18	book	book	NOUN
futech-324	87	19	platform	platform	NOUN
futech-324	87	20	.	.	PUNCT
futech-324	88	1	compared	compare	VERB
futech-324	88	2	to	to	ADP
futech-324	88	3	the	the	DET
futech-324	88	4	other	other	ADJ
futech-324	88	5	models	model	NOUN
futech-324	88	6	,	,	PUNCT
futech-324	88	7	svm	svm	PROPN
futech-324	88	8	proved	prove	VERB
futech-324	88	9	to	to	PART
futech-324	88	10	be	be	AUX
futech-324	88	11	more	more	ADV
futech-324	88	12	efficient	efficient	ADJ
futech-324	88	13	for	for	ADP
futech-324	88	14	sentiment	sentiment	NOUN
futech-324	88	15	classification	classification	NOUN
futech-324	88	16	,	,	PUNCT
futech-324	88	17	making	make	VERB
futech-324	88	18	it	it	PRON
futech-324	88	19	a	a	DET
futech-324	88	20	recommended	recommend	VERB
futech-324	88	21	approach	approach	NOUN
futech-324	88	22	for	for	ADP
futech-324	88	23	analyzing	analyze	VERB
futech-324	88	24	consumers	consumer	NOUN
futech-324	88	25	'	'	PART
futech-324	88	26	feelings	feeling	NOUN
futech-324	88	27	,	,	PUNCT
futech-324	88	28	which	which	PRON
futech-324	88	29	can	can	AUX
futech-324	88	30	help	help	VERB
futech-324	88	31	brands	brand	NOUN
futech-324	88	32	manage	manage	VERB
futech-324	88	33	their	their	PRON
futech-324	88	34	products	product	NOUN
futech-324	88	35	better	well	ADV
futech-324	88	36	.	.	PUNCT
futech-324	89	1	m.	m.	NOUN
futech-324	89	2	aamir	aamir	PROPN
futech-324	89	3	et	et	PROPN
futech-324	89	4	al	al	PROPN
futech-324	89	5	.	.	PUNCT
futech-324	90	1	[	[	X
futech-324	90	2	10	10	NUM
futech-324	90	3	]	]	PUNCT
futech-324	90	4	conducted	conduct	VERB
futech-324	90	5	a	a	DET
futech-324	90	6	similar	similar	ADJ
futech-324	90	7	study	study	NOUN
futech-324	90	8	using	use	VERB
futech-324	90	9	different	different	ADJ
futech-324	90	10	ml	ml	NOUN
futech-324	90	11	and	and	CCONJ
futech-324	90	12	dl	dl	PROPN
futech-324	90	13	techniques	technique	NOUN
futech-324	90	14	to	to	PART
futech-324	90	15	find	find	VERB
futech-324	90	16	public	public	ADJ
futech-324	90	17	sentiment	sentiment	NOUN
futech-324	90	18	on	on	ADP
futech-324	90	19	tweets	tweet	NOUN
futech-324	90	20	about	about	ADP
futech-324	90	21	ola	ola	PROPN
futech-324	90	22	and	and	CCONJ
futech-324	90	23	uber	uber	ADJ
futech-324	90	24	.	.	PUNCT
futech-324	91	1	in	in	ADP
futech-324	91	2	this	this	DET
futech-324	91	3	study	study	NOUN
futech-324	91	4	,	,	PUNCT
futech-324	91	5	different	different	ADJ
futech-324	91	6	algorithms	algorithm	NOUN
futech-324	91	7	were	be	AUX
futech-324	91	8	evaluated	evaluate	VERB
futech-324	91	9	to	to	PART
futech-324	91	10	understand	understand	VERB
futech-324	91	11	customer	customer	NOUN
futech-324	91	12	feedback	feedback	NOUN
futech-324	91	13	for	for	ADP
futech-324	91	14	these	these	DET
futech-324	91	15	ride	ride	NOUN
futech-324	91	16	-	-	PUNCT
futech-324	91	17	hailing	hail	VERB
futech-324	91	18	services	service	NOUN
futech-324	91	19	.	.	PUNCT
futech-324	92	1	the	the	DET
futech-324	92	2	study	study	NOUN
futech-324	92	3	assisted	assist	VERB
futech-324	92	4	in	in	ADP
futech-324	92	5	improving	improve	VERB
futech-324	92	6	the	the	DET
futech-324	92	7	accuracy	accuracy	NOUN
futech-324	92	8	of	of	ADP
futech-324	92	9	algorithms	algorithm	NOUN
futech-324	92	10	employed	employ	VERB
futech-324	92	11	and	and	CCONJ
futech-324	92	12	facilitated	facilitate	VERB
futech-324	92	13	companies	company	NOUN
futech-324	92	14	to	to	PART
futech-324	92	15	tailor	tailor	VERB
futech-324	92	16	their	their	PRON
futech-324	92	17	services	service	NOUN
futech-324	92	18	according	accord	VERB
futech-324	92	19	to	to	ADP
futech-324	92	20	user	user	NOUN
futech-324	92	21	sentiments	sentiment	NOUN
futech-324	92	22	,	,	PUNCT
futech-324	92	23	leading	lead	VERB
futech-324	92	24	to	to	ADP
futech-324	92	25	a	a	DET
futech-324	92	26	healthier	healthy	ADJ
futech-324	92	27	online	online	ADJ
futech-324	92	28	ecosystem	ecosystem	NOUN
futech-324	92	29	.	.	PUNCT
futech-324	93	1	q.	q.	PROPN
futech-324	93	2	zeng	zeng	PROPN
futech-324	93	3	et	et	PROPN
futech-324	93	4	al	al	PROPN
futech-324	93	5	.	.	PUNCT
futech-324	94	1	[	[	X
futech-324	94	2	11	11	NUM
futech-324	94	3	]	]	PUNCT
futech-324	94	4	proposed	propose	VERB
futech-324	94	5	a	a	DET
futech-324	94	6	neural	neural	ADJ
futech-324	94	7	network	network	NOUN
futech-324	94	8	model	model	NOUN
futech-324	94	9	for	for	ADP
futech-324	94	10	aspect	aspect	NOUN
futech-324	94	11	-	-	PUNCT
futech-324	94	12	based	base	VERB
futech-324	94	13	sentiment	sentiment	NOUN
futech-324	94	14	classification	classification	NOUN
futech-324	94	15	,	,	PUNCT
futech-324	94	16	integrating	integrate	VERB
futech-324	94	17	a	a	DET
futech-324	94	18	selfattention	selfattention	NOUN
futech-324	94	19	mechanism	mechanism	NOUN
futech-324	94	20	to	to	PART
futech-324	94	21	capture	capture	VERB
futech-324	94	22	contextual	contextual	ADJ
futech-324	94	23	semantic	semantic	ADJ
futech-324	94	24	information	information	NOUN
futech-324	94	25	.	.	PUNCT
futech-324	95	1	relative	relative	ADJ
futech-324	95	2	position	position	NOUN
futech-324	95	3	representations	representation	NOUN
futech-324	95	4	(	(	PUNCT
futech-324	95	5	prp	prp	PROPN
futech-324	95	6	)	)	PUNCT
futech-324	95	7	were	be	AUX
futech-324	95	8	produced	produce	VERB
futech-324	95	9	by	by	ADP
futech-324	95	10	attending	attend	VERB
futech-324	95	11	to	to	ADP
futech-324	95	12	a	a	DET
futech-324	95	13	global	global	ADJ
futech-324	95	14	context	context	NOUN
futech-324	95	15	,	,	PUNCT
futech-324	95	16	while	while	SCONJ
futech-324	95	17	pairs	pair	NOUN
futech-324	95	18	of	of	ADP
futech-324	95	19	words	word	NOUN
futech-324	95	20	were	be	AUX
futech-324	95	21	given	give	VERB
futech-324	95	22	relu	relu	NOUN
futech-324	95	23	gated	gate	VERB
futech-324	95	24	convolutional	convolutional	ADJ
futech-324	95	25	networks	network	NOUN
futech-324	95	26	for	for	ADP
futech-324	95	27	sentiment	sentiment	NOUN
futech-324	95	28	feature	feature	NOUN
futech-324	95	29	extraction	extraction	NOUN
futech-324	95	30	.	.	PUNCT
futech-324	96	1	it	it	PRON
futech-324	96	2	has	have	AUX
futech-324	96	3	also	also	ADV
futech-324	96	4	been	be	AUX
futech-324	96	5	shown	show	VERB
futech-324	96	6	that	that	SCONJ
futech-324	96	7	their	their	PRON
futech-324	96	8	model	model	NOUN
futech-324	96	9	outperforms	outperform	VERB
futech-324	96	10	every	every	DET
futech-324	96	11	other	other	ADJ
futech-324	96	12	model	model	NOUN
futech-324	96	13	in	in	ADP
futech-324	96	14	terms	term	NOUN
futech-324	96	15	of	of	ADP
futech-324	96	16	predicting	predict	VERB
futech-324	96	17	valence	valence	NOUN
futech-324	96	18	and	and	CCONJ
futech-324	96	19	arousal	arousal	NOUN
futech-324	96	20	on	on	ADP
futech-324	96	21	the	the	DET
futech-324	96	22	semeval	semeval	NOUN
futech-324	96	23	dataset	dataset	NOUN
futech-324	96	24	,	,	PUNCT
futech-324	96	25	tweets	tweet	NOUN
futech-324	96	26	,	,	PUNCT
futech-324	96	27	and	and	CCONJ
futech-324	96	28	cvat	cvat	PROPN
futech-324	96	29	,	,	PUNCT
futech-324	96	30	confirming	confirm	VERB
futech-324	96	31	its	its	PRON
futech-324	96	32	robustness	robustness	NOUN
futech-324	96	33	for	for	ADP
futech-324	96	34	sentiment	sentiment	NOUN
futech-324	96	35	analysis	analysis	NOUN
futech-324	96	36	.	.	PUNCT
futech-324	97	1	m.	m.	NOUN
futech-324	97	2	khalid	khalid	PROPN
futech-324	97	3	et	et	PROPN
futech-324	97	4	al	al	PROPN
futech-324	97	5	.	.	PUNCT
futech-324	98	1	[	[	X
futech-324	98	2	12	12	NUM
futech-324	98	3	]	]	PUNCT
futech-324	98	4	proposed	propose	VERB
futech-324	98	5	a	a	DET
futech-324	98	6	sentiment	sentiment	NOUN
futech-324	98	7	majority	majority	NOUN
futech-324	98	8	voting	voting	NOUN
futech-324	98	9	classifier	classifier	NOUN
futech-324	98	10	(	(	PUNCT
futech-324	98	11	smvc	smvc	PROPN
futech-324	98	12	)	)	PUNCT
futech-324	98	13	to	to	PART
futech-324	98	14	analyze	analyze	VERB
futech-324	98	15	the	the	DET
futech-324	98	16	sentiment	sentiment	NOUN
futech-324	98	17	of	of	ADP
futech-324	98	18	deepfake	deepfake	NOUN
futech-324	98	19	technology	technology	NOUN
futech-324	98	20	-	-	PUNCT
futech-324	98	21	related	relate	VERB
futech-324	98	22	tweets	tweet	NOUN
futech-324	98	23	.	.	PUNCT
futech-324	99	1	the	the	DET
futech-324	99	2	authors	author	NOUN
futech-324	99	3	used	use	VERB
futech-324	99	4	majority	majority	NOUN
futech-324	99	5	voting	vote	VERB
futech-324	99	6	to	to	PART
futech-324	99	7	aggregate	aggregate	VERB
futech-324	99	8	the	the	DET
futech-324	99	9	predictions	prediction	NOUN
futech-324	99	10	from	from	ADP
futech-324	99	11	several	several	ADJ
futech-324	99	12	lexicon	lexicon	NOUN
futech-324	99	13	-	-	PUNCT
futech-324	99	14	based	base	VERB
futech-324	99	15	models	model	NOUN
futech-324	99	16	and	and	CCONJ
futech-324	99	17	used	use	VERB
futech-324	99	18	transfer	transfer	NOUN
futech-324	99	19	learning	learn	VERB
futech-324	99	20	with	with	ADP
futech-324	99	21	lstm	lstm	NOUN
futech-324	99	22	and	and	CCONJ
futech-324	99	23	decision	decision	NOUN
futech-324	99	24	tree	tree	NOUN
futech-324	99	25	models	model	NOUN
futech-324	99	26	.	.	PUNCT
futech-324	100	1	the	the	DET
futech-324	100	2	method	method	NOUN
futech-324	100	3	reached	reach	VERB
futech-324	100	4	the	the	DET
futech-324	100	5	best	good	ADJ
futech-324	100	6	accuracy	accuracy	NOUN
futech-324	100	7	of	of	ADP
futech-324	100	8	98.9	98.9	NUM
futech-324	100	9	%	%	NOUN
futech-324	100	10	,	,	PUNCT
futech-324	100	11	demonstrating	demonstrate	VERB
futech-324	100	12	robustness	robustness	NOUN
futech-324	100	13	in	in	ADP
futech-324	100	14	sentiment	sentiment	NOUN
futech-324	100	15	classification	classification	NOUN
futech-324	100	16	.	.	PUNCT
futech-324	101	1	a.l.rao	a.l.rao	NOUN
futech-324	101	2	et	et	PROPN
futech-324	101	3	al	al	PROPN
futech-324	101	4	.	.	PUNCT
futech-324	102	1	[	[	X
futech-324	102	2	13	13	NUM
futech-324	102	3	]	]	PUNCT
futech-324	102	4	focused	focus	VERB
futech-324	102	5	on	on	ADP
futech-324	102	6	sentiment	sentiment	NOUN
futech-324	102	7	analysis	analysis	NOUN
futech-324	102	8	of	of	ADP
futech-324	102	9	the	the	DET
futech-324	102	10	airline	airline	NOUN
futech-324	102	11	tweet	tweet	NOUN
futech-324	102	12	or	or	CCONJ
futech-324	102	13	airline	airline	NOUN
futech-324	102	14	reviews	review	NOUN
futech-324	102	15	kaggle	kaggle	VERB
futech-324	102	16	dataset	dataset	NOUN
futech-324	102	17	.	.	PUNCT
futech-324	103	1	they	they	PRON
futech-324	103	2	suggested	suggest	VERB
futech-324	103	3	an	an	DET
futech-324	103	4	ensemble	ensemble	ADJ
futech-324	103	5	architecture	architecture	NOUN
futech-324	103	6	for	for	ADP
futech-324	103	7	cnn	cnn	PROPN
futech-324	103	8	and	and	CCONJ
futech-324	103	9	lstm	lstm	NOUN
futech-324	103	10	for	for	ADP
futech-324	103	11	improving	improve	VERB
futech-324	103	12	sentiment	sentiment	NOUN
futech-324	103	13	classification	classification	NOUN
futech-324	103	14	.	.	PUNCT
futech-324	104	1	they	they	PRON
futech-324	104	2	benchmarked	benchmarke	VERB
futech-324	104	3	this	this	DET
futech-324	104	4	ensemble	ensemble	ADJ
futech-324	104	5	model	model	NOUN
futech-324	104	6	against	against	ADP
futech-324	104	7	a	a	DET
futech-324	104	8	standalone	standalone	ADJ
futech-324	104	9	lstm	lstm	NOUN
futech-324	104	10	model	model	NOUN
futech-324	104	11	,	,	PUNCT
futech-324	104	12	which	which	PRON
futech-324	104	13	was	be	AUX
futech-324	104	14	also	also	ADV
futech-324	104	15	trained	train	VERB
futech-324	104	16	on	on	ADP
futech-324	104	17	the	the	DET
futech-324	104	18	same	same	ADJ
futech-324	104	19	dataset	dataset	NOUN
futech-324	104	20	,	,	PUNCT
futech-324	104	21	and	and	CCONJ
futech-324	104	22	reported	report	VERB
futech-324	104	23	the	the	DET
futech-324	104	24	performance	performance	NOUN
futech-324	104	25	of	of	ADP
futech-324	104	26	each	each	PRON
futech-324	104	27	and	and	CCONJ
futech-324	104	28	stated	state	VERB
futech-324	104	29	that	that	SCONJ
futech-324	104	30	the	the	DET
futech-324	104	31	ensemble	ensemble	ADJ
futech-324	104	32	approach	approach	NOUN
futech-324	104	33	provided	provide	VERB
futech-324	104	34	superior	superior	ADJ
futech-324	104	35	performance	performance	NOUN
futech-324	104	36	,	,	PUNCT
futech-324	104	37	providing	provide	VERB
futech-324	104	38	an	an	DET
futech-324	104	39	improvement	improvement	NOUN
futech-324	104	40	on	on	ADP
futech-324	104	41	standard	standard	ADJ
futech-324	104	42	methods	method	NOUN
futech-324	104	43	.	.	PUNCT
futech-324	105	1	y.	y.	PROPN
futech-324	105	2	matrane	matrane	PROPN
futech-324	105	3	et	et	PROPN
futech-324	105	4	al	al	PROPN
futech-324	105	5	.	.	PUNCT
futech-324	106	1	[	[	X
futech-324	106	2	14	14	NUM
futech-324	106	3	]	]	PUNCT
futech-324	106	4	performed	perform	VERB
futech-324	106	5	the	the	DET
futech-324	106	6	sentiment	sentiment	NOUN
futech-324	106	7	analysis	analysis	NOUN
futech-324	106	8	of	of	ADP
futech-324	106	9	md	md	PROPN
futech-324	106	10	,	,	PUNCT
futech-324	106	11	discussing	discuss	VERB
futech-324	106	12	specifically	specifically	ADV
futech-324	106	13	the	the	DET
futech-324	106	14	issues	issue	NOUN
futech-324	106	15	related	relate	VERB
futech-324	106	16	to	to	PART
futech-324	106	17	dialect	dialect	VERB
futech-324	106	18	-	-	PUNCT
futech-324	106	19	specific	specific	ADJ
futech-324	106	20	preprocessing	preprocessing	NOUN
futech-324	106	21	techniques	technique	NOUN
futech-324	106	22	.	.	PUNCT
futech-324	107	1	they	they	PRON
futech-324	107	2	reported	report	VERB
futech-324	107	3	better	well	ADJ
futech-324	107	4	results	result	NOUN
futech-324	107	5	by	by	ADP
futech-324	107	6	not	not	PART
futech-324	107	7	utilizing	utilize	VERB
futech-324	107	8	traditional	traditional	ADJ
futech-324	107	9	techniques	technique	NOUN
futech-324	107	10	like	like	ADP
futech-324	107	11	stemming	stem	VERB
futech-324	107	12	and	and	CCONJ
futech-324	107	13	by	by	ADP
futech-324	107	14	using	use	VERB
futech-324	107	15	the	the	DET
futech-324	107	16	qarib	qarib	NOUN
futech-324	107	17	feature	feature	NOUN
futech-324	107	18	extractor	extractor	NOUN
futech-324	107	19	with	with	ADP
futech-324	107	20	bigru	bigru	NOUN
futech-324	107	21	.	.	PUNCT
futech-324	108	1	the	the	DET
futech-324	108	2	results	result	NOUN
futech-324	108	3	of	of	ADP
futech-324	108	4	darijabert	darijabert	NOUN
futech-324	108	5	on	on	ADP
futech-324	108	6	the	the	DET
futech-324	108	7	fb	fb	NOUN
futech-324	108	8	dataset	dataset	NOUN
futech-324	108	9	showed	show	VERB
futech-324	108	10	their	their	PRON
futech-324	108	11	fine	fine	ADJ
futech-324	108	12	-	-	PUNCT
futech-324	108	13	tuning	tune	VERB
futech-324	108	14	approach	approach	NOUN
futech-324	108	15	to	to	PART
futech-324	108	16	be	be	AUX
futech-324	108	17	effective	effective	ADJ
futech-324	108	18	,	,	PUNCT
futech-324	108	19	signifying	signify	VERB
futech-324	108	20	the	the	DET
futech-324	108	21	need	need	NOUN
futech-324	108	22	for	for	ADP
futech-324	108	23	dialect	dialect	VERB
futech-324	108	24	-	-	PUNCT
futech-324	108	25	specific	specific	ADJ
futech-324	108	26	techniques	technique	NOUN
futech-324	108	27	in	in	ADP
futech-324	108	28	dealing	deal	VERB
futech-324	108	29	with	with	ADP
futech-324	108	30	the	the	DET
futech-324	108	31	arabic	arabic	ADJ
futech-324	108	32	dialect	dialect	NOUN
futech-324	108	33	.	.	PUNCT
futech-324	109	1	h.	h.	PROPN
futech-324	109	2	shuqin	shuqin	PROPN
futech-324	109	3	et	et	PROPN
futech-324	109	4	al	al	PROPN
futech-324	109	5	.	.	PUNCT
futech-324	110	1	[	[	X
futech-324	110	2	15	15	NUM
futech-324	110	3	]	]	PUNCT
futech-324	110	4	implemented	implement	VERB
futech-324	110	5	a	a	DET
futech-324	110	6	bert	bert	NOUN
futech-324	110	7	-	-	PUNCT
futech-324	110	8	bilstmattention	bilstmattention	NOUN
futech-324	110	9	(	(	PUNCT
futech-324	110	10	bba	bba	NOUN
futech-324	110	11	)	)	PUNCT
futech-324	110	12	model	model	NOUN
futech-324	110	13	to	to	PART
futech-324	110	14	recognize	recognize	VERB
futech-324	110	15	sentiment	sentiment	NOUN
futech-324	110	16	for	for	ADP
futech-324	110	17	course	course	NOUN
futech-324	110	18	evaluation	evaluation	NOUN
futech-324	110	19	.	.	PUNCT
futech-324	111	1	the	the	DET
futech-324	111	2	model	model	NOUN
futech-324	111	3	was	be	AUX
futech-324	111	4	built	build	VERB
futech-324	111	5	by	by	ADP
futech-324	111	6	leveraging	leverage	VERB
futech-324	111	7	the	the	DET
futech-324	111	8	bert	bert	PROPN
futech-324	111	9	model	model	NOUN
futech-324	111	10	for	for	ADP
futech-324	111	11	context	context	NOUN
futech-324	111	12	,	,	PUNCT
futech-324	111	13	along	along	ADP
futech-324	111	14	with	with	ADP
futech-324	111	15	bilstm	bilstm	NOUN
futech-324	111	16	and	and	CCONJ
futech-324	111	17	attention	attention	NOUN
futech-324	111	18	mechanisms	mechanism	NOUN
futech-324	111	19	to	to	PART
futech-324	111	20	better	well	ADV
futech-324	111	21	refine	refine	VERB
futech-324	111	22	the	the	DET
futech-324	111	23	focus	focus	NOUN
futech-324	111	24	across	across	ADP
futech-324	111	25	the	the	DET
futech-324	111	26	relevant	relevant	ADJ
futech-324	111	27	components	component	NOUN
futech-324	111	28	of	of	ADP
futech-324	111	29	text	text	NOUN
futech-324	111	30	.	.	PUNCT
futech-324	112	1	the	the	DET
futech-324	112	2	bba	bba	PROPN
futech-324	112	3	model	model	NOUN
futech-324	112	4	beat	beat	VERB
futech-324	112	5	the	the	DET
futech-324	112	6	existing	exist	VERB
futech-324	112	7	methods	method	NOUN
futech-324	112	8	,	,	PUNCT
futech-324	112	9	showing	show	VERB
futech-324	112	10	that	that	SCONJ
futech-324	112	11	the	the	DET
futech-324	112	12	deep	deep	ADJ
futech-324	112	13	semantic	semantic	ADJ
futech-324	112	14	representation	representation	NOUN
futech-324	112	15	of	of	ADP
futech-324	112	16	education	education	NOUN
futech-324	112	17	-	-	PUNCT
futech-324	112	18	based	base	VERB
futech-324	112	19	feedback	feedback	NOUN
futech-324	112	20	could	could	AUX
futech-324	112	21	be	be	AUX
futech-324	112	22	achieved	achieve	VERB
futech-324	112	23	through	through	ADP
futech-324	112	24	this	this	DET
futech-324	112	25	model	model	NOUN
futech-324	112	26	.	.	PUNCT
futech-324	113	1	a.	a.	NOUN
futech-324	113	2	he	he	PRON
futech-324	113	3	et	et	PROPN
futech-324	113	4	al	al	PROPN
futech-324	113	5	.	.	PUNCT
futech-324	114	1	[	[	X
futech-324	114	2	16	16	NUM
futech-324	114	3	]	]	PUNCT
futech-324	114	4	introduced	introduce	VERB
futech-324	114	5	a	a	DET
futech-324	114	6	novel	novel	ADJ
futech-324	114	7	deep	deep	ADJ
futech-324	114	8	tensor	tensor	NOUN
futech-324	114	9	evidence	evidence	NOUN
futech-324	114	10	fusion	fusion	NOUN
futech-324	114	11	network	network	NOUN
futech-324	114	12	for	for	ADP
futech-324	114	13	multimodal	multimodal	ADJ
futech-324	114	14	sentiment	sentiment	NOUN
futech-324	114	15	classification	classification	NOUN
futech-324	114	16	.	.	PUNCT
futech-324	115	1	they	they	PRON
futech-324	115	2	introduced	introduce	VERB
futech-324	115	3	a	a	DET
futech-324	115	4	joint	joint	ADJ
futech-324	115	5	view	view	NOUN
futech-324	115	6	scoring	score	VERB
futech-324	115	7	net	net	NOUN
futech-324	115	8	that	that	PRON
futech-324	115	9	integrates	integrate	VERB
futech-324	115	10	lstm	lstm	NOUN
futech-324	115	11	and	and	CCONJ
futech-324	115	12	tensor	tensor	NOUN
futech-324	115	13	neural	neural	ADJ
futech-324	115	14	networks	network	NOUN
futech-324	115	15	to	to	PART
futech-324	115	16	extract	extract	VERB
futech-324	115	17	the	the	DET
futech-324	115	18	intermodal	intermodal	ADJ
futech-324	115	19	and	and	CCONJ
futech-324	115	20	intramodal	intramodal	ADJ
futech-324	115	21	rich	rich	ADJ
futech-324	115	22	information	information	NOUN
futech-324	115	23	.	.	PUNCT
futech-324	116	1	they	they	PRON
futech-324	116	2	also	also	ADV
futech-324	116	3	proposed	propose	VERB
futech-324	116	4	a	a	DET
futech-324	116	5	temporal	temporal	ADJ
futech-324	116	6	cue	cue	NOUN
futech-324	116	7	evaluation	evaluation	NOUN
futech-324	116	8	network	network	NOUN
futech-324	116	9	based	base	VERB
futech-324	116	10	on	on	ADP
futech-324	116	11	temporal	temporal	ADJ
futech-324	116	12	granularity	granularity	NOUN
futech-324	116	13	and	and	CCONJ
futech-324	116	14	a	a	DET
futech-324	116	15	trustworthy	trustworthy	ADJ
futech-324	116	16	fusion	fusion	NOUN
futech-324	116	17	layer	layer	NOUN
futech-324	116	18	to	to	PART
futech-324	116	19	enhance	enhance	VERB
futech-324	116	20	the	the	DET
futech-324	116	21	accuracy	accuracy	NOUN
futech-324	116	22	and	and	CCONJ
futech-324	116	23	robustness	robustness	NOUN
futech-324	116	24	of	of	ADP
futech-324	116	25	decisions	decision	NOUN
futech-324	116	26	.	.	PUNCT
futech-324	117	1	on	on	ADP
futech-324	117	2	the	the	DET
futech-324	117	3	cmu	cmu	ADV
futech-324	117	4	-	-	PUNCT
futech-324	117	5	mosei	mosei	ADJ
futech-324	117	6	and	and	CCONJ
futech-324	117	7	cmu	cmu	ADV
futech-324	117	8	-	-	PUNCT
futech-324	117	9	mosi	mosi	NOUN
futech-324	117	10	datasets	dataset	NOUN
futech-324	117	11	,	,	PUNCT
futech-324	117	12	the	the	DET
futech-324	117	13	results	result	NOUN
futech-324	117	14	were	be	AUX
futech-324	117	15	better	well	ADJ
futech-324	117	16	than	than	ADP
futech-324	117	17	those	those	PRON
futech-324	117	18	of	of	ADP
futech-324	117	19	sota	sota	NOUN
futech-324	117	20	methods	method	NOUN
futech-324	117	21	.	.	PUNCT
futech-324	118	1	z.	z.	PROPN
futech-324	118	2	wang	wang	PROPN
futech-324	118	3	et	et	PROPN
futech-324	118	4	al	al	PROPN
futech-324	118	5	.	.	PUNCT
futech-324	119	1	[	[	X
futech-324	119	2	17	17	NUM
futech-324	119	3	]	]	PUNCT
futech-324	119	4	proposed	propose	VERB
futech-324	119	5	a	a	DET
futech-324	119	6	multi	multi	ADJ
futech-324	119	7	-	-	ADJ
futech-324	119	8	label	label	ADJ
futech-324	119	9	classification	classification	NOUN
futech-324	119	10	approach	approach	NOUN
futech-324	119	11	for	for	ADP
futech-324	119	12	handling	handle	VERB
futech-324	119	13	toxic	toxic	ADJ
futech-324	119	14	comments	comment	NOUN
futech-324	119	15	in	in	ADP
futech-324	119	16	social	social	ADJ
futech-324	119	17	media	medium	NOUN
futech-324	119	18	,	,	PUNCT
futech-324	119	19	specifically	specifically	ADV
futech-324	119	20	focusing	focus	VERB
futech-324	119	21	on	on	ADP
futech-324	119	22	indonesia	indonesia	PROPN
futech-324	119	23	's	's	PART
futech-324	119	24	twitter	twitter	NOUN
futech-324	119	25	platform	platform	NOUN
futech-324	119	26	.	.	PUNCT
futech-324	120	1	using	use	VERB
futech-324	120	2	two	two	NUM
futech-324	120	3	kinds	kind	NOUN
futech-324	120	4	of	of	ADP
futech-324	120	5	word	word	NOUN
futech-324	120	6	vectors	vector	NOUN
futech-324	120	7	from	from	ADP
futech-324	120	8	bert	bert	PROPN
futech-324	120	9	's	's	PART
futech-324	120	10	hidden	hidden	ADJ
futech-324	120	11	layers	layer	NOUN
futech-324	120	12	and	and	CCONJ
futech-324	120	13	combining	combine	VERB
futech-324	120	14	both	both	CCONJ
futech-324	120	15	bert	bert	PROPN
futech-324	120	16	+	+	CCONJ
futech-324	120	17	bilstm	bilstm	NOUN
futech-324	120	18	method	method	NOUN
futech-324	120	19	,	,	PUNCT
futech-324	120	20	the	the	DET
futech-324	120	21	model	model	NOUN
futech-324	120	22	achieved	achieve	VERB
futech-324	120	23	better	well	ADJ
futech-324	120	24	performance	performance	NOUN
futech-324	120	25	.	.	PUNCT
futech-324	121	1	in	in	ADP
futech-324	121	2	such	such	DET
futech-324	121	3	a	a	DET
futech-324	121	4	complex	complex	ADJ
futech-324	121	5	form	form	NOUN
futech-324	121	6	of	of	ADP
futech-324	121	7	sentiment	sentiment	NOUN
futech-324	121	8	analysis	analysis	NOUN
futech-324	121	9	task	task	NOUN
futech-324	121	10	,	,	PUNCT
futech-324	121	11	the	the	DET
futech-324	121	12	proposed	propose	VERB
futech-324	121	13	model	model	NOUN
futech-324	121	14	achieved	achieve	VERB
futech-324	121	15	an	an	DET
futech-324	121	16	accuracy	accuracy	NOUN
futech-324	121	17	of	of	ADP
futech-324	121	18	0.889	0.889	NUM
futech-324	121	19	,	,	PUNCT
futech-324	121	20	precision	precision	NOUN
futech-324	121	21	of	of	ADP
futech-324	121	22	0.925	0.925	NUM
futech-324	121	23	,	,	PUNCT
futech-324	121	24	recall	recall	NOUN
futech-324	121	25	of	of	ADP
futech-324	121	26	0.917	0.917	NUM
futech-324	121	27	,	,	PUNCT
futech-324	121	28	and	and	CCONJ
futech-324	121	29	f1	f1	ADJ
futech-324	121	30	score	score	NOUN
futech-324	121	31	of	of	ADP
futech-324	121	32	0.91	0.91	NUM
futech-324	121	33	,	,	PUNCT
futech-324	121	34	thus	thus	ADV
futech-324	121	35	proving	prove	VERB
futech-324	121	36	the	the	DET
futech-324	121	37	efficacy	efficacy	NOUN
futech-324	121	38	of	of	ADP
futech-324	121	39	task	task	NOUN
futech-324	121	40	-	-	PUNCT
futech-324	121	41	specific	specific	ADJ
futech-324	121	42	semantic	semantic	ADJ
futech-324	121	43	embedding	embedding	NOUN
futech-324	121	44	and	and	CCONJ
futech-324	121	45	sequential	sequential	ADJ
futech-324	121	46	learning	learning	NOUN
futech-324	121	47	through	through	ADP
futech-324	121	48	bilstm	bilstm	NOUN
futech-324	121	49	.	.	PUNCT
futech-324	122	1	s.	s.	PROPN
futech-324	122	2	k.	k.	PROPN
futech-324	122	3	putri	putri	PROPN
futech-324	122	4	et	et	PROPN
futech-324	122	5	al	al	PROPN
futech-324	122	6	.	.	PUNCT
futech-324	123	1	[	[	X
futech-324	123	2	18	18	NUM
futech-324	123	3	]	]	PUNCT
futech-324	123	4	applied	apply	VERB
futech-324	123	5	various	various	ADJ
futech-324	123	6	ml	ml	NOUN
futech-324	123	7	models	model	NOUN
futech-324	123	8	to	to	PART
futech-324	123	9	predict	predict	VERB
futech-324	123	10	toxic	toxic	ADJ
futech-324	123	11	leadership	leadership	NOUN
futech-324	123	12	in	in	ADP
futech-324	123	13	the	the	DET
futech-324	123	14	moroccan	moroccan	NOUN
futech-324	123	15	it	it	PRON
futech-324	123	16	sector	sector	NOUN
futech-324	123	17	.	.	PUNCT
futech-324	124	1	in	in	ADP
futech-324	124	2	this	this	DET
futech-324	124	3	regard	regard	NOUN
futech-324	124	4	,	,	PUNCT
futech-324	124	5	the	the	DET
futech-324	124	6	study	study	NOUN
futech-324	124	7	identified	identify	VERB
futech-324	124	8	undermining	undermine	VERB
futech-324	124	9	behavior	behavior	NOUN
futech-324	124	10	,	,	PUNCT
futech-324	124	11	narcissistic	narcissistic	ADJ
futech-324	124	12	traits	trait	NOUN
futech-324	124	13	,	,	PUNCT
futech-324	124	14	unjust	unjust	ADJ
futech-324	124	15	treatment	treatment	NOUN
futech-324	124	16	,	,	PUNCT
futech-324	124	17	and	and	CCONJ
futech-324	124	18	fear	fear	NOUN
futech-324	124	19	of	of	ADP
futech-324	124	20	retribution	retribution	NOUN
futech-324	124	21	as	as	ADP
futech-324	124	22	the	the	DET
futech-324	124	23	main	main	ADJ
futech-324	124	24	contributors	contributor	NOUN
futech-324	124	25	to	to	ADP
futech-324	124	26	toxic	toxic	ADJ
futech-324	124	27	leadership	leadership	NOUN
futech-324	124	28	,	,	PUNCT
futech-324	124	29	as	as	SCONJ
futech-324	124	30	it	it	PRON
futech-324	124	31	sought	seek	VERB
futech-324	124	32	to	to	PART
futech-324	124	33	provide	provide	VERB
futech-324	124	34	a	a	DET
futech-324	124	35	comprehensive	comprehensive	ADJ
futech-324	124	36	analysis	analysis	NOUN
futech-324	124	37	of	of	ADP
futech-324	124	38	how	how	SCONJ
futech-324	124	39	toxic	toxic	ADJ
futech-324	124	40	leadership	leadership	NOUN
futech-324	124	41	could	could	AUX
futech-324	124	42	be	be	AUX
futech-324	124	43	predicted	predict	VERB
futech-324	124	44	using	use	VERB
futech-324	124	45	different	different	ADJ
futech-324	124	46	types	type	NOUN
futech-324	124	47	of	of	ADP
futech-324	124	48	ml	ml	NOUN
futech-324	124	49	algorithms	algorithm	NOUN
futech-324	124	50	.	.	PUNCT
futech-324	125	1	a.	a.	PROPN
futech-324	125	2	lakshmanarao	lakshmanarao	PROPN
futech-324	125	3	et	et	PROPN
futech-324	125	4	al	al	PROPN
futech-324	125	5	.	.	PUNCT
futech-324	126	1	[	[	X
futech-324	126	2	19	19	NUM
futech-324	126	3	]	]	PUNCT
futech-324	126	4	sentiment	sentiment	NOUN
futech-324	126	5	analysis	analysis	NOUN
futech-324	126	6	of	of	ADP
futech-324	126	7	airlines	airline	NOUN
futech-324	126	8	tweet	tweet	PROPN
futech-324	126	9	gathered	gather	VERB
futech-324	126	10	from	from	ADP
futech-324	126	11	kaggle	kaggle	PROPN
futech-324	126	12	.	.	PUNCT
futech-324	127	1	they	they	PRON
futech-324	127	2	applied	apply	VERB
futech-324	127	3	different	different	ADJ
futech-324	127	4	neural	neural	ADJ
futech-324	127	5	network	network	NOUN
futech-324	127	6	approaches	approach	NOUN
futech-324	127	7	to	to	PART
futech-324	127	8	classify	classify	VERB
futech-324	127	9	tweets	tweet	NOUN
futech-324	127	10	into	into	ADP
futech-324	127	11	different	different	ADJ
futech-324	127	12	categories	category	NOUN
futech-324	127	13	.	.	PUNCT
futech-324	128	1	s.	s.	PROPN
futech-324	128	2	dutta	dutta	PROPN
futech-324	128	3	et	et	PROPN
futech-324	128	4	al	al	PROPN
futech-324	128	5	.	.	PUNCT
futech-324	129	1	[	[	X
futech-324	129	2	20	20	NUM
futech-324	129	3	]	]	PUNCT
futech-324	129	4	addressed	address	VERB
futech-324	129	5	the	the	DET
futech-324	129	6	toxic	toxic	ADJ
futech-324	129	7	comment	comment	NOUN
futech-324	129	8	detection	detection	NOUN
futech-324	129	9	problem	problem	NOUN
futech-324	129	10	in	in	ADP
futech-324	129	11	assamese	assamese	PROPN
futech-324	129	12	,	,	PUNCT
futech-324	129	13	a	a	DET
futech-324	129	14	morphologically	morphologically	ADV
futech-324	129	15	rich	rich	ADJ
futech-324	129	16	and	and	CCONJ
futech-324	129	17	ambiguous	ambiguous	ADJ
futech-324	129	18	language	language	NOUN
futech-324	129	19	,	,	PUNCT
futech-324	129	20	which	which	PRON
futech-324	129	21	poses	pose	VERB
futech-324	129	22	a	a	DET
futech-324	129	23	challenge	challenge	NOUN
futech-324	129	24	to	to	PART
futech-324	129	25	sentiment	sentiment	VERB
futech-324	129	26	analysis	analysis	NOUN
futech-324	129	27	.	.	PUNCT
futech-324	130	1	this	this	DET
futech-324	130	2	paper	paper	NOUN
futech-324	130	3	is	be	AUX
futech-324	130	4	part	part	NOUN
futech-324	130	5	of	of	ADP
futech-324	130	6	a	a	DET
futech-324	130	7	very	very	ADV
futech-324	130	8	large	large	ADJ
futech-324	130	9	study	study	NOUN
futech-324	130	10	of	of	ADP
futech-324	130	11	general	general	ADJ
futech-324	130	12	nmf	nmf	PROPN
futech-324	130	13	topic	topic	NOUN
futech-324	130	14	modeling	modeling	NOUN
futech-324	130	15	,	,	PUNCT
futech-324	130	16	which	which	PRON
futech-324	130	17	indicates	indicate	VERB
futech-324	130	18	preceding	precede	VERB
futech-324	130	19	work	work	NOUN
futech-324	130	20	by	by	ADP
futech-324	130	21	manually	manually	ADV
futech-324	130	22	collecting	collect	VERB
futech-324	130	23	19,550	19,550	NUM
futech-324	130	24	comments	comment	NOUN
futech-324	130	25	from	from	ADP
futech-324	130	26	social	social	ADJ
futech-324	130	27	media	medium	NOUN
futech-324	130	28	sites	site	NOUN
futech-324	130	29	and	and	CCONJ
futech-324	130	30	testing	test	VERB
futech-324	130	31	a	a	DET
futech-324	130	32	number	number	NOUN
futech-324	130	33	of	of	ADP
futech-324	130	34	ml	ml	NOUN
futech-324	130	35	models	model	NOUN
futech-324	130	36	.	.	PUNCT
futech-324	131	1	all	all	DET
futech-324	131	2	these	these	DET
futech-324	131	3	models	model	NOUN
futech-324	131	4	have	have	AUX
futech-324	131	5	been	be	AUX
futech-324	131	6	trained	train	VERB
futech-324	131	7	against	against	ADP
futech-324	131	8	various	various	ADJ
futech-324	131	9	text	text	NOUN
futech-324	131	10	representations	representation	NOUN
futech-324	131	11	(	(	PUNCT
futech-324	131	12	count	count	NOUN
futech-324	131	13	vector	vector	NOUN
futech-324	131	14	,	,	PUNCT
futech-324	131	15	count	count	NOUN
futech-324	131	16	vector	vector	NOUN
futech-324	131	17	+	+	CCONJ
futech-324	131	18	tf	tf	PROPN
futech-324	131	19	-	-	PUNCT
futech-324	131	20	idf	idf	PROPN
futech-324	131	21	,	,	PUNCT
futech-324	131	22	n	n	CCONJ
futech-324	131	23	-	-	PUNCT
futech-324	131	24	gram	gram	NOUN
futech-324	131	25	,	,	PUNCT
futech-324	131	26	etc	etc	X
futech-324	131	27	.	.	X
futech-324	131	28	)	)	PUNCT
futech-324	132	1	the	the	DET
futech-324	132	2	best	good	ADJ
futech-324	132	3	f1	f1	NOUN
futech-324	132	4	-	-	PUNCT
futech-324	132	5	score	score	NOUN
futech-324	132	6	was	be	AUX
futech-324	132	7	94	94	NUM
futech-324	132	8	%	%	NOUN
futech-324	132	9	with	with	ADP
futech-324	132	10	svm	svm	PROPN
futech-324	132	11	+	+	CCONJ
futech-324	132	12	count	count	NOUN
futech-324	132	13	vector	vector	NOUN
futech-324	132	14	+	+	CCONJ
futech-324	132	15	tf	tf	PROPN
futech-324	132	16	-	-	PUNCT
futech-324	132	17	idf	idf	PROPN
futech-324	132	18	compared	compare	VERB
futech-324	132	19	to	to	ADP
futech-324	132	20	the	the	DET
futech-324	132	21	other	other	ADJ
futech-324	132	22	models	model	NOUN
futech-324	132	23	.	.	PUNCT
futech-324	133	1	y.	y.	PROPN
futech-324	133	2	mamani	mamani	PROPN
futech-324	133	3	et	et	PROPN
futech-324	133	4	al	al	PROPN
futech-324	133	5	.	.	PUNCT
futech-324	134	1	[	[	X
futech-324	134	2	21	21	NUM
futech-324	134	3	]	]	PUNCT
futech-324	134	4	provided	provide	VERB
futech-324	134	5	a	a	DET
futech-324	134	6	summary	summary	NOUN
futech-324	134	7	of	of	ADP
futech-324	134	8	recent	recent	ADJ
futech-324	134	9	ml	ml	NOUN
futech-324	134	10	techniques	technique	NOUN
futech-324	134	11	used	use	VERB
futech-324	134	12	for	for	ADP
futech-324	134	13	sentiment	sentiment	NOUN
futech-324	134	14	analysis	analysis	NOUN
futech-324	134	15	,	,	PUNCT
futech-324	134	16	developing	develop	VERB
futech-324	134	17	a	a	DET
futech-324	134	18	framework	framework	NOUN
futech-324	134	19	to	to	PART
futech-324	134	20	classify	classify	VERB
futech-324	134	21	sentiment	sentiment	NOUN
futech-324	134	22	models	model	NOUN
futech-324	134	23	according	accord	VERB
futech-324	134	24	to	to	ADP
futech-324	134	25	their	their	PRON
futech-324	134	26	structure	structure	NOUN
futech-324	134	27	.	.	PUNCT
futech-324	135	1	the	the	DET
futech-324	135	2	paper	paper	NOUN
futech-324	135	3	also	also	ADV
futech-324	135	4	discussed	discuss	VERB
futech-324	135	5	challenges	challenge	NOUN
futech-324	135	6	faced	face	VERB
futech-324	135	7	by	by	ADP
futech-324	135	8	the	the	DET
futech-324	135	9	community	community	NOUN
futech-324	135	10	and	and	CCONJ
futech-324	135	11	emerging	emerge	VERB
futech-324	135	12	trends	trend	NOUN
futech-324	135	13	,	,	PUNCT
futech-324	135	14	providing	provide	VERB
futech-324	135	15	future	future	ADJ
futech-324	135	16	directions	direction	NOUN
futech-324	135	17	for	for	ADP
futech-324	135	18	the	the	DET
futech-324	135	19	research	research	NOUN
futech-324	135	20	in	in	ADP
futech-324	135	21	sentiment	sentiment	NOUN
futech-324	135	22	analysis	analysis	NOUN
futech-324	135	23	.	.	PUNCT
futech-324	136	1	rahul	rahul	PROPN
futech-324	136	2	et	et	PROPN
futech-324	136	3	al	al	PROPN
futech-324	136	4	.	.	PUNCT
futech-324	137	1	[	[	X
futech-324	137	2	22	22	NUM
futech-324	137	3	]	]	PUNCT
futech-324	137	4	addressed	address	VERB
futech-324	137	5	the	the	DET
futech-324	137	6	classification	classification	NOUN
futech-324	137	7	problem	problem	NOUN
futech-324	137	8	for	for	ADP
futech-324	137	9	toxic	toxic	ADJ
futech-324	137	10	comments	comment	NOUN
futech-324	137	11	,	,	PUNCT
futech-324	137	12	which	which	PRON
futech-324	137	13	can	can	AUX
futech-324	137	14	be	be	AUX
futech-324	137	15	used	use	VERB
futech-324	137	16	to	to	PART
futech-324	137	17	measure	measure	VERB
futech-324	137	18	the	the	DET
futech-324	137	19	severity	severity	NOUN
futech-324	137	20	of	of	ADP
futech-324	137	21	online	online	ADJ
futech-324	137	22	harassment	harassment	NOUN
futech-324	137	23	.	.	PUNCT
futech-324	138	1	they	they	PRON
futech-324	138	2	examined	examine	VERB
futech-324	138	3	online	online	ADJ
futech-324	138	4	comments	comment	NOUN
futech-324	138	5	without	without	ADP
futech-324	138	6	a	a	DET
futech-324	138	7	focus	focus	NOUN
futech-324	138	8	on	on	ADP
futech-324	138	9	toxicity	toxicity	NOUN
futech-324	138	10	using	use	VERB
futech-324	138	11	six	six	NUM
futech-324	138	12	machine	machine	NOUN
futech-324	138	13	learning	learning	NOUN
futech-324	138	14	(	(	PUNCT
futech-324	138	15	ml	ml	NOUN
futech-324	138	16	)	)	PUNCT
futech-324	138	17	algorithms	algorithm	NOUN
futech-324	138	18	.	.	PUNCT
futech-324	139	1	they	they	PRON
futech-324	139	2	worked	work	VERB
futech-324	139	3	to	to	PART
futech-324	139	4	enhance	enhance	VERB
futech-324	139	5	the	the	DET
futech-324	139	6	classification	classification	NOUN
futech-324	139	7	of	of	ADP
futech-324	139	8	negative	negative	ADJ
futech-324	139	9	information	information	NOUN
futech-324	139	10	within	within	ADP
futech-324	139	11	textual	textual	ADJ
futech-324	139	12	comments	comment	NOUN
futech-324	139	13	,	,	PUNCT
futech-324	139	14	utilizing	utilize	VERB
futech-324	139	15	an	an	DET
futech-324	139	16	enhanced	enhanced	ADJ
futech-324	139	17	classification	classification	NOUN
futech-324	139	18	to	to	PART
futech-324	139	19	mitigate	mitigate	VERB
futech-324	139	20	harmful	harmful	ADJ
futech-324	139	21	information	information	NOUN
futech-324	139	22	.	.	PUNCT
futech-324	140	1	this	this	PRON
futech-324	140	2	enables	enable	VERB
futech-324	140	3	organizations	organization	NOUN
futech-324	140	4	to	to	PART
futech-324	140	5	identify	identify	VERB
futech-324	140	6	toxins	toxin	NOUN
futech-324	140	7	in	in	ADP
futech-324	140	8	discussions	discussion	NOUN
futech-324	140	9	and	and	CCONJ
futech-324	140	10	take	take	VERB
futech-324	140	11	action	action	NOUN
futech-324	140	12	to	to	PART
futech-324	140	13	lead	lead	VERB
futech-324	140	14	to	to	ADP
futech-324	140	15	better	well	ADJ
futech-324	140	16	environments	environment	NOUN
futech-324	140	17	.	.	PUNCT
futech-324	141	1	m.	m.	NOUN
futech-324	141	2	aquino	aquino	PROPN
futech-324	141	3	et	et	PROPN
futech-324	141	4	al	al	PROPN
futech-324	141	5	.	.	PUNCT
futech-324	142	1	in	in	ADP
futech-324	142	2	ref	ref	NOUN
futech-324	142	3	[	[	X
futech-324	142	4	23	23	NUM
futech-324	142	5	]	]	PUNCT
futech-324	142	6	,	,	PUNCT
futech-324	142	7	the	the	DET
futech-324	142	8	authors	author	NOUN
futech-324	142	9	explored	explore	VERB
futech-324	142	10	a	a	DET
futech-324	142	11	new	new	ADJ
futech-324	142	12	ml	ml	NOUN
futech-324	142	13	-	-	PUNCT
futech-324	142	14	based	base	VERB
futech-324	142	15	method	method	NOUN
futech-324	142	16	to	to	PART
futech-324	142	17	detect	detect	VERB
futech-324	142	18	comment	comment	NOUN
futech-324	142	19	toxicity	toxicity	NOUN
futech-324	142	20	using	use	VERB
futech-324	142	21	text	text	NOUN
futech-324	142	22	and	and	CCONJ
futech-324	142	23	emojis	emoji	NOUN
futech-324	142	24	.	.	PUNCT
futech-324	143	1	they	they	PRON
futech-324	143	2	trained	train	VERB
futech-324	143	3	a	a	DET
futech-324	143	4	bidirectional	bidirectional	ADJ
futech-324	143	5	lstm	lstm	NOUN
futech-324	143	6	model	model	NOUN
futech-324	143	7	using	use	VERB
futech-324	143	8	glove	glove	NOUN
futech-324	143	9	and	and	CCONJ
futech-324	143	10	emoji2vec	emoji2vec	NOUN
futech-324	143	11	combined	combine	VERB
futech-324	143	12	word	word	NOUN
futech-324	143	13	embeddings	embedding	NOUN
futech-324	143	14	.	.	PUNCT
futech-324	144	1	it	it	PRON
futech-324	144	2	provides	provide	VERB
futech-324	144	3	a	a	DET
futech-324	144	4	new	new	ADJ
futech-324	144	5	labeled	label	VERB
futech-324	144	6	dataset	dataset	NOUN
futech-324	144	7	of	of	ADP
futech-324	144	8	text	text	NOUN
futech-324	144	9	and	and	CCONJ
futech-324	144	10	emoji	emoji	PROPN
futech-324	144	11	data	data	PROPN
futech-324	144	12	,	,	PUNCT
futech-324	144	13	providing	provide	VERB
futech-324	144	14	an	an	DET
futech-324	144	15	effective	effective	ADJ
futech-324	144	16	means	mean	NOUN
futech-324	144	17	of	of	ADP
futech-324	144	18	comment	comment	NOUN
futech-324	144	19	toxicity	toxicity	NOUN
futech-324	144	20	detection	detection	NOUN
futech-324	144	21	.	.	PUNCT
futech-324	145	1	t.	t.	PROPN
futech-324	145	2	v.	v.	PROPN
futech-324	145	3	sai	sai	PROPN
futech-324	145	4	krishna	krishna	PROPN
futech-324	145	5	et	et	PROPN
futech-324	145	6	al	al	PROPN
futech-324	145	7	.	.	PUNCT
futech-324	145	8	using	use	VERB
futech-324	145	9	a	a	DET
futech-324	145	10	range	range	NOUN
futech-324	145	11	of	of	ADP
futech-324	145	12	features	feature	NOUN
futech-324	145	13	and	and	CCONJ
futech-324	145	14	both	both	DET
futech-324	145	15	ml	ml	NOUN
futech-324	145	16	&	&	CCONJ
futech-324	145	17	dl	dl	PROPN
futech-324	145	18	methods	method	NOUN
futech-324	145	19	,	,	PUNCT
futech-324	145	20	ref	ref	X
futech-324	146	1	[	[	X
futech-324	146	2	24	24	NUM
futech-324	146	3	]	]	PUNCT
futech-324	146	4	proposes	propose	VERB
futech-324	146	5	sentiment	sentiment	NOUN
futech-324	146	6	classification	classification	NOUN
futech-324	146	7	on	on	ADP
futech-324	146	8	twitter	twitter	NOUN
futech-324	146	9	.	.	PUNCT
futech-324	147	1	they	they	PRON
futech-324	147	2	used	use	VERB
futech-324	147	3	several	several	ADJ
futech-324	147	4	ml	ml	NOUN
futech-324	147	5	models	model	NOUN
futech-324	147	6	.	.	PUNCT
futech-324	148	1	they	they	PRON
futech-324	148	2	also	also	ADV
futech-324	148	3	proposed	propose	VERB
futech-324	148	4	a	a	DET
futech-324	148	5	new	new	ADJ
futech-324	148	6	end	end	NOUN
futech-324	148	7	-	-	PUNCT
futech-324	148	8	to	to	ADP
futech-324	148	9	-	-	PUNCT
futech-324	148	10	end	end	NOUN
futech-324	148	11	ensemble	ensemble	ADJ
futech-324	148	12	approach	approach	NOUN
futech-324	148	13	of	of	ADP
futech-324	148	14	using	use	VERB
futech-324	148	15	ml	ml	NOUN
futech-324	148	16	and	and	CCONJ
futech-324	148	17	dl	dl	PROPN
futech-324	148	18	models	model	NOUN
futech-324	148	19	,	,	PUNCT
futech-324	148	20	which	which	PRON
futech-324	148	21	yields	yield	VERB
futech-324	148	22	higher	high	ADJ
futech-324	148	23	accuracy	accuracy	NOUN
futech-324	148	24	vs.	vs.	ADP
futech-324	148	25	these	these	DET
futech-324	148	26	techniques	technique	NOUN
futech-324	148	27	when	when	SCONJ
futech-324	148	28	applied	apply	VERB
futech-324	148	29	to	to	ADP
futech-324	148	30	i.e.	i.e.	X
futech-324	148	31	,	,	PUNCT
futech-324	148	32	realtime	realtime	ADJ
futech-324	148	33	twitter	twitter	NOUN
futech-324	148	34	data	datum	NOUN
futech-324	148	35	for	for	ADP
futech-324	148	36	sentiment	sentiment	NOUN
futech-324	148	37	classification	classification	NOUN
futech-324	148	38	.	.	PUNCT
futech-324	149	1	for	for	ADP
futech-324	149	2	example	example	NOUN
futech-324	149	3	,	,	PUNCT
futech-324	149	4	singh	singh	PROPN
futech-324	149	5	et	et	PROPN
futech-324	149	6	al	al	PROPN
futech-324	149	7	.	.	PUNCT
futech-324	150	1	[	[	X
futech-324	150	2	25	25	NUM
futech-324	150	3	]	]	PUNCT
futech-324	150	4	detected	detect	VERB
futech-324	150	5	levels	level	NOUN
futech-324	150	6	of	of	ADP
futech-324	150	7	toxicity	toxicity	NOUN
futech-324	150	8	in	in	ADP
futech-324	150	9	comments	comment	NOUN
futech-324	150	10	on	on	ADP
futech-324	150	11	social	social	ADJ
futech-324	150	12	media	medium	NOUN
futech-324	150	13	using	use	VERB
futech-324	150	14	the	the	DET
futech-324	150	15	jigsaw	jigsaw	NOUN
futech-324	150	16	dataset	dataset	VERB
futech-324	150	17	from	from	ADP
futech-324	150	18	google	google	PROPN
futech-324	150	19	.	.	PUNCT
futech-324	151	1	the	the	DET
futech-324	151	2	dataset	dataset	NOUN
futech-324	151	3	is	be	AUX
futech-324	151	4	a	a	DET
futech-324	151	5	multilabel	multilabel	NOUN
futech-324	151	6	classification	classification	NOUN
futech-324	151	7	task	task	NOUN
futech-324	151	8	with	with	ADP
futech-324	151	9	several	several	ADJ
futech-324	151	10	classes	class	NOUN
futech-324	151	11	.	.	PUNCT
futech-324	152	1	the	the	DET
futech-324	152	2	logistic	logistic	ADJ
futech-324	152	3	regression	regression	NOUN
futech-324	152	4	model	model	NOUN
futech-324	152	5	performed	perform	VERB
futech-324	152	6	the	the	DET
futech-324	152	7	best	good	ADJ
futech-324	152	8	in	in	ADP
futech-324	152	9	terms	term	NOUN
futech-324	152	10	of	of	ADP
futech-324	152	11	accuracy	accuracy	NOUN
futech-324	152	12	and	and	CCONJ
futech-324	152	13	hamming	ham	VERB
futech-324	152	14	loss	loss	NOUN
futech-324	152	15	among	among	ADP
futech-324	152	16	the	the	DET
futech-324	152	17	other	other	ADJ
futech-324	152	18	models	model	NOUN
futech-324	152	19	,	,	PUNCT
futech-324	152	20	indicating	indicate	VERB
futech-324	152	21	that	that	SCONJ
futech-324	152	22	it	it	PRON
futech-324	152	23	is	be	AUX
futech-324	152	24	well	well	ADV
futech-324	152	25	-	-	PUNCT
futech-324	152	26	suited	suit	VERB
futech-324	152	27	to	to	ADP
futech-324	152	28	the	the	DET
futech-324	152	29	task	task	NOUN
futech-324	152	30	of	of	ADP
futech-324	152	31	s.	s.	PROPN
futech-324	152	32	sushma	sushma	PROPN
futech-324	152	33	et	et	PROPN
futech-324	152	34	al	al	PROPN
futech-324	152	35	.	.	PROPN
futech-324	152	36	/future	/future	PUNCT
futech-324	152	37	technology	technology	PROPN
futech-324	152	38	august	august	PROPN
futech-324	152	39	2025|	2025|	NUM
futech-324	153	1	volume	volume	NOUN
futech-324	153	2	04	04	NUM
futech-324	154	1	|	|	NOUN
futech-324	154	2	issue	issue	NOUN
futech-324	154	3	03	03	NUM
futech-324	155	1	|	|	ADV
futech-324	155	2	pages	page	VERB
futech-324	155	3	76	76	NUM
futech-324	155	4	-	-	SYM
futech-324	155	5	84	84	NUM
futech-324	155	6	79	79	NUM
futech-324	155	7	identifying	identify	VERB
futech-324	155	8	toxic	toxic	ADJ
futech-324	155	9	comments	comment	NOUN
futech-324	155	10	.	.	PUNCT
futech-324	156	1	venugopal	venugopal	PROPN
futech-324	156	2	[	[	X
futech-324	156	3	26	26	NUM
futech-324	156	4	]	]	PUNCT
futech-324	156	5	explored	explore	VERB
futech-324	156	6	the	the	DET
futech-324	156	7	classification	classification	NOUN
futech-324	156	8	of	of	ADP
futech-324	156	9	toxic	toxic	ADJ
futech-324	156	10	comments	comment	NOUN
futech-324	156	11	and	and	CCONJ
futech-324	156	12	identified	identify	VERB
futech-324	156	13	challenges	challenge	NOUN
futech-324	156	14	of	of	ADP
futech-324	156	15	detecting	detect	VERB
futech-324	156	16	the	the	DET
futech-324	156	17	toxicity	toxicity	NOUN
futech-324	156	18	of	of	ADP
futech-324	156	19	comments	comment	NOUN
futech-324	156	20	in	in	ADP
futech-324	156	21	multiple	multiple	ADJ
futech-324	156	22	languages	language	NOUN
futech-324	156	23	in	in	ADP
futech-324	156	24	a	a	DET
futech-324	156	25	centralized	centralized	ADJ
futech-324	156	26	system	system	NOUN
futech-324	156	27	for	for	ADP
futech-324	156	28	comment	comment	NOUN
futech-324	156	29	detection	detection	NOUN
futech-324	156	30	across	across	ADP
futech-324	156	31	social	social	ADJ
futech-324	156	32	media	medium	NOUN
futech-324	156	33	platforms	platform	NOUN
futech-324	156	34	.	.	PUNCT
futech-324	157	1	they	they	PRON
futech-324	157	2	used	use	VERB
futech-324	157	3	state	state	NOUN
futech-324	157	4	-	-	PUNCT
futech-324	157	5	of	of	ADP
futech-324	157	6	-	-	PUNCT
futech-324	157	7	the	the	DET
futech-324	157	8	-	-	PUNCT
futech-324	157	9	art	art	NOUN
futech-324	157	10	dl	dl	PROPN
futech-324	157	11	architectures	architecture	NOUN
futech-324	157	12	such	such	ADJ
futech-324	157	13	as	as	ADP
futech-324	157	14	bert	bert	PROPN
futech-324	157	15	or	or	CCONJ
futech-324	157	16	xlm	xlm	PROPN
futech-324	157	17	-	-	PUNCT
futech-324	157	18	roberta	roberta	PROPN
futech-324	157	19	for	for	ADP
futech-324	157	20	multilingual	multilingual	ADJ
futech-324	157	21	toxicity	toxicity	NOUN
futech-324	157	22	detection	detection	NOUN
futech-324	157	23	.	.	PUNCT
futech-324	158	1	the	the	DET
futech-324	158	2	paper	paper	NOUN
futech-324	158	3	noted	note	VERB
futech-324	158	4	that	that	SCONJ
futech-324	158	5	these	these	DET
futech-324	158	6	models	model	NOUN
futech-324	158	7	,	,	PUNCT
futech-324	158	8	when	when	SCONJ
futech-324	158	9	paired	pair	VERB
futech-324	158	10	with	with	ADP
futech-324	158	11	appropriate	appropriate	ADJ
futech-324	158	12	preprocessing	preprocessing	NOUN
futech-324	158	13	of	of	ADP
futech-324	158	14	datasets	dataset	NOUN
futech-324	158	15	and	and	CCONJ
futech-324	158	16	tuning	tuning	NOUN
futech-324	158	17	of	of	ADP
futech-324	158	18	hyperparameters	hyperparameter	NOUN
futech-324	158	19	,	,	PUNCT
futech-324	158	20	can	can	AUX
futech-324	158	21	provide	provide	VERB
futech-324	158	22	significant	significant	ADJ
futech-324	158	23	improvement	improvement	NOUN
futech-324	158	24	in	in	ADP
futech-324	158	25	the	the	DET
futech-324	158	26	accuracy	accuracy	NOUN
futech-324	158	27	of	of	ADP
futech-324	158	28	detecting	detect	VERB
futech-324	158	29	toxic	toxic	ADJ
futech-324	158	30	comments	comment	NOUN
futech-324	158	31	compared	compare	VERB
futech-324	158	32	to	to	ADP
futech-324	158	33	traditional	traditional	ADJ
futech-324	158	34	models	model	NOUN
futech-324	158	35	such	such	ADJ
futech-324	158	36	as	as	ADP
futech-324	158	37	svm	svm	PROPN
futech-324	158	38	.	.	PROPN
futech-324	158	39	n.	n.	PROPN
futech-324	158	40	boudjani	boudjani	PROPN
futech-324	158	41	et	et	PROPN
futech-324	158	42	al	al	PROPN
futech-324	158	43	.	.	PUNCT
futech-324	159	1	[	[	X
futech-324	159	2	27	27	NUM
futech-324	159	3	]	]	PUNCT
futech-324	159	4	used	use	VERB
futech-324	159	5	n	n	PRON
futech-324	159	6	-	-	PUNCT
futech-324	159	7	grams	gram	NOUN
futech-324	159	8	,	,	PUNCT
futech-324	159	9	linguistic	linguistic	ADJ
futech-324	159	10	features	feature	NOUN
futech-324	159	11	,	,	PUNCT
futech-324	159	12	and	and	CCONJ
futech-324	159	13	a	a	DET
futech-324	159	14	lexicon	lexicon	NOUN
futech-324	159	15	of	of	ADP
futech-324	159	16	insulting	insult	VERB
futech-324	159	17	words	word	NOUN
futech-324	159	18	to	to	PART
futech-324	159	19	create	create	VERB
futech-324	159	20	a	a	DET
futech-324	159	21	supervised	supervised	ADJ
futech-324	159	22	method	method	NOUN
futech-324	159	23	for	for	ADP
futech-324	159	24	the	the	DET
futech-324	159	25	classification	classification	NOUN
futech-324	159	26	of	of	ADP
futech-324	159	27	french	french	ADJ
futech-324	159	28	toxic	toxic	ADJ
futech-324	159	29	comments	comment	NOUN
futech-324	159	30	.	.	PUNCT
futech-324	160	1	using	use	VERB
futech-324	160	2	linear	linear	ADJ
futech-324	160	3	svm	svm	NOUN
futech-324	160	4	and	and	CCONJ
futech-324	160	5	decision	decision	NOUN
futech-324	160	6	tree	tree	NOUN
futech-324	160	7	classifiers	classifier	NOUN
futech-324	160	8	,	,	PUNCT
futech-324	160	9	their	their	PRON
futech-324	160	10	approach	approach	NOUN
futech-324	160	11	yielded	yield	VERB
futech-324	160	12	precision	precision	NOUN
futech-324	160	13	,	,	PUNCT
futech-324	160	14	recall	recall	NOUN
futech-324	160	15	,	,	PUNCT
futech-324	160	16	and	and	CCONJ
futech-324	160	17	f1	f1	ADJ
futech-324	160	18	-	-	PUNCT
futech-324	160	19	score	score	NOUN
futech-324	160	20	values	value	NOUN
futech-324	160	21	of	of	ADP
futech-324	160	22	87	87	NUM
futech-324	160	23	%	%	NOUN
futech-324	160	24	,	,	PUNCT
futech-324	160	25	83	83	NUM
futech-324	160	26	%	%	NOUN
futech-324	160	27	,	,	PUNCT
futech-324	160	28	and	and	CCONJ
futech-324	160	29	78	78	NUM
futech-324	160	30	%	%	NOUN
futech-324	160	31	,	,	PUNCT
futech-324	160	32	respectively	respectively	ADV
futech-324	160	33	.	.	PUNCT
futech-324	161	1	a.	a.	PROPN
futech-324	161	2	jessica	jessica	PROPN
futech-324	161	3	et	et	PROPN
futech-324	161	4	al	al	PROPN
futech-324	161	5	.	.	PROPN
futech-324	161	6	applied	apply	VERB
futech-324	161	7	bert	bert	PROPN
futech-324	161	8	-	-	PUNCT
futech-324	161	9	cnn	cnn	PROPN
futech-324	161	10	and	and	CCONJ
futech-324	161	11	bert	bert	PROPN
futech-324	161	12	-	-	PUNCT
futech-324	161	13	lstm	lstm	ADJ
futech-324	161	14	hybrid	hybrid	NOUN
futech-324	161	15	models	model	NOUN
futech-324	161	16	for	for	ADP
futech-324	161	17	detecting	detect	VERB
futech-324	161	18	cyberbullying	cyberbullye	VERB
futech-324	161	19	in	in	ADP
futech-324	161	20	online	online	ADJ
futech-324	161	21	comments	comment	NOUN
futech-324	161	22	.	.	PUNCT
futech-324	162	1	this	this	DET
futech-324	162	2	model	model	NOUN
futech-324	162	3	consists	consist	VERB
futech-324	162	4	of	of	ADP
futech-324	162	5	bert	bert	PROPN
futech-324	162	6	and	and	CCONJ
futech-324	162	7	cnn	cnn	PROPN
futech-324	162	8	,	,	PUNCT
futech-324	162	9	which	which	PRON
futech-324	162	10	are	be	AUX
futech-324	162	11	created	create	VERB
futech-324	162	12	through	through	ADP
futech-324	162	13	the	the	DET
futech-324	162	14	combination	combination	NOUN
futech-324	162	15	of	of	ADP
futech-324	162	16	sentences	sentence	NOUN
futech-324	162	17	by	by	ADP
futech-324	162	18	using	use	VERB
futech-324	162	19	bert	bert	PROPN
futech-324	162	20	so	so	SCONJ
futech-324	162	21	bert	bert	PROPN
futech-324	162	22	has	have	VERB
futech-324	162	23	good	good	ADJ
futech-324	162	24	language	language	NOUN
futech-324	162	25	understanding	understanding	NOUN
futech-324	162	26	ability	ability	NOUN
futech-324	162	27	,	,	PUNCT
futech-324	162	28	and	and	CCONJ
futech-324	162	29	using	use	VERB
futech-324	162	30	cnn	cnn	PROPN
futech-324	162	31	for	for	ADP
futech-324	162	32	feature	feature	NOUN
futech-324	162	33	extraction	extraction	NOUN
futech-324	162	34	work	work	NOUN
futech-324	162	35	.	.	PUNCT
futech-324	163	1	3	3	X
futech-324	163	2	.	.	X
futech-324	163	3	methodology	methodology	NOUN
futech-324	163	4	the	the	DET
futech-324	163	5	proposed	propose	VERB
futech-324	163	6	method	method	NOUN
futech-324	163	7	for	for	ADP
futech-324	163	8	toxic	toxic	ADJ
futech-324	163	9	comment	comment	NOUN
futech-324	163	10	classification	classification	NOUN
futech-324	163	11	is	be	AUX
futech-324	163	12	shown	show	VERB
futech-324	163	13	in	in	ADP
futech-324	163	14	figure	figure	NOUN
futech-324	163	15	1	1	NUM
futech-324	163	16	.	.	PUNCT
futech-324	164	1	the	the	DET
futech-324	164	2	proposed	propose	VERB
futech-324	164	3	method	method	NOUN
futech-324	164	4	for	for	ADP
futech-324	164	5	toxic	toxic	ADJ
futech-324	164	6	comment	comment	NOUN
futech-324	164	7	detection	detection	NOUN
futech-324	164	8	is	be	AUX
futech-324	164	9	designed	design	VERB
futech-324	164	10	to	to	PART
futech-324	164	11	evaluate	evaluate	VERB
futech-324	164	12	various	various	ADJ
futech-324	164	13	dl	dl	PROPN
futech-324	164	14	models	model	NOUN
futech-324	164	15	in	in	ADP
futech-324	164	16	combination	combination	NOUN
futech-324	164	17	with	with	ADP
futech-324	164	18	word	word	NOUN
futech-324	164	19	embedding	embed	VERB
futech-324	164	20	techniques	technique	NOUN
futech-324	164	21	such	such	ADJ
futech-324	164	22	as	as	ADP
futech-324	164	23	word2vec	word2vec	X
futech-324	164	24	,	,	PUNCT
futech-324	164	25	tf	tf	PROPN
futech-324	164	26	-	-	PUNCT
futech-324	164	27	idf	idf	PROPN
futech-324	164	28	,	,	PUNCT
futech-324	164	29	and	and	CCONJ
futech-324	164	30	bert	bert	PROPN
futech-324	164	31	.	.	PUNCT
futech-324	165	1	the	the	DET
futech-324	165	2	method	method	NOUN
futech-324	165	3	is	be	AUX
futech-324	165	4	applied	apply	VERB
futech-324	165	5	to	to	ADP
futech-324	165	6	the	the	DET
futech-324	165	7	jigsaw	jigsaw	NOUN
futech-324	165	8	toxic	toxic	ADJ
futech-324	165	9	comment	comment	NOUN
futech-324	165	10	dataset	dataset	NOUN
futech-324	165	11	,	,	PUNCT
futech-324	165	12	which	which	PRON
futech-324	165	13	contains	contain	VERB
futech-324	165	14	labeled	label	VERB
futech-324	165	15	english	english	ADJ
futech-324	165	16	comments	comment	NOUN
futech-324	165	17	categorized	categorize	VERB
futech-324	165	18	as	as	ADP
futech-324	165	19	toxic	toxic	ADJ
futech-324	165	20	or	or	CCONJ
futech-324	165	21	non	non	ADJ
futech-324	165	22	-	-	ADJ
futech-324	165	23	toxic	toxic	ADJ
futech-324	165	24	.	.	PUNCT
futech-324	166	1	the	the	DET
futech-324	166	2	dataset	dataset	NOUN
futech-324	166	3	undergoes	undergoe	NOUN
futech-324	166	4	preprocessing	preprocesse	VERB
futech-324	166	5	to	to	PART
futech-324	166	6	clean	clean	VERB
futech-324	166	7	the	the	DET
futech-324	166	8	text	text	NOUN
futech-324	166	9	by	by	ADP
futech-324	166	10	handling	handle	VERB
futech-324	166	11	inconsistencies	inconsistency	NOUN
futech-324	166	12	such	such	ADJ
futech-324	166	13	as	as	ADP
futech-324	166	14	missing	miss	VERB
futech-324	166	15	values	value	NOUN
futech-324	166	16	and	and	CCONJ
futech-324	166	17	ensuring	ensure	VERB
futech-324	166	18	that	that	SCONJ
futech-324	166	19	the	the	DET
futech-324	166	20	comment	comment	NOUN
futech-324	166	21	texts	text	NOUN
futech-324	166	22	are	be	AUX
futech-324	166	23	appropriately	appropriately	ADV
futech-324	166	24	formatted	format	VERB
futech-324	166	25	for	for	ADP
futech-324	166	26	further	further	ADJ
futech-324	166	27	analysis	analysis	NOUN
futech-324	166	28	.	.	PUNCT
futech-324	167	1	figure	figure	NOUN
futech-324	167	2	1	1	NUM
futech-324	167	3	.	.	PUNCT
futech-324	167	4	proposed	propose	VERB
futech-324	167	5	methodology	methodology	NOUN
futech-324	167	6	for	for	ADP
futech-324	167	7	toxic	toxic	ADJ
futech-324	167	8	comment	comment	NOUN
futech-324	167	9	classification	classification	NOUN
futech-324	167	10	tf	tf	PROPN
futech-324	167	11	-	-	PUNCT
futech-324	167	12	idf	idf	PROPN
futech-324	167	13	and	and	CCONJ
futech-324	167	14	word2vec	word2vec	PRON
futech-324	167	15	are	be	AUX
futech-324	167	16	used	use	VERB
futech-324	167	17	to	to	PART
futech-324	167	18	extract	extract	VERB
futech-324	167	19	text	text	NOUN
futech-324	167	20	features	feature	NOUN
futech-324	167	21	in	in	ADP
futech-324	167	22	two	two	NUM
futech-324	167	23	distinct	distinct	ADJ
futech-324	167	24	ways	way	NOUN
futech-324	167	25	.	.	PUNCT
futech-324	168	1	tf	tf	X
futech-324	168	2	-	-	PUNCT
futech-324	168	3	idf	idf	PROPN
futech-324	168	4	computes	compute	VERB
futech-324	168	5	the	the	DET
futech-324	168	6	relative	relative	ADJ
futech-324	168	7	importance	importance	NOUN
futech-324	168	8	of	of	ADP
futech-324	168	9	a	a	DET
futech-324	168	10	word	word	NOUN
futech-324	168	11	within	within	ADP
futech-324	168	12	a	a	DET
futech-324	168	13	document	document	NOUN
futech-324	168	14	by	by	ADP
futech-324	168	15	analyzing	analyze	VERB
futech-324	168	16	its	its	PRON
futech-324	168	17	frequency	frequency	NOUN
futech-324	168	18	in	in	ADP
futech-324	168	19	relation	relation	NOUN
futech-324	168	20	to	to	ADP
futech-324	168	21	the	the	DET
futech-324	168	22	entire	entire	ADJ
futech-324	168	23	corpus	corpus	NOUN
futech-324	168	24	,	,	PUNCT
futech-324	168	25	generating	generate	VERB
futech-324	168	26	sparse	sparse	ADJ
futech-324	168	27	numerical	numerical	ADJ
futech-324	168	28	vectors	vector	NOUN
futech-324	168	29	.	.	PUNCT
futech-324	169	1	in	in	ADP
futech-324	169	2	contrast	contrast	NOUN
futech-324	169	3	,	,	PUNCT
futech-324	169	4	word2vec	word2vec	PRON
futech-324	169	5	learns	learn	VERB
futech-324	169	6	distributed	distribute	VERB
futech-324	169	7	word	word	NOUN
futech-324	169	8	representations	representation	NOUN
futech-324	169	9	by	by	ADP
futech-324	169	10	training	training	NOUN
futech-324	169	11	on	on	ADP
futech-324	169	12	the	the	DET
futech-324	169	13	tokenized	tokenized	ADJ
futech-324	169	14	text	text	NOUN
futech-324	169	15	data	datum	NOUN
futech-324	169	16	,	,	PUNCT
futech-324	169	17	capturing	capture	VERB
futech-324	169	18	semantic	semantic	ADJ
futech-324	169	19	relationships	relationship	NOUN
futech-324	169	20	between	between	ADP
futech-324	169	21	words	word	NOUN
futech-324	169	22	through	through	ADP
futech-324	169	23	cooccurrence	cooccurrence	NOUN
futech-324	169	24	patterns	pattern	NOUN
futech-324	169	25	.	.	PUNCT
futech-324	170	1	each	each	DET
futech-324	170	2	comment	comment	NOUN
futech-324	170	3	is	be	AUX
futech-324	170	4	then	then	ADV
futech-324	170	5	represented	represent	VERB
futech-324	170	6	as	as	ADP
futech-324	170	7	a	a	DET
futech-324	170	8	fixed	fix	VERB
futech-324	170	9	-	-	PUNCT
futech-324	170	10	length	length	NOUN
futech-324	170	11	vector	vector	NOUN
futech-324	170	12	by	by	ADP
futech-324	170	13	averaging	average	VERB
futech-324	170	14	the	the	DET
futech-324	170	15	embeddings	embedding	NOUN
futech-324	170	16	of	of	ADP
futech-324	170	17	its	its	PRON
futech-324	170	18	words	word	NOUN
futech-324	170	19	.	.	PUNCT
futech-324	171	1	these	these	DET
futech-324	171	2	two	two	NUM
futech-324	171	3	methods	method	NOUN
futech-324	171	4	provide	provide	VERB
futech-324	171	5	complementary	complementary	ADJ
futech-324	171	6	insights	insight	NOUN
futech-324	171	7	—	—	PUNCT
futech-324	171	8	tfidf	tfidf	NOUN
futech-324	171	9	emphasizes	emphasize	VERB
futech-324	171	10	statistical	statistical	ADJ
futech-324	171	11	significance	significance	NOUN
futech-324	171	12	,	,	PUNCT
futech-324	171	13	while	while	SCONJ
futech-324	171	14	word2vec	word2vec	X
futech-324	171	15	captures	capture	VERB
futech-324	171	16	contextual	contextual	ADJ
futech-324	171	17	meaning	meaning	NOUN
futech-324	171	18	.	.	PUNCT
futech-324	172	1	the	the	DET
futech-324	172	2	next	next	ADJ
futech-324	172	3	step	step	NOUN
futech-324	172	4	involves	involve	VERB
futech-324	172	5	applying	apply	VERB
futech-324	172	6	different	different	ADJ
futech-324	172	7	dl	dl	PROPN
futech-324	172	8	models	model	NOUN
futech-324	172	9	,	,	PUNCT
futech-324	172	10	namely	namely	ADV
futech-324	172	11	rnn	rnn	PROPN
futech-324	172	12	,	,	PUNCT
futech-324	172	13	lstm	lstm	NOUN
futech-324	172	14	,	,	PUNCT
futech-324	172	15	and	and	CCONJ
futech-324	172	16	gru	gru	NOUN
futech-324	172	17	,	,	PUNCT
futech-324	172	18	to	to	PART
futech-324	172	19	classify	classify	VERB
futech-324	172	20	the	the	DET
futech-324	172	21	comments	comment	NOUN
futech-324	172	22	as	as	ADP
futech-324	172	23	either	either	CCONJ
futech-324	172	24	toxic	toxic	ADJ
futech-324	172	25	or	or	CCONJ
futech-324	172	26	non	non	ADJ
futech-324	172	27	-	-	ADJ
futech-324	172	28	toxic	toxic	ADJ
futech-324	172	29	.	.	PUNCT
futech-324	173	1	each	each	DET
futech-324	173	2	model	model	NOUN
futech-324	173	3	architecture	architecture	NOUN
futech-324	173	4	consists	consist	VERB
futech-324	173	5	of	of	ADP
futech-324	173	6	an	an	DET
futech-324	173	7	embedding	embed	VERB
futech-324	173	8	or	or	CCONJ
futech-324	173	9	input	input	NOUN
futech-324	173	10	layer	layer	NOUN
futech-324	173	11	,	,	PUNCT
futech-324	173	12	a	a	DET
futech-324	173	13	recurrent	recurrent	ADJ
futech-324	173	14	layer	layer	NOUN
futech-324	173	15	(	(	PUNCT
futech-324	173	16	rnn	rnn	PROPN
futech-324	173	17	,	,	PUNCT
futech-324	173	18	lstm	lstm	NOUN
futech-324	173	19	,	,	PUNCT
futech-324	173	20	or	or	CCONJ
futech-324	173	21	gru	gru	NOUN
futech-324	173	22	)	)	PUNCT
futech-324	173	23	,	,	PUNCT
futech-324	173	24	and	and	CCONJ
futech-324	173	25	a	a	DET
futech-324	173	26	dense	dense	ADJ
futech-324	173	27	output	output	NOUN
futech-324	173	28	layer	layer	NOUN
futech-324	173	29	for	for	ADP
futech-324	173	30	binary	binary	ADJ
futech-324	173	31	classification	classification	NOUN
futech-324	173	32	.	.	PUNCT
futech-324	174	1	the	the	DET
futech-324	174	2	structure	structure	NOUN
futech-324	174	3	remains	remain	VERB
futech-324	174	4	consistent	consistent	ADJ
futech-324	174	5	across	across	ADP
futech-324	174	6	experiments	experiment	NOUN
futech-324	174	7	,	,	PUNCT
futech-324	174	8	with	with	ADP
futech-324	174	9	only	only	ADV
futech-324	174	10	the	the	DET
futech-324	174	11	embedding	embed	VERB
futech-324	174	12	source	source	NOUN
futech-324	174	13	(	(	PUNCT
futech-324	174	14	tf	tf	PROPN
futech-324	174	15	-	-	PUNCT
futech-324	174	16	idf	idf	PROPN
futech-324	174	17	,	,	PUNCT
futech-324	174	18	word2vec	word2vec	X
futech-324	174	19	,	,	PUNCT
futech-324	174	20	or	or	CCONJ
futech-324	174	21	bert	bert	PROPN
futech-324	174	22	)	)	PUNCT
futech-324	174	23	varying	varying	NOUN
futech-324	174	24	.	.	PUNCT
futech-324	175	1	bert	bert	PROPN
futech-324	175	2	,	,	PUNCT
futech-324	175	3	a	a	DET
futech-324	175	4	transformer	transformer	NOUN
futech-324	175	5	-	-	PUNCT
futech-324	175	6	based	base	VERB
futech-324	175	7	model	model	NOUN
futech-324	175	8	,	,	PUNCT
futech-324	175	9	is	be	AUX
futech-324	175	10	also	also	ADV
futech-324	175	11	employed	employ	VERB
futech-324	175	12	for	for	ADP
futech-324	175	13	richer	rich	ADJ
futech-324	175	14	feature	feature	NOUN
futech-324	175	15	extraction	extraction	NOUN
futech-324	175	16	.	.	PUNCT
futech-324	176	1	this	this	DET
futech-324	176	2	pre	pre	ADJ
futech-324	176	3	-	-	ADJ
futech-324	176	4	trained	train	VERB
futech-324	176	5	bert	bert	PROPN
futech-324	176	6	model	model	NOUN
futech-324	176	7	uses	use	VERB
futech-324	176	8	a	a	DET
futech-324	176	9	large	large	ADJ
futech-324	176	10	corpus	corpus	NOUN
futech-324	176	11	to	to	PART
futech-324	176	12	train	train	VERB
futech-324	176	13	and	and	CCONJ
futech-324	176	14	fine	fine	ADJ
futech-324	176	15	-	-	PUNCT
futech-324	176	16	tunes	tune	NOUN
futech-324	176	17	itself	itself	PRON
futech-324	176	18	on	on	ADP
futech-324	176	19	the	the	DET
futech-324	176	20	toxic	toxic	ADJ
futech-324	176	21	comment	comment	NOUN
futech-324	176	22	dataset	dataset	NOUN
futech-324	176	23	.	.	PUNCT
futech-324	177	1	in	in	ADP
futech-324	177	2	this	this	DET
futech-324	177	3	setup	setup	NOUN
futech-324	177	4	,	,	PUNCT
futech-324	177	5	token	token	ADJ
futech-324	177	6	embeddings	embedding	NOUN
futech-324	177	7	are	be	AUX
futech-324	177	8	extracted	extract	VERB
futech-324	177	9	using	use	VERB
futech-324	177	10	the	the	DET
futech-324	177	11	distilbert	distilbert	ADJ
futech-324	177	12	tokenizer	tokenizer	NOUN
futech-324	177	13	,	,	PUNCT
futech-324	177	14	particularly	particularly	ADV
futech-324	177	15	from	from	ADP
futech-324	177	16	the	the	DET
futech-324	177	17	[	[	X
futech-324	177	18	cls	cls	X
futech-324	177	19	]	]	PUNCT
futech-324	177	20	token	token	VERB
futech-324	177	21	.	.	PUNCT
futech-324	178	1	these	these	DET
futech-324	178	2	embeddings	embedding	NOUN
futech-324	178	3	are	be	AUX
futech-324	178	4	then	then	ADV
futech-324	178	5	passed	pass	VERB
futech-324	178	6	through	through	ADP
futech-324	178	7	the	the	DET
futech-324	178	8	recurrent	recurrent	ADJ
futech-324	178	9	models	model	NOUN
futech-324	178	10	to	to	PART
futech-324	178	11	improve	improve	VERB
futech-324	178	12	contextual	contextual	ADJ
futech-324	178	13	interpretation	interpretation	NOUN
futech-324	178	14	[	[	X
futech-324	178	15	28	28	NUM
futech-324	178	16	]	]	PUNCT
futech-324	178	17	.	.	PUNCT
futech-324	179	1	this	this	DET
futech-324	179	2	architecture	architecture	NOUN
futech-324	179	3	produces	produce	VERB
futech-324	179	4	hybrid	hybrid	ADJ
futech-324	179	5	models	model	NOUN
futech-324	179	6	such	such	ADJ
futech-324	179	7	as	as	ADP
futech-324	179	8	rnn+bert	rnn+bert	NOUN
futech-324	179	9	,	,	PUNCT
futech-324	179	10	lstm+bert	lstm+bert	PROPN
futech-324	179	11	,	,	PUNCT
futech-324	179	12	and	and	CCONJ
futech-324	179	13	gru+bert	gru+bert	NOUN
futech-324	179	14	,	,	PUNCT
futech-324	179	15	where	where	SCONJ
futech-324	179	16	bert	bert	PROPN
futech-324	179	17	serves	serve	VERB
futech-324	179	18	solely	solely	ADV
futech-324	179	19	as	as	ADP
futech-324	179	20	a	a	DET
futech-324	179	21	feature	feature	NOUN
futech-324	179	22	extractor	extractor	NOUN
futech-324	179	23	,	,	PUNCT
futech-324	179	24	providing	provide	VERB
futech-324	179	25	contextual	contextual	ADJ
futech-324	179	26	embeddings	embedding	NOUN
futech-324	179	27	.	.	PUNCT
futech-324	180	1	these	these	DET
futech-324	180	2	embeddings	embedding	NOUN
futech-324	180	3	are	be	AUX
futech-324	180	4	fed	feed	VERB
futech-324	180	5	as	as	ADP
futech-324	180	6	input	input	NOUN
futech-324	180	7	to	to	ADP
futech-324	180	8	the	the	DET
futech-324	180	9	corresponding	corresponding	ADJ
futech-324	180	10	recurrent	recurrent	ADJ
futech-324	180	11	layers	layer	NOUN
futech-324	180	12	for	for	ADP
futech-324	180	13	learning	learn	VERB
futech-324	180	14	temporal	temporal	ADJ
futech-324	180	15	dependencies	dependency	NOUN
futech-324	180	16	prior	prior	ADV
futech-324	180	17	to	to	ADP
futech-324	180	18	classification	classification	NOUN
futech-324	180	19	.	.	PUNCT
futech-324	181	1	bert	bert	PROPN
futech-324	181	2	is	be	AUX
futech-324	181	3	not	not	PART
futech-324	181	4	used	use	VERB
futech-324	181	5	as	as	ADP
futech-324	181	6	a	a	DET
futech-324	181	7	classifier	classifier	NOUN
futech-324	181	8	in	in	ADP
futech-324	181	9	this	this	DET
futech-324	181	10	work	work	NOUN
futech-324	181	11	,	,	PUNCT
futech-324	181	12	but	but	CCONJ
futech-324	181	13	it	it	PRON
futech-324	181	14	significantly	significantly	ADV
futech-324	181	15	enhances	enhance	VERB
futech-324	181	16	the	the	DET
futech-324	181	17	input	input	NOUN
futech-324	181	18	representation	representation	NOUN
futech-324	181	19	for	for	ADP
futech-324	181	20	downstream	downstream	ADJ
futech-324	181	21	learning	learning	NOUN
futech-324	181	22	.	.	PUNCT
futech-324	182	1	this	this	DET
futech-324	182	2	approach	approach	NOUN
futech-324	182	3	enables	enable	VERB
futech-324	182	4	the	the	DET
futech-324	182	5	models	model	NOUN
futech-324	182	6	to	to	PART
futech-324	182	7	recognize	recognize	VERB
futech-324	182	8	complex	complex	ADJ
futech-324	182	9	and	and	CCONJ
futech-324	182	10	contextual	contextual	ADJ
futech-324	182	11	relationships	relationship	NOUN
futech-324	182	12	between	between	ADP
futech-324	182	13	words	word	NOUN
futech-324	182	14	in	in	ADP
futech-324	182	15	comments	comment	NOUN
futech-324	182	16	.	.	PUNCT
futech-324	183	1	final	final	ADJ
futech-324	183	2	classification	classification	NOUN
futech-324	183	3	is	be	AUX
futech-324	183	4	performed	perform	VERB
futech-324	183	5	by	by	ADP
futech-324	183	6	the	the	DET
futech-324	183	7	dense	dense	ADJ
futech-324	183	8	output	output	NOUN
futech-324	183	9	layer	layer	NOUN
futech-324	183	10	.	.	PUNCT
futech-324	184	1	the	the	DET
futech-324	184	2	entire	entire	ADJ
futech-324	184	3	framework	framework	NOUN
futech-324	184	4	is	be	AUX
futech-324	184	5	designed	design	VERB
futech-324	184	6	to	to	PART
futech-324	184	7	output	output	VERB
futech-324	184	8	the	the	DET
futech-324	184	9	probability	probability	NOUN
futech-324	184	10	of	of	ADP
futech-324	184	11	a	a	DET
futech-324	184	12	comment	comment	NOUN
futech-324	184	13	being	be	AUX
futech-324	184	14	toxic	toxic	ADJ
futech-324	184	15	based	base	VERB
futech-324	184	16	on	on	ADP
futech-324	184	17	the	the	DET
futech-324	184	18	learned	learn	VERB
futech-324	184	19	features	feature	NOUN
futech-324	184	20	.	.	PUNCT
futech-324	185	1	all	all	DET
futech-324	185	2	models	model	NOUN
futech-324	185	3	are	be	AUX
futech-324	185	4	evaluated	evaluate	VERB
futech-324	185	5	on	on	ADP
futech-324	185	6	a	a	DET
futech-324	185	7	range	range	NOUN
futech-324	185	8	of	of	ADP
futech-324	185	9	performance	performance	NOUN
futech-324	185	10	metrics	metric	NOUN
futech-324	185	11	,	,	PUNCT
futech-324	185	12	including	include	VERB
futech-324	185	13	accuracy	accuracy	NOUN
futech-324	185	14	,	,	PUNCT
futech-324	185	15	precision	precision	NOUN
futech-324	185	16	,	,	PUNCT
futech-324	185	17	recall	recall	NOUN
futech-324	185	18	,	,	PUNCT
futech-324	185	19	and	and	CCONJ
futech-324	185	20	f1	f1	PROPN
futech-324	185	21	score	score	NOUN
futech-324	185	22	.	.	PUNCT
futech-324	186	1	the	the	DET
futech-324	186	2	outcome	outcome	NOUN
futech-324	186	3	of	of	ADP
futech-324	186	4	each	each	DET
futech-324	186	5	model	model	NOUN
futech-324	186	6	is	be	AUX
futech-324	186	7	compared	compare	VERB
futech-324	186	8	to	to	PART
futech-324	186	9	decide	decide	VERB
futech-324	186	10	which	which	PRON
futech-324	186	11	is	be	AUX
futech-324	186	12	the	the	DET
futech-324	186	13	best	good	ADJ
futech-324	186	14	way	way	NOUN
futech-324	186	15	to	to	PART
futech-324	186	16	detect	detect	VERB
futech-324	186	17	toxicity	toxicity	NOUN
futech-324	186	18	in	in	ADP
futech-324	186	19	online	online	ADJ
futech-324	186	20	commentary	commentary	NOUN
futech-324	186	21	.	.	PUNCT
futech-324	187	1	this	this	DET
futech-324	187	2	experimental	experimental	ADJ
futech-324	187	3	design	design	NOUN
futech-324	187	4	enables	enable	VERB
futech-324	187	5	a	a	DET
futech-324	187	6	comparative	comparative	ADJ
futech-324	187	7	evaluation	evaluation	NOUN
futech-324	187	8	of	of	ADP
futech-324	187	9	different	different	ADJ
futech-324	187	10	embedding	embed	VERB
futech-324	187	11	-	-	PUNCT
futech-324	187	12	model	model	NOUN
futech-324	187	13	combinations	combination	NOUN
futech-324	187	14	.	.	PUNCT
futech-324	188	1	using	use	VERB
futech-324	188	2	traditional	traditional	ADJ
futech-324	188	3	feature	feature	NOUN
futech-324	188	4	extraction	extraction	NOUN
futech-324	188	5	techniques	technique	NOUN
futech-324	188	6	such	such	ADJ
futech-324	188	7	as	as	ADP
futech-324	188	8	word2vec	word2vec	X
futech-324	188	9	and	and	CCONJ
futech-324	188	10	tf	tf	PROPN
futech-324	188	11	-	-	PUNCT
futech-324	188	12	idf	idf	PROPN
futech-324	188	13	together	together	ADV
futech-324	188	14	with	with	ADP
futech-324	188	15	state	state	NOUN
futech-324	188	16	-	-	PUNCT
futech-324	188	17	of	of	ADP
futech-324	188	18	-	-	PUNCT
futech-324	188	19	the	the	DET
futech-324	188	20	-	-	PUNCT
futech-324	188	21	art	art	NOUN
futech-324	188	22	transformer	transformer	NOUN
futech-324	188	23	-	-	PUNCT
futech-324	188	24	based	base	VERB
futech-324	188	25	models	model	NOUN
futech-324	188	26	like	like	ADP
futech-324	188	27	bert	bert	PROPN
futech-324	188	28	,	,	PUNCT
futech-324	188	29	the	the	DET
futech-324	188	30	proposed	propose	VERB
futech-324	188	31	algorithm	algorithm	NOUN
futech-324	188	32	has	have	VERB
futech-324	188	33	high	high	ADJ
futech-324	188	34	classification	classification	NOUN
futech-324	188	35	accuracy	accuracy	NOUN
futech-324	188	36	.	.	PUNCT
futech-324	189	1	it	it	PRON
futech-324	189	2	also	also	ADV
futech-324	189	3	verifies	verify	VERB
futech-324	189	4	the	the	DET
futech-324	189	5	potential	potential	NOUN
futech-324	189	6	for	for	ADP
futech-324	189	7	hybrid	hybrid	ADJ
futech-324	189	8	dl	dl	PROPN
futech-324	189	9	models	model	NOUN
futech-324	189	10	as	as	ADP
futech-324	189	11	a	a	DET
futech-324	189	12	means	means	NOUN
futech-324	189	13	of	of	ADP
futech-324	189	14	fulfilling	fulfil	VERB
futech-324	189	15	text	text	NOUN
futech-324	189	16	categorisation	categorisation	NOUN
futech-324	189	17	work	work	NOUN
futech-324	189	18	.	.	PUNCT
futech-324	190	1	3.1	3.1	NUM
futech-324	190	2	data	datum	NOUN
futech-324	190	3	collection	collection	NOUN
futech-324	190	4	this	this	DET
futech-324	190	5	paper	paper	NOUN
futech-324	190	6	used	use	VERB
futech-324	190	7	the	the	DET
futech-324	190	8	jigsaw	jigsaw	NOUN
futech-324	190	9	toxic	toxic	ADJ
futech-324	190	10	comment	comment	NOUN
futech-324	190	11	dataset	dataset	NOUN
futech-324	190	12	,	,	PUNCT
futech-324	190	13	collected	collect	VERB
futech-324	190	14	from	from	ADP
futech-324	190	15	kaggle	kaggle	PROPN
futech-324	190	16	[	[	X
futech-324	190	17	29	29	NUM
futech-324	190	18	]	]	PUNCT
futech-324	190	19	.	.	PUNCT
futech-324	191	1	the	the	DET
futech-324	191	2	dataset	dataset	NOUN
futech-324	191	3	contains	contain	VERB
futech-324	191	4	a	a	DET
futech-324	191	5	collection	collection	NOUN
futech-324	191	6	of	of	ADP
futech-324	191	7	english	english	ADJ
futech-324	191	8	language	language	NOUN
futech-324	191	9	comments	comment	NOUN
futech-324	191	10	,	,	PUNCT
futech-324	191	11	each	each	PRON
futech-324	191	12	labeled	label	VERB
futech-324	191	13	as	as	ADP
futech-324	191	14	either	either	CCONJ
futech-324	191	15	toxic	toxic	ADJ
futech-324	191	16	or	or	CCONJ
futech-324	191	17	non	non	ADJ
futech-324	191	18	-	-	ADJ
futech-324	191	19	toxic	toxic	ADJ
futech-324	191	20	.	.	PUNCT
futech-324	192	1	the	the	DET
futech-324	192	2	dataset	dataset	NOUN
futech-324	192	3	consists	consist	VERB
futech-324	192	4	of	of	ADP
futech-324	192	5	two	two	NUM
futech-324	192	6	important	important	ADJ
futech-324	192	7	columns	column	NOUN
futech-324	192	8	:	:	PUNCT
futech-324	192	9	comment_text	comment_text	NOUN
futech-324	192	10	,	,	PUNCT
futech-324	192	11	which	which	PRON
futech-324	192	12	contains	contain	VERB
futech-324	192	13	the	the	DET
futech-324	192	14	actual	actual	ADJ
futech-324	192	15	comment	comment	NOUN
futech-324	192	16	text	text	NOUN
futech-324	192	17	,	,	PUNCT
futech-324	192	18	and	and	CCONJ
futech-324	192	19	toxic	toxic	ADJ
futech-324	192	20	,	,	PUNCT
futech-324	192	21	a	a	DET
futech-324	192	22	binary	binary	ADJ
futech-324	192	23	target	target	NOUN
futech-324	192	24	variable	variable	NOUN
futech-324	192	25	(	(	PUNCT
futech-324	192	26	1	1	NUM
futech-324	192	27	refers	refer	VERB
futech-324	192	28	to	to	ADP
futech-324	192	29	toxic	toxic	ADJ
futech-324	192	30	comments	comment	NOUN
futech-324	192	31	,	,	PUNCT
futech-324	192	32	while	while	SCONJ
futech-324	192	33	0	0	NUM
futech-324	192	34	refers	refer	VERB
futech-324	192	35	to	to	ADP
futech-324	192	36	all	all	DET
futech-324	192	37	other	other	ADJ
futech-324	192	38	types	type	NOUN
futech-324	192	39	of	of	ADP
futech-324	192	40	comments	comment	NOUN
futech-324	192	41	)	)	PUNCT
futech-324	192	42	.	.	PUNCT
futech-324	193	1	this	this	DET
futech-324	193	2	data	data	NOUN
futech-324	193	3	comes	come	VERB
futech-324	193	4	from	from	ADP
futech-324	193	5	a	a	DET
futech-324	193	6	variety	variety	NOUN
futech-324	193	7	of	of	ADP
futech-324	193	8	online	online	ADJ
futech-324	193	9	sources	source	NOUN
futech-324	193	10	,	,	PUNCT
futech-324	193	11	including	include	VERB
futech-324	193	12	civil	civil	ADJ
futech-324	193	13	comments	comment	NOUN
futech-324	193	14	and	and	CCONJ
futech-324	193	15	wikipedia	wikipedia	PROPN
futech-324	193	16	talk	talk	NOUN
futech-324	193	17	page	page	NOUN
futech-324	193	18	edits	edit	VERB
futech-324	193	19	.	.	PUNCT
futech-324	194	1	text	text	NOUN
futech-324	194	2	-	-	PUNCT
futech-324	194	3	based	base	VERB
futech-324	194	4	prediction	prediction	NOUN
futech-324	194	5	of	of	ADP
futech-324	194	6	whether	whether	SCONJ
futech-324	194	7	a	a	DET
futech-324	194	8	comment	comment	NOUN
futech-324	194	9	is	be	AUX
futech-324	194	10	toxic	toxic	ADJ
futech-324	194	11	or	or	CCONJ
futech-324	194	12	not	not	PART
futech-324	194	13	.	.	PUNCT
futech-324	195	1	s.	s.	PROPN
futech-324	195	2	sushma	sushma	PROPN
futech-324	195	3	et	et	PROPN
futech-324	195	4	al	al	PROPN
futech-324	195	5	.	.	PROPN
futech-324	195	6	/future	/future	PUNCT
futech-324	195	7	technology	technology	PROPN
futech-324	195	8	august	august	PROPN
futech-324	195	9	2025|	2025|	NUM
futech-324	195	10	volume	volume	NOUN
futech-324	195	11	04	04	NUM
futech-324	196	1	|	|	NOUN
futech-324	196	2	issue	issue	NOUN
futech-324	196	3	03	03	NUM
futech-324	197	1	|	|	ADV
futech-324	197	2	pages	page	VERB
futech-324	197	3	76	76	NUM
futech-324	197	4	-	-	SYM
futech-324	197	5	84	84	NUM
futech-324	197	6	80	80	NUM
futech-324	197	7	3.2	3.2	NUM
futech-324	197	8	preprocessing	preprocessing	NOUN
futech-324	197	9	in	in	ADP
futech-324	197	10	the	the	DET
futech-324	197	11	preprocessing	preprocessing	NOUN
futech-324	197	12	phase	phase	NOUN
futech-324	197	13	,	,	PUNCT
futech-324	197	14	a	a	DET
futech-324	197	15	variety	variety	NOUN
futech-324	197	16	of	of	ADP
futech-324	197	17	procedures	procedure	NOUN
futech-324	197	18	were	be	AUX
futech-324	197	19	used	use	VERB
futech-324	197	20	to	to	PART
futech-324	197	21	clean	clean	VERB
futech-324	197	22	and	and	CCONJ
futech-324	197	23	prepare	prepare	VERB
futech-324	197	24	the	the	DET
futech-324	197	25	data	datum	NOUN
futech-324	197	26	for	for	ADP
futech-324	197	27	modeling	modeling	NOUN
futech-324	197	28	.	.	PUNCT
futech-324	198	1	this	this	PRON
futech-324	198	2	involved	involved	ADJ
futech-324	198	3	eliminating	eliminate	VERB
futech-324	198	4	all	all	DET
futech-324	198	5	kinds	kind	NOUN
futech-324	198	6	of	of	ADP
futech-324	198	7	inconsistencies	inconsistency	NOUN
futech-324	198	8	,	,	PUNCT
futech-324	198	9	such	such	ADJ
futech-324	198	10	as	as	ADP
futech-324	198	11	missing	miss	VERB
futech-324	198	12	values	value	NOUN
futech-324	198	13	in	in	ADP
futech-324	198	14	required	required	ADJ
futech-324	198	15	columns	column	NOUN
futech-324	198	16	(	(	PUNCT
futech-324	198	17	comment_text	comment_text	NOUN
futech-324	198	18	and	and	CCONJ
futech-324	198	19	toxic	toxic	ADJ
futech-324	198	20	)	)	PUNCT
futech-324	198	21	.	.	PUNCT
futech-324	199	1	missing	miss	VERB
futech-324	199	2	entries	entry	NOUN
futech-324	199	3	like	like	ADP
futech-324	199	4	these	these	PRON
futech-324	199	5	are	be	AUX
futech-324	199	6	corrected	correct	VERB
futech-324	199	7	to	to	PART
futech-324	199	8	preserve	preserve	VERB
futech-324	199	9	the	the	DET
futech-324	199	10	completeness	completeness	NOUN
futech-324	199	11	of	of	ADP
futech-324	199	12	the	the	DET
futech-324	199	13	dataset	dataset	NOUN
futech-324	199	14	and	and	CCONJ
futech-324	199	15	its	its	PRON
futech-324	199	16	suitability	suitability	NOUN
futech-324	199	17	for	for	ADP
futech-324	199	18	analysis	analysis	NOUN
futech-324	199	19	.	.	PUNCT
futech-324	200	1	later	later	ADV
futech-324	200	2	,	,	PUNCT
futech-324	200	3	the	the	DET
futech-324	200	4	data	datum	NOUN
futech-324	200	5	was	be	AUX
futech-324	200	6	tokenized	tokenize	VERB
futech-324	200	7	and	and	CCONJ
futech-324	200	8	transformed	transform	VERB
futech-324	200	9	into	into	ADP
futech-324	200	10	a	a	DET
futech-324	200	11	form	form	NOUN
futech-324	200	12	suitable	suitable	ADJ
futech-324	200	13	for	for	ADP
futech-324	200	14	input	input	NOUN
futech-324	200	15	into	into	ADP
futech-324	200	16	ml	ml	NOUN
futech-324	200	17	models	model	NOUN
futech-324	200	18	,	,	PUNCT
futech-324	200	19	optimizing	optimize	VERB
futech-324	200	20	both	both	CCONJ
futech-324	200	21	training	training	NOUN
futech-324	200	22	and	and	CCONJ
futech-324	200	23	validation	validation	NOUN
futech-324	200	24	datasets	dataset	NOUN
futech-324	200	25	so	so	SCONJ
futech-324	200	26	that	that	SCONJ
futech-324	200	27	they	they	PRON
futech-324	200	28	would	would	AUX
futech-324	200	29	work	work	VERB
futech-324	200	30	perfectly	perfectly	ADV
futech-324	200	31	with	with	ADP
futech-324	200	32	the	the	DET
futech-324	200	33	model	model	NOUN
futech-324	200	34	.	.	PUNCT
futech-324	201	1	4	4	X
futech-324	201	2	.	.	NOUN
futech-324	201	3	results	result	NOUN
futech-324	201	4	and	and	CCONJ
futech-324	201	5	discussion	discussion	NOUN
futech-324	201	6	4.1	4.1	NUM
futech-324	201	7	applying	apply	VERB
futech-324	201	8	rnn	rnn	NOUN
futech-324	201	9	with	with	ADP
futech-324	201	10	tf	tf	PROPN
futech-324	201	11	-	-	PUNCT
futech-324	201	12	idf	idf	PROPN
futech-324	201	13	in	in	ADP
futech-324	201	14	this	this	DET
futech-324	201	15	phase	phase	NOUN
futech-324	201	16	,	,	PUNCT
futech-324	201	17	the	the	DET
futech-324	201	18	tf	tf	PROPN
futech-324	201	19	-	-	PUNCT
futech-324	201	20	idf	idf	PROPN
futech-324	201	21	method	method	NOUN
futech-324	201	22	is	be	AUX
futech-324	201	23	applied	apply	VERB
futech-324	201	24	in	in	ADP
futech-324	201	25	combination	combination	NOUN
futech-324	201	26	with	with	ADP
futech-324	201	27	a	a	DET
futech-324	201	28	simplernn	simplernn	NOUN
futech-324	201	29	model	model	NOUN
futech-324	201	30	for	for	ADP
futech-324	201	31	toxic	toxic	ADJ
futech-324	201	32	comment	comment	NOUN
futech-324	201	33	classification	classification	NOUN
futech-324	201	34	.	.	PUNCT
futech-324	202	1	the	the	DET
futech-324	202	2	text	text	NOUN
futech-324	202	3	data	datum	NOUN
futech-324	202	4	is	be	AUX
futech-324	202	5	first	first	ADV
futech-324	202	6	vectorized	vectorize	VERB
futech-324	202	7	using	use	VERB
futech-324	202	8	the	the	DET
futech-324	202	9	tf	tf	PROPN
futech-324	202	10	-	-	PUNCT
futech-324	202	11	idf	idf	PROPN
futech-324	202	12	technique	technique	NOUN
futech-324	202	13	,	,	PUNCT
futech-324	202	14	which	which	PRON
futech-324	202	15	converts	convert	VERB
futech-324	202	16	the	the	DET
futech-324	202	17	text	text	NOUN
futech-324	202	18	into	into	ADP
futech-324	202	19	a	a	DET
futech-324	202	20	numerical	numerical	ADJ
futech-324	202	21	representation	representation	NOUN
futech-324	202	22	based	base	VERB
futech-324	202	23	on	on	ADP
futech-324	202	24	the	the	DET
futech-324	202	25	frequency	frequency	NOUN
futech-324	202	26	of	of	ADP
futech-324	202	27	terms	term	NOUN
futech-324	202	28	in	in	ADP
futech-324	202	29	the	the	DET
futech-324	202	30	dataset	dataset	NOUN
futech-324	202	31	.	.	PUNCT
futech-324	203	1	to	to	PART
futech-324	203	2	perform	perform	VERB
futech-324	203	3	dimensionality	dimensionality	NOUN
futech-324	203	4	reduction	reduction	NOUN
futech-324	203	5	,	,	PUNCT
futech-324	203	6	5,000	5,000	NUM
futech-324	203	7	input	input	NOUN
futech-324	203	8	features	feature	NOUN
futech-324	203	9	are	be	AUX
futech-324	203	10	extracted	extract	VERB
futech-324	203	11	from	from	ADP
futech-324	203	12	the	the	DET
futech-324	203	13	text	text	NOUN
futech-324	203	14	data	datum	NOUN
futech-324	203	15	so	so	SCONJ
futech-324	203	16	that	that	SCONJ
futech-324	203	17	the	the	DET
futech-324	203	18	model	model	NOUN
futech-324	203	19	can	can	AUX
futech-324	203	20	concentrate	concentrate	VERB
futech-324	203	21	on	on	ADP
futech-324	203	22	the	the	DET
futech-324	203	23	most	most	ADV
futech-324	203	24	pertinent	pertinent	ADJ
futech-324	203	25	terms	term	NOUN
futech-324	203	26	.	.	PUNCT
futech-324	204	1	we	we	PRON
futech-324	204	2	first	first	ADV
futech-324	204	3	split	split	VERB
futech-324	204	4	our	our	PRON
futech-324	204	5	data	datum	NOUN
futech-324	204	6	into	into	ADP
futech-324	204	7	training	training	NOUN
futech-324	204	8	and	and	CCONJ
futech-324	204	9	test	test	NOUN
futech-324	204	10	sets	set	NOUN
futech-324	204	11	as	as	SCONJ
futech-324	204	12	follows	follow	VERB
futech-324	204	13	:	:	PUNCT
futech-324	204	14	80	80	NUM
futech-324	204	15	%	%	NOUN
futech-324	204	16	for	for	ADP
futech-324	204	17	training	training	NOUN
futech-324	204	18	and	and	CCONJ
futech-324	204	19	20	20	NUM
futech-324	204	20	%	%	NOUN
futech-324	204	21	for	for	ADP
futech-324	204	22	testing	testing	NOUN
futech-324	204	23	.	.	PUNCT
futech-324	205	1	the	the	DET
futech-324	205	2	model	model	NOUN
futech-324	205	3	consists	consist	VERB
futech-324	205	4	of	of	ADP
futech-324	205	5	an	an	DET
futech-324	205	6	input	input	NOUN
futech-324	205	7	embedding	embed	VERB
futech-324	205	8	layer	layer	NOUN
futech-324	205	9	,	,	PUNCT
futech-324	205	10	an	an	DET
futech-324	205	11	rnn	rnn	NOUN
futech-324	205	12	layer	layer	NOUN
futech-324	205	13	with	with	ADP
futech-324	205	14	64	64	NUM
futech-324	205	15	units	unit	NOUN
futech-324	205	16	,	,	PUNCT
futech-324	205	17	followed	follow	VERB
futech-324	205	18	by	by	ADP
futech-324	205	19	an	an	DET
futech-324	205	20	output	output	NOUN
futech-324	205	21	dense	dense	ADJ
futech-324	205	22	layer	layer	NOUN
futech-324	205	23	with	with	ADP
futech-324	205	24	a	a	DET
futech-324	205	25	unit	unit	NOUN
futech-324	205	26	for	for	ADP
futech-324	205	27	binary	binary	ADJ
futech-324	205	28	classification	classification	NOUN
futech-324	205	29	.	.	PUNCT
futech-324	206	1	to	to	PART
futech-324	206	2	accelerate	accelerate	VERB
futech-324	206	3	the	the	DET
futech-324	206	4	training	training	NOUN
futech-324	206	5	process	process	NOUN
futech-324	206	6	,	,	PUNCT
futech-324	206	7	the	the	DET
futech-324	206	8	model	model	NOUN
futech-324	206	9	is	be	AUX
futech-324	206	10	trained	train	VERB
futech-324	206	11	for	for	ADP
futech-324	206	12	five	five	NUM
futech-324	206	13	epochs	epoch	NOUN
futech-324	206	14	with	with	ADP
futech-324	206	15	a	a	DET
futech-324	206	16	batch	batch	NOUN
futech-324	206	17	size	size	NOUN
futech-324	206	18	of	of	ADP
futech-324	206	19	256	256	NUM
futech-324	206	20	.	.	PUNCT
futech-324	207	1	the	the	DET
futech-324	207	2	final	final	ADJ
futech-324	207	3	evaluation	evaluation	NOUN
futech-324	207	4	on	on	ADP
futech-324	207	5	the	the	DET
futech-324	207	6	test	test	NOUN
futech-324	207	7	set	set	NOUN
futech-324	207	8	yields	yield	VERB
futech-324	207	9	a	a	DET
futech-324	207	10	validation	validation	NOUN
futech-324	207	11	accuracy	accuracy	NOUN
futech-324	207	12	of	of	ADP
futech-324	207	13	around	around	ADP
futech-324	207	14	89	89	NUM
futech-324	207	15	%	%	NOUN
futech-324	207	16	,	,	PUNCT
futech-324	207	17	indicating	indicate	VERB
futech-324	207	18	that	that	SCONJ
futech-324	207	19	the	the	DET
futech-324	207	20	model	model	NOUN
futech-324	207	21	successfully	successfully	ADV
futech-324	207	22	captures	capture	VERB
futech-324	207	23	the	the	DET
futech-324	207	24	sequential	sequential	ADJ
futech-324	207	25	nature	nature	NOUN
futech-324	207	26	of	of	ADP
futech-324	207	27	the	the	DET
futech-324	207	28	text	text	NOUN
futech-324	207	29	data	datum	NOUN
futech-324	207	30	and	and	CCONJ
futech-324	207	31	performs	perform	VERB
futech-324	207	32	adequately	adequately	ADV
futech-324	207	33	in	in	ADP
futech-324	207	34	classifying	classify	VERB
futech-324	207	35	toxic	toxic	ADJ
futech-324	207	36	comments	comment	NOUN
futech-324	207	37	.	.	PUNCT
futech-324	208	1	4.2	4.2	NUM
futech-324	208	2	applying	apply	VERB
futech-324	208	3	lstm	lstm	NOUN
futech-324	208	4	with	with	ADP
futech-324	208	5	tf	tf	PROPN
futech-324	208	6	-	-	PUNCT
futech-324	208	7	idf	idf	PROPN
futech-324	208	8	in	in	ADP
futech-324	208	9	this	this	DET
futech-324	208	10	step	step	NOUN
futech-324	208	11	,	,	PUNCT
futech-324	208	12	an	an	DET
futech-324	208	13	lstm	lstm	NOUN
futech-324	208	14	model	model	NOUN
futech-324	208	15	is	be	AUX
futech-324	208	16	utilized	utilize	VERB
futech-324	208	17	in	in	ADP
futech-324	208	18	conjunction	conjunction	NOUN
futech-324	208	19	with	with	ADP
futech-324	208	20	tf	tf	PROPN
futech-324	208	21	-	-	PUNCT
futech-324	208	22	idf	idf	NOUN
futech-324	208	23	for	for	ADP
futech-324	208	24	classifying	classify	VERB
futech-324	208	25	toxic	toxic	ADJ
futech-324	208	26	comments	comment	NOUN
futech-324	208	27	.	.	PUNCT
futech-324	209	1	the	the	DET
futech-324	209	2	text	text	NOUN
futech-324	209	3	data	datum	NOUN
futech-324	209	4	is	be	AUX
futech-324	209	5	vectorized	vectorize	VERB
futech-324	209	6	using	use	VERB
futech-324	209	7	the	the	DET
futech-324	209	8	tf	tf	PROPN
futech-324	209	9	-	-	PUNCT
futech-324	209	10	idf	idf	PROPN
futech-324	209	11	method	method	NOUN
futech-324	209	12	,	,	PUNCT
futech-324	209	13	which	which	PRON
futech-324	209	14	extracts	extract	VERB
futech-324	209	15	relevant	relevant	ADJ
futech-324	209	16	features	feature	NOUN
futech-324	209	17	based	base	VERB
futech-324	209	18	on	on	ADP
futech-324	209	19	the	the	DET
futech-324	209	20	term	term	NOUN
futech-324	209	21	frequency	frequency	NOUN
futech-324	209	22	-	-	PUNCT
futech-324	209	23	inverse	inverse	NOUN
futech-324	209	24	document	document	NOUN
futech-324	209	25	frequency	frequency	NOUN
futech-324	209	26	.	.	PUNCT
futech-324	210	1	the	the	DET
futech-324	210	2	neural	neural	ADJ
futech-324	210	3	network	network	NOUN
futech-324	210	4	is	be	AUX
futech-324	210	5	constructed	construct	VERB
futech-324	210	6	in	in	ADP
futech-324	210	7	the	the	DET
futech-324	210	8	form	form	NOUN
futech-324	210	9	of	of	ADP
futech-324	210	10	an	an	DET
futech-324	210	11	embedding	embed	VERB
futech-324	210	12	layer	layer	NOUN
futech-324	210	13	that	that	PRON
futech-324	210	14	takes	take	VERB
futech-324	210	15	the	the	DET
futech-324	210	16	input	input	NOUN
futech-324	210	17	dimensions	dimension	NOUN
futech-324	210	18	,	,	PUNCT
futech-324	210	19	and	and	CCONJ
futech-324	210	20	then	then	ADV
futech-324	210	21	an	an	DET
futech-324	210	22	lstm	lstm	ADJ
futech-324	210	23	layer	layer	NOUN
futech-324	210	24	,	,	PUNCT
futech-324	210	25	which	which	PRON
futech-324	210	26	has	have	VERB
futech-324	210	27	64	64	NUM
futech-324	210	28	units	unit	NOUN
futech-324	210	29	.	.	PUNCT
futech-324	211	1	we	we	PRON
futech-324	211	2	add	add	VERB
futech-324	211	3	a	a	DET
futech-324	211	4	dense	dense	ADJ
futech-324	211	5	output	output	NOUN
futech-324	211	6	layer	layer	NOUN
futech-324	211	7	for	for	ADP
futech-324	211	8	binary	binary	ADJ
futech-324	211	9	classification	classification	NOUN
futech-324	211	10	(	(	PUNCT
futech-324	211	11	toxic	toxic	ADJ
futech-324	211	12	vs	vs	ADP
futech-324	211	13	non	non	ADJ
futech-324	211	14	-	-	ADJ
futech-324	211	15	toxic	toxic	ADJ
futech-324	211	16	comments	comment	NOUN
futech-324	211	17	)	)	PUNCT
futech-324	211	18	.	.	PUNCT
futech-324	212	1	the	the	DET
futech-324	212	2	training	training	NOUN
futech-324	212	3	process	process	NOUN
futech-324	212	4	gradually	gradually	ADV
futech-324	212	5	improves	improve	VERB
futech-324	212	6	accuracy	accuracy	NOUN
futech-324	212	7	,	,	PUNCT
futech-324	212	8	with	with	ADP
futech-324	212	9	the	the	DET
futech-324	212	10	training	training	NOUN
futech-324	212	11	done	do	VERB
futech-324	212	12	over	over	ADP
futech-324	212	13	3	3	NUM
futech-324	212	14	epochs	epoch	NOUN
futech-324	212	15	using	use	VERB
futech-324	212	16	a	a	DET
futech-324	212	17	function	function	NOUN
futech-324	212	18	to	to	PART
futech-324	212	19	train	train	VERB
futech-324	212	20	on	on	ADP
futech-324	212	21	one	one	NUM
futech-324	212	22	set	set	NOUN
futech-324	212	23	of	of	ADP
futech-324	212	24	the	the	DET
futech-324	212	25	dataset	dataset	NOUN
futech-324	212	26	and	and	CCONJ
futech-324	212	27	the	the	DET
futech-324	212	28	given	give	VERB
futech-324	212	29	statements	statement	NOUN
futech-324	212	30	,	,	PUNCT
futech-324	212	31	with	with	ADP
futech-324	212	32	a	a	DET
futech-324	212	33	batch	batch	NOUN
futech-324	212	34	size	size	NOUN
futech-324	212	35	of	of	ADP
futech-324	212	36	16	16	NUM
futech-324	212	37	.	.	PUNCT
futech-324	213	1	the	the	DET
futech-324	213	2	trained	train	VERB
futech-324	213	3	model	model	NOUN
futech-324	213	4	is	be	AUX
futech-324	213	5	then	then	ADV
futech-324	213	6	evaluated	evaluate	VERB
futech-324	213	7	on	on	ADP
futech-324	213	8	the	the	DET
futech-324	213	9	test	test	NOUN
futech-324	213	10	set	set	NOUN
futech-324	213	11	,	,	PUNCT
futech-324	213	12	giving	give	VERB
futech-324	213	13	us	we	PRON
futech-324	213	14	a	a	DET
futech-324	213	15	validation	validation	NOUN
futech-324	213	16	accuracy	accuracy	NOUN
futech-324	213	17	of	of	ADP
futech-324	213	18	90.46	90.46	NUM
futech-324	213	19	%	%	NOUN
futech-324	213	20	,	,	PUNCT
futech-324	213	21	demonstrating	demonstrate	VERB
futech-324	213	22	the	the	DET
futech-324	213	23	model	model	NOUN
futech-324	213	24	’s	’s	PART
futech-324	213	25	ability	ability	NOUN
futech-324	213	26	to	to	PART
futech-324	213	27	capture	capture	VERB
futech-324	213	28	longterm	longterm	ADJ
futech-324	213	29	dependencies	dependency	NOUN
futech-324	213	30	within	within	ADP
futech-324	213	31	the	the	DET
futech-324	213	32	data	datum	NOUN
futech-324	213	33	whilst	whilst	SCONJ
futech-324	213	34	providing	provide	VERB
futech-324	213	35	a	a	DET
futech-324	213	36	baseline	baseline	NOUN
futech-324	213	37	for	for	ADP
futech-324	213	38	the	the	DET
futech-324	213	39	classification	classification	NOUN
futech-324	213	40	of	of	ADP
futech-324	213	41	toxic	toxic	ADJ
futech-324	213	42	comments	comment	NOUN
futech-324	213	43	.	.	PUNCT
futech-324	214	1	4.3	4.3	NUM
futech-324	214	2	applying	apply	VERB
futech-324	214	3	gru	gru	NOUN
futech-324	214	4	with	with	ADP
futech-324	214	5	tf	tf	PROPN
futech-324	214	6	-	-	PUNCT
futech-324	214	7	idf	idf	PROPN
futech-324	214	8	this	this	DET
futech-324	214	9	section	section	NOUN
futech-324	214	10	demonstrates	demonstrate	VERB
futech-324	214	11	the	the	DET
futech-324	214	12	application	application	NOUN
futech-324	214	13	of	of	ADP
futech-324	214	14	a	a	DET
futech-324	214	15	gru	gru	NOUN
futech-324	214	16	(	(	PUNCT
futech-324	214	17	gated	gate	VERB
futech-324	214	18	recurrent	recurrent	ADJ
futech-324	214	19	unit	unit	NOUN
futech-324	214	20	)	)	PUNCT
futech-324	214	21	model	model	NOUN
futech-324	214	22	with	with	ADP
futech-324	214	23	tf	tf	PROPN
futech-324	214	24	-	-	PUNCT
futech-324	214	25	idf	idf	PROPN
futech-324	214	26	for	for	ADP
futech-324	214	27	the	the	DET
futech-324	214	28	task	task	NOUN
futech-324	214	29	of	of	ADP
futech-324	214	30	toxic	toxic	ADJ
futech-324	214	31	comment	comment	NOUN
futech-324	214	32	classification	classification	NOUN
futech-324	214	33	.	.	PUNCT
futech-324	215	1	similar	similar	ADJ
futech-324	215	2	to	to	ADP
futech-324	215	3	the	the	DET
futech-324	215	4	lstm	lstm	PROPN
futech-324	215	5	model	model	NOUN
futech-324	215	6	,	,	PUNCT
futech-324	215	7	tf	tf	PROPN
futech-324	215	8	-	-	PUNCT
futech-324	215	9	idf	idf	PROPN
futech-324	215	10	is	be	AUX
futech-324	215	11	used	use	VERB
futech-324	215	12	in	in	ADP
futech-324	215	13	this	this	DET
futech-324	215	14	model	model	NOUN
futech-324	215	15	.	.	PUNCT
futech-324	216	1	the	the	DET
futech-324	216	2	gru	gru	NOUN
futech-324	216	3	model	model	NOUN
futech-324	216	4	consists	consist	VERB
futech-324	216	5	of	of	ADP
futech-324	216	6	an	an	DET
futech-324	216	7	embedding	embed	VERB
futech-324	216	8	layer	layer	NOUN
futech-324	216	9	,	,	PUNCT
futech-324	216	10	single	single	ADJ
futech-324	216	11	gru	gru	NOUN
futech-324	216	12	layer	layer	NOUN
futech-324	216	13	of	of	ADP
futech-324	216	14	size	size	NOUN
futech-324	216	15	64	64	NUM
futech-324	216	16	,	,	PUNCT
futech-324	216	17	followed	follow	VERB
futech-324	216	18	by	by	ADP
futech-324	216	19	a	a	DET
futech-324	216	20	dense	dense	ADJ
futech-324	216	21	output	output	NOUN
futech-324	216	22	layer	layer	NOUN
futech-324	216	23	for	for	ADP
futech-324	216	24	predicting	predict	VERB
futech-324	216	25	binary	binary	ADJ
futech-324	216	26	classes	class	NOUN
futech-324	216	27	.	.	PUNCT
futech-324	217	1	the	the	DET
futech-324	217	2	model	model	NOUN
futech-324	217	3	is	be	AUX
futech-324	217	4	evaluated	evaluate	VERB
futech-324	217	5	on	on	ADP
futech-324	217	6	the	the	DET
futech-324	217	7	test	test	NOUN
futech-324	217	8	data	datum	NOUN
futech-324	217	9	after	after	ADP
futech-324	217	10	training	train	VERB
futech-324	217	11	with	with	ADP
futech-324	217	12	a	a	DET
futech-324	217	13	batch	batch	NOUN
futech-324	217	14	size	size	NOUN
futech-324	217	15	of	of	ADP
futech-324	217	16	16	16	NUM
futech-324	217	17	.	.	PUNCT
futech-324	218	1	the	the	DET
futech-324	218	2	final	final	ADJ
futech-324	218	3	validation	validation	NOUN
futech-324	218	4	accuracy	accuracy	NOUN
futech-324	218	5	~90	~90	PRON
futech-324	218	6	%	%	NOUN
futech-324	218	7	generally	generally	ADV
futech-324	218	8	indicates	indicate	VERB
futech-324	218	9	that	that	SCONJ
futech-324	218	10	the	the	DET
futech-324	218	11	gru	gru	NOUN
futech-324	218	12	model	model	NOUN
futech-324	218	13	also	also	ADV
futech-324	218	14	effectively	effectively	ADV
futech-324	218	15	captured	capture	VERB
futech-324	218	16	the	the	DET
futech-324	218	17	relevant	relevant	ADJ
futech-324	218	18	patterns	pattern	NOUN
futech-324	218	19	from	from	ADP
futech-324	218	20	the	the	DET
futech-324	218	21	text	text	NOUN
futech-324	218	22	and	and	CCONJ
futech-324	218	23	achieved	achieve	VERB
futech-324	218	24	comparable	comparable	ADJ
futech-324	218	25	classification	classification	NOUN
futech-324	218	26	accuracy	accuracy	NOUN
futech-324	218	27	with	with	ADP
futech-324	218	28	the	the	DET
futech-324	218	29	lstm	lstm	NOUN
futech-324	218	30	model	model	NOUN
futech-324	218	31	for	for	ADP
futech-324	218	32	classifying	classify	VERB
futech-324	218	33	a	a	DET
futech-324	218	34	toxic	toxic	ADJ
futech-324	218	35	comment	comment	NOUN
futech-324	218	36	.	.	PUNCT
futech-324	219	1	4.4	4.4	NUM
futech-324	219	2	applying	apply	VERB
futech-324	219	3	rnn	rnn	NOUN
futech-324	219	4	with	with	ADP
futech-324	219	5	word2vec	word2vec	PRON
futech-324	219	6	in	in	ADP
futech-324	219	7	this	this	DET
futech-324	219	8	approach	approach	NOUN
futech-324	219	9	,	,	PUNCT
futech-324	219	10	word2vec	word2vec	X
futech-324	219	11	embeddings	embedding	NOUN
futech-324	219	12	were	be	AUX
futech-324	219	13	combined	combine	VERB
futech-324	219	14	with	with	ADP
futech-324	219	15	an	an	DET
futech-324	219	16	rnn	rnn	NOUN
futech-324	219	17	to	to	PART
futech-324	219	18	classify	classify	VERB
futech-324	219	19	toxic	toxic	ADJ
futech-324	219	20	and	and	CCONJ
futech-324	219	21	non	non	ADJ
futech-324	219	22	-	-	ADJ
futech-324	219	23	toxic	toxic	ADJ
futech-324	219	24	comments	comment	NOUN
futech-324	219	25	.	.	PUNCT
futech-324	220	1	word2vec	word2vec	X
futech-324	220	2	embeddings	embedding	NOUN
futech-324	220	3	were	be	AUX
futech-324	220	4	first	first	ADV
futech-324	220	5	generated	generate	VERB
futech-324	220	6	by	by	ADP
futech-324	220	7	training	training	NOUN
futech-324	220	8	on	on	ADP
futech-324	220	9	the	the	DET
futech-324	220	10	tokenized	tokenized	ADJ
futech-324	220	11	comments	comment	NOUN
futech-324	220	12	from	from	ADP
futech-324	220	13	the	the	DET
futech-324	220	14	jigsaw	jigsaw	NOUN
futech-324	220	15	toxic	toxic	ADJ
futech-324	220	16	comment	comment	NOUN
futech-324	220	17	dataset	dataset	NOUN
futech-324	220	18	.	.	PUNCT
futech-324	221	1	each	each	DET
futech-324	221	2	comment	comment	NOUN
futech-324	221	3	was	be	AUX
futech-324	221	4	represented	represent	VERB
futech-324	221	5	by	by	ADP
futech-324	221	6	a	a	DET
futech-324	221	7	fixed	fix	VERB
futech-324	221	8	-	-	PUNCT
futech-324	221	9	length	length	NOUN
futech-324	221	10	vector	vector	NOUN
futech-324	221	11	by	by	ADP
futech-324	221	12	averaging	average	VERB
futech-324	221	13	the	the	DET
futech-324	221	14	embeddings	embedding	NOUN
futech-324	221	15	of	of	ADP
futech-324	221	16	its	its	PRON
futech-324	221	17	words	word	NOUN
futech-324	221	18	,	,	PUNCT
futech-324	221	19	with	with	ADP
futech-324	221	20	zero	zero	NUM
futech-324	221	21	vectors	vector	NOUN
futech-324	221	22	assigned	assign	VERB
futech-324	221	23	to	to	ADP
futech-324	221	24	out	out	ADP
futech-324	221	25	-	-	PUNCT
futech-324	221	26	of	of	ADP
futech-324	221	27	-	-	PUNCT
futech-324	221	28	vocabulary	vocabulary	ADJ
futech-324	221	29	words	word	NOUN
futech-324	221	30	.	.	PUNCT
futech-324	222	1	the	the	DET
futech-324	222	2	padding	padding	NOUN
futech-324	222	3	was	be	AUX
futech-324	222	4	applied	apply	VERB
futech-324	222	5	to	to	PART
futech-324	222	6	ensure	ensure	VERB
futech-324	222	7	uniform	uniform	ADJ
futech-324	222	8	input	input	NOUN
futech-324	222	9	lengths	length	NOUN
futech-324	222	10	for	for	ADP
futech-324	222	11	the	the	DET
futech-324	222	12	rnn	rnn	NOUN
futech-324	222	13	.	.	PUNCT
futech-324	223	1	the	the	DET
futech-324	223	2	model	model	NOUN
futech-324	223	3	was	be	AUX
futech-324	223	4	trained	train	VERB
futech-324	223	5	for	for	ADP
futech-324	223	6	20	20	NUM
futech-324	223	7	epochs	epoch	NOUN
futech-324	223	8	with	with	ADP
futech-324	223	9	a	a	DET
futech-324	223	10	batch	batch	NOUN
futech-324	223	11	size	size	NOUN
futech-324	223	12	of	of	ADP
futech-324	223	13	16	16	NUM
futech-324	223	14	.	.	PUNCT
futech-324	224	1	figure	figure	NOUN
futech-324	224	2	2	2	NUM
futech-324	224	3	shows	show	VERB
futech-324	224	4	epoch	epoch	NOUN
futech-324	224	5	-	-	PUNCT
futech-324	224	6	wise	wise	ADJ
futech-324	224	7	accuracy	accuracy	NOUN
futech-324	224	8	,	,	PUNCT
futech-324	224	9	and	and	CCONJ
futech-324	224	10	figure	figure	VERB
futech-324	224	11	3	3	NUM
futech-324	224	12	shows	show	VERB
futech-324	224	13	loss	loss	NOUN
futech-324	224	14	values	value	NOUN
futech-324	224	15	with	with	ADP
futech-324	224	16	the	the	DET
futech-324	224	17	model	model	NOUN
futech-324	224	18	.	.	PUNCT
futech-324	225	1	it	it	PRON
futech-324	225	2	achieved	achieve	VERB
futech-324	225	3	a	a	DET
futech-324	225	4	validation	validation	NOUN
futech-324	225	5	accuracy	accuracy	NOUN
futech-324	225	6	of	of	ADP
futech-324	225	7	approximately	approximately	ADV
futech-324	225	8	90.76	90.76	NUM
futech-324	225	9	%	%	NOUN
futech-324	225	10	,	,	PUNCT
futech-324	225	11	demonstrating	demonstrate	VERB
futech-324	225	12	the	the	DET
futech-324	225	13	effectiveness	effectiveness	NOUN
futech-324	225	14	of	of	ADP
futech-324	225	15	using	use	VERB
futech-324	225	16	word2vec	word2vec	PROPN
futech-324	225	17	embeddings	embedding	NOUN
futech-324	225	18	and	and	CCONJ
futech-324	225	19	rnns	rnn	NOUN
futech-324	225	20	for	for	ADP
futech-324	225	21	toxic	toxic	ADJ
futech-324	225	22	comment	comment	NOUN
futech-324	225	23	classification	classification	NOUN
futech-324	225	24	.	.	PUNCT
futech-324	226	1	figure	figure	NOUN
futech-324	226	2	2	2	NUM
futech-324	226	3	.	.	X
futech-324	227	1	rnn+	rnn+	PROPN
futech-324	227	2	word2vec	word2vec	X
futech-324	227	3	model	model	VERB
futech-324	227	4	epoch	epoch	PROPN
futech-324	227	5	wise	wise	ADJ
futech-324	227	6	accuracy	accuracy	NOUN
futech-324	227	7	figure	figure	NOUN
futech-324	227	8	3	3	NUM
futech-324	227	9	.	.	PUNCT
futech-324	228	1	rnn+	rnn+	PROPN
futech-324	228	2	word2vec	word2vec	PRON
futech-324	228	3	model	model	VERB
futech-324	228	4	epoch	epoch	PROPN
futech-324	228	5	wise	wise	ADJ
futech-324	228	6	loss	loss	NOUN
futech-324	228	7	s.	s.	PROPN
futech-324	228	8	sushma	sushma	PROPN
futech-324	228	9	et	et	PROPN
futech-324	228	10	al	al	PROPN
futech-324	228	11	.	.	PROPN
futech-324	228	12	/future	/future	PUNCT
futech-324	228	13	technology	technology	PROPN
futech-324	228	14	august	august	PROPN
futech-324	228	15	2025|	2025|	NUM
futech-324	229	1	volume	volume	NOUN
futech-324	229	2	04	04	NUM
futech-324	230	1	|	|	NOUN
futech-324	230	2	issue	issue	NOUN
futech-324	230	3	03	03	NUM
futech-324	231	1	|	|	ADV
futech-324	231	2	pages	page	VERB
futech-324	231	3	76	76	NUM
futech-324	231	4	-	-	SYM
futech-324	231	5	84	84	NUM
futech-324	231	6	81	81	NUM
futech-324	231	7	4.5	4.5	NUM
futech-324	231	8	applying	apply	VERB
futech-324	231	9	lstm	lstm	NOUN
futech-324	231	10	with	with	ADP
futech-324	231	11	word2vec	word2vec	PRON
futech-324	231	12	in	in	ADP
futech-324	231	13	this	this	DET
futech-324	231	14	approach	approach	NOUN
futech-324	231	15	,	,	PUNCT
futech-324	231	16	the	the	DET
futech-324	231	17	lstm	lstm	PROPN
futech-324	231	18	model	model	NOUN
futech-324	231	19	was	be	AUX
futech-324	231	20	combined	combine	VERB
futech-324	231	21	with	with	ADP
futech-324	231	22	word2vec	word2vec	X
futech-324	231	23	embeddings	embedding	NOUN
futech-324	231	24	.	.	PUNCT
futech-324	232	1	the	the	DET
futech-324	232	2	word2vec	word2vec	PROPN
futech-324	232	3	embeddings	embedding	NOUN
futech-324	232	4	were	be	AUX
futech-324	232	5	first	first	ADV
futech-324	232	6	generated	generate	VERB
futech-324	232	7	by	by	ADP
futech-324	232	8	training	train	VERB
futech-324	232	9	a	a	DET
futech-324	232	10	word2vec	word2vec	X
futech-324	232	11	model	model	NOUN
futech-324	232	12	on	on	ADP
futech-324	232	13	the	the	DET
futech-324	232	14	tokenized	tokenized	ADJ
futech-324	232	15	comments	comment	NOUN
futech-324	232	16	from	from	ADP
futech-324	232	17	the	the	DET
futech-324	232	18	jigsaw	jigsaw	NOUN
futech-324	232	19	toxic	toxic	ADJ
futech-324	232	20	comment	comment	NOUN
futech-324	232	21	dataset	dataset	NOUN
futech-324	232	22	.	.	PUNCT
futech-324	233	1	each	each	DET
futech-324	233	2	comment	comment	NOUN
futech-324	233	3	was	be	AUX
futech-324	233	4	then	then	ADV
futech-324	233	5	represented	represent	VERB
futech-324	233	6	as	as	ADP
futech-324	233	7	a	a	DET
futech-324	233	8	fixed	fix	VERB
futech-324	233	9	-	-	PUNCT
futech-324	233	10	length	length	NOUN
futech-324	233	11	vector	vector	NOUN
futech-324	233	12	,	,	PUNCT
futech-324	233	13	computed	compute	VERB
futech-324	233	14	by	by	ADP
futech-324	233	15	averaging	average	VERB
futech-324	233	16	the	the	DET
futech-324	233	17	word	word	NOUN
futech-324	233	18	embeddings	embedding	NOUN
futech-324	233	19	in	in	ADP
futech-324	233	20	the	the	DET
futech-324	233	21	comment	comment	NOUN
futech-324	233	22	.	.	PUNCT
futech-324	234	1	the	the	DET
futech-324	234	2	data	datum	NOUN
futech-324	234	3	was	be	AUX
futech-324	234	4	divided	divide	VERB
futech-324	234	5	into	into	ADP
futech-324	234	6	80	80	NUM
futech-324	234	7	%	%	NOUN
futech-324	234	8	training	training	NOUN
futech-324	234	9	and	and	CCONJ
futech-324	234	10	20	20	NUM
futech-324	234	11	%	%	NOUN
futech-324	234	12	testing	testing	NOUN
futech-324	234	13	.	.	PUNCT
futech-324	235	1	next	next	ADV
futech-324	235	2	,	,	PUNCT
futech-324	235	3	padding	padding	NOUN
futech-324	235	4	was	be	AUX
futech-324	235	5	performed	perform	VERB
futech-324	235	6	to	to	PART
futech-324	235	7	ensure	ensure	VERB
futech-324	235	8	uniform	uniform	ADJ
futech-324	235	9	input	input	NOUN
futech-324	235	10	lengths	length	NOUN
futech-324	235	11	for	for	ADP
futech-324	235	12	the	the	DET
futech-324	235	13	lstm	lstm	PROPN
futech-324	235	14	model	model	NOUN
futech-324	235	15	.	.	PUNCT
futech-324	236	1	the	the	DET
futech-324	236	2	model	model	NOUN
futech-324	236	3	was	be	AUX
futech-324	236	4	trained	train	VERB
futech-324	236	5	for	for	ADP
futech-324	236	6	20	20	NUM
futech-324	236	7	epochs	epoch	NOUN
futech-324	236	8	with	with	ADP
futech-324	236	9	a	a	DET
futech-324	236	10	batch	batch	NOUN
futech-324	236	11	size	size	NOUN
futech-324	236	12	of	of	ADP
futech-324	236	13	16	16	NUM
futech-324	236	14	,	,	PUNCT
futech-324	236	15	achieving	achieve	VERB
futech-324	236	16	a	a	DET
futech-324	236	17	92.25	92.25	NUM
futech-324	236	18	%	%	NOUN
futech-324	236	19	validation	validation	NOUN
futech-324	236	20	accuracy	accuracy	NOUN
futech-324	236	21	.	.	PUNCT
futech-324	237	1	epoch	epoch	NOUN
futech-324	237	2	-	-	PUNCT
futech-324	237	3	wise	wise	ADJ
futech-324	237	4	accuracy	accuracy	NOUN
futech-324	237	5	values	value	NOUN
futech-324	237	6	with	with	ADP
futech-324	237	7	the	the	DET
futech-324	237	8	model	model	NOUN
futech-324	237	9	are	be	AUX
futech-324	237	10	shown	show	VERB
futech-324	237	11	in	in	ADP
futech-324	237	12	figure	figure	NOUN
futech-324	237	13	4	4	NUM
futech-324	237	14	.	.	PUNCT
futech-324	238	1	epoch	epoch	NOUN
futech-324	238	2	-	-	PUNCT
futech-324	238	3	wise	wise	ADJ
futech-324	238	4	loss	loss	NOUN
futech-324	238	5	values	value	NOUN
futech-324	238	6	with	with	ADP
futech-324	238	7	the	the	DET
futech-324	238	8	model	model	NOUN
futech-324	238	9	are	be	AUX
futech-324	238	10	shown	show	VERB
futech-324	238	11	in	in	ADP
futech-324	238	12	figure	figure	NOUN
futech-324	238	13	5	5	NUM
futech-324	238	14	.	.	PUNCT
futech-324	239	1	the	the	DET
futech-324	239	2	fact	fact	NOUN
futech-324	239	3	that	that	SCONJ
futech-324	239	4	lstm	lstm	NOUN
futech-324	239	5	with	with	ADP
futech-324	239	6	word2vec	word2vec	X
futech-324	239	7	embeddings	embedding	NOUN
futech-324	239	8	outperforms	outperform	VERB
futech-324	239	9	all	all	DET
futech-324	239	10	other	other	ADJ
futech-324	239	11	models	model	NOUN
futech-324	239	12	and	and	CCONJ
futech-324	239	13	proves	prove	VERB
futech-324	239	14	efficient	efficient	ADJ
futech-324	239	15	demonstrates	demonstrate	VERB
futech-324	239	16	the	the	DET
futech-324	239	17	potential	potential	NOUN
futech-324	239	18	of	of	ADP
futech-324	239	19	this	this	DET
futech-324	239	20	combination	combination	NOUN
futech-324	239	21	.	.	PUNCT
futech-324	240	1	figure	figure	NOUN
futech-324	240	2	4	4	NUM
futech-324	240	3	.	.	PUNCT
futech-324	240	4	lstm+	lstm+	PART
futech-324	241	1	word2vec	word2vec	PROPN
futech-324	241	2	model	model	PROPN
futech-324	241	3	epoch	epoch	PROPN
futech-324	241	4	-	-	PUNCT
futech-324	241	5	wise	wise	ADJ
futech-324	241	6	accuracy	accuracy	NOUN
futech-324	241	7	figure	figure	NOUN
futech-324	241	8	5	5	NUM
futech-324	241	9	.	.	PUNCT
futech-324	241	10	lstm+	lstm+	PART
futech-324	242	1	word2vec	word2vec	PROPN
futech-324	242	2	model	model	PROPN
futech-324	242	3	epoch	epoch	PROPN
futech-324	242	4	wise	wise	ADJ
futech-324	242	5	loss	loss	NOUN
futech-324	242	6	4.6	4.6	NUM
futech-324	242	7	applying	apply	VERB
futech-324	242	8	gru	gru	NOUN
futech-324	242	9	with	with	ADP
futech-324	242	10	word2vec	word2vec	PRON
futech-324	242	11	in	in	ADP
futech-324	242	12	this	this	DET
futech-324	242	13	section	section	NOUN
futech-324	242	14	,	,	PUNCT
futech-324	242	15	the	the	DET
futech-324	242	16	gru	gru	NOUN
futech-324	242	17	model	model	NOUN
futech-324	242	18	is	be	AUX
futech-324	242	19	applied	apply	VERB
futech-324	242	20	with	with	ADP
futech-324	242	21	word2vec	word2vec	X
futech-324	242	22	embeddings	embedding	NOUN
futech-324	242	23	for	for	ADP
futech-324	242	24	toxic	toxic	ADJ
futech-324	242	25	comment	comment	NOUN
futech-324	242	26	classification	classification	NOUN
futech-324	242	27	.	.	PUNCT
futech-324	243	1	similar	similar	ADJ
futech-324	243	2	to	to	ADP
futech-324	243	3	the	the	DET
futech-324	243	4	previous	previous	ADJ
futech-324	243	5	methods	method	NOUN
futech-324	243	6	,	,	PUNCT
futech-324	243	7	the	the	DET
futech-324	243	8	word2vec	word2vec	PROPN
futech-324	243	9	model	model	NOUN
futech-324	243	10	was	be	AUX
futech-324	243	11	trained	train	VERB
futech-324	243	12	on	on	ADP
futech-324	243	13	the	the	DET
futech-324	243	14	tokenized	tokenized	ADJ
futech-324	243	15	comments	comment	NOUN
futech-324	243	16	from	from	ADP
futech-324	243	17	the	the	DET
futech-324	243	18	jigsaw	jigsaw	NOUN
futech-324	243	19	toxic	toxic	ADJ
futech-324	243	20	comment	comment	NOUN
futech-324	243	21	dataset	dataset	NOUN
futech-324	243	22	.	.	PUNCT
futech-324	244	1	each	each	DET
futech-324	244	2	comment	comment	NOUN
futech-324	244	3	was	be	AUX
futech-324	244	4	then	then	ADV
futech-324	244	5	transformed	transform	VERB
futech-324	244	6	into	into	ADP
futech-324	244	7	a	a	DET
futech-324	244	8	fixed	fix	VERB
futech-324	244	9	-	-	PUNCT
futech-324	244	10	length	length	NOUN
futech-324	244	11	vector	vector	NOUN
futech-324	244	12	using	use	VERB
futech-324	244	13	the	the	DET
futech-324	244	14	average	average	NOUN
futech-324	244	15	of	of	ADP
futech-324	244	16	word	word	NOUN
futech-324	244	17	embeddings	embedding	NOUN
futech-324	244	18	in	in	ADP
futech-324	244	19	the	the	DET
futech-324	244	20	comment	comment	NOUN
futech-324	244	21	.	.	PUNCT
futech-324	245	1	the	the	DET
futech-324	245	2	model	model	NOUN
futech-324	245	3	was	be	AUX
futech-324	245	4	trained	train	VERB
futech-324	245	5	for	for	ADP
futech-324	245	6	20	20	NUM
futech-324	245	7	epochs	epoch	NOUN
futech-324	245	8	with	with	ADP
futech-324	245	9	a	a	DET
futech-324	245	10	batch	batch	NOUN
futech-324	245	11	size	size	NOUN
futech-324	245	12	of	of	ADP
futech-324	245	13	16	16	NUM
futech-324	245	14	.	.	PUNCT
futech-324	246	1	the	the	DET
futech-324	246	2	gru	gru	NOUN
futech-324	246	3	model	model	NOUN
futech-324	246	4	achieved	achieve	VERB
futech-324	246	5	a	a	DET
futech-324	246	6	validation	validation	NOUN
futech-324	246	7	accuracy	accuracy	NOUN
futech-324	246	8	of	of	ADP
futech-324	246	9	92.11	92.11	NUM
futech-324	246	10	%	%	NOUN
futech-324	246	11	,	,	PUNCT
futech-324	246	12	showcasing	showcase	VERB
futech-324	246	13	its	its	PRON
futech-324	246	14	ability	ability	NOUN
futech-324	246	15	to	to	PART
futech-324	246	16	capture	capture	VERB
futech-324	246	17	sequential	sequential	ADJ
futech-324	246	18	patterns	pattern	NOUN
futech-324	246	19	in	in	ADP
futech-324	246	20	the	the	DET
futech-324	246	21	text	text	NOUN
futech-324	246	22	and	and	CCONJ
futech-324	246	23	effectively	effectively	ADV
futech-324	246	24	classify	classify	VERB
futech-324	246	25	toxic	toxic	ADJ
futech-324	246	26	and	and	CCONJ
futech-324	246	27	non	non	ADJ
futech-324	246	28	-	-	ADJ
futech-324	246	29	toxic	toxic	ADJ
futech-324	246	30	comments	comment	NOUN
futech-324	246	31	.	.	PUNCT
futech-324	247	1	figure	figure	VERB
futech-324	247	2	6	6	NUM
futech-324	247	3	and	and	CCONJ
futech-324	247	4	figure	figure	VERB
futech-324	247	5	7	7	NUM
futech-324	247	6	show	show	VERB
futech-324	247	7	epoch	epoch	NOUN
futech-324	247	8	-	-	PUNCT
futech-324	247	9	wise	wise	ADJ
futech-324	247	10	accuracy	accuracy	NOUN
futech-324	247	11	and	and	CCONJ
futech-324	247	12	loss	loss	NOUN
futech-324	247	13	values	value	NOUN
futech-324	247	14	with	with	ADP
futech-324	247	15	the	the	DET
futech-324	247	16	model	model	NOUN
futech-324	247	17	,	,	PUNCT
futech-324	247	18	respectively	respectively	ADV
futech-324	247	19	.	.	PUNCT
futech-324	248	1	figure	figure	NOUN
futech-324	248	2	6	6	NUM
futech-324	248	3	.	.	PUNCT
futech-324	249	1	gru+	gru+	NOUN
futech-324	249	2	word2vec	word2vec	X
futech-324	249	3	model	model	VERB
futech-324	249	4	epoch	epoch	PROPN
futech-324	249	5	-	-	PUNCT
futech-324	249	6	wise	wise	ADJ
futech-324	249	7	accuracy	accuracy	NOUN
futech-324	249	8	figure	figure	NOUN
futech-324	249	9	7	7	NUM
futech-324	249	10	.	.	PUNCT
futech-324	250	1	gru+	gru+	NOUN
futech-324	250	2	word2vec	word2vec	X
futech-324	250	3	model	model	VERB
futech-324	250	4	epoch	epoch	PROPN
futech-324	250	5	-	-	PUNCT
futech-324	250	6	wise	wise	ADJ
futech-324	250	7	loss	loss	NOUN
futech-324	250	8	4.7	4.7	NUM
futech-324	250	9	applying	apply	VERB
futech-324	250	10	rnn	rnn	NOUN
futech-324	250	11	with	with	ADP
futech-324	250	12	bert	bert	PROPN
futech-324	250	13	an	an	DET
futech-324	250	14	rnn	rnn	NOUN
futech-324	250	15	model	model	NOUN
futech-324	250	16	is	be	AUX
futech-324	250	17	combined	combine	VERB
futech-324	250	18	with	with	ADP
futech-324	250	19	bert	bert	PROPN
futech-324	250	20	embeddings	embedding	NOUN
futech-324	250	21	to	to	PART
futech-324	250	22	classify	classify	VERB
futech-324	250	23	toxic	toxic	ADJ
futech-324	250	24	comments	comment	NOUN
futech-324	250	25	.	.	PUNCT
futech-324	251	1	the	the	DET
futech-324	251	2	dataset	dataset	NOUN
futech-324	251	3	is	be	AUX
futech-324	251	4	preprocessed	preprocesse	VERB
futech-324	251	5	using	use	VERB
futech-324	251	6	the	the	DET
futech-324	251	7	distilbert	distilbert	ADJ
futech-324	251	8	tokenizer	tokenizer	NOUN
futech-324	251	9	,	,	PUNCT
futech-324	251	10	and	and	CCONJ
futech-324	251	11	embeddings	embedding	NOUN
futech-324	251	12	are	be	AUX
futech-324	251	13	extracted	extract	VERB
futech-324	251	14	from	from	ADP
futech-324	251	15	the	the	DET
futech-324	251	16	[	[	X
futech-324	251	17	cls	cls	X
futech-324	251	18	]	]	PUNCT
futech-324	251	19	token	token	VERB
futech-324	251	20	.	.	PUNCT
futech-324	252	1	these	these	DET
futech-324	252	2	embeddings	embedding	NOUN
futech-324	252	3	are	be	AUX
futech-324	252	4	passed	pass	VERB
futech-324	252	5	through	through	ADP
futech-324	252	6	a	a	DET
futech-324	252	7	simplernn	simplernn	NOUN
futech-324	252	8	layer	layer	NOUN
futech-324	252	9	and	and	CCONJ
futech-324	252	10	a	a	DET
futech-324	252	11	dense	dense	ADJ
futech-324	252	12	layer	layer	NOUN
futech-324	252	13	for	for	ADP
futech-324	252	14	binary	binary	ADJ
futech-324	252	15	classification	classification	NOUN
futech-324	252	16	.	.	PUNCT
futech-324	253	1	after	after	ADP
futech-324	253	2	training	train	VERB
futech-324	253	3	for	for	ADP
futech-324	253	4	10	10	NUM
futech-324	253	5	epochs	epoch	NOUN
futech-324	253	6	,	,	PUNCT
futech-324	253	7	the	the	DET
futech-324	253	8	model	model	NOUN
futech-324	253	9	achieved	achieve	VERB
futech-324	253	10	a	a	DET
futech-324	253	11	validation	validation	NOUN
futech-324	253	12	accuracy	accuracy	NOUN
futech-324	253	13	of	of	ADP
futech-324	253	14	approximately	approximately	ADV
futech-324	253	15	75.7	75.7	NUM
futech-324	253	16	%	%	NOUN
futech-324	253	17	,	,	PUNCT
futech-324	253	18	demonstrating	demonstrate	VERB
futech-324	253	19	the	the	DET
futech-324	253	20	effectiveness	effectiveness	NOUN
futech-324	253	21	of	of	ADP
futech-324	253	22	rnns	rnn	NOUN
futech-324	253	23	with	with	ADP
futech-324	253	24	bert	bert	PROPN
futech-324	253	25	embeddings	embedding	NOUN
futech-324	253	26	for	for	ADP
futech-324	253	27	toxic	toxic	ADJ
futech-324	253	28	comment	comment	NOUN
futech-324	253	29	classification	classification	NOUN
futech-324	253	30	.	.	PUNCT
futech-324	254	1	figure	figure	NOUN
futech-324	254	2	8	8	NUM
futech-324	254	3	and	and	CCONJ
futech-324	254	4	figure	figure	VERB
futech-324	254	5	9	9	NUM
futech-324	254	6	depict	depict	NOUN
futech-324	254	7	epochwise	epochwise	NOUN
futech-324	254	8	accuracy	accuracy	NOUN
futech-324	254	9	and	and	CCONJ
futech-324	254	10	loss	loss	NOUN
futech-324	254	11	for	for	ADP
futech-324	254	12	this	this	DET
futech-324	254	13	model	model	NOUN
futech-324	254	14	.	.	PUNCT
futech-324	255	1	s.	s.	PROPN
futech-324	255	2	sushma	sushma	PROPN
futech-324	255	3	et	et	PROPN
futech-324	255	4	al	al	PROPN
futech-324	255	5	.	.	PROPN
futech-324	255	6	/future	/future	PUNCT
futech-324	255	7	technology	technology	PROPN
futech-324	255	8	august	august	PROPN
futech-324	255	9	2025|	2025|	NUM
futech-324	255	10	volume	volume	NOUN
futech-324	255	11	04	04	NUM
futech-324	256	1	|	|	NOUN
futech-324	256	2	issue	issue	NOUN
futech-324	256	3	03	03	NUM
futech-324	257	1	|	|	ADV
futech-324	257	2	pages	page	VERB
futech-324	257	3	76	76	NUM
futech-324	257	4	-	-	SYM
futech-324	257	5	84	84	NUM
futech-324	257	6	82	82	NUM
futech-324	257	7	figure	figure	NOUN
futech-324	257	8	8	8	NUM
futech-324	257	9	.	.	PUNCT
futech-324	258	1	rnn+	rnn+	PROPN
futech-324	258	2	bert	bert	PROPN
futech-324	258	3	model	model	PROPN
futech-324	258	4	epoch	epoch	PROPN
futech-324	258	5	wise	wise	ADJ
futech-324	258	6	accuracy	accuracy	NOUN
futech-324	258	7	figure	figure	NOUN
futech-324	258	8	9	9	NUM
futech-324	258	9	.	.	PUNCT
futech-324	259	1	rnn+	rnn+	PROPN
futech-324	259	2	bert	bert	PROPN
futech-324	259	3	model	model	PROPN
futech-324	259	4	epoch	epoch	PROPN
futech-324	259	5	wise	wise	ADJ
futech-324	259	6	loss	loss	NOUN
futech-324	259	7	4.8	4.8	NUM
futech-324	259	8	applying	apply	VERB
futech-324	259	9	lstm	lstm	NOUN
futech-324	259	10	and	and	CCONJ
futech-324	259	11	gru	gru	VERB
futech-324	259	12	with	with	ADP
futech-324	259	13	bert	bert	PROPN
futech-324	259	14	in	in	ADP
futech-324	259	15	addition	addition	NOUN
futech-324	259	16	to	to	ADP
futech-324	259	17	the	the	DET
futech-324	259	18	rnn	rnn	PROPN
futech-324	259	19	-	-	PUNCT
futech-324	259	20	based	base	VERB
futech-324	259	21	model	model	NOUN
futech-324	259	22	,	,	PUNCT
futech-324	259	23	the	the	DET
futech-324	259	24	effectiveness	effectiveness	NOUN
futech-324	259	25	of	of	ADP
futech-324	259	26	lstm	lstm	PROPN
futech-324	259	27	and	and	CCONJ
futech-324	259	28	gru	gru	NOUN
futech-324	259	29	models	model	NOUN
futech-324	259	30	combined	combine	VERB
futech-324	259	31	with	with	ADP
futech-324	259	32	bert	bert	NOUN
futech-324	259	33	embeddings	embedding	NOUN
futech-324	259	34	for	for	ADP
futech-324	259	35	toxic	toxic	ADJ
futech-324	259	36	comment	comment	NOUN
futech-324	259	37	classification	classification	NOUN
futech-324	259	38	is	be	AUX
futech-324	259	39	also	also	ADV
futech-324	259	40	evaluated	evaluate	VERB
futech-324	259	41	.	.	PUNCT
futech-324	260	1	both	both	DET
futech-324	260	2	models	model	NOUN
futech-324	260	3	leverage	leverage	VERB
futech-324	260	4	bert	bert	PROPN
futech-324	260	5	's	's	PART
futech-324	260	6	pre	pre	ADJ
futech-324	260	7	-	-	ADJ
futech-324	260	8	trained	train	VERB
futech-324	260	9	embeddings	embedding	NOUN
futech-324	260	10	,	,	PUNCT
futech-324	260	11	which	which	PRON
futech-324	260	12	are	be	AUX
futech-324	260	13	extracted	extract	VERB
futech-324	260	14	using	use	VERB
futech-324	260	15	the	the	DET
futech-324	260	16	distilbert	distilbert	ADJ
futech-324	260	17	tokenizer	tokenizer	NOUN
futech-324	260	18	.	.	PUNCT
futech-324	261	1	the	the	DET
futech-324	261	2	embeddings	embedding	NOUN
futech-324	261	3	are	be	AUX
futech-324	261	4	then	then	ADV
futech-324	261	5	processed	process	VERB
futech-324	261	6	through	through	ADP
futech-324	261	7	lstm	lstm	PROPN
futech-324	261	8	and	and	CCONJ
futech-324	261	9	gru	gru	NOUN
futech-324	261	10	layers	layer	NOUN
futech-324	261	11	,	,	PUNCT
futech-324	261	12	followed	follow	VERB
futech-324	261	13	by	by	ADP
futech-324	261	14	dense	dense	ADJ
futech-324	261	15	layers	layer	NOUN
futech-324	261	16	for	for	ADP
futech-324	261	17	binary	binary	ADJ
futech-324	261	18	classification	classification	NOUN
futech-324	261	19	.	.	PUNCT
futech-324	262	1	for	for	ADP
futech-324	262	2	the	the	DET
futech-324	262	3	lstm	lstm	PROPN
futech-324	262	4	model	model	NOUN
futech-324	262	5	,	,	PUNCT
futech-324	262	6	a	a	DET
futech-324	262	7	validation	validation	NOUN
futech-324	262	8	accuracy	accuracy	NOUN
futech-324	262	9	of	of	ADP
futech-324	262	10	approximately	approximately	ADV
futech-324	262	11	80	80	NUM
futech-324	262	12	%	%	NOUN
futech-324	262	13	was	be	AUX
futech-324	262	14	achieved	achieve	VERB
futech-324	262	15	,	,	PUNCT
futech-324	262	16	while	while	SCONJ
futech-324	262	17	the	the	DET
futech-324	262	18	gru	gru	NOUN
futech-324	262	19	model	model	NOUN
futech-324	262	20	demonstrated	demonstrate	VERB
futech-324	262	21	a	a	DET
futech-324	262	22	slightly	slightly	ADV
futech-324	262	23	lower	low	ADJ
futech-324	262	24	validation	validation	NOUN
futech-324	262	25	accuracy	accuracy	NOUN
futech-324	262	26	of	of	ADP
futech-324	262	27	79	79	NUM
futech-324	262	28	%	%	NOUN
futech-324	262	29	.	.	PUNCT
futech-324	263	1	4.9	4.9	NUM
futech-324	263	2	comparison	comparison	NOUN
futech-324	263	3	of	of	ADP
futech-324	263	4	applied	apply	VERB
futech-324	263	5	models	model	NOUN
futech-324	263	6	the	the	DET
futech-324	263	7	performance	performance	NOUN
futech-324	263	8	of	of	ADP
futech-324	263	9	the	the	DET
futech-324	263	10	proposed	propose	VERB
futech-324	263	11	models	model	NOUN
futech-324	263	12	was	be	AUX
futech-324	263	13	evaluated	evaluate	VERB
futech-324	263	14	using	use	VERB
futech-324	263	15	various	various	ADJ
futech-324	263	16	algorithms	algorithm	NOUN
futech-324	263	17	in	in	ADP
futech-324	263	18	combination	combination	NOUN
futech-324	263	19	with	with	ADP
futech-324	263	20	different	different	ADJ
futech-324	263	21	feature	feature	NOUN
futech-324	263	22	extraction	extraction	NOUN
futech-324	263	23	techniques	technique	NOUN
futech-324	263	24	,	,	PUNCT
futech-324	263	25	including	include	VERB
futech-324	263	26	tf	tf	PROPN
futech-324	263	27	-	-	PUNCT
futech-324	263	28	idf	idf	PROPN
futech-324	263	29	,	,	PUNCT
futech-324	263	30	word2vec	word2vec	X
futech-324	263	31	,	,	PUNCT
futech-324	263	32	and	and	CCONJ
futech-324	263	33	bert	bert	PROPN
futech-324	263	34	.	.	PUNCT
futech-324	264	1	the	the	DET
futech-324	264	2	accuracies	accuracy	NOUN
futech-324	264	3	achieved	achieve	VERB
futech-324	264	4	by	by	ADP
futech-324	264	5	each	each	DET
futech-324	264	6	model	model	NOUN
futech-324	264	7	are	be	AUX
futech-324	264	8	shown	show	VERB
futech-324	264	9	in	in	ADP
futech-324	264	10	table	table	NOUN
futech-324	264	11	1	1	NUM
futech-324	264	12	and	and	CCONJ
futech-324	264	13	figure	figure	VERB
futech-324	264	14	10	10	NUM
futech-324	264	15	.	.	PUNCT
futech-324	265	1	as	as	SCONJ
futech-324	265	2	seen	see	VERB
futech-324	265	3	from	from	ADP
futech-324	265	4	the	the	DET
futech-324	265	5	table	table	NOUN
futech-324	265	6	,	,	PUNCT
futech-324	265	7	the	the	DET
futech-324	265	8	combination	combination	NOUN
futech-324	265	9	of	of	ADP
futech-324	265	10	lstm	lstm	NOUN
futech-324	265	11	with	with	ADP
futech-324	265	12	word2vec	word2vec	PRON
futech-324	265	13	achieved	achieve	VERB
futech-324	265	14	the	the	DET
futech-324	265	15	highest	high	ADJ
futech-324	265	16	accuracy	accuracy	NOUN
futech-324	265	17	of	of	ADP
futech-324	265	18	92.25	92.25	NUM
futech-324	265	19	%	%	NOUN
futech-324	265	20	,	,	PUNCT
futech-324	265	21	followed	follow	VERB
futech-324	265	22	by	by	ADP
futech-324	265	23	lstm	lstm	NOUN
futech-324	265	24	+	+	CCONJ
futech-324	265	25	bert	bert	PROPN
futech-324	265	26	at	at	ADP
futech-324	265	27	92.11	92.11	NUM
futech-324	265	28	%	%	NOUN
futech-324	265	29	.	.	PUNCT
futech-324	266	1	rnn	rnn	PROPN
futech-324	266	2	-	-	PUNCT
futech-324	266	3	based	base	VERB
futech-324	266	4	models	model	NOUN
futech-324	266	5	demonstrated	demonstrate	VERB
futech-324	266	6	competitive	competitive	ADJ
futech-324	266	7	performance	performance	NOUN
futech-324	266	8	,	,	PUNCT
futech-324	266	9	with	with	ADP
futech-324	266	10	rnn	rnn	NOUN
futech-324	266	11	+	+	CCONJ
futech-324	266	12	word2vec	word2vec	VERB
futech-324	266	13	reaching	reach	VERB
futech-324	266	14	an	an	DET
futech-324	266	15	accuracy	accuracy	NOUN
futech-324	266	16	of	of	ADP
futech-324	266	17	90.76	90.76	NUM
futech-324	266	18	%	%	NOUN
futech-324	266	19	,	,	PUNCT
futech-324	266	20	and	and	CCONJ
futech-324	266	21	rnn	rnn	VERB
futech-324	266	22	+	+	CCONJ
futech-324	266	23	tfidf	tfidf	NOUN
futech-324	266	24	yielding	yield	VERB
futech-324	266	25	89.46	89.46	NUM
futech-324	266	26	%	%	NOUN
futech-324	266	27	.	.	PUNCT
futech-324	267	1	gru	gru	NOUN
futech-324	267	2	models	model	NOUN
futech-324	267	3	exhibited	exhibit	VERB
futech-324	267	4	lower	low	ADJ
futech-324	267	5	accuracies	accuracy	NOUN
futech-324	267	6	in	in	ADP
futech-324	267	7	comparison	comparison	NOUN
futech-324	267	8	,	,	PUNCT
futech-324	267	9	with	with	ADP
futech-324	267	10	gru	gru	NOUN
futech-324	267	11	+	+	CCONJ
futech-324	267	12	tfidf	tfidf	NOUN
futech-324	267	13	achieving	achieve	VERB
futech-324	267	14	75.70	75.70	NUM
futech-324	267	15	%	%	NOUN
futech-324	267	16	and	and	CCONJ
futech-324	267	17	gru	gru	NOUN
futech-324	267	18	+	+	CCONJ
futech-324	267	19	word2vec	word2vec	X
futech-324	267	20	and	and	CCONJ
futech-324	267	21	gru	gru	VERB
futech-324	267	22	+	+	CCONJ
futech-324	267	23	bert	bert	NOUN
futech-324	267	24	showing	show	VERB
futech-324	267	25	accuracies	accuracy	NOUN
futech-324	267	26	of	of	ADP
futech-324	267	27	80.00	80.00	NUM
futech-324	267	28	%	%	NOUN
futech-324	267	29	and	and	CCONJ
futech-324	267	30	79.00	79.00	NUM
futech-324	267	31	%	%	NOUN
futech-324	267	32	,	,	PUNCT
futech-324	267	33	respectively	respectively	ADV
futech-324	267	34	.	.	PUNCT
futech-324	267	35	table	table	NOUN
futech-324	267	36	1	1	NUM
futech-324	267	37	.	.	PUNCT
futech-324	267	38	comparison	comparison	NOUN
futech-324	267	39	of	of	ADP
futech-324	267	40	applied	apply	VERB
futech-324	267	41	models	model	NOUN
futech-324	267	42	figure	figure	VERB
futech-324	267	43	10	10	NUM
futech-324	267	44	.	.	PUNCT
futech-324	268	1	accuracy	accuracy	NOUN
futech-324	268	2	comparison	comparison	NOUN
futech-324	268	3	of	of	ADP
futech-324	268	4	applied	apply	VERB
futech-324	268	5	models	model	NOUN
futech-324	268	6	5	5	NUM
futech-324	268	7	.	.	PUNCT
futech-324	268	8	conclusion	conclusion	NOUN
futech-324	268	9	in	in	ADP
futech-324	268	10	this	this	DET
futech-324	268	11	paper	paper	NOUN
futech-324	268	12	,	,	PUNCT
futech-324	268	13	various	various	ADJ
futech-324	268	14	deep	deep	ADJ
futech-324	268	15	learning	learning	NOUN
futech-324	268	16	models	model	NOUN
futech-324	268	17	combined	combine	VERB
futech-324	268	18	with	with	ADP
futech-324	268	19	different	different	ADJ
futech-324	268	20	feature	feature	NOUN
futech-324	268	21	extraction	extraction	NOUN
futech-324	268	22	techniques	technique	NOUN
futech-324	268	23	,	,	PUNCT
futech-324	268	24	such	such	ADJ
futech-324	268	25	as	as	ADP
futech-324	268	26	tf	tf	PROPN
futech-324	268	27	-	-	PUNCT
futech-324	268	28	idf	idf	PROPN
futech-324	268	29	,	,	PUNCT
futech-324	268	30	word2vec	word2vec	X
futech-324	268	31	,	,	PUNCT
futech-324	268	32	and	and	CCONJ
futech-324	268	33	bert	bert	PROPN
futech-324	268	34	,	,	PUNCT
futech-324	268	35	were	be	AUX
futech-324	268	36	evaluated	evaluate	VERB
futech-324	268	37	for	for	ADP
futech-324	268	38	the	the	DET
futech-324	268	39	task	task	NOUN
futech-324	268	40	of	of	ADP
futech-324	268	41	toxic	toxic	ADJ
futech-324	268	42	comment	comment	NOUN
futech-324	268	43	classification	classification	NOUN
futech-324	268	44	.	.	PUNCT
futech-324	269	1	the	the	DET
futech-324	269	2	results	result	NOUN
futech-324	269	3	demonstrated	demonstrate	VERB
futech-324	269	4	that	that	SCONJ
futech-324	269	5	deep	deep	ADJ
futech-324	269	6	learning	learning	NOUN
futech-324	269	7	models	model	NOUN
futech-324	269	8	,	,	PUNCT
futech-324	269	9	particularly	particularly	ADV
futech-324	269	10	lstm	lstm	NOUN
futech-324	269	11	-	-	PUNCT
futech-324	269	12	based	base	VERB
futech-324	269	13	architectures	architecture	NOUN
futech-324	269	14	,	,	PUNCT
futech-324	269	15	outperformed	outperform	VERB
futech-324	269	16	others	other	NOUN
futech-324	269	17	in	in	ADP
futech-324	269	18	terms	term	NOUN
futech-324	269	19	of	of	ADP
futech-324	269	20	accuracy	accuracy	NOUN
futech-324	269	21	.	.	PUNCT
futech-324	270	1	the	the	DET
futech-324	270	2	highest	high	ADJ
futech-324	270	3	accuracy	accuracy	NOUN
futech-324	270	4	of	of	ADP
futech-324	270	5	92.25	92.25	NUM
futech-324	270	6	%	%	NOUN
futech-324	270	7	was	be	AUX
futech-324	270	8	achieved	achieve	VERB
futech-324	270	9	with	with	ADP
futech-324	270	10	the	the	DET
futech-324	270	11	combination	combination	NOUN
futech-324	270	12	of	of	ADP
futech-324	270	13	lstm	lstm	NOUN
futech-324	270	14	and	and	CCONJ
futech-324	270	15	word2vec	word2vec	X
futech-324	270	16	,	,	PUNCT
futech-324	270	17	closely	closely	ADV
futech-324	270	18	followed	follow	VERB
futech-324	270	19	by	by	ADP
futech-324	270	20	lstm	lstm	NOUN
futech-324	270	21	+	+	CCONJ
futech-324	270	22	bert	bert	NOUN
futech-324	270	23	at	at	ADP
futech-324	270	24	method	method	NOUN
futech-324	270	25	accuracy	accuracy	NOUN
futech-324	270	26	rnn	rnn	NOUN
futech-324	270	27	+	+	NUM
futech-324	270	28	tfidf	tfidf	NOUN
futech-324	270	29	89.46	89.46	NUM
futech-324	270	30	lstm	lstm	NOUN
futech-324	270	31	+	+	CCONJ
futech-324	270	32	tfidf	tfidf	NOUN
futech-324	270	33	90.46	90.46	NUM
futech-324	270	34	rnn	rnn	VERB
futech-324	270	35	+	+	CCONJ
futech-324	270	36	word2vec	word2vec	X
futech-324	270	37	90.76	90.76	NUM
futech-324	270	38	lstm	lstm	NOUN
futech-324	270	39	+	+	CCONJ
futech-324	270	40	word2vec	word2vec	X
futech-324	270	41	92.25	92.25	NUM
futech-324	270	42	rnn	rnn	NOUN
futech-324	270	43	+	+	CCONJ
futech-324	270	44	bert	bert	PROPN
futech-324	270	45	90.76	90.76	NUM
futech-324	270	46	lstm	lstm	NOUN
futech-324	270	47	+	+	CCONJ
futech-324	270	48	bert	bert	PROPN
futech-324	270	49	92.11	92.11	NUM
futech-324	270	50	gru	gru	NOUN
futech-324	270	51	+	+	CCONJ
futech-324	270	52	tfidf	tfidf	NOUN
futech-324	270	53	75.70	75.70	NUM
futech-324	270	54	gru	gru	NOUN
futech-324	270	55	+	+	CCONJ
futech-324	270	56	word2vec	word2vec	X
futech-324	270	57	80.00	80.00	NUM
futech-324	270	58	gru	gru	NOUN
futech-324	270	59	+	+	CCONJ
futech-324	270	60	bert	bert	PROPN
futech-324	270	61	79.00	79.00	NUM
futech-324	270	62	0	0	NUM
futech-324	270	63	20	20	NUM
futech-324	270	64	40	40	NUM
futech-324	270	65	60	60	NUM
futech-324	270	66	80	80	NUM
futech-324	270	67	100	100	NUM
futech-324	270	68	rnn	rnn	NOUN
futech-324	270	69	+	+	NUM
futech-324	270	70	tfidf	tfidf	NOUN
futech-324	270	71	lstm	lstm	NOUN
futech-324	270	72	+	+	CCONJ
futech-324	270	73	tfidf	tfidf	NOUN
futech-324	270	74	rnn	rnn	VERB
futech-324	270	75	+	+	CCONJ
futech-324	270	76	word2vec	word2vec	X
futech-324	270	77	lstm	lstm	NOUN
futech-324	270	78	+	+	CCONJ
futech-324	270	79	word2vec	word2vec	X
futech-324	270	80	rnn	rnn	VERB
futech-324	270	81	+	+	CCONJ
futech-324	270	82	bert	bert	PROPN
futech-324	270	83	lstm	lstm	PROPN
futech-324	270	84	+	+	CCONJ
futech-324	270	85	bert	bert	PROPN
futech-324	270	86	gru	gru	PROPN
futech-324	270	87	+	+	CCONJ
futech-324	270	88	tfidf	tfidf	PROPN
futech-324	270	89	gru	gru	PROPN
futech-324	270	90	+	+	CCONJ
futech-324	271	1	word2vec	word2vec	PROPN
futech-324	271	2	gru	gru	X
futech-324	271	3	+	+	CCONJ
futech-324	271	4	bert	bert	PROPN
futech-324	271	5	accuracy	accuracy	PROPN
futech-324	271	6	comparison	comparison	NOUN
futech-324	271	7	s.	s.	PROPN
futech-324	271	8	sushma	sushma	PROPN
futech-324	271	9	et	et	PROPN
futech-324	271	10	al	al	PROPN
futech-324	271	11	.	.	PROPN
futech-324	271	12	/future	/future	PUNCT
futech-324	271	13	technology	technology	PROPN
futech-324	271	14	august	august	PROPN
futech-324	271	15	2025|	2025|	NUM
futech-324	271	16	volume	volume	NOUN
futech-324	271	17	04	04	NUM
futech-324	272	1	|	|	NOUN
futech-324	272	2	issue	issue	NOUN
futech-324	272	3	03	03	NUM
futech-324	273	1	|	|	ADV
futech-324	273	2	pages	page	VERB
futech-324	273	3	76	76	NUM
futech-324	273	4	-	-	SYM
futech-324	273	5	84	84	NUM
futech-324	273	6	83	83	NUM
futech-324	273	7	92.11	92.11	NUM
futech-324	273	8	%	%	NOUN
futech-324	273	9	.	.	PUNCT
futech-324	274	1	these	these	DET
futech-324	274	2	findings	finding	NOUN
futech-324	274	3	indicate	indicate	VERB
futech-324	274	4	that	that	SCONJ
futech-324	274	5	lstm	lstm	PROPN
futech-324	274	6	networks	network	NOUN
futech-324	274	7	,	,	PUNCT
futech-324	274	8	with	with	ADP
futech-324	274	9	their	their	PRON
futech-324	274	10	ability	ability	NOUN
futech-324	274	11	to	to	PART
futech-324	274	12	capture	capture	VERB
futech-324	274	13	long	long	ADJ
futech-324	274	14	-	-	PUNCT
futech-324	274	15	term	term	NOUN
futech-324	274	16	dependencies	dependency	NOUN
futech-324	274	17	,	,	PUNCT
futech-324	274	18	are	be	AUX
futech-324	274	19	particularly	particularly	ADV
futech-324	274	20	well	well	ADV
futech-324	274	21	-	-	PUNCT
futech-324	274	22	suited	suit	VERB
futech-324	274	23	for	for	ADP
futech-324	274	24	handling	handle	VERB
futech-324	274	25	complex	complex	ADJ
futech-324	274	26	text	text	NOUN
futech-324	274	27	data	datum	NOUN
futech-324	274	28	in	in	ADP
futech-324	274	29	sentiment	sentiment	NOUN
futech-324	274	30	analysis	analysis	NOUN
futech-324	274	31	tasks	task	NOUN
futech-324	274	32	.	.	PUNCT
futech-324	275	1	on	on	ADP
futech-324	275	2	the	the	DET
futech-324	275	3	other	other	ADJ
futech-324	275	4	hand	hand	NOUN
futech-324	275	5	,	,	PUNCT
futech-324	275	6	rnn	rnn	VERB
futech-324	275	7	and	and	CCONJ
futech-324	275	8	gru	gru	NOUN
futech-324	275	9	models	model	NOUN
futech-324	275	10	showed	show	VERB
futech-324	275	11	slightly	slightly	ADV
futech-324	275	12	lower	low	ADJ
futech-324	275	13	performance	performance	NOUN
futech-324	275	14	but	but	CCONJ
futech-324	275	15	still	still	ADV
futech-324	275	16	provided	provide	VERB
futech-324	275	17	competitive	competitive	ADJ
futech-324	275	18	results	result	NOUN
futech-324	275	19	,	,	PUNCT
futech-324	275	20	with	with	ADP
futech-324	275	21	rnn	rnn	NOUN
futech-324	275	22	+	+	CCONJ
futech-324	275	23	word2vec	word2vec	VERB
futech-324	275	24	achieving	achieve	VERB
futech-324	275	25	a	a	DET
futech-324	275	26	90.76	90.76	NUM
futech-324	275	27	%	%	NOUN
futech-324	275	28	accuracy	accuracy	NOUN
futech-324	275	29	.	.	PUNCT
futech-324	276	1	while	while	SCONJ
futech-324	276	2	gru	gru	NOUN
futech-324	276	3	models	model	NOUN
futech-324	276	4	demonstrated	demonstrate	VERB
futech-324	276	5	lower	low	ADJ
futech-324	276	6	accuracy	accuracy	NOUN
futech-324	276	7	compared	compare	VERB
futech-324	276	8	to	to	ADP
futech-324	276	9	lstm	lstm	PROPN
futech-324	276	10	,	,	PUNCT
futech-324	276	11	they	they	PRON
futech-324	276	12	still	still	ADV
futech-324	276	13	hold	hold	VERB
futech-324	276	14	potential	potential	NOUN
futech-324	276	15	for	for	ADP
futech-324	276	16	future	future	ADJ
futech-324	276	17	optimization	optimization	NOUN
futech-324	276	18	and	and	CCONJ
futech-324	276	19	experimentation	experimentation	NOUN
futech-324	276	20	.	.	PUNCT
futech-324	277	1	future	future	ADJ
futech-324	277	2	work	work	NOUN
futech-324	277	3	could	could	AUX
futech-324	277	4	focus	focus	VERB
futech-324	277	5	on	on	ADP
futech-324	277	6	further	further	ADJ
futech-324	277	7	refining	refine	VERB
futech-324	277	8	these	these	DET
futech-324	277	9	models	model	NOUN
futech-324	277	10	,	,	PUNCT
futech-324	277	11	experimenting	experiment	VERB
futech-324	277	12	with	with	ADP
futech-324	277	13	larger	large	ADJ
futech-324	277	14	datasets	dataset	NOUN
futech-324	277	15	,	,	PUNCT
futech-324	277	16	and	and	CCONJ
futech-324	277	17	exploring	explore	VERB
futech-324	277	18	more	more	ADV
futech-324	277	19	advanced	advanced	ADJ
futech-324	277	20	techniques	technique	NOUN
futech-324	277	21	,	,	PUNCT
futech-324	277	22	such	such	ADJ
futech-324	277	23	as	as	ADP
futech-324	277	24	fine	fine	ADV
futech-324	277	25	-	-	PUNCT
futech-324	277	26	tuning	tune	VERB
futech-324	277	27	pre	pre	ADJ
futech-324	277	28	-	-	ADJ
futech-324	277	29	trained	train	VERB
futech-324	277	30	models	model	NOUN
futech-324	277	31	or	or	CCONJ
futech-324	277	32	combining	combine	VERB
futech-324	277	33	multiple	multiple	ADJ
futech-324	277	34	models	model	NOUN
futech-324	277	35	,	,	PUNCT
futech-324	277	36	to	to	PART
futech-324	277	37	improve	improve	VERB
futech-324	277	38	classification	classification	NOUN
futech-324	277	39	accuracy	accuracy	NOUN
futech-324	277	40	.	.	PUNCT
futech-324	278	1	this	this	DET
futech-324	278	2	paper	paper	NOUN
futech-324	278	3	demonstrates	demonstrate	VERB
futech-324	278	4	the	the	DET
futech-324	278	5	potential	potential	NOUN
futech-324	278	6	of	of	ADP
futech-324	278	7	using	use	VERB
futech-324	278	8	advanced	advanced	ADJ
futech-324	278	9	dl	dl	NOUN
futech-324	278	10	techniques	technique	NOUN
futech-324	278	11	and	and	CCONJ
futech-324	278	12	pre	pre	ADJ
futech-324	278	13	-	-	ADJ
futech-324	278	14	trained	train	VERB
futech-324	278	15	models	model	NOUN
futech-324	278	16	for	for	ADP
futech-324	278	17	effectively	effectively	ADV
futech-324	278	18	addressing	address	VERB
futech-324	278	19	the	the	DET
futech-324	278	20	challenge	challenge	NOUN
futech-324	278	21	of	of	ADP
futech-324	278	22	toxic	toxic	ADJ
futech-324	278	23	comment	comment	NOUN
futech-324	278	24	classification	classification	NOUN
futech-324	278	25	.	.	PUNCT
futech-324	279	1	ethical	ethical	ADJ
futech-324	279	2	issue	issue	NOUN
futech-324	279	3	the	the	DET
futech-324	279	4	authors	author	NOUN
futech-324	279	5	are	be	AUX
futech-324	279	6	aware	aware	ADJ
futech-324	279	7	of	of	ADP
futech-324	279	8	and	and	CCONJ
futech-324	279	9	comply	comply	ADJ
futech-324	279	10	with	with	ADP
futech-324	279	11	best	good	ADJ
futech-324	279	12	practices	practice	NOUN
futech-324	279	13	in	in	ADP
futech-324	279	14	publication	publication	NOUN
futech-324	279	15	ethics	ethic	NOUN
futech-324	279	16	,	,	PUNCT
futech-324	279	17	specifically	specifically	ADV
futech-324	279	18	with	with	ADP
futech-324	279	19	regard	regard	NOUN
futech-324	279	20	to	to	ADP
futech-324	279	21	authorship	authorship	NOUN
futech-324	279	22	(	(	PUNCT
futech-324	279	23	avoidance	avoidance	NOUN
futech-324	279	24	of	of	ADP
futech-324	279	25	guest	guest	NOUN
futech-324	279	26	authorship	authorship	NOUN
futech-324	279	27	)	)	PUNCT
futech-324	279	28	,	,	PUNCT
futech-324	279	29	dual	dual	ADJ
futech-324	279	30	submission	submission	NOUN
futech-324	279	31	,	,	PUNCT
futech-324	279	32	manipulation	manipulation	NOUN
futech-324	279	33	of	of	ADP
futech-324	279	34	figures	figure	NOUN
futech-324	279	35	,	,	PUNCT
futech-324	279	36	competing	compete	VERB
futech-324	279	37	interests	interest	NOUN
futech-324	279	38	,	,	PUNCT
futech-324	279	39	and	and	CCONJ
futech-324	279	40	compliance	compliance	NOUN
futech-324	279	41	with	with	ADP
futech-324	279	42	policies	policy	NOUN
futech-324	279	43	on	on	ADP
futech-324	279	44	research	research	NOUN
futech-324	279	45	ethics	ethic	NOUN
futech-324	279	46	.	.	PUNCT
futech-324	280	1	the	the	DET
futech-324	280	2	author	author	NOUN
futech-324	280	3	adheres	adhere	VERB
futech-324	280	4	to	to	ADP
futech-324	280	5	publication	publication	NOUN
futech-324	280	6	requirements	requirement	NOUN
futech-324	280	7	that	that	SCONJ
futech-324	280	8	the	the	DET
futech-324	280	9	submitted	submit	VERB
futech-324	280	10	work	work	NOUN
futech-324	280	11	is	be	AUX
futech-324	280	12	original	original	ADJ
futech-324	280	13	and	and	CCONJ
futech-324	280	14	has	have	AUX
futech-324	280	15	not	not	PART
futech-324	280	16	been	be	AUX
futech-324	280	17	published	publish	VERB
futech-324	280	18	elsewhere	elsewhere	ADV
futech-324	280	19	.	.	PUNCT
futech-324	281	1	data	datum	NOUN
futech-324	281	2	availability	availability	NOUN
futech-324	281	3	statement	statement	NOUN
futech-324	281	4	the	the	DET
futech-324	281	5	manuscript	manuscript	NOUN
futech-324	281	6	contains	contain	VERB
futech-324	281	7	all	all	DET
futech-324	281	8	the	the	DET
futech-324	281	9	data	datum	NOUN
futech-324	281	10	.	.	PUNCT
futech-324	282	1	however	however	ADV
futech-324	282	2	,	,	PUNCT
futech-324	282	3	more	more	ADJ
futech-324	282	4	data	datum	NOUN
futech-324	282	5	will	will	AUX
futech-324	282	6	be	be	AUX
futech-324	282	7	available	available	ADJ
futech-324	282	8	upon	upon	SCONJ
futech-324	282	9	request	request	NOUN
futech-324	282	10	from	from	ADP
futech-324	282	11	the	the	DET
futech-324	282	12	authors	author	NOUN
futech-324	282	13	.	.	PUNCT
futech-324	283	1	conflict	conflict	NOUN
futech-324	283	2	of	of	ADP
futech-324	283	3	interest	interest	NOUN
futech-324	283	4	the	the	DET
futech-324	283	5	authors	author	NOUN
futech-324	283	6	declare	declare	VERB
futech-324	283	7	no	no	DET
futech-324	283	8	potential	potential	ADJ
futech-324	283	9	conflict	conflict	NOUN
futech-324	283	10	of	of	ADP
futech-324	283	11	interest	interest	NOUN
futech-324	283	12	.	.	PUNCT
futech-324	284	1	references	reference	NOUN
futech-324	284	2	[	[	X
futech-324	284	3	1	1	NUM
futech-324	284	4	]	]	PUNCT
futech-324	284	5	a.	a.	NOUN
futech-324	284	6	albladi	albladi	NOUN
futech-324	284	7	,	,	PUNCT
futech-324	284	8	m.	m.	NOUN
futech-324	284	9	islam	islam	PROPN
futech-324	284	10	and	and	CCONJ
futech-324	284	11	c.	c.	PROPN
futech-324	284	12	seals	seal	NOUN
futech-324	284	13	,	,	PUNCT
futech-324	284	14	"	"	PUNCT
futech-324	284	15	sentiment	sentiment	NOUN
futech-324	284	16	analysis	analysis	NOUN
futech-324	284	17	of	of	ADP
futech-324	284	18	twitter	twitter	NOUN
futech-324	284	19	data	datum	NOUN
futech-324	284	20	using	use	VERB
futech-324	284	21	nlp	nlp	NOUN
futech-324	284	22	models	model	NOUN
futech-324	284	23	:	:	PUNCT
futech-324	284	24	a	a	DET
futech-324	284	25	comprehensive	comprehensive	ADJ
futech-324	284	26	review	review	NOUN
futech-324	284	27	,	,	PUNCT
futech-324	284	28	"	"	PUNCT
futech-324	284	29	in	in	ADP
futech-324	284	30	ieee	ieee	NOUN
futech-324	284	31	access	access	NOUN
futech-324	284	32	,	,	PUNCT
futech-324	284	33	vol	vol	NOUN
futech-324	284	34	.	.	PROPN
futech-324	284	35	13	13	NUM
futech-324	284	36	,	,	PUNCT
futech-324	284	37	pp	pp	ADJ
futech-324	284	38	.	.	PUNCT
futech-324	284	39	30444	30444	NUM
futech-324	284	40	-	-	SYM
futech-324	284	41	30468	30468	NUM
futech-324	284	42	,	,	PUNCT
futech-324	284	43	2025	2025	NUM
futech-324	284	44	,	,	PUNCT
futech-324	284	45	doi	doi	NOUN
futech-324	284	46	:	:	PUNCT
futech-324	284	47	10.1109	10.1109	NUM
futech-324	284	48	/	/	SYM
futech-324	284	49	access.2025.3541494	access.2025.3541494	PROPN
futech-324	284	50	.	.	PUNCT
futech-324	285	1	[	[	X
futech-324	285	2	2	2	NUM
futech-324	285	3	]	]	PUNCT
futech-324	285	4	z.	z.	PROPN
futech-324	285	5	hao	hao	PROPN
futech-324	285	6	et	et	PROPN
futech-324	285	7	al	al	PROPN
futech-324	285	8	.	.	PROPN
futech-324	285	9	,	,	PUNCT
futech-324	285	10	"	"	PUNCT
futech-324	285	11	a	a	DET
futech-324	285	12	novel	novel	ADJ
futech-324	285	13	public	public	ADJ
futech-324	285	14	sentiment	sentiment	NOUN
futech-324	285	15	analysis	analysis	NOUN
futech-324	285	16	method	method	NOUN
futech-324	285	17	based	base	VERB
futech-324	285	18	on	on	ADP
futech-324	285	19	an	an	DET
futech-324	285	20	isomerism	isomerism	NOUN
futech-324	285	21	learning	learn	VERB
futech-324	285	22	model	model	NOUN
futech-324	285	23	via	via	ADP
futech-324	285	24	multiphase	multiphase	NOUN
futech-324	285	25	processing	processing	NOUN
futech-324	285	26	,	,	PUNCT
futech-324	285	27	"	"	PUNCT
futech-324	285	28	in	in	ADP
futech-324	285	29	ieee	ieee	NOUN
futech-324	285	30	transactions	transaction	NOUN
futech-324	285	31	on	on	ADP
futech-324	285	32	neural	neural	ADJ
futech-324	285	33	networks	network	NOUN
futech-324	285	34	and	and	CCONJ
futech-324	285	35	learning	learning	NOUN
futech-324	285	36	systems	system	NOUN
futech-324	285	37	,	,	PUNCT
futech-324	285	38	vol	vol	NOUN
futech-324	285	39	.	.	PROPN
futech-324	286	1	36	36	NUM
futech-324	286	2	,	,	PUNCT
futech-324	286	3	no	no	INTJ
futech-324	286	4	.	.	NOUN
futech-324	286	5	1	1	NUM
futech-324	286	6	,	,	PUNCT
futech-324	286	7	pp	pp	ADJ
futech-324	286	8	.	.	PUNCT
futech-324	287	1	249	249	NUM
futech-324	287	2	-	-	SYM
futech-324	287	3	259	259	NUM
futech-324	287	4	,	,	PUNCT
futech-324	287	5	jan	jan	PROPN
futech-324	287	6	.	.	PROPN
futech-324	287	7	2025	2025	NUM
futech-324	287	8	,	,	PUNCT
futech-324	287	9	doi	doi	NOUN
futech-324	287	10	:	:	PUNCT
futech-324	287	11	10.1109	10.1109	NUM
futech-324	287	12	/	/	SYM
futech-324	287	13	tnnls.2023.3274912	tnnls.2023.3274912	NOUN
futech-324	287	14	.	.	PUNCT
futech-324	288	1	[	[	X
futech-324	288	2	3	3	NUM
futech-324	288	3	]	]	PUNCT
futech-324	288	4	x.	x.	NOUN
futech-324	288	5	wang	wang	PROPN
futech-324	288	6	,	,	PUNCT
futech-324	288	7	j.	j.	PROPN
futech-324	288	8	lyu	lyu	PROPN
futech-324	288	9	,	,	PUNCT
futech-324	288	10	b.	b.	PROPN
futech-324	288	11	-g	-g	PROPN
futech-324	288	12	.	.	PUNCT
futech-324	289	1	kim	kim	PROPN
futech-324	289	2	,	,	PUNCT
futech-324	289	3	b.	b.	PROPN
futech-324	289	4	d.	d.	PROPN
futech-324	289	5	parameshachari	parameshachari	PROPN
futech-324	289	6	,	,	PUNCT
futech-324	289	7	k.	k.	PROPN
futech-324	289	8	li	li	PROPN
futech-324	289	9	and	and	CCONJ
futech-324	289	10	q.	q.	PROPN
futech-324	289	11	li	li	PROPN
futech-324	289	12	,	,	PUNCT
futech-324	289	13	"	"	PUNCT
futech-324	289	14	exploring	explore	VERB
futech-324	289	15	multimodal	multimodal	ADJ
futech-324	289	16	multiscale	multiscale	NOUN
futech-324	289	17	features	feature	NOUN
futech-324	289	18	for	for	ADP
futech-324	289	19	sentiment	sentiment	NOUN
futech-324	289	20	analysis	analysis	NOUN
futech-324	289	21	using	use	VERB
futech-324	289	22	fuzzy	fuzzy	ADJ
futech-324	289	23	-	-	PUNCT
futech-324	289	24	deep	deep	ADJ
futech-324	289	25	neural	neural	ADJ
futech-324	289	26	network	network	NOUN
futech-324	289	27	learning	learning	NOUN
futech-324	289	28	,	,	PUNCT
futech-324	289	29	"	"	PUNCT
futech-324	289	30	in	in	ADP
futech-324	289	31	ieee	ieee	NOUN
futech-324	289	32	transactions	transaction	NOUN
futech-324	289	33	on	on	ADP
futech-324	289	34	fuzzy	fuzzy	ADJ
futech-324	289	35	systems	system	NOUN
futech-324	289	36	,	,	PUNCT
futech-324	289	37	vol	vol	NOUN
futech-324	289	38	.	.	PROPN
futech-324	289	39	33	33	NUM
futech-324	289	40	,	,	PUNCT
futech-324	289	41	no	no	INTJ
futech-324	289	42	.	.	NOUN
futech-324	289	43	1	1	NUM
futech-324	289	44	,	,	PUNCT
futech-324	289	45	pp	pp	ADJ
futech-324	289	46	.	.	PUNCT
futech-324	290	1	28	28	NUM
futech-324	290	2	-	-	SYM
futech-324	290	3	42	42	NUM
futech-324	290	4	,	,	PUNCT
futech-324	290	5	jan	jan	PROPN
futech-324	290	6	.	.	PROPN
futech-324	290	7	2025	2025	NUM
futech-324	290	8	,	,	PUNCT
futech-324	290	9	doi	doi	NOUN
futech-324	290	10	:	:	PUNCT
futech-324	290	11	10.1109	10.1109	NUM
futech-324	290	12	/	/	SYM
futech-324	290	13	tfuzz.2024.3419140	tfuzz.2024.3419140	ADJ
futech-324	290	14	.	.	PUNCT
futech-324	291	1	[	[	X
futech-324	291	2	4	4	X
futech-324	291	3	]	]	X
futech-324	291	4	h.	h.	PROPN
futech-324	291	5	t.	t.	PROPN
futech-324	291	6	phan	phan	PROPN
futech-324	291	7	,	,	PUNCT
futech-324	291	8	v.	v.	PROPN
futech-324	291	9	d.	d.	PROPN
futech-324	291	10	nguyen	nguyen	PROPN
futech-324	291	11	and	and	CCONJ
futech-324	291	12	n.	n.	PROPN
futech-324	291	13	t.	t.	PROPN
futech-324	291	14	nguyen	nguyen	NOUN
futech-324	291	15	,	,	PUNCT
futech-324	291	16	"	"	PUNCT
futech-324	291	17	mulgcn	mulgcn	PROPN
futech-324	291	18	:	:	PUNCT
futech-324	291	19	multigraph	multigraph	VERB
futech-324	291	20	convolutional	convolutional	ADJ
futech-324	291	21	network	network	NOUN
futech-324	291	22	for	for	ADP
futech-324	291	23	aspect	aspect	NOUN
futech-324	291	24	-	-	PUNCT
futech-324	291	25	level	level	NOUN
futech-324	291	26	sentiment	sentiment	NOUN
futech-324	291	27	analysis	analysis	NOUN
futech-324	291	28	,	,	PUNCT
futech-324	291	29	"	"	PUNCT
futech-324	291	30	in	in	ADP
futech-324	291	31	ieee	ieee	NOUN
futech-324	291	32	access	access	NOUN
futech-324	291	33	,	,	PUNCT
futech-324	291	34	vol	vol	NOUN
futech-324	291	35	.	.	PROPN
futech-324	291	36	13	13	NUM
futech-324	291	37	,	,	PUNCT
futech-324	291	38	pp	pp	ADJ
futech-324	291	39	.	.	PUNCT
futech-324	291	40	26304	26304	NUM
futech-324	291	41	-	-	SYM
futech-324	291	42	26317	26317	NUM
futech-324	291	43	,	,	PUNCT
futech-324	291	44	2025	2025	NUM
futech-324	291	45	,	,	PUNCT
futech-324	291	46	doi	doi	NOUN
futech-324	291	47	:	:	PUNCT
futech-324	291	48	10.1109	10.1109	NUM
futech-324	291	49	/	/	SYM
futech-324	291	50	access.2025.3537340	access.2025.3537340	NOUN
futech-324	291	51	.	.	PUNCT
futech-324	292	1	[	[	X
futech-324	292	2	5	5	NUM
futech-324	292	3	]	]	PUNCT
futech-324	292	4	s.	s.	PROPN
futech-324	292	5	ali	ali	PROPN
futech-324	292	6	,	,	PUNCT
futech-324	292	7	u.	u.	PROPN
futech-324	292	8	jamil	jamil	PROPN
futech-324	292	9	,	,	PUNCT
futech-324	292	10	m.	m.	NOUN
futech-324	292	11	younas	younas	PROPN
futech-324	292	12	,	,	PUNCT
futech-324	292	13	b.	b.	PROPN
futech-324	292	14	zafar	zafar	PROPN
futech-324	292	15	and	and	CCONJ
futech-324	292	16	m.	m.	PROPN
futech-324	292	17	kashif	kashif	PROPN
futech-324	292	18	hanif	hanif	PROPN
futech-324	292	19	,	,	PUNCT
futech-324	292	20	"	"	PUNCT
futech-324	292	21	optimized	optimize	VERB
futech-324	292	22	identification	identification	NOUN
futech-324	292	23	of	of	ADP
futech-324	292	24	sentence	sentence	NOUN
futech-324	292	25	-	-	PUNCT
futech-324	292	26	level	level	NOUN
futech-324	292	27	multiclass	multiclass	ADJ
futech-324	292	28	events	event	NOUN
futech-324	292	29	on	on	ADP
futech-324	292	30	urdu	urdu	NOUN
futech-324	292	31	-	-	PUNCT
futech-324	292	32	language	language	NOUN
futech-324	292	33	-	-	PUNCT
futech-324	292	34	text	text	NOUN
futech-324	292	35	using	use	VERB
futech-324	292	36	machine	machine	NOUN
futech-324	292	37	learning	learning	NOUN
futech-324	292	38	techniques	technique	NOUN
futech-324	292	39	,	,	PUNCT
futech-324	292	40	"	"	PUNCT
futech-324	292	41	in	in	ADP
futech-324	292	42	ieee	ieee	NOUN
futech-324	292	43	access	access	NOUN
futech-324	292	44	,	,	PUNCT
futech-324	292	45	vol	vol	NOUN
futech-324	292	46	.	.	PROPN
futech-324	292	47	13	13	NUM
futech-324	292	48	,	,	PUNCT
futech-324	292	49	pp	pp	ADJ
futech-324	292	50	.	.	PUNCT
futech-324	293	1	1	1	NUM
futech-324	293	2	-	-	SYM
futech-324	293	3	25	25	NUM
futech-324	293	4	,	,	PUNCT
futech-324	293	5	2025	2025	NUM
futech-324	293	6	,	,	PUNCT
futech-324	293	7	doi	doi	NOUN
futech-324	293	8	:	:	PUNCT
futech-324	293	9	10.1109	10.1109	NUM
futech-324	293	10	/	/	SYM
futech-324	293	11	access.2024.3522992	access.2024.3522992	ADJ
futech-324	293	12	.	.	PUNCT
futech-324	294	1	[	[	X
futech-324	294	2	6	6	NUM
futech-324	294	3	]	]	PUNCT
futech-324	294	4	w.	w.	PROPN
futech-324	294	5	gong	gong	PROPN
futech-324	294	6	,	,	PUNCT
futech-324	294	7	"	"	PUNCT
futech-324	294	8	text	text	NOUN
futech-324	294	9	sentiment	sentiment	NOUN
futech-324	294	10	classification	classification	NOUN
futech-324	294	11	algorithm	algorithm	NOUN
futech-324	294	12	based	base	VERB
futech-324	294	13	on	on	ADP
futech-324	294	14	bilstm	bilstm	NOUN
futech-324	294	15	deep	deep	ADJ
futech-324	294	16	learning	learning	NOUN
futech-324	294	17	,	,	PUNCT
futech-324	294	18	"	"	PUNCT
futech-324	294	19	2024	2024	NUM
futech-324	294	20	international	international	ADJ
futech-324	294	21	conference	conference	NOUN
futech-324	294	22	on	on	ADP
futech-324	294	23	industrial	industrial	ADJ
futech-324	294	24	iot	iot	NOUN
futech-324	294	25	,	,	PUNCT
futech-324	294	26	big	big	ADJ
futech-324	294	27	data	datum	NOUN
futech-324	294	28	and	and	CCONJ
futech-324	294	29	supply	supply	NOUN
futech-324	294	30	chain	chain	NOUN
futech-324	294	31	(	(	PUNCT
futech-324	294	32	iiotbdsc	iiotbdsc	NOUN
futech-324	294	33	)	)	PUNCT
futech-324	294	34	,	,	PUNCT
futech-324	294	35	wuhan	wuhan	PROPN
futech-324	294	36	,	,	PUNCT
futech-324	294	37	china	china	PROPN
futech-324	294	38	,	,	PUNCT
futech-324	294	39	2024	2024	NUM
futech-324	294	40	,	,	PUNCT
futech-324	294	41	pp	pp	ADV
futech-324	294	42	.	.	PUNCT
futech-324	295	1	83	83	NUM
futech-324	295	2	-	-	SYM
futech-324	295	3	87	87	NUM
futech-324	295	4	,	,	PUNCT
futech-324	295	5	doi	doi	NOUN
futech-324	295	6	:	:	PUNCT
futech-324	295	7	10.1109	10.1109	NUM
futech-324	295	8	/	/	SYM
futech-324	295	9	iiotbdsc64371.2024.00025	iiotbdsc64371.2024.00025	NOUN
futech-324	295	10	.	.	PUNCT
futech-324	296	1	[	[	X
futech-324	296	2	7	7	X
futech-324	296	3	]	]	PUNCT
futech-324	296	4	s.	s.	PROPN
futech-324	296	5	a.	a.	PROPN
futech-324	296	6	mostafa	mostafa	PROPN
futech-324	296	7	,	,	PUNCT
futech-324	296	8	w.	w.	PROPN
futech-324	296	9	s.	s.	PROPN
futech-324	296	10	al	al	PROPN
futech-324	296	11	-	-	PUNCT
futech-324	296	12	dayyeni	dayyeni	ADJ
futech-324	296	13	,	,	PUNCT
futech-324	296	14	a.	a.	PROPN
futech-324	296	15	n.	n.	PROPN
futech-324	296	16	kareem	kareem	PROPN
futech-324	296	17	,	,	PUNCT
futech-324	296	18	m.	m.	NOUN
futech-324	296	19	a.	a.	NOUN
futech-324	296	20	jubair	jubair	PROPN
futech-324	296	21	,	,	PUNCT
futech-324	296	22	m.	m.	NOUN
futech-324	296	23	m.	m.	PROPN
futech-324	296	24	jaber	jaber	PROPN
futech-324	296	25	and	and	CCONJ
futech-324	296	26	b.	b.	PROPN
futech-324	296	27	a.	a.	PROPN
futech-324	296	28	khalaf	khalaf	PROPN
futech-324	296	29	,	,	PUNCT
futech-324	296	30	"	"	PUNCT
futech-324	296	31	classification	classification	NOUN
futech-324	296	32	and	and	CCONJ
futech-324	296	33	sentiment	sentiment	NOUN
futech-324	296	34	analysis	analysis	NOUN
futech-324	296	35	of	of	ADP
futech-324	296	36	amazon	amazon	PROPN
futech-324	296	37	alexa	alexa	NOUN
futech-324	296	38	reviews	review	NOUN
futech-324	296	39	,	,	PUNCT
futech-324	296	40	"	"	PUNCT
futech-324	296	41	2024	2024	NUM
futech-324	296	42	1st	1st	ADJ
futech-324	296	43	international	international	ADJ
futech-324	296	44	conference	conference	NOUN
futech-324	296	45	on	on	ADP
futech-324	296	46	logistics	logistic	NOUN
futech-324	296	47	(	(	PUNCT
futech-324	296	48	icl	icl	PROPN
futech-324	296	49	)	)	PUNCT
futech-324	296	50	,	,	PUNCT
futech-324	296	51	jeddah	jeddah	PROPN
futech-324	296	52	,	,	PUNCT
futech-324	296	53	saudi	saudi	PROPN
futech-324	296	54	arabia	arabia	PROPN
futech-324	296	55	,	,	PUNCT
futech-324	296	56	2024	2024	NUM
futech-324	296	57	,	,	PUNCT
futech-324	296	58	pp	pp	ADV
futech-324	296	59	.	.	PUNCT
futech-324	297	1	1	1	NUM
futech-324	297	2	-	-	SYM
futech-324	297	3	5	5	NUM
futech-324	297	4	,	,	PUNCT
futech-324	297	5	doi	doi	NOUN
futech-324	297	6	:	:	PUNCT
futech-324	297	7	10.1109	10.1109	NUM
futech-324	297	8	/	/	SYM
futech-324	297	9	icl62932.2024.10788570	icl62932.2024.10788570	NOUN
futech-324	297	10	.	.	PUNCT
futech-324	298	1	[	[	X
futech-324	298	2	8	8	NUM
futech-324	298	3	]	]	PUNCT
futech-324	298	4	s.	s.	PROPN
futech-324	298	5	mehta	mehta	PROPN
futech-324	298	6	and	and	CCONJ
futech-324	298	7	a.	a.	NOUN
futech-324	298	8	bhalla	bhalla	PROPN
futech-324	298	9	,	,	PUNCT
futech-324	298	10	"	"	PUNCT
futech-324	298	11	enhanced	enhance	VERB
futech-324	298	12	sentiment	sentiment	NOUN
futech-324	298	13	classification	classification	NOUN
futech-324	298	14	with	with	ADP
futech-324	298	15	federated	federated	ADJ
futech-324	298	16	learning	learning	NOUN
futech-324	298	17	cnns	cnn	NOUN
futech-324	298	18	:	:	PUNCT
futech-324	298	19	exploring	explore	VERB
futech-324	298	20	five	five	NUM
futech-324	298	21	sentiment	sentiment	NOUN
futech-324	298	22	categories	category	NOUN
futech-324	298	23	,	,	PUNCT
futech-324	298	24	"	"	PUNCT
futech-324	298	25	2024	2024	NUM
futech-324	298	26	3rd	3rd	ADJ
futech-324	298	27	international	international	ADJ
futech-324	298	28	conference	conference	NOUN
futech-324	298	29	for	for	ADP
futech-324	298	30	advancement	advancement	NOUN
futech-324	298	31	in	in	ADP
futech-324	298	32	technology	technology	NOUN
futech-324	298	33	(	(	PUNCT
futech-324	298	34	iconat	iconat	NOUN
futech-324	298	35	)	)	PUNCT
futech-324	298	36	,	,	PUNCT
futech-324	298	37	goa	goa	PROPN
futech-324	298	38	,	,	PUNCT
futech-324	298	39	india	india	PROPN
futech-324	298	40	,	,	PUNCT
futech-324	298	41	2024	2024	NUM
futech-324	298	42	,	,	PUNCT
futech-324	298	43	pp	pp	ADJ
futech-324	298	44	.	.	PUNCT
futech-324	298	45	1	1	NUM
futech-324	298	46	-	-	SYM
futech-324	298	47	5	5	NUM
futech-324	298	48	,	,	PUNCT
futech-324	298	49	doi	doi	NOUN
futech-324	298	50	:	:	PUNCT
futech-324	298	51	10.1109	10.1109	NUM
futech-324	298	52	/	/	SYM
futech-324	298	53	iconat61936.2024.10774751	iconat61936.2024.10774751	PROPN
futech-324	298	54	.	.	PUNCT
futech-324	299	1	[	[	X
futech-324	299	2	9	9	NUM
futech-324	299	3	]	]	PUNCT
futech-324	299	4	x.	x.	NOUN
futech-324	299	5	he	he	PRON
futech-324	299	6	,	,	PUNCT
futech-324	299	7	"	"	PUNCT
futech-324	299	8	sentiment	sentiment	NOUN
futech-324	299	9	classification	classification	NOUN
futech-324	299	10	of	of	ADP
futech-324	299	11	social	social	ADJ
futech-324	299	12	media	medium	NOUN
futech-324	299	13	user	user	NOUN
futech-324	299	14	comments	comment	NOUN
futech-324	299	15	using	use	VERB
futech-324	299	16	svm	svm	ADJ
futech-324	299	17	models	model	NOUN
futech-324	299	18	,	,	PUNCT
futech-324	299	19	"	"	PUNCT
futech-324	299	20	2024	2024	NUM
futech-324	299	21	5th	5th	ADJ
futech-324	299	22	international	international	ADJ
futech-324	299	23	seminar	seminar	NOUN
futech-324	299	24	on	on	ADP
futech-324	299	25	artificial	artificial	ADJ
futech-324	299	26	intelligence	intelligence	NOUN
futech-324	299	27	,	,	PUNCT
futech-324	299	28	networking	networking	NOUN
futech-324	299	29	and	and	CCONJ
futech-324	299	30	information	information	NOUN
futech-324	299	31	technology	technology	NOUN
futech-324	299	32	(	(	PUNCT
futech-324	299	33	ainit	ainit	PROPN
futech-324	299	34	)	)	PUNCT
futech-324	299	35	,	,	PUNCT
futech-324	299	36	nanjing	nanjing	PROPN
futech-324	299	37	,	,	PUNCT
futech-324	299	38	china	china	PROPN
futech-324	299	39	,	,	PUNCT
futech-324	299	40	2024	2024	NUM
futech-324	299	41	,	,	PUNCT
futech-324	299	42	pp	pp	ADJ
futech-324	299	43	.	.	PUNCT
futech-324	300	1	1755	1755	NUM
futech-324	300	2	-	-	SYM
futech-324	300	3	1759	1759	NUM
futech-324	300	4	,	,	PUNCT
futech-324	300	5	doi	doi	NOUN
futech-324	300	6	:	:	PUNCT
futech-324	300	7	10.1109	10.1109	NUM
futech-324	300	8	/	/	SYM
futech-324	300	9	ainit61980.2024.10581547	ainit61980.2024.10581547	NOUN
futech-324	300	10	.	.	PUNCT
futech-324	301	1	[	[	X
futech-324	301	2	10	10	NUM
futech-324	301	3	]	]	PUNCT
futech-324	301	4	m.	m.	NOUN
futech-324	301	5	aamir	aamir	PROPN
futech-324	301	6	,	,	PUNCT
futech-324	301	7	l.	l.	PROPN
futech-324	301	8	j	j	PROPN
futech-324	301	9	and	and	CCONJ
futech-324	301	10	sweety	sweety	NOUN
futech-324	301	11	,	,	PUNCT
futech-324	301	12	"	"	PUNCT
futech-324	301	13	a	a	DET
futech-324	301	14	comparative	comparative	ADJ
futech-324	301	15	study	study	NOUN
futech-324	301	16	of	of	ADP
futech-324	301	17	ml	ml	NOUN
futech-324	301	18	and	and	CCONJ
futech-324	301	19	dl	dl	PROPN
futech-324	301	20	approaches	approach	NOUN
futech-324	301	21	for	for	ADP
futech-324	301	22	twitter	twitter	NOUN
futech-324	301	23	sentiment	sentiment	NOUN
futech-324	301	24	classification	classification	NOUN
futech-324	301	25	,	,	PUNCT
futech-324	301	26	"	"	PUNCT
futech-324	301	27	2024	2024	NUM
futech-324	301	28	international	international	ADJ
futech-324	301	29	conference	conference	NOUN
futech-324	301	30	on	on	ADP
futech-324	301	31	electrical	electrical	ADJ
futech-324	301	32	electronics	electronic	NOUN
futech-324	301	33	and	and	CCONJ
futech-324	301	34	computing	computing	NOUN
futech-324	301	35	technologies	technology	NOUN
futech-324	301	36	(	(	PUNCT
futech-324	301	37	iceect	iceect	NOUN
futech-324	301	38	)	)	PUNCT
futech-324	301	39	,	,	PUNCT
futech-324	301	40	greater	great	ADJ
futech-324	301	41	noida	noida	PROPN
futech-324	301	42	,	,	PUNCT
futech-324	301	43	india	india	PROPN
futech-324	301	44	,	,	PUNCT
futech-324	301	45	2024	2024	NUM
futech-324	301	46	,	,	PUNCT
futech-324	301	47	pp	pp	ADJ
futech-324	301	48	.	.	PUNCT
futech-324	301	49	1	1	NUM
futech-324	301	50	-	-	SYM
futech-324	301	51	5	5	NUM
futech-324	301	52	,	,	PUNCT
futech-324	301	53	doi	doi	NOUN
futech-324	301	54	:	:	PUNCT
futech-324	301	55	10.1109	10.1109	NUM
futech-324	301	56	/	/	SYM
futech-324	301	57	iceect61758.2024.10738938	iceect61758.2024.10738938	PROPN
futech-324	301	58	.	.	PUNCT
futech-324	302	1	[	[	X
futech-324	302	2	11	11	NUM
futech-324	302	3	]	]	PUNCT
futech-324	302	4	q.	q.	PROPN
futech-324	302	5	zeng	zeng	PROPN
futech-324	302	6	,	,	PUNCT
futech-324	302	7	"	"	PUNCT
futech-324	302	8	design	design	NOUN
futech-324	302	9	of	of	ADP
futech-324	302	10	intelligent	intelligent	ADJ
futech-324	302	11	sentiment	sentiment	NOUN
futech-324	302	12	classification	classification	NOUN
futech-324	302	13	model	model	NOUN
futech-324	302	14	based	base	VERB
futech-324	302	15	on	on	ADP
futech-324	302	16	deep	deep	ADJ
futech-324	302	17	neural	neural	ADJ
futech-324	302	18	network	network	NOUN
futech-324	302	19	algorithm	algorithm	NOUN
futech-324	302	20	in	in	ADP
futech-324	302	21	social	social	ADJ
futech-324	302	22	media	medium	NOUN
futech-324	302	23	,	,	PUNCT
futech-324	302	24	"	"	PUNCT
futech-324	302	25	in	in	ADP
futech-324	302	26	ieee	ieee	NOUN
futech-324	302	27	access	access	NOUN
futech-324	302	28	,	,	PUNCT
futech-324	302	29	vol	vol	NOUN
futech-324	302	30	.	.	PROPN
futech-324	302	31	12	12	NUM
futech-324	302	32	,	,	PUNCT
futech-324	302	33	pp	pp	ADJ
futech-324	302	34	.	.	PUNCT
futech-324	302	35	8104781056	8104781056	NUM
futech-324	302	36	,	,	PUNCT
futech-324	302	37	2024	2024	NUM
futech-324	302	38	,	,	PUNCT
futech-324	302	39	doi	doi	NOUN
futech-324	302	40	:	:	PUNCT
futech-324	302	41	10.1109	10.1109	NUM
futech-324	302	42	/	/	SYM
futech-324	302	43	access.2024.3409818	access.2024.3409818	NUM
futech-324	302	44	.	.	PUNCT
futech-324	303	1	[	[	X
futech-324	303	2	12	12	NUM
futech-324	303	3	]	]	PUNCT
futech-324	303	4	m.	m.	NOUN
futech-324	303	5	khalid	khalid	PROPN
futech-324	303	6	et	et	PROPN
futech-324	303	7	al	al	PROPN
futech-324	303	8	.	.	PROPN
futech-324	303	9	,	,	PUNCT
futech-324	303	10	"	"	PUNCT
futech-324	303	11	novel	novel	ADJ
futech-324	303	12	sentiment	sentiment	NOUN
futech-324	303	13	majority	majority	NOUN
futech-324	303	14	voting	voting	NOUN
futech-324	303	15	classifier	classifier	NOUN
futech-324	303	16	and	and	CCONJ
futech-324	303	17	transfer	transfer	VERB
futech-324	303	18	learning	learning	NOUN
futech-324	303	19	-	-	PUNCT
futech-324	303	20	based	base	VERB
futech-324	303	21	feature	feature	NOUN
futech-324	303	22	engineering	engineering	NOUN
futech-324	303	23	for	for	ADP
futech-324	303	24	sentiment	sentiment	NOUN
futech-324	303	25	analysis	analysis	NOUN
futech-324	303	26	of	of	ADP
futech-324	303	27	deepfake	deepfake	NOUN
futech-324	303	28	tweets	tweet	NOUN
futech-324	303	29	,	,	PUNCT
futech-324	303	30	"	"	PUNCT
futech-324	303	31	in	in	ADP
futech-324	303	32	ieee	ieee	NOUN
futech-324	303	33	access	access	NOUN
futech-324	303	34	,	,	PUNCT
futech-324	303	35	vol	vol	NOUN
futech-324	303	36	.	.	PROPN
futech-324	303	37	12	12	NUM
futech-324	303	38	,	,	PUNCT
futech-324	303	39	pp	pp	ADJ
futech-324	303	40	.	.	PUNCT
futech-324	303	41	67117	67117	NUM
futech-324	303	42	-	-	SYM
futech-324	303	43	67129	67129	NUM
futech-324	303	44	,	,	PUNCT
futech-324	303	45	2024	2024	NUM
futech-324	303	46	,	,	PUNCT
futech-324	303	47	doi	doi	NOUN
futech-324	303	48	:	:	PUNCT
futech-324	303	49	10.1109	10.1109	NUM
futech-324	303	50	/	/	SYM
futech-324	303	51	access.2024.3398582	access.2024.3398582	NOUN
futech-324	303	52	.	.	PUNCT
futech-324	304	1	[	[	X
futech-324	304	2	13	13	NUM
futech-324	304	3	]	]	PUNCT
futech-324	304	4	a.	a.	NOUN
futech-324	304	5	lakshmanarao	lakshmanarao	PROPN
futech-324	304	6	,	,	PUNCT
futech-324	304	7	c.	c.	PROPN
futech-324	304	8	gupta	gupta	PROPN
futech-324	304	9	and	and	CCONJ
futech-324	304	10	t.	t.	PROPN
futech-324	304	11	s.	s.	PROPN
futech-324	304	12	r.	r.	PROPN
futech-324	304	13	kiran	kiran	PROPN
futech-324	304	14	,	,	PUNCT
futech-324	304	15	"	"	PUNCT
futech-324	304	16	airline	airline	NOUN
futech-324	304	17	twitter	twitter	NOUN
futech-324	304	18	sentiment	sentiment	NOUN
futech-324	304	19	classification	classification	NOUN
futech-324	304	20	using	use	VERB
futech-324	304	21	deep	deep	ADJ
futech-324	304	22	learning	learning	NOUN
futech-324	304	23	fusion	fusion	NOUN
futech-324	304	24	,	,	PUNCT
futech-324	304	25	"	"	PUNCT
futech-324	304	26	2022	2022	NUM
futech-324	304	27	international	international	ADJ
futech-324	304	28	conference	conference	NOUN
futech-324	304	29	on	on	ADP
futech-324	304	30	smart	smart	ADJ
futech-324	304	31	generation	generation	NOUN
futech-324	304	32	computing	computing	NOUN
futech-324	304	33	,	,	PUNCT
futech-324	304	34	communication	communication	NOUN
futech-324	304	35	and	and	CCONJ
futech-324	304	36	networking	networking	NOUN
futech-324	304	37	(	(	PUNCT
futech-324	304	38	smart	smart	ADJ
futech-324	304	39	gencon	gencon	NOUN
futech-324	304	40	)	)	PUNCT
futech-324	304	41	,	,	PUNCT
futech-324	304	42	bangalore	bangalore	PROPN
futech-324	304	43	,	,	PUNCT
futech-324	304	44	india	india	PROPN
futech-324	304	45	,	,	PUNCT
futech-324	304	46	2022	2022	NUM
futech-324	304	47	,	,	PUNCT
futech-324	304	48	pp	pp	ADJ
futech-324	304	49	.	.	PUNCT
futech-324	304	50	1	1	NUM
futech-324	304	51	-	-	SYM
futech-324	304	52	4	4	NUM
futech-324	304	53	,	,	PUNCT
futech-324	304	54	doi	doi	NOUN
futech-324	304	55	:	:	PUNCT
futech-324	304	56	10.1109	10.1109	NUM
futech-324	304	57	/	/	SYM
futech-324	304	58	smartgencon56628.2022.10084207	smartgencon56628.2022.10084207	PROPN
futech-324	304	59	.	.	PUNCT
futech-324	305	1	[	[	X
futech-324	305	2	14	14	NUM
futech-324	305	3	]	]	X
futech-324	305	4	y.	y.	PROPN
futech-324	305	5	matrane	matrane	PROPN
futech-324	305	6	,	,	PUNCT
futech-324	305	7	f.	f.	PROPN
futech-324	305	8	benabbou	benabbou	PROPN
futech-324	305	9	and	and	CCONJ
futech-324	305	10	z.	z.	PROPN
futech-324	305	11	ellaky	ellaky	PROPN
futech-324	305	12	,	,	PUNCT
futech-324	305	13	"	"	PUNCT
futech-324	305	14	enhancing	enhance	VERB
futech-324	305	15	moroccan	moroccan	ADJ
futech-324	305	16	dialect	dialect	NOUN
futech-324	305	17	sentiment	sentiment	NOUN
futech-324	305	18	analysis	analysis	NOUN
futech-324	305	19	through	through	ADP
futech-324	305	20	optimized	optimize	VERB
futech-324	305	21	preprocessing	preprocessing	NOUN
futech-324	305	22	and	and	CCONJ
futech-324	305	23	transfer	transfer	VERB
futech-324	305	24	learning	learning	NOUN
futech-324	305	25	techniques	technique	NOUN
futech-324	305	26	,	,	PUNCT
futech-324	305	27	"	"	PUNCT
futech-324	305	28	in	in	ADP
futech-324	305	29	ieee	ieee	NOUN
futech-324	305	30	access	access	NOUN
futech-324	305	31	,	,	PUNCT
futech-324	305	32	vol	vol	NOUN
futech-324	305	33	.	.	PROPN
futech-324	305	34	12	12	NUM
futech-324	305	35	,	,	PUNCT
futech-324	305	36	pp	pp	ADJ
futech-324	305	37	.	.	PUNCT
futech-324	305	38	187756187777	187756187777	NUM
futech-324	305	39	,	,	PUNCT
futech-324	305	40	2024	2024	NUM
futech-324	305	41	,	,	PUNCT
futech-324	305	42	doi	doi	NOUN
futech-324	305	43	:	:	PUNCT
futech-324	305	44	10.1109	10.1109	NUM
futech-324	305	45	/	/	SYM
futech-324	305	46	access.2024.3514934	access.2024.3514934	NOUN
futech-324	305	47	.	.	PUNCT
futech-324	306	1	[	[	X
futech-324	306	2	15	15	NUM
futech-324	306	3	]	]	X
futech-324	306	4	h.	h.	PROPN
futech-324	306	5	shuqin	shuqin	PROPN
futech-324	306	6	and	and	CCONJ
futech-324	306	7	r.	r.	PROPN
futech-324	306	8	c.	c.	PROPN
futech-324	306	9	raga	raga	PROPN
futech-324	306	10	,	,	PUNCT
futech-324	306	11	"	"	PUNCT
futech-324	306	12	a	a	DET
futech-324	306	13	deep	deep	ADJ
futech-324	306	14	learning	learning	NOUN
futech-324	306	15	model	model	NOUN
futech-324	306	16	for	for	ADP
futech-324	306	17	student	student	NOUN
futech-324	306	18	sentiment	sentiment	NOUN
futech-324	306	19	analysis	analysis	NOUN
futech-324	306	20	on	on	ADP
futech-324	306	21	course	course	NOUN
futech-324	306	22	reviews	review	NOUN
futech-324	306	23	,	,	PUNCT
futech-324	306	24	"	"	PUNCT
futech-324	306	25	in	in	ADP
futech-324	306	26	ieee	ieee	NOUN
futech-324	306	27	access	access	NOUN
futech-324	306	28	,	,	PUNCT
futech-324	306	29	vol	vol	NOUN
futech-324	306	30	.	.	PROPN
futech-324	306	31	12	12	NUM
futech-324	306	32	,	,	PUNCT
futech-324	306	33	pp	pp	ADJ
futech-324	306	34	.	.	PUNCT
futech-324	306	35	136747	136747	NUM
futech-324	306	36	-	-	SYM
futech-324	306	37	136758	136758	NUM
futech-324	306	38	,	,	PUNCT
futech-324	306	39	2024	2024	NUM
futech-324	306	40	,	,	PUNCT
futech-324	306	41	doi	doi	NOUN
futech-324	306	42	:	:	PUNCT
futech-324	306	43	10.1109	10.1109	NUM
futech-324	306	44	/	/	SYM
futech-324	306	45	access.2024.3463793	access.2024.3463793	ADJ
futech-324	306	46	.	.	PUNCT
futech-324	307	1	[	[	X
futech-324	307	2	16	16	NUM
futech-324	307	3	]	]	PUNCT
futech-324	307	4	a.	a.	NOUN
futech-324	307	5	he	he	PROPN
futech-324	307	6	and	and	CCONJ
futech-324	307	7	m.	m.	NOUN
futech-324	307	8	abisado	abisado	NOUN
futech-324	307	9	,	,	PUNCT
futech-324	307	10	"	"	PUNCT
futech-324	307	11	text	text	NOUN
futech-324	307	12	sentiment	sentiment	NOUN
futech-324	307	13	analysis	analysis	NOUN
futech-324	307	14	of	of	ADP
futech-324	307	15	douban	douban	NOUN
futech-324	307	16	film	film	NOUN
futech-324	307	17	short	short	ADJ
futech-324	307	18	comments	comment	NOUN
futech-324	307	19	based	base	VERB
futech-324	307	20	on	on	ADP
futech-324	307	21	bert	bert	PROPN
futech-324	307	22	-	-	PUNCT
futech-324	307	23	cnnbilstm	cnnbilstm	PROPN
futech-324	307	24	-	-	PUNCT
futech-324	307	25	att	att	PROPN
futech-324	307	26	model	model	NOUN
futech-324	307	27	,	,	PUNCT
futech-324	307	28	"	"	PUNCT
futech-324	307	29	in	in	ADP
futech-324	307	30	ieee	ieee	NOUN
futech-324	307	31	access	access	NOUN
futech-324	307	32	,	,	PUNCT
futech-324	307	33	vol	vol	NOUN
futech-324	307	34	.	.	PROPN
futech-324	307	35	12	12	NUM
futech-324	307	36	,	,	PUNCT
futech-324	307	37	pp	pp	ADJ
futech-324	307	38	.	.	PUNCT
futech-324	307	39	4522945237	4522945237	NUM
futech-324	307	40	,	,	PUNCT
futech-324	307	41	2024	2024	NUM
futech-324	307	42	,	,	PUNCT
futech-324	307	43	doi	doi	NOUN
futech-324	307	44	:	:	PUNCT
futech-324	307	45	10.1109	10.1109	NUM
futech-324	307	46	/	/	SYM
futech-324	307	47	access.2024.3381515	access.2024.3381515	NOUN
futech-324	307	48	.	.	PUNCT
futech-324	308	1	[	[	X
futech-324	308	2	17	17	NUM
futech-324	308	3	]	]	PUNCT
futech-324	308	4	z.	z.	PROPN
futech-324	308	5	wang	wang	PROPN
futech-324	308	6	,	,	PUNCT
futech-324	308	7	g.	g.	PROPN
futech-324	308	8	xu	xu	PROPN
futech-324	308	9	,	,	PUNCT
futech-324	308	10	x.	x.	PROPN
futech-324	308	11	zhou	zhou	PROPN
futech-324	308	12	,	,	PUNCT
futech-324	308	13	j.	j.	PROPN
futech-324	308	14	y.	y.	PROPN
futech-324	308	15	kim	kim	PROPN
futech-324	308	16	,	,	PUNCT
futech-324	308	17	h.	h.	PROPN
futech-324	308	18	zhu	zhu	PROPN
futech-324	308	19	and	and	CCONJ
futech-324	308	20	l.	l.	PROPN
futech-324	308	21	deng	deng	PROPN
futech-324	308	22	,	,	PUNCT
futech-324	308	23	"	"	PUNCT
futech-324	308	24	deep	deep	ADJ
futech-324	308	25	tensor	tensor	NOUN
futech-324	308	26	evidence	evidence	NOUN
futech-324	308	27	fusion	fusion	NOUN
futech-324	308	28	network	network	NOUN
futech-324	308	29	for	for	ADP
futech-324	308	30	sentiment	sentiment	NOUN
futech-324	308	31	classification	classification	NOUN
futech-324	308	32	,	,	PUNCT
futech-324	308	33	"	"	PUNCT
futech-324	308	34	in	in	ADP
futech-324	308	35	ieee	ieee	NOUN
futech-324	308	36	transactions	transaction	NOUN
futech-324	308	37	on	on	ADP
futech-324	308	38	computational	computational	ADJ
futech-324	308	39	social	social	ADJ
futech-324	308	40	systems	system	NOUN
futech-324	308	41	,	,	PUNCT
futech-324	308	42	vol	vol	NOUN
futech-324	308	43	.	.	PROPN
futech-324	308	44	11	11	NUM
futech-324	308	45	,	,	PUNCT
futech-324	308	46	no	no	INTJ
futech-324	308	47	.	.	NOUN
futech-324	308	48	4	4	NUM
futech-324	308	49	,	,	PUNCT
futech-324	308	50	pp	pp	ADJ
futech-324	308	51	.	.	PUNCT
futech-324	308	52	4605	4605	NUM
futech-324	308	53	-	-	SYM
futech-324	308	54	4613	4613	NUM
futech-324	308	55	,	,	PUNCT
futech-324	308	56	aug	aug	PROPN
futech-324	308	57	.	.	PROPN
futech-324	308	58	2024	2024	NUM
futech-324	308	59	,	,	PUNCT
futech-324	308	60	doi	doi	NOUN
futech-324	308	61	:	:	PUNCT
futech-324	308	62	10.1109	10.1109	NUM
futech-324	308	63	/	/	SYM
futech-324	308	64	tcss.2022.3197994	tcss.2022.3197994	NUM
futech-324	308	65	.	.	PUNCT
futech-324	309	1	[	[	X
futech-324	309	2	18	18	NUM
futech-324	309	3	]	]	PUNCT
futech-324	309	4	s.	s.	PROPN
futech-324	309	5	k.	k.	PROPN
futech-324	309	6	putri	putri	PROPN
futech-324	309	7	,	,	PUNCT
futech-324	309	8	a.	a.	PROPN
futech-324	309	9	amalia	amalia	PROPN
futech-324	309	10	and	and	CCONJ
futech-324	309	11	t.	t.	PROPN
futech-324	309	12	f.	f.	PROPN
futech-324	309	13	abidin	abidin	PROPN
futech-324	309	14	,	,	PUNCT
futech-324	309	15	"	"	PUNCT
futech-324	309	16	sentiment	sentiment	NOUN
futech-324	309	17	analysis	analysis	NOUN
futech-324	309	18	multi	multi	ADJ
futech-324	309	19	-	-	NOUN
futech-324	309	20	label	label	NOUN
futech-324	309	21	of	of	ADP
futech-324	309	22	toxic	toxic	ADJ
futech-324	309	23	comments	comment	NOUN
futech-324	309	24	using	use	VERB
futech-324	309	25	bertbilstm	bertbilstm	PROPN
futech-324	309	26	methods	method	NOUN
futech-324	309	27	,	,	PUNCT
futech-324	309	28	"	"	PUNCT
futech-324	309	29	2024	2024	NUM
futech-324	309	30	international	international	ADJ
futech-324	309	31	conference	conference	NOUN
futech-324	309	32	on	on	ADP
futech-324	309	33	s.	s.	PROPN
futech-324	309	34	sushma	sushma	PROPN
futech-324	309	35	et	et	PROPN
futech-324	309	36	al	al	PROPN
futech-324	309	37	.	.	PROPN
futech-324	309	38	/future	/future	PUNCT
futech-324	309	39	technology	technology	PROPN
futech-324	309	40	august	august	PROPN
futech-324	309	41	2025|	2025|	NUM
futech-324	309	42	volume	volume	NOUN
futech-324	309	43	04	04	NUM
futech-324	310	1	|	|	NOUN
futech-324	310	2	issue	issue	NOUN
futech-324	310	3	03	03	NUM
futech-324	311	1	|	|	ADV
futech-324	311	2	pages	page	VERB
futech-324	311	3	76	76	NUM
futech-324	311	4	-	-	SYM
futech-324	311	5	84	84	NUM
futech-324	311	6	84	84	NUM
futech-324	311	7	electrical	electrical	ADJ
futech-324	311	8	engineering	engineering	NOUN
futech-324	311	9	and	and	CCONJ
futech-324	311	10	informatics	informatics	PROPN
futech-324	311	11	(	(	PUNCT
futech-324	311	12	iceltics	iceltic	NOUN
futech-324	311	13	)	)	PUNCT
futech-324	311	14	,	,	PUNCT
futech-324	311	15	banda	banda	PROPN
futech-324	311	16	aceh	aceh	PROPN
futech-324	311	17	,	,	PUNCT
futech-324	311	18	indonesia	indonesia	PROPN
futech-324	311	19	,	,	PUNCT
futech-324	311	20	2024	2024	NUM
futech-324	311	21	,	,	PUNCT
futech-324	311	22	pp	pp	ADJ
futech-324	311	23	.	.	PUNCT
futech-324	312	1	120	120	NUM
futech-324	312	2	-	-	SYM
futech-324	312	3	124	124	NUM
futech-324	312	4	,	,	PUNCT
futech-324	312	5	doi	doi	NOUN
futech-324	312	6	:	:	PUNCT
futech-324	312	7	10.1109	10.1109	NUM
futech-324	312	8	/	/	SYM
futech-324	312	9	iceltics62730.2024.10776338	iceltics62730.2024.10776338	NOUN
futech-324	312	10	.	.	PUNCT
futech-324	313	1	[	[	X
futech-324	313	2	19	19	NUM
futech-324	313	3	]	]	PUNCT
futech-324	313	4	a.	a.	NOUN
futech-324	313	5	lakshmanarao	lakshmanarao	PROPN
futech-324	313	6	,	,	PUNCT
futech-324	313	7	a.	a.	NOUN
futech-324	313	8	srisaila	srisaila	PROPN
futech-324	313	9	and	and	CCONJ
futech-324	313	10	t.	t.	PROPN
futech-324	313	11	s.	s.	PROPN
futech-324	313	12	r.	r.	PROPN
futech-324	313	13	kiran	kiran	PROPN
futech-324	313	14	,	,	PUNCT
futech-324	313	15	"	"	PUNCT
futech-324	313	16	twitter	twitter	NOUN
futech-324	313	17	sentiment	sentiment	NOUN
futech-324	313	18	classification	classification	NOUN
futech-324	313	19	with	with	ADP
futech-324	313	20	deep	deep	ADJ
futech-324	313	21	learning	learning	NOUN
futech-324	313	22	lstm	lstm	NOUN
futech-324	313	23	for	for	ADP
futech-324	313	24	airline	airline	NOUN
futech-324	313	25	tweets	tweet	NOUN
futech-324	313	26	,	,	PUNCT
futech-324	313	27	"	"	PUNCT
futech-324	313	28	2022	2022	NUM
futech-324	313	29	8th	8th	ADJ
futech-324	313	30	international	international	ADJ
futech-324	313	31	conference	conference	NOUN
futech-324	313	32	on	on	ADP
futech-324	313	33	advanced	advanced	ADJ
futech-324	313	34	computing	computing	NOUN
futech-324	313	35	and	and	CCONJ
futech-324	313	36	communication	communication	NOUN
futech-324	313	37	systems	system	NOUN
futech-324	313	38	(	(	PUNCT
futech-324	313	39	icaccs	icaccs	NOUN
futech-324	313	40	)	)	PUNCT
futech-324	313	41	,	,	PUNCT
futech-324	313	42	coimbatore	coimbatore	PROPN
futech-324	313	43	,	,	PUNCT
futech-324	313	44	india	india	PROPN
futech-324	313	45	,	,	PUNCT
futech-324	313	46	2022	2022	NUM
futech-324	313	47	,	,	PUNCT
futech-324	313	48	pp	pp	ADJ
futech-324	313	49	.	.	PUNCT
futech-324	314	1	520	520	NUM
futech-324	314	2	-	-	SYM
futech-324	314	3	524	524	NUM
futech-324	314	4	,	,	PUNCT
futech-324	314	5	doi	doi	NOUN
futech-324	314	6	:	:	PUNCT
futech-324	314	7	10.1109	10.1109	NUM
futech-324	314	8	/	/	SYM
futech-324	314	9	icaccs54159.2022.9785208	icaccs54159.2022.9785208	NOUN
futech-324	314	10	.	.	PUNCT
futech-324	315	1	[	[	X
futech-324	315	2	20	20	NUM
futech-324	315	3	]	]	PUNCT
futech-324	315	4	s.	s.	PROPN
futech-324	315	5	dutta	dutta	PROPN
futech-324	315	6	,	,	PUNCT
futech-324	315	7	m.	m.	NOUN
futech-324	315	8	neog	neog	NOUN
futech-324	315	9	and	and	CCONJ
futech-324	315	10	n.	n.	PROPN
futech-324	315	11	baruah	baruah	PROPN
futech-324	315	12	,	,	PUNCT
futech-324	315	13	"	"	PUNCT
futech-324	315	14	assamese	assamese	ADJ
futech-324	315	15	toxic	toxic	ADJ
futech-324	315	16	comment	comment	NOUN
futech-324	315	17	detection	detection	NOUN
futech-324	315	18	on	on	ADP
futech-324	315	19	social	social	ADJ
futech-324	315	20	media	medium	NOUN
futech-324	315	21	using	use	VERB
futech-324	315	22	machine	machine	NOUN
futech-324	315	23	learning	learning	NOUN
futech-324	315	24	methods	method	NOUN
futech-324	315	25	,	,	PUNCT
futech-324	315	26	"	"	PUNCT
futech-324	315	27	2024	2024	NUM
futech-324	315	28	second	second	ADJ
futech-324	315	29	international	international	ADJ
futech-324	315	30	conference	conference	NOUN
futech-324	315	31	on	on	ADP
futech-324	315	32	emerging	emerge	VERB
futech-324	315	33	trends	trend	NOUN
futech-324	315	34	in	in	ADP
futech-324	315	35	information	information	NOUN
futech-324	315	36	technology	technology	NOUN
futech-324	315	37	and	and	CCONJ
futech-324	315	38	engineering	engineering	NOUN
futech-324	315	39	(	(	PUNCT
futech-324	315	40	icetite	icetite	NOUN
futech-324	315	41	)	)	PUNCT
futech-324	315	42	,	,	PUNCT
futech-324	315	43	vellore	vellore	PROPN
futech-324	315	44	,	,	PUNCT
futech-324	315	45	india	india	PROPN
futech-324	315	46	,	,	PUNCT
futech-324	315	47	2024	2024	NUM
futech-324	315	48	,	,	PUNCT
futech-324	315	49	pp	pp	ADJ
futech-324	315	50	.	.	PUNCT
futech-324	316	1	1	1	NUM
futech-324	316	2	-	-	SYM
futech-324	316	3	8	8	NUM
futech-324	316	4	,	,	PUNCT
futech-324	316	5	doi	doi	NOUN
futech-324	316	6	:	:	PUNCT
futech-324	316	7	10.1109	10.1109	NUM
futech-324	316	8	/	/	SYM
futech-324	316	9	icetite58242.2024.10493331	icetite58242.2024.10493331	PROPN
futech-324	316	10	.	.	PUNCT
futech-324	317	1	[	[	X
futech-324	317	2	21	21	NUM
futech-324	317	3	]	]	X
futech-324	317	4	y.	y.	PROPN
futech-324	317	5	mamani	mamani	PROPN
futech-324	317	6	-	-	PUNCT
futech-324	317	7	coaquira	coaquira	PROPN
futech-324	317	8	and	and	CCONJ
futech-324	317	9	e.	e.	PROPN
futech-324	317	10	villanueva	villanueva	PROPN
futech-324	317	11	,	,	PUNCT
futech-324	317	12	"	"	PUNCT
futech-324	317	13	a	a	DET
futech-324	317	14	review	review	NOUN
futech-324	317	15	on	on	ADP
futech-324	317	16	text	text	NOUN
futech-324	317	17	sentiment	sentiment	NOUN
futech-324	317	18	analysis	analysis	NOUN
futech-324	317	19	with	with	ADP
futech-324	317	20	machine	machine	NOUN
futech-324	317	21	learning	learning	NOUN
futech-324	317	22	and	and	CCONJ
futech-324	317	23	deep	deep	ADJ
futech-324	317	24	learning	learning	NOUN
futech-324	317	25	techniques	technique	NOUN
futech-324	317	26	,	,	PUNCT
futech-324	317	27	"	"	PUNCT
futech-324	317	28	in	in	ADP
futech-324	317	29	ieee	ieee	NOUN
futech-324	317	30	access	access	NOUN
futech-324	317	31	,	,	PUNCT
futech-324	317	32	vol	vol	NOUN
futech-324	317	33	.	.	PROPN
futech-324	317	34	12	12	NUM
futech-324	317	35	,	,	PUNCT
futech-324	317	36	pp	pp	ADJ
futech-324	317	37	.	.	PUNCT
futech-324	317	38	193115	193115	NUM
futech-324	317	39	-	-	SYM
futech-324	317	40	193130	193130	NUM
futech-324	317	41	,	,	PUNCT
futech-324	317	42	2024	2024	NUM
futech-324	317	43	,	,	PUNCT
futech-324	317	44	doi	doi	NOUN
futech-324	317	45	:	:	PUNCT
futech-324	317	46	10.1109	10.1109	NUM
futech-324	317	47	/	/	SYM
futech-324	317	48	access.2024.3513321	access.2024.3513321	NOUN
futech-324	317	49	.	.	PUNCT
futech-324	318	1	[	[	X
futech-324	318	2	22	22	NUM
futech-324	318	3	]	]	SYM
futech-324	318	4	rahul	rahul	PROPN
futech-324	318	5	,	,	PUNCT
futech-324	318	6	h.	h.	PROPN
futech-324	318	7	kajla	kajla	PROPN
futech-324	318	8	,	,	PUNCT
futech-324	318	9	j.	j.	PROPN
futech-324	318	10	hooda	hooda	PROPN
futech-324	318	11	and	and	CCONJ
futech-324	318	12	g.	g.	PROPN
futech-324	318	13	saini	saini	PROPN
futech-324	318	14	,	,	PUNCT
futech-324	318	15	"	"	PUNCT
futech-324	318	16	classification	classification	NOUN
futech-324	318	17	of	of	ADP
futech-324	318	18	online	online	ADJ
futech-324	318	19	toxic	toxic	ADJ
futech-324	318	20	comments	comment	NOUN
futech-324	318	21	using	use	VERB
futech-324	318	22	machine	machine	NOUN
futech-324	318	23	learning	learning	NOUN
futech-324	318	24	algorithms	algorithm	NOUN
futech-324	318	25	,	,	PUNCT
futech-324	318	26	"	"	PUNCT
futech-324	318	27	2020	2020	NUM
futech-324	318	28	4th	4th	ADJ
futech-324	318	29	international	international	ADJ
futech-324	318	30	conference	conference	NOUN
futech-324	318	31	on	on	ADP
futech-324	318	32	intelligent	intelligent	ADJ
futech-324	318	33	computing	computing	NOUN
futech-324	318	34	and	and	CCONJ
futech-324	318	35	control	control	NOUN
futech-324	318	36	systems	system	NOUN
futech-324	318	37	(	(	PUNCT
futech-324	318	38	iciccs	iciccs	PROPN
futech-324	318	39	)	)	PUNCT
futech-324	318	40	,	,	PUNCT
futech-324	318	41	madurai	madurai	NOUN
futech-324	318	42	,	,	PUNCT
futech-324	318	43	india	india	PROPN
futech-324	318	44	,	,	PUNCT
futech-324	318	45	2020	2020	NUM
futech-324	318	46	,	,	PUNCT
futech-324	318	47	pp	pp	ADV
futech-324	318	48	.	.	PUNCT
futech-324	318	49	1119	1119	NUM
futech-324	318	50	-	-	SYM
futech-324	318	51	1123	1123	NUM
futech-324	318	52	,	,	PUNCT
futech-324	318	53	doi	doi	NOUN
futech-324	318	54	:	:	PUNCT
futech-324	318	55	10.1109	10.1109	NUM
futech-324	318	56	/	/	SYM
futech-324	318	57	iciccs48265.2020.9120939	iciccs48265.2020.9120939	PROPN
futech-324	318	58	.	.	PUNCT
futech-324	319	1	[	[	X
futech-324	319	2	23	23	NUM
futech-324	319	3	]	]	PUNCT
futech-324	319	4	m.	m.	NOUN
futech-324	319	5	aquino	aquino	PROPN
futech-324	319	6	et	et	PROPN
futech-324	319	7	al	al	PROPN
futech-324	319	8	.	.	PROPN
futech-324	319	9	,	,	PUNCT
futech-324	319	10	"	"	PUNCT
futech-324	319	11	toxic	toxic	ADJ
futech-324	319	12	comment	comment	NOUN
futech-324	319	13	detection	detection	NOUN
futech-324	319	14	:	:	PUNCT
futech-324	319	15	analyzing	analyze	VERB
futech-324	319	16	the	the	DET
futech-324	319	17	combination	combination	NOUN
futech-324	319	18	of	of	ADP
futech-324	319	19	text	text	NOUN
futech-324	319	20	and	and	CCONJ
futech-324	319	21	emojis	emoji	NOUN
futech-324	319	22	,	,	PUNCT
futech-324	319	23	"	"	PUNCT
futech-324	319	24	2021	2021	NUM
futech-324	319	25	ieee	ieee	NOUN
futech-324	319	26	18th	18th	ADJ
futech-324	319	27	international	international	ADJ
futech-324	319	28	conference	conference	NOUN
futech-324	319	29	on	on	ADP
futech-324	319	30	mobile	mobile	ADJ
futech-324	319	31	ad	ad	X
futech-324	319	32	hoc	hoc	X
futech-324	319	33	and	and	CCONJ
futech-324	319	34	smart	smart	ADJ
futech-324	319	35	systems	system	NOUN
futech-324	319	36	(	(	PUNCT
futech-324	319	37	mass	mass	PROPN
futech-324	319	38	)	)	PUNCT
futech-324	319	39	,	,	PUNCT
futech-324	319	40	denver	denver	PROPN
futech-324	319	41	,	,	PUNCT
futech-324	319	42	co	co	PROPN
futech-324	319	43	,	,	PUNCT
futech-324	319	44	usa	usa	PROPN
futech-324	319	45	,	,	PUNCT
futech-324	319	46	2021	2021	NUM
futech-324	319	47	,	,	PUNCT
futech-324	319	48	pp	pp	ADV
futech-324	319	49	.	.	PUNCT
futech-324	320	1	661	661	NUM
futech-324	320	2	-	-	SYM
futech-324	320	3	662	662	NUM
futech-324	320	4	,	,	PUNCT
futech-324	320	5	doi	doi	NOUN
futech-324	320	6	:	:	PUNCT
futech-324	320	7	10.1109	10.1109	NUM
futech-324	320	8	/	/	SYM
futech-324	320	9	mass52906.2021.00097	mass52906.2021.00097	PROPN
futech-324	320	10	.	.	PUNCT
futech-324	321	1	[	[	X
futech-324	321	2	24	24	NUM
futech-324	321	3	]	]	PUNCT
futech-324	321	4	t.	t.	PROPN
futech-324	321	5	v.	v.	PROPN
futech-324	321	6	sai	sai	PROPN
futech-324	321	7	krishna	krishna	PROPN
futech-324	321	8	,	,	PUNCT
futech-324	321	9	t.	t.	PROPN
futech-324	321	10	s.	s.	PROPN
futech-324	321	11	rama	rama	PROPN
futech-324	321	12	krishna	krishna	PROPN
futech-324	321	13	,	,	PUNCT
futech-324	321	14	s.	s.	PROPN
futech-324	321	15	kalime	kalime	PROPN
futech-324	321	16	,	,	PUNCT
futech-324	321	17	c.	c.	PROPN
futech-324	321	18	v.	v.	PROPN
futech-324	321	19	murali	murali	PROPN
futech-324	321	20	krishna	krishna	PROPN
futech-324	321	21	,	,	PUNCT
futech-324	321	22	s.	s.	PROPN
futech-324	321	23	neelima	neelima	PROPN
futech-324	321	24	,	,	PUNCT
futech-324	321	25	and	and	CCONJ
futech-324	321	26	r.	r.	PROPN
futech-324	321	27	r.	r.	PROPN
futech-324	321	28	pbv	pbv	VERB
futech-324	321	29	,	,	PUNCT
futech-324	321	30	“	"	PUNCT
futech-324	321	31	a	a	DET
futech-324	321	32	novel	novel	ADJ
futech-324	321	33	ensemble	ensemble	ADJ
futech-324	321	34	approach	approach	NOUN
futech-324	321	35	for	for	ADP
futech-324	321	36	twitter	twitter	NOUN
futech-324	321	37	sentiment	sentiment	NOUN
futech-324	321	38	classification	classification	NOUN
futech-324	321	39	with	with	ADP
futech-324	321	40	ml	ml	NOUN
futech-324	321	41	and	and	CCONJ
futech-324	321	42	lstm	lstm	ADJ
futech-324	321	43	algorithms	algorithm	NOUN
futech-324	321	44	for	for	ADP
futech-324	321	45	realtime	realtime	ADJ
futech-324	321	46	tweets	tweet	NOUN
futech-324	321	47	analysis	analysis	NOUN
futech-324	321	48	,	,	PUNCT
futech-324	321	49	”	"	PUNCT
futech-324	321	50	indonesian	indonesian	ADJ
futech-324	321	51	journal	journal	NOUN
futech-324	321	52	of	of	ADP
futech-324	321	53	electrical	electrical	ADJ
futech-324	321	54	engineering	engineering	NOUN
futech-324	321	55	and	and	CCONJ
futech-324	321	56	computer	computer	NOUN
futech-324	321	57	science	science	NOUN
futech-324	321	58	,	,	PUNCT
futech-324	321	59	vol	vol	NOUN
futech-324	321	60	.	.	PROPN
futech-324	322	1	34	34	NUM
futech-324	322	2	,	,	PUNCT
futech-324	322	3	no	no	INTJ
futech-324	322	4	.	.	NOUN
futech-324	323	1	3	3	X
futech-324	323	2	.	.	X
futech-324	323	3	institute	institute	PROPN
futech-324	323	4	of	of	ADP
futech-324	323	5	advanced	advanced	ADJ
futech-324	323	6	engineering	engineering	NOUN
futech-324	323	7	and	and	CCONJ
futech-324	323	8	science	science	NOUN
futech-324	323	9	,	,	PUNCT
futech-324	323	10	p.	p.	NOUN
futech-324	323	11	1904	1904	NUM
futech-324	323	12	,	,	PUNCT
futech-324	323	13	jun	jun	PROPN
futech-324	323	14	.	.	PROPN
futech-324	323	15	01	01	NUM
futech-324	323	16	,	,	PUNCT
futech-324	323	17	2024	2024	NUM
futech-324	323	18	.	.	PUNCT
futech-324	324	1	doi	doi	NOUN
futech-324	324	2	:	:	PUNCT
futech-324	324	3	10.11591	10.11591	NUM
futech-324	324	4	/	/	SYM
futech-324	324	5	ijeecs.v34.i3.pp1904	ijeecs.v34.i3.pp1904	NOUN
futech-324	324	6	-	-	PUNCT
futech-324	324	7	1914	1914	NUM
futech-324	324	8	.	.	PUNCT
futech-324	325	1	[	[	X
futech-324	325	2	25	25	NUM
futech-324	325	3	]	]	X
futech-324	325	4	n.	n.	PROPN
futech-324	325	5	k.	k.	PROPN
futech-324	325	6	singh	singh	PROPN
futech-324	325	7	and	and	CCONJ
futech-324	325	8	s.	s.	PROPN
futech-324	325	9	chand	chand	PROPN
futech-324	325	10	,	,	PUNCT
futech-324	325	11	"	"	PUNCT
futech-324	325	12	machine	machine	NOUN
futech-324	325	13	learning	learning	NOUN
futech-324	325	14	-	-	PUNCT
futech-324	325	15	based	base	VERB
futech-324	325	16	multilabel	multilabel	NOUN
futech-324	325	17	toxic	toxic	ADJ
futech-324	325	18	comment	comment	NOUN
futech-324	325	19	classification	classification	NOUN
futech-324	325	20	,	,	PUNCT
futech-324	325	21	"	"	PUNCT
futech-324	325	22	2022	2022	NUM
futech-324	325	23	international	international	ADJ
futech-324	325	24	conference	conference	NOUN
futech-324	325	25	on	on	ADP
futech-324	325	26	computing	computing	NOUN
futech-324	325	27	,	,	PUNCT
futech-324	325	28	communication	communication	NOUN
futech-324	325	29	,	,	PUNCT
futech-324	325	30	and	and	CCONJ
futech-324	325	31	intelligent	intelligent	ADJ
futech-324	325	32	systems	system	NOUN
futech-324	325	33	(	(	PUNCT
futech-324	325	34	icccis	icccis	NOUN
futech-324	325	35	)	)	PUNCT
futech-324	325	36	,	,	PUNCT
futech-324	325	37	greater	great	ADJ
futech-324	325	38	noida	noida	PROPN
futech-324	325	39	,	,	PUNCT
futech-324	325	40	india	india	PROPN
futech-324	325	41	,	,	PUNCT
futech-324	325	42	2022	2022	NUM
futech-324	325	43	,	,	PUNCT
futech-324	325	44	pp	pp	ADJ
futech-324	325	45	.	.	PUNCT
futech-324	326	1	435	435	NUM
futech-324	326	2	-	-	SYM
futech-324	326	3	439	439	NUM
futech-324	326	4	,	,	PUNCT
futech-324	326	5	doi	doi	NOUN
futech-324	326	6	:	:	PUNCT
futech-324	326	7	10.1109	10.1109	NUM
futech-324	326	8	/	/	SYM
futech-324	326	9	icccis56430.2022.10037626	icccis56430.2022.10037626	NOUN
futech-324	326	10	.	.	PUNCT
futech-324	327	1	[	[	X
futech-324	327	2	26	26	NUM
futech-324	327	3	]	]	X
futech-324	327	4	n.	n.	PROPN
futech-324	327	5	l.	l.	PROPN
futech-324	327	6	v.	v.	PROPN
futech-324	327	7	venugopal	venugopal	PROPN
futech-324	327	8	,	,	PUNCT
futech-324	327	9	p.	p.	PROPN
futech-324	327	10	kanchanamala	kanchanamala	PROPN
futech-324	327	11	,	,	PUNCT
futech-324	327	12	s.	s.	PROPN
futech-324	327	13	muppidi	muppidi	PROPN
futech-324	327	14	,	,	PUNCT
futech-324	327	15	t.	t.	PROPN
futech-324	327	16	b.	b.	PROPN
futech-324	327	17	prakash	prakash	PROPN
futech-324	327	18	,	,	PUNCT
futech-324	327	19	t.	t.	PROPN
futech-324	327	20	neelima	neelima	PROPN
futech-324	327	21	and	and	CCONJ
futech-324	327	22	s.	s.	PROPN
futech-324	327	23	a.	a.	PROPN
futech-324	327	24	devi	devi	PROPN
futech-324	327	25	,	,	PUNCT
futech-324	327	26	"	"	PUNCT
futech-324	327	27	multilingual	multilingual	ADJ
futech-324	327	28	toxic	toxic	ADJ
futech-324	327	29	comment	comment	NOUN
futech-324	327	30	classification	classification	NOUN
futech-324	327	31	using	use	VERB
futech-324	327	32	deep	deep	ADJ
futech-324	327	33	learning	learning	NOUN
futech-324	327	34	,	,	PUNCT
futech-324	327	35	"	"	PUNCT
futech-324	327	36	2024	2024	NUM
futech-324	327	37	2nd	2nd	ADJ
futech-324	327	38	international	international	ADJ
futech-324	327	39	conference	conference	NOUN
futech-324	327	40	on	on	ADP
futech-324	327	41	self	self	NOUN
futech-324	327	42	sustainable	sustainable	ADJ
futech-324	327	43	artificial	artificial	ADJ
futech-324	327	44	intelligence	intelligence	NOUN
futech-324	327	45	systems	system	NOUN
futech-324	327	46	(	(	PUNCT
futech-324	327	47	icssas	icssa	NOUN
futech-324	327	48	)	)	PUNCT
futech-324	327	49	,	,	PUNCT
futech-324	327	50	erode	erode	VERB
futech-324	327	51	,	,	PUNCT
futech-324	327	52	india	india	PROPN
futech-324	327	53	,	,	PUNCT
futech-324	327	54	2024	2024	NUM
futech-324	327	55	,	,	PUNCT
futech-324	327	56	pp	pp	ADJ
futech-324	327	57	.	.	PUNCT
futech-324	328	1	752	752	NUM
futech-324	328	2	-	-	SYM
futech-324	328	3	757	757	NOUN
futech-324	328	4	,	,	PUNCT
futech-324	328	5	doi	doi	NOUN
futech-324	328	6	:	:	PUNCT
futech-324	328	7	10.1109	10.1109	NUM
futech-324	328	8	/	/	SYM
futech-324	328	9	icssas64001.2024.10760913	icssas64001.2024.10760913	PROPN
futech-324	328	10	.	.	PUNCT
futech-324	329	1	[	[	X
futech-324	329	2	27	27	NUM
futech-324	329	3	]	]	X
futech-324	329	4	n.	n.	PROPN
futech-324	329	5	boudjani	boudjani	PROPN
futech-324	329	6	,	,	PUNCT
futech-324	329	7	y.	y.	PROPN
futech-324	329	8	haralambous	haralambous	PROPN
futech-324	329	9	and	and	CCONJ
futech-324	329	10	i.	i.	PROPN
futech-324	329	11	lyubareva	lyubareva	PROPN
futech-324	329	12	,	,	PUNCT
futech-324	329	13	"	"	PUNCT
futech-324	329	14	toxic	toxic	ADJ
futech-324	329	15	comment	comment	NOUN
futech-324	329	16	classification	classification	NOUN
futech-324	329	17	for	for	ADP
futech-324	329	18	french	french	ADJ
futech-324	329	19	online	online	ADJ
futech-324	329	20	comments	comment	NOUN
futech-324	329	21	,	,	PUNCT
futech-324	329	22	"	"	PUNCT
futech-324	329	23	2020	2020	NUM
futech-324	329	24	19th	19th	ADJ
futech-324	329	25	ieee	ieee	NOUN
futech-324	329	26	international	international	ADJ
futech-324	329	27	conference	conference	NOUN
futech-324	329	28	on	on	ADP
futech-324	329	29	machine	machine	NOUN
futech-324	329	30	learning	learning	NOUN
futech-324	329	31	and	and	CCONJ
futech-324	329	32	applications	application	NOUN
futech-324	329	33	(	(	PUNCT
futech-324	329	34	icmla	icmla	NOUN
futech-324	329	35	)	)	PUNCT
futech-324	329	36	,	,	PUNCT
futech-324	329	37	miami	miami	PROPN
futech-324	329	38	,	,	PUNCT
futech-324	329	39	fl	fl	PROPN
futech-324	329	40	,	,	PUNCT
futech-324	329	41	usa	usa	PROPN
futech-324	329	42	,	,	PUNCT
futech-324	329	43	2020	2020	NUM
futech-324	329	44	,	,	PUNCT
futech-324	329	45	pp	pp	ADV
futech-324	329	46	.	.	PUNCT
futech-324	330	1	1010	1010	NUM
futech-324	330	2	-	-	SYM
futech-324	330	3	1014	1014	NUM
futech-324	330	4	,	,	PUNCT
futech-324	330	5	doi	doi	NOUN
futech-324	330	6	:	:	PUNCT
futech-324	330	7	10.1109	10.1109	NUM
futech-324	330	8	/	/	SYM
futech-324	330	9	icmla51294.2020.00164	icmla51294.2020.00164	NOUN
futech-324	330	10	.	.	PUNCT
futech-324	331	1	[	[	X
futech-324	331	2	28	28	NUM
futech-324	331	3	]	]	X
futech-324	331	4	a.	a.	NOUN
futech-324	331	5	jessica	jessica	PROPN
futech-324	331	6	,	,	PUNCT
futech-324	331	7	m.	m.	PROPN
futech-324	331	8	s.	s.	PROPN
futech-324	331	9	sugiarto	sugiarto	PROPN
futech-324	331	10	,	,	PUNCT
futech-324	331	11	jerry	jerry	PROPN
futech-324	331	12	,	,	PUNCT
futech-324	331	13	s.	s.	PROPN
futech-324	331	14	achmad	achmad	PROPN
futech-324	331	15	and	and	CCONJ
futech-324	331	16	r.	r.	PROPN
futech-324	331	17	sutoyo	sutoyo	PROPN
futech-324	331	18	,	,	PUNCT
futech-324	331	19	"	"	PUNCT
futech-324	331	20	a	a	DET
futech-324	331	21	hybrid	hybrid	ADJ
futech-324	331	22	deep	deep	ADJ
futech-324	331	23	learning	learning	NOUN
futech-324	331	24	techniques	technique	NOUN
futech-324	331	25	using	use	VERB
futech-324	331	26	bert	bert	PROPN
futech-324	331	27	and	and	CCONJ
futech-324	331	28	cnn	cnn	PROPN
futech-324	331	29	for	for	ADP
futech-324	331	30	toxic	toxic	ADJ
futech-324	331	31	comments	comment	NOUN
futech-324	331	32	classification	classification	NOUN
futech-324	331	33	,	,	PUNCT
futech-324	331	34	"	"	PUNCT
futech-324	331	35	2024	2024	NUM
futech-324	331	36	international	international	ADJ
futech-324	331	37	conference	conference	NOUN
futech-324	331	38	on	on	ADP
futech-324	331	39	information	information	NOUN
futech-324	331	40	management	management	NOUN
futech-324	331	41	and	and	CCONJ
futech-324	331	42	technology	technology	NOUN
futech-324	331	43	(	(	PUNCT
futech-324	331	44	icimtech	icimtech	PROPN
futech-324	331	45	)	)	PUNCT
futech-324	331	46	,	,	PUNCT
futech-324	331	47	bali	bali	PROPN
futech-324	331	48	,	,	PUNCT
futech-324	331	49	indonesia	indonesia	PROPN
futech-324	331	50	,	,	PUNCT
futech-324	331	51	2024	2024	NUM
futech-324	331	52	,	,	PUNCT
futech-324	331	53	pp	pp	ADJ
futech-324	331	54	.	.	PUNCT
futech-324	332	1	393	393	NUM
futech-324	332	2	-	-	SYM
futech-324	332	3	398	398	NUM
futech-324	332	4	,	,	PUNCT
futech-324	332	5	doi	doi	NOUN
futech-324	332	6	:	:	PUNCT
futech-324	332	7	10.1109	10.1109	NUM
futech-324	332	8	/	/	SYM
futech-324	332	9	icimtech63123.2024.10780934	icimtech63123.2024.10780934	PROPN
futech-324	332	10	.	.	PUNCT
futech-324	333	1	[	[	X
futech-324	333	2	29	29	NUM
futech-324	333	3	]	]	AUX
futech-324	333	4	https://www.kaggle.com/c/jigsaw-multilingual-toxiccomment-classification/data	https://www.kaggle.com/c/jigsaw-multilingual-toxiccomment-classification/data	X
futech-324	333	5	this	this	DET
futech-324	333	6	article	article	NOUN
futech-324	333	7	is	be	AUX
futech-324	333	8	an	an	DET
futech-324	333	9	open	open	ADJ
futech-324	333	10	-	-	PUNCT
futech-324	333	11	access	access	NOUN
futech-324	333	12	article	article	NOUN
futech-324	333	13	distributed	distribute	VERB
futech-324	333	14	under	under	ADP
futech-324	333	15	the	the	DET
futech-324	333	16	terms	term	NOUN
futech-324	333	17	and	and	CCONJ
futech-324	333	18	conditions	condition	NOUN
futech-324	333	19	of	of	ADP
futech-324	333	20	the	the	DET
futech-324	333	21	creative	creative	ADJ
futech-324	333	22	commons	common	NOUN
futech-324	333	23	attribution	attribution	NOUN
futech-324	333	24	(	(	PUNCT
futech-324	333	25	cc	cc	NOUN
futech-324	333	26	by	by	ADP
futech-324	333	27	)	)	PUNCT
futech-324	333	28	license	license	NOUN
futech-324	333	29	(	(	PUNCT
futech-324	333	30	https://creativecommons.org/licenses/by/4.0/	https://creativecommons.org/licenses/by/4.0/	PROPN
futech-324	333	31	)	)	PUNCT
futech-324	333	32	.	.	PUNCT
futech-324	334	1	https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/data	https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/data	PROPN
futech-324	334	2	https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/data	https://www.kaggle.com/c/jigsaw-multilingual-toxic-comment-classification/data	PROPN
futech-324	334	3	https://creativecommons.org/licenses/by/4.0/	https://creativecommons.org/licenses/by/4.0/	PROPN