id	sid	tid	token	lemma	pos
easat-10389	1	1	edelweiss	edelweiss	PROPN
easat-10389	1	2	applied	apply	VERB
easat-10389	1	3	science	science	NOUN
easat-10389	1	4	and	and	CCONJ
easat-10389	1	5	technology	technology	NOUN
easat-10389	1	6	issn	issn	PROPN
easat-10389	1	7	:	:	PUNCT
easat-10389	1	8	2576	2576	NUM
easat-10389	1	9	-	-	SYM
easat-10389	1	10	8484	8484	NUM
easat-10389	1	11	vol	vol	NOUN
easat-10389	1	12	.	.	PROPN
easat-10389	2	1	9	9	NUM
easat-10389	2	2	,	,	PUNCT
easat-10389	2	3	no	no	INTJ
easat-10389	2	4	.	.	NOUN
easat-10389	2	5	10	10	NUM
easat-10389	2	6	,	,	PUNCT
easat-10389	2	7	256	256	NUM
easat-10389	2	8	-	-	SYM
easat-10389	2	9	271	271	NUM
easat-10389	2	10	2025	2025	NUM
easat-10389	2	11	publisher	publisher	NOUN
easat-10389	2	12	:	:	PUNCT
easat-10389	2	13	learning	learn	VERB
easat-10389	2	14	gate	gate	NOUN
easat-10389	2	15	doi	doi	PROPN
easat-10389	2	16	:	:	PUNCT
easat-10389	2	17	10.55214/2576	10.55214/2576	NUM
easat-10389	2	18	-	-	PUNCT
easat-10389	2	19	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	3	1	©	©	ADP
easat-10389	3	2	2025	2025	NUM
easat-10389	3	3	by	by	ADP
easat-10389	3	4	the	the	DET
easat-10389	3	5	authors	author	NOUN
easat-10389	3	6	;	;	PUNCT
easat-10389	3	7	licensee	licensee	PROPN
easat-10389	3	8	learning	learning	NOUN
easat-10389	3	9	gate	gate	NOUN
easat-10389	3	10	©	©	PROPN
easat-10389	3	11	2025	2025	NUM
easat-10389	3	12	by	by	ADP
easat-10389	3	13	the	the	DET
easat-10389	3	14	authors	author	NOUN
easat-10389	3	15	;	;	PUNCT
easat-10389	3	16	licensee	licensee	PROPN
easat-10389	3	17	learning	learning	NOUN
easat-10389	3	18	gate	gate	NOUN
easat-10389	3	19	history	history	NOUN
easat-10389	3	20	:	:	PUNCT
easat-10389	3	21	received	receive	VERB
easat-10389	3	22	:	:	PUNCT
easat-10389	3	23	8	8	NUM
easat-10389	3	24	august	august	PROPN
easat-10389	3	25	2025	2025	NUM
easat-10389	3	26	;	;	PUNCT
easat-10389	3	27	revised	revise	VERB
easat-10389	3	28	:	:	PUNCT
easat-10389	3	29	11	11	NUM
easat-10389	3	30	september	september	PROPN
easat-10389	3	31	2025	2025	NUM
easat-10389	3	32	;	;	PUNCT
easat-10389	3	33	accepted	accept	VERB
easat-10389	3	34	:	:	PUNCT
easat-10389	3	35	15	15	NUM
easat-10389	3	36	september	september	PROPN
easat-10389	3	37	2025	2025	NUM
easat-10389	3	38	;	;	PUNCT
easat-10389	3	39	published	publish	VERB
easat-10389	3	40	:	:	PUNCT
easat-10389	4	1	7	7	NUM
easat-10389	4	2	october	october	NOUN
easat-10389	4	3	2025	2025	NUM
easat-10389	4	4	*	*	PUNCT
easat-10389	4	5	correspondence	correspondence	NOUN
easat-10389	4	6	:	:	PUNCT
easat-10389	4	7	vothanhha.ktd@utc.edu.vn	vothanhha.ktd@utc.edu.vn	X
easat-10389	4	8	hybrid	hybrid	PROPN
easat-10389	4	9	cnn	cnn	PROPN
easat-10389	4	10	–	–	PUNCT
easat-10389	4	11	gru	gru	NOUN
easat-10389	4	12	attention	attention	NOUN
easat-10389	4	13	network	network	NOUN
easat-10389	4	14	for	for	ADP
easat-10389	4	15	predicting	predict	VERB
easat-10389	4	16	asthma	asthma	NOUN
easat-10389	4	17	incidence	incidence	NOUN
easat-10389	4	18	in	in	ADP
easat-10389	4	19	urban	urban	ADJ
easat-10389	4	20	environments	environment	NOUN
easat-10389	4	21	vo	vo	X
easat-10389	4	22	thanh	thanh	PROPN
easat-10389	4	23	ha1	ha1	PROPN
easat-10389	4	24	*	*	PROPN
easat-10389	4	25	,	,	PUNCT
easat-10389	4	26	pham	pham	PROPN
easat-10389	4	27	tam	tam	PROPN
easat-10389	4	28	thanh2	thanh2	PROPN
easat-10389	4	29	,	,	PUNCT
easat-10389	4	30	le	le	X
easat-10389	4	31	an	an	DET
easat-10389	4	32	khanh3	khanh3	NOUN
easat-10389	4	33	1,3faculty	1,3faculty	NUM
easat-10389	4	34	of	of	ADP
easat-10389	4	35	electrical	electrical	ADJ
easat-10389	4	36	and	and	CCONJ
easat-10389	4	37	electronics	electronic	NOUN
easat-10389	4	38	engineering	engineering	NOUN
easat-10389	4	39	,	,	PUNCT
easat-10389	4	40	university	university	NOUN
easat-10389	4	41	of	of	ADP
easat-10389	4	42	transport	transport	NOUN
easat-10389	4	43	and	and	CCONJ
easat-10389	4	44	communications	communication	NOUN
easat-10389	4	45	,	,	PUNCT
easat-10389	4	46	hanoi	hanoi	PROPN
easat-10389	4	47	,	,	PUNCT
easat-10389	4	48	vietnam	vietnam	PROPN
easat-10389	4	49	;	;	PUNCT
easat-10389	4	50	vothanhha.ktd@utc.edu.vn	vothanhha.ktd@utc.edu.vn	X
easat-10389	4	51	(	(	PUNCT
easat-10389	4	52	v.t.h	v.t.h	NOUN
easat-10389	4	53	.	.	PUNCT
easat-10389	4	54	)	)	PUNCT
easat-10389	4	55	.	.	PUNCT
easat-10389	5	1	2vietnam	2vietnam	NUM
easat-10389	5	2	maritime	maritime	ADJ
easat-10389	5	3	university	university	NOUN
easat-10389	5	4	,	,	PUNCT
easat-10389	5	5	hai	hai	PROPN
easat-10389	5	6	phong	phong	PROPN
easat-10389	5	7	,	,	PUNCT
easat-10389	5	8	vietnam	vietnam	PROPN
easat-10389	5	9	.	.	PUNCT
easat-10389	6	1	abstract	abstract	ADJ
easat-10389	6	2	:	:	PUNCT
easat-10389	6	3	this	this	DET
easat-10389	6	4	study	study	NOUN
easat-10389	6	5	aims	aim	VERB
easat-10389	6	6	to	to	PART
easat-10389	6	7	develop	develop	VERB
easat-10389	6	8	a	a	DET
easat-10389	6	9	robust	robust	ADJ
easat-10389	6	10	deep	deep	ADJ
easat-10389	6	11	learning	learning	NOUN
easat-10389	6	12	framework	framework	NOUN
easat-10389	6	13	for	for	ADP
easat-10389	6	14	predicting	predict	VERB
easat-10389	6	15	asthma	asthma	NOUN
easat-10389	6	16	incidence	incidence	NOUN
easat-10389	6	17	by	by	ADP
easat-10389	6	18	utilizing	utilize	VERB
easat-10389	6	19	the	the	DET
easat-10389	6	20	air	air	NOUN
easat-10389	6	21	quality	quality	NOUN
easat-10389	6	22	index	index	NOUN
easat-10389	6	23	(	(	PUNCT
easat-10389	6	24	aqi	aqi	PROPN
easat-10389	6	25	)	)	PUNCT
easat-10389	6	26	and	and	CCONJ
easat-10389	6	27	environmental	environmental	ADJ
easat-10389	6	28	data	datum	NOUN
easat-10389	6	29	,	,	PUNCT
easat-10389	6	30	thereby	thereby	ADV
easat-10389	6	31	enhancing	enhance	VERB
easat-10389	6	32	proactive	proactive	ADJ
easat-10389	6	33	public	public	ADJ
easat-10389	6	34	health	health	NOUN
easat-10389	6	35	monitoring	monitoring	NOUN
easat-10389	6	36	.	.	PUNCT
easat-10389	7	1	a	a	DET
easat-10389	7	2	dual	dual	ADJ
easat-10389	7	3	-	-	PUNCT
easat-10389	7	4	branch	branch	NOUN
easat-10389	7	5	architecture	architecture	NOUN
easat-10389	7	6	is	be	AUX
easat-10389	7	7	proposed	propose	VERB
easat-10389	7	8	,	,	PUNCT
easat-10389	7	9	combining	combine	VERB
easat-10389	7	10	a	a	DET
easat-10389	7	11	convolutional	convolutional	ADJ
easat-10389	7	12	neural	neural	ADJ
easat-10389	7	13	network	network	NOUN
easat-10389	7	14	(	(	PUNCT
easat-10389	7	15	cnn	cnn	PROPN
easat-10389	7	16	)	)	PUNCT
easat-10389	7	17	to	to	PART
easat-10389	7	18	extract	extract	VERB
easat-10389	7	19	spatial	spatial	ADJ
easat-10389	7	20	features	feature	NOUN
easat-10389	7	21	of	of	ADP
easat-10389	7	22	air	air	NOUN
easat-10389	7	23	pollution	pollution	NOUN
easat-10389	7	24	with	with	ADP
easat-10389	7	25	a	a	DET
easat-10389	7	26	gated	gate	VERB
easat-10389	7	27	recurrent	recurrent	ADJ
easat-10389	7	28	unit	unit	NOUN
easat-10389	7	29	(	(	PUNCT
easat-10389	7	30	gru	gru	PROPN
easat-10389	7	31	)	)	PUNCT
easat-10389	7	32	to	to	PART
easat-10389	7	33	model	model	VERB
easat-10389	7	34	temporal	temporal	ADJ
easat-10389	7	35	changes	change	NOUN
easat-10389	7	36	in	in	ADP
easat-10389	7	37	weather	weather	NOUN
easat-10389	7	38	conditions	condition	NOUN
easat-10389	7	39	.	.	PUNCT
easat-10389	8	1	an	an	DET
easat-10389	8	2	attention	attention	NOUN
easat-10389	8	3	fusion	fusion	NOUN
easat-10389	8	4	mechanism	mechanism	NOUN
easat-10389	8	5	adaptively	adaptively	ADV
easat-10389	8	6	emphasizes	emphasize	VERB
easat-10389	8	7	critical	critical	ADJ
easat-10389	8	8	environmental	environmental	ADJ
easat-10389	8	9	factors	factor	NOUN
easat-10389	8	10	contributing	contribute	VERB
easat-10389	8	11	to	to	ADP
easat-10389	8	12	asthma	asthma	NOUN
easat-10389	8	13	onset	onset	VERB
easat-10389	8	14	.	.	PUNCT
easat-10389	9	1	the	the	DET
easat-10389	9	2	model	model	NOUN
easat-10389	9	3	was	be	AUX
easat-10389	9	4	trained	train	VERB
easat-10389	9	5	and	and	CCONJ
easat-10389	9	6	validated	validate	VERB
easat-10389	9	7	using	use	VERB
easat-10389	9	8	datasets	dataset	NOUN
easat-10389	9	9	from	from	ADP
easat-10389	9	10	seoul	seoul	PROPN
easat-10389	9	11	,	,	PUNCT
easat-10389	9	12	los	los	PROPN
easat-10389	9	13	angeles	angeles	PROPN
easat-10389	9	14	,	,	PUNCT
easat-10389	9	15	and	and	CCONJ
easat-10389	9	16	hanoi	hanoi	PROPN
easat-10389	9	17	,	,	PUNCT
easat-10389	9	18	covering	cover	VERB
easat-10389	9	19	diverse	diverse	ADJ
easat-10389	9	20	climatic	climatic	ADJ
easat-10389	9	21	and	and	CCONJ
easat-10389	9	22	pollution	pollution	NOUN
easat-10389	9	23	patterns	pattern	NOUN
easat-10389	9	24	.	.	PUNCT
easat-10389	10	1	experimental	experimental	ADJ
easat-10389	10	2	results	result	NOUN
easat-10389	10	3	demonstrate	demonstrate	VERB
easat-10389	10	4	that	that	SCONJ
easat-10389	10	5	the	the	DET
easat-10389	10	6	proposed	propose	VERB
easat-10389	10	7	cnn	cnn	PROPN
easat-10389	10	8	–	–	PUNCT
easat-10389	10	9	gru	gru	PROPN
easat-10389	10	10	–	–	PUNCT
easat-10389	10	11	attention	attention	NOUN
easat-10389	10	12	model	model	NOUN
easat-10389	10	13	consistently	consistently	ADV
easat-10389	10	14	outperforms	outperform	VERB
easat-10389	10	15	traditional	traditional	ADJ
easat-10389	10	16	machine	machine	NOUN
easat-10389	10	17	learning	learning	NOUN
easat-10389	10	18	and	and	CCONJ
easat-10389	10	19	single	single	ADJ
easat-10389	10	20	-	-	PUNCT
easat-10389	10	21	branch	branch	NOUN
easat-10389	10	22	deep	deep	ADJ
easat-10389	10	23	learning	learning	NOUN
easat-10389	10	24	models	model	NOUN
easat-10389	10	25	,	,	PUNCT
easat-10389	10	26	achieving	achieve	VERB
easat-10389	10	27	an	an	DET
easat-10389	10	28	area	area	NOUN
easat-10389	10	29	under	under	ADP
easat-10389	10	30	the	the	DET
easat-10389	10	31	curve	curve	NOUN
easat-10389	10	32	(	(	PUNCT
easat-10389	10	33	auc	auc	NOUN
easat-10389	10	34	)	)	PUNCT
easat-10389	10	35	of	of	ADP
easat-10389	10	36	0.89	0.89	NUM
easat-10389	10	37	and	and	CCONJ
easat-10389	10	38	an	an	DET
easat-10389	10	39	f1	f1	NOUN
easat-10389	10	40	-	-	PUNCT
easat-10389	10	41	score	score	NOUN
easat-10389	10	42	of	of	ADP
easat-10389	10	43	0.84	0.84	NUM
easat-10389	10	44	.	.	PUNCT
easat-10389	11	1	these	these	DET
easat-10389	11	2	findings	finding	NOUN
easat-10389	11	3	highlight	highlight	VERB
easat-10389	11	4	the	the	DET
easat-10389	11	5	model	model	NOUN
easat-10389	11	6	’s	’s	PART
easat-10389	11	7	ability	ability	NOUN
easat-10389	11	8	to	to	PART
easat-10389	11	9	capture	capture	VERB
easat-10389	11	10	complex	complex	ADJ
easat-10389	11	11	spatiotemporal	spatiotemporal	ADJ
easat-10389	11	12	dependencies	dependency	NOUN
easat-10389	11	13	between	between	ADP
easat-10389	11	14	pollution	pollution	NOUN
easat-10389	11	15	and	and	CCONJ
easat-10389	11	16	weather	weather	NOUN
easat-10389	11	17	.	.	PUNCT
easat-10389	12	1	the	the	DET
easat-10389	12	2	approach	approach	NOUN
easat-10389	12	3	provides	provide	VERB
easat-10389	12	4	a	a	DET
easat-10389	12	5	scalable	scalable	ADJ
easat-10389	12	6	,	,	PUNCT
easat-10389	12	7	data	datum	NOUN
easat-10389	12	8	-	-	PUNCT
easat-10389	12	9	driven	drive	VERB
easat-10389	12	10	foundation	foundation	NOUN
easat-10389	12	11	for	for	ADP
easat-10389	12	12	early	early	ADJ
easat-10389	12	13	asthma	asthma	NOUN
easat-10389	12	14	risk	risk	NOUN
easat-10389	12	15	warning	warning	NOUN
easat-10389	12	16	systems	system	NOUN
easat-10389	12	17	and	and	CCONJ
easat-10389	12	18	urban	urban	PROPN
easat-10389	12	19	environmental	environmental	ADJ
easat-10389	12	20	health	health	NOUN
easat-10389	12	21	monitoring	monitoring	NOUN
easat-10389	12	22	applications	application	NOUN
easat-10389	12	23	.	.	PUNCT
easat-10389	13	1	keywords	keyword	NOUN
easat-10389	13	2	:	:	PUNCT
easat-10389	13	3	aqi	aqi	PROPN
easat-10389	13	4	,	,	PUNCT
easat-10389	13	5	cnn	cnn	PROPN
easat-10389	13	6	,	,	PUNCT
easat-10389	13	7	deep	deep	ADJ
easat-10389	13	8	learning	learning	NOUN
easat-10389	13	9	,	,	PUNCT
easat-10389	13	10	gru	gru	PROPN
easat-10389	13	11	,	,	PUNCT
easat-10389	13	12	machine	machine	NOUN
easat-10389	13	13	learning	learning	NOUN
easat-10389	13	14	.	.	PUNCT
easat-10389	14	1	1	1	X
easat-10389	14	2	.	.	X
easat-10389	14	3	introduction	introduction	NOUN
easat-10389	14	4	over	over	ADP
easat-10389	14	5	260	260	NUM
easat-10389	14	6	million	million	NUM
easat-10389	14	7	people	people	NOUN
easat-10389	14	8	worldwide	worldwide	ADV
easat-10389	14	9	suffer	suffer	VERB
easat-10389	14	10	from	from	ADP
easat-10389	14	11	asthma	asthma	NOUN
easat-10389	14	12	,	,	PUNCT
easat-10389	14	13	a	a	DET
easat-10389	14	14	chronic	chronic	ADJ
easat-10389	14	15	respiratory	respiratory	ADJ
easat-10389	14	16	disease	disease	NOUN
easat-10389	14	17	significantly	significantly	ADV
easat-10389	14	18	influenced	influence	VERB
easat-10389	14	19	by	by	ADP
easat-10389	14	20	environmental	environmental	ADJ
easat-10389	14	21	factors	factor	NOUN
easat-10389	14	22	,	,	PUNCT
easat-10389	14	23	including	include	VERB
easat-10389	14	24	air	air	NOUN
easat-10389	14	25	quality	quality	NOUN
easat-10389	14	26	,	,	PUNCT
easat-10389	14	27	industrial	industrial	ADJ
easat-10389	14	28	development	development	NOUN
easat-10389	14	29	,	,	PUNCT
easat-10389	14	30	and	and	CCONJ
easat-10389	14	31	weather	weather	NOUN
easat-10389	14	32	fluctuations	fluctuation	NOUN
easat-10389	14	33	[	[	X
easat-10389	14	34	1	1	NUM
easat-10389	14	35	]	]	PUNCT
easat-10389	14	36	.	.	PUNCT
easat-10389	15	1	in	in	ADP
easat-10389	15	2	recent	recent	ADJ
easat-10389	15	3	decades	decade	NOUN
easat-10389	15	4	,	,	PUNCT
easat-10389	15	5	hospital	hospital	NOUN
easat-10389	15	6	admissions	admission	NOUN
easat-10389	15	7	and	and	CCONJ
easat-10389	15	8	emergency	emergency	NOUN
easat-10389	15	9	room	room	NOUN
easat-10389	15	10	visits	visit	NOUN
easat-10389	15	11	for	for	ADP
easat-10389	15	12	asthma	asthma	NOUN
easat-10389	15	13	have	have	AUX
easat-10389	15	14	increased	increase	VERB
easat-10389	15	15	significantly	significantly	ADV
easat-10389	15	16	,	,	PUNCT
easat-10389	15	17	particularly	particularly	ADV
easat-10389	15	18	in	in	ADP
easat-10389	15	19	urban	urban	ADJ
easat-10389	15	20	and	and	CCONJ
easat-10389	15	21	polluted	polluted	ADJ
easat-10389	15	22	areas	area	NOUN
easat-10389	15	23	[	[	X
easat-10389	15	24	2	2	NUM
easat-10389	15	25	,	,	PUNCT
easat-10389	15	26	3	3	NUM
easat-10389	15	27	]	]	PUNCT
easat-10389	15	28	.	.	PUNCT
easat-10389	16	1	this	this	DET
easat-10389	16	2	increase	increase	NOUN
easat-10389	16	3	is	be	AUX
easat-10389	16	4	mainly	mainly	ADV
easat-10389	16	5	due	due	ADJ
easat-10389	16	6	to	to	ADP
easat-10389	16	7	industrial	industrial	ADJ
easat-10389	16	8	growth	growth	NOUN
easat-10389	16	9	,	,	PUNCT
easat-10389	16	10	urban	urban	ADJ
easat-10389	16	11	transportation	transportation	NOUN
easat-10389	16	12	,	,	PUNCT
easat-10389	16	13	and	and	CCONJ
easat-10389	16	14	climate	climate	NOUN
easat-10389	16	15	instability	instability	NOUN
easat-10389	16	16	[	[	X
easat-10389	16	17	4	4	NUM
easat-10389	16	18	,	,	PUNCT
easat-10389	16	19	5	5	NUM
easat-10389	16	20	]	]	PUNCT
easat-10389	16	21	.	.	PUNCT
easat-10389	17	1	traditional	traditional	ADJ
easat-10389	17	2	monitoring	monitoring	NOUN
easat-10389	17	3	methods	method	NOUN
easat-10389	17	4	primarily	primarily	ADV
easat-10389	17	5	rely	rely	VERB
easat-10389	17	6	on	on	ADP
easat-10389	17	7	statistical	statistical	ADJ
easat-10389	17	8	or	or	CCONJ
easat-10389	17	9	rule	rule	NOUN
easat-10389	17	10	-	-	PUNCT
easat-10389	17	11	based	base	VERB
easat-10389	17	12	models	model	NOUN
easat-10389	17	13	that	that	PRON
easat-10389	17	14	analyze	analyze	VERB
easat-10389	17	15	past	past	ADJ
easat-10389	17	16	hospital	hospital	NOUN
easat-10389	17	17	visits	visit	NOUN
easat-10389	17	18	and	and	CCONJ
easat-10389	17	19	pollutant	pollutant	ADJ
easat-10389	17	20	levels	level	NOUN
easat-10389	17	21	.	.	PUNCT
easat-10389	18	1	while	while	SCONJ
easat-10389	18	2	these	these	DET
easat-10389	18	3	approaches	approach	NOUN
easat-10389	18	4	offer	offer	VERB
easat-10389	18	5	useful	useful	ADJ
easat-10389	18	6	insights	insight	NOUN
easat-10389	18	7	into	into	ADP
easat-10389	18	8	overall	overall	ADJ
easat-10389	18	9	exposure	exposure	NOUN
easat-10389	18	10	–	–	PUNCT
easat-10389	18	11	response	response	NOUN
easat-10389	18	12	patterns	pattern	NOUN
easat-10389	18	13	,	,	PUNCT
easat-10389	18	14	they	they	PRON
easat-10389	18	15	often	often	ADV
easat-10389	18	16	fail	fail	VERB
easat-10389	18	17	to	to	PART
easat-10389	18	18	capture	capture	VERB
easat-10389	18	19	the	the	DET
easat-10389	18	20	nonlinear	nonlinear	ADJ
easat-10389	18	21	and	and	CCONJ
easat-10389	18	22	dynamic	dynamic	ADJ
easat-10389	18	23	interactions	interaction	NOUN
easat-10389	18	24	among	among	ADP
easat-10389	18	25	various	various	ADJ
easat-10389	18	26	environmental	environmental	ADJ
easat-10389	18	27	factors	factor	NOUN
easat-10389	18	28	,	,	PUNCT
easat-10389	18	29	such	such	ADJ
easat-10389	18	30	as	as	ADP
easat-10389	18	31	temperature	temperature	NOUN
easat-10389	18	32	changes	change	NOUN
easat-10389	18	33	,	,	PUNCT
easat-10389	18	34	humidity	humidity	NOUN
easat-10389	18	35	,	,	PUNCT
easat-10389	18	36	and	and	CCONJ
easat-10389	18	37	seasonal	seasonal	ADJ
easat-10389	18	38	shifts	shift	NOUN
easat-10389	18	39	[	[	X
easat-10389	18	40	2	2	NUM
easat-10389	18	41	,	,	PUNCT
easat-10389	18	42	6	6	NUM
easat-10389	18	43	]	]	PUNCT
easat-10389	18	44	.	.	PUNCT
easat-10389	19	1	recent	recent	ADJ
easat-10389	19	2	research	research	NOUN
easat-10389	19	3	shows	show	VERB
easat-10389	19	4	that	that	SCONJ
easat-10389	19	5	short	short	ADJ
easat-10389	19	6	-	-	PUNCT
easat-10389	19	7	term	term	NOUN
easat-10389	19	8	exposure	exposure	NOUN
easat-10389	19	9	to	to	ADP
easat-10389	19	10	particulate	particulate	NOUN
easat-10389	19	11	matter	matter	NOUN
easat-10389	19	12	(	(	PUNCT
easat-10389	19	13	e.g.	e.g.	ADV
easat-10389	19	14	,	,	PUNCT
easat-10389	19	15	pm₂.₅	pm₂.₅	NUM
easat-10389	19	16	)	)	PUNCT
easat-10389	19	17	and	and	CCONJ
easat-10389	19	18	other	other	ADJ
easat-10389	19	19	pollutants	pollutant	NOUN
easat-10389	19	20	greatly	greatly	ADV
easat-10389	19	21	raises	raise	VERB
easat-10389	19	22	the	the	DET
easat-10389	19	23	risk	risk	NOUN
easat-10389	19	24	of	of	ADP
easat-10389	19	25	asthma	asthma	NOUN
easat-10389	19	26	flareups	flareup	VERB
easat-10389	19	27	,	,	PUNCT
easat-10389	19	28	emphasizing	emphasize	VERB
easat-10389	19	29	the	the	DET
easat-10389	19	30	need	need	NOUN
easat-10389	19	31	for	for	ADP
easat-10389	19	32	more	more	ADV
easat-10389	19	33	advanced	advanced	ADJ
easat-10389	19	34	predictive	predictive	ADJ
easat-10389	19	35	tools	tool	NOUN
easat-10389	19	36	[	[	X
easat-10389	19	37	5	5	NUM
easat-10389	19	38	,	,	PUNCT
easat-10389	19	39	6	6	NUM
easat-10389	19	40	]	]	PUNCT
easat-10389	19	41	.	.	PUNCT
easat-10389	20	1	with	with	ADP
easat-10389	20	2	the	the	DET
easat-10389	20	3	rise	rise	NOUN
easat-10389	20	4	of	of	ADP
easat-10389	20	5	machine	machine	NOUN
easat-10389	20	6	learning	learning	NOUN
easat-10389	20	7	,	,	PUNCT
easat-10389	20	8	scientists	scientist	NOUN
easat-10389	20	9	have	have	AUX
easat-10389	20	10	begun	begin	VERB
easat-10389	20	11	using	use	VERB
easat-10389	20	12	deep	deep	ADJ
easat-10389	20	13	neural	neural	ADJ
easat-10389	20	14	networks	network	NOUN
easat-10389	20	15	to	to	PART
easat-10389	20	16	uncover	uncover	VERB
easat-10389	20	17	complex	complex	ADJ
easat-10389	20	18	spatiotemporal	spatiotemporal	ADJ
easat-10389	20	19	relationships	relationship	NOUN
easat-10389	20	20	in	in	ADP
easat-10389	20	21	health	health	NOUN
easat-10389	20	22	and	and	CCONJ
easat-10389	20	23	environmental	environmental	ADJ
easat-10389	20	24	data	datum	NOUN
easat-10389	21	1	[	[	X
easat-10389	21	2	7	7	NUM
easat-10389	21	3	,	,	PUNCT
easat-10389	21	4	8	8	NUM
easat-10389	21	5	]	]	PUNCT
easat-10389	21	6	.	.	PUNCT
easat-10389	22	1	for	for	ADP
easat-10389	22	2	instance	instance	NOUN
easat-10389	22	3	,	,	PUNCT
easat-10389	22	4	lstm	lstm	NOUN
easat-10389	22	5	-	-	PUNCT
easat-10389	22	6	based	base	VERB
easat-10389	22	7	models	model	NOUN
easat-10389	22	8	have	have	AUX
easat-10389	22	9	been	be	AUX
easat-10389	22	10	employed	employ	VERB
easat-10389	22	11	to	to	PART
easat-10389	22	12	forecast	forecast	VERB
easat-10389	22	13	daily	daily	ADJ
easat-10389	22	14	counts	count	NOUN
easat-10389	22	15	of	of	ADP
easat-10389	22	16	asthma	asthma	NOUN
easat-10389	22	17	patients	patient	NOUN
easat-10389	22	18	using	use	VERB
easat-10389	22	19	weather	weather	NOUN
easat-10389	22	20	and	and	CCONJ
easat-10389	22	21	air	air	NOUN
easat-10389	22	22	quality	quality	NOUN
easat-10389	22	23	information	information	NOUN
easat-10389	23	1	[	[	X
easat-10389	23	2	9	9	NUM
easat-10389	23	3	]	]	PUNCT
easat-10389	23	4	.	.	PUNCT
easat-10389	24	1	more	more	ADV
easat-10389	24	2	sophisticated	sophisticated	ADJ
easat-10389	24	3	dual	dual	ADJ
easat-10389	24	4	-	-	PUNCT
easat-10389	24	5	path	path	NOUN
easat-10389	24	6	deep	deep	ADJ
easat-10389	24	7	learning	learning	NOUN
easat-10389	24	8	architectures	architecture	NOUN
easat-10389	24	9	and	and	CCONJ
easat-10389	24	10	attention	attention	NOUN
easat-10389	24	11	mechanisms	mechanism	NOUN
easat-10389	24	12	enhance	enhance	VERB
easat-10389	24	13	sensor	sensor	NOUN
easat-10389	24	14	calibration	calibration	NOUN
easat-10389	24	15	and	and	CCONJ
easat-10389	24	16	multimodal	multimodal	NOUN
easat-10389	24	17	data	datum	NOUN
easat-10389	24	18	integration	integration	NOUN
easat-10389	24	19	,	,	PUNCT
easat-10389	24	20	boosting	boost	VERB
easat-10389	24	21	the	the	DET
easat-10389	24	22	accuracy	accuracy	NOUN
easat-10389	24	23	of	of	ADP
easat-10389	24	24	asthma	asthma	NOUN
easat-10389	24	25	risk	risk	NOUN
easat-10389	24	26	forecasts	forecast	NOUN
easat-10389	24	27	[	[	X
easat-10389	24	28	10	10	NUM
easat-10389	24	29	]	]	PUNCT
easat-10389	24	30	.	.	PUNCT
easat-10389	25	1	additionally	additionally	ADV
easat-10389	25	2	,	,	PUNCT
easat-10389	25	3	spatiotemporal	spatiotemporal	ADJ
easat-10389	25	4	graph	graph	NOUN
easat-10389	25	5	neural	neural	ADJ
easat-10389	25	6	networks	network	NOUN
easat-10389	25	7	and	and	CCONJ
easat-10389	25	8	hybrid	hybrid	NOUN
easat-10389	25	9	models	model	NOUN
easat-10389	25	10	that	that	PRON
easat-10389	25	11	combine	combine	AUX
easat-10389	25	12	graph	graph	NOUN
easat-10389	25	13	structures	structure	NOUN
easat-10389	25	14	with	with	ADP
easat-10389	25	15	interpretable	interpretable	ADJ
easat-10389	25	16	neural	neural	ADJ
easat-10389	25	17	networks	network	NOUN
easat-10389	25	18	have	have	AUX
easat-10389	25	19	shown	show	VERB
easat-10389	25	20	promising	promising	ADJ
easat-10389	25	21	results	result	NOUN
easat-10389	25	22	in	in	ADP
easat-10389	25	23	identifying	identify	VERB
easat-10389	25	24	regional	regional	ADJ
easat-10389	25	25	asthma	asthma	NOUN
easat-10389	25	26	patterns	pattern	NOUN
easat-10389	25	27	influenced	influence	VERB
easat-10389	25	28	by	by	ADP
easat-10389	25	29	environmental	environmental	ADJ
easat-10389	25	30	factors	factor	NOUN
easat-10389	25	31	[	[	X
easat-10389	25	32	11	11	NUM
easat-10389	25	33	,	,	PUNCT
easat-10389	25	34	12	12	NUM
easat-10389	25	35	]	]	PUNCT
easat-10389	25	36	.	.	PUNCT
easat-10389	26	1	https://orcid.org/0000-0003-4023-260x	https://orcid.org/0000-0003-4023-260x	PROPN
easat-10389	26	2	https://orcid.org/0000-0001-9656-617x	https://orcid.org/0000-0001-9656-617x	PROPN
easat-10389	26	3	257	257	NUM
easat-10389	26	4	edelweiss	edelweiss	PROPN
easat-10389	26	5	applied	apply	VERB
easat-10389	26	6	science	science	NOUN
easat-10389	26	7	and	and	CCONJ
easat-10389	26	8	technology	technology	NOUN
easat-10389	26	9	issn	issn	PROPN
easat-10389	26	10	:	:	PUNCT
easat-10389	26	11	2576	2576	NUM
easat-10389	26	12	-	-	SYM
easat-10389	26	13	8484	8484	NUM
easat-10389	26	14	vol	vol	NOUN
easat-10389	26	15	.	.	PROPN
easat-10389	27	1	9	9	NUM
easat-10389	27	2	,	,	PUNCT
easat-10389	27	3	no	no	INTJ
easat-10389	27	4	.	.	NOUN
easat-10389	27	5	10	10	NUM
easat-10389	27	6	:	:	PUNCT
easat-10389	27	7	256	256	NUM
easat-10389	27	8	-	-	SYM
easat-10389	27	9	271	271	NUM
easat-10389	27	10	,	,	PUNCT
easat-10389	27	11	2025	2025	NUM
easat-10389	27	12	doi	doi	NOUN
easat-10389	27	13	:	:	PUNCT
easat-10389	27	14	10.55214/2576	10.55214/2576	NUM
easat-10389	27	15	-	-	PUNCT
easat-10389	27	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	27	17	©	©	ADP
easat-10389	27	18	2025	2025	NUM
easat-10389	27	19	by	by	ADP
easat-10389	27	20	the	the	DET
easat-10389	27	21	authors	author	NOUN
easat-10389	27	22	;	;	PUNCT
easat-10389	27	23	licensee	licensee	PROPN
easat-10389	27	24	learning	learn	VERB
easat-10389	27	25	gate	gate	NOUN
easat-10389	27	26	beyond	beyond	ADP
easat-10389	27	27	asthma	asthma	NOUN
easat-10389	27	28	,	,	PUNCT
easat-10389	27	29	researchers	researcher	NOUN
easat-10389	27	30	have	have	AUX
easat-10389	27	31	applied	apply	VERB
easat-10389	27	32	graph	graph	NOUN
easat-10389	27	33	neural	neural	ADJ
easat-10389	27	34	network	network	NOUN
easat-10389	27	35	frameworks	framework	NOUN
easat-10389	27	36	for	for	ADP
easat-10389	27	37	public	public	ADJ
easat-10389	27	38	health	health	NOUN
easat-10389	27	39	surveillance	surveillance	NOUN
easat-10389	27	40	tasks	task	NOUN
easat-10389	27	41	,	,	PUNCT
easat-10389	27	42	such	such	ADJ
easat-10389	27	43	as	as	ADP
easat-10389	27	44	influenza	influenza	NOUN
easat-10389	27	45	-	-	PUNCT
easat-10389	27	46	like	like	ADJ
easat-10389	27	47	illness	illness	NOUN
easat-10389	27	48	nowcasting	nowcasting	NOUN
easat-10389	27	49	,	,	PUNCT
easat-10389	27	50	and	and	CCONJ
easat-10389	27	51	demonstrated	demonstrate	VERB
easat-10389	27	52	their	their	PRON
easat-10389	27	53	ability	ability	NOUN
easat-10389	27	54	to	to	PART
easat-10389	27	55	exploit	exploit	VERB
easat-10389	27	56	the	the	DET
easat-10389	27	57	interplay	interplay	NOUN
easat-10389	27	58	between	between	ADP
easat-10389	27	59	temporal	temporal	ADJ
easat-10389	27	60	,	,	PUNCT
easat-10389	27	61	geographical	geographical	ADJ
easat-10389	27	62	,	,	PUNCT
easat-10389	27	63	and	and	CCONJ
easat-10389	27	64	functional	functional	ADJ
easat-10389	27	65	spatial	spatial	ADJ
easat-10389	27	66	features	feature	NOUN
easat-10389	27	67	[	[	X
easat-10389	27	68	13	13	NUM
easat-10389	27	69	,	,	PUNCT
easat-10389	27	70	14	14	NUM
easat-10389	27	71	]	]	PUNCT
easat-10389	27	72	.	.	PUNCT
easat-10389	28	1	these	these	DET
easat-10389	28	2	innovations	innovation	NOUN
easat-10389	28	3	,	,	PUNCT
easat-10389	28	4	coupled	couple	VERB
easat-10389	28	5	with	with	ADP
easat-10389	28	6	advances	advance	NOUN
easat-10389	28	7	in	in	ADP
easat-10389	28	8	deep	deep	ADJ
easat-10389	28	9	learning	learning	NOUN
easat-10389	28	10	-	-	PUNCT
easat-10389	28	11	based	base	VERB
easat-10389	28	12	biological	biological	ADJ
easat-10389	28	13	modeling	modeling	NOUN
easat-10389	28	14	[	[	X
easat-10389	28	15	15	15	NUM
easat-10389	28	16	]	]	PUNCT
easat-10389	28	17	,	,	PUNCT
easat-10389	28	18	further	far	ADV
easat-10389	28	19	highlight	highlight	VERB
easat-10389	28	20	the	the	DET
easat-10389	28	21	potential	potential	NOUN
easat-10389	28	22	of	of	ADP
easat-10389	28	23	integrating	integrate	VERB
easat-10389	28	24	multi	multi	ADJ
easat-10389	28	25	-	-	NOUN
easat-10389	28	26	source	source	NOUN
easat-10389	28	27	and	and	CCONJ
easat-10389	28	28	multi	multi	ADJ
easat-10389	28	29	-	-	NOUN
easat-10389	28	30	task	task	ADJ
easat-10389	28	31	ai	ai	ADJ
easat-10389	28	32	frameworks	framework	NOUN
easat-10389	28	33	for	for	ADP
easat-10389	28	34	complex	complex	ADJ
easat-10389	28	35	health	health	NOUN
easat-10389	28	36	forecasting	forecasting	NOUN
easat-10389	28	37	problems	problem	NOUN
easat-10389	28	38	.	.	PUNCT
easat-10389	29	1	building	build	VERB
easat-10389	29	2	on	on	ADP
easat-10389	29	3	this	this	DET
easat-10389	29	4	progress	progress	NOUN
easat-10389	29	5	,	,	PUNCT
easat-10389	29	6	recent	recent	ADJ
easat-10389	29	7	work	work	NOUN
easat-10389	29	8	has	have	AUX
easat-10389	29	9	emphasized	emphasize	VERB
easat-10389	29	10	the	the	DET
easat-10389	29	11	importance	importance	NOUN
easat-10389	29	12	of	of	ADP
easat-10389	29	13	spatial	spatial	ADJ
easat-10389	29	14	risk	risk	NOUN
easat-10389	29	15	mapping	mapping	NOUN
easat-10389	29	16	optimized	optimize	VERB
easat-10389	29	17	by	by	ADP
easat-10389	29	18	metaheuristic	metaheuristic	ADJ
easat-10389	29	19	algorithms	algorithm	NOUN
easat-10389	29	20	to	to	PART
easat-10389	29	21	identify	identify	VERB
easat-10389	29	22	asthma	asthma	NOUN
easat-10389	29	23	-	-	PUNCT
easat-10389	29	24	prone	prone	ADJ
easat-10389	29	25	areas	area	NOUN
easat-10389	29	26	and	and	CCONJ
easat-10389	29	27	improve	improve	VERB
easat-10389	29	28	targeted	target	VERB
easat-10389	29	29	intervention	intervention	NOUN
easat-10389	29	30	strategies	strategy	NOUN
easat-10389	29	31	[	[	X
easat-10389	29	32	16	16	NUM
easat-10389	29	33	]	]	PUNCT
easat-10389	29	34	.	.	PUNCT
easat-10389	30	1	2	2	X
easat-10389	30	2	.	.	X
easat-10389	30	3	literature	literature	NOUN
easat-10389	30	4	review	review	VERB
easat-10389	30	5	2.1	2.1	NUM
easat-10389	30	6	.	.	PUNCT
easat-10389	30	7	old	old	ADJ
easat-10389	30	8	-	-	PUNCT
easat-10389	30	9	fashioned	fashioned	ADJ
easat-10389	30	10	and	and	CCONJ
easat-10389	30	11	machine	machine	NOUN
easat-10389	30	12	learning	learning	NOUN
easat-10389	30	13	methods	method	NOUN
easat-10389	30	14	early	early	ADJ
easat-10389	30	15	projections	projection	NOUN
easat-10389	30	16	employed	employ	VERB
easat-10389	30	17	basic	basic	ADJ
easat-10389	30	18	statistical	statistical	ADJ
easat-10389	30	19	methods	method	NOUN
easat-10389	30	20	,	,	PUNCT
easat-10389	30	21	such	such	ADJ
easat-10389	30	22	as	as	ADP
easat-10389	30	23	time	time	NOUN
easat-10389	30	24	series	series	PROPN
easat-10389	30	25	regression	regression	PROPN
easat-10389	30	26	and	and	CCONJ
easat-10389	30	27	arima	arima	PROPN
easat-10389	30	28	,	,	PUNCT
easat-10389	30	29	which	which	PRON
easat-10389	30	30	struggled	struggle	VERB
easat-10389	30	31	to	to	PART
easat-10389	30	32	capture	capture	VERB
easat-10389	30	33	the	the	DET
easat-10389	30	34	complex	complex	ADJ
easat-10389	30	35	patterns	pattern	NOUN
easat-10389	30	36	and	and	CCONJ
easat-10389	30	37	changes	change	NOUN
easat-10389	30	38	in	in	ADP
easat-10389	30	39	environmental	environmental	ADJ
easat-10389	30	40	data	datum	NOUN
easat-10389	30	41	over	over	ADP
easat-10389	30	42	time	time	NOUN
easat-10389	30	43	[	[	X
easat-10389	30	44	17	17	NUM
easat-10389	30	45	]	]	PUNCT
easat-10389	30	46	.	.	PUNCT
easat-10389	31	1	support	support	NOUN
easat-10389	31	2	vector	vector	NOUN
easat-10389	31	3	machines	machine	NOUN
easat-10389	31	4	,	,	PUNCT
easat-10389	31	5	random	random	ADJ
easat-10389	31	6	forest	forest	NOUN
easat-10389	31	7	,	,	PUNCT
easat-10389	31	8	and	and	CCONJ
easat-10389	31	9	xgboost	xgboost	X
easat-10389	31	10	are	be	AUX
easat-10389	31	11	some	some	PRON
easat-10389	31	12	of	of	ADP
easat-10389	31	13	the	the	DET
easat-10389	31	14	newer	new	ADJ
easat-10389	31	15	machine	machine	NOUN
easat-10389	31	16	learning	learning	NOUN
easat-10389	31	17	models	model	NOUN
easat-10389	31	18	that	that	PRON
easat-10389	31	19	perform	perform	VERB
easat-10389	31	20	better	well	ADV
easat-10389	31	21	when	when	SCONJ
easat-10389	31	22	aqi	aqi	ADJ
easat-10389	31	23	and	and	CCONJ
easat-10389	31	24	meteorological	meteorological	ADJ
easat-10389	31	25	data	datum	NOUN
easat-10389	31	26	are	be	AUX
easat-10389	31	27	combined	combine	VERB
easat-10389	31	28	[	[	X
easat-10389	31	29	18	18	NUM
easat-10389	31	30	]	]	PUNCT
easat-10389	31	31	.	.	PUNCT
easat-10389	32	1	however	however	ADV
easat-10389	32	2	,	,	PUNCT
easat-10389	32	3	they	they	PRON
easat-10389	32	4	required	require	VERB
easat-10389	32	5	manual	manual	ADJ
easat-10389	32	6	feature	feature	NOUN
easat-10389	32	7	engineering	engineering	NOUN
easat-10389	32	8	and	and	CCONJ
easat-10389	32	9	did	do	AUX
easat-10389	32	10	not	not	PART
easat-10389	32	11	work	work	VERB
easat-10389	32	12	well	well	ADV
easat-10389	32	13	with	with	ADP
easat-10389	32	14	high	high	ADV
easat-10389	32	15	-	-	PUNCT
easat-10389	32	16	dimensional	dimensional	ADJ
easat-10389	32	17	or	or	CCONJ
easat-10389	32	18	multivariate	multivariate	NOUN
easat-10389	32	19	time	time	NOUN
easat-10389	32	20	series	series	PROPN
easat-10389	32	21	data	data	PROPN
easat-10389	32	22	.	.	PUNCT
easat-10389	33	1	2.2	2.2	NUM
easat-10389	33	2	.	.	PUNCT
easat-10389	33	3	frameworks	framework	NOUN
easat-10389	33	4	based	base	VERB
easat-10389	33	5	on	on	ADP
easat-10389	33	6	deep	deep	ADJ
easat-10389	33	7	learning	learning	NOUN
easat-10389	33	8	deep	deep	ADJ
easat-10389	33	9	learning	learning	NOUN
easat-10389	33	10	has	have	AUX
easat-10389	33	11	gained	gain	VERB
easat-10389	33	12	popularity	popularity	NOUN
easat-10389	33	13	because	because	SCONJ
easat-10389	33	14	it	it	PRON
easat-10389	33	15	can	can	AUX
easat-10389	33	16	learn	learn	VERB
easat-10389	33	17	from	from	ADP
easat-10389	33	18	raw	raw	ADJ
easat-10389	33	19	data	datum	NOUN
easat-10389	33	20	.	.	PUNCT
easat-10389	34	1	huang	huang	PROPN
easat-10389	34	2	et	et	PROPN
easat-10389	34	3	al	al	PROPN
easat-10389	34	4	.	.	PUNCT
easat-10389	35	1	[	[	X
easat-10389	35	2	19	19	NUM
easat-10389	35	3	]	]	PUNCT
easat-10389	35	4	,	,	PUNCT
easat-10389	35	5	for	for	ADP
easat-10389	35	6	example	example	NOUN
easat-10389	35	7	,	,	PUNCT
easat-10389	35	8	used	use	VERB
easat-10389	35	9	lstm	lstm	NOUN
easat-10389	35	10	to	to	PART
easat-10389	35	11	estimate	estimate	VERB
easat-10389	35	12	changes	change	NOUN
easat-10389	35	13	in	in	ADP
easat-10389	35	14	air	air	NOUN
easat-10389	35	15	pollution	pollution	NOUN
easat-10389	35	16	and	and	CCONJ
easat-10389	35	17	temperature	temperature	NOUN
easat-10389	35	18	over	over	ADP
easat-10389	35	19	time	time	NOUN
easat-10389	35	20	,	,	PUNCT
easat-10389	35	21	which	which	PRON
easat-10389	35	22	outperformed	outperform	VERB
easat-10389	35	23	traditional	traditional	ADJ
easat-10389	35	24	ml	ml	NOUN
easat-10389	35	25	models	model	NOUN
easat-10389	35	26	.	.	PUNCT
easat-10389	36	1	lee	lee	PROPN
easat-10389	36	2	et	et	PROPN
easat-10389	36	3	al	al	PROPN
easat-10389	36	4	.	.	PUNCT
easat-10389	37	1	[	[	X
easat-10389	37	2	20	20	NUM
easat-10389	37	3	]	]	PUNCT
easat-10389	37	4	built	build	VERB
easat-10389	37	5	a	a	DET
easat-10389	37	6	cnn	cnn	PROPN
easat-10389	37	7	-	-	PUNCT
easat-10389	37	8	lstm	lstm	ADJ
easat-10389	37	9	hybrid	hybrid	NOUN
easat-10389	37	10	to	to	PART
easat-10389	37	11	predict	predict	VERB
easat-10389	37	12	asthma	asthma	NOUN
easat-10389	37	13	hospitalizations	hospitalization	NOUN
easat-10389	37	14	using	use	VERB
easat-10389	37	15	aqi	aqi	PROPN
easat-10389	37	16	and	and	CCONJ
easat-10389	37	17	weather	weather	NOUN
easat-10389	37	18	data	datum	NOUN
easat-10389	37	19	.	.	PUNCT
easat-10389	38	1	it	it	PRON
easat-10389	38	2	performed	perform	VERB
easat-10389	38	3	better	well	ADV
easat-10389	38	4	on	on	ADP
easat-10389	38	5	some	some	DET
easat-10389	38	6	metrics	metric	NOUN
easat-10389	38	7	but	but	CCONJ
easat-10389	38	8	struggled	struggle	VERB
easat-10389	38	9	to	to	PART
easat-10389	38	10	generalize	generalize	VERB
easat-10389	38	11	across	across	ADP
easat-10389	38	12	different	different	ADJ
easat-10389	38	13	areas	area	NOUN
easat-10389	38	14	.	.	PUNCT
easat-10389	39	1	zhao	zhao	PROPN
easat-10389	39	2	et	et	PROPN
easat-10389	39	3	al	al	PROPN
easat-10389	39	4	.	.	PUNCT
easat-10389	40	1	[	[	X
easat-10389	40	2	21	21	NUM
easat-10389	40	3	]	]	PUNCT
easat-10389	40	4	developed	develop	VERB
easat-10389	40	5	an	an	DET
easat-10389	40	6	attention	attention	NOUN
easat-10389	40	7	-	-	PUNCT
easat-10389	40	8	based	base	VERB
easat-10389	40	9	bi	bi	ADJ
easat-10389	40	10	-	-	ADJ
easat-10389	40	11	lstm	lstm	ADJ
easat-10389	40	12	model	model	NOUN
easat-10389	40	13	for	for	ADP
easat-10389	40	14	predicting	predict	VERB
easat-10389	40	15	asthma	asthma	NOUN
easat-10389	40	16	risk	risk	NOUN
easat-10389	40	17	,	,	PUNCT
easat-10389	40	18	thereby	thereby	ADV
easat-10389	40	19	improving	improve	VERB
easat-10389	40	20	performance	performance	NOUN
easat-10389	40	21	and	and	CCONJ
easat-10389	40	22	facilitating	facilitate	VERB
easat-10389	40	23	a	a	DET
easat-10389	40	24	deeper	deep	ADJ
easat-10389	40	25	understanding	understanding	NOUN
easat-10389	40	26	of	of	ADP
easat-10389	40	27	the	the	DET
easat-10389	40	28	underlying	underlie	VERB
easat-10389	40	29	mechanisms	mechanism	NOUN
easat-10389	40	30	.	.	PUNCT
easat-10389	41	1	however	however	ADV
easat-10389	41	2	,	,	PUNCT
easat-10389	41	3	it	it	PRON
easat-10389	41	4	only	only	ADV
easat-10389	41	5	analyzes	analyze	VERB
easat-10389	41	6	one	one	NUM
easat-10389	41	7	-	-	PUNCT
easat-10389	41	8	dimensional	dimensional	ADJ
easat-10389	41	9	time	time	NOUN
easat-10389	41	10	series	series	NOUN
easat-10389	41	11	(	(	PUNCT
easat-10389	41	12	such	such	ADJ
easat-10389	41	13	as	as	ADP
easat-10389	41	14	pm2.5	pm2.5	PART
easat-10389	41	15	concentration	concentration	NOUN
easat-10389	41	16	)	)	PUNCT
easat-10389	41	17	,	,	PUNCT
easat-10389	41	18	thus	thus	ADV
easat-10389	41	19	ignoring	ignore	VERB
easat-10389	41	20	spatial	spatial	ADJ
easat-10389	41	21	relationships	relationship	NOUN
easat-10389	41	22	.	.	PUNCT
easat-10389	42	1	2.3	2.3	NUM
easat-10389	42	2	.	.	PUNCT
easat-10389	43	1	multi	multi	ADJ
easat-10389	43	2	-	-	ADJ
easat-10389	43	3	modal	modal	ADJ
easat-10389	43	4	and	and	CCONJ
easat-10389	43	5	hybrid	hybrid	ADJ
easat-10389	43	6	architectures	architecture	NOUN
easat-10389	43	7	new	new	ADJ
easat-10389	43	8	research	research	NOUN
easat-10389	43	9	has	have	AUX
easat-10389	43	10	developed	develop	VERB
easat-10389	43	11	multi	multi	ADJ
easat-10389	43	12	-	-	ADJ
easat-10389	43	13	input	input	ADJ
easat-10389	43	14	models	model	NOUN
easat-10389	43	15	that	that	PRON
easat-10389	43	16	utilize	utilize	VERB
easat-10389	43	17	geographical	geographical	ADJ
easat-10389	43	18	and	and	CCONJ
easat-10389	43	19	temporal	temporal	ADJ
easat-10389	43	20	data	datum	NOUN
easat-10389	43	21	to	to	PART
easat-10389	43	22	address	address	VERB
easat-10389	43	23	the	the	DET
easat-10389	43	24	challenges	challenge	NOUN
easat-10389	43	25	faced	face	VERB
easat-10389	43	26	by	by	ADP
easat-10389	43	27	current	current	ADJ
easat-10389	43	28	models	model	NOUN
easat-10389	43	29	.	.	PUNCT
easat-10389	44	1	kim	kim	PROPN
easat-10389	44	2	and	and	CCONJ
easat-10389	44	3	park	park	PROPN
easat-10389	44	4	[	[	X
easat-10389	44	5	22	22	NUM
easat-10389	44	6	]	]	PUNCT
easat-10389	44	7	developed	develop	VERB
easat-10389	44	8	a	a	DET
easat-10389	44	9	two	two	NUM
easat-10389	44	10	-	-	PUNCT
easat-10389	44	11	part	part	NOUN
easat-10389	44	12	system	system	NOUN
easat-10389	44	13	that	that	PRON
easat-10389	44	14	utilizes	utilize	VERB
easat-10389	44	15	a	a	DET
easat-10389	44	16	cnn	cnn	NOUN
easat-10389	44	17	to	to	PART
easat-10389	44	18	analyze	analyze	VERB
easat-10389	44	19	aqi	aqi	ADJ
easat-10389	44	20	variations	variation	NOUN
easat-10389	44	21	across	across	ADP
easat-10389	44	22	different	different	ADJ
easat-10389	44	23	locations	location	NOUN
easat-10389	44	24	and	and	CCONJ
easat-10389	44	25	a	a	DET
easat-10389	44	26	gru	gru	NOUN
easat-10389	44	27	to	to	PART
easat-10389	44	28	comprehend	comprehend	VERB
easat-10389	44	29	weather	weather	NOUN
easat-10389	44	30	changes	change	NOUN
easat-10389	44	31	.	.	PUNCT
easat-10389	45	1	they	they	PRON
easat-10389	45	2	also	also	ADV
easat-10389	45	3	added	add	VERB
easat-10389	45	4	an	an	DET
easat-10389	45	5	attention	attention	NOUN
easat-10389	45	6	fusion	fusion	NOUN
easat-10389	45	7	module	module	NOUN
easat-10389	45	8	to	to	PART
easat-10389	45	9	boost	boost	VERB
easat-10389	45	10	performance	performance	NOUN
easat-10389	45	11	in	in	ADP
easat-10389	45	12	various	various	ADJ
easat-10389	45	13	situations	situation	NOUN
easat-10389	45	14	.	.	PUNCT
easat-10389	46	1	li	li	PROPN
easat-10389	46	2	et	et	PROPN
easat-10389	46	3	al	al	PROPN
easat-10389	46	4	.	.	PUNCT
easat-10389	47	1	[	[	X
easat-10389	47	2	23	23	NUM
easat-10389	47	3	]	]	PUNCT
easat-10389	47	4	built	build	VERB
easat-10389	47	5	a	a	DET
easat-10389	47	6	transformer	transformer	NOUN
easat-10389	47	7	-	-	PUNCT
easat-10389	47	8	based	base	VERB
easat-10389	47	9	encoder	encoder	NOUN
easat-10389	47	10	-	-	PUNCT
easat-10389	47	11	decoder	decoder	NOUN
easat-10389	47	12	model	model	NOUN
easat-10389	47	13	that	that	PRON
easat-10389	47	14	combines	combine	VERB
easat-10389	47	15	real	real	ADJ
easat-10389	47	16	-	-	PUNCT
easat-10389	47	17	time	time	NOUN
easat-10389	47	18	sensor	sensor	NOUN
easat-10389	47	19	data	datum	NOUN
easat-10389	47	20	with	with	ADP
easat-10389	47	21	opensource	opensource	PROPN
easat-10389	47	22	aqi	aqi	PROPN
easat-10389	47	23	indices	index	NOUN
easat-10389	47	24	.	.	PUNCT
easat-10389	48	1	the	the	DET
easat-10389	48	2	model	model	NOUN
easat-10389	48	3	works	work	VERB
easat-10389	48	4	best	well	ADV
easat-10389	48	5	on	on	ADP
easat-10389	48	6	urban	urban	ADJ
easat-10389	48	7	datasets	dataset	NOUN
easat-10389	48	8	;	;	PUNCT
easat-10389	48	9	however	however	ADV
easat-10389	48	10	,	,	PUNCT
easat-10389	48	11	its	its	PRON
easat-10389	48	12	complexity	complexity	NOUN
easat-10389	48	13	and	and	CCONJ
easat-10389	48	14	high	high	ADJ
easat-10389	48	15	training	training	NOUN
easat-10389	48	16	requirements	requirement	NOUN
easat-10389	48	17	make	make	VERB
easat-10389	48	18	it	it	PRON
easat-10389	48	19	challenging	challenging	ADJ
easat-10389	48	20	to	to	PART
easat-10389	48	21	use	use	VERB
easat-10389	48	22	in	in	ADP
easat-10389	48	23	real	real	ADJ
easat-10389	48	24	-	-	PUNCT
easat-10389	48	25	world	world	NOUN
easat-10389	48	26	settings	setting	NOUN
easat-10389	48	27	.	.	PUNCT
easat-10389	49	1	zhang	zhang	PROPN
easat-10389	49	2	et	et	PROPN
easat-10389	49	3	al	al	PROPN
easat-10389	49	4	.	.	PUNCT
easat-10389	50	1	[	[	X
easat-10389	50	2	24	24	NUM
easat-10389	50	3	]	]	PUNCT
easat-10389	50	4	improved	improve	VERB
easat-10389	50	5	a	a	DET
easat-10389	50	6	gnn	gnn	NOUN
easat-10389	50	7	framework	framework	NOUN
easat-10389	50	8	to	to	PART
easat-10389	50	9	capture	capture	VERB
easat-10389	50	10	complex	complex	ADJ
easat-10389	50	11	spatial	spatial	ADJ
easat-10389	50	12	correlations	correlation	NOUN
easat-10389	50	13	between	between	ADP
easat-10389	50	14	air	air	NOUN
easat-10389	50	15	quality	quality	NOUN
easat-10389	50	16	sensors	sensor	NOUN
easat-10389	50	17	.	.	PUNCT
easat-10389	51	1	the	the	DET
easat-10389	51	2	new	new	ADJ
easat-10389	51	3	approach	approach	NOUN
easat-10389	51	4	excels	excel	VERB
easat-10389	51	5	at	at	ADP
easat-10389	51	6	tracking	track	VERB
easat-10389	51	7	the	the	DET
easat-10389	51	8	spread	spread	NOUN
easat-10389	51	9	of	of	ADP
easat-10389	51	10	contaminants	contaminant	NOUN
easat-10389	51	11	using	use	VERB
easat-10389	51	12	a	a	DET
easat-10389	51	13	spatiotemporal	spatiotemporal	ADJ
easat-10389	51	14	graph	graph	NOUN
easat-10389	51	15	attention	attention	NOUN
easat-10389	51	16	mechanism	mechanism	NOUN
easat-10389	51	17	;	;	PUNCT
easat-10389	51	18	however	however	ADV
easat-10389	51	19	,	,	PUNCT
easat-10389	51	20	it	it	PRON
easat-10389	51	21	is	be	AUX
easat-10389	51	22	less	less	ADV
easat-10389	51	23	effective	effective	ADJ
easat-10389	51	24	in	in	ADP
easat-10389	51	25	densely	densely	ADV
easat-10389	51	26	populated	populated	ADJ
easat-10389	51	27	city	city	NOUN
easat-10389	51	28	areas	area	NOUN
easat-10389	51	29	.	.	PUNCT
easat-10389	52	1	there	there	PRON
easat-10389	52	2	is	be	VERB
easat-10389	52	3	considerable	considerable	ADJ
easat-10389	52	4	interest	interest	NOUN
easat-10389	52	5	in	in	ADP
easat-10389	52	6	hybrid	hybrid	ADJ
easat-10389	52	7	designs	design	NOUN
easat-10389	52	8	that	that	PRON
easat-10389	52	9	combine	combine	VERB
easat-10389	52	10	deep	deep	ADJ
easat-10389	52	11	learning	learning	NOUN
easat-10389	52	12	with	with	ADP
easat-10389	52	13	traditional	traditional	ADJ
easat-10389	52	14	statistical	statistical	ADJ
easat-10389	52	15	models	model	NOUN
easat-10389	52	16	.	.	PUNCT
easat-10389	53	1	for	for	ADP
easat-10389	53	2	instance	instance	NOUN
easat-10389	53	3	,	,	PUNCT
easat-10389	53	4	chen	chen	PROPN
easat-10389	53	5	et	et	PROPN
easat-10389	53	6	al	al	PROPN
easat-10389	53	7	.	.	PUNCT
easat-10389	54	1	[	[	X
easat-10389	54	2	25	25	NUM
easat-10389	54	3	]	]	PUNCT
easat-10389	54	4	developed	develop	VERB
easat-10389	54	5	a	a	DET
easat-10389	54	6	hybrid	hybrid	ADJ
easat-10389	54	7	lstm	lstm	ADJ
easat-10389	54	8	-	-	PUNCT
easat-10389	54	9	arima	arima	NOUN
easat-10389	54	10	model	model	NOUN
easat-10389	54	11	that	that	PRON
easat-10389	54	12	leverages	leverage	VERB
easat-10389	54	13	both	both	DET
easat-10389	54	14	statistical	statistical	ADJ
easat-10389	54	15	trend	trend	NOUN
easat-10389	54	16	analysis	analysis	NOUN
easat-10389	54	17	and	and	CCONJ
easat-10389	54	18	deep	deep	ADJ
easat-10389	54	19	learning	learning	NOUN
easat-10389	54	20	's	's	PART
easat-10389	54	21	ability	ability	NOUN
easat-10389	54	22	to	to	PART
easat-10389	54	23	adapt	adapt	VERB
easat-10389	54	24	to	to	ADP
easat-10389	54	25	irregular	irregular	ADJ
easat-10389	54	26	patterns	pattern	NOUN
easat-10389	54	27	.	.	PUNCT
easat-10389	55	1	this	this	DET
easat-10389	55	2	method	method	NOUN
easat-10389	55	3	works	work	VERB
easat-10389	55	4	well	well	ADV
easat-10389	55	5	in	in	ADP
easat-10389	55	6	areas	area	NOUN
easat-10389	55	7	where	where	SCONJ
easat-10389	55	8	pollution	pollution	NOUN
easat-10389	55	9	levels	level	NOUN
easat-10389	55	10	fluctuate	fluctuate	VERB
easat-10389	55	11	with	with	ADP
easat-10389	55	12	the	the	DET
easat-10389	55	13	seasons	season	NOUN
easat-10389	55	14	,	,	PUNCT
easat-10389	55	15	but	but	CCONJ
easat-10389	55	16	it	it	PRON
easat-10389	55	17	requires	require	VERB
easat-10389	55	18	extensive	extensive	ADJ
easat-10389	55	19	preprocessing	preprocessing	NOUN
easat-10389	55	20	,	,	PUNCT
easat-10389	55	21	which	which	PRON
easat-10389	55	22	slows	slow	VERB
easat-10389	55	23	down	down	ADP
easat-10389	55	24	real	real	ADJ
easat-10389	55	25	-	-	PUNCT
easat-10389	55	26	time	time	NOUN
easat-10389	55	27	performance	performance	NOUN
easat-10389	55	28	.	.	PUNCT
easat-10389	56	1	these	these	DET
easat-10389	56	2	changes	change	NOUN
easat-10389	56	3	underscore	underscore	VERB
easat-10389	56	4	the	the	DET
easat-10389	56	5	importance	importance	NOUN
easat-10389	56	6	of	of	ADP
easat-10389	56	7	developing	develop	VERB
easat-10389	56	8	models	model	NOUN
easat-10389	56	9	that	that	PRON
easat-10389	56	10	strike	strike	VERB
easat-10389	56	11	a	a	DET
easat-10389	56	12	balance	balance	NOUN
easat-10389	56	13	between	between	ADP
easat-10389	56	14	accuracy	accuracy	NOUN
easat-10389	56	15	and	and	CCONJ
easat-10389	56	16	computational	computational	ADJ
easat-10389	56	17	efficiency	efficiency	NOUN
easat-10389	56	18	,	,	PUNCT
easat-10389	56	19	particularly	particularly	ADV
easat-10389	56	20	in	in	ADP
easat-10389	56	21	cities	city	NOUN
easat-10389	56	22	where	where	SCONJ
easat-10389	56	23	data	datum	NOUN
easat-10389	56	24	quality	quality	NOUN
easat-10389	56	25	and	and	CCONJ
easat-10389	56	26	availability	availability	NOUN
easat-10389	56	27	can	can	AUX
easat-10389	56	28	vary	vary	VERB
easat-10389	56	29	significantly	significantly	ADV
easat-10389	56	30	.	.	PUNCT
easat-10389	57	1	future	future	ADJ
easat-10389	57	2	research	research	NOUN
easat-10389	57	3	should	should	AUX
easat-10389	57	4	focus	focus	VERB
easat-10389	57	5	on	on	ADP
easat-10389	57	6	creating	create	VERB
easat-10389	57	7	lightweight	lightweight	ADJ
easat-10389	57	8	designs	design	NOUN
easat-10389	57	9	that	that	PRON
easat-10389	57	10	use	use	VERB
easat-10389	57	11	less	less	ADJ
easat-10389	57	12	energy	energy	NOUN
easat-10389	57	13	and	and	CCONJ
easat-10389	57	14	can	can	AUX
easat-10389	57	15	be	be	AUX
easat-10389	57	16	utilized	utilize	VERB
easat-10389	57	17	in	in	ADP
easat-10389	57	18	more	more	ADJ
easat-10389	57	19	locations	location	NOUN
easat-10389	57	20	,	,	PUNCT
easat-10389	57	21	particularly	particularly	ADV
easat-10389	57	22	in	in	ADP
easat-10389	57	23	areas	area	NOUN
easat-10389	57	24	with	with	ADP
easat-10389	57	25	limited	limited	ADJ
easat-10389	57	26	resources	resource	NOUN
easat-10389	57	27	.	.	PUNCT
easat-10389	58	1	258	258	NUM
easat-10389	58	2	edelweiss	edelweiss	PROPN
easat-10389	58	3	applied	apply	VERB
easat-10389	58	4	science	science	NOUN
easat-10389	58	5	and	and	CCONJ
easat-10389	58	6	technology	technology	NOUN
easat-10389	58	7	issn	issn	PROPN
easat-10389	58	8	:	:	PUNCT
easat-10389	58	9	2576	2576	NUM
easat-10389	58	10	-	-	SYM
easat-10389	58	11	8484	8484	NUM
easat-10389	58	12	vol	vol	NOUN
easat-10389	58	13	.	.	PROPN
easat-10389	59	1	9	9	NUM
easat-10389	59	2	,	,	PUNCT
easat-10389	59	3	no	no	INTJ
easat-10389	59	4	.	.	NOUN
easat-10389	59	5	10	10	NUM
easat-10389	59	6	:	:	PUNCT
easat-10389	59	7	256	256	NUM
easat-10389	59	8	-	-	SYM
easat-10389	59	9	271	271	NUM
easat-10389	59	10	,	,	PUNCT
easat-10389	59	11	2025	2025	NUM
easat-10389	59	12	doi	doi	NOUN
easat-10389	59	13	:	:	PUNCT
easat-10389	59	14	10.55214/2576	10.55214/2576	NUM
easat-10389	59	15	-	-	PUNCT
easat-10389	59	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	59	17	©	©	ADP
easat-10389	59	18	2025	2025	NUM
easat-10389	59	19	by	by	ADP
easat-10389	59	20	the	the	DET
easat-10389	59	21	authors	author	NOUN
easat-10389	59	22	;	;	PUNCT
easat-10389	59	23	licensee	licensee	PROPN
easat-10389	59	24	learning	learning	NOUN
easat-10389	59	25	gate	gate	VERB
easat-10389	59	26	2.4	2.4	NUM
easat-10389	59	27	.	.	PUNCT
easat-10389	60	1	important	important	ADJ
easat-10389	60	2	gaps	gap	NOUN
easat-10389	60	3	and	and	CCONJ
easat-10389	60	4	problems	problem	NOUN
easat-10389	60	5	although	although	SCONJ
easat-10389	60	6	there	there	PRON
easat-10389	60	7	have	have	AUX
easat-10389	60	8	been	be	AUX
easat-10389	60	9	notable	notable	ADJ
easat-10389	60	10	improvements	improvement	NOUN
easat-10389	60	11	,	,	PUNCT
easat-10389	60	12	modern	modern	ADJ
easat-10389	60	13	models	model	NOUN
easat-10389	60	14	still	still	ADV
easat-10389	60	15	encounter	encounter	VERB
easat-10389	60	16	several	several	ADJ
easat-10389	60	17	challenges	challenge	NOUN
easat-10389	60	18	.	.	PUNCT
easat-10389	61	1	first	first	ADV
easat-10389	61	2	,	,	PUNCT
easat-10389	61	3	generalizability	generalizability	NOUN
easat-10389	61	4	is	be	AUX
easat-10389	61	5	a	a	DET
easat-10389	61	6	significant	significant	ADJ
easat-10389	61	7	concern	concern	NOUN
easat-10389	61	8	,	,	PUNCT
easat-10389	61	9	as	as	SCONJ
easat-10389	61	10	many	many	ADJ
easat-10389	61	11	prediction	prediction	NOUN
easat-10389	61	12	frameworks	framework	NOUN
easat-10389	61	13	are	be	AUX
easat-10389	61	14	trained	train	VERB
easat-10389	61	15	and	and	CCONJ
easat-10389	61	16	validated	validate	VERB
easat-10389	61	17	only	only	ADV
easat-10389	61	18	on	on	ADP
easat-10389	61	19	data	datum	NOUN
easat-10389	61	20	from	from	ADP
easat-10389	61	21	a	a	DET
easat-10389	61	22	single	single	ADJ
easat-10389	61	23	city	city	NOUN
easat-10389	61	24	or	or	CCONJ
easat-10389	61	25	region	region	NOUN
easat-10389	61	26	,	,	PUNCT
easat-10389	61	27	which	which	PRON
easat-10389	61	28	limits	limit	VERB
easat-10389	61	29	their	their	PRON
easat-10389	61	30	effectiveness	effectiveness	NOUN
easat-10389	61	31	when	when	SCONJ
easat-10389	61	32	applied	apply	VERB
easat-10389	61	33	to	to	ADP
easat-10389	61	34	other	other	ADJ
easat-10389	61	35	areas	area	NOUN
easat-10389	61	36	with	with	ADP
easat-10389	61	37	different	different	ADJ
easat-10389	61	38	environmental	environmental	ADJ
easat-10389	61	39	conditions	condition	NOUN
easat-10389	61	40	[	[	X
easat-10389	61	41	6	6	NUM
easat-10389	61	42	,	,	PUNCT
easat-10389	61	43	9	9	NUM
easat-10389	61	44	]	]	PUNCT
easat-10389	61	45	.	.	PUNCT
easat-10389	62	1	second	second	ADJ
easat-10389	62	2	,	,	PUNCT
easat-10389	62	3	proper	proper	ADJ
easat-10389	62	4	integration	integration	NOUN
easat-10389	62	5	of	of	ADP
easat-10389	62	6	spatial	spatial	ADJ
easat-10389	62	7	and	and	CCONJ
easat-10389	62	8	temporal	temporal	ADJ
easat-10389	62	9	dependencies	dependency	NOUN
easat-10389	62	10	is	be	AUX
easat-10389	62	11	often	often	ADV
easat-10389	62	12	lacking	lack	VERB
easat-10389	62	13	.	.	PUNCT
easat-10389	63	1	many	many	ADJ
easat-10389	63	2	studies	study	NOUN
easat-10389	63	3	have	have	AUX
easat-10389	63	4	shown	show	VERB
easat-10389	63	5	that	that	SCONJ
easat-10389	63	6	models	model	NOUN
easat-10389	63	7	typically	typically	ADV
easat-10389	63	8	analyze	analyze	VERB
easat-10389	63	9	geographical	geographical	ADJ
easat-10389	63	10	and	and	CCONJ
easat-10389	63	11	temporal	temporal	ADJ
easat-10389	63	12	data	datum	NOUN
easat-10389	63	13	separately	separately	ADV
easat-10389	63	14	,	,	PUNCT
easat-10389	63	15	thereby	thereby	ADV
easat-10389	63	16	ignoring	ignore	VERB
easat-10389	63	17	the	the	DET
easat-10389	63	18	complex	complex	ADJ
easat-10389	63	19	interconnections	interconnection	NOUN
easat-10389	63	20	that	that	PRON
easat-10389	63	21	heavily	heavily	ADV
easat-10389	63	22	influence	influence	NOUN
easat-10389	63	23	asthma	asthma	NOUN
easat-10389	63	24	risk	risk	NOUN
easat-10389	64	1	[	[	X
easat-10389	64	2	11	11	NUM
easat-10389	64	3	-	-	SYM
easat-10389	64	4	13	13	NUM
easat-10389	64	5	]	]	PUNCT
easat-10389	64	6	.	.	PUNCT
easat-10389	65	1	third	third	ADJ
easat-10389	65	2	,	,	PUNCT
easat-10389	65	3	interpretability	interpretability	NOUN
easat-10389	65	4	remains	remain	VERB
easat-10389	65	5	a	a	DET
easat-10389	65	6	persistent	persistent	ADJ
easat-10389	65	7	issue	issue	NOUN
easat-10389	65	8	.	.	PUNCT
easat-10389	66	1	although	although	SCONJ
easat-10389	66	2	deep	deep	ADJ
easat-10389	66	3	learning	learning	NOUN
easat-10389	66	4	approaches	approach	NOUN
easat-10389	66	5	have	have	AUX
easat-10389	66	6	demonstrated	demonstrate	VERB
easat-10389	66	7	impressive	impressive	ADJ
easat-10389	66	8	predictive	predictive	ADJ
easat-10389	66	9	power	power	NOUN
easat-10389	66	10	,	,	PUNCT
easat-10389	66	11	their	their	PRON
easat-10389	66	12	“	"	PUNCT
easat-10389	66	13	black	black	ADJ
easat-10389	66	14	box	box	NOUN
easat-10389	66	15	”	"	PUNCT
easat-10389	66	16	nature	nature	NOUN
easat-10389	66	17	hampers	hamper	VERB
easat-10389	66	18	the	the	DET
easat-10389	66	19	ability	ability	NOUN
easat-10389	66	20	to	to	PART
easat-10389	66	21	derive	derive	VERB
easat-10389	66	22	actionable	actionable	ADJ
easat-10389	66	23	insights	insight	NOUN
easat-10389	66	24	for	for	ADP
easat-10389	66	25	clinical	clinical	ADJ
easat-10389	66	26	decision	decision	NOUN
easat-10389	66	27	-	-	PUNCT
easat-10389	66	28	making	making	NOUN
easat-10389	66	29	and	and	CCONJ
easat-10389	66	30	the	the	DET
easat-10389	66	31	design	design	NOUN
easat-10389	66	32	of	of	ADP
easat-10389	66	33	public	public	ADJ
easat-10389	66	34	health	health	NOUN
easat-10389	66	35	policies	policy	NOUN
easat-10389	66	36	[	[	X
easat-10389	66	37	7	7	NUM
easat-10389	66	38	,	,	PUNCT
easat-10389	66	39	8	8	NUM
easat-10389	66	40	]	]	PUNCT
easat-10389	66	41	.	.	PUNCT
easat-10389	67	1	lastly	lastly	ADV
easat-10389	67	2	,	,	PUNCT
easat-10389	67	3	the	the	DET
easat-10389	67	4	practical	practical	ADJ
easat-10389	67	5	implementation	implementation	NOUN
easat-10389	67	6	remains	remain	VERB
easat-10389	67	7	limited	limited	ADJ
easat-10389	67	8	.	.	PUNCT
easat-10389	68	1	most	most	ADJ
easat-10389	68	2	existing	exist	VERB
easat-10389	68	3	research	research	NOUN
easat-10389	68	4	focuses	focus	VERB
easat-10389	68	5	on	on	ADP
easat-10389	68	6	offline	offline	ADJ
easat-10389	68	7	simulations	simulation	NOUN
easat-10389	68	8	and	and	CCONJ
easat-10389	68	9	retrospective	retrospective	ADJ
easat-10389	68	10	validation	validation	NOUN
easat-10389	68	11	,	,	PUNCT
easat-10389	68	12	with	with	ADP
easat-10389	68	13	very	very	ADV
easat-10389	68	14	few	few	ADJ
easat-10389	68	15	efforts	effort	NOUN
easat-10389	68	16	to	to	PART
easat-10389	68	17	incorporate	incorporate	VERB
easat-10389	68	18	these	these	DET
easat-10389	68	19	predictive	predictive	ADJ
easat-10389	68	20	systems	system	NOUN
easat-10389	68	21	into	into	ADP
easat-10389	68	22	real	real	ADJ
easat-10389	68	23	-	-	PUNCT
easat-10389	68	24	time	time	NOUN
easat-10389	68	25	monitoring	monitoring	NOUN
easat-10389	68	26	platforms	platform	NOUN
easat-10389	68	27	or	or	CCONJ
easat-10389	68	28	mobile	mobile	ADJ
easat-10389	68	29	health	health	NOUN
easat-10389	68	30	alert	alert	NOUN
easat-10389	68	31	apps	app	NOUN
easat-10389	68	32	[	[	X
easat-10389	68	33	6	6	NUM
easat-10389	68	34	,	,	PUNCT
easat-10389	68	35	10	10	NUM
easat-10389	68	36	,	,	PUNCT
easat-10389	68	37	16	16	NUM
easat-10389	68	38	]	]	PUNCT
easat-10389	68	39	.	.	PUNCT
easat-10389	69	1	this	this	DET
easat-10389	69	2	gap	gap	NOUN
easat-10389	69	3	restricts	restrict	VERB
easat-10389	69	4	the	the	DET
easat-10389	69	5	immediate	immediate	ADJ
easat-10389	69	6	use	use	NOUN
easat-10389	69	7	of	of	ADP
easat-10389	69	8	advanced	advanced	ADJ
easat-10389	69	9	ai	ai	ADJ
easat-10389	69	10	-	-	PUNCT
easat-10389	69	11	driven	drive	VERB
easat-10389	69	12	models	model	NOUN
easat-10389	69	13	in	in	ADP
easat-10389	69	14	large	large	ADJ
easat-10389	69	15	-	-	PUNCT
easat-10389	69	16	scale	scale	NOUN
easat-10389	69	17	public	public	ADJ
easat-10389	69	18	health	health	NOUN
easat-10389	69	19	surveillance	surveillance	NOUN
easat-10389	69	20	.	.	PUNCT
easat-10389	70	1	3	3	X
easat-10389	70	2	.	.	X
easat-10389	70	3	materials	material	NOUN
easat-10389	70	4	and	and	CCONJ
easat-10389	70	5	methods	method	NOUN
easat-10389	70	6	this	this	DET
easat-10389	70	7	work	work	NOUN
easat-10389	70	8	introduces	introduce	VERB
easat-10389	70	9	a	a	DET
easat-10389	70	10	dual	dual	ADJ
easat-10389	70	11	-	-	PUNCT
easat-10389	70	12	branch	branch	NOUN
easat-10389	70	13	deep	deep	ADJ
easat-10389	70	14	learning	learning	NOUN
easat-10389	70	15	model	model	NOUN
easat-10389	70	16	trained	train	VERB
easat-10389	70	17	on	on	ADP
easat-10389	70	18	a	a	DET
easat-10389	70	19	diverse	diverse	ADJ
easat-10389	70	20	range	range	NOUN
easat-10389	70	21	of	of	ADP
easat-10389	70	22	organized	organized	ADJ
easat-10389	70	23	and	and	CCONJ
easat-10389	70	24	unstructured	unstructured	ADJ
easat-10389	70	25	data	datum	NOUN
easat-10389	70	26	.	.	PUNCT
easat-10389	71	1	one	one	NUM
easat-10389	71	2	branch	branch	NOUN
easat-10389	71	3	uses	use	VERB
easat-10389	71	4	rnns	rnn	NOUN
easat-10389	71	5	for	for	ADP
easat-10389	71	6	sequential	sequential	ADJ
easat-10389	71	7	data	datum	NOUN
easat-10389	71	8	,	,	PUNCT
easat-10389	71	9	while	while	SCONJ
easat-10389	71	10	the	the	DET
easat-10389	71	11	other	other	ADJ
easat-10389	71	12	employs	employ	VERB
easat-10389	71	13	cnns	cnn	NOUN
easat-10389	71	14	for	for	ADP
easat-10389	71	15	spatial	spatial	ADJ
easat-10389	71	16	data	datum	NOUN
easat-10389	71	17	.	.	PUNCT
easat-10389	72	1	the	the	DET
easat-10389	72	2	two	two	NUM
easat-10389	72	3	branches	branch	NOUN
easat-10389	72	4	merge	merge	VERB
easat-10389	72	5	in	in	ADP
easat-10389	72	6	a	a	DET
easat-10389	72	7	fully	fully	ADV
easat-10389	72	8	integrated	integrate	VERB
easat-10389	72	9	layer	layer	NOUN
easat-10389	72	10	,	,	PUNCT
easat-10389	72	11	offering	offer	VERB
easat-10389	72	12	enhanced	enhanced	ADJ
easat-10389	72	13	learning	learn	VERB
easat-10389	72	14	capabilities	capability	NOUN
easat-10389	72	15	.	.	PUNCT
easat-10389	73	1	the	the	DET
easat-10389	73	2	model	model	NOUN
easat-10389	73	3	was	be	AUX
easat-10389	73	4	trained	train	VERB
easat-10389	73	5	using	use	VERB
easat-10389	73	6	supervised	supervised	ADJ
easat-10389	73	7	learning	learning	NOUN
easat-10389	73	8	,	,	PUNCT
easat-10389	73	9	with	with	ADP
easat-10389	73	10	data	datum	NOUN
easat-10389	73	11	divided	divide	VERB
easat-10389	73	12	into	into	ADP
easat-10389	73	13	three	three	NUM
easat-10389	73	14	sets	set	NOUN
easat-10389	73	15	:	:	PUNCT
easat-10389	73	16	training	training	NOUN
easat-10389	73	17	,	,	PUNCT
easat-10389	73	18	validation	validation	NOUN
easat-10389	73	19	,	,	PUNCT
easat-10389	73	20	and	and	CCONJ
easat-10389	73	21	testing	testing	NOUN
easat-10389	73	22	.	.	PUNCT
easat-10389	74	1	dropout	dropout	NOUN
easat-10389	74	2	,	,	PUNCT
easat-10389	74	3	weight	weight	NOUN
easat-10389	74	4	decay	decay	NOUN
easat-10389	74	5	,	,	PUNCT
easat-10389	74	6	early	early	ADJ
easat-10389	74	7	stopping	stopping	NOUN
easat-10389	74	8	,	,	PUNCT
easat-10389	74	9	and	and	CCONJ
easat-10389	74	10	data	datum	NOUN
easat-10389	74	11	augmentation	augmentation	NOUN
easat-10389	74	12	are	be	AUX
easat-10389	74	13	among	among	ADP
easat-10389	74	14	the	the	DET
easat-10389	74	15	techniques	technique	NOUN
easat-10389	74	16	used	use	VERB
easat-10389	74	17	to	to	PART
easat-10389	74	18	prevent	prevent	VERB
easat-10389	74	19	overfitting	overfitting	NOUN
easat-10389	74	20	.	.	PUNCT
easat-10389	75	1	the	the	DET
easat-10389	75	2	adam	adam	PROPN
easat-10389	75	3	optimizer	optimizer	NOUN
easat-10389	75	4	,	,	PUNCT
easat-10389	75	5	with	with	ADP
easat-10389	75	6	a	a	DET
easat-10389	75	7	carefully	carefully	ADV
easat-10389	75	8	set	set	VERB
easat-10389	75	9	learning	learning	NOUN
easat-10389	75	10	rate	rate	NOUN
easat-10389	75	11	,	,	PUNCT
easat-10389	75	12	maintained	maintain	VERB
easat-10389	75	13	stability	stability	NOUN
easat-10389	75	14	.	.	PUNCT
easat-10389	76	1	we	we	PRON
easat-10389	76	2	evaluated	evaluate	VERB
easat-10389	76	3	performance	performance	NOUN
easat-10389	76	4	using	use	VERB
easat-10389	76	5	metrics	metric	NOUN
easat-10389	76	6	such	such	ADJ
easat-10389	76	7	as	as	ADP
easat-10389	76	8	accuracy	accuracy	NOUN
easat-10389	76	9	,	,	PUNCT
easat-10389	76	10	precision	precision	NOUN
easat-10389	76	11	,	,	PUNCT
easat-10389	76	12	recall	recall	NOUN
easat-10389	76	13	,	,	PUNCT
easat-10389	76	14	f1	f1	NOUN
easat-10389	76	15	score	score	NOUN
easat-10389	76	16	,	,	PUNCT
easat-10389	76	17	and	and	CCONJ
easat-10389	76	18	auc	auc	VERB
easat-10389	76	19	across	across	ADP
easat-10389	76	20	different	different	ADJ
easat-10389	76	21	subsets	subset	NOUN
easat-10389	76	22	to	to	PART
easat-10389	76	23	ensure	ensure	VERB
easat-10389	76	24	relevance	relevance	NOUN
easat-10389	76	25	to	to	ADP
easat-10389	76	26	real	real	ADJ
easat-10389	76	27	-	-	PUNCT
easat-10389	76	28	world	world	NOUN
easat-10389	76	29	applications	application	NOUN
easat-10389	76	30	.	.	PUNCT
easat-10389	77	1	3.1	3.1	NUM
easat-10389	77	2	.	.	PUNCT
easat-10389	78	1	the	the	DET
easat-10389	78	2	structure	structure	NOUN
easat-10389	78	3	of	of	ADP
easat-10389	78	4	the	the	DET
easat-10389	78	5	model	model	NOUN
easat-10389	78	6	the	the	DET
easat-10389	78	7	proposed	propose	VERB
easat-10389	78	8	framework	framework	NOUN
easat-10389	78	9	employs	employ	VERB
easat-10389	78	10	a	a	DET
easat-10389	78	11	dual	dual	ADJ
easat-10389	78	12	-	-	PUNCT
easat-10389	78	13	branch	branch	NOUN
easat-10389	78	14	hybrid	hybrid	ADJ
easat-10389	78	15	architecture	architecture	NOUN
easat-10389	78	16	to	to	PART
easat-10389	78	17	capture	capture	VERB
easat-10389	78	18	both	both	DET
easat-10389	78	19	evolving	evolve	VERB
easat-10389	78	20	pollution	pollution	NOUN
easat-10389	78	21	patterns	pattern	NOUN
easat-10389	78	22	and	and	CCONJ
easat-10389	78	23	spatial	spatial	ADJ
easat-10389	78	24	patterns	pattern	NOUN
easat-10389	78	25	,	,	PUNCT
easat-10389	78	26	as	as	ADV
easat-10389	78	27	well	well	ADV
easat-10389	78	28	as	as	ADP
easat-10389	78	29	changing	change	VERB
easat-10389	78	30	weather	weather	NOUN
easat-10389	78	31	patterns	pattern	NOUN
easat-10389	78	32	over	over	ADP
easat-10389	78	33	time	time	NOUN
easat-10389	78	34	.	.	PUNCT
easat-10389	79	1	•	•	NUM
easat-10389	79	2	the	the	DET
easat-10389	79	3	first	first	ADJ
easat-10389	79	4	branch	branch	NOUN
easat-10389	79	5	processes	process	VERB
easat-10389	79	6	air	air	NOUN
easat-10389	79	7	pollutant	pollutant	ADJ
easat-10389	79	8	sequences	sequence	NOUN
easat-10389	79	9	with	with	ADP
easat-10389	79	10	a	a	DET
easat-10389	79	11	convolutional	convolutional	ADJ
easat-10389	79	12	neural	neural	ADJ
easat-10389	79	13	network	network	NOUN
easat-10389	79	14	(	(	PUNCT
easat-10389	79	15	cnn	cnn	PROPN
easat-10389	79	16	)	)	PUNCT
easat-10389	79	17	.	.	PUNCT
easat-10389	80	1	it	it	PRON
easat-10389	80	2	identifies	identify	VERB
easat-10389	80	3	localized	localized	ADJ
easat-10389	80	4	spatial	spatial	ADJ
easat-10389	80	5	-	-	PUNCT
easat-10389	80	6	temporal	temporal	ADJ
easat-10389	80	7	patterns	pattern	NOUN
easat-10389	80	8	and	and	CCONJ
easat-10389	80	9	short	short	ADJ
easat-10389	80	10	-	-	PUNCT
easat-10389	80	11	term	term	NOUN
easat-10389	80	12	trends	trend	NOUN
easat-10389	80	13	in	in	ADP
easat-10389	80	14	air	air	NOUN
easat-10389	80	15	quality	quality	NOUN
easat-10389	80	16	.	.	PUNCT
easat-10389	81	1	•	•	NUM
easat-10389	81	2	the	the	DET
easat-10389	81	3	second	second	ADJ
easat-10389	81	4	branch	branch	NOUN
easat-10389	81	5	uses	use	VERB
easat-10389	81	6	a	a	DET
easat-10389	81	7	gated	gate	VERB
easat-10389	81	8	recurrent	recurrent	ADJ
easat-10389	81	9	unit	unit	NOUN
easat-10389	81	10	(	(	PUNCT
easat-10389	81	11	gru	gru	NOUN
easat-10389	81	12	)	)	PUNCT
easat-10389	81	13	network	network	NOUN
easat-10389	81	14	to	to	PART
easat-10389	81	15	simulate	simulate	VERB
easat-10389	81	16	weather	weather	NOUN
easat-10389	81	17	sequences	sequence	NOUN
easat-10389	81	18	that	that	PRON
easat-10389	81	19	change	change	VERB
easat-10389	81	20	over	over	ADP
easat-10389	81	21	time	time	NOUN
easat-10389	81	22	.	.	PUNCT
easat-10389	82	1	it	it	PRON
easat-10389	82	2	does	do	VERB
easat-10389	82	3	this	this	PRON
easat-10389	82	4	by	by	ADP
easat-10389	82	5	capturing	capture	VERB
easat-10389	82	6	long	long	ADJ
easat-10389	82	7	-	-	PUNCT
easat-10389	82	8	term	term	NOUN
easat-10389	82	9	dependencies	dependency	NOUN
easat-10389	82	10	and	and	CCONJ
easat-10389	82	11	seasonal	seasonal	ADJ
easat-10389	82	12	changes	change	NOUN
easat-10389	82	13	in	in	ADP
easat-10389	82	14	weather	weather	NOUN
easat-10389	82	15	that	that	PRON
easat-10389	82	16	impact	impact	VERB
easat-10389	82	17	respiratory	respiratory	ADJ
easat-10389	82	18	health	health	NOUN
easat-10389	82	19	.	.	PUNCT
easat-10389	83	1	•	•	NUM
easat-10389	83	2	the	the	DET
easat-10389	83	3	third	third	ADJ
easat-10389	83	4	attention	attention	NOUN
easat-10389	83	5	method	method	NOUN
easat-10389	83	6	is	be	AUX
easat-10389	83	7	applied	apply	VERB
easat-10389	83	8	to	to	ADP
easat-10389	83	9	the	the	DET
easat-10389	83	10	combined	combine	VERB
easat-10389	83	11	outputs	output	NOUN
easat-10389	83	12	of	of	ADP
easat-10389	83	13	both	both	DET
easat-10389	83	14	branches	branch	NOUN
easat-10389	83	15	,	,	PUNCT
easat-10389	83	16	assigning	assign	VERB
easat-10389	83	17	adaptive	adaptive	ADJ
easat-10389	83	18	weights	weight	NOUN
easat-10389	83	19	to	to	ADP
easat-10389	83	20	the	the	DET
easat-10389	83	21	most	most	ADV
easat-10389	83	22	critical	critical	ADJ
easat-10389	83	23	information	information	NOUN
easat-10389	83	24	for	for	ADP
easat-10389	83	25	predicting	predict	VERB
easat-10389	83	26	asthma	asthma	NOUN
easat-10389	83	27	.	.	PUNCT
easat-10389	84	1	•	•	NUM
easat-10389	84	2	the	the	DET
easat-10389	84	3	final	final	ADJ
easat-10389	84	4	fused	fuse	VERB
easat-10389	84	5	representation	representation	NOUN
easat-10389	84	6	is	be	AUX
easat-10389	84	7	processed	process	VERB
easat-10389	84	8	through	through	ADP
easat-10389	84	9	two	two	NUM
easat-10389	84	10	fully	fully	ADV
easat-10389	84	11	connected	connect	VERB
easat-10389	84	12	layers	layer	NOUN
easat-10389	84	13	and	and	CCONJ
easat-10389	84	14	a	a	DET
easat-10389	84	15	sigmoid	sigmoid	NOUN
easat-10389	84	16	output	output	NOUN
easat-10389	84	17	layer	layer	NOUN
easat-10389	84	18	to	to	PART
easat-10389	84	19	make	make	VERB
easat-10389	84	20	a	a	DET
easat-10389	84	21	prediction	prediction	NOUN
easat-10389	84	22	.	.	PUNCT
easat-10389	85	1	this	this	DET
easat-10389	85	2	lightweight	lightweight	ADJ
easat-10389	85	3	structure	structure	NOUN
easat-10389	85	4	can	can	AUX
easat-10389	85	5	perform	perform	VERB
easat-10389	85	6	real	real	ADJ
easat-10389	85	7	-	-	PUNCT
easat-10389	85	8	time	time	NOUN
easat-10389	85	9	inference	inference	NOUN
easat-10389	85	10	while	while	SCONJ
easat-10389	85	11	maintaining	maintain	VERB
easat-10389	85	12	a	a	DET
easat-10389	85	13	good	good	ADJ
easat-10389	85	14	model	model	NOUN
easat-10389	85	15	fit	fit	PROPN
easat-10389	85	16	.	.	PUNCT
easat-10389	86	1	the	the	DET
easat-10389	86	2	dual	dual	ADJ
easat-10389	86	3	-	-	PUNCT
easat-10389	86	4	branch	branch	NOUN
easat-10389	86	5	architecture	architecture	NOUN
easat-10389	86	6	is	be	AUX
easat-10389	86	7	mathematically	mathematically	ADV
easat-10389	86	8	defined	define	VERB
easat-10389	86	9	to	to	PART
easat-10389	86	10	facilitate	facilitate	VERB
easat-10389	86	11	easier	easy	ADJ
easat-10389	86	12	understanding	understanding	NOUN
easat-10389	86	13	and	and	CCONJ
easat-10389	86	14	comprehension	comprehension	NOUN
easat-10389	86	15	.	.	PUNCT
easat-10389	87	1	it	it	PRON
easat-10389	87	2	explains	explain	VERB
easat-10389	87	3	the	the	DET
easat-10389	87	4	roles	role	NOUN
easat-10389	87	5	of	of	ADP
easat-10389	87	6	the	the	DET
easat-10389	87	7	convolutional	convolutional	ADJ
easat-10389	87	8	encoder	encoder	NOUN
easat-10389	87	9	,	,	PUNCT
easat-10389	87	10	recurrent	recurrent	ADJ
easat-10389	87	11	sequence	sequence	NOUN
easat-10389	87	12	processor	processor	NOUN
easat-10389	87	13	,	,	PUNCT
easat-10389	87	14	attention	attention	NOUN
easat-10389	87	15	mechanism	mechanism	NOUN
easat-10389	87	16	,	,	PUNCT
easat-10389	87	17	and	and	CCONJ
easat-10389	87	18	classification	classification	NOUN
easat-10389	87	19	head	head	NOUN
easat-10389	87	20	,	,	PUNCT
easat-10389	87	21	and	and	CCONJ
easat-10389	87	22	demonstrates	demonstrate	VERB
easat-10389	87	23	how	how	SCONJ
easat-10389	87	24	various	various	ADJ
easat-10389	87	25	environmental	environmental	ADJ
easat-10389	87	26	factors	factor	NOUN
easat-10389	87	27	influence	influence	VERB
easat-10389	87	28	the	the	DET
easat-10389	87	29	prediction	prediction	NOUN
easat-10389	87	30	of	of	ADP
easat-10389	87	31	asthma	asthma	NOUN
easat-10389	87	32	risk	risk	NOUN
easat-10389	87	33	.	.	PUNCT
easat-10389	88	1	the	the	DET
easat-10389	88	2	cnn	cnn	PROPN
easat-10389	88	3	branch	branch	NOUN
easat-10389	88	4	finds	find	VERB
easat-10389	88	5	localized	localized	ADJ
easat-10389	88	6	patterns	pattern	NOUN
easat-10389	88	7	in	in	ADP
easat-10389	88	8	aqi	aqi	PROPN
easat-10389	88	9	time	time	NOUN
easat-10389	88	10	series	series	NOUN
easat-10389	88	11	using	use	VERB
easat-10389	88	12	1d	1d	NUM
easat-10389	88	13	convolution	convolution	NOUN
easat-10389	88	14	(	(	PUNCT
easat-10389	88	15	eq	eq	NOUN
easat-10389	88	16	.	.	PROPN
easat-10389	88	17	1	1	NUM
easat-10389	88	18	)	)	PUNCT
easat-10389	88	19	.	.	PUNCT
easat-10389	89	1	ℎ𝑡	ℎ𝑡	NOUN
easat-10389	89	2	(	(	PUNCT
easat-10389	89	3	𝐶𝑁𝑁	𝐶𝑁𝑁	PROPN
easat-10389	89	4	)	)	PUNCT
easat-10389	89	5	=	=	PUNCT
easat-10389	90	1	𝑅𝑒𝐿𝑈(𝑊𝑐𝑥𝑡:𝑡+𝑘−1	𝑅𝑒𝐿𝑈(𝑊𝑐𝑥𝑡:𝑡+𝑘−1	PROPN
easat-10389	90	2	+	+	CCONJ
easat-10389	90	3	𝑏𝑐	𝑏𝑐	NOUN
easat-10389	90	4	)	)	PUNCT
easat-10389	90	5	(	(	PUNCT
easat-10389	90	6	1	1	X
easat-10389	90	7	)	)	PUNCT
easat-10389	91	1	where	where	SCONJ
easat-10389	91	2	:	:	PUNCT
easat-10389	91	3	𝑥𝑡:𝑡+𝑘−1	𝑥𝑡:𝑡+𝑘−1	X
easat-10389	91	4	is	be	AUX
easat-10389	91	5	a	a	DET
easat-10389	91	6	local	local	ADJ
easat-10389	91	7	aqi	aqi	ADJ
easat-10389	91	8	segment	segment	NOUN
easat-10389	91	9	,	,	PUNCT
easat-10389	91	10	𝑊𝑐	𝑊𝑐	PROPN
easat-10389	91	11	is	be	AUX
easat-10389	91	12	the	the	DET
easat-10389	91	13	convolution	convolution	NOUN
easat-10389	91	14	kernel	kernel	NOUN
easat-10389	91	15	,	,	PUNCT
easat-10389	91	16	and	and	CCONJ
easat-10389	91	17	*	*	PUNCT
easat-10389	91	18	denotes	denote	NOUN
easat-10389	91	19	1d	1d	NUM
easat-10389	91	20	convolution	convolution	NOUN
easat-10389	91	21	.	.	PUNCT
easat-10389	92	1	the	the	DET
easat-10389	92	2	gru	gru	PROPN
easat-10389	92	3	branch	branch	NOUN
easat-10389	92	4	models	model	VERB
easat-10389	92	5	sequential	sequential	ADJ
easat-10389	92	6	dependencies	dependency	NOUN
easat-10389	92	7	in	in	ADP
easat-10389	92	8	meteorological	meteorological	ADJ
easat-10389	92	9	variables	variable	NOUN
easat-10389	92	10	(	(	PUNCT
easat-10389	92	11	eqs	eqs	X
easat-10389	92	12	.	.	PROPN
easat-10389	92	13	2	2	NUM
easat-10389	92	14	-	-	SYM
easat-10389	92	15	5	5	NUM
easat-10389	92	16	)	)	PUNCT
easat-10389	92	17	.	.	PUNCT
easat-10389	93	1	𝑧𝑡	𝑧𝑡	X
easat-10389	93	2	=	=	SYM
easat-10389	93	3	𝛿(𝑊𝑧𝑥𝑡	𝛿(𝑊𝑧𝑥𝑡	PROPN
easat-10389	94	1	+	+	NUM
easat-10389	94	2	𝑈𝑧ℎ𝑡−1	𝑈𝑧ℎ𝑡−1	PRON
easat-10389	94	3	+	+	X
easat-10389	94	4	𝑏𝑧	𝑏𝑧	ADJ
easat-10389	94	5	(	(	PUNCT
easat-10389	94	6	2	2	NUM
easat-10389	94	7	)	)	PUNCT
easat-10389	94	8	𝑟𝑡	𝑟𝑡	PROPN
easat-10389	94	9	=	=	PUNCT
easat-10389	94	10	𝛿(𝑊𝑟𝑥𝑡	𝛿(𝑊𝑟𝑥𝑡	PROPN
easat-10389	95	1	+	+	CCONJ
easat-10389	95	2	𝑈𝑟ℎ𝑡−1	𝑈𝑟ℎ𝑡−1	X
easat-10389	95	3	+	+	CCONJ
easat-10389	95	4	𝑏𝑟	𝑏𝑟	X
easat-10389	95	5	(	(	PUNCT
easat-10389	95	6	3	3	NUM
easat-10389	95	7	)	)	PUNCT
easat-10389	95	8	259	259	NUM
easat-10389	95	9	edelweiss	edelweiss	PROPN
easat-10389	95	10	applied	apply	VERB
easat-10389	95	11	science	science	NOUN
easat-10389	95	12	and	and	CCONJ
easat-10389	95	13	technology	technology	NOUN
easat-10389	95	14	issn	issn	PROPN
easat-10389	95	15	:	:	PUNCT
easat-10389	95	16	2576	2576	NUM
easat-10389	95	17	-	-	SYM
easat-10389	95	18	8484	8484	NUM
easat-10389	95	19	vol	vol	NOUN
easat-10389	95	20	.	.	PROPN
easat-10389	96	1	9	9	NUM
easat-10389	96	2	,	,	PUNCT
easat-10389	96	3	no	no	INTJ
easat-10389	96	4	.	.	NOUN
easat-10389	96	5	10	10	NUM
easat-10389	96	6	:	:	PUNCT
easat-10389	96	7	256	256	NUM
easat-10389	96	8	-	-	SYM
easat-10389	96	9	271	271	NUM
easat-10389	96	10	,	,	PUNCT
easat-10389	96	11	2025	2025	NUM
easat-10389	96	12	doi	doi	NOUN
easat-10389	96	13	:	:	PUNCT
easat-10389	96	14	10.55214/2576	10.55214/2576	NUM
easat-10389	96	15	-	-	PUNCT
easat-10389	96	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	96	17	©	©	ADP
easat-10389	96	18	2025	2025	NUM
easat-10389	96	19	by	by	ADP
easat-10389	96	20	the	the	DET
easat-10389	96	21	authors	author	NOUN
easat-10389	96	22	;	;	PUNCT
easat-10389	96	23	licensee	licensee	PROPN
easat-10389	96	24	learning	learn	VERB
easat-10389	96	25	gate	gate	PROPN
easat-10389	96	26	ℎ	ℎ	PROPN
easat-10389	96	27	�	�	PROPN
easat-10389	96	28	̃	̃	PROPN
easat-10389	96	29	�	�	NOUN
easat-10389	96	30	=	=	SYM
easat-10389	96	31	𝑡𝑎𝑛ℎ(𝑊ℎ𝑥𝑡	𝑡𝑎𝑛ℎ(𝑊ℎ𝑥𝑡	PROPN
easat-10389	96	32	+	+	CCONJ
easat-10389	96	33	𝑈ℎ(𝑟𝑡⨀ℎ𝑡−1	𝑈ℎ(𝑟𝑡⨀ℎ𝑡−1	PROPN
easat-10389	96	34	+	+	CCONJ
easat-10389	96	35	𝑏ℎ	𝑏ℎ	NOUN
easat-10389	96	36	(	(	PUNCT
easat-10389	96	37	4	4	NUM
easat-10389	96	38	)	)	PUNCT
easat-10389	96	39	ℎ𝑡	ℎ𝑡	NOUN
easat-10389	96	40	=	=	SYM
easat-10389	96	41	(	(	PUNCT
easat-10389	96	42	1	1	NUM
easat-10389	96	43	−	−	NOUN
easat-10389	96	44	𝑧𝑡)⨀ℎ𝑡−1	𝑧𝑡)⨀ℎ𝑡−1	VERB
easat-10389	97	1	+	+	CCONJ
easat-10389	97	2	𝑧𝑡⨀ℎ	𝑧𝑡⨀ℎ	PROPN
easat-10389	97	3	�	�	PROPN
easat-10389	97	4	̃	̃	PROPN
easat-10389	97	5	�	�	PROPN
easat-10389	97	6	(	(	PUNCT
easat-10389	97	7	5	5	NUM
easat-10389	97	8	)	)	PUNCT
easat-10389	97	9	where	where	SCONJ
easat-10389	97	10	:	:	PUNCT
easat-10389	97	11	𝑥𝑡is	𝑥𝑡is	VERB
easat-10389	97	12	the	the	DET
easat-10389	97	13	input	input	NOUN
easat-10389	97	14	vector	vector	NOUN
easat-10389	97	15	at	at	ADP
easat-10389	97	16	time	time	NOUN
easat-10389	97	17	step	step	NOUN
easat-10389	97	18	t	t	PROPN
easat-10389	97	19	;	;	PUNCT
easat-10389	97	20	ℎ𝑡−1	ℎ𝑡−1	VERB
easat-10389	97	21	ℎ𝑡	ℎ𝑡	NOUN
easat-10389	97	22	are	be	AUX
easat-10389	97	23	hidden	hide	VERB
easat-10389	97	24	states	state	NOUN
easat-10389	97	25	at	at	ADP
easat-10389	97	26	time	time	NOUN
easat-10389	97	27	steps	step	VERB
easat-10389	97	28	t-1	t-1	PROPN
easat-10389	97	29	and	and	CCONJ
easat-10389	97	30	t	t	PROPN
easat-10389	97	31	,	,	PUNCT
easat-10389	97	32	respectively	respectively	ADV
easat-10389	97	33	;	;	PUNCT
easat-10389	97	34	𝑧𝑡	𝑧𝑡	X
easat-10389	97	35	,	,	PUNCT
easat-10389	97	36	𝑟𝑡	𝑟𝑡	PROPN
easat-10389	97	37	are	be	AUX
easat-10389	97	38	update	update	NOUN
easat-10389	97	39	gate	gate	NOUN
easat-10389	97	40	and	and	CCONJ
easat-10389	97	41	reset	reset	ADJ
easat-10389	97	42	gate	gate	NOUN
easat-10389	97	43	vectors	vector	NOUN
easat-10389	97	44	;	;	PUNCT
easat-10389	97	45	ℎ	ℎ	PROPN
easat-10389	97	46	�	�	PROPN
easat-10389	97	47	̃	̃	PROPN
easat-10389	97	48	�	�	NOUN
easat-10389	97	49	is	be	AUX
easat-10389	97	50	the	the	DET
easat-10389	97	51	candidate	candidate	NOUN
easat-10389	97	52	hidden	hide	VERB
easat-10389	97	53	state	state	NOUN
easat-10389	97	54	;	;	PUNCT
easat-10389	97	55	w	w	X
easat-10389	97	56	,	,	PUNCT
easat-10389	97	57	u	u	PROPN
easat-10389	97	58	,	,	PUNCT
easat-10389	97	59	b	b	PROPN
easat-10389	97	60	are	be	AUX
easat-10389	97	61	weight	weight	NOUN
easat-10389	97	62	matrices	matrix	NOUN
easat-10389	97	63	and	and	CCONJ
easat-10389	97	64	bias	bias	NOUN
easat-10389	97	65	vectors	vector	NOUN
easat-10389	97	66	for	for	ADP
easat-10389	97	67	respective	respective	ADJ
easat-10389	97	68	gates	gate	NOUN
easat-10389	97	69	and	and	CCONJ
easat-10389	97	70	activations	activation	NOUN
easat-10389	97	71	;	;	PUNCT
easat-10389	97	72	𝛿	𝛿	PRON
easat-10389	97	73	is	be	AUX
easat-10389	97	74	the	the	DET
easat-10389	97	75	sigmoid	sigmoid	NOUN
easat-10389	97	76	activation	activation	NOUN
easat-10389	97	77	function	function	NOUN
easat-10389	97	78	;	;	PUNCT
easat-10389	97	79	tank	tank	NOUN
easat-10389	97	80	is	be	AUX
easat-10389	97	81	the	the	DET
easat-10389	97	82	hyperbolic	hyperbolic	ADJ
easat-10389	97	83	tangent	tangent	NOUN
easat-10389	97	84	activation	activation	NOUN
easat-10389	97	85	function	function	NOUN
easat-10389	97	86	;	;	PUNCT
easat-10389	97	87	⨀	⨀	NOUN
easat-10389	97	88	is	be	AUX
easat-10389	97	89	element	element	ADJ
easat-10389	97	90	-	-	ADJ
easat-10389	97	91	wise	wise	ADJ
easat-10389	97	92	(	(	PUNCT
easat-10389	97	93	hadamard	hadamard	ADJ
easat-10389	97	94	)	)	PUNCT
easat-10389	97	95	multiplication	multiplication	NOUN
easat-10389	97	96	.	.	PUNCT
easat-10389	98	1	the	the	DET
easat-10389	98	2	attention	attention	NOUN
easat-10389	98	3	mechanism	mechanism	NOUN
easat-10389	98	4	assigns	assign	NOUN
easat-10389	98	5	adaptive	adaptive	ADJ
easat-10389	98	6	weights	weight	NOUN
easat-10389	98	7	to	to	ADP
easat-10389	98	8	features	feature	NOUN
easat-10389	98	9	from	from	ADP
easat-10389	98	10	both	both	DET
easat-10389	98	11	branches	branch	NOUN
easat-10389	98	12	(	(	PUNCT
easat-10389	98	13	eq.6	eq.6	PROPN
easat-10389	98	14	)	)	PUNCT
easat-10389	98	15	.	.	PUNCT
easat-10389	99	1	𝑒𝑖	𝑒𝑖	X
easat-10389	99	2	=	=	NOUN
easat-10389	99	3	tanh(𝑊𝑒ℎ𝑒	tanh(𝑊𝑒ℎ𝑒	PROPN
easat-10389	99	4	+	+	CCONJ
easat-10389	99	5	𝑏𝑒	𝑏𝑒	NOUN
easat-10389	99	6	)	)	PUNCT
easat-10389	99	7	;	;	PUNCT
easat-10389	99	8	𝛼𝑖	𝛼𝑖	PROPN
easat-10389	99	9	=	=	SYM
easat-10389	99	10	exp	exp	X
easat-10389	99	11	(	(	PUNCT
easat-10389	99	12	𝑒𝑖	𝑒𝑖	PROPN
easat-10389	99	13	)	)	PUNCT
easat-10389	99	14	∑	∑	NOUN
easat-10389	99	15	exp	exp	X
easat-10389	99	16	(	(	PUNCT
easat-10389	99	17	𝑒𝑖)𝑗	𝑒𝑖)𝑗	PROPN
easat-10389	99	18	;	;	PUNCT
easat-10389	99	19	𝑐	𝑐	PROPN
easat-10389	99	20	=	=	SYM
easat-10389	99	21	∑	∑	PUNCT
easat-10389	99	22	𝛼𝑖ℎ𝑖𝑖	𝛼𝑖ℎ𝑖𝑖	NOUN
easat-10389	99	23	(	(	PUNCT
easat-10389	99	24	6	6	NUM
easat-10389	99	25	)	)	PUNCT
easat-10389	99	26	where	where	SCONJ
easat-10389	99	27	:	:	PUNCT
easat-10389	99	28	.	.	PUNCT
easat-10389	100	1	𝛼𝑖denotes	𝛼𝑖denote	NOUN
easat-10389	100	2	attention	attention	NOUN
easat-10389	100	3	weight	weight	NOUN
easat-10389	100	4	and	and	CCONJ
easat-10389	100	5	c	c	NOUN
easat-10389	100	6	is	be	AUX
easat-10389	100	7	the	the	DET
easat-10389	100	8	final	final	ADJ
easat-10389	100	9	fused	fuse	VERB
easat-10389	100	10	context	context	NOUN
easat-10389	100	11	vector	vector	NOUN
easat-10389	100	12	.	.	PUNCT
easat-10389	101	1	the	the	DET
easat-10389	101	2	output	output	NOUN
easat-10389	101	3	layer	layer	NOUN
easat-10389	101	4	generates	generate	VERB
easat-10389	101	5	the	the	DET
easat-10389	101	6	probability	probability	NOUN
easat-10389	101	7	of	of	ADP
easat-10389	101	8	asthma	asthma	NOUN
easat-10389	101	9	risk	risk	NOUN
easat-10389	101	10	from	from	ADP
easat-10389	101	11	the	the	DET
easat-10389	101	12	combined	combine	VERB
easat-10389	101	13	representation	representation	NOUN
easat-10389	101	14	(	(	PUNCT
easat-10389	101	15	eq.7	eq.7	PROPN
easat-10389	101	16	)	)	PUNCT
easat-10389	101	17	.	.	PUNCT
easat-10389	102	1	�	�	PROPN
easat-10389	102	2	̂	̂	VERB
easat-10389	102	3	�	�	NOUN
easat-10389	102	4	=	=	SYM
easat-10389	102	5	𝛿(𝑊0𝑐	𝛿(𝑊0𝑐	PROPN
easat-10389	102	6	+	+	CCONJ
easat-10389	102	7	𝑏0	𝑏0	NOUN
easat-10389	102	8	)	)	PUNCT
easat-10389	102	9	(	(	PUNCT
easat-10389	102	10	7	7	X
easat-10389	102	11	)	)	PUNCT
easat-10389	102	12	a	a	DET
easat-10389	102	13	binary	binary	ADJ
easat-10389	102	14	cross	cross	NOUN
easat-10389	102	15	-	-	ADJ
easat-10389	102	16	entropy	entropy	ADJ
easat-10389	102	17	loss	loss	NOUN
easat-10389	102	18	function	function	NOUN
easat-10389	102	19	to	to	PART
easat-10389	102	20	train	train	VERB
easat-10389	102	21	the	the	DET
easat-10389	102	22	model	model	NOUN
easat-10389	102	23	,	,	PUNCT
easat-10389	102	24	which	which	PRON
easat-10389	102	25	is	be	AUX
easat-10389	102	26	perfect	perfect	ADJ
easat-10389	102	27	for	for	ADP
easat-10389	102	28	classifying	classify	VERB
easat-10389	102	29	health	health	NOUN
easat-10389	102	30	risks	risk	NOUN
easat-10389	102	31	(	(	PUNCT
easat-10389	102	32	eq.8	eq.8	PROPN
easat-10389	102	33	)	)	PUNCT
easat-10389	102	34	.	.	PUNCT
easat-10389	103	1	ℒ	ℒ	NOUN
easat-10389	103	2	=	=	PUNCT
easat-10389	103	3	−	−	PROPN
easat-10389	103	4	1	1	NUM
easat-10389	103	5	𝑁	𝑁	PROPN
easat-10389	103	6	∑	∑	PROPN
easat-10389	104	1	[	[	X
easat-10389	104	2	𝑦𝑖	𝑦𝑖	ADP
easat-10389	104	3	𝑁	𝑁	PROPN
easat-10389	104	4	𝑖=1	𝑖=1	PROPN
easat-10389	104	5	log(	log(	ADJ
easat-10389	104	6	�	�	PROPN
easat-10389	104	7	̂	̂	VERB
easat-10389	104	8	�	�	NOUN
easat-10389	104	9	𝑖	𝑖	NUM
easat-10389	104	10	)	)	PUNCT
easat-10389	104	11	+	+	CCONJ
easat-10389	104	12	(	(	PUNCT
easat-10389	104	13	1	1	NUM
easat-10389	104	14	−	−	NOUN
easat-10389	104	15	𝑦𝑖)log	𝑦𝑖)log	NOUN
easat-10389	104	16	(	(	PUNCT
easat-10389	104	17	1	1	NUM
easat-10389	104	18	−	−	PRON
easat-10389	104	19	�	�	PROPN
easat-10389	104	20	̂	̂	VERB
easat-10389	104	21	�	�	NOUN
easat-10389	104	22	𝑖	𝑖	NUM
easat-10389	104	23	)	)	PUNCT
easat-10389	104	24	]	]	PUNCT
easat-10389	104	25	(	(	PUNCT
easat-10389	104	26	8)	8)	NUM
easat-10389	104	27	where	where	SCONJ
easat-10389	104	28	:	:	PUNCT
easat-10389	104	29	𝑦𝑖	𝑦𝑖	PROPN
easat-10389	104	30	∈	∈	PROPN
easat-10389	104	31	{	{	PUNCT
easat-10389	104	32	0,1	0,1	NOUN
easat-10389	104	33	}	}	PUNCT
easat-10389	104	34	is	be	AUX
easat-10389	104	35	the	the	DET
easat-10389	104	36	ground	ground	NOUN
easat-10389	104	37	truth	truth	NOUN
easat-10389	104	38	label	label	NOUN
easat-10389	104	39	for	for	ADP
easat-10389	104	40	asthma	asthma	NOUN
easat-10389	104	41	risk	risk	NOUN
easat-10389	104	42	,	,	PUNCT
easat-10389	104	43	and	and	CCONJ
easat-10389	104	44	�	�	PROPN
easat-10389	104	45	̂	̂	NOUN
easat-10389	104	46	�	�	NOUN
easat-10389	104	47	𝑖is	𝑖is	PROPN
easat-10389	104	48	the	the	DET
easat-10389	104	49	predicted	predict	VERB
easat-10389	104	50	probability	probability	NOUN
easat-10389	104	51	.	.	PUNCT
easat-10389	105	1	the	the	DET
easat-10389	105	2	setup	setup	NOUN
easat-10389	105	3	is	be	AUX
easat-10389	105	4	illustrated	illustrate	VERB
easat-10389	105	5	in	in	ADP
easat-10389	105	6	figure	figure	NOUN
easat-10389	105	7	1	1	NUM
easat-10389	105	8	.	.	PUNCT
easat-10389	106	1	a	a	DET
easat-10389	106	2	cnn	cnn	PROPN
easat-10389	106	3	captures	capture	VERB
easat-10389	106	4	spatial	spatial	ADJ
easat-10389	106	5	-	-	PUNCT
easat-10389	106	6	temporal	temporal	ADJ
easat-10389	106	7	features	feature	NOUN
easat-10389	106	8	from	from	ADP
easat-10389	106	9	air	air	NOUN
easat-10389	106	10	pollution	pollution	NOUN
easat-10389	106	11	data	datum	NOUN
easat-10389	106	12	,	,	PUNCT
easat-10389	106	13	while	while	SCONJ
easat-10389	106	14	a	a	DET
easat-10389	106	15	gru	gru	NOUN
easat-10389	106	16	processes	process	VERB
easat-10389	106	17	sequential	sequential	ADJ
easat-10389	106	18	meteorological	meteorological	ADJ
easat-10389	106	19	data	datum	NOUN
easat-10389	106	20	.	.	PUNCT
easat-10389	107	1	an	an	DET
easat-10389	107	2	attention	attention	NOUN
easat-10389	107	3	mechanism	mechanism	NOUN
easat-10389	107	4	merges	merge	VERB
easat-10389	107	5	the	the	DET
easat-10389	107	6	outputs	output	NOUN
easat-10389	107	7	,	,	PUNCT
easat-10389	107	8	emphasizing	emphasize	VERB
easat-10389	107	9	the	the	DET
easat-10389	107	10	most	most	ADV
easat-10389	107	11	critical	critical	ADJ
easat-10389	107	12	aspects	aspect	NOUN
easat-10389	107	13	for	for	ADP
easat-10389	107	14	asthma	asthma	NOUN
easat-10389	107	15	risk	risk	NOUN
easat-10389	107	16	,	,	PUNCT
easat-10389	107	17	with	with	ADP
easat-10389	107	18	fully	fully	ADV
easat-10389	107	19	connected	connected	ADJ
easat-10389	107	20	layers	layer	NOUN
easat-10389	107	21	and	and	CCONJ
easat-10389	107	22	sigmoid	sigmoid	NOUN
easat-10389	107	23	activation	activation	NOUN
easat-10389	107	24	for	for	ADP
easat-10389	107	25	binary	binary	ADJ
easat-10389	107	26	risk	risk	NOUN
easat-10389	107	27	classification	classification	NOUN
easat-10389	107	28	.	.	PUNCT
easat-10389	108	1	figure	figure	NOUN
easat-10389	108	2	1	1	NUM
easat-10389	108	3	illustrates	illustrate	VERB
easat-10389	108	4	the	the	DET
easat-10389	108	5	complete	complete	ADJ
easat-10389	108	6	architecture	architecture	NOUN
easat-10389	108	7	of	of	ADP
easat-10389	108	8	the	the	DET
easat-10389	108	9	proposed	propose	VERB
easat-10389	108	10	dual	dual	ADJ
easat-10389	108	11	-	-	PUNCT
easat-10389	108	12	branch	branch	NOUN
easat-10389	108	13	deep	deep	ADJ
easat-10389	108	14	learning	learning	NOUN
easat-10389	108	15	model	model	NOUN
easat-10389	108	16	.	.	PUNCT
easat-10389	109	1	the	the	DET
easat-10389	109	2	framework	framework	NOUN
easat-10389	109	3	is	be	AUX
easat-10389	109	4	designed	design	VERB
easat-10389	109	5	to	to	PART
easat-10389	109	6	integrate	integrate	VERB
easat-10389	109	7	diverse	diverse	ADJ
easat-10389	109	8	environmental	environmental	ADJ
easat-10389	109	9	factors	factor	NOUN
easat-10389	109	10	for	for	ADP
easat-10389	109	11	predicting	predict	VERB
easat-10389	109	12	asthma	asthma	NOUN
easat-10389	109	13	risk	risk	NOUN
easat-10389	109	14	.	.	PUNCT
easat-10389	110	1	specifically	specifically	ADV
easat-10389	110	2	,	,	PUNCT
easat-10389	110	3	air	air	NOUN
easat-10389	110	4	quality	quality	NOUN
easat-10389	110	5	measurements	measurement	NOUN
easat-10389	110	6	(	(	PUNCT
easat-10389	110	7	e.g.	e.g.	ADV
easat-10389	110	8	,	,	PUNCT
easat-10389	110	9	pm2.5	pm2.5	PROPN
easat-10389	110	10	,	,	PUNCT
easat-10389	110	11	no2	no2	NOUN
easat-10389	110	12	)	)	PUNCT
easat-10389	110	13	are	be	AUX
easat-10389	110	14	processed	process	VERB
easat-10389	110	15	by	by	ADP
easat-10389	110	16	a	a	DET
easat-10389	110	17	1d	1d	NUM
easat-10389	110	18	cnn	cnn	NOUN
easat-10389	110	19	to	to	PART
easat-10389	110	20	detect	detect	VERB
easat-10389	110	21	local	local	ADJ
easat-10389	110	22	spatial	spatial	ADJ
easat-10389	110	23	patterns	pattern	NOUN
easat-10389	110	24	.	.	PUNCT
easat-10389	111	1	meanwhile	meanwhile	ADV
easat-10389	111	2	,	,	PUNCT
easat-10389	111	3	meteorological	meteorological	ADJ
easat-10389	111	4	variables	variable	NOUN
easat-10389	111	5	(	(	PUNCT
easat-10389	111	6	e.g.	e.g.	ADV
easat-10389	111	7	,	,	PUNCT
easat-10389	111	8	temperature	temperature	NOUN
easat-10389	111	9	,	,	PUNCT
easat-10389	111	10	humidity	humidity	NOUN
easat-10389	111	11	,	,	PUNCT
easat-10389	111	12	wind	wind	NOUN
easat-10389	111	13	speed	speed	NOUN
easat-10389	111	14	)	)	PUNCT
easat-10389	111	15	are	be	AUX
easat-10389	111	16	fed	feed	VERB
easat-10389	111	17	into	into	ADP
easat-10389	111	18	a	a	DET
easat-10389	111	19	recurrent	recurrent	ADJ
easat-10389	111	20	neural	neural	ADJ
easat-10389	111	21	network	network	NOUN
easat-10389	111	22	(	(	PUNCT
easat-10389	111	23	gru	gru	PROPN
easat-10389	111	24	)	)	PUNCT
easat-10389	111	25	to	to	PART
easat-10389	111	26	capture	capture	VERB
easat-10389	111	27	long	long	ADJ
easat-10389	111	28	-	-	PUNCT
easat-10389	111	29	term	term	NOUN
easat-10389	111	30	temporal	temporal	ADJ
easat-10389	111	31	dependencies	dependency	NOUN
easat-10389	111	32	.	.	PUNCT
easat-10389	112	1	the	the	DET
easat-10389	112	2	outputs	output	NOUN
easat-10389	112	3	from	from	ADP
easat-10389	112	4	both	both	DET
easat-10389	112	5	branches	branch	NOUN
easat-10389	112	6	are	be	AUX
easat-10389	112	7	integrated	integrate	VERB
easat-10389	112	8	through	through	ADP
easat-10389	112	9	an	an	DET
easat-10389	112	10	attention	attention	NOUN
easat-10389	112	11	mechanism	mechanism	NOUN
easat-10389	112	12	,	,	PUNCT
easat-10389	112	13	which	which	PRON
easat-10389	112	14	adaptively	adaptively	ADV
easat-10389	112	15	assigns	assign	VERB
easat-10389	112	16	weights	weight	NOUN
easat-10389	112	17	to	to	PART
easat-10389	112	18	emphasize	emphasize	VERB
easat-10389	112	19	the	the	DET
easat-10389	112	20	most	most	ADV
easat-10389	112	21	relevant	relevant	ADJ
easat-10389	112	22	features	feature	NOUN
easat-10389	112	23	.	.	PUNCT
easat-10389	113	1	the	the	DET
easat-10389	113	2	final	final	ADJ
easat-10389	113	3	fused	fused	ADJ
easat-10389	113	4	representation	representation	NOUN
easat-10389	113	5	is	be	AUX
easat-10389	113	6	then	then	ADV
easat-10389	113	7	passed	pass	VERB
easat-10389	113	8	through	through	ADP
easat-10389	113	9	fully	fully	ADV
easat-10389	113	10	connected	connected	ADJ
easat-10389	113	11	layers	layer	NOUN
easat-10389	113	12	for	for	ADP
easat-10389	113	13	binary	binary	ADJ
easat-10389	113	14	classification	classification	NOUN
easat-10389	113	15	(	(	PUNCT
easat-10389	113	16	high	high	ADJ
easat-10389	113	17	vs.	vs.	ADP
easat-10389	113	18	low	low	ADJ
easat-10389	113	19	asthma	asthma	NOUN
easat-10389	113	20	risk	risk	NOUN
easat-10389	113	21	)	)	PUNCT
easat-10389	113	22	.	.	PUNCT
easat-10389	114	1	this	this	DET
easat-10389	114	2	architecture	architecture	NOUN
easat-10389	114	3	enables	enable	VERB
easat-10389	114	4	the	the	DET
easat-10389	114	5	model	model	NOUN
easat-10389	114	6	to	to	PART
easat-10389	114	7	learn	learn	VERB
easat-10389	114	8	from	from	ADP
easat-10389	114	9	both	both	CCONJ
easat-10389	114	10	spatial	spatial	ADJ
easat-10389	114	11	and	and	CCONJ
easat-10389	114	12	temporal	temporal	ADJ
easat-10389	114	13	cues	cue	NOUN
easat-10389	114	14	simultaneously	simultaneously	ADV
easat-10389	114	15	,	,	PUNCT
easat-10389	114	16	thereby	thereby	ADV
easat-10389	114	17	enhancing	enhance	VERB
easat-10389	114	18	interpretability	interpretability	NOUN
easat-10389	114	19	and	and	CCONJ
easat-10389	114	20	facilitating	facilitate	VERB
easat-10389	114	21	deployment	deployment	ADJ
easat-10389	114	22	readiness	readiness	NOUN
easat-10389	114	23	.	.	PUNCT
easat-10389	115	1	260	260	NUM
easat-10389	115	2	edelweiss	edelweiss	PROPN
easat-10389	115	3	applied	apply	VERB
easat-10389	115	4	science	science	NOUN
easat-10389	115	5	and	and	CCONJ
easat-10389	115	6	technology	technology	NOUN
easat-10389	115	7	issn	issn	PROPN
easat-10389	115	8	:	:	PUNCT
easat-10389	115	9	2576	2576	NUM
easat-10389	115	10	-	-	SYM
easat-10389	115	11	8484	8484	NUM
easat-10389	115	12	vol	vol	NOUN
easat-10389	115	13	.	.	PROPN
easat-10389	116	1	9	9	NUM
easat-10389	116	2	,	,	PUNCT
easat-10389	116	3	no	no	INTJ
easat-10389	116	4	.	.	NOUN
easat-10389	116	5	10	10	NUM
easat-10389	116	6	:	:	PUNCT
easat-10389	116	7	256	256	NUM
easat-10389	116	8	-	-	SYM
easat-10389	116	9	271	271	NUM
easat-10389	116	10	,	,	PUNCT
easat-10389	116	11	2025	2025	NUM
easat-10389	116	12	doi	doi	NOUN
easat-10389	116	13	:	:	PUNCT
easat-10389	116	14	10.55214/2576	10.55214/2576	NUM
easat-10389	116	15	-	-	PUNCT
easat-10389	116	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	116	17	©	©	ADP
easat-10389	116	18	2025	2025	NUM
easat-10389	116	19	by	by	ADP
easat-10389	116	20	the	the	DET
easat-10389	116	21	authors	author	NOUN
easat-10389	116	22	;	;	PUNCT
easat-10389	116	23	licensee	licensee	PROPN
easat-10389	116	24	learning	learn	VERB
easat-10389	116	25	gate	gate	NOUN
easat-10389	116	26	figure	figure	NOUN
easat-10389	116	27	1	1	NUM
easat-10389	116	28	.	.	PUNCT
easat-10389	116	29	deep	deep	ADJ
easat-10389	116	30	learning	learning	NOUN
easat-10389	116	31	architecture	architecture	NOUN
easat-10389	116	32	with	with	ADP
easat-10389	116	33	two	two	NUM
easat-10389	116	34	branches	branch	NOUN
easat-10389	116	35	for	for	ADP
easat-10389	116	36	predicting	predict	VERB
easat-10389	116	37	asthma	asthma	NOUN
easat-10389	116	38	risk	risk	NOUN
easat-10389	116	39	.	.	PUNCT
easat-10389	117	1	3.2	3.2	NUM
easat-10389	117	2	.	.	X
easat-10389	118	1	setting	set	VERB
easat-10389	118	2	up	up	ADP
easat-10389	118	3	training	train	VERB
easat-10389	118	4	the	the	DET
easat-10389	118	5	suggested	suggest	VERB
easat-10389	118	6	model	model	NOUN
easat-10389	118	7	was	be	AUX
easat-10389	118	8	created	create	VERB
easat-10389	118	9	and	and	CCONJ
easat-10389	118	10	tested	test	VERB
easat-10389	118	11	only	only	ADV
easat-10389	118	12	in	in	ADP
easat-10389	118	13	a	a	DET
easat-10389	118	14	simulation	simulation	NOUN
easat-10389	118	15	environment	environment	NOUN
easat-10389	118	16	using	use	VERB
easat-10389	118	17	python	python	NOUN
easat-10389	118	18	and	and	CCONJ
easat-10389	118	19	tensorflow	tensorflow	NOUN
easat-10389	118	20	2.13	2.13	NUM
easat-10389	118	21	.	.	PUNCT
easat-10389	119	1	it	it	PRON
easat-10389	119	2	was	be	AUX
easat-10389	119	3	not	not	PART
easat-10389	119	4	deployed	deploy	VERB
easat-10389	119	5	on	on	ADP
easat-10389	119	6	real	real	ADJ
easat-10389	119	7	-	-	PUNCT
easat-10389	119	8	time	time	NOUN
easat-10389	119	9	systems	system	NOUN
easat-10389	119	10	or	or	CCONJ
easat-10389	119	11	hardware	hardware	NOUN
easat-10389	119	12	platforms	platform	NOUN
easat-10389	119	13	.	.	PUNCT
easat-10389	120	1	we	we	PRON
easat-10389	120	2	used	use	VERB
easat-10389	120	3	a	a	DET
easat-10389	120	4	workstation	workstation	NOUN
easat-10389	120	5	equipped	equip	VERB
easat-10389	120	6	with	with	ADP
easat-10389	120	7	an	an	DET
easat-10389	120	8	nvidia	nvidia	PROPN
easat-10389	120	9	rtx	rtx	PROPN
easat-10389	120	10	3090	3090	NUM
easat-10389	120	11	gpu	gpu	NOUN
easat-10389	120	12	for	for	ADP
easat-10389	120	13	all	all	DET
easat-10389	120	14	trials	trial	NOUN
easat-10389	120	15	to	to	PART
easat-10389	120	16	ensure	ensure	VERB
easat-10389	120	17	that	that	SCONJ
easat-10389	120	18	the	the	DET
easat-10389	120	19	model	model	NOUN
easat-10389	120	20	training	training	NOUN
easat-10389	120	21	was	be	AUX
easat-10389	120	22	sufficiently	sufficiently	ADV
easat-10389	120	23	fast	fast	ADJ
easat-10389	120	24	.	.	PUNCT
easat-10389	121	1	the	the	DET
easat-10389	121	2	training	training	NOUN
easat-10389	121	3	procedure	procedure	NOUN
easat-10389	121	4	involved	involve	VERB
easat-10389	121	5	predicting	predict	VERB
easat-10389	121	6	the	the	DET
easat-10389	121	7	risk	risk	NOUN
easat-10389	121	8	of	of	ADP
easat-10389	121	9	asthma	asthma	NOUN
easat-10389	121	10	as	as	ADP
easat-10389	121	11	a	a	DET
easat-10389	121	12	binary	binary	ADJ
easat-10389	121	13	classification	classification	NOUN
easat-10389	121	14	problem	problem	NOUN
easat-10389	121	15	(	(	PUNCT
easat-10389	121	16	high	high	ADJ
easat-10389	121	17	vs.	vs.	ADP
easat-10389	121	18	low	low	ADJ
easat-10389	121	19	risk	risk	NOUN
easat-10389	121	20	)	)	PUNCT
easat-10389	121	21	,	,	PUNCT
easat-10389	121	22	using	use	VERB
easat-10389	121	23	binary	binary	PROPN
easat-10389	121	24	cross	cross	NOUN
easat-10389	121	25	-	-	NOUN
easat-10389	121	26	entropy	entropy	NOUN
easat-10389	121	27	as	as	ADP
easat-10389	121	28	the	the	DET
easat-10389	121	29	loss	loss	NOUN
easat-10389	121	30	function	function	NOUN
easat-10389	121	31	.	.	PUNCT
easat-10389	122	1	this	this	DET
easat-10389	122	2	paper	paper	NOUN
easat-10389	122	3	presents	present	VERB
easat-10389	122	4	the	the	DET
easat-10389	122	5	model	model	NOUN
easat-10389	122	6	using	use	VERB
easat-10389	122	7	the	the	DET
easat-10389	122	8	following	follow	VERB
easat-10389	122	9	setup	setup	NOUN
easat-10389	122	10	:	:	PUNCT
easat-10389	122	11	•	•	NUM
easat-10389	122	12	adam	adam	PROPN
easat-10389	122	13	optimizer	optimizer	NOUN
easat-10389	122	14	with	with	ADP
easat-10389	122	15	an	an	DET
easat-10389	122	16	initial	initial	ADJ
easat-10389	122	17	learning	learning	NOUN
easat-10389	122	18	rate	rate	NOUN
easat-10389	122	19	of	of	ADP
easat-10389	122	20	0.001	0.001	NUM
easat-10389	122	21	.	.	PUNCT
easat-10389	123	1	•	•	NOUN
easat-10389	123	2	the	the	DET
easat-10389	123	3	batch	batch	NOUN
easat-10389	123	4	size	size	NOUN
easat-10389	123	5	is	be	AUX
easat-10389	123	6	64	64	NUM
easat-10389	123	7	.	.	NOUN
easat-10389	124	1	•	•	NOUN
easat-10389	124	2	epochs	epoch	NOUN
easat-10389	124	3	:	:	PUNCT
easat-10389	124	4	up	up	ADP
easat-10389	124	5	to	to	PART
easat-10389	124	6	100	100	NUM
easat-10389	124	7	,	,	PUNCT
easat-10389	124	8	with	with	ADP
easat-10389	124	9	early	early	ADJ
easat-10389	124	10	stopping	stopping	NOUN
easat-10389	124	11	if	if	SCONJ
easat-10389	124	12	the	the	DET
easat-10389	124	13	validation	validation	NOUN
easat-10389	124	14	loss	loss	NOUN
easat-10389	124	15	did	do	AUX
easat-10389	124	16	not	not	PART
easat-10389	124	17	improve	improve	VERB
easat-10389	124	18	over	over	ADP
easat-10389	124	19	10	10	NUM
easat-10389	124	20	consecutive	consecutive	ADJ
easat-10389	124	21	epochs	epoch	NOUN
easat-10389	124	22	.	.	PUNCT
easat-10389	125	1	•	•	NUM
easat-10389	125	2	validation	validation	NOUN
easat-10389	125	3	split	split	NOUN
easat-10389	125	4	:	:	PUNCT
easat-10389	125	5	20	20	NUM
easat-10389	125	6	%	%	NOUN
easat-10389	125	7	of	of	ADP
easat-10389	125	8	the	the	DET
easat-10389	125	9	data	datum	NOUN
easat-10389	125	10	used	use	VERB
easat-10389	125	11	for	for	ADP
easat-10389	125	12	training	training	NOUN
easat-10389	125	13	.	.	PUNCT
easat-10389	126	1	•	•	NUM
easat-10389	126	2	regularization	regularization	NOUN
easat-10389	126	3	:	:	PUNCT
easat-10389	126	4	use	use	VERB
easat-10389	126	5	dropout	dropout	NOUN
easat-10389	126	6	with	with	ADP
easat-10389	126	7	a	a	DET
easat-10389	126	8	rate	rate	NOUN
easat-10389	126	9	of	of	ADP
easat-10389	126	10	0.3	0.3	NUM
easat-10389	126	11	and	and	CCONJ
easat-10389	126	12	l2	l2	NOUN
easat-10389	126	13	kernel	kernel	NOUN
easat-10389	126	14	regularization	regularization	NOUN
easat-10389	126	15	to	to	PART
easat-10389	126	16	prevent	prevent	VERB
easat-10389	126	17	the	the	DET
easat-10389	126	18	model	model	NOUN
easat-10389	126	19	from	from	ADP
easat-10389	126	20	overfitting	overfitte	VERB
easat-10389	126	21	.	.	PUNCT
easat-10389	127	1	•	•	NOUN
easat-10389	127	2	the	the	DET
easat-10389	127	3	measures	measure	NOUN
easat-10389	127	4	used	use	VERB
easat-10389	127	5	for	for	ADP
easat-10389	127	6	evaluation	evaluation	NOUN
easat-10389	127	7	are	be	AUX
easat-10389	127	8	accuracy	accuracy	NOUN
easat-10389	127	9	,	,	PUNCT
easat-10389	127	10	precision	precision	NOUN
easat-10389	127	11	,	,	PUNCT
easat-10389	127	12	recall	recall	NOUN
easat-10389	127	13	,	,	PUNCT
easat-10389	127	14	f1	f1	NOUN
easat-10389	127	15	-	-	PUNCT
easat-10389	127	16	score	score	NOUN
easat-10389	127	17	,	,	PUNCT
easat-10389	127	18	and	and	CCONJ
easat-10389	127	19	the	the	DET
easat-10389	127	20	area	area	NOUN
easat-10389	127	21	under	under	ADP
easat-10389	127	22	the	the	DET
easat-10389	127	23	curve	curve	NOUN
easat-10389	127	24	(	(	PUNCT
easat-10389	127	25	auc	auc	NOUN
easat-10389	127	26	)	)	PUNCT
easat-10389	127	27	.	.	PUNCT
easat-10389	128	1	this	this	DET
easat-10389	128	2	study	study	NOUN
easat-10389	128	3	assessed	assess	VERB
easat-10389	128	4	the	the	DET
easat-10389	128	5	model	model	NOUN
easat-10389	128	6	's	's	PART
easat-10389	128	7	strength	strength	NOUN
easat-10389	128	8	using	use	VERB
easat-10389	128	9	a	a	DET
easat-10389	128	10	method	method	NOUN
easat-10389	128	11	called	call	VERB
easat-10389	128	12	5	5	NUM
easat-10389	128	13	-	-	ADJ
easat-10389	128	14	fold	fold	ADJ
easat-10389	128	15	cross	cross	NOUN
easat-10389	128	16	-	-	NOUN
easat-10389	128	17	validation	validation	NOUN
easat-10389	128	18	in	in	ADP
easat-10389	128	19	different	different	ADJ
easat-10389	128	20	cities	city	NOUN
easat-10389	128	21	and	and	CCONJ
easat-10389	128	22	then	then	ADV
easat-10389	128	23	again	again	ADV
easat-10389	128	24	using	use	VERB
easat-10389	128	25	2024	2024	NUM
easat-10389	128	26	data	datum	NOUN
easat-10389	128	27	that	that	PRON
easat-10389	128	28	had	have	AUX
easat-10389	128	29	never	never	ADV
easat-10389	128	30	been	be	AUX
easat-10389	128	31	seen	see	VERB
easat-10389	128	32	before	before	ADV
easat-10389	128	33	.	.	PUNCT
easat-10389	129	1	then	then	ADV
easat-10389	129	2	,	,	PUNCT
easat-10389	129	3	a	a	DET
easat-10389	129	4	grid	grid	NOUN
easat-10389	129	5	search	search	NOUN
easat-10389	129	6	was	be	AUX
easat-10389	129	7	used	use	VERB
easat-10389	129	8	to	to	PART
easat-10389	129	9	find	find	VERB
easat-10389	129	10	the	the	DET
easat-10389	129	11	best	good	ADJ
easat-10389	129	12	hyperparameters	hyperparameter	NOUN
easat-10389	129	13	,	,	PUNCT
easat-10389	129	14	and	and	CCONJ
easat-10389	129	15	the	the	DET
easat-10389	129	16	performance	performance	NOUN
easat-10389	129	17	was	be	AUX
easat-10389	129	18	averaged	average	VERB
easat-10389	129	19	across	across	ADP
easat-10389	129	20	the	the	DET
easat-10389	129	21	folds	fold	NOUN
easat-10389	129	22	.	.	PUNCT
easat-10389	130	1	the	the	DET
easat-10389	130	2	results	result	NOUN
easat-10389	130	3	are	be	AUX
easat-10389	130	4	intriguing	intriguing	ADJ
easat-10389	130	5	,	,	PUNCT
easat-10389	130	6	but	but	CCONJ
easat-10389	130	7	the	the	DET
easat-10389	130	8	study	study	NOUN
easat-10389	130	9	is	be	AUX
easat-10389	130	10	still	still	ADV
easat-10389	130	11	confined	confine	VERB
easat-10389	130	12	to	to	ADP
easat-10389	130	13	simulations	simulation	NOUN
easat-10389	130	14	and	and	CCONJ
easat-10389	130	15	offline	offline	ADJ
easat-10389	130	16	validation	validation	NOUN
easat-10389	130	17	.	.	PUNCT
easat-10389	131	1	real	real	ADJ
easat-10389	131	2	-	-	PUNCT
easat-10389	131	3	world	world	NOUN
easat-10389	131	4	deployment	deployment	NOUN
easat-10389	131	5	and	and	CCONJ
easat-10389	131	6	live	live	ADJ
easat-10389	131	7	data	data	NOUN
easat-10389	131	8	integration	integration	NOUN
easat-10389	131	9	will	will	AUX
easat-10389	131	10	need	need	VERB
easat-10389	131	11	to	to	PART
easat-10389	131	12	be	be	AUX
easat-10389	131	13	postponed	postpone	VERB
easat-10389	131	14	for	for	ADP
easat-10389	131	15	further	further	ADJ
easat-10389	131	16	work	work	NOUN
easat-10389	131	17	.	.	PUNCT
easat-10389	132	1	3.3	3.3	NUM
easat-10389	132	2	.	.	PUNCT
easat-10389	133	1	putting	put	VERB
easat-10389	133	2	it	it	PRON
easat-10389	133	3	into	into	ADP
easat-10389	133	4	action	action	NOUN
easat-10389	133	5	and	and	CCONJ
easat-10389	133	6	simulating	simulate	VERB
easat-10389	133	7	3.3.1	3.3.1	NUM
easat-10389	133	8	.	.	PUNCT
easat-10389	133	9	making	make	VERB
easat-10389	133	10	and	and	CCONJ
easat-10389	133	11	preparing	prepare	VERB
easat-10389	133	12	the	the	DET
easat-10389	133	13	data	datum	NOUN
easat-10389	133	14	261	261	NUM
easat-10389	133	15	edelweiss	edelweiss	PROPN
easat-10389	133	16	applied	apply	VERB
easat-10389	133	17	science	science	NOUN
easat-10389	133	18	and	and	CCONJ
easat-10389	133	19	technology	technology	NOUN
easat-10389	133	20	issn	issn	PROPN
easat-10389	133	21	:	:	PUNCT
easat-10389	133	22	2576	2576	NUM
easat-10389	133	23	-	-	SYM
easat-10389	133	24	8484	8484	NUM
easat-10389	133	25	vol	vol	NOUN
easat-10389	133	26	.	.	PROPN
easat-10389	134	1	9	9	NUM
easat-10389	134	2	,	,	PUNCT
easat-10389	134	3	no	no	INTJ
easat-10389	134	4	.	.	NOUN
easat-10389	134	5	10	10	NUM
easat-10389	134	6	:	:	PUNCT
easat-10389	134	7	256	256	NUM
easat-10389	134	8	-	-	SYM
easat-10389	134	9	271	271	NUM
easat-10389	134	10	,	,	PUNCT
easat-10389	134	11	2025	2025	NUM
easat-10389	134	12	doi	doi	NOUN
easat-10389	134	13	:	:	PUNCT
easat-10389	134	14	10.55214/2576	10.55214/2576	NUM
easat-10389	134	15	-	-	PUNCT
easat-10389	134	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	134	17	©	©	ADP
easat-10389	134	18	2025	2025	NUM
easat-10389	134	19	by	by	ADP
easat-10389	134	20	the	the	DET
easat-10389	134	21	authors	author	NOUN
easat-10389	134	22	;	;	PUNCT
easat-10389	134	23	licensee	licensee	PROPN
easat-10389	134	24	learning	learning	NOUN
easat-10389	134	25	gate	gate	VERB
easat-10389	134	26	the	the	DET
easat-10389	134	27	study	study	NOUN
easat-10389	134	28	utilized	utilize	VERB
easat-10389	134	29	real	real	ADJ
easat-10389	134	30	-	-	PUNCT
easat-10389	134	31	world	world	NOUN
easat-10389	134	32	environmental	environmental	ADJ
easat-10389	134	33	data	datum	NOUN
easat-10389	134	34	from	from	ADP
easat-10389	134	35	seoul	seoul	PROPN
easat-10389	134	36	(	(	PUNCT
easat-10389	134	37	south	south	PROPN
easat-10389	134	38	korea	korea	PROPN
easat-10389	134	39	)	)	PUNCT
easat-10389	134	40	,	,	PUNCT
easat-10389	134	41	los	los	PROPN
easat-10389	134	42	angeles	angeles	PROPN
easat-10389	134	43	(	(	PUNCT
easat-10389	134	44	usa	usa	PROPN
easat-10389	134	45	)	)	PUNCT
easat-10389	134	46	,	,	PUNCT
easat-10389	134	47	and	and	CCONJ
easat-10389	134	48	hanoi	hanoi	PROPN
easat-10389	134	49	(	(	PUNCT
easat-10389	134	50	vietnam	vietnam	PROPN
easat-10389	134	51	)	)	PUNCT
easat-10389	134	52	to	to	PART
easat-10389	134	53	train	train	VERB
easat-10389	134	54	and	and	CCONJ
easat-10389	134	55	test	test	VERB
easat-10389	134	56	the	the	DET
easat-10389	134	57	proposed	propose	VERB
easat-10389	134	58	model	model	NOUN
easat-10389	134	59	.	.	PUNCT
easat-10389	135	1	these	these	DET
easat-10389	135	2	cities	city	NOUN
easat-10389	135	3	were	be	AUX
easat-10389	135	4	selected	select	VERB
easat-10389	135	5	because	because	SCONJ
easat-10389	135	6	they	they	PRON
easat-10389	135	7	have	have	VERB
easat-10389	135	8	large	large	ADJ
easat-10389	135	9	populations	population	NOUN
easat-10389	135	10	and	and	CCONJ
easat-10389	135	11	their	their	PRON
easat-10389	135	12	air	air	NOUN
easat-10389	135	13	quality	quality	NOUN
easat-10389	135	14	and	and	CCONJ
easat-10389	135	15	weather	weather	NOUN
easat-10389	135	16	conditions	condition	NOUN
easat-10389	135	17	differ	differ	VERB
easat-10389	135	18	significantly	significantly	ADV
easat-10389	135	19	.	.	PUNCT
easat-10389	136	1	the	the	DET
easat-10389	136	2	data	datum	NOUN
easat-10389	136	3	collection	collection	NOUN
easat-10389	136	4	,	,	PUNCT
easat-10389	136	5	spanning	span	VERB
easat-10389	136	6	from	from	ADP
easat-10389	136	7	2020	2020	NUM
easat-10389	136	8	to	to	ADP
easat-10389	136	9	2024	2024	NUM
easat-10389	136	10	,	,	PUNCT
easat-10389	136	11	was	be	AUX
easat-10389	136	12	compiled	compile	VERB
easat-10389	136	13	using	use	VERB
easat-10389	136	14	official	official	ADJ
easat-10389	136	15	government	government	NOUN
easat-10389	136	16	sources	source	NOUN
easat-10389	136	17	and	and	CCONJ
easat-10389	136	18	open	open	ADJ
easat-10389	136	19	-	-	PUNCT
easat-10389	136	20	access	access	NOUN
easat-10389	136	21	apis	apis	ADJ
easat-10389	136	22	.	.	PUNCT
easat-10389	137	1	some	some	PRON
easat-10389	137	2	of	of	ADP
easat-10389	137	3	the	the	DET
easat-10389	137	4	input	input	NOUN
easat-10389	137	5	features	feature	NOUN
easat-10389	137	6	include	include	VERB
easat-10389	137	7	:	:	PUNCT
easat-10389	137	8	•	•	NUM
easat-10389	137	9	air	air	NOUN
easat-10389	137	10	quality	quality	NOUN
easat-10389	137	11	measurements	measurement	NOUN
easat-10389	137	12	(	(	PUNCT
easat-10389	137	13	pm2.5	pm2.5	PROPN
easat-10389	137	14	,	,	PUNCT
easat-10389	137	15	pm10	pm10	PROPN
easat-10389	137	16	,	,	PUNCT
easat-10389	137	17	no₂	no₂	PROPN
easat-10389	137	18	,	,	PUNCT
easat-10389	137	19	so₂	so₂	PROPN
easat-10389	137	20	,	,	PUNCT
easat-10389	137	21	co	co	NOUN
easat-10389	137	22	,	,	PUNCT
easat-10389	137	23	o₃	o₃	PROPN
easat-10389	137	24	)	)	PUNCT
easat-10389	137	25	from	from	ADP
easat-10389	137	26	national	national	ADJ
easat-10389	137	27	air	air	NOUN
easat-10389	137	28	monitoring	monitoring	NOUN
easat-10389	137	29	groups	group	NOUN
easat-10389	137	30	such	such	ADJ
easat-10389	137	31	as	as	ADP
easat-10389	137	32	the	the	DET
easat-10389	137	33	us	us	PROPN
easat-10389	137	34	epa	epa	PROPN
easat-10389	137	35	,	,	PUNCT
easat-10389	137	36	korea	korea	PROPN
easat-10389	137	37	air	air	PROPN
easat-10389	137	38	,	,	PUNCT
easat-10389	137	39	and	and	CCONJ
easat-10389	137	40	vietnam	vietnam	PROPN
easat-10389	137	41	monre	monre	PROPN
easat-10389	137	42	.	.	PUNCT
easat-10389	138	1	•	•	NUM
easat-10389	138	2	weather	weather	NOUN
easat-10389	138	3	data	datum	NOUN
easat-10389	138	4	(	(	PUNCT
easat-10389	138	5	temperature	temperature	NOUN
easat-10389	138	6	,	,	PUNCT
easat-10389	138	7	humidity	humidity	NOUN
easat-10389	138	8	,	,	PUNCT
easat-10389	138	9	wind	wind	NOUN
easat-10389	138	10	speed	speed	NOUN
easat-10389	138	11	,	,	PUNCT
easat-10389	138	12	rainfall	rainfall	NOUN
easat-10389	138	13	,	,	PUNCT
easat-10389	138	14	and	and	CCONJ
easat-10389	138	15	air	air	NOUN
easat-10389	138	16	pressure	pressure	NOUN
easat-10389	138	17	)	)	PUNCT
easat-10389	138	18	from	from	ADP
easat-10389	138	19	meteorological	meteorological	ADJ
easat-10389	138	20	services	service	NOUN
easat-10389	138	21	and	and	CCONJ
easat-10389	138	22	local	local	ADJ
easat-10389	138	23	stations	station	NOUN
easat-10389	138	24	.	.	PUNCT
easat-10389	139	1	•	•	NUM
easat-10389	139	2	temporal	temporal	ADJ
easat-10389	139	3	data	datum	NOUN
easat-10389	139	4	includes	include	VERB
easat-10389	139	5	the	the	DET
easat-10389	139	6	date	date	NOUN
easat-10389	139	7	,	,	PUNCT
easat-10389	139	8	day	day	NOUN
easat-10389	139	9	of	of	ADP
easat-10389	139	10	the	the	DET
easat-10389	139	11	week	week	NOUN
easat-10389	139	12	,	,	PUNCT
easat-10389	139	13	and	and	CCONJ
easat-10389	139	14	seasonal	seasonal	ADJ
easat-10389	139	15	designation	designation	NOUN
easat-10389	139	16	.	.	PUNCT
easat-10389	140	1	the	the	DET
easat-10389	140	2	goal	goal	NOUN
easat-10389	140	3	variable	variable	NOUN
easat-10389	140	4	indicates	indicate	VERB
easat-10389	140	5	the	the	DET
easat-10389	140	6	number	number	NOUN
easat-10389	140	7	of	of	ADP
easat-10389	140	8	people	people	NOUN
easat-10389	140	9	hospitalized	hospitalize	VERB
easat-10389	140	10	each	each	DET
easat-10389	140	11	day	day	NOUN
easat-10389	140	12	due	due	ADP
easat-10389	140	13	to	to	ADP
easat-10389	140	14	asthma	asthma	NOUN
easat-10389	140	15	.	.	PUNCT
easat-10389	141	1	this	this	DET
easat-10389	141	2	data	data	NOUN
easat-10389	141	3	is	be	AUX
easat-10389	141	4	sourced	source	VERB
easat-10389	141	5	from	from	ADP
easat-10389	141	6	publicly	publicly	ADV
easat-10389	141	7	available	available	ADJ
easat-10389	141	8	,	,	PUNCT
easat-10389	141	9	anonymized	anonymize	VERB
easat-10389	141	10	health	health	NOUN
easat-10389	141	11	records	record	NOUN
easat-10389	141	12	.	.	PUNCT
easat-10389	142	1	missing	miss	VERB
easat-10389	142	2	values	value	NOUN
easat-10389	142	3	were	be	AUX
easat-10389	142	4	filled	fill	VERB
easat-10389	142	5	using	use	VERB
easat-10389	142	6	forward	forward	ADV
easat-10389	142	7	filling	filling	NOUN
easat-10389	142	8	,	,	PUNCT
easat-10389	142	9	and	and	CCONJ
easat-10389	142	10	the	the	DET
easat-10389	142	11	data	datum	NOUN
easat-10389	142	12	were	be	AUX
easat-10389	142	13	then	then	ADV
easat-10389	142	14	standardized	standardize	VERB
easat-10389	142	15	to	to	PART
easat-10389	142	16	have	have	VERB
easat-10389	142	17	a	a	DET
easat-10389	142	18	mean	mean	NOUN
easat-10389	142	19	of	of	ADP
easat-10389	142	20	zero	zero	NUM
easat-10389	142	21	and	and	CCONJ
easat-10389	142	22	a	a	DET
easat-10389	142	23	variance	variance	NOUN
easat-10389	142	24	of	of	ADP
easat-10389	142	25	one	one	NUM
easat-10389	142	26	.	.	PUNCT
easat-10389	143	1	for	for	ADP
easat-10389	143	2	the	the	DET
easat-10389	143	3	experiment	experiment	NOUN
easat-10389	143	4	,	,	PUNCT
easat-10389	143	5	the	the	DET
easat-10389	143	6	study	study	NOUN
easat-10389	143	7	randomly	randomly	ADV
easat-10389	143	8	assigned	assign	VERB
easat-10389	143	9	binary	binary	ADJ
easat-10389	143	10	labels	label	NOUN
easat-10389	143	11	to	to	ADP
easat-10389	143	12	asthma	asthma	NOUN
easat-10389	143	13	days	day	NOUN
easat-10389	143	14	,	,	PUNCT
easat-10389	143	15	marking	mark	VERB
easat-10389	143	16	them	they	PRON
easat-10389	143	17	as	as	ADP
easat-10389	143	18	either	either	CCONJ
easat-10389	143	19	highrisk	highrisk	NOUN
easat-10389	143	20	(	(	PUNCT
easat-10389	143	21	1	1	NUM
easat-10389	143	22	)	)	PUNCT
easat-10389	143	23	or	or	CCONJ
easat-10389	143	24	low	low	ADJ
easat-10389	143	25	-	-	PUNCT
easat-10389	143	26	risk	risk	NOUN
easat-10389	143	27	(	(	PUNCT
easat-10389	143	28	0	0	NUM
easat-10389	143	29	)	)	PUNCT
easat-10389	143	30	.	.	PUNCT
easat-10389	144	1	before	before	ADP
easat-10389	144	2	training	train	VERB
easat-10389	144	3	the	the	DET
easat-10389	144	4	model	model	NOUN
easat-10389	144	5	,	,	PUNCT
easat-10389	144	6	input	input	NOUN
easat-10389	144	7	features	feature	NOUN
easat-10389	144	8	were	be	AUX
easat-10389	144	9	also	also	ADV
easat-10389	144	10	normalized	normalize	VERB
easat-10389	144	11	to	to	PART
easat-10389	144	12	have	have	VERB
easat-10389	144	13	a	a	DET
easat-10389	144	14	mean	mean	NOUN
easat-10389	144	15	of	of	ADP
easat-10389	144	16	zero	zero	NUM
easat-10389	144	17	and	and	CCONJ
easat-10389	144	18	a	a	DET
easat-10389	144	19	variance	variance	NOUN
easat-10389	144	20	of	of	ADP
easat-10389	144	21	one	one	NUM
easat-10389	144	22	.	.	PUNCT
easat-10389	145	1	standard	standard	ADJ
easat-10389	145	2	imputation	imputation	NOUN
easat-10389	145	3	methods	method	NOUN
easat-10389	145	4	,	,	PUNCT
easat-10389	145	5	like	like	ADP
easat-10389	145	6	forward	forward	ADV
easat-10389	145	7	filling	filling	NOUN
easat-10389	145	8	,	,	PUNCT
easat-10389	145	9	were	be	AUX
easat-10389	145	10	used	use	VERB
easat-10389	145	11	to	to	PART
easat-10389	145	12	address	address	VERB
easat-10389	145	13	missing	miss	VERB
easat-10389	145	14	data	datum	NOUN
easat-10389	145	15	.	.	PUNCT
easat-10389	146	1	3.3.2	3.3.2	X
easat-10389	146	2	.	.	PUNCT
easat-10389	147	1	putting	put	VERB
easat-10389	147	2	the	the	DET
easat-10389	147	3	model	model	NOUN
easat-10389	147	4	into	into	ADP
easat-10389	147	5	action	action	NOUN
easat-10389	147	6	this	this	DET
easat-10389	147	7	research	research	NOUN
easat-10389	147	8	created	create	VERB
easat-10389	147	9	two	two	NUM
easat-10389	147	10	key	key	ADJ
easat-10389	147	11	scripts	script	NOUN
easat-10389	147	12	for	for	ADP
easat-10389	147	13	simulation	simulation	NOUN
easat-10389	147	14	.	.	PUNCT
easat-10389	148	1	•	•	NUM
easat-10389	148	2	full	full	ADJ
easat-10389	148	3	dual	dual	ADJ
easat-10389	148	4	-	-	PUNCT
easat-10389	148	5	branch	branch	NOUN
easat-10389	148	6	model	model	NOUN
easat-10389	148	7	simulation	simulation	NOUN
easat-10389	148	8	:	:	PUNCT
easat-10389	148	9	this	this	DET
easat-10389	148	10	script	script	NOUN
easat-10389	148	11	utilizes	utilize	VERB
easat-10389	148	12	the	the	DET
easat-10389	148	13	entire	entire	ADJ
easat-10389	148	14	model	model	NOUN
easat-10389	148	15	,	,	PUNCT
easat-10389	148	16	which	which	PRON
easat-10389	148	17	comprises	comprise	VERB
easat-10389	148	18	a	a	DET
easat-10389	148	19	cnn	cnn	NOUN
easat-10389	148	20	for	for	ADP
easat-10389	148	21	air	air	NOUN
easat-10389	148	22	quality	quality	PROPN
easat-10389	148	23	data	datum	NOUN
easat-10389	148	24	,	,	PUNCT
easat-10389	148	25	a	a	DET
easat-10389	148	26	gru	gru	NOUN
easat-10389	148	27	for	for	ADP
easat-10389	148	28	meteorological	meteorological	ADJ
easat-10389	148	29	data	datum	NOUN
easat-10389	148	30	,	,	PUNCT
easat-10389	148	31	and	and	CCONJ
easat-10389	148	32	a	a	DET
easat-10389	148	33	fusion	fusion	NOUN
easat-10389	148	34	layer	layer	NOUN
easat-10389	148	35	that	that	PRON
easat-10389	148	36	combines	combine	VERB
easat-10389	148	37	the	the	DET
easat-10389	148	38	two	two	NUM
easat-10389	148	39	to	to	PART
easat-10389	148	40	determine	determine	VERB
easat-10389	148	41	the	the	DET
easat-10389	148	42	likelihood	likelihood	NOUN
easat-10389	148	43	of	of	ADP
easat-10389	148	44	someone	someone	PRON
easat-10389	148	45	having	have	VERB
easat-10389	148	46	asthma	asthma	NOUN
easat-10389	148	47	.	.	PUNCT
easat-10389	149	1	there	there	PRON
easat-10389	149	2	are	be	VERB
easat-10389	149	3	dropouts	dropout	NOUN
easat-10389	149	4	and	and	CCONJ
easat-10389	149	5	l2	l2	NOUN
easat-10389	149	6	regularization	regularization	NOUN
easat-10389	149	7	in	in	ADP
easat-10389	149	8	the	the	DET
easat-10389	149	9	architecture	architecture	NOUN
easat-10389	149	10	to	to	PART
easat-10389	149	11	prevent	prevent	VERB
easat-10389	149	12	overfitting	overfitting	NOUN
easat-10389	149	13	.	.	PUNCT
easat-10389	150	1	•	•	NOUN
easat-10389	150	2	this	this	DET
easat-10389	150	3	study	study	NOUN
easat-10389	150	4	employed	employ	VERB
easat-10389	150	5	this	this	DET
easat-10389	150	6	smaller	small	ADJ
easat-10389	150	7	version	version	NOUN
easat-10389	150	8	for	for	ADP
easat-10389	150	9	ablation	ablation	NOUN
easat-10389	150	10	study	study	NOUN
easat-10389	150	11	reasons	reason	NOUN
easat-10389	150	12	.	.	PUNCT
easat-10389	151	1	it	it	PRON
easat-10389	151	2	separates	separate	VERB
easat-10389	151	3	the	the	DET
easat-10389	151	4	gru	gru	NOUN
easat-10389	151	5	branch	branch	NOUN
easat-10389	151	6	to	to	PART
easat-10389	151	7	examine	examine	VERB
easat-10389	151	8	how	how	SCONJ
easat-10389	151	9	weather	weather	NOUN
easat-10389	151	10	conditions	condition	NOUN
easat-10389	151	11	independently	independently	ADV
easat-10389	151	12	impact	impact	VERB
easat-10389	151	13	the	the	DET
easat-10389	151	14	prediction	prediction	NOUN
easat-10389	151	15	of	of	ADP
easat-10389	151	16	asthma	asthma	NOUN
easat-10389	151	17	risk	risk	NOUN
easat-10389	151	18	.	.	PUNCT
easat-10389	152	1	this	this	DET
easat-10389	152	2	study	study	NOUN
easat-10389	152	3	used	use	VERB
easat-10389	152	4	the	the	DET
easat-10389	152	5	adam	adam	PROPN
easat-10389	152	6	optimizer	optimizer	NOUN
easat-10389	152	7	with	with	ADP
easat-10389	152	8	a	a	DET
easat-10389	152	9	learning	learn	VERB
easat-10389	152	10	rate	rate	NOUN
easat-10389	152	11	of	of	ADP
easat-10389	152	12	0.001	0.001	NUM
easat-10389	152	13	to	to	PART
easat-10389	152	14	construct	construct	VERB
easat-10389	152	15	each	each	DET
easat-10389	152	16	model	model	NOUN
easat-10389	152	17	and	and	CCONJ
easat-10389	152	18	trained	train	VERB
easat-10389	152	19	them	they	PRON
easat-10389	152	20	using	use	VERB
easat-10389	152	21	binary	binary	PROPN
easat-10389	152	22	cross	cross	ADJ
easat-10389	152	23	-	-	ADJ
easat-10389	152	24	entropy	entropy	ADJ
easat-10389	152	25	loss	loss	NOUN
easat-10389	152	26	.	.	PUNCT
easat-10389	153	1	this	this	DET
easat-10389	153	2	study	study	NOUN
easat-10389	153	3	trained	train	VERB
easat-10389	153	4	the	the	DET
easat-10389	153	5	models	model	NOUN
easat-10389	153	6	for	for	ADP
easat-10389	153	7	20	20	NUM
easat-10389	153	8	epochs	epoch	NOUN
easat-10389	153	9	with	with	ADP
easat-10389	153	10	a	a	DET
easat-10389	153	11	batch	batch	NOUN
easat-10389	153	12	size	size	NOUN
easat-10389	153	13	of	of	ADP
easat-10389	153	14	64	64	NUM
easat-10389	153	15	and	and	CCONJ
easat-10389	153	16	used	use	VERB
easat-10389	153	17	a	a	DET
easat-10389	153	18	20	20	NUM
easat-10389	153	19	%	%	NOUN
easat-10389	153	20	validation	validation	NOUN
easat-10389	153	21	split	split	NOUN
easat-10389	153	22	to	to	PART
easat-10389	153	23	monitor	monitor	VERB
easat-10389	153	24	their	their	PRON
easat-10389	153	25	performance	performance	NOUN
easat-10389	153	26	.	.	PUNCT
easat-10389	154	1	this	this	DET
easat-10389	154	2	simulation	simulation	NOUN
easat-10389	154	3	did	do	AUX
easat-10389	154	4	not	not	PART
easat-10389	154	5	employ	employ	VERB
easat-10389	154	6	early	early	ADJ
easat-10389	154	7	stopping	stopping	NOUN
easat-10389	154	8	,	,	PUNCT
easat-10389	154	9	allowing	allow	VERB
easat-10389	154	10	all	all	DET
easat-10389	154	11	models	model	NOUN
easat-10389	154	12	to	to	PART
easat-10389	154	13	be	be	AUX
easat-10389	154	14	compared	compare	VERB
easat-10389	154	15	fairly	fairly	ADV
easat-10389	154	16	and	and	CCONJ
easat-10389	154	17	accurately	accurately	ADV
easat-10389	154	18	.	.	PUNCT
easat-10389	155	1	flowcharts	flowchart	NOUN
easat-10389	155	2	for	for	ADP
easat-10389	155	3	implementing	implement	VERB
easat-10389	155	4	models	model	NOUN
easat-10389	155	5	are	be	AUX
easat-10389	155	6	shown	show	VERB
easat-10389	155	7	in	in	ADP
easat-10389	155	8	figure	figure	NOUN
easat-10389	155	9	2	2	NUM
easat-10389	155	10	.	.	NUM
easat-10389	155	11	262	262	NUM
easat-10389	155	12	edelweiss	edelweiss	PROPN
easat-10389	155	13	applied	apply	VERB
easat-10389	155	14	science	science	NOUN
easat-10389	155	15	and	and	CCONJ
easat-10389	155	16	technology	technology	NOUN
easat-10389	155	17	issn	issn	PROPN
easat-10389	155	18	:	:	PUNCT
easat-10389	155	19	2576	2576	NUM
easat-10389	155	20	-	-	SYM
easat-10389	155	21	8484	8484	NUM
easat-10389	155	22	vol	vol	NOUN
easat-10389	155	23	.	.	PROPN
easat-10389	156	1	9	9	NUM
easat-10389	156	2	,	,	PUNCT
easat-10389	156	3	no	no	INTJ
easat-10389	156	4	.	.	NOUN
easat-10389	156	5	10	10	NUM
easat-10389	156	6	:	:	PUNCT
easat-10389	156	7	256	256	NUM
easat-10389	156	8	-	-	SYM
easat-10389	156	9	271	271	NUM
easat-10389	156	10	,	,	PUNCT
easat-10389	156	11	2025	2025	NUM
easat-10389	156	12	doi	doi	NOUN
easat-10389	156	13	:	:	PUNCT
easat-10389	156	14	10.55214/2576	10.55214/2576	NUM
easat-10389	156	15	-	-	PUNCT
easat-10389	156	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	156	17	©	©	ADP
easat-10389	156	18	2025	2025	NUM
easat-10389	156	19	by	by	ADP
easat-10389	156	20	the	the	DET
easat-10389	156	21	authors	author	NOUN
easat-10389	156	22	;	;	PUNCT
easat-10389	156	23	licensee	licensee	PROPN
easat-10389	156	24	learning	learning	NOUN
easat-10389	156	25	gate	gate	VERB
easat-10389	156	26	a.the	a.the	DET
easat-10389	156	27	suggested	suggest	VERB
easat-10389	156	28	cnn	cnn	PROPN
easat-10389	156	29	–	–	PUNCT
easat-10389	156	30	gru	gru	PROPN
easat-10389	156	31	–	–	PUNCT
easat-10389	156	32	attention	attention	NOUN
easat-10389	156	33	model	model	NOUN
easat-10389	156	34	combines	combine	VERB
easat-10389	156	35	spatial	spatial	ADJ
easat-10389	156	36	and	and	CCONJ
easat-10389	156	37	temporal	temporal	ADJ
easat-10389	156	38	data	datum	NOUN
easat-10389	156	39	.	.	PUNCT
easat-10389	157	1	b.the	b.the	DET
easat-10389	157	2	ablation	ablation	NOUN
easat-10389	157	3	version	version	NOUN
easat-10389	157	4	separates	separate	VERB
easat-10389	157	5	the	the	DET
easat-10389	157	6	gru	gru	NOUN
easat-10389	157	7	-	-	PUNCT
easat-10389	157	8	based	base	VERB
easat-10389	157	9	temporal	temporal	ADJ
easat-10389	157	10	branch	branch	NOUN
easat-10389	157	11	figure	figure	NOUN
easat-10389	157	12	2	2	NUM
easat-10389	157	13	.	.	PUNCT
easat-10389	157	14	flowcharts	flowchart	NOUN
easat-10389	157	15	for	for	ADP
easat-10389	157	16	implementing	implement	VERB
easat-10389	157	17	models	model	NOUN
easat-10389	157	18	.	.	PUNCT
easat-10389	158	1	the	the	DET
easat-10389	158	2	flowcharts	flowchart	NOUN
easat-10389	158	3	for	for	ADP
easat-10389	158	4	implementing	implement	VERB
easat-10389	158	5	the	the	DET
easat-10389	158	6	model	model	NOUN
easat-10389	158	7	are	be	AUX
easat-10389	158	8	shown	show	VERB
easat-10389	158	9	in	in	ADP
easat-10389	158	10	figure	figure	NOUN
easat-10389	158	11	2	2	NUM
easat-10389	158	12	.	.	PUNCT
easat-10389	158	13	figure	figure	NOUN
easat-10389	158	14	2a	2a	PROPN
easat-10389	158	15	illustrates	illustrate	VERB
easat-10389	158	16	the	the	DET
easat-10389	158	17	entire	entire	ADJ
easat-10389	158	18	cnn	cnn	PROPN
easat-10389	158	19	–	–	PUNCT
easat-10389	158	20	gru	gru	PROPN
easat-10389	158	21	–	–	PUNCT
easat-10389	158	22	attention	attention	NOUN
easat-10389	158	23	configuration	configuration	NOUN
easat-10389	158	24	.	.	PUNCT
easat-10389	159	1	the	the	DET
easat-10389	159	2	conv1d	conv1d	PROPN
easat-10389	159	3	layers	layer	NOUN
easat-10389	159	4	process	process	VERB
easat-10389	159	5	air	air	NOUN
easat-10389	159	6	quality	quality	PROPN
easat-10389	159	7	data	datum	NOUN
easat-10389	159	8	,	,	PUNCT
easat-10389	159	9	the	the	DET
easat-10389	159	10	gru	gru	NOUN
easat-10389	159	11	layers	layer	NOUN
easat-10389	159	12	handle	handle	VERB
easat-10389	159	13	meteorological	meteorological	ADJ
easat-10389	159	14	data	datum	NOUN
easat-10389	159	15	,	,	PUNCT
easat-10389	159	16	and	and	CCONJ
easat-10389	159	17	the	the	DET
easat-10389	159	18	attention	attention	NOUN
easat-10389	159	19	module	module	NOUN
easat-10389	159	20	combines	combine	VERB
easat-10389	159	21	the	the	DET
easat-10389	159	22	findings	finding	NOUN
easat-10389	159	23	.	.	PUNCT
easat-10389	160	1	the	the	DET
easat-10389	160	2	fused	fuse	VERB
easat-10389	160	3	output	output	NOUN
easat-10389	160	4	is	be	AUX
easat-10389	160	5	concatenated	concatenate	VERB
easat-10389	160	6	and	and	CCONJ
easat-10389	160	7	then	then	ADV
easat-10389	160	8	passed	pass	VERB
easat-10389	160	9	through	through	ADP
easat-10389	160	10	fully	fully	ADV
easat-10389	160	11	connected	connect	VERB
easat-10389	160	12	dense	dense	ADJ
easat-10389	160	13	layers	layer	NOUN
easat-10389	160	14	with	with	ADP
easat-10389	160	15	dropout	dropout	NOUN
easat-10389	160	16	and	and	CCONJ
easat-10389	160	17	l2	l2	NOUN
easat-10389	160	18	regularization	regularization	NOUN
easat-10389	160	19	.	.	PUNCT
easat-10389	161	1	figure	figure	NOUN
easat-10389	161	2	2b	2b	NUM
easat-10389	161	3	depicts	depict	VERB
easat-10389	161	4	the	the	DET
easat-10389	161	5	gru	gru	NOUN
easat-10389	161	6	-	-	ADJ
easat-10389	161	7	only	only	ADJ
easat-10389	161	8	ablation	ablation	NOUN
easat-10389	161	9	model	model	NOUN
easat-10389	161	10	,	,	PUNCT
easat-10389	161	11	which	which	PRON
easat-10389	161	12	examines	examine	VERB
easat-10389	161	13	only	only	ADV
easat-10389	161	14	weather	weather	NOUN
easat-10389	161	15	data	datum	NOUN
easat-10389	161	16	and	and	CCONJ
easat-10389	161	17	employs	employ	VERB
easat-10389	161	18	a	a	DET
easat-10389	161	19	temporal	temporal	ADJ
easat-10389	161	20	attention	attention	NOUN
easat-10389	161	21	mechanism	mechanism	NOUN
easat-10389	161	22	.	.	PUNCT
easat-10389	162	1	this	this	DET
easat-10389	162	2	simplified	simplified	ADJ
easat-10389	162	3	version	version	NOUN
easat-10389	162	4	omits	omit	VERB
easat-10389	162	5	the	the	DET
easat-10389	162	6	spatial	spatial	ADJ
easat-10389	162	7	branch	branch	NOUN
easat-10389	162	8	,	,	PUNCT
easat-10389	162	9	demonstrating	demonstrate	VERB
easat-10389	162	10	how	how	SCONJ
easat-10389	162	11	weather	weather	NOUN
easat-10389	162	12	patterns	pattern	NOUN
easat-10389	162	13	independently	independently	ADV
easat-10389	162	14	influence	influence	VERB
easat-10389	162	15	asthma	asthma	NOUN
easat-10389	162	16	risk	risk	NOUN
easat-10389	162	17	.	.	PUNCT
easat-10389	163	1	3.3.3	3.3.3	NUM
easat-10389	163	2	.	.	PUNCT
easat-10389	163	3	evaluation	evaluation	NOUN
easat-10389	163	4	protocol	protocol	NOUN
easat-10389	163	5	this	this	DET
easat-10389	163	6	study	study	NOUN
easat-10389	163	7	used	use	VERB
easat-10389	163	8	standard	standard	ADJ
easat-10389	163	9	classification	classification	NOUN
easat-10389	163	10	metrics	metric	NOUN
easat-10389	163	11	,	,	PUNCT
easat-10389	163	12	including	include	VERB
easat-10389	163	13	accuracy	accuracy	NOUN
easat-10389	163	14	,	,	PUNCT
easat-10389	163	15	precision	precision	NOUN
easat-10389	163	16	,	,	PUNCT
easat-10389	163	17	recall	recall	NOUN
easat-10389	163	18	,	,	PUNCT
easat-10389	163	19	f1	f1	NOUN
easat-10389	163	20	-	-	PUNCT
easat-10389	163	21	score	score	NOUN
easat-10389	163	22	,	,	PUNCT
easat-10389	163	23	and	and	CCONJ
easat-10389	163	24	auc	auc	NOUN
easat-10389	163	25	,	,	PUNCT
easat-10389	163	26	to	to	PART
easat-10389	163	27	evaluate	evaluate	VERB
easat-10389	163	28	the	the	DET
easat-10389	163	29	model	model	NOUN
easat-10389	163	30	's	's	PART
easat-10389	163	31	performance	performance	NOUN
easat-10389	163	32	.	.	PUNCT
easat-10389	164	1	predictions	prediction	NOUN
easat-10389	164	2	were	be	AUX
easat-10389	164	3	made	make	VERB
easat-10389	164	4	on	on	ADP
easat-10389	164	5	a	a	DET
easat-10389	164	6	test	test	NOUN
easat-10389	164	7	set	set	VERB
easat-10389	164	8	separate	separate	ADJ
easat-10389	164	9	from	from	ADP
easat-10389	164	10	the	the	DET
easat-10389	164	11	training	training	NOUN
easat-10389	164	12	data	datum	NOUN
easat-10389	164	13	(	(	PUNCT
easat-10389	164	14	20	20	NUM
easat-10389	164	15	%	%	NOUN
easat-10389	164	16	of	of	ADP
easat-10389	164	17	the	the	DET
easat-10389	164	18	total	total	NOUN
easat-10389	164	19	)	)	PUNCT
easat-10389	164	20	,	,	PUNCT
easat-10389	164	21	with	with	ADP
easat-10389	164	22	results	result	NOUN
easat-10389	164	23	averaged	average	VERB
easat-10389	164	24	over	over	ADP
easat-10389	164	25	five	five	NUM
easat-10389	164	26	cross	cross	ADJ
easat-10389	164	27	-	-	ADJ
easat-10389	164	28	validation	validation	ADJ
easat-10389	164	29	folds	fold	NOUN
easat-10389	164	30	to	to	PART
easat-10389	164	31	improve	improve	VERB
easat-10389	164	32	statistical	statistical	ADJ
easat-10389	164	33	reliability	reliability	NOUN
easat-10389	164	34	.	.	PUNCT
easat-10389	165	1	python	python	NOUN
easat-10389	165	2	modules	module	NOUN
easat-10389	165	3	such	such	ADJ
easat-10389	165	4	as	as	ADP
easat-10389	165	5	matplotlib	matplotlib	PROPN
easat-10389	165	6	,	,	PUNCT
easat-10389	165	7	scikit	scikit	NOUN
easat-10389	165	8	-	-	PUNCT
easat-10389	165	9	learn	learn	VERB
easat-10389	165	10	,	,	PUNCT
easat-10389	165	11	and	and	CCONJ
easat-10389	165	12	pandas	panda	NOUN
easat-10389	165	13	were	be	AUX
easat-10389	165	14	employed	employ	VERB
easat-10389	165	15	to	to	PART
easat-10389	165	16	analyze	analyze	VERB
easat-10389	165	17	and	and	CCONJ
easat-10389	165	18	evaluate	evaluate	VERB
easat-10389	165	19	all	all	DET
easat-10389	165	20	simulation	simulation	NOUN
easat-10389	165	21	data	datum	NOUN
easat-10389	165	22	.	.	PUNCT
easat-10389	166	1	the	the	DET
easat-10389	166	2	complete	complete	ADJ
easat-10389	166	3	code	code	NOUN
easat-10389	166	4	is	be	AUX
easat-10389	166	5	available	available	ADJ
easat-10389	166	6	as	as	ADP
easat-10389	166	7	supplementary	supplementary	ADJ
easat-10389	166	8	material	material	NOUN
easat-10389	166	9	for	for	ADP
easat-10389	166	10	repeated	repeat	VERB
easat-10389	166	11	testing	testing	NOUN
easat-10389	166	12	.	.	PUNCT
easat-10389	167	1	a	a	DET
easat-10389	167	2	gpu	gpu	NOUN
easat-10389	167	3	with	with	ADP
easat-10389	167	4	24	24	NUM
easat-10389	167	5	gb	gb	NOUN
easat-10389	167	6	of	of	ADP
easat-10389	167	7	vram	vram	NOUN
easat-10389	167	8	was	be	AUX
easat-10389	167	9	utilized	utilize	VERB
easat-10389	167	10	to	to	PART
easat-10389	167	11	ensure	ensure	VERB
easat-10389	167	12	efficient	efficient	ADJ
easat-10389	167	13	training	training	NOUN
easat-10389	167	14	and	and	CCONJ
easat-10389	167	15	testing	testing	NOUN
easat-10389	167	16	.	.	PUNCT
easat-10389	168	1	the	the	DET
easat-10389	168	2	goal	goal	NOUN
easat-10389	168	3	was	be	AUX
easat-10389	168	4	to	to	PART
easat-10389	168	5	validate	validate	VERB
easat-10389	168	6	the	the	DET
easat-10389	168	7	conceptual	conceptual	ADJ
easat-10389	168	8	framework	framework	NOUN
easat-10389	168	9	through	through	ADP
easat-10389	168	10	simulations	simulation	NOUN
easat-10389	168	11	that	that	PRON
easat-10389	168	12	could	could	AUX
easat-10389	168	13	be	be	AUX
easat-10389	168	14	repeated	repeat	VERB
easat-10389	168	15	without	without	ADP
easat-10389	168	16	requiring	require	VERB
easat-10389	168	17	real	real	ADJ
easat-10389	168	18	-	-	PUNCT
easat-10389	168	19	time	time	NOUN
easat-10389	168	20	deployment	deployment	NOUN
easat-10389	168	21	or	or	CCONJ
easat-10389	168	22	physical	physical	ADJ
easat-10389	168	23	hardware	hardware	NOUN
easat-10389	168	24	.	.	PUNCT
easat-10389	169	1	3.3.4	3.3.4	X
easat-10389	169	2	.	.	PUNCT
easat-10389	169	3	figure	figure	NOUN
easat-10389	169	4	integration	integration	NOUN
easat-10389	169	5	this	this	DET
easat-10389	169	6	study	study	NOUN
easat-10389	169	7	presents	present	VERB
easat-10389	169	8	two	two	NUM
easat-10389	169	9	models	model	NOUN
easat-10389	169	10	that	that	PRON
easat-10389	169	11	can	can	AUX
easat-10389	169	12	be	be	AUX
easat-10389	169	13	used	use	VERB
easat-10389	169	14	to	to	PART
easat-10389	169	15	evaluate	evaluate	VERB
easat-10389	169	16	the	the	DET
easat-10389	169	17	effectiveness	effectiveness	NOUN
easat-10389	169	18	of	of	ADP
easat-10389	169	19	alternative	alternative	ADJ
easat-10389	169	20	designs	design	NOUN
easat-10389	169	21	in	in	ADP
easat-10389	169	22	processing	process	VERB
easat-10389	169	23	sequential	sequential	ADJ
easat-10389	169	24	data	datum	NOUN
easat-10389	169	25	.	.	PUNCT
easat-10389	170	1	figure	figure	NOUN
easat-10389	170	2	3	3	NUM
easat-10389	170	3	illustrates	illustrate	VERB
easat-10389	170	4	the	the	DET
easat-10389	170	5	test	test	NOUN
easat-10389	170	6	results	result	NOUN
easat-10389	170	7	.	.	PUNCT
easat-10389	171	1	the	the	DET
easat-10389	171	2	dual	dual	ADJ
easat-10389	171	3	-	-	PUNCT
easat-10389	171	4	branch	branch	NOUN
easat-10389	171	5	model	model	NOUN
easat-10389	171	6	features	feature	VERB
easat-10389	171	7	a	a	DET
easat-10389	171	8	new	new	ADJ
easat-10389	171	9	parallel	parallel	ADJ
easat-10389	171	10	structure	structure	NOUN
easat-10389	171	11	that	that	PRON
easat-10389	171	12	enables	enable	VERB
easat-10389	171	13	it	it	PRON
easat-10389	171	14	to	to	PART
easat-10389	171	15	handle	handle	VERB
easat-10389	171	16	both	both	CCONJ
easat-10389	171	17	temporal	temporal	ADJ
easat-10389	171	18	and	and	CCONJ
easat-10389	171	19	contextual	contextual	ADJ
easat-10389	171	20	information	information	NOUN
easat-10389	171	21	simultaneously	simultaneously	ADV
easat-10389	171	22	.	.	PUNCT
easat-10389	172	1	the	the	DET
easat-10389	172	2	gru	gru	NOUN
easat-10389	172	3	-	-	ADJ
easat-10389	172	4	alone	alone	ADJ
easat-10389	172	5	ablation	ablation	NOUN
easat-10389	172	6	model	model	NOUN
easat-10389	172	7	,	,	PUNCT
easat-10389	172	8	on	on	ADP
easat-10389	172	9	the	the	DET
easat-10389	172	10	other	other	ADJ
easat-10389	172	11	hand	hand	NOUN
easat-10389	172	12	,	,	PUNCT
easat-10389	172	13	focuses	focus	VERB
easat-10389	172	14	exclusively	exclusively	ADV
easat-10389	172	15	on	on	ADP
easat-10389	172	16	temporal	temporal	ADJ
easat-10389	172	17	dynamics	dynamic	NOUN
easat-10389	172	18	with	with	ADP
easat-10389	172	19	a	a	DET
easat-10389	172	20	conventional	conventional	ADJ
easat-10389	172	21	gru	gru	NOUN
easat-10389	172	22	configuration	configuration	NOUN
easat-10389	172	23	.	.	PUNCT
easat-10389	173	1	tests	test	NOUN
easat-10389	173	2	on	on	ADP
easat-10389	173	3	specific	specific	ADJ
easat-10389	173	4	simulation	simulation	NOUN
easat-10389	173	5	tasks	task	NOUN
easat-10389	173	6	demonstrate	demonstrate	VERB
easat-10389	173	7	that	that	SCONJ
easat-10389	173	8	the	the	DET
easat-10389	173	9	dual	dual	ADJ
easat-10389	173	10	-	-	PUNCT
easat-10389	173	11	branch	branch	NOUN
easat-10389	173	12	architecture	architecture	NOUN
easat-10389	173	13	excels	excel	VERB
easat-10389	173	14	at	at	ADP
easat-10389	173	15	handling	handle	VERB
easat-10389	173	16	complex	complex	ADJ
easat-10389	173	17	data	datum	NOUN
easat-10389	173	18	,	,	PUNCT
easat-10389	173	19	as	as	SCONJ
easat-10389	173	20	indicated	indicate	VERB
easat-10389	173	21	by	by	ADP
easat-10389	173	22	this	this	DET
easat-10389	173	23	difference	difference	NOUN
easat-10389	173	24	.	.	PUNCT
easat-10389	174	1	the	the	DET
easat-10389	174	2	left	left	ADJ
easat-10389	174	3	side	side	NOUN
easat-10389	174	4	of	of	ADP
easat-10389	174	5	the	the	DET
easat-10389	174	6	diagram	diagram	NOUN
easat-10389	174	7	illustrates	illustrate	VERB
easat-10389	174	8	the	the	DET
easat-10389	174	9	proposed	propose	VERB
easat-10389	174	10	dual	dual	ADJ
easat-10389	174	11	-	-	PUNCT
easat-10389	174	12	branch	branch	NOUN
easat-10389	174	13	structure	structure	NOUN
easat-10389	174	14	,	,	PUNCT
easat-10389	174	15	comprising	comprise	VERB
easat-10389	174	16	two	two	NUM
easat-10389	174	17	parallel	parallel	ADJ
easat-10389	174	18	paths	path	NOUN
easat-10389	174	19	:	:	PUNCT
easat-10389	174	20	a	a	DET
easat-10389	174	21	cnn	cnn	NOUN
easat-10389	174	22	that	that	PRON
easat-10389	174	23	extracts	extract	VERB
easat-10389	174	24	key	key	ADJ
easat-10389	174	25	information	information	NOUN
easat-10389	174	26	from	from	ADP
easat-10389	174	27	air	air	NOUN
easat-10389	174	28	pollution	pollution	NOUN
easat-10389	174	29	data	datum	NOUN
easat-10389	174	30	and	and	CCONJ
easat-10389	174	31	a	a	DET
easat-10389	174	32	gru	gru	NOUN
easat-10389	174	33	network	network	NOUN
easat-10389	174	34	that	that	PRON
easat-10389	174	35	identifies	identify	VERB
easat-10389	174	36	patterns	pattern	NOUN
easat-10389	174	37	in	in	ADP
easat-10389	174	38	meteorological	meteorological	ADJ
easat-10389	174	39	data	datum	NOUN
easat-10389	174	40	over	over	ADP
easat-10389	174	41	time	time	NOUN
easat-10389	174	42	.	.	PUNCT
easat-10389	175	1	a	a	DET
easat-10389	175	2	fusion	fusion	NOUN
easat-10389	175	3	layer	layer	NOUN
easat-10389	175	4	combines	combine	VERB
easat-10389	175	5	the	the	DET
easat-10389	175	6	results	result	NOUN
easat-10389	175	7	from	from	ADP
easat-10389	175	8	both	both	DET
easat-10389	175	9	branches	branch	NOUN
easat-10389	175	10	.	.	PUNCT
easat-10389	176	1	263	263	NUM
easat-10389	176	2	edelweiss	edelweiss	PROPN
easat-10389	176	3	applied	apply	VERB
easat-10389	176	4	science	science	NOUN
easat-10389	176	5	and	and	CCONJ
easat-10389	176	6	technology	technology	NOUN
easat-10389	176	7	issn	issn	PROPN
easat-10389	176	8	:	:	PUNCT
easat-10389	176	9	2576	2576	NUM
easat-10389	176	10	-	-	SYM
easat-10389	176	11	8484	8484	NUM
easat-10389	176	12	vol	vol	NOUN
easat-10389	176	13	.	.	PROPN
easat-10389	177	1	9	9	NUM
easat-10389	177	2	,	,	PUNCT
easat-10389	177	3	no	no	INTJ
easat-10389	177	4	.	.	NOUN
easat-10389	177	5	10	10	NUM
easat-10389	177	6	:	:	PUNCT
easat-10389	177	7	256	256	NUM
easat-10389	177	8	-	-	SYM
easat-10389	177	9	271	271	NUM
easat-10389	177	10	,	,	PUNCT
easat-10389	177	11	2025	2025	NUM
easat-10389	177	12	doi	doi	NOUN
easat-10389	177	13	:	:	PUNCT
easat-10389	177	14	10.55214/2576	10.55214/2576	NUM
easat-10389	177	15	-	-	PUNCT
easat-10389	177	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	177	17	©	©	ADP
easat-10389	177	18	2025	2025	NUM
easat-10389	177	19	by	by	ADP
easat-10389	177	20	the	the	DET
easat-10389	177	21	authors	author	NOUN
easat-10389	177	22	;	;	PUNCT
easat-10389	177	23	licensee	licensee	PROPN
easat-10389	177	24	learning	learning	NOUN
easat-10389	177	25	gate	gate	NOUN
easat-10389	177	26	then	then	ADV
easat-10389	177	27	,	,	PUNCT
easat-10389	177	28	the	the	DET
easat-10389	177	29	data	data	NOUN
easat-10389	177	30	is	be	AUX
easat-10389	177	31	analyzed	analyze	VERB
easat-10389	177	32	again	again	ADV
easat-10389	177	33	to	to	PART
easat-10389	177	34	assess	assess	VERB
easat-10389	177	35	the	the	DET
easat-10389	177	36	risk	risk	NOUN
easat-10389	177	37	of	of	ADP
easat-10389	177	38	asthma	asthma	NOUN
easat-10389	177	39	.	.	PUNCT
easat-10389	178	1	the	the	DET
easat-10389	178	2	right	right	ADJ
easat-10389	178	3	side	side	NOUN
easat-10389	178	4	of	of	ADP
easat-10389	178	5	the	the	DET
easat-10389	178	6	diagram	diagram	NOUN
easat-10389	178	7	,	,	PUNCT
easat-10389	178	8	however	however	ADV
easat-10389	178	9	,	,	PUNCT
easat-10389	178	10	illustrates	illustrate	VERB
easat-10389	178	11	the	the	DET
easat-10389	178	12	simplified	simplified	ADJ
easat-10389	178	13	gru	gru	NOUN
easat-10389	178	14	-	-	ADJ
easat-10389	178	15	only	only	ADV
easat-10389	178	16	design	design	NOUN
easat-10389	178	17	used	use	VERB
easat-10389	178	18	in	in	ADP
easat-10389	178	19	the	the	DET
easat-10389	178	20	ablation	ablation	NOUN
easat-10389	178	21	study	study	NOUN
easat-10389	178	22	.	.	PUNCT
easat-10389	179	1	this	this	DET
easat-10389	179	2	version	version	NOUN
easat-10389	179	3	employs	employ	VERB
easat-10389	179	4	a	a	DET
easat-10389	179	5	gru	gru	NOUN
easat-10389	179	6	network	network	NOUN
easat-10389	179	7	to	to	PART
easat-10389	179	8	process	process	VERB
easat-10389	179	9	weather	weather	NOUN
easat-10389	179	10	data	datum	NOUN
easat-10389	179	11	.	.	PUNCT
easat-10389	180	1	it	it	PRON
easat-10389	180	2	does	do	AUX
easat-10389	180	3	not	not	PART
easat-10389	180	4	include	include	VERB
easat-10389	180	5	air	air	NOUN
easat-10389	180	6	quality	quality	NOUN
easat-10389	180	7	input	input	NOUN
easat-10389	180	8	or	or	CCONJ
easat-10389	180	9	dual	dual	ADJ
easat-10389	180	10	-	-	PUNCT
easat-10389	180	11	branch	branch	NOUN
easat-10389	180	12	fusion	fusion	NOUN
easat-10389	180	13	.	.	PUNCT
easat-10389	181	1	this	this	DET
easat-10389	181	2	design	design	NOUN
easat-10389	181	3	facilitates	facilitate	VERB
easat-10389	181	4	a	a	DET
easat-10389	181	5	clearer	clear	ADJ
easat-10389	181	6	understanding	understanding	NOUN
easat-10389	181	7	of	of	ADP
easat-10389	181	8	how	how	SCONJ
easat-10389	181	9	meteorological	meteorological	ADJ
easat-10389	181	10	factors	factor	NOUN
easat-10389	181	11	impact	impact	VERB
easat-10389	181	12	the	the	DET
easat-10389	181	13	forecast	forecast	NOUN
easat-10389	181	14	output	output	NOUN
easat-10389	181	15	.	.	PUNCT
easat-10389	182	1	figure	figure	NOUN
easat-10389	182	2	3	3	NUM
easat-10389	182	3	compares	compare	VERB
easat-10389	182	4	the	the	DET
easat-10389	182	5	complete	complete	ADJ
easat-10389	182	6	dual	dual	ADJ
easat-10389	182	7	-	-	PUNCT
easat-10389	182	8	branch	branch	NOUN
easat-10389	182	9	model	model	NOUN
easat-10389	182	10	and	and	CCONJ
easat-10389	182	11	the	the	DET
easat-10389	182	12	gru	gru	PROPN
easat-10389	182	13	-	-	ADJ
easat-10389	182	14	only	only	ADJ
easat-10389	182	15	ablation	ablation	NOUN
easat-10389	182	16	model	model	NOUN
easat-10389	182	17	in	in	ADP
easat-10389	182	18	terms	term	NOUN
easat-10389	182	19	of	of	ADP
easat-10389	182	20	their	their	PRON
easat-10389	182	21	architecture	architecture	NOUN
easat-10389	182	22	for	for	ADP
easat-10389	182	23	predicting	predict	VERB
easat-10389	182	24	asthma	asthma	NOUN
easat-10389	182	25	risk	risk	NOUN
easat-10389	182	26	.	.	PUNCT
easat-10389	183	1	figure	figure	NOUN
easat-10389	183	2	3	3	NUM
easat-10389	183	3	.	.	PUNCT
easat-10389	183	4	implementation	implementation	NOUN
easat-10389	183	5	diagram	diagram	NOUN
easat-10389	183	6	of	of	ADP
easat-10389	183	7	the	the	DET
easat-10389	183	8	two	two	NUM
easat-10389	183	9	simulation	simulation	NOUN
easat-10389	183	10	scripts	script	NOUN
easat-10389	183	11	.	.	PUNCT
easat-10389	184	1	4	4	X
easat-10389	184	2	.	.	X
easat-10389	184	3	experimental	experimental	ADJ
easat-10389	184	4	setup	setup	NOUN
easat-10389	184	5	this	this	DET
easat-10389	184	6	section	section	NOUN
easat-10389	184	7	provides	provide	VERB
easat-10389	184	8	a	a	DET
easat-10389	184	9	comprehensive	comprehensive	ADJ
easat-10389	184	10	review	review	NOUN
easat-10389	184	11	of	of	ADP
easat-10389	184	12	the	the	DET
easat-10389	184	13	proposed	propose	VERB
easat-10389	184	14	dual	dual	ADJ
easat-10389	184	15	-	-	PUNCT
easat-10389	184	16	branch	branch	NOUN
easat-10389	184	17	deep	deep	ADJ
easat-10389	184	18	learning	learning	NOUN
easat-10389	184	19	model	model	NOUN
easat-10389	184	20	,	,	PUNCT
easat-10389	184	21	examining	examine	VERB
easat-10389	184	22	several	several	ADJ
easat-10389	184	23	key	key	ADJ
easat-10389	184	24	components	component	NOUN
easat-10389	184	25	in	in	ADP
easat-10389	184	26	detail	detail	NOUN
easat-10389	184	27	.	.	PUNCT
easat-10389	185	1	it	it	PRON
easat-10389	185	2	discusses	discuss	VERB
easat-10389	185	3	the	the	DET
easat-10389	185	4	assessment	assessment	NOUN
easat-10389	185	5	measures	measure	NOUN
easat-10389	185	6	used	use	VERB
easat-10389	185	7	to	to	PART
easat-10389	185	8	evaluate	evaluate	VERB
easat-10389	185	9	the	the	DET
easat-10389	185	10	model	model	NOUN
easat-10389	185	11	's	's	PART
easat-10389	185	12	performance	performance	NOUN
easat-10389	185	13	and	and	CCONJ
easat-10389	185	14	accuracy	accuracy	NOUN
easat-10389	185	15	.	.	PUNCT
easat-10389	186	1	it	it	PRON
easat-10389	186	2	also	also	ADV
easat-10389	186	3	discusses	discuss	VERB
easat-10389	186	4	the	the	DET
easat-10389	186	5	baseline	baseline	NOUN
easat-10389	186	6	models	model	NOUN
easat-10389	186	7	used	use	VERB
easat-10389	186	8	for	for	ADP
easat-10389	186	9	comparison	comparison	NOUN
easat-10389	186	10	to	to	PART
easat-10389	186	11	establish	establish	VERB
easat-10389	186	12	a	a	DET
easat-10389	186	13	benchmark	benchmark	NOUN
easat-10389	186	14	.	.	PUNCT
easat-10389	187	1	the	the	DET
easat-10389	187	2	discussion	discussion	NOUN
easat-10389	187	3	also	also	ADV
easat-10389	187	4	delves	delve	VERB
easat-10389	187	5	into	into	ADP
easat-10389	187	6	great	great	ADJ
easat-10389	187	7	depth	depth	NOUN
easat-10389	187	8	on	on	ADP
easat-10389	187	9	the	the	DET
easat-10389	187	10	outcomes	outcome	NOUN
easat-10389	187	11	achieved	achieve	VERB
easat-10389	187	12	by	by	ADP
easat-10389	187	13	the	the	DET
easat-10389	187	14	model	model	NOUN
easat-10389	187	15	on	on	ADP
easat-10389	187	16	different	different	ADJ
easat-10389	187	17	datasets	dataset	NOUN
easat-10389	187	18	,	,	PUNCT
easat-10389	187	19	ensuring	ensure	VERB
easat-10389	187	20	that	that	SCONJ
easat-10389	187	21	the	the	DET
easat-10389	187	22	assessment	assessment	NOUN
easat-10389	187	23	takes	take	VERB
easat-10389	187	24	into	into	ADP
easat-10389	187	25	account	account	NOUN
easat-10389	187	26	the	the	DET
easat-10389	187	27	variety	variety	NOUN
easat-10389	187	28	of	of	ADP
easat-10389	187	29	data	data	NOUN
easat-10389	187	30	scenarios	scenario	NOUN
easat-10389	187	31	and	and	CCONJ
easat-10389	187	32	circumstances	circumstance	NOUN
easat-10389	187	33	.	.	PUNCT
easat-10389	188	1	the	the	DET
easat-10389	188	2	study	study	NOUN
easat-10389	188	3	also	also	ADV
easat-10389	188	4	examines	examine	VERB
easat-10389	188	5	the	the	DET
easat-10389	188	6	results	result	NOUN
easat-10389	188	7	of	of	ADP
easat-10389	188	8	several	several	ADJ
easat-10389	188	9	experimental	experimental	ADJ
easat-10389	188	10	setups	setup	NOUN
easat-10389	188	11	,	,	PUNCT
easat-10389	188	12	providing	provide	VERB
easat-10389	188	13	a	a	DET
easat-10389	188	14	deeper	deep	ADJ
easat-10389	188	15	understanding	understanding	NOUN
easat-10389	188	16	of	of	ADP
easat-10389	188	17	how	how	SCONJ
easat-10389	188	18	the	the	DET
easat-10389	188	19	model	model	NOUN
easat-10389	188	20	behaves	behave	VERB
easat-10389	188	21	and	and	CCONJ
easat-10389	188	22	its	its	PRON
easat-10389	188	23	ability	ability	NOUN
easat-10389	188	24	to	to	PART
easat-10389	188	25	adapt	adapt	VERB
easat-10389	188	26	to	to	ADP
easat-10389	188	27	diverse	diverse	ADJ
easat-10389	188	28	situations	situation	NOUN
easat-10389	188	29	.	.	PUNCT
easat-10389	189	1	overall	overall	ADV
easat-10389	189	2	,	,	PUNCT
easat-10389	189	3	this	this	DET
easat-10389	189	4	section	section	NOUN
easat-10389	189	5	aims	aim	VERB
easat-10389	189	6	to	to	PART
easat-10389	189	7	provide	provide	VERB
easat-10389	189	8	a	a	DET
easat-10389	189	9	comprehensive	comprehensive	ADJ
easat-10389	189	10	examination	examination	NOUN
easat-10389	189	11	of	of	ADP
easat-10389	189	12	the	the	DET
easat-10389	189	13	model	model	NOUN
easat-10389	189	14	's	's	PART
easat-10389	189	15	effectiveness	effectiveness	NOUN
easat-10389	189	16	,	,	PUNCT
easat-10389	189	17	utilizing	utilize	VERB
easat-10389	189	18	various	various	ADJ
easat-10389	189	19	assessment	assessment	NOUN
easat-10389	189	20	criteria	criterion	NOUN
easat-10389	189	21	,	,	PUNCT
easat-10389	189	22	comparison	comparison	NOUN
easat-10389	189	23	points	point	NOUN
easat-10389	189	24	,	,	PUNCT
easat-10389	189	25	and	and	CCONJ
easat-10389	189	26	insights	insight	NOUN
easat-10389	189	27	derived	derive	VERB
easat-10389	189	28	from	from	ADP
easat-10389	189	29	extensive	extensive	ADJ
easat-10389	189	30	testing	testing	NOUN
easat-10389	189	31	with	with	ADP
easat-10389	189	32	diverse	diverse	ADJ
easat-10389	189	33	datasets	dataset	NOUN
easat-10389	189	34	and	and	CCONJ
easat-10389	189	35	settings	setting	NOUN
easat-10389	189	36	.	.	PUNCT
easat-10389	190	1	4.1	4.1	NUM
easat-10389	190	2	.	.	PUNCT
easat-10389	190	3	metrics	metric	NOUN
easat-10389	190	4	for	for	ADP
easat-10389	190	5	evaluation	evaluation	NOUN
easat-10389	190	6	using	use	VERB
easat-10389	190	7	conventional	conventional	ADJ
easat-10389	190	8	classification	classification	NOUN
easat-10389	190	9	measures	measure	NOUN
easat-10389	190	10	,	,	PUNCT
easat-10389	190	11	this	this	DET
easat-10389	190	12	article	article	NOUN
easat-10389	190	13	examined	examine	VERB
easat-10389	190	14	how	how	SCONJ
easat-10389	190	15	well	well	ADV
easat-10389	190	16	the	the	DET
easat-10389	190	17	proposed	propose	VERB
easat-10389	190	18	model	model	NOUN
easat-10389	190	19	might	might	AUX
easat-10389	190	20	predict	predict	VERB
easat-10389	190	21	the	the	DET
easat-10389	190	22	risk	risk	NOUN
easat-10389	190	23	of	of	ADP
easat-10389	190	24	asthma	asthma	NOUN
easat-10389	190	25	.	.	PUNCT
easat-10389	191	1	•	•	NUM
easat-10389	191	2	accuracy	accuracy	NOUN
easat-10389	191	3	is	be	AUX
easat-10389	191	4	the	the	DET
easat-10389	191	5	percentage	percentage	NOUN
easat-10389	191	6	of	of	ADP
easat-10389	191	7	correct	correct	ADJ
easat-10389	191	8	predictions	prediction	NOUN
easat-10389	191	9	.	.	PUNCT
easat-10389	192	1	•	•	NUM
easat-10389	192	2	precision	precision	NOUN
easat-10389	192	3	demonstrates	demonstrate	VERB
easat-10389	192	4	the	the	DET
easat-10389	192	5	reliability	reliability	NOUN
easat-10389	192	6	of	of	ADP
easat-10389	192	7	high	high	ADJ
easat-10389	192	8	-	-	PUNCT
easat-10389	192	9	risk	risk	NOUN
easat-10389	192	10	forecasts	forecast	NOUN
easat-10389	192	11	in	in	ADP
easat-10389	192	12	identifying	identify	VERB
easat-10389	192	13	actual	actual	ADJ
easat-10389	192	14	asthma	asthma	NOUN
easat-10389	192	15	patients	patient	NOUN
easat-10389	192	16	.	.	PUNCT
easat-10389	193	1	•	•	NUM
easat-10389	193	2	recall	recall	NOUN
easat-10389	193	3	(	(	PUNCT
easat-10389	193	4	sensitivity	sensitivity	NOUN
easat-10389	193	5	)	)	PUNCT
easat-10389	193	6	tests	test	VERB
easat-10389	193	7	how	how	SCONJ
easat-10389	193	8	well	well	ADV
easat-10389	193	9	the	the	DET
easat-10389	193	10	model	model	NOUN
easat-10389	193	11	can	can	AUX
easat-10389	193	12	identify	identify	VERB
easat-10389	193	13	actual	actual	ADJ
easat-10389	193	14	high	high	ADJ
easat-10389	193	15	-	-	PUNCT
easat-10389	193	16	risk	risk	NOUN
easat-10389	193	17	instances	instance	NOUN
easat-10389	193	18	.	.	PUNCT
easat-10389	194	1	•	•	NUM
easat-10389	194	2	the	the	DET
easat-10389	194	3	f1	f1	NOUN
easat-10389	194	4	score	score	NOUN
easat-10389	194	5	is	be	AUX
easat-10389	194	6	the	the	DET
easat-10389	194	7	harmonic	harmonic	ADJ
easat-10389	194	8	mean	mean	NOUN
easat-10389	194	9	of	of	ADP
easat-10389	194	10	accuracy	accuracy	NOUN
easat-10389	194	11	and	and	CCONJ
easat-10389	194	12	recall	recall	NOUN
easat-10389	194	13	.	.	PUNCT
easat-10389	195	1	it	it	PRON
easat-10389	195	2	balances	balance	VERB
easat-10389	195	3	performance	performance	NOUN
easat-10389	195	4	,	,	PUNCT
easat-10389	195	5	especially	especially	ADV
easat-10389	195	6	when	when	SCONJ
easat-10389	195	7	the	the	DET
easat-10389	195	8	data	data	NOUN
easat-10389	195	9	is	be	AUX
easat-10389	195	10	not	not	PART
easat-10389	195	11	evenly	evenly	ADV
easat-10389	195	12	distributed	distribute	VERB
easat-10389	195	13	.	.	PUNCT
easat-10389	196	1	•	•	NUM
easat-10389	196	2	the	the	DET
easat-10389	196	3	auc	auc	NOUN
easat-10389	196	4	-	-	PUNCT
easat-10389	196	5	roc	roc	NOUN
easat-10389	196	6	indicates	indicate	VERB
easat-10389	196	7	how	how	SCONJ
easat-10389	196	8	well	well	ADV
easat-10389	196	9	the	the	DET
easat-10389	196	10	model	model	NOUN
easat-10389	196	11	distinguishes	distinguish	VERB
easat-10389	196	12	between	between	ADP
easat-10389	196	13	highand	highand	ADJ
easat-10389	196	14	low	low	ADJ
easat-10389	196	15	-	-	PUNCT
easat-10389	196	16	risk	risk	NOUN
easat-10389	196	17	groups	group	NOUN
easat-10389	196	18	at	at	ADP
easat-10389	196	19	various	various	ADJ
easat-10389	196	20	thresholds	threshold	NOUN
easat-10389	196	21	.	.	PUNCT
easat-10389	197	1	264	264	NUM
easat-10389	197	2	edelweiss	edelweiss	PROPN
easat-10389	197	3	applied	apply	VERB
easat-10389	197	4	science	science	NOUN
easat-10389	197	5	and	and	CCONJ
easat-10389	197	6	technology	technology	NOUN
easat-10389	197	7	issn	issn	PROPN
easat-10389	197	8	:	:	PUNCT
easat-10389	197	9	2576	2576	NUM
easat-10389	197	10	-	-	SYM
easat-10389	197	11	8484	8484	NUM
easat-10389	197	12	vol	vol	NOUN
easat-10389	197	13	.	.	PROPN
easat-10389	198	1	9	9	NUM
easat-10389	198	2	,	,	PUNCT
easat-10389	198	3	no	no	INTJ
easat-10389	198	4	.	.	NOUN
easat-10389	198	5	10	10	NUM
easat-10389	198	6	:	:	PUNCT
easat-10389	198	7	256	256	NUM
easat-10389	198	8	-	-	SYM
easat-10389	198	9	271	271	NUM
easat-10389	198	10	,	,	PUNCT
easat-10389	198	11	2025	2025	NUM
easat-10389	198	12	doi	doi	NOUN
easat-10389	198	13	:	:	PUNCT
easat-10389	198	14	10.55214/2576	10.55214/2576	NUM
easat-10389	198	15	-	-	PUNCT
easat-10389	198	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	198	17	©	©	ADP
easat-10389	198	18	2025	2025	NUM
easat-10389	198	19	by	by	ADP
easat-10389	198	20	the	the	DET
easat-10389	198	21	authors	author	NOUN
easat-10389	198	22	;	;	PUNCT
easat-10389	198	23	licensee	licensee	PROPN
easat-10389	198	24	learning	learning	NOUN
easat-10389	198	25	gate	gate	VERB
easat-10389	198	26	•	•	ADP
easat-10389	198	27	these	these	DET
easat-10389	198	28	measurements	measurement	NOUN
easat-10389	198	29	,	,	PUNCT
easat-10389	198	30	when	when	SCONJ
easat-10389	198	31	taken	take	VERB
easat-10389	198	32	together	together	ADV
easat-10389	198	33	,	,	PUNCT
easat-10389	198	34	provide	provide	VERB
easat-10389	198	35	a	a	DET
easat-10389	198	36	comprehensive	comprehensive	ADJ
easat-10389	198	37	picture	picture	NOUN
easat-10389	198	38	of	of	ADP
easat-10389	198	39	how	how	SCONJ
easat-10389	198	40	well	well	ADV
easat-10389	198	41	the	the	DET
easat-10389	198	42	model	model	NOUN
easat-10389	198	43	performs	perform	VERB
easat-10389	198	44	and	and	CCONJ
easat-10389	198	45	how	how	SCONJ
easat-10389	198	46	consistently	consistently	ADV
easat-10389	198	47	it	it	PRON
easat-10389	198	48	identifies	identify	VERB
easat-10389	198	49	asthma	asthma	NOUN
easat-10389	198	50	risk	risk	NOUN
easat-10389	198	51	in	in	ADP
easat-10389	198	52	diverse	diverse	ADJ
easat-10389	198	53	environments	environment	NOUN
easat-10389	198	54	.	.	PUNCT
easat-10389	199	1	4.2	4.2	NUM
easat-10389	199	2	.	.	PUNCT
easat-10389	199	3	baseline	baseline	NOUN
easat-10389	199	4	models	model	NOUN
easat-10389	199	5	to	to	PART
easat-10389	199	6	use	use	VERB
easat-10389	199	7	as	as	ADP
easat-10389	199	8	a	a	DET
easat-10389	199	9	point	point	NOUN
easat-10389	199	10	of	of	ADP
easat-10389	199	11	reference	reference	NOUN
easat-10389	199	12	this	this	DET
easat-10389	199	13	study	study	NOUN
easat-10389	199	14	examined	examine	VERB
easat-10389	199	15	the	the	DET
easat-10389	199	16	suggested	suggest	VERB
easat-10389	199	17	method	method	NOUN
easat-10389	199	18	against	against	ADP
easat-10389	199	19	baseline	baseline	NOUN
easat-10389	199	20	models	model	NOUN
easat-10389	199	21	to	to	PART
easat-10389	199	22	ensure	ensure	VERB
easat-10389	199	23	the	the	DET
easat-10389	199	24	results	result	NOUN
easat-10389	199	25	were	be	AUX
easat-10389	199	26	fair	fair	ADJ
easat-10389	199	27	,	,	PUNCT
easat-10389	199	28	which	which	PRON
easat-10389	199	29	were	be	AUX
easat-10389	199	30	trained	train	VERB
easat-10389	199	31	and	and	CCONJ
easat-10389	199	32	tested	test	VERB
easat-10389	199	33	under	under	ADP
easat-10389	199	34	the	the	DET
easat-10389	199	35	same	same	ADJ
easat-10389	199	36	settings	setting	NOUN
easat-10389	199	37	.	.	PUNCT
easat-10389	200	1	the	the	DET
easat-10389	200	2	baselines	baseline	NOUN
easat-10389	200	3	included	include	VERB
easat-10389	200	4	logistic	logistic	ADJ
easat-10389	200	5	regression	regression	NOUN
easat-10389	200	6	(	(	PUNCT
easat-10389	200	7	lr	lr	NOUN
easat-10389	200	8	)	)	PUNCT
easat-10389	200	9	,	,	PUNCT
easat-10389	200	10	a	a	DET
easat-10389	200	11	simple	simple	ADJ
easat-10389	200	12	statistical	statistical	ADJ
easat-10389	200	13	classifier	classifier	NOUN
easat-10389	200	14	;	;	PUNCT
easat-10389	200	15	support	support	NOUN
easat-10389	200	16	vector	vector	NOUN
easat-10389	200	17	machines	machine	NOUN
easat-10389	200	18	(	(	PUNCT
easat-10389	200	19	svm	svm	PROPN
easat-10389	200	20	)	)	PUNCT
easat-10389	200	21	,	,	PUNCT
easat-10389	200	22	which	which	PRON
easat-10389	200	23	model	model	VERB
easat-10389	200	24	nonlinear	nonlinear	ADJ
easat-10389	200	25	boundaries	boundary	NOUN
easat-10389	200	26	;	;	PUNCT
easat-10389	200	27	random	random	ADJ
easat-10389	200	28	forests	forest	NOUN
easat-10389	200	29	(	(	PUNCT
easat-10389	200	30	rf	rf	NOUN
easat-10389	200	31	)	)	PUNCT
easat-10389	200	32	,	,	PUNCT
easat-10389	200	33	a	a	DET
easat-10389	200	34	flexible	flexible	ADJ
easat-10389	200	35	ensemble	ensemble	ADJ
easat-10389	200	36	algorithm	algorithm	NOUN
easat-10389	200	37	;	;	PUNCT
easat-10389	200	38	and	and	CCONJ
easat-10389	200	39	a	a	DET
easat-10389	200	40	long	long	ADJ
easat-10389	200	41	short	short	ADJ
easat-10389	200	42	-	-	PUNCT
easat-10389	200	43	term	term	NOUN
easat-10389	200	44	memory	memory	NOUN
easat-10389	200	45	(	(	PUNCT
easat-10389	200	46	lstm	lstm	NOUN
easat-10389	200	47	)	)	PUNCT
easat-10389	200	48	network	network	NOUN
easat-10389	200	49	,	,	PUNCT
easat-10389	200	50	which	which	PRON
easat-10389	200	51	is	be	AUX
easat-10389	200	52	well	well	ADV
easat-10389	200	53	-	-	PUNCT
easat-10389	200	54	suited	suit	VERB
easat-10389	200	55	for	for	ADP
easat-10389	200	56	analyzing	analyze	VERB
easat-10389	200	57	time	time	NOUN
easat-10389	200	58	series	series	NOUN
easat-10389	200	59	and	and	CCONJ
easat-10389	200	60	predicting	predict	VERB
easat-10389	200	61	health	health	NOUN
easat-10389	200	62	outcomes	outcome	NOUN
easat-10389	200	63	.	.	PUNCT
easat-10389	201	1	this	this	DET
easat-10389	201	2	study	study	NOUN
easat-10389	201	3	examined	examine	VERB
easat-10389	201	4	two	two	NUM
easat-10389	201	5	ablated	ablated	ADJ
easat-10389	201	6	versions	version	NOUN
easat-10389	201	7	to	to	PART
easat-10389	201	8	determine	determine	VERB
easat-10389	201	9	the	the	DET
easat-10389	201	10	contribution	contribution	NOUN
easat-10389	201	11	of	of	ADP
easat-10389	201	12	each	each	DET
easat-10389	201	13	framework	framework	NOUN
easat-10389	201	14	branch	branch	NOUN
easat-10389	201	15	:	:	PUNCT
easat-10389	201	16	one	one	NUM
easat-10389	201	17	that	that	PRON
easat-10389	201	18	utilized	utilize	VERB
easat-10389	201	19	only	only	ADV
easat-10389	201	20	the	the	DET
easat-10389	201	21	cnn	cnn	PROPN
easat-10389	201	22	branch	branch	NOUN
easat-10389	201	23	and	and	CCONJ
easat-10389	201	24	the	the	DET
easat-10389	201	25	other	other	ADJ
easat-10389	201	26	that	that	PRON
easat-10389	201	27	used	use	VERB
easat-10389	201	28	only	only	ADV
easat-10389	201	29	the	the	DET
easat-10389	201	30	gru	gru	NOUN
easat-10389	201	31	branch	branch	NOUN
easat-10389	201	32	.	.	PUNCT
easat-10389	202	1	table	table	NOUN
easat-10389	202	2	1	1	NUM
easat-10389	202	3	presents	present	VERB
easat-10389	202	4	the	the	DET
easat-10389	202	5	results	result	NOUN
easat-10389	202	6	,	,	PUNCT
easat-10389	202	7	which	which	PRON
easat-10389	202	8	were	be	AUX
easat-10389	202	9	averaged	average	VERB
easat-10389	202	10	over	over	ADP
easat-10389	202	11	three	three	NUM
easat-10389	202	12	metropolitan	metropolitan	ADJ
easat-10389	202	13	datasets	dataset	NOUN
easat-10389	202	14	(	(	PUNCT
easat-10389	202	15	seoul	seoul	PROPN
easat-10389	202	16	,	,	PUNCT
easat-10389	202	17	los	los	PROPN
easat-10389	202	18	angeles	angeles	PROPN
easat-10389	202	19	,	,	PUNCT
easat-10389	202	20	and	and	CCONJ
easat-10389	202	21	hanoi	hanoi	PROPN
easat-10389	202	22	)	)	PUNCT
easat-10389	202	23	and	and	CCONJ
easat-10389	202	24	obtained	obtain	VERB
easat-10389	202	25	using	use	VERB
easat-10389	202	26	five	five	NUM
easat-10389	202	27	-	-	ADJ
easat-10389	202	28	fold	fold	ADJ
easat-10389	202	29	cross	cross	ADJ
easat-10389	202	30	-	-	ADJ
easat-10389	202	31	validation	validation	ADJ
easat-10389	202	32	and	and	CCONJ
easat-10389	202	33	temporal	temporal	ADJ
easat-10389	202	34	hold	hold	VERB
easat-10389	202	35	-	-	PUNCT
easat-10389	202	36	out	out	ADP
easat-10389	202	37	assessments	assessment	NOUN
easat-10389	202	38	on	on	ADP
easat-10389	202	39	the	the	DET
easat-10389	202	40	2024	2024	NUM
easat-10389	202	41	data	datum	NOUN
easat-10389	202	42	.	.	PUNCT
easat-10389	203	1	4.3	4.3	NUM
easat-10389	203	2	.	.	PUNCT
easat-10389	203	3	results	result	NOUN
easat-10389	203	4	and	and	CCONJ
easat-10389	203	5	study	study	VERB
easat-10389	203	6	4.3.1	4.3.1	NUM
easat-10389	203	7	model	model	NOUN
easat-10389	203	8	performance	performance	NOUN
easat-10389	203	9	table	table	NOUN
easat-10389	203	10	1	1	NUM
easat-10389	203	11	shows	show	VERB
easat-10389	203	12	that	that	SCONJ
easat-10389	203	13	our	our	PRON
easat-10389	203	14	dual	dual	ADJ
easat-10389	203	15	-	-	PUNCT
easat-10389	203	16	branch	branch	NOUN
easat-10389	203	17	model	model	NOUN
easat-10389	203	18	outperforms	outperform	VERB
easat-10389	203	19	both	both	PRON
easat-10389	203	20	standard	standard	ADJ
easat-10389	203	21	machine	machine	NOUN
easat-10389	203	22	learning	learning	NOUN
easat-10389	203	23	methods	method	NOUN
easat-10389	203	24	and	and	CCONJ
easat-10389	203	25	single	single	ADJ
easat-10389	203	26	-	-	PUNCT
easat-10389	203	27	branch	branch	NOUN
easat-10389	203	28	deep	deep	ADJ
easat-10389	203	29	learning	learn	VERB
easat-10389	203	30	baselines	baseline	NOUN
easat-10389	203	31	on	on	ADP
easat-10389	203	32	all	all	DET
easat-10389	203	33	assessment	assessment	NOUN
easat-10389	203	34	measures	measure	NOUN
easat-10389	203	35	.	.	PUNCT
easat-10389	204	1	the	the	DET
easat-10389	204	2	auc	auc	NOUN
easat-10389	204	3	of	of	ADP
easat-10389	204	4	0.89	0.89	NUM
easat-10389	204	5	indicates	indicate	VERB
easat-10389	204	6	that	that	SCONJ
easat-10389	204	7	our	our	PRON
easat-10389	204	8	model	model	NOUN
easat-10389	204	9	performs	perform	VERB
easat-10389	204	10	well	well	ADV
easat-10389	204	11	in	in	ADP
easat-10389	204	12	identifying	identify	VERB
easat-10389	204	13	days	day	NOUN
easat-10389	204	14	when	when	SCONJ
easat-10389	204	15	asthma	asthma	NOUN
easat-10389	204	16	is	be	AUX
easat-10389	204	17	likely	likely	ADJ
easat-10389	204	18	to	to	PART
easat-10389	204	19	be	be	AUX
easat-10389	204	20	high	high	ADJ
easat-10389	204	21	-	-	PUNCT
easat-10389	204	22	risk	risk	NOUN
easat-10389	204	23	,	,	PUNCT
easat-10389	204	24	and	and	CCONJ
easat-10389	204	25	the	the	DET
easat-10389	204	26	f1	f1	ADJ
easat-10389	204	27	score	score	NOUN
easat-10389	204	28	of	of	ADP
easat-10389	204	29	0.84	0.84	NUM
easat-10389	204	30	suggests	suggest	VERB
easat-10389	204	31	that	that	SCONJ
easat-10389	204	32	it	it	PRON
easat-10389	204	33	strikes	strike	VERB
easat-10389	204	34	a	a	DET
easat-10389	204	35	good	good	ADJ
easat-10389	204	36	balance	balance	NOUN
easat-10389	204	37	between	between	ADP
easat-10389	204	38	accuracy	accuracy	NOUN
easat-10389	204	39	and	and	CCONJ
easat-10389	204	40	completeness	completeness	NOUN
easat-10389	204	41	.	.	PUNCT
easat-10389	205	1	further	further	ADJ
easat-10389	205	2	research	research	NOUN
easat-10389	205	3	revealed	reveal	VERB
easat-10389	205	4	that	that	SCONJ
easat-10389	205	5	the	the	DET
easat-10389	205	6	attention	attention	NOUN
easat-10389	205	7	mechanism	mechanism	NOUN
easat-10389	205	8	had	have	VERB
easat-10389	205	9	a	a	DET
easat-10389	205	10	significant	significant	ADJ
easat-10389	205	11	impact	impact	NOUN
easat-10389	205	12	by	by	ADP
easat-10389	205	13	dynamically	dynamically	ADV
easat-10389	205	14	prioritizing	prioritize	VERB
easat-10389	205	15	the	the	DET
easat-10389	205	16	most	most	ADV
easat-10389	205	17	critical	critical	ADJ
easat-10389	205	18	elements	element	NOUN
easat-10389	205	19	in	in	ADP
easat-10389	205	20	both	both	DET
easat-10389	205	21	branches	branch	NOUN
easat-10389	205	22	.	.	PUNCT
easat-10389	206	1	the	the	DET
easat-10389	206	2	method	method	NOUN
easat-10389	206	3	worked	work	VERB
easat-10389	206	4	well	well	ADV
easat-10389	206	5	,	,	PUNCT
easat-10389	206	6	regardless	regardless	ADV
easat-10389	206	7	of	of	ADP
easat-10389	206	8	whether	whether	SCONJ
easat-10389	206	9	the	the	DET
easat-10389	206	10	weather	weather	NOUN
easat-10389	206	11	changed	change	VERB
easat-10389	206	12	quickly	quickly	ADV
easat-10389	206	13	or	or	CCONJ
easat-10389	206	14	pollution	pollution	NOUN
easat-10389	206	15	levels	level	NOUN
easat-10389	206	16	rose	rise	VERB
easat-10389	206	17	rapidly	rapidly	ADV
easat-10389	206	18	,	,	PUNCT
easat-10389	206	19	both	both	PRON
easat-10389	206	20	of	of	ADP
easat-10389	206	21	which	which	PRON
easat-10389	206	22	are	be	AUX
easat-10389	206	23	times	time	NOUN
easat-10389	206	24	when	when	SCONJ
easat-10389	206	25	asthma	asthma	NOUN
easat-10389	206	26	rates	rate	NOUN
easat-10389	206	27	tend	tend	VERB
easat-10389	206	28	to	to	PART
easat-10389	206	29	increase	increase	VERB
easat-10389	206	30	.	.	PUNCT
easat-10389	207	1	table	table	NOUN
easat-10389	207	2	1	1	NUM
easat-10389	207	3	.	.	PUNCT
easat-10389	207	4	summarizes	summarize	VERB
easat-10389	207	5	the	the	DET
easat-10389	207	6	results	result	NOUN
easat-10389	207	7	,	,	PUNCT
easat-10389	207	8	averaged	average	VERB
easat-10389	207	9	over	over	ADP
easat-10389	207	10	three	three	NUM
easat-10389	207	11	urban	urban	ADJ
easat-10389	207	12	datasets	dataset	NOUN
easat-10389	207	13	.	.	PUNCT
easat-10389	208	1	model	model	NOUN
easat-10389	208	2	accuracy	accuracy	PROPN
easat-10389	208	3	precision	precision	NOUN
easat-10389	208	4	recall	recall	VERB
easat-10389	208	5	f1	f1	NOUN
easat-10389	208	6	-	-	PUNCT
easat-10389	208	7	score	score	NOUN
easat-10389	208	8	auc	auc	NOUN
easat-10389	208	9	logistic	logistic	ADJ
easat-10389	208	10	regression	regression	NOUN
easat-10389	208	11	0.71	0.71	NUM
easat-10389	208	12	0.68	0.68	NUM
easat-10389	208	13	0.65	0.65	NUM
easat-10389	208	14	0.66	0.66	NUM
easat-10389	208	15	0.73	0.73	NUM
easat-10389	208	16	svm	svm	PROPN
easat-10389	208	17	0.75	0.75	NUM
easat-10389	208	18	0.72	0.72	NUM
easat-10389	208	19	0.70	0.70	NUM
easat-10389	208	20	0.71	0.71	NUM
easat-10389	208	21	0.77	0.77	NUM
easat-10389	208	22	random	random	ADJ
easat-10389	208	23	forest	forest	NOUN
easat-10389	208	24	0.79	0.79	NUM
easat-10389	208	25	0.76	0.76	NUM
easat-10389	208	26	0.74	0.74	NUM
easat-10389	208	27	0.75	0.75	NUM
easat-10389	208	28	0.81	0.81	NUM
easat-10389	208	29	lstm	lstm	NOUN
easat-10389	208	30	0.82	0.82	NUM
easat-10389	208	31	0.80	0.80	NUM
easat-10389	208	32	0.78	0.78	NUM
easat-10389	208	33	0.79	0.79	NUM
easat-10389	208	34	0.84	0.84	NUM
easat-10389	208	35	cnn	cnn	PROPN
easat-10389	208	36	-	-	PUNCT
easat-10389	208	37	only	only	ADV
easat-10389	208	38	0.80	0.80	NUM
easat-10389	208	39	0.78	0.78	NUM
easat-10389	208	40	0.75	0.75	NUM
easat-10389	208	41	0.76	0.76	NUM
easat-10389	208	42	0.82	0.82	NUM
easat-10389	208	43	gru	gru	NOUN
easat-10389	208	44	-	-	PUNCT
easat-10389	208	45	only	only	ADV
easat-10389	208	46	0.81	0.81	NUM
easat-10389	208	47	0.79	0.79	NUM
easat-10389	208	48	0.77	0.77	NUM
easat-10389	208	49	0.78	0.78	NUM
easat-10389	208	50	0.83	0.83	NUM
easat-10389	208	51	proposed	propose	VERB
easat-10389	208	52	model	model	NOUN
easat-10389	208	53	(	(	PUNCT
easat-10389	208	54	cnn	cnn	PROPN
easat-10389	208	55	+	+	PROPN
easat-10389	208	56	gru	gru	PROPN
easat-10389	208	57	+	+	CCONJ
easat-10389	208	58	attention	attention	NOUN
easat-10389	208	59	)	)	PUNCT
easat-10389	208	60	0.87	0.87	NUM
easat-10389	208	61	0.85	0.85	NUM
easat-10389	208	62	0.84	0.84	NUM
easat-10389	208	63	0.84	0.84	NUM
easat-10389	208	64	0.89	0.89	NUM
easat-10389	208	65	table	table	NOUN
easat-10389	208	66	1	1	NUM
easat-10389	208	67	shows	show	VERB
easat-10389	208	68	that	that	SCONJ
easat-10389	208	69	our	our	PRON
easat-10389	208	70	dual	dual	ADJ
easat-10389	208	71	-	-	PUNCT
easat-10389	208	72	branch	branch	NOUN
easat-10389	208	73	model	model	NOUN
easat-10389	208	74	outperforms	outperform	VERB
easat-10389	208	75	both	both	PRON
easat-10389	208	76	standard	standard	ADJ
easat-10389	208	77	machine	machine	NOUN
easat-10389	208	78	learning	learning	NOUN
easat-10389	208	79	methods	method	NOUN
easat-10389	208	80	and	and	CCONJ
easat-10389	208	81	single	single	ADJ
easat-10389	208	82	-	-	PUNCT
easat-10389	208	83	branch	branch	NOUN
easat-10389	208	84	deep	deep	ADJ
easat-10389	208	85	learning	learn	VERB
easat-10389	208	86	baselines	baseline	NOUN
easat-10389	208	87	on	on	ADP
easat-10389	208	88	all	all	DET
easat-10389	208	89	assessment	assessment	NOUN
easat-10389	208	90	measures	measure	NOUN
easat-10389	208	91	.	.	PUNCT
easat-10389	209	1	the	the	DET
easat-10389	209	2	auc	auc	NOUN
easat-10389	209	3	of	of	ADP
easat-10389	209	4	0.89	0.89	NUM
easat-10389	209	5	indicates	indicate	VERB
easat-10389	209	6	that	that	SCONJ
easat-10389	209	7	our	our	PRON
easat-10389	209	8	model	model	NOUN
easat-10389	209	9	performs	perform	VERB
easat-10389	209	10	well	well	ADV
easat-10389	209	11	in	in	ADP
easat-10389	209	12	identifying	identify	VERB
easat-10389	209	13	days	day	NOUN
easat-10389	209	14	when	when	SCONJ
easat-10389	209	15	asthma	asthma	NOUN
easat-10389	209	16	is	be	AUX
easat-10389	209	17	likely	likely	ADJ
easat-10389	209	18	to	to	PART
easat-10389	209	19	be	be	AUX
easat-10389	209	20	high	high	ADJ
easat-10389	209	21	-	-	PUNCT
easat-10389	209	22	risk	risk	NOUN
easat-10389	209	23	,	,	PUNCT
easat-10389	209	24	and	and	CCONJ
easat-10389	209	25	the	the	DET
easat-10389	209	26	f1	f1	NOUN
easat-10389	209	27	-	-	PUNCT
easat-10389	209	28	score	score	NOUN
easat-10389	209	29	of	of	ADP
easat-10389	209	30	0.84	0.84	NUM
easat-10389	209	31	suggests	suggest	VERB
easat-10389	209	32	that	that	SCONJ
easat-10389	209	33	it	it	PRON
easat-10389	209	34	strikes	strike	VERB
easat-10389	209	35	a	a	DET
easat-10389	209	36	good	good	ADJ
easat-10389	209	37	balance	balance	NOUN
easat-10389	209	38	between	between	ADP
easat-10389	209	39	accuracy	accuracy	NOUN
easat-10389	209	40	and	and	CCONJ
easat-10389	209	41	completeness	completeness	NOUN
easat-10389	209	42	.	.	PUNCT
easat-10389	210	1	more	more	ADJ
easat-10389	210	2	research	research	NOUN
easat-10389	210	3	showed	show	VERB
easat-10389	210	4	that	that	SCONJ
easat-10389	210	5	the	the	DET
easat-10389	210	6	attention	attention	NOUN
easat-10389	210	7	mechanism	mechanism	NOUN
easat-10389	210	8	made	make	VERB
easat-10389	210	9	a	a	DET
easat-10389	210	10	significant	significant	ADJ
easat-10389	210	11	difference	difference	NOUN
easat-10389	210	12	by	by	ADP
easat-10389	210	13	dynamically	dynamically	ADV
easat-10389	210	14	prioritising	prioritise	VERB
easat-10389	210	15	the	the	DET
easat-10389	210	16	most	most	ADV
easat-10389	210	17	critical	critical	ADJ
easat-10389	210	18	elements	element	NOUN
easat-10389	210	19	in	in	ADP
easat-10389	210	20	both	both	DET
easat-10389	210	21	branches	branch	NOUN
easat-10389	210	22	.	.	PUNCT
easat-10389	211	1	the	the	DET
easat-10389	211	2	method	method	NOUN
easat-10389	211	3	worked	work	VERB
easat-10389	211	4	well	well	ADV
easat-10389	211	5	,	,	PUNCT
easat-10389	211	6	regardless	regardless	ADV
easat-10389	211	7	of	of	ADP
easat-10389	211	8	whether	whether	SCONJ
easat-10389	211	9	the	the	DET
easat-10389	211	10	weather	weather	NOUN
easat-10389	211	11	changed	change	VERB
easat-10389	211	12	quickly	quickly	ADV
easat-10389	211	13	or	or	CCONJ
easat-10389	211	14	pollution	pollution	NOUN
easat-10389	211	15	levels	level	NOUN
easat-10389	211	16	rose	rise	VERB
easat-10389	211	17	rapidly	rapidly	ADV
easat-10389	211	18	,	,	PUNCT
easat-10389	211	19	both	both	PRON
easat-10389	211	20	of	of	ADP
easat-10389	211	21	which	which	PRON
easat-10389	211	22	are	be	AUX
easat-10389	211	23	times	time	NOUN
easat-10389	211	24	when	when	SCONJ
easat-10389	211	25	asthma	asthma	NOUN
easat-10389	211	26	rates	rate	NOUN
easat-10389	211	27	tend	tend	VERB
easat-10389	211	28	to	to	PART
easat-10389	211	29	increase	increase	VERB
easat-10389	211	30	.	.	PUNCT
easat-10389	212	1	additionally	additionally	ADV
easat-10389	212	2	,	,	PUNCT
easat-10389	212	3	the	the	DET
easat-10389	212	4	model	model	NOUN
easat-10389	212	5	was	be	AUX
easat-10389	212	6	able	able	ADJ
easat-10389	212	7	to	to	PART
easat-10389	212	8	generalise	generalise	VERB
easat-10389	212	9	well	well	ADV
easat-10389	212	10	across	across	ADP
easat-10389	212	11	multiple	multiple	ADJ
easat-10389	212	12	cities	city	NOUN
easat-10389	212	13	,	,	PUNCT
easat-10389	212	14	even	even	ADV
easat-10389	212	15	when	when	SCONJ
easat-10389	212	16	trained	train	VERB
easat-10389	212	17	in	in	ADP
easat-10389	212	18	one	one	NUM
easat-10389	212	19	location	location	NOUN
easat-10389	212	20	and	and	CCONJ
easat-10389	212	21	evaluated	evaluate	VERB
easat-10389	212	22	in	in	ADP
easat-10389	212	23	another	another	PRON
easat-10389	212	24	.	.	PUNCT
easat-10389	213	1	the	the	DET
easat-10389	213	2	outcome	outcome	NOUN
easat-10389	213	3	supports	support	VERB
easat-10389	213	4	the	the	DET
easat-10389	213	5	framework	framework	NOUN
easat-10389	213	6	's	's	PART
easat-10389	213	7	ability	ability	NOUN
easat-10389	213	8	to	to	PART
easat-10389	213	9	adapt	adapt	VERB
easat-10389	213	10	to	to	ADP
easat-10389	213	11	various	various	ADJ
easat-10389	213	12	situations	situation	NOUN
easat-10389	213	13	and	and	CCONJ
easat-10389	213	14	its	its	PRON
easat-10389	213	15	potential	potential	NOUN
easat-10389	213	16	for	for	ADP
easat-10389	213	17	real	real	ADJ
easat-10389	213	18	-	-	PUNCT
easat-10389	213	19	world	world	NOUN
easat-10389	213	20	application	application	NOUN
easat-10389	213	21	,	,	PUNCT
easat-10389	213	22	provided	provide	VERB
easat-10389	213	23	it	it	PRON
easat-10389	213	24	can	can	AUX
easat-10389	213	25	be	be	AUX
easat-10389	213	26	connected	connect	VERB
easat-10389	213	27	to	to	PART
easat-10389	213	28	live	live	VERB
easat-10389	213	29	data	datum	NOUN
easat-10389	213	30	sources	source	NOUN
easat-10389	213	31	.	.	PUNCT
easat-10389	214	1	4.3.2	4.3.2	X
easat-10389	214	2	.	.	PUNCT
easat-10389	215	1	roc	roc	NOUN
easat-10389	215	2	curve	curve	NOUN
easat-10389	215	3	analysis	analysis	NOUN
easat-10389	215	4	figure	figure	NOUN
easat-10389	215	5	4	4	NUM
easat-10389	215	6	displays	display	VERB
easat-10389	215	7	the	the	DET
easat-10389	215	8	roc	roc	PROPN
easat-10389	215	9	curves	curve	NOUN
easat-10389	215	10	for	for	ADP
easat-10389	215	11	all	all	DET
easat-10389	215	12	the	the	DET
easat-10389	215	13	evaluated	evaluated	ADJ
easat-10389	215	14	models	model	NOUN
easat-10389	215	15	.	.	PUNCT
easat-10389	216	1	the	the	DET
easat-10389	216	2	dual	dual	ADJ
easat-10389	216	3	-	-	PUNCT
easat-10389	216	4	branch	branch	NOUN
easat-10389	216	5	cnn	cnn	PROPN
easat-10389	216	6	–	–	PUNCT
easat-10389	216	7	gru	gru	NOUN
easat-10389	216	8	framework	framework	NOUN
easat-10389	216	9	outperforms	outperform	VERB
easat-10389	216	10	basic	basic	ADJ
easat-10389	216	11	approaches	approach	NOUN
easat-10389	216	12	in	in	ADP
easat-10389	216	13	making	make	VERB
easat-10389	216	14	predictions	prediction	NOUN
easat-10389	216	15	.	.	PUNCT
easat-10389	217	1	the	the	DET
easat-10389	217	2	proposed	propose	VERB
easat-10389	217	3	model	model	NOUN
easat-10389	217	4	distinguishes	distinguish	VERB
easat-10389	217	5	between	between	ADP
easat-10389	217	6	highand	highand	ADJ
easat-10389	217	7	low	low	ADJ
easat-10389	217	8	-	-	PUNCT
easat-10389	217	9	risk	risk	NOUN
easat-10389	217	10	asthma	asthma	NOUN
easat-10389	217	11	patients	patient	NOUN
easat-10389	217	12	at	at	ADP
easat-10389	217	13	different	different	ADJ
easat-10389	217	14	thresholds	threshold	NOUN
easat-10389	217	15	,	,	PUNCT
easat-10389	217	16	with	with	ADP
easat-10389	217	17	an	an	DET
easat-10389	217	18	auc	auc	NOUN
easat-10389	217	19	of	of	ADP
easat-10389	217	20	0.89	0.89	NUM
easat-10389	217	21	.	.	PUNCT
easat-10389	218	1	it	it	PRON
easat-10389	218	2	surpasses	surpass	VERB
easat-10389	218	3	classic	classic	ADJ
easat-10389	218	4	machine	machine	NOUN
easat-10389	218	5	learning	learning	NOUN
easat-10389	218	6	models	model	NOUN
easat-10389	218	7	such	such	ADJ
easat-10389	218	8	as	as	ADP
easat-10389	218	9	logistic	logistic	ADJ
easat-10389	218	10	regression	regression	NOUN
easat-10389	218	11	(	(	PUNCT
easat-10389	218	12	auc	auc	NOUN
easat-10389	218	13	=	=	NOUN
easat-10389	218	14	0.73	0.73	NUM
easat-10389	218	15	)	)	PUNCT
easat-10389	218	16	,	,	PUNCT
easat-10389	218	17	svm	svm	PROPN
easat-10389	218	18	(	(	PUNCT
easat-10389	218	19	0.77	0.77	NUM
easat-10389	218	20	)	)	PUNCT
easat-10389	218	21	,	,	PUNCT
easat-10389	218	22	and	and	CCONJ
easat-10389	218	23	random	random	ADJ
easat-10389	218	24	265	265	NUM
easat-10389	218	25	edelweiss	edelweiss	PROPN
easat-10389	218	26	applied	apply	VERB
easat-10389	218	27	science	science	NOUN
easat-10389	218	28	and	and	CCONJ
easat-10389	218	29	technology	technology	NOUN
easat-10389	218	30	issn	issn	PROPN
easat-10389	218	31	:	:	PUNCT
easat-10389	218	32	2576	2576	NUM
easat-10389	218	33	-	-	SYM
easat-10389	218	34	8484	8484	NUM
easat-10389	218	35	vol	vol	NOUN
easat-10389	218	36	.	.	PROPN
easat-10389	219	1	9	9	NUM
easat-10389	219	2	,	,	PUNCT
easat-10389	219	3	no	no	INTJ
easat-10389	219	4	.	.	NOUN
easat-10389	219	5	10	10	NUM
easat-10389	219	6	:	:	PUNCT
easat-10389	219	7	256	256	NUM
easat-10389	219	8	-	-	SYM
easat-10389	219	9	271	271	NUM
easat-10389	219	10	,	,	PUNCT
easat-10389	219	11	2025	2025	NUM
easat-10389	219	12	doi	doi	NOUN
easat-10389	219	13	:	:	PUNCT
easat-10389	219	14	10.55214/2576	10.55214/2576	NUM
easat-10389	219	15	-	-	PUNCT
easat-10389	219	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	219	17	©	©	ADP
easat-10389	219	18	2025	2025	NUM
easat-10389	219	19	by	by	ADP
easat-10389	219	20	the	the	DET
easat-10389	219	21	authors	author	NOUN
easat-10389	219	22	;	;	PUNCT
easat-10389	219	23	licensee	licensee	PROPN
easat-10389	219	24	learning	learn	VERB
easat-10389	219	25	gate	gate	PROPN
easat-10389	219	26	forest	forest	PROPN
easat-10389	219	27	(	(	PUNCT
easat-10389	219	28	0.81	0.81	NUM
easat-10389	219	29	)	)	PUNCT
easat-10389	219	30	,	,	PUNCT
easat-10389	219	31	as	as	ADV
easat-10389	219	32	well	well	ADV
easat-10389	219	33	as	as	ADP
easat-10389	219	34	deep	deep	ADJ
easat-10389	219	35	learning	learning	NOUN
easat-10389	219	36	models	model	NOUN
easat-10389	219	37	that	that	PRON
easat-10389	219	38	solely	solely	ADV
easat-10389	219	39	utilize	utilize	VERB
easat-10389	219	40	cnn	cnn	PROPN
easat-10389	219	41	(	(	PUNCT
easat-10389	219	42	0.82	0.82	NUM
easat-10389	219	43	)	)	PUNCT
easat-10389	219	44	or	or	CCONJ
easat-10389	219	45	gru	gru	X
easat-10389	219	46	(	(	PUNCT
easat-10389	219	47	0.83	0.83	NUM
easat-10389	219	48	)	)	PUNCT
easat-10389	219	49	.	.	PUNCT
easat-10389	220	1	this	this	PRON
easat-10389	220	2	demonstrates	demonstrate	VERB
easat-10389	220	3	the	the	DET
easat-10389	220	4	value	value	NOUN
easat-10389	220	5	of	of	ADP
easat-10389	220	6	integrating	integrate	VERB
easat-10389	220	7	spatial	spatial	ADJ
easat-10389	220	8	elements	element	NOUN
easat-10389	220	9	from	from	ADP
easat-10389	220	10	air	air	NOUN
easat-10389	220	11	pollution	pollution	NOUN
easat-10389	220	12	data	datum	NOUN
easat-10389	220	13	with	with	ADP
easat-10389	220	14	temporal	temporal	ADJ
easat-10389	220	15	meteorological	meteorological	ADJ
easat-10389	220	16	dynamics	dynamic	NOUN
easat-10389	220	17	.	.	PUNCT
easat-10389	221	1	the	the	DET
easat-10389	221	2	gru	gru	NOUN
easat-10389	221	3	-	-	ADJ
easat-10389	221	4	only	only	ADJ
easat-10389	221	5	model	model	NOUN
easat-10389	221	6	and	and	CCONJ
easat-10389	221	7	the	the	DET
easat-10389	221	8	entire	entire	ADJ
easat-10389	221	9	model	model	NOUN
easat-10389	221	10	do	do	AUX
easat-10389	221	11	not	not	PART
easat-10389	221	12	perform	perform	VERB
easat-10389	221	13	very	very	ADV
easat-10389	221	14	differently	differently	ADV
easat-10389	221	15	,	,	PUNCT
easat-10389	221	16	but	but	CCONJ
easat-10389	221	17	this	this	PRON
easat-10389	221	18	suggests	suggest	VERB
easat-10389	221	19	that	that	SCONJ
easat-10389	221	20	weather	weather	NOUN
easat-10389	221	21	data	datum	NOUN
easat-10389	221	22	may	may	AUX
easat-10389	221	23	be	be	AUX
easat-10389	221	24	more	more	ADV
easat-10389	221	25	important	important	ADJ
easat-10389	221	26	for	for	ADP
easat-10389	221	27	the	the	DET
easat-10389	221	28	predictions	prediction	NOUN
easat-10389	221	29	.	.	PUNCT
easat-10389	222	1	this	this	PRON
easat-10389	222	2	indicates	indicate	VERB
easat-10389	222	3	the	the	DET
easat-10389	222	4	need	need	NOUN
easat-10389	222	5	to	to	PART
easat-10389	222	6	find	find	VERB
easat-10389	222	7	more	more	ADV
easat-10389	222	8	effective	effective	ADJ
easat-10389	222	9	ways	way	NOUN
easat-10389	222	10	to	to	PART
easat-10389	222	11	learn	learn	VERB
easat-10389	222	12	from	from	ADP
easat-10389	222	13	spatial	spatial	ADJ
easat-10389	222	14	characteristics	characteristic	NOUN
easat-10389	222	15	or	or	CCONJ
easat-10389	222	16	to	to	PART
easat-10389	222	17	better	well	ADV
easat-10389	222	18	integrate	integrate	VERB
easat-10389	222	19	the	the	DET
easat-10389	222	20	data	datum	NOUN
easat-10389	222	21	.	.	PUNCT
easat-10389	223	1	the	the	DET
easat-10389	223	2	dualbranch	dualbranch	ADJ
easat-10389	223	3	architecture	architecture	NOUN
easat-10389	223	4	performs	perform	VERB
easat-10389	223	5	well	well	ADV
easat-10389	223	6	overall	overall	ADJ
easat-10389	223	7	,	,	PUNCT
easat-10389	223	8	although	although	SCONJ
easat-10389	223	9	further	further	ADJ
easat-10389	223	10	fine	fine	ADV
easat-10389	223	11	-	-	PUNCT
easat-10389	223	12	tuning	tuning	NOUN
easat-10389	223	13	may	may	AUX
easat-10389	223	14	be	be	AUX
easat-10389	223	15	necessary	necessary	ADJ
easat-10389	223	16	.	.	PUNCT
easat-10389	224	1	figure	figure	VERB
easat-10389	224	2	4	4	NUM
easat-10389	224	3	.	.	PUNCT
easat-10389	225	1	roc	roc	PROPN
easat-10389	225	2	curves	curve	NOUN
easat-10389	225	3	that	that	PRON
easat-10389	225	4	illustrate	illustrate	VERB
easat-10389	225	5	the	the	DET
easat-10389	225	6	predictive	predictive	ADJ
easat-10389	225	7	accuracy	accuracy	NOUN
easat-10389	225	8	of	of	ADP
easat-10389	225	9	the	the	DET
easat-10389	225	10	baseline	baseline	NOUN
easat-10389	225	11	and	and	CCONJ
easat-10389	225	12	suggested	suggest	VERB
easat-10389	225	13	models	model	NOUN
easat-10389	225	14	.	.	PUNCT
easat-10389	226	1	figure	figure	NOUN
easat-10389	226	2	5	5	NUM
easat-10389	226	3	illustrates	illustrate	VERB
easat-10389	226	4	the	the	DET
easat-10389	226	5	performance	performance	NOUN
easat-10389	226	6	of	of	ADP
easat-10389	226	7	all	all	DET
easat-10389	226	8	models	model	NOUN
easat-10389	226	9	based	base	VERB
easat-10389	226	10	on	on	ADP
easat-10389	226	11	the	the	DET
easat-10389	226	12	f1	f1	NOUN
easat-10389	226	13	score	score	NOUN
easat-10389	226	14	.	.	PUNCT
easat-10389	227	1	the	the	DET
easat-10389	227	2	proposed	propose	VERB
easat-10389	227	3	dual	dual	ADJ
easat-10389	227	4	-	-	PUNCT
easat-10389	227	5	branch	branch	NOUN
easat-10389	227	6	cnn	cnn	PROPN
easat-10389	227	7	–	–	PUNCT
easat-10389	227	8	gru	gru	NOUN
easat-10389	227	9	architecture	architecture	NOUN
easat-10389	227	10	achieved	achieve	VERB
easat-10389	227	11	the	the	DET
easat-10389	227	12	highest	high	ADJ
easat-10389	227	13	score	score	NOUN
easat-10389	227	14	of	of	ADP
easat-10389	227	15	0.84	0.84	NUM
easat-10389	227	16	,	,	PUNCT
easat-10389	227	17	indicating	indicate	VERB
easat-10389	227	18	a	a	DET
easat-10389	227	19	better	well	ADJ
easat-10389	227	20	balance	balance	NOUN
easat-10389	227	21	of	of	ADP
easat-10389	227	22	precision	precision	NOUN
easat-10389	227	23	and	and	CCONJ
easat-10389	227	24	recall	recall	NOUN
easat-10389	227	25	,	,	PUNCT
easat-10389	227	26	which	which	PRON
easat-10389	227	27	is	be	AUX
easat-10389	227	28	crucial	crucial	ADJ
easat-10389	227	29	for	for	ADP
easat-10389	227	30	health	health	NOUN
easat-10389	227	31	-	-	PUNCT
easat-10389	227	32	related	relate	VERB
easat-10389	227	33	predictions	prediction	NOUN
easat-10389	227	34	,	,	PUNCT
easat-10389	227	35	as	as	SCONJ
easat-10389	227	36	errors	error	NOUN
easat-10389	227	37	can	can	AUX
easat-10389	227	38	have	have	VERB
easat-10389	227	39	severe	severe	ADJ
easat-10389	227	40	consequences	consequence	NOUN
easat-10389	227	41	.	.	PUNCT
easat-10389	228	1	deep	deep	ADJ
easat-10389	228	2	learning	learn	VERB
easat-10389	228	3	baselines	baseline	NOUN
easat-10389	228	4	,	,	PUNCT
easat-10389	228	5	such	such	ADJ
easat-10389	228	6	as	as	ADP
easat-10389	228	7	the	the	DET
easat-10389	228	8	gru	gru	VERB
easat-10389	228	9	-	-	ADJ
easat-10389	228	10	only	only	ADJ
easat-10389	228	11	model	model	NOUN
easat-10389	228	12	(	(	PUNCT
easat-10389	228	13	0.78	0.78	NUM
easat-10389	228	14	)	)	PUNCT
easat-10389	228	15	and	and	CCONJ
easat-10389	228	16	cnn	cnn	PROPN
easat-10389	228	17	-	-	PUNCT
easat-10389	228	18	only	only	ADV
easat-10389	228	19	model	model	NOUN
easat-10389	228	20	(	(	PUNCT
easat-10389	228	21	0.76	0.76	NUM
easat-10389	228	22	)	)	PUNCT
easat-10389	228	23	,	,	PUNCT
easat-10389	228	24	performed	perform	VERB
easat-10389	228	25	well	well	ADV
easat-10389	228	26	but	but	CCONJ
easat-10389	228	27	not	not	PART
easat-10389	228	28	sufficiently	sufficiently	ADV
easat-10389	228	29	,	,	PUNCT
easat-10389	228	30	suggesting	suggest	VERB
easat-10389	228	31	that	that	SCONJ
easat-10389	228	32	combining	combine	VERB
easat-10389	228	33	spatial	spatial	ADJ
easat-10389	228	34	and	and	CCONJ
easat-10389	228	35	temporal	temporal	ADJ
easat-10389	228	36	variables	variable	NOUN
easat-10389	228	37	may	may	AUX
easat-10389	228	38	be	be	AUX
easat-10389	228	39	advantageous	advantageous	ADJ
easat-10389	228	40	.	.	PUNCT
easat-10389	229	1	traditional	traditional	ADJ
easat-10389	229	2	machine	machine	NOUN
easat-10389	229	3	learning	learning	NOUN
easat-10389	229	4	models	model	NOUN
easat-10389	229	5	,	,	PUNCT
easat-10389	229	6	including	include	VERB
easat-10389	229	7	logistic	logistic	ADJ
easat-10389	229	8	regression	regression	NOUN
easat-10389	229	9	(	(	PUNCT
easat-10389	229	10	0.66	0.66	NUM
easat-10389	229	11	)	)	PUNCT
easat-10389	229	12	,	,	PUNCT
easat-10389	229	13	svm	svm	PROPN
easat-10389	229	14	(	(	PUNCT
easat-10389	229	15	0.71	0.71	NUM
easat-10389	229	16	)	)	PUNCT
easat-10389	229	17	,	,	PUNCT
easat-10389	229	18	and	and	CCONJ
easat-10389	229	19	random	random	ADJ
easat-10389	229	20	forests	forest	NOUN
easat-10389	229	21	(	(	PUNCT
easat-10389	229	22	0.75	0.75	NUM
easat-10389	229	23	)	)	PUNCT
easat-10389	229	24	,	,	PUNCT
easat-10389	229	25	underperformed	underperform	VERB
easat-10389	229	26	relative	relative	ADJ
easat-10389	229	27	to	to	ADP
easat-10389	229	28	deep	deep	ADJ
easat-10389	229	29	learning	learning	NOUN
easat-10389	229	30	models	model	NOUN
easat-10389	229	31	.	.	PUNCT
easat-10389	230	1	this	this	DET
easat-10389	230	2	highlights	highlight	VERB
easat-10389	230	3	the	the	DET
easat-10389	230	4	limitations	limitation	NOUN
easat-10389	230	5	of	of	ADP
easat-10389	230	6	handcrafted	handcraft	VERB
easat-10389	230	7	features	feature	NOUN
easat-10389	230	8	in	in	ADP
easat-10389	230	9	capturing	capture	VERB
easat-10389	230	10	complex	complex	ADJ
easat-10389	230	11	correlations	correlation	NOUN
easat-10389	230	12	between	between	ADP
easat-10389	230	13	environmental	environmental	ADJ
easat-10389	230	14	health	health	NOUN
easat-10389	230	15	factors	factor	NOUN
easat-10389	230	16	and	and	CCONJ
easat-10389	230	17	other	other	ADJ
easat-10389	230	18	variables	variable	NOUN
easat-10389	230	19	.	.	PUNCT
easat-10389	231	1	the	the	DET
easat-10389	231	2	slight	slight	ADJ
easat-10389	231	3	performance	performance	NOUN
easat-10389	231	4	difference	difference	NOUN
easat-10389	231	5	between	between	ADP
easat-10389	231	6	the	the	DET
easat-10389	231	7	gru	gru	VERB
easat-10389	231	8	-	-	ADJ
easat-10389	231	9	only	only	ADJ
easat-10389	231	10	model	model	NOUN
easat-10389	231	11	and	and	CCONJ
easat-10389	231	12	the	the	DET
easat-10389	231	13	complete	complete	ADJ
easat-10389	231	14	model	model	NOUN
easat-10389	231	15	indicates	indicate	VERB
easat-10389	231	16	that	that	SCONJ
easat-10389	231	17	weather	weather	NOUN
easat-10389	231	18	patterns	pattern	NOUN
easat-10389	231	19	significantly	significantly	ADV
easat-10389	231	20	influence	influence	VERB
easat-10389	231	21	prediction	prediction	NOUN
easat-10389	231	22	accuracy	accuracy	NOUN
easat-10389	231	23	.	.	PUNCT
easat-10389	232	1	this	this	PRON
easat-10389	232	2	suggests	suggest	VERB
easat-10389	232	3	that	that	SCONJ
easat-10389	232	4	the	the	DET
easat-10389	232	5	spatial	spatial	ADJ
easat-10389	232	6	component	component	NOUN
easat-10389	232	7	of	of	ADP
easat-10389	232	8	the	the	DET
easat-10389	232	9	model	model	NOUN
easat-10389	232	10	could	could	AUX
easat-10389	232	11	benefit	benefit	VERB
easat-10389	232	12	from	from	ADP
easat-10389	232	13	deeper	deep	ADJ
easat-10389	232	14	or	or	CCONJ
easat-10389	232	15	attention	attention	NOUN
easat-10389	232	16	-	-	PUNCT
easat-10389	232	17	based	base	VERB
easat-10389	232	18	architectures	architecture	NOUN
easat-10389	232	19	.	.	PUNCT
easat-10389	233	1	the	the	DET
easat-10389	233	2	f1	f1	ADJ
easat-10389	233	3	-	-	PUNCT
easat-10389	233	4	score	score	NOUN
easat-10389	233	5	analysis	analysis	NOUN
easat-10389	233	6	demonstrates	demonstrate	VERB
easat-10389	233	7	that	that	SCONJ
easat-10389	233	8	the	the	DET
easat-10389	233	9	proposed	propose	VERB
easat-10389	233	10	model	model	NOUN
easat-10389	233	11	is	be	AUX
easat-10389	233	12	robust	robust	ADJ
easat-10389	233	13	and	and	CCONJ
easat-10389	233	14	identifies	identify	VERB
easat-10389	233	15	potential	potential	ADJ
easat-10389	233	16	areas	area	NOUN
easat-10389	233	17	for	for	ADP
easat-10389	233	18	further	further	ADJ
easat-10389	233	19	improvement	improvement	NOUN
easat-10389	233	20	.	.	PUNCT
easat-10389	234	1	266	266	NUM
easat-10389	234	2	edelweiss	edelweiss	PROPN
easat-10389	234	3	applied	apply	VERB
easat-10389	234	4	science	science	NOUN
easat-10389	234	5	and	and	CCONJ
easat-10389	234	6	technology	technology	NOUN
easat-10389	234	7	issn	issn	PROPN
easat-10389	234	8	:	:	PUNCT
easat-10389	234	9	2576	2576	NUM
easat-10389	234	10	-	-	SYM
easat-10389	234	11	8484	8484	NUM
easat-10389	234	12	vol	vol	NOUN
easat-10389	234	13	.	.	PROPN
easat-10389	235	1	9	9	NUM
easat-10389	235	2	,	,	PUNCT
easat-10389	235	3	no	no	INTJ
easat-10389	235	4	.	.	NOUN
easat-10389	235	5	10	10	NUM
easat-10389	235	6	:	:	PUNCT
easat-10389	235	7	256	256	NUM
easat-10389	235	8	-	-	SYM
easat-10389	235	9	271	271	NUM
easat-10389	235	10	,	,	PUNCT
easat-10389	235	11	2025	2025	NUM
easat-10389	235	12	doi	doi	NOUN
easat-10389	235	13	:	:	PUNCT
easat-10389	235	14	10.55214/2576	10.55214/2576	NUM
easat-10389	235	15	-	-	PUNCT
easat-10389	235	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	235	17	©	©	ADP
easat-10389	235	18	2025	2025	NUM
easat-10389	235	19	by	by	ADP
easat-10389	235	20	the	the	DET
easat-10389	235	21	authors	author	NOUN
easat-10389	235	22	;	;	PUNCT
easat-10389	235	23	licensee	licensee	PROPN
easat-10389	235	24	learning	learn	VERB
easat-10389	235	25	gate	gate	PROPN
easat-10389	235	26	figure	figure	NOUN
easat-10389	235	27	5	5	NUM
easat-10389	235	28	.	.	PUNCT
easat-10389	235	29	f1	f1	NOUN
easat-10389	235	30	-	-	PUNCT
easat-10389	235	31	score	score	NOUN
easat-10389	235	32	comparison	comparison	NOUN
easat-10389	235	33	across	across	ADP
easat-10389	235	34	all	all	DET
easat-10389	235	35	evaluated	evaluated	ADJ
easat-10389	235	36	models	model	NOUN
easat-10389	235	37	.	.	PUNCT
easat-10389	236	1	4.3.3	4.3.3	X
easat-10389	236	2	.	.	PUNCT
easat-10389	236	3	confusion	confusion	NOUN
easat-10389	236	4	matrix	matrix	NOUN
easat-10389	236	5	interpretation	interpretation	NOUN
easat-10389	236	6	figure	figure	NOUN
easat-10389	236	7	6	6	NUM
easat-10389	236	8	.	.	PUNCT
easat-10389	237	1	the	the	DET
easat-10389	237	2	confusion	confusion	NOUN
easat-10389	237	3	matrix	matrix	NOUN
easat-10389	237	4	of	of	ADP
easat-10389	237	5	the	the	DET
easat-10389	237	6	proposed	propose	VERB
easat-10389	237	7	cnn	cnn	PROPN
easat-10389	237	8	-	-	PUNCT
easat-10389	237	9	gru	gru	NOUN
easat-10389	237	10	-	-	ADJ
easat-10389	237	11	attention	attention	NOUN
easat-10389	237	12	mode	mode	NOUN
easat-10389	237	13	.	.	PUNCT
easat-10389	238	1	figure	figure	NOUN
easat-10389	238	2	6	6	NUM
easat-10389	238	3	shows	show	VERB
easat-10389	238	4	the	the	DET
easat-10389	238	5	confusion	confusion	NOUN
easat-10389	238	6	matrix	matrix	NOUN
easat-10389	238	7	for	for	ADP
easat-10389	238	8	the	the	DET
easat-10389	238	9	suggested	suggest	VERB
easat-10389	238	10	cnn	cnn	PROPN
easat-10389	238	11	-	-	PUNCT
easat-10389	238	12	gru	gru	PROPN
easat-10389	238	13	-	-	PUNCT
easat-10389	238	14	attention	attention	NOUN
easat-10389	238	15	model	model	NOUN
easat-10389	238	16	.	.	PUNCT
easat-10389	239	1	it	it	PRON
easat-10389	239	2	demonstrates	demonstrate	VERB
easat-10389	239	3	its	its	PRON
easat-10389	239	4	ability	ability	NOUN
easat-10389	239	5	to	to	PART
easat-10389	239	6	distinguish	distinguish	VERB
easat-10389	239	7	between	between	ADP
easat-10389	239	8	high	high	ADJ
easat-10389	239	9	-	-	PUNCT
easat-10389	239	10	risk	risk	NOUN
easat-10389	239	11	and	and	CCONJ
easat-10389	239	12	low	low	ADJ
easat-10389	239	13	-	-	PUNCT
easat-10389	239	14	risk	risk	NOUN
easat-10389	239	15	asthma	asthma	NOUN
easat-10389	239	16	patients	patient	NOUN
easat-10389	239	17	effectively	effectively	ADV
easat-10389	239	18	.	.	PUNCT
easat-10389	240	1	the	the	DET
easat-10389	240	2	model	model	NOUN
easat-10389	240	3	successfully	successfully	ADV
easat-10389	240	4	identified	identify	VERB
easat-10389	240	5	45	45	NUM
easat-10389	240	6	out	out	ADP
easat-10389	240	7	of	of	ADP
easat-10389	240	8	50	50	NUM
easat-10389	240	9	high	high	ADJ
easat-10389	240	10	-	-	PUNCT
easat-10389	240	11	risk	risk	NOUN
easat-10389	240	12	samples	sample	NOUN
easat-10389	240	13	and	and	CCONJ
easat-10389	240	14	42	42	NUM
easat-10389	240	15	out	out	ADP
easat-10389	240	16	of	of	ADP
easat-10389	240	17	50	50	NUM
easat-10389	240	18	low	low	ADJ
easat-10389	240	19	-	-	PUNCT
easat-10389	240	20	risk	risk	NOUN
easat-10389	240	21	samples	sample	NOUN
easat-10389	240	22	,	,	PUNCT
easat-10389	240	23	indicating	indicate	VERB
easat-10389	240	24	that	that	SCONJ
easat-10389	240	25	it	it	PRON
easat-10389	240	26	was	be	AUX
easat-10389	240	27	both	both	CCONJ
easat-10389	240	28	sensitive	sensitive	ADJ
easat-10389	240	29	and	and	CCONJ
easat-10389	240	30	specific	specific	ADJ
easat-10389	240	31	for	for	ADP
easat-10389	240	32	both	both	DET
easat-10389	240	33	classifications	classification	NOUN
easat-10389	240	34	.	.	PUNCT
easat-10389	241	1	267	267	NUM
easat-10389	241	2	edelweiss	edelweiss	PROPN
easat-10389	241	3	applied	apply	VERB
easat-10389	241	4	science	science	NOUN
easat-10389	241	5	and	and	CCONJ
easat-10389	241	6	technology	technology	NOUN
easat-10389	241	7	issn	issn	PROPN
easat-10389	241	8	:	:	PUNCT
easat-10389	241	9	2576	2576	NUM
easat-10389	241	10	-	-	SYM
easat-10389	241	11	8484	8484	NUM
easat-10389	241	12	vol	vol	NOUN
easat-10389	241	13	.	.	PROPN
easat-10389	242	1	9	9	NUM
easat-10389	242	2	,	,	PUNCT
easat-10389	242	3	no	no	INTJ
easat-10389	242	4	.	.	NOUN
easat-10389	242	5	10	10	NUM
easat-10389	242	6	:	:	PUNCT
easat-10389	242	7	256	256	NUM
easat-10389	242	8	-	-	SYM
easat-10389	242	9	271	271	NUM
easat-10389	242	10	,	,	PUNCT
easat-10389	242	11	2025	2025	NUM
easat-10389	242	12	doi	doi	NOUN
easat-10389	242	13	:	:	PUNCT
easat-10389	242	14	10.55214/2576	10.55214/2576	NUM
easat-10389	242	15	-	-	PUNCT
easat-10389	242	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	242	17	©	©	ADP
easat-10389	242	18	2025	2025	NUM
easat-10389	242	19	by	by	ADP
easat-10389	242	20	the	the	DET
easat-10389	242	21	authors	author	NOUN
easat-10389	242	22	;	;	PUNCT
easat-10389	242	23	licensee	licensee	PROPN
easat-10389	242	24	learning	learning	NOUN
easat-10389	242	25	gate	gate	VERB
easat-10389	242	26	the	the	DET
easat-10389	242	27	confusion	confusion	NOUN
easat-10389	242	28	matrix	matrix	NOUN
easat-10389	242	29	demonstrates	demonstrate	VERB
easat-10389	242	30	that	that	SCONJ
easat-10389	242	31	the	the	DET
easat-10389	242	32	model	model	NOUN
easat-10389	242	33	performs	perform	VERB
easat-10389	242	34	well	well	ADV
easat-10389	242	35	in	in	ADP
easat-10389	242	36	classifying	classify	VERB
easat-10389	242	37	high	high	ADJ
easat-10389	242	38	-	-	PUNCT
easat-10389	242	39	risk	risk	NOUN
easat-10389	242	40	asthma	asthma	NOUN
easat-10389	242	41	cases	case	NOUN
easat-10389	242	42	.	.	PUNCT
easat-10389	243	1	it	it	PRON
easat-10389	243	2	correctly	correctly	ADV
easat-10389	243	3	identified	identify	VERB
easat-10389	243	4	45	45	NUM
easat-10389	243	5	high	high	ADJ
easat-10389	243	6	-	-	PUNCT
easat-10389	243	7	risk	risk	NOUN
easat-10389	243	8	instances	instance	NOUN
easat-10389	243	9	(	(	PUNCT
easat-10389	243	10	true	true	ADJ
easat-10389	243	11	positives	positive	NOUN
easat-10389	243	12	)	)	PUNCT
easat-10389	243	13	and	and	CCONJ
easat-10389	243	14	42	42	NUM
easat-10389	243	15	low	low	ADJ
easat-10389	243	16	-	-	PUNCT
easat-10389	243	17	risk	risk	NOUN
easat-10389	243	18	cases	case	NOUN
easat-10389	243	19	(	(	PUNCT
easat-10389	243	20	true	true	ADJ
easat-10389	243	21	negatives	negative	NOUN
easat-10389	243	22	)	)	PUNCT
easat-10389	243	23	out	out	ADP
easat-10389	243	24	of	of	ADP
easat-10389	243	25	100	100	NUM
easat-10389	243	26	test	test	NOUN
easat-10389	243	27	samples	sample	NOUN
easat-10389	243	28	.	.	PUNCT
easat-10389	244	1	however	however	ADV
easat-10389	244	2	,	,	PUNCT
easat-10389	244	3	it	it	PRON
easat-10389	244	4	also	also	ADV
easat-10389	244	5	misclassified	misclassifie	VERB
easat-10389	244	6	eight	eight	NUM
easat-10389	244	7	low	low	ADJ
easat-10389	244	8	-	-	PUNCT
easat-10389	244	9	risk	risk	NOUN
easat-10389	244	10	cases	case	NOUN
easat-10389	244	11	as	as	ADP
easat-10389	244	12	high	high	ADJ
easat-10389	244	13	-	-	PUNCT
easat-10389	244	14	risk	risk	NOUN
easat-10389	244	15	(	(	PUNCT
easat-10389	244	16	false	false	ADJ
easat-10389	244	17	positives	positive	NOUN
easat-10389	244	18	)	)	PUNCT
easat-10389	244	19	and	and	CCONJ
easat-10389	244	20	five	five	NUM
easat-10389	244	21	high	high	ADJ
easat-10389	244	22	-	-	PUNCT
easat-10389	244	23	risk	risk	NOUN
easat-10389	244	24	cases	case	NOUN
easat-10389	244	25	as	as	ADP
easat-10389	244	26	low	low	ADJ
easat-10389	244	27	-	-	PUNCT
easat-10389	244	28	risk	risk	NOUN
easat-10389	244	29	(	(	PUNCT
easat-10389	244	30	false	false	ADJ
easat-10389	244	31	negatives	negative	NOUN
easat-10389	244	32	)	)	PUNCT
easat-10389	244	33	.	.	PUNCT
easat-10389	245	1	the	the	DET
easat-10389	245	2	model	model	NOUN
easat-10389	245	3	achieved	achieve	VERB
easat-10389	245	4	87	87	NUM
easat-10389	245	5	%	%	NOUN
easat-10389	245	6	accuracy	accuracy	NOUN
easat-10389	245	7	,	,	PUNCT
easat-10389	245	8	with	with	ADP
easat-10389	245	9	84.9	84.9	NUM
easat-10389	245	10	%	%	NOUN
easat-10389	245	11	of	of	ADP
easat-10389	245	12	high	high	ADJ
easat-10389	245	13	-	-	PUNCT
easat-10389	245	14	risk	risk	NOUN
easat-10389	245	15	predictions	prediction	NOUN
easat-10389	245	16	and	and	CCONJ
easat-10389	245	17	90	90	NUM
easat-10389	245	18	%	%	NOUN
easat-10389	245	19	of	of	ADP
easat-10389	245	20	low	low	ADJ
easat-10389	245	21	-	-	PUNCT
easat-10389	245	22	risk	risk	NOUN
easat-10389	245	23	predictions	prediction	NOUN
easat-10389	245	24	being	be	AUX
easat-10389	245	25	correct	correct	ADJ
easat-10389	245	26	.	.	PUNCT
easat-10389	246	1	a	a	DET
easat-10389	246	2	macro	macro	NOUN
easat-10389	246	3	-	-	ADJ
easat-10389	246	4	averaged	average	VERB
easat-10389	246	5	f1	f1	NOUN
easat-10389	246	6	score	score	NOUN
easat-10389	246	7	of	of	ADP
easat-10389	246	8	0.87	0.87	NUM
easat-10389	246	9	indicates	indicate	VERB
easat-10389	246	10	that	that	SCONJ
easat-10389	246	11	the	the	DET
easat-10389	246	12	test	test	NOUN
easat-10389	246	13	effectively	effectively	ADV
easat-10389	246	14	balances	balance	VERB
easat-10389	246	15	sensitivity	sensitivity	NOUN
easat-10389	246	16	and	and	CCONJ
easat-10389	246	17	specificity	specificity	NOUN
easat-10389	246	18	.	.	PUNCT
easat-10389	247	1	nonetheless	nonetheless	ADV
easat-10389	247	2	,	,	PUNCT
easat-10389	247	3	the	the	DET
easat-10389	247	4	higher	high	ADJ
easat-10389	247	5	false	false	ADJ
easat-10389	247	6	positive	positive	ADJ
easat-10389	247	7	rate	rate	NOUN
easat-10389	247	8	suggests	suggest	VERB
easat-10389	247	9	that	that	SCONJ
easat-10389	247	10	people	people	NOUN
easat-10389	247	11	tend	tend	VERB
easat-10389	247	12	to	to	PART
easat-10389	247	13	overestimate	overestimate	VERB
easat-10389	247	14	danger	danger	NOUN
easat-10389	247	15	to	to	PART
easat-10389	247	16	avoid	avoid	VERB
easat-10389	247	17	missing	miss	VERB
easat-10389	247	18	essential	essential	ADJ
easat-10389	247	19	cases	case	NOUN
easat-10389	247	20	.	.	PUNCT
easat-10389	248	1	the	the	DET
easat-10389	248	2	five	five	NUM
easat-10389	248	3	false	false	ADJ
easat-10389	248	4	negatives	negative	NOUN
easat-10389	248	5	underscore	underscore	VERB
easat-10389	248	6	the	the	DET
easat-10389	248	7	need	need	NOUN
easat-10389	248	8	to	to	PART
easat-10389	248	9	enhance	enhance	VERB
easat-10389	248	10	the	the	DET
easat-10389	248	11	test	test	NOUN
easat-10389	248	12	's	's	PART
easat-10389	248	13	sensitivity	sensitivity	NOUN
easat-10389	248	14	,	,	PUNCT
easat-10389	248	15	particularly	particularly	ADV
easat-10389	248	16	in	in	ADP
easat-10389	248	17	cases	case	NOUN
easat-10389	248	18	with	with	ADP
easat-10389	248	19	borderline	borderline	NOUN
easat-10389	248	20	risk	risk	NOUN
easat-10389	248	21	levels	level	NOUN
easat-10389	248	22	.	.	PUNCT
easat-10389	249	1	4.4	4.4	NUM
easat-10389	249	2	.	.	PUNCT
easat-10389	250	1	analysis	analysis	NOUN
easat-10389	250	2	of	of	ADP
easat-10389	250	3	stability	stability	NOUN
easat-10389	250	4	and	and	CCONJ
easat-10389	250	5	error	error	NOUN
easat-10389	250	6	this	this	DET
easat-10389	250	7	article	article	NOUN
easat-10389	250	8	employed	employ	VERB
easat-10389	250	9	five	five	NUM
easat-10389	250	10	-	-	ADJ
easat-10389	250	11	fold	fold	ADJ
easat-10389	250	12	cross	cross	NOUN
easat-10389	250	13	-	-	NOUN
easat-10389	250	14	validation	validation	NOUN
easat-10389	250	15	to	to	PART
easat-10389	250	16	assess	assess	VERB
easat-10389	250	17	the	the	DET
easat-10389	250	18	model	model	NOUN
easat-10389	250	19	's	's	PART
easat-10389	250	20	coherence	coherence	NOUN
easat-10389	250	21	and	and	CCONJ
easat-10389	250	22	examined	examine	VERB
easat-10389	250	23	the	the	DET
easat-10389	250	24	distributions	distribution	NOUN
easat-10389	250	25	of	of	ADP
easat-10389	250	26	the	the	DET
easat-10389	250	27	f1	f1	NOUN
easat-10389	250	28	-	-	PUNCT
easat-10389	250	29	score	score	NOUN
easat-10389	250	30	(	(	PUNCT
easat-10389	250	31	figure	figure	NOUN
easat-10389	250	32	7	7	NUM
easat-10389	250	33	)	)	PUNCT
easat-10389	250	34	.	.	PUNCT
easat-10389	251	1	the	the	DET
easat-10389	251	2	cnn+gru+attention	cnn+gru+attention	NOUN
easat-10389	251	3	framework	framework	NOUN
easat-10389	251	4	achieved	achieve	VERB
easat-10389	251	5	the	the	DET
easat-10389	251	6	highest	high	ADJ
easat-10389	251	7	median	median	ADJ
easat-10389	251	8	f1	f1	NOUN
easat-10389	251	9	score	score	NOUN
easat-10389	251	10	(	(	PUNCT
easat-10389	251	11	approximately	approximately	ADV
easat-10389	251	12	0.84	0.84	NUM
easat-10389	251	13	)	)	PUNCT
easat-10389	251	14	,	,	PUNCT
easat-10389	251	15	with	with	ADP
easat-10389	251	16	minimal	minimal	ADJ
easat-10389	251	17	variation	variation	NOUN
easat-10389	251	18	,	,	PUNCT
easat-10389	251	19	indicating	indicate	VERB
easat-10389	251	20	that	that	SCONJ
easat-10389	251	21	it	it	PRON
easat-10389	251	22	performed	perform	VERB
easat-10389	251	23	well	well	ADV
easat-10389	251	24	and	and	CCONJ
easat-10389	251	25	consistently	consistently	ADV
easat-10389	251	26	.	.	PUNCT
easat-10389	252	1	on	on	ADP
easat-10389	252	2	the	the	DET
easat-10389	252	3	other	other	ADJ
easat-10389	252	4	hand	hand	NOUN
easat-10389	252	5	,	,	PUNCT
easat-10389	252	6	classic	classic	ADJ
easat-10389	252	7	models	model	NOUN
easat-10389	252	8	like	like	ADP
easat-10389	252	9	logistic	logistic	ADJ
easat-10389	252	10	regression	regression	NOUN
easat-10389	252	11	and	and	CCONJ
easat-10389	252	12	svm	svm	PROPN
easat-10389	252	13	showed	show	VERB
easat-10389	252	14	lower	low	ADJ
easat-10389	252	15	medians	median	NOUN
easat-10389	252	16	and	and	CCONJ
easat-10389	252	17	increased	increase	VERB
easat-10389	252	18	variability	variability	NOUN
easat-10389	252	19	,	,	PUNCT
easat-10389	252	20	which	which	PRON
easat-10389	252	21	indicates	indicate	VERB
easat-10389	252	22	that	that	SCONJ
easat-10389	252	23	they	they	PRON
easat-10389	252	24	are	be	AUX
easat-10389	252	25	sensitive	sensitive	ADJ
easat-10389	252	26	to	to	ADP
easat-10389	252	27	data	datum	NOUN
easat-10389	252	28	splits	split	NOUN
easat-10389	252	29	.	.	PUNCT
easat-10389	253	1	the	the	DET
easat-10389	253	2	gru	gru	NOUN
easat-10389	253	3	-	-	PUNCT
easat-10389	253	4	only	only	ADV
easat-10389	253	5	and	and	CCONJ
easat-10389	253	6	cnn	cnn	PROPN
easat-10389	253	7	-	-	PUNCT
easat-10389	253	8	only	only	ADJ
easat-10389	253	9	models	model	NOUN
easat-10389	253	10	were	be	AUX
easat-10389	253	11	stable	stable	ADJ
easat-10389	253	12	,	,	PUNCT
easat-10389	253	13	but	but	CCONJ
easat-10389	253	14	they	they	PRON
easat-10389	253	15	varied	vary	VERB
easat-10389	253	16	more	more	ADJ
easat-10389	253	17	than	than	ADP
easat-10389	253	18	the	the	DET
easat-10389	253	19	combined	combined	ADJ
easat-10389	253	20	model	model	NOUN
easat-10389	253	21	.	.	PUNCT
easat-10389	254	1	this	this	PRON
easat-10389	254	2	indicates	indicate	VERB
easat-10389	254	3	that	that	SCONJ
easat-10389	254	4	collaboration	collaboration	NOUN
easat-10389	254	5	helps	help	VERB
easat-10389	254	6	prevent	prevent	VERB
easat-10389	254	7	models	model	NOUN
easat-10389	254	8	from	from	ADP
easat-10389	254	9	overfitting	overfitte	VERB
easat-10389	254	10	and	and	CCONJ
easat-10389	254	11	improves	improve	VERB
easat-10389	254	12	their	their	PRON
easat-10389	254	13	performance	performance	NOUN
easat-10389	254	14	.	.	PUNCT
easat-10389	255	1	the	the	DET
easat-10389	255	2	suggested	suggest	VERB
easat-10389	255	3	model	model	NOUN
easat-10389	255	4	's	's	PART
easat-10389	255	5	limited	limited	ADJ
easat-10389	255	6	interquartile	interquartile	NOUN
easat-10389	255	7	range	range	NOUN
easat-10389	255	8	and	and	CCONJ
easat-10389	255	9	lack	lack	NOUN
easat-10389	255	10	of	of	ADP
easat-10389	255	11	outliers	outlier	NOUN
easat-10389	255	12	show	show	VERB
easat-10389	255	13	that	that	SCONJ
easat-10389	255	14	it	it	PRON
easat-10389	255	15	is	be	AUX
easat-10389	255	16	even	even	ADV
easat-10389	255	17	more	more	ADV
easat-10389	255	18	reliable	reliable	ADJ
easat-10389	255	19	for	for	ADP
easat-10389	255	20	use	use	NOUN
easat-10389	255	21	in	in	ADP
easat-10389	255	22	the	the	DET
easat-10389	255	23	real	real	ADJ
easat-10389	255	24	world	world	NOUN
easat-10389	255	25	.	.	PUNCT
easat-10389	256	1	these	these	DET
easat-10389	256	2	results	result	NOUN
easat-10389	256	3	demonstrate	demonstrate	VERB
easat-10389	256	4	that	that	SCONJ
easat-10389	256	5	the	the	DET
easat-10389	256	6	model	model	NOUN
easat-10389	256	7	is	be	AUX
easat-10389	256	8	consistent	consistent	ADJ
easat-10389	256	9	and	and	CCONJ
easat-10389	256	10	reliable	reliable	ADJ
easat-10389	256	11	,	,	PUNCT
easat-10389	256	12	which	which	PRON
easat-10389	256	13	is	be	AUX
easat-10389	256	14	crucial	crucial	ADJ
easat-10389	256	15	for	for	ADP
easat-10389	256	16	healthcare	healthcare	NOUN
easat-10389	256	17	systems	system	NOUN
easat-10389	256	18	that	that	PRON
easat-10389	256	19	require	require	VERB
easat-10389	256	20	warnings	warning	NOUN
easat-10389	256	21	that	that	PRON
easat-10389	256	22	are	be	AUX
easat-10389	256	23	effective	effective	ADJ
easat-10389	256	24	in	in	ADP
easat-10389	256	25	various	various	ADJ
easat-10389	256	26	situations	situation	NOUN
easat-10389	256	27	.	.	PUNCT
easat-10389	257	1	a	a	DET
easat-10389	257	2	seasonal	seasonal	ADJ
easat-10389	257	3	error	error	NOUN
easat-10389	257	4	study	study	NOUN
easat-10389	257	5	showed	show	VERB
easat-10389	257	6	that	that	SCONJ
easat-10389	257	7	there	there	PRON
easat-10389	257	8	were	be	VERB
easat-10389	257	9	higher	high	ADJ
easat-10389	257	10	false	false	ADJ
easat-10389	257	11	negatives	negative	NOUN
easat-10389	257	12	in	in	ADP
easat-10389	257	13	spring	spring	NOUN
easat-10389	257	14	and	and	CCONJ
easat-10389	257	15	autumn	autumn	NOUN
easat-10389	257	16	.	.	PUNCT
easat-10389	258	1	this	this	PRON
easat-10389	258	2	is	be	AUX
easat-10389	258	3	likely	likely	ADJ
easat-10389	258	4	due	due	ADJ
easat-10389	258	5	to	to	ADP
easat-10389	258	6	allergens	allergen	NOUN
easat-10389	258	7	that	that	PRON
easat-10389	258	8	were	be	AUX
easat-10389	258	9	not	not	PART
easat-10389	258	10	taken	take	VERB
easat-10389	258	11	into	into	ADP
easat-10389	258	12	consideration	consideration	NOUN
easat-10389	258	13	,	,	PUNCT
easat-10389	258	14	such	such	ADJ
easat-10389	258	15	as	as	ADP
easat-10389	258	16	pollen	pollen	NOUN
easat-10389	258	17	.	.	PUNCT
easat-10389	259	1	additionally	additionally	ADV
easat-10389	259	2	,	,	PUNCT
easat-10389	259	3	the	the	DET
easat-10389	259	4	model	model	NOUN
easat-10389	259	5	's	's	PART
easat-10389	259	6	recall	recall	NOUN
easat-10389	259	7	decreased	decrease	VERB
easat-10389	259	8	when	when	SCONJ
easat-10389	259	9	the	the	DET
easat-10389	259	10	air	air	NOUN
easat-10389	259	11	quality	quality	NOUN
easat-10389	259	12	was	be	AUX
easat-10389	259	13	inferior	inferior	ADJ
easat-10389	259	14	(	(	PUNCT
easat-10389	259	15	aqi	aqi	PROPN
easat-10389	259	16	>	>	X
easat-10389	259	17	150	150	NUM
easat-10389	259	18	)	)	PUNCT
easat-10389	259	19	,	,	PUNCT
easat-10389	259	20	indicating	indicate	VERB
easat-10389	259	21	that	that	SCONJ
easat-10389	259	22	the	the	DET
easat-10389	259	23	training	training	NOUN
easat-10389	259	24	set	set	NOUN
easat-10389	259	25	requires	require	VERB
easat-10389	259	26	more	more	ADJ
easat-10389	259	27	high	high	ADJ
easat-10389	259	28	-	-	PUNCT
easat-10389	259	29	risk	risk	NOUN
easat-10389	259	30	data	datum	NOUN
easat-10389	259	31	.	.	PUNCT
easat-10389	260	1	adding	add	VERB
easat-10389	260	2	pollen	pollen	NOUN
easat-10389	260	3	indicators	indicator	NOUN
easat-10389	260	4	or	or	CCONJ
easat-10389	260	5	training	training	NOUN
easat-10389	260	6	in	in	ADP
easat-10389	260	7	different	different	ADJ
easat-10389	260	8	seasons	season	NOUN
easat-10389	260	9	and	and	CCONJ
easat-10389	260	10	aqi	aqi	NOUN
easat-10389	260	11	levels	level	NOUN
easat-10389	260	12	might	might	AUX
easat-10389	260	13	make	make	VERB
easat-10389	260	14	it	it	PRON
easat-10389	260	15	more	more	ADV
easat-10389	260	16	resilient	resilient	ADJ
easat-10389	260	17	.	.	PUNCT
easat-10389	261	1	4.4.1	4.4.1	X
easat-10389	261	2	.	.	PUNCT
easat-10389	261	3	seasonal	seasonal	ADJ
easat-10389	261	4	error	error	NOUN
easat-10389	261	5	analysis	analysis	NOUN
easat-10389	261	6	this	this	DET
easat-10389	261	7	study	study	NOUN
easat-10389	261	8	examined	examine	VERB
easat-10389	261	9	the	the	DET
easat-10389	261	10	model	model	NOUN
easat-10389	261	11	's	's	PART
easat-10389	261	12	errors	error	NOUN
easat-10389	261	13	throughout	throughout	ADP
easat-10389	261	14	spring	spring	NOUN
easat-10389	261	15	,	,	PUNCT
easat-10389	261	16	summer	summer	NOUN
easat-10389	261	17	,	,	PUNCT
easat-10389	261	18	autumn	autumn	NOUN
easat-10389	261	19	,	,	PUNCT
easat-10389	261	20	and	and	CCONJ
easat-10389	261	21	winter	winter	NOUN
easat-10389	261	22	by	by	ADP
easat-10389	261	23	utilizing	utilize	VERB
easat-10389	261	24	time	time	NOUN
easat-10389	261	25	-	-	PUNCT
easat-10389	261	26	stamped	stamp	VERB
easat-10389	261	27	test	test	NOUN
easat-10389	261	28	data	datum	NOUN
easat-10389	261	29	grouped	group	VERB
easat-10389	261	30	by	by	ADP
easat-10389	261	31	local	local	ADJ
easat-10389	261	32	weather	weather	NOUN
easat-10389	261	33	definitions	definition	NOUN
easat-10389	261	34	(	(	PUNCT
easat-10389	261	35	fig	fig	NOUN
easat-10389	261	36	.	.	PUNCT
easat-10389	262	1	8)	8)	NUM
easat-10389	262	2	.	.	PUNCT
easat-10389	263	1	the	the	DET
easat-10389	263	2	aim	aim	NOUN
easat-10389	263	3	was	be	AUX
easat-10389	263	4	to	to	PART
easat-10389	263	5	identify	identify	VERB
easat-10389	263	6	the	the	DET
easat-10389	263	7	model	model	NOUN
easat-10389	263	8	's	's	PART
easat-10389	263	9	limitations	limitation	NOUN
easat-10389	263	10	and	and	CCONJ
easat-10389	263	11	behavior	behavior	NOUN
easat-10389	263	12	in	in	ADP
easat-10389	263	13	different	different	ADJ
easat-10389	263	14	contexts	contexts	NOUN
easat-10389	263	15	.	.	PUNCT
easat-10389	264	1	figure	figure	VERB
easat-10389	264	2	7	7	NUM
easat-10389	264	3	.	.	PUNCT
easat-10389	265	1	seasonal	seasonal	ADJ
easat-10389	265	2	variation	variation	NOUN
easat-10389	265	3	in	in	ADP
easat-10389	265	4	f1	f1	NOUN
easat-10389	265	5	-	-	PUNCT
easat-10389	265	6	score	score	NOUN
easat-10389	265	7	for	for	ADP
easat-10389	265	8	asthma	asthma	NOUN
easat-10389	265	9	risk	risk	NOUN
easat-10389	265	10	prediction	prediction	NOUN
easat-10389	265	11	.	.	PUNCT
easat-10389	266	1	268	268	NUM
easat-10389	266	2	edelweiss	edelweiss	PROPN
easat-10389	266	3	applied	apply	VERB
easat-10389	266	4	science	science	NOUN
easat-10389	266	5	and	and	CCONJ
easat-10389	266	6	technology	technology	NOUN
easat-10389	266	7	issn	issn	PROPN
easat-10389	266	8	:	:	PUNCT
easat-10389	266	9	2576	2576	NUM
easat-10389	266	10	-	-	SYM
easat-10389	266	11	8484	8484	NUM
easat-10389	266	12	vol	vol	NOUN
easat-10389	266	13	.	.	PROPN
easat-10389	267	1	9	9	NUM
easat-10389	267	2	,	,	PUNCT
easat-10389	267	3	no	no	INTJ
easat-10389	267	4	.	.	NOUN
easat-10389	267	5	10	10	NUM
easat-10389	267	6	:	:	PUNCT
easat-10389	267	7	256	256	NUM
easat-10389	267	8	-	-	SYM
easat-10389	267	9	271	271	NUM
easat-10389	267	10	,	,	PUNCT
easat-10389	267	11	2025	2025	NUM
easat-10389	267	12	doi	doi	NOUN
easat-10389	267	13	:	:	PUNCT
easat-10389	267	14	10.55214/2576	10.55214/2576	NUM
easat-10389	267	15	-	-	PUNCT
easat-10389	267	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	267	17	©	©	ADP
easat-10389	267	18	2025	2025	NUM
easat-10389	267	19	by	by	ADP
easat-10389	267	20	the	the	DET
easat-10389	267	21	authors	author	NOUN
easat-10389	267	22	;	;	PUNCT
easat-10389	267	23	licensee	licensee	PROPN
easat-10389	267	24	learning	learning	NOUN
easat-10389	267	25	gate	gate	VERB
easat-10389	267	26	the	the	DET
easat-10389	267	27	study	study	NOUN
easat-10389	267	28	found	find	VERB
easat-10389	267	29	that	that	SCONJ
easat-10389	267	30	forecast	forecast	NOUN
easat-10389	267	31	accuracy	accuracy	NOUN
easat-10389	267	32	varied	varied	ADJ
easat-10389	267	33	significantly	significantly	ADV
easat-10389	267	34	between	between	ADP
easat-10389	267	35	seasons	season	NOUN
easat-10389	267	36	,	,	PUNCT
easat-10389	267	37	especially	especially	ADV
easat-10389	267	38	in	in	ADP
easat-10389	267	39	regions	region	NOUN
easat-10389	267	40	where	where	SCONJ
easat-10389	267	41	climate	climate	NOUN
easat-10389	267	42	change	change	NOUN
easat-10389	267	43	was	be	AUX
easat-10389	267	44	more	more	ADV
easat-10389	267	45	evident	evident	ADJ
easat-10389	267	46	.	.	PUNCT
easat-10389	268	1	summer	summer	NOUN
easat-10389	268	2	had	have	VERB
easat-10389	268	3	the	the	DET
easat-10389	268	4	highest	high	ADJ
easat-10389	268	5	f1	f1	NOUN
easat-10389	268	6	score	score	NOUN
easat-10389	268	7	(	(	PUNCT
easat-10389	268	8	0.93	0.93	NUM
easat-10389	268	9	)	)	PUNCT
easat-10389	268	10	because	because	SCONJ
easat-10389	268	11	air	air	NOUN
easat-10389	268	12	pollution	pollution	NOUN
easat-10389	268	13	levels	level	NOUN
easat-10389	268	14	remained	remain	VERB
easat-10389	268	15	stable	stable	ADJ
easat-10389	268	16	and	and	CCONJ
easat-10389	268	17	outdoor	outdoor	ADJ
easat-10389	268	18	activities	activity	NOUN
easat-10389	268	19	were	be	AUX
easat-10389	268	20	easier	easy	ADJ
easat-10389	268	21	to	to	PART
easat-10389	268	22	plan	plan	VERB
easat-10389	268	23	.	.	PUNCT
easat-10389	269	1	the	the	DET
easat-10389	269	2	algorithm	algorithm	NOUN
easat-10389	269	3	was	be	AUX
easat-10389	269	4	less	less	ADV
easat-10389	269	5	effective	effective	ADJ
easat-10389	269	6	in	in	ADP
easat-10389	269	7	winter	winter	NOUN
easat-10389	269	8	,	,	PUNCT
easat-10389	269	9	as	as	SCONJ
easat-10389	269	10	it	it	PRON
easat-10389	269	11	missed	miss	VERB
easat-10389	269	12	more	more	ADJ
easat-10389	269	13	high	high	ADJ
easat-10389	269	14	-	-	PUNCT
easat-10389	269	15	risk	risk	NOUN
easat-10389	269	16	cases	case	NOUN
easat-10389	269	17	,	,	PUNCT
easat-10389	269	18	probably	probably	ADV
easat-10389	269	19	due	due	ADJ
easat-10389	269	20	to	to	ADP
easat-10389	269	21	rapid	rapid	ADJ
easat-10389	269	22	temperature	temperature	NOUN
easat-10389	269	23	shifts	shift	NOUN
easat-10389	269	24	,	,	PUNCT
easat-10389	269	25	increased	increase	VERB
easat-10389	269	26	indoor	indoor	ADJ
easat-10389	269	27	pollution	pollution	NOUN
easat-10389	269	28	,	,	PUNCT
easat-10389	269	29	and	and	CCONJ
easat-10389	269	30	the	the	DET
easat-10389	269	31	presence	presence	NOUN
easat-10389	269	32	of	of	ADP
easat-10389	269	33	flu	flu	NOUN
easat-10389	269	34	-	-	PUNCT
easat-10389	269	35	like	like	ADJ
easat-10389	269	36	symptoms	symptom	NOUN
easat-10389	269	37	,	,	PUNCT
easat-10389	269	38	which	which	PRON
easat-10389	269	39	made	make	VERB
easat-10389	269	40	accurate	accurate	ADJ
easat-10389	269	41	forecasting	forecasting	NOUN
easat-10389	269	42	more	more	ADV
easat-10389	269	43	challenging	challenging	ADJ
easat-10389	269	44	.	.	PUNCT
easat-10389	270	1	spring	spring	NOUN
easat-10389	270	2	and	and	CCONJ
easat-10389	270	3	autumn	autumn	NOUN
easat-10389	270	4	yielded	yield	VERB
easat-10389	270	5	mixed	mixed	ADJ
easat-10389	270	6	results	result	NOUN
easat-10389	270	7	,	,	PUNCT
easat-10389	270	8	with	with	ADP
easat-10389	270	9	lower	low	ADJ
easat-10389	270	10	recall	recall	NOUN
easat-10389	270	11	due	due	ADP
easat-10389	270	12	to	to	ADP
easat-10389	270	13	unpredictable	unpredictable	ADJ
easat-10389	270	14	weather	weather	NOUN
easat-10389	270	15	and	and	CCONJ
easat-10389	270	16	occasional	occasional	ADJ
easat-10389	270	17	spikes	spike	NOUN
easat-10389	270	18	in	in	ADP
easat-10389	270	19	contaminants	contaminant	NOUN
easat-10389	270	20	such	such	ADJ
easat-10389	270	21	as	as	ADP
easat-10389	270	22	pollen	pollen	NOUN
easat-10389	270	23	and	and	CCONJ
easat-10389	270	24	pm2.5	pm2.5	PROPN
easat-10389	270	25	.	.	PUNCT
easat-10389	271	1	these	these	DET
easat-10389	271	2	findings	finding	NOUN
easat-10389	271	3	suggest	suggest	VERB
easat-10389	271	4	that	that	SCONJ
easat-10389	271	5	future	future	ADJ
easat-10389	271	6	models	model	NOUN
easat-10389	271	7	should	should	AUX
easat-10389	271	8	include	include	VERB
easat-10389	271	9	season	season	NOUN
easat-10389	271	10	-	-	PUNCT
easat-10389	271	11	specific	specific	ADJ
easat-10389	271	12	adjustments	adjustment	NOUN
easat-10389	271	13	or	or	CCONJ
easat-10389	271	14	add	add	VERB
easat-10389	271	15	modules	module	NOUN
easat-10389	271	16	that	that	PRON
easat-10389	271	17	consider	consider	VERB
easat-10389	271	18	seasonal	seasonal	ADJ
easat-10389	271	19	context	context	NOUN
easat-10389	271	20	to	to	PART
easat-10389	271	21	improve	improve	VERB
easat-10389	271	22	performance	performance	NOUN
easat-10389	271	23	in	in	ADP
easat-10389	271	24	real	real	ADJ
easat-10389	271	25	-	-	PUNCT
easat-10389	271	26	world	world	NOUN
easat-10389	271	27	settings	setting	NOUN
easat-10389	271	28	,	,	PUNCT
easat-10389	271	29	especially	especially	ADV
easat-10389	271	30	in	in	ADP
easat-10389	271	31	clinical	clinical	ADJ
easat-10389	271	32	environments	environment	NOUN
easat-10389	271	33	or	or	CCONJ
easat-10389	271	34	asthma	asthma	NOUN
easat-10389	271	35	alert	alert	NOUN
easat-10389	271	36	systems	system	NOUN
easat-10389	271	37	during	during	ADP
easat-10389	271	38	high	high	ADJ
easat-10389	271	39	-	-	PUNCT
easat-10389	271	40	risk	risk	NOUN
easat-10389	271	41	times	time	NOUN
easat-10389	271	42	.	.	PUNCT
easat-10389	272	1	4.4.2	4.4.2	NUM
easat-10389	272	2	.	.	PUNCT
easat-10389	272	3	aqi	aqi	ADJ
easat-10389	272	4	-	-	PUNCT
easat-10389	272	5	level	level	NOUN
easat-10389	272	6	error	error	NOUN
easat-10389	272	7	analysis	analysis	NOUN
easat-10389	272	8	in	in	ADP
easat-10389	272	9	addition	addition	NOUN
easat-10389	272	10	to	to	ADP
easat-10389	272	11	seasonal	seasonal	ADJ
easat-10389	272	12	trends	trend	NOUN
easat-10389	272	13	,	,	PUNCT
easat-10389	272	14	we	we	PRON
easat-10389	272	15	examined	examine	VERB
easat-10389	272	16	model	model	NOUN
easat-10389	272	17	performance	performance	NOUN
easat-10389	272	18	across	across	ADP
easat-10389	272	19	various	various	ADJ
easat-10389	272	20	air	air	NOUN
easat-10389	272	21	quality	quality	NOUN
easat-10389	272	22	index	index	NOUN
easat-10389	272	23	(	(	PUNCT
easat-10389	272	24	aqi	aqi	NOUN
easat-10389	272	25	)	)	PUNCT
easat-10389	272	26	levels	level	NOUN
easat-10389	272	27	to	to	PART
easat-10389	272	28	understand	understand	VERB
easat-10389	272	29	how	how	SCONJ
easat-10389	272	30	the	the	DET
easat-10389	272	31	severity	severity	NOUN
easat-10389	272	32	of	of	ADP
easat-10389	272	33	pollutants	pollutant	NOUN
easat-10389	272	34	influences	influence	VERB
easat-10389	272	35	prediction	prediction	NOUN
easat-10389	272	36	accuracy	accuracy	NOUN
easat-10389	272	37	.	.	PUNCT
easat-10389	273	1	aqi	aqi	PROPN
easat-10389	273	2	values	value	NOUN
easat-10389	273	3	were	be	AUX
easat-10389	273	4	categorized	categorize	VERB
easat-10389	273	5	into	into	ADP
easat-10389	273	6	four	four	NUM
easat-10389	273	7	standard	standard	ADJ
easat-10389	273	8	groups	group	NOUN
easat-10389	273	9	:	:	PUNCT
easat-10389	273	10	good	good	ADJ
easat-10389	273	11	(	(	PUNCT
easat-10389	273	12	0–50	0–50	NOUN
easat-10389	273	13	)	)	PUNCT
easat-10389	273	14	,	,	PUNCT
easat-10389	273	15	moderate	moderate	ADJ
easat-10389	273	16	(	(	PUNCT
easat-10389	273	17	51–100	51–100	NUM
easat-10389	273	18	)	)	PUNCT
easat-10389	273	19	,	,	PUNCT
easat-10389	273	20	unhealthy	unhealthy	ADJ
easat-10389	273	21	for	for	ADP
easat-10389	273	22	sensitive	sensitive	ADJ
easat-10389	273	23	groups	group	NOUN
easat-10389	273	24	(	(	PUNCT
easat-10389	273	25	101–150	101–150	NUM
easat-10389	273	26	)	)	PUNCT
easat-10389	273	27	,	,	PUNCT
easat-10389	273	28	and	and	CCONJ
easat-10389	273	29	unhealthy	unhealthy	ADJ
easat-10389	273	30	or	or	CCONJ
easat-10389	273	31	worse	bad	ADJ
easat-10389	273	32	(	(	PUNCT
easat-10389	273	33	>	>	NOUN
easat-10389	273	34	150	150	NUM
easat-10389	273	35	)	)	PUNCT
easat-10389	273	36	,	,	PUNCT
easat-10389	273	37	following	follow	VERB
easat-10389	273	38	guidelines	guideline	NOUN
easat-10389	273	39	from	from	ADP
easat-10389	273	40	the	the	DET
easat-10389	273	41	who	who	PRON
easat-10389	273	42	and	and	CCONJ
easat-10389	273	43	epa	epa	PROPN
easat-10389	273	44	.	.	PUNCT
easat-10389	274	1	the	the	DET
easat-10389	274	2	model	model	NOUN
easat-10389	274	3	achieved	achieve	VERB
easat-10389	274	4	better	well	ADJ
easat-10389	274	5	results	result	NOUN
easat-10389	274	6	in	in	ADP
easat-10389	274	7	the	the	DET
easat-10389	274	8	moderate	moderate	ADJ
easat-10389	274	9	and	and	CCONJ
easat-10389	274	10	unhealthy	unhealthy	ADJ
easat-10389	274	11	for	for	ADP
easat-10389	274	12	sensitive	sensitive	ADJ
easat-10389	274	13	groups	group	NOUN
easat-10389	274	14	categories	category	NOUN
easat-10389	274	15	,	,	PUNCT
easat-10389	274	16	where	where	SCONJ
easat-10389	274	17	pollution	pollution	NOUN
easat-10389	274	18	levels	level	NOUN
easat-10389	274	19	were	be	AUX
easat-10389	274	20	sufficiently	sufficiently	ADV
easat-10389	274	21	high	high	ADJ
easat-10389	274	22	to	to	PART
easat-10389	274	23	be	be	AUX
easat-10389	274	24	strongly	strongly	ADV
easat-10389	274	25	associated	associate	VERB
easat-10389	274	26	with	with	ADP
easat-10389	274	27	asthma	asthma	NOUN
easat-10389	274	28	cases	case	NOUN
easat-10389	274	29	.	.	PUNCT
easat-10389	275	1	notably	notably	ADV
easat-10389	275	2	,	,	PUNCT
easat-10389	275	3	the	the	DET
easat-10389	275	4	f1score	f1score	NOUN
easat-10389	275	5	reached	reach	VERB
easat-10389	275	6	a	a	DET
easat-10389	275	7	peak	peak	NOUN
easat-10389	275	8	of	of	ADP
easat-10389	275	9	0.91	0.91	NUM
easat-10389	275	10	within	within	ADP
easat-10389	275	11	the	the	DET
easat-10389	275	12	101–150	101–150	NUM
easat-10389	275	13	range	range	NOUN
easat-10389	275	14	,	,	PUNCT
easat-10389	275	15	likely	likely	ADJ
easat-10389	275	16	due	due	ADJ
easat-10389	275	17	to	to	PART
easat-10389	275	18	clearer	clear	ADJ
easat-10389	275	19	signal	signal	VERB
easat-10389	275	20	patterns	pattern	NOUN
easat-10389	275	21	in	in	ADP
easat-10389	275	22	the	the	DET
easat-10389	275	23	data	datum	NOUN
easat-10389	275	24	.	.	PUNCT
easat-10389	276	1	conversely	conversely	ADV
easat-10389	276	2	,	,	PUNCT
easat-10389	276	3	prediction	prediction	NOUN
easat-10389	276	4	accuracy	accuracy	NOUN
easat-10389	276	5	fell	fall	VERB
easat-10389	276	6	significantly	significantly	ADV
easat-10389	276	7	within	within	ADP
easat-10389	276	8	the	the	DET
easat-10389	276	9	“	"	PUNCT
easat-10389	276	10	good	good	ADJ
easat-10389	276	11	”	"	PUNCT
easat-10389	276	12	aqi	aqi	ADJ
easat-10389	276	13	range	range	NOUN
easat-10389	276	14	(	(	PUNCT
easat-10389	276	15	f1	f1	NOUN
easat-10389	276	16	-	-	PUNCT
easat-10389	276	17	score	score	NOUN
easat-10389	276	18	approximately	approximately	ADV
easat-10389	276	19	0.68	0.68	NUM
easat-10389	276	20	)	)	PUNCT
easat-10389	276	21	,	,	PUNCT
easat-10389	276	22	often	often	ADV
easat-10389	276	23	due	due	ADP
easat-10389	276	24	to	to	ADP
easat-10389	276	25	false	false	ADJ
easat-10389	276	26	positive	positive	ADJ
easat-10389	276	27	instances	instance	NOUN
easat-10389	276	28	where	where	SCONJ
easat-10389	276	29	asthma	asthma	NOUN
easat-10389	276	30	attacks	attack	NOUN
easat-10389	276	31	occurred	occur	VERB
easat-10389	276	32	despite	despite	SCONJ
easat-10389	276	33	relatively	relatively	ADV
easat-10389	276	34	clean	clean	ADJ
easat-10389	276	35	air	air	NOUN
easat-10389	276	36	.	.	PUNCT
easat-10389	277	1	these	these	DET
easat-10389	277	2	cases	case	NOUN
easat-10389	277	3	may	may	AUX
easat-10389	277	4	be	be	AUX
easat-10389	277	5	influenced	influence	VERB
easat-10389	277	6	by	by	ADP
easat-10389	277	7	non	non	ADJ
easat-10389	277	8	-	-	ADJ
easat-10389	277	9	environmental	environmental	ADJ
easat-10389	277	10	triggers	trigger	NOUN
easat-10389	277	11	such	such	ADJ
easat-10389	277	12	as	as	ADP
easat-10389	277	13	stress	stress	NOUN
easat-10389	277	14	or	or	CCONJ
easat-10389	277	15	indoor	indoor	ADJ
easat-10389	277	16	allergens	allergen	NOUN
easat-10389	277	17	,	,	PUNCT
easat-10389	277	18	which	which	PRON
easat-10389	277	19	were	be	AUX
easat-10389	277	20	not	not	PART
easat-10389	277	21	captured	capture	VERB
easat-10389	277	22	in	in	ADP
easat-10389	277	23	the	the	DET
easat-10389	277	24	dataset	dataset	NOUN
easat-10389	277	25	.	.	PUNCT
easat-10389	278	1	performance	performance	NOUN
easat-10389	278	2	also	also	ADV
easat-10389	278	3	declined	decline	VERB
easat-10389	278	4	slightly	slightly	ADV
easat-10389	278	5	in	in	ADP
easat-10389	278	6	the	the	DET
easat-10389	278	7	unhealthy	unhealthy	ADJ
easat-10389	278	8	or	or	CCONJ
easat-10389	278	9	worse	bad	ADJ
easat-10389	278	10	group	group	NOUN
easat-10389	278	11	(	(	PUNCT
easat-10389	278	12	>	>	NOUN
easat-10389	278	13	150	150	NUM
easat-10389	278	14	)	)	PUNCT
easat-10389	278	15	,	,	PUNCT
easat-10389	278	16	with	with	ADP
easat-10389	278	17	an	an	DET
easat-10389	278	18	increase	increase	NOUN
easat-10389	278	19	in	in	ADP
easat-10389	278	20	false	false	ADJ
easat-10389	278	21	negatives	negative	NOUN
easat-10389	278	22	,	,	PUNCT
easat-10389	278	23	likely	likely	ADJ
easat-10389	278	24	due	due	ADP
easat-10389	278	25	to	to	ADP
easat-10389	278	26	overlapping	overlap	VERB
easat-10389	278	27	health	health	NOUN
easat-10389	278	28	conditions	condition	NOUN
easat-10389	278	29	or	or	CCONJ
easat-10389	278	30	delays	delay	NOUN
easat-10389	278	31	in	in	ADP
easat-10389	278	32	emergency	emergency	NOUN
easat-10389	278	33	response	response	NOUN
easat-10389	278	34	in	in	ADP
easat-10389	278	35	real	real	ADJ
easat-10389	278	36	-	-	PUNCT
easat-10389	278	37	world	world	NOUN
easat-10389	278	38	reports	report	NOUN
easat-10389	278	39	.	.	PUNCT
easat-10389	279	1	these	these	DET
easat-10389	279	2	findings	finding	NOUN
easat-10389	279	3	suggest	suggest	VERB
easat-10389	279	4	that	that	SCONJ
easat-10389	279	5	the	the	DET
easat-10389	279	6	model	model	NOUN
easat-10389	279	7	performs	perform	VERB
easat-10389	279	8	best	good	ADJ
easat-10389	279	9	when	when	SCONJ
easat-10389	279	10	air	air	NOUN
easat-10389	279	11	quality	quality	NOUN
easat-10389	279	12	is	be	AUX
easat-10389	279	13	moderate	moderate	ADJ
easat-10389	279	14	.	.	PUNCT
easat-10389	280	1	however	however	ADV
easat-10389	280	2	,	,	PUNCT
easat-10389	280	3	it	it	PRON
easat-10389	280	4	may	may	AUX
easat-10389	280	5	require	require	VERB
easat-10389	280	6	additional	additional	ADJ
easat-10389	280	7	specific	specific	ADJ
easat-10389	280	8	information	information	NOUN
easat-10389	280	9	about	about	ADP
easat-10389	280	10	patients	patient	NOUN
easat-10389	280	11	or	or	CCONJ
easat-10389	280	12	indoor	indoor	ADJ
easat-10389	280	13	air	air	NOUN
easat-10389	280	14	conditions	condition	NOUN
easat-10389	280	15	to	to	PART
easat-10389	280	16	provide	provide	VERB
easat-10389	280	17	more	more	ADV
easat-10389	280	18	accurate	accurate	ADJ
easat-10389	280	19	classifications	classification	NOUN
easat-10389	280	20	in	in	ADP
easat-10389	280	21	environments	environment	NOUN
easat-10389	280	22	that	that	PRON
easat-10389	280	23	are	be	AUX
easat-10389	280	24	either	either	CCONJ
easat-10389	280	25	very	very	ADV
easat-10389	280	26	poor	poor	ADJ
easat-10389	280	27	or	or	CCONJ
easat-10389	280	28	optimal	optimal	ADJ
easat-10389	280	29	.	.	PUNCT
easat-10389	281	1	figure	figure	NOUN
easat-10389	281	2	8	8	NUM
easat-10389	281	3	.	.	PUNCT
easat-10389	282	1	f1	f1	NOUN
easat-10389	282	2	-	-	PUNCT
easat-10389	282	3	score	score	NOUN
easat-10389	282	4	across	across	ADP
easat-10389	282	5	different	different	ADJ
easat-10389	282	6	aqi	aqi	ADJ
easat-10389	282	7	levels	level	NOUN
easat-10389	282	8	in	in	ADP
easat-10389	282	9	asthma	asthma	PROPN
easat-10389	282	10	risk	risk	NOUN
easat-10389	282	11	prediction	prediction	NOUN
easat-10389	282	12	.	.	PUNCT
easat-10389	283	1	4.4.3	4.4.3	NUM
easat-10389	283	2	.	.	NOUN
easat-10389	283	3	results	result	NOUN
easat-10389	283	4	and	and	CCONJ
easat-10389	283	5	discussion	discussion	NOUN
easat-10389	283	6	the	the	DET
easat-10389	283	7	experimental	experimental	ADJ
easat-10389	283	8	results	result	NOUN
easat-10389	283	9	demonstrated	demonstrate	VERB
easat-10389	283	10	that	that	SCONJ
easat-10389	283	11	the	the	DET
easat-10389	283	12	proposed	propose	VERB
easat-10389	283	13	cnn	cnn	PROPN
easat-10389	283	14	+	+	CCONJ
easat-10389	283	15	gru	gru	PROPN
easat-10389	283	16	+	+	CCONJ
easat-10389	283	17	attention	attention	NOUN
easat-10389	283	18	model	model	NOUN
easat-10389	283	19	outperformed	outperform	VERB
easat-10389	283	20	traditional	traditional	ADJ
easat-10389	283	21	methods	method	NOUN
easat-10389	283	22	,	,	PUNCT
easat-10389	283	23	including	include	VERB
easat-10389	283	24	logistic	logistic	ADJ
easat-10389	283	25	regression	regression	NOUN
easat-10389	283	26	,	,	PUNCT
easat-10389	283	27	svm	svm	NOUN
easat-10389	283	28	,	,	PUNCT
easat-10389	283	29	and	and	CCONJ
easat-10389	283	30	single	single	ADJ
easat-10389	283	31	-	-	PUNCT
easat-10389	283	32	branch	branch	NOUN
easat-10389	283	33	deep	deep	ADJ
easat-10389	283	34	learning	learning	NOUN
easat-10389	283	35	models	model	NOUN
easat-10389	283	36	.	.	PUNCT
easat-10389	284	1	it	it	PRON
easat-10389	284	2	achieved	achieve	VERB
easat-10389	284	3	an	an	DET
easat-10389	284	4	auc	auc	NOUN
easat-10389	284	5	of	of	ADP
easat-10389	284	6	0.89	0.89	NUM
easat-10389	284	7	and	and	CCONJ
easat-10389	284	8	an	an	DET
easat-10389	284	9	f1	f1	NOUN
easat-10389	284	10	-	-	PUNCT
easat-10389	284	11	score	score	NOUN
easat-10389	284	12	of	of	ADP
easat-10389	284	13	0.84	0.84	NUM
easat-10389	284	14	(	(	PUNCT
easat-10389	284	15	figures	figure	NOUN
easat-10389	284	16	5	5	NUM
easat-10389	284	17	&	&	CCONJ
easat-10389	284	18	6	6	NUM
easat-10389	284	19	)	)	PUNCT
easat-10389	284	20	,	,	PUNCT
easat-10389	284	21	demonstrating	demonstrate	VERB
easat-10389	284	22	robust	robust	ADJ
easat-10389	284	23	269	269	NUM
easat-10389	284	24	edelweiss	edelweiss	PROPN
easat-10389	284	25	applied	apply	VERB
easat-10389	284	26	science	science	NOUN
easat-10389	284	27	and	and	CCONJ
easat-10389	284	28	technology	technology	NOUN
easat-10389	284	29	issn	issn	PROPN
easat-10389	284	30	:	:	PUNCT
easat-10389	284	31	2576	2576	NUM
easat-10389	284	32	-	-	SYM
easat-10389	284	33	8484	8484	NUM
easat-10389	284	34	vol	vol	NOUN
easat-10389	284	35	.	.	PROPN
easat-10389	285	1	9	9	NUM
easat-10389	285	2	,	,	PUNCT
easat-10389	285	3	no	no	INTJ
easat-10389	285	4	.	.	NOUN
easat-10389	285	5	10	10	NUM
easat-10389	285	6	:	:	PUNCT
easat-10389	285	7	256	256	NUM
easat-10389	285	8	-	-	SYM
easat-10389	285	9	271	271	NUM
easat-10389	285	10	,	,	PUNCT
easat-10389	285	11	2025	2025	NUM
easat-10389	285	12	doi	doi	NOUN
easat-10389	285	13	:	:	PUNCT
easat-10389	285	14	10.55214/2576	10.55214/2576	NUM
easat-10389	285	15	-	-	PUNCT
easat-10389	285	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	285	17	©	©	ADP
easat-10389	285	18	2025	2025	NUM
easat-10389	285	19	by	by	ADP
easat-10389	285	20	the	the	DET
easat-10389	285	21	authors	author	NOUN
easat-10389	285	22	;	;	PUNCT
easat-10389	285	23	licensee	licensee	PROPN
easat-10389	285	24	learning	learn	VERB
easat-10389	285	25	gate	gate	NOUN
easat-10389	285	26	discrimination	discrimination	NOUN
easat-10389	285	27	between	between	ADP
easat-10389	285	28	high	high	ADJ
easat-10389	285	29	-	-	PUNCT
easat-10389	285	30	risk	risk	NOUN
easat-10389	285	31	and	and	CCONJ
easat-10389	285	32	low	low	ADJ
easat-10389	285	33	-	-	PUNCT
easat-10389	285	34	risk	risk	NOUN
easat-10389	285	35	asthma	asthma	NOUN
easat-10389	285	36	cases	case	NOUN
easat-10389	285	37	.	.	PUNCT
easat-10389	286	1	this	this	PRON
easat-10389	286	2	underscores	underscore	VERB
easat-10389	286	3	the	the	DET
easat-10389	286	4	effectiveness	effectiveness	NOUN
easat-10389	286	5	of	of	ADP
easat-10389	286	6	integrating	integrate	VERB
easat-10389	286	7	air	air	NOUN
easat-10389	286	8	quality	quality	NOUN
easat-10389	286	9	(	(	PUNCT
easat-10389	286	10	aqi	aqi	PROPN
easat-10389	286	11	)	)	PUNCT
easat-10389	286	12	and	and	CCONJ
easat-10389	286	13	weather	weather	NOUN
easat-10389	286	14	data	datum	NOUN
easat-10389	286	15	using	use	VERB
easat-10389	286	16	parallel	parallel	ADJ
easat-10389	286	17	processing	processing	NOUN
easat-10389	286	18	and	and	CCONJ
easat-10389	286	19	attention	attention	NOUN
easat-10389	286	20	mechanisms	mechanism	NOUN
easat-10389	286	21	.	.	PUNCT
easat-10389	287	1	seasonal	seasonal	ADJ
easat-10389	287	2	analysis	analysis	NOUN
easat-10389	287	3	(	(	PUNCT
easat-10389	287	4	figure	figure	NOUN
easat-10389	287	5	7	7	NUM
easat-10389	287	6	)	)	PUNCT
easat-10389	287	7	revealed	reveal	VERB
easat-10389	287	8	superior	superior	ADJ
easat-10389	287	9	performance	performance	NOUN
easat-10389	287	10	in	in	ADP
easat-10389	287	11	winter	winter	NOUN
easat-10389	287	12	and	and	CCONJ
easat-10389	287	13	spring	spring	NOUN
easat-10389	287	14	,	,	PUNCT
easat-10389	287	15	likely	likely	ADJ
easat-10389	287	16	due	due	ADP
easat-10389	287	17	to	to	ADP
easat-10389	287	18	higher	high	ADJ
easat-10389	287	19	pollution	pollution	NOUN
easat-10389	287	20	and	and	CCONJ
easat-10389	287	21	dynamic	dynamic	ADJ
easat-10389	287	22	weather	weather	NOUN
easat-10389	287	23	providing	provide	VERB
easat-10389	287	24	richer	rich	ADJ
easat-10389	287	25	features	feature	NOUN
easat-10389	287	26	for	for	ADP
easat-10389	287	27	training	training	NOUN
easat-10389	287	28	.	.	PUNCT
easat-10389	288	1	in	in	ADP
easat-10389	288	2	contrast	contrast	NOUN
easat-10389	288	3	,	,	PUNCT
easat-10389	288	4	poorer	poor	ADJ
easat-10389	288	5	results	result	NOUN
easat-10389	288	6	in	in	ADP
easat-10389	288	7	summer	summer	NOUN
easat-10389	288	8	may	may	AUX
easat-10389	288	9	stem	stem	VERB
easat-10389	288	10	from	from	ADP
easat-10389	288	11	less	less	ADV
easat-10389	288	12	distinct	distinct	ADJ
easat-10389	288	13	signals	signal	NOUN
easat-10389	288	14	during	during	ADP
easat-10389	288	15	that	that	DET
easat-10389	288	16	season	season	NOUN
easat-10389	288	17	.	.	PUNCT
easat-10389	289	1	aqi	aqi	ADJ
easat-10389	289	2	-	-	PUNCT
easat-10389	289	3	level	level	NOUN
easat-10389	289	4	analysis	analysis	NOUN
easat-10389	289	5	(	(	PUNCT
easat-10389	289	6	figure	figure	NOUN
easat-10389	289	7	y	y	NOUN
easat-10389	289	8	)	)	PUNCT
easat-10389	289	9	indicated	indicate	VERB
easat-10389	289	10	the	the	DET
easat-10389	289	11	model	model	NOUN
easat-10389	289	12	performs	perform	VERB
easat-10389	289	13	best	good	ADJ
easat-10389	289	14	in	in	ADP
easat-10389	289	15	the	the	DET
easat-10389	289	16	usg	usg	PROPN
easat-10389	289	17	(	(	PUNCT
easat-10389	289	18	101–150	101–150	NUM
easat-10389	289	19	)	)	PUNCT
easat-10389	289	20	and	and	CCONJ
easat-10389	289	21	moderate	moderate	ADJ
easat-10389	289	22	(	(	PUNCT
easat-10389	289	23	51–100	51–100	NUM
easat-10389	289	24	)	)	PUNCT
easat-10389	289	25	categories	category	NOUN
easat-10389	289	26	,	,	PUNCT
easat-10389	289	27	where	where	SCONJ
easat-10389	289	28	patterns	pattern	NOUN
easat-10389	289	29	are	be	AUX
easat-10389	289	30	more	more	ADV
easat-10389	289	31	distinguishable	distinguishable	ADJ
easat-10389	289	32	.	.	PUNCT
easat-10389	290	1	lower	low	ADJ
easat-10389	290	2	performance	performance	NOUN
easat-10389	290	3	in	in	ADP
easat-10389	290	4	the	the	DET
easat-10389	290	5	“	"	PUNCT
easat-10389	290	6	good	good	ADJ
easat-10389	290	7	(	(	PUNCT
easat-10389	290	8	0–50	0–50	NOUN
easat-10389	290	9	)	)	PUNCT
easat-10389	290	10	”	"	PUNCT
easat-10389	290	11	range	range	NOUN
easat-10389	290	12	may	may	AUX
easat-10389	290	13	be	be	AUX
easat-10389	290	14	due	due	ADJ
easat-10389	290	15	to	to	ADP
easat-10389	290	16	weak	weak	ADJ
easat-10389	290	17	signal	signal	NOUN
easat-10389	290	18	contrasts	contrast	VERB
easat-10389	290	19	,	,	PUNCT
easat-10389	290	20	while	while	SCONJ
easat-10389	290	21	reduced	reduced	ADJ
easat-10389	290	22	accuracy	accuracy	NOUN
easat-10389	290	23	in	in	ADP
easat-10389	290	24	the	the	DET
easat-10389	290	25	“	"	PUNCT
easat-10389	290	26	unhealthy+	unhealthy+	PROPN
easat-10389	290	27	(	(	PUNCT
easat-10389	290	28	>	>	NOUN
easat-10389	290	29	150	150	NUM
easat-10389	290	30	)	)	PUNCT
easat-10389	290	31	”	"	PUNCT
easat-10389	290	32	category	category	NOUN
easat-10389	290	33	could	could	AUX
easat-10389	290	34	result	result	VERB
easat-10389	290	35	from	from	ADP
easat-10389	290	36	data	datum	NOUN
easat-10389	290	37	imbalance	imbalance	NOUN
easat-10389	290	38	and	and	CCONJ
easat-10389	290	39	limited	limited	ADJ
easat-10389	290	40	samples	sample	NOUN
easat-10389	290	41	.	.	PUNCT
easat-10389	291	1	two	two	NUM
easat-10389	291	2	key	key	ADJ
easat-10389	291	3	areas	area	NOUN
easat-10389	291	4	require	require	VERB
easat-10389	291	5	improvement	improvement	NOUN
easat-10389	291	6	:	:	PUNCT
easat-10389	291	7	(	(	PUNCT
easat-10389	291	8	i	i	NOUN
easat-10389	291	9	)	)	PUNCT
easat-10389	291	10	enhancing	enhance	VERB
easat-10389	291	11	seasonal	seasonal	ADJ
easat-10389	291	12	stability	stability	NOUN
easat-10389	291	13	through	through	ADP
easat-10389	291	14	strategies	strategy	NOUN
easat-10389	291	15	such	such	ADJ
easat-10389	291	16	as	as	ADP
easat-10389	291	17	seasonal	seasonal	ADJ
easat-10389	291	18	weighting	weighting	NOUN
easat-10389	291	19	or	or	CCONJ
easat-10389	291	20	synthetic	synthetic	ADJ
easat-10389	291	21	data	datum	NOUN
easat-10389	291	22	augmentation	augmentation	NOUN
easat-10389	291	23	,	,	PUNCT
easat-10389	291	24	and	and	CCONJ
easat-10389	291	25	(	(	PUNCT
easat-10389	291	26	ii	ii	NOUN
easat-10389	291	27	)	)	PUNCT
easat-10389	291	28	strengthening	strengthen	VERB
easat-10389	291	29	generalization	generalization	NOUN
easat-10389	291	30	in	in	ADP
easat-10389	291	31	low	low	ADJ
easat-10389	291	32	-	-	PUNCT
easat-10389	291	33	aqi	aqi	NOUN
easat-10389	291	34	settings	setting	NOUN
easat-10389	291	35	by	by	ADP
easat-10389	291	36	better	well	ADV
easat-10389	291	37	detecting	detect	VERB
easat-10389	291	38	subtle	subtle	ADJ
easat-10389	291	39	clinical	clinical	ADJ
easat-10389	291	40	risk	risk	NOUN
easat-10389	291	41	signals	signal	NOUN
easat-10389	291	42	.	.	PUNCT
easat-10389	292	1	lastly	lastly	ADV
easat-10389	292	2	,	,	PUNCT
easat-10389	292	3	the	the	DET
easat-10389	292	4	confusion	confusion	NOUN
easat-10389	292	5	matrix	matrix	NOUN
easat-10389	292	6	(	(	PUNCT
easat-10389	292	7	figure	figure	NOUN
easat-10389	292	8	8)	8)	NUM
easat-10389	292	9	highlights	highlight	NOUN
easat-10389	292	10	some	some	DET
easat-10389	292	11	false	false	ADJ
easat-10389	292	12	positives	positive	NOUN
easat-10389	292	13	,	,	PUNCT
easat-10389	292	14	mainly	mainly	ADV
easat-10389	292	15	for	for	ADP
easat-10389	292	16	low	low	ADJ
easat-10389	292	17	aqi	aqi	ADJ
easat-10389	292	18	cases	case	NOUN
easat-10389	292	19	where	where	SCONJ
easat-10389	292	20	individuals	individual	NOUN
easat-10389	292	21	still	still	ADV
easat-10389	292	22	exhibited	exhibit	VERB
easat-10389	292	23	asthma	asthma	NOUN
easat-10389	292	24	risk	risk	NOUN
easat-10389	292	25	.	.	PUNCT
easat-10389	293	1	future	future	ADJ
easat-10389	293	2	models	model	NOUN
easat-10389	293	3	could	could	AUX
easat-10389	293	4	integrate	integrate	VERB
easat-10389	293	5	personal	personal	ADJ
easat-10389	293	6	health	health	NOUN
easat-10389	293	7	data	datum	NOUN
easat-10389	293	8	,	,	PUNCT
easat-10389	293	9	including	include	VERB
easat-10389	293	10	medical	medical	ADJ
easat-10389	293	11	history	history	NOUN
easat-10389	293	12	and	and	CCONJ
easat-10389	293	13	inflammation	inflammation	NOUN
easat-10389	293	14	biomarkers	biomarker	NOUN
easat-10389	293	15	,	,	PUNCT
easat-10389	293	16	to	to	PART
easat-10389	293	17	enhance	enhance	VERB
easat-10389	293	18	prediction	prediction	NOUN
easat-10389	293	19	accuracy	accuracy	NOUN
easat-10389	293	20	.	.	PUNCT
easat-10389	294	1	5	5	X
easat-10389	294	2	.	.	X
easat-10389	294	3	ablation	ablation	NOUN
easat-10389	294	4	study	study	NOUN
easat-10389	294	5	this	this	DET
easat-10389	294	6	paper	paper	NOUN
easat-10389	294	7	conducted	conduct	VERB
easat-10389	294	8	an	an	DET
easat-10389	294	9	ablation	ablation	NOUN
easat-10389	294	10	study	study	NOUN
easat-10389	294	11	by	by	ADP
easat-10389	294	12	isolating	isolate	VERB
easat-10389	294	13	model	model	NOUN
easat-10389	294	14	branches	branch	NOUN
easat-10389	294	15	and	and	CCONJ
easat-10389	294	16	attention	attention	NOUN
easat-10389	294	17	layers	layer	NOUN
easat-10389	294	18	to	to	PART
easat-10389	294	19	assess	assess	VERB
easat-10389	294	20	the	the	DET
easat-10389	294	21	contribution	contribution	NOUN
easat-10389	294	22	of	of	ADP
easat-10389	294	23	each	each	DET
easat-10389	294	24	architectural	architectural	ADJ
easat-10389	294	25	component	component	NOUN
easat-10389	294	26	.	.	PUNCT
easat-10389	295	1	5.1	5.1	NUM
easat-10389	295	2	.	.	PUNCT
easat-10389	296	1	experimental	experimental	ADJ
easat-10389	296	2	variants	variant	NOUN
easat-10389	296	3	•	•	ADP
easat-10389	296	4	cnn	cnn	PROPN
easat-10389	296	5	-	-	PUNCT
easat-10389	296	6	only	only	ADJ
easat-10389	296	7	model	model	NOUN
easat-10389	296	8	:	:	PUNCT
easat-10389	296	9	captures	capture	VERB
easat-10389	296	10	local	local	ADJ
easat-10389	296	11	spatial	spatial	ADJ
easat-10389	296	12	patterns	pattern	NOUN
easat-10389	296	13	in	in	ADP
easat-10389	296	14	aqi	aqi	PROPN
easat-10389	296	15	data	datum	NOUN
easat-10389	296	16	using	use	VERB
easat-10389	296	17	1d	1d	NUM
easat-10389	296	18	convolution	convolution	NOUN
easat-10389	296	19	,	,	PUNCT
easat-10389	296	20	without	without	ADP
easat-10389	296	21	meteorological	meteorological	ADJ
easat-10389	296	22	time	time	NOUN
easat-10389	296	23	series	series	PROPN
easat-10389	296	24	.	.	PUNCT
easat-10389	297	1	•	•	NUM
easat-10389	297	2	gru	gru	VERB
easat-10389	297	3	-	-	ADJ
easat-10389	297	4	only	only	ADJ
easat-10389	297	5	model	model	NOUN
easat-10389	297	6	:	:	PUNCT
easat-10389	297	7	learns	learn	VERB
easat-10389	297	8	temporal	temporal	ADJ
easat-10389	297	9	sequences	sequence	NOUN
easat-10389	297	10	from	from	ADP
easat-10389	297	11	weather	weather	NOUN
easat-10389	297	12	data	datum	NOUN
easat-10389	297	13	,	,	PUNCT
easat-10389	297	14	excluding	exclude	VERB
easat-10389	297	15	pollutant	pollutant	ADJ
easat-10389	297	16	patterns	pattern	NOUN
easat-10389	297	17	.	.	PUNCT
easat-10389	298	1	•	•	NUM
easat-10389	298	2	cnn+gru	cnn+gru	NOUN
easat-10389	298	3	(	(	PUNCT
easat-10389	298	4	no	no	DET
easat-10389	298	5	attention	attention	NOUN
easat-10389	298	6	):	):	PUNCT
easat-10389	298	7	combines	combine	VERB
easat-10389	298	8	both	both	DET
easat-10389	298	9	branches	branch	NOUN
easat-10389	298	10	without	without	ADP
easat-10389	298	11	attention	attention	NOUN
easat-10389	298	12	fusion	fusion	NOUN
easat-10389	298	13	.	.	PUNCT
easat-10389	299	1	•	•	NUM
easat-10389	299	2	complete	complete	ADJ
easat-10389	299	3	model	model	NOUN
easat-10389	299	4	(	(	PUNCT
easat-10389	299	5	cnn+gru+attention	cnn+gru+attention	NOUN
easat-10389	299	6	):	):	PUNCT
easat-10389	299	7	employs	employ	VERB
easat-10389	299	8	a	a	DET
easat-10389	299	9	dual	dual	ADJ
easat-10389	299	10	-	-	PUNCT
easat-10389	299	11	branch	branch	NOUN
easat-10389	299	12	architecture	architecture	NOUN
easat-10389	299	13	with	with	ADP
easat-10389	299	14	dynamic	dynamic	ADJ
easat-10389	299	15	attention	attention	NOUN
easat-10389	299	16	weights	weight	NOUN
easat-10389	299	17	.	.	PUNCT
easat-10389	300	1	table	table	NOUN
easat-10389	300	2	2	2	NUM
easat-10389	300	3	.	.	PUNCT
easat-10389	300	4	ablation	ablation	NOUN
easat-10389	300	5	study	study	NOUN
easat-10389	300	6	results	result	NOUN
easat-10389	300	7	(	(	PUNCT
easat-10389	300	8	average	average	ADJ
easat-10389	300	9	across	across	ADP
easat-10389	300	10	datasets	dataset	NOUN
easat-10389	300	11	)	)	PUNCT
easat-10389	300	12	model	model	NOUN
easat-10389	300	13	variant	variant	ADJ
easat-10389	300	14	accuracy	accuracy	NOUN
easat-10389	300	15	f1	f1	NOUN
easat-10389	300	16	-	-	PUNCT
easat-10389	300	17	score	score	NOUN
easat-10389	300	18	auc	auc	NOUN
easat-10389	300	19	cnn	cnn	PROPN
easat-10389	300	20	-	-	PUNCT
easat-10389	300	21	only	only	ADV
easat-10389	300	22	0.80	0.80	NUM
easat-10389	300	23	0.76	0.76	NUM
easat-10389	300	24	0.82	0.82	NUM
easat-10389	300	25	gru	gru	NOUN
easat-10389	300	26	-	-	PUNCT
easat-10389	300	27	only	only	ADV
easat-10389	300	28	0.81	0.81	NUM
easat-10389	300	29	0.78	0.78	NUM
easat-10389	300	30	0.83	0.83	NUM
easat-10389	300	31	cnn+gru	cnn+gru	NOUN
easat-10389	300	32	(	(	PUNCT
easat-10389	300	33	no	no	DET
easat-10389	300	34	attention	attention	NOUN
easat-10389	300	35	)	)	PUNCT
easat-10389	300	36	0.84	0.84	NUM
easat-10389	300	37	0.81	0.81	NUM
easat-10389	300	38	0.86	0.86	NUM
easat-10389	300	39	full	full	ADJ
easat-10389	300	40	model	model	NOUN
easat-10389	300	41	0.87	0.87	NUM
easat-10389	300	42	0.84	0.84	NUM
easat-10389	300	43	0.89	0.89	NUM
easat-10389	300	44	5.2	5.2	NUM
easat-10389	300	45	.	.	PUNCT
easat-10389	301	1	key	key	ADJ
easat-10389	301	2	observations	observation	NOUN
easat-10389	301	3	the	the	DET
easat-10389	301	4	ablation	ablation	NOUN
easat-10389	301	5	study	study	NOUN
easat-10389	301	6	highlights	highlight	VERB
easat-10389	301	7	the	the	DET
easat-10389	301	8	contributions	contribution	NOUN
easat-10389	301	9	of	of	ADP
easat-10389	301	10	each	each	DET
easat-10389	301	11	component	component	NOUN
easat-10389	301	12	in	in	ADP
easat-10389	301	13	the	the	DET
easat-10389	301	14	proposed	propose	VERB
easat-10389	301	15	model	model	NOUN
easat-10389	301	16	:	:	PUNCT
easat-10389	301	17	•	•	ADP
easat-10389	301	18	both	both	CCONJ
easat-10389	301	19	the	the	DET
easat-10389	301	20	cnn	cnn	PROPN
easat-10389	301	21	and	and	CCONJ
easat-10389	301	22	gru	gru	PROPN
easat-10389	301	23	branches	branch	NOUN
easat-10389	301	24	are	be	AUX
easat-10389	301	25	essential	essential	ADJ
easat-10389	301	26	.	.	PUNCT
easat-10389	302	1	the	the	DET
easat-10389	302	2	cnn	cnn	PROPN
easat-10389	302	3	-	-	PUNCT
easat-10389	302	4	only	only	ADV
easat-10389	302	5	model	model	NOUN
easat-10389	302	6	(	(	PUNCT
easat-10389	302	7	f1	f1	NOUN
easat-10389	302	8	=	=	SYM
easat-10389	302	9	0.76	0.76	NUM
easat-10389	302	10	)	)	PUNCT
easat-10389	302	11	captures	capture	VERB
easat-10389	302	12	local	local	ADJ
easat-10389	302	13	air	air	NOUN
easat-10389	302	14	quality	quality	NOUN
easat-10389	302	15	fluctuations	fluctuation	NOUN
easat-10389	302	16	but	but	CCONJ
easat-10389	302	17	struggles	struggle	NOUN
easat-10389	302	18	with	with	ADP
easat-10389	302	19	sequential	sequential	ADJ
easat-10389	302	20	dependencies	dependency	NOUN
easat-10389	302	21	.	.	PUNCT
easat-10389	303	1	in	in	ADP
easat-10389	303	2	contrast	contrast	NOUN
easat-10389	303	3	,	,	PUNCT
easat-10389	303	4	the	the	DET
easat-10389	303	5	gru	gru	NOUN
easat-10389	303	6	-	-	ADJ
easat-10389	303	7	only	only	ADJ
easat-10389	303	8	model	model	NOUN
easat-10389	303	9	(	(	PUNCT
easat-10389	303	10	f1	f1	NOUN
easat-10389	303	11	=	=	SYM
easat-10389	303	12	0.78	0.78	NUM
easat-10389	303	13	)	)	PUNCT
easat-10389	303	14	captures	capture	VERB
easat-10389	303	15	temporal	temporal	ADJ
easat-10389	303	16	weather	weather	NOUN
easat-10389	303	17	patterns	pattern	NOUN
easat-10389	303	18	,	,	PUNCT
easat-10389	303	19	demonstrating	demonstrate	VERB
easat-10389	303	20	a	a	DET
easat-10389	303	21	stronger	strong	ADJ
easat-10389	303	22	link	link	NOUN
easat-10389	303	23	between	between	ADP
easat-10389	303	24	meteorological	meteorological	ADJ
easat-10389	303	25	data	datum	NOUN
easat-10389	303	26	and	and	CCONJ
easat-10389	303	27	asthma	asthma	NOUN
easat-10389	303	28	incidence	incidence	NOUN
easat-10389	303	29	.	.	PUNCT
easat-10389	304	1	however	however	ADV
easat-10389	304	2	,	,	PUNCT
easat-10389	304	3	their	their	PRON
easat-10389	304	4	combined	combine	VERB
easat-10389	304	5	architecture	architecture	NOUN
easat-10389	304	6	outperforms	outperform	NOUN
easat-10389	304	7	both	both	PRON
easat-10389	304	8	,	,	PUNCT
easat-10389	304	9	showing	show	VERB
easat-10389	304	10	the	the	DET
easat-10389	304	11	complementary	complementary	ADJ
easat-10389	304	12	value	value	NOUN
easat-10389	304	13	of	of	ADP
easat-10389	304	14	spatial	spatial	ADJ
easat-10389	304	15	and	and	CCONJ
easat-10389	304	16	temporal	temporal	ADJ
easat-10389	304	17	information	information	NOUN
easat-10389	304	18	.	.	PUNCT
easat-10389	305	1	•	•	NUM
easat-10389	305	2	the	the	DET
easat-10389	305	3	attention	attention	NOUN
easat-10389	305	4	mechanism	mechanism	NOUN
easat-10389	305	5	improves	improve	VERB
easat-10389	305	6	accuracy	accuracy	NOUN
easat-10389	305	7	and	and	CCONJ
easat-10389	305	8	interpretability	interpretability	NOUN
easat-10389	305	9	.	.	PUNCT
easat-10389	306	1	removing	remove	VERB
easat-10389	306	2	it	it	PRON
easat-10389	306	3	lowers	lower	VERB
easat-10389	306	4	the	the	DET
easat-10389	306	5	f1	f1	NOUN
easat-10389	306	6	score	score	NOUN
easat-10389	306	7	from	from	ADP
easat-10389	306	8	0.84	0.84	NUM
easat-10389	306	9	to	to	ADP
easat-10389	306	10	0.81	0.81	NUM
easat-10389	306	11	,	,	PUNCT
easat-10389	306	12	underscoring	underscore	VERB
easat-10389	306	13	its	its	PRON
easat-10389	306	14	importance	importance	NOUN
easat-10389	306	15	in	in	ADP
easat-10389	306	16	feature	feature	NOUN
easat-10389	306	17	weighting	weighting	NOUN
easat-10389	306	18	.	.	PUNCT
easat-10389	307	1	attention	attention	NOUN
easat-10389	307	2	fusion	fusion	NOUN
easat-10389	307	3	dynamically	dynamically	ADV
easat-10389	307	4	prioritizes	prioritize	VERB
easat-10389	307	5	key	key	ADJ
easat-10389	307	6	inputs	input	NOUN
easat-10389	307	7	,	,	PUNCT
easat-10389	307	8	especially	especially	ADV
easat-10389	307	9	in	in	ADP
easat-10389	307	10	noisy	noisy	ADJ
easat-10389	307	11	or	or	CCONJ
easat-10389	307	12	complex	complex	ADJ
easat-10389	307	13	conditions	condition	NOUN
easat-10389	307	14	,	,	PUNCT
easat-10389	307	15	while	while	SCONJ
easat-10389	307	16	enhancing	enhance	VERB
easat-10389	307	17	transparency	transparency	NOUN
easat-10389	307	18	by	by	ADP
easat-10389	307	19	highlighting	highlight	VERB
easat-10389	307	20	significant	significant	ADJ
easat-10389	307	21	factors	factor	NOUN
easat-10389	307	22	,	,	PUNCT
easat-10389	307	23	such	such	ADJ
easat-10389	307	24	as	as	ADP
easat-10389	307	25	pm₂.₅	pm₂.₅	NUM
easat-10389	307	26	spikes	spike	NOUN
easat-10389	307	27	or	or	CCONJ
easat-10389	307	28	abrupt	abrupt	ADJ
easat-10389	307	29	humidity	humidity	NOUN
easat-10389	307	30	changes	change	NOUN
easat-10389	307	31	.	.	PUNCT
easat-10389	308	1	270	270	NUM
easat-10389	308	2	edelweiss	edelweiss	PROPN
easat-10389	308	3	applied	apply	VERB
easat-10389	308	4	science	science	NOUN
easat-10389	308	5	and	and	CCONJ
easat-10389	308	6	technology	technology	NOUN
easat-10389	308	7	issn	issn	PROPN
easat-10389	308	8	:	:	PUNCT
easat-10389	308	9	2576	2576	NUM
easat-10389	308	10	-	-	SYM
easat-10389	308	11	8484	8484	NUM
easat-10389	308	12	vol	vol	NOUN
easat-10389	308	13	.	.	PROPN
easat-10389	309	1	9	9	NUM
easat-10389	309	2	,	,	PUNCT
easat-10389	309	3	no	no	INTJ
easat-10389	309	4	.	.	NOUN
easat-10389	309	5	10	10	NUM
easat-10389	309	6	:	:	PUNCT
easat-10389	309	7	256	256	NUM
easat-10389	309	8	-	-	SYM
easat-10389	309	9	271	271	NUM
easat-10389	309	10	,	,	PUNCT
easat-10389	309	11	2025	2025	NUM
easat-10389	309	12	doi	doi	NOUN
easat-10389	309	13	:	:	PUNCT
easat-10389	309	14	10.55214/2576	10.55214/2576	NUM
easat-10389	309	15	-	-	PUNCT
easat-10389	309	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	309	17	©	©	ADP
easat-10389	309	18	2025	2025	NUM
easat-10389	309	19	by	by	ADP
easat-10389	309	20	the	the	DET
easat-10389	309	21	authors	author	NOUN
easat-10389	309	22	;	;	PUNCT
easat-10389	309	23	licensee	licensee	PROPN
easat-10389	309	24	learning	learning	NOUN
easat-10389	309	25	gate	gate	NOUN
easat-10389	309	26	•	•	ADP
easat-10389	309	27	the	the	DET
easat-10389	309	28	complete	complete	ADJ
easat-10389	309	29	dual	dual	ADJ
easat-10389	309	30	-	-	PUNCT
easat-10389	309	31	branch	branch	NOUN
easat-10389	309	32	architecture	architecture	NOUN
easat-10389	309	33	with	with	ADP
easat-10389	309	34	attention	attention	NOUN
easat-10389	309	35	achieves	achieve	VERB
easat-10389	309	36	the	the	DET
easat-10389	309	37	highest	high	ADJ
easat-10389	309	38	auc	auc	NOUN
easat-10389	309	39	(	(	PUNCT
easat-10389	309	40	0.89	0.89	NUM
easat-10389	309	41	)	)	PUNCT
easat-10389	309	42	and	and	CCONJ
easat-10389	309	43	f1	f1	PROPN
easat-10389	309	44	score	score	NOUN
easat-10389	309	45	(	(	PUNCT
easat-10389	309	46	0.84	0.84	NUM
easat-10389	309	47	)	)	PUNCT
easat-10389	309	48	,	,	PUNCT
easat-10389	309	49	excelling	excel	VERB
easat-10389	309	50	particularly	particularly	ADV
easat-10389	309	51	in	in	ADP
easat-10389	309	52	high	high	ADJ
easat-10389	309	53	-	-	PUNCT
easat-10389	309	54	aqi	aqi	NOUN
easat-10389	309	55	scenarios	scenario	NOUN
easat-10389	309	56	and	and	CCONJ
easat-10389	309	57	during	during	ADP
easat-10389	309	58	seasonal	seasonal	ADJ
easat-10389	309	59	transitions	transition	NOUN
easat-10389	309	60	,	,	PUNCT
easat-10389	309	61	where	where	SCONJ
easat-10389	309	62	simpler	simple	ADJ
easat-10389	309	63	models	model	NOUN
easat-10389	309	64	tend	tend	VERB
easat-10389	309	65	to	to	PART
easat-10389	309	66	falter	falter	VERB
easat-10389	309	67	.	.	PUNCT
easat-10389	310	1	•	•	NUM
easat-10389	310	2	·	·	PUNCT
easat-10389	310	3	the	the	DET
easat-10389	310	4	model	model	NOUN
easat-10389	310	5	's	's	PART
easat-10389	310	6	modularity	modularity	NOUN
easat-10389	310	7	supports	support	VERB
easat-10389	310	8	flexible	flexible	ADJ
easat-10389	310	9	deployment	deployment	NOUN
easat-10389	310	10	.	.	PUNCT
easat-10389	311	1	when	when	SCONJ
easat-10389	311	2	one	one	NUM
easat-10389	311	3	data	data	NOUN
easat-10389	311	4	stream	stream	NOUN
easat-10389	311	5	(	(	PUNCT
easat-10389	311	6	e.g.	e.g.	ADV
easat-10389	311	7	,	,	PUNCT
easat-10389	311	8	aqi	aqi	ADJ
easat-10389	311	9	or	or	CCONJ
easat-10389	311	10	meteorological	meteorological	ADJ
easat-10389	311	11	)	)	PUNCT
easat-10389	311	12	is	be	AUX
easat-10389	311	13	unavailable	unavailable	ADJ
easat-10389	311	14	,	,	PUNCT
easat-10389	311	15	either	either	CCONJ
easat-10389	311	16	branch	branch	NOUN
easat-10389	311	17	still	still	ADV
easat-10389	311	18	provides	provide	VERB
easat-10389	311	19	accurate	accurate	ADJ
easat-10389	311	20	predictions	prediction	NOUN
easat-10389	311	21	.	.	PUNCT
easat-10389	312	1	this	this	DET
easat-10389	312	2	flexibility	flexibility	NOUN
easat-10389	312	3	makes	make	VERB
easat-10389	312	4	it	it	PRON
easat-10389	312	5	ideal	ideal	ADJ
easat-10389	312	6	for	for	ADP
easat-10389	312	7	edge	edge	NOUN
easat-10389	312	8	computing	computing	NOUN
easat-10389	312	9	and	and	CCONJ
easat-10389	312	10	resource	resource	NOUN
easat-10389	312	11	-	-	PUNCT
easat-10389	312	12	constrained	constrain	VERB
easat-10389	312	13	settings	setting	NOUN
easat-10389	312	14	,	,	PUNCT
easat-10389	312	15	such	such	ADJ
easat-10389	312	16	as	as	ADP
easat-10389	312	17	mobile	mobile	ADJ
easat-10389	312	18	or	or	CCONJ
easat-10389	312	19	wearable	wearable	ADJ
easat-10389	312	20	health	health	NOUN
easat-10389	312	21	systems	system	NOUN
easat-10389	312	22	.	.	PUNCT
easat-10389	313	1	overall	overall	ADV
easat-10389	313	2	,	,	PUNCT
easat-10389	313	3	the	the	DET
easat-10389	313	4	cnn	cnn	PROPN
easat-10389	313	5	,	,	PUNCT
easat-10389	313	6	gru	gru	PROPN
easat-10389	313	7	,	,	PUNCT
easat-10389	313	8	and	and	CCONJ
easat-10389	313	9	attention	attention	NOUN
easat-10389	313	10	mechanisms	mechanism	NOUN
easat-10389	313	11	each	each	PRON
easat-10389	313	12	contribute	contribute	VERB
easat-10389	313	13	significantly	significantly	ADV
easat-10389	313	14	to	to	ADP
easat-10389	313	15	the	the	DET
easat-10389	313	16	model	model	NOUN
easat-10389	313	17	’s	’s	PART
easat-10389	313	18	performance	performance	NOUN
easat-10389	313	19	,	,	PUNCT
easat-10389	313	20	robustness	robustness	NOUN
easat-10389	313	21	,	,	PUNCT
easat-10389	313	22	and	and	CCONJ
easat-10389	313	23	interpretability	interpretability	NOUN
easat-10389	313	24	,	,	PUNCT
easat-10389	313	25	creating	create	VERB
easat-10389	313	26	an	an	DET
easat-10389	313	27	adaptable	adaptable	ADJ
easat-10389	313	28	architecture	architecture	NOUN
easat-10389	313	29	for	for	ADP
easat-10389	313	30	real	real	ADJ
easat-10389	313	31	-	-	PUNCT
easat-10389	313	32	time	time	NOUN
easat-10389	313	33	environmental	environmental	ADJ
easat-10389	313	34	health	health	NOUN
easat-10389	313	35	monitoring	monitoring	NOUN
easat-10389	313	36	.	.	PUNCT
easat-10389	314	1	6	6	X
easat-10389	314	2	.	.	X
easat-10389	314	3	conclusion	conclusion	NOUN
easat-10389	314	4	and	and	CCONJ
easat-10389	314	5	future	future	ADJ
easat-10389	314	6	work	work	NOUN
easat-10389	314	7	this	this	DET
easat-10389	314	8	paper	paper	NOUN
easat-10389	314	9	introduces	introduce	NOUN
easat-10389	314	10	a	a	DET
easat-10389	314	11	deep	deep	ADJ
easat-10389	314	12	learning	learning	NOUN
easat-10389	314	13	system	system	NOUN
easat-10389	314	14	that	that	PRON
easat-10389	314	15	combines	combine	VERB
easat-10389	314	16	convolutional	convolutional	ADJ
easat-10389	314	17	(	(	PUNCT
easat-10389	314	18	conv1d	conv1d	NOUN
easat-10389	314	19	)	)	PUNCT
easat-10389	314	20	,	,	PUNCT
easat-10389	314	21	recurrent	recurrent	NOUN
easat-10389	314	22	(	(	PUNCT
easat-10389	314	23	gru	gru	NOUN
easat-10389	314	24	)	)	PUNCT
easat-10389	314	25	,	,	PUNCT
easat-10389	314	26	and	and	CCONJ
easat-10389	314	27	attention	attention	NOUN
easat-10389	314	28	mechanisms	mechanism	NOUN
easat-10389	314	29	to	to	PART
easat-10389	314	30	predict	predict	VERB
easat-10389	314	31	asthma	asthma	NOUN
easat-10389	314	32	risk	risk	NOUN
easat-10389	314	33	based	base	VERB
easat-10389	314	34	on	on	ADP
easat-10389	314	35	environmental	environmental	ADJ
easat-10389	314	36	factors	factor	NOUN
easat-10389	314	37	,	,	PUNCT
easat-10389	314	38	such	such	ADJ
easat-10389	314	39	as	as	ADP
easat-10389	314	40	air	air	NOUN
easat-10389	314	41	quality	quality	NOUN
easat-10389	314	42	and	and	CCONJ
easat-10389	314	43	weather	weather	NOUN
easat-10389	314	44	.	.	PUNCT
easat-10389	315	1	the	the	DET
easat-10389	315	2	dual	dual	ADJ
easat-10389	315	3	-	-	PUNCT
easat-10389	315	4	branch	branch	NOUN
easat-10389	315	5	design	design	NOUN
easat-10389	315	6	supports	support	VERB
easat-10389	315	7	different	different	ADJ
easat-10389	315	8	input	input	NOUN
easat-10389	315	9	types	type	NOUN
easat-10389	315	10	,	,	PUNCT
easat-10389	315	11	with	with	ADP
easat-10389	315	12	the	the	DET
easat-10389	315	13	attention	attention	NOUN
easat-10389	315	14	mechanism	mechanism	NOUN
easat-10389	315	15	highlighting	highlight	VERB
easat-10389	315	16	key	key	ADJ
easat-10389	315	17	features	feature	NOUN
easat-10389	315	18	.	.	PUNCT
easat-10389	316	1	the	the	DET
easat-10389	316	2	cnn+gru+attention	cnn+gru+attention	NOUN
easat-10389	316	3	model	model	NOUN
easat-10389	316	4	surpasses	surpass	VERB
easat-10389	316	5	previous	previous	ADJ
easat-10389	316	6	machine	machine	NOUN
easat-10389	316	7	learning	learning	NOUN
easat-10389	316	8	methods	method	NOUN
easat-10389	316	9	(	(	PUNCT
easat-10389	316	10	such	such	ADJ
easat-10389	316	11	as	as	ADP
easat-10389	316	12	logistic	logistic	ADJ
easat-10389	316	13	regression	regression	NOUN
easat-10389	316	14	,	,	PUNCT
easat-10389	316	15	svm	svm	NOUN
easat-10389	316	16	,	,	PUNCT
easat-10389	316	17	and	and	CCONJ
easat-10389	316	18	random	random	ADJ
easat-10389	316	19	forest	forest	NOUN
easat-10389	316	20	)	)	PUNCT
easat-10389	316	21	and	and	CCONJ
easat-10389	316	22	single	single	ADJ
easat-10389	316	23	deep	deep	ADJ
easat-10389	316	24	learning	learning	NOUN
easat-10389	316	25	models	model	NOUN
easat-10389	316	26	(	(	PUNCT
easat-10389	316	27	including	include	VERB
easat-10389	316	28	cnn	cnn	PROPN
easat-10389	316	29	-	-	PUNCT
easat-10389	316	30	only	only	ADV
easat-10389	316	31	,	,	PUNCT
easat-10389	316	32	gru	gru	NOUN
easat-10389	316	33	-	-	PUNCT
easat-10389	316	34	only	only	ADV
easat-10389	316	35	,	,	PUNCT
easat-10389	316	36	and	and	CCONJ
easat-10389	316	37	lstm	lstm	ADJ
easat-10389	316	38	)	)	PUNCT
easat-10389	316	39	in	in	ADP
easat-10389	316	40	terms	term	NOUN
easat-10389	316	41	of	of	ADP
easat-10389	316	42	auc	auc	NOUN
easat-10389	316	43	and	and	CCONJ
easat-10389	316	44	f1	f1	NOUN
easat-10389	316	45	-	-	PUNCT
easat-10389	316	46	score	score	NOUN
easat-10389	316	47	.	.	PUNCT
easat-10389	317	1	seasonal	seasonal	ADJ
easat-10389	317	2	and	and	CCONJ
easat-10389	317	3	aqi	aqi	NOUN
easat-10389	317	4	-	-	PUNCT
easat-10389	317	5	based	base	VERB
easat-10389	317	6	evaluations	evaluation	NOUN
easat-10389	317	7	show	show	VERB
easat-10389	317	8	that	that	SCONJ
easat-10389	317	9	the	the	DET
easat-10389	317	10	model	model	NOUN
easat-10389	317	11	responds	respond	VERB
easat-10389	317	12	to	to	ADP
easat-10389	317	13	environmental	environmental	ADJ
easat-10389	317	14	changes	change	NOUN
easat-10389	317	15	,	,	PUNCT
easat-10389	317	16	emphasizing	emphasize	VERB
easat-10389	317	17	the	the	DET
easat-10389	317	18	importance	importance	NOUN
easat-10389	317	19	of	of	ADP
easat-10389	317	20	dynamic	dynamic	ADJ
easat-10389	317	21	data	datum	NOUN
easat-10389	317	22	for	for	ADP
easat-10389	317	23	health	health	NOUN
easat-10389	317	24	risk	risk	NOUN
easat-10389	317	25	prediction	prediction	NOUN
easat-10389	317	26	.	.	PUNCT
easat-10389	318	1	however	however	ADV
easat-10389	318	2	,	,	PUNCT
easat-10389	318	3	the	the	DET
easat-10389	318	4	model	model	NOUN
easat-10389	318	5	encounters	encounter	VERB
easat-10389	318	6	difficulties	difficulty	NOUN
easat-10389	318	7	with	with	ADP
easat-10389	318	8	low	low	ADJ
easat-10389	318	9	pollution	pollution	NOUN
easat-10389	318	10	levels	level	NOUN
easat-10389	318	11	and	and	CCONJ
easat-10389	318	12	imbalanced	imbalanced	ADJ
easat-10389	318	13	data	datum	NOUN
easat-10389	318	14	over	over	ADP
easat-10389	318	15	time	time	NOUN
easat-10389	318	16	.	.	PUNCT
easat-10389	319	1	future	future	ADJ
easat-10389	319	2	work	work	NOUN
easat-10389	319	3	involves	involve	VERB
easat-10389	319	4	adding	add	VERB
easat-10389	319	5	more	more	ADJ
easat-10389	319	6	personal	personal	ADJ
easat-10389	319	7	health	health	NOUN
easat-10389	319	8	data	datum	NOUN
easat-10389	319	9	and	and	CCONJ
easat-10389	319	10	developing	develop	VERB
easat-10389	319	11	advanced	advanced	ADJ
easat-10389	319	12	models	model	NOUN
easat-10389	319	13	,	,	PUNCT
easat-10389	319	14	such	such	ADJ
easat-10389	319	15	as	as	ADP
easat-10389	319	16	transformers	transformer	NOUN
easat-10389	319	17	.	.	PUNCT
easat-10389	320	1	the	the	DET
easat-10389	320	2	model	model	NOUN
easat-10389	320	3	will	will	AUX
easat-10389	320	4	also	also	ADV
easat-10389	320	5	be	be	AUX
easat-10389	320	6	optimized	optimize	VERB
easat-10389	320	7	for	for	ADP
easat-10389	320	8	realtime	realtime	ADJ
easat-10389	320	9	use	use	NOUN
easat-10389	320	10	on	on	ADP
easat-10389	320	11	mobile	mobile	ADJ
easat-10389	320	12	devices	device	NOUN
easat-10389	320	13	and	and	CCONJ
easat-10389	320	14	calibrated	calibrate	VERB
easat-10389	320	15	for	for	ADP
easat-10389	320	16	improved	improved	ADJ
easat-10389	320	17	performance	performance	NOUN
easat-10389	320	18	across	across	ADP
easat-10389	320	19	various	various	ADJ
easat-10389	320	20	scenarios	scenario	NOUN
easat-10389	320	21	.	.	PUNCT
easat-10389	321	1	this	this	DET
easat-10389	321	2	research	research	NOUN
easat-10389	321	3	establishes	establish	VERB
easat-10389	321	4	a	a	DET
easat-10389	321	5	foundation	foundation	NOUN
easat-10389	321	6	for	for	ADP
easat-10389	321	7	predicting	predict	VERB
easat-10389	321	8	individual	individual	ADJ
easat-10389	321	9	asthma	asthma	NOUN
easat-10389	321	10	risk	risk	NOUN
easat-10389	321	11	,	,	PUNCT
easat-10389	321	12	ultimately	ultimately	ADV
easat-10389	321	13	aiding	aid	VERB
easat-10389	321	14	early	early	ADJ
easat-10389	321	15	intervention	intervention	NOUN
easat-10389	321	16	and	and	CCONJ
easat-10389	321	17	public	public	ADJ
easat-10389	321	18	health	health	NOUN
easat-10389	321	19	monitoring	monitoring	NOUN
easat-10389	321	20	.	.	PUNCT
easat-10389	322	1	6.1	6.1	NUM
easat-10389	322	2	.	.	PUNCT
easat-10389	322	3	disclosure	disclosure	NOUN
easat-10389	322	4	of	of	ADP
easat-10389	322	5	the	the	DET
easat-10389	322	6	ai	ai	NOUN
easat-10389	322	7	tool	tool	NOUN
easat-10389	322	8	sage	sage	VERB
easat-10389	322	9	the	the	DET
easat-10389	322	10	writers	writer	NOUN
easat-10389	322	11	utilized	utilize	VERB
easat-10389	322	12	openai	openai	PROPN
easat-10389	322	13	's	's	PART
easat-10389	322	14	chatgpt	chatgpt	NOUN
easat-10389	322	15	to	to	PART
easat-10389	322	16	clarify	clarify	VERB
easat-10389	322	17	points	point	NOUN
easat-10389	322	18	,	,	PUNCT
easat-10389	322	19	refine	refine	ADJ
easat-10389	322	20	structure	structure	NOUN
easat-10389	322	21	,	,	PUNCT
easat-10389	322	22	and	and	CCONJ
easat-10389	322	23	select	select	VERB
easat-10389	322	24	more	more	ADV
easat-10389	322	25	effective	effective	ADJ
easat-10389	322	26	words	word	NOUN
easat-10389	322	27	.	.	PUNCT
easat-10389	323	1	ai	ai	VERB
easat-10389	323	2	tools	tool	NOUN
easat-10389	323	3	were	be	AUX
easat-10389	323	4	only	only	ADV
easat-10389	323	5	employed	employ	VERB
easat-10389	323	6	for	for	ADP
easat-10389	323	7	language	language	NOUN
easat-10389	323	8	and	and	CCONJ
easat-10389	323	9	formatting	formatting	NOUN
easat-10389	323	10	,	,	PUNCT
easat-10389	323	11	not	not	PART
easat-10389	323	12	for	for	ADP
easat-10389	323	13	designing	design	VERB
easat-10389	323	14	research	research	NOUN
easat-10389	323	15	or	or	CCONJ
easat-10389	323	16	generating	generating	NOUN
easat-10389	323	17	results	result	NOUN
easat-10389	323	18	.	.	PUNCT
easat-10389	324	1	we	we	PRON
easat-10389	324	2	thoroughly	thoroughly	ADV
easat-10389	324	3	reviewed	review	VERB
easat-10389	324	4	all	all	DET
easat-10389	324	5	ai	ai	NOUN
easat-10389	324	6	-	-	PUNCT
easat-10389	324	7	assisted	assist	VERB
easat-10389	324	8	material	material	NOUN
easat-10389	324	9	to	to	PART
easat-10389	324	10	ensure	ensure	VERB
easat-10389	324	11	its	its	PRON
easat-10389	324	12	accuracy	accuracy	NOUN
easat-10389	324	13	and	and	CCONJ
easat-10389	324	14	originality	originality	NOUN
easat-10389	324	15	.	.	PUNCT
easat-10389	325	1	transparency	transparency	NOUN
easat-10389	325	2	:	:	PUNCT
easat-10389	325	3	the	the	DET
easat-10389	325	4	authors	author	NOUN
easat-10389	325	5	confirm	confirm	VERB
easat-10389	325	6	that	that	SCONJ
easat-10389	325	7	the	the	DET
easat-10389	325	8	manuscript	manuscript	NOUN
easat-10389	325	9	is	be	AUX
easat-10389	325	10	an	an	DET
easat-10389	325	11	honest	honest	ADJ
easat-10389	325	12	,	,	PUNCT
easat-10389	325	13	accurate	accurate	ADJ
easat-10389	325	14	,	,	PUNCT
easat-10389	325	15	and	and	CCONJ
easat-10389	325	16	transparent	transparent	ADJ
easat-10389	325	17	account	account	NOUN
easat-10389	325	18	of	of	ADP
easat-10389	325	19	the	the	DET
easat-10389	325	20	study	study	NOUN
easat-10389	325	21	;	;	PUNCT
easat-10389	325	22	that	that	SCONJ
easat-10389	325	23	no	no	DET
easat-10389	325	24	vital	vital	ADJ
easat-10389	325	25	features	feature	NOUN
easat-10389	325	26	of	of	ADP
easat-10389	325	27	the	the	DET
easat-10389	325	28	study	study	NOUN
easat-10389	325	29	have	have	AUX
easat-10389	325	30	been	be	AUX
easat-10389	325	31	omitted	omit	VERB
easat-10389	325	32	;	;	PUNCT
easat-10389	325	33	and	and	CCONJ
easat-10389	325	34	that	that	SCONJ
easat-10389	325	35	any	any	DET
easat-10389	325	36	discrepancies	discrepancy	NOUN
easat-10389	325	37	from	from	ADP
easat-10389	325	38	the	the	DET
easat-10389	325	39	study	study	NOUN
easat-10389	325	40	as	as	SCONJ
easat-10389	325	41	planned	plan	VERB
easat-10389	325	42	have	have	AUX
easat-10389	325	43	been	be	AUX
easat-10389	325	44	explained	explain	VERB
easat-10389	325	45	.	.	PUNCT
easat-10389	326	1	this	this	DET
easat-10389	326	2	study	study	NOUN
easat-10389	326	3	followed	follow	VERB
easat-10389	326	4	all	all	DET
easat-10389	326	5	ethical	ethical	ADJ
easat-10389	326	6	practices	practice	NOUN
easat-10389	326	7	during	during	ADP
easat-10389	326	8	writing	writing	NOUN
easat-10389	326	9	.	.	PUNCT
easat-10389	327	1	copyright	copyright	NOUN
easat-10389	327	2	:	:	PUNCT
easat-10389	327	3	©	©	PROPN
easat-10389	327	4	2025	2025	NUM
easat-10389	327	5	by	by	ADP
easat-10389	327	6	the	the	DET
easat-10389	327	7	authors	author	NOUN
easat-10389	327	8	.	.	PUNCT
easat-10389	328	1	this	this	DET
easat-10389	328	2	open	open	ADJ
easat-10389	328	3	-	-	PUNCT
easat-10389	328	4	access	access	NOUN
easat-10389	328	5	article	article	NOUN
easat-10389	328	6	is	be	AUX
easat-10389	328	7	distributed	distribute	VERB
easat-10389	328	8	under	under	ADP
easat-10389	328	9	the	the	DET
easat-10389	328	10	terms	term	NOUN
easat-10389	328	11	and	and	CCONJ
easat-10389	328	12	conditions	condition	NOUN
easat-10389	328	13	of	of	ADP
easat-10389	328	14	the	the	DET
easat-10389	328	15	creative	creative	ADJ
easat-10389	328	16	commons	common	NOUN
easat-10389	328	17	attribution	attribution	NOUN
easat-10389	328	18	(	(	PUNCT
easat-10389	328	19	cc	cc	NOUN
easat-10389	328	20	by	by	ADP
easat-10389	328	21	)	)	PUNCT
easat-10389	328	22	license	license	NOUN
easat-10389	328	23	(	(	PUNCT
easat-10389	328	24	https://creativecommons.org/licenses/by/4.0/	https://creativecommons.org/licenses/by/4.0/	PROPN
easat-10389	328	25	)	)	PUNCT
easat-10389	328	26	.	.	PUNCT
easat-10389	329	1	references	reference	NOUN
easat-10389	329	2	[	[	X
easat-10389	329	3	1	1	NUM
easat-10389	329	4	]	]	PUNCT
easat-10389	329	5	a.	a.	NOUN
easat-10389	329	6	j.	j.	PROPN
easat-10389	329	7	cohen	cohen	PROPN
easat-10389	329	8	et	et	PROPN
easat-10389	329	9	al	al	PROPN
easat-10389	329	10	.	.	PROPN
easat-10389	329	11	,	,	PUNCT
easat-10389	329	12	"	"	PUNCT
easat-10389	329	13	estimates	estimate	NOUN
easat-10389	329	14	and	and	CCONJ
easat-10389	329	15	25	25	NUM
easat-10389	329	16	-	-	PUNCT
easat-10389	329	17	year	year	NOUN
easat-10389	329	18	trends	trend	NOUN
easat-10389	329	19	of	of	ADP
easat-10389	329	20	the	the	DET
easat-10389	329	21	global	global	ADJ
easat-10389	329	22	burden	burden	NOUN
easat-10389	329	23	of	of	ADP
easat-10389	329	24	disease	disease	NOUN
easat-10389	329	25	attributable	attributable	ADJ
easat-10389	329	26	to	to	ADP
easat-10389	329	27	ambient	ambient	ADJ
easat-10389	329	28	air	air	NOUN
easat-10389	329	29	pollution	pollution	NOUN
easat-10389	329	30	:	:	PUNCT
easat-10389	329	31	an	an	DET
easat-10389	329	32	analysis	analysis	NOUN
easat-10389	329	33	of	of	ADP
easat-10389	329	34	data	datum	NOUN
easat-10389	329	35	from	from	ADP
easat-10389	329	36	the	the	DET
easat-10389	329	37	global	global	ADJ
easat-10389	329	38	burden	burden	NOUN
easat-10389	329	39	of	of	ADP
easat-10389	329	40	diseases	disease	NOUN
easat-10389	329	41	study	study	VERB
easat-10389	329	42	2015	2015	NUM
easat-10389	329	43	,	,	PUNCT
easat-10389	329	44	"	"	PUNCT
easat-10389	329	45	the	the	DET
easat-10389	329	46	lancet	lancet	NOUN
easat-10389	329	47	,	,	PUNCT
easat-10389	329	48	vol	vol	NOUN
easat-10389	329	49	.	.	PROPN
easat-10389	329	50	389	389	NUM
easat-10389	329	51	,	,	PUNCT
easat-10389	329	52	no	no	INTJ
easat-10389	329	53	.	.	NOUN
easat-10389	329	54	10082	10082	NUM
easat-10389	329	55	,	,	PUNCT
easat-10389	329	56	pp	pp	ADJ
easat-10389	329	57	.	.	PUNCT
easat-10389	330	1	1907	1907	NUM
easat-10389	330	2	-	-	SYM
easat-10389	330	3	1918	1918	NUM
easat-10389	330	4	,	,	PUNCT
easat-10389	330	5	2017	2017	NUM
easat-10389	330	6	.	.	PUNCT
easat-10389	331	1	https://doi.org/10.1016/s0140-6736(17)30505-6	https://doi.org/10.1016/s0140-6736(17)30505-6	NOUN
easat-10389	332	1	[	[	X
easat-10389	332	2	2	2	NUM
easat-10389	332	3	]	]	X
easat-10389	332	4	x.-y	x.-y	NOUN
easat-10389	332	5	.	.	PUNCT
easat-10389	333	1	zheng	zheng	PROPN
easat-10389	333	2	et	et	PROPN
easat-10389	333	3	al	al	PROPN
easat-10389	333	4	.	.	PROPN
easat-10389	333	5	,	,	PUNCT
easat-10389	333	6	"	"	PUNCT
easat-10389	333	7	association	association	NOUN
easat-10389	333	8	between	between	ADP
easat-10389	333	9	air	air	NOUN
easat-10389	333	10	pollutants	pollutant	NOUN
easat-10389	333	11	and	and	CCONJ
easat-10389	333	12	asthma	asthma	NOUN
easat-10389	333	13	emergency	emergency	NOUN
easat-10389	333	14	room	room	NOUN
easat-10389	333	15	visits	visit	NOUN
easat-10389	333	16	and	and	CCONJ
easat-10389	333	17	hospital	hospital	NOUN
easat-10389	333	18	admissions	admission	NOUN
easat-10389	333	19	in	in	ADP
easat-10389	333	20	time	time	NOUN
easat-10389	333	21	series	series	PROPN
easat-10389	333	22	studies	study	NOUN
easat-10389	333	23	:	:	PUNCT
easat-10389	333	24	a	a	DET
easat-10389	333	25	systematic	systematic	ADJ
easat-10389	333	26	review	review	NOUN
easat-10389	333	27	and	and	CCONJ
easat-10389	333	28	meta	meta	ADJ
easat-10389	333	29	-	-	PUNCT
easat-10389	333	30	analysis	analysis	NOUN
easat-10389	333	31	,	,	PUNCT
easat-10389	333	32	"	"	PUNCT
easat-10389	333	33	plos	plos	PROPN
easat-10389	333	34	one	one	NUM
easat-10389	333	35	,	,	PUNCT
easat-10389	333	36	vol	vol	NOUN
easat-10389	333	37	.	.	PROPN
easat-10389	333	38	10	10	NUM
easat-10389	333	39	,	,	PUNCT
easat-10389	333	40	no	no	INTJ
easat-10389	333	41	.	.	NOUN
easat-10389	333	42	9	9	NUM
easat-10389	333	43	,	,	PUNCT
easat-10389	333	44	p.	p.	NOUN
easat-10389	333	45	e0138146	e0138146	PROPN
easat-10389	333	46	,	,	PUNCT
easat-10389	333	47	2015	2015	NUM
easat-10389	333	48	.	.	PUNCT
easat-10389	334	1	https://doi.org/10.1371/journal.pone.0138146	https://doi.org/10.1371/journal.pone.0138146	PROPN
easat-10389	335	1	[	[	X
easat-10389	335	2	3	3	X
easat-10389	335	3	]	]	X
easat-10389	335	4	d.	d.	PROPN
easat-10389	335	5	espejo	espejo	PROPN
easat-10389	335	6	,	,	PUNCT
easat-10389	335	7	v.	v.	PROPN
easat-10389	335	8	plaza	plaza	PROPN
easat-10389	335	9	,	,	PUNCT
easat-10389	335	10	s.	s.	PROPN
easat-10389	335	11	quirce	quirce	PROPN
easat-10389	335	12	,	,	PUNCT
easat-10389	335	13	j.	j.	PROPN
easat-10389	335	14	a.	a.	PROPN
easat-10389	335	15	trigueros	trigueros	PROPN
easat-10389	335	16	,	,	PUNCT
easat-10389	335	17	and	and	CCONJ
easat-10389	335	18	x.	x.	PROPN
easat-10389	335	19	muñoz	muñoz	PROPN
easat-10389	335	20	,	,	PUNCT
easat-10389	335	21	"	"	PUNCT
easat-10389	335	22	influence	influence	NOUN
easat-10389	335	23	of	of	ADP
easat-10389	335	24	outdoor	outdoor	ADJ
easat-10389	335	25	air	air	NOUN
easat-10389	335	26	pollutants	pollutant	NOUN
easat-10389	335	27	on	on	ADP
easat-10389	335	28	asthma	asthma	NOUN
easat-10389	335	29	:	:	PUNCT
easat-10389	335	30	a	a	DET
easat-10389	335	31	narrative	narrative	NOUN
easat-10389	335	32	review	review	NOUN
easat-10389	335	33	,	,	PUNCT
easat-10389	335	34	"	"	PUNCT
easat-10389	335	35	open	open	ADJ
easat-10389	335	36	respiratory	respiratory	ADJ
easat-10389	335	37	archives	archive	NOUN
easat-10389	335	38	,	,	PUNCT
easat-10389	335	39	vol	vol	NOUN
easat-10389	335	40	.	.	PROPN
easat-10389	335	41	7	7	NUM
easat-10389	335	42	,	,	PUNCT
easat-10389	335	43	no	no	INTJ
easat-10389	335	44	.	.	NOUN
easat-10389	335	45	3	3	NUM
easat-10389	335	46	,	,	PUNCT
easat-10389	335	47	p.	p.	NOUN
easat-10389	335	48	100448	100448	NUM
easat-10389	335	49	,	,	PUNCT
easat-10389	335	50	2025	2025	NUM
easat-10389	335	51	.	.	PUNCT
easat-10389	336	1	https://doi.org/10.1016/j.opresp.2025.100448	https://doi.org/10.1016/j.opresp.2025.100448	PROPN
easat-10389	336	2	https://creativecommons.org/licenses/by/4.0/	https://creativecommons.org/licenses/by/4.0/	PROPN
easat-10389	336	3	https://doi.org/10.1016/s0140-6736(17)30505-6	https://doi.org/10.1016/s0140-6736(17)30505-6	PROPN
easat-10389	336	4	https://doi.org/10.1371/journal.pone.0138146	https://doi.org/10.1371/journal.pone.0138146	PROPN
easat-10389	336	5	https://doi.org/10.1016/j.opresp.2025.100448	https://doi.org/10.1016/j.opresp.2025.100448	VERB
easat-10389	336	6	271	271	NUM
easat-10389	336	7	edelweiss	edelweiss	PROPN
easat-10389	336	8	applied	apply	VERB
easat-10389	336	9	science	science	NOUN
easat-10389	336	10	and	and	CCONJ
easat-10389	336	11	technology	technology	NOUN
easat-10389	336	12	issn	issn	PROPN
easat-10389	336	13	:	:	PUNCT
easat-10389	336	14	2576	2576	NUM
easat-10389	336	15	-	-	SYM
easat-10389	336	16	8484	8484	NUM
easat-10389	336	17	vol	vol	NOUN
easat-10389	336	18	.	.	PROPN
easat-10389	337	1	9	9	NUM
easat-10389	337	2	,	,	PUNCT
easat-10389	337	3	no	no	INTJ
easat-10389	337	4	.	.	NOUN
easat-10389	337	5	10	10	NUM
easat-10389	337	6	:	:	PUNCT
easat-10389	337	7	256	256	NUM
easat-10389	337	8	-	-	SYM
easat-10389	337	9	271	271	NUM
easat-10389	337	10	,	,	PUNCT
easat-10389	337	11	2025	2025	NUM
easat-10389	337	12	doi	doi	NOUN
easat-10389	337	13	:	:	PUNCT
easat-10389	337	14	10.55214/2576	10.55214/2576	NUM
easat-10389	337	15	-	-	PUNCT
easat-10389	337	16	8484.v9i10.10389	8484.v9i10.10389	NOUN
easat-10389	337	17	©	©	ADP
easat-10389	337	18	2025	2025	NUM
easat-10389	337	19	by	by	ADP
easat-10389	337	20	the	the	DET
easat-10389	337	21	authors	author	NOUN
easat-10389	337	22	;	;	PUNCT
easat-10389	337	23	licensee	licensee	PROPN
easat-10389	337	24	learning	learning	NOUN
easat-10389	337	25	gate	gate	NOUN
easat-10389	338	1	[	[	X
easat-10389	338	2	4	4	NUM
easat-10389	338	3	]	]	PUNCT
easat-10389	338	4	d.	d.	PROPN
easat-10389	338	5	evangelopoulos	evangelopoulos	PROPN
easat-10389	338	6	et	et	PROPN
easat-10389	338	7	al	al	PROPN
easat-10389	338	8	.	.	PROPN
easat-10389	338	9	,	,	PUNCT
easat-10389	338	10	"	"	PUNCT
easat-10389	338	11	the	the	DET
easat-10389	338	12	role	role	NOUN
easat-10389	338	13	of	of	ADP
easat-10389	338	14	burden	burden	NOUN
easat-10389	338	15	of	of	ADP
easat-10389	338	16	disease	disease	NOUN
easat-10389	338	17	assessment	assessment	NOUN
easat-10389	338	18	in	in	ADP
easat-10389	338	19	tracking	track	VERB
easat-10389	338	20	progress	progress	NOUN
easat-10389	338	21	towards	towards	ADP
easat-10389	338	22	achieving	achieve	VERB
easat-10389	338	23	who	who	PRON
easat-10389	338	24	global	global	ADJ
easat-10389	338	25	air	air	NOUN
easat-10389	338	26	quality	quality	NOUN
easat-10389	338	27	guidelines	guideline	NOUN
easat-10389	338	28	,	,	PUNCT
easat-10389	338	29	"	"	PUNCT
easat-10389	338	30	international	international	ADJ
easat-10389	338	31	journal	journal	NOUN
easat-10389	338	32	of	of	ADP
easat-10389	338	33	public	public	ADJ
easat-10389	338	34	health	health	NOUN
easat-10389	338	35	,	,	PUNCT
easat-10389	338	36	vol	vol	NOUN
easat-10389	338	37	.	.	PROPN
easat-10389	338	38	65	65	NUM
easat-10389	338	39	,	,	PUNCT
easat-10389	338	40	pp	pp	ADJ
easat-10389	338	41	.	.	PUNCT
easat-10389	339	1	1455	1455	NUM
easat-10389	339	2	-	-	SYM
easat-10389	339	3	1465	1465	NUM
easat-10389	339	4	,	,	PUNCT
easat-10389	339	5	2020	2020	NUM
easat-10389	339	6	.	.	PUNCT
easat-10389	340	1	https://doi.org/10.1007/s00038-020-01479-z	https://doi.org/10.1007/s00038-020-01479-z	PROPN
easat-10389	341	1	[	[	X
easat-10389	341	2	5	5	NUM
easat-10389	341	3	]	]	X
easat-10389	341	4	y.	y.	PROPN
easat-10389	341	5	guo	guo	PROPN
easat-10389	341	6	,	,	PUNCT
easat-10389	341	7	y.	y.	PROPN
easat-10389	341	8	wu	wu	PROPN
easat-10389	341	9	,	,	PUNCT
easat-10389	341	10	t.	t.	PROPN
easat-10389	341	11	ye	ye	PROPN
easat-10389	341	12	,	,	PUNCT
easat-10389	341	13	l.	l.	PROPN
easat-10389	341	14	zhang	zhang	PROPN
easat-10389	341	15	,	,	PUNCT
easat-10389	341	16	a.	a.	PROPN
easat-10389	341	17	johnson	johnson	PROPN
easat-10389	341	18	,	,	PUNCT
easat-10389	341	19	and	and	CCONJ
easat-10389	341	20	s.	s.	PROPN
easat-10389	341	21	li	li	PROPN
easat-10389	341	22	,	,	PUNCT
easat-10389	341	23	"	"	PUNCT
easat-10389	341	24	a	a	DET
easat-10389	341	25	causal	causal	NOUN
easat-10389	341	26	modelling	model	VERB
easat-10389	341	27	framework	framework	NOUN
easat-10389	341	28	for	for	ADP
easat-10389	341	29	short	short	ADJ
easat-10389	341	30	-	-	PUNCT
easat-10389	341	31	term	term	NOUN
easat-10389	341	32	effects	effect	NOUN
easat-10389	341	33	of	of	ADP
easat-10389	341	34	pm2	pm2	NOUN
easat-10389	341	35	.	.	PUNCT
easat-10389	341	36	5	5	NUM
easat-10389	341	37	on	on	ADP
easat-10389	341	38	hospitalisations	hospitalisation	NOUN
easat-10389	341	39	:	:	PUNCT
easat-10389	341	40	a	a	DET
easat-10389	341	41	nationwide	nationwide	ADJ
easat-10389	341	42	time	time	NOUN
easat-10389	341	43	series	series	PROPN
easat-10389	341	44	study	study	PROPN
easat-10389	341	45	in	in	ADP
easat-10389	341	46	brazil	brazil	PROPN
easat-10389	341	47	,	,	PUNCT
easat-10389	341	48	"	"	PUNCT
easat-10389	341	49	environment	environment	NOUN
easat-10389	341	50	international	international	ADJ
easat-10389	341	51	,	,	PUNCT
easat-10389	341	52	vol	vol	NOUN
easat-10389	341	53	.	.	PROPN
easat-10389	341	54	171	171	NUM
easat-10389	341	55	,	,	PUNCT
easat-10389	341	56	p.	p.	NOUN
easat-10389	341	57	107688	107688	NUM
easat-10389	341	58	,	,	PUNCT
easat-10389	341	59	2023	2023	NUM
easat-10389	341	60	.	.	PUNCT
easat-10389	342	1	https://doi.org/10.1016/j.envint.2022.107688	https://doi.org/10.1016/j.envint.2022.107688	PROPN
easat-10389	342	2	[	[	X
easat-10389	342	3	6	6	X
easat-10389	342	4	]	]	PUNCT
easat-10389	342	5	j.	j.	PROPN
easat-10389	342	6	s.	s.	PROPN
easat-10389	342	7	d.	d.	PROPN
easat-10389	342	8	reis	reis	PROPN
easat-10389	342	9	et	et	PROPN
easat-10389	342	10	al	al	PROPN
easat-10389	342	11	.	.	PROPN
easat-10389	342	12	,	,	PUNCT
easat-10389	342	13	"	"	PUNCT
easat-10389	342	14	predicting	predict	VERB
easat-10389	342	15	asthma	asthma	NOUN
easat-10389	342	16	hospitalizations	hospitalization	NOUN
easat-10389	342	17	from	from	ADP
easat-10389	342	18	climate	climate	NOUN
easat-10389	342	19	and	and	CCONJ
easat-10389	342	20	air	air	NOUN
easat-10389	342	21	pollution	pollution	NOUN
easat-10389	342	22	data	datum	NOUN
easat-10389	342	23	:	:	PUNCT
easat-10389	342	24	a	a	DET
easat-10389	342	25	machine	machine	NOUN
easat-10389	342	26	learning	learning	NOUN
easat-10389	342	27	-	-	PUNCT
easat-10389	342	28	based	base	VERB
easat-10389	342	29	approach	approach	NOUN
easat-10389	342	30	,	,	PUNCT
easat-10389	342	31	"	"	PUNCT
easat-10389	342	32	climate	climate	NOUN
easat-10389	342	33	,	,	PUNCT
easat-10389	342	34	vol	vol	NOUN
easat-10389	342	35	.	.	PROPN
easat-10389	342	36	13	13	NUM
easat-10389	342	37	,	,	PUNCT
easat-10389	342	38	no	no	INTJ
easat-10389	342	39	.	.	NOUN
easat-10389	342	40	2	2	NUM
easat-10389	342	41	,	,	PUNCT
easat-10389	342	42	p.	p.	NOUN
easat-10389	342	43	23	23	NUM
easat-10389	342	44	,	,	PUNCT
easat-10389	342	45	2025	2025	NUM
easat-10389	342	46	.	.	PUNCT
easat-10389	343	1	https://doi.org/10.3390/cli13020023	https://doi.org/10.3390/cli13020023	NOUN
easat-10389	344	1	[	[	X
easat-10389	344	2	7	7	NUM
easat-10389	344	3	]	]	X
easat-10389	344	4	m.	m.	NOUN
easat-10389	344	5	i.	i.	PROPN
easat-10389	344	6	jordan	jordan	PROPN
easat-10389	344	7	and	and	CCONJ
easat-10389	344	8	t.	t.	PROPN
easat-10389	344	9	m.	m.	PROPN
easat-10389	344	10	mitchell	mitchell	PROPN
easat-10389	344	11	,	,	PUNCT
easat-10389	344	12	"	"	PUNCT
easat-10389	344	13	machine	machine	NOUN
easat-10389	344	14	learning	learning	NOUN
easat-10389	344	15	:	:	PUNCT
easat-10389	344	16	trends	trend	NOUN
easat-10389	344	17	,	,	PUNCT
easat-10389	344	18	perspectives	perspective	NOUN
easat-10389	344	19	,	,	PUNCT
easat-10389	344	20	and	and	CCONJ
easat-10389	344	21	prospects	prospect	NOUN
easat-10389	344	22	,	,	PUNCT
easat-10389	344	23	"	"	PUNCT
easat-10389	344	24	science	science	NOUN
easat-10389	344	25	,	,	PUNCT
easat-10389	344	26	vol	vol	NOUN
easat-10389	344	27	.	.	PROPN
easat-10389	345	1	349	349	NUM
easat-10389	345	2	,	,	PUNCT
easat-10389	345	3	no	no	INTJ
easat-10389	345	4	.	.	NOUN
easat-10389	346	1	6245	6245	NUM
easat-10389	346	2	,	,	PUNCT
easat-10389	346	3	pp	pp	ADV
easat-10389	346	4	.	.	PUNCT
easat-10389	347	1	255	255	NUM
easat-10389	347	2	-	-	SYM
easat-10389	347	3	260	260	NUM
easat-10389	347	4	,	,	PUNCT
easat-10389	347	5	2015	2015	NUM
easat-10389	347	6	.	.	PUNCT
easat-10389	348	1	https://doi.org/10.1126/science.aaa8415	https://doi.org/10.1126/science.aaa8415	NUM
easat-10389	349	1	[	[	X
easat-10389	349	2	8	8	NUM
easat-10389	349	3	]	]	PUNCT
easat-10389	349	4	a.	a.	NOUN
easat-10389	349	5	botchkarev	botchkarev	PROPN
easat-10389	349	6	,	,	PUNCT
easat-10389	349	7	"	"	PUNCT
easat-10389	349	8	a	a	DET
easat-10389	349	9	new	new	ADJ
easat-10389	349	10	typology	typology	NOUN
easat-10389	349	11	design	design	NOUN
easat-10389	349	12	of	of	ADP
easat-10389	349	13	performance	performance	NOUN
easat-10389	349	14	metrics	metric	NOUN
easat-10389	349	15	to	to	PART
easat-10389	349	16	measure	measure	VERB
easat-10389	349	17	errors	error	NOUN
easat-10389	349	18	in	in	ADP
easat-10389	349	19	machine	machine	NOUN
easat-10389	349	20	learning	learn	VERB
easat-10389	349	21	regression	regression	NOUN
easat-10389	349	22	algorithms	algorithm	NOUN
easat-10389	349	23	,	,	PUNCT
easat-10389	349	24	"	"	PUNCT
easat-10389	349	25	interdisciplinary	interdisciplinary	ADJ
easat-10389	349	26	journal	journal	NOUN
easat-10389	349	27	of	of	ADP
easat-10389	349	28	information	information	NOUN
easat-10389	349	29	,	,	PUNCT
easat-10389	349	30	knowledge	knowledge	NOUN
easat-10389	349	31	,	,	PUNCT
easat-10389	349	32	and	and	CCONJ
easat-10389	349	33	management	management	NOUN
easat-10389	349	34	,	,	PUNCT
easat-10389	349	35	vol	vol	NOUN
easat-10389	349	36	.	.	PROPN
easat-10389	349	37	14	14	NUM
easat-10389	349	38	,	,	PUNCT
easat-10389	349	39	pp	pp	ADJ
easat-10389	349	40	.	.	PUNCT
easat-10389	350	1	45–76	45–76	NUM
easat-10389	350	2	,	,	PUNCT
easat-10389	350	3	2019	2019	NUM
easat-10389	350	4	.	.	PUNCT
easat-10389	351	1	https://doi.org/10.28945/4184	https://doi.org/10.28945/4184	NOUN
easat-10389	351	2	[	[	X
easat-10389	351	3	9	9	NUM
easat-10389	351	4	]	]	PUNCT
easat-10389	351	5	m.	m.	NOUN
easat-10389	351	6	chang	chang	PROPN
easat-10389	351	7	and	and	CCONJ
easat-10389	351	8	y.	y.	PROPN
easat-10389	351	9	ku	ku	PROPN
easat-10389	351	10	,	,	PUNCT
easat-10389	351	11	"	"	PUNCT
easat-10389	351	12	lstm	lstm	ADJ
easat-10389	351	13	model	model	NOUN
easat-10389	351	14	for	for	ADP
easat-10389	351	15	predicting	predict	VERB
easat-10389	351	16	the	the	DET
easat-10389	351	17	daily	daily	ADJ
easat-10389	351	18	number	number	NOUN
easat-10389	351	19	of	of	ADP
easat-10389	351	20	asthma	asthma	NOUN
easat-10389	351	21	patients	patient	NOUN
easat-10389	351	22	in	in	ADP
easat-10389	351	23	seoul	seoul	PROPN
easat-10389	351	24	,	,	PUNCT
easat-10389	351	25	south	south	PROPN
easat-10389	351	26	korea	korea	PROPN
easat-10389	351	27	,	,	PUNCT
easat-10389	351	28	using	use	VERB
easat-10389	351	29	meteorological	meteorological	ADJ
easat-10389	351	30	and	and	CCONJ
easat-10389	351	31	air	air	NOUN
easat-10389	351	32	pollution	pollution	NOUN
easat-10389	351	33	data	datum	NOUN
easat-10389	351	34	,	,	PUNCT
easat-10389	351	35	"	"	PUNCT
easat-10389	351	36	environmental	environmental	ADJ
easat-10389	351	37	science	science	NOUN
easat-10389	351	38	and	and	CCONJ
easat-10389	351	39	pollution	pollution	NOUN
easat-10389	351	40	research	research	NOUN
easat-10389	351	41	,	,	PUNCT
easat-10389	351	42	vol	vol	NOUN
easat-10389	351	43	.	.	PROPN
easat-10389	351	44	30	30	NUM
easat-10389	351	45	,	,	PUNCT
easat-10389	351	46	pp	pp	ADJ
easat-10389	351	47	.	.	PUNCT
easat-10389	351	48	37440	37440	NUM
easat-10389	351	49	-	-	SYM
easat-10389	351	50	37448	37448	NUM
easat-10389	351	51	,	,	PUNCT
easat-10389	351	52	2023	2023	NUM
easat-10389	351	53	.	.	PUNCT
easat-10389	352	1	https://doi.org/10.1007/s11356-022-24956-9	https://doi.org/10.1007/s11356-022-24956-9	NUM
easat-10389	352	2	[	[	SYM
easat-10389	352	3	10	10	NUM
easat-10389	352	4	]	]	PUNCT
easat-10389	352	5	p.	p.	PROPN
easat-10389	352	6	c.	c.	PROPN
easat-10389	352	7	liu	liu	PROPN
easat-10389	352	8	,	,	PUNCT
easat-10389	352	9	t.	t.	PROPN
easat-10389	352	10	i.	i.	PROPN
easat-10389	352	11	chou	chou	PROPN
easat-10389	352	12	,	,	PUNCT
easat-10389	352	13	s.	s.	PROPN
easat-10389	352	14	w.	w.	PROPN
easat-10389	352	15	chiu	chiu	PROPN
easat-10389	352	16	,	,	PUNCT
easat-10389	352	17	and	and	CCONJ
easat-10389	352	18	k.	k.	PROPN
easat-10389	352	19	t.	t.	PROPN
easat-10389	352	20	tang	tang	PROPN
easat-10389	352	21	,	,	PUNCT
easat-10389	352	22	"	"	PUNCT
easat-10389	352	23	a	a	DET
easat-10389	352	24	dual	dual	ADJ
easat-10389	352	25	-	-	PUNCT
easat-10389	352	26	path	path	NOUN
easat-10389	352	27	deep	deep	ADJ
easat-10389	352	28	learning	learning	NOUN
easat-10389	352	29	model	model	NOUN
easat-10389	352	30	for	for	ADP
easat-10389	352	31	low	low	ADJ
easat-10389	352	32	-	-	PUNCT
easat-10389	352	33	cost	cost	NOUN
easat-10389	352	34	air	air	NOUN
easat-10389	352	35	quality	quality	NOUN
easat-10389	352	36	sensor	sensor	NOUN
easat-10389	352	37	calibration	calibration	NOUN
easat-10389	352	38	,	,	PUNCT
easat-10389	352	39	"	"	PUNCT
easat-10389	352	40	ieee	ieee	NOUN
easat-10389	352	41	sensors	sensor	NOUN
easat-10389	352	42	journal	journal	PROPN
easat-10389	352	43	,	,	PUNCT
easat-10389	352	44	vol	vol	NOUN
easat-10389	352	45	.	.	PROPN
easat-10389	352	46	24	24	NUM
easat-10389	352	47	,	,	PUNCT
easat-10389	352	48	no	no	INTJ
easat-10389	352	49	.	.	NOUN
easat-10389	352	50	23	23	NUM
easat-10389	352	51	,	,	PUNCT
easat-10389	352	52	pp	pp	ADJ
easat-10389	352	53	.	.	PUNCT
easat-10389	352	54	39914	39914	NUM
easat-10389	352	55	-	-	SYM
easat-10389	352	56	39922	39922	NUM
easat-10389	352	57	,	,	PUNCT
easat-10389	352	58	2024	2024	NUM
easat-10389	352	59	.	.	PUNCT
easat-10389	353	1	https://doi.org/10.1109/jsen.2024.3472291	https://doi.org/10.1109/jsen.2024.3472291	PROPN
easat-10389	354	1	[	[	X
easat-10389	354	2	11	11	NUM
easat-10389	354	3	]	]	PUNCT
easat-10389	354	4	s.	s.	PROPN
easat-10389	354	5	sui	sui	PROPN
easat-10389	354	6	and	and	CCONJ
easat-10389	354	7	q.	q.	PROPN
easat-10389	354	8	han	han	PROPN
easat-10389	354	9	,	,	PUNCT
easat-10389	354	10	"	"	PUNCT
easat-10389	354	11	multi	multi	ADJ
easat-10389	354	12	-	-	ADJ
easat-10389	354	13	view	view	ADJ
easat-10389	354	14	multi	multi	ADJ
easat-10389	354	15	-	-	ADJ
easat-10389	354	16	task	task	ADJ
easat-10389	354	17	spatiotemporal	spatiotemporal	ADJ
easat-10389	354	18	graph	graph	NOUN
easat-10389	354	19	convolutional	convolutional	ADJ
easat-10389	354	20	network	network	NOUN
easat-10389	354	21	for	for	ADP
easat-10389	354	22	air	air	NOUN
easat-10389	354	23	quality	quality	NOUN
easat-10389	354	24	prediction	prediction	NOUN
easat-10389	354	25	,	,	PUNCT
easat-10389	354	26	"	"	PUNCT
easat-10389	354	27	science	science	NOUN
easat-10389	354	28	of	of	ADP
easat-10389	354	29	the	the	DET
easat-10389	354	30	total	total	ADJ
easat-10389	354	31	environment	environment	NOUN
easat-10389	354	32	,	,	PUNCT
easat-10389	354	33	vol	vol	NOUN
easat-10389	354	34	.	.	PROPN
easat-10389	354	35	893	893	NUM
easat-10389	354	36	,	,	PUNCT
easat-10389	354	37	p.	p.	NOUN
easat-10389	354	38	164699	164699	NUM
easat-10389	354	39	,	,	PUNCT
easat-10389	354	40	2023	2023	NUM
easat-10389	354	41	.	.	PUNCT
easat-10389	355	1	https://doi.org/10.1016/j.scitotenv.2023.164699	https://doi.org/10.1016/j.scitotenv.2023.164699	ADJ
easat-10389	355	2	[	[	X
easat-10389	355	3	12	12	NUM
easat-10389	355	4	]	]	X
easat-10389	355	5	h.	h.	PROPN
easat-10389	355	6	ding	ding	PROPN
easat-10389	355	7	and	and	CCONJ
easat-10389	355	8	g.	g.	PROPN
easat-10389	355	9	noh	noh	PROPN
easat-10389	355	10	,	,	PUNCT
easat-10389	355	11	"	"	PUNCT
easat-10389	355	12	a	a	DET
easat-10389	355	13	hybrid	hybrid	ADJ
easat-10389	355	14	model	model	NOUN
easat-10389	355	15	for	for	ADP
easat-10389	355	16	spatiotemporal	spatiotemporal	ADJ
easat-10389	355	17	air	air	NOUN
easat-10389	355	18	quality	quality	NOUN
easat-10389	355	19	prediction	prediction	NOUN
easat-10389	355	20	based	base	VERB
easat-10389	355	21	on	on	ADP
easat-10389	355	22	interpretable	interpretable	ADJ
easat-10389	355	23	neural	neural	ADJ
easat-10389	355	24	networks	network	NOUN
easat-10389	355	25	and	and	CCONJ
easat-10389	355	26	a	a	DET
easat-10389	355	27	graph	graph	NOUN
easat-10389	355	28	neural	neural	ADJ
easat-10389	355	29	network	network	NOUN
easat-10389	355	30	,	,	PUNCT
easat-10389	355	31	"	"	PUNCT
easat-10389	355	32	atmosphere	atmosphere	NOUN
easat-10389	355	33	,	,	PUNCT
easat-10389	355	34	vol	vol	NOUN
easat-10389	355	35	.	.	PROPN
easat-10389	356	1	14	14	NUM
easat-10389	356	2	,	,	PUNCT
easat-10389	356	3	no	no	INTJ
easat-10389	356	4	.	.	NOUN
easat-10389	356	5	12	12	NUM
easat-10389	356	6	,	,	PUNCT
easat-10389	356	7	p.	p.	NOUN
easat-10389	356	8	1807	1807	NUM
easat-10389	356	9	,	,	PUNCT
easat-10389	356	10	2023	2023	NUM
easat-10389	356	11	.	.	PUNCT
easat-10389	357	1	https://doi.org/10.3390/atmos14121807	https://doi.org/10.3390/atmos14121807	PROPN
easat-10389	358	1	[	[	X
easat-10389	358	2	13	13	NUM
easat-10389	358	3	]	]	PUNCT
easat-10389	358	4	j.	j.	PROPN
easat-10389	358	5	luo	luo	PROPN
easat-10389	358	6	et	et	PROPN
easat-10389	358	7	al	al	PROPN
easat-10389	358	8	.	.	PROPN
easat-10389	358	9	,	,	PUNCT
easat-10389	358	10	"	"	PUNCT
easat-10389	358	11	a	a	DET
easat-10389	358	12	novel	novel	ADJ
easat-10389	358	13	graph	graph	NOUN
easat-10389	358	14	neural	neural	ADJ
easat-10389	358	15	network	network	NOUN
easat-10389	358	16	based	base	VERB
easat-10389	358	17	approach	approach	NOUN
easat-10389	358	18	for	for	ADP
easat-10389	358	19	influenza	influenza	NOUN
easat-10389	358	20	-	-	PUNCT
easat-10389	358	21	like	like	ADJ
easat-10389	358	22	illness	illness	NOUN
easat-10389	358	23	nowcasting	nowcasting	NOUN
easat-10389	358	24	:	:	PUNCT
easat-10389	358	25	exploring	explore	VERB
easat-10389	358	26	the	the	DET
easat-10389	358	27	interplay	interplay	NOUN
easat-10389	358	28	of	of	ADP
easat-10389	358	29	temporal	temporal	ADJ
easat-10389	358	30	,	,	PUNCT
easat-10389	358	31	geographical	geographical	ADJ
easat-10389	358	32	,	,	PUNCT
easat-10389	358	33	and	and	CCONJ
easat-10389	358	34	functional	functional	ADJ
easat-10389	358	35	spatial	spatial	ADJ
easat-10389	358	36	features	feature	NOUN
easat-10389	358	37	,	,	PUNCT
easat-10389	358	38	"	"	PUNCT
easat-10389	358	39	bmc	bmc	ADJ
easat-10389	358	40	public	public	ADJ
easat-10389	358	41	health	health	NOUN
easat-10389	358	42	,	,	PUNCT
easat-10389	358	43	vol	vol	NOUN
easat-10389	358	44	.	.	PROPN
easat-10389	358	45	25	25	NUM
easat-10389	358	46	,	,	PUNCT
easat-10389	358	47	p.	p.	NOUN
easat-10389	358	48	408	408	NUM
easat-10389	358	49	,	,	PUNCT
easat-10389	358	50	2025	2025	NUM
easat-10389	358	51	.	.	PUNCT
easat-10389	359	1	https://doi.org/10.1186/s12889-025-21618-6	https://doi.org/10.1186/s12889-025-21618-6	NUM
easat-10389	360	1	[	[	X
easat-10389	360	2	14	14	NUM
easat-10389	360	3	]	]	PUNCT
easat-10389	360	4	a.	a.	NOUN
easat-10389	360	5	kara	kara	PROPN
easat-10389	360	6	,	,	PUNCT
easat-10389	360	7	"	"	PUNCT
easat-10389	360	8	multi	multi	ADJ
easat-10389	360	9	-	-	ADJ
easat-10389	360	10	step	step	ADJ
easat-10389	360	11	influenza	influenza	NOUN
easat-10389	360	12	outbreak	outbreak	NOUN
easat-10389	360	13	forecasting	forecasting	NOUN
easat-10389	360	14	using	use	VERB
easat-10389	360	15	deep	deep	ADJ
easat-10389	360	16	lstm	lstm	NOUN
easat-10389	360	17	network	network	NOUN
easat-10389	360	18	and	and	CCONJ
easat-10389	360	19	genetic	genetic	ADJ
easat-10389	360	20	algorithm	algorithm	NOUN
easat-10389	360	21	,	,	PUNCT
easat-10389	360	22	"	"	PUNCT
easat-10389	360	23	expert	expert	NOUN
easat-10389	360	24	systems	system	NOUN
easat-10389	360	25	with	with	ADP
easat-10389	360	26	applications	application	NOUN
easat-10389	360	27	,	,	PUNCT
easat-10389	360	28	vol	vol	NOUN
easat-10389	360	29	.	.	PROPN
easat-10389	360	30	180	180	NUM
easat-10389	360	31	,	,	PUNCT
easat-10389	360	32	p.	p.	NOUN
easat-10389	360	33	115153	115153	NUM
easat-10389	360	34	,	,	PUNCT
easat-10389	360	35	2021	2021	NUM
easat-10389	360	36	.	.	PUNCT
easat-10389	361	1	https://doi.org/10.1016/j.eswa.2021.115153	https://doi.org/10.1016/j.eswa.2021.115153	PROPN
easat-10389	361	2	[	[	X
easat-10389	361	3	15	15	NUM
easat-10389	361	4	]	]	X
easat-10389	361	5	j.	j.	PROPN
easat-10389	361	6	dauparas	dauparas	PROPN
easat-10389	361	7	et	et	PROPN
easat-10389	361	8	al	al	PROPN
easat-10389	361	9	.	.	PROPN
easat-10389	361	10	,	,	PUNCT
easat-10389	361	11	"	"	PUNCT
easat-10389	361	12	robust	robust	ADJ
easat-10389	361	13	deep	deep	ADJ
easat-10389	361	14	learning	learning	NOUN
easat-10389	361	15	–	–	PUNCT
easat-10389	361	16	based	base	VERB
easat-10389	361	17	protein	protein	NOUN
easat-10389	361	18	sequence	sequence	NOUN
easat-10389	361	19	design	design	NOUN
easat-10389	361	20	using	use	VERB
easat-10389	361	21	proteinmpnn	proteinmpnn	NOUN
easat-10389	361	22	,	,	PUNCT
easat-10389	361	23	"	"	PUNCT
easat-10389	361	24	science	science	NOUN
easat-10389	361	25	,	,	PUNCT
easat-10389	361	26	vol	vol	NOUN
easat-10389	361	27	.	.	PROPN
easat-10389	361	28	378	378	NUM
easat-10389	361	29	,	,	PUNCT
easat-10389	361	30	no	no	INTJ
easat-10389	361	31	.	.	NOUN
easat-10389	361	32	6615	6615	NUM
easat-10389	361	33	,	,	PUNCT
easat-10389	361	34	pp	pp	ADV
easat-10389	361	35	.	.	PUNCT
easat-10389	362	1	49	49	NUM
easat-10389	362	2	-	-	SYM
easat-10389	362	3	56	56	NUM
easat-10389	362	4	,	,	PUNCT
easat-10389	362	5	2022	2022	NUM
easat-10389	362	6	.	.	PUNCT
easat-10389	363	1	https://doi.org/10.1126/science.add2187	https://doi.org/10.1126/science.add2187	PROPN
easat-10389	364	1	[	[	X
easat-10389	364	2	16	16	NUM
easat-10389	364	3	]	]	PUNCT
easat-10389	364	4	s.	s.	PROPN
easat-10389	364	5	v.	v.	ADP
easat-10389	364	6	razavi	razavi	PROPN
easat-10389	364	7	-	-	PUNCT
easat-10389	364	8	termeh	termeh	NOUN
easat-10389	364	9	,	,	PUNCT
easat-10389	364	10	a.	a.	PROPN
easat-10389	364	11	sadeghi	sadeghi	PROPN
easat-10389	364	12	-	-	PUNCT
easat-10389	364	13	niaraki	niaraki	PROPN
easat-10389	364	14	,	,	PUNCT
easat-10389	364	15	and	and	CCONJ
easat-10389	364	16	s.-m	s.-m	NOUN
easat-10389	364	17	.	.	PUNCT
easat-10389	365	1	choi	choi	NOUN
easat-10389	365	2	,	,	PUNCT
easat-10389	365	3	"	"	PUNCT
easat-10389	365	4	spatio	spatio	NOUN
easat-10389	365	5	-	-	PUNCT
easat-10389	365	6	temporal	temporal	ADJ
easat-10389	365	7	modelling	modelling	NOUN
easat-10389	365	8	of	of	ADP
easat-10389	365	9	asthma	asthma	NOUN
easat-10389	365	10	-	-	PUNCT
easat-10389	365	11	prone	prone	ADJ
easat-10389	365	12	areas	area	NOUN
easat-10389	365	13	using	use	VERB
easat-10389	365	14	a	a	DET
easat-10389	365	15	machine	machine	NOUN
easat-10389	365	16	learning	learning	NOUN
easat-10389	365	17	optimized	optimize	VERB
easat-10389	365	18	with	with	ADP
easat-10389	365	19	metaheuristic	metaheuristic	ADJ
easat-10389	365	20	algorithms	algorithm	NOUN
easat-10389	365	21	,	,	PUNCT
easat-10389	365	22	"	"	PUNCT
easat-10389	365	23	geocarto	geocarto	PROPN
easat-10389	365	24	international	international	PROPN
easat-10389	365	25	,	,	PUNCT
easat-10389	365	26	vol	vol	NOUN
easat-10389	365	27	.	.	PROPN
easat-10389	365	28	37	37	NUM
easat-10389	365	29	,	,	PUNCT
easat-10389	365	30	no	no	INTJ
easat-10389	365	31	.	.	NOUN
easat-10389	365	32	25	25	NUM
easat-10389	365	33	,	,	PUNCT
easat-10389	365	34	pp	pp	ADJ
easat-10389	365	35	.	.	PUNCT
easat-10389	365	36	9917	9917	NUM
easat-10389	365	37	-	-	SYM
easat-10389	365	38	9942	9942	NUM
easat-10389	365	39	,	,	PUNCT
easat-10389	365	40	2022	2022	NUM
easat-10389	365	41	.	.	PUNCT
easat-10389	366	1	https://doi.org/10.1080/10106049.2022.2028903	https://doi.org/10.1080/10106049.2022.2028903	PROPN
easat-10389	367	1	[	[	X
easat-10389	367	2	17	17	NUM
easat-10389	367	3	]	]	X
easat-10389	367	4	y.	y.	PROPN
easat-10389	367	5	zheng	zheng	PROPN
easat-10389	367	6	,	,	PUNCT
easat-10389	367	7	"	"	PUNCT
easat-10389	367	8	a	a	DET
easat-10389	367	9	new	new	ADJ
easat-10389	367	10	perspective	perspective	NOUN
easat-10389	367	11	on	on	ADP
easat-10389	367	12	air	air	NOUN
easat-10389	367	13	quality	quality	PROPN
easat-10389	367	14	index	index	NOUN
easat-10389	367	15	time	time	NOUN
easat-10389	367	16	series	series	PROPN
easat-10389	367	17	forecasting	forecasting	PROPN
easat-10389	367	18	,	,	PUNCT
easat-10389	367	19	"	"	PUNCT
easat-10389	367	20	science	science	NOUN
easat-10389	367	21	of	of	ADP
easat-10389	367	22	the	the	DET
easat-10389	367	23	total	total	ADJ
easat-10389	367	24	environment	environment	NOUN
easat-10389	367	25	,	,	PUNCT
easat-10389	367	26	vol	vol	NOUN
easat-10389	367	27	.	.	PROPN
easat-10389	367	28	823	823	NUM
easat-10389	367	29	,	,	PUNCT
easat-10389	367	30	p.	p.	NOUN
easat-10389	367	31	153697	153697	NUM
easat-10389	367	32	,	,	PUNCT
easat-10389	367	33	2021	2021	NUM
easat-10389	367	34	.	.	PUNCT
easat-10389	368	1	[	[	X
easat-10389	368	2	18	18	NUM
easat-10389	368	3	]	]	X
easat-10389	368	4	y.	y.	PROPN
easat-10389	368	5	wang	wang	PROPN
easat-10389	368	6	et	et	PROPN
easat-10389	368	7	al	al	PROPN
easat-10389	368	8	.	.	PROPN
easat-10389	368	9	,	,	PUNCT
easat-10389	368	10	"	"	PUNCT
easat-10389	368	11	self	self	NOUN
easat-10389	368	12	-	-	PUNCT
easat-10389	368	13	instruct	instruct	NOUN
easat-10389	368	14	:	:	PUNCT
easat-10389	368	15	aligning	align	VERB
easat-10389	368	16	language	language	NOUN
easat-10389	368	17	models	model	NOUN
easat-10389	368	18	with	with	ADP
easat-10389	368	19	self	self	NOUN
easat-10389	368	20	-	-	PUNCT
easat-10389	368	21	generated	generate	VERB
easat-10389	368	22	instructions	instruction	NOUN
easat-10389	368	23	,	,	PUNCT
easat-10389	368	24	"	"	PUNCT
easat-10389	368	25	arxiv	arxiv	PROPN
easat-10389	368	26	preprint	preprint	VERB
easat-10389	368	27	arxiv:2212.10560	arxiv:2212.10560	ADV
easat-10389	368	28	,	,	PUNCT
easat-10389	368	29	2022	2022	NUM
easat-10389	368	30	.	.	PUNCT
easat-10389	369	1	[	[	X
easat-10389	369	2	19	19	NUM
easat-10389	369	3	]	]	X
easat-10389	369	4	j.	j.	PROPN
easat-10389	369	5	huang	huang	PROPN
easat-10389	369	6	,	,	PUNCT
easat-10389	369	7	y.	y.	PROPN
easat-10389	369	8	zhang	zhang	PROPN
easat-10389	369	9	,	,	PUNCT
easat-10389	369	10	x.	x.	PROPN
easat-10389	369	11	li	li	PROPN
easat-10389	369	12	,	,	PUNCT
easat-10389	369	13	and	and	CCONJ
easat-10389	369	14	l.	l.	PROPN
easat-10389	369	15	wang	wang	PROPN
easat-10389	369	16	,	,	PUNCT
easat-10389	369	17	"	"	PUNCT
easat-10389	369	18	a	a	DET
easat-10389	369	19	deep	deep	ADJ
easat-10389	369	20	learning	learning	NOUN
easat-10389	369	21	–	–	PUNCT
easat-10389	369	22	random	random	ADJ
easat-10389	369	23	forest	forest	NOUN
easat-10389	369	24	hybrid	hybrid	NOUN
easat-10389	369	25	model	model	NOUN
easat-10389	369	26	for	for	ADP
easat-10389	369	27	predicting	predict	VERB
easat-10389	369	28	historical	historical	ADJ
easat-10389	369	29	temperature	temperature	NOUN
easat-10389	369	30	variations	variation	NOUN
easat-10389	369	31	driven	drive	VERB
easat-10389	369	32	by	by	ADP
easat-10389	369	33	air	air	NOUN
easat-10389	369	34	pollution	pollution	NOUN
easat-10389	369	35	:	:	PUNCT
easat-10389	369	36	methodological	methodological	ADJ
easat-10389	369	37	insights	insight	NOUN
easat-10389	369	38	from	from	ADP
easat-10389	369	39	wuhan	wuhan	PROPN
easat-10389	369	40	,	,	PUNCT
easat-10389	369	41	"	"	PUNCT
easat-10389	369	42	atmosphere	atmosphere	NOUN
easat-10389	369	43	,	,	PUNCT
easat-10389	369	44	vol	vol	NOUN
easat-10389	369	45	.	.	PROPN
easat-10389	369	46	16	16	NUM
easat-10389	369	47	,	,	PUNCT
easat-10389	369	48	no	no	INTJ
easat-10389	369	49	.	.	NOUN
easat-10389	369	50	9	9	NUM
easat-10389	369	51	,	,	PUNCT
easat-10389	369	52	p.	p.	NOUN
easat-10389	369	53	1056	1056	NUM
easat-10389	369	54	,	,	PUNCT
easat-10389	369	55	2021	2021	NUM
easat-10389	369	56	.	.	PUNCT
easat-10389	370	1	[	[	X
easat-10389	370	2	20	20	NUM
easat-10389	370	3	]	]	X
easat-10389	370	4	h.	h.	PROPN
easat-10389	370	5	lee	lee	PROPN
easat-10389	370	6	,	,	PUNCT
easat-10389	370	7	s.	s.	PROPN
easat-10389	370	8	kim	kim	PROPN
easat-10389	370	9	,	,	PUNCT
easat-10389	370	10	and	and	CCONJ
easat-10389	370	11	j.	j.	PROPN
easat-10389	370	12	park	park	PROPN
easat-10389	370	13	,	,	PUNCT
easat-10389	370	14	"	"	PUNCT
easat-10389	370	15	air	air	NOUN
easat-10389	370	16	-	-	PUNCT
easat-10389	370	17	quality	quality	NOUN
easat-10389	370	18	prediction	prediction	NOUN
easat-10389	370	19	based	base	VERB
easat-10389	370	20	on	on	ADP
easat-10389	370	21	the	the	DET
easat-10389	370	22	arima	arima	PROPN
easat-10389	370	23	-	-	PUNCT
easat-10389	370	24	cnn	cnn	PROPN
easat-10389	370	25	-	-	PUNCT
easat-10389	370	26	lstm	lstm	ADJ
easat-10389	370	27	combination	combination	NOUN
easat-10389	370	28	model	model	NOUN
easat-10389	370	29	optimized	optimize	VERB
easat-10389	370	30	by	by	ADP
easat-10389	370	31	dung	dung	NOUN
easat-10389	370	32	beetle	beetle	NOUN
easat-10389	370	33	optimizer	optimizer	NOUN
easat-10389	370	34	,	,	PUNCT
easat-10389	370	35	"	"	PUNCT
easat-10389	370	36	scientific	scientific	ADJ
easat-10389	370	37	reports	report	NOUN
easat-10389	370	38	,	,	PUNCT
easat-10389	370	39	vol	vol	NOUN
easat-10389	370	40	.	.	PROPN
easat-10389	370	41	13	13	NUM
easat-10389	370	42	,	,	PUNCT
easat-10389	370	43	no	no	INTJ
easat-10389	370	44	.	.	NOUN
easat-10389	370	45	1	1	NUM
easat-10389	370	46	,	,	PUNCT
easat-10389	370	47	p.	p.	NOUN
easat-10389	370	48	36620	36620	NUM
easat-10389	370	49	,	,	PUNCT
easat-10389	370	50	2022	2022	NUM
easat-10389	370	51	.	.	PUNCT
easat-10389	371	1	[	[	X
easat-10389	371	2	21	21	NUM
easat-10389	371	3	]	]	X
easat-10389	371	4	y.	y.	PROPN
easat-10389	371	5	zhao	zhao	PROPN
easat-10389	371	6	,	,	PUNCT
easat-10389	371	7	r.	r.	PROPN
easat-10389	371	8	chen	chen	PROPN
easat-10389	371	9	,	,	PUNCT
easat-10389	371	10	h.	h.	PROPN
easat-10389	371	11	li	li	PROPN
easat-10389	371	12	,	,	PUNCT
easat-10389	371	13	and	and	CCONJ
easat-10389	371	14	x.	x.	NOUN
easat-10389	371	15	sun	sun	PROPN
easat-10389	371	16	,	,	PUNCT
easat-10389	371	17	"	"	PUNCT
easat-10389	371	18	enhanced	enhance	VERB
easat-10389	371	19	air	air	NOUN
easat-10389	371	20	quality	quality	NOUN
easat-10389	371	21	prediction	prediction	NOUN
easat-10389	371	22	using	use	VERB
easat-10389	371	23	adaptive	adaptive	ADJ
easat-10389	371	24	residual	residual	ADJ
easat-10389	371	25	bi	bi	NOUN
easat-10389	371	26	-	-	NOUN
easat-10389	371	27	lstm	lstm	NOUN
easat-10389	371	28	with	with	ADP
easat-10389	371	29	pyramid	pyramid	NOUN
easat-10389	371	30	dilation	dilation	NOUN
easat-10389	371	31	and	and	CCONJ
easat-10389	371	32	optimal	optimal	ADJ
easat-10389	371	33	weighted	weight	VERB
easat-10389	371	34	feature	feature	NOUN
easat-10389	371	35	selection	selection	NOUN
easat-10389	371	36	,	,	PUNCT
easat-10389	371	37	"	"	PUNCT
easat-10389	371	38	scientific	scientific	ADJ
easat-10389	371	39	reports	report	NOUN
easat-10389	371	40	,	,	PUNCT
easat-10389	371	41	vol	vol	NOUN
easat-10389	371	42	.	.	PROPN
easat-10389	371	43	15	15	NUM
easat-10389	371	44	,	,	PUNCT
easat-10389	371	45	no	no	INTJ
easat-10389	371	46	.	.	NOUN
easat-10389	371	47	1	1	NUM
easat-10389	371	48	,	,	PUNCT
easat-10389	371	49	p.	p.	NOUN
easat-10389	371	50	14668	14668	NUM
easat-10389	371	51	,	,	PUNCT
easat-10389	371	52	2023	2023	NUM
easat-10389	371	53	.	.	PUNCT
easat-10389	372	1	[	[	X
easat-10389	372	2	22	22	NUM
easat-10389	372	3	]	]	PUNCT
easat-10389	372	4	j.	j.	PROPN
easat-10389	372	5	kim	kim	PROPN
easat-10389	372	6	and	and	CCONJ
easat-10389	372	7	s.	s.	PROPN
easat-10389	372	8	park	park	PROPN
easat-10389	372	9	,	,	PUNCT
easat-10389	372	10	"	"	PUNCT
easat-10389	372	11	a	a	DET
easat-10389	372	12	hybrid	hybrid	ADJ
easat-10389	372	13	deep	deep	ADJ
easat-10389	372	14	learning	learning	NOUN
easat-10389	372	15	model	model	NOUN
easat-10389	372	16	for	for	ADP
easat-10389	372	17	air	air	NOUN
easat-10389	372	18	quality	quality	NOUN
easat-10389	372	19	prediction	prediction	NOUN
easat-10389	372	20	using	use	VERB
easat-10389	372	21	cnn	cnn	PROPN
easat-10389	372	22	and	and	CCONJ
easat-10389	372	23	gru	gru	VERB
easat-10389	372	24	with	with	ADP
easat-10389	372	25	attention	attention	NOUN
easat-10389	372	26	fusion	fusion	NOUN
easat-10389	372	27	,	,	PUNCT
easat-10389	372	28	"	"	PUNCT
easat-10389	372	29	environmental	environmental	ADJ
easat-10389	372	30	science	science	NOUN
easat-10389	372	31	and	and	CCONJ
easat-10389	372	32	pollution	pollution	NOUN
easat-10389	372	33	research	research	NOUN
easat-10389	372	34	,	,	PUNCT
easat-10389	372	35	vol	vol	NOUN
easat-10389	372	36	.	.	PROPN
easat-10389	372	37	31	31	NUM
easat-10389	372	38	,	,	PUNCT
easat-10389	372	39	no	no	INTJ
easat-10389	372	40	.	.	NOUN
easat-10389	372	41	12	12	NUM
easat-10389	372	42	,	,	PUNCT
easat-10389	372	43	pp	pp	ADJ
easat-10389	372	44	.	.	PUNCT
easat-10389	373	1	12345–12356	12345–12356	NUM
easat-10389	373	2	,	,	PUNCT
easat-10389	373	3	2024	2024	NUM
easat-10389	373	4	.	.	PUNCT
easat-10389	374	1	[	[	X
easat-10389	374	2	23	23	NUM
easat-10389	374	3	]	]	X
easat-10389	374	4	y.	y.	PROPN
easat-10389	374	5	li	li	PROPN
easat-10389	374	6	,	,	PUNCT
easat-10389	374	7	h.	h.	PROPN
easat-10389	374	8	zhang	zhang	PROPN
easat-10389	374	9	,	,	PUNCT
easat-10389	374	10	and	and	CCONJ
easat-10389	374	11	l.	l.	PROPN
easat-10389	374	12	wang	wang	PROPN
easat-10389	374	13	,	,	PUNCT
easat-10389	374	14	"	"	PUNCT
easat-10389	374	15	transformer	transformer	NOUN
easat-10389	374	16	-	-	PUNCT
easat-10389	374	17	based	base	VERB
easat-10389	374	18	encoder	encoder	NOUN
easat-10389	374	19	-	-	PUNCT
easat-10389	374	20	decoder	decoder	NOUN
easat-10389	374	21	model	model	NOUN
easat-10389	374	22	for	for	ADP
easat-10389	374	23	real	real	ADJ
easat-10389	374	24	-	-	PUNCT
easat-10389	374	25	time	time	NOUN
easat-10389	374	26	air	air	NOUN
easat-10389	374	27	quality	quality	NOUN
easat-10389	374	28	prediction	prediction	NOUN
easat-10389	374	29	using	use	VERB
easat-10389	374	30	sensor	sensor	NOUN
easat-10389	374	31	data	datum	NOUN
easat-10389	374	32	,	,	PUNCT
easat-10389	374	33	"	"	PUNCT
easat-10389	374	34	atmospheric	atmospheric	ADJ
easat-10389	374	35	environment	environment	NOUN
easat-10389	374	36	,	,	PUNCT
easat-10389	374	37	vol	vol	NOUN
easat-10389	374	38	.	.	PUNCT
easat-10389	375	1	259	259	NUM
easat-10389	375	2	,	,	PUNCT
easat-10389	375	3	p.	p.	NOUN
easat-10389	375	4	118456	118456	NUM
easat-10389	375	5	,	,	PUNCT
easat-10389	375	6	2025	2025	NUM
easat-10389	375	7	.	.	PUNCT
easat-10389	376	1	[	[	X
easat-10389	376	2	24	24	NUM
easat-10389	376	3	]	]	PUNCT
easat-10389	376	4	x.	x.	PROPN
easat-10389	376	5	zhang	zhang	PROPN
easat-10389	376	6	,	,	PUNCT
easat-10389	376	7	j.	j.	PROPN
easat-10389	376	8	li	li	PROPN
easat-10389	376	9	,	,	PUNCT
easat-10389	376	10	and	and	CCONJ
easat-10389	376	11	y.	y.	PROPN
easat-10389	376	12	chen	chen	PROPN
easat-10389	376	13	,	,	PUNCT
easat-10389	376	14	"	"	PUNCT
easat-10389	376	15	spatiotemporal	spatiotemporal	ADJ
easat-10389	376	16	graph	graph	NOUN
easat-10389	376	17	attention	attention	NOUN
easat-10389	376	18	network	network	NOUN
easat-10389	376	19	for	for	ADP
easat-10389	376	20	urban	urban	ADJ
easat-10389	376	21	air	air	PROPN
easat-10389	376	22	quality	quality	PROPN
easat-10389	376	23	forecasting	forecasting	NOUN
easat-10389	376	24	,	,	PUNCT
easat-10389	376	25	"	"	PUNCT
easat-10389	376	26	environmental	environmental	ADJ
easat-10389	376	27	modelling	modelling	NOUN
easat-10389	376	28	&	&	CCONJ
easat-10389	376	29	software	software	NOUN
easat-10389	376	30	,	,	PUNCT
easat-10389	376	31	vol	vol	NOUN
easat-10389	376	32	.	.	PROPN
easat-10389	376	33	150	150	NUM
easat-10389	376	34	,	,	PUNCT
easat-10389	376	35	p.	p.	NOUN
easat-10389	376	36	105332	105332	NUM
easat-10389	376	37	,	,	PUNCT
easat-10389	376	38	2025	2025	NUM
easat-10389	376	39	.	.	PUNCT
easat-10389	377	1	[	[	X
easat-10389	377	2	25	25	NUM
easat-10389	377	3	]	]	PUNCT
easat-10389	377	4	x.	x.	PROPN
easat-10389	377	5	chen	chen	PROPN
easat-10389	377	6	,	,	PUNCT
easat-10389	377	7	x.	x.	PROPN
easat-10389	377	8	zheng	zheng	PROPN
easat-10389	377	9	,	,	PUNCT
easat-10389	377	10	and	and	CCONJ
easat-10389	377	11	b.-x	b.-x	NOUN
easat-10389	377	12	.	.	PUNCT
easat-10389	378	1	liu	liu	PROPN
easat-10389	378	2	,	,	PUNCT
easat-10389	378	3	"	"	PUNCT
easat-10389	378	4	stock	stock	NOUN
easat-10389	378	5	price	price	NOUN
easat-10389	378	6	forecasting	forecasting	NOUN
easat-10389	378	7	model	model	NOUN
easat-10389	378	8	based	base	VERB
easat-10389	378	9	on	on	ADP
easat-10389	378	10	arima	arima	NOUN
easat-10389	378	11	-	-	PUNCT
easat-10389	378	12	tcn	tcn	PROPN
easat-10389	378	13	,	,	PUNCT
easat-10389	378	14	"	"	PUNCT
easat-10389	378	15	in	in	ADP
easat-10389	378	16	2025	2025	NUM
easat-10389	378	17	8th	8th	ADJ
easat-10389	378	18	world	world	NOUN
easat-10389	378	19	conference	conference	NOUN
easat-10389	378	20	on	on	ADP
easat-10389	378	21	computing	computing	NOUN
easat-10389	378	22	and	and	CCONJ
easat-10389	378	23	communication	communication	NOUN
easat-10389	378	24	technologies	technology	NOUN
easat-10389	378	25	(	(	PUNCT
easat-10389	378	26	wccct	wccct	NOUN
easat-10389	378	27	)	)	PUNCT
easat-10389	378	28	(	(	PUNCT
easat-10389	378	29	pp	pp	ADJ
easat-10389	378	30	.	.	PUNCT
easat-10389	379	1	425	425	NUM
easat-10389	379	2	-	-	NUM
easat-10389	379	3	431	431	NUM
easat-10389	379	4	)	)	PUNCT
easat-10389	379	5	.	.	PUNCT
easat-10389	380	1	ieee	ieee	NOUN
easat-10389	380	2	,	,	PUNCT
easat-10389	380	3	2025	2025	NUM
easat-10389	380	4	.	.	PUNCT
easat-10389	381	1	https://doi.org/10.1007/s00038-020-01479-z	https://doi.org/10.1007/s00038-020-01479-z	PROPN
easat-10389	381	2	https://doi.org/10.1016/j.envint.2022.107688	https://doi.org/10.1016/j.envint.2022.107688	PROPN
easat-10389	381	3	https://doi.org/10.3390/cli13020023	https://doi.org/10.3390/cli13020023	NOUN
easat-10389	381	4	https://doi.org/10.1126/science.aaa8415	https://doi.org/10.1126/science.aaa8415	PROPN
easat-10389	381	5	https://doi.org/10.28945/4184	https://doi.org/10.28945/4184	VERB
easat-10389	381	6	https://doi.org/10.1007/s11356-022-24956-9	https://doi.org/10.1007/s11356-022-24956-9	NOUN
easat-10389	381	7	https://doi.org/10.1109/jsen.2024.3472291	https://doi.org/10.1109/jsen.2024.3472291	PROPN
easat-10389	381	8	https://doi.org/10.1016/j.scitotenv.2023.164699	https://doi.org/10.1016/j.scitotenv.2023.164699	ADJ
easat-10389	381	9	https://doi.org/10.3390/atmos14121807	https://doi.org/10.3390/atmos14121807	NOUN
easat-10389	381	10	https://doi.org/10.1186/s12889-025-21618-6	https://doi.org/10.1186/s12889-025-21618-6	NUM
easat-10389	381	11	https://doi.org/10.1016/j.eswa.2021.115153	https://doi.org/10.1016/j.eswa.2021.115153	VERB
easat-10389	381	12	https://doi.org/10.1126/science.add2187	https://doi.org/10.1126/science.add2187	PROPN
easat-10389	381	13	https://doi.org/10.1080/10106049.2022.2028903	https://doi.org/10.1080/10106049.2022.2028903	PROPN