id	sid	tid	token	lemma	pos
cana-6072	1	1	communications	communication	NOUN
cana-6072	1	2	on	on	ADP
cana-6072	1	3	applied	apply	VERB
cana-6072	1	4	nonlinear	nonlinear	ADJ
cana-6072	1	5	analysis	analysis	NOUN
cana-6072	1	6	issn	issn	NOUN
cana-6072	1	7	:	:	PUNCT
cana-6072	1	8	1074	1074	NUM
cana-6072	1	9	-	-	PUNCT
cana-6072	1	10	133x	133x	NUM
cana-6072	1	11	vol	vol	NOUN
cana-6072	1	12	32	32	NUM
cana-6072	1	13	no	no	NOUN
cana-6072	1	14	.	.	PUNCT
cana-6072	2	1	4s	4s	NUM
cana-6072	2	2	(	(	PUNCT
cana-6072	2	3	2025	2025	NUM
cana-6072	2	4	)	)	PUNCT
cana-6072	2	5	757	757	NUM
cana-6072	2	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	2	7	sentiment	sentiment	NOUN
cana-6072	2	8	classification	classification	NOUN
cana-6072	2	9	of	of	ADP
cana-6072	2	10	emoji	emoji	PROPN
cana-6072	2	11	and	and	CCONJ
cana-6072	2	12	text	text	NOUN
cana-6072	2	13	:	:	PUNCT
cana-6072	2	14	an	an	DET
cana-6072	2	15	analysis	analysis	NOUN
cana-6072	2	16	of	of	ADP
cana-6072	2	17	model	model	NOUN
cana-6072	2	18	families	family	NOUN
cana-6072	2	19	,	,	PUNCT
cana-6072	2	20	fusion	fusion	NOUN
cana-6072	2	21	strategies	strategy	NOUN
cana-6072	2	22	,	,	PUNCT
cana-6072	2	23	and	and	CCONJ
cana-6072	2	24	performance	performance	NOUN
cana-6072	2	25	gains	gain	NOUN
cana-6072	2	26	sapandeep	sapandeep	VERB
cana-6072	2	27	singh	singh	PROPN
cana-6072	2	28	sandhu	sandhu	PROPN
cana-6072	2	29	research	research	PROPN
cana-6072	2	30	scholar	scholar	NOUN
cana-6072	2	31	,	,	PUNCT
cana-6072	2	32	university	university	NOUN
cana-6072	2	33	institute	institute	NOUN
cana-6072	2	34	of	of	ADP
cana-6072	2	35	computing	computing	PROPN
cana-6072	2	36	,	,	PUNCT
cana-6072	2	37	chandigarh	chandigarh	PROPN
cana-6072	2	38	university	university	NOUN
cana-6072	2	39	,	,	PUNCT
cana-6072	2	40	gharuan	gharuan	PROPN
cana-6072	2	41	,	,	PUNCT
cana-6072	2	42	mohali	mohali	PROPN
cana-6072	2	43	,	,	PUNCT
cana-6072	2	44	punjab	punjab	PROPN
cana-6072	2	45	140413	140413	NUM
cana-6072	2	46	,	,	PUNCT
cana-6072	2	47	india	india	PROPN
cana-6072	2	48	.	.	PUNCT
cana-6072	3	1	sapandeepsinghsandhu@gmail.com	sapandeepsinghsandhu@gmail.com	PROPN
cana-6072	3	2	,	,	PUNCT
cana-6072	3	3	orcid	orcid	PROPN
cana-6072	3	4	:	:	PUNCT
cana-6072	3	5	https://orcid.org/0000-0002-6865-7493	https://orcid.org/0000-0002-6865-7493	PROPN
cana-6072	3	6	dr	dr	PROPN
cana-6072	3	7	.	.	PROPN
cana-6072	3	8	amanpreet	amanpreet	PROPN
cana-6072	3	9	kaur	kaur	PROPN
cana-6072	3	10	sandhu	sandhu	PROPN
cana-6072	3	11	professor	professor	PROPN
cana-6072	3	12	,	,	PUNCT
cana-6072	3	13	university	university	NOUN
cana-6072	3	14	institute	institute	NOUN
cana-6072	3	15	of	of	ADP
cana-6072	3	16	computing	computing	PROPN
cana-6072	3	17	,	,	PUNCT
cana-6072	3	18	chandigarh	chandigarh	PROPN
cana-6072	3	19	university	university	NOUN
cana-6072	3	20	,	,	PUNCT
cana-6072	3	21	gharuan	gharuan	PROPN
cana-6072	3	22	,	,	PUNCT
cana-6072	3	23	mohali	mohali	PROPN
cana-6072	3	24	,	,	PUNCT
cana-6072	3	25	punjab	punjab	PROPN
cana-6072	3	26	140413	140413	NUM
cana-6072	3	27	,	,	PUNCT
cana-6072	3	28	india	india	PROPN
cana-6072	3	29	amanpreet.e9962@cumail.in	amanpreet.e9962@cumail.in	PART
cana-6072	3	30	article	article	NOUN
cana-6072	3	31	history	history	NOUN
cana-6072	3	32	:	:	PUNCT
cana-6072	3	33	received:01/02/2025	received:01/02/2025	PROPN
cana-6072	3	34	revised	revise	VERB
cana-6072	3	35	:	:	PUNCT
cana-6072	3	36	06/03/2025	06/03/2025	NUM
cana-6072	3	37	accepted	accept	VERB
cana-6072	3	38	:	:	PUNCT
cana-6072	3	39	30/04/2025	30/04/2025	NUM
cana-6072	3	40	abstract	abstract	NOUN
cana-6072	3	41	:	:	PUNCT
cana-6072	3	42	social	social	ADJ
cana-6072	3	43	media	medium	NOUN
cana-6072	3	44	posts	post	NOUN
cana-6072	3	45	often	often	ADV
cana-6072	3	46	combine	combine	VERB
cana-6072	3	47	words	word	NOUN
cana-6072	3	48	with	with	ADP
cana-6072	3	49	emojis	emoji	NOUN
cana-6072	3	50	that	that	PRON
cana-6072	3	51	intensify	intensify	VERB
cana-6072	3	52	,	,	PUNCT
cana-6072	3	53	invert	invert	NOUN
cana-6072	3	54	,	,	PUNCT
cana-6072	3	55	or	or	CCONJ
cana-6072	3	56	clarify	clarify	VERB
cana-6072	3	57	sentiment	sentiment	NOUN
cana-6072	3	58	.	.	PUNCT
cana-6072	4	1	because	because	SCONJ
cana-6072	4	2	emojis	emoji	NOUN
cana-6072	4	3	add	add	VERB
cana-6072	4	4	information	information	NOUN
cana-6072	4	5	that	that	PRON
cana-6072	4	6	is	be	AUX
cana-6072	4	7	not	not	PART
cana-6072	4	8	present	present	ADJ
cana-6072	4	9	in	in	ADP
cana-6072	4	10	text	text	NOUN
cana-6072	4	11	alone	alone	ADV
cana-6072	4	12	,	,	PUNCT
cana-6072	4	13	removing	remove	VERB
cana-6072	4	14	them	they	PRON
cana-6072	4	15	can	can	AUX
cana-6072	4	16	not	not	PART
cana-6072	4	17	,	,	PUNCT
cana-6072	4	18	in	in	ADP
cana-6072	4	19	principle	principle	NOUN
cana-6072	4	20	,	,	PUNCT
cana-6072	4	21	reduce	reduce	VERB
cana-6072	4	22	optimal	optimal	ADJ
cana-6072	4	23	classification	classification	NOUN
cana-6072	4	24	error	error	NOUN
cana-6072	4	25	;	;	PUNCT
cana-6072	4	26	conversely	conversely	ADV
cana-6072	4	27	,	,	PUNCT
cana-6072	4	28	naïve	naïve	ADJ
cana-6072	4	29	handling	handling	NOUN
cana-6072	4	30	can	can	AUX
cana-6072	4	31	misinterpret	misinterpret	VERB
cana-6072	4	32	polarity	polarity	NOUN
cana-6072	4	33	in	in	ADP
cana-6072	4	34	practice	practice	NOUN
cana-6072	4	35	.	.	PUNCT
cana-6072	5	1	this	this	DET
cana-6072	5	2	study	study	NOUN
cana-6072	5	3	examines	examine	VERB
cana-6072	5	4	emoji	emoji	NOUN
cana-6072	5	5	-	-	PUNCT
cana-6072	5	6	aware	aware	ADJ
cana-6072	5	7	sentiment	sentiment	NOUN
cana-6072	5	8	classification	classification	NOUN
cana-6072	5	9	,	,	PUNCT
cana-6072	5	10	including	include	VERB
cana-6072	5	11	the	the	DET
cana-6072	5	12	model	model	NOUN
cana-6072	5	13	families	family	NOUN
cana-6072	5	14	,	,	PUNCT
cana-6072	5	15	fusion	fusion	NOUN
cana-6072	5	16	strategies	strategy	NOUN
cana-6072	5	17	,	,	PUNCT
cana-6072	5	18	emoji	emoji	PROPN
cana-6072	5	19	representations	representation	NOUN
cana-6072	5	20	,	,	PUNCT
cana-6072	5	21	datasets	dataset	NOUN
cana-6072	5	22	,	,	PUNCT
cana-6072	5	23	metrics	metric	NOUN
cana-6072	5	24	,	,	PUNCT
cana-6072	5	25	performance	performance	NOUN
cana-6072	5	26	gains	gain	NOUN
cana-6072	5	27	over	over	ADP
cana-6072	5	28	text	text	NOUN
cana-6072	5	29	-	-	PUNCT
cana-6072	5	30	only	only	ADJ
cana-6072	5	31	baselines	baseline	NOUN
cana-6072	5	32	,	,	PUNCT
cana-6072	5	33	and	and	CCONJ
cana-6072	5	34	recurring	recur	VERB
cana-6072	5	35	failure	failure	NOUN
cana-6072	5	36	modes	mode	NOUN
cana-6072	5	37	.	.	PUNCT
cana-6072	6	1	we	we	PRON
cana-6072	6	2	reviewed	review	VERB
cana-6072	6	3	20	20	NUM
cana-6072	6	4	english	english	ADJ
cana-6072	6	5	-	-	PUNCT
cana-6072	6	6	language	language	NOUN
cana-6072	6	7	studies	study	NOUN
cana-6072	6	8	(	(	PUNCT
cana-6072	6	9	2015–2025	2015–2025	NUM
cana-6072	6	10	)	)	PUNCT
cana-6072	6	11	evaluated	evaluate	VERB
cana-6072	6	12	via	via	ADP
cana-6072	6	13	a	a	DET
cana-6072	6	14	predefined	predefine	VERB
cana-6072	6	15	schema	schema	NOUN
cana-6072	6	16	for	for	ADP
cana-6072	6	17	reproducibility	reproducibility	NOUN
cana-6072	6	18	,	,	PUNCT
cana-6072	6	19	baseline	baseline	NOUN
cana-6072	6	20	parity	parity	NOUN
cana-6072	6	21	,	,	PUNCT
cana-6072	6	22	ablations	ablation	NOUN
cana-6072	6	23	,	,	PUNCT
cana-6072	6	24	and	and	CCONJ
cana-6072	6	25	statistical	statistical	ADJ
cana-6072	6	26	reporting	reporting	NOUN
cana-6072	6	27	.	.	PUNCT
cana-6072	7	1	emoji	emoji	PROPN
cana-6072	7	2	representations	representation	NOUN
cana-6072	7	3	and	and	CCONJ
cana-6072	7	4	fusion	fusion	NOUN
cana-6072	7	5	strategies	strategy	NOUN
cana-6072	7	6	form	form	VERB
cana-6072	7	7	a	a	DET
cana-6072	7	8	multimodal	multimodal	NOUN
cana-6072	7	9	taxonomy	taxonomy	NOUN
cana-6072	7	10	.	.	PUNCT
cana-6072	8	1	of	of	ADP
cana-6072	8	2	the	the	DET
cana-6072	8	3	fusion	fusion	NOUN
cana-6072	8	4	strategies	strategy	NOUN
cana-6072	8	5	,	,	PUNCT
cana-6072	8	6	intermediate	intermediate	ADJ
cana-6072	8	7	fusion	fusion	NOUN
cana-6072	8	8	most	most	ADV
cana-6072	8	9	often	often	ADV
cana-6072	8	10	outperforms	outperform	VERB
cana-6072	8	11	early	early	ADJ
cana-6072	8	12	and	and	CCONJ
cana-6072	8	13	late	late	ADJ
cana-6072	8	14	variants	variant	NOUN
cana-6072	8	15	despite	despite	SCONJ
cana-6072	8	16	higher	high	ADJ
cana-6072	8	17	resource	resource	NOUN
cana-6072	8	18	costs	cost	NOUN
cana-6072	8	19	,	,	PUNCT
cana-6072	8	20	with	with	ADP
cana-6072	8	21	reported	report	VERB
cana-6072	8	22	results	result	NOUN
cana-6072	8	23	including	include	VERB
cana-6072	8	24	95	95	NUM
cana-6072	8	25	%	%	NOUN
cana-6072	8	26	accuracy	accuracy	NOUN
cana-6072	8	27	(	(	PUNCT
cana-6072	8	28	early	early	ADJ
cana-6072	8	29	-	-	PUNCT
cana-6072	8	30	fusion	fusion	NOUN
cana-6072	8	31	bilstm	bilstm	NOUN
cana-6072	8	32	on	on	ADP
cana-6072	8	33	weibo	weibo	NOUN
cana-6072	8	34	)	)	PUNCT
cana-6072	8	35	,	,	PUNCT
cana-6072	8	36	gains	gain	NOUN
cana-6072	8	37	of	of	ADP
cana-6072	8	38	2.3	2.3	NUM
cana-6072	8	39	%	%	NOUN
cana-6072	8	40	,	,	PUNCT
cana-6072	8	41	10.9	10.9	NUM
cana-6072	8	42	%	%	NOUN
cana-6072	8	43	,	,	PUNCT
cana-6072	8	44	and	and	CCONJ
cana-6072	8	45	2.7	2.7	NUM
cana-6072	8	46	%	%	NOUN
cana-6072	8	47	(	(	PUNCT
cana-6072	8	48	cross	cross	ADJ
cana-6072	8	49	-	-	ADJ
cana-6072	8	50	attention	attention	ADJ
cana-6072	8	51	model	model	NOUN
cana-6072	8	52	on	on	ADP
cana-6072	8	53	msd	msd	PROPN
cana-6072	8	54	/	/	SYM
cana-6072	8	55	td	td	PROPN
cana-6072	8	56	/	/	SYM
cana-6072	8	57	erd	erd	NOUN
cana-6072	8	58	)	)	PUNCT
cana-6072	8	59	,	,	PUNCT
cana-6072	8	60	80.42	80.42	NUM
cana-6072	8	61	%	%	NOUN
cana-6072	8	62	macro	macro	NOUN
cana-6072	8	63	-	-	NOUN
cana-6072	8	64	f1	f1	NOUN
cana-6072	8	65	and	and	CCONJ
cana-6072	8	66	87.95	87.95	NUM
cana-6072	8	67	%	%	NOUN
cana-6072	8	68	accuracy	accuracy	NOUN
cana-6072	8	69	(	(	PUNCT
cana-6072	8	70	mutual	mutual	ADJ
cana-6072	8	71	-	-	PUNCT
cana-6072	8	72	attention	attention	NOUN
cana-6072	8	73	text	text	NOUN
cana-6072	8	74	-	-	PUNCT
cana-6072	8	75	emoji	emoji	ADJ
cana-6072	8	76	-	-	PUNCT
cana-6072	8	77	image	image	NOUN
cana-6072	8	78	model	model	NOUN
cana-6072	8	79	)	)	PUNCT
cana-6072	8	80	,	,	PUNCT
cana-6072	8	81	and	and	CCONJ
cana-6072	8	82	approximately	approximately	ADV
cana-6072	8	83	9point	9point	NUM
cana-6072	8	84	accuracy	accuracy	NOUN
cana-6072	8	85	improvement	improvement	NOUN
cana-6072	8	86	on	on	ADP
cana-6072	8	87	twitter	twitter	NOUN
cana-6072	8	88	airline	airline	NOUN
cana-6072	8	89	.	.	PUNCT
cana-6072	9	1	generalization	generalization	NOUN
cana-6072	9	2	to	to	ADP
cana-6072	9	3	unseen	unseen	ADJ
cana-6072	9	4	emojis	emoji	NOUN
cana-6072	9	5	remains	remain	VERB
cana-6072	9	6	fragile	fragile	ADJ
cana-6072	9	7	without	without	ADP
cana-6072	9	8	robust	robust	ADJ
cana-6072	9	9	encoders	encoder	NOUN
cana-6072	9	10	,	,	PUNCT
cana-6072	9	11	necessitating	necessitate	VERB
cana-6072	9	12	validation	validation	NOUN
cana-6072	9	13	via	via	ADP
cana-6072	9	14	paired	pair	VERB
cana-6072	9	15	testing	testing	NOUN
cana-6072	9	16	and	and	CCONJ
cana-6072	9	17	confidence	confidence	NOUN
cana-6072	9	18	intervals	interval	NOUN
cana-6072	9	19	against	against	ADP
cana-6072	9	20	threats	threat	NOUN
cana-6072	9	21	like	like	ADP
cana-6072	9	22	data	datum	NOUN
cana-6072	9	23	leakage	leakage	NOUN
cana-6072	9	24	.	.	PUNCT
cana-6072	10	1	keywords	keyword	NOUN
cana-6072	10	2	:	:	PUNCT
cana-6072	10	3	sentiment	sentiment	NOUN
cana-6072	10	4	analysis	analysis	NOUN
cana-6072	10	5	,	,	PUNCT
cana-6072	10	6	emojis	emoji	NOUN
cana-6072	10	7	,	,	PUNCT
cana-6072	10	8	multimodal	multimodal	ADJ
cana-6072	10	9	fusion	fusion	NOUN
cana-6072	10	10	,	,	PUNCT
cana-6072	10	11	bert	bert	NOUN
cana-6072	10	12	,	,	PUNCT
cana-6072	10	13	bertweet	bertweet	NOUN
cana-6072	10	14	,	,	PUNCT
cana-6072	10	15	cross	cross	NOUN
cana-6072	10	16	-	-	ADJ
cana-6072	10	17	attention	attention	ADJ
cana-6072	10	18	,	,	PUNCT
cana-6072	10	19	reproducibility	reproducibility	NOUN
cana-6072	10	20	,	,	PUNCT
cana-6072	10	21	prisma	prisma	PROPN
cana-6072	10	22	introduction	introduction	NOUN
cana-6072	10	23	emojis	emoji	NOUN
cana-6072	10	24	have	have	AUX
cana-6072	10	25	become	become	VERB
cana-6072	10	26	ubiquitous	ubiquitous	ADJ
cana-6072	10	27	in	in	ADP
cana-6072	10	28	online	online	ADJ
cana-6072	10	29	communication	communication	NOUN
cana-6072	10	30	,	,	PUNCT
cana-6072	10	31	serving	serve	VERB
cana-6072	10	32	as	as	ADP
cana-6072	10	33	paralinguistic	paralinguistic	ADJ
cana-6072	10	34	cues	cue	NOUN
cana-6072	10	35	that	that	PRON
cana-6072	10	36	intensify	intensify	VERB
cana-6072	10	37	,	,	PUNCT
cana-6072	10	38	invert	invert	NOUN
cana-6072	10	39	,	,	PUNCT
cana-6072	10	40	or	or	CCONJ
cana-6072	10	41	clarify	clarify	VERB
cana-6072	10	42	the	the	DET
cana-6072	10	43	sentiment	sentiment	NOUN
cana-6072	10	44	of	of	ADP
cana-6072	10	45	textual	textual	ADJ
cana-6072	10	46	messages	message	NOUN
cana-6072	10	47	.	.	PUNCT
cana-6072	11	1	in	in	ADP
cana-6072	11	2	informal	informal	ADJ
cana-6072	11	3	social	social	ADJ
cana-6072	11	4	media	medium	NOUN
cana-6072	11	5	posts	post	NOUN
cana-6072	11	6	,	,	PUNCT
cana-6072	11	7	a	a	DET
cana-6072	11	8	single	single	ADJ
cana-6072	11	9	emoji	emoji	NOUN
cana-6072	11	10	can	can	AUX
cana-6072	11	11	dramatically	dramatically	ADV
cana-6072	11	12	alter	alter	VERB
cana-6072	11	13	the	the	DET
cana-6072	11	14	tone	tone	NOUN
cana-6072	11	15	–	–	PUNCT
cana-6072	11	16	for	for	ADP
cana-6072	11	17	example	example	NOUN
cana-6072	11	18	,	,	PUNCT
cana-6072	11	19	the	the	DET
cana-6072	11	20	word	word	NOUN
cana-6072	11	21	“	"	PUNCT
cana-6072	11	22	great	great	ADJ
cana-6072	11	23	”	"	PUNCT
cana-6072	11	24	can	can	AUX
cana-6072	11	25	be	be	AUX
cana-6072	11	26	sarcastically	sarcastically	ADV
cana-6072	11	27	negated	negate	VERB
cana-6072	11	28	by	by	ADP
cana-6072	11	29	adding	add	VERB
cana-6072	11	30	(	(	PUNCT
cana-6072	11	31	upside	upside	ADV
cana-6072	11	32	-	-	PUNCT
cana-6072	11	33	down	down	ADP
cana-6072	11	34	face	face	NOUN
cana-6072	11	35	)	)	PUNCT
cana-6072	11	36	emoji	emoji	PROPN
cana-6072	12	1	[	[	X
cana-6072	12	2	lou	lou	PROPN
cana-6072	12	3	,	,	PUNCT
cana-6072	12	4	y.	y.	PROPN
cana-6072	12	5	et	et	PROPN
cana-6072	12	6	al	al	PROPN
cana-6072	12	7	.	.	PROPN
cana-6072	12	8	,	,	PUNCT
cana-6072	12	9	2024	2024	NUM
cana-6072	12	10	]	]	PUNCT
cana-6072	12	11	.	.	PUNCT
cana-6072	13	1	early	early	ADJ
cana-6072	13	2	sentiment	sentiment	NOUN
cana-6072	13	3	analysis	analysis	NOUN
cana-6072	13	4	systems	system	NOUN
cana-6072	13	5	struggled	struggle	VERB
cana-6072	13	6	with	with	ADP
cana-6072	13	7	such	such	ADJ
cana-6072	13	8	cases	case	NOUN
cana-6072	13	9	,	,	PUNCT
cana-6072	13	10	often	often	ADV
cana-6072	13	11	treating	treat	VERB
cana-6072	13	12	emojis	emoji	NOUN
cana-6072	13	13	as	as	ADP
cana-6072	13	14	noise	noise	NOUN
cana-6072	13	15	and	and	CCONJ
cana-6072	13	16	simply	simply	ADV
cana-6072	13	17	removing	remove	VERB
cana-6072	13	18	them	they	PRON
cana-6072	13	19	during	during	ADP
cana-6072	13	20	preprocessing	preprocessing	NOUN
cana-6072	13	21	,	,	PUNCT
cana-6072	13	22	which	which	PRON
cana-6072	13	23	discards	discard	VERB
cana-6072	13	24	valuable	valuable	ADJ
cana-6072	13	25	affective	affective	ADJ
cana-6072	13	26	signals	signal	NOUN
cana-6072	14	1	[	[	X
cana-6072	14	2	khan	khan	PROPN
cana-6072	14	3	,	,	PUNCT
cana-6072	14	4	a.	a.	PROPN
cana-6072	14	5	et	et	PROPN
cana-6072	14	6	al	al	PROPN
cana-6072	14	7	.	.	PROPN
cana-6072	14	8	,	,	PUNCT
cana-6072	14	9	2025	2025	NUM
cana-6072	14	10	]	]	PUNCT
cana-6072	14	11	.	.	PUNCT
cana-6072	15	1	this	this	DET
cana-6072	15	2	practice	practice	NOUN
cana-6072	15	3	leads	lead	VERB
cana-6072	15	4	to	to	ADP
cana-6072	15	5	significant	significant	ADJ
cana-6072	15	6	misinterpretations	misinterpretation	NOUN
cana-6072	15	7	and	and	CCONJ
cana-6072	15	8	missed	miss	VERB
cana-6072	15	9	nuances	nuance	NOUN
cana-6072	15	10	,	,	PUNCT
cana-6072	15	11	driving	drive	VERB
cana-6072	15	12	a	a	DET
cana-6072	15	13	surge	surge	NOUN
cana-6072	15	14	of	of	ADP
cana-6072	15	15	interest	interest	NOUN
cana-6072	16	1	mailto:sapandeepsinghsandhu@gmail.com	mailto:sapandeepsinghsandhu@gmail.com	X
cana-6072	16	2	https://orcid.org/0000-0002-6865-7493	https://orcid.org/0000-0002-6865-7493	ADV
cana-6072	16	3	mailto:amanpreet.e9962@cumail.in	mailto:amanpreet.e9962@cumail.in	PUNCT
cana-6072	16	4	communications	communication	NOUN
cana-6072	16	5	on	on	ADP
cana-6072	16	6	applied	apply	VERB
cana-6072	16	7	nonlinear	nonlinear	ADJ
cana-6072	16	8	analysis	analysis	NOUN
cana-6072	16	9	issn	issn	NOUN
cana-6072	16	10	:	:	PUNCT
cana-6072	16	11	1074	1074	NUM
cana-6072	16	12	-	-	PUNCT
cana-6072	16	13	133x	133x	NUM
cana-6072	16	14	vol	vol	NOUN
cana-6072	16	15	32	32	NUM
cana-6072	16	16	no	no	NOUN
cana-6072	16	17	.	.	PUNCT
cana-6072	17	1	4s	4s	NUM
cana-6072	17	2	(	(	PUNCT
cana-6072	17	3	2025	2025	NUM
cana-6072	17	4	)	)	PUNCT
cana-6072	17	5	758	758	NUM
cana-6072	17	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	17	7	in	in	ADP
cana-6072	17	8	developing	develop	VERB
cana-6072	17	9	emoji	emoji	NOUN
cana-6072	17	10	-	-	PUNCT
cana-6072	17	11	aware	aware	ADJ
cana-6072	17	12	sentiment	sentiment	NOUN
cana-6072	17	13	analysis	analysis	NOUN
cana-6072	17	14	models	model	NOUN
cana-6072	17	15	that	that	PRON
cana-6072	17	16	jointly	jointly	ADV
cana-6072	17	17	leverage	leverage	VERB
cana-6072	17	18	text	text	NOUN
cana-6072	17	19	and	and	CCONJ
cana-6072	17	20	emoji	emoji	PROPN
cana-6072	17	21	information	information	NOUN
cana-6072	17	22	for	for	ADP
cana-6072	17	23	more	more	ADV
cana-6072	17	24	robust	robust	ADJ
cana-6072	17	25	sentiment	sentiment	NOUN
cana-6072	17	26	classification	classification	NOUN
cana-6072	18	1	[	[	X
cana-6072	18	2	khan	khan	PROPN
cana-6072	18	3	,	,	PUNCT
cana-6072	18	4	a.	a.	PROPN
cana-6072	18	5	et	et	PROPN
cana-6072	18	6	al	al	PROPN
cana-6072	18	7	.	.	PROPN
cana-6072	18	8	,	,	PUNCT
cana-6072	18	9	2025	2025	NUM
cana-6072	18	10	;	;	PUNCT
cana-6072	18	11	lou	lou	PROPN
cana-6072	18	12	,	,	PUNCT
cana-6072	18	13	y.	y.	PROPN
cana-6072	18	14	et	et	PROPN
cana-6072	18	15	al	al	PROPN
cana-6072	18	16	.	.	PROPN
cana-6072	18	17	,	,	PUNCT
cana-6072	18	18	2020	2020	NUM
cana-6072	18	19	]	]	PUNCT
cana-6072	18	20	.	.	PUNCT
cana-6072	19	1	the	the	DET
cana-6072	19	2	objective	objective	NOUN
cana-6072	19	3	of	of	ADP
cana-6072	19	4	this	this	DET
cana-6072	19	5	systematic	systematic	ADJ
cana-6072	19	6	literature	literature	NOUN
cana-6072	19	7	review	review	NOUN
cana-6072	19	8	(	(	PUNCT
cana-6072	19	9	slr	slr	NOUN
cana-6072	19	10	)	)	PUNCT
cana-6072	19	11	is	be	AUX
cana-6072	19	12	to	to	PART
cana-6072	19	13	analyze	analyze	VERB
cana-6072	19	14	the	the	DET
cana-6072	19	15	effectiveness	effectiveness	NOUN
cana-6072	19	16	of	of	ADP
cana-6072	19	17	these	these	DET
cana-6072	19	18	emoji	emoji	NOUN
cana-6072	19	19	-	-	PUNCT
cana-6072	19	20	enhanced	enhance	VERB
cana-6072	19	21	sentiment	sentiment	NOUN
cana-6072	19	22	analysis	analysis	NOUN
cana-6072	19	23	models	model	NOUN
cana-6072	19	24	and	and	CCONJ
cana-6072	19	25	explore	explore	VERB
cana-6072	19	26	the	the	DET
cana-6072	19	27	range	range	NOUN
cana-6072	19	28	of	of	ADP
cana-6072	19	29	model	model	NOUN
cana-6072	19	30	architectures	architecture	NOUN
cana-6072	19	31	proposed	propose	VERB
cana-6072	19	32	to	to	PART
cana-6072	19	33	incorporate	incorporate	VERB
cana-6072	19	34	emojis	emoji	NOUN
cana-6072	19	35	.	.	PUNCT
cana-6072	20	1	we	we	PRON
cana-6072	20	2	focus	focus	VERB
cana-6072	20	3	on	on	ADP
cana-6072	20	4	understanding	understand	VERB
cana-6072	20	5	how	how	SCONJ
cana-6072	20	6	integrating	integrating	ADJ
cana-6072	20	7	emojis	emoji	NOUN
cana-6072	20	8	as	as	ADP
cana-6072	20	9	textual	textual	ADJ
cana-6072	20	10	features	feature	NOUN
cana-6072	20	11	or	or	CCONJ
cana-6072	20	12	as	as	ADP
cana-6072	20	13	visual	visual	ADJ
cana-6072	20	14	images	image	NOUN
cana-6072	20	15	influences	influence	VERB
cana-6072	20	16	classification	classification	NOUN
cana-6072	20	17	performance	performance	NOUN
cana-6072	20	18	and	and	CCONJ
cana-6072	20	19	the	the	DET
cana-6072	20	20	fusion	fusion	NOUN
cana-6072	20	21	strategies	strategy	NOUN
cana-6072	20	22	used	use	VERB
cana-6072	20	23	by	by	ADP
cana-6072	20	24	state	state	NOUN
cana-6072	20	25	-	-	PUNCT
cana-6072	20	26	of	of	ADP
cana-6072	20	27	-	-	PUNCT
cana-6072	20	28	the	the	DET
cana-6072	20	29	-	-	PUNCT
cana-6072	20	30	art	art	NOUN
cana-6072	20	31	models	model	NOUN
cana-6072	20	32	to	to	PART
cana-6072	20	33	combine	combine	VERB
cana-6072	20	34	these	these	DET
cana-6072	20	35	multimodal	multimodal	NOUN
cana-6072	20	36	signals	signal	NOUN
cana-6072	20	37	[	[	X
cana-6072	20	38	lou	lou	PROPN
cana-6072	20	39	,	,	PUNCT
cana-6072	20	40	y.	y.	PROPN
cana-6072	20	41	et	et	PROPN
cana-6072	20	42	al	al	PROPN
cana-6072	20	43	.	.	PROPN
cana-6072	20	44	,	,	PUNCT
cana-6072	20	45	2020	2020	NUM
cana-6072	20	46	]	]	PUNCT
cana-6072	20	47	,	,	PUNCT
cana-6072	20	48	[	[	X
cana-6072	20	49	xu	xu	INTJ
cana-6072	20	50	,	,	PUNCT
cana-6072	20	51	q.	q.	PROPN
cana-6072	20	52	et	et	PROPN
cana-6072	20	53	al	al	PROPN
cana-6072	20	54	.	.	PROPN
cana-6072	20	55	,	,	PUNCT
cana-6072	20	56	2024	2024	NUM
cana-6072	20	57	]	]	PUNCT
cana-6072	20	58	.	.	PUNCT
cana-6072	21	1	the	the	DET
cana-6072	21	2	primary	primary	ADJ
cana-6072	21	3	domain	domain	NOUN
cana-6072	21	4	of	of	ADP
cana-6072	21	5	interest	interest	NOUN
cana-6072	21	6	is	be	AUX
cana-6072	21	7	social	social	ADJ
cana-6072	21	8	media	medium	NOUN
cana-6072	21	9	(	(	PUNCT
cana-6072	21	10	e.g.	e.g.	ADV
cana-6072	21	11	,	,	PUNCT
cana-6072	21	12	twitter	twitter	NOUN
cana-6072	21	13	,	,	PUNCT
cana-6072	21	14	weibo	weibo	PROPN
cana-6072	21	15	)	)	PUNCT
cana-6072	21	16	,	,	PUNCT
cana-6072	21	17	where	where	SCONJ
cana-6072	21	18	emoji	emoji	PROPN
cana-6072	21	19	usage	usage	NOUN
cana-6072	21	20	is	be	AUX
cana-6072	21	21	pervasive	pervasive	ADJ
cana-6072	21	22	,	,	PUNCT
cana-6072	21	23	to	to	PART
cana-6072	21	24	gauge	gauge	VERB
cana-6072	21	25	the	the	DET
cana-6072	21	26	generality	generality	NOUN
cana-6072	21	27	of	of	ADP
cana-6072	21	28	these	these	DET
cana-6072	21	29	methods	method	NOUN
cana-6072	21	30	and	and	CCONJ
cana-6072	21	31	identify	identify	VERB
cana-6072	21	32	strategies	strategy	NOUN
cana-6072	21	33	that	that	PRON
cana-6072	21	34	consistently	consistently	ADV
cana-6072	21	35	improve	improve	VERB
cana-6072	21	36	sentiment	sentiment	NOUN
cana-6072	21	37	detection	detection	NOUN
cana-6072	21	38	[	[	X
cana-6072	21	39	eisner	eisner	NOUN
cana-6072	21	40	,	,	PUNCT
cana-6072	21	41	b.	b.	PROPN
cana-6072	21	42	,	,	PUNCT
cana-6072	21	43	et	et	PROPN
cana-6072	21	44	al	al	PROPN
cana-6072	21	45	.	.	PROPN
cana-6072	21	46	,	,	PUNCT
cana-6072	21	47	2016	2016	NUM
cana-6072	21	48	]	]	PUNCT
cana-6072	21	49	.	.	PUNCT
cana-6072	22	1	the	the	DET
cana-6072	22	2	motivation	motivation	NOUN
cana-6072	22	3	for	for	ADP
cana-6072	22	4	this	this	DET
cana-6072	22	5	work	work	NOUN
cana-6072	22	6	stems	stem	VERB
cana-6072	22	7	from	from	ADP
cana-6072	22	8	the	the	DET
cana-6072	22	9	limitations	limitation	NOUN
cana-6072	22	10	of	of	ADP
cana-6072	22	11	text	text	NOUN
cana-6072	22	12	-	-	PUNCT
cana-6072	22	13	only	only	ADJ
cana-6072	22	14	models	model	NOUN
cana-6072	22	15	.	.	PUNCT
cana-6072	23	1	seminal	seminal	ADJ
cana-6072	23	2	research	research	NOUN
cana-6072	23	3	by	by	ADP
cana-6072	23	4	kralj	kralj	PROPN
cana-6072	23	5	novak	novak	PROPN
cana-6072	23	6	et	et	PROPN
cana-6072	23	7	al	al	PROPN
cana-6072	23	8	.	.	PROPN
cana-6072	23	9	constructed	construct	VERB
cana-6072	23	10	the	the	DET
cana-6072	23	11	first	first	ADJ
cana-6072	23	12	large	large	ADJ
cana-6072	23	13	-	-	PUNCT
cana-6072	23	14	scale	scale	NOUN
cana-6072	23	15	emoji	emoji	NOUN
cana-6072	23	16	sentiment	sentiment	NOUN
cana-6072	23	17	lexicon	lexicon	NOUN
cana-6072	23	18	,	,	PUNCT
cana-6072	23	19	revealing	reveal	VERB
cana-6072	23	20	that	that	SCONJ
cana-6072	23	21	emojis	emoji	NOUN
cana-6072	23	22	carry	carry	VERB
cana-6072	23	23	consistent	consistent	ADJ
cana-6072	23	24	sentiment	sentiment	NOUN
cana-6072	23	25	polarity	polarity	NOUN
cana-6072	23	26	across	across	ADP
cana-6072	23	27	languages	language	NOUN
cana-6072	23	28	,	,	PUNCT
cana-6072	23	29	establishing	establish	VERB
cana-6072	23	30	them	they	PRON
cana-6072	23	31	as	as	ADP
cana-6072	23	32	a	a	DET
cana-6072	23	33	legitimate	legitimate	ADJ
cana-6072	23	34	sentiment	sentiment	NOUN
cana-6072	23	35	resource	resource	NOUN
cana-6072	23	36	in	in	ADP
cana-6072	23	37	their	their	PRON
cana-6072	23	38	own	own	ADJ
cana-6072	23	39	right	right	NOUN
cana-6072	24	1	[	[	X
cana-6072	24	2	lou	lou	PROPN
cana-6072	24	3	,	,	PUNCT
cana-6072	24	4	y.	y.	PROPN
cana-6072	24	5	et	et	PROPN
cana-6072	24	6	al	al	PROPN
cana-6072	24	7	.	.	PROPN
cana-6072	24	8	,	,	PUNCT
cana-6072	24	9	2024	2024	NUM
cana-6072	24	10	]	]	PUNCT
cana-6072	24	11	.	.	PUNCT
cana-6072	25	1	this	this	PRON
cana-6072	25	2	was	be	AUX
cana-6072	25	3	followed	follow	VERB
cana-6072	25	4	by	by	ADP
cana-6072	25	5	the	the	DET
cana-6072	25	6	groundbreaking	groundbreake	VERB
cana-6072	25	7	deepmoji	deepmoji	ADJ
cana-6072	25	8	model	model	NOUN
cana-6072	26	1	[	[	X
cana-6072	26	2	miller	miller	PROPN
cana-6072	26	3	,	,	PUNCT
cana-6072	26	4	h.	h.	PROPN
cana-6072	26	5	et	et	PROPN
cana-6072	26	6	al	al	PROPN
cana-6072	26	7	.	.	PROPN
cana-6072	26	8	,	,	PUNCT
cana-6072	26	9	2016	2016	NUM
cana-6072	26	10	]	]	PUNCT
cana-6072	26	11	,	,	PUNCT
cana-6072	26	12	which	which	PRON
cana-6072	26	13	used	use	VERB
cana-6072	26	14	distant	distant	ADJ
cana-6072	26	15	supervision	supervision	NOUN
cana-6072	26	16	from	from	ADP
cana-6072	26	17	millions	million	NOUN
cana-6072	26	18	of	of	ADP
cana-6072	26	19	emoji	emoji	NOUN
cana-6072	26	20	-	-	PUNCT
cana-6072	26	21	annotated	annotate	VERB
cana-6072	26	22	tweets	tweet	NOUN
cana-6072	26	23	to	to	PART
cana-6072	26	24	learn	learn	VERB
cana-6072	26	25	rich	rich	ADJ
cana-6072	26	26	emotional	emotional	ADJ
cana-6072	26	27	representations	representation	NOUN
cana-6072	26	28	,	,	PUNCT
cana-6072	26	29	confirming	confirm	VERB
cana-6072	26	30	the	the	DET
cana-6072	26	31	power	power	NOUN
cana-6072	26	32	of	of	ADP
cana-6072	26	33	emojis	emoji	NOUN
cana-6072	26	34	as	as	ADP
cana-6072	26	35	sentiment	sentiment	NOUN
cana-6072	26	36	cues	cue	NOUN
cana-6072	26	37	and	and	CCONJ
cana-6072	26	38	enabling	enable	VERB
cana-6072	26	39	effective	effective	ADJ
cana-6072	26	40	transfer	transfer	NOUN
cana-6072	26	41	learning	learning	NOUN
cana-6072	26	42	across	across	ADP
cana-6072	26	43	domains	domain	NOUN
cana-6072	26	44	.	.	PUNCT
cana-6072	27	1	subsequent	subsequent	ADJ
cana-6072	27	2	research	research	NOUN
cana-6072	27	3	has	have	AUX
cana-6072	27	4	evolved	evolve	VERB
cana-6072	27	5	into	into	ADP
cana-6072	27	6	sophisticated	sophisticated	ADJ
cana-6072	27	7	deep	deep	ADJ
cana-6072	27	8	learning	learning	NOUN
cana-6072	27	9	architectures	architecture	NOUN
cana-6072	27	10	that	that	PRON
cana-6072	27	11	can	can	AUX
cana-6072	27	12	be	be	AUX
cana-6072	27	13	categorized	categorize	VERB
cana-6072	27	14	by	by	ADP
cana-6072	27	15	their	their	PRON
cana-6072	27	16	representation	representation	NOUN
cana-6072	27	17	and	and	CCONJ
cana-6072	27	18	fusion	fusion	NOUN
cana-6072	27	19	of	of	ADP
cana-6072	27	20	emoji	emoji	PROPN
cana-6072	27	21	information	information	NOUN
cana-6072	27	22	.	.	PUNCT
cana-6072	28	1	one	one	NUM
cana-6072	28	2	dominant	dominant	ADJ
cana-6072	28	3	approach	approach	NOUN
cana-6072	28	4	treats	treat	NOUN
cana-6072	28	5	emojis	emoji	NOUN
cana-6072	28	6	as	as	ADP
cana-6072	28	7	additional	additional	ADJ
cana-6072	28	8	textual	textual	ADJ
cana-6072	28	9	features	feature	NOUN
cana-6072	28	10	,	,	PUNCT
cana-6072	28	11	learning	learn	VERB
cana-6072	28	12	emoji	emoji	NOUN
cana-6072	28	13	embeddings	embedding	NOUN
cana-6072	28	14	analogous	analogous	ADJ
cana-6072	28	15	to	to	ADP
cana-6072	28	16	word	word	NOUN
cana-6072	28	17	embeddings	embedding	NOUN
cana-6072	28	18	.	.	PUNCT
cana-6072	29	1	the	the	DET
cana-6072	29	2	emoji2vec	emoji2vec	NOUN
cana-6072	29	3	model	model	NOUN
cana-6072	29	4	[	[	X
cana-6072	29	5	novak	novak	PROPN
cana-6072	29	6	et	et	PROPN
cana-6072	29	7	al	al	PROPN
cana-6072	29	8	.	.	PROPN
cana-6072	29	9	,	,	PUNCT
cana-6072	29	10	2015	2015	NUM
cana-6072	29	11	]	]	PUNCT
cana-6072	29	12	learned	learn	VERB
cana-6072	29	13	emoji	emoji	PROPN
cana-6072	29	14	vectors	vector	NOUN
cana-6072	29	15	from	from	ADP
cana-6072	29	16	their	their	PRON
cana-6072	29	17	unicode	unicode	ADJ
cana-6072	29	18	descriptions	description	NOUN
cana-6072	29	19	to	to	PART
cana-6072	29	20	address	address	VERB
cana-6072	29	21	the	the	DET
cana-6072	29	22	lack	lack	NOUN
cana-6072	29	23	of	of	ADP
cana-6072	29	24	emoji	emoji	PROPN
cana-6072	29	25	coverage	coverage	NOUN
cana-6072	29	26	in	in	ADP
cana-6072	29	27	standard	standard	ADJ
cana-6072	29	28	nlp	nlp	ADJ
cana-6072	29	29	embeddings	embedding	NOUN
cana-6072	29	30	like	like	ADP
cana-6072	29	31	word2vec	word2vec	PROPN
cana-6072	29	32	.	.	PUNCT
cana-6072	30	1	a	a	DET
cana-6072	30	2	significant	significant	ADJ
cana-6072	30	3	advancement	advancement	NOUN
cana-6072	30	4	was	be	AUX
cana-6072	30	5	the	the	DET
cana-6072	30	6	development	development	NOUN
cana-6072	30	7	of	of	ADP
cana-6072	30	8	bi	bi	ADJ
cana-6072	30	9	-	-	ADJ
cana-6072	30	10	sense	sense	ADJ
cana-6072	30	11	emoji	emoji	NOUN
cana-6072	30	12	embeddings	embedding	NOUN
cana-6072	30	13	,	,	PUNCT
cana-6072	30	14	which	which	PRON
cana-6072	30	15	capture	capture	VERB
cana-6072	30	16	distinct	distinct	ADJ
cana-6072	30	17	positive	positive	ADJ
cana-6072	30	18	and	and	CCONJ
cana-6072	30	19	negative	negative	ADJ
cana-6072	30	20	usages	usage	NOUN
cana-6072	30	21	of	of	ADP
cana-6072	30	22	the	the	DET
cana-6072	30	23	same	same	ADJ
cana-6072	30	24	emoji	emoji	NOUN
cana-6072	30	25	(	(	PUNCT
cana-6072	30	26	e.g.	e.g.	ADV
cana-6072	30	27	,	,	PUNCT
cana-6072	30	28	a	a	DET
cana-6072	30	29	face	face	NOUN
cana-6072	30	30	versus	versus	ADP
cana-6072	30	31	a	a	DET
cana-6072	30	32	face	face	NOUN
cana-6072	30	33	)	)	PUNCT
cana-6072	31	1	[	[	X
cana-6072	31	2	lou	lou	X
cana-6072	31	3	,	,	PUNCT
cana-6072	31	4	y.	y.	PROPN
cana-6072	31	5	,	,	PUNCT
cana-6072	31	6	et	et	PROPN
cana-6072	31	7	al	al	PROPN
cana-6072	31	8	.	.	PROPN
cana-6072	31	9	,	,	PUNCT
cana-6072	31	10	2024	2024	NUM
cana-6072	31	11	]	]	PUNCT
cana-6072	31	12	.	.	PUNCT
cana-6072	32	1	these	these	DET
cana-6072	32	2	embeddings	embedding	NOUN
cana-6072	32	3	are	be	AUX
cana-6072	32	4	often	often	ADV
cana-6072	32	5	integrated	integrate	VERB
cana-6072	32	6	via	via	ADP
cana-6072	32	7	attention	attention	NOUN
cana-6072	32	8	-	-	PUNCT
cana-6072	32	9	based	base	VERB
cana-6072	32	10	lstm	lstm	NOUN
cana-6072	32	11	networks	network	NOUN
cana-6072	32	12	,	,	PUNCT
cana-6072	32	13	where	where	SCONJ
cana-6072	32	14	the	the	DET
cana-6072	32	15	emoji	emoji	PROPN
cana-6072	32	16	context	context	NOUN
cana-6072	32	17	guides	guide	VERB
cana-6072	32	18	the	the	DET
cana-6072	32	19	attention	attention	NOUN
cana-6072	32	20	mechanism	mechanism	NOUN
cana-6072	32	21	to	to	PART
cana-6072	32	22	weight	weight	VERB
cana-6072	32	23	the	the	DET
cana-6072	32	24	importance	importance	NOUN
cana-6072	32	25	of	of	ADP
cana-6072	32	26	specific	specific	ADJ
cana-6072	32	27	words	word	NOUN
cana-6072	32	28	,	,	PUNCT
cana-6072	32	29	significantly	significantly	ADV
cana-6072	32	30	outperforming	outperform	VERB
cana-6072	32	31	text	text	NOUN
cana-6072	32	32	-	-	PUNCT
cana-6072	32	33	only	only	ADV
cana-6072	32	34	baselines	baseline	NOUN
cana-6072	33	1	[	[	X
cana-6072	33	2	lou	lou	PROPN
cana-6072	33	3	,	,	PUNCT
cana-6072	33	4	y.	y.	PROPN
cana-6072	33	5	,	,	PUNCT
cana-6072	33	6	et	et	PROPN
cana-6072	33	7	al	al	PROPN
cana-6072	33	8	.	.	PROPN
cana-6072	33	9	,	,	PUNCT
cana-6072	33	10	2024	2024	NUM
cana-6072	33	11	]	]	PUNCT
cana-6072	33	12	,	,	PUNCT
cana-6072	33	13	[	[	X
cana-6072	33	14	khan	khan	PROPN
cana-6072	33	15	,	,	PUNCT
cana-6072	33	16	a.	a.	PROPN
cana-6072	33	17	et	et	PROPN
cana-6072	33	18	al	al	PROPN
cana-6072	33	19	.	.	PROPN
cana-6072	33	20	,	,	PUNCT
cana-6072	33	21	2025	2025	NUM
cana-6072	33	22	]	]	PUNCT
cana-6072	33	23	.	.	PUNCT
cana-6072	34	1	studies	study	NOUN
cana-6072	34	2	consistently	consistently	ADV
cana-6072	34	3	show	show	VERB
cana-6072	34	4	that	that	SCONJ
cana-6072	34	5	such	such	ADJ
cana-6072	34	6	emoji	emoji	NOUN
cana-6072	34	7	-	-	PUNCT
cana-6072	34	8	enhanced	enhance	VERB
cana-6072	34	9	models	model	NOUN
cana-6072	34	10	outperform	outperform	VERB
cana-6072	34	11	their	their	PRON
cana-6072	34	12	emoji	emoji	ADJ
cana-6072	34	13	-	-	PUNCT
cana-6072	34	14	blind	blind	ADJ
cana-6072	34	15	counterparts	counterpart	NOUN
cana-6072	34	16	by	by	ADP
cana-6072	34	17	2–10	2–10	NOUN
cana-6072	34	18	%	%	NOUN
cana-6072	34	19	in	in	ADP
cana-6072	34	20	accuracy	accuracy	NOUN
cana-6072	35	1	[	[	X
cana-6072	35	2	khan	khan	PROPN
cana-6072	35	3	,	,	PUNCT
cana-6072	35	4	a.	a.	PROPN
cana-6072	35	5	et	et	PROPN
cana-6072	35	6	al	al	PROPN
cana-6072	35	7	.	.	PROPN
cana-6072	35	8	,	,	PUNCT
cana-6072	35	9	2025	2025	NUM
cana-6072	35	10	]	]	PUNCT
cana-6072	35	11	,	,	PUNCT
cana-6072	36	1	[	[	X
cana-6072	36	2	lou	lou	PROPN
cana-6072	36	3	,	,	PUNCT
cana-6072	36	4	y.	y.	PROPN
cana-6072	36	5	,	,	PUNCT
cana-6072	36	6	et	et	PROPN
cana-6072	36	7	al	al	PROPN
cana-6072	36	8	.	.	PROPN
cana-6072	36	9	,	,	PUNCT
cana-6072	36	10	2024	2024	NUM
cana-6072	36	11	]	]	PUNCT
cana-6072	36	12	,	,	PUNCT
cana-6072	36	13	with	with	ADP
cana-6072	36	14	recent	recent	ADJ
cana-6072	36	15	bert	bert	NOUN
cana-6072	36	16	-	-	PUNCT
cana-6072	36	17	based	base	VERB
cana-6072	36	18	classifiers	classifier	NOUN
cana-6072	36	19	also	also	ADV
cana-6072	36	20	seeing	see	VERB
cana-6072	36	21	substantial	substantial	ADJ
cana-6072	36	22	gains	gain	NOUN
cana-6072	36	23	when	when	SCONJ
cana-6072	36	24	explicitly	explicitly	ADV
cana-6072	36	25	handling	handle	VERB
cana-6072	36	26	emojis	emoji	NOUN
cana-6072	37	1	[	[	X
cana-6072	37	2	chen	chen	PROPN
cana-6072	37	3	,	,	PUNCT
cana-6072	37	4	y.	y.	PROPN
cana-6072	37	5	et	et	PROPN
cana-6072	37	6	al	al	PROPN
cana-6072	37	7	.	.	PROPN
cana-6072	37	8	,	,	PUNCT
cana-6072	37	9	2018	2018	NUM
cana-6072	37	10	]	]	PUNCT
cana-6072	37	11	.	.	PUNCT
cana-6072	38	1	a	a	DET
cana-6072	38	2	more	more	ADV
cana-6072	38	3	recent	recent	ADJ
cana-6072	38	4	and	and	CCONJ
cana-6072	38	5	complex	complex	ADJ
cana-6072	38	6	avenue	avenue	NOUN
cana-6072	38	7	of	of	ADP
cana-6072	38	8	research	research	NOUN
cana-6072	38	9	considers	consider	VERB
cana-6072	38	10	the	the	DET
cana-6072	38	11	visual	visual	ADJ
cana-6072	38	12	modality	modality	NOUN
cana-6072	38	13	of	of	ADP
cana-6072	38	14	emojis	emoji	NOUN
cana-6072	38	15	,	,	PUNCT
cana-6072	38	16	treating	treat	VERB
cana-6072	38	17	them	they	PRON
cana-6072	38	18	as	as	ADP
cana-6072	38	19	images	image	NOUN
cana-6072	38	20	to	to	PART
cana-6072	38	21	extract	extract	VERB
cana-6072	38	22	features	feature	NOUN
cana-6072	38	23	from	from	ADP
cana-6072	38	24	their	their	PRON
cana-6072	38	25	graphic	graphic	ADJ
cana-6072	38	26	rendering	rendering	NOUN
cana-6072	38	27	.	.	PUNCT
cana-6072	39	1	models	model	NOUN
cana-6072	39	2	like	like	ADP
cana-6072	39	3	the	the	DET
cana-6072	39	4	attentionbased	attentionbase	VERB
cana-6072	39	5	modality	modality	NOUN
cana-6072	39	6	-	-	PUNCT
cana-6072	39	7	gated	gate	VERB
cana-6072	39	8	networks	network	NOUN
cana-6072	39	9	(	(	PUNCT
cana-6072	39	10	amgn	amgn	NOUN
cana-6072	39	11	)	)	PUNCT
cana-6072	40	1	[	[	X
cana-6072	40	2	lou	lou	X
cana-6072	40	3	,	,	PUNCT
cana-6072	40	4	y.	y.	PROPN
cana-6072	40	5	,	,	PUNCT
cana-6072	40	6	et	et	PROPN
cana-6072	40	7	al	al	PROPN
cana-6072	40	8	.	.	PROPN
cana-6072	40	9	,	,	PUNCT
cana-6072	40	10	2024	2024	NUM
cana-6072	40	11	]	]	PUNCT
cana-6072	40	12	and	and	CCONJ
cana-6072	40	13	the	the	DET
cana-6072	40	14	mutual	mutual	ADJ
cana-6072	40	15	attention	attention	NOUN
cana-6072	40	16	mechanism	mechanism	NOUN
cana-6072	40	17	-	-	PUNCT
cana-6072	40	18	based	base	VERB
cana-6072	40	19	emoji	emoji	NOUN
cana-6072	40	20	multimodal	multimodal	ADJ
cana-6072	40	21	microblog	microblog	NOUN
cana-6072	40	22	sentiment	sentiment	NOUN
cana-6072	40	23	analysis	analysis	NOUN
cana-6072	40	24	(	(	PUNCT
cana-6072	40	25	mam	mam	NOUN
cana-6072	40	26	-	-	PUNCT
cana-6072	40	27	emmsa	emmsa	NOUN
cana-6072	40	28	)	)	PUNCT
cana-6072	41	1	[	[	X
cana-6072	41	2	6	6	NUM
cana-6072	41	3	]	]	PUNCT
cana-6072	41	4	use	use	NOUN
cana-6072	41	5	cnns	cnn	NOUN
cana-6072	41	6	or	or	CCONJ
cana-6072	41	7	transformers	transformer	NOUN
cana-6072	41	8	to	to	PART
cana-6072	41	9	encode	encode	VERB
cana-6072	41	10	the	the	DET
cana-6072	41	11	emoji	emoji	PROPN
cana-6072	41	12	icon	icon	NOUN
cana-6072	41	13	and	and	CCONJ
cana-6072	41	14	employ	employ	VERB
cana-6072	41	15	attention	attention	NOUN
cana-6072	41	16	-	-	PUNCT
cana-6072	41	17	based	base	VERB
cana-6072	41	18	fusion	fusion	NOUN
cana-6072	41	19	to	to	PART
cana-6072	41	20	learn	learn	VERB
cana-6072	41	21	the	the	DET
cana-6072	41	22	alignment	alignment	NOUN
cana-6072	41	23	between	between	ADP
cana-6072	41	24	visual	visual	ADJ
cana-6072	41	25	and	and	CCONJ
cana-6072	41	26	textual	textual	ADJ
cana-6072	41	27	features	feature	NOUN
cana-6072	41	28	.	.	PUNCT
cana-6072	42	1	this	this	DET
cana-6072	42	2	approach	approach	NOUN
cana-6072	42	3	captures	capture	VERB
cana-6072	42	4	sentiment	sentiment	NOUN
cana-6072	42	5	expressed	express	VERB
cana-6072	42	6	through	through	ADP
cana-6072	42	7	appearance	appearance	NOUN
cana-6072	42	8	(	(	PUNCT
cana-6072	42	9	e.g.	e.g.	ADV
cana-6072	42	10	,	,	PUNCT
cana-6072	42	11	a	a	DET
cana-6072	42	12	face	face	NOUN
cana-6072	42	13	versus	versus	ADP
cana-6072	42	14	a	a	DET
cana-6072	42	15	face	face	NOUN
cana-6072	42	16	)	)	PUNCT
cana-6072	42	17	that	that	SCONJ
cana-6072	42	18	pure	pure	ADJ
cana-6072	42	19	text	text	NOUN
cana-6072	42	20	embeddings	embedding	NOUN
cana-6072	42	21	might	might	AUX
cana-6072	42	22	miss	miss	VERB
cana-6072	42	23	,	,	PUNCT
cana-6072	42	24	yielding	yield	VERB
cana-6072	42	25	accuracy	accuracy	NOUN
cana-6072	42	26	gains	gain	NOUN
cana-6072	42	27	of	of	ADP
cana-6072	42	28	1	1	NUM
cana-6072	42	29	-	-	SYM
cana-6072	42	30	3	3	NUM
cana-6072	42	31	%	%	NOUN
cana-6072	43	1	[	[	X
cana-6072	43	2	xu	xu	PROPN
cana-6072	43	3	,	,	PUNCT
cana-6072	43	4	q.	q.	PROPN
cana-6072	43	5	a.	a.	PROPN
cana-6072	43	6	et	et	PROPN
cana-6072	43	7	al	al	PROPN
cana-6072	43	8	.	.	PROPN
cana-6072	43	9	,	,	PUNCT
cana-6072	43	10	2024	2024	NUM
cana-6072	43	11	]	]	PUNCT
cana-6072	43	12	.	.	PUNCT
cana-6072	44	1	these	these	DET
cana-6072	44	2	multimodal	multimodal	ADJ
cana-6072	44	3	fusion	fusion	NOUN
cana-6072	44	4	networks	network	NOUN
cana-6072	44	5	represent	represent	VERB
cana-6072	44	6	the	the	DET
cana-6072	44	7	cutting	cut	VERB
cana-6072	44	8	edge	edge	NOUN
cana-6072	44	9	,	,	PUNCT
cana-6072	44	10	utilizing	utilize	VERB
cana-6072	44	11	multi	multi	ADJ
cana-6072	44	12	-	-	ADJ
cana-6072	44	13	head	head	ADJ
cana-6072	44	14	self	self	NOUN
cana-6072	44	15	-	-	PUNCT
cana-6072	44	16	attention	attention	NOUN
cana-6072	44	17	mechanisms	mechanism	NOUN
cana-6072	44	18	to	to	PART
cana-6072	44	19	minimize	minimize	VERB
cana-6072	44	20	noise	noise	NOUN
cana-6072	44	21	interference	interference	NOUN
cana-6072	44	22	and	and	CCONJ
cana-6072	44	23	exploit	exploit	VERB
cana-6072	44	24	fine	fine	ADV
cana-6072	44	25	-	-	PUNCT
cana-6072	44	26	grained	grain	VERB
cana-6072	44	27	correlations	correlation	NOUN
cana-6072	44	28	between	between	ADP
cana-6072	44	29	heterogeneous	heterogeneous	ADJ
cana-6072	44	30	data	datum	NOUN
cana-6072	44	31	modalities	modality	NOUN
cana-6072	44	32	[	[	X
cana-6072	44	33	lou	lou	PROPN
cana-6072	44	34	,	,	PUNCT
cana-6072	44	35	y.	y.	PROPN
cana-6072	44	36	,	,	PUNCT
cana-6072	44	37	et	et	PROPN
cana-6072	44	38	al	al	PROPN
cana-6072	44	39	.	.	PROPN
cana-6072	44	40	,	,	PUNCT
cana-6072	44	41	2020	2020	NUM
cana-6072	44	42	]	]	PUNCT
cana-6072	44	43	.	.	PUNCT
cana-6072	45	1	communications	communication	NOUN
cana-6072	45	2	on	on	ADP
cana-6072	45	3	applied	apply	VERB
cana-6072	45	4	nonlinear	nonlinear	ADJ
cana-6072	45	5	analysis	analysis	NOUN
cana-6072	45	6	issn	issn	NOUN
cana-6072	45	7	:	:	PUNCT
cana-6072	45	8	1074	1074	NUM
cana-6072	45	9	-	-	PUNCT
cana-6072	45	10	133x	133x	NUM
cana-6072	45	11	vol	vol	NOUN
cana-6072	45	12	32	32	NUM
cana-6072	45	13	no	no	NOUN
cana-6072	45	14	.	.	PUNCT
cana-6072	46	1	4s	4s	NUM
cana-6072	46	2	(	(	PUNCT
cana-6072	46	3	2025	2025	NUM
cana-6072	46	4	)	)	PUNCT
cana-6072	46	5	759	759	NUM
cana-6072	46	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	46	7	despite	despite	SCONJ
cana-6072	46	8	promising	promising	ADJ
cana-6072	46	9	progress	progress	NOUN
cana-6072	46	10	,	,	PUNCT
cana-6072	46	11	several	several	ADJ
cana-6072	46	12	challenges	challenge	NOUN
cana-6072	46	13	persist	persist	VERB
cana-6072	46	14	.	.	PUNCT
cana-6072	47	1	emoji	emoji	PROPN
cana-6072	47	2	polysemy	polysemy	PROPN
cana-6072	47	3	and	and	CCONJ
cana-6072	47	4	context	context	NOUN
cana-6072	47	5	dependence	dependence	NOUN
cana-6072	47	6	(	(	PUNCT
cana-6072	47	7	e.g.	e.g.	ADV
cana-6072	47	8	,	,	PUNCT
cana-6072	47	9	or	or	CCONJ
cana-6072	47	10	for	for	ADP
cana-6072	47	11	gratitude	gratitude	NOUN
cana-6072	47	12	or	or	CCONJ
cana-6072	47	13	prayer	prayer	NOUN
cana-6072	47	14	)	)	PUNCT
cana-6072	47	15	remain	remain	VERB
cana-6072	47	16	a	a	DET
cana-6072	47	17	significant	significant	ADJ
cana-6072	47	18	hurdle	hurdle	NOUN
cana-6072	47	19	,	,	PUNCT
cana-6072	47	20	though	though	SCONJ
cana-6072	47	21	bisense	bisense	NOUN
cana-6072	47	22	embeddings	embedding	NOUN
cana-6072	47	23	aim	aim	VERB
cana-6072	47	24	to	to	PART
cana-6072	47	25	address	address	VERB
cana-6072	47	26	this	this	PRON
cana-6072	47	27	[	[	X
cana-6072	47	28	lou	lou	PROPN
cana-6072	47	29	,	,	PUNCT
cana-6072	47	30	y.	y.	PROPN
cana-6072	47	31	,	,	PUNCT
cana-6072	47	32	et	et	PROPN
cana-6072	47	33	al	al	PROPN
cana-6072	47	34	.	.	PROPN
cana-6072	47	35	,	,	PUNCT
cana-6072	47	36	2024	2024	NUM
cana-6072	47	37	]	]	PUNCT
cana-6072	47	38	.	.	PUNCT
cana-6072	48	1	furthermore	furthermore	ADV
cana-6072	48	2	,	,	PUNCT
cana-6072	48	3	cultural	cultural	ADJ
cana-6072	48	4	variance	variance	NOUN
cana-6072	48	5	and	and	CCONJ
cana-6072	48	6	platform	platform	NOUN
cana-6072	48	7	-	-	PUNCT
cana-6072	48	8	dependent	dependent	ADJ
cana-6072	48	9	rendering	rendering	NOUN
cana-6072	48	10	of	of	ADP
cana-6072	48	11	emojis	emojis	NOUN
cana-6072	48	12	can	can	AUX
cana-6072	48	13	lead	lead	VERB
cana-6072	48	14	to	to	ADP
cana-6072	48	15	misunderstandings	misunderstanding	NOUN
cana-6072	48	16	[	[	X
cana-6072	48	17	liu	liu	PROPN
cana-6072	48	18	,	,	PUNCT
cana-6072	48	19	c.	c.	PROPN
cana-6072	48	20	,	,	PUNCT
cana-6072	48	21	et	et	PROPN
cana-6072	48	22	al	al	PROPN
cana-6072	48	23	.	.	PROPN
cana-6072	48	24	,	,	PUNCT
cana-6072	48	25	2021	2021	NUM
cana-6072	48	26	]	]	PUNCT
cana-6072	48	27	,	,	PUNCT
cana-6072	48	28	and	and	CCONJ
cana-6072	48	29	sarcasm	sarcasm	NOUN
cana-6072	48	30	signalled	signal	VERB
cana-6072	48	31	by	by	ADP
cana-6072	48	32	emojis	emoji	NOUN
cana-6072	48	33	(	(	PUNCT
cana-6072	48	34	e.g.	e.g.	ADV
cana-6072	48	35	,	,	PUNCT
cana-6072	48	36	or	or	CCONJ
cana-6072	48	37	)	)	PUNCT
cana-6072	48	38	is	be	AUX
cana-6072	48	39	difficult	difficult	ADJ
cana-6072	48	40	to	to	PART
cana-6072	48	41	detect	detect	VERB
cana-6072	48	42	reliably	reliably	ADV
cana-6072	48	43	.	.	PUNCT
cana-6072	49	1	practical	practical	ADJ
cana-6072	49	2	issues	issue	NOUN
cana-6072	49	3	like	like	ADP
cana-6072	49	4	class	class	NOUN
cana-6072	49	5	imbalance	imbalance	NOUN
cana-6072	49	6	in	in	ADP
cana-6072	49	7	datasets	dataset	NOUN
cana-6072	49	8	and	and	CCONJ
cana-6072	49	9	the	the	DET
cana-6072	49	10	computational	computational	ADJ
cana-6072	49	11	overhead	overhead	NOUN
cana-6072	49	12	of	of	ADP
cana-6072	49	13	complex	complex	ADJ
cana-6072	49	14	fusion	fusion	NOUN
cana-6072	49	15	networks	network	NOUN
cana-6072	49	16	also	also	ADV
cana-6072	49	17	pose	pose	VERB
cana-6072	49	18	obstacles	obstacle	NOUN
cana-6072	49	19	[	[	X
cana-6072	49	20	khan	khan	PROPN
cana-6072	49	21	,	,	PUNCT
cana-6072	49	22	a.	a.	PROPN
cana-6072	49	23	et	et	PROPN
cana-6072	49	24	al	al	PROPN
cana-6072	49	25	.	.	PROPN
cana-6072	49	26	,	,	PUNCT
cana-6072	49	27	2025	2025	NUM
cana-6072	49	28	]	]	PUNCT
cana-6072	49	29	,	,	PUNCT
cana-6072	50	1	[	[	X
cana-6072	50	2	lou	lou	PROPN
cana-6072	50	3	,	,	PUNCT
cana-6072	50	4	y.	y.	PROPN
cana-6072	50	5	,	,	PUNCT
cana-6072	50	6	et	et	PROPN
cana-6072	50	7	al	al	PROPN
cana-6072	50	8	.	.	PROPN
cana-6072	50	9	,	,	PUNCT
cana-6072	50	10	2020	2020	NUM
cana-6072	50	11	]	]	PUNCT
cana-6072	50	12	.	.	PUNCT
cana-6072	51	1	finally	finally	ADV
cana-6072	51	2	,	,	PUNCT
cana-6072	51	3	there	there	PRON
cana-6072	51	4	is	be	VERB
cana-6072	51	5	a	a	DET
cana-6072	51	6	growing	grow	VERB
cana-6072	51	7	need	need	NOUN
cana-6072	51	8	for	for	ADP
cana-6072	51	9	explainable	explainable	ADJ
cana-6072	51	10	ai	ai	NOUN
cana-6072	51	11	(	(	PUNCT
cana-6072	51	12	xai	xai	PROPN
cana-6072	51	13	)	)	PUNCT
cana-6072	51	14	in	in	ADP
cana-6072	51	15	high	high	ADJ
cana-6072	51	16	-	-	PUNCT
cana-6072	51	17	stakes	stake	NOUN
cana-6072	51	18	applications	application	NOUN
cana-6072	51	19	to	to	PART
cana-6072	51	20	ensure	ensure	VERB
cana-6072	51	21	models	model	NOUN
cana-6072	51	22	remain	remain	VERB
cana-6072	51	23	interpretable	interpretable	ADJ
cana-6072	52	1	[	[	X
cana-6072	52	2	lou	lou	PROPN
cana-6072	52	3	,	,	PUNCT
cana-6072	52	4	y.	y.	PROPN
cana-6072	52	5	,	,	PUNCT
cana-6072	52	6	et	et	PROPN
cana-6072	52	7	al	al	PROPN
cana-6072	52	8	.	.	PROPN
cana-6072	52	9	,	,	PUNCT
cana-6072	52	10	2020	2020	NUM
cana-6072	52	11	]	]	PUNCT
cana-6072	52	12	.	.	PUNCT
cana-6072	53	1	this	this	DET
cana-6072	53	2	slr	slr	NOUN
cana-6072	53	3	investigates	investigate	VERB
cana-6072	53	4	the	the	DET
cana-6072	53	5	state	state	NOUN
cana-6072	53	6	-	-	PUNCT
cana-6072	53	7	of	of	ADP
cana-6072	53	8	-	-	PUNCT
cana-6072	53	9	the	the	DET
cana-6072	53	10	-	-	PUNCT
cana-6072	53	11	art	art	NOUN
cana-6072	53	12	to	to	PART
cana-6072	53	13	answer	answer	VERB
cana-6072	53	14	key	key	ADJ
cana-6072	53	15	research	research	NOUN
cana-6072	53	16	questions	question	NOUN
cana-6072	53	17	on	on	ADP
cana-6072	53	18	model	model	NOUN
cana-6072	53	19	architectures	architecture	NOUN
cana-6072	53	20	,	,	PUNCT
cana-6072	53	21	representation	representation	NOUN
cana-6072	53	22	strategies	strategy	NOUN
cana-6072	53	23	,	,	PUNCT
cana-6072	53	24	performance	performance	NOUN
cana-6072	53	25	gains	gain	NOUN
cana-6072	53	26	,	,	PUNCT
cana-6072	53	27	and	and	CCONJ
cana-6072	53	28	persistent	persistent	ADJ
cana-6072	53	29	limitations	limitation	NOUN
cana-6072	53	30	.	.	PUNCT
cana-6072	54	1	by	by	ADP
cana-6072	54	2	synthesizing	synthesize	VERB
cana-6072	54	3	findings	finding	NOUN
cana-6072	54	4	from	from	ADP
cana-6072	54	5	2015	2015	NUM
cana-6072	54	6	to	to	ADP
cana-6072	54	7	2025	2025	NUM
cana-6072	54	8	,	,	PUNCT
cana-6072	54	9	this	this	DET
cana-6072	54	10	review	review	NOUN
cana-6072	54	11	provides	provide	VERB
cana-6072	54	12	a	a	DET
cana-6072	54	13	comprehensive	comprehensive	ADJ
cana-6072	54	14	overview	overview	NOUN
cana-6072	54	15	of	of	ADP
cana-6072	54	16	how	how	SCONJ
cana-6072	54	17	integrating	integrating	ADJ
cana-6072	54	18	emojis	emojis	NOUN
cana-6072	54	19	has	have	AUX
cana-6072	54	20	revolutionized	revolutionize	VERB
cana-6072	54	21	sentiment	sentiment	NOUN
cana-6072	54	22	analysis	analysis	NOUN
cana-6072	54	23	,	,	PUNCT
cana-6072	54	24	outlining	outline	VERB
cana-6072	54	25	best	good	ADJ
cana-6072	54	26	practices	practice	NOUN
cana-6072	54	27	and	and	CCONJ
cana-6072	54	28	charting	chart	VERB
cana-6072	54	29	a	a	DET
cana-6072	54	30	course	course	NOUN
cana-6072	54	31	for	for	ADP
cana-6072	54	32	future	future	ADJ
cana-6072	54	33	research	research	NOUN
cana-6072	54	34	.	.	PUNCT
cana-6072	55	1	figure	figure	NOUN
cana-6072	55	2	1	1	NUM
cana-6072	55	3	.	.	PUNCT
cana-6072	55	4	timeline	timeline	NOUN
cana-6072	55	5	of	of	ADP
cana-6072	55	6	emoji	emoji	NOUN
cana-6072	55	7	-	-	PUNCT
cana-6072	55	8	aware	aware	ADJ
cana-6072	55	9	sentiment	sentiment	NOUN
cana-6072	55	10	research	research	NOUN
cana-6072	55	11	(	(	PUNCT
cana-6072	55	12	1997–2025	1997–2025	NUM
cana-6072	55	13	)	)	PUNCT
cana-6072	55	14	.	.	PUNCT
cana-6072	56	1	methods	method	NOUN
cana-6072	56	2	and	and	CCONJ
cana-6072	56	3	tools	tool	NOUN
cana-6072	56	4	prisma	prisma	NOUN
cana-6072	56	5	protocol	protocol	NOUN
cana-6072	56	6	:	:	PUNCT
cana-6072	56	7	we	we	PRON
cana-6072	56	8	followed	follow	VERB
cana-6072	56	9	the	the	DET
cana-6072	56	10	prisma	prisma	PROPN
cana-6072	56	11	2020	2020	NUM
cana-6072	56	12	guidelines	guideline	NOUN
cana-6072	56	13	to	to	PART
cana-6072	56	14	ensure	ensure	VERB
cana-6072	56	15	transparent	transparent	ADJ
cana-6072	56	16	identification	identification	NOUN
cana-6072	56	17	,	,	PUNCT
cana-6072	56	18	screening	screening	NOUN
cana-6072	56	19	,	,	PUNCT
cana-6072	56	20	eligibility	eligibility	NOUN
cana-6072	56	21	assessment	assessment	NOUN
cana-6072	56	22	,	,	PUNCT
cana-6072	56	23	and	and	CCONJ
cana-6072	56	24	inclusion	inclusion	NOUN
cana-6072	56	25	.	.	PUNCT
cana-6072	57	1	information	information	NOUN
cana-6072	57	2	sources	source	NOUN
cana-6072	57	3	:	:	PUNCT
cana-6072	57	4	peer	peer	NOUN
cana-6072	57	5	-	-	PUNCT
cana-6072	57	6	reviewed	review	VERB
cana-6072	57	7	venues	venue	NOUN
cana-6072	57	8	indexed	index	VERB
cana-6072	57	9	in	in	ADP
cana-6072	57	10	scopus	scopus	PROPN
cana-6072	57	11	,	,	PUNCT
cana-6072	57	12	web	web	NOUN
cana-6072	57	13	of	of	ADP
cana-6072	57	14	science	science	NOUN
cana-6072	57	15	,	,	PUNCT
cana-6072	57	16	ieee	ieee	PROPN
cana-6072	57	17	xplore	xplore	PROPN
cana-6072	57	18	,	,	PUNCT
cana-6072	57	19	acm	acm	PROPN
cana-6072	57	20	digital	digital	PROPN
cana-6072	57	21	library	library	NOUN
cana-6072	57	22	,	,	PUNCT
cana-6072	57	23	acl	acl	PROPN
cana-6072	57	24	anthology	anthology	NOUN
cana-6072	57	25	,	,	PUNCT
cana-6072	57	26	and	and	CCONJ
cana-6072	57	27	google	google	PROPN
cana-6072	57	28	scholar	scholar	NOUN
cana-6072	57	29	;	;	PUNCT
cana-6072	57	30	plus	plus	CCONJ
cana-6072	57	31	hand	hand	NOUN
cana-6072	57	32	-	-	PUNCT
cana-6072	57	33	search	search	NOUN
cana-6072	57	34	(	(	PUNCT
cana-6072	57	35	backward	backward	ADV
cana-6072	57	36	/	/	SYM
cana-6072	57	37	forward	forward	ADJ
cana-6072	57	38	snowballing	snowballing	NOUN
cana-6072	57	39	)	)	PUNCT
cana-6072	57	40	.	.	PUNCT
cana-6072	58	1	search	search	NOUN
cana-6072	58	2	strategy	strategy	NOUN
cana-6072	58	3	:	:	PUNCT
cana-6072	58	4	boolean	boolean	ADJ
cana-6072	58	5	strings	string	NOUN
cana-6072	58	6	combining	combine	VERB
cana-6072	58	7	sentiment	sentiment	NOUN
cana-6072	58	8	terms	term	NOUN
cana-6072	58	9	,	,	PUNCT
cana-6072	58	10	emoji	emoji	PROPN
cana-6072	58	11	terms	term	NOUN
cana-6072	58	12	,	,	PUNCT
cana-6072	58	13	and	and	CCONJ
cana-6072	58	14	microblog	microblog	NOUN
cana-6072	58	15	platforms	platform	NOUN
cana-6072	58	16	,	,	PUNCT
cana-6072	58	17	e.g.	e.g.	ADV
cana-6072	58	18	,	,	PUNCT
cana-6072	58	19	(	(	PUNCT
cana-6072	58	20	sentiment	sentiment	NOUN
cana-6072	58	21	or	or	CCONJ
cana-6072	58	22	polarity	polarity	NOUN
cana-6072	58	23	or	or	CCONJ
cana-6072	58	24	“	"	PUNCT
cana-6072	58	25	sentiment	sentiment	NOUN
cana-6072	58	26	classification	classification	NOUN
cana-6072	58	27	”	"	PUNCT
cana-6072	58	28	)	)	PUNCT
cana-6072	58	29	and	and	CCONJ
cana-6072	58	30	(	(	PUNCT
cana-6072	58	31	emoji	emoji	PROPN
cana-6072	58	32	or	or	CCONJ
cana-6072	58	33	emoticon	emoticon	PROPN
cana-6072	58	34	or	or	CCONJ
cana-6072	58	35	“	"	PUNCT
cana-6072	58	36	emoji	emoji	PROPN
cana-6072	58	37	embedding	embedding	NOUN
cana-6072	58	38	”	"	PUNCT
cana-6072	58	39	or	or	CCONJ
cana-6072	58	40	“	"	PUNCT
cana-6072	58	41	emoji	emoji	X
cana-6072	58	42	lexicon	lexicon	NOUN
cana-6072	58	43	”	"	PUNCT
cana-6072	58	44	)	)	PUNCT
cana-6072	58	45	and	and	CCONJ
cana-6072	58	46	(	(	PUNCT
cana-6072	58	47	twitter	twitter	NOUN
cana-6072	58	48	or	or	CCONJ
cana-6072	58	49	x	x	NOUN
cana-6072	58	50	or	or	CCONJ
cana-6072	58	51	microblog	microblog	NOUN
cana-6072	58	52	)	)	PUNCT
cana-6072	58	53	.	.	PUNCT
cana-6072	59	1	database	database	NOUN
cana-6072	59	2	-	-	PUNCT
cana-6072	59	3	specific	specific	ADJ
cana-6072	59	4	variants	variant	NOUN
cana-6072	59	5	(	(	PUNCT
cana-6072	59	6	field	field	NOUN
cana-6072	59	7	filters	filter	NOUN
cana-6072	59	8	,	,	PUNCT
cana-6072	59	9	metadata	metadata	ADJ
cana-6072	59	10	constraints	constraint	NOUN
cana-6072	59	11	)	)	PUNCT
cana-6072	59	12	were	be	AUX
cana-6072	59	13	applied	apply	VERB
cana-6072	59	14	.	.	PUNCT
cana-6072	60	1	searches	search	NOUN
cana-6072	60	2	were	be	AUX
cana-6072	60	3	limited	limit	VERB
cana-6072	60	4	to	to	ADP
cana-6072	60	5	2015–2025	2015–2025	NUM
cana-6072	60	6	;	;	PUNCT
cana-6072	60	7	article	article	NOUN
cana-6072	60	8	language	language	NOUN
cana-6072	60	9	was	be	AUX
cana-6072	60	10	english	english	PROPN
cana-6072	60	11	(	(	PUNCT
cana-6072	60	12	dataset	dataset	NOUN
cana-6072	60	13	language	language	NOUN
cana-6072	60	14	unrestricted	unrestricted	ADJ
cana-6072	60	15	)	)	PUNCT
cana-6072	60	16	.	.	PUNCT
cana-6072	61	1	communications	communication	NOUN
cana-6072	61	2	on	on	ADP
cana-6072	61	3	applied	apply	VERB
cana-6072	61	4	nonlinear	nonlinear	ADJ
cana-6072	61	5	analysis	analysis	NOUN
cana-6072	61	6	issn	issn	NOUN
cana-6072	61	7	:	:	PUNCT
cana-6072	61	8	1074	1074	NUM
cana-6072	61	9	-	-	PUNCT
cana-6072	61	10	133x	133x	NUM
cana-6072	61	11	vol	vol	NOUN
cana-6072	61	12	32	32	NUM
cana-6072	61	13	no	no	NOUN
cana-6072	61	14	.	.	PUNCT
cana-6072	62	1	4s	4s	NUM
cana-6072	62	2	(	(	PUNCT
cana-6072	62	3	2025	2025	NUM
cana-6072	62	4	)	)	PUNCT
cana-6072	62	5	760	760	NUM
cana-6072	62	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	62	7	table	table	NOUN
cana-6072	62	8	1	1	NUM
cana-6072	62	9	.	.	PUNCT
cana-6072	63	1	prisma	prisma	PROPN
cana-6072	63	2	2020	2020	NUM
cana-6072	63	3	flow	flow	NOUN
cana-6072	63	4	counts	count	NOUN
cana-6072	63	5	(	(	PUNCT
cana-6072	63	6	minimal	minimal	ADJ
cana-6072	63	7	n	n	CCONJ
cana-6072	63	8	-	-	PUNCT
cana-6072	63	9	values	value	NOUN
cana-6072	63	10	)	)	PUNCT
cana-6072	63	11	.	.	PUNCT
cana-6072	64	1	stage	stage	NOUN
cana-6072	64	2	n	n	PROPN
cana-6072	64	3	records	record	NOUN
cana-6072	64	4	identified	identify	VERB
cana-6072	64	5	(	(	PUNCT
cana-6072	64	6	all	all	DET
cana-6072	64	7	sources	source	NOUN
cana-6072	64	8	)	)	PUNCT
cana-6072	64	9	duplicates	duplicate	NOUN
cana-6072	64	10	removed	remove	VERB
cana-6072	64	11	records	record	NOUN
cana-6072	64	12	screened	screen	VERB
cana-6072	64	13	(	(	PUNCT
cana-6072	64	14	titles	title	NOUN
cana-6072	64	15	/	/	SYM
cana-6072	64	16	abstracts	abstract	NOUN
cana-6072	64	17	)	)	PUNCT
cana-6072	64	18	excluded	exclude	VERB
cana-6072	64	19	at	at	ADP
cana-6072	64	20	screening	screen	VERB
cana-6072	64	21	full	full	ADJ
cana-6072	64	22	-	-	PUNCT
cana-6072	64	23	text	text	NOUN
cana-6072	64	24	assessed	assess	VERB
cana-6072	64	25	for	for	ADP
cana-6072	64	26	eligibility	eligibility	NOUN
cana-6072	64	27	full	full	ADJ
cana-6072	64	28	-	-	PUNCT
cana-6072	64	29	text	text	NOUN
cana-6072	64	30	excluded	exclude	VERB
cana-6072	64	31	studies	study	NOUN
cana-6072	64	32	included	include	VERB
cana-6072	64	33	in	in	ADP
cana-6072	64	34	synthesis	synthesis	NOUN
cana-6072	64	35	350	350	NUM
cana-6072	64	36	50	50	NUM
cana-6072	64	37	300	300	NUM
cana-6072	64	38	230	230	NUM
cana-6072	64	39	70	70	NUM
cana-6072	64	40	50	50	NUM
cana-6072	64	41	20	20	NUM
cana-6072	64	42	selection	selection	NOUN
cana-6072	64	43	process	process	NOUN
cana-6072	64	44	&	&	CCONJ
cana-6072	64	45	prisma	prisma	PROPN
cana-6072	64	46	tracking	tracking	NOUN
cana-6072	64	47	:	:	PUNCT
cana-6072	64	48	both	both	DET
cana-6072	64	49	reviewers	reviewer	NOUN
cana-6072	64	50	independently	independently	ADV
cana-6072	64	51	screened	screen	VERB
cana-6072	64	52	titles	title	NOUN
cana-6072	64	53	/	/	SYM
cana-6072	64	54	abstracts	abstract	NOUN
cana-6072	64	55	,	,	PUNCT
cana-6072	64	56	then	then	ADV
cana-6072	64	57	full	full	ADJ
cana-6072	64	58	texts	text	NOUN
cana-6072	64	59	,	,	PUNCT
cana-6072	64	60	logging	log	VERB
cana-6072	64	61	reasons	reason	NOUN
cana-6072	64	62	for	for	ADP
cana-6072	64	63	exclusion	exclusion	NOUN
cana-6072	64	64	;	;	PUNCT
cana-6072	64	65	disagreements	disagreement	NOUN
cana-6072	64	66	were	be	AUX
cana-6072	64	67	resolved	resolve	VERB
cana-6072	64	68	by	by	ADP
cana-6072	64	69	discussion	discussion	NOUN
cana-6072	64	70	.	.	PUNCT
cana-6072	65	1	after	after	ADP
cana-6072	65	2	merging	merge	VERB
cana-6072	65	3	exports	export	NOUN
cana-6072	65	4	,	,	PUNCT
cana-6072	65	5	we	we	PRON
cana-6072	65	6	identified	identify	VERB
cana-6072	65	7	350	350	NUM
cana-6072	65	8	records	record	NOUN
cana-6072	65	9	in	in	ADP
cana-6072	65	10	total	total	ADJ
cana-6072	65	11	(	(	PUNCT
cana-6072	65	12	databases	database	NOUN
cana-6072	65	13	=	=	SYM
cana-6072	65	14	320	320	NUM
cana-6072	65	15	;	;	PUNCT
cana-6072	65	16	other	other	ADJ
cana-6072	65	17	methods	method	NOUN
cana-6072	65	18	=	=	SYM
cana-6072	65	19	30	30	NUM
cana-6072	65	20	)	)	PUNCT
cana-6072	65	21	,	,	PUNCT
cana-6072	65	22	removed	remove	VERB
cana-6072	65	23	50	50	NUM
cana-6072	65	24	duplicates	duplicate	NOUN
cana-6072	65	25	,	,	PUNCT
cana-6072	65	26	and	and	CCONJ
cana-6072	65	27	screened	screen	VERB
cana-6072	65	28	300	300	NUM
cana-6072	65	29	titles	title	NOUN
cana-6072	65	30	/	/	SYM
cana-6072	65	31	abstracts	abstract	NOUN
cana-6072	65	32	.	.	PUNCT
cana-6072	66	1	we	we	PRON
cana-6072	66	2	excluded	exclude	VERB
cana-6072	66	3	230	230	NUM
cana-6072	66	4	at	at	ADP
cana-6072	66	5	screening	screen	VERB
cana-6072	66	6	as	as	ADV
cana-6072	66	7	clearly	clearly	ADV
cana-6072	66	8	out	out	ADP
cana-6072	66	9	of	of	ADP
cana-6072	66	10	scope	scope	NOUN
cana-6072	66	11	.	.	PUNCT
cana-6072	67	1	we	we	PRON
cana-6072	67	2	assessed	assess	VERB
cana-6072	67	3	70	70	NUM
cana-6072	67	4	full	full	ADJ
cana-6072	67	5	texts	text	NOUN
cana-6072	67	6	for	for	ADP
cana-6072	67	7	eligibility	eligibility	NOUN
cana-6072	67	8	and	and	CCONJ
cana-6072	67	9	excluded	exclude	VERB
cana-6072	67	10	50	50	NUM
cana-6072	67	11	(	(	PUNCT
cana-6072	67	12	e.g.	e.g.	ADV
cana-6072	67	13	,	,	PUNCT
cana-6072	67	14	no	no	DET
cana-6072	67	15	experiments	experiment	NOUN
cana-6072	67	16	,	,	PUNCT
cana-6072	67	17	off	off	ADP
cana-6072	67	18	-	-	PUNCT
cana-6072	67	19	topic	topic	NOUN
cana-6072	67	20	,	,	PUNCT
cana-6072	67	21	or	or	CCONJ
cana-6072	67	22	insufficient	insufficient	ADJ
cana-6072	67	23	methodological	methodological	ADJ
cana-6072	67	24	detail	detail	NOUN
cana-6072	67	25	)	)	PUNCT
cana-6072	67	26	.	.	PUNCT
cana-6072	68	1	20	20	NUM
cana-6072	68	2	studies	study	NOUN
cana-6072	68	3	were	be	AUX
cana-6072	68	4	included	include	VERB
cana-6072	68	5	in	in	ADP
cana-6072	68	6	the	the	DET
cana-6072	68	7	qualitative	qualitative	ADJ
cana-6072	68	8	synthesis	synthesis	NOUN
cana-6072	68	9	.	.	PUNCT
cana-6072	69	1	figure	figure	NOUN
cana-6072	69	2	?	?	PUNCT
cana-6072	69	3	?	?	PUNCT
cana-6072	70	1	and	and	CCONJ
cana-6072	70	2	table	table	NOUN
cana-6072	70	3	1	1	NUM
cana-6072	70	4	summarize	summarize	VERB
cana-6072	70	5	these	these	DET
cana-6072	70	6	counts	count	NOUN
cana-6072	70	7	.	.	PUNCT
cana-6072	71	1	figure	figure	NOUN
cana-6072	71	2	2	2	NUM
cana-6072	71	3	.	.	PUNCT
cana-6072	72	1	prisma	prisma	PROPN
cana-6072	72	2	2020	2020	NUM
cana-6072	72	3	flow	flow	NOUN
cana-6072	72	4	of	of	ADP
cana-6072	72	5	study	study	NOUN
cana-6072	72	6	selection	selection	NOUN
cana-6072	72	7	.	.	PUNCT
cana-6072	73	1	communications	communication	NOUN
cana-6072	73	2	on	on	ADP
cana-6072	73	3	applied	apply	VERB
cana-6072	73	4	nonlinear	nonlinear	ADJ
cana-6072	73	5	analysis	analysis	NOUN
cana-6072	73	6	issn	issn	NOUN
cana-6072	73	7	:	:	PUNCT
cana-6072	73	8	1074	1074	NUM
cana-6072	73	9	-	-	PUNCT
cana-6072	73	10	133x	133x	NUM
cana-6072	73	11	vol	vol	NOUN
cana-6072	73	12	32	32	NUM
cana-6072	73	13	no	no	NOUN
cana-6072	73	14	.	.	PUNCT
cana-6072	74	1	4s	4s	NUM
cana-6072	74	2	(	(	PUNCT
cana-6072	74	3	2025	2025	NUM
cana-6072	74	4	)	)	PUNCT
cana-6072	74	5	761	761	NUM
cana-6072	74	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	74	7	as	as	SCONJ
cana-6072	74	8	summarized	summarize	VERB
cana-6072	74	9	in	in	ADP
cana-6072	74	10	table	table	NOUN
cana-6072	74	11	1	1	NUM
cana-6072	74	12	,	,	PUNCT
cana-6072	74	13	our	our	PRON
cana-6072	74	14	screening	screening	NOUN
cana-6072	74	15	yielded	yield	VERB
cana-6072	74	16	20	20	NUM
cana-6072	74	17	studies	study	NOUN
cana-6072	74	18	.	.	PUNCT
cana-6072	75	1	data	datum	NOUN
cana-6072	75	2	items	item	NOUN
cana-6072	75	3	:	:	PUNCT
cana-6072	75	4	for	for	ADP
cana-6072	75	5	each	each	DET
cana-6072	75	6	included	include	VERB
cana-6072	75	7	study	study	NOUN
cana-6072	75	8	,	,	PUNCT
cana-6072	75	9	we	we	PRON
cana-6072	75	10	extracted	extract	VERB
cana-6072	75	11	:	:	PUNCT
cana-6072	75	12	task	task	NOUN
cana-6072	75	13	setup	setup	NOUN
cana-6072	75	14	,	,	PUNCT
cana-6072	75	15	datasets	dataset	NOUN
cana-6072	75	16	,	,	PUNCT
cana-6072	75	17	emoji	emoji	PROPN
cana-6072	75	18	representation	representation	NOUN
cana-6072	75	19	,	,	PUNCT
cana-6072	75	20	fusion	fusion	NOUN
cana-6072	75	21	type	type	NOUN
cana-6072	75	22	,	,	PUNCT
cana-6072	75	23	model	model	NOUN
cana-6072	75	24	family	family	NOUN
cana-6072	75	25	,	,	PUNCT
cana-6072	75	26	baselines	baseline	NOUN
cana-6072	75	27	,	,	PUNCT
cana-6072	75	28	metrics	metric	NOUN
cana-6072	75	29	,	,	PUNCT
cana-6072	75	30	gains	gain	NOUN
cana-6072	75	31	over	over	ADP
cana-6072	75	32	text	text	NOUN
cana-6072	75	33	-	-	PUNCT
cana-6072	75	34	only	only	ADV
cana-6072	75	35	,	,	PUNCT
cana-6072	75	36	ablations	ablation	NOUN
cana-6072	75	37	,	,	PUNCT
cana-6072	75	38	and	and	CCONJ
cana-6072	75	39	reproducibility	reproducibility	NOUN
cana-6072	75	40	artifacts	artifact	NOUN
cana-6072	75	41	.	.	PUNCT
cana-6072	76	1	risk	risk	NOUN
cana-6072	76	2	of	of	ADP
cana-6072	76	3	bias	bias	NOUN
cana-6072	76	4	assessment	assessment	NOUN
cana-6072	76	5	:	:	PUNCT
cana-6072	77	1	we	we	PRON
cana-6072	77	2	appraised	appraise	VERB
cana-6072	77	3	reporting	report	VERB
cana-6072	77	4	quality	quality	NOUN
cana-6072	77	5	(	(	PUNCT
cana-6072	77	6	baseline	baseline	NOUN
cana-6072	77	7	/	/	SYM
cana-6072	77	8	tuning	tuning	NOUN
cana-6072	77	9	parity	parity	NOUN
cana-6072	77	10	;	;	PUNCT
cana-6072	77	11	metric	metric	ADJ
cana-6072	77	12	and	and	CCONJ
cana-6072	77	13	statistical	statistical	ADJ
cana-6072	77	14	rigor	rigor	NOUN
cana-6072	77	15	;	;	PUNCT
cana-6072	77	16	ablation	ablation	NOUN
cana-6072	77	17	breadth	breadth	NOUN
cana-6072	77	18	)	)	PUNCT
cana-6072	77	19	and	and	CCONJ
cana-6072	77	20	transparency	transparency	NOUN
cana-6072	77	21	.	.	PUNCT
cana-6072	78	1	any	any	DET
cana-6072	78	2	concerns	concern	NOUN
cana-6072	78	3	were	be	AUX
cana-6072	78	4	noted	note	VERB
cana-6072	78	5	at	at	ADP
cana-6072	78	6	study	study	NOUN
cana-6072	78	7	level	level	NOUN
cana-6072	78	8	and	and	CCONJ
cana-6072	78	9	reflected	reflect	VERB
cana-6072	78	10	in	in	ADP
cana-6072	78	11	the	the	DET
cana-6072	78	12	synthesis	synthesis	NOUN
cana-6072	78	13	.	.	PUNCT
cana-6072	79	1	research	research	NOUN
cana-6072	79	2	questions	question	NOUN
cana-6072	79	3	(	(	PUNCT
cana-6072	79	4	rqs	rqs	PROPN
cana-6072	79	5	)	)	PUNCT
cana-6072	79	6	rq1	rq1	NOUN
cana-6072	79	7	:	:	PUNCT
cana-6072	79	8	which	which	PRON
cana-6072	79	9	model	model	NOUN
cana-6072	79	10	families	family	NOUN
cana-6072	79	11	and	and	CCONJ
cana-6072	79	12	fusion	fusion	NOUN
cana-6072	79	13	strategies	strategy	NOUN
cana-6072	79	14	are	be	AUX
cana-6072	79	15	used	use	VERB
cana-6072	79	16	to	to	PART
cana-6072	79	17	jointly	jointly	ADV
cana-6072	79	18	model	model	VERB
cana-6072	79	19	emojis	emoji	NOUN
cana-6072	79	20	and	and	CCONJ
cana-6072	79	21	text	text	NOUN
cana-6072	79	22	?	?	PUNCT
cana-6072	80	1	rq2	rq2	NOUN
cana-6072	80	2	:	:	PUNCT
cana-6072	80	3	how	how	SCONJ
cana-6072	80	4	are	be	AUX
cana-6072	80	5	emojis	emoji	NOUN
cana-6072	80	6	represented	represent	VERB
cana-6072	80	7	(	(	PUNCT
cana-6072	80	8	lexicon	lexicon	NOUN
cana-6072	80	9	,	,	PUNCT
cana-6072	80	10	textual	textual	ADJ
cana-6072	80	11	gloss	gloss	NOUN
cana-6072	80	12	,	,	PUNCT
cana-6072	80	13	embeddings	embedding	NOUN
cana-6072	80	14	/	/	SYM
cana-6072	80	15	special	special	ADJ
cana-6072	80	16	tokens	token	NOUN
cana-6072	80	17	,	,	PUNCT
cana-6072	80	18	visual	visual	ADJ
cana-6072	80	19	glyph	glyph	NOUN
cana-6072	80	20	encoders	encoder	NOUN
cana-6072	80	21	)	)	PUNCT
cana-6072	80	22	,	,	PUNCT
cana-6072	80	23	and	and	CCONJ
cana-6072	80	24	what	what	PRON
cana-6072	80	25	trade	trade	NOUN
cana-6072	80	26	-	-	PUNCT
cana-6072	80	27	offs	off	NOUN
cana-6072	80	28	arise	arise	NOUN
cana-6072	80	29	?	?	PUNCT
cana-6072	81	1	rq3	rq3	NOUN
cana-6072	81	2	:	:	PUNCT
cana-6072	81	3	which	which	PRON
cana-6072	81	4	datasets	dataset	NOUN
cana-6072	81	5	and	and	CCONJ
cana-6072	81	6	metrics	metric	NOUN
cana-6072	81	7	dominate	dominate	VERB
cana-6072	81	8	evaluation	evaluation	NOUN
cana-6072	81	9	,	,	PUNCT
cana-6072	81	10	and	and	CCONJ
cana-6072	81	11	how	how	SCONJ
cana-6072	81	12	emoji	emoji	ADJ
cana-6072	81	13	density	density	NOUN
cana-6072	81	14	affects	affect	VERB
cana-6072	81	15	performance	performance	NOUN
cana-6072	81	16	?	?	PUNCT
cana-6072	82	1	rq4	rq4	NOUN
cana-6072	82	2	:	:	PUNCT
cana-6072	82	3	when	when	SCONJ
cana-6072	82	4	(	(	PUNCT
cana-6072	82	5	and	and	CCONJ
cana-6072	82	6	by	by	ADP
cana-6072	82	7	how	how	SCONJ
cana-6072	82	8	much	much	ADJ
cana-6072	82	9	)	)	PUNCT
cana-6072	82	10	do	do	AUX
cana-6072	82	11	emoji	emoji	NOUN
cana-6072	82	12	-	-	PUNCT
cana-6072	82	13	aware	aware	ADJ
cana-6072	82	14	models	model	NOUN
cana-6072	82	15	outperform	outperform	VERB
cana-6072	82	16	text	text	NOUN
cana-6072	82	17	-	-	PUNCT
cana-6072	82	18	only	only	ADV
cana-6072	82	19	baselines	baseline	NOUN
cana-6072	82	20	across	across	ADP
cana-6072	82	21	settings	setting	NOUN
cana-6072	82	22	?	?	PUNCT
cana-6072	83	1	rq5	rq5	NOUN
cana-6072	83	2	:	:	PUNCT
cana-6072	83	3	what	what	DET
cana-6072	83	4	failure	failure	NOUN
cana-6072	83	5	modes	mode	NOUN
cana-6072	83	6	and	and	CCONJ
cana-6072	83	7	reproducibility	reproducibility	NOUN
cana-6072	83	8	pitfalls	pitfall	NOUN
cana-6072	83	9	recur	recur	VERB
cana-6072	83	10	,	,	PUNCT
cana-6072	83	11	and	and	CCONJ
cana-6072	83	12	what	what	PRON
cana-6072	83	13	remedies	remedy	NOUN
cana-6072	83	14	exist	exist	VERB
cana-6072	83	15	?	?	PUNCT
cana-6072	84	1	eligibility	eligibility	NOUN
cana-6072	84	2	criteria	criterion	NOUN
cana-6072	84	3	inclusion	inclusion	NOUN
cana-6072	84	4	.	.	PUNCT
cana-6072	85	1	(	(	PUNCT
cana-6072	85	2	i	i	NOUN
cana-6072	85	3	)	)	PUNCT
cana-6072	85	4	english	english	ADJ
cana-6072	85	5	-	-	PUNCT
cana-6072	85	6	language	language	NOUN
cana-6072	85	7	articles	article	NOUN
cana-6072	85	8	(	(	PUNCT
cana-6072	85	9	peer	peer	NOUN
cana-6072	85	10	-	-	PUNCT
cana-6072	85	11	reviewed	review	VERB
cana-6072	85	12	)	)	PUNCT
cana-6072	85	13	;	;	PUNCT
cana-6072	85	14	(	(	PUNCT
cana-6072	85	15	ii	ii	NOUN
cana-6072	85	16	)	)	PUNCT
cana-6072	85	17	primary	primary	ADJ
cana-6072	85	18	studies	study	NOUN
cana-6072	85	19	on	on	ADP
cana-6072	85	20	sentiment	sentiment	NOUN
cana-6072	85	21	classification	classification	NOUN
cana-6072	85	22	that	that	PRON
cana-6072	85	23	jointly	jointly	ADV
cana-6072	85	24	use	use	VERB
cana-6072	85	25	emojis	emoji	NOUN
cana-6072	85	26	and	and	CCONJ
cana-6072	85	27	text	text	NOUN
cana-6072	85	28	;	;	PUNCT
cana-6072	85	29	(	(	PUNCT
cana-6072	85	30	iii	iii	X
cana-6072	85	31	)	)	PUNCT
cana-6072	85	32	report	report	VERB
cana-6072	85	33	quantitative	quantitative	ADJ
cana-6072	85	34	evaluation	evaluation	NOUN
cana-6072	85	35	with	with	ADP
cana-6072	85	36	identifiable	identifiable	ADJ
cana-6072	85	37	datasets	dataset	NOUN
cana-6072	85	38	/	/	SYM
cana-6072	85	39	metrics	metric	NOUN
cana-6072	85	40	;	;	PUNCT
cana-6072	85	41	(	(	PUNCT
cana-6072	85	42	iv	iv	X
cana-6072	85	43	)	)	PUNCT
cana-6072	85	44	2015–2025	2015–2025	NUM
cana-6072	85	45	.	.	PUNCT
cana-6072	86	1	exclusion	exclusion	NOUN
cana-6072	86	2	.	.	PUNCT
cana-6072	87	1	(	(	PUNCT
cana-6072	87	2	i	i	NOUN
cana-6072	87	3	)	)	PUNCT
cana-6072	87	4	purely	purely	ADV
cana-6072	87	5	text	text	NOUN
cana-6072	87	6	-	-	PUNCT
cana-6072	87	7	only	only	ADJ
cana-6072	87	8	sentiment	sentiment	NOUN
cana-6072	87	9	;	;	PUNCT
cana-6072	87	10	(	(	PUNCT
cana-6072	87	11	ii	ii	NOUN
cana-6072	87	12	)	)	PUNCT
cana-6072	87	13	emoji	emoji	PROPN
cana-6072	87	14	prediction	prediction	NOUN
cana-6072	87	15	/	/	SYM
cana-6072	87	16	tagging	tagging	NOUN
cana-6072	87	17	without	without	ADP
cana-6072	87	18	sentiment	sentiment	NOUN
cana-6072	87	19	;	;	PUNCT
cana-6072	87	20	(	(	PUNCT
cana-6072	87	21	iii	iii	X
cana-6072	87	22	)	)	PUNCT
cana-6072	87	23	non	non	ADJ
cana-6072	87	24	-	-	ADJ
cana-6072	87	25	peer	peer	NOUN
cana-6072	87	26	-	-	PUNCT
cana-6072	87	27	reviewed	review	VERB
cana-6072	87	28	(	(	PUNCT
cana-6072	87	29	unless	unless	SCONJ
cana-6072	87	30	the	the	DET
cana-6072	87	31	venue	venue	NOUN
cana-6072	87	32	explicitly	explicitly	ADV
cana-6072	87	33	peer	peer	NOUN
cana-6072	87	34	-	-	PUNCT
cana-6072	87	35	reviews	review	NOUN
cana-6072	87	36	)	)	PUNCT
cana-6072	87	37	or	or	CCONJ
cana-6072	87	38	non	non	ADJ
cana-6072	87	39	-	-	ADJ
cana-6072	87	40	archival	archival	ADJ
cana-6072	87	41	posters	poster	NOUN
cana-6072	87	42	;	;	PUNCT
cana-6072	87	43	(	(	PUNCT
cana-6072	87	44	iv	iv	X
cana-6072	87	45	)	)	PUNCT
cana-6072	87	46	non	non	ADJ
cana-6072	87	47	-	-	ADJ
cana-6072	87	48	english	english	ADJ
cana-6072	87	49	articles	article	NOUN
cana-6072	87	50	;	;	PUNCT
cana-6072	87	51	(	(	PUNCT
cana-6072	87	52	v	v	NOUN
cana-6072	87	53	)	)	PUNCT
cana-6072	87	54	inaccessible	inaccessible	ADJ
cana-6072	87	55	full	full	ADJ
cana-6072	87	56	text	text	NOUN
cana-6072	87	57	;	;	PUNCT
cana-6072	87	58	(	(	PUNCT
cana-6072	87	59	vi	vi	NOUN
cana-6072	87	60	)	)	PUNCT
cana-6072	87	61	duplicates	duplicate	NOUN
cana-6072	87	62	/	/	SYM
cana-6072	87	63	extended	extend	VERB
cana-6072	87	64	versions	version	NOUN
cana-6072	87	65	retain	retain	VERB
cana-6072	87	66	the	the	DET
cana-6072	87	67	most	most	ADV
cana-6072	87	68	complete	complete	ADJ
cana-6072	87	69	.	.	PUNCT
cana-6072	88	1	results	result	VERB
cana-6072	88	2	taxonomy	taxonomy	NOUN
cana-6072	88	3	we	we	PRON
cana-6072	88	4	identify	identify	VERB
cana-6072	88	5	four	four	NUM
cana-6072	88	6	primary	primary	ADJ
cana-6072	88	7	representation	representation	NOUN
cana-6072	88	8	families	family	NOUN
cana-6072	88	9	for	for	ADP
cana-6072	88	10	emojis	emoji	NOUN
cana-6072	88	11	in	in	ADP
cana-6072	88	12	sentiment	sentiment	NOUN
cana-6072	88	13	models	model	NOUN
cana-6072	88	14	:	:	PUNCT
cana-6072	89	1	1	1	X
cana-6072	89	2	.	.	X
cana-6072	89	3	lexicon	lexicon	PROPN
cana-6072	89	4	scores	score	NOUN
cana-6072	89	5	map	map	VERB
cana-6072	89	6	each	each	DET
cana-6072	89	7	emoji	emoji	NOUN
cana-6072	89	8	to	to	ADP
cana-6072	89	9	a	a	DET
cana-6072	89	10	prior	prior	ADJ
cana-6072	89	11	polarity	polarity	NOUN
cana-6072	89	12	/	/	SYM
cana-6072	89	13	affect	affect	VERB
cana-6072	89	14	score	score	NOUN
cana-6072	89	15	(	(	PUNCT
cana-6072	89	16	interpretable	interpretable	ADJ
cana-6072	89	17	,	,	PUNCT
cana-6072	89	18	low	low	ADJ
cana-6072	89	19	cost	cost	NOUN
cana-6072	89	20	)	)	PUNCT
cana-6072	90	1	[	[	X
cana-6072	90	2	novak	novak	PROPN
cana-6072	90	3	,	,	PUNCT
cana-6072	90	4	k.	k.	PROPN
cana-6072	90	5	et	et	PROPN
cana-6072	90	6	al	al	PROPN
cana-6072	90	7	,	,	PUNCT
cana-6072	90	8	2015	2015	NUM
cana-6072	90	9	]	]	PUNCT
cana-6072	90	10	.	.	PUNCT
cana-6072	91	1	2	2	X
cana-6072	91	2	.	.	X
cana-6072	91	3	textual	textual	ADJ
cana-6072	91	4	glossing	glossing	NOUN
cana-6072	91	5	replaces	replace	VERB
cana-6072	91	6	an	an	DET
cana-6072	91	7	emoji	emoji	NOUN
cana-6072	91	8	with	with	ADP
cana-6072	91	9	a	a	DET
cana-6072	91	10	descriptive	descriptive	ADJ
cana-6072	91	11	token	token	NOUN
cana-6072	91	12	(	(	PUNCT
cana-6072	91	13	e.g.	e.g.	ADV
cana-6072	91	14	,	,	PUNCT
cana-6072	91	15	“	"	PUNCT
cana-6072	91	16	:	:	PUNCT
cana-6072	91	17	face_with_tears_of_joy	face_with_tears_of_joy	NOUN
cana-6072	91	18	:	:	PUNCT
cana-6072	91	19	”	"	PUNCT
cana-6072	91	20	)	)	PUNCT
cana-6072	91	21	so	so	SCONJ
cana-6072	91	22	it	it	PRON
cana-6072	91	23	can	can	AUX
cana-6072	91	24	be	be	AUX
cana-6072	91	25	ingested	ingest	VERB
cana-6072	91	26	by	by	ADP
cana-6072	91	27	text	text	NOUN
cana-6072	91	28	models	model	NOUN
cana-6072	91	29	(	(	PUNCT
cana-6072	91	30	lou	lou	PROPN
cana-6072	91	31	,	,	PUNCT
cana-6072	91	32	y.	y.	PROPN
cana-6072	91	33	et	et	PROPN
cana-6072	91	34	al	al	PROPN
cana-6072	91	35	,	,	PUNCT
cana-6072	91	36	2024	2024	NUM
cana-6072	91	37	)	)	PUNCT
cana-6072	91	38	.	.	PUNCT
cana-6072	92	1	3	3	X
cana-6072	92	2	.	.	X
cana-6072	92	3	token	token	VERB
cana-6072	92	4	embeddings	embedding	NOUN
cana-6072	92	5	treat	treat	VERB
cana-6072	92	6	emojis	emoji	NOUN
cana-6072	92	7	as	as	ADP
cana-6072	92	8	tokens	token	NOUN
cana-6072	92	9	and	and	CCONJ
cana-6072	92	10	learn	learn	VERB
cana-6072	92	11	vectors	vector	NOUN
cana-6072	92	12	by	by	ADP
cana-6072	92	13	extending	extend	VERB
cana-6072	92	14	the	the	DET
cana-6072	92	15	vocabulary	vocabulary	NOUN
cana-6072	92	16	or	or	CCONJ
cana-6072	92	17	training	train	VERB
cana-6072	92	18	specialized	specialized	ADJ
cana-6072	92	19	emoji	emoji	NOUN
cana-6072	92	20	embeddings	embedding	NOUN
cana-6072	92	21	(	(	PUNCT
cana-6072	92	22	e.g.	e.g.	ADV
cana-6072	92	23	,	,	PUNCT
cana-6072	92	24	emoji2vec	emoji2vec	NOUN
cana-6072	92	25	/	/	SYM
cana-6072	92	26	usage	usage	NOUN
cana-6072	92	27	-	-	PUNCT
cana-6072	92	28	derived	derive	VERB
cana-6072	92	29	)	)	PUNCT
cana-6072	93	1	[	[	X
cana-6072	93	2	chen	chen	PROPN
cana-6072	93	3	,	,	PUNCT
cana-6072	93	4	y.	y.	PROPN
cana-6072	93	5	et	et	PROPN
cana-6072	93	6	al	al	PROPN
cana-6072	93	7	.	.	PROPN
cana-6072	93	8	,	,	PUNCT
cana-6072	93	9	2018	2018	NUM
cana-6072	93	10	]	]	PUNCT
cana-6072	93	11	.	.	PUNCT
cana-6072	94	1	4	4	X
cana-6072	94	2	.	.	X
cana-6072	94	3	visual	visual	ADJ
cana-6072	94	4	glyph	glyph	PROPN
cana-6072	94	5	features	feature	NOUN
cana-6072	94	6	treat	treat	VERB
cana-6072	94	7	the	the	DET
cana-6072	94	8	emoji	emoji	NOUN
cana-6072	94	9	as	as	ADP
cana-6072	94	10	an	an	DET
cana-6072	94	11	image	image	NOUN
cana-6072	94	12	and	and	CCONJ
cana-6072	94	13	extract	extract	VERB
cana-6072	94	14	features	feature	NOUN
cana-6072	94	15	with	with	ADP
cana-6072	94	16	cnns	cnns	ADJ
cana-6072	94	17	or	or	CCONJ
cana-6072	94	18	vision	vision	NOUN
cana-6072	94	19	transformers	transformer	NOUN
cana-6072	94	20	,	,	PUNCT
cana-6072	94	21	mitigating	mitigate	VERB
cana-6072	94	22	platform	platform	NOUN
cana-6072	94	23	-	-	PUNCT
cana-6072	94	24	specific	specific	ADJ
cana-6072	94	25	rendering	rendering	NOUN
cana-6072	94	26	issues	issue	NOUN
cana-6072	94	27	[	[	X
cana-6072	94	28	lou	lou	PROPN
cana-6072	94	29	,	,	PUNCT
cana-6072	94	30	y.	y.	PROPN
cana-6072	94	31	et	et	PROPN
cana-6072	94	32	al	al	PROPN
cana-6072	94	33	,	,	PUNCT
cana-6072	94	34	2024	2024	NUM
cana-6072	94	35	]	]	PUNCT
cana-6072	94	36	.	.	PUNCT
cana-6072	95	1	early	early	ADJ
cana-6072	95	2	work	work	NOUN
cana-6072	95	3	favored	favor	VERB
cana-6072	95	4	lexicons	lexicon	NOUN
cana-6072	95	5	for	for	ADP
cana-6072	95	6	simplicity	simplicity	NOUN
cana-6072	95	7	and	and	CCONJ
cana-6072	95	8	interpretability	interpretability	NOUN
cana-6072	95	9	[	[	X
cana-6072	95	10	novak	novak	PROPN
cana-6072	95	11	,	,	PUNCT
cana-6072	95	12	k.	k.	PROPN
cana-6072	95	13	et	et	PROPN
cana-6072	95	14	al	al	PROPN
cana-6072	95	15	,	,	PUNCT
cana-6072	95	16	2015	2015	NUM
cana-6072	95	17	]	]	PUNCT
cana-6072	95	18	.	.	PUNCT
cana-6072	96	1	glossing	gloss	VERB
cana-6072	96	2	remains	remain	VERB
cana-6072	96	3	common	common	ADJ
cana-6072	96	4	because	because	SCONJ
cana-6072	96	5	it	it	PRON
cana-6072	96	6	integrates	integrate	VERB
cana-6072	96	7	seamlessly	seamlessly	ADV
cana-6072	96	8	with	with	ADP
cana-6072	96	9	pretrained	pretraine	VERB
cana-6072	96	10	language	language	NOUN
cana-6072	96	11	models	model	NOUN
cana-6072	96	12	.	.	PUNCT
cana-6072	97	1	communications	communication	NOUN
cana-6072	97	2	on	on	ADP
cana-6072	97	3	applied	apply	VERB
cana-6072	97	4	nonlinear	nonlinear	ADJ
cana-6072	97	5	analysis	analysis	NOUN
cana-6072	97	6	issn	issn	NOUN
cana-6072	97	7	:	:	PUNCT
cana-6072	97	8	1074	1074	NUM
cana-6072	97	9	-	-	PUNCT
cana-6072	97	10	133x	133x	NUM
cana-6072	97	11	vol	vol	NOUN
cana-6072	97	12	32	32	NUM
cana-6072	97	13	no	no	NOUN
cana-6072	97	14	.	.	PUNCT
cana-6072	98	1	4s	4s	NUM
cana-6072	98	2	(	(	PUNCT
cana-6072	98	3	2025	2025	NUM
cana-6072	98	4	)	)	PUNCT
cana-6072	98	5	762	762	NUM
cana-6072	98	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	98	7	embedding	embed	VERB
cana-6072	98	8	-	-	PUNCT
cana-6072	98	9	based	base	VERB
cana-6072	98	10	methods	method	NOUN
cana-6072	98	11	extend	extend	VERB
cana-6072	98	12	word2vec	word2vec	NOUN
cana-6072	98	13	/	/	SYM
cana-6072	98	14	glove	glove	NOUN
cana-6072	98	15	or	or	CCONJ
cana-6072	98	16	learn	learn	VERB
cana-6072	98	17	emoji	emoji	ADJ
cana-6072	98	18	-	-	PUNCT
cana-6072	98	19	specific	specific	ADJ
cana-6072	98	20	vectors	vector	NOUN
cana-6072	98	21	from	from	ADP
cana-6072	98	22	social	social	ADJ
cana-6072	98	23	media	medium	NOUN
cana-6072	98	24	usage	usage	NOUN
cana-6072	98	25	,	,	PUNCT
cana-6072	98	26	typically	typically	ADV
cana-6072	98	27	yielding	yield	VERB
cana-6072	98	28	gains	gain	NOUN
cana-6072	98	29	over	over	ADP
cana-6072	98	30	text	text	NOUN
cana-6072	98	31	-	-	PUNCT
cana-6072	98	32	only	only	ADV
cana-6072	98	33	baselines	baseline	NOUN
cana-6072	98	34	[	[	X
cana-6072	98	35	lou	lou	PROPN
cana-6072	98	36	,	,	PUNCT
cana-6072	98	37	y.	y.	PROPN
cana-6072	98	38	et	et	PROPN
cana-6072	98	39	al	al	PROPN
cana-6072	98	40	,	,	PUNCT
cana-6072	98	41	2024	2024	NUM
cana-6072	98	42	]	]	PUNCT
cana-6072	98	43	.	.	PUNCT
cana-6072	99	1	visual	visual	ADJ
cana-6072	99	2	approaches	approach	NOUN
cana-6072	99	3	are	be	AUX
cana-6072	99	4	emerging	emerge	VERB
cana-6072	99	5	,	,	PUNCT
cana-6072	99	6	using	use	VERB
cana-6072	99	7	resnet	resnet	NOUN
cana-6072	99	8	/	/	SYM
cana-6072	99	9	vit	vit	NOUN
cana-6072	99	10	encoders	encoder	NOUN
cana-6072	99	11	to	to	PART
cana-6072	99	12	model	model	VERB
cana-6072	99	13	glyph	glyph	ADJ
cana-6072	99	14	appearance	appearance	NOUN
cana-6072	99	15	variation	variation	NOUN
cana-6072	100	1	[	[	X
cana-6072	100	2	chen	chen	PROPN
cana-6072	100	3	,	,	PUNCT
cana-6072	100	4	y.	y.	PROPN
cana-6072	100	5	et	et	PROPN
cana-6072	100	6	al	al	PROPN
cana-6072	100	7	.	.	PROPN
cana-6072	100	8	,	,	PUNCT
cana-6072	100	9	2018	2018	NUM
cana-6072	100	10	]	]	PUNCT
cana-6072	100	11	.	.	PUNCT
cana-6072	101	1	each	each	DET
cana-6072	101	2	choice	choice	NOUN
cana-6072	101	3	trades	trade	VERB
cana-6072	101	4	off	off	ADP
cana-6072	101	5	coverage	coverage	NOUN
cana-6072	101	6	,	,	PUNCT
cana-6072	101	7	domain	domain	NOUN
cana-6072	101	8	shift	shift	NOUN
cana-6072	101	9	robustness	robustness	NOUN
cana-6072	101	10	(	(	PUNCT
cana-6072	101	11	platform	platform	NOUN
cana-6072	101	12	/	/	SYM
cana-6072	101	13	culture	culture	NOUN
cana-6072	101	14	)	)	PUNCT
cana-6072	101	15	,	,	PUNCT
cana-6072	101	16	and	and	CCONJ
cana-6072	101	17	cost	cost	NOUN
cana-6072	101	18	;	;	PUNCT
cana-6072	101	19	the	the	DET
cana-6072	101	20	optimal	optimal	ADJ
cana-6072	101	21	setting	setting	NOUN
cana-6072	101	22	depends	depend	VERB
cana-6072	101	23	on	on	ADP
cana-6072	101	24	resource	resource	NOUN
cana-6072	101	25	constraints	constraint	NOUN
cana-6072	101	26	and	and	CCONJ
cana-6072	101	27	emoji	emoji	NOUN
cana-6072	101	28	density	density	NOUN
cana-6072	101	29	in	in	ADP
cana-6072	101	30	the	the	DET
cana-6072	101	31	target	target	NOUN
cana-6072	101	32	corpus	corpus	NOUN
cana-6072	101	33	.	.	PUNCT
cana-6072	102	1	as	as	SCONJ
cana-6072	102	2	summarized	summarize	VERB
cana-6072	102	3	in	in	ADP
cana-6072	102	4	table	table	NOUN
cana-6072	102	5	2	2	NUM
cana-6072	102	6	,	,	PUNCT
cana-6072	102	7	we	we	PRON
cana-6072	102	8	code	code	VERB
cana-6072	102	9	each	each	DET
cana-6072	102	10	study	study	NOUN
cana-6072	102	11	by	by	ADP
cana-6072	102	12	its	its	PRON
cana-6072	102	13	representation	representation	NOUN
cana-6072	102	14	choice	choice	NOUN
cana-6072	102	15	and	and	CCONJ
cana-6072	102	16	note	note	VERB
cana-6072	102	17	principal	principal	ADJ
cana-6072	102	18	trade	trade	NOUN
cana-6072	102	19	-	-	PUNCT
cana-6072	102	20	offs	off	NOUN
cana-6072	102	21	(	(	PUNCT
cana-6072	102	22	interpretability	interpretability	NOUN
cana-6072	102	23	,	,	PUNCT
cana-6072	102	24	coverage	coverage	NOUN
cana-6072	102	25	,	,	PUNCT
cana-6072	102	26	robustness	robustness	NOUN
cana-6072	102	27	,	,	PUNCT
cana-6072	102	28	compute	compute	NOUN
cana-6072	102	29	)	)	PUNCT
cana-6072	102	30	.	.	PUNCT
cana-6072	103	1	figure	figure	VERB
cana-6072	103	2	3	3	NUM
cana-6072	103	3	.	.	PUNCT
cana-6072	103	4	emoji	emoji	PROPN
cana-6072	103	5	representation	representation	NOUN
cana-6072	103	6	families	family	NOUN
cana-6072	103	7	:	:	PUNCT
cana-6072	103	8	(	(	PUNCT
cana-6072	103	9	1	1	X
cana-6072	103	10	)	)	PUNCT
cana-6072	103	11	lexicon	lexicon	NOUN
cana-6072	103	12	scores	score	NOUN
cana-6072	103	13	,	,	PUNCT
cana-6072	103	14	(	(	PUNCT
cana-6072	103	15	2	2	X
cana-6072	103	16	)	)	PUNCT
cana-6072	103	17	textual	textual	ADJ
cana-6072	103	18	glossing	glossing	NOUN
cana-6072	103	19	,	,	PUNCT
cana-6072	103	20	(	(	PUNCT
cana-6072	103	21	3	3	X
cana-6072	103	22	)	)	PUNCT
cana-6072	103	23	token	token	ADJ
cana-6072	103	24	embeddings	embedding	NOUN
cana-6072	103	25	,	,	PUNCT
cana-6072	103	26	and	and	CCONJ
cana-6072	103	27	(	(	PUNCT
cana-6072	103	28	4	4	X
cana-6072	103	29	)	)	PUNCT
cana-6072	103	30	visual	visual	ADJ
cana-6072	103	31	glyph	glyph	NOUN
cana-6072	103	32	features	feature	NOUN
cana-6072	103	33	.	.	PUNCT
cana-6072	104	1	communications	communication	NOUN
cana-6072	104	2	on	on	ADP
cana-6072	104	3	applied	apply	VERB
cana-6072	104	4	nonlinear	nonlinear	ADJ
cana-6072	104	5	analysis	analysis	NOUN
cana-6072	104	6	issn	issn	NOUN
cana-6072	104	7	:	:	PUNCT
cana-6072	104	8	1074	1074	NUM
cana-6072	104	9	-	-	PUNCT
cana-6072	104	10	133x	133x	NUM
cana-6072	104	11	vol	vol	NOUN
cana-6072	104	12	32	32	NUM
cana-6072	104	13	no	no	NOUN
cana-6072	104	14	.	.	PUNCT
cana-6072	105	1	4s	4s	NUM
cana-6072	105	2	(	(	PUNCT
cana-6072	105	3	2025	2025	NUM
cana-6072	105	4	)	)	PUNCT
cana-6072	105	5	763	763	NUM
cana-6072	105	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	105	7	table	table	NOUN
cana-6072	105	8	2	2	NUM
cana-6072	105	9	.	.	PUNCT
cana-6072	105	10	emoji	emoji	PROPN
cana-6072	105	11	representation	representation	NOUN
cana-6072	105	12	families	family	NOUN
cana-6072	105	13	and	and	CCONJ
cana-6072	105	14	trade	trade	NOUN
cana-6072	105	15	-	-	PUNCT
cana-6072	105	16	offs	off	NOUN
cana-6072	105	17	.	.	PUNCT
cana-6072	106	1	family	family	NOUN
cana-6072	106	2	strengths	strength	NOUN
cana-6072	106	3	/	/	SYM
cana-6072	106	4	typical	typical	ADJ
cana-6072	106	5	use	use	NOUN
cana-6072	106	6	cove	cove	NOUN
cana-6072	106	7	rage	rage	NOUN
cana-6072	106	8	robus	robus	PROPN
cana-6072	106	9	tness	tness	PROPN
cana-6072	106	10	cost	cost	VERB
cana-6072	106	11	lexicon	lexicon	PROPN
cana-6072	106	12	scores	score	NOUN
cana-6072	106	13	interpretable	interpretable	ADJ
cana-6072	106	14	prior	prior	ADJ
cana-6072	106	15	polarity	polarity	NOUN
cana-6072	106	16	;	;	PUNCT
cana-6072	106	17	quick	quick	ADJ
cana-6072	106	18	baselines	baseline	NOUN
cana-6072	106	19	;	;	PUNCT
cana-6072	106	20	zero	zero	NUM
cana-6072	106	21	training	training	NOUN
cana-6072	106	22	;	;	PUNCT
cana-6072	106	23	good	good	ADJ
cana-6072	106	24	for	for	ADP
cana-6072	106	25	rule	rule	NOUN
cana-6072	106	26	hybrids	hybrid	NOUN
cana-6072	106	27	low	low	ADJ
cana-6072	106	28	–	–	PUNCT
cana-6072	106	29	med	med	ADJ
cana-6072	106	30	low	low	ADJ
cana-6072	106	31	low	low	ADJ
cana-6072	106	32	textual	textual	ADJ
cana-6072	106	33	glossing	gloss	VERB
cana-6072	106	34	drops	drop	NOUN
cana-6072	106	35	into	into	ADP
cana-6072	106	36	plms	plm	NOUN
cana-6072	106	37	;	;	PUNCT
cana-6072	106	38	leverages	leverage	NOUN
cana-6072	106	39	textual	textual	ADJ
cana-6072	106	40	pretraining	pretraining	NOUN
cana-6072	106	41	;	;	PUNCT
cana-6072	106	42	easy	easy	ADJ
cana-6072	106	43	data	datum	NOUN
cana-6072	106	44	pipeline	pipeline	NOUN
cana-6072	106	45	(:	(:	NOUN
cana-6072	106	46	smiling_face	smiling_face	PROPN
cana-6072	106	47	:)	:)	NOUN
cana-6072	106	48	med	me	VERB
cana-6072	106	49	med	med	ADJ
cana-6072	106	50	low	low	ADV
cana-6072	106	51	token	token	VERB
cana-6072	106	52	embeddings	embedding	NOUN
cana-6072	106	53	learns	learn	VERB
cana-6072	106	54	usage	usage	NOUN
cana-6072	106	55	-	-	PUNCT
cana-6072	106	56	specific	specific	ADJ
cana-6072	106	57	semantics	semantic	NOUN
cana-6072	106	58	;	;	PUNCT
cana-6072	106	59	integrates	integrate	NOUN
cana-6072	106	60	with	with	ADP
cana-6072	106	61	rnn	rnn	PROPN
cana-6072	106	62	/	/	SYM
cana-6072	106	63	cnn	cnn	PROPN
cana-6072	106	64	/	/	SYM
cana-6072	106	65	transformer	transformer	NOUN
cana-6072	106	66	text	text	NOUN
cana-6072	106	67	encoders	encoder	NOUN
cana-6072	106	68	med	me	VERB
cana-6072	106	69	–	–	PUNCT
cana-6072	106	70	high	high	ADJ
cana-6072	106	71	med	med	ADJ
cana-6072	106	72	med	med	ADJ
cana-6072	106	73	visual	visual	ADJ
cana-6072	106	74	glyph	glyph	NOUN
cana-6072	106	75	features	feature	VERB
cana-6072	106	76	captures	capture	VERB
cana-6072	106	77	platform	platform	NOUN
cana-6072	106	78	-	-	PUNCT
cana-6072	106	79	specific	specific	ADJ
cana-6072	106	80	appearance	appearance	NOUN
cana-6072	106	81	;	;	PUNCT
cana-6072	106	82	complements	complement	VERB
cana-6072	106	83	text	text	NOUN
cana-6072	106	84	;	;	PUNCT
cana-6072	106	85	helps	help	VERB
cana-6072	106	86	disambiguate	disambiguate	ADJ
cana-6072	106	87	polysemy	polysemy	NOUN
cana-6072	106	88	high	high	ADV
cana-6072	106	89	*	*	PUNCT
cana-6072	106	90	med	med	ADJ
cana-6072	106	91	–	–	PUNCT
cana-6072	106	92	high	high	ADJ
cana-6072	106	93	med	med	ADJ
cana-6072	106	94	–	–	PUNCT
cana-6072	106	95	high	high	ADJ
cana-6072	106	96	*	*	PUNCT
cana-6072	106	97	coverage	coverage	NOUN
cana-6072	106	98	reflects	reflect	VERB
cana-6072	106	99	glyph	glyph	NOUN
cana-6072	106	100	availability	availability	NOUN
cana-6072	106	101	(	(	PUNCT
cana-6072	106	102	full	full	ADJ
cana-6072	106	103	unicode	unicode	NOUN
cana-6072	106	104	set	set	NOUN
cana-6072	106	105	)	)	PUNCT
cana-6072	106	106	rather	rather	ADV
cana-6072	106	107	than	than	ADP
cana-6072	106	108	sense	sense	VERB
cana-6072	106	109	coverage	coverage	NOUN
cana-6072	106	110	;	;	PUNCT
cana-6072	106	111	robustness	robustness	NOUN
cana-6072	106	112	considers	consider	VERB
cana-6072	106	113	culture	culture	NOUN
cana-6072	106	114	/	/	SYM
cana-6072	106	115	platform	platform	NOUN
cana-6072	106	116	shift	shift	NOUN
cana-6072	106	117	tolerance	tolerance	NOUN
cana-6072	106	118	.	.	PUNCT
cana-6072	107	1	cost	cost	NOUN
cana-6072	107	2	reflects	reflect	VERB
cana-6072	107	3	compute	compute	NOUN
cana-6072	107	4	/	/	SYM
cana-6072	107	5	data	data	NOUN
cana-6072	107	6	/	/	SYM
cana-6072	107	7	engineering	engineering	NOUN
cana-6072	107	8	overhead	overhead	NOUN
cana-6072	107	9	.	.	PUNCT
cana-6072	108	1	fusion	fusion	NOUN
cana-6072	108	2	strategies	strategy	NOUN
cana-6072	108	3	emoji	emoji	NOUN
cana-6072	108	4	information	information	NOUN
cana-6072	108	5	can	can	AUX
cana-6072	108	6	enter	enter	VERB
cana-6072	108	7	model	model	NOUN
cana-6072	108	8	architectures	architecture	NOUN
cana-6072	108	9	at	at	ADP
cana-6072	108	10	different	different	ADJ
cana-6072	108	11	stages	stage	NOUN
cana-6072	108	12	.	.	PUNCT
cana-6072	109	1	we	we	PRON
cana-6072	109	2	categorize	categorize	VERB
cana-6072	109	3	fusion	fusion	NOUN
cana-6072	109	4	strategies	strategy	NOUN
cana-6072	109	5	as	as	ADV
cana-6072	109	6	early	early	ADJ
cana-6072	109	7	(	(	PUNCT
cana-6072	109	8	feature	feature	NOUN
cana-6072	109	9	-	-	PUNCT
cana-6072	109	10	level	level	NOUN
cana-6072	109	11	)	)	PUNCT
cana-6072	109	12	,	,	PUNCT
cana-6072	109	13	late	late	ADV
cana-6072	109	14	(	(	PUNCT
cana-6072	109	15	decision	decision	NOUN
cana-6072	109	16	-	-	PUNCT
cana-6072	109	17	level	level	NOUN
cana-6072	109	18	)	)	PUNCT
cana-6072	109	19	,	,	PUNCT
cana-6072	109	20	intermediate	intermediate	ADJ
cana-6072	109	21	(	(	PUNCT
cana-6072	109	22	attention	attention	NOUN
cana-6072	109	23	/	/	SYM
cana-6072	109	24	interaction	interaction	NOUN
cana-6072	109	25	)	)	PUNCT
cana-6072	109	26	,	,	PUNCT
cana-6072	109	27	or	or	CCONJ
cana-6072	109	28	hybrid	hybrid	NOUN
cana-6072	109	29	.	.	PUNCT
cana-6072	110	1	architectures	architecture	NOUN
cana-6072	110	2	are	be	AUX
cana-6072	110	3	sketched	sketch	VERB
cana-6072	110	4	in	in	ADP
cana-6072	110	5	fig	fig	NOUN
cana-6072	110	6	.	.	PUNCT
cana-6072	111	1	4	4	NUM
cana-6072	111	2	,	,	PUNCT
cana-6072	111	3	and	and	CCONJ
cana-6072	111	4	capabilities	capability	NOUN
cana-6072	111	5	/	/	SYM
cana-6072	111	6	costs	cost	NOUN
cana-6072	111	7	are	be	AUX
cana-6072	111	8	summarized	summarize	VERB
cana-6072	111	9	in	in	ADP
cana-6072	111	10	table	table	NOUN
cana-6072	111	11	3	3	NUM
cana-6072	111	12	.	.	PUNCT
cana-6072	111	13	table	table	NOUN
cana-6072	111	14	3	3	NUM
cana-6072	111	15	.	.	PUNCT
cana-6072	111	16	fusion	fusion	NOUN
cana-6072	111	17	strategies	strategy	NOUN
cana-6072	111	18	:	:	PUNCT
cana-6072	111	19	capabilities	capability	NOUN
cana-6072	111	20	and	and	CCONJ
cana-6072	111	21	costs	cost	NOUN
cana-6072	111	22	.	.	PUNCT
cana-6072	112	1	fusion	fusion	NOUN
cana-6072	112	2	type	type	NOUN
cana-6072	112	3	mechanism	mechanism	NOUN
cana-6072	112	4	pros	pro	VERB
cana-6072	112	5	cons	con	NOUN
cana-6072	112	6	typical	typical	ADJ
cana-6072	112	7	use	use	NOUN
cana-6072	112	8	cases	case	NOUN
cana-6072	112	9	early	early	ADV
cana-6072	112	10	(	(	PUNCT
cana-6072	112	11	featurelevel	featurelevel	NOUN
cana-6072	112	12	)	)	PUNCT
cana-6072	112	13	integrate	integrate	VERB
cana-6072	112	14	emoji	emoji	PROPN
cana-6072	112	15	features	feature	NOUN
cana-6072	112	16	into	into	ADP
cana-6072	112	17	text	text	NOUN
cana-6072	112	18	vector	vector	NOUN
cana-6072	112	19	(	(	PUNCT
cana-6072	112	20	e.g.	e.g.	ADV
cana-6072	112	21	,	,	PUNCT
cana-6072	112	22	embedding	embed	VERB
cana-6072	112	23	concat	concat	NOUN
cana-6072	112	24	)	)	PUNCT
cana-6072	112	25	.	.	PUNCT
cana-6072	113	1	fine	fine	ADJ
cana-6072	113	2	-	-	PUNCT
cana-6072	113	3	grained	grain	VERB
cana-6072	113	4	word	word	NOUN
cana-6072	113	5	–	–	PUNCT
cana-6072	113	6	emoji	emoji	NOUN
cana-6072	113	7	interactions	interaction	NOUN
cana-6072	113	8	;	;	PUNCT
cana-6072	113	9	single	single	ADJ
cana-6072	113	10	model	model	NOUN
cana-6072	113	11	learns	learn	VERB
cana-6072	113	12	jointly	jointly	ADV
cana-6072	113	13	.	.	PUNCT
cana-6072	114	1	higher	high	ADJ
cana-6072	114	2	-	-	PUNCT
cana-6072	114	3	dim	dim	ADJ
cana-6072	114	4	input	input	NOUN
cana-6072	114	5	;	;	PUNCT
cana-6072	114	6	brittle	brittle	ADJ
cana-6072	114	7	with	with	ADP
cana-6072	114	8	missing	miss	VERB
cana-6072	114	9	modality	modality	NOUN
cana-6072	114	10	;	;	PUNCT
cana-6072	114	11	risk	risk	NOUN
cana-6072	114	12	of	of	ADP
cana-6072	114	13	overfitting	overfitte	VERB
cana-6072	114	14	.	.	PUNCT
cana-6072	115	1	short	short	ADJ
cana-6072	115	2	texts	text	NOUN
cana-6072	115	3	where	where	SCONJ
cana-6072	115	4	emojis	emoji	NOUN
cana-6072	115	5	alter	alter	VERB
cana-6072	115	6	specific	specific	ADJ
cana-6072	115	7	words	word	NOUN
cana-6072	115	8	(	(	PUNCT
cana-6072	115	9	e.g.	e.g.	ADV
cana-6072	115	10	,	,	PUNCT
cana-6072	115	11	sarcasm	sarcasm	NOUN
cana-6072	115	12	cues	cue	NOUN
cana-6072	115	13	)	)	PUNCT
cana-6072	115	14	.	.	PUNCT
cana-6072	116	1	late	late	ADV
cana-6072	116	2	(	(	PUNCT
cana-6072	116	3	decisionlevel	decisionlevel	ADJ
cana-6072	116	4	)	)	PUNCT
cana-6072	116	5	separate	separate	ADJ
cana-6072	116	6	text	text	NOUN
cana-6072	116	7	/	/	SYM
cana-6072	116	8	emoji	emoji	PROPN
cana-6072	116	9	classifiers	classifier	NOUN
cana-6072	116	10	;	;	PUNCT
cana-6072	116	11	combine	combine	VERB
cana-6072	116	12	outputs	output	NOUN
cana-6072	116	13	(	(	PUNCT
cana-6072	116	14	e.g.	e.g.	ADV
cana-6072	116	15	,	,	PUNCT
cana-6072	116	16	averaging	averaging	NOUN
cana-6072	116	17	)	)	PUNCT
cana-6072	116	18	.	.	PUNCT
cana-6072	117	1	modular	modular	ADJ
cana-6072	117	2	;	;	PUNCT
cana-6072	117	3	robust	robust	ADJ
cana-6072	117	4	to	to	ADP
cana-6072	117	5	missing	miss	VERB
cana-6072	117	6	modality	modality	NOUN
cana-6072	117	7	.	.	PUNCT
cana-6072	118	1	loses	lose	VERB
cana-6072	118	2	cross	cross	ADJ
cana-6072	118	3	-	-	ADJ
cana-6072	118	4	modal	modal	ADJ
cana-6072	118	5	nuance	nuance	NOUN
cana-6072	118	6	;	;	PUNCT
cana-6072	118	7	no	no	DET
cana-6072	118	8	joint	joint	ADJ
cana-6072	118	9	feature	feature	NOUN
cana-6072	118	10	learning	learning	NOUN
cana-6072	118	11	.	.	PUNCT
cana-6072	119	1	multi	multi	ADJ
cana-6072	119	2	-	-	ADJ
cana-6072	119	3	source	source	NOUN
cana-6072	119	4	setups	setup	NOUN
cana-6072	119	5	;	;	PUNCT
cana-6072	119	6	external	external	ADJ
cana-6072	119	7	emoji	emoji	PROPN
cana-6072	119	8	analysis	analysis	NOUN
cana-6072	119	9	.	.	PUNCT
cana-6072	120	1	intermediate	intermediate	ADJ
cana-6072	120	2	(	(	PUNCT
cana-6072	120	3	attention	attention	NOUN
cana-6072	120	4	)	)	PUNCT
cana-6072	120	5	cross-/co	cross-/co	NOUN
cana-6072	120	6	-	-	PUNCT
cana-6072	120	7	attention	attention	NOUN
cana-6072	120	8	between	between	ADP
cana-6072	120	9	text	text	NOUN
cana-6072	120	10	and	and	CCONJ
cana-6072	120	11	emoji	emoji	PROPN
cana-6072	120	12	reps	rep	NOUN
cana-6072	120	13	.	.	PROPN
cana-6072	120	14	captures	captures	PROPN
cana-6072	120	15	contextual	contextual	ADJ
cana-6072	120	16	interplay	interplay	NOUN
cana-6072	120	17	;	;	PUNCT
cana-6072	120	18	often	often	ADV
cana-6072	120	19	best	good	ADJ
cana-6072	120	20	accuracy	accuracy	NOUN
cana-6072	120	21	.	.	PUNCT
cana-6072	121	1	more	more	ADV
cana-6072	121	2	complex	complex	ADJ
cana-6072	121	3	;	;	PUNCT
cana-6072	121	4	higher	high	ADJ
cana-6072	121	5	training	training	NOUN
cana-6072	121	6	cost	cost	NOUN
cana-6072	121	7	/	/	SYM
cana-6072	121	8	data	datum	NOUN
cana-6072	121	9	.	.	PUNCT
cana-6072	122	1	context	context	NOUN
cana-6072	122	2	-	-	PUNCT
cana-6072	122	3	dependent	dependent	ADJ
cana-6072	122	4	emoji	emoji	NOUN
cana-6072	122	5	meaning	meaning	NOUN
cana-6072	122	6	(	(	PUNCT
cana-6072	122	7	irony	irony	NOUN
cana-6072	122	8	,	,	PUNCT
cana-6072	122	9	tone	tone	NOUN
cana-6072	122	10	)	)	PUNCT
cana-6072	122	11	.	.	PUNCT
cana-6072	123	1	hybrid	hybrid	ADJ
cana-6072	123	2	multi	multi	ADJ
cana-6072	123	3	-	-	ADJ
cana-6072	123	4	step	step	ADJ
cana-6072	123	5	/	/	SYM
cana-6072	123	6	multimodel	multimodel	PROPN
cana-6072	123	7	(	(	PUNCT
cana-6072	123	8	e.g.	e.g.	ADV
cana-6072	123	9	,	,	PUNCT
cana-6072	123	10	early+attention	early+attention	NOUN
cana-6072	123	11	;	;	PUNCT
cana-6072	123	12	ensembling	ensemble	VERB
cana-6072	123	13	)	)	PUNCT
cana-6072	123	14	.	.	PUNCT
cana-6072	124	1	leverages	leverage	VERB
cana-6072	124	2	multiple	multiple	ADJ
cana-6072	124	3	strengths	strength	NOUN
cana-6072	124	4	;	;	PUNCT
cana-6072	124	5	often	often	ADV
cana-6072	124	6	sota	sota	NOUN
cana-6072	124	7	.	.	PUNCT
cana-6072	125	1	most	most	ADV
cana-6072	125	2	complex	complex	ADJ
cana-6072	125	3	;	;	PUNCT
cana-6072	125	4	harder	hard	ADJ
cana-6072	125	5	to	to	PART
cana-6072	125	6	interpret	interpret	VERB
cana-6072	125	7	;	;	PUNCT
cana-6072	125	8	careful	careful	ADJ
cana-6072	125	9	tuning	tuning	NOUN
cana-6072	125	10	.	.	PUNCT
cana-6072	126	1	highstakes	highstake	NOUN
cana-6072	126	2	/	/	SYM
cana-6072	126	3	production	production	NOUN
cana-6072	126	4	robustness	robustness	NOUN
cana-6072	126	5	;	;	PUNCT
cana-6072	126	6	fallback	fallback	NOUN
cana-6072	126	7	models	model	NOUN
cana-6072	126	8	.	.	PUNCT
cana-6072	127	1	early	early	ADJ
cana-6072	127	2	fusion	fusion	NOUN
cana-6072	127	3	.	.	PUNCT
cana-6072	128	1	concatenate	concatenate	NOUN
cana-6072	128	2	or	or	CCONJ
cana-6072	128	3	combine	combine	VERB
cana-6072	128	4	emoji	emoji	PROPN
cana-6072	128	5	features	feature	NOUN
cana-6072	128	6	with	with	ADP
cana-6072	128	7	text	text	NOUN
cana-6072	128	8	features	feature	NOUN
cana-6072	128	9	at	at	ADP
cana-6072	128	10	input	input	NOUN
cana-6072	128	11	or	or	CCONJ
cana-6072	128	12	intermediate	intermediate	ADJ
cana-6072	128	13	layers	layer	NOUN
cana-6072	128	14	,	,	PUNCT
cana-6072	128	15	then	then	ADV
cana-6072	128	16	feed	feed	VERB
cana-6072	128	17	into	into	ADP
cana-6072	128	18	one	one	NUM
cana-6072	128	19	model	model	NOUN
cana-6072	128	20	.	.	PUNCT
cana-6072	129	1	for	for	ADP
cana-6072	129	2	example	example	NOUN
cana-6072	129	3	,	,	PUNCT
cana-6072	129	4	liu	liu	PROPN
cana-6072	129	5	et	et	PROPN
cana-6072	129	6	al	al	PROPN
cana-6072	129	7	.	.	PROPN
cana-6072	129	8	embed	embed	PROPN
cana-6072	129	9	emojis	emoji	NOUN
cana-6072	129	10	in	in	ADP
cana-6072	129	11	the	the	DET
cana-6072	129	12	same	same	ADJ
cana-6072	129	13	vector	vector	NOUN
cana-6072	129	14	space	space	NOUN
cana-6072	129	15	as	as	ADP
cana-6072	129	16	words	word	NOUN
cana-6072	129	17	and	and	CCONJ
cana-6072	129	18	input	input	VERB
cana-6072	129	19	them	they	PRON
cana-6072	129	20	to	to	ADP
cana-6072	129	21	a	a	DET
cana-6072	129	22	bilstm	bilstm	NOUN
cana-6072	129	23	(	(	PUNCT
cana-6072	129	24	cemo	cemo	NOUN
cana-6072	129	25	-	-	PUNCT
cana-6072	129	26	lstm	lstm	NOUN
cana-6072	129	27	)	)	PUNCT
cana-6072	130	1	[	[	X
cana-6072	130	2	liu	liu	PROPN
cana-6072	130	3	,	,	PUNCT
cana-6072	130	4	c.	c.	PROPN
cana-6072	130	5	,	,	PUNCT
cana-6072	130	6	et	et	PROPN
cana-6072	130	7	al	al	PROPN
cana-6072	130	8	.	.	PROPN
cana-6072	130	9	,	,	PUNCT
cana-6072	130	10	2021	2021	NUM
cana-6072	130	11	]	]	PUNCT
cana-6072	130	12	,	,	PUNCT
cana-6072	130	13	effectively	effectively	ADV
cana-6072	130	14	an	an	DET
cana-6072	130	15	early	early	ADJ
cana-6072	130	16	fusion	fusion	NOUN
cana-6072	130	17	of	of	ADP
cana-6072	130	18	modalities	modality	NOUN
cana-6072	130	19	.	.	PUNCT
cana-6072	131	1	this	this	PRON
cana-6072	131	2	enables	enable	VERB
cana-6072	131	3	fine	fine	ADJ
cana-6072	131	4	emoji	emoji	ADJ
cana-6072	131	5	–	–	PUNCT
cana-6072	131	6	word	word	NOUN
cana-6072	131	7	interactions	interaction	NOUN
cana-6072	131	8	but	but	CCONJ
cana-6072	131	9	increases	increase	NOUN
cana-6072	131	10	feature	feature	VERB
cana-6072	131	11	dimensionality	dimensionality	NOUN
cana-6072	131	12	and	and	CCONJ
cana-6072	131	13	risks	risk	NOUN
cana-6072	131	14	overfitting	overfitte	VERB
cana-6072	131	15	if	if	SCONJ
cana-6072	131	16	one	one	NUM
cana-6072	131	17	modality	modality	NOUN
cana-6072	131	18	is	be	AUX
cana-6072	131	19	sparse	sparse	ADJ
cana-6072	131	20	.	.	PUNCT
cana-6072	132	1	communications	communication	NOUN
cana-6072	132	2	on	on	ADP
cana-6072	132	3	applied	apply	VERB
cana-6072	132	4	nonlinear	nonlinear	ADJ
cana-6072	132	5	analysis	analysis	NOUN
cana-6072	132	6	issn	issn	NOUN
cana-6072	132	7	:	:	PUNCT
cana-6072	132	8	1074	1074	NUM
cana-6072	132	9	-	-	PUNCT
cana-6072	132	10	133x	133x	NUM
cana-6072	132	11	vol	vol	NOUN
cana-6072	132	12	32	32	NUM
cana-6072	132	13	no	no	NOUN
cana-6072	132	14	.	.	PUNCT
cana-6072	133	1	4s	4s	NUM
cana-6072	133	2	(	(	PUNCT
cana-6072	133	3	2025	2025	NUM
cana-6072	133	4	)	)	PUNCT
cana-6072	133	5	764	764	NUM
cana-6072	133	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	133	7	late	late	ADJ
cana-6072	133	8	fusion	fusion	NOUN
cana-6072	133	9	.	.	PUNCT
cana-6072	134	1	compute	compute	VERB
cana-6072	134	2	separate	separate	ADJ
cana-6072	134	3	predictions	prediction	NOUN
cana-6072	134	4	from	from	ADP
cana-6072	134	5	text	text	NOUN
cana-6072	134	6	and	and	CCONJ
cana-6072	134	7	emoji	emoji	PROPN
cana-6072	134	8	pipelines	pipeline	NOUN
cana-6072	134	9	,	,	PUNCT
cana-6072	134	10	then	then	ADV
cana-6072	134	11	combine	combine	VERB
cana-6072	134	12	outputs	output	NOUN
cana-6072	134	13	(	(	PUNCT
cana-6072	134	14	e.g.	e.g.	ADV
cana-6072	134	15	,	,	PUNCT
cana-6072	134	16	probability	probability	NOUN
cana-6072	134	17	averaging	averaging	NOUN
cana-6072	134	18	or	or	CCONJ
cana-6072	134	19	a	a	DET
cana-6072	134	20	meta	meta	ADJ
cana-6072	134	21	-	-	PUNCT
cana-6072	134	22	classifier	classifier	NOUN
cana-6072	134	23	)	)	PUNCT
cana-6072	134	24	.	.	PUNCT
cana-6072	135	1	this	this	DET
cana-6072	135	2	modular	modular	ADJ
cana-6072	135	3	design	design	NOUN
cana-6072	135	4	is	be	AUX
cana-6072	135	5	robust	robust	ADJ
cana-6072	135	6	to	to	ADP
cana-6072	135	7	missing	miss	VERB
cana-6072	135	8	modalities	modality	NOUN
cana-6072	135	9	but	but	CCONJ
cana-6072	135	10	can	can	AUX
cana-6072	135	11	not	not	PART
cana-6072	135	12	capture	capture	VERB
cana-6072	135	13	nuanced	nuanced	ADJ
cana-6072	135	14	cross	cross	ADJ
cana-6072	135	15	-	-	ADJ
cana-6072	135	16	modal	modal	ADJ
cana-6072	135	17	interactions	interaction	NOUN
cana-6072	135	18	.	.	PUNCT
cana-6072	136	1	explicit	explicit	ADJ
cana-6072	136	2	late	late	ADJ
cana-6072	136	3	-	-	PUNCT
cana-6072	136	4	fusion	fusion	NOUN
cana-6072	136	5	examples	example	NOUN
cana-6072	136	6	are	be	AUX
cana-6072	136	7	rarer	rare	ADJ
cana-6072	136	8	in	in	ADP
cana-6072	136	9	emoji	emoji	PROPN
cana-6072	136	10	sentiment	sentiment	NOUN
cana-6072	136	11	specifically	specifically	ADV
cana-6072	136	12	;	;	PUNCT
cana-6072	136	13	they	they	PRON
cana-6072	136	14	are	be	AUX
cana-6072	136	15	more	more	ADV
cana-6072	136	16	common	common	ADJ
cana-6072	136	17	in	in	ADP
cana-6072	136	18	multimodal	multimodal	ADJ
cana-6072	136	19	sentiment	sentiment	NOUN
cana-6072	136	20	involving	involve	VERB
cana-6072	136	21	images	image	NOUN
cana-6072	136	22	[	[	X
cana-6072	136	23	huang	huang	PROPN
cana-6072	136	24	,	,	PUNCT
cana-6072	136	25	f.	f.	PROPN
cana-6072	136	26	et	et	PROPN
cana-6072	136	27	al	al	PROPN
cana-6072	136	28	.	.	PROPN
cana-6072	136	29	,	,	PUNCT
cana-6072	136	30	2020	2020	NUM
cana-6072	136	31	]	]	PUNCT
cana-6072	136	32	.	.	PUNCT
cana-6072	137	1	intermediate	intermediate	ADJ
cana-6072	137	2	fusion	fusion	NOUN
cana-6072	137	3	(	(	PUNCT
cana-6072	137	4	attention	attention	NOUN
cana-6072	137	5	)	)	PUNCT
cana-6072	137	6	.	.	PUNCT
cana-6072	138	1	merge	merge	VERB
cana-6072	138	2	modalities	modality	NOUN
cana-6072	138	3	via	via	ADP
cana-6072	138	4	attention	attention	NOUN
cana-6072	138	5	mechanisms	mechanism	NOUN
cana-6072	138	6	that	that	PRON
cana-6072	138	7	learn	learn	VERB
cana-6072	138	8	crossmodal	crossmodal	ADJ
cana-6072	138	9	alignments	alignment	NOUN
cana-6072	138	10	.	.	PUNCT
cana-6072	139	1	recent	recent	ADJ
cana-6072	139	2	works	work	NOUN
cana-6072	139	3	adopt	adopt	VERB
cana-6072	139	4	this	this	PRON
cana-6072	139	5	extensively	extensively	ADV
cana-6072	139	6	;	;	PUNCT
cana-6072	139	7	e.g.	e.g.	ADV
cana-6072	139	8	,	,	PUNCT
cana-6072	139	9	lou	lou	PROPN
cana-6072	139	10	et	et	PROPN
cana-6072	139	11	al	al	PROPN
cana-6072	139	12	.	.	PROPN
cana-6072	139	13	apply	apply	VERB
cana-6072	139	14	emoji	emoji	NOUN
cana-6072	139	15	-	-	PUNCT
cana-6072	139	16	based	base	VERB
cana-6072	139	17	attention	attention	NOUN
cana-6072	139	18	to	to	ADP
cana-6072	139	19	weight	weight	NOUN
cana-6072	139	20	words	word	NOUN
cana-6072	139	21	depending	depend	VERB
cana-6072	139	22	on	on	ADP
cana-6072	139	23	emoji	emoji	PROPN
cana-6072	139	24	presence	presence	NOUN
cana-6072	140	1	[	[	X
cana-6072	140	2	lou	lou	PROPN
cana-6072	140	3	,	,	PUNCT
cana-6072	140	4	y.	y.	PROPN
cana-6072	140	5	et	et	PROPN
cana-6072	140	6	al	al	PROPN
cana-6072	140	7	.	.	PROPN
cana-6072	140	8	,	,	PUNCT
cana-6072	140	9	2020	2020	NUM
cana-6072	140	10	]	]	PUNCT
cana-6072	140	11	,	,	PUNCT
cana-6072	140	12	while	while	SCONJ
cana-6072	140	13	tang	tang	PROPN
cana-6072	140	14	et	et	PROPN
cana-6072	140	15	al	al	PROPN
cana-6072	140	16	.	.	PUNCT
cana-6072	141	1	use	use	VERB
cana-6072	141	2	cross	cross	NOUN
cana-6072	141	3	-	-	NOUN
cana-6072	141	4	attention	attention	NOUN
cana-6072	141	5	to	to	PART
cana-6072	141	6	align	align	VERB
cana-6072	141	7	emoji	emoji	PROPN
cana-6072	141	8	semantics	semantic	NOUN
cana-6072	141	9	with	with	ADP
cana-6072	141	10	text	text	NOUN
cana-6072	141	11	[	[	X
cana-6072	141	12	tang	tang	NOUN
cana-6072	141	13	,	,	PUNCT
cana-6072	141	14	h.	h.	PROPN
cana-6072	141	15	et	et	PROPN
cana-6072	141	16	al	al	PROPN
cana-6072	141	17	.	.	PROPN
cana-6072	141	18	,	,	PUNCT
cana-6072	141	19	2024	2024	NUM
cana-6072	141	20	]	]	PUNCT
cana-6072	141	21	.	.	PUNCT
cana-6072	142	1	such	such	ADJ
cana-6072	142	2	models	model	NOUN
cana-6072	142	3	capture	capture	VERB
cana-6072	142	4	subtle	subtle	ADJ
cana-6072	142	5	interactions	interaction	NOUN
cana-6072	142	6	and	and	CCONJ
cana-6072	142	7	often	often	ADV
cana-6072	142	8	yield	yield	VERB
cana-6072	142	9	the	the	DET
cana-6072	142	10	best	good	ADJ
cana-6072	142	11	results	result	NOUN
cana-6072	142	12	,	,	PUNCT
cana-6072	142	13	though	though	SCONJ
cana-6072	142	14	with	with	ADP
cana-6072	142	15	increased	increase	VERB
cana-6072	142	16	training	training	NOUN
cana-6072	142	17	cost	cost	NOUN
cana-6072	142	18	and	and	CCONJ
cana-6072	142	19	data	datum	NOUN
cana-6072	142	20	requirements	requirement	NOUN
cana-6072	142	21	.	.	PUNCT
cana-6072	143	1	hybrid	hybrid	ADJ
cana-6072	143	2	fusion	fusion	NOUN
cana-6072	143	3	.	.	PUNCT
cana-6072	144	1	combine	combine	VERB
cana-6072	144	2	multiple	multiple	ADJ
cana-6072	144	3	fusion	fusion	NOUN
cana-6072	144	4	approaches	approach	NOUN
cana-6072	144	5	or	or	CCONJ
cana-6072	144	6	stages	stage	NOUN
cana-6072	144	7	.	.	PUNCT
cana-6072	145	1	some	some	DET
cana-6072	145	2	models	model	NOUN
cana-6072	145	3	first	first	ADV
cana-6072	145	4	perform	perform	VERB
cana-6072	145	5	early	early	ADJ
cana-6072	145	6	feature	feature	NOUN
cana-6072	145	7	concatenation	concatenation	NOUN
cana-6072	145	8	,	,	PUNCT
cana-6072	145	9	then	then	ADV
cana-6072	145	10	apply	apply	VERB
cana-6072	145	11	attention	attention	NOUN
cana-6072	145	12	,	,	PUNCT
cana-6072	145	13	or	or	CCONJ
cana-6072	145	14	ensemble	ensemble	VERB
cana-6072	145	15	different	different	ADJ
cana-6072	145	16	fusion	fusion	NOUN
cana-6072	145	17	models	model	NOUN
cana-6072	145	18	.	.	PUNCT
cana-6072	146	1	for	for	ADP
cana-6072	146	2	instance	instance	NOUN
cana-6072	146	3	,	,	PUNCT
cana-6072	146	4	gan	gan	PROPN
cana-6072	146	5	et	et	PROPN
cana-6072	146	6	al	al	PROPN
cana-6072	146	7	.	.	PUNCT
cana-6072	147	1	propose	propose	VERB
cana-6072	147	2	a	a	DET
cana-6072	147	3	multimodal	multimodal	ADJ
cana-6072	147	4	fusion	fusion	NOUN
cana-6072	147	5	network	network	NOUN
cana-6072	147	6	(	(	PUNCT
cana-6072	147	7	mfn	mfn	NOUN
cana-6072	147	8	)	)	PUNCT
cana-6072	147	9	with	with	ADP
cana-6072	147	10	both	both	CCONJ
cana-6072	147	11	multi	multi	ADJ
cana-6072	147	12	-	-	ADJ
cana-6072	147	13	head	head	ADJ
cana-6072	147	14	selfattention	selfattention	NOUN
cana-6072	147	15	and	and	CCONJ
cana-6072	147	16	feature	feature	NOUN
cana-6072	147	17	concatenation	concatenation	NOUN
cana-6072	147	18	,	,	PUNCT
cana-6072	147	19	effectively	effectively	ADV
cana-6072	147	20	a	a	DET
cana-6072	147	21	hybrid	hybrid	ADJ
cana-6072	147	22	design	design	NOUN
cana-6072	147	23	[	[	X
cana-6072	147	24	gan	gan	NOUN
cana-6072	147	25	,	,	PUNCT
cana-6072	147	26	c.	c.	PROPN
cana-6072	147	27	et	et	PROPN
cana-6072	147	28	al	al	PROPN
cana-6072	147	29	.	.	PROPN
cana-6072	147	30	,	,	PUNCT
cana-6072	147	31	2024	2024	NUM
cana-6072	147	32	]	]	PUNCT
cana-6072	147	33	.	.	PUNCT
cana-6072	148	1	these	these	PRON
cana-6072	148	2	maximize	maximize	VERB
cana-6072	148	3	information	information	NOUN
cana-6072	148	4	use	use	NOUN
cana-6072	148	5	but	but	CCONJ
cana-6072	148	6	are	be	AUX
cana-6072	148	7	complex	complex	ADJ
cana-6072	148	8	to	to	PART
cana-6072	148	9	train	train	VERB
cana-6072	148	10	and	and	CCONJ
cana-6072	148	11	interpret	interpret	VERB
cana-6072	148	12	.	.	PUNCT
cana-6072	149	1	figure	figure	NOUN
cana-6072	149	2	4	4	NUM
cana-6072	149	3	.	.	NOUN
cana-6072	149	4	representative	representative	ADJ
cana-6072	149	5	fusion	fusion	NOUN
cana-6072	149	6	schematics	schematic	NOUN
cana-6072	149	7	:	:	PUNCT
cana-6072	149	8	early	early	ADJ
cana-6072	149	9	(	(	PUNCT
cana-6072	149	10	feature	feature	NOUN
cana-6072	149	11	-	-	PUNCT
cana-6072	149	12	level	level	NOUN
cana-6072	149	13	)	)	PUNCT
cana-6072	149	14	,	,	PUNCT
cana-6072	149	15	late	late	ADV
cana-6072	149	16	(	(	PUNCT
cana-6072	149	17	decision	decision	NOUN
cana-6072	149	18	-	-	PUNCT
cana-6072	149	19	level	level	NOUN
cana-6072	149	20	)	)	PUNCT
cana-6072	149	21	,	,	PUNCT
cana-6072	149	22	intermediate	intermediate	ADJ
cana-6072	149	23	(	(	PUNCT
cana-6072	149	24	attention	attention	NOUN
cana-6072	149	25	)	)	PUNCT
cana-6072	149	26	,	,	PUNCT
cana-6072	149	27	and	and	CCONJ
cana-6072	149	28	hybrid	hybrid	NOUN
cana-6072	149	29	.	.	PUNCT
cana-6072	150	1	communications	communication	NOUN
cana-6072	150	2	on	on	ADP
cana-6072	150	3	applied	apply	VERB
cana-6072	150	4	nonlinear	nonlinear	ADJ
cana-6072	150	5	analysis	analysis	NOUN
cana-6072	150	6	issn	issn	NOUN
cana-6072	150	7	:	:	PUNCT
cana-6072	150	8	1074	1074	NUM
cana-6072	150	9	-	-	PUNCT
cana-6072	150	10	133x	133x	NUM
cana-6072	150	11	vol	vol	NOUN
cana-6072	150	12	32	32	NUM
cana-6072	150	13	no	no	NOUN
cana-6072	150	14	.	.	PUNCT
cana-6072	151	1	4s	4s	NUM
cana-6072	151	2	(	(	PUNCT
cana-6072	151	3	2025	2025	NUM
cana-6072	151	4	)	)	PUNCT
cana-6072	151	5	765	765	NUM
cana-6072	151	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	151	7	datasets	dataset	NOUN
cana-6072	151	8	&	&	CCONJ
cana-6072	151	9	metrics	metric	NOUN
cana-6072	151	10	a	a	DET
cana-6072	151	11	wide	wide	ADJ
cana-6072	151	12	range	range	NOUN
cana-6072	151	13	of	of	ADP
cana-6072	151	14	datasets	dataset	NOUN
cana-6072	151	15	has	have	AUX
cana-6072	151	16	been	be	AUX
cana-6072	151	17	employed	employ	VERB
cana-6072	151	18	for	for	ADP
cana-6072	151	19	emoji	emoji	NOUN
cana-6072	151	20	-	-	PUNCT
cana-6072	151	21	aware	aware	ADJ
cana-6072	151	22	sentiment	sentiment	NOUN
cana-6072	151	23	analysis	analysis	NOUN
cana-6072	151	24	,	,	PUNCT
cana-6072	151	25	with	with	ADP
cana-6072	151	26	twitterbased	twitterbase	VERB
cana-6072	151	27	corpora	corpora	PROPN
cana-6072	151	28	dominating	dominate	VERB
cana-6072	151	29	the	the	DET
cana-6072	151	30	field	field	NOUN
cana-6072	151	31	due	due	ADP
cana-6072	151	32	to	to	ADP
cana-6072	151	33	the	the	DET
cana-6072	151	34	prevalence	prevalence	NOUN
cana-6072	151	35	of	of	ADP
cana-6072	151	36	emojis	emoji	NOUN
cana-6072	151	37	in	in	ADP
cana-6072	151	38	microblogging	microblogge	VERB
cana-6072	151	39	contexts	contexts	NOUN
cana-6072	151	40	.	.	PUNCT
cana-6072	152	1	datasets	dataset	NOUN
cana-6072	152	2	such	such	ADJ
cana-6072	152	3	as	as	ADP
cana-6072	152	4	the	the	DET
cana-6072	152	5	twitter	twitter	NOUN
cana-6072	152	6	airline	airline	NOUN
cana-6072	152	7	corpus	corpu	NOUN
cana-6072	152	8	provide	provide	VERB
cana-6072	152	9	short	short	ADJ
cana-6072	152	10	english	english	ADJ
cana-6072	152	11	-	-	PUNCT
cana-6072	152	12	language	language	NOUN
cana-6072	152	13	reviews	review	NOUN
cana-6072	152	14	with	with	ADP
cana-6072	152	15	frequent	frequent	ADJ
cana-6072	152	16	use	use	NOUN
cana-6072	152	17	of	of	ADP
cana-6072	152	18	emojis	emoji	NOUN
cana-6072	152	19	in	in	ADP
cana-6072	152	20	complaints	complaint	NOUN
cana-6072	152	21	and	and	CCONJ
cana-6072	152	22	praise	praise	NOUN
cana-6072	152	23	,	,	PUNCT
cana-6072	152	24	and	and	CCONJ
cana-6072	152	25	are	be	AUX
cana-6072	152	26	commonly	commonly	ADV
cana-6072	152	27	used	use	VERB
cana-6072	152	28	benchmarks	benchmark	NOUN
cana-6072	152	29	for	for	ADP
cana-6072	152	30	threeclass	threeclass	ADJ
cana-6072	152	31	sentiment	sentiment	NOUN
cana-6072	152	32	tasks	task	NOUN
cana-6072	152	33	.	.	PUNCT
cana-6072	153	1	similarly	similarly	ADV
cana-6072	153	2	,	,	PUNCT
cana-6072	153	3	sentiment140	sentiment140	PROPN
cana-6072	153	4	remains	remain	VERB
cana-6072	153	5	one	one	NUM
cana-6072	153	6	of	of	ADP
cana-6072	153	7	the	the	DET
cana-6072	153	8	largest	large	ADJ
cana-6072	153	9	and	and	CCONJ
cana-6072	153	10	most	most	ADV
cana-6072	153	11	-	-	PUNCT
cana-6072	153	12	cited	cite	VERB
cana-6072	153	13	resources	resource	NOUN
cana-6072	153	14	,	,	PUNCT
cana-6072	153	15	containing	contain	VERB
cana-6072	153	16	∼1.6	∼1.6	NOUN
cana-6072	153	17	million	million	NUM
cana-6072	153	18	tweets	tweet	NOUN
cana-6072	153	19	automatically	automatically	ADV
cana-6072	153	20	labeled	label	VERB
cana-6072	153	21	via	via	ADP
cana-6072	153	22	emoticons	emoticon	NOUN
cana-6072	153	23	.	.	PUNCT
cana-6072	154	1	although	although	SCONJ
cana-6072	154	2	it	it	PRON
cana-6072	154	3	includes	include	VERB
cana-6072	154	4	emojis	emoji	NOUN
cana-6072	154	5	,	,	PUNCT
cana-6072	154	6	its	its	PRON
cana-6072	154	7	distant	distant	ADJ
cana-6072	154	8	-	-	PUNCT
cana-6072	154	9	supervision	supervision	NOUN
cana-6072	154	10	labeling	labeling	NOUN
cana-6072	154	11	introduces	introduce	VERB
cana-6072	154	12	potential	potential	ADJ
cana-6072	154	13	noise	noise	NOUN
cana-6072	154	14	and	and	CCONJ
cana-6072	154	15	train	train	NOUN
cana-6072	154	16	–	–	PUNCT
cana-6072	154	17	test	test	NOUN
cana-6072	154	18	leakage	leakage	NOUN
cana-6072	154	19	if	if	SCONJ
cana-6072	154	20	not	not	PART
cana-6072	154	21	carefully	carefully	ADV
cana-6072	154	22	managed	manage	VERB
cana-6072	154	23	.	.	PUNCT
cana-6072	155	1	beyond	beyond	ADP
cana-6072	155	2	twitter	twitter	NOUN
cana-6072	155	3	,	,	PUNCT
cana-6072	155	4	several	several	ADJ
cana-6072	155	5	studies	study	NOUN
cana-6072	155	6	draw	draw	VERB
cana-6072	155	7	on	on	ADP
cana-6072	155	8	chinese	chinese	ADJ
cana-6072	155	9	microblog	microblog	NOUN
cana-6072	155	10	data	datum	NOUN
cana-6072	155	11	from	from	ADP
cana-6072	155	12	sina	sina	PROPN
cana-6072	155	13	weibo	weibo	PROPN
cana-6072	155	14	,	,	PUNCT
cana-6072	155	15	where	where	SCONJ
cana-6072	155	16	emojis	emoji	NOUN
cana-6072	155	17	are	be	AUX
cana-6072	155	18	often	often	ADV
cana-6072	155	19	appended	append	VERB
cana-6072	155	20	at	at	ADP
cana-6072	155	21	the	the	DET
cana-6072	155	22	end	end	NOUN
cana-6072	155	23	of	of	ADP
cana-6072	155	24	posts	post	NOUN
cana-6072	155	25	.	.	PUNCT
cana-6072	156	1	these	these	DET
cana-6072	156	2	datasets	dataset	NOUN
cana-6072	156	3	typically	typically	ADV
cana-6072	156	4	involve	involve	VERB
cana-6072	156	5	polarity	polarity	NOUN
cana-6072	156	6	classification	classification	NOUN
cana-6072	156	7	but	but	CCONJ
cana-6072	156	8	are	be	AUX
cana-6072	156	9	known	know	VERB
cana-6072	156	10	to	to	PART
cana-6072	156	11	suffer	suffer	VERB
cana-6072	156	12	from	from	ADP
cana-6072	156	13	class	class	NOUN
cana-6072	156	14	imbalance	imbalance	NOUN
cana-6072	156	15	and	and	CCONJ
cana-6072	156	16	often	often	ADV
cana-6072	156	17	require	require	VERB
cana-6072	156	18	manual	manual	ADJ
cana-6072	156	19	labeling	labeling	NOUN
cana-6072	156	20	.	.	PUNCT
cana-6072	157	1	more	more	ADV
cana-6072	157	2	fine	fine	ADV
cana-6072	157	3	-	-	PUNCT
cana-6072	157	4	grained	grain	VERB
cana-6072	157	5	emotion	emotion	NOUN
cana-6072	157	6	corpora	corpora	VERB
cana-6072	157	7	such	such	ADJ
cana-6072	157	8	as	as	ADP
cana-6072	157	9	goemotions	goemotion	NOUN
cana-6072	157	10	(	(	PUNCT
cana-6072	157	11	reddit	reddit	NOUN
cana-6072	157	12	)	)	PUNCT
cana-6072	157	13	have	have	AUX
cana-6072	157	14	also	also	ADV
cana-6072	157	15	been	be	AUX
cana-6072	157	16	used	use	VERB
cana-6072	157	17	,	,	PUNCT
cana-6072	157	18	although	although	SCONJ
cana-6072	157	19	their	their	PRON
cana-6072	157	20	relatively	relatively	ADV
cana-6072	157	21	low	low	ADJ
cana-6072	157	22	emoji	emoji	NOUN
cana-6072	157	23	frequency	frequency	NOUN
cana-6072	157	24	makes	make	VERB
cana-6072	157	25	them	they	PRON
cana-6072	157	26	less	less	ADV
cana-6072	157	27	suitable	suitable	ADJ
cana-6072	157	28	for	for	ADP
cana-6072	157	29	directly	directly	ADV
cana-6072	157	30	assessing	assess	VERB
cana-6072	157	31	emoji	emoji	ADJ
cana-6072	157	32	-	-	PUNCT
cana-6072	157	33	specific	specific	ADJ
cana-6072	157	34	modeling	modeling	NOUN
cana-6072	157	35	.	.	PUNCT
cana-6072	158	1	table	table	NOUN
cana-6072	158	2	4	4	NUM
cana-6072	158	3	.	.	PUNCT
cana-6072	159	1	datasets	dataset	NOUN
cana-6072	159	2	commonly	commonly	ADV
cana-6072	159	3	used	use	VERB
cana-6072	159	4	for	for	ADP
cana-6072	159	5	emoji+text	emoji+text	PROPN
cana-6072	159	6	sentiment	sentiment	NOUN
cana-6072	159	7	dataset	dataset	NOUN
cana-6072	159	8	domain	domain	NOUN
cana-6072	159	9	/	/	SYM
cana-6072	159	10	language	language	NOUN
cana-6072	159	11	emoji	emoji	NOUN
cana-6072	159	12	prevalence	prevalence	NOUN
cana-6072	159	13	typical	typical	ADJ
cana-6072	159	14	task	task	PROPN
cana-6072	159	15	licensing	licensing	NOUN
cana-6072	159	16	/	/	SYM
cana-6072	159	17	access	access	NOUN
cana-6072	159	18	twitter	twitter	NOUN
cana-6072	159	19	airline	airline	NOUN
cana-6072	159	20	twitter	twitter	NOUN
cana-6072	159	21	(	(	PUNCT
cana-6072	159	22	en	en	X
cana-6072	159	23	)	)	PUNCT
cana-6072	159	24	,	,	PUNCT
cana-6072	159	25	airline	airline	NOUN
cana-6072	159	26	reviews	review	NOUN
cana-6072	159	27	med	me	VERB
cana-6072	159	28	–	–	PUNCT
cana-6072	159	29	high	high	ADJ
cana-6072	159	30	(	(	PUNCT
cana-6072	159	31	frequent	frequent	ADJ
cana-6072	159	32	emojis	emoji	NOUN
cana-6072	159	33	in	in	ADP
cana-6072	159	34	complaints	complaint	NOUN
cana-6072	159	35	/	/	SYM
cana-6072	159	36	praise	praise	NOUN
cana-6072	159	37	)	)	PUNCT
cana-6072	159	38	3	3	NUM
cana-6072	159	39	-	-	PUNCT
cana-6072	159	40	class	class	NOUN
cana-6072	159	41	polarity	polarity	NOUN
cana-6072	159	42	(	(	PUNCT
cana-6072	159	43	pos	pos	NOUN
cana-6072	159	44	/	/	SYM
cana-6072	159	45	neg	neg	PROPN
cana-6072	159	46	/	/	SYM
cana-6072	159	47	neu	neu	NOUN
cana-6072	159	48	)	)	PUNCT
cana-6072	159	49	public	public	NOUN
cana-6072	159	50	(	(	PUNCT
cana-6072	159	51	e.g.	e.g.	ADV
cana-6072	159	52	,	,	PUNCT
cana-6072	159	53	kaggle	kaggle	NOUN
cana-6072	159	54	)	)	PUNCT
cana-6072	159	55	;	;	PUNCT
cana-6072	159	56	must	must	AUX
cana-6072	159	57	respect	respect	VERB
cana-6072	159	58	twitter	twitter	NOUN
cana-6072	159	59	tos	tos	NOUN
cana-6072	159	60	sentiment140	sentiment140	PROPN
cana-6072	159	61	twitter	twitter	NOUN
cana-6072	159	62	(	(	PUNCT
cana-6072	159	63	en	en	X
cana-6072	159	64	)	)	PUNCT
cana-6072	159	65	,	,	PUNCT
cana-6072	159	66	general	general	ADJ
cana-6072	159	67	tweets	tweet	NOUN
cana-6072	159	68	medium	medium	NOUN
cana-6072	159	69	(	(	PUNCT
cana-6072	159	70	emoticons	emoticon	NOUN
cana-6072	159	71	replaced	replace	VERB
cana-6072	159	72	by	by	ADP
cana-6072	159	73	labels	label	NOUN
cana-6072	159	74	during	during	ADP
cana-6072	159	75	distant	distant	ADJ
cana-6072	159	76	supervision	supervision	NOUN
cana-6072	159	77	)	)	PUNCT
cana-6072	159	78	2	2	NUM
cana-6072	159	79	-	-	PUNCT
cana-6072	159	80	class	class	NOUN
cana-6072	159	81	polarity	polarity	NOUN
cana-6072	159	82	(	(	PUNCT
cana-6072	159	83	pos	pos	NOUN
cana-6072	159	84	/	/	SYM
cana-6072	159	85	neg	neg	ADJ
cana-6072	159	86	)	)	PUNCT
cana-6072	159	87	public	public	NOUN
cana-6072	159	88	(	(	PUNCT
cana-6072	159	89	distantsupervised	distantsupervised	ADJ
cana-6072	159	90	)	)	PUNCT
cana-6072	159	91	weibo	weibo	PROPN
cana-6072	159	92	sina	sina	PROPN
cana-6072	159	93	weibo	weibo	PROPN
cana-6072	159	94	(	(	PUNCT
cana-6072	159	95	zh	zh	PROPN
cana-6072	159	96	-	-	PUNCT
cana-6072	159	97	cn	cn	PROPN
cana-6072	159	98	)	)	PUNCT
cana-6072	159	99	,	,	PUNCT
cana-6072	159	100	microblogs	microblog	NOUN
cana-6072	159	101	high	high	ADV
cana-6072	159	102	(	(	PUNCT
cana-6072	159	103	emojis	emoji	NOUN
cana-6072	159	104	often	often	ADV
cana-6072	159	105	at	at	ADP
cana-6072	159	106	end	end	NOUN
cana-6072	159	107	of	of	ADP
cana-6072	159	108	post	post	NOUN
cana-6072	159	109	)	)	PUNCT
cana-6072	159	110	polarity	polarity	NOUN
cana-6072	159	111	(	(	PUNCT
cana-6072	159	112	pos	pos	NOUN
cana-6072	159	113	/	/	SYM
cana-6072	159	114	neg	neg	PROPN
cana-6072	159	115	/	/	SYM
cana-6072	159	116	neu	neu	NOUN
cana-6072	159	117	)	)	PUNCT
cana-6072	159	118	semi	semi	ADJ
cana-6072	159	119	-	-	ADJ
cana-6072	159	120	public	public	ADJ
cana-6072	159	121	(	(	PUNCT
cana-6072	159	122	varies	varie	NOUN
cana-6072	159	123	by	by	ADP
cana-6072	159	124	source	source	NOUN
cana-6072	159	125	)	)	PUNCT
cana-6072	159	126	goemotions	goemotion	NOUN
cana-6072	159	127	reddit	reddit	NOUN
cana-6072	159	128	(	(	PUNCT
cana-6072	159	129	en	en	X
cana-6072	159	130	)	)	PUNCT
cana-6072	159	131	,	,	PUNCT
cana-6072	159	132	general	general	ADJ
cana-6072	159	133	content	content	NOUN
cana-6072	159	134	low	low	ADJ
cana-6072	159	135	(	(	PUNCT
cana-6072	159	136	few	few	ADJ
cana-6072	159	137	emojis	emoji	NOUN
cana-6072	159	138	in	in	ADP
cana-6072	159	139	reddit	reddit	NOUN
cana-6072	159	140	text	text	NOUN
cana-6072	159	141	)	)	PUNCT
cana-6072	159	142	27	27	NUM
cana-6072	159	143	-	-	PUNCT
cana-6072	159	144	class	class	NOUN
cana-6072	159	145	emotions	emotion	NOUN
cana-6072	159	146	(	(	PUNCT
cana-6072	159	147	multi	multi	ADJ
cana-6072	159	148	-	-	ADJ
cana-6072	159	149	label	label	ADJ
cana-6072	159	150	)	)	PUNCT
cana-6072	159	151	public	public	NOUN
cana-6072	159	152	(	(	PUNCT
cana-6072	159	153	google	google	PROPN
cana-6072	159	154	ai	ai	NOUN
cana-6072	159	155	,	,	PUNCT
cana-6072	159	156	apache-2.0	apache-2.0	ADJ
cana-6072	159	157	)	)	PUNCT
cana-6072	159	158	evaluation	evaluation	NOUN
cana-6072	159	159	metrics	metric	NOUN
cana-6072	159	160	largely	largely	ADV
cana-6072	159	161	mirror	mirror	VERB
cana-6072	159	162	broader	broad	ADJ
cana-6072	159	163	sentiment	sentiment	NOUN
cana-6072	159	164	research	research	NOUN
cana-6072	159	165	.	.	PUNCT
cana-6072	160	1	accuracy	accuracy	NOUN
cana-6072	160	2	is	be	AUX
cana-6072	160	3	frequently	frequently	ADV
cana-6072	160	4	reported	report	VERB
cana-6072	160	5	,	,	PUNCT
cana-6072	160	6	but	but	CCONJ
cana-6072	160	7	macro	macro	ADJ
cana-6072	160	8	-	-	ADJ
cana-6072	160	9	averaged	average	VERB
cana-6072	160	10	f1	f1	NOUN
cana-6072	160	11	has	have	AUX
cana-6072	160	12	emerged	emerge	VERB
cana-6072	160	13	as	as	ADP
cana-6072	160	14	the	the	DET
cana-6072	160	15	preferred	preferred	ADJ
cana-6072	160	16	metric	metric	NOUN
cana-6072	160	17	in	in	ADP
cana-6072	160	18	class	class	NOUN
cana-6072	160	19	-	-	PUNCT
cana-6072	160	20	imbalanced	imbalance	VERB
cana-6072	160	21	scenarios	scenario	NOUN
cana-6072	160	22	(	(	PUNCT
cana-6072	160	23	e.g.	e.g.	ADV
cana-6072	160	24	,	,	PUNCT
cana-6072	160	25	when	when	SCONJ
cana-6072	160	26	neutral	neutral	ADJ
cana-6072	160	27	posts	post	NOUN
cana-6072	160	28	dominate	dominate	VERB
cana-6072	160	29	)	)	PUNCT
cana-6072	160	30	.	.	PUNCT
cana-6072	161	1	macro	macro	ADJ
cana-6072	161	2	-	-	NOUN
cana-6072	161	3	f1	f1	NOUN
cana-6072	161	4	better	well	ADV
cana-6072	161	5	reflects	reflect	VERB
cana-6072	161	6	performance	performance	NOUN
cana-6072	161	7	across	across	ADP
cana-6072	161	8	minority	minority	NOUN
cana-6072	161	9	classes	class	NOUN
cana-6072	161	10	,	,	PUNCT
cana-6072	161	11	whereas	whereas	SCONJ
cana-6072	161	12	accuracy	accuracy	NOUN
cana-6072	161	13	can	can	AUX
cana-6072	161	14	be	be	AUX
cana-6072	161	15	misleading	misleading	ADJ
cana-6072	161	16	if	if	SCONJ
cana-6072	161	17	a	a	DET
cana-6072	161	18	model	model	NOUN
cana-6072	161	19	predicts	predict	VERB
cana-6072	161	20	the	the	DET
cana-6072	161	21	majority	majority	NOUN
cana-6072	161	22	class	class	NOUN
cana-6072	161	23	.	.	PUNCT
cana-6072	162	1	several	several	ADJ
cana-6072	162	2	studies	study	NOUN
cana-6072	162	3	also	also	ADV
cana-6072	162	4	report	report	VERB
cana-6072	162	5	per	per	ADP
cana-6072	162	6	-	-	PUNCT
cana-6072	162	7	class	class	NOUN
cana-6072	162	8	precision	precision	NOUN
cana-6072	162	9	/	/	SYM
cana-6072	162	10	recall	recall	NOUN
cana-6072	162	11	,	,	PUNCT
cana-6072	162	12	though	though	SCONJ
cana-6072	162	13	detailed	detailed	ADJ
cana-6072	162	14	error	error	NOUN
cana-6072	162	15	analyses	analysis	NOUN
cana-6072	162	16	remain	remain	VERB
cana-6072	162	17	limited	limited	ADJ
cana-6072	162	18	.	.	PUNCT
cana-6072	163	1	a	a	DET
cana-6072	163	2	notable	notable	ADJ
cana-6072	163	3	shortcoming	shortcoming	NOUN
cana-6072	163	4	is	be	AUX
cana-6072	163	5	the	the	DET
cana-6072	163	6	scarcity	scarcity	NOUN
cana-6072	163	7	of	of	ADP
cana-6072	163	8	statistical	statistical	ADJ
cana-6072	163	9	testing	testing	NOUN
cana-6072	163	10	(	(	PUNCT
cana-6072	163	11	e.g.	e.g.	ADV
cana-6072	163	12	,	,	PUNCT
cana-6072	163	13	confidence	confidence	NOUN
cana-6072	163	14	intervals	interval	NOUN
cana-6072	163	15	,	,	PUNCT
cana-6072	163	16	paired	pair	VERB
cana-6072	163	17	tests	test	NOUN
cana-6072	163	18	across	across	ADP
cana-6072	163	19	runs	run	NOUN
cana-6072	163	20	)	)	PUNCT
cana-6072	163	21	.	.	PUNCT
cana-6072	164	1	for	for	ADP
cana-6072	164	2	greater	great	ADJ
cana-6072	164	3	rigor	rigor	NOUN
cana-6072	164	4	,	,	PUNCT
cana-6072	164	5	future	future	ADJ
cana-6072	164	6	work	work	NOUN
cana-6072	164	7	should	should	AUX
cana-6072	164	8	report	report	VERB
cana-6072	164	9	bootstrap	bootstrap	NOUN
cana-6072	164	10	cis	cis	NOUN
cana-6072	164	11	for	for	ADP
cana-6072	164	12	f1	f1	NOUN
cana-6072	164	13	/	/	SYM
cana-6072	164	14	accuracy	accuracy	NOUN
cana-6072	164	15	and	and	CCONJ
cana-6072	164	16	use	use	VERB
cana-6072	164	17	paired	pair	VERB
cana-6072	164	18	tests	test	NOUN
cana-6072	164	19	to	to	PART
cana-6072	164	20	establish	establish	VERB
cana-6072	164	21	whether	whether	SCONJ
cana-6072	164	22	improvements	improvement	NOUN
cana-6072	164	23	are	be	AUX
cana-6072	164	24	statistically	statistically	ADV
cana-6072	164	25	meaningful	meaningful	ADJ
cana-6072	164	26	rather	rather	ADV
cana-6072	164	27	than	than	ADP
cana-6072	164	28	due	due	ADJ
cana-6072	164	29	to	to	ADP
cana-6072	164	30	chance	chance	NOUN
cana-6072	164	31	.	.	PUNCT
cana-6072	165	1	as	as	SCONJ
cana-6072	165	2	summarized	summarize	VERB
cana-6072	165	3	in	in	ADP
cana-6072	165	4	table	table	NOUN
cana-6072	165	5	4	4	NUM
cana-6072	165	6	,	,	PUNCT
cana-6072	165	7	frequently	frequently	ADV
cana-6072	165	8	used	use	VERB
cana-6072	165	9	datasets	dataset	NOUN
cana-6072	165	10	vary	vary	VERB
cana-6072	165	11	in	in	ADP
cana-6072	165	12	emoji	emoji	PROPN
cana-6072	165	13	prevalence	prevalence	NOUN
cana-6072	165	14	,	,	PUNCT
cana-6072	165	15	label	label	NOUN
cana-6072	165	16	schemes	scheme	NOUN
cana-6072	165	17	,	,	PUNCT
cana-6072	165	18	and	and	CCONJ
cana-6072	165	19	licensing	licensing	NOUN
cana-6072	165	20	/	/	SYM
cana-6072	165	21	access	access	NOUN
cana-6072	165	22	.	.	PUNCT
cana-6072	166	1	communications	communication	NOUN
cana-6072	166	2	on	on	ADP
cana-6072	166	3	applied	apply	VERB
cana-6072	166	4	nonlinear	nonlinear	ADJ
cana-6072	166	5	analysis	analysis	NOUN
cana-6072	166	6	issn	issn	NOUN
cana-6072	166	7	:	:	PUNCT
cana-6072	166	8	1074	1074	NUM
cana-6072	166	9	-	-	PUNCT
cana-6072	166	10	133x	133x	NUM
cana-6072	166	11	vol	vol	NOUN
cana-6072	166	12	32	32	NUM
cana-6072	166	13	no	no	NOUN
cana-6072	166	14	.	.	PUNCT
cana-6072	167	1	4s	4s	NUM
cana-6072	167	2	(	(	PUNCT
cana-6072	167	3	2025	2025	NUM
cana-6072	167	4	)	)	PUNCT
cana-6072	167	5	766	766	NUM
cana-6072	167	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	167	7	findings	finding	VERB
cana-6072	167	8	model	model	NOUN
cana-6072	167	9	families	family	NOUN
cana-6072	167	10	&	&	CCONJ
cana-6072	167	11	notable	notable	ADJ
cana-6072	167	12	results	result	NOUN
cana-6072	167	13	.	.	PUNCT
cana-6072	168	1	the	the	DET
cana-6072	168	2	reviewed	review	VERB
cana-6072	168	3	studies	study	NOUN
cana-6072	168	4	fall	fall	VERB
cana-6072	168	5	into	into	ADP
cana-6072	168	6	four	four	NUM
cana-6072	168	7	families	family	NOUN
cana-6072	168	8	:	:	PUNCT
cana-6072	168	9	(	(	PUNCT
cana-6072	168	10	i	i	NOUN
cana-6072	168	11	)	)	PUNCT
cana-6072	168	12	lexicon	lexicon	NOUN
cana-6072	168	13	/	/	SYM
cana-6072	168	14	rule	rule	NOUN
cana-6072	168	15	-	-	PUNCT
cana-6072	168	16	based	base	VERB
cana-6072	168	17	(	(	PUNCT
cana-6072	168	18	early	early	ADJ
cana-6072	168	19	emoji	emoji	PROPN
cana-6072	168	20	dictionaries	dictionary	NOUN
cana-6072	168	21	or	or	CCONJ
cana-6072	168	22	handcrafted	handcraft	VERB
cana-6072	168	23	cues	cue	NOUN
cana-6072	168	24	)	)	PUNCT
cana-6072	168	25	,	,	PUNCT
cana-6072	168	26	(	(	PUNCT
cana-6072	168	27	ii	ii	NOUN
cana-6072	168	28	)	)	PUNCT
cana-6072	168	29	classical	classical	ADJ
cana-6072	168	30	ml	ml	NOUN
cana-6072	168	31	(	(	PUNCT
cana-6072	168	32	svm	svm	ADJ
cana-6072	168	33	,	,	PUNCT
cana-6072	168	34	logistic	logistic	ADJ
cana-6072	168	35	regression	regression	NOUN
cana-6072	168	36	,	,	PUNCT
cana-6072	168	37	ensembles	ensemble	NOUN
cana-6072	168	38	)	)	PUNCT
cana-6072	168	39	using	use	VERB
cana-6072	168	40	engineered	engineer	VERB
cana-6072	168	41	emoji	emoji	PROPN
cana-6072	168	42	features	feature	NOUN
cana-6072	168	43	,	,	PUNCT
cana-6072	168	44	(	(	PUNCT
cana-6072	168	45	iii	iii	X
cana-6072	168	46	)	)	PUNCT
cana-6072	168	47	neural	neural	ADJ
cana-6072	168	48	networks	network	NOUN
cana-6072	168	49	,	,	PUNCT
cana-6072	168	50	and	and	CCONJ
cana-6072	168	51	(	(	PUNCT
cana-6072	168	52	iv	iv	X
cana-6072	168	53	)	)	PUNCT
cana-6072	168	54	transformer	transformer	NOUN
cana-6072	168	55	-	-	PUNCT
cana-6072	168	56	based	base	VERB
cana-6072	168	57	models	model	NOUN
cana-6072	168	58	.	.	PUNCT
cana-6072	169	1	the	the	DET
cana-6072	169	2	field	field	NOUN
cana-6072	169	3	has	have	AUX
cana-6072	169	4	shifted	shift	VERB
cana-6072	169	5	decisively	decisively	ADV
cana-6072	169	6	toward	toward	ADP
cana-6072	169	7	deep	deep	ADJ
cana-6072	169	8	learning	learning	NOUN
cana-6072	169	9	especially	especially	ADV
cana-6072	169	10	transformers	transformer	NOUN
cana-6072	169	11	that	that	PRON
cana-6072	169	12	explicitly	explicitly	ADV
cana-6072	169	13	fuse	fuse	VERB
cana-6072	169	14	emoji	emoji	PROPN
cana-6072	169	15	and	and	CCONJ
cana-6072	169	16	text	text	NOUN
cana-6072	169	17	within	within	ADP
cana-6072	169	18	unified	unified	ADJ
cana-6072	169	19	architectures	architecture	NOUN
cana-6072	169	20	.	.	PUNCT
cana-6072	170	1	among	among	ADP
cana-6072	170	2	neural	neural	ADJ
cana-6072	170	3	approaches	approach	NOUN
cana-6072	170	4	,	,	PUNCT
cana-6072	170	5	a	a	DET
cana-6072	170	6	representative	representative	ADJ
cana-6072	170	7	example	example	NOUN
cana-6072	170	8	is	be	AUX
cana-6072	170	9	cemo	cemo	NOUN
cana-6072	170	10	-	-	PUNCT
cana-6072	170	11	lstm	lstm	NOUN
cana-6072	170	12	,	,	PUNCT
cana-6072	170	13	which	which	PRON
cana-6072	170	14	augments	augment	VERB
cana-6072	170	15	a	a	DET
cana-6072	170	16	bilstm	bilstm	NOUN
cana-6072	170	17	with	with	ADP
cana-6072	170	18	emoji	emoji	PROPN
cana-6072	170	19	embeddings	embedding	NOUN
cana-6072	170	20	for	for	ADP
cana-6072	170	21	weibo	weibo	NOUN
cana-6072	170	22	posts	post	NOUN
cana-6072	170	23	during	during	ADP
cana-6072	170	24	covid-19	covid-19	PROPN
cana-6072	170	25	discussions	discussion	NOUN
cana-6072	170	26	.	.	PUNCT
cana-6072	171	1	this	this	PRON
cana-6072	171	2	is	be	AUX
cana-6072	171	3	a	a	DET
cana-6072	171	4	classic	classic	ADJ
cana-6072	171	5	early	early	ADJ
cana-6072	171	6	fusion	fusion	NOUN
cana-6072	171	7	design	design	NOUN
cana-6072	171	8	emoji	emoji	PROPN
cana-6072	171	9	tokens	token	NOUN
cana-6072	171	10	are	be	AUX
cana-6072	171	11	integrated	integrate	VERB
cana-6072	171	12	directly	directly	ADV
cana-6072	171	13	into	into	ADP
cana-6072	171	14	the	the	DET
cana-6072	171	15	text	text	NOUN
cana-6072	171	16	pipeline	pipeline	NOUN
cana-6072	171	17	,	,	PUNCT
cana-6072	171	18	yielding	yield	VERB
cana-6072	171	19	clear	clear	ADJ
cana-6072	171	20	gains	gain	NOUN
cana-6072	171	21	over	over	ADP
cana-6072	171	22	text	text	NOUN
cana-6072	171	23	-	-	PUNCT
cana-6072	171	24	only	only	ADJ
cana-6072	171	25	baselines	baseline	NOUN
cana-6072	171	26	.	.	PUNCT
cana-6072	172	1	more	more	ADV
cana-6072	172	2	sophisticated	sophisticated	ADJ
cana-6072	172	3	architectures	architecture	NOUN
cana-6072	172	4	adopt	adopt	VERB
cana-6072	172	5	cross	cross	NOUN
cana-6072	172	6	-	-	NOUN
cana-6072	172	7	attention	attention	NOUN
cana-6072	172	8	to	to	PART
cana-6072	172	9	align	align	VERB
cana-6072	172	10	textual	textual	ADJ
cana-6072	172	11	and	and	CCONJ
cana-6072	172	12	emoji	emoji	PROPN
cana-6072	172	13	features	feature	NOUN
cana-6072	172	14	.	.	PUNCT
cana-6072	173	1	for	for	ADP
cana-6072	173	2	instance	instance	NOUN
cana-6072	173	3	,	,	PUNCT
cana-6072	173	4	emfsa	emfsa	VERB
cana-6072	173	5	integrates	integrate	NOUN
cana-6072	173	6	topic	topic	NOUN
cana-6072	173	7	,	,	PUNCT
cana-6072	173	8	text	text	NOUN
cana-6072	173	9	,	,	PUNCT
cana-6072	173	10	and	and	CCONJ
cana-6072	173	11	emoji	emoji	PROPN
cana-6072	173	12	information	information	NOUN
cana-6072	173	13	with	with	ADP
cana-6072	173	14	a	a	DET
cana-6072	173	15	bert	bert	NOUN
cana-6072	173	16	-	-	PUNCT
cana-6072	173	17	based	base	VERB
cana-6072	173	18	encoder	encoder	NOUN
cana-6072	173	19	plus	plus	CCONJ
cana-6072	173	20	cross	cross	NOUN
cana-6072	173	21	-	-	NOUN
cana-6072	173	22	attention	attention	ADJ
cana-6072	173	23	,	,	PUNCT
cana-6072	173	24	achieving	achieve	VERB
cana-6072	173	25	accuracy	accuracy	NOUN
cana-6072	173	26	improvements	improvement	NOUN
cana-6072	173	27	up	up	ADP
cana-6072	173	28	to	to	ADP
cana-6072	173	29	+10.9	+10.9	NOUN
cana-6072	173	30	%	%	NOUN
cana-6072	173	31	across	across	ADP
cana-6072	173	32	three	three	NUM
cana-6072	173	33	chinese	chinese	ADJ
cana-6072	173	34	microblog	microblog	NOUN
cana-6072	173	35	datasets	dataset	NOUN
cana-6072	173	36	.	.	PUNCT
cana-6072	174	1	these	these	DET
cana-6072	174	2	results	result	NOUN
cana-6072	174	3	highlight	highlight	VERB
cana-6072	174	4	the	the	DET
cana-6072	174	5	value	value	NOUN
cana-6072	174	6	of	of	ADP
cana-6072	174	7	modeling	model	VERB
cana-6072	174	8	fine	fine	ADV
cana-6072	174	9	-	-	PUNCT
cana-6072	174	10	grained	grain	VERB
cana-6072	174	11	interactions	interaction	NOUN
cana-6072	174	12	between	between	ADP
cana-6072	174	13	modalities	modality	NOUN
cana-6072	174	14	.	.	PUNCT
cana-6072	175	1	table	table	NOUN
cana-6072	175	2	5	5	NUM
cana-6072	175	3	.	.	PUNCT
cana-6072	175	4	selected	select	VERB
cana-6072	175	5	findings	finding	NOUN
cana-6072	175	6	(	(	PUNCT
cana-6072	175	7	emoji	emoji	NOUN
cana-6072	175	8	-	-	PUNCT
cana-6072	175	9	aware	aware	ADJ
cana-6072	175	10	vs.	vs.	ADP
cana-6072	175	11	text	text	NOUN
cana-6072	175	12	-	-	PUNCT
cana-6072	175	13	only	only	ADV
cana-6072	175	14	)	)	PUNCT
cana-6072	175	15	liu	liu	PROPN
cana-6072	175	16	et	et	PROPN
cana-6072	175	17	al	al	PROPN
cana-6072	175	18	.	.	PUNCT
cana-6072	176	1	cemo	cemo	PROPN
cana-6072	176	2	-	-	PUNCT
cana-6072	176	3	lstm	lstm	ADJ
cana-6072	176	4	model	model	NOUN
cana-6072	176	5	(	(	PUNCT
cana-6072	176	6	emoji	emoji	PROPN
cana-6072	176	7	repr	repr	PROPN
cana-6072	176	8	.	.	PUNCT
cana-6072	176	9	)	)	PUNCT
cana-6072	177	1	cemo	cemo	NOUN
cana-6072	177	2	-	-	PUNCT
cana-6072	177	3	lstm	lstm	PROPN
cana-6072	177	4	(	(	PUNCT
cana-6072	177	5	emoji	emoji	NOUN
cana-6072	177	6	embeddings	embedding	NOUN
cana-6072	177	7	as	as	ADP
cana-6072	177	8	tokens	token	NOUN
cana-6072	177	9	)	)	PUNCT
cana-6072	177	10	;	;	PUNCT
cana-6072	177	11	emojis	emoji	NOUN
cana-6072	177	12	and	and	CCONJ
cana-6072	177	13	words	word	NOUN
cana-6072	177	14	embedded	embed	VERB
cana-6072	177	15	jointly	jointly	ADV
cana-6072	177	16	.	.	PUNCT
cana-6072	178	1	fusion	fusion	NOUN
cana-6072	178	2	early	early	ADV
cana-6072	178	3	:	:	PUNCT
cana-6072	178	4	concatenate	concatenate	NOUN
cana-6072	178	5	embeddings	embedding	NOUN
cana-6072	178	6	into	into	ADP
cana-6072	178	7	one	one	NUM
cana-6072	178	8	sequence	sequence	NOUN
cana-6072	178	9	.	.	PUNCT
cana-6072	179	1	dataset(s	dataset(s	NUM
cana-6072	179	2	)	)	PUNCT
cana-6072	179	3	weibo	weibo	NOUN
cana-6072	179	4	(	(	PUNCT
cana-6072	179	5	chinese	chinese	ADJ
cana-6072	179	6	microblogs	microblog	NOUN
cana-6072	179	7	)	)	PUNCT
cana-6072	179	8	.	.	PUNCT
cana-6072	180	1	metric(s	metric(s	X
cana-6072	180	2	)	)	PUNCT
cana-6072	180	3	accuracy	accuracy	NOUN
cana-6072	180	4	(	(	PUNCT
cana-6072	180	5	reported	report	VERB
cana-6072	180	6	best	good	ADJ
cana-6072	180	7	≈	≈	NOUN
cana-6072	180	8	0.95	0.95	NUM
cana-6072	180	9	)	)	PUNCT
cana-6072	180	10	.	.	PUNCT
cana-6072	181	1	headline	headline	NOUN
cana-6072	181	2	result	result	VERB
cana-6072	181	3	highest	high	ADJ
cana-6072	181	4	accuracy	accuracy	NOUN
cana-6072	181	5	(	(	PUNCT
cana-6072	181	6	~0.95	~0.95	NOUN
cana-6072	181	7	)	)	PUNCT
cana-6072	181	8	on	on	ADP
cana-6072	181	9	chinese	chinese	ADJ
cana-6072	181	10	online	online	ADJ
cana-6072	181	11	texts	text	NOUN
cana-6072	181	12	/	/	SYM
cana-6072	181	13	weibo	weibo	NOUN
cana-6072	181	14	;	;	PUNCT
cana-6072	181	15	emoji	emoji	NOUN
cana-6072	181	16	embeddings	embedding	NOUN
cana-6072	181	17	improve	improve	VERB
cana-6072	181	18	over	over	ADP
cana-6072	181	19	text	text	NOUN
cana-6072	181	20	-	-	PUNCT
cana-6072	181	21	only	only	ADJ
cana-6072	181	22	baselines	baseline	NOUN
cana-6072	181	23	.	.	PUNCT
cana-6072	182	1	tang	tang	PROPN
cana-6072	182	2	et	et	PROPN
cana-6072	182	3	al	al	PROPN
cana-6072	182	4	.	.	PROPN
cana-6072	182	5	emfsa	emfsa	PROPN
cana-6072	182	6	model	model	PROPN
cana-6072	182	7	(	(	PUNCT
cana-6072	182	8	emoji	emoji	PROPN
cana-6072	182	9	repr	repr	PROPN
cana-6072	182	10	.	.	PUNCT
cana-6072	182	11	)	)	PUNCT
cana-6072	183	1	emoji	emoji	NOUN
cana-6072	183	2	-	-	ADJ
cana-6072	183	3	aware	aware	ADJ
cana-6072	183	4	pretraining	pretraine	VERB
cana-6072	183	5	+	+	CCONJ
cana-6072	183	6	emoji	emoji	NOUN
cana-6072	183	7	/	/	SYM
cana-6072	183	8	text	text	NOUN
cana-6072	183	9	/	/	SYM
cana-6072	183	10	topic	topic	NOUN
cana-6072	183	11	features	feature	NOUN
cana-6072	183	12	;	;	PUNCT
cana-6072	183	13	multifeature	multifeature	NOUN
cana-6072	183	14	fusion	fusion	NOUN
cana-6072	183	15	module	module	NOUN
cana-6072	183	16	with	with	ADP
cana-6072	183	17	cross	cross	NOUN
cana-6072	183	18	-	-	NOUN
cana-6072	183	19	attention	attention	ADJ
cana-6072	183	20	.	.	PUNCT
cana-6072	184	1	fusion	fusion	NOUN
cana-6072	184	2	intermediate	intermediate	ADJ
cana-6072	184	3	:	:	PUNCT
cana-6072	184	4	cross	cross	ADJ
cana-6072	184	5	-	-	ADJ
cana-6072	184	6	attention	attention	ADJ
cana-6072	184	7	multifeature	multifeature	NOUN
cana-6072	184	8	fusion	fusion	NOUN
cana-6072	184	9	.	.	PUNCT
cana-6072	185	1	dataset(s	dataset(s	NUM
cana-6072	185	2	)	)	PUNCT
cana-6072	185	3	msd_emoji	msd_emoji	NOUN
cana-6072	185	4	,	,	PUNCT
cana-6072	185	5	td_emoji	td_emoji	NOUN
cana-6072	185	6	,	,	PUNCT
cana-6072	185	7	erd_emoji	erd_emoji	PROPN
cana-6072	185	8	(	(	PUNCT
cana-6072	185	9	3	3	NUM
cana-6072	185	10	-	-	PUNCT
cana-6072	185	11	class	class	NOUN
cana-6072	185	12	)	)	PUNCT
cana-6072	185	13	.	.	PUNCT
cana-6072	186	1	metric(s	metric(s	X
cana-6072	186	2	)	)	PUNCT
cana-6072	186	3	accuracy	accuracy	NOUN
cana-6072	186	4	,	,	PUNCT
cana-6072	186	5	macro	macro	NOUN
cana-6072	186	6	-	-	NOUN
cana-6072	186	7	precision	precision	NOUN
cana-6072	186	8	/	/	SYM
cana-6072	186	9	recall	recall	NOUN
cana-6072	186	10	/	/	SYM
cana-6072	186	11	f1	f1	NOUN
cana-6072	186	12	.	.	PUNCT
cana-6072	187	1	headline	headline	NOUN
cana-6072	187	2	result	result	VERB
cana-6072	187	3	vs	vs	ADP
cana-6072	187	4	baselines	baseline	NOUN
cana-6072	187	5	(	(	PUNCT
cana-6072	187	6	accuracy	accuracy	NOUN
cana-6072	187	7	):	):	PUNCT
cana-6072	187	8	+2.3	+2.3	NUM
cana-6072	187	9	%	%	NOUN
cana-6072	187	10	(	(	PUNCT
cana-6072	187	11	msd_emoji	msd_emoji	NOUN
cana-6072	187	12	)	)	PUNCT
cana-6072	187	13	,	,	PUNCT
cana-6072	187	14	+10.9	+10.9	PROPN
cana-6072	187	15	%	%	NOUN
cana-6072	187	16	(	(	PUNCT
cana-6072	187	17	td_emoji	td_emoji	NOUN
cana-6072	187	18	)	)	PUNCT
cana-6072	187	19	,	,	PUNCT
cana-6072	187	20	+2.7	+2.7	PROPN
cana-6072	187	21	%	%	NOUN
cana-6072	187	22	(	(	PUNCT
cana-6072	187	23	erd_emoji	erd_emoji	PROPN
cana-6072	187	24	)	)	PUNCT
cana-6072	187	25	;	;	PUNCT
cana-6072	187	26	consistent	consistent	ADJ
cana-6072	187	27	macro	macro	ADJ
cana-6072	187	28	-	-	ADJ
cana-6072	187	29	metric	metric	ADJ
cana-6072	187	30	gains	gain	NOUN
cana-6072	187	31	.	.	PUNCT
cana-6072	188	1	lou	lou	PROPN
cana-6072	188	2	et	et	PROPN
cana-6072	188	3	al	al	PROPN
cana-6072	188	4	.	.	PUNCT
cana-6072	189	1	mam	mam	PROPN
cana-6072	189	2	-	-	PUNCT
cana-6072	189	3	emmsa	emmsa	NOUN
cana-6072	189	4	model	model	NOUN
cana-6072	189	5	(	(	PUNCT
cana-6072	189	6	emoji	emoji	PROPN
cana-6072	189	7	repr	repr	PROPN
cana-6072	189	8	.	.	PUNCT
cana-6072	189	9	)	)	PUNCT
cana-6072	190	1	emoji	emoji	PROPN
cana-6072	190	2	images	image	NOUN
cana-6072	190	3	(	(	PUNCT
cana-6072	190	4	visual	visual	ADJ
cana-6072	190	5	glyphs	glyph	NOUN
cana-6072	190	6	)	)	PUNCT
cana-6072	190	7	+	+	NUM
cana-6072	190	8	textual	textual	ADJ
cana-6072	190	9	emoji	emoji	PROPN
cana-6072	190	10	tokens	token	NOUN
cana-6072	190	11	;	;	PUNCT
cana-6072	190	12	mutual	mutual	ADJ
cana-6072	190	13	attention	attention	NOUN
cana-6072	190	14	.	.	PUNCT
cana-6072	191	1	fusion	fusion	NOUN
cana-6072	191	2	intermediate	intermediate	ADJ
cana-6072	191	3	:	:	PUNCT
cana-6072	191	4	mutual	mutual	ADJ
cana-6072	191	5	attention	attention	NOUN
cana-6072	191	6	between	between	ADP
cana-6072	191	7	emoji	emoji	NOUN
cana-6072	191	8	images	image	NOUN
cana-6072	191	9	and	and	CCONJ
cana-6072	191	10	text	text	NOUN
cana-6072	191	11	.	.	PUNCT
cana-6072	192	1	dataset(s	dataset(s	NUM
cana-6072	192	2	)	)	PUNCT
cana-6072	192	3	weibo	weibo	PROPN
cana-6072	192	4	emoji	emoji	PROPN
cana-6072	192	5	–	–	PUNCT
cana-6072	192	6	text	text	NOUN
cana-6072	192	7	corpus	corpus	NOUN
cana-6072	192	8	(	(	PUNCT
cana-6072	192	9	12,278	12,278	NUM
cana-6072	192	10	posts	post	NOUN
cana-6072	192	11	)	)	PUNCT
cana-6072	192	12	.	.	PUNCT
cana-6072	193	1	metric(s	metric(s	X
cana-6072	193	2	)	)	PUNCT
cana-6072	193	3	macro	macro	NOUN
cana-6072	193	4	-	-	NOUN
cana-6072	193	5	f1	f1	NOUN
cana-6072	193	6	,	,	PUNCT
cana-6072	193	7	accuracy	accuracy	NOUN
cana-6072	193	8	(	(	PUNCT
cana-6072	193	9	5×	5×	NOUN
cana-6072	193	10	repeated	repeat	VERB
cana-6072	193	11	runs	run	NOUN
cana-6072	193	12	averaged	average	VERB
cana-6072	193	13	)	)	PUNCT
cana-6072	193	14	.	.	PUNCT
cana-6072	194	1	headline	headline	PROPN
cana-6072	194	2	result	result	VERB
cana-6072	194	3	macro	macro	NOUN
cana-6072	194	4	-	-	NOUN
cana-6072	194	5	f1	f1	NOUN
cana-6072	194	6	80.42	80.42	NUM
cana-6072	194	7	,	,	PUNCT
cana-6072	194	8	acc	acc	PROPN
cana-6072	194	9	87.95	87.95	NUM
cana-6072	194	10	;	;	PUNCT
cana-6072	194	11	beats	beat	NOUN
cana-6072	194	12	prior	prior	ADJ
cana-6072	194	13	baselines	baseline	NOUN
cana-6072	194	14	;	;	PUNCT
cana-6072	194	15	mutual	mutual	ADJ
cana-6072	194	16	attention	attention	NOUN
cana-6072	194	17	boosts	boost	VERB
cana-6072	194	18	cross	cross	ADJ
cana-6072	194	19	-	-	ADJ
cana-6072	194	20	modal	modal	ADJ
cana-6072	194	21	interactions	interaction	NOUN
cana-6072	194	22	.	.	PUNCT
cana-6072	195	1	gan	gan	VERB
cana-6072	195	2	et	et	PROPN
cana-6072	195	3	al	al	PROPN
cana-6072	195	4	.	.	PUNCT
cana-6072	195	5	mfn	mfn	PROPN
cana-6072	195	6	model	model	NOUN
cana-6072	195	7	(	(	PUNCT
cana-6072	195	8	emoji	emoji	PROPN
cana-6072	195	9	repr	repr	PROPN
cana-6072	195	10	.	.	PUNCT
cana-6072	195	11	)	)	PUNCT
cana-6072	196	1	visual	visual	ADJ
cana-6072	196	2	–	–	PUNCT
cana-6072	196	3	textual	textual	ADJ
cana-6072	196	4	fine	fine	ADV
cana-6072	196	5	-	-	PUNCT
cana-6072	196	6	grained	grain	VERB
cana-6072	196	7	correlation	correlation	NOUN
cana-6072	196	8	fusion	fusion	NOUN
cana-6072	196	9	(	(	PUNCT
cana-6072	196	10	not	not	PART
cana-6072	196	11	emoji	emoji	NOUN
cana-6072	196	12	-	-	PUNCT
cana-6072	196	13	specific	specific	ADJ
cana-6072	196	14	)	)	PUNCT
cana-6072	196	15	.	.	PUNCT
cana-6072	197	1	fusion	fusion	NOUN
cana-6072	197	2	intermediate	intermediate	ADJ
cana-6072	197	3	:	:	PUNCT
cana-6072	197	4	correlation	correlation	NOUN
cana-6072	197	5	fusion	fusion	NOUN
cana-6072	197	6	;	;	PUNCT
cana-6072	197	7	outperforms	outperform	VERB
cana-6072	197	8	early	early	ADV
cana-6072	197	9	/	/	SYM
cana-6072	197	10	late	late	ADV
cana-6072	197	11	in	in	ADP
cana-6072	197	12	ablations	ablation	NOUN
cana-6072	197	13	.	.	PUNCT
cana-6072	198	1	dataset(s	dataset(s	NUM
cana-6072	198	2	)	)	PUNCT
cana-6072	198	3	twitter	twitter	NOUN
cana-6072	198	4	,	,	PUNCT
cana-6072	198	5	getty	getty	NOUN
cana-6072	198	6	images	image	NOUN
cana-6072	198	7	,	,	PUNCT
cana-6072	198	8	flickr	flickr	PROPN
cana-6072	198	9	.	.	PUNCT
cana-6072	199	1	communications	communication	NOUN
cana-6072	199	2	on	on	ADP
cana-6072	199	3	applied	apply	VERB
cana-6072	199	4	nonlinear	nonlinear	ADJ
cana-6072	199	5	analysis	analysis	NOUN
cana-6072	199	6	issn	issn	NOUN
cana-6072	199	7	:	:	PUNCT
cana-6072	199	8	1074	1074	NUM
cana-6072	199	9	-	-	PUNCT
cana-6072	199	10	133x	133x	NUM
cana-6072	199	11	vol	vol	NOUN
cana-6072	199	12	32	32	NUM
cana-6072	199	13	no	no	NOUN
cana-6072	199	14	.	.	PUNCT
cana-6072	200	1	4s	4s	NUM
cana-6072	200	2	(	(	PUNCT
cana-6072	200	3	2025	2025	NUM
cana-6072	200	4	)	)	PUNCT
cana-6072	200	5	767	767	NUM
cana-6072	200	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	200	7	metric(s	metric(s	PROPN
cana-6072	200	8	)	)	PUNCT
cana-6072	200	9	f1	f1	NOUN
cana-6072	200	10	,	,	PUNCT
cana-6072	200	11	accuracy	accuracy	NOUN
cana-6072	200	12	(	(	PUNCT
cana-6072	200	13	5	5	NUM
cana-6072	200	14	-	-	NOUN
cana-6072	200	15	fold	fold	ADJ
cana-6072	200	16	)	)	PUNCT
cana-6072	200	17	.	.	PUNCT
cana-6072	201	1	headline	headline	NOUN
cana-6072	201	2	result	result	NOUN
cana-6072	201	3	twitter	twitter	NOUN
cana-6072	201	4	:	:	PUNCT
cana-6072	201	5	f1	f1	NOUN
cana-6072	201	6	97.60	97.60	NUM
cana-6072	201	7	/	/	SYM
cana-6072	201	8	acc	acc	PROPN
cana-6072	201	9	97.48	97.48	NUM
cana-6072	201	10	;	;	PUNCT
cana-6072	202	1	getty	getty	PROPN
cana-6072	202	2	:	:	PUNCT
cana-6072	202	3	f1	f1	NOUN
cana-6072	202	4	94.13	94.13	NUM
cana-6072	202	5	/	/	SYM
cana-6072	202	6	acc	acc	PROPN
cana-6072	202	7	95.60	95.60	NUM
cana-6072	202	8	;	;	PUNCT
cana-6072	202	9	flickr	flickr	PROPN
cana-6072	202	10	:	:	PUNCT
cana-6072	202	11	f1	f1	PROPN
cana-6072	202	12	92.29	92.29	NUM
cana-6072	202	13	/	/	SYM
cana-6072	202	14	acc	acc	PROPN
cana-6072	202	15	88.68	88.68	NUM
cana-6072	202	16	;	;	PUNCT
cana-6072	202	17	beats	beat	VERB
cana-6072	202	18	11	11	NUM
cana-6072	202	19	baselines	baseline	NOUN
cana-6072	202	20	.	.	PUNCT
cana-6072	203	1	multimodal	multimodal	NOUN
cana-6072	203	2	extensions	extension	NOUN
cana-6072	203	3	leverage	leverage	VERB
cana-6072	203	4	the	the	DET
cana-6072	203	5	visual	visual	ADJ
cana-6072	203	6	properties	property	NOUN
cana-6072	203	7	of	of	ADP
cana-6072	203	8	emoji	emoji	PROPN
cana-6072	203	9	glyphs	glyphs	PROPN
cana-6072	203	10	.	.	PUNCT
cana-6072	204	1	mam	mam	NOUN
cana-6072	204	2	-	-	PUNCT
cana-6072	204	3	emmsa	emmsa	NOUN
cana-6072	204	4	(	(	PUNCT
cana-6072	204	5	s3	s3	PROPN
cana-6072	204	6	)	)	PUNCT
cana-6072	204	7	combines	combine	VERB
cana-6072	204	8	text	text	NOUN
cana-6072	204	9	(	(	PUNCT
cana-6072	204	10	bert	bert	NOUN
cana-6072	204	11	with	with	ADP
cana-6072	204	12	emoji	emoji	PROPN
cana-6072	204	13	glosses	gloss	NOUN
cana-6072	204	14	)	)	PUNCT
cana-6072	204	15	and	and	CCONJ
cana-6072	204	16	emoji	emoji	PROPN
cana-6072	204	17	images	image	NOUN
cana-6072	204	18	(	(	PUNCT
cana-6072	204	19	resnet-50	resnet-50	NUM
cana-6072	204	20	)	)	PUNCT
cana-6072	204	21	,	,	PUNCT
cana-6072	204	22	fused	fuse	VERB
cana-6072	204	23	via	via	ADP
cana-6072	204	24	mutual	mutual	ADJ
cana-6072	204	25	attention	attention	NOUN
cana-6072	204	26	;	;	PUNCT
cana-6072	204	27	removing	remove	VERB
cana-6072	204	28	either	either	CCONJ
cana-6072	204	29	modality	modality	NOUN
cana-6072	204	30	notably	notably	ADV
cana-6072	204	31	degrades	degrade	VERB
cana-6072	204	32	performance	performance	NOUN
cana-6072	204	33	(	(	PUNCT
cana-6072	204	34	macro	macro	NOUN
cana-6072	204	35	-	-	NOUN
cana-6072	204	36	f1	f1	NOUN
cana-6072	204	37	≈	≈	PROPN
cana-6072	204	38	0.804	0.804	NUM
cana-6072	204	39	;	;	PUNCT
cana-6072	205	1	acc	acc	PROPN
cana-6072	205	2	≈	≈	PROPN
cana-6072	205	3	87.95	87.95	NUM
cana-6072	205	4	)	)	PUNCT
cana-6072	205	5	.	.	PUNCT
cana-6072	206	1	at	at	ADP
cana-6072	206	2	the	the	DET
cana-6072	206	3	high	high	ADJ
cana-6072	206	4	end	end	NOUN
cana-6072	206	5	of	of	ADP
cana-6072	206	6	complexity	complexity	NOUN
cana-6072	206	7	,	,	PUNCT
cana-6072	206	8	hybrid	hybrid	ADJ
cana-6072	206	9	designs	design	NOUN
cana-6072	206	10	have	have	AUX
cana-6072	206	11	emerged	emerge	VERB
cana-6072	206	12	as	as	ADP
cana-6072	206	13	state	state	NOUN
cana-6072	206	14	of	of	ADP
cana-6072	206	15	the	the	DET
cana-6072	206	16	art	art	NOUN
cana-6072	206	17	.	.	PUNCT
cana-6072	207	1	mfn	mfn	PROPN
cana-6072	207	2	(	(	PUNCT
cana-6072	207	3	s4	s4	PROPN
cana-6072	207	4	)	)	PUNCT
cana-6072	207	5	mixes	mix	VERB
cana-6072	207	6	multi	multi	ADJ
cana-6072	207	7	-	-	ADJ
cana-6072	207	8	head	head	ADJ
cana-6072	207	9	self	self	NOUN
cana-6072	207	10	-	-	PUNCT
cana-6072	207	11	attention	attention	NOUN
cana-6072	207	12	with	with	ADP
cana-6072	207	13	early	early	ADJ
cana-6072	207	14	fusion	fusion	NOUN
cana-6072	207	15	for	for	ADP
cana-6072	207	16	multimodal	multimodal	NOUN
cana-6072	207	17	twitter	twitter	NOUN
cana-6072	207	18	sentiment	sentiment	NOUN
cana-6072	207	19	,	,	PUNCT
cana-6072	207	20	attaining	attain	VERB
cana-6072	207	21	near	near	ADV
cana-6072	207	22	-	-	PUNCT
cana-6072	207	23	perfect	perfect	ADJ
cana-6072	207	24	macro	macro	NOUN
cana-6072	207	25	-	-	NOUN
cana-6072	207	26	f1	f1	ADJ
cana-6072	207	27	(	(	PUNCT
cana-6072	207	28	≈97.6	≈97.6	NOUN
cana-6072	207	29	)	)	PUNCT
cana-6072	207	30	and	and	CCONJ
cana-6072	207	31	surpassing	surpass	VERB
cana-6072	207	32	eleven	eleven	NUM
cana-6072	207	33	baselines	baseline	NOUN
cana-6072	207	34	.	.	PUNCT
cana-6072	208	1	overall	overall	ADV
cana-6072	208	2	,	,	PUNCT
cana-6072	208	3	the	the	DET
cana-6072	208	4	trajectory	trajectory	NOUN
cana-6072	208	5	is	be	AUX
cana-6072	208	6	clear	clear	ADJ
cana-6072	208	7	:	:	PUNCT
cana-6072	208	8	earlier	early	ADJ
cana-6072	208	9	rnn	rnn	NOUN
cana-6072	208	10	-	-	PUNCT
cana-6072	208	11	based	base	VERB
cana-6072	208	12	models	model	NOUN
cana-6072	208	13	and	and	CCONJ
cana-6072	208	14	simple	simple	ADJ
cana-6072	208	15	embeddings	embedding	NOUN
cana-6072	208	16	have	have	AUX
cana-6072	208	17	given	give	VERB
cana-6072	208	18	way	way	NOUN
cana-6072	208	19	to	to	ADP
cana-6072	208	20	transformer	transformer	NOUN
cana-6072	208	21	architectures	architecture	NOUN
cana-6072	208	22	with	with	ADP
cana-6072	208	23	cross	cross	ADJ
cana-6072	208	24	-	-	ADJ
cana-6072	208	25	modal	modal	ADJ
cana-6072	208	26	attention	attention	NOUN
cana-6072	208	27	.	.	PUNCT
cana-6072	209	1	nearly	nearly	ADV
cana-6072	209	2	all	all	DET
cana-6072	209	3	emoji	emoji	ADJ
cana-6072	209	4	-	-	PUNCT
cana-6072	209	5	aware	aware	ADJ
cana-6072	209	6	models	model	NOUN
cana-6072	209	7	outperform	outperform	VERB
cana-6072	209	8	text	text	NOUN
cana-6072	209	9	-	-	PUNCT
cana-6072	209	10	only	only	ADV
cana-6072	209	11	counterparts	counterpart	NOUN
cana-6072	209	12	,	,	PUNCT
cana-6072	209	13	but	but	CCONJ
cana-6072	209	14	absolute	absolute	ADJ
cana-6072	209	15	scores	score	NOUN
cana-6072	209	16	vary	vary	VERB
cana-6072	209	17	with	with	ADP
cana-6072	209	18	dataset	dataset	ADJ
cana-6072	209	19	characteristics	characteristic	NOUN
cana-6072	209	20	:	:	PUNCT
cana-6072	209	21	noisy	noisy	ADJ
cana-6072	209	22	twitter	twitter	NOUN
cana-6072	209	23	data	datum	NOUN
cana-6072	209	24	often	often	ADV
cana-6072	209	25	yields	yield	VERB
cana-6072	209	26	70–80	70–80	NUM
cana-6072	209	27	%	%	NOUN
cana-6072	209	28	accuracy	accuracy	NOUN
cana-6072	209	29	,	,	PUNCT
cana-6072	209	30	while	while	SCONJ
cana-6072	209	31	curated	curate	VERB
cana-6072	209	32	product	product	NOUN
cana-6072	209	33	-	-	PUNCT
cana-6072	209	34	review	review	NOUN
cana-6072	209	35	or	or	CCONJ
cana-6072	209	36	annotated	annotate	VERB
cana-6072	209	37	weibo	weibo	NOUN
cana-6072	209	38	corpora	corpora	PROPN
cana-6072	209	39	can	can	AUX
cana-6072	209	40	exceed	exceed	VERB
cana-6072	209	41	95	95	NUM
cana-6072	209	42	%	%	NOUN
cana-6072	209	43	.	.	PUNCT
cana-6072	210	1	crucially	crucially	ADV
cana-6072	210	2	,	,	PUNCT
cana-6072	210	3	the	the	DET
cana-6072	210	4	relative	relative	ADJ
cana-6072	210	5	gains	gain	NOUN
cana-6072	210	6	from	from	ADP
cana-6072	210	7	modeling	model	VERB
cana-6072	210	8	emoji	emoji	NOUN
cana-6072	210	9	matter	matter	NOUN
cana-6072	210	10	most	most	ADV
cana-6072	210	11	,	,	PUNCT
cana-6072	210	12	consistently	consistently	ADV
cana-6072	210	13	showing	show	VERB
cana-6072	210	14	complementary	complementary	ADJ
cana-6072	210	15	affective	affective	NOUN
cana-6072	210	16	cues	cue	NOUN
cana-6072	210	17	beyond	beyond	ADP
cana-6072	210	18	text	text	NOUN
cana-6072	210	19	alone	alone	ADV
cana-6072	210	20	.	.	PUNCT
cana-6072	211	1	table	table	NOUN
cana-6072	211	2	6	6	NUM
cana-6072	211	3	:	:	PUNCT
cana-6072	211	4	common	common	ADJ
cana-6072	211	5	threats	threat	NOUN
cana-6072	211	6	to	to	ADP
cana-6072	211	7	validity	validity	NOUN
cana-6072	211	8	and	and	CCONJ
cana-6072	211	9	mitigation	mitigation	NOUN
cana-6072	211	10	strategies	strategy	NOUN
cana-6072	211	11	hazards	hazards	AUX
cana-6072	211	12	impact	impact	VERB
cana-6072	211	13	best	good	ADJ
cana-6072	211	14	practice	practice	NOUN
cana-6072	211	15	apples	apple	NOUN
cana-6072	211	16	-	-	PUNCT
cana-6072	211	17	to	to	ADP
cana-6072	211	18	-	-	PUNCT
cana-6072	211	19	oranges	orange	NOUN
cana-6072	211	20	evaluation	evaluation	NOUN
cana-6072	211	21	(	(	PUNCT
cana-6072	211	22	inconsistent	inconsistent	ADJ
cana-6072	211	23	splits	split	VERB
cana-6072	211	24	or	or	CCONJ
cana-6072	211	25	metrics	metric	NOUN
cana-6072	211	26	)	)	PUNCT
cana-6072	211	27	unfair	unfair	ADJ
cana-6072	211	28	comparisons	comparison	NOUN
cana-6072	211	29	;	;	PUNCT
cana-6072	211	30	inflated	inflated	ADJ
cana-6072	211	31	claims	claim	NOUN
cana-6072	211	32	(	(	PUNCT
cana-6072	211	33	e.g.	e.g.	ADV
cana-6072	211	34	,	,	PUNCT
cana-6072	211	35	testing	test	VERB
cana-6072	211	36	on	on	ADP
cana-6072	211	37	an	an	DET
cana-6072	211	38	easier	easy	ADJ
cana-6072	211	39	split	split	NOUN
cana-6072	211	40	)	)	PUNCT
cana-6072	211	41	.	.	PUNCT
cana-6072	212	1	use	use	VERB
cana-6072	212	2	standard	standard	NOUN
cana-6072	212	3	splits	split	NOUN
cana-6072	212	4	or	or	CCONJ
cana-6072	212	5	clearly	clearly	ADV
cana-6072	212	6	disclose	disclose	VERB
cana-6072	212	7	custom	custom	NOUN
cana-6072	212	8	ones	one	NOUN
cana-6072	212	9	;	;	PUNCT
cana-6072	212	10	re	re	VERB
cana-6072	212	11	-	-	ADJ
cana-6072	212	12	run	run	ADJ
cana-6072	212	13	baselines	baseline	NOUN
cana-6072	212	14	on	on	ADP
cana-6072	212	15	the	the	DET
cana-6072	212	16	same	same	ADJ
cana-6072	212	17	splits	split	NOUN
cana-6072	212	18	/	/	SYM
cana-6072	212	19	metrics	metric	NOUN
cana-6072	212	20	;	;	PUNCT
cana-6072	212	21	publish	publish	VERB
cana-6072	212	22	splits	split	VERB
cana-6072	212	23	for	for	ADP
cana-6072	212	24	reuse	reuse	NOUN
cana-6072	212	25	.	.	PUNCT
cana-6072	213	1	distant	distant	ADJ
cana-6072	213	2	-	-	PUNCT
cana-6072	213	3	supervision	supervision	NOUN
cana-6072	213	4	leakage	leakage	NOUN
cana-6072	213	5	(	(	PUNCT
cana-6072	213	6	emoji	emoji	PROPN
cana-6072	213	7	used	use	VERB
cana-6072	213	8	as	as	ADP
cana-6072	213	9	labels	label	NOUN
cana-6072	213	10	and	and	CCONJ
cana-6072	213	11	as	as	ADP
cana-6072	213	12	features	feature	NOUN
cana-6072	213	13	)	)	PUNCT
cana-6072	213	14	inflated	inflate	VERB
cana-6072	213	15	performance	performance	NOUN
cana-6072	213	16	model	model	NOUN
cana-6072	213	17	“	"	PUNCT
cana-6072	213	18	cheats	cheat	NOUN
cana-6072	213	19	”	"	PUNCT
cana-6072	213	20	by	by	ADP
cana-6072	213	21	keying	key	VERB
cana-6072	213	22	on	on	ADP
cana-6072	213	23	label	label	NOUN
cana-6072	213	24	emojis	emoji	NOUN
cana-6072	213	25	in	in	ADP
cana-6072	213	26	text	text	NOUN
cana-6072	213	27	.	.	PUNCT
cana-6072	214	1	remove	remove	VERB
cana-6072	214	2	label	label	NOUN
cana-6072	214	3	emojis	emoji	NOUN
cana-6072	214	4	from	from	ADP
cana-6072	214	5	inputs	input	NOUN
cana-6072	214	6	at	at	ADP
cana-6072	214	7	test	test	NOUN
cana-6072	214	8	time	time	NOUN
cana-6072	214	9	,	,	PUNCT
cana-6072	214	10	or	or	CCONJ
cana-6072	214	11	use	use	VERB
cana-6072	214	12	different	different	ADJ
cana-6072	214	13	emojis	emoji	NOUN
cana-6072	214	14	for	for	ADP
cana-6072	214	15	labelling	labelling	NOUN
cana-6072	214	16	vs.	vs.	ADP
cana-6072	214	17	features	feature	NOUN
cana-6072	214	18	;	;	PUNCT
cana-6072	214	19	manually	manually	ADV
cana-6072	214	20	audit	audit	NOUN
cana-6072	214	21	test	test	NOUN
cana-6072	214	22	sets	set	NOUN
cana-6072	214	23	for	for	ADP
cana-6072	214	24	label	label	NOUN
cana-6072	214	25	tokens	token	NOUN
cana-6072	214	26	.	.	PUNCT
cana-6072	215	1	vendor	vendor	NOUN
cana-6072	215	2	glyph	glyph	PROPN
cana-6072	215	3	variance	variance	NOUN
cana-6072	215	4	(	(	PUNCT
cana-6072	215	5	platform	platform	NOUN
cana-6072	215	6	rendering	rendering	NOUN
cana-6072	215	7	differences	difference	NOUN
cana-6072	215	8	)	)	PUNCT
cana-6072	215	9	inconsistent	inconsistent	ADJ
cana-6072	215	10	sentiment	sentiment	NOUN
cana-6072	215	11	for	for	ADP
cana-6072	215	12	the	the	DET
cana-6072	215	13	same	same	ADJ
cana-6072	215	14	emoji	emoji	NOUN
cana-6072	215	15	across	across	ADP
cana-6072	215	16	platforms	platform	NOUN
cana-6072	215	17	;	;	PUNCT
cana-6072	215	18	accuracy	accuracy	NOUN
cana-6072	215	19	drops	drop	NOUN
cana-6072	215	20	when	when	SCONJ
cana-6072	215	21	the	the	DET
cana-6072	215	22	platform	platform	NOUN
cana-6072	215	23	changes	change	NOUN
cana-6072	215	24	.	.	PUNCT
cana-6072	216	1	train	train	VERB
cana-6072	216	2	with	with	ADP
cana-6072	216	3	multiple	multiple	ADJ
cana-6072	216	4	renderings	rendering	NOUN
cana-6072	216	5	when	when	SCONJ
cana-6072	216	6	possible	possible	ADJ
cana-6072	216	7	;	;	PUNCT
cana-6072	216	8	consider	consider	VERB
cana-6072	216	9	image	image	NOUN
cana-6072	216	10	-	-	PUNCT
cana-6072	216	11	based	base	VERB
cana-6072	216	12	emoji	emoji	NOUN
cana-6072	216	13	representation	representation	NOUN
cana-6072	216	14	to	to	PART
cana-6072	216	15	capture	capture	VERB
cana-6072	216	16	visual	visual	ADJ
cana-6072	216	17	differences	difference	NOUN
cana-6072	216	18	;	;	PUNCT
cana-6072	216	19	acknowledge	acknowledge	VERB
cana-6072	216	20	limits	limit	NOUN
cana-6072	216	21	and	and	CCONJ
cana-6072	216	22	avoid	avoid	VERB
cana-6072	216	23	over	over	ADP
cana-6072	216	24	-	-	PUNCT
cana-6072	216	25	generalization	generalization	NOUN
cana-6072	216	26	.	.	PUNCT
cana-6072	217	1	cultural	cultural	ADJ
cana-6072	217	2	/	/	SYM
cana-6072	217	3	linguistic	linguistic	ADJ
cana-6072	217	4	drift	drift	NOUN
cana-6072	217	5	(	(	PUNCT
cana-6072	217	6	meanings	meaning	NOUN
cana-6072	217	7	differ	differ	VERB
cana-6072	217	8	across	across	ADP
cana-6072	217	9	languages	language	NOUN
cana-6072	217	10	/	/	SYM
cana-6072	217	11	communities	community	NOUN
cana-6072	217	12	)	)	PUNCT
cana-6072	217	13	poor	poor	ADJ
cana-6072	217	14	transfer	transfer	NOUN
cana-6072	217	15	to	to	ADP
cana-6072	217	16	other	other	ADJ
cana-6072	217	17	languages	language	NOUN
cana-6072	217	18	or	or	CCONJ
cana-6072	217	19	domains	domain	NOUN
cana-6072	217	20	;	;	PUNCT
cana-6072	217	21	meaning	mean	VERB
cana-6072	217	22	varies	vary	VERB
cana-6072	217	23	by	by	ADP
cana-6072	217	24	community	community	NOUN
cana-6072	217	25	.	.	PUNCT
cana-6072	218	1	include	include	VERB
cana-6072	218	2	diverse	diverse	ADJ
cana-6072	218	3	language	language	NOUN
cana-6072	218	4	data	datum	NOUN
cana-6072	218	5	;	;	PUNCT
cana-6072	218	6	build	build	VERB
cana-6072	218	7	crosslingual	crosslingual	ADJ
cana-6072	218	8	models	model	NOUN
cana-6072	218	9	;	;	PUNCT
cana-6072	218	10	use	use	VERB
cana-6072	218	11	translation	translation	NOUN
cana-6072	218	12	carefully	carefully	ADV
cana-6072	218	13	;	;	PUNCT
cana-6072	218	14	finetune	finetune	NOUN
cana-6072	218	15	on	on	ADP
cana-6072	218	16	target	target	NOUN
cana-6072	218	17	language	language	NOUN
cana-6072	218	18	/	/	SYM
cana-6072	218	19	domain	domain	NOUN
cana-6072	218	20	when	when	SCONJ
cana-6072	218	21	available	available	ADJ
cana-6072	218	22	.	.	PUNCT
cana-6072	219	1	lack	lack	NOUN
cana-6072	219	2	of	of	ADP
cana-6072	219	3	parity	parity	NOUN
cana-6072	219	4	in	in	ADP
cana-6072	219	5	baselines	baseline	NOUN
cana-6072	219	6	(	(	PUNCT
cana-6072	219	7	unequal	unequal	ADJ
cana-6072	219	8	tuning	tuning	NOUN
cana-6072	219	9	/	/	SYM
cana-6072	219	10	training	training	NOUN
cana-6072	219	11	)	)	PUNCT
cana-6072	219	12	overstated	overstate	VERB
cana-6072	219	13	gains	gain	NOUN
cana-6072	219	14	;	;	PUNCT
cana-6072	219	15	unclear	unclear	ADJ
cana-6072	219	16	if	if	SCONJ
cana-6072	219	17	improvements	improvement	NOUN
cana-6072	219	18	come	come	VERB
cana-6072	219	19	from	from	ADP
cana-6072	219	20	emoji	emoji	PROPN
cana-6072	219	21	or	or	CCONJ
cana-6072	219	22	better	well	ADJ
cana-6072	219	23	tuning	tuning	NOUN
cana-6072	219	24	.	.	PUNCT
cana-6072	220	1	train	train	NOUN
cana-6072	220	2	/	/	SYM
cana-6072	220	3	evaluate	evaluate	VERB
cana-6072	220	4	text	text	NOUN
cana-6072	220	5	-	-	PUNCT
cana-6072	220	6	only	only	ADJ
cana-6072	220	7	baselines	baseline	NOUN
cana-6072	220	8	under	under	ADP
cana-6072	220	9	identical	identical	ADJ
cana-6072	220	10	settings	setting	NOUN
cana-6072	220	11	(	(	PUNCT
cana-6072	220	12	epochs	epoch	NOUN
cana-6072	220	13	,	,	PUNCT
cana-6072	220	14	data	datum	NOUN
cana-6072	220	15	,	,	PUNCT
cana-6072	220	16	optimizer	optimizer	NOUN
cana-6072	220	17	)	)	PUNCT
cana-6072	220	18	;	;	PUNCT
cana-6072	220	19	run	run	VERB
cana-6072	220	20	ablations	ablation	NOUN
cana-6072	220	21	with	with	ADP
cana-6072	220	22	emoji	emoji	PROPN
cana-6072	220	23	off	off	ADP
cana-6072	220	24	vs.	vs.	X
cana-6072	220	25	on	on	ADV
cana-6072	220	26	;	;	PUNCT
cana-6072	220	27	use	use	VERB
cana-6072	220	28	strong	strong	ADJ
cana-6072	220	29	,	,	PUNCT
cana-6072	220	30	current	current	ADJ
cana-6072	220	31	baselines	baseline	NOUN
cana-6072	220	32	.	.	PUNCT
cana-6072	221	1	communications	communication	NOUN
cana-6072	221	2	on	on	ADP
cana-6072	221	3	applied	apply	VERB
cana-6072	221	4	nonlinear	nonlinear	ADJ
cana-6072	221	5	analysis	analysis	NOUN
cana-6072	221	6	issn	issn	NOUN
cana-6072	221	7	:	:	PUNCT
cana-6072	221	8	1074	1074	NUM
cana-6072	221	9	-	-	PUNCT
cana-6072	221	10	133x	133x	NUM
cana-6072	221	11	vol	vol	NOUN
cana-6072	221	12	32	32	NUM
cana-6072	221	13	no	no	NOUN
cana-6072	221	14	.	.	PUNCT
cana-6072	222	1	4s	4s	NUM
cana-6072	222	2	(	(	PUNCT
cana-6072	222	3	2025	2025	NUM
cana-6072	222	4	)	)	PUNCT
cana-6072	223	1	768	768	NUM
cana-6072	223	2	https://internationalpubls.com	https://internationalpubls.com	NUM
cana-6072	223	3	discussion	discussion	NOUN
cana-6072	223	4	emoji	emoji	NOUN
cana-6072	223	5	-	-	PUNCT
cana-6072	223	6	aware	aware	ADJ
cana-6072	223	7	sentiment	sentiment	NOUN
cana-6072	223	8	classification	classification	NOUN
cana-6072	223	9	has	have	AUX
cana-6072	223	10	become	become	VERB
cana-6072	223	11	mature	mature	ADJ
cana-6072	223	12	yet	yet	ADV
cana-6072	223	13	methodologically	methodologically	ADV
cana-6072	223	14	fragile	fragile	ADJ
cana-6072	223	15	.	.	PUNCT
cana-6072	224	1	on	on	ADP
cana-6072	224	2	one	one	NUM
cana-6072	224	3	hand	hand	NOUN
cana-6072	224	4	,	,	PUNCT
cana-6072	224	5	there	there	PRON
cana-6072	224	6	is	be	VERB
cana-6072	224	7	a	a	DET
cana-6072	224	8	broad	broad	ADJ
cana-6072	224	9	consensus	consensus	NOUN
cana-6072	224	10	that	that	PRON
cana-6072	224	11	emojis	emojis	NOUN
cana-6072	224	12	can	can	AUX
cana-6072	224	13	enhance	enhance	VERB
cana-6072	224	14	sentiment	sentiment	NOUN
cana-6072	224	15	analysis	analysis	NOUN
cana-6072	224	16	:	:	PUNCT
cana-6072	224	17	virtually	virtually	ADV
cana-6072	224	18	all	all	PRON
cana-6072	224	19	studies	study	NOUN
cana-6072	224	20	we	we	PRON
cana-6072	224	21	reviewed	review	VERB
cana-6072	224	22	found	find	VERB
cana-6072	224	23	some	some	DET
cana-6072	224	24	improvement	improvement	NOUN
cana-6072	224	25	when	when	SCONJ
cana-6072	224	26	emojis	emoji	NOUN
cana-6072	224	27	are	be	AUX
cana-6072	224	28	modeled	model	VERB
cana-6072	224	29	,	,	PUNCT
cana-6072	224	30	especially	especially	ADV
cana-6072	224	31	in	in	ADP
cana-6072	224	32	informal	informal	ADJ
cana-6072	224	33	domains	domain	NOUN
cana-6072	224	34	like	like	ADP
cana-6072	224	35	twitter	twitter	NOUN
cana-6072	224	36	(	(	PUNCT
cana-6072	224	37	often	often	ADV
cana-6072	224	38	a	a	DET
cana-6072	224	39	few	few	ADJ
cana-6072	224	40	f1	f1	NOUN
cana-6072	224	41	points	point	NOUN
cana-6072	224	42	gain	gain	NOUN
cana-6072	224	43	)	)	PUNCT
cana-6072	225	1	[	[	X
cana-6072	225	2	xu	xu	INTJ
cana-6072	225	3	,	,	PUNCT
cana-6072	225	4	q.	q.	PROPN
cana-6072	225	5	a.	a.	PROPN
cana-6072	225	6	,	,	PUNCT
cana-6072	225	7	et	et	PROPN
cana-6072	225	8	al	al	PROPN
cana-6072	225	9	.	.	PROPN
cana-6072	225	10	,	,	PUNCT
cana-6072	225	11	2024	2024	NUM
cana-6072	225	12	]	]	PUNCT
cana-6072	225	13	.	.	PUNCT
cana-6072	226	1	the	the	DET
cana-6072	226	2	inclusion	inclusion	NOUN
cana-6072	226	3	of	of	ADP
cana-6072	226	4	emojis	emojis	NOUN
cana-6072	226	5	helps	help	VERB
cana-6072	226	6	disambiguate	disambiguate	ADJ
cana-6072	226	7	sentiment	sentiment	NOUN
cana-6072	226	8	,	,	PUNCT
cana-6072	226	9	detect	detect	VERB
cana-6072	226	10	sarcasm	sarcasm	NOUN
cana-6072	226	11	,	,	PUNCT
cana-6072	226	12	and	and	CCONJ
cana-6072	226	13	intensify	intensify	VERB
cana-6072	226	14	emotional	emotional	ADJ
cana-6072	226	15	tones	tone	NOUN
cana-6072	226	16	that	that	PRON
cana-6072	226	17	plain	plain	ADJ
cana-6072	226	18	text	text	NOUN
cana-6072	226	19	might	might	AUX
cana-6072	226	20	underrepresent	underrepresent	VERB
cana-6072	226	21	[	[	X
cana-6072	226	22	huang	huang	PROPN
cana-6072	226	23	,	,	PUNCT
cana-6072	226	24	f.	f.	PROPN
cana-6072	226	25	et	et	PROPN
cana-6072	226	26	al	al	PROPN
cana-6072	226	27	.	.	PROPN
cana-6072	226	28	,	,	PUNCT
cana-6072	226	29	2020	2020	NUM
cana-6072	226	30	]	]	PUNCT
cana-6072	226	31	.	.	PUNCT
cana-6072	227	1	on	on	ADP
cana-6072	227	2	the	the	DET
cana-6072	227	3	other	other	ADJ
cana-6072	227	4	hand	hand	NOUN
cana-6072	227	5	,	,	PUNCT
cana-6072	227	6	the	the	DET
cana-6072	227	7	field	field	NOUN
cana-6072	227	8	is	be	AUX
cana-6072	227	9	still	still	ADV
cana-6072	227	10	consolidating	consolidate	VERB
cana-6072	227	11	best	good	ADJ
cana-6072	227	12	practices	practice	NOUN
cana-6072	227	13	;	;	PUNCT
cana-6072	227	14	different	different	ADJ
cana-6072	227	15	works	work	NOUN
cana-6072	227	16	use	use	VERB
cana-6072	227	17	different	different	ADJ
cana-6072	227	18	emoji	emoji	PROPN
cana-6072	227	19	representations	representation	NOUN
cana-6072	227	20	,	,	PUNCT
cana-6072	227	21	fusion	fusion	NOUN
cana-6072	227	22	strategies	strategy	NOUN
cana-6072	227	23	,	,	PUNCT
cana-6072	227	24	and	and	CCONJ
cana-6072	227	25	evaluation	evaluation	NOUN
cana-6072	227	26	setups	setup	NOUN
cana-6072	227	27	,	,	PUNCT
cana-6072	227	28	making	make	VERB
cana-6072	227	29	it	it	PRON
cana-6072	227	30	challenging	challenging	ADJ
cana-6072	227	31	to	to	PART
cana-6072	227	32	compare	compare	VERB
cana-6072	227	33	results	result	NOUN
cana-6072	227	34	directly	directly	ADV
cana-6072	227	35	.	.	PUNCT
cana-6072	228	1	this	this	DET
cana-6072	228	2	study	study	NOUN
cana-6072	228	3	indicates	indicate	VERB
cana-6072	228	4	that	that	SCONJ
cana-6072	228	5	the	the	DET
cana-6072	228	6	strongest	strong	ADJ
cana-6072	228	7	results	result	NOUN
cana-6072	228	8	come	come	VERB
cana-6072	228	9	from	from	ADP
cana-6072	228	10	context	context	NOUN
cana-6072	228	11	-	-	PUNCT
cana-6072	228	12	aware	aware	ADJ
cana-6072	228	13	fusion	fusion	NOUN
cana-6072	228	14	mechanisms	mechanism	NOUN
cana-6072	228	15	–	–	PUNCT
cana-6072	228	16	particularly	particularly	ADV
cana-6072	228	17	attention	attention	NOUN
cana-6072	228	18	models	model	NOUN
cana-6072	228	19	that	that	PRON
cana-6072	228	20	allow	allow	VERB
cana-6072	228	21	emojis	emoji	NOUN
cana-6072	228	22	to	to	PART
cana-6072	228	23	dynamically	dynamically	ADV
cana-6072	228	24	modulate	modulate	VERB
cana-6072	228	25	text	text	NOUN
cana-6072	228	26	representation	representation	NOUN
cana-6072	229	1	[	[	X
cana-6072	229	2	lou	lou	PROPN
cana-6072	229	3	,	,	PUNCT
cana-6072	229	4	y.	y.	PROPN
cana-6072	229	5	et	et	PROPN
cana-6072	229	6	al	al	PROPN
cana-6072	229	7	.	.	PROPN
cana-6072	229	8	,	,	PUNCT
cana-6072	229	9	2020	2020	NUM
cana-6072	229	10	]	]	PUNCT
cana-6072	229	11	.	.	PUNCT
cana-6072	230	1	for	for	ADP
cana-6072	230	2	example	example	NOUN
cana-6072	230	3	,	,	PUNCT
cana-6072	230	4	the	the	DET
cana-6072	230	5	attention	attention	NOUN
cana-6072	230	6	-	-	PUNCT
cana-6072	230	7	based	base	VERB
cana-6072	230	8	models	model	NOUN
cana-6072	230	9	(	(	PUNCT
cana-6072	230	10	see	see	VERB
cana-6072	230	11	table	table	NOUN
cana-6072	230	12	5	5	NUM
cana-6072	230	13	)	)	PUNCT
cana-6072	230	14	outperformed	outperform	VERB
cana-6072	230	15	early	early	ADJ
cana-6072	230	16	fusion	fusion	NOUN
cana-6072	230	17	lstms	lstms	NOUN
cana-6072	230	18	by	by	ADP
cana-6072	230	19	focusing	focus	VERB
cana-6072	230	20	on	on	ADP
cana-6072	230	21	the	the	DET
cana-6072	230	22	most	most	ADJ
cana-6072	230	23	sentiment	sentiment	NOUN
cana-6072	230	24	-	-	PUNCT
cana-6072	230	25	relevant	relevant	ADJ
cana-6072	230	26	words	word	NOUN
cana-6072	230	27	and	and	CCONJ
cana-6072	230	28	emojis	emoji	NOUN
cana-6072	230	29	in	in	ADP
cana-6072	230	30	tandem	tandem	PROPN
cana-6072	231	1	[	[	X
cana-6072	231	2	lou	lou	PROPN
cana-6072	231	3	,	,	PUNCT
cana-6072	231	4	y.	y.	PROPN
cana-6072	231	5	et	et	PROPN
cana-6072	231	6	al	al	PROPN
cana-6072	231	7	.	.	PROPN
cana-6072	231	8	,	,	PUNCT
cana-6072	231	9	2020	2020	NUM
cana-6072	231	10	]	]	PUNCT
cana-6072	231	11	additionally	additionally	ADV
cana-6072	231	12	,	,	PUNCT
cana-6072	231	13	approaches	approach	VERB
cana-6072	231	14	that	that	PRON
cana-6072	231	15	incorporate	incorporate	VERB
cana-6072	231	16	visual	visual	ADJ
cana-6072	231	17	information	information	NOUN
cana-6072	231	18	of	of	ADP
cana-6072	231	19	emojis	emojis	NOUN
cana-6072	231	20	(	(	PUNCT
cana-6072	231	21	treating	treat	VERB
cana-6072	231	22	emojis	emoji	NOUN
cana-6072	231	23	as	as	ADP
cana-6072	231	24	tiny	tiny	ADJ
cana-6072	231	25	images	image	NOUN
cana-6072	231	26	)	)	PUNCT
cana-6072	231	27	set	set	VERB
cana-6072	231	28	a	a	DET
cana-6072	231	29	new	new	ADJ
cana-6072	231	30	state	state	NOUN
cana-6072	231	31	-	-	PUNCT
cana-6072	231	32	of	of	ADP
cana-6072	231	33	-	-	PUNCT
cana-6072	231	34	the	the	DET
cana-6072	231	35	-	-	PUNCT
cana-6072	231	36	art	art	NOUN
cana-6072	231	37	in	in	ADP
cana-6072	231	38	specific	specific	ADJ
cana-6072	231	39	scenarios	scenario	NOUN
cana-6072	231	40	[	[	X
cana-6072	231	41	lou	lou	PROPN
cana-6072	231	42	,	,	PUNCT
cana-6072	231	43	y.	y.	PROPN
cana-6072	231	44	et	et	PROPN
cana-6072	231	45	al	al	PROPN
cana-6072	231	46	.	.	PROPN
cana-6072	231	47	,	,	PUNCT
cana-6072	231	48	2024	2024	NUM
cana-6072	231	49	]	]	PUNCT
cana-6072	231	50	,	,	PUNCT
cana-6072	231	51	especially	especially	ADV
cana-6072	231	52	to	to	PART
cana-6072	231	53	handle	handle	VERB
cana-6072	231	54	scenarios	scenario	NOUN
cana-6072	231	55	where	where	SCONJ
cana-6072	231	56	an	an	DET
cana-6072	231	57	emoji	emoji	NOUN
cana-6072	231	58	’s	’s	PART
cana-6072	231	59	appearance	appearance	NOUN
cana-6072	231	60	carries	carry	VERB
cana-6072	231	61	meaning	meaning	NOUN
cana-6072	231	62	(	(	PUNCT
cana-6072	231	63	e.g.	e.g.	ADV
cana-6072	231	64	,	,	PUNCT
cana-6072	231	65	the	the	DET
cana-6072	231	66	difference	difference	NOUN
cana-6072	231	67	between	between	ADP
cana-6072	231	68	and	and	CCONJ
cana-6072	231	69	which	which	PRON
cana-6072	231	70	are	be	AUX
cana-6072	231	71	subtle	subtle	ADJ
cana-6072	231	72	in	in	ADP
cana-6072	231	73	text	text	NOUN
cana-6072	231	74	but	but	CCONJ
cana-6072	231	75	clearer	clear	ADJ
cana-6072	231	76	in	in	ADP
cana-6072	231	77	image	image	NOUN
cana-6072	231	78	form	form	NOUN
cana-6072	231	79	)	)	PUNCT
cana-6072	231	80	.	.	PUNCT
cana-6072	232	1	visual	visual	ADJ
cana-6072	232	2	modeling	modeling	NOUN
cana-6072	232	3	also	also	ADV
cana-6072	232	4	offers	offer	VERB
cana-6072	232	5	a	a	DET
cana-6072	232	6	way	way	NOUN
cana-6072	232	7	to	to	PART
cana-6072	232	8	be	be	AUX
cana-6072	232	9	vendor	vendor	NOUN
cana-6072	232	10	-	-	PUNCT
cana-6072	232	11	aware	aware	ADJ
cana-6072	232	12	:	:	PUNCT
cana-6072	232	13	by	by	ADP
cana-6072	232	14	training	training	NOUN
cana-6072	232	15	on	on	ADP
cana-6072	232	16	apple	apple	NOUN
cana-6072	232	17	,	,	PUNCT
cana-6072	232	18	google	google	PROPN
cana-6072	232	19	,	,	PUNCT
cana-6072	232	20	etc	etc	X
cana-6072	232	21	.	.	X
cana-6072	232	22	glyphs	glyphs	PROPN
cana-6072	232	23	,	,	PUNCT
cana-6072	232	24	a	a	DET
cana-6072	232	25	model	model	NOUN
cana-6072	232	26	could	could	AUX
cana-6072	232	27	potentially	potentially	ADV
cana-6072	232	28	account	account	VERB
cana-6072	232	29	for	for	ADP
cana-6072	232	30	different	different	ADJ
cana-6072	232	31	user	user	NOUN
cana-6072	232	32	experiences	experience	NOUN
cana-6072	232	33	(	(	PUNCT
cana-6072	232	34	though	though	SCONJ
cana-6072	232	35	we	we	PRON
cana-6072	232	36	did	do	AUX
cana-6072	232	37	n’t	not	PART
cana-6072	232	38	find	find	VERB
cana-6072	232	39	a	a	DET
cana-6072	232	40	study	study	NOUN
cana-6072	232	41	that	that	PRON
cana-6072	232	42	explicitly	explicitly	ADV
cana-6072	232	43	did	do	VERB
cana-6072	232	44	this	this	PRON
cana-6072	232	45	,	,	PUNCT
cana-6072	232	46	it	it	PRON
cana-6072	232	47	’s	’	VERB
cana-6072	232	48	a	a	DET
cana-6072	232	49	logical	logical	ADJ
cana-6072	232	50	extension	extension	NOUN
cana-6072	232	51	of	of	ADP
cana-6072	232	52	mam	mam	NOUN
cana-6072	232	53	-	-	PUNCT
cana-6072	232	54	emmsa	emmsa	NOUN
cana-6072	233	1	[	[	X
cana-6072	233	2	lou	lou	PROPN
cana-6072	233	3	,	,	PUNCT
cana-6072	233	4	y.	y.	PROPN
cana-6072	233	5	et	et	PROPN
cana-6072	233	6	al	al	PROPN
cana-6072	233	7	.	.	PROPN
cana-6072	233	8	,	,	PUNCT
cana-6072	233	9	2024	2024	NUM
cana-6072	233	10	]	]	PUNCT
cana-6072	233	11	.	.	PUNCT
cana-6072	234	1	despite	despite	SCONJ
cana-6072	234	2	overall	overall	ADJ
cana-6072	234	3	positive	positive	ADJ
cana-6072	234	4	findings	finding	NOUN
cana-6072	234	5	,	,	PUNCT
cana-6072	234	6	there	there	PRON
cana-6072	234	7	are	be	VERB
cana-6072	234	8	failure	failure	NOUN
cana-6072	234	9	modes	mode	NOUN
cana-6072	234	10	.	.	PUNCT
cana-6072	235	1	ironically	ironically	ADV
cana-6072	235	2	,	,	PUNCT
cana-6072	235	3	emojis	emoji	NOUN
cana-6072	235	4	can	can	AUX
cana-6072	235	5	introduce	introduce	VERB
cana-6072	235	6	noise	noise	NOUN
cana-6072	235	7	if	if	SCONJ
cana-6072	235	8	not	not	PART
cana-6072	235	9	handled	handle	VERB
cana-6072	235	10	properly	properly	ADV
cana-6072	235	11	–	–	PUNCT
cana-6072	235	12	e.g.	e.g.	ADV
cana-6072	235	13	,	,	PUNCT
cana-6072	235	14	treating	treat	VERB
cana-6072	235	15	all	all	DET
cana-6072	235	16	emoji	emoji	NOUN
cana-6072	235	17	unicode	unicode	NOUN
cana-6072	235	18	codes	code	NOUN
cana-6072	235	19	the	the	DET
cana-6072	235	20	same	same	ADJ
cana-6072	235	21	regardless	regardless	ADV
cana-6072	235	22	of	of	ADP
cana-6072	235	23	context	context	NOUN
cana-6072	235	24	can	can	AUX
cana-6072	235	25	confuse	confuse	VERB
cana-6072	235	26	models	model	NOUN
cana-6072	235	27	(	(	PUNCT
cana-6072	235	28	the	the	DET
cana-6072	235	29	“	"	PUNCT
cana-6072	235	30	one	one	NUM
cana-6072	235	31	-	-	PUNCT
cana-6072	235	32	hot	hot	ADJ
cana-6072	235	33	emoji	emoji	NOUN
cana-6072	235	34	embedding	embed	VERB
cana-6072	235	35	”	"	PUNCT
cana-6072	235	36	issue	issue	NOUN
cana-6072	235	37	)	)	PUNCT
cana-6072	235	38	.	.	PUNCT
cana-6072	236	1	some	some	DET
cana-6072	236	2	models	model	NOUN
cana-6072	236	3	struggled	struggle	VERB
cana-6072	236	4	with	with	ADP
cana-6072	236	5	emoji	emoji	PROPN
cana-6072	236	6	polysemy	polysemy	NOUN
cana-6072	236	7	:	:	PUNCT
cana-6072	236	8	emojis	emoji	NOUN
cana-6072	236	9	like	like	ADP
cana-6072	236	10	(	(	PUNCT
cana-6072	236	11	folded	fold	VERB
cana-6072	236	12	hands	hand	NOUN
cana-6072	236	13	)	)	PUNCT
cana-6072	236	14	can	can	AUX
cana-6072	236	15	express	express	VERB
cana-6072	236	16	gratitude	gratitude	NOUN
cana-6072	236	17	or	or	CCONJ
cana-6072	236	18	a	a	DET
cana-6072	236	19	plea	plea	NOUN
cana-6072	236	20	,	,	PUNCT
cana-6072	236	21	and	and	CCONJ
cana-6072	236	22	models	model	NOUN
cana-6072	236	23	misclassify	misclassify	VERB
cana-6072	236	24	if	if	SCONJ
cana-6072	236	25	they	they	PRON
cana-6072	236	26	do	do	AUX
cana-6072	236	27	n’t	not	PART
cana-6072	236	28	have	have	VERB
cana-6072	236	29	context	context	NOUN
cana-6072	236	30	.	.	PUNCT
cana-6072	237	1	a	a	DET
cana-6072	237	2	promising	promising	ADJ
cana-6072	237	3	direction	direction	NOUN
cana-6072	237	4	is	be	AUX
cana-6072	237	5	the	the	DET
cana-6072	237	6	“	"	PUNCT
cana-6072	237	7	bi	bi	ADJ
cana-6072	237	8	-	-	ADJ
cana-6072	237	9	sense	sense	ADJ
cana-6072	237	10	emoji	emoji	NOUN
cana-6072	237	11	embedding	embed	VERB
cana-6072	237	12	”	"	PUNCT
cana-6072	237	13	idea	idea	NOUN
cana-6072	237	14	by	by	ADP
cana-6072	237	15	chen	chen	PROPN
cana-6072	237	16	et	et	PROPN
cana-6072	237	17	al	al	PROPN
cana-6072	237	18	.	.	PROPN
cana-6072	238	1	(	(	PUNCT
cana-6072	238	2	2018	2018	NUM
cana-6072	238	3	)	)	PUNCT
cana-6072	238	4	,	,	PUNCT
cana-6072	238	5	who	who	PRON
cana-6072	238	6	explicitly	explicitly	ADV
cana-6072	238	7	learned	learn	VERB
cana-6072	238	8	two	two	NUM
cana-6072	238	9	representations	representation	NOUN
cana-6072	238	10	per	per	ADP
cana-6072	238	11	emoji	emoji	NOUN
cana-6072	238	12	(	(	PUNCT
cana-6072	238	13	one	one	NUM
cana-6072	238	14	for	for	ADP
cana-6072	238	15	positive	positive	ADJ
cana-6072	238	16	contexts	contexts	NOUN
cana-6072	238	17	,	,	PUNCT
cana-6072	238	18	one	one	NUM
cana-6072	238	19	for	for	ADP
cana-6072	238	20	negative	negative	ADJ
cana-6072	238	21	)	)	PUNCT
cana-6072	239	1	[	[	X
cana-6072	239	2	chen	chen	PROPN
cana-6072	239	3	,	,	PUNCT
cana-6072	239	4	y.	y.	PROPN
cana-6072	239	5	et	et	PROPN
cana-6072	239	6	al	al	PROPN
cana-6072	239	7	.	.	PROPN
cana-6072	239	8	,	,	PUNCT
cana-6072	239	9	2018	2018	NUM
cana-6072	239	10	]	]	PUNCT
cana-6072	239	11	.	.	PUNCT
cana-6072	240	1	this	this	PRON
cana-6072	240	2	helped	help	VERB
cana-6072	240	3	their	their	PRON
cana-6072	240	4	attention	attention	NOUN
cana-6072	240	5	model	model	NOUN
cana-6072	240	6	pick	pick	VERB
cana-6072	240	7	the	the	DET
cana-6072	240	8	right	right	ADJ
cana-6072	240	9	sense	sense	NOUN
cana-6072	240	10	and	and	CCONJ
cana-6072	240	11	improved	improve	VERB
cana-6072	240	12	robustness	robustness	NOUN
cana-6072	240	13	on	on	ADP
cana-6072	240	14	twitter	twitter	NOUN
cana-6072	240	15	sentiment	sentiment	NOUN
cana-6072	240	16	.	.	PUNCT
cana-6072	241	1	future	future	ADJ
cana-6072	241	2	research	research	NOUN
cana-6072	241	3	could	could	AUX
cana-6072	241	4	generalize	generalize	VERB
cana-6072	241	5	this	this	PRON
cana-6072	241	6	to	to	ADP
cana-6072	241	7	multi	multi	ADJ
cana-6072	241	8	-	-	ADJ
cana-6072	241	9	sense	sense	ADJ
cana-6072	241	10	embeddings	embedding	NOUN
cana-6072	241	11	or	or	CCONJ
cana-6072	241	12	even	even	ADV
cana-6072	241	13	dynamic	dynamic	ADJ
cana-6072	241	14	sense	sense	NOUN
cana-6072	241	15	inference	inference	NOUN
cana-6072	241	16	(	(	PUNCT
cana-6072	241	17	depending	depend	VERB
cana-6072	241	18	on	on	ADP
cana-6072	241	19	neighboring	neighboring	NOUN
cana-6072	241	20	words	word	NOUN
cana-6072	241	21	)	)	PUNCT
cana-6072	241	22	.	.	PUNCT
cana-6072	242	1	reproducibility	reproducibility	NOUN
cana-6072	242	2	and	and	CCONJ
cana-6072	242	3	evaluation	evaluation	NOUN
cana-6072	242	4	issues	issue	NOUN
cana-6072	242	5	are	be	AUX
cana-6072	242	6	a	a	DET
cana-6072	242	7	notable	notable	ADJ
cana-6072	242	8	concern	concern	NOUN
cana-6072	242	9	.	.	PUNCT
cana-6072	243	1	many	many	ADJ
cana-6072	243	2	papers	paper	NOUN
cana-6072	243	3	still	still	ADV
cana-6072	243	4	do	do	AUX
cana-6072	243	5	not	not	PART
cana-6072	243	6	release	release	VERB
cana-6072	243	7	code	code	NOUN
cana-6072	243	8	or	or	CCONJ
cana-6072	243	9	complete	complete	ADJ
cana-6072	243	10	details	detail	NOUN
cana-6072	243	11	,	,	PUNCT
cana-6072	243	12	making	make	VERB
cana-6072	243	13	it	it	PRON
cana-6072	243	14	hard	hard	ADJ
cana-6072	243	15	to	to	PART
cana-6072	243	16	replicate	replicate	VERB
cana-6072	243	17	results	result	NOUN
cana-6072	243	18	exactly	exactly	ADV
cana-6072	243	19	.	.	PUNCT
cana-6072	244	1	w	w	NOUN
cana-6072	244	2	without	without	ADP
cana-6072	244	3	such	such	ADJ
cana-6072	244	4	rigor	rigor	NOUN
cana-6072	244	5	,	,	PUNCT
cana-6072	244	6	there	there	PRON
cana-6072	244	7	’s	’	VERB
cana-6072	244	8	a	a	DET
cana-6072	244	9	risk	risk	NOUN
cana-6072	244	10	of	of	ADP
cana-6072	244	11	overestimating	overestimate	VERB
cana-6072	244	12	the	the	DET
cana-6072	244	13	benefits	benefit	NOUN
cana-6072	244	14	of	of	ADP
cana-6072	244	15	emojis	emoji	NOUN
cana-6072	244	16	.	.	PUNCT
cana-6072	245	1	we	we	PRON
cana-6072	245	2	found	find	VERB
cana-6072	245	3	at	at	ADP
cana-6072	245	4	least	least	ADV
cana-6072	245	5	one	one	NUM
cana-6072	245	6	case	case	NOUN
cana-6072	245	7	where	where	SCONJ
cana-6072	245	8	an	an	DET
cana-6072	245	9	emoji	emoji	NOUN
cana-6072	245	10	-	-	PUNCT
cana-6072	245	11	aware	aware	ADJ
cana-6072	245	12	model	model	NOUN
cana-6072	245	13	showed	show	VERB
cana-6072	245	14	no	no	DET
cana-6072	245	15	gain	gain	NOUN
cana-6072	245	16	versus	versus	ADP
cana-6072	245	17	a	a	DET
cana-6072	245	18	properly	properly	ADV
cana-6072	245	19	tuned	tune	VERB
cana-6072	245	20	roberta	roberta	PROPN
cana-6072	245	21	baseline	baseline	PROPN
cana-6072	245	22	–	–	PUNCT
cana-6072	245	23	a	a	DET
cana-6072	245	24	result	result	NOUN
cana-6072	245	25	that	that	PRON
cana-6072	245	26	was	be	AUX
cana-6072	245	27	only	only	ADV
cana-6072	245	28	clear	clear	ADJ
cana-6072	245	29	because	because	SCONJ
cana-6072	245	30	the	the	DET
cana-6072	245	31	authors	author	NOUN
cana-6072	245	32	did	do	VERB
cana-6072	245	33	a	a	DET
cana-6072	245	34	fair	fair	ADJ
cana-6072	245	35	comparison	comparison	NOUN
cana-6072	245	36	.	.	PUNCT
cana-6072	246	1	this	this	DET
cana-6072	246	2	kind	kind	NOUN
cana-6072	246	3	of	of	ADP
cana-6072	246	4	honesty	honesty	NOUN
cana-6072	246	5	is	be	AUX
cana-6072	246	6	crucial	crucial	ADJ
cana-6072	246	7	for	for	ADP
cana-6072	246	8	the	the	DET
cana-6072	246	9	field	field	NOUN
cana-6072	246	10	’s	’s	PART
cana-6072	246	11	credibility	credibility	NOUN
cana-6072	246	12	.	.	PUNCT
cana-6072	247	1	generality	generality	NOUN
cana-6072	247	2	vs.	vs.	ADP
cana-6072	247	3	domain	domain	NOUN
cana-6072	247	4	-	-	PUNCT
cana-6072	247	5	specificity	specificity	NOUN
cana-6072	247	6	.	.	PUNCT
cana-6072	248	1	another	another	DET
cana-6072	248	2	discussion	discussion	NOUN
cana-6072	248	3	point	point	NOUN
cana-6072	248	4	is	be	AUX
cana-6072	248	5	how	how	SCONJ
cana-6072	248	6	well	well	ADV
cana-6072	248	7	emoji	emoji	PROPN
cana-6072	248	8	models	model	NOUN
cana-6072	248	9	generalize	generalize	VERB
cana-6072	248	10	.	.	PUNCT
cana-6072	249	1	a	a	DET
cana-6072	249	2	model	model	NOUN
cana-6072	249	3	trained	train	VERB
cana-6072	249	4	on	on	ADP
cana-6072	249	5	twitter	twitter	NOUN
cana-6072	249	6	might	might	AUX
cana-6072	249	7	flounder	flounder	VERB
cana-6072	249	8	on	on	ADP
cana-6072	249	9	reddit	reddit	NOUN
cana-6072	249	10	or	or	CCONJ
cana-6072	249	11	in	in	ADP
cana-6072	249	12	business	business	NOUN
cana-6072	249	13	reviews	review	NOUN
cana-6072	249	14	,	,	PUNCT
cana-6072	249	15	not	not	PART
cana-6072	249	16	just	just	ADV
cana-6072	249	17	because	because	SCONJ
cana-6072	249	18	of	of	ADP
cana-6072	249	19	vocabulary	vocabulary	NOUN
cana-6072	249	20	but	but	CCONJ
cana-6072	249	21	because	because	SCONJ
cana-6072	249	22	emoji	emoji	PROPN
cana-6072	249	23	usage	usage	NOUN
cana-6072	249	24	conventions	convention	NOUN
cana-6072	249	25	differ	differ	VERB
cana-6072	249	26	.	.	PUNCT
cana-6072	250	1	crossdomain	crossdomain	ADJ
cana-6072	250	2	studies	study	NOUN
cana-6072	250	3	are	be	AUX
cana-6072	250	4	scarce	scarce	ADJ
cana-6072	250	5	;	;	PUNCT
cana-6072	250	6	one	one	NUM
cana-6072	250	7	path	path	NOUN
cana-6072	250	8	forward	forward	ADV
cana-6072	250	9	is	be	AUX
cana-6072	250	10	to	to	PART
cana-6072	250	11	build	build	VERB
cana-6072	250	12	models	model	NOUN
cana-6072	250	13	aware	aware	ADJ
cana-6072	250	14	of	of	ADP
cana-6072	250	15	domain	domain	NOUN
cana-6072	250	16	emoji	emoji	NOUN
cana-6072	250	17	frequency	frequency	NOUN
cana-6072	250	18	(	(	PUNCT
cana-6072	250	19	some	some	DET
cana-6072	250	20	emojis	emoji	NOUN
cana-6072	250	21	are	be	AUX
cana-6072	250	22	more	more	ADV
cana-6072	250	23	common	common	ADJ
cana-6072	250	24	in	in	ADP
cana-6072	250	25	certain	certain	ADJ
cana-6072	250	26	communities	community	NOUN
cana-6072	250	27	,	,	PUNCT
cana-6072	250	28	carrying	carry	VERB
cana-6072	250	29	in	in	ADP
cana-6072	250	30	-	-	PUNCT
cana-6072	250	31	jokes	joke	NOUN
cana-6072	250	32	or	or	CCONJ
cana-6072	250	33	specific	specific	ADJ
cana-6072	250	34	meanings	meaning	NOUN
cana-6072	250	35	)	)	PUNCT
cana-6072	250	36	.	.	PUNCT
cana-6072	251	1	similarly	similarly	ADV
cana-6072	251	2	,	,	PUNCT
cana-6072	251	3	cross	cross	ADJ
cana-6072	251	4	-	-	ADJ
cana-6072	251	5	lingual	lingual	ADJ
cana-6072	251	6	sentiment	sentiment	NOUN
cana-6072	251	7	using	use	VERB
cana-6072	251	8	emojis	emoji	NOUN
cana-6072	251	9	is	be	AUX
cana-6072	251	10	a	a	DET
cana-6072	251	11	fascinating	fascinating	ADJ
cana-6072	251	12	area	area	NOUN
cana-6072	251	13	:	:	PUNCT
cana-6072	251	14	zhenpeng	zhenpeng	PROPN
cana-6072	251	15	chen	chen	PROPN
cana-6072	251	16	et	et	PROPN
cana-6072	251	17	al	al	PROPN
cana-6072	251	18	.	.	PROPN
cana-6072	252	1	(	(	PUNCT
cana-6072	252	2	2019	2019	NUM
cana-6072	252	3	)	)	PUNCT
cana-6072	252	4	showed	show	VERB
cana-6072	252	5	that	that	SCONJ
cana-6072	252	6	using	use	VERB
cana-6072	252	7	emojis	emoji	NOUN
cana-6072	252	8	as	as	ADP
cana-6072	252	9	a	a	DET
cana-6072	252	10	“	"	PUNCT
cana-6072	252	11	bridge	bridge	NOUN
cana-6072	252	12	”	"	PUNCT
cana-6072	252	13	can	can	AUX
cana-6072	252	14	transfer	transfer	VERB
cana-6072	252	15	sentiment	sentiment	NOUN
cana-6072	252	16	knowledge	knowledge	NOUN
cana-6072	252	17	between	between	ADP
cana-6072	252	18	english	english	PROPN
cana-6072	252	19	and	and	CCONJ
cana-6072	252	20	spanish	spanish	ADJ
cana-6072	252	21	without	without	ADP
cana-6072	252	22	direct	direct	ADJ
cana-6072	252	23	translation	translation	NOUN
cana-6072	252	24	.	.	PUNCT
cana-6072	253	1	this	this	PRON
cana-6072	253	2	leverages	leverage	VERB
cana-6072	253	3	the	the	DET
cana-6072	253	4	fact	fact	NOUN
cana-6072	253	5	that	that	PRON
cana-6072	253	6	is	be	AUX
cana-6072	253	7	universally	universally	ADV
cana-6072	253	8	understood	understand	VERB
cana-6072	253	9	,	,	PUNCT
cana-6072	253	10	even	even	ADV
cana-6072	253	11	if	if	SCONJ
cana-6072	253	12	“	"	PUNCT
cana-6072	253	13	lol	lol	VERB
cana-6072	253	14	”	"	PUNCT
cana-6072	253	15	vs	vs	ADP
cana-6072	253	16	“	"	PUNCT
cana-6072	253	17	jaja	jaja	PROPN
cana-6072	253	18	”	"	PUNCT
cana-6072	253	19	are	be	AUX
cana-6072	253	20	language	language	NOUN
cana-6072	253	21	-	-	PUNCT
cana-6072	253	22	specific	specific	NOUN
cana-6072	253	23	.	.	PUNCT
cana-6072	254	1	such	such	ADJ
cana-6072	254	2	approaches	approach	NOUN
cana-6072	254	3	could	could	AUX
cana-6072	254	4	democratize	democratize	VERB
cana-6072	254	5	sentiment	sentiment	NOUN
cana-6072	254	6	analysis	analysis	NOUN
cana-6072	254	7	for	for	ADP
cana-6072	254	8	low	low	ADJ
cana-6072	254	9	-	-	PUNCT
cana-6072	254	10	resource	resource	NOUN
cana-6072	254	11	languages	language	NOUN
cana-6072	254	12	by	by	ADP
cana-6072	254	13	using	use	VERB
cana-6072	254	14	emojis	emoji	NOUN
cana-6072	254	15	as	as	ADP
cana-6072	254	16	a	a	DET
cana-6072	254	17	pseudouniversal	pseudouniversal	ADJ
cana-6072	254	18	label	label	NOUN
cana-6072	254	19	.	.	PUNCT
cana-6072	255	1	communications	communication	NOUN
cana-6072	255	2	on	on	ADP
cana-6072	255	3	applied	apply	VERB
cana-6072	255	4	nonlinear	nonlinear	ADJ
cana-6072	255	5	analysis	analysis	NOUN
cana-6072	255	6	issn	issn	NOUN
cana-6072	255	7	:	:	PUNCT
cana-6072	255	8	1074	1074	NUM
cana-6072	255	9	-	-	PUNCT
cana-6072	255	10	133x	133x	NUM
cana-6072	255	11	vol	vol	NOUN
cana-6072	255	12	32	32	NUM
cana-6072	255	13	no	no	NOUN
cana-6072	255	14	.	.	PUNCT
cana-6072	256	1	4s	4s	NUM
cana-6072	256	2	(	(	PUNCT
cana-6072	256	3	2025	2025	NUM
cana-6072	256	4	)	)	PUNCT
cana-6072	256	5	769	769	NUM
cana-6072	257	1	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	257	2	ethical	ethical	ADJ
cana-6072	257	3	and	and	CCONJ
cana-6072	257	4	social	social	ADJ
cana-6072	257	5	considerations	consideration	NOUN
cana-6072	257	6	.	.	PUNCT
cana-6072	258	1	emojis	emoji	NOUN
cana-6072	258	2	are	be	AUX
cana-6072	258	3	fun	fun	ADJ
cana-6072	258	4	and	and	CCONJ
cana-6072	258	5	often	often	ADV
cana-6072	258	6	convey	convey	VERB
cana-6072	258	7	positive	positive	ADJ
cana-6072	258	8	emotions	emotion	NOUN
cana-6072	258	9	,	,	PUNCT
cana-6072	258	10	but	but	CCONJ
cana-6072	258	11	they	they	PRON
cana-6072	258	12	can	can	AUX
cana-6072	258	13	also	also	ADV
cana-6072	258	14	encode	encode	VERB
cana-6072	258	15	sensitive	sensitive	ADJ
cana-6072	258	16	information	information	NOUN
cana-6072	258	17	(	(	PUNCT
cana-6072	258	18	identity	identity	NOUN
cana-6072	258	19	,	,	PUNCT
cana-6072	258	20	sarcasm	sarcasm	NOUN
cana-6072	258	21	directed	direct	VERB
cana-6072	258	22	at	at	ADP
cana-6072	258	23	protected	protect	VERB
cana-6072	258	24	groups	group	NOUN
cana-6072	258	25	,	,	PUNCT
cana-6072	258	26	etc	etc	X
cana-6072	258	27	.	.	X
cana-6072	258	28	)	)	PUNCT
cana-6072	258	29	.	.	PUNCT
cana-6072	259	1	one	one	PRON
cana-6072	259	2	should	should	AUX
cana-6072	259	3	consider	consider	VERB
cana-6072	259	4	biases	bias	NOUN
cana-6072	259	5	–	–	PUNCT
cana-6072	259	6	e.g.	e.g.	ADV
cana-6072	259	7	,	,	PUNCT
cana-6072	259	8	the	the	DET
cana-6072	259	9	emoji	emoji	NOUN
cana-6072	259	10	might	might	AUX
cana-6072	259	11	be	be	AUX
cana-6072	259	12	used	use	VERB
cana-6072	259	13	more	more	ADJ
cana-6072	259	14	by	by	ADP
cana-6072	259	15	younger	young	ADJ
cana-6072	259	16	demographics	demographic	NOUN
cana-6072	259	17	,	,	PUNCT
cana-6072	259	18	or	or	CCONJ
cana-6072	259	19	certain	certain	ADJ
cana-6072	259	20	emojis	emoji	NOUN
cana-6072	259	21	might	might	AUX
cana-6072	259	22	be	be	AUX
cana-6072	259	23	slang	slang	ADJ
cana-6072	259	24	in	in	ADP
cana-6072	259	25	certain	certain	ADJ
cana-6072	259	26	cultures	culture	NOUN
cana-6072	259	27	.	.	PUNCT
cana-6072	260	1	if	if	SCONJ
cana-6072	260	2	a	a	DET
cana-6072	260	3	sentiment	sentiment	NOUN
cana-6072	260	4	model	model	NOUN
cana-6072	260	5	misinterprets	misinterpret	VERB
cana-6072	260	6	these	these	PRON
cana-6072	260	7	,	,	PUNCT
cana-6072	260	8	it	it	PRON
cana-6072	260	9	could	could	AUX
cana-6072	260	10	lead	lead	VERB
cana-6072	260	11	to	to	ADP
cana-6072	260	12	skewed	skewed	ADJ
cana-6072	260	13	analyses	analysis	NOUN
cana-6072	260	14	(	(	PUNCT
cana-6072	260	15	imagine	imagine	VERB
cana-6072	260	16	a	a	DET
cana-6072	260	17	system	system	NOUN
cana-6072	260	18	analyzing	analyze	VERB
cana-6072	260	19	mental	mental	ADJ
cana-6072	260	20	health	health	NOUN
cana-6072	260	21	tweets	tweet	NOUN
cana-6072	260	22	:	:	PUNCT
cana-6072	260	23	interpreting	interpret	VERB
cana-6072	260	24	a	a	DET
cana-6072	260	25	(	(	PUNCT
cana-6072	260	26	crying	cry	VERB
cana-6072	260	27	)	)	PUNCT
cana-6072	260	28	emoji	emoji	PROPN
cana-6072	260	29	always	always	ADV
cana-6072	260	30	as	as	ADP
cana-6072	260	31	negative	negative	ADJ
cana-6072	260	32	might	might	AUX
cana-6072	260	33	overlook	overlook	VERB
cana-6072	260	34	that	that	SCONJ
cana-6072	260	35	it	it	PRON
cana-6072	260	36	’s	’s	AUX
cana-6072	260	37	also	also	ADV
cana-6072	260	38	used	use	VERB
cana-6072	260	39	in	in	ADP
cana-6072	260	40	joking	joke	VERB
cana-6072	260	41	hyperbole	hyperbole	NOUN
cana-6072	260	42	)	)	PUNCT
cana-6072	260	43	.	.	PUNCT
cana-6072	261	1	transparency	transparency	NOUN
cana-6072	261	2	in	in	ADP
cana-6072	261	3	how	how	SCONJ
cana-6072	261	4	models	model	NOUN
cana-6072	261	5	handle	handle	VERB
cana-6072	261	6	emojis	emoji	NOUN
cana-6072	261	7	(	(	PUNCT
cana-6072	261	8	perhaps	perhaps	ADV
cana-6072	261	9	via	via	ADP
cana-6072	261	10	attention	attention	NOUN
cana-6072	261	11	weight	weight	NOUN
cana-6072	261	12	visualizations	visualization	NOUN
cana-6072	261	13	[	[	X
cana-6072	261	14	chen	chen	PROPN
cana-6072	261	15	,	,	PUNCT
cana-6072	261	16	y.	y.	PROPN
cana-6072	261	17	et	et	PROPN
cana-6072	261	18	al	al	PROPN
cana-6072	261	19	.	.	PROPN
cana-6072	261	20	,	,	PUNCT
cana-6072	261	21	2018	2018	NUM
cana-6072	261	22	]	]	PUNCT
cana-6072	261	23	or	or	CCONJ
cana-6072	261	24	example	example	NOUN
cana-6072	261	25	-	-	PUNCT
cana-6072	261	26	based	base	VERB
cana-6072	261	27	explanations	explanation	NOUN
cana-6072	261	28	)	)	PUNCT
cana-6072	261	29	can	can	AUX
cana-6072	261	30	help	help	VERB
cana-6072	261	31	users	user	NOUN
cana-6072	261	32	trust	trust	VERB
cana-6072	261	33	and	and	CCONJ
cana-6072	261	34	appropriately	appropriately	ADV
cana-6072	261	35	use	use	VERB
cana-6072	261	36	these	these	DET
cana-6072	261	37	systems	system	NOUN
cana-6072	261	38	.	.	PUNCT
cana-6072	262	1	future	future	ADJ
cana-6072	262	2	directions	direction	NOUN
cana-6072	262	3	•	•	PRON
cana-6072	262	4	emoji	emoji	PROPN
cana-6072	262	5	sense	sense	PROPN
cana-6072	262	6	disambiguation	disambiguation	NOUN
cana-6072	262	7	:	:	PUNCT
cana-6072	262	8	going	go	VERB
cana-6072	262	9	beyond	beyond	ADP
cana-6072	262	10	bi	bi	NOUN
cana-6072	262	11	-	-	NOUN
cana-6072	262	12	sense	sense	NOUN
cana-6072	262	13	,	,	PUNCT
cana-6072	262	14	possibly	possibly	ADV
cana-6072	262	15	leveraging	leverage	VERB
cana-6072	262	16	emoji	emoji	NOUN
cana-6072	262	17	definition	definition	NOUN
cana-6072	262	18	corpora	corpora	NOUN
cana-6072	262	19	or	or	CCONJ
cana-6072	262	20	user	user	NOUN
cana-6072	262	21	feedback	feedback	NOUN
cana-6072	262	22	to	to	PART
cana-6072	262	23	distinguish	distinguish	VERB
cana-6072	262	24	meanings	meaning	NOUN
cana-6072	262	25	.	.	PUNCT
cana-6072	263	1	incorporating	incorporate	VERB
cana-6072	263	2	context	context	NOUN
cana-6072	263	3	(	(	PUNCT
cana-6072	263	4	topic	topic	NOUN
cana-6072	263	5	or	or	CCONJ
cana-6072	263	6	adjacent	adjacent	ADJ
cana-6072	263	7	words	word	NOUN
cana-6072	263	8	)	)	PUNCT
cana-6072	263	9	to	to	PART
cana-6072	263	10	infer	infer	VERB
cana-6072	263	11	which	which	DET
cana-6072	263	12	“	"	PUNCT
cana-6072	263	13	sense	sense	NOUN
cana-6072	263	14	”	"	PUNCT
cana-6072	263	15	of	of	ADP
cana-6072	263	16	emoji	emoji	PROPN
cana-6072	263	17	is	be	AUX
cana-6072	263	18	intended	intend	VERB
cana-6072	263	19	could	could	AUX
cana-6072	263	20	improve	improve	VERB
cana-6072	263	21	sentiment	sentiment	NOUN
cana-6072	263	22	accuracy	accuracy	NOUN
cana-6072	263	23	for	for	ADP
cana-6072	263	24	ambivalent	ambivalent	ADJ
cana-6072	263	25	emojis	emoji	NOUN
cana-6072	263	26	.	.	PUNCT
cana-6072	264	1	•	•	NUM
cana-6072	264	2	vendor	vendor	NOUN
cana-6072	264	3	-	-	PUNCT
cana-6072	264	4	aware	aware	ADJ
cana-6072	264	5	models	model	NOUN
cana-6072	264	6	:	:	PUNCT
cana-6072	264	7	as	as	SCONJ
cana-6072	264	8	noted	note	VERB
cana-6072	264	9	,	,	PUNCT
cana-6072	264	10	integrating	integrate	VERB
cana-6072	264	11	the	the	DET
cana-6072	264	12	actual	actual	ADJ
cana-6072	264	13	rendered	render	VERB
cana-6072	264	14	image	image	NOUN
cana-6072	264	15	or	or	CCONJ
cana-6072	264	16	at	at	ADP
cana-6072	264	17	least	least	ADJ
cana-6072	264	18	an	an	DET
cana-6072	264	19	embedding	embedding	NOUN
cana-6072	264	20	of	of	ADP
cana-6072	264	21	the	the	DET
cana-6072	264	22	vendor	vendor	NOUN
cana-6072	264	23	i	i	PRON
cana-6072	264	24	d.	d.	VERB
cana-6072	264	25	a	a	DET
cana-6072	264	26	truly	truly	ADV
cana-6072	264	27	robust	robust	ADJ
cana-6072	264	28	model	model	NOUN
cana-6072	264	29	might	might	AUX
cana-6072	264	30	take	take	VERB
cana-6072	264	31	as	as	ADP
cana-6072	264	32	input	input	NOUN
cana-6072	264	33	:	:	PUNCT
cana-6072	264	34	text	text	NOUN
cana-6072	264	35	,	,	PUNCT
cana-6072	264	36	emoji	emoji	PROPN
cana-6072	264	37	unicode	unicode	NOUN
cana-6072	264	38	,	,	PUNCT
cana-6072	264	39	and	and	CCONJ
cana-6072	264	40	a	a	DET
cana-6072	264	41	specifier	specifi	ADJ
cana-6072	264	42	like	like	ADP
cana-6072	264	43	“	"	PUNCT
cana-6072	264	44	platform	platform	NOUN
cana-6072	264	45	=	=	NOUN
cana-6072	264	46	apple	apple	NOUN
cana-6072	264	47	”	"	PUNCT
cana-6072	264	48	to	to	PART
cana-6072	264	49	adjust	adjust	VERB
cana-6072	264	50	sentiment	sentiment	NOUN
cana-6072	264	51	inference	inference	NOUN
cana-6072	264	52	accordingly	accordingly	ADV
cana-6072	264	53	.	.	PUNCT
cana-6072	265	1	•	•	NOUN
cana-6072	265	2	dynamic	dynamic	ADJ
cana-6072	265	3	and	and	CCONJ
cana-6072	265	4	emerging	emerge	VERB
cana-6072	265	5	emojis	emoji	NOUN
cana-6072	265	6	:	:	PUNCT
cana-6072	265	7	the	the	DET
cana-6072	265	8	emoji	emoji	PROPN
cana-6072	265	9	lexicon	lexicon	PROPN
cana-6072	265	10	updates	update	NOUN
cana-6072	265	11	annually	annually	ADV
cana-6072	265	12	.	.	PUNCT
cana-6072	266	1	how	how	SCONJ
cana-6072	266	2	to	to	PART
cana-6072	266	3	continuously	continuously	ADV
cana-6072	266	4	learn	learn	VERB
cana-6072	266	5	new	new	ADJ
cana-6072	266	6	emoji	emoji	PROPN
cana-6072	266	7	sentiment	sentiment	NOUN
cana-6072	266	8	without	without	ADP
cana-6072	266	9	retraining	retrain	VERB
cana-6072	266	10	from	from	ADP
cana-6072	266	11	scratch	scratch	NOUN
cana-6072	266	12	?	?	PUNCT
cana-6072	267	1	few	few	ADJ
cana-6072	267	2	-	-	PUNCT
cana-6072	267	3	shot	shot	NOUN
cana-6072	267	4	learning	learning	NOUN
cana-6072	267	5	or	or	CCONJ
cana-6072	267	6	meta	meta	ADJ
cana-6072	267	7	-	-	PUNCT
cana-6072	267	8	learning	learn	VERB
cana-6072	267	9	approaches	approach	NOUN
cana-6072	267	10	could	could	AUX
cana-6072	267	11	watch	watch	VERB
cana-6072	267	12	new	new	ADJ
cana-6072	267	13	emoji	emoji	PROPN
cana-6072	267	14	usages	usage	NOUN
cana-6072	267	15	for	for	ADP
cana-6072	267	16	a	a	DET
cana-6072	267	17	few	few	ADJ
cana-6072	267	18	weeks	week	NOUN
cana-6072	267	19	and	and	CCONJ
cana-6072	267	20	update	update	VERB
cana-6072	267	21	a	a	DET
cana-6072	267	22	model	model	NOUN
cana-6072	267	23	.	.	PUNCT
cana-6072	268	1	this	this	PRON
cana-6072	268	2	also	also	ADV
cana-6072	268	3	ties	tie	VERB
cana-6072	268	4	to	to	ADP
cana-6072	268	5	timestamp	timestamp	NOUN
cana-6072	268	6	-	-	PUNCT
cana-6072	268	7	awareness	awareness	NOUN
cana-6072	268	8	–	–	PUNCT
cana-6072	268	9	e.g.	e.g.	ADV
cana-6072	268	10	,	,	PUNCT
cana-6072	268	11	(	(	PUNCT
cana-6072	268	12	mask	mask	NOUN
cana-6072	268	13	)	)	PUNCT
cana-6072	268	14	was	be	AUX
cana-6072	268	15	neutral	neutral	ADJ
cana-6072	268	16	pre-2020	pre-2020	NOUN
cana-6072	268	17	,	,	PUNCT
cana-6072	268	18	but	but	CCONJ
cana-6072	268	19	during	during	ADP
cana-6072	268	20	covid	covid	NOUN
cana-6072	268	21	it	it	PRON
cana-6072	268	22	took	take	VERB
cana-6072	268	23	on	on	ADP
cana-6072	268	24	new	new	ADJ
cana-6072	268	25	sentiment	sentiment	NOUN
cana-6072	268	26	connotations	connotation	NOUN
cana-6072	268	27	.	.	PUNCT
cana-6072	269	1	•	•	NUM
cana-6072	269	2	multilingual	multilingual	ADJ
cana-6072	269	3	and	and	CCONJ
cana-6072	269	4	cross	cross	ADJ
cana-6072	269	5	-	-	ADJ
cana-6072	269	6	cultural	cultural	ADJ
cana-6072	269	7	studies	study	NOUN
cana-6072	269	8	:	:	PUNCT
cana-6072	269	9	expanding	expand	VERB
cana-6072	269	10	on	on	ADP
cana-6072	269	11	cross	cross	ADJ
cana-6072	269	12	-	-	ADJ
cana-6072	269	13	lingual	lingual	ADJ
cana-6072	269	14	transfer	transfer	NOUN
cana-6072	269	15	,	,	PUNCT
cana-6072	269	16	one	one	PRON
cana-6072	269	17	could	could	AUX
cana-6072	269	18	use	use	VERB
cana-6072	269	19	emojis	emoji	NOUN
cana-6072	269	20	as	as	ADP
cana-6072	269	21	a	a	DET
cana-6072	269	22	pivot	pivot	NOUN
cana-6072	269	23	to	to	PART
cana-6072	269	24	align	align	VERB
cana-6072	269	25	embeddings	embedding	NOUN
cana-6072	269	26	across	across	ADP
cana-6072	269	27	languages	language	NOUN
cana-6072	269	28	(	(	PUNCT
cana-6072	269	29	since	since	SCONJ
cana-6072	269	30	an	an	DET
cana-6072	269	31	emoji	emoji	NOUN
cana-6072	269	32	“	"	PUNCT
cana-6072	269	33	”	"	PUNCT
cana-6072	269	34	has	have	VERB
cana-6072	269	35	the	the	DET
cana-6072	269	36	same	same	ADJ
cana-6072	269	37	meaning	meaning	NOUN
cana-6072	269	38	regardless	regardless	ADV
cana-6072	269	39	of	of	ADP
cana-6072	269	40	surrounding	surround	VERB
cana-6072	269	41	language	language	NOUN
cana-6072	269	42	text	text	NOUN
cana-6072	269	43	)	)	PUNCT
cana-6072	269	44	.	.	PUNCT
cana-6072	270	1	this	this	PRON
cana-6072	270	2	might	might	AUX
cana-6072	270	3	assist	assist	VERB
cana-6072	270	4	in	in	ADP
cana-6072	270	5	creating	create	VERB
cana-6072	270	6	language	language	NOUN
cana-6072	270	7	-	-	PUNCT
cana-6072	270	8	agnostic	agnostic	ADJ
cana-6072	270	9	sentiment	sentiment	NOUN
cana-6072	270	10	models	model	NOUN
cana-6072	270	11	or	or	CCONJ
cana-6072	270	12	in	in	ADP
cana-6072	270	13	understanding	understand	VERB
cana-6072	270	14	cultural	cultural	ADJ
cana-6072	270	15	sentiment	sentiment	NOUN
cana-6072	270	16	differences	difference	NOUN
cana-6072	270	17	(	(	PUNCT
cana-6072	270	18	e.g.	e.g.	ADV
cana-6072	270	19	,	,	PUNCT
cana-6072	270	20	does	do	AUX
cana-6072	270	21	skew	skew	VERB
cana-6072	270	22	positive	positive	ADJ
cana-6072	270	23	in	in	ADP
cana-6072	270	24	one	one	NUM
cana-6072	270	25	language	language	NOUN
cana-6072	270	26	and	and	CCONJ
cana-6072	270	27	neutral	neutral	ADJ
cana-6072	270	28	in	in	ADP
cana-6072	270	29	another	another	PRON
cana-6072	270	30	?	?	PUNCT
cana-6072	270	31	)	)	PUNCT
cana-6072	270	32	.	.	PUNCT
cana-6072	271	1	•	•	NUM
cana-6072	271	2	causal	causal	ADJ
cana-6072	271	3	analysis	analysis	NOUN
cana-6072	271	4	:	:	PUNCT
cana-6072	271	5	are	be	AUX
cana-6072	271	6	emojis	emoji	NOUN
cana-6072	271	7	simply	simply	ADV
cana-6072	271	8	correlated	correlate	VERB
cana-6072	271	9	with	with	ADP
cana-6072	271	10	sentiment	sentiment	NOUN
cana-6072	271	11	or	or	CCONJ
cana-6072	271	12	actually	actually	ADV
cana-6072	271	13	causal	causal	ADV
cana-6072	271	14	in	in	ADP
cana-6072	271	15	influencing	influence	VERB
cana-6072	271	16	readers	reader	NOUN
cana-6072	271	17	’	’	PART
cana-6072	271	18	perceptions	perception	NOUN
cana-6072	271	19	?	?	PUNCT
cana-6072	272	1	some	some	DET
cana-6072	272	2	user	user	NOUN
cana-6072	272	3	studies	study	NOUN
cana-6072	272	4	could	could	AUX
cana-6072	272	5	be	be	AUX
cana-6072	272	6	done	do	VERB
cana-6072	272	7	to	to	PART
cana-6072	272	8	see	see	VERB
cana-6072	272	9	if	if	SCONJ
cana-6072	272	10	removing	remove	VERB
cana-6072	272	11	emojis	emoji	NOUN
cana-6072	272	12	changes	change	NOUN
cana-6072	272	13	how	how	SCONJ
cana-6072	272	14	humans	human	NOUN
cana-6072	272	15	perceive	perceive	VERB
cana-6072	272	16	the	the	DET
cana-6072	272	17	sentiment	sentiment	NOUN
cana-6072	272	18	of	of	ADP
cana-6072	272	19	a	a	DET
cana-6072	272	20	message	message	NOUN
cana-6072	272	21	.	.	PUNCT
cana-6072	273	1	that	that	PRON
cana-6072	273	2	would	would	AUX
cana-6072	273	3	inform	inform	VERB
cana-6072	273	4	how	how	SCONJ
cana-6072	273	5	models	model	NOUN
cana-6072	273	6	should	should	AUX
cana-6072	273	7	weigh	weigh	VERB
cana-6072	273	8	emoji	emoji	PROPN
cana-6072	273	9	vs	vs	ADP
cana-6072	273	10	text	text	NOUN
cana-6072	273	11	information	information	NOUN
cana-6072	273	12	.	.	PUNCT
cana-6072	274	1	in	in	ADP
cana-6072	274	2	summary	summary	NOUN
cana-6072	274	3	,	,	PUNCT
cana-6072	274	4	the	the	DET
cana-6072	274	5	field	field	NOUN
cana-6072	274	6	has	have	AUX
cana-6072	274	7	progressed	progress	VERB
cana-6072	274	8	from	from	ADP
cana-6072	274	9	treating	treat	VERB
cana-6072	274	10	emojis	emoji	NOUN
cana-6072	274	11	as	as	ADP
cana-6072	274	12	a	a	DET
cana-6072	274	13	curiosity	curiosity	NOUN
cana-6072	274	14	to	to	ADP
cana-6072	274	15	integrating	integrate	VERB
cana-6072	274	16	them	they	PRON
cana-6072	274	17	as	as	ADP
cana-6072	274	18	a	a	DET
cana-6072	274	19	first	first	ADJ
cana-6072	274	20	-	-	PUNCT
cana-6072	274	21	class	class	NOUN
cana-6072	274	22	signal	signal	NOUN
cana-6072	274	23	in	in	ADP
cana-6072	274	24	sentiment	sentiment	NOUN
cana-6072	274	25	analysis	analysis	NOUN
cana-6072	274	26	.	.	PUNCT
cana-6072	275	1	the	the	DET
cana-6072	275	2	best	well	ADV
cana-6072	275	3	-	-	PUNCT
cana-6072	275	4	performing	perform	VERB
cana-6072	275	5	current	current	ADJ
cana-6072	275	6	models	model	NOUN
cana-6072	275	7	use	use	VERB
cana-6072	275	8	advanced	advanced	ADJ
cana-6072	275	9	fusion	fusion	NOUN
cana-6072	275	10	to	to	PART
cana-6072	275	11	capture	capture	VERB
cana-6072	275	12	nuance	nuance	NOUN
cana-6072	275	13	,	,	PUNCT
cana-6072	275	14	and	and	CCONJ
cana-6072	275	15	as	as	ADP
cana-6072	275	16	a	a	DET
cana-6072	275	17	result	result	NOUN
cana-6072	275	18	,	,	PUNCT
cana-6072	275	19	sentiment	sentiment	NOUN
cana-6072	275	20	analysis	analysis	NOUN
cana-6072	275	21	is	be	AUX
cana-6072	275	22	more	more	ADV
cana-6072	275	23	accurate	accurate	ADJ
cana-6072	275	24	on	on	ADP
cana-6072	275	25	emoji	emoji	ADJ
cana-6072	275	26	-	-	ADJ
cana-6072	275	27	rich	rich	ADJ
cana-6072	275	28	data	datum	NOUN
cana-6072	275	29	than	than	ADP
cana-6072	275	30	ever	ever	ADV
cana-6072	275	31	before	before	ADV
cana-6072	275	32	.	.	PUNCT
cana-6072	276	1	however	however	ADV
cana-6072	276	2	,	,	PUNCT
cana-6072	276	3	careful	careful	ADJ
cana-6072	276	4	evaluation	evaluation	NOUN
cana-6072	276	5	is	be	AUX
cana-6072	276	6	necessary	necessary	ADJ
cana-6072	276	7	to	to	PART
cana-6072	276	8	ensure	ensure	VERB
cana-6072	276	9	that	that	SCONJ
cana-6072	276	10	these	these	DET
cana-6072	276	11	gains	gain	NOUN
cana-6072	276	12	are	be	AUX
cana-6072	276	13	real	real	ADJ
cana-6072	276	14	and	and	CCONJ
cana-6072	276	15	generalizable	generalizable	ADJ
cana-6072	276	16	.	.	PUNCT
cana-6072	277	1	by	by	ADP
cana-6072	277	2	following	follow	VERB
cana-6072	277	3	rigorous	rigorous	ADJ
cana-6072	277	4	evaluation	evaluation	NOUN
cana-6072	277	5	practices	practice	NOUN
cana-6072	277	6	and	and	CCONJ
cana-6072	277	7	exploring	explore	VERB
cana-6072	277	8	the	the	DET
cana-6072	277	9	open	open	ADJ
cana-6072	277	10	challenges	challenge	NOUN
cana-6072	277	11	identified	identify	VERB
cana-6072	277	12	,	,	PUNCT
cana-6072	277	13	researchers	researcher	NOUN
cana-6072	277	14	can	can	AUX
cana-6072	277	15	build	build	VERB
cana-6072	277	16	on	on	ADP
cana-6072	277	17	a	a	DET
cana-6072	277	18	solid	solid	ADJ
cana-6072	277	19	foundation	foundation	NOUN
cana-6072	277	20	toward	toward	ADP
cana-6072	277	21	truly	truly	ADV
cana-6072	277	22	robust	robust	ADJ
cana-6072	277	23	sentiment	sentiment	NOUN
cana-6072	277	24	understanding	understanding	NOUN
cana-6072	277	25	in	in	ADP
cana-6072	277	26	our	our	PRON
cana-6072	277	27	emoji	emoji	NOUN
cana-6072	277	28	-	-	PUNCT
cana-6072	277	29	filled	fill	VERB
cana-6072	277	30	communication	communication	NOUN
cana-6072	277	31	landscape	landscape	NOUN
cana-6072	277	32	.	.	PUNCT
cana-6072	278	1	limitations	limitation	NOUN
cana-6072	278	2	of	of	ADP
cana-6072	278	3	the	the	DET
cana-6072	278	4	study	study	NOUN
cana-6072	278	5	the	the	DET
cana-6072	278	6	set	set	NOUN
cana-6072	278	7	of	of	ADP
cana-6072	278	8	included	include	VERB
cana-6072	278	9	studies	study	NOUN
cana-6072	278	10	and	and	CCONJ
cana-6072	278	11	our	our	PRON
cana-6072	278	12	focus	focus	NOUN
cana-6072	278	13	on	on	ADP
cana-6072	278	14	english	english	ADJ
cana-6072	278	15	-	-	PUNCT
cana-6072	278	16	language	language	NOUN
cana-6072	278	17	publications	publication	NOUN
cana-6072	278	18	limit	limit	VERB
cana-6072	278	19	this	this	DET
cana-6072	278	20	review	review	NOUN
cana-6072	278	21	.	.	PUNCT
cana-6072	279	1	while	while	SCONJ
cana-6072	279	2	we	we	PRON
cana-6072	279	3	attempted	attempt	VERB
cana-6072	279	4	to	to	PART
cana-6072	279	5	cover	cover	VERB
cana-6072	279	6	a	a	DET
cana-6072	279	7	comprehensive	comprehensive	ADJ
cana-6072	279	8	range	range	NOUN
cana-6072	279	9	of	of	ADP
cana-6072	279	10	venues	venue	NOUN
cana-6072	279	11	(	(	PUNCT
cana-6072	279	12	including	include	VERB
cana-6072	279	13	those	those	PRON
cana-6072	279	14	indexed	index	VERB
cana-6072	279	15	by	by	ADP
cana-6072	279	16	scopus	scopus	PROPN
cana-6072	279	17	,	,	PUNCT
cana-6072	279	18	ieee	ieee	NOUN
cana-6072	279	19	,	,	PUNCT
cana-6072	279	20	acm	acm	PROPN
cana-6072	279	21	,	,	PUNCT
cana-6072	279	22	acl	acl	PROPN
cana-6072	279	23	,	,	PUNCT
cana-6072	279	24	etc	etc	X
cana-6072	279	25	.	.	X
cana-6072	279	26	)	)	PUNCT
cana-6072	279	27	,	,	PUNCT
cana-6072	279	28	it	it	PRON
cana-6072	279	29	’s	’	VERB
cana-6072	279	30	possible	possible	ADJ
cana-6072	279	31	that	that	SCONJ
cana-6072	279	32	relevant	relevant	ADJ
cana-6072	279	33	industry	industry	NOUN
cana-6072	279	34	developments	development	NOUN
cana-6072	279	35	or	or	CCONJ
cana-6072	279	36	nonenglish	nonenglish	ADJ
cana-6072	279	37	works	work	NOUN
cana-6072	279	38	were	be	AUX
cana-6072	279	39	missed	miss	VERB
cana-6072	279	40	.	.	PUNCT
cana-6072	280	1	the	the	DET
cana-6072	280	2	synthesis	synthesis	NOUN
cana-6072	280	3	is	be	AUX
cana-6072	280	4	qualitative	qualitative	ADJ
cana-6072	280	5	;	;	PUNCT
cana-6072	280	6	we	we	PRON
cana-6072	280	7	did	do	AUX
cana-6072	280	8	not	not	PART
cana-6072	280	9	perform	perform	VERB
cana-6072	280	10	a	a	DET
cana-6072	280	11	meta	meta	ADJ
cana-6072	280	12	-	-	PUNCT
cana-6072	280	13	analysis	analysis	NOUN
cana-6072	280	14	communications	communication	NOUN
cana-6072	280	15	on	on	ADP
cana-6072	280	16	applied	apply	VERB
cana-6072	280	17	nonlinear	nonlinear	ADJ
cana-6072	280	18	analysis	analysis	NOUN
cana-6072	280	19	issn	issn	NOUN
cana-6072	280	20	:	:	PUNCT
cana-6072	280	21	1074	1074	NUM
cana-6072	280	22	-	-	PUNCT
cana-6072	280	23	133x	133x	NUM
cana-6072	280	24	vol	vol	NOUN
cana-6072	280	25	32	32	NUM
cana-6072	280	26	no	no	NOUN
cana-6072	280	27	.	.	PUNCT
cana-6072	281	1	4s	4s	NUM
cana-6072	281	2	(	(	PUNCT
cana-6072	281	3	2025	2025	NUM
cana-6072	281	4	)	)	PUNCT
cana-6072	281	5	770	770	NUM
cana-6072	281	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	281	7	since	since	SCONJ
cana-6072	281	8	evaluation	evaluation	NOUN
cana-6072	281	9	metrics	metric	NOUN
cana-6072	281	10	and	and	CCONJ
cana-6072	281	11	datasets	dataset	NOUN
cana-6072	281	12	varied	vary	VERB
cana-6072	281	13	widely	widely	ADV
cana-6072	281	14	.	.	PUNCT
cana-6072	282	1	however	however	ADV
cana-6072	282	2	,	,	PUNCT
cana-6072	282	3	the	the	DET
cana-6072	282	4	prisma	prisma	NOUN
cana-6072	282	5	-	-	PUNCT
cana-6072	282	6	guided	guide	VERB
cana-6072	282	7	methodology	methodology	NOUN
cana-6072	282	8	and	and	CCONJ
cana-6072	282	9	explicit	explicit	ADJ
cana-6072	282	10	data	datum	NOUN
cana-6072	282	11	extraction	extraction	NOUN
cana-6072	282	12	schema	schema	NOUN
cana-6072	282	13	(	(	PUNCT
cana-6072	282	14	table	table	NOUN
cana-6072	282	15	1	1	NUM
cana-6072	282	16	)	)	PUNCT
cana-6072	282	17	strengthen	strengthen	VERB
cana-6072	282	18	the	the	DET
cana-6072	282	19	reliability	reliability	NOUN
cana-6072	282	20	of	of	ADP
cana-6072	282	21	our	our	PRON
cana-6072	282	22	conclusions	conclusion	NOUN
cana-6072	282	23	.	.	PUNCT
cana-6072	283	1	another	another	DET
cana-6072	283	2	limitation	limitation	NOUN
cana-6072	283	3	is	be	AUX
cana-6072	283	4	that	that	SCONJ
cana-6072	283	5	the	the	DET
cana-6072	283	6	fast	fast	ADV
cana-6072	283	7	-	-	PUNCT
cana-6072	283	8	moving	move	VERB
cana-6072	283	9	nature	nature	NOUN
cana-6072	283	10	of	of	ADP
cana-6072	283	11	nlp	nlp	NOUN
cana-6072	283	12	means	mean	VERB
cana-6072	283	13	new	new	ADJ
cana-6072	283	14	emoji	emoji	PROPN
cana-6072	283	15	-	-	PUNCT
cana-6072	283	16	related	relate	VERB
cana-6072	283	17	innovations	innovation	NOUN
cana-6072	283	18	(	(	PUNCT
cana-6072	283	19	e.g.	e.g.	ADV
cana-6072	283	20	,	,	PUNCT
cana-6072	283	21	in	in	ADP
cana-6072	283	22	2025	2025	NUM
cana-6072	283	23	and	and	CCONJ
cana-6072	283	24	beyond	beyond	ADP
cana-6072	283	25	)	)	PUNCT
cana-6072	283	26	might	might	AUX
cana-6072	283	27	not	not	PART
cana-6072	283	28	be	be	AUX
cana-6072	283	29	captured	capture	VERB
cana-6072	283	30	here	here	ADV
cana-6072	283	31	.	.	PUNCT
cana-6072	284	1	we	we	PRON
cana-6072	284	2	encourage	encourage	VERB
cana-6072	284	3	readers	reader	NOUN
cana-6072	284	4	to	to	PART
cana-6072	284	5	view	view	VERB
cana-6072	284	6	this	this	DET
cana-6072	284	7	review	review	NOUN
cana-6072	284	8	as	as	ADP
cana-6072	284	9	a	a	DET
cana-6072	284	10	snapshot	snapshot	NOUN
cana-6072	284	11	up	up	ADP
cana-6072	284	12	to	to	ADP
cana-6072	284	13	2024	2024	NUM
cana-6072	284	14	.	.	PUNCT
cana-6072	285	1	our	our	PRON
cana-6072	285	2	risk	risk	NOUN
cana-6072	285	3	-	-	PUNCT
cana-6072	285	4	of	of	ADP
cana-6072	285	5	-	-	PUNCT
cana-6072	285	6	bias	bias	NOUN
cana-6072	285	7	appraisal	appraisal	NOUN
cana-6072	285	8	(	(	PUNCT
cana-6072	285	9	table	table	NOUN
cana-6072	285	10	6	6	NUM
cana-6072	285	11	)	)	PUNCT
cana-6072	285	12	suggests	suggest	VERB
cana-6072	285	13	many	many	ADJ
cana-6072	285	14	studies	study	NOUN
cana-6072	285	15	have	have	VERB
cana-6072	285	16	a	a	DET
cana-6072	285	17	medium	medium	ADJ
cana-6072	285	18	risk	risk	NOUN
cana-6072	285	19	;	;	PUNCT
cana-6072	285	20	thus	thus	ADV
cana-6072	285	21	,	,	PUNCT
cana-6072	285	22	some	some	DET
cana-6072	285	23	reported	report	VERB
cana-6072	285	24	improvements	improvement	NOUN
cana-6072	285	25	should	should	AUX
cana-6072	285	26	be	be	AUX
cana-6072	285	27	interpreted	interpret	VERB
cana-6072	285	28	with	with	ADP
cana-6072	285	29	caution	caution	NOUN
cana-6072	285	30	.	.	PUNCT
cana-6072	286	1	expanding	expand	VERB
cana-6072	286	2	this	this	DET
cana-6072	286	3	review	review	NOUN
cana-6072	286	4	to	to	ADP
cana-6072	286	5	more	more	ADJ
cana-6072	286	6	languages	language	NOUN
cana-6072	286	7	or	or	CCONJ
cana-6072	286	8	to	to	ADP
cana-6072	286	9	closely	closely	ADV
cana-6072	286	10	related	relate	VERB
cana-6072	286	11	tasks	task	NOUN
cana-6072	286	12	(	(	PUNCT
cana-6072	286	13	like	like	ADP
cana-6072	286	14	emotion	emotion	NOUN
cana-6072	286	15	intensity	intensity	NOUN
cana-6072	286	16	prediction	prediction	NOUN
cana-6072	286	17	with	with	ADP
cana-6072	286	18	emojis	emoji	NOUN
cana-6072	286	19	)	)	PUNCT
cana-6072	286	20	would	would	AUX
cana-6072	286	21	be	be	AUX
cana-6072	286	22	a	a	DET
cana-6072	286	23	valuable	valuable	ADJ
cana-6072	286	24	extension	extension	NOUN
cana-6072	286	25	.	.	PUNCT
cana-6072	287	1	conclusion	conclusion	NOUN
cana-6072	287	2	emojis	emoji	NOUN
cana-6072	287	3	are	be	AUX
cana-6072	287	4	no	no	ADV
cana-6072	287	5	longer	long	ADV
cana-6072	287	6	just	just	ADV
cana-6072	287	7	cute	cute	VERB
cana-6072	287	8	add	add	ADJ
cana-6072	287	9	-	-	PUNCT
cana-6072	287	10	ons	ons	NOUN
cana-6072	287	11	to	to	PART
cana-6072	287	12	text	text	NOUN
cana-6072	287	13	;	;	PUNCT
cana-6072	287	14	they	they	PRON
cana-6072	287	15	carry	carry	VERB
cana-6072	287	16	sentiment	sentiment	NOUN
cana-6072	287	17	and	and	CCONJ
cana-6072	287	18	thus	thus	ADV
cana-6072	287	19	demand	demand	VERB
cana-6072	287	20	attention	attention	NOUN
cana-6072	287	21	in	in	ADP
cana-6072	287	22	sentiment	sentiment	NOUN
cana-6072	287	23	analysis	analysis	NOUN
cana-6072	287	24	.	.	PUNCT
cana-6072	288	1	this	this	DET
cana-6072	288	2	systematic	systematic	ADJ
cana-6072	288	3	review	review	NOUN
cana-6072	288	4	,	,	PUNCT
cana-6072	288	5	following	follow	VERB
cana-6072	288	6	a	a	DET
cana-6072	288	7	prisma	prisma	NOUN
cana-6072	288	8	protocol	protocol	NOUN
cana-6072	288	9	,	,	PUNCT
cana-6072	288	10	shows	show	VERB
cana-6072	288	11	that	that	SCONJ
cana-6072	288	12	incorporating	incorporate	VERB
cana-6072	288	13	emojis	emoji	NOUN
cana-6072	288	14	can	can	AUX
cana-6072	288	15	indeed	indeed	ADV
cana-6072	288	16	improve	improve	VERB
cana-6072	288	17	sentiment	sentiment	NOUN
cana-6072	288	18	classification	classification	NOUN
cana-6072	288	19	–	–	PUNCT
cana-6072	288	20	but	but	CCONJ
cana-6072	288	21	doing	do	VERB
cana-6072	288	22	so	so	ADV
cana-6072	288	23	robustly	robustly	ADV
cana-6072	288	24	requires	require	VERB
cana-6072	288	25	careful	careful	ADJ
cana-6072	288	26	consideration	consideration	NOUN
cana-6072	288	27	of	of	ADP
cana-6072	288	28	representation	representation	NOUN
cana-6072	288	29	,	,	PUNCT
cana-6072	288	30	fusion	fusion	NOUN
cana-6072	288	31	,	,	PUNCT
cana-6072	288	32	and	and	CCONJ
cana-6072	288	33	evaluation	evaluation	NOUN
cana-6072	288	34	.	.	PUNCT
cana-6072	289	1	lexicons	lexicon	NOUN
cana-6072	289	2	and	and	CCONJ
cana-6072	289	3	simple	simple	ADJ
cana-6072	289	4	techniques	technique	NOUN
cana-6072	289	5	can	can	AUX
cana-6072	289	6	offer	offer	VERB
cana-6072	289	7	quick	quick	ADJ
cana-6072	289	8	wins	win	NOUN
cana-6072	289	9	,	,	PUNCT
cana-6072	289	10	but	but	CCONJ
cana-6072	289	11	more	more	ADV
cana-6072	289	12	sophisticated	sophisticated	ADJ
cana-6072	289	13	methods	method	NOUN
cana-6072	289	14	(	(	PUNCT
cana-6072	289	15	attention	attention	NOUN
cana-6072	289	16	mechanisms	mechanism	NOUN
cana-6072	289	17	,	,	PUNCT
cana-6072	289	18	multimodal	multimodal	ADJ
cana-6072	289	19	fusion	fusion	NOUN
cana-6072	289	20	)	)	PUNCT
cana-6072	289	21	unlock	unlock	NOUN
cana-6072	289	22	the	the	DET
cana-6072	289	23	full	full	ADJ
cana-6072	289	24	potential	potential	NOUN
cana-6072	289	25	of	of	ADP
cana-6072	289	26	emoji	emoji	PROPN
cana-6072	289	27	signals	signal	NOUN
cana-6072	289	28	.	.	PUNCT
cana-6072	290	1	we	we	PRON
cana-6072	290	2	found	find	VERB
cana-6072	290	3	that	that	SCONJ
cana-6072	290	4	the	the	DET
cana-6072	290	5	best	good	ADJ
cana-6072	290	6	emojiaware	emojiaware	NOUN
cana-6072	290	7	models	model	NOUN
cana-6072	290	8	leverage	leverage	VERB
cana-6072	290	9	interactions	interaction	NOUN
cana-6072	290	10	between	between	ADP
cana-6072	290	11	text	text	NOUN
cana-6072	290	12	and	and	CCONJ
cana-6072	290	13	emoji	emoji	NOUN
cana-6072	290	14	,	,	PUNCT
cana-6072	290	15	often	often	ADV
cana-6072	290	16	through	through	ADP
cana-6072	290	17	attention	attention	NOUN
cana-6072	290	18	,	,	PUNCT
cana-6072	290	19	and	and	CCONJ
cana-6072	290	20	sometimes	sometimes	ADV
cana-6072	290	21	even	even	ADV
cana-6072	290	22	integrate	integrate	VERB
cana-6072	290	23	the	the	DET
cana-6072	290	24	visual	visual	ADJ
cana-6072	290	25	emoji	emoji	PROPN
cana-6072	290	26	rendering	rendering	NOUN
cana-6072	290	27	for	for	ADP
cana-6072	290	28	additional	additional	ADJ
cana-6072	290	29	context	context	NOUN
cana-6072	290	30	.	.	PUNCT
cana-6072	291	1	we	we	PRON
cana-6072	291	2	also	also	ADV
cana-6072	291	3	highlighted	highlight	VERB
cana-6072	291	4	common	common	ADJ
cana-6072	291	5	pitfalls	pitfall	NOUN
cana-6072	291	6	:	:	PUNCT
cana-6072	291	7	without	without	ADP
cana-6072	291	8	apples	apple	NOUN
cana-6072	291	9	-	-	PUNCT
cana-6072	291	10	to	to	ADP
cana-6072	291	11	-	-	PUNCT
cana-6072	291	12	apples	apples	NOUN
cana-6072	291	13	evaluation	evaluation	NOUN
cana-6072	291	14	,	,	PUNCT
cana-6072	291	15	improvements	improvement	NOUN
cana-6072	291	16	may	may	AUX
cana-6072	291	17	be	be	AUX
cana-6072	291	18	overstated	overstate	VERB
cana-6072	291	19	,	,	PUNCT
cana-6072	291	20	and	and	CCONJ
cana-6072	291	21	without	without	ADP
cana-6072	291	22	parity	parity	NOUN
cana-6072	291	23	in	in	ADP
cana-6072	291	24	baselines	baseline	NOUN
cana-6072	291	25	,	,	PUNCT
cana-6072	291	26	one	one	PRON
cana-6072	291	27	might	might	AUX
cana-6072	291	28	credit	credit	NOUN
cana-6072	291	29	emojis	emoji	NOUN
cana-6072	291	30	for	for	ADP
cana-6072	291	31	what	what	PRON
cana-6072	291	32	is	be	AUX
cana-6072	291	33	actually	actually	ADV
cana-6072	291	34	an	an	DET
cana-6072	291	35	architecture	architecture	NOUN
cana-6072	291	36	advantage	advantage	NOUN
cana-6072	291	37	.	.	PUNCT
cana-6072	292	1	in	in	ADP
cana-6072	292	2	practical	practical	ADJ
cana-6072	292	3	terms	term	NOUN
cana-6072	292	4	,	,	PUNCT
cana-6072	292	5	sentiment	sentiment	NOUN
cana-6072	292	6	analysis	analysis	NOUN
cana-6072	292	7	systems	system	NOUN
cana-6072	292	8	should	should	AUX
cana-6072	292	9	start	start	VERB
cana-6072	292	10	treating	treat	VERB
cana-6072	292	11	emojis	emoji	NOUN
cana-6072	292	12	as	as	ADP
cana-6072	292	13	firstclass	firstclass	ADP
cana-6072	292	14	citizens	citizen	NOUN
cana-6072	292	15	.	.	PUNCT
cana-6072	293	1	a	a	DET
cana-6072	293	2	modern	modern	ADJ
cana-6072	293	3	tweet	tweet	NOUN
cana-6072	293	4	sentiment	sentiment	NOUN
cana-6072	293	5	classifier	classifier	NOUN
cana-6072	293	6	,	,	PUNCT
cana-6072	293	7	for	for	ADP
cana-6072	293	8	instance	instance	NOUN
cana-6072	293	9	,	,	PUNCT
cana-6072	293	10	could	could	AUX
cana-6072	293	11	use	use	VERB
cana-6072	293	12	a	a	DET
cana-6072	293	13	bert	bert	NOUN
cana-6072	293	14	-	-	PUNCT
cana-6072	293	15	based	base	VERB
cana-6072	293	16	model	model	NOUN
cana-6072	293	17	that	that	PRON
cana-6072	293	18	knows	know	VERB
cana-6072	293	19	how	how	SCONJ
cana-6072	293	20	to	to	PART
cana-6072	293	21	handle	handle	VERB
cana-6072	293	22	“	"	PUNCT
cana-6072	293	23	”	"	PUNCT
cana-6072	293	24	or	or	CCONJ
cana-6072	293	25	“	"	PUNCT
cana-6072	293	26	”	"	PUNCT
cana-6072	293	27	in	in	ADP
cana-6072	293	28	context	context	NOUN
cana-6072	293	29	rather	rather	ADV
cana-6072	293	30	than	than	ADP
cana-6072	293	31	dropping	drop	VERB
cana-6072	293	32	them	they	PRON
cana-6072	293	33	.	.	PUNCT
cana-6072	294	1	such	such	DET
cana-6072	294	2	a	a	DET
cana-6072	294	3	model	model	NOUN
cana-6072	294	4	will	will	AUX
cana-6072	294	5	be	be	AUX
cana-6072	294	6	more	more	ADV
cana-6072	294	7	attuned	attuned	ADJ
cana-6072	294	8	to	to	ADP
cana-6072	294	9	the	the	DET
cana-6072	294	10	genuine	genuine	ADJ
cana-6072	294	11	tone	tone	NOUN
cana-6072	294	12	of	of	ADP
cana-6072	294	13	user	user	NOUN
cana-6072	294	14	messages	message	NOUN
cana-6072	294	15	.	.	PUNCT
cana-6072	295	1	as	as	SCONJ
cana-6072	295	2	we	we	PRON
cana-6072	295	3	move	move	VERB
cana-6072	295	4	toward	toward	ADP
cana-6072	295	5	more	more	ADJ
cana-6072	295	6	human	human	NOUN
cana-6072	295	7	-	-	PUNCT
cana-6072	295	8	centered	center	VERB
cana-6072	295	9	nlp	nlp	NOUN
cana-6072	295	10	,	,	PUNCT
cana-6072	295	11	recognizing	recognize	VERB
cana-6072	295	12	the	the	DET
cana-6072	295	13	role	role	NOUN
cana-6072	295	14	of	of	ADP
cana-6072	295	15	paralinguistic	paralinguistic	ADJ
cana-6072	295	16	elements	element	NOUN
cana-6072	295	17	like	like	ADP
cana-6072	295	18	emojis	emoji	NOUN
cana-6072	295	19	,	,	PUNCT
cana-6072	295	20	emoticons	emoticon	NOUN
cana-6072	295	21	,	,	PUNCT
cana-6072	295	22	and	and	CCONJ
cana-6072	295	23	even	even	ADV
cana-6072	295	24	gifs	gif	NOUN
cana-6072	295	25	or	or	CCONJ
cana-6072	295	26	stickers	sticker	NOUN
cana-6072	295	27	will	will	AUX
cana-6072	295	28	be	be	AUX
cana-6072	295	29	crucial	crucial	ADJ
cana-6072	295	30	for	for	SCONJ
cana-6072	295	31	systems	system	NOUN
cana-6072	295	32	to	to	PART
cana-6072	295	33	truly	truly	ADV
cana-6072	295	34	grasp	grasp	VERB
cana-6072	295	35	user	user	NOUN
cana-6072	295	36	sentiment	sentiment	NOUN
cana-6072	295	37	and	and	CCONJ
cana-6072	295	38	intent	intent	NOUN
cana-6072	295	39	.	.	PUNCT
cana-6072	296	1	in	in	ADP
cana-6072	296	2	conclusion	conclusion	NOUN
cana-6072	296	3	,	,	PUNCT
cana-6072	296	4	our	our	PRON
cana-6072	296	5	review	review	NOUN
cana-6072	296	6	underscores	underscore	VERB
cana-6072	296	7	that	that	SCONJ
cana-6072	296	8	robust	robust	ADJ
cana-6072	296	9	,	,	PUNCT
cana-6072	296	10	reproducible	reproducible	ADJ
cana-6072	296	11	gains	gain	NOUN
cana-6072	296	12	from	from	ADP
cana-6072	296	13	emojis	emoji	NOUN
cana-6072	296	14	require	require	VERB
cana-6072	296	15	principled	principled	ADJ
cana-6072	296	16	representation	representation	NOUN
cana-6072	296	17	,	,	PUNCT
cana-6072	296	18	contextual	contextual	ADJ
cana-6072	296	19	fusion	fusion	NOUN
cana-6072	296	20	,	,	PUNCT
cana-6072	296	21	and	and	CCONJ
cana-6072	296	22	rigorous	rigorous	ADJ
cana-6072	296	23	evaluation	evaluation	NOUN
cana-6072	296	24	.	.	PUNCT
cana-6072	297	1	we	we	PRON
cana-6072	297	2	are	be	AUX
cana-6072	297	3	optimistic	optimistic	ADJ
cana-6072	297	4	that	that	SCONJ
cana-6072	297	5	with	with	ADP
cana-6072	297	6	the	the	DET
cana-6072	297	7	roadmap	roadmap	NOUN
cana-6072	297	8	laid	lay	VERB
cana-6072	297	9	out	out	ADP
cana-6072	297	10	here	here	ADV
cana-6072	297	11	and	and	CCONJ
cana-6072	297	12	a	a	DET
cana-6072	297	13	continued	continue	VERB
cana-6072	297	14	emphasis	emphasis	NOUN
cana-6072	297	15	on	on	ADP
cana-6072	297	16	reproducibility	reproducibility	NOUN
cana-6072	297	17	emoji	emoji	PROPN
cana-6072	297	18	-	-	PUNCT
cana-6072	297	19	aware	aware	ADJ
cana-6072	297	20	sentiment	sentiment	NOUN
cana-6072	297	21	analysis	analysis	NOUN
cana-6072	297	22	will	will	AUX
cana-6072	297	23	transition	transition	VERB
cana-6072	297	24	from	from	ADP
cana-6072	297	25	a	a	DET
cana-6072	297	26	niche	niche	NOUN
cana-6072	297	27	topic	topic	NOUN
cana-6072	297	28	to	to	ADP
cana-6072	297	29	a	a	DET
cana-6072	297	30	standard	standard	ADJ
cana-6072	297	31	component	component	NOUN
cana-6072	297	32	of	of	ADP
cana-6072	297	33	sentiment	sentiment	NOUN
cana-6072	297	34	analysis	analysis	NOUN
cana-6072	297	35	toolkits	toolkit	NOUN
cana-6072	297	36	,	,	PUNCT
cana-6072	297	37	enriching	enrich	VERB
cana-6072	297	38	the	the	DET
cana-6072	297	39	fidelity	fidelity	NOUN
cana-6072	297	40	with	with	ADP
cana-6072	297	41	which	which	PRON
cana-6072	297	42	we	we	PRON
cana-6072	297	43	can	can	AUX
cana-6072	297	44	computationally	computationally	ADV
cana-6072	297	45	understand	understand	VERB
cana-6072	297	46	human	human	ADJ
cana-6072	297	47	expression	expression	NOUN
cana-6072	297	48	in	in	ADP
cana-6072	297	49	the	the	DET
cana-6072	297	50	digital	digital	ADJ
cana-6072	297	51	age	age	NOUN
cana-6072	297	52	.	.	PUNCT
cana-6072	298	1	references	reference	NOUN
cana-6072	298	2	1	1	NUM
cana-6072	298	3	.	.	PUNCT
cana-6072	299	1	al	al	PROPN
cana-6072	299	2	-	-	PUNCT
cana-6072	299	3	azani	azani	PROPN
cana-6072	299	4	,	,	PUNCT
cana-6072	299	5	sadam	sadam	PROPN
cana-6072	299	6	,	,	PUNCT
cana-6072	299	7	and	and	CCONJ
cana-6072	299	8	el	el	PROPN
cana-6072	299	9	sayed	sayed	PROPN
cana-6072	299	10	m.	m.	PROPN
cana-6072	300	1	el	el	PROPN
cana-6072	300	2	-	-	PROPN
cana-6072	300	3	alfy	alfy	PROPN
cana-6072	300	4	.	.	PUNCT
cana-6072	300	5	2021	2021	NUM
cana-6072	300	6	.	.	PUNCT
cana-6072	301	1	“	"	PUNCT
cana-6072	301	2	early	early	ADJ
cana-6072	301	3	and	and	CCONJ
cana-6072	301	4	late	late	ADJ
cana-6072	301	5	fusion	fusion	NOUN
cana-6072	301	6	of	of	ADP
cana-6072	301	7	emojis	emoji	NOUN
cana-6072	301	8	and	and	CCONJ
cana-6072	301	9	text	text	NOUN
cana-6072	301	10	to	to	PART
cana-6072	301	11	enhance	enhance	VERB
cana-6072	301	12	opinion	opinion	NOUN
cana-6072	301	13	mining	mining	NOUN
cana-6072	301	14	.	.	PUNCT
cana-6072	301	15	”	"	PUNCT
cana-6072	302	1	ieee	ieee	NOUN
cana-6072	302	2	access	access	NOUN
cana-6072	302	3	9:121031–45	9:121031–45	PROPN
cana-6072	302	4	.	.	PUNCT
cana-6072	303	1	doi	doi	NOUN
cana-6072	303	2	:	:	PUNCT
cana-6072	303	3	10.1109	10.1109	NUM
cana-6072	303	4	/	/	SYM
cana-6072	303	5	access.2021.3108502	access.2021.3108502	NOUN
cana-6072	303	6	.	.	PUNCT
cana-6072	304	1	2	2	X
cana-6072	304	2	.	.	X
cana-6072	304	3	alshenqeeti	alshenqeeti	PROPN
cana-6072	304	4	,	,	PUNCT
cana-6072	304	5	hamza	hamza	PROPN
cana-6072	304	6	.	.	PUNCT
cana-6072	304	7	2016	2016	NUM
cana-6072	304	8	.	.	PUNCT
cana-6072	305	1	“	"	PUNCT
cana-6072	305	2	are	be	AUX
cana-6072	305	3	emojis	emoji	NOUN
cana-6072	305	4	creating	create	VERB
cana-6072	305	5	a	a	DET
cana-6072	305	6	new	new	ADJ
cana-6072	305	7	or	or	CCONJ
cana-6072	305	8	old	old	ADJ
cana-6072	305	9	visual	visual	ADJ
cana-6072	305	10	language	language	NOUN
cana-6072	305	11	for	for	ADP
cana-6072	305	12	new	new	ADJ
cana-6072	305	13	generations	generation	NOUN
cana-6072	305	14	?	?	PUNCT
cana-6072	306	1	a	a	DET
cana-6072	306	2	socio	socio	NOUN
cana-6072	306	3	-	-	PUNCT
cana-6072	306	4	semiotic	semiotic	ADJ
cana-6072	306	5	study	study	NOUN
cana-6072	306	6	.	.	PUNCT
cana-6072	306	7	”	"	PUNCT
cana-6072	306	8	.	.	PUNCT
cana-6072	307	1	advances	advance	NOUN
cana-6072	307	2	in	in	ADP
cana-6072	307	3	language	language	NOUN
cana-6072	307	4	and	and	CCONJ
cana-6072	307	5	literary	literary	ADJ
cana-6072	307	6	studies	study	NOUN
cana-6072	307	7	7(6	7(6	NUM
cana-6072	307	8	)	)	PUNCT
cana-6072	307	9	.	.	PUNCT
cana-6072	308	1	doi:10.7575	doi:10.7575	ADJ
cana-6072	308	2	/	/	SYM
cana-6072	308	3	aiac.alls.v.7n.6p.56	aiac.alls.v.7n.6p.56	DET
cana-6072	308	4	3	3	NUM
cana-6072	308	5	.	.	PUNCT
cana-6072	308	6	barbieri	barbieri	PROPN
cana-6072	308	7	,	,	PUNCT
cana-6072	308	8	francesco	francesco	PROPN
cana-6072	308	9	,	,	PUNCT
cana-6072	308	10	german	german	ADJ
cana-6072	308	11	kruszewski	kruszewski	NOUN
cana-6072	308	12	,	,	PUNCT
cana-6072	308	13	francesco	francesco	PROPN
cana-6072	308	14	ronzano	ronzano	PROPN
cana-6072	308	15	,	,	PUNCT
cana-6072	308	16	and	and	CCONJ
cana-6072	308	17	horacio	horacio	PROPN
cana-6072	308	18	saggion	saggion	NOUN
cana-6072	308	19	.	.	PUNCT
cana-6072	309	1	2016	2016	NUM
cana-6072	309	2	.	.	PUNCT
cana-6072	310	1	“	"	PUNCT
cana-6072	310	2	how	how	SCONJ
cana-6072	310	3	cosmopolitan	cosmopolitan	ADJ
cana-6072	310	4	are	be	AUX
cana-6072	310	5	emojis	emoji	NOUN
cana-6072	310	6	?	?	PUNCT
cana-6072	310	7	”	"	PUNCT
cana-6072	311	1	531–35	531–35	NUM
cana-6072	311	2	.	.	PUNCT
cana-6072	312	1	doi	doi	NOUN
cana-6072	312	2	:	:	PUNCT
cana-6072	312	3	10.1145/2964284.2967278	10.1145/2964284.2967278	NUM
cana-6072	312	4	.	.	NOUN
cana-6072	312	5	4	4	NUM
cana-6072	312	6	.	.	X
cana-6072	312	7	barbieri	barbieri	PROPN
cana-6072	312	8	,	,	PUNCT
cana-6072	312	9	francesco	francesco	PROPN
cana-6072	312	10	,	,	PUNCT
cana-6072	312	11	francesco	francesco	PROPN
cana-6072	312	12	ronzano	ronzano	PROPN
cana-6072	312	13	,	,	PUNCT
cana-6072	312	14	and	and	CCONJ
cana-6072	312	15	horacio	horacio	PROPN
cana-6072	312	16	saggion	saggion	NOUN
cana-6072	312	17	.	.	PUNCT
cana-6072	313	1	2016	2016	NUM
cana-6072	313	2	.	.	PUNCT
cana-6072	314	1	“	"	PUNCT
cana-6072	314	2	what	what	PRON
cana-6072	314	3	does	do	AUX
cana-6072	314	4	this	this	DET
cana-6072	314	5	emoji	emoji	NOUN
cana-6072	314	6	mean	mean	VERB
cana-6072	314	7	?	?	PUNCT
cana-6072	315	1	a	a	DET
cana-6072	315	2	vector	vector	NOUN
cana-6072	315	3	space	space	NOUN
cana-6072	315	4	skip	skip	NOUN
cana-6072	315	5	-	-	PUNCT
cana-6072	315	6	gram	gram	NOUN
cana-6072	315	7	model	model	NOUN
cana-6072	315	8	for	for	ADP
cana-6072	315	9	twitter	twitter	NOUN
cana-6072	315	10	emojis	emoji	NOUN
cana-6072	315	11	.	.	PUNCT
cana-6072	315	12	”	"	PUNCT
cana-6072	316	1	proceedings	proceeding	NOUN
cana-6072	316	2	of	of	ADP
cana-6072	316	3	the	the	DET
cana-6072	316	4	10th	10th	ADJ
cana-6072	316	5	international	international	ADJ
cana-6072	316	6	conference	conference	NOUN
cana-6072	316	7	on	on	ADP
cana-6072	316	8	language	language	NOUN
cana-6072	316	9	resources	resource	NOUN
cana-6072	316	10	and	and	CCONJ
cana-6072	316	11	evaluation	evaluation	NOUN
cana-6072	316	12	,	,	PUNCT
cana-6072	316	13	lrec	lrec	VERB
cana-6072	316	14	2016	2016	NUM
cana-6072	316	15	3967–72	3967–72	NUM
cana-6072	316	16	.	.	PUNCT
cana-6072	317	1	communications	communication	NOUN
cana-6072	317	2	on	on	ADP
cana-6072	317	3	applied	apply	VERB
cana-6072	317	4	nonlinear	nonlinear	ADJ
cana-6072	317	5	analysis	analysis	NOUN
cana-6072	317	6	issn	issn	NOUN
cana-6072	317	7	:	:	PUNCT
cana-6072	317	8	1074	1074	NUM
cana-6072	317	9	-	-	PUNCT
cana-6072	317	10	133x	133x	NUM
cana-6072	317	11	vol	vol	NOUN
cana-6072	317	12	32	32	NUM
cana-6072	317	13	no	no	NOUN
cana-6072	317	14	.	.	PUNCT
cana-6072	318	1	4s	4s	NUM
cana-6072	318	2	(	(	PUNCT
cana-6072	318	3	2025	2025	NUM
cana-6072	318	4	)	)	PUNCT
cana-6072	318	5	771	771	NUM
cana-6072	318	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	318	7	5	5	NUM
cana-6072	318	8	.	.	PUNCT
cana-6072	319	1	chen	chen	PROPN
cana-6072	319	2	,	,	PUNCT
cana-6072	319	3	y.	y.	PROPN
cana-6072	319	4	,	,	PUNCT
cana-6072	319	5	yuan	yuan	PROPN
cana-6072	319	6	,	,	PUNCT
cana-6072	319	7	j.	j.	PROPN
cana-6072	319	8	,	,	PUNCT
cana-6072	319	9	you	you	PRON
cana-6072	319	10	,	,	PUNCT
cana-6072	319	11	q.	q.	PROPN
cana-6072	319	12	,	,	PUNCT
cana-6072	319	13	&	&	CCONJ
cana-6072	319	14	luo	luo	PROPN
cana-6072	319	15	,	,	PUNCT
cana-6072	319	16	j.	j.	PROPN
cana-6072	319	17	(	(	PUNCT
cana-6072	319	18	2018	2018	NUM
cana-6072	319	19	,	,	PUNCT
cana-6072	319	20	october	october	PROPN
cana-6072	319	21	)	)	PUNCT
cana-6072	319	22	.	.	PUNCT
cana-6072	320	1	twitter	twitter	NOUN
cana-6072	320	2	sentiment	sentiment	NOUN
cana-6072	320	3	analysis	analysis	NOUN
cana-6072	320	4	via	via	ADP
cana-6072	320	5	bi	bi	ADJ
cana-6072	320	6	-	-	ADJ
cana-6072	320	7	sense	sense	ADJ
cana-6072	320	8	emoji	emoji	NOUN
cana-6072	320	9	embedding	embed	VERB
cana-6072	320	10	and	and	CCONJ
cana-6072	320	11	attention	attention	NOUN
cana-6072	320	12	-	-	PUNCT
cana-6072	320	13	based	base	VERB
cana-6072	320	14	lstm	lstm	NOUN
cana-6072	320	15	.	.	PUNCT
cana-6072	321	1	in	in	ADP
cana-6072	321	2	proceedings	proceeding	NOUN
cana-6072	321	3	of	of	ADP
cana-6072	321	4	the	the	DET
cana-6072	321	5	26th	26th	ADJ
cana-6072	321	6	acm	acm	PROPN
cana-6072	321	7	international	international	ADJ
cana-6072	321	8	conference	conference	NOUN
cana-6072	321	9	on	on	ADP
cana-6072	321	10	multimedia	multimedia	NOUN
cana-6072	321	11	(	(	PUNCT
cana-6072	321	12	pp	pp	ADJ
cana-6072	321	13	.	.	PUNCT
cana-6072	321	14	117	117	NUM
cana-6072	321	15	-	-	SYM
cana-6072	321	16	125	125	NUM
cana-6072	321	17	)	)	PUNCT
cana-6072	321	18	.	.	PUNCT
cana-6072	322	1	6	6	X
cana-6072	322	2	.	.	X
cana-6072	322	3	chen	chen	PROPN
cana-6072	322	4	,	,	PUNCT
cana-6072	322	5	z.	z.	PROPN
cana-6072	322	6	,	,	PUNCT
cana-6072	322	7	shen	shen	PROPN
cana-6072	322	8	,	,	PUNCT
cana-6072	322	9	s.	s.	PROPN
cana-6072	322	10	,	,	PUNCT
cana-6072	322	11	hu	hu	PROPN
cana-6072	322	12	,	,	PUNCT
cana-6072	322	13	z.	z.	PROPN
cana-6072	322	14	,	,	PUNCT
cana-6072	322	15	lu	lu	PROPN
cana-6072	322	16	,	,	PUNCT
cana-6072	322	17	x.	x.	PROPN
cana-6072	322	18	,	,	PUNCT
cana-6072	322	19	mei	mei	PROPN
cana-6072	322	20	,	,	PUNCT
cana-6072	322	21	q.	q.	PROPN
cana-6072	322	22	,	,	PUNCT
cana-6072	322	23	&	&	CCONJ
cana-6072	322	24	liu	liu	PROPN
cana-6072	322	25	,	,	PUNCT
cana-6072	322	26	x.	x.	NOUN
cana-6072	322	27	(	(	PUNCT
cana-6072	322	28	2019	2019	NUM
cana-6072	322	29	,	,	PUNCT
cana-6072	322	30	may	may	AUX
cana-6072	322	31	)	)	PUNCT
cana-6072	322	32	.	.	PUNCT
cana-6072	323	1	emoji	emoji	NOUN
cana-6072	323	2	-	-	PUNCT
cana-6072	323	3	powered	power	VERB
cana-6072	323	4	representation	representation	NOUN
cana-6072	323	5	learning	learn	VERB
cana-6072	323	6	for	for	ADP
cana-6072	323	7	cross	cross	ADJ
cana-6072	323	8	-	-	ADJ
cana-6072	323	9	lingual	lingual	ADJ
cana-6072	323	10	sentiment	sentiment	NOUN
cana-6072	323	11	classification	classification	NOUN
cana-6072	323	12	.	.	PUNCT
cana-6072	324	1	in	in	ADP
cana-6072	324	2	the	the	DET
cana-6072	324	3	world	world	NOUN
cana-6072	324	4	wide	wide	ADJ
cana-6072	324	5	web	web	NOUN
cana-6072	324	6	conference	conference	NOUN
cana-6072	324	7	(	(	PUNCT
cana-6072	324	8	pp	pp	ADJ
cana-6072	324	9	.	.	PUNCT
cana-6072	325	1	251	251	NUM
cana-6072	325	2	-	-	SYM
cana-6072	325	3	262	262	NUM
cana-6072	325	4	)	)	PUNCT
cana-6072	325	5	.	.	PUNCT
cana-6072	326	1	7	7	X
cana-6072	326	2	.	.	X
cana-6072	326	3	eisner	eisner	PROPN
cana-6072	326	4	,	,	PUNCT
cana-6072	326	5	b.	b.	PROPN
cana-6072	326	6	,	,	PUNCT
cana-6072	326	7	rocktäschel	rocktäschel	PROPN
cana-6072	326	8	,	,	PUNCT
cana-6072	326	9	t.	t.	PROPN
cana-6072	326	10	,	,	PUNCT
cana-6072	326	11	augenstein	augenstein	PROPN
cana-6072	326	12	,	,	PUNCT
cana-6072	326	13	i.	i.	NOUN
cana-6072	326	14	,	,	PUNCT
cana-6072	326	15	bošnjak	bošnjak	ADV
cana-6072	326	16	,	,	PUNCT
cana-6072	326	17	m.	m.	NOUN
cana-6072	326	18	,	,	PUNCT
cana-6072	326	19	&	&	CCONJ
cana-6072	326	20	riedel	riedel	PROPN
cana-6072	326	21	,	,	PUNCT
cana-6072	326	22	s.	s.	PROPN
cana-6072	326	23	(	(	PUNCT
cana-6072	326	24	2016	2016	NUM
cana-6072	326	25	)	)	PUNCT
cana-6072	326	26	.	.	PUNCT
cana-6072	327	1	emoji2vec	emoji2vec	NOUN
cana-6072	327	2	:	:	PUNCT
cana-6072	327	3	learning	learn	VERB
cana-6072	327	4	emoji	emoji	NOUN
cana-6072	327	5	representations	representation	NOUN
cana-6072	327	6	from	from	ADP
cana-6072	327	7	their	their	PRON
cana-6072	327	8	description	description	NOUN
cana-6072	327	9	.	.	PUNCT
cana-6072	328	1	arxiv	arxiv	PROPN
cana-6072	328	2	preprint	preprint	PROPN
cana-6072	328	3	arxiv:1609.08359	arxiv:1609.08359	NOUN
cana-6072	328	4	.	.	PUNCT
cana-6072	329	1	8	8	NUM
cana-6072	329	2	.	.	X
cana-6072	329	3	felbo	felbo	PROPN
cana-6072	329	4	,	,	PUNCT
cana-6072	329	5	b.	b.	PROPN
cana-6072	329	6	,	,	PUNCT
cana-6072	329	7	mislove	mislove	NOUN
cana-6072	329	8	,	,	PUNCT
cana-6072	329	9	a.	a.	NOUN
cana-6072	329	10	,	,	PUNCT
cana-6072	329	11	søgaard	søgaard	NOUN
cana-6072	329	12	,	,	PUNCT
cana-6072	329	13	a.	a.	NOUN
cana-6072	329	14	,	,	PUNCT
cana-6072	329	15	rahwan	rahwan	PROPN
cana-6072	329	16	,	,	PUNCT
cana-6072	329	17	i.	i.	PROPN
cana-6072	329	18	,	,	PUNCT
cana-6072	329	19	&	&	CCONJ
cana-6072	329	20	lehmann	lehmann	PROPN
cana-6072	329	21	,	,	PUNCT
cana-6072	329	22	s.	s.	PROPN
cana-6072	329	23	(	(	PUNCT
cana-6072	329	24	2017	2017	NUM
cana-6072	329	25	)	)	PUNCT
cana-6072	329	26	.	.	PUNCT
cana-6072	330	1	using	use	VERB
cana-6072	330	2	millions	million	NOUN
cana-6072	330	3	of	of	ADP
cana-6072	330	4	emoji	emoji	PROPN
cana-6072	330	5	occurrences	occurrence	NOUN
cana-6072	330	6	to	to	PART
cana-6072	330	7	learn	learn	VERB
cana-6072	330	8	any	any	DET
cana-6072	330	9	-	-	PUNCT
cana-6072	330	10	domain	domain	NOUN
cana-6072	330	11	representations	representation	NOUN
cana-6072	330	12	for	for	ADP
cana-6072	330	13	detecting	detect	VERB
cana-6072	330	14	sentiment	sentiment	NOUN
cana-6072	330	15	,	,	PUNCT
cana-6072	330	16	emotion	emotion	NOUN
cana-6072	330	17	and	and	CCONJ
cana-6072	330	18	sarcasm	sarcasm	NOUN
cana-6072	330	19	.	.	PUNCT
cana-6072	331	1	arxiv	arxiv	PROPN
cana-6072	331	2	preprint	preprint	VERB
cana-6072	331	3	arxiv:1708.00524	arxiv:1708.00524	ADV
cana-6072	331	4	.	.	PUNCT
cana-6072	332	1	9	9	X
cana-6072	332	2	.	.	X
cana-6072	332	3	gan	gan	PROPN
cana-6072	332	4	,	,	PUNCT
cana-6072	332	5	c.	c.	PROPN
cana-6072	332	6	,	,	PUNCT
cana-6072	332	7	fu	fu	PROPN
cana-6072	332	8	,	,	PUNCT
cana-6072	332	9	x.	x.	PROPN
cana-6072	332	10	,	,	PUNCT
cana-6072	332	11	feng	feng	PROPN
cana-6072	332	12	,	,	PUNCT
cana-6072	332	13	q.	q.	PROPN
cana-6072	332	14	,	,	PUNCT
cana-6072	332	15	zhu	zhu	PROPN
cana-6072	332	16	,	,	PUNCT
cana-6072	332	17	q.	q.	PROPN
cana-6072	332	18	,	,	PUNCT
cana-6072	332	19	cao	cao	PROPN
cana-6072	332	20	,	,	PUNCT
cana-6072	332	21	y.	y.	PROPN
cana-6072	332	22	,	,	PUNCT
cana-6072	332	23	&	&	CCONJ
cana-6072	332	24	zhu	zhu	PROPN
cana-6072	332	25	,	,	PUNCT
cana-6072	332	26	y.	y.	PROPN
cana-6072	332	27	(	(	PUNCT
cana-6072	332	28	2024	2024	NUM
cana-6072	332	29	)	)	PUNCT
cana-6072	332	30	.	.	PUNCT
cana-6072	333	1	a	a	DET
cana-6072	333	2	multimodal	multimodal	ADJ
cana-6072	333	3	fusion	fusion	NOUN
cana-6072	333	4	network	network	NOUN
cana-6072	333	5	with	with	ADP
cana-6072	333	6	attention	attention	NOUN
cana-6072	333	7	mechanisms	mechanism	NOUN
cana-6072	333	8	for	for	ADP
cana-6072	333	9	visual	visual	ADJ
cana-6072	333	10	–	–	PUNCT
cana-6072	333	11	textual	textual	ADJ
cana-6072	333	12	sentiment	sentiment	NOUN
cana-6072	333	13	analysis	analysis	NOUN
cana-6072	333	14	.	.	PUNCT
cana-6072	334	1	expert	expert	NOUN
cana-6072	334	2	systems	system	NOUN
cana-6072	334	3	with	with	ADP
cana-6072	334	4	applications	application	NOUN
cana-6072	334	5	,	,	PUNCT
cana-6072	334	6	242	242	NUM
cana-6072	334	7	,	,	PUNCT
cana-6072	334	8	122731	122731	NUM
cana-6072	334	9	.	.	PUNCT
cana-6072	335	1	10	10	NUM
cana-6072	335	2	.	.	X
cana-6072	336	1	huang	huang	PROPN
cana-6072	336	2	,	,	PUNCT
cana-6072	336	3	f.	f.	PROPN
cana-6072	336	4	,	,	PUNCT
cana-6072	336	5	wei	wei	PROPN
cana-6072	336	6	,	,	PUNCT
cana-6072	336	7	k.	k.	PROPN
cana-6072	336	8	,	,	PUNCT
cana-6072	336	9	weng	weng	PROPN
cana-6072	336	10	,	,	PUNCT
cana-6072	336	11	j.	j.	PROPN
cana-6072	336	12	,	,	PUNCT
cana-6072	336	13	&	&	CCONJ
cana-6072	336	14	li	li	PROPN
cana-6072	336	15	,	,	PUNCT
cana-6072	336	16	z.	z.	PROPN
cana-6072	336	17	(	(	PUNCT
cana-6072	336	18	2020	2020	NUM
cana-6072	336	19	)	)	PUNCT
cana-6072	336	20	.	.	PUNCT
cana-6072	337	1	attention	attention	NOUN
cana-6072	337	2	-	-	PUNCT
cana-6072	337	3	based	base	VERB
cana-6072	337	4	modality	modality	NOUN
cana-6072	337	5	-	-	PUNCT
cana-6072	337	6	gated	gate	VERB
cana-6072	337	7	networks	network	NOUN
cana-6072	337	8	for	for	ADP
cana-6072	337	9	image	image	NOUN
cana-6072	337	10	-	-	PUNCT
cana-6072	337	11	text	text	NOUN
cana-6072	337	12	sentiment	sentiment	NOUN
cana-6072	337	13	analysis	analysis	NOUN
cana-6072	337	14	.	.	PUNCT
cana-6072	338	1	acm	acm	PROPN
cana-6072	338	2	transactions	transaction	NOUN
cana-6072	338	3	on	on	ADP
cana-6072	338	4	multimedia	multimedia	NOUN
cana-6072	338	5	computing	computing	NOUN
cana-6072	338	6	,	,	PUNCT
cana-6072	338	7	communications	communication	NOUN
cana-6072	338	8	,	,	PUNCT
cana-6072	338	9	and	and	CCONJ
cana-6072	338	10	applications	application	NOUN
cana-6072	338	11	(	(	PUNCT
cana-6072	338	12	tomm	tomm	NOUN
cana-6072	338	13	)	)	PUNCT
cana-6072	338	14	,	,	PUNCT
cana-6072	338	15	16(3	16(3	NOUN
cana-6072	338	16	)	)	PUNCT
cana-6072	338	17	,	,	PUNCT
cana-6072	338	18	1	1	NUM
cana-6072	338	19	-	-	SYM
cana-6072	338	20	19	19	NUM
cana-6072	338	21	.	.	NOUN
cana-6072	338	22	11	11	NUM
cana-6072	338	23	.	.	PUNCT
cana-6072	339	1	kastrati	kastrati	NOUN
cana-6072	339	2	,	,	PUNCT
cana-6072	339	3	muhamet	muhamet	NOUN
cana-6072	339	4	,	,	PUNCT
cana-6072	339	5	marenglen	marenglen	PROPN
cana-6072	339	6	biba	biba	PROPN
cana-6072	339	7	,	,	PUNCT
cana-6072	339	8	ali	ali	PROPN
cana-6072	339	9	shariq	shariq	PROPN
cana-6072	339	10	imran	imran	PROPN
cana-6072	339	11	,	,	PUNCT
cana-6072	339	12	and	and	CCONJ
cana-6072	339	13	zenun	zenun	PROPN
cana-6072	339	14	kastrati	kastrati	NOUN
cana-6072	339	15	.	.	PUNCT
cana-6072	339	16	2022	2022	NUM
cana-6072	339	17	.	.	PUNCT
cana-6072	340	1	“	"	PUNCT
cana-6072	340	2	sentiment	sentiment	NOUN
cana-6072	340	3	polarity	polarity	NOUN
cana-6072	340	4	and	and	CCONJ
cana-6072	340	5	emotion	emotion	NOUN
cana-6072	340	6	detection	detection	NOUN
cana-6072	340	7	from	from	ADP
cana-6072	340	8	tweets	tweet	NOUN
cana-6072	340	9	using	use	VERB
cana-6072	340	10	distant	distant	ADJ
cana-6072	340	11	supervision	supervision	NOUN
cana-6072	340	12	and	and	CCONJ
cana-6072	340	13	deep	deep	ADJ
cana-6072	340	14	learning	learning	NOUN
cana-6072	340	15	models	model	NOUN
cana-6072	340	16	.	.	PUNCT
cana-6072	340	17	”	"	PUNCT
cana-6072	341	1	lecture	lecture	NOUN
cana-6072	341	2	notes	note	NOUN
cana-6072	341	3	in	in	ADP
cana-6072	341	4	computer	computer	NOUN
cana-6072	341	5	science	science	NOUN
cana-6072	341	6	(	(	PUNCT
cana-6072	341	7	including	include	VERB
cana-6072	341	8	subseries	subserie	NOUN
cana-6072	341	9	lecture	lecture	VERB
cana-6072	341	10	notes	note	NOUN
cana-6072	341	11	in	in	ADP
cana-6072	341	12	artificial	artificial	ADJ
cana-6072	341	13	intelligence	intelligence	NOUN
cana-6072	341	14	and	and	CCONJ
cana-6072	341	15	lecture	lecture	NOUN
cana-6072	341	16	notes	note	NOUN
cana-6072	341	17	in	in	ADP
cana-6072	341	18	bioinformatics	bioinformatics	NOUN
cana-6072	341	19	)	)	PUNCT
cana-6072	341	20	13515	13515	NUM
cana-6072	341	21	lnai:13–23	lnai:13–23	PROPN
cana-6072	341	22	.	.	PUNCT
cana-6072	342	1	doi	doi	NOUN
cana-6072	342	2	:	:	PUNCT
cana-6072	342	3	10.1007/978	10.1007/978	NUM
cana-6072	342	4	-	-	SYM
cana-6072	342	5	3	3	NUM
cana-6072	342	6	-	-	PUNCT
cana-6072	342	7	031	031	NUM
cana-6072	342	8	-	-	PUNCT
cana-6072	342	9	16564	16564	NUM
cana-6072	342	10	-	-	SYM
cana-6072	342	11	1_2	1_2	NUM
cana-6072	342	12	.	.	PUNCT
cana-6072	343	1	12	12	NUM
cana-6072	343	2	.	.	PUNCT
cana-6072	344	1	khan	khan	PROPN
cana-6072	344	2	,	,	PUNCT
cana-6072	344	3	a.	a.	PROPN
cana-6072	344	4	,	,	PUNCT
cana-6072	344	5	majumdar	majumdar	PROPN
cana-6072	344	6	,	,	PUNCT
cana-6072	344	7	d.	d.	PROPN
cana-6072	344	8	&	&	CCONJ
cana-6072	344	9	mondal	mondal	PROPN
cana-6072	344	10	,	,	PUNCT
cana-6072	344	11	b.	b.	PROPN
cana-6072	344	12	sentiment	sentiment	PROPN
cana-6072	344	13	analysis	analysis	NOUN
cana-6072	344	14	of	of	ADP
cana-6072	344	15	emoji	emoji	PROPN
cana-6072	344	16	fused	fuse	VERB
cana-6072	344	17	reviews	review	NOUN
cana-6072	344	18	using	use	VERB
cana-6072	344	19	machine	machine	NOUN
cana-6072	344	20	learning	learning	NOUN
cana-6072	344	21	and	and	CCONJ
cana-6072	344	22	bert	bert	PROPN
cana-6072	344	23	.	.	PUNCT
cana-6072	345	1	sci	sci	PROPN
cana-6072	345	2	rep	rep	PROPN
cana-6072	345	3	15	15	NUM
cana-6072	345	4	,	,	PUNCT
cana-6072	345	5	7538	7538	NUM
cana-6072	345	6	(	(	PUNCT
cana-6072	345	7	2025	2025	NUM
cana-6072	345	8	)	)	PUNCT
cana-6072	345	9	.	.	PUNCT
cana-6072	346	1	https://doi.org/10.1038/s41598025-92286-0	https://doi.org/10.1038/s41598025-92286-0	PROPN
cana-6072	346	2	13	13	NUM
cana-6072	346	3	.	.	PUNCT
cana-6072	347	1	kralj	kralj	PROPN
cana-6072	347	2	novak	novak	PROPN
cana-6072	347	3	p	p	PROPN
cana-6072	347	4	,	,	PUNCT
cana-6072	347	5	smailović	smailović	NOUN
cana-6072	347	6	j	j	PROPN
cana-6072	347	7	,	,	PUNCT
cana-6072	347	8	sluban	sluban	PROPN
cana-6072	347	9	b	b	PROPN
cana-6072	347	10	,	,	PUNCT
cana-6072	347	11	mozetič	mozetič	PROPN
cana-6072	347	12	i.	i.	PROPN
cana-6072	347	13	sentiment	sentiment	PROPN
cana-6072	347	14	of	of	ADP
cana-6072	347	15	emojis	emoji	NOUN
cana-6072	347	16	.	.	PUNCT
cana-6072	348	1	plos	plos	PROPN
cana-6072	348	2	one	one	NUM
cana-6072	348	3	.	.	PUNCT
cana-6072	349	1	2015	2015	NUM
cana-6072	349	2	dec	dec	PROPN
cana-6072	349	3	7;10(12):e0144296	7;10(12):e0144296	PROPN
cana-6072	349	4	.	.	PUNCT
cana-6072	349	5	doi	doi	NOUN
cana-6072	349	6	:	:	PUNCT
cana-6072	349	7	10.1371	10.1371	NUM
cana-6072	349	8	/	/	SYM
cana-6072	349	9	journal.pone.0144296	journal.pone.0144296	PROPN
cana-6072	349	10	.	.	PUNCT
cana-6072	350	1	pmid	pmid	NOUN
cana-6072	350	2	:	:	PUNCT
cana-6072	350	3	26641093	26641093	NUM
cana-6072	350	4	;	;	PUNCT
cana-6072	350	5	pmcid	pmcid	PROPN
cana-6072	350	6	:	:	PUNCT
cana-6072	350	7	pmc4671607	pmc4671607	ADJ
cana-6072	350	8	.	.	PUNCT
cana-6072	351	1	14	14	NUM
cana-6072	351	2	.	.	PUNCT
cana-6072	352	1	liu	liu	PROPN
cana-6072	352	2	,	,	PUNCT
cana-6072	352	3	c.	c.	PROPN
cana-6072	352	4	,	,	PUNCT
cana-6072	352	5	fang	fang	PROPN
cana-6072	352	6	,	,	PUNCT
cana-6072	352	7	f.	f.	PROPN
cana-6072	352	8	,	,	PUNCT
cana-6072	352	9	lin	lin	PROPN
cana-6072	352	10	,	,	PUNCT
cana-6072	352	11	x.	x.	PROPN
cana-6072	352	12	,	,	PUNCT
cana-6072	352	13	cai	cai	PROPN
cana-6072	352	14	,	,	PUNCT
cana-6072	352	15	t.	t.	PROPN
cana-6072	352	16	,	,	PUNCT
cana-6072	352	17	tan	tan	PROPN
cana-6072	352	18	,	,	PUNCT
cana-6072	352	19	x.	x.	PROPN
cana-6072	352	20	,	,	PUNCT
cana-6072	352	21	liu	liu	PROPN
cana-6072	352	22	,	,	PUNCT
cana-6072	352	23	j.	j.	PROPN
cana-6072	352	24	,	,	PUNCT
cana-6072	352	25	&	&	CCONJ
cana-6072	352	26	lu	lu	PROPN
cana-6072	352	27	,	,	PUNCT
cana-6072	352	28	x.	x.	NOUN
cana-6072	352	29	(	(	PUNCT
cana-6072	352	30	2021	2021	NUM
cana-6072	352	31	)	)	PUNCT
cana-6072	352	32	.	.	PUNCT
cana-6072	353	1	improving	improve	VERB
cana-6072	353	2	sentiment	sentiment	NOUN
cana-6072	353	3	analysis	analysis	NOUN
cana-6072	353	4	accuracy	accuracy	NOUN
cana-6072	353	5	with	with	ADP
cana-6072	353	6	emoji	emoji	PROPN
cana-6072	353	7	embedding	embed	VERB
cana-6072	353	8	.	.	PUNCT
cana-6072	354	1	journal	journal	NOUN
cana-6072	354	2	of	of	ADP
cana-6072	354	3	safety	safety	NOUN
cana-6072	354	4	science	science	NOUN
cana-6072	354	5	and	and	CCONJ
cana-6072	354	6	resilience	resilience	NOUN
cana-6072	354	7	,	,	PUNCT
cana-6072	354	8	2(4	2(4	NUM
cana-6072	354	9	)	)	PUNCT
cana-6072	354	10	,	,	PUNCT
cana-6072	354	11	246	246	NUM
cana-6072	354	12	-	-	SYM
cana-6072	354	13	252	252	NUM
cana-6072	354	14	.	.	PUNCT
cana-6072	354	15	15	15	NUM
cana-6072	354	16	.	.	PUNCT
cana-6072	355	1	lou	lou	PROPN
cana-6072	355	2	,	,	PUNCT
cana-6072	355	3	y.	y.	PROPN
cana-6072	355	4	,	,	PUNCT
cana-6072	355	5	zhang	zhang	PROPN
cana-6072	355	6	,	,	PUNCT
cana-6072	355	7	y.	y.	PROPN
cana-6072	355	8	,	,	PUNCT
cana-6072	355	9	li	li	PROPN
cana-6072	355	10	,	,	PUNCT
cana-6072	355	11	f.	f.	PROPN
cana-6072	355	12	,	,	PUNCT
cana-6072	355	13	qian	qian	PROPN
cana-6072	355	14	,	,	PUNCT
cana-6072	355	15	t.	t.	PROPN
cana-6072	355	16	,	,	PUNCT
cana-6072	355	17	&	&	CCONJ
cana-6072	355	18	ji	ji	PROPN
cana-6072	355	19	,	,	PUNCT
cana-6072	355	20	d.	d.	PROPN
cana-6072	355	21	(	(	PUNCT
cana-6072	355	22	2020	2020	NUM
cana-6072	355	23	)	)	PUNCT
cana-6072	355	24	.	.	PUNCT
cana-6072	356	1	emoji	emoji	NOUN
cana-6072	356	2	-	-	PUNCT
cana-6072	356	3	based	base	VERB
cana-6072	356	4	sentiment	sentiment	NOUN
cana-6072	356	5	analysis	analysis	NOUN
cana-6072	356	6	using	use	VERB
cana-6072	356	7	attention	attention	NOUN
cana-6072	356	8	networks	network	NOUN
cana-6072	356	9	.	.	PUNCT
cana-6072	357	1	acm	acm	NOUN
cana-6072	357	2	transactions	transaction	NOUN
cana-6072	357	3	on	on	ADP
cana-6072	357	4	asian	asian	ADJ
cana-6072	357	5	and	and	CCONJ
cana-6072	357	6	low	low	ADJ
cana-6072	357	7	-	-	PUNCT
cana-6072	357	8	resource	resource	NOUN
cana-6072	357	9	language	language	NOUN
cana-6072	357	10	information	information	NOUN
cana-6072	357	11	processing	processing	NOUN
cana-6072	357	12	(	(	PUNCT
cana-6072	357	13	tallip	tallip	NOUN
cana-6072	357	14	)	)	PUNCT
cana-6072	357	15	,	,	PUNCT
cana-6072	357	16	19(5	19(5	NUM
cana-6072	357	17	)	)	PUNCT
cana-6072	357	18	,	,	PUNCT
cana-6072	357	19	1	1	NUM
cana-6072	357	20	-	-	SYM
cana-6072	357	21	13	13	NUM
cana-6072	357	22	.	.	PUNCT
cana-6072	358	1	16	16	NUM
cana-6072	358	2	.	.	PUNCT
cana-6072	359	1	lou	lou	PROPN
cana-6072	359	2	,	,	PUNCT
cana-6072	359	3	y.	y.	PROPN
cana-6072	359	4	,	,	PUNCT
cana-6072	359	5	zhou	zhou	PROPN
cana-6072	359	6	,	,	PUNCT
cana-6072	359	7	j.	j.	PROPN
cana-6072	359	8	,	,	PUNCT
cana-6072	359	9	zhou	zhou	PROPN
cana-6072	359	10	,	,	PUNCT
cana-6072	360	1	j.	j.	PROPN
cana-6072	360	2	et	et	PROPN
cana-6072	360	3	al	al	PROPN
cana-6072	360	4	.	.	PUNCT
cana-6072	360	5	emoji	emoji	PROPN
cana-6072	360	6	multimodal	multimodal	ADJ
cana-6072	360	7	microblog	microblog	NOUN
cana-6072	360	8	sentiment	sentiment	NOUN
cana-6072	360	9	analysis	analysis	NOUN
cana-6072	360	10	based	base	VERB
cana-6072	360	11	on	on	ADP
cana-6072	360	12	mutual	mutual	ADJ
cana-6072	360	13	attention	attention	NOUN
cana-6072	360	14	mechanism	mechanism	NOUN
cana-6072	360	15	.	.	PUNCT
cana-6072	361	1	sci	sci	PROPN
cana-6072	361	2	rep	rep	PROPN
cana-6072	361	3	14	14	NUM
cana-6072	361	4	,	,	PUNCT
cana-6072	361	5	29314	29314	NUM
cana-6072	361	6	(	(	PUNCT
cana-6072	361	7	2024	2024	NUM
cana-6072	361	8	)	)	PUNCT
cana-6072	361	9	.	.	PUNCT
cana-6072	362	1	https://doi.org/10.1038/s41598-024-80167-x	https://doi.org/10.1038/s41598-024-80167-x	PROPN
cana-6072	362	2	17	17	NUM
cana-6072	362	3	.	.	PUNCT
cana-6072	363	1	miller	miller	PROPN
cana-6072	363	2	,	,	PUNCT
cana-6072	363	3	h.	h.	PROPN
cana-6072	363	4	,	,	PUNCT
cana-6072	363	5	thebault	thebault	NOUN
cana-6072	363	6	-	-	PUNCT
cana-6072	363	7	spieker	spieker	PROPN
cana-6072	363	8	,	,	PUNCT
cana-6072	363	9	j.	j.	PROPN
cana-6072	363	10	,	,	PUNCT
cana-6072	363	11	chang	chang	PROPN
cana-6072	363	12	,	,	PUNCT
cana-6072	363	13	s.	s.	PROPN
cana-6072	363	14	,	,	PUNCT
cana-6072	363	15	johnson	johnson	PROPN
cana-6072	363	16	,	,	PUNCT
cana-6072	363	17	i.	i.	PROPN
cana-6072	363	18	,	,	PUNCT
cana-6072	363	19	terveen	terveen	PROPN
cana-6072	363	20	,	,	PUNCT
cana-6072	363	21	l.	l.	PROPN
cana-6072	363	22	,	,	PUNCT
cana-6072	363	23	&	&	CCONJ
cana-6072	363	24	hecht	hecht	PROPN
cana-6072	363	25	,	,	PUNCT
cana-6072	363	26	b.	b.	PROPN
cana-6072	363	27	(	(	PUNCT
cana-6072	363	28	2016	2016	NUM
cana-6072	363	29	)	)	PUNCT
cana-6072	363	30	.	.	PUNCT
cana-6072	364	1	“	"	PUNCT
cana-6072	364	2	blissfully	blissfully	ADV
cana-6072	364	3	happy	happy	ADJ
cana-6072	364	4	”	"	PUNCT
cana-6072	364	5	or	or	CCONJ
cana-6072	364	6	“	"	PUNCT
cana-6072	364	7	ready	ready	ADJ
cana-6072	364	8	tofight	tofight	NOUN
cana-6072	364	9	”	"	PUNCT
cana-6072	364	10	:	:	PUNCT
cana-6072	364	11	varying	vary	VERB
cana-6072	364	12	interpretations	interpretation	NOUN
cana-6072	364	13	of	of	ADP
cana-6072	364	14	emoji	emoji	PROPN
cana-6072	364	15	.	.	PUNCT
cana-6072	365	1	in	in	ADP
cana-6072	365	2	proceedings	proceeding	NOUN
cana-6072	365	3	of	of	ADP
cana-6072	365	4	the	the	DET
cana-6072	365	5	international	international	ADJ
cana-6072	365	6	aaai	aaai	PROPN
cana-6072	365	7	conference	conference	PROPN
cana-6072	365	8	on	on	ADP
cana-6072	365	9	web	web	NOUN
cana-6072	365	10	and	and	CCONJ
cana-6072	365	11	social	social	ADJ
cana-6072	365	12	media	medium	NOUN
cana-6072	365	13	(	(	PUNCT
cana-6072	365	14	vol	vol	NOUN
cana-6072	365	15	.	.	PROPN
cana-6072	365	16	10	10	NUM
cana-6072	365	17	,	,	PUNCT
cana-6072	365	18	no	no	INTJ
cana-6072	365	19	.	.	NOUN
cana-6072	365	20	1	1	NUM
cana-6072	365	21	,	,	PUNCT
cana-6072	365	22	pp	pp	ADJ
cana-6072	365	23	.	.	PUNCT
cana-6072	365	24	259268	259268	NUM
cana-6072	365	25	)	)	PUNCT
cana-6072	365	26	.	.	PUNCT
cana-6072	366	1	18	18	NUM
cana-6072	366	2	.	.	X
cana-6072	366	3	pang	pang	NOUN
cana-6072	366	4	,	,	PUNCT
cana-6072	366	5	b.	b.	PROPN
cana-6072	366	6	,	,	PUNCT
cana-6072	366	7	lillian	lillian	PROPN
cana-6072	366	8	lee	lee	PROPN
cana-6072	366	9	,	,	PUNCT
cana-6072	366	10	and	and	CCONJ
cana-6072	366	11	shivakumar	shivakumar	PROPN
cana-6072	366	12	vaithyanathan	vaithyanathan	PROPN
cana-6072	366	13	.	.	PUNCT
cana-6072	366	14	2002	2002	NUM
cana-6072	366	15	.	.	PUNCT
cana-6072	367	1	“	"	PUNCT
cana-6072	367	2	thumbs	thumb	VERB
cana-6072	367	3	up	up	ADP
cana-6072	367	4	?	?	PUNCT
cana-6072	368	1	sentiment	sentiment	NOUN
cana-6072	368	2	classification	classification	NOUN
cana-6072	368	3	using	use	VERB
cana-6072	368	4	machine	machine	NOUN
cana-6072	368	5	learning	learn	VERB
cana-6072	368	6	techniques	technique	NOUN
cana-6072	368	7	.	.	PUNCT
cana-6072	368	8	”	"	PUNCT
cana-6072	369	1	arxiv	arxiv	PROPN
cana-6072	369	2	cs.cl/0205:null	cs.cl/0205:null	PROPN
cana-6072	369	3	.	.	PUNCT
cana-6072	369	4	doi	doi	PROPN
cana-6072	369	5	:	:	PUNCT
cana-6072	369	6	10.3115/1118693.1118704	10.3115/1118693.1118704	NUM
cana-6072	369	7	.	.	X
cana-6072	370	1	https://doi.org/10.1038/s41598-025-92286-0	https://doi.org/10.1038/s41598-025-92286-0	NUM
cana-6072	370	2	https://doi.org/10.1038/s41598-025-92286-0	https://doi.org/10.1038/s41598-025-92286-0	NUM
cana-6072	370	3	https://doi.org/10.1038/s41598-024-80167-x	https://doi.org/10.1038/s41598-024-80167-x	PROPN
cana-6072	370	4	communications	communication	NOUN
cana-6072	370	5	on	on	ADP
cana-6072	370	6	applied	apply	VERB
cana-6072	370	7	nonlinear	nonlinear	ADJ
cana-6072	370	8	analysis	analysis	NOUN
cana-6072	370	9	issn	issn	NOUN
cana-6072	370	10	:	:	PUNCT
cana-6072	370	11	1074	1074	NUM
cana-6072	370	12	-	-	PUNCT
cana-6072	370	13	133x	133x	NUM
cana-6072	370	14	vol	vol	NOUN
cana-6072	370	15	32	32	NUM
cana-6072	370	16	no	no	NOUN
cana-6072	370	17	.	.	PUNCT
cana-6072	371	1	4s	4s	NUM
cana-6072	371	2	(	(	PUNCT
cana-6072	371	3	2025	2025	NUM
cana-6072	371	4	)	)	PUNCT
cana-6072	371	5	772	772	NUM
cana-6072	371	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-6072	371	7	19	19	NUM
cana-6072	371	8	.	.	PUNCT
cana-6072	372	1	peng	peng	PROPN
cana-6072	372	2	,	,	PUNCT
cana-6072	372	3	dunlu	dunlu	PROPN
cana-6072	372	4	,	,	PUNCT
cana-6072	372	5	and	and	CCONJ
cana-6072	372	6	huimin	huimin	PROPN
cana-6072	372	7	zhao	zhao	PROPN
cana-6072	372	8	.	.	PUNCT
cana-6072	372	9	2021	2021	NUM
cana-6072	372	10	.	.	PUNCT
cana-6072	373	1	“	"	PUNCT
cana-6072	373	2	seq2emoji	seq2emoji	NOUN
cana-6072	373	3	:	:	PUNCT
cana-6072	373	4	a	a	DET
cana-6072	373	5	hybrid	hybrid	ADJ
cana-6072	373	6	sequence	sequence	NOUN
cana-6072	373	7	generation	generation	NOUN
cana-6072	373	8	model	model	NOUN
cana-6072	373	9	for	for	ADP
cana-6072	373	10	short	short	ADJ
cana-6072	373	11	text	text	NOUN
cana-6072	373	12	emoji	emoji	PROPN
cana-6072	373	13	prediction	prediction	NOUN
cana-6072	373	14	.	.	PUNCT
cana-6072	373	15	”	"	PUNCT
cana-6072	374	1	knowledge	knowledge	NOUN
cana-6072	374	2	-	-	PUNCT
cana-6072	374	3	based	base	VERB
cana-6072	374	4	systems	system	NOUN
cana-6072	374	5	214:106727	214:106727	PRON
cana-6072	374	6	.	.	PUNCT
cana-6072	375	1	doi	doi	NOUN
cana-6072	375	2	:	:	PUNCT
cana-6072	375	3	10.1016	10.1016	NUM
cana-6072	375	4	/	/	SYM
cana-6072	375	5	j.knosys.2020.106727	j.knosys.2020.106727	NOUN
cana-6072	375	6	.	.	PUNCT
cana-6072	376	1	20	20	NUM
cana-6072	376	2	.	.	X
cana-6072	376	3	pratibha	pratibha	NOUN
cana-6072	376	4	,	,	PUNCT
cana-6072	376	5	amandeep	amandeep	NOUN
cana-6072	376	6	kaur	kaur	PROPN
cana-6072	376	7	,	,	PUNCT
cana-6072	376	8	and	and	CCONJ
cana-6072	376	9	meenu	meenu	PROPN
cana-6072	376	10	khurana	khurana	PROPN
cana-6072	376	11	.	.	PUNCT
cana-6072	377	1	2024	2024	NUM
cana-6072	377	2	.	.	PUNCT
cana-6072	378	1	“	"	PUNCT
cana-6072	378	2	multimodal	multimodal	ADJ
cana-6072	378	3	sentiments	sentiment	NOUN
cana-6072	378	4	:	:	PUNCT
cana-6072	378	5	unraveling	unravel	VERB
cana-6072	378	6	text	text	NOUN
cana-6072	378	7	and	and	CCONJ
cana-6072	378	8	emoji	emoji	PROPN
cana-6072	378	9	dynamics	dynamic	NOUN
cana-6072	378	10	through	through	ADP
cana-6072	378	11	deep	deep	ADJ
cana-6072	378	12	learning	learning	NOUN
cana-6072	378	13	.	.	PUNCT
cana-6072	378	14	”	"	PUNCT
cana-6072	379	1	2024	2024	NUM
cana-6072	379	2	11th	11th	ADJ
cana-6072	379	3	international	international	ADJ
cana-6072	379	4	conference	conference	NOUN
cana-6072	379	5	on	on	ADP
cana-6072	379	6	reliability	reliability	NOUN
cana-6072	379	7	,	,	PUNCT
cana-6072	379	8	infocom	infocom	NOUN
cana-6072	379	9	technologies	technology	NOUN
cana-6072	379	10	and	and	CCONJ
cana-6072	379	11	optimization	optimization	NOUN
cana-6072	379	12	(	(	PUNCT
cana-6072	379	13	trends	trend	NOUN
cana-6072	379	14	and	and	CCONJ
cana-6072	379	15	future	future	ADJ
cana-6072	379	16	directions	direction	NOUN
cana-6072	379	17	)	)	PUNCT
cana-6072	379	18	,	,	PUNCT
cana-6072	379	19	icrito	icrito	NOUN
cana-6072	379	20	2024	2024	NUM
cana-6072	379	21	1–6	1–6	NUM
cana-6072	379	22	.	.	PUNCT
cana-6072	380	1	doi	doi	NOUN
cana-6072	380	2	:	:	PUNCT
cana-6072	380	3	10.1109	10.1109	NUM
cana-6072	380	4	/	/	SYM
cana-6072	380	5	icrito61523.2024.10522265	icrito61523.2024.10522265	NOUN
cana-6072	380	6	.	.	NOUN
cana-6072	380	7	21	21	NUM
cana-6072	380	8	.	.	PUNCT
cana-6072	381	1	stamatov	stamatov	NOUN
cana-6072	381	2	,	,	PUNCT
cana-6072	381	3	elian	elian	PROPN
cana-6072	381	4	.	.	PROPN
cana-6072	381	5	2017	2017	NUM
cana-6072	381	6	.	.	PUNCT
cana-6072	382	1	“	"	PUNCT
cana-6072	382	2	do	do	AUX
cana-6072	382	3	emoji	emoji	PROPN
cana-6072	382	4	use	use	VERB
cana-6072	382	5	a	a	DET
cana-6072	382	6	grammar	grammar	NOUN
cana-6072	382	7	 	 	SPACE
cana-6072	382	8	?	?	PUNCT
cana-6072	383	1	emergent	emergent	ADJ
cana-6072	383	2	structure	structure	NOUN
cana-6072	383	3	in	in	ADP
cana-6072	383	4	non	non	ADJ
cana-6072	383	5	-	-	ADJ
cana-6072	383	6	verbal	verbal	ADJ
cana-6072	383	7	digital	digital	ADJ
cana-6072	383	8	communication	communication	NOUN
cana-6072	383	9	.	.	PUNCT
cana-6072	383	10	”	"	PUNCT
cana-6072	384	1	(	(	PUNCT
cana-6072	384	2	february):1–54	february):1–54	PROPN
cana-6072	384	3	.	.	NOUN
cana-6072	384	4	22	22	NUM
cana-6072	384	5	.	.	X
cana-6072	384	6	tang	tang	PROPN
cana-6072	384	7	,	,	PUNCT
cana-6072	384	8	h.	h.	PROPN
cana-6072	384	9	,	,	PUNCT
cana-6072	384	10	tang	tang	PROPN
cana-6072	384	11	,	,	PUNCT
cana-6072	384	12	w.	w.	PROPN
cana-6072	384	13	,	,	PUNCT
cana-6072	384	14	zhu	zhu	PROPN
cana-6072	384	15	,	,	PUNCT
cana-6072	384	16	d.	d.	PROPN
cana-6072	384	17	,	,	PUNCT
cana-6072	384	18	wang	wang	PROPN
cana-6072	384	19	,	,	PUNCT
cana-6072	384	20	s.	s.	PROPN
cana-6072	384	21	,	,	PUNCT
cana-6072	384	22	wang	wang	PROPN
cana-6072	384	23	,	,	PUNCT
cana-6072	384	24	y.	y.	PROPN
cana-6072	384	25	,	,	PUNCT
cana-6072	384	26	&	&	CCONJ
cana-6072	384	27	wang	wang	PROPN
cana-6072	384	28	,	,	PUNCT
cana-6072	384	29	l.	l.	PROPN
cana-6072	384	30	(	(	PUNCT
cana-6072	384	31	2024	2024	NUM
cana-6072	384	32	)	)	PUNCT
cana-6072	384	33	.	.	PUNCT
cana-6072	385	1	emfsa	emfsa	PROPN
cana-6072	385	2	:	:	PUNCT
cana-6072	385	3	emojibased	emojibase	VERB
cana-6072	385	4	multifeature	multifeature	NOUN
cana-6072	385	5	fusion	fusion	NOUN
cana-6072	385	6	sentiment	sentiment	NOUN
cana-6072	385	7	analysis	analysis	NOUN
cana-6072	385	8	.	.	PUNCT
cana-6072	386	1	plos	plos	PROPN
cana-6072	386	2	one	one	NUM
cana-6072	386	3	,	,	PUNCT
cana-6072	386	4	19(9	19(9	NUM
cana-6072	386	5	)	)	PUNCT
cana-6072	386	6	,	,	PUNCT
cana-6072	386	7	e0310715	e0310715	NOUN
cana-6072	386	8	.	.	PUNCT
cana-6072	386	9	23	23	NUM
cana-6072	386	10	.	.	PUNCT
cana-6072	387	1	wahyuni	wahyuni	PROPN
cana-6072	387	2	,	,	PUNCT
cana-6072	387	3	rinda	rinda	NOUN
cana-6072	387	4	,	,	PUNCT
cana-6072	387	5	and	and	CCONJ
cana-6072	387	6	indra	indra	PROPN
cana-6072	387	7	budi	budi	PROPN
cana-6072	387	8	.	.	PROPN
cana-6072	387	9	2018	2018	NUM
cana-6072	387	10	.	.	PUNCT
cana-6072	388	1	“	"	PUNCT
cana-6072	388	2	combining	combine	VERB
cana-6072	388	3	linguistic	linguistic	ADJ
cana-6072	388	4	,	,	PUNCT
cana-6072	388	5	semantic	semantic	ADJ
cana-6072	388	6	and	and	CCONJ
cana-6072	388	7	lexicon	lexicon	ADJ
cana-6072	388	8	feature	feature	NOUN
cana-6072	388	9	for	for	ADP
cana-6072	388	10	emoji	emoji	PROPN
cana-6072	388	11	classification	classification	NOUN
cana-6072	388	12	in	in	ADP
cana-6072	388	13	twitter	twitter	NOUN
cana-6072	388	14	10.1016	10.1016	NUM
cana-6072	388	15	/	/	SYM
cana-6072	388	16	j.procs.2018.08.166	j.procs.2018.08.166	PROPN
cana-6072	388	17	.	.	PUNCT
cana-6072	389	1	24	24	NUM
cana-6072	389	2	.	.	PUNCT
cana-6072	390	1	waltman	waltman	PROPN
cana-6072	390	2	,	,	PUNCT
cana-6072	390	3	ludo	ludo	PROPN
cana-6072	390	4	,	,	PUNCT
cana-6072	390	5	clara	clara	PROPN
cana-6072	390	6	calero	calero	PROPN
cana-6072	390	7	-	-	PUNCT
cana-6072	390	8	medina	medina	PROPN
cana-6072	390	9	,	,	PUNCT
cana-6072	390	10	joost	joost	PROPN
cana-6072	390	11	kosten	kosten	PROPN
cana-6072	390	12	,	,	PUNCT
cana-6072	390	13	ed	ed	PROPN
cana-6072	390	14	c.	c.	PROPN
cana-6072	390	15	m.	m.	PROPN
cana-6072	390	16	noyons	noyons	PROPN
cana-6072	390	17	,	,	PUNCT
cana-6072	390	18	robert	robert	PROPN
cana-6072	390	19	j.	j.	PROPN
cana-6072	390	20	w.	w.	PROPN
cana-6072	390	21	tijssen	tijssen	PROPN
cana-6072	390	22	,	,	PUNCT
cana-6072	390	23	nees	nees	PROPN
cana-6072	390	24	jan	jan	PROPN
cana-6072	390	25	van	van	PROPN
cana-6072	390	26	eck	eck	PROPN
cana-6072	390	27	,	,	PUNCT
cana-6072	390	28	thed	the	VERB
cana-6072	390	29	n.	n.	PROPN
cana-6072	390	30	van	van	PROPN
cana-6072	390	31	leeuwen	leeuwen	PROPN
cana-6072	390	32	,	,	PUNCT
cana-6072	390	33	anthony	anthony	PROPN
cana-6072	390	34	f.	f.	PROPN
cana-6072	390	35	j.	j.	PROPN
cana-6072	390	36	van	van	PROPN
cana-6072	390	37	raan	raan	PROPN
cana-6072	390	38	,	,	PUNCT
cana-6072	390	39	martijn	martijn	PROPN
cana-6072	390	40	s.	s.	PROPN
cana-6072	390	41	visser	visser	PROPN
cana-6072	390	42	,	,	PUNCT
cana-6072	390	43	and	and	CCONJ
cana-6072	390	44	paul	paul	PROPN
cana-6072	390	45	wouters	wouter	NOUN
cana-6072	390	46	.	.	PUNCT
cana-6072	390	47	2012	2012	NUM
cana-6072	390	48	.	.	PUNCT
cana-6072	391	1	“	"	PUNCT
cana-6072	391	2	the	the	DET
cana-6072	391	3	leiden	leiden	PROPN
cana-6072	391	4	ranking	ranking	ADJ
cana-6072	391	5	2011/2012	2011/2012	NUM
cana-6072	391	6	:	:	PUNCT
cana-6072	391	7	data	data	NOUN
cana-6072	391	8	collection	collection	NOUN
cana-6072	391	9	,	,	PUNCT
cana-6072	391	10	indicators	indicator	NOUN
cana-6072	391	11	,	,	PUNCT
cana-6072	391	12	and	and	CCONJ
cana-6072	391	13	interpretation	interpretation	NOUN
cana-6072	391	14	.	.	PUNCT
cana-6072	391	15	”	"	PUNCT
cana-6072	391	16	journal	journal	NOUN
cana-6072	391	17	of	of	ADP
cana-6072	391	18	the	the	DET
cana-6072	391	19	american	american	ADJ
cana-6072	391	20	society	society	NOUN
cana-6072	391	21	for	for	ADP
cana-6072	391	22	information	information	NOUN
cana-6072	391	23	science	science	NOUN
cana-6072	391	24	and	and	CCONJ
cana-6072	391	25	technology	technology	NOUN
cana-6072	391	26	63(12):2419–32	63(12):2419–32	NOUN
cana-6072	391	27	.	.	PUNCT
cana-6072	392	1	doi	doi	NOUN
cana-6072	392	2	:	:	PUNCT
cana-6072	392	3	10.1002	10.1002	NUM
cana-6072	392	4	/	/	SYM
cana-6072	392	5	asi.22708	asi.22708	NOUN
cana-6072	392	6	.	.	PROPN
cana-6072	392	7	25	25	NUM
cana-6072	392	8	.	.	PUNCT
cana-6072	393	1	xu	xu	PROPN
cana-6072	393	2	,	,	PUNCT
cana-6072	393	3	q.	q.	PROPN
cana-6072	393	4	a.	a.	PROPN
cana-6072	393	5	,	,	PUNCT
cana-6072	393	6	jayne	jayne	PROPN
cana-6072	393	7	,	,	PUNCT
cana-6072	393	8	c.	c.	PROPN
cana-6072	393	9	,	,	PUNCT
cana-6072	393	10	&	&	CCONJ
cana-6072	393	11	chang	chang	PROPN
cana-6072	393	12	,	,	PUNCT
cana-6072	393	13	v.	v.	PROPN
cana-6072	393	14	(	(	PUNCT
cana-6072	393	15	2024	2024	NUM
cana-6072	393	16	)	)	PUNCT
cana-6072	393	17	.	.	PUNCT
cana-6072	394	1	an	an	DET
cana-6072	394	2	emoji	emoji	PROPN
cana-6072	394	3	feature	feature	NOUN
cana-6072	394	4	-	-	PUNCT
cana-6072	394	5	incorporated	incorporate	VERB
cana-6072	394	6	multi	multi	ADJ
cana-6072	394	7	-	-	ADJ
cana-6072	394	8	view	view	ADJ
cana-6072	394	9	deep	deep	ADJ
cana-6072	394	10	learning	learning	NOUN
cana-6072	394	11	for	for	ADP
cana-6072	394	12	explainable	explainable	ADJ
cana-6072	394	13	sentiment	sentiment	NOUN
cana-6072	394	14	classification	classification	NOUN
cana-6072	394	15	of	of	ADP
cana-6072	394	16	social	social	ADJ
cana-6072	394	17	media	medium	NOUN
cana-6072	394	18	reviews	review	NOUN
cana-6072	394	19	.	.	PUNCT
cana-6072	395	1	technological	technological	ADJ
cana-6072	395	2	forecasting	forecasting	NOUN
cana-6072	395	3	and	and	CCONJ
cana-6072	395	4	social	social	ADJ
cana-6072	395	5	change	change	NOUN
cana-6072	395	6	,	,	PUNCT
cana-6072	395	7	202	202	NUM
cana-6072	395	8	,	,	PUNCT
cana-6072	395	9	123326	123326	NUM
cana-6072	395	10	.	.	PUNCT
cana-6072	396	1	https://doi.org/10.1016/j.techfore.2024.123326	https://doi.org/10.1016/j.techfore.2024.123326	ADJ
cana-6072	396	2	26	26	NUM
cana-6072	396	3	.	.	PUNCT
cana-6072	397	1	zhao	zhao	PROPN
cana-6072	397	2	,	,	PUNCT
cana-6072	397	3	jichang	jichang	PROPN
cana-6072	397	4	,	,	PUNCT
cana-6072	397	5	li	li	PROPN
cana-6072	397	6	dong	dong	PROPN
cana-6072	397	7	,	,	PUNCT
cana-6072	397	8	junjie	junjie	PROPN
cana-6072	397	9	wu	wu	PROPN
cana-6072	397	10	,	,	PUNCT
cana-6072	397	11	and	and	CCONJ
cana-6072	397	12	ke	ke	PROPN
cana-6072	397	13	xu	xu	PROPN
cana-6072	397	14	.	.	PUNCT
cana-6072	397	15	2012	2012	NUM
cana-6072	397	16	.	.	PUNCT
cana-6072	398	1	“	"	PUNCT
cana-6072	398	2	moodlens	moodlen	NOUN
cana-6072	398	3	:	:	PUNCT
cana-6072	398	4	an	an	DET
cana-6072	398	5	emoticon	emoticon	NOUN
cana-6072	398	6	-	-	PUNCT
cana-6072	398	7	based	base	VERB
cana-6072	398	8	sentiment	sentiment	NOUN
cana-6072	398	9	analysis	analysis	NOUN
cana-6072	398	10	system	system	NOUN
cana-6072	398	11	for	for	ADP
cana-6072	398	12	chinese	chinese	ADJ
cana-6072	398	13	tweets	tweet	NOUN
cana-6072	398	14	.	.	PUNCT
cana-6072	398	15	”	"	PUNCT
cana-6072	399	1	proceedings	proceeding	NOUN
cana-6072	399	2	of	of	ADP
cana-6072	399	3	the	the	DET
cana-6072	399	4	acm	acm	PROPN
cana-6072	399	5	sigkdd	sigkdd	PROPN
cana-6072	399	6	international	international	ADJ
cana-6072	399	7	conference	conference	NOUN
cana-6072	399	8	on	on	ADP
cana-6072	399	9	knowledge	knowledge	NOUN
cana-6072	399	10	discovery	discovery	PROPN
cana-6072	399	11	and	and	CCONJ
cana-6072	399	12	data	datum	NOUN
cana-6072	399	13	mining	mining	NOUN
cana-6072	399	14	(	(	PUNCT
cana-6072	399	15	august	august	PROPN
cana-6072	399	16	2016):1528–31	2016):1528–31	PROPN
cana-6072	399	17	.	.	PUNCT
cana-6072	400	1	doi	doi	PROPN
cana-6072	400	2	:	:	PUNCT
cana-6072	400	3	10.1145/2339530.2339772	10.1145/2339530.2339772	NUM
cana-6072	400	4	.	.	PUNCT
cana-6072	401	1	27	27	NUM
cana-6072	401	2	.	.	PUNCT
cana-6072	402	1	zhao	zhao	NOUN
cana-6072	402	2	,	,	PUNCT
cana-6072	402	3	peijun	peijun	PROPN
cana-6072	402	4	,	,	PUNCT
cana-6072	402	5	jia	jia	PROPN
cana-6072	402	6	jia	jia	PROPN
cana-6072	402	7	,	,	PUNCT
cana-6072	402	8	yongsheng	yongsheng	PROPN
cana-6072	402	9	an	an	PROPN
cana-6072	402	10	,	,	PUNCT
cana-6072	402	11	jie	jie	PROPN
cana-6072	402	12	liang	liang	PROPN
cana-6072	402	13	,	,	PUNCT
cana-6072	402	14	lexing	lexing	PROPN
cana-6072	402	15	xie	xie	PROPN
cana-6072	402	16	,	,	PUNCT
cana-6072	402	17	and	and	CCONJ
cana-6072	402	18	jiebo	jiebo	PROPN
cana-6072	402	19	luo	luo	PROPN
cana-6072	402	20	.	.	PROPN
cana-6072	402	21	2018	2018	NUM
cana-6072	402	22	.	.	PUNCT
cana-6072	403	1	“	"	PUNCT
cana-6072	403	2	analyzing	analyze	VERB
cana-6072	403	3	and	and	CCONJ
cana-6072	403	4	predicting	predict	VERB
cana-6072	403	5	emoji	emoji	PROPN
cana-6072	403	6	usages	usage	NOUN
cana-6072	403	7	in	in	ADP
cana-6072	403	8	social	social	ADJ
cana-6072	403	9	media	medium	NOUN
cana-6072	403	10	.	.	PUNCT
cana-6072	403	11	”	"	PUNCT
cana-6072	404	1	the	the	DET
cana-6072	404	2	web	web	NOUN
cana-6072	404	3	conference	conference	NOUN
cana-6072	404	4	2018	2018	NUM
cana-6072	404	5	companion	companion	NOUN
cana-6072	404	6	of	of	ADP
cana-6072	404	7	the	the	DET
cana-6072	404	8	world	world	NOUN
cana-6072	404	9	wide	wide	ADJ
cana-6072	404	10	web	web	NOUN
cana-6072	404	11	conference	conference	NOUN
cana-6072	404	12	,	,	PUNCT
cana-6072	404	13	www	www	PROPN
cana-6072	404	14	2018	2018	NUM
cana-6072	404	15	327–34	327–34	NUM
cana-6072	404	16	.	.	PUNCT
cana-6072	405	1	doi	doi	NOUN
cana-6072	405	2	:	:	PUNCT
cana-6072	405	3	10.1145/3184558.3186344	10.1145/3184558.3186344	NUM
cana-6072	405	4	.	.	NOUN
cana-6072	406	1	28	28	NUM
cana-6072	406	2	.	.	PUNCT
cana-6072	407	1	zou	zou	PROPN
cana-6072	407	2	,	,	PUNCT
cana-6072	407	3	haochen	haochen	ADV
cana-6072	407	4	,	,	PUNCT
cana-6072	407	5	and	and	CCONJ
cana-6072	407	6	kun	kun	PROPN
cana-6072	407	7	xiang	xiang	PROPN
cana-6072	407	8	.	.	PROPN
cana-6072	407	9	2022	2022	NUM
cana-6072	407	10	.	.	PUNCT
cana-6072	408	1	“	"	PUNCT
cana-6072	408	2	sentiment	sentiment	NOUN
cana-6072	408	3	classification	classification	NOUN
cana-6072	408	4	method	method	NOUN
cana-6072	408	5	based	base	VERB
cana-6072	408	6	on	on	ADP
cana-6072	408	7	blending	blending	NOUN
cana-6072	408	8	of	of	ADP
cana-6072	408	9	emoticons	emoticon	NOUN
cana-6072	408	10	and	and	CCONJ
cana-6072	408	11	short	short	ADJ
cana-6072	408	12	texts	text	NOUN
cana-6072	408	13	.	.	PUNCT
cana-6072	408	14	”	"	PUNCT
cana-6072	408	15	entropy	entropy	PROPN
cana-6072	408	16	24(3	24(3	NUM
cana-6072	408	17	)	)	PUNCT
cana-6072	408	18	.	.	PUNCT
cana-6072	409	1	doi	doi	NOUN
cana-6072	409	2	:	:	PUNCT
cana-6072	409	3	10.3390	10.3390	NUM
cana-6072	409	4	/	/	SYM
cana-6072	409	5	e24030398	e24030398	PROPN
cana-6072	409	6	.	.	PUNCT
cana-6072	410	1	https://doi.org/10.1016/j.techfore.2024.123326	https://doi.org/10.1016/j.techfore.2024.123326	ADJ