id	sid	tid	token	lemma	pos
fcis-10302	1	1	frontiers	frontier	NOUN
fcis-10302	1	2	in	in	ADP
fcis-10302	1	3	computing	computing	NOUN
fcis-10302	1	4	and	and	CCONJ
fcis-10302	1	5	intelligent	intelligent	ADJ
fcis-10302	1	6	systems	system	NOUN
fcis-10302	1	7	issn	issn	VERB
fcis-10302	1	8	:	:	PUNCT
fcis-10302	1	9	2832	2832	NUM
fcis-10302	1	10	-	-	SYM
fcis-10302	1	11	6024	6024	NUM
fcis-10302	1	12	|	|	NOUN
fcis-10302	1	13	vol	vol	NOUN
fcis-10302	1	14	.	.	PROPN
fcis-10302	2	1	4	4	NUM
fcis-10302	2	2	,	,	PUNCT
fcis-10302	2	3	no	no	INTJ
fcis-10302	2	4	.	.	NOUN
fcis-10302	2	5	2	2	NUM
fcis-10302	2	6	,	,	PUNCT
fcis-10302	2	7	2023	2023	NUM
fcis-10302	2	8	90	90	NUM
fcis-10302	2	9	iot	iot	NOUN
fcis-10302	2	10	intrusion	intrusion	NOUN
fcis-10302	2	11	detection	detection	NOUN
fcis-10302	2	12	model	model	NOUN
fcis-10302	2	13	based	base	VERB
fcis-10302	2	14	on	on	ADP
fcis-10302	2	15	cnn‐gru	cnn‐gru	X
fcis-10302	2	16	zhaolian	zhaolian	PROPN
fcis-10302	2	17	wang	wang	PROPN
fcis-10302	2	18	,	,	PUNCT
fcis-10302	2	19	hong	hong	PROPN
fcis-10302	2	20	huang	huang	PROPN
fcis-10302	3	1	*	*	PROPN
fcis-10302	3	2	,	,	PUNCT
fcis-10302	3	3	rui	rui	PROPN
fcis-10302	3	4	du	du	PROPN
fcis-10302	3	5	,	,	PUNCT
fcis-10302	3	6	xing	xing	PROPN
fcis-10302	3	7	li	li	PROPN
fcis-10302	3	8	,	,	PUNCT
fcis-10302	3	9	guotao	guotao	PROPN
fcis-10302	3	10	yuan	yuan	PROPN
fcis-10302	3	11	school	school	NOUN
fcis-10302	3	12	of	of	ADP
fcis-10302	3	13	materials	material	NOUN
fcis-10302	3	14	science	science	NOUN
fcis-10302	3	15	and	and	CCONJ
fcis-10302	3	16	technology	technology	NOUN
fcis-10302	3	17	,	,	PUNCT
fcis-10302	3	18	university	university	NOUN
fcis-10302	3	19	of	of	ADP
fcis-10302	3	20	science	science	NOUN
fcis-10302	3	21	and	and	CCONJ
fcis-10302	3	22	engineering	engineering	NOUN
fcis-10302	3	23	,	,	PUNCT
fcis-10302	3	24	yibin	yibin	PROPN
fcis-10302	3	25	644000	644000	NUM
fcis-10302	3	26	,	,	PUNCT
fcis-10302	3	27	china	china	PROPN
fcis-10302	3	28	*	*	PUNCT
fcis-10302	3	29	corresponding	correspond	VERB
fcis-10302	3	30	author	author	NOUN
fcis-10302	3	31	:	:	PUNCT
fcis-10302	3	32	hong	hong	PROPN
fcis-10302	3	33	huang	huang	PROPN
fcis-10302	3	34	(	(	PUNCT
fcis-10302	3	35	email	email	NOUN
fcis-10302	3	36	:	:	PUNCT
fcis-10302	3	37	huanghong@suse.edu.cn	huanghong@suse.edu.cn	NOUN
fcis-10302	3	38	)	)	PUNCT
fcis-10302	3	39	abstract	abstract	NOUN
fcis-10302	3	40	:	:	PUNCT
fcis-10302	3	41	with	with	ADP
fcis-10302	3	42	the	the	DET
fcis-10302	3	43	rapid	rapid	ADJ
fcis-10302	3	44	development	development	NOUN
fcis-10302	3	45	of	of	ADP
fcis-10302	3	46	iot	iot	PROPN
fcis-10302	3	47	technology	technology	NOUN
fcis-10302	3	48	,	,	PUNCT
fcis-10302	3	49	security	security	NOUN
fcis-10302	3	50	concerns	concern	NOUN
fcis-10302	3	51	surrounding	surround	VERB
fcis-10302	3	52	iot	iot	NOUN
fcis-10302	3	53	devices	device	NOUN
fcis-10302	3	54	have	have	AUX
fcis-10302	3	55	gained	gain	VERB
fcis-10302	3	56	attention	attention	NOUN
fcis-10302	3	57	.	.	PUNCT
fcis-10302	4	1	an	an	DET
fcis-10302	4	2	intrusion	intrusion	NOUN
fcis-10302	4	3	detection	detection	NOUN
fcis-10302	4	4	system	system	NOUN
fcis-10302	4	5	for	for	ADP
fcis-10302	4	6	iot	iot	NOUN
fcis-10302	4	7	can	can	AUX
fcis-10302	4	8	quickly	quickly	ADV
fcis-10302	4	9	and	and	CCONJ
fcis-10302	4	10	accurately	accurately	ADV
fcis-10302	4	11	identify	identify	VERB
fcis-10302	4	12	highly	highly	ADV
fcis-10302	4	13	redundant	redundant	ADJ
fcis-10302	4	14	data	data	NOUN
fcis-10302	4	15	features	feature	NOUN
fcis-10302	4	16	in	in	ADP
fcis-10302	4	17	iot	iot	ADJ
fcis-10302	4	18	traffic	traffic	NOUN
fcis-10302	4	19	categories	category	NOUN
fcis-10302	4	20	.	.	PUNCT
fcis-10302	5	1	to	to	PART
fcis-10302	5	2	reduce	reduce	VERB
fcis-10302	5	3	data	datum	NOUN
fcis-10302	5	4	,	,	PUNCT
fcis-10302	5	5	feature	feature	NOUN
fcis-10302	5	6	redundancy	redundancy	NOUN
fcis-10302	5	7	during	during	ADP
fcis-10302	5	8	the	the	DET
fcis-10302	5	9	identification	identification	NOUN
fcis-10302	5	10	process	process	NOUN
fcis-10302	5	11	,	,	PUNCT
fcis-10302	5	12	this	this	DET
fcis-10302	5	13	study	study	NOUN
fcis-10302	5	14	proposes	propose	VERB
fcis-10302	5	15	the	the	DET
fcis-10302	5	16	use	use	NOUN
fcis-10302	5	17	of	of	ADP
fcis-10302	5	18	extreme	extreme	ADJ
fcis-10302	5	19	gradient	gradient	NOUN
fcis-10302	5	20	boosting	boost	VERB
fcis-10302	5	21	(	(	PUNCT
fcis-10302	5	22	xgboost	xgboost	ADV
fcis-10302	5	23	)	)	PUNCT
fcis-10302	5	24	for	for	ADP
fcis-10302	5	25	feature	feature	NOUN
fcis-10302	5	26	selection	selection	NOUN
fcis-10302	5	27	to	to	PART
fcis-10302	5	28	obtain	obtain	VERB
fcis-10302	5	29	an	an	DET
fcis-10302	5	30	optimal	optimal	ADJ
fcis-10302	5	31	feature	feature	NOUN
fcis-10302	5	32	subset	subset	NOUN
fcis-10302	5	33	.	.	PUNCT
fcis-10302	6	1	additionally	additionally	ADV
fcis-10302	6	2	,	,	PUNCT
fcis-10302	6	3	to	to	PART
fcis-10302	6	4	improve	improve	VERB
fcis-10302	6	5	the	the	DET
fcis-10302	6	6	accuracy	accuracy	NOUN
fcis-10302	6	7	of	of	ADP
fcis-10302	6	8	identifying	identify	VERB
fcis-10302	6	9	malicious	malicious	ADJ
fcis-10302	6	10	traffic	traffic	NOUN
fcis-10302	6	11	in	in	ADP
fcis-10302	6	12	iot	iot	PROPN
fcis-10302	6	13	devices	device	NOUN
fcis-10302	6	14	,	,	PUNCT
fcis-10302	6	15	a	a	DET
fcis-10302	6	16	fusion	fusion	NOUN
fcis-10302	6	17	model	model	NOUN
fcis-10302	6	18	combining	combine	VERB
fcis-10302	6	19	convolutional	convolutional	ADJ
fcis-10302	6	20	neural	neural	ADJ
fcis-10302	6	21	networks	network	NOUN
fcis-10302	6	22	(	(	PUNCT
fcis-10302	6	23	cnn	cnn	PROPN
fcis-10302	6	24	)	)	PUNCT
fcis-10302	6	25	and	and	CCONJ
fcis-10302	6	26	gated	gate	VERB
fcis-10302	6	27	recurrent	recurrent	ADJ
fcis-10302	6	28	units	unit	NOUN
fcis-10302	6	29	(	(	PUNCT
fcis-10302	6	30	gru	gru	PROPN
fcis-10302	6	31	)	)	PUNCT
fcis-10302	6	32	for	for	ADP
fcis-10302	6	33	iot	iot	PROPN
fcis-10302	6	34	intrusion	intrusion	NOUN
fcis-10302	6	35	detection	detection	NOUN
fcis-10302	6	36	is	be	AUX
fcis-10302	6	37	proposed	propose	VERB
fcis-10302	6	38	.	.	PUNCT
fcis-10302	7	1	finally	finally	ADV
fcis-10302	7	2	,	,	PUNCT
fcis-10302	7	3	a	a	DET
fcis-10302	7	4	comparative	comparative	ADJ
fcis-10302	7	5	analysis	analysis	NOUN
fcis-10302	7	6	experiment	experiment	NOUN
fcis-10302	7	7	is	be	AUX
fcis-10302	7	8	conducted	conduct	VERB
fcis-10302	7	9	between	between	ADP
fcis-10302	7	10	cnn	cnn	PROPN
fcis-10302	7	11	-	-	PUNCT
fcis-10302	7	12	gru	gru	PROPN
fcis-10302	7	13	and	and	CCONJ
fcis-10302	7	14	cnn	cnn	PROPN
fcis-10302	7	15	-	-	PUNCT
fcis-10302	7	16	lstm	lstm	PROPN
fcis-10302	7	17	,	,	PUNCT
fcis-10302	7	18	demonstrating	demonstrate	VERB
fcis-10302	7	19	that	that	SCONJ
fcis-10302	7	20	the	the	DET
fcis-10302	7	21	proposed	propose	VERB
fcis-10302	7	22	model	model	NOUN
fcis-10302	7	23	achieves	achieve	VERB
fcis-10302	7	24	lower	low	ADJ
fcis-10302	7	25	processing	processing	NOUN
fcis-10302	7	26	time	time	NOUN
fcis-10302	7	27	while	while	SCONJ
fcis-10302	7	28	ensuring	ensure	VERB
fcis-10302	7	29	accuracy	accuracy	NOUN
fcis-10302	7	30	.	.	PUNCT
fcis-10302	8	1	furthermore	furthermore	ADV
fcis-10302	8	2	,	,	PUNCT
fcis-10302	8	3	the	the	DET
fcis-10302	8	4	proposed	propose	VERB
fcis-10302	8	5	method	method	NOUN
fcis-10302	8	6	outperforms	outperform	VERB
fcis-10302	8	7	classical	classical	ADJ
fcis-10302	8	8	iot	iot	PROPN
fcis-10302	8	9	intrusion	intrusion	NOUN
fcis-10302	8	10	detection	detection	NOUN
fcis-10302	8	11	algorithms	algorithm	NOUN
fcis-10302	8	12	in	in	ADP
fcis-10302	8	13	terms	term	NOUN
fcis-10302	8	14	of	of	ADP
fcis-10302	8	15	precision	precision	NOUN
fcis-10302	8	16	and	and	CCONJ
fcis-10302	8	17	recall	recall	NOUN
fcis-10302	8	18	.	.	PUNCT
fcis-10302	9	1	keywords	keyword	NOUN
fcis-10302	9	2	:	:	PUNCT
fcis-10302	9	3	iot	iot	PROPN
fcis-10302	9	4	intrusion	intrusion	PROPN
fcis-10302	9	5	detection	detection	NOUN
fcis-10302	9	6	;	;	PUNCT
fcis-10302	9	7	cnn	cnn	PROPN
fcis-10302	9	8	;	;	PUNCT
fcis-10302	9	9	gru	gru	PROPN
fcis-10302	9	10	.	.	PROPN
fcis-10302	9	11	1	1	X
fcis-10302	9	12	.	.	X
fcis-10302	9	13	introduction	introduction	NOUN
fcis-10302	9	14	in	in	ADP
fcis-10302	9	15	recent	recent	ADJ
fcis-10302	9	16	years	year	NOUN
fcis-10302	9	17	,	,	PUNCT
fcis-10302	9	18	with	with	ADP
fcis-10302	9	19	the	the	DET
fcis-10302	9	20	development	development	NOUN
fcis-10302	9	21	of	of	ADP
fcis-10302	9	22	iot	iot	PROPN
fcis-10302	9	23	technology	technology	NOUN
fcis-10302	9	24	,	,	PUNCT
fcis-10302	9	25	the	the	DET
fcis-10302	9	26	number	number	NOUN
fcis-10302	9	27	and	and	CCONJ
fcis-10302	9	28	scope	scope	NOUN
fcis-10302	9	29	of	of	ADP
fcis-10302	9	30	iot	iot	PROPN
fcis-10302	9	31	devices	device	NOUN
fcis-10302	9	32	have	have	AUX
fcis-10302	9	33	been	be	AUX
fcis-10302	9	34	continuously	continuously	ADV
fcis-10302	9	35	expanding	expand	VERB
fcis-10302	9	36	,	,	PUNCT
fcis-10302	9	37	becoming	become	VERB
fcis-10302	9	38	an	an	DET
fcis-10302	9	39	important	important	ADJ
fcis-10302	9	40	driving	driving	NOUN
fcis-10302	9	41	force	force	NOUN
fcis-10302	9	42	for	for	ADP
fcis-10302	9	43	digital	digital	ADJ
fcis-10302	9	44	transformation	transformation	NOUN
fcis-10302	9	45	and	and	CCONJ
fcis-10302	9	46	intelligent	intelligent	ADJ
fcis-10302	9	47	development	development	NOUN
fcis-10302	9	48	.	.	PUNCT
fcis-10302	10	1	at	at	ADP
fcis-10302	10	2	the	the	DET
fcis-10302	10	3	same	same	ADJ
fcis-10302	10	4	time	time	NOUN
fcis-10302	10	5	,	,	PUNCT
fcis-10302	10	6	ensuring	ensure	VERB
fcis-10302	10	7	the	the	DET
fcis-10302	10	8	security	security	NOUN
fcis-10302	10	9	of	of	ADP
fcis-10302	10	10	iot	iot	PROPN
fcis-10302	10	11	devices	device	NOUN
fcis-10302	10	12	and	and	CCONJ
fcis-10302	10	13	applications	application	NOUN
fcis-10302	10	14	in	in	ADP
fcis-10302	10	15	small	small	ADJ
fcis-10302	10	16	and	and	CCONJ
fcis-10302	10	17	large	large	ADJ
fcis-10302	10	18	networks	network	NOUN
fcis-10302	10	19	composed	compose	VERB
fcis-10302	10	20	of	of	ADP
fcis-10302	10	21	physical	physical	ADJ
fcis-10302	10	22	and	and	CCONJ
fcis-10302	10	23	virtual	virtual	ADJ
fcis-10302	10	24	infrastructure	infrastructure	NOUN
fcis-10302	10	25	has	have	AUX
fcis-10302	10	26	become	become	VERB
fcis-10302	10	27	crucial	crucial	ADJ
fcis-10302	10	28	.	.	PUNCT
fcis-10302	11	1	current	current	ADJ
fcis-10302	11	2	iot	iot	PROPN
fcis-10302	11	3	devices	device	NOUN
fcis-10302	11	4	are	be	AUX
fcis-10302	11	5	susceptible	susceptible	ADJ
fcis-10302	11	6	to	to	ADP
fcis-10302	11	7	network	network	NOUN
fcis-10302	11	8	attacks	attack	NOUN
fcis-10302	11	9	such	such	ADJ
fcis-10302	11	10	as	as	ADP
fcis-10302	11	11	denial	denial	NOUN
fcis-10302	11	12	of	of	ADP
fcis-10302	11	13	service	service	NOUN
fcis-10302	11	14	[	[	X
fcis-10302	11	15	1	1	NUM
fcis-10302	11	16	]	]	PUNCT
fcis-10302	11	17	,	,	PUNCT
fcis-10302	11	18	botnets	botnet	NOUN
fcis-10302	11	19	[	[	X
fcis-10302	11	20	2	2	NUM
fcis-10302	11	21	]	]	PUNCT
fcis-10302	11	22	,	,	PUNCT
fcis-10302	11	23	and	and	CCONJ
fcis-10302	11	24	penetration	penetration	NOUN
fcis-10302	11	25	attacks	attack	NOUN
fcis-10302	11	26	[	[	X
fcis-10302	11	27	3	3	NUM
fcis-10302	11	28	]	]	PUNCT
fcis-10302	11	29	.	.	PUNCT
fcis-10302	12	1	the	the	DET
fcis-10302	12	2	mirai	mirai	PROPN
fcis-10302	12	3	botnet	botnet	NOUN
fcis-10302	12	4	attack	attack	NOUN
fcis-10302	12	5	is	be	AUX
fcis-10302	12	6	a	a	DET
fcis-10302	12	7	distributed	distribute	VERB
fcis-10302	12	8	denial	denial	NOUN
fcis-10302	12	9	-	-	PUNCT
fcis-10302	12	10	of	of	ADP
fcis-10302	12	11	-	-	PUNCT
fcis-10302	12	12	service	service	NOUN
fcis-10302	12	13	(	(	PUNCT
fcis-10302	12	14	ddos	ddos	NOUN
fcis-10302	12	15	)	)	PUNCT
fcis-10302	12	16	attack	attack	NOUN
fcis-10302	12	17	on	on	ADP
fcis-10302	12	18	iot	iot	PROPN
fcis-10302	12	19	devices	device	NOUN
fcis-10302	12	20	[	[	X
fcis-10302	12	21	4	4	NUM
fcis-10302	12	22	]	]	PUNCT
fcis-10302	12	23	.	.	PUNCT
fcis-10302	13	1	mirai	mirai	PROPN
fcis-10302	13	2	can	can	AUX
fcis-10302	13	3	incorporate	incorporate	VERB
fcis-10302	13	4	connected	connected	ADJ
fcis-10302	13	5	devices	device	NOUN
fcis-10302	13	6	extracted	extract	VERB
fcis-10302	13	7	from	from	ADP
fcis-10302	13	8	smart	smart	ADJ
fcis-10302	13	9	homes	home	NOUN
fcis-10302	13	10	into	into	ADP
fcis-10302	13	11	a	a	DET
fcis-10302	13	12	"	"	PUNCT
fcis-10302	13	13	botnet	botnet	NOUN
fcis-10302	13	14	"	"	PUNCT
fcis-10302	14	1	[	[	X
fcis-10302	14	2	2	2	NUM
fcis-10302	14	3	]	]	PUNCT
fcis-10302	14	4	,	,	PUNCT
fcis-10302	14	5	leading	lead	VERB
fcis-10302	14	6	to	to	ADP
fcis-10302	14	7	the	the	DET
fcis-10302	14	8	shutdown	shutdown	NOUN
fcis-10302	14	9	of	of	ADP
fcis-10302	14	10	dyn	dyn	NOUN
fcis-10302	14	11	's	's	PART
fcis-10302	14	12	infrastructure	infrastructure	NOUN
fcis-10302	14	13	,	,	PUNCT
fcis-10302	14	14	a	a	DET
fcis-10302	14	15	prominent	prominent	ADJ
fcis-10302	14	16	us	us	PROPN
fcis-10302	14	17	domain	domain	NOUN
fcis-10302	14	18	name	name	NOUN
fcis-10302	14	19	service	service	NOUN
fcis-10302	14	20	provider	provider	NOUN
fcis-10302	14	21	,	,	PUNCT
fcis-10302	14	22	affecting	affect	VERB
fcis-10302	14	23	websites	website	NOUN
fcis-10302	14	24	such	such	ADJ
fcis-10302	14	25	as	as	ADP
fcis-10302	14	26	twitter	twitter	NOUN
fcis-10302	14	27	,	,	PUNCT
fcis-10302	14	28	paypal	paypal	NOUN
fcis-10302	14	29	,	,	PUNCT
fcis-10302	14	30	and	and	CCONJ
fcis-10302	14	31	github	github	PROPN
fcis-10302	14	32	.	.	PUNCT
fcis-10302	15	1	in	in	ADP
fcis-10302	15	2	terms	term	NOUN
fcis-10302	15	3	of	of	ADP
fcis-10302	15	4	feature	feature	NOUN
fcis-10302	15	5	redundancy	redundancy	NOUN
fcis-10302	15	6	,	,	PUNCT
fcis-10302	15	7	iot	iot	PROPN
fcis-10302	15	8	intrusion	intrusion	NOUN
fcis-10302	15	9	detection	detection	NOUN
fcis-10302	15	10	techniques	technique	NOUN
fcis-10302	15	11	aim	aim	VERB
fcis-10302	15	12	to	to	PART
fcis-10302	15	13	reduce	reduce	VERB
fcis-10302	15	14	the	the	DET
fcis-10302	15	15	number	number	NOUN
fcis-10302	15	16	of	of	ADP
fcis-10302	15	17	features	feature	NOUN
fcis-10302	15	18	to	to	PART
fcis-10302	15	19	not	not	PART
fcis-10302	15	20	only	only	ADV
fcis-10302	15	21	shorten	shorten	VERB
fcis-10302	15	22	the	the	DET
fcis-10302	15	23	model	model	NOUN
fcis-10302	15	24	training	training	NOUN
fcis-10302	15	25	time	time	NOUN
fcis-10302	15	26	but	but	CCONJ
fcis-10302	15	27	also	also	ADV
fcis-10302	15	28	improve	improve	VERB
fcis-10302	15	29	detection	detection	NOUN
fcis-10302	15	30	effectiveness	effectiveness	NOUN
fcis-10302	15	31	.	.	PUNCT
fcis-10302	16	1	liu	liu	PROPN
fcis-10302	16	2	et	et	PROPN
fcis-10302	16	3	al	al	PROPN
fcis-10302	16	4	.	.	PUNCT
fcis-10302	17	1	[	[	X
fcis-10302	17	2	11	11	NUM
fcis-10302	17	3	]	]	PUNCT
fcis-10302	17	4	proposed	propose	VERB
fcis-10302	17	5	an	an	DET
fcis-10302	17	6	intrusion	intrusion	NOUN
fcis-10302	17	7	detection	detection	NOUN
fcis-10302	17	8	method	method	NOUN
fcis-10302	17	9	based	base	VERB
fcis-10302	17	10	on	on	ADP
fcis-10302	17	11	principal	principal	ADJ
fcis-10302	17	12	component	component	NOUN
fcis-10302	17	13	analysis	analysis	NOUN
fcis-10302	17	14	(	(	PUNCT
fcis-10302	17	15	pca	pca	NOUN
fcis-10302	17	16	)	)	PUNCT
fcis-10302	17	17	and	and	CCONJ
fcis-10302	17	18	recurrent	recurrent	ADJ
fcis-10302	17	19	neural	neural	ADJ
fcis-10302	17	20	networks	network	NOUN
fcis-10302	17	21	(	(	PUNCT
fcis-10302	17	22	rnn	rnn	PROPN
fcis-10302	17	23	)	)	PUNCT
fcis-10302	17	24	.	.	PUNCT
fcis-10302	18	1	pca	pca	PROPN
fcis-10302	18	2	was	be	AUX
fcis-10302	18	3	used	use	VERB
fcis-10302	18	4	to	to	PART
fcis-10302	18	5	reduce	reduce	VERB
fcis-10302	18	6	data	data	NOUN
fcis-10302	18	7	dimensionality	dimensionality	NOUN
fcis-10302	18	8	and	and	CCONJ
fcis-10302	18	9	noise	noise	NOUN
fcis-10302	18	10	and	and	CCONJ
fcis-10302	18	11	identify	identify	VERB
fcis-10302	18	12	the	the	DET
fcis-10302	18	13	most	most	ADV
fcis-10302	18	14	informative	informative	ADJ
fcis-10302	18	15	subset	subset	NOUN
fcis-10302	18	16	of	of	ADP
fcis-10302	18	17	principal	principal	ADJ
fcis-10302	18	18	features	feature	NOUN
fcis-10302	18	19	.	.	PUNCT
fcis-10302	19	1	the	the	DET
fcis-10302	19	2	processed	process	VERB
fcis-10302	19	3	data	datum	NOUN
fcis-10302	19	4	was	be	AUX
fcis-10302	19	5	then	then	ADV
fcis-10302	19	6	trained	train	VERB
fcis-10302	19	7	using	use	VERB
fcis-10302	19	8	an	an	DET
fcis-10302	19	9	rnn	rnn	NOUN
fcis-10302	19	10	network	network	NOUN
fcis-10302	19	11	with	with	ADP
fcis-10302	19	12	high	high	ADJ
fcis-10302	19	13	classification	classification	NOUN
fcis-10302	19	14	accuracy	accuracy	NOUN
fcis-10302	19	15	.	.	PUNCT
fcis-10302	20	1	zhou	zhou	PROPN
fcis-10302	20	2	et	et	PROPN
fcis-10302	20	3	al	al	PROPN
fcis-10302	20	4	.	.	PUNCT
fcis-10302	21	1	[	[	X
fcis-10302	21	2	12	12	NUM
fcis-10302	21	3	]	]	PUNCT
fcis-10302	21	4	proposed	propose	VERB
fcis-10302	21	5	a	a	DET
fcis-10302	21	6	feature	feature	NOUN
fcis-10302	21	7	selection	selection	NOUN
fcis-10302	21	8	method	method	NOUN
fcis-10302	21	9	using	use	VERB
fcis-10302	21	10	gradient	gradient	NOUN
fcis-10302	21	11	boosting	boost	VERB
fcis-10302	21	12	trees	tree	NOUN
fcis-10302	21	13	.	.	PUNCT
fcis-10302	22	1	xgboost	xgboost	PROPN
fcis-10302	22	2	is	be	AUX
fcis-10302	22	3	a	a	DET
fcis-10302	22	4	nonlinear	nonlinear	ADJ
fcis-10302	22	5	method	method	NOUN
fcis-10302	22	6	of	of	ADP
fcis-10302	22	7	gradient	gradient	NOUN
fcis-10302	22	8	boosting	boost	VERB
fcis-10302	22	9	that	that	SCONJ
fcis-10302	22	10	not	not	PART
fcis-10302	22	11	only	only	ADV
fcis-10302	22	12	calculates	calculate	VERB
fcis-10302	22	13	feature	feature	NOUN
fcis-10302	22	14	importance	importance	NOUN
fcis-10302	22	15	for	for	ADP
fcis-10302	22	16	feature	feature	NOUN
fcis-10302	22	17	selection	selection	NOUN
fcis-10302	22	18	but	but	CCONJ
fcis-10302	22	19	also	also	ADV
fcis-10302	22	20	prevents	prevent	VERB
fcis-10302	22	21	overfitting	overfitte	VERB
fcis-10302	22	22	by	by	ADP
fcis-10302	22	23	adding	add	VERB
fcis-10302	22	24	an	an	DET
fcis-10302	22	25	l2	l2	NOUN
fcis-10302	22	26	regularization	regularization	NOUN
fcis-10302	22	27	term	term	NOUN
fcis-10302	22	28	to	to	ADP
fcis-10302	22	29	the	the	DET
fcis-10302	22	30	objective	objective	ADJ
fcis-10302	22	31	function	function	NOUN
fcis-10302	22	32	,	,	PUNCT
fcis-10302	22	33	effectively	effectively	ADV
fcis-10302	22	34	addressing	address	VERB
fcis-10302	22	35	the	the	DET
fcis-10302	22	36	problem	problem	NOUN
fcis-10302	22	37	of	of	ADP
fcis-10302	22	38	feature	feature	NOUN
fcis-10302	22	39	redundancy	redundancy	NOUN
fcis-10302	22	40	.	.	PUNCT
fcis-10302	23	1	in	in	ADP
fcis-10302	23	2	the	the	DET
fcis-10302	23	3	exploration	exploration	NOUN
fcis-10302	23	4	of	of	ADP
fcis-10302	23	5	iot	iot	PROPN
fcis-10302	23	6	intrusion	intrusion	NOUN
fcis-10302	23	7	detection	detection	NOUN
fcis-10302	23	8	models	model	NOUN
fcis-10302	23	9	,	,	PUNCT
fcis-10302	23	10	several	several	ADJ
fcis-10302	23	11	methods	method	NOUN
fcis-10302	23	12	and	and	CCONJ
fcis-10302	23	13	techniques	technique	NOUN
fcis-10302	23	14	have	have	AUX
fcis-10302	23	15	been	be	AUX
fcis-10302	23	16	proposed	propose	VERB
fcis-10302	23	17	by	by	ADP
fcis-10302	23	18	researchers	researcher	NOUN
fcis-10302	23	19	.	.	PUNCT
fcis-10302	24	1	gurusamy	gurusamy	NOUN
fcis-10302	25	1	[	[	X
fcis-10302	25	2	5	5	NUM
fcis-10302	25	3	]	]	PUNCT
fcis-10302	25	4	proposed	propose	VERB
fcis-10302	25	5	a	a	DET
fcis-10302	25	6	secure	secure	ADJ
fcis-10302	25	7	iot	iot	NOUN
fcis-10302	25	8	framework	framework	NOUN
fcis-10302	25	9	using	use	VERB
fcis-10302	25	10	software	software	NOUN
fcis-10302	25	11	-	-	PUNCT
fcis-10302	25	12	defined	define	VERB
fcis-10302	25	13	networking	networking	NOUN
fcis-10302	25	14	(	(	PUNCT
fcis-10302	25	15	sdn	sdn	PROPN
fcis-10302	25	16	)	)	PUNCT
fcis-10302	25	17	in	in	ADP
fcis-10302	25	18	iot	iot	PROPN
fcis-10302	25	19	traffic	traffic	NOUN
fcis-10302	25	20	.	.	PUNCT
fcis-10302	26	1	svm	svm	ADJ
fcis-10302	26	2	algorithms	algorithm	NOUN
fcis-10302	26	3	were	be	AUX
fcis-10302	26	4	used	use	VERB
fcis-10302	26	5	on	on	ADP
fcis-10302	26	6	the	the	DET
fcis-10302	26	7	sdn	sdn	NOUN
fcis-10302	26	8	controller	controller	NOUN
fcis-10302	26	9	to	to	PART
fcis-10302	26	10	monitor	monitor	VERB
fcis-10302	26	11	and	and	CCONJ
fcis-10302	26	12	learn	learn	VERB
fcis-10302	26	13	iot	iot	ADJ
fcis-10302	26	14	device	device	NOUN
fcis-10302	26	15	behavior	behavior	NOUN
fcis-10302	26	16	and	and	CCONJ
fcis-10302	26	17	detect	detect	NOUN
fcis-10302	26	18	attacks	attack	NOUN
fcis-10302	26	19	,	,	PUNCT
fcis-10302	26	20	achieving	achieve	VERB
fcis-10302	26	21	high	high	ADJ
fcis-10302	26	22	precision	precision	NOUN
fcis-10302	26	23	,	,	PUNCT
fcis-10302	26	24	recall	recall	NOUN
fcis-10302	26	25	,	,	PUNCT
fcis-10302	26	26	and	and	CCONJ
fcis-10302	26	27	fast	fast	ADJ
fcis-10302	26	28	response	response	NOUN
fcis-10302	26	29	time	time	NOUN
fcis-10302	26	30	.	.	PUNCT
fcis-10302	27	1	hodo	hodo	NOUN
fcis-10302	27	2	et	et	PROPN
fcis-10302	27	3	al	al	PROPN
fcis-10302	27	4	.	.	PUNCT
fcis-10302	28	1	[	[	X
fcis-10302	28	2	6	6	NUM
fcis-10302	28	3	]	]	PUNCT
fcis-10302	28	4	applied	apply	VERB
fcis-10302	28	5	artificial	artificial	ADJ
fcis-10302	28	6	neural	neural	ADJ
fcis-10302	28	7	network	network	NOUN
fcis-10302	28	8	(	(	PUNCT
fcis-10302	28	9	ann	ann	PROPN
fcis-10302	28	10	)	)	PUNCT
fcis-10302	28	11	algorithms	algorithm	NOUN
fcis-10302	28	12	to	to	PART
fcis-10302	28	13	detect	detect	VERB
fcis-10302	28	14	ddos	ddo	NOUN
fcis-10302	28	15	/	/	SYM
fcis-10302	28	16	dos	do	NOUN
fcis-10302	28	17	attacks	attack	NOUN
fcis-10302	28	18	based	base	VERB
fcis-10302	28	19	on	on	ADP
fcis-10302	28	20	the	the	DET
fcis-10302	28	21	characteristics	characteristic	NOUN
fcis-10302	28	22	of	of	ADP
fcis-10302	28	23	host	host	NOUN
fcis-10302	28	24	-	-	PUNCT
fcis-10302	28	25	based	base	VERB
fcis-10302	28	26	ids	id	NOUN
fcis-10302	28	27	and	and	CCONJ
fcis-10302	28	28	network	network	NOUN
fcis-10302	28	29	-	-	PUNCT
fcis-10302	28	30	based	base	VERB
fcis-10302	28	31	ids	id	NOUN
fcis-10302	28	32	,	,	PUNCT
fcis-10302	28	33	demonstrating	demonstrate	VERB
fcis-10302	28	34	good	good	ADJ
fcis-10302	28	35	accuracy	accuracy	NOUN
fcis-10302	28	36	in	in	ADP
fcis-10302	28	37	traffic	traffic	NOUN
fcis-10302	28	38	classification	classification	NOUN
fcis-10302	28	39	.	.	PUNCT
fcis-10302	29	1	debastrita	debastrita	PROPN
fcis-10302	29	2	et	et	PROPN
fcis-10302	29	3	al	al	PROPN
fcis-10302	29	4	.	.	PUNCT
fcis-10302	30	1	[	[	X
fcis-10302	30	2	7	7	X
fcis-10302	30	3	]	]	PUNCT
fcis-10302	30	4	proposed	propose	VERB
fcis-10302	30	5	a	a	DET
fcis-10302	30	6	set	set	ADJ
fcis-10302	30	7	learning	learning	NOUN
fcis-10302	30	8	method	method	NOUN
fcis-10302	30	9	for	for	ADP
fcis-10302	30	10	anomaly	anomaly	NOUN
fcis-10302	30	11	detection	detection	NOUN
fcis-10302	30	12	,	,	PUNCT
fcis-10302	30	13	utilizing	utilize	VERB
fcis-10302	30	14	stacked	stack	VERB
fcis-10302	30	15	autoencoders	autoencoder	NOUN
fcis-10302	30	16	(	(	PUNCT
fcis-10302	30	17	sae	sae	PROPN
fcis-10302	30	18	)	)	PUNCT
fcis-10302	30	19	to	to	PART
fcis-10302	30	20	extract	extract	VERB
fcis-10302	30	21	deep	deep	ADJ
fcis-10302	30	22	features	feature	NOUN
fcis-10302	30	23	and	and	CCONJ
fcis-10302	30	24	inputting	inputte	VERB
fcis-10302	30	25	them	they	PRON
fcis-10302	30	26	into	into	ADP
fcis-10302	30	27	a	a	DET
fcis-10302	30	28	probabilistic	probabilistic	ADJ
fcis-10302	30	29	neural	neural	ADJ
fcis-10302	30	30	network	network	NOUN
fcis-10302	30	31	ensemble	ensemble	ADJ
fcis-10302	30	32	for	for	ADP
fcis-10302	30	33	single	single	ADJ
fcis-10302	30	34	and	and	CCONJ
fcis-10302	30	35	multiple	multiple	ADJ
fcis-10302	30	36	anomaly	anomaly	NOUN
fcis-10302	30	37	detection	detection	NOUN
fcis-10302	30	38	.	.	PUNCT
fcis-10302	31	1	the	the	DET
fcis-10302	31	2	use	use	NOUN
fcis-10302	31	3	of	of	ADP
fcis-10302	31	4	autoencoders	autoencoder	NOUN
fcis-10302	31	5	resulted	result	VERB
fcis-10302	31	6	in	in	ADP
fcis-10302	31	7	good	good	ADJ
fcis-10302	31	8	and	and	CCONJ
fcis-10302	31	9	stable	stable	ADJ
fcis-10302	31	10	performance	performance	NOUN
fcis-10302	31	11	.	.	PUNCT
fcis-10302	32	1	wang	wang	PROPN
fcis-10302	32	2	yun	yun	PROPN
fcis-10302	33	1	[	[	X
fcis-10302	33	2	8	8	NUM
fcis-10302	33	3	]	]	PUNCT
fcis-10302	33	4	constructed	construct	VERB
fcis-10302	33	5	a	a	DET
fcis-10302	33	6	home	home	NOUN
fcis-10302	33	7	iot	iot	NOUN
fcis-10302	33	8	intrusion	intrusion	PROPN
fcis-10302	33	9	detection	detection	NOUN
fcis-10302	33	10	model	model	NOUN
fcis-10302	33	11	based	base	VERB
fcis-10302	33	12	on	on	ADP
fcis-10302	33	13	recurrent	recurrent	ADJ
fcis-10302	33	14	neural	neural	ADJ
fcis-10302	33	15	networks	network	NOUN
fcis-10302	33	16	(	(	PUNCT
fcis-10302	33	17	rnn	rnn	PROPN
fcis-10302	33	18	)	)	PUNCT
fcis-10302	33	19	using	use	VERB
fcis-10302	33	20	the	the	DET
fcis-10302	33	21	nsl	nsl	NOUN
fcis-10302	33	22	-	-	PUNCT
fcis-10302	33	23	kdd	kdd	PROPN
fcis-10302	33	24	dataset	dataset	NOUN
fcis-10302	33	25	,	,	PUNCT
fcis-10302	33	26	reducing	reduce	VERB
fcis-10302	33	27	false	false	ADJ
fcis-10302	33	28	positives	positive	NOUN
fcis-10302	33	29	and	and	CCONJ
fcis-10302	33	30	improving	improve	VERB
fcis-10302	33	31	accuracy	accuracy	NOUN
fcis-10302	33	32	.	.	PUNCT
fcis-10302	34	1	alkahtani	alkahtani	PROPN
fcis-10302	34	2	et	et	PROPN
fcis-10302	34	3	al	al	PROPN
fcis-10302	34	4	.	.	PUNCT
fcis-10302	35	1	[	[	X
fcis-10302	35	2	9	9	NUM
fcis-10302	35	3	]	]	PUNCT
fcis-10302	35	4	achieved	achieve	VERB
fcis-10302	35	5	certain	certain	ADJ
fcis-10302	35	6	results	result	NOUN
fcis-10302	35	7	in	in	ADP
fcis-10302	35	8	detecting	detect	VERB
fcis-10302	35	9	iot	iot	ADJ
fcis-10302	35	10	device	device	NOUN
fcis-10302	35	11	botnet	botnet	NOUN
fcis-10302	35	12	attacks	attack	NOUN
fcis-10302	35	13	using	use	VERB
fcis-10302	35	14	convolutional	convolutional	ADJ
fcis-10302	35	15	neural	neural	ADJ
fcis-10302	35	16	networks	network	NOUN
fcis-10302	35	17	(	(	PUNCT
fcis-10302	35	18	cnn	cnn	PROPN
fcis-10302	35	19	)	)	PUNCT
fcis-10302	35	20	and	and	CCONJ
fcis-10302	35	21	long	long	ADJ
fcis-10302	35	22	short	short	ADJ
fcis-10302	35	23	-	-	PUNCT
fcis-10302	35	24	term	term	NOUN
fcis-10302	35	25	memory	memory	NOUN
fcis-10302	35	26	(	(	PUNCT
fcis-10302	35	27	lstm	lstm	ADJ
fcis-10302	35	28	)	)	PUNCT
fcis-10302	35	29	algorithms	algorithm	NOUN
fcis-10302	35	30	,	,	PUNCT
fcis-10302	35	31	showing	show	VERB
fcis-10302	35	32	promising	promising	ADJ
fcis-10302	35	33	accuracy	accuracy	NOUN
fcis-10302	35	34	.	.	PUNCT
fcis-10302	36	1	li	li	PROPN
fcis-10302	36	2	xiaojia	xiaojia	PROPN
fcis-10302	36	3	et	et	PROPN
fcis-10302	36	4	al	al	PROPN
fcis-10302	36	5	.	.	PUNCT
fcis-10302	37	1	[	[	X
fcis-10302	37	2	10	10	NUM
fcis-10302	37	3	]	]	PUNCT
fcis-10302	37	4	utilized	utilize	VERB
fcis-10302	37	5	cnn	cnn	NOUN
fcis-10302	37	6	and	and	CCONJ
fcis-10302	37	7	recurrent	recurrent	ADJ
fcis-10302	37	8	neural	neural	ADJ
fcis-10302	37	9	networks	network	NOUN
fcis-10302	37	10	(	(	PUNCT
fcis-10302	37	11	rnn	rnn	PROPN
fcis-10302	37	12	)	)	PUNCT
fcis-10302	37	13	for	for	ADP
fcis-10302	37	14	iot	iot	PROPN
fcis-10302	37	15	intrusion	intrusion	NOUN
fcis-10302	37	16	detection	detection	NOUN
fcis-10302	37	17	using	use	VERB
fcis-10302	37	18	the	the	DET
fcis-10302	37	19	ton	ton	NOUN
fcis-10302	37	20	-	-	PUNCT
fcis-10302	37	21	iot	iot	NOUN
fcis-10302	37	22	dataset	dataset	NOUN
fcis-10302	37	23	,	,	PUNCT
fcis-10302	37	24	achieving	achieve	VERB
fcis-10302	37	25	good	good	ADJ
fcis-10302	37	26	precision	precision	NOUN
fcis-10302	37	27	and	and	CCONJ
fcis-10302	37	28	f1	f1	NOUN
fcis-10302	37	29	score	score	NOUN
fcis-10302	37	30	.	.	PUNCT
fcis-10302	38	1	1dcnn	1dcnn	NUM
fcis-10302	38	2	has	have	VERB
fcis-10302	38	3	the	the	DET
fcis-10302	38	4	ability	ability	NOUN
fcis-10302	38	5	to	to	PART
fcis-10302	38	6	capture	capture	VERB
fcis-10302	38	7	local	local	ADJ
fcis-10302	38	8	parallel	parallel	ADJ
fcis-10302	38	9	features	feature	NOUN
fcis-10302	38	10	in	in	ADP
fcis-10302	38	11	iot	iot	ADJ
fcis-10302	38	12	device	device	NOUN
fcis-10302	38	13	traffic	traffic	NOUN
fcis-10302	38	14	quickly	quickly	ADV
fcis-10302	38	15	,	,	PUNCT
fcis-10302	38	16	while	while	SCONJ
fcis-10302	38	17	rnn	rnn	NOUN
fcis-10302	38	18	,	,	PUNCT
fcis-10302	38	19	specifically	specifically	ADV
fcis-10302	38	20	gru	gru	VERB
fcis-10302	38	21	,	,	PUNCT
fcis-10302	38	22	can	can	AUX
fcis-10302	38	23	extract	extract	VERB
fcis-10302	38	24	long	long	ADJ
fcis-10302	38	25	-	-	PUNCT
fcis-10302	38	26	term	term	NOUN
fcis-10302	38	27	dependency	dependency	NOUN
fcis-10302	38	28	learning	learning	NOUN
fcis-10302	38	29	features	feature	NOUN
fcis-10302	38	30	.	.	PUNCT
fcis-10302	39	1	gru	gru	PROPN
fcis-10302	39	2	has	have	VERB
fcis-10302	39	3	a	a	DET
fcis-10302	39	4	simpler	simple	ADJ
fcis-10302	39	5	internal	internal	ADJ
fcis-10302	39	6	structure	structure	NOUN
fcis-10302	39	7	and	and	CCONJ
fcis-10302	39	8	fewer	few	ADJ
fcis-10302	39	9	parameters	parameter	NOUN
fcis-10302	39	10	compared	compare	VERB
fcis-10302	39	11	to	to	ADP
fcis-10302	39	12	lstm	lstm	PROPN
fcis-10302	39	13	,	,	PUNCT
fcis-10302	39	14	making	make	VERB
fcis-10302	39	15	it	it	PRON
fcis-10302	39	16	suitable	suitable	ADJ
fcis-10302	39	17	for	for	ADP
fcis-10302	39	18	accurately	accurately	ADV
fcis-10302	39	19	and	and	CCONJ
fcis-10302	39	20	efficiently	efficiently	ADV
fcis-10302	39	21	identifying	identify	VERB
fcis-10302	39	22	traffic	traffic	NOUN
fcis-10302	39	23	within	within	ADP
fcis-10302	39	24	a	a	DET
fcis-10302	39	25	short	short	ADJ
fcis-10302	39	26	period	period	NOUN
fcis-10302	39	27	of	of	ADP
fcis-10302	39	28	time	time	NOUN
fcis-10302	39	29	.	.	PUNCT
fcis-10302	40	1	this	this	DET
fcis-10302	40	2	paper	paper	NOUN
fcis-10302	40	3	proposes	propose	VERB
fcis-10302	40	4	a	a	DET
fcis-10302	40	5	hybrid	hybrid	ADJ
fcis-10302	40	6	model	model	NOUN
fcis-10302	40	7	combining	combine	VERB
fcis-10302	40	8	convolutional	convolutional	ADJ
fcis-10302	40	9	neural	neural	ADJ
fcis-10302	40	10	networks	network	NOUN
fcis-10302	40	11	(	(	PUNCT
fcis-10302	40	12	cnn	cnn	PROPN
fcis-10302	40	13	)	)	PUNCT
fcis-10302	40	14	and	and	CCONJ
fcis-10302	40	15	gated	gate	VERB
fcis-10302	40	16	recurrent	recurrent	ADJ
fcis-10302	40	17	units	unit	NOUN
fcis-10302	40	18	(	(	PUNCT
fcis-10302	40	19	gru	gru	PROPN
fcis-10302	40	20	)	)	PUNCT
fcis-10302	40	21	for	for	ADP
fcis-10302	40	22	the	the	DET
fcis-10302	40	23	detection	detection	NOUN
fcis-10302	40	24	of	of	ADP
fcis-10302	40	25	malicious	malicious	ADJ
fcis-10302	40	26	attacks	attack	NOUN
fcis-10302	40	27	on	on	ADP
fcis-10302	40	28	selected	select	VERB
fcis-10302	40	29	iot	iot	NOUN
fcis-10302	40	30	devices	device	NOUN
fcis-10302	40	31	.	.	PUNCT
fcis-10302	41	1	the	the	DET
fcis-10302	41	2	n	n	NUM
fcis-10302	41	3	-	-	PUNCT
fcis-10302	41	4	baiot	baiot	NOUN
fcis-10302	41	5	dataset	dataset	NOUN
fcis-10302	41	6	is	be	AUX
fcis-10302	41	7	used	use	VERB
fcis-10302	41	8	as	as	ADP
fcis-10302	41	9	the	the	DET
fcis-10302	41	10	base	base	NOUN
fcis-10302	41	11	dataset	dataset	NOUN
fcis-10302	41	12	.	.	PUNCT
fcis-10302	42	1	firstly	firstly	ADV
fcis-10302	42	2	,	,	PUNCT
fcis-10302	42	3	the	the	DET
fcis-10302	42	4	dataset	dataset	NOUN
fcis-10302	42	5	is	be	AUX
fcis-10302	42	6	normalized	normalize	VERB
fcis-10302	42	7	and	and	CCONJ
fcis-10302	42	8	one	one	NUM
fcis-10302	42	9	-	-	PUNCT
fcis-10302	42	10	hot	hot	ADJ
fcis-10302	42	11	encoded	encode	VERB
fcis-10302	42	12	.	.	PUNCT
fcis-10302	43	1	xgboost	xgboost	PROPN
fcis-10302	43	2	is	be	AUX
fcis-10302	43	3	then	then	ADV
fcis-10302	43	4	used	use	VERB
fcis-10302	43	5	for	for	ADP
fcis-10302	43	6	feature	feature	NOUN
fcis-10302	43	7	selection	selection	NOUN
fcis-10302	43	8	to	to	PART
fcis-10302	43	9	reduce	reduce	VERB
fcis-10302	43	10	data	data	NOUN
fcis-10302	43	11	redundancy	redundancy	NOUN
fcis-10302	43	12	.	.	PUNCT
fcis-10302	44	1	finally	finally	ADV
fcis-10302	44	2	,	,	PUNCT
fcis-10302	44	3	a	a	DET
fcis-10302	44	4	one	one	NUM
fcis-10302	44	5	-	-	PUNCT
fcis-10302	44	6	dimensional	dimensional	ADJ
fcis-10302	44	7	cnn	cnn	NOUN
fcis-10302	44	8	is	be	AUX
fcis-10302	44	9	used	use	VERB
fcis-10302	44	10	to	to	PART
fcis-10302	44	11	extract	extract	VERB
fcis-10302	44	12	spatial	spatial	ADJ
fcis-10302	44	13	features	feature	NOUN
fcis-10302	44	14	,	,	PUNCT
fcis-10302	44	15	and	and	CCONJ
fcis-10302	44	16	gru	gru	NOUN
fcis-10302	44	17	is	be	AUX
fcis-10302	44	18	employed	employ	VERB
fcis-10302	44	19	to	to	PART
fcis-10302	44	20	extract	extract	VERB
fcis-10302	44	21	longterm	longterm	ADJ
fcis-10302	44	22	dependency	dependency	NOUN
fcis-10302	44	23	information	information	NOUN
fcis-10302	44	24	features	feature	NOUN
fcis-10302	44	25	and	and	CCONJ
fcis-10302	44	26	learn	learn	VERB
fcis-10302	44	27	the	the	DET
fcis-10302	44	28	interdependencies	interdependency	NOUN
fcis-10302	44	29	between	between	ADP
fcis-10302	44	30	features	feature	NOUN
fcis-10302	44	31	.	.	PUNCT
fcis-10302	45	1	the	the	DET
fcis-10302	45	2	softmax	softmax	NOUN
fcis-10302	45	3	function	function	NOUN
fcis-10302	45	4	is	be	AUX
fcis-10302	45	5	used	use	VERB
fcis-10302	45	6	for	for	ADP
fcis-10302	45	7	classification	classification	NOUN
fcis-10302	45	8	.	.	PUNCT
fcis-10302	46	1	comparative	comparative	ADJ
fcis-10302	46	2	analysis	analysis	NOUN
fcis-10302	46	3	experiments	experiment	NOUN
fcis-10302	46	4	demonstrate	demonstrate	VERB
fcis-10302	46	5	the	the	DET
fcis-10302	46	6	applicability	applicability	NOUN
fcis-10302	46	7	of	of	ADP
fcis-10302	46	8	the	the	DET
fcis-10302	46	9	proposed	propose	VERB
fcis-10302	46	10	model	model	NOUN
fcis-10302	46	11	.	.	PUNCT
fcis-10302	47	1	the	the	DET
fcis-10302	47	2	results	result	NOUN
fcis-10302	47	3	show	show	VERB
fcis-10302	47	4	that	that	SCONJ
fcis-10302	47	5	this	this	DET
fcis-10302	47	6	method	method	NOUN
fcis-10302	47	7	outperforms	outperform	VERB
fcis-10302	47	8	recent	recent	ADJ
fcis-10302	47	9	iot	iot	PROPN
fcis-10302	47	10	intrusion	intrusion	NOUN
fcis-10302	47	11	detection	detection	NOUN
fcis-10302	47	12	methods	method	NOUN
fcis-10302	47	13	in	in	ADP
fcis-10302	47	14	terms	term	NOUN
fcis-10302	47	15	of	of	ADP
fcis-10302	47	16	accuracy	accuracy	NOUN
fcis-10302	47	17	,	,	PUNCT
fcis-10302	47	18	recall	recall	NOUN
fcis-10302	47	19	,	,	PUNCT
fcis-10302	47	20	and	and	CCONJ
fcis-10302	47	21	loss	loss	NOUN
fcis-10302	47	22	based	base	VERB
fcis-10302	47	23	on	on	ADP
fcis-10302	47	24	the	the	DET
fcis-10302	47	25	aforementioned	aforementioned	ADJ
fcis-10302	47	26	dataset	dataset	NOUN
fcis-10302	47	27	,	,	PUNCT
fcis-10302	47	28	effectively	effectively	ADV
fcis-10302	47	29	improving	improve	VERB
fcis-10302	47	30	the	the	DET
fcis-10302	47	31	accuracy	accuracy	NOUN
fcis-10302	47	32	of	of	ADP
fcis-10302	47	33	iot	iot	PROPN
fcis-10302	47	34	intrusion	intrusion	NOUN
fcis-10302	47	35	detection	detection	NOUN
fcis-10302	47	36	.	.	PUNCT
fcis-10302	48	1	91	91	NUM
fcis-10302	48	2	2	2	NUM
fcis-10302	48	3	.	.	PUNCT
fcis-10302	48	4	related	relate	VERB
fcis-10302	48	5	work	work	NOUN
fcis-10302	48	6	2.1	2.1	NUM
fcis-10302	48	7	.	.	PUNCT
fcis-10302	49	1	xgboost	xgboost	X
fcis-10302	50	1	xgboost	xgboost	X
fcis-10302	50	2	is	be	AUX
fcis-10302	50	3	a	a	DET
fcis-10302	50	4	boosting	boost	VERB
fcis-10302	50	5	-	-	PUNCT
fcis-10302	50	6	based	base	VERB
fcis-10302	50	7	ensemble	ensemble	ADJ
fcis-10302	50	8	method	method	NOUN
fcis-10302	50	9	that	that	PRON
fcis-10302	50	10	utilizes	utilize	VERB
fcis-10302	50	11	decision	decision	NOUN
fcis-10302	50	12	trees	tree	NOUN
fcis-10302	50	13	.	.	PUNCT
fcis-10302	51	1	the	the	DET
fcis-10302	51	2	algorithm	algorithm	NOUN
fcis-10302	51	3	follows	follow	VERB
fcis-10302	51	4	the	the	DET
fcis-10302	51	5	principle	principle	NOUN
fcis-10302	51	6	of	of	ADP
fcis-10302	51	7	optimizing	optimize	VERB
fcis-10302	51	8	the	the	DET
fcis-10302	51	9	empirical	empirical	ADJ
fcis-10302	51	10	loss	loss	NOUN
fcis-10302	51	11	function	function	NOUN
fcis-10302	51	12	by	by	ADP
fcis-10302	51	13	iteratively	iteratively	ADV
fcis-10302	51	14	fitting	fit	VERB
fcis-10302	51	15	the	the	DET
fcis-10302	51	16	negative	negative	ADJ
fcis-10302	51	17	gradient	gradient	NOUN
fcis-10302	51	18	of	of	ADP
fcis-10302	51	19	the	the	DET
fcis-10302	51	20	loss	loss	NOUN
fcis-10302	51	21	function	function	NOUN
fcis-10302	51	22	.	.	PUNCT
fcis-10302	52	1	it	it	PRON
fcis-10302	52	2	then	then	ADV
fcis-10302	52	3	generates	generate	VERB
fcis-10302	52	4	the	the	DET
fcis-10302	52	5	optimal	optimal	ADJ
fcis-10302	52	6	weak	weak	ADJ
fcis-10302	52	7	learner	learner	NOUN
fcis-10302	52	8	using	use	VERB
fcis-10302	52	9	a	a	DET
fcis-10302	52	10	linear	linear	ADJ
fcis-10302	52	11	search	search	NOUN
fcis-10302	52	12	method	method	NOUN
fcis-10302	52	13	.	.	PUNCT
fcis-10302	53	1	the	the	DET
fcis-10302	53	2	objective	objective	ADJ
fcis-10302	53	3	function	function	NOUN
fcis-10302	53	4	for	for	ADP
fcis-10302	53	5	the	the	DET
fcis-10302	53	6	t	t	PROPN
fcis-10302	53	7	-	-	PUNCT
fcis-10302	53	8	th	th	VERB
fcis-10302	53	9	tree	tree	NOUN
fcis-10302	53	10	is	be	AUX
fcis-10302	53	11	as	as	SCONJ
fcis-10302	53	12	follows	follow	VERB
fcis-10302	53	13	:	:	PUNCT
fcis-10302	53	14			ADJ
fcis-10302	53	15	(	(	PUNCT
fcis-10302	53	16	)	)	PUNCT
fcis-10302	53	17	1	1	NUM
fcis-10302	53	18	1	1	NUM
fcis-10302	53	19	obj	obj	PROPN
fcis-10302	53	20	(	(	PUNCT
fcis-10302	53	21	,	,	PUNCT
fcis-10302	53	22	)	)	PUNCT
fcis-10302	53	23	(	(	PUNCT
fcis-10302	53	24	)	)	PUNCT
fcis-10302	54	1	n	n	CCONJ
fcis-10302	54	2	k	k	PROPN
fcis-10302	54	3	t	t	PROPN
fcis-10302	55	1	t	t	X
fcis-10302	56	1	i	i	PRON
fcis-10302	56	2	ti	ti	VERB
fcis-10302	57	1	i	i	NOUN
fcis-10302	57	2	k	k	PROPN
fcis-10302	58	1	l	l	NOUN
fcis-10302	58	2	y	y	PROPN
fcis-10302	58	3	y	y	PROPN
fcis-10302	59	1	f	f	PROPN
fcis-10302	60	1			NUM
fcis-10302	61	1			NUM
fcis-10302	61	2			PROPN
fcis-10302	61	3			PUNCT
fcis-10302	61	4			PUNCT
fcis-10302	61	5			X
fcis-10302	61	6	(	(	PUNCT
fcis-10302	61	7	1	1	X
fcis-10302	61	8	)	)	PUNCT
fcis-10302	61	9	where	where	SCONJ
fcis-10302	61	10			PROPN
fcis-10302	61	11	(	(	PUNCT
fcis-10302	61	12	,	,	PUNCT
fcis-10302	61	13	)	)	PUNCT
fcis-10302	61	14	t	t	NOUN
fcis-10302	62	1	i	i	PRON
fcis-10302	62	2	il	il	PROPN
fcis-10302	62	3	y	y	PROPN
fcis-10302	62	4	y	y	PROPN
fcis-10302	62	5	is	be	AUX
fcis-10302	62	6	the	the	DET
fcis-10302	62	7	loss	loss	NOUN
fcis-10302	62	8	function	function	NOUN
fcis-10302	62	9	used	use	VERB
fcis-10302	62	10	to	to	PART
fcis-10302	62	11	evaluate	evaluate	VERB
fcis-10302	62	12	model	model	NOUN
fcis-10302	62	13	error	error	NOUN
fcis-10302	62	14	and	and	CCONJ
fcis-10302	62	15	(	(	PUNCT
fcis-10302	62	16	)	)	PUNCT
fcis-10302	62	17	tf	tf	PROPN
fcis-10302	62	18	is	be	AUX
fcis-10302	62	19	the	the	DET
fcis-10302	62	20	regularization	regularization	NOUN
fcis-10302	62	21	term	term	NOUN
fcis-10302	62	22	used	use	VERB
fcis-10302	62	23	to	to	PART
fcis-10302	62	24	control	control	VERB
fcis-10302	62	25	the	the	DET
fcis-10302	62	26	complexity	complexity	NOUN
fcis-10302	62	27	of	of	ADP
fcis-10302	62	28	the	the	DET
fcis-10302	62	29	tree	tree	NOUN
fcis-10302	62	30	and	and	CCONJ
fcis-10302	62	31	prevent	prevent	VERB
fcis-10302	62	32	overfitting	overfitting	NOUN
fcis-10302	62	33	.	.	PUNCT
fcis-10302	63	1	to	to	PART
fcis-10302	63	2	find	find	VERB
fcis-10302	63	3	the	the	DET
fcis-10302	63	4	minimum	minimum	NOUN
fcis-10302	63	5	of	of	ADP
fcis-10302	63	6	the	the	DET
fcis-10302	63	7	objective	objective	ADJ
fcis-10302	63	8	function	function	NOUN
fcis-10302	63	9	,	,	PUNCT
fcis-10302	63	10	the	the	DET
fcis-10302	63	11	objective	objective	ADJ
fcis-10302	63	12	function	function	NOUN
fcis-10302	63	13	is	be	AUX
fcis-10302	63	14	expanded	expand	VERB
fcis-10302	63	15	using	use	VERB
fcis-10302	63	16	the	the	DET
fcis-10302	63	17	second	second	ADJ
fcis-10302	63	18	-	-	PUNCT
fcis-10302	63	19	order	order	NOUN
fcis-10302	63	20	taylor	taylor	PROPN
fcis-10302	63	21	series	series	PROPN
fcis-10302	63	22	approximation	approximation	PROPN
fcis-10302	63	23	,	,	PUNCT
fcis-10302	63	24	and	and	CCONJ
fcis-10302	63	25	the	the	DET
fcis-10302	63	26	optimal	optimal	ADJ
fcis-10302	63	27	weights	weight	NOUN
fcis-10302	63	28			PROPN
fcis-10302	63	29	and	and	CCONJ
fcis-10302	63	30	the	the	DET
fcis-10302	63	31	corresponding	corresponding	ADJ
fcis-10302	63	32	objective	objective	ADJ
fcis-10302	63	33	function	function	NOUN
fcis-10302	63	34	value	value	NOUN
fcis-10302	63	35	can	can	AUX
fcis-10302	63	36	be	be	AUX
fcis-10302	63	37	obtained	obtain	VERB
fcis-10302	63	38	:	:	PUNCT
fcis-10302	63	39	*	*	PUNCT
fcis-10302	64	1	j	j	PROPN
fcis-10302	64	2	j	j	PROPN
fcis-10302	64	3	j	j	PROPN
fcis-10302	64	4	g	g	PROPN
fcis-10302	64	5	h	h	PROPN
fcis-10302	64	6			PROPN
fcis-10302	64	7			X
fcis-10302	64	8			PROPN
fcis-10302	64	9			PROPN
fcis-10302	64	10			X
fcis-10302	64	11	(	(	PUNCT
fcis-10302	64	12	2	2	NUM
fcis-10302	64	13	)	)	PUNCT
fcis-10302	64	14	1	1	NUM
fcis-10302	64	15	1	1	NUM
fcis-10302	64	16	2	2	NUM
fcis-10302	64	17	t	t	NOUN
fcis-10302	64	18	j	j	PROPN
fcis-10302	64	19	obj	obj	PROPN
fcis-10302	64	20	j	j	PROPN
fcis-10302	65	1	j	j	PROPN
fcis-10302	65	2	g	g	PROPN
fcis-10302	65	3	x	x	PROPN
fcis-10302	65	4	t	t	PROPN
fcis-10302	65	5	h	h	PROPN
fcis-10302	65	6			ADJ
fcis-10302	65	7			PROPN
fcis-10302	65	8			PROPN
fcis-10302	65	9			PROPN
fcis-10302	65	10			SYM
fcis-10302	65	11			NOUN
fcis-10302	65	12	(	(	PUNCT
fcis-10302	65	13	3	3	NUM
fcis-10302	65	14	)	)	PUNCT
fcis-10302	66	1	where	where	SCONJ
fcis-10302	66	2	j	j	PROPN
fcis-10302	66	3	j	j	PROPN
fcis-10302	66	4	i	i	PRON
fcis-10302	66	5	i	i	PRON
fcis-10302	67	1	i	i	PRON
fcis-10302	67	2	g	g	PROPN
fcis-10302	67	3	g	g	PROPN
fcis-10302	67	4			PROPN
fcis-10302	67	5			PROPN
fcis-10302	67	6	represents	represent	VERB
fcis-10302	67	7	the	the	DET
fcis-10302	67	8	sum	sum	NOUN
fcis-10302	67	9	of	of	ADP
fcis-10302	67	10	the	the	DET
fcis-10302	67	11	first	first	ADJ
fcis-10302	67	12	-	-	PUNCT
fcis-10302	67	13	order	order	NOUN
fcis-10302	67	14	gradients	gradient	NOUN
fcis-10302	67	15	of	of	ADP
fcis-10302	67	16	the	the	DET
fcis-10302	67	17	loss	loss	NOUN
fcis-10302	67	18	function	function	NOUN
fcis-10302	67	19	for	for	ADP
fcis-10302	67	20	all	all	DET
fcis-10302	67	21	samples	sample	NOUN
fcis-10302	67	22	in	in	ADP
fcis-10302	67	23	the	the	DET
fcis-10302	67	24	j	j	PROPN
fcis-10302	67	25	-	-	PUNCT
fcis-10302	67	26	th	th	X
fcis-10302	67	27	node	node	NOUN
fcis-10302	67	28	,	,	PUNCT
fcis-10302	67	29	and	and	CCONJ
fcis-10302	68	1	j	j	PROPN
fcis-10302	68	2	j	j	PROPN
fcis-10302	68	3	j	j	PROPN
fcis-10302	69	1	i	i	PRON
fcis-10302	70	1	i	i	PRON
fcis-10302	70	2	h	h	NOUN
fcis-10302	70	3	h	h	NOUN
fcis-10302	70	4			NOUN
fcis-10302	70	5			PROPN
fcis-10302	70	6	represents	represent	VERB
fcis-10302	70	7	the	the	DET
fcis-10302	70	8	sum	sum	NOUN
fcis-10302	70	9	of	of	ADP
fcis-10302	70	10	the	the	DET
fcis-10302	70	11	second	second	ADJ
fcis-10302	70	12	-	-	PUNCT
fcis-10302	70	13	order	order	NOUN
fcis-10302	70	14	gradients	gradient	NOUN
fcis-10302	70	15	of	of	ADP
fcis-10302	70	16	the	the	DET
fcis-10302	70	17	loss	loss	NOUN
fcis-10302	70	18	function	function	NOUN
fcis-10302	70	19	for	for	ADP
fcis-10302	70	20	all	all	DET
fcis-10302	70	21	samples	sample	NOUN
fcis-10302	70	22	in	in	ADP
fcis-10302	70	23	the	the	DET
fcis-10302	70	24	j	j	PROPN
fcis-10302	70	25	-	-	PUNCT
fcis-10302	70	26	th	th	VERB
fcis-10302	70	27	node	node	NOUN
fcis-10302	70	28	.	.	PUNCT
fcis-10302	71	1	2.2	2.2	NUM
fcis-10302	71	2	.	.	PUNCT
fcis-10302	71	3	cnn	cnn	PROPN
fcis-10302	71	4	in	in	ADP
fcis-10302	71	5	the	the	DET
fcis-10302	71	6	proposed	propose	VERB
fcis-10302	71	7	model	model	NOUN
fcis-10302	71	8	architecture	architecture	NOUN
fcis-10302	71	9	,	,	PUNCT
fcis-10302	71	10	we	we	PRON
fcis-10302	71	11	have	have	AUX
fcis-10302	71	12	introduced	introduce	VERB
fcis-10302	71	13	a	a	DET
fcis-10302	71	14	one	one	NUM
fcis-10302	71	15	-	-	PUNCT
fcis-10302	71	16	dimensional	dimensional	ADJ
fcis-10302	71	17	convolutional	convolutional	ADJ
fcis-10302	71	18	neural	neural	ADJ
fcis-10302	71	19	network	network	NOUN
fcis-10302	71	20	(	(	PUNCT
fcis-10302	71	21	1dcnn	1dcnn	NUM
fcis-10302	71	22	)	)	PUNCT
fcis-10302	71	23	component	component	NOUN
fcis-10302	71	24	for	for	ADP
fcis-10302	71	25	the	the	DET
fcis-10302	71	26	identification	identification	NOUN
fcis-10302	71	27	of	of	ADP
fcis-10302	71	28	malicious	malicious	ADJ
fcis-10302	71	29	iot	iot	NOUN
fcis-10302	71	30	traffic	traffic	NOUN
fcis-10302	71	31	.	.	PUNCT
fcis-10302	72	1	1dcnn	1dcnn	NUM
fcis-10302	72	2	is	be	AUX
fcis-10302	72	3	capable	capable	ADJ
fcis-10302	72	4	of	of	ADP
fcis-10302	72	5	extracting	extract	VERB
fcis-10302	72	6	morphological	morphological	ADJ
fcis-10302	72	7	features	feature	NOUN
fcis-10302	72	8	from	from	ADP
fcis-10302	72	9	traffic	traffic	NOUN
fcis-10302	72	10	data	datum	NOUN
fcis-10302	72	11	to	to	PART
fcis-10302	72	12	enhance	enhance	VERB
fcis-10302	72	13	the	the	DET
fcis-10302	72	14	understanding	understanding	NOUN
fcis-10302	72	15	of	of	ADP
fcis-10302	72	16	the	the	DET
fcis-10302	72	17	traffic	traffic	NOUN
fcis-10302	72	18	.	.	PUNCT
fcis-10302	73	1	a	a	DET
fcis-10302	73	2	typical	typical	ADJ
fcis-10302	73	3	1dcnn	1dcnn	NUM
fcis-10302	73	4	architecture	architecture	NOUN
fcis-10302	73	5	consists	consist	VERB
fcis-10302	73	6	of	of	ADP
fcis-10302	73	7	convolutional	convolutional	ADJ
fcis-10302	73	8	layers	layer	NOUN
fcis-10302	73	9	,	,	PUNCT
fcis-10302	73	10	pooling	pool	VERB
fcis-10302	73	11	layers	layer	NOUN
fcis-10302	73	12	,	,	PUNCT
fcis-10302	73	13	and	and	CCONJ
fcis-10302	73	14	fully	fully	ADV
fcis-10302	73	15	connected	connected	ADJ
fcis-10302	73	16	layers	layer	NOUN
fcis-10302	73	17	.	.	PUNCT
fcis-10302	74	1	(	(	PUNCT
fcis-10302	74	2	1	1	X
fcis-10302	74	3	)	)	PUNCT
fcis-10302	74	4	convolutional	convolutional	ADJ
fcis-10302	74	5	layers	layer	NOUN
fcis-10302	74	6	:	:	PUNCT
fcis-10302	74	7	the	the	DET
fcis-10302	74	8	convolutional	convolutional	ADJ
fcis-10302	74	9	layers	layer	NOUN
fcis-10302	74	10	in	in	ADP
fcis-10302	74	11	1dcnn	1dcnn	NUM
fcis-10302	74	12	overcome	overcome	VERB
fcis-10302	74	13	the	the	DET
fcis-10302	74	14	slow	slow	ADJ
fcis-10302	74	15	convergence	convergence	NOUN
fcis-10302	74	16	of	of	ADP
fcis-10302	74	17	conventional	conventional	ADJ
fcis-10302	74	18	neural	neural	ADJ
fcis-10302	74	19	networks	network	NOUN
fcis-10302	74	20	.	.	PUNCT
fcis-10302	75	1	this	this	DET
fcis-10302	75	2	layer	layer	NOUN
fcis-10302	75	3	consists	consist	VERB
fcis-10302	75	4	of	of	ADP
fcis-10302	75	5	a	a	DET
fcis-10302	75	6	set	set	NOUN
fcis-10302	75	7	of	of	ADP
fcis-10302	75	8	time	time	NOUN
fcis-10302	75	9	series	series	PROPN
fcis-10302	75	10	graphs	graph	NOUN
fcis-10302	75	11	,	,	PUNCT
fcis-10302	75	12	kernels	kernel	NOUN
fcis-10302	75	13	,	,	PUNCT
fcis-10302	75	14	filters	filter	NOUN
fcis-10302	75	15	,	,	PUNCT
fcis-10302	75	16	strides	stride	NOUN
fcis-10302	75	17	,	,	PUNCT
fcis-10302	75	18	and	and	CCONJ
fcis-10302	75	19	neurons	neuron	NOUN
fcis-10302	75	20	.	.	PUNCT
fcis-10302	76	1	each	each	DET
fcis-10302	76	2	neuron	neuron	NOUN
fcis-10302	76	3	in	in	ADP
fcis-10302	76	4	this	this	DET
fcis-10302	76	5	layer	layer	NOUN
fcis-10302	76	6	is	be	AUX
fcis-10302	76	7	connected	connect	VERB
fcis-10302	76	8	to	to	ADP
fcis-10302	76	9	adjacent	adjacent	ADJ
fcis-10302	76	10	neurons	neuron	NOUN
fcis-10302	76	11	.	.	PUNCT
fcis-10302	77	1	the	the	DET
fcis-10302	77	2	convolution	convolution	NOUN
fcis-10302	77	3	operation	operation	NOUN
fcis-10302	77	4	computes	compute	VERB
fcis-10302	77	5	the	the	DET
fcis-10302	77	6	dot	dot	NOUN
fcis-10302	77	7	product	product	NOUN
fcis-10302	77	8	between	between	ADP
fcis-10302	77	9	the	the	DET
fcis-10302	77	10	corresponding	corresponding	ADJ
fcis-10302	77	11	convolution	convolution	NOUN
fcis-10302	77	12	filter	filter	NOUN
fcis-10302	77	13	and	and	CCONJ
fcis-10302	77	14	the	the	DET
fcis-10302	77	15	input	input	NOUN
fcis-10302	77	16	time	time	NOUN
fcis-10302	77	17	series	series	PROPN
fcis-10302	77	18	graph	graph	NOUN
fcis-10302	77	19	.	.	PUNCT
fcis-10302	78	1	through	through	ADP
fcis-10302	78	2	the	the	DET
fcis-10302	78	3	kernels	kernel	NOUN
fcis-10302	78	4	,	,	PUNCT
fcis-10302	78	5	it	it	PRON
fcis-10302	78	6	learns	learn	VERB
fcis-10302	78	7	the	the	DET
fcis-10302	78	8	features	feature	NOUN
fcis-10302	78	9	of	of	ADP
fcis-10302	78	10	the	the	DET
fcis-10302	78	11	input	input	NOUN
fcis-10302	78	12	mapping	mapping	NOUN
fcis-10302	78	13	.	.	PUNCT
fcis-10302	79	1	the	the	DET
fcis-10302	79	2	time	time	NOUN
fcis-10302	79	3	series	series	PROPN
fcis-10302	79	4	input	input	NOUN
fcis-10302	79	5	is	be	AUX
fcis-10302	79	6	first	first	ADV
fcis-10302	79	7	fed	feed	VERB
fcis-10302	79	8	into	into	ADP
fcis-10302	79	9	the	the	DET
fcis-10302	79	10	input	input	NOUN
fcis-10302	79	11	layer	layer	NOUN
fcis-10302	79	12	.	.	PUNCT
fcis-10302	80	1	then	then	ADV
fcis-10302	80	2	,	,	PUNCT
fcis-10302	80	3	the	the	DET
fcis-10302	80	4	output	output	NOUN
fcis-10302	80	5	of	of	ADP
fcis-10302	80	6	the	the	DET
fcis-10302	80	7	convolution	convolution	NOUN
fcis-10302	80	8	process	process	NOUN
fcis-10302	80	9	is	be	AUX
fcis-10302	80	10	computed	compute	VERB
fcis-10302	80	11	as	as	SCONJ
fcis-10302	80	12	follows	follow	VERB
fcis-10302	80	13	:	:	PUNCT
fcis-10302	80	14			PROPN
fcis-10302	80	15			PROPN
fcis-10302	81	1	m	m	VERB
fcis-10302	81	2	l-1l	l-1l	NOUN
fcis-10302	81	3	l-1l	l-1l	VERB
fcis-10302	81	4	immm	immm	ADV
fcis-10302	82	1	i	i	PRON
fcis-10302	83	1	i	i	VERB
fcis-10302	83	2	l	l	NOUN
fcis-10302	83	3	y	y	PROPN
fcis-10302	83	4	b	b	PROPN
fcis-10302	83	5	conv1d	conv1d	PROPN
fcis-10302	83	6	y	y	PROPN
fcis-10302	83	7			PROPN
fcis-10302	83	8			NUM
fcis-10302	83	9			ADV
fcis-10302	83	10			X
fcis-10302	83	11	，	，	PUNCT
fcis-10302	83	12	(	(	PUNCT
fcis-10302	83	13	4	4	X
fcis-10302	83	14	)	)	PUNCT
fcis-10302	83	15	where	where	SCONJ
fcis-10302	83	16	l	l	NOUN
fcis-10302	83	17	my	my	PRON
fcis-10302	83	18	and	and	CCONJ
fcis-10302	83	19	l	l	NOUN
fcis-10302	83	20	mb	mb	PROPN
fcis-10302	83	21	represent	represent	VERB
fcis-10302	83	22	the	the	DET
fcis-10302	83	23	output	output	NOUN
fcis-10302	83	24	and	and	CCONJ
fcis-10302	83	25	bias	bias	NOUN
fcis-10302	83	26	of	of	ADP
fcis-10302	83	27	the	the	DET
fcis-10302	83	28	mth	mth	NOUN
fcis-10302	83	29	neuron	neuron	NOUN
fcis-10302	83	30	in	in	ADP
fcis-10302	83	31	the	the	DET
fcis-10302	83	32	lth	lth	NOUN
fcis-10302	83	33	layer	layer	NOUN
fcis-10302	83	34	,	,	PUNCT
fcis-10302	83	35	respectively	respectively	ADV
fcis-10302	83	36	.	.	PUNCT
fcis-10302	84	1	l-1	l-1	PROPN
fcis-10302	84	2	imw	imw	PROPN
fcis-10302	84	3	and	and	CCONJ
fcis-10302	84	4			PROPN
fcis-10302	84	5	l-1	l-1	PROPN
fcis-10302	84	6	iy	iy	PROPN
fcis-10302	84	7	represent	represent	VERB
fcis-10302	84	8	the	the	DET
fcis-10302	84	9	kernel	kernel	NOUN
fcis-10302	84	10	weights	weight	VERB
fcis-10302	84	11	from	from	ADP
fcis-10302	84	12	the	the	DET
fcis-10302	84	13	ith	ith	PROPN
fcis-10302	84	14	neuron	neuron	NOUN
fcis-10302	84	15	in	in	ADP
fcis-10302	84	16	the	the	DET
fcis-10302	84	17	(	(	PUNCT
fcis-10302	84	18	l-1)th	l-1)th	NOUN
fcis-10302	84	19	layer	layer	NOUN
fcis-10302	84	20	to	to	ADP
fcis-10302	84	21	the	the	DET
fcis-10302	84	22	m	m	PROPN
fcis-10302	84	23	-	-	PUNCT
fcis-10302	84	24	th	th	VERB
fcis-10302	84	25	neuron	neuron	NOUN
fcis-10302	84	26	in	in	ADP
fcis-10302	84	27	the	the	DET
fcis-10302	84	28	lth	lth	NOUN
fcis-10302	84	29	layer	layer	NOUN
fcis-10302	84	30	and	and	CCONJ
fcis-10302	84	31	the	the	DET
fcis-10302	84	32	input	input	NOUN
fcis-10302	84	33	to	to	ADP
fcis-10302	84	34	the	the	DET
fcis-10302	84	35	mth	mth	NOUN
fcis-10302	84	36	neuron	neuron	NOUN
fcis-10302	84	37	in	in	ADP
fcis-10302	84	38	the	the	DET
fcis-10302	84	39	lth	lth	NOUN
fcis-10302	84	40	layer	layer	NOUN
fcis-10302	84	41	from	from	ADP
fcis-10302	84	42	the	the	DET
fcis-10302	84	43	ith	ith	PROPN
fcis-10302	84	44	neuron	neuron	NOUN
fcis-10302	84	45	in	in	ADP
fcis-10302	84	46	the	the	DET
fcis-10302	84	47	(	(	PUNCT
fcis-10302	84	48	l-1)th	l-1)th	NOUN
fcis-10302	84	49	layer	layer	NOUN
fcis-10302	84	50	,	,	PUNCT
fcis-10302	84	51	respectively	respectively	ADV
fcis-10302	84	52	.	.	PUNCT
fcis-10302	85	1	ml	ml	X
fcis-10302	85	2	is	be	AUX
fcis-10302	85	3	the	the	DET
fcis-10302	85	4	activation	activation	NOUN
fcis-10302	85	5	function	function	NOUN
fcis-10302	85	6	applied	apply	VERB
fcis-10302	85	7	to	to	ADP
fcis-10302	85	8	the	the	DET
fcis-10302	85	9	input	input	NOUN
fcis-10302	85	10	mapping	mapping	NOUN
fcis-10302	85	11	.	.	PUNCT
fcis-10302	86	1	(	(	PUNCT
fcis-10302	86	2	2	2	X
fcis-10302	86	3	)	)	PUNCT
fcis-10302	86	4	pooling	pool	VERB
fcis-10302	86	5	layers	layer	NOUN
fcis-10302	86	6	:	:	PUNCT
fcis-10302	86	7	the	the	DET
fcis-10302	86	8	pooling	pool	VERB
fcis-10302	86	9	operation	operation	NOUN
fcis-10302	86	10	in	in	ADP
fcis-10302	86	11	a	a	DET
fcis-10302	86	12	1dcnn	1dcnn	NUM
fcis-10302	86	13	model	model	NOUN
fcis-10302	86	14	is	be	AUX
fcis-10302	86	15	a	a	DET
fcis-10302	86	16	resampling	resample	VERB
fcis-10302	86	17	process	process	NOUN
fcis-10302	86	18	that	that	PRON
fcis-10302	86	19	converts	convert	VERB
fcis-10302	86	20	multiple	multiple	ADJ
fcis-10302	86	21	elements	element	NOUN
fcis-10302	86	22	into	into	ADP
fcis-10302	86	23	a	a	DET
fcis-10302	86	24	single	single	ADJ
fcis-10302	86	25	element	element	NOUN
fcis-10302	86	26	.	.	PUNCT
fcis-10302	87	1	this	this	DET
fcis-10302	87	2	operation	operation	NOUN
fcis-10302	87	3	minimizes	minimize	VERB
fcis-10302	87	4	computational	computational	ADJ
fcis-10302	87	5	costs	cost	NOUN
fcis-10302	87	6	,	,	PUNCT
fcis-10302	87	7	retains	retain	VERB
fcis-10302	87	8	important	important	ADJ
fcis-10302	87	9	information	information	NOUN
fcis-10302	87	10	,	,	PUNCT
fcis-10302	87	11	and	and	CCONJ
fcis-10302	87	12	helps	help	VERB
fcis-10302	87	13	prevent	prevent	VERB
fcis-10302	87	14	overfitting	overfitting	NOUN
fcis-10302	87	15	.	.	PUNCT
fcis-10302	88	1	in	in	ADP
fcis-10302	88	2	this	this	DET
fcis-10302	88	3	study	study	NOUN
fcis-10302	88	4	,	,	PUNCT
fcis-10302	88	5	we	we	PRON
fcis-10302	88	6	utilize	utilize	VERB
fcis-10302	88	7	the	the	DET
fcis-10302	88	8	max	max	PROPN
fcis-10302	88	9	pooling	pooling	NOUN
fcis-10302	88	10	model	model	NOUN
fcis-10302	88	11	,	,	PUNCT
fcis-10302	88	12	which	which	PRON
fcis-10302	88	13	selects	select	VERB
fcis-10302	88	14	the	the	DET
fcis-10302	88	15	maximum	maximum	ADJ
fcis-10302	88	16	value	value	NOUN
fcis-10302	88	17	in	in	ADP
fcis-10302	88	18	an	an	DET
fcis-10302	88	19	array	array	NOUN
fcis-10302	88	20	as	as	ADP
fcis-10302	88	21	the	the	DET
fcis-10302	88	22	output	output	NOUN
fcis-10302	88	23	.	.	PUNCT
fcis-10302	89	1	the	the	DET
fcis-10302	89	2	formula	formula	NOUN
fcis-10302	89	3	for	for	ADP
fcis-10302	89	4	reducing	reduce	VERB
fcis-10302	89	5	the	the	DET
fcis-10302	89	6	size	size	NOUN
fcis-10302	89	7	of	of	ADP
fcis-10302	89	8	time	time	NOUN
fcis-10302	89	9	series	series	NOUN
fcis-10302	89	10	graphs	graph	NOUN
fcis-10302	89	11	while	while	SCONJ
fcis-10302	89	12	maintaining	maintain	VERB
fcis-10302	89	13	discriminative	discriminative	NOUN
fcis-10302	89	14	information	information	NOUN
fcis-10302	89	15	is	be	AUX
fcis-10302	89	16	as	as	SCONJ
fcis-10302	89	17	follows	follow	VERB
fcis-10302	89	18	:	:	PUNCT
fcis-10302	89	19			NOUN
fcis-10302	89	20	p	p	VERB
fcis-10302	89	21	max	max	PROPN
fcis-10302	89	22	yr	yr	PROPN
fcis-10302	89	23	(	(	PUNCT
fcis-10302	89	24	5	5	NUM
fcis-10302	89	25	)	)	PUNCT
fcis-10302	89	26	where	where	SCONJ
fcis-10302	89	27	p	p	NOUN
fcis-10302	89	28	represents	represent	VERB
fcis-10302	89	29	the	the	DET
fcis-10302	89	30	output	output	NOUN
fcis-10302	89	31	of	of	ADP
fcis-10302	89	32	the	the	DET
fcis-10302	89	33	max	max	PROPN
fcis-10302	89	34	pooling	pool	VERB
fcis-10302	89	35	layer	layer	NOUN
fcis-10302	89	36	,	,	PUNCT
fcis-10302	89	37	and	and	CCONJ
fcis-10302	89	38	yr	yr	NOUN
fcis-10302	89	39	represents	represent	VERB
fcis-10302	89	40	the	the	DET
fcis-10302	89	41	corresponding	corresponding	ADJ
fcis-10302	89	42	set	set	NOUN
fcis-10302	89	43	of	of	ADP
fcis-10302	89	44	elements	element	NOUN
fcis-10302	89	45	within	within	ADP
fcis-10302	89	46	the	the	DET
fcis-10302	89	47	pooling	pool	VERB
fcis-10302	89	48	region	region	NOUN
fcis-10302	89	49	.	.	PUNCT
fcis-10302	90	1	(	(	PUNCT
fcis-10302	90	2	3	3	X
fcis-10302	90	3	)	)	PUNCT
fcis-10302	90	4	fully	fully	ADV
fcis-10302	90	5	connected	connected	ADJ
fcis-10302	90	6	layers	layer	NOUN
fcis-10302	90	7	:	:	PUNCT
fcis-10302	90	8	following	follow	VERB
fcis-10302	90	9	the	the	DET
fcis-10302	90	10	pooling	pool	VERB
fcis-10302	90	11	layer	layer	NOUN
fcis-10302	90	12	,	,	PUNCT
fcis-10302	90	13	the	the	DET
fcis-10302	90	14	input	input	NOUN
fcis-10302	90	15	is	be	AUX
fcis-10302	90	16	flattened	flatten	VERB
fcis-10302	90	17	into	into	ADP
fcis-10302	90	18	a	a	DET
fcis-10302	90	19	one	one	NUM
fcis-10302	90	20	-	-	PUNCT
fcis-10302	90	21	dimensional	dimensional	ADJ
fcis-10302	90	22	array	array	NOUN
fcis-10302	90	23	and	and	CCONJ
fcis-10302	90	24	connected	connect	VERB
fcis-10302	90	25	to	to	ADP
fcis-10302	90	26	a	a	DET
fcis-10302	90	27	layer	layer	NOUN
fcis-10302	90	28	where	where	SCONJ
fcis-10302	90	29	each	each	DET
fcis-10302	90	30	input	input	NOUN
fcis-10302	90	31	is	be	AUX
fcis-10302	90	32	connected	connect	VERB
fcis-10302	90	33	to	to	ADP
fcis-10302	90	34	the	the	DET
fcis-10302	90	35	output	output	NOUN
fcis-10302	90	36	with	with	ADP
fcis-10302	90	37	weights	weight	NOUN
fcis-10302	90	38	.	.	PUNCT
fcis-10302	91	1	the	the	DET
fcis-10302	91	2	structure	structure	NOUN
fcis-10302	91	3	of	of	ADP
fcis-10302	91	4	this	this	DET
fcis-10302	91	5	layer	layer	NOUN
fcis-10302	91	6	is	be	AUX
fcis-10302	91	7	similar	similar	ADJ
fcis-10302	91	8	to	to	ADP
fcis-10302	91	9	a	a	DET
fcis-10302	91	10	conventional	conventional	ADJ
fcis-10302	91	11	artificial	artificial	ADJ
fcis-10302	91	12	neural	neural	ADJ
fcis-10302	91	13	network	network	NOUN
fcis-10302	91	14	,	,	PUNCT
fcis-10302	91	15	and	and	CCONJ
fcis-10302	91	16	the	the	DET
fcis-10302	91	17	neurons	neuron	NOUN
fcis-10302	91	18	in	in	ADP
fcis-10302	91	19	this	this	DET
fcis-10302	91	20	layer	layer	NOUN
fcis-10302	91	21	can	can	AUX
fcis-10302	91	22	be	be	AUX
fcis-10302	91	23	computed	compute	VERB
fcis-10302	91	24	as	as	SCONJ
fcis-10302	91	25	follows	follow	VERB
fcis-10302	91	26	:	:	PUNCT
fcis-10302	92	1	k	k	X
fcis-10302	92	2	,	,	PUNCT
fcis-10302	92	3	1	1	NUM
fcis-10302	92	4	u	u	NOUN
fcis-10302	92	5	m	m	VERB
fcis-10302	92	6	i	i	INTJ
fcis-10302	92	7	k	k	INTJ
fcis-10302	93	1	i	i	PRON
fcis-10302	93	2	k	k	VERB
fcis-10302	94	1	i	i	INTJ
fcis-10302	94	2	w	w	VERB
fcis-10302	94	3	p	p	PROPN
fcis-10302	94	4	b	b	PROPN
fcis-10302	94	5			PROPN
fcis-10302	94	6			PROPN
fcis-10302	94	7			NOUN
fcis-10302	94	8	(	(	PUNCT
fcis-10302	94	9	6	6	NUM
fcis-10302	94	10	)	)	PUNCT
fcis-10302	94	11	where	where	SCONJ
fcis-10302	94	12	uk	uk	PROPN
fcis-10302	94	13	represents	represent	VERB
fcis-10302	94	14	the	the	DET
fcis-10302	94	15	output	output	NOUN
fcis-10302	94	16	value	value	NOUN
fcis-10302	94	17	of	of	ADP
fcis-10302	94	18	the	the	DET
fcis-10302	94	19	kth	kth	PROPN
fcis-10302	94	20	neuron	neuron	PROPN
fcis-10302	94	21	,	,	PUNCT
fcis-10302	94	22	wi	wi	PROPN
fcis-10302	94	23	,	,	PUNCT
fcis-10302	94	24	k	k	PROPN
fcis-10302	94	25	is	be	AUX
fcis-10302	94	26	the	the	DET
fcis-10302	94	27	weight	weight	NOUN
fcis-10302	94	28	value	value	NOUN
fcis-10302	94	29	of	of	ADP
fcis-10302	94	30	the	the	DET
fcis-10302	94	31	ith	ith	PROPN
fcis-10302	94	32	input	input	NOUN
fcis-10302	94	33	for	for	ADP
fcis-10302	94	34	the	the	DET
fcis-10302	94	35	kth	kth	PROPN
fcis-10302	94	36	neuron	neuron	PROPN
fcis-10302	94	37	,	,	PUNCT
fcis-10302	94	38	pi	pi	PROPN
fcis-10302	94	39	is	be	AUX
fcis-10302	94	40	the	the	DET
fcis-10302	94	41	ith	ith	PROPN
fcis-10302	94	42	input	input	NOUN
fcis-10302	94	43	value	value	NOUN
fcis-10302	94	44	,	,	PUNCT
fcis-10302	94	45	and	and	CCONJ
fcis-10302	94	46	bk	bk	PRON
fcis-10302	94	47	is	be	AUX
fcis-10302	94	48	the	the	DET
fcis-10302	94	49	bias	bias	NOUN
fcis-10302	94	50	value	value	NOUN
fcis-10302	94	51	of	of	ADP
fcis-10302	94	52	the	the	DET
fcis-10302	94	53	kth	kth	PROPN
fcis-10302	94	54	neuron	neuron	PROPN
fcis-10302	94	55	.	.	PUNCT
fcis-10302	95	1	after	after	ADP
fcis-10302	95	2	obtaining	obtain	VERB
fcis-10302	95	3	the	the	DET
fcis-10302	95	4	output	output	NOUN
fcis-10302	95	5	from	from	ADP
fcis-10302	95	6	the	the	DET
fcis-10302	95	7	fully	fully	ADV
fcis-10302	95	8	connected	connect	VERB
fcis-10302	95	9	layer	layer	NOUN
fcis-10302	95	10	,	,	PUNCT
fcis-10302	95	11	an	an	DET
fcis-10302	95	12	activation	activation	NOUN
fcis-10302	95	13	function	function	NOUN
fcis-10302	95	14	is	be	AUX
fcis-10302	95	15	applied	apply	VERB
fcis-10302	95	16	.	.	PUNCT
fcis-10302	96	1	in	in	ADP
fcis-10302	96	2	this	this	DET
fcis-10302	96	3	model	model	NOUN
fcis-10302	96	4	,	,	PUNCT
fcis-10302	96	5	the	the	DET
fcis-10302	96	6	chosen	choose	VERB
fcis-10302	96	7	activation	activation	NOUN
fcis-10302	96	8	function	function	NOUN
fcis-10302	96	9	is	be	AUX
fcis-10302	96	10	softmax	softmax	NOUN
fcis-10302	96	11	,	,	PUNCT
fcis-10302	96	12	which	which	PRON
fcis-10302	96	13	is	be	AUX
fcis-10302	96	14	used	use	VERB
fcis-10302	96	15	for	for	ADP
fcis-10302	96	16	multi	multi	ADJ
fcis-10302	96	17	-	-	ADJ
fcis-10302	96	18	class	class	ADJ
fcis-10302	96	19	classification	classification	NOUN
fcis-10302	96	20	.	.	PUNCT
fcis-10302	97	1	2.3	2.3	NUM
fcis-10302	97	2	.	.	PUNCT
fcis-10302	97	3	gru	gru	PROPN
fcis-10302	97	4	gru	gru	PROPN
fcis-10302	97	5	(	(	PUNCT
fcis-10302	97	6	gated	gate	VERB
fcis-10302	97	7	recurrent	recurrent	ADJ
fcis-10302	97	8	unit	unit	NOUN
fcis-10302	97	9	)	)	PUNCT
fcis-10302	97	10	is	be	AUX
fcis-10302	97	11	a	a	DET
fcis-10302	97	12	type	type	NOUN
fcis-10302	97	13	of	of	ADP
fcis-10302	97	14	recurrent	recurrent	ADJ
fcis-10302	97	15	neural	neural	ADJ
fcis-10302	97	16	network	network	NOUN
fcis-10302	97	17	that	that	PRON
fcis-10302	97	18	is	be	AUX
fcis-10302	97	19	simpler	simple	ADJ
fcis-10302	97	20	compared	compare	VERB
fcis-10302	97	21	to	to	ADP
fcis-10302	97	22	the	the	DET
fcis-10302	97	23	lstm	lstm	NOUN
fcis-10302	97	24	structure	structure	NOUN
fcis-10302	97	25	because	because	SCONJ
fcis-10302	97	26	it	it	PRON
fcis-10302	97	27	does	do	AUX
fcis-10302	97	28	not	not	PART
fcis-10302	97	29	have	have	VERB
fcis-10302	97	30	a	a	DET
fcis-10302	97	31	memory	memory	NOUN
fcis-10302	97	32	cell	cell	NOUN
fcis-10302	97	33	.	.	PUNCT
fcis-10302	98	1	gru	gru	PROPN
fcis-10302	98	2	offers	offer	VERB
fcis-10302	98	3	a	a	DET
fcis-10302	98	4	lower	low	ADJ
fcis-10302	98	5	computational	computational	ADJ
fcis-10302	98	6	cost	cost	NOUN
fcis-10302	98	7	while	while	SCONJ
fcis-10302	98	8	delivering	deliver	VERB
fcis-10302	98	9	performance	performance	NOUN
fcis-10302	98	10	comparable	comparable	ADJ
fcis-10302	98	11	to	to	ADP
fcis-10302	98	12	lstm	lstm	PROPN
fcis-10302	98	13	.	.	PUNCT
fcis-10302	99	1	the	the	DET
fcis-10302	99	2	gru	gru	NOUN
fcis-10302	99	3	structure	structure	NOUN
fcis-10302	99	4	consists	consist	VERB
fcis-10302	99	5	of	of	ADP
fcis-10302	99	6	two	two	NUM
fcis-10302	99	7	units	unit	NOUN
fcis-10302	99	8	,	,	PUNCT
fcis-10302	99	9	namely	namely	ADV
fcis-10302	99	10	the	the	DET
fcis-10302	99	11	update	update	NOUN
fcis-10302	99	12	gate	gate	NOUN
fcis-10302	99	13	and	and	CCONJ
fcis-10302	99	14	the	the	DET
fcis-10302	99	15	reset	reset	NOUN
fcis-10302	99	16	gate	gate	NOUN
fcis-10302	99	17	.	.	PUNCT
fcis-10302	100	1	the	the	DET
fcis-10302	100	2	update	update	NOUN
fcis-10302	100	3	gate	gate	NOUN
fcis-10302	100	4	determines	determine	VERB
fcis-10302	100	5	the	the	DET
fcis-10302	100	6	past	past	ADJ
fcis-10302	100	7	information	information	NOUN
fcis-10302	100	8	that	that	PRON
fcis-10302	100	9	needs	need	VERB
fcis-10302	100	10	to	to	PART
fcis-10302	100	11	be	be	AUX
fcis-10302	100	12	retained	retain	VERB
fcis-10302	100	13	or	or	CCONJ
fcis-10302	100	14	forgotten	forget	VERB
fcis-10302	100	15	in	in	ADP
fcis-10302	100	16	the	the	DET
fcis-10302	100	17	time	time	NOUN
fcis-10302	100	18	series	series	NOUN
fcis-10302	100	19	to	to	PART
fcis-10302	100	20	meet	meet	VERB
fcis-10302	100	21	the	the	DET
fcis-10302	100	22	current	current	ADJ
fcis-10302	100	23	prediction	prediction	NOUN
fcis-10302	100	24	requirements	requirement	NOUN
fcis-10302	100	25	.	.	PUNCT
fcis-10302	101	1	the	the	DET
fcis-10302	101	2	reset	reset	NOUN
fcis-10302	101	3	gate	gate	NOUN
fcis-10302	101	4	,	,	PUNCT
fcis-10302	101	5	on	on	ADP
fcis-10302	101	6	the	the	DET
fcis-10302	101	7	other	other	ADJ
fcis-10302	101	8	hand	hand	NOUN
fcis-10302	101	9	,	,	PUNCT
fcis-10302	101	10	decides	decide	VERB
fcis-10302	101	11	the	the	DET
fcis-10302	101	12	amount	amount	NOUN
fcis-10302	101	13	of	of	ADP
fcis-10302	101	14	information	information	NOUN
fcis-10302	101	15	that	that	PRON
fcis-10302	101	16	needs	need	VERB
fcis-10302	101	17	to	to	PART
fcis-10302	101	18	be	be	AUX
fcis-10302	101	19	memorized	memorize	VERB
fcis-10302	101	20	.	.	PUNCT
fcis-10302	102	1	the	the	DET
fcis-10302	102	2	illustration	illustration	NOUN
fcis-10302	102	3	in	in	ADP
fcis-10302	102	4	figure	figure	NOUN
fcis-10302	102	5	1	1	NUM
fcis-10302	102	6	demonstrates	demonstrate	VERB
fcis-10302	102	7	this	this	DET
fcis-10302	102	8	process	process	NOUN
fcis-10302	102	9	,	,	PUNCT
fcis-10302	102	10	as	as	ADP
fcis-10302	102	11	the	the	DET
fcis-10302	102	12	figure	figure	NOUN
fcis-10302	102	13	1	1	NUM
fcis-10302	102	14	.	.	PUNCT
fcis-10302	103	1	in	in	ADP
fcis-10302	103	2	gru	gru	PROPN
fcis-10302	103	3	,	,	PUNCT
fcis-10302	103	4	the	the	DET
fcis-10302	103	5	current	current	ADJ
fcis-10302	103	6	input	input	NOUN
fcis-10302	103	7	tx	tx	PROPN
fcis-10302	103	8	and	and	CCONJ
fcis-10302	103	9	the	the	DET
fcis-10302	103	10	previous	previous	ADJ
fcis-10302	103	11	output	output	NOUN
fcis-10302	103	12	ht-1	ht-1	VERB
fcis-10302	103	13	are	be	AUX
fcis-10302	103	14	first	first	ADV
fcis-10302	103	15	concatenated	concatenate	VERB
fcis-10302	103	16	as	as	ADP
fcis-10302	103	17	the	the	DET
fcis-10302	103	18	input	input	NOUN
fcis-10302	103	19	to	to	ADP
fcis-10302	103	20	the	the	DET
fcis-10302	103	21	update	update	NOUN
fcis-10302	103	22	gate	gate	NOUN
fcis-10302	103	23	and	and	CCONJ
fcis-10302	103	24	the	the	DET
fcis-10302	103	25	reset	reset	NOUN
fcis-10302	103	26	gate	gate	NOUN
fcis-10302	103	27	.	.	PUNCT
fcis-10302	104	1	additionally	additionally	ADV
fcis-10302	104	2	,	,	PUNCT
fcis-10302	104	3	each	each	DET
fcis-10302	104	4	value	value	NOUN
fcis-10302	104	5	in	in	ADP
fcis-10302	104	6	every	every	DET
fcis-10302	104	7	unit	unit	NOUN
fcis-10302	104	8	or	or	CCONJ
fcis-10302	104	9	component	component	NOUN
fcis-10302	104	10	of	of	ADP
fcis-10302	104	11	the	the	DET
fcis-10302	104	12	gru	gru	NOUN
fcis-10302	104	13	can	can	AUX
fcis-10302	104	14	be	be	AUX
fcis-10302	104	15	obtained	obtain	VERB
fcis-10302	104	16	as	as	ADP
fcis-10302	104	17	follows	follow	VERB
fcis-10302	104	18	:	:	PUNCT
fcis-10302	104	19			PROPN
fcis-10302	104	20	tz	tz	PROPN
fcis-10302	104	21	t-1	t-1	PROPN
fcis-10302	104	22	zz	zz	PROPN
fcis-10302	105	1	w	w	PROPN
fcis-10302	105	2	h	h	PROPN
fcis-10302	105	3	x	x	VERB
fcis-10302	105	4	bt	bt	PROPN
fcis-10302	105	5			PROPN
fcis-10302	105	6			PROPN
fcis-10302	105	7			PROPN
fcis-10302	105	8			NOUN
fcis-10302	105	9			PROPN
fcis-10302	105	10	，	，	PUNCT
fcis-10302	105	11	(	(	PUNCT
fcis-10302	105	12	7	7	X
fcis-10302	105	13	)	)	PUNCT
fcis-10302	105	14	t	t	NOUN
fcis-10302	106	1	tt-1r	tt-1r	VERB
fcis-10302	106	2	r	r	NOUN
fcis-10302	106	3	r	r	NOUN
fcis-10302	106	4	w	w	NOUN
fcis-10302	106	5	h	h	NOUN
fcis-10302	106	6	x	x	PROPN
fcis-10302	106	7	b	b	NOUN
fcis-10302	106	8			PROPN
fcis-10302	107	1			NOUN
fcis-10302	107	2			VERB
fcis-10302	107	3			PROPN
fcis-10302	107	4	（	（	PUNCT
fcis-10302	107	5	，	，	PROPN
fcis-10302	108	1	+	+	PUNCT
fcis-10302	108	2	）	）	PUNCT
fcis-10302	108	3	(	(	PUNCT
fcis-10302	108	4	8)	8)	NUM
fcis-10302	108	5	-1tanh	-1tanh	NOUN
fcis-10302	108	6	tt	tt	PROPN
fcis-10302	108	7	tth	tth	PROPN
fcis-10302	108	8	h	h	PROPN
fcis-10302	108	9	h	h	PROPN
fcis-10302	109	1	w	w	PROPN
fcis-10302	109	2	h	h	PROPN
fcis-10302	109	3	x	x	X
fcis-10302	109	4	b	b	VERB
fcis-10302	109	5			NOUN
fcis-10302	109	6			PUNCT
fcis-10302	109	7			NOUN
fcis-10302	109	8			PROPN
fcis-10302	109	9			NOUN
fcis-10302	109	10	（	（	PUNCT
fcis-10302	110	1	，	，	PROPN
fcis-10302	110	2	）	）	PUNCT
fcis-10302	110	3	(	(	PUNCT
fcis-10302	110	4	9	9	NUM
fcis-10302	110	5	)	)	PUNCT
fcis-10302	110	6	where	where	SCONJ
fcis-10302	110	7	tz	tz	NOUN
fcis-10302	110	8	,	,	PUNCT
fcis-10302	110	9	tr	tr	VERB
fcis-10302	110	10	,	,	PUNCT
fcis-10302	110	11	and	and	CCONJ
fcis-10302	110	12	th	th	X
fcis-10302	110	13	represent	represent	VERB
fcis-10302	110	14	the	the	DET
fcis-10302	110	15	outputs	output	NOUN
fcis-10302	110	16	of	of	ADP
fcis-10302	110	17	the	the	DET
fcis-10302	110	18	update	update	NOUN
fcis-10302	110	19	gate	gate	NOUN
fcis-10302	110	20	,	,	PUNCT
fcis-10302	110	21	reset	reset	NOUN
fcis-10302	110	22	gate	gate	NOUN
fcis-10302	110	23	,	,	PUNCT
fcis-10302	110	24	and	and	CCONJ
fcis-10302	110	25	candidate	candidate	NOUN
fcis-10302	110	26	hidden	hide	VERB
fcis-10302	110	27	state	state	NOUN
fcis-10302	110	28	,	,	PUNCT
fcis-10302	110	29	respectively	respectively	ADV
fcis-10302	110	30	.	.	PUNCT
fcis-10302	111	1			PRON
fcis-10302	111	2	and	and	CCONJ
fcis-10302	111	3	tanh	tanh	PROPN
fcis-10302	111	4	represent	represent	VERB
fcis-10302	111	5	the	the	DET
fcis-10302	111	6	activation	activation	NOUN
fcis-10302	111	7	functions	function	NOUN
fcis-10302	111	8	sigmoid	sigmoid	NOUN
fcis-10302	111	9	and	and	CCONJ
fcis-10302	111	10	hyperbolic	hyperbolic	ADJ
fcis-10302	111	11	tangent	tangent	NOUN
fcis-10302	111	12	function	function	NOUN
fcis-10302	111	13	,	,	PUNCT
fcis-10302	111	14	92	92	NUM
fcis-10302	111	15	respectively	respectively	ADV
fcis-10302	111	16	.	.	PUNCT
fcis-10302	112	1	zw	zw	PROPN
fcis-10302	112	2	,	,	PUNCT
fcis-10302	112	3	w	w	PROPN
fcis-10302	112	4			PUNCT
fcis-10302	112	5	,	,	PUNCT
fcis-10302	112	6	h	h	PROPN
fcis-10302	112	7	w	w	PROPN
fcis-10302	112	8	and	and	CCONJ
fcis-10302	112	9	zb	zb	PROPN
fcis-10302	112	10	,	,	PUNCT
fcis-10302	112	11	b	b	PROPN
fcis-10302	112	12			NUM
fcis-10302	112	13	,	,	PUNCT
fcis-10302	112	14	h	h	PROPN
fcis-10302	112	15	b	b	PROPN
fcis-10302	112	16	,	,	PUNCT
fcis-10302	112	17	and	and	CCONJ
fcis-10302	112	18	b	b	X
fcis-10302	112	19	are	be	AUX
fcis-10302	112	20	the	the	DET
fcis-10302	112	21	weight	weight	NOUN
fcis-10302	112	22	matrices	matrix	NOUN
fcis-10302	112	23	and	and	CCONJ
fcis-10302	112	24	bias	bias	NOUN
fcis-10302	112	25	vectors	vector	NOUN
fcis-10302	112	26	associated	associate	VERB
fcis-10302	112	27	with	with	ADP
fcis-10302	112	28	the	the	DET
fcis-10302	112	29	update	update	NOUN
fcis-10302	112	30	gate	gate	NOUN
fcis-10302	112	31	,	,	PUNCT
fcis-10302	112	32	reset	reset	NOUN
fcis-10302	112	33	gate	gate	NOUN
fcis-10302	112	34	,	,	PUNCT
fcis-10302	112	35	and	and	CCONJ
fcis-10302	112	36	candidate	candidate	NOUN
fcis-10302	112	37	hidden	hide	VERB
fcis-10302	112	38	state	state	NOUN
fcis-10302	112	39	.	.	PUNCT
fcis-10302	113	1	the	the	DET
fcis-10302	113	2	output	output	NOUN
fcis-10302	113	3	of	of	ADP
fcis-10302	113	4	the	the	DET
fcis-10302	113	5	tth	tth	PROPN
fcis-10302	113	6	gru	gru	PROPN
fcis-10302	113	7	can	can	AUX
fcis-10302	113	8	be	be	AUX
fcis-10302	113	9	obtained	obtain	VERB
fcis-10302	113	10	as	as	ADP
fcis-10302	113	11	follows	follow	VERB
fcis-10302	113	12	:	:	PUNCT
fcis-10302	113	13	tt-1(1	tt-1(1	DET
fcis-10302	113	14	)	)	PUNCT
fcis-10302	113	15	h	h	NOUN
fcis-10302	113	16	z	z	NOUN
fcis-10302	114	1	ht	ht	INTJ
fcis-10302	114	2	t	t	PROPN
fcis-10302	114	3	th	th	X
fcis-10302	114	4	z	z	PROPN
fcis-10302	114	5			PROPN
fcis-10302	114	6			PROPN
fcis-10302	114	7			PROPN
fcis-10302	114	8			PROPN
fcis-10302	114	9	(	(	PUNCT
fcis-10302	114	10	10	10	NUM
fcis-10302	114	11	)	)	PUNCT
fcis-10302	114	12	×	×	NOUN
fcis-10302	114	13	××	××	NOUN
fcis-10302	114	14	＋	＋	PUNCT
fcis-10302	114	15			NOUN
fcis-10302	114	16	tx	tx	ADP
fcis-10302	114	17	t-1h	t-1h	PROPN
fcis-10302	114	18	th	th	X
fcis-10302	114	19	tanh	tanh	PROPN
fcis-10302	114	20	tr	tr	ADV
fcis-10302	114	21	thtz	thtz	ADV
fcis-10302	114	22	t1	t1	NOUN
fcis-10302	114	23	z	z	NOUN
fcis-10302	114	24	figure	figure	NOUN
fcis-10302	114	25	1	1	NUM
fcis-10302	114	26	.	.	PUNCT
fcis-10302	114	27	gru	gru	NOUN
fcis-10302	114	28	architecture	architecture	NOUN
fcis-10302	114	29	3	3	NUM
fcis-10302	114	30	.	.	PUNCT
fcis-10302	114	31	model	model	NOUN
fcis-10302	114	32	design	design	NOUN
fcis-10302	114	33	to	to	PART
fcis-10302	114	34	improve	improve	VERB
fcis-10302	114	35	the	the	DET
fcis-10302	114	36	detection	detection	NOUN
fcis-10302	114	37	performance	performance	NOUN
fcis-10302	114	38	of	of	ADP
fcis-10302	114	39	iot	iot	PROPN
fcis-10302	114	40	intrusion	intrusion	PROPN
fcis-10302	114	41	detection	detection	NOUN
fcis-10302	114	42	,	,	PUNCT
fcis-10302	114	43	a	a	DET
fcis-10302	114	44	deep	deep	ADJ
fcis-10302	114	45	learning	learning	NOUN
fcis-10302	114	46	approach	approach	NOUN
fcis-10302	114	47	based	base	VERB
fcis-10302	114	48	on	on	ADP
fcis-10302	114	49	xgboost	xgboost	X
fcis-10302	114	50	feature	feature	NOUN
fcis-10302	114	51	selection	selection	NOUN
fcis-10302	114	52	and	and	CCONJ
fcis-10302	114	53	cnn	cnn	PROPN
fcis-10302	114	54	-	-	PUNCT
fcis-10302	114	55	gru	gru	PROPN
fcis-10302	114	56	is	be	AUX
fcis-10302	114	57	proposed	propose	VERB
fcis-10302	114	58	.	.	PUNCT
fcis-10302	115	1	the	the	DET
fcis-10302	115	2	model	model	NOUN
fcis-10302	115	3	consists	consist	VERB
fcis-10302	115	4	of	of	ADP
fcis-10302	115	5	three	three	NUM
fcis-10302	115	6	modules	module	NOUN
fcis-10302	115	7	:	:	PUNCT
fcis-10302	115	8	data	datum	NOUN
fcis-10302	115	9	preprocessing	preprocessing	NOUN
fcis-10302	115	10	module	module	NOUN
fcis-10302	115	11	,	,	PUNCT
fcis-10302	115	12	feature	feature	NOUN
fcis-10302	115	13	selection	selection	NOUN
fcis-10302	115	14	module	module	NOUN
fcis-10302	115	15	,	,	PUNCT
fcis-10302	115	16	and	and	CCONJ
fcis-10302	115	17	classifier	classifier	NOUN
fcis-10302	115	18	module	module	NOUN
fcis-10302	115	19	.	.	PUNCT
fcis-10302	116	1	the	the	DET
fcis-10302	116	2	design	design	NOUN
fcis-10302	116	3	process	process	NOUN
fcis-10302	116	4	of	of	ADP
fcis-10302	116	5	the	the	DET
fcis-10302	116	6	model	model	NOUN
fcis-10302	116	7	is	be	AUX
fcis-10302	116	8	as	as	SCONJ
fcis-10302	116	9	follows	follow	VERB
fcis-10302	116	10	:	:	PUNCT
fcis-10302	116	11	input	input	NOUN
fcis-10302	116	12	:	:	PUNCT
fcis-10302	116	13	relevant	relevant	ADJ
fcis-10302	116	14	traffic	traffic	NOUN
fcis-10302	116	15	data	datum	NOUN
fcis-10302	116	16	collected	collect	VERB
fcis-10302	116	17	from	from	ADP
fcis-10302	116	18	iot	iot	PROPN
fcis-10302	116	19	devices	device	NOUN
fcis-10302	116	20	.	.	PUNCT
fcis-10302	117	1	the	the	DET
fcis-10302	117	2	publicly	publicly	ADV
fcis-10302	117	3	available	available	ADJ
fcis-10302	117	4	n	n	CCONJ
fcis-10302	117	5	-	-	PUNCT
fcis-10302	117	6	baiot	baiot	NOUN
fcis-10302	117	7	dataset	dataset	NOUN
fcis-10302	117	8	,	,	PUNCT
fcis-10302	117	9	specifically	specifically	ADV
fcis-10302	117	10	the	the	DET
fcis-10302	117	11	data	datum	NOUN
fcis-10302	117	12	related	relate	VERB
fcis-10302	117	13	to	to	ADP
fcis-10302	117	14	baby	baby	NOUN
fcis-10302	117	15	monitor	monitor	NOUN
fcis-10302	117	16	devices	device	NOUN
fcis-10302	117	17	,	,	PUNCT
fcis-10302	117	18	is	be	AUX
fcis-10302	117	19	used	use	VERB
fcis-10302	117	20	in	in	ADP
fcis-10302	117	21	this	this	DET
fcis-10302	117	22	experiment	experiment	NOUN
fcis-10302	117	23	.	.	PUNCT
fcis-10302	118	1	output	output	NOUN
fcis-10302	118	2	:	:	PUNCT
fcis-10302	118	3	attack	attack	NOUN
fcis-10302	118	4	categories	category	NOUN
fcis-10302	118	5	targeting	target	VERB
fcis-10302	118	6	iot	iot	NOUN
fcis-10302	118	7	devices	device	NOUN
fcis-10302	118	8	.	.	PUNCT
fcis-10302	119	1	step	step	NOUN
fcis-10302	119	2	1	1	NUM
fcis-10302	119	3	:	:	PUNCT
fcis-10302	119	4	perform	perform	VERB
fcis-10302	119	5	missing	miss	VERB
fcis-10302	119	6	value	value	NOUN
fcis-10302	119	7	handling	handling	NOUN
fcis-10302	119	8	on	on	ADP
fcis-10302	119	9	the	the	DET
fcis-10302	119	10	initial	initial	ADJ
fcis-10302	119	11	input	input	NOUN
fcis-10302	119	12	dataset	dataset	NOUN
fcis-10302	119	13	.	.	PUNCT
fcis-10302	120	1	step	step	NOUN
fcis-10302	120	2	2	2	NUM
fcis-10302	120	3	:	:	PUNCT
fcis-10302	120	4	normalize	normalize	VERB
fcis-10302	120	5	the	the	DET
fcis-10302	120	6	dataset	dataset	NOUN
fcis-10302	120	7	n	n	CCONJ
fcis-10302	120	8	,	,	PUNCT
fcis-10302	120	9	convert	convert	VERB
fcis-10302	120	10	non	non	ADJ
fcis-10302	120	11	-	-	ADJ
fcis-10302	120	12	numeric	numeric	ADJ
fcis-10302	120	13	features	feature	NOUN
fcis-10302	120	14	to	to	ADP
fcis-10302	120	15	numerical	numerical	ADJ
fcis-10302	120	16	values	value	NOUN
fcis-10302	120	17	,	,	PUNCT
fcis-10302	120	18	and	and	CCONJ
fcis-10302	120	19	encode	encode	VERB
fcis-10302	120	20	the	the	DET
fcis-10302	120	21	labels	label	NOUN
fcis-10302	120	22	to	to	PART
fcis-10302	120	23	obtain	obtain	VERB
fcis-10302	120	24	the	the	DET
fcis-10302	120	25	preprocessed	preprocesse	VERB
fcis-10302	120	26	dataset	dataset	NOUN
fcis-10302	120	27	n.	n.	NOUN
fcis-10302	120	28	step	step	NOUN
fcis-10302	120	29	3	3	NUM
fcis-10302	120	30	:	:	PUNCT
fcis-10302	120	31	split	split	VERB
fcis-10302	120	32	the	the	DET
fcis-10302	120	33	dataset	dataset	NOUN
fcis-10302	120	34	into	into	ADP
fcis-10302	120	35	training	training	NOUN
fcis-10302	120	36	set	set	VERB
fcis-10302	120	37	n1	n1	NOUN
fcis-10302	120	38	and	and	CCONJ
fcis-10302	120	39	test	test	NOUN
fcis-10302	120	40	set	set	VERB
fcis-10302	120	41	n2	n2	PROPN
fcis-10302	120	42	.	.	PUNCT
fcis-10302	121	1	perform	perform	VERB
fcis-10302	121	2	feature	feature	NOUN
fcis-10302	121	3	selection	selection	NOUN
fcis-10302	121	4	using	use	VERB
fcis-10302	121	5	xgboost	xgboost	ADP
fcis-10302	121	6	feature	feature	NOUN
fcis-10302	121	7	importance	importance	NOUN
fcis-10302	121	8	scoring	score	VERB
fcis-10302	121	9	to	to	PART
fcis-10302	121	10	obtain	obtain	VERB
fcis-10302	121	11	the	the	DET
fcis-10302	121	12	optimal	optimal	ADJ
fcis-10302	121	13	feature	feature	NOUN
fcis-10302	121	14	subset	subset	VERB
fcis-10302	121	15	f.	f.	PROPN
fcis-10302	121	16	step	step	PROPN
fcis-10302	121	17	4	4	NUM
fcis-10302	121	18	:	:	PUNCT
fcis-10302	121	19	utilize	utilize	VERB
fcis-10302	121	20	the	the	DET
fcis-10302	121	21	selected	select	VERB
fcis-10302	121	22	optimal	optimal	ADJ
fcis-10302	121	23	feature	feature	NOUN
fcis-10302	121	24	set	set	VERB
fcis-10302	121	25	f	f	PROPN
fcis-10302	121	26	to	to	PART
fcis-10302	121	27	obtain	obtain	VERB
fcis-10302	121	28	the	the	DET
fcis-10302	121	29	training	training	NOUN
fcis-10302	121	30	set	set	VERB
fcis-10302	121	31	q1	q1	PROPN
fcis-10302	121	32	and	and	CCONJ
fcis-10302	121	33	test	test	NOUN
fcis-10302	121	34	set	set	VERB
fcis-10302	121	35	q2	q2	NOUN
fcis-10302	121	36	,	,	PUNCT
fcis-10302	121	37	removing	remove	VERB
fcis-10302	121	38	redundant	redundant	ADJ
fcis-10302	121	39	features	feature	NOUN
fcis-10302	121	40	.	.	PUNCT
fcis-10302	122	1	step	step	NOUN
fcis-10302	122	2	5	5	NUM
fcis-10302	122	3	:	:	PUNCT
fcis-10302	122	4	train	train	VERB
fcis-10302	122	5	the	the	DET
fcis-10302	122	6	cnn	cnn	PROPN
fcis-10302	122	7	-	-	PUNCT
fcis-10302	122	8	gru	gru	PROPN
fcis-10302	122	9	model	model	NOUN
fcis-10302	122	10	using	use	VERB
fcis-10302	122	11	the	the	DET
fcis-10302	122	12	adam	adam	PROPN
fcis-10302	122	13	optimizer	optimizer	NOUN
fcis-10302	122	14	to	to	PART
fcis-10302	122	15	optimize	optimize	VERB
fcis-10302	122	16	the	the	DET
fcis-10302	122	17	constructed	construct	VERB
fcis-10302	122	18	cnn	cnn	PROPN
fcis-10302	122	19	-	-	PUNCT
fcis-10302	122	20	gru	gru	PROPN
fcis-10302	122	21	classifier	classifier	NOUN
fcis-10302	122	22	.	.	PUNCT
fcis-10302	123	1	step	step	NOUN
fcis-10302	123	2	6	6	NUM
fcis-10302	123	3	:	:	PUNCT
fcis-10302	123	4	evaluate	evaluate	VERB
fcis-10302	123	5	the	the	DET
fcis-10302	123	6	proposed	propose	VERB
fcis-10302	123	7	cnn	cnn	PROPN
fcis-10302	123	8	-	-	PUNCT
fcis-10302	123	9	gru	gru	PROPN
fcis-10302	123	10	classifier	classifier	PROPN
fcis-10302	123	11	model	model	NOUN
fcis-10302	123	12	using	use	VERB
fcis-10302	123	13	metrics	metric	NOUN
fcis-10302	123	14	such	such	ADJ
fcis-10302	123	15	as	as	ADP
fcis-10302	123	16	accuracy	accuracy	NOUN
fcis-10302	123	17	,	,	PUNCT
fcis-10302	123	18	precision	precision	NOUN
fcis-10302	123	19	,	,	PUNCT
fcis-10302	123	20	recall	recall	NOUN
fcis-10302	123	21	,	,	PUNCT
fcis-10302	123	22	and	and	CCONJ
fcis-10302	123	23	f1	f1	NOUN
fcis-10302	123	24	-	-	PUNCT
fcis-10302	123	25	score	score	NOUN
fcis-10302	123	26	.	.	PUNCT
fcis-10302	124	1	3.1	3.1	NUM
fcis-10302	124	2	.	.	PUNCT
fcis-10302	124	3	data	datum	NOUN
fcis-10302	124	4	preprocessing	preprocessing	NOUN
fcis-10302	124	5	module	module	NOUN
fcis-10302	124	6	non	non	ADJ
fcis-10302	124	7	-	-	ADJ
fcis-10302	124	8	numeric	numeric	ADJ
fcis-10302	124	9	features	feature	NOUN
fcis-10302	124	10	are	be	AUX
fcis-10302	124	11	encoded	encode	VERB
fcis-10302	124	12	using	use	VERB
fcis-10302	124	13	one	one	NUM
fcis-10302	124	14	-	-	PUNCT
fcis-10302	124	15	hot	hot	ADJ
fcis-10302	124	16	encoding	encoding	NOUN
fcis-10302	124	17	to	to	PART
fcis-10302	124	18	convert	convert	VERB
fcis-10302	124	19	them	they	PRON
fcis-10302	124	20	into	into	ADP
fcis-10302	124	21	numerical	numerical	ADJ
fcis-10302	124	22	values	value	NOUN
fcis-10302	124	23	.	.	PUNCT
fcis-10302	125	1	then	then	ADV
fcis-10302	125	2	,	,	PUNCT
fcis-10302	125	3	the	the	DET
fcis-10302	125	4	feature	feature	NOUN
fcis-10302	125	5	data	datum	NOUN
fcis-10302	125	6	is	be	AUX
fcis-10302	125	7	standardized	standardize	VERB
fcis-10302	125	8	to	to	PART
fcis-10302	125	9	eliminate	eliminate	VERB
fcis-10302	125	10	the	the	DET
fcis-10302	125	11	scale	scale	NOUN
fcis-10302	125	12	differences	difference	NOUN
fcis-10302	125	13	between	between	ADP
fcis-10302	125	14	features	feature	NOUN
fcis-10302	125	15	.	.	PUNCT
fcis-10302	126	1	the	the	DET
fcis-10302	126	2	standardization	standardization	NOUN
fcis-10302	126	3	process	process	NOUN
fcis-10302	126	4	is	be	AUX
fcis-10302	126	5	shown	show	VERB
fcis-10302	126	6	in	in	ADP
fcis-10302	126	7	equations	equation	NOUN
fcis-10302	126	8	(	(	PUNCT
fcis-10302	126	9	13)-(15	13)-(15	NUM
fcis-10302	126	10	):	):	PUNCT
fcis-10302	126	11	'	'	PUNCT
fcis-10302	126	12	t	t	X
fcis-10302	126	13	ij	ij	INTJ
fcis-10302	127	1	j	j	PROPN
fcis-10302	127	2	ij	ij	INTJ
fcis-10302	127	3	j	j	PROPN
fcis-10302	127	4	t	t	PROPN
fcis-10302	127	5	avg	avg	PROPN
fcis-10302	127	6			PROPN
fcis-10302	127	7			PROPN
fcis-10302	127	8			NUM
fcis-10302	127	9	(	(	PUNCT
fcis-10302	127	10	11	11	NUM
fcis-10302	127	11	)	)	PUNCT
fcis-10302	127	12	j	j	NOUN
fcis-10302	127	13	1	1	NUM
fcis-10302	127	14	2	2	NUM
fcis-10302	127	15	nj	nj	PROPN
fcis-10302	127	16	1	1	NUM
fcis-10302	127	17	(	(	PUNCT
fcis-10302	127	18	t	t	PROPN
fcis-10302	127	19	)	)	PUNCT
fcis-10302	128	1	j	j	PROPN
fcis-10302	128	2	javg	javg	PROPN
fcis-10302	128	3	t	t	PROPN
fcis-10302	128	4	t	t	PROPN
fcis-10302	128	5	n	n	PROPN
fcis-10302	128	6			PROPN
fcis-10302	128	7			ADJ
fcis-10302	128	8			ADV
fcis-10302	128	9	…	…	PUNCT
fcis-10302	128	10	+	+	CCONJ
fcis-10302	128	11	(	(	PUNCT
fcis-10302	128	12	12	12	NUM
fcis-10302	128	13	)	)	SYM
fcis-10302	128	14	2	2	NUM
fcis-10302	128	15	2	2	NUM
fcis-10302	128	16	2	2	NUM
fcis-10302	128	17	j	j	NOUN
fcis-10302	128	18	1	1	NUM
fcis-10302	128	19	j	j	PROPN
fcis-10302	128	20	2	2	NUM
fcis-10302	128	21	j	j	PROPN
fcis-10302	128	22	n	n	CCONJ
fcis-10302	128	23	j	j	PROPN
fcis-10302	128	24	1	1	NUM
fcis-10302	128	25	(	(	PUNCT
fcis-10302	128	26	(	(	PUNCT
fcis-10302	128	27	(	(	PUNCT
fcis-10302	128	28	j	j	PROPN
fcis-10302	128	29	j	j	PROPN
fcis-10302	128	30	jt	jt	PROPN
fcis-10302	128	31	avg	avg	PROPN
fcis-10302	128	32	t	t	PROPN
fcis-10302	128	33	avg	avg	PROPN
fcis-10302	128	34	t	t	PROPN
fcis-10302	128	35	avg	avg	PROPN
fcis-10302	128	36	n	n	PROPN
fcis-10302	128	37			PROPN
fcis-10302	128	38			PROPN
fcis-10302	128	39			PROPN
fcis-10302	128	40			PROPN
fcis-10302	128	41			VERB
fcis-10302	128	42			PROPN
fcis-10302	128	43			ADV
fcis-10302	128	44			NOUN
fcis-10302	128	45			PROPN
fcis-10302	128	46	）	）	SYM
fcis-10302	128	47	）	）	PUNCT
fcis-10302	129	1	…	…	PUNCT
fcis-10302	129	2	+	+	NUM
fcis-10302	129	3	）	）	PUNCT
fcis-10302	129	4	(	(	PUNCT
fcis-10302	129	5	13	13	NUM
fcis-10302	129	6	)	)	PUNCT
fcis-10302	129	7	here	here	ADV
fcis-10302	129	8	,	,	PUNCT
fcis-10302	129	9	tij	tij	PROPN
fcis-10302	129	10	represents	represent	VERB
fcis-10302	129	11	the	the	DET
fcis-10302	129	12	feature	feature	NOUN
fcis-10302	129	13	value	value	NOUN
fcis-10302	129	14	,	,	PUNCT
fcis-10302	129	15	tij	tij	PROPN
fcis-10302	129	16	’	'	PUNCT
fcis-10302	129	17	represents	represent	VERB
fcis-10302	129	18	the	the	DET
fcis-10302	129	19	standardized	standardized	ADJ
fcis-10302	129	20	feature	feature	NOUN
fcis-10302	129	21	value	value	NOUN
fcis-10302	129	22	,	,	PUNCT
fcis-10302	129	23	avgj	avgj	PROPN
fcis-10302	129	24	represents	represent	VERB
fcis-10302	129	25	the	the	DET
fcis-10302	129	26	average	average	ADJ
fcis-10302	129	27	value	value	NOUN
fcis-10302	129	28	of	of	ADP
fcis-10302	129	29	the	the	DET
fcis-10302	129	30	feature	feature	NOUN
fcis-10302	129	31	,	,	PUNCT
fcis-10302	129	32	and	and	CCONJ
fcis-10302	129	33	σj	σj	VERB
fcis-10302	129	34	represents	represent	VERB
fcis-10302	129	35	the	the	DET
fcis-10302	129	36	standard	standard	ADJ
fcis-10302	129	37	deviation	deviation	NOUN
fcis-10302	129	38	of	of	ADP
fcis-10302	129	39	the	the	DET
fcis-10302	129	40	feature	feature	NOUN
fcis-10302	129	41	.	.	PUNCT
fcis-10302	130	1	3.2	3.2	NUM
fcis-10302	130	2	.	.	PUNCT
fcis-10302	131	1	feature	feature	NOUN
fcis-10302	131	2	selection	selection	NOUN
fcis-10302	131	3	module	module	NOUN
fcis-10302	131	4	in	in	ADP
fcis-10302	131	5	this	this	DET
fcis-10302	131	6	study	study	NOUN
fcis-10302	131	7	,	,	PUNCT
fcis-10302	131	8	the	the	DET
fcis-10302	131	9	xgboost	xgboost	PROPN
fcis-10302	131	10	model	model	NOUN
fcis-10302	131	11	is	be	AUX
fcis-10302	131	12	used	use	VERB
fcis-10302	131	13	as	as	ADP
fcis-10302	131	14	a	a	DET
fcis-10302	131	15	feature	feature	NOUN
fcis-10302	131	16	selector	selector	NOUN
fcis-10302	131	17	.	.	PUNCT
fcis-10302	132	1	information	information	NOUN
fcis-10302	132	2	gain	gain	NOUN
fcis-10302	132	3	is	be	AUX
fcis-10302	132	4	employed	employ	VERB
fcis-10302	132	5	to	to	PART
fcis-10302	132	6	filter	filter	VERB
fcis-10302	132	7	the	the	DET
fcis-10302	132	8	features	feature	NOUN
fcis-10302	132	9	,	,	PUNCT
fcis-10302	132	10	and	and	CCONJ
fcis-10302	132	11	its	its	PRON
fcis-10302	132	12	formula	formula	NOUN
fcis-10302	132	13	is	be	AUX
fcis-10302	132	14	as	as	SCONJ
fcis-10302	132	15	follows	follow	VERB
fcis-10302	132	16	:	:	PUNCT
fcis-10302	132	17	2	2	NUM
fcis-10302	132	18	2	2	NUM
fcis-10302	132	19	2	2	NUM
fcis-10302	132	20	(	(	PUNCT
fcis-10302	132	21	)	)	PUNCT
fcis-10302	132	22	1	1	NUM
fcis-10302	132	23	2	2	NUM
fcis-10302	132	24	l	l	NOUN
fcis-10302	132	25	r	r	NOUN
fcis-10302	132	26	l	l	NOUN
fcis-10302	132	27	r	r	NOUN
fcis-10302	132	28	l	l	NOUN
fcis-10302	132	29	r	r	NOUN
fcis-10302	132	30	l	l	NOUN
fcis-10302	132	31	r	r	NOUN
fcis-10302	133	1	g	g	NOUN
fcis-10302	133	2	g	g	NOUN
fcis-10302	133	3	g	g	NOUN
fcis-10302	133	4	g	g	PROPN
fcis-10302	133	5	gain	gain	NOUN
fcis-10302	133	6	h	h	PROPN
fcis-10302	133	7	h	h	NOUN
fcis-10302	133	8	h	h	PROPN
fcis-10302	133	9	h	h	PROPN
fcis-10302	133	10			PROPN
fcis-10302	133	11			X
fcis-10302	133	12			ADJ
fcis-10302	133	13			ADJ
fcis-10302	133	14			NOUN
fcis-10302	133	15			VERB
fcis-10302	133	16			PROPN
fcis-10302	133	17			ADJ
fcis-10302	133	18			PROPN
fcis-10302	133	19			PROPN
fcis-10302	133	20			PROPN
fcis-10302	133	21			ADV
fcis-10302	133	22			PUNCT
fcis-10302	133	23			NUM
fcis-10302	133	24			NOUN
fcis-10302	133	25	(	(	PUNCT
fcis-10302	133	26	14	14	NUM
fcis-10302	133	27	)	)	PUNCT
fcis-10302	133	28	where	where	SCONJ
fcis-10302	133	29	gl	gl	PROPN
fcis-10302	133	30	represents	represent	VERB
fcis-10302	133	31	the	the	DET
fcis-10302	133	32	sum	sum	NOUN
fcis-10302	133	33	of	of	ADP
fcis-10302	133	34	the	the	DET
fcis-10302	133	35	first	first	ADJ
fcis-10302	133	36	-	-	PUNCT
fcis-10302	133	37	order	order	NOUN
fcis-10302	133	38	gradients	gradient	NOUN
fcis-10302	133	39	of	of	ADP
fcis-10302	133	40	the	the	DET
fcis-10302	133	41	loss	loss	NOUN
fcis-10302	133	42	function	function	NOUN
fcis-10302	133	43	for	for	ADP
fcis-10302	133	44	the	the	DET
fcis-10302	133	45	left	left	ADJ
fcis-10302	133	46	leaf	leaf	NOUN
fcis-10302	133	47	node	node	NOUN
fcis-10302	133	48	,	,	PUNCT
fcis-10302	133	49	hl	hl	PROPN
fcis-10302	133	50	represents	represent	VERB
fcis-10302	133	51	the	the	DET
fcis-10302	133	52	sum	sum	NOUN
fcis-10302	133	53	of	of	ADP
fcis-10302	133	54	the	the	DET
fcis-10302	133	55	second	second	ADJ
fcis-10302	133	56	-	-	PUNCT
fcis-10302	133	57	order	order	NOUN
fcis-10302	133	58	gradients	gradient	NOUN
fcis-10302	133	59	of	of	ADP
fcis-10302	133	60	the	the	DET
fcis-10302	133	61	loss	loss	NOUN
fcis-10302	133	62	function	function	NOUN
fcis-10302	133	63	for	for	ADP
fcis-10302	133	64	the	the	DET
fcis-10302	133	65	left	left	ADJ
fcis-10302	133	66	leaf	leaf	NOUN
fcis-10302	133	67	node	node	NOUN
fcis-10302	133	68	,	,	PUNCT
fcis-10302	133	69	λ	λ	PROPN
fcis-10302	133	70	controls	control	VERB
fcis-10302	133	71	the	the	DET
fcis-10302	133	72	score	score	NOUN
fcis-10302	133	73	of	of	ADP
fcis-10302	133	74	the	the	DET
fcis-10302	133	75	leaf	leaf	NOUN
fcis-10302	133	76	node	node	NOUN
fcis-10302	133	77	,	,	PUNCT
fcis-10302	133	78	and	and	CCONJ
fcis-10302	133	79	γ	γ	PROPN
fcis-10302	133	80	controls	control	VERB
fcis-10302	133	81	the	the	DET
fcis-10302	133	82	number	number	NOUN
fcis-10302	133	83	of	of	ADP
fcis-10302	133	84	leaf	leaf	NOUN
fcis-10302	133	85	nodes	node	NOUN
fcis-10302	133	86	.	.	PUNCT
fcis-10302	134	1	by	by	ADP
fcis-10302	134	2	using	use	VERB
fcis-10302	134	3	information	information	NOUN
fcis-10302	134	4	gain	gain	NOUN
fcis-10302	134	5	,	,	PUNCT
fcis-10302	134	6	the	the	DET
fcis-10302	134	7	model	model	NOUN
fcis-10302	134	8	performs	perform	VERB
fcis-10302	134	9	feature	feature	NOUN
fcis-10302	134	10	selection	selection	NOUN
fcis-10302	134	11	.	.	PUNCT
fcis-10302	135	1	a	a	DET
fcis-10302	135	2	higher	high	ADJ
fcis-10302	135	3	information	information	NOUN
fcis-10302	135	4	gain	gain	NOUN
fcis-10302	135	5	indicates	indicate	VERB
fcis-10302	135	6	a	a	DET
fcis-10302	135	7	larger	large	ADJ
fcis-10302	135	8	decrease	decrease	NOUN
fcis-10302	135	9	in	in	ADP
fcis-10302	135	10	loss	loss	NOUN
fcis-10302	135	11	,	,	PUNCT
fcis-10302	135	12	indicating	indicate	VERB
fcis-10302	135	13	a	a	DET
fcis-10302	135	14	better	well	ADJ
fcis-10302	135	15	split	split	NOUN
fcis-10302	135	16	at	at	ADP
fcis-10302	135	17	the	the	DET
fcis-10302	135	18	current	current	ADJ
fcis-10302	135	19	node	node	NOUN
fcis-10302	135	20	.	.	PUNCT
fcis-10302	136	1	by	by	ADP
fcis-10302	136	2	calculating	calculate	VERB
fcis-10302	136	3	the	the	DET
fcis-10302	136	4	information	information	NOUN
fcis-10302	136	5	gain	gain	NOUN
fcis-10302	136	6	for	for	ADP
fcis-10302	136	7	each	each	DET
fcis-10302	136	8	feature	feature	NOUN
fcis-10302	136	9	,	,	PUNCT
fcis-10302	136	10	the	the	DET
fcis-10302	136	11	model	model	NOUN
fcis-10302	136	12	identifies	identify	VERB
fcis-10302	136	13	the	the	DET
fcis-10302	136	14	optimal	optimal	ADJ
fcis-10302	136	15	split	split	NOUN
fcis-10302	136	16	point	point	NOUN
fcis-10302	136	17	that	that	PRON
fcis-10302	136	18	maximizes	maximize	VERB
fcis-10302	136	19	the	the	DET
fcis-10302	136	20	information	information	NOUN
fcis-10302	136	21	gain	gain	NOUN
fcis-10302	136	22	.	.	PUNCT
fcis-10302	137	1	this	this	DET
fcis-10302	137	2	process	process	NOUN
fcis-10302	137	3	iteratively	iteratively	ADV
fcis-10302	137	4	selects	select	VERB
fcis-10302	137	5	features	feature	NOUN
fcis-10302	137	6	,	,	PUNCT
fcis-10302	137	7	retaining	retain	VERB
fcis-10302	137	8	those	those	PRON
fcis-10302	137	9	with	with	ADP
fcis-10302	137	10	the	the	DET
fcis-10302	137	11	highest	high	ADJ
fcis-10302	137	12	discriminative	discriminative	NOUN
fcis-10302	137	13	power	power	NOUN
fcis-10302	137	14	while	while	SCONJ
fcis-10302	137	15	removing	remove	VERB
fcis-10302	137	16	unnecessary	unnecessary	ADJ
fcis-10302	137	17	features	feature	NOUN
fcis-10302	137	18	,	,	PUNCT
fcis-10302	137	19	aiming	aim	VERB
fcis-10302	137	20	to	to	PART
fcis-10302	137	21	improve	improve	VERB
fcis-10302	137	22	classification	classification	NOUN
fcis-10302	137	23	performance	performance	NOUN
fcis-10302	137	24	and	and	CCONJ
fcis-10302	137	25	reduce	reduce	VERB
fcis-10302	137	26	computational	computational	ADJ
fcis-10302	137	27	complexity	complexity	NOUN
fcis-10302	137	28	.	.	PUNCT
fcis-10302	138	1	in	in	ADP
fcis-10302	138	2	this	this	DET
fcis-10302	138	3	study	study	NOUN
fcis-10302	138	4	,	,	PUNCT
fcis-10302	138	5	xgboost	xgboost	X
fcis-10302	138	6	feature	feature	NOUN
fcis-10302	138	7	selection	selection	NOUN
fcis-10302	138	8	is	be	AUX
fcis-10302	138	9	employed	employ	VERB
fcis-10302	138	10	to	to	PART
fcis-10302	138	11	select	select	VERB
fcis-10302	138	12	the	the	DET
fcis-10302	138	13	top	top	ADJ
fcis-10302	138	14	30	30	NUM
fcis-10302	138	15	features	feature	NOUN
fcis-10302	138	16	,	,	PUNCT
fcis-10302	138	17	addressing	address	VERB
fcis-10302	138	18	the	the	DET
fcis-10302	138	19	challenge	challenge	NOUN
fcis-10302	138	20	of	of	ADP
fcis-10302	138	21	complex	complex	ADJ
fcis-10302	138	22	and	and	CCONJ
fcis-10302	138	23	diverse	diverse	ADJ
fcis-10302	138	24	malicious	malicious	ADJ
fcis-10302	138	25	traffic	traffic	NOUN
fcis-10302	138	26	features	feature	NOUN
fcis-10302	138	27	in	in	ADP
fcis-10302	138	28	iot	iot	NOUN
fcis-10302	138	29	.	.	PROPN
fcis-10302	138	30	3.3	3.3	NUM
fcis-10302	138	31	.	.	PUNCT
fcis-10302	139	1	fused	fuse	VERB
fcis-10302	139	2	classifier	classifier	NOUN
fcis-10302	139	3	module	module	NOUN
fcis-10302	139	4	(	(	PUNCT
fcis-10302	139	5	1	1	X
fcis-10302	139	6	)	)	PUNCT
fcis-10302	139	7	cnn	cnn	PROPN
fcis-10302	139	8	-	-	PUNCT
fcis-10302	139	9	gru	gru	PROPN
fcis-10302	139	10	hybrid	hybrid	ADJ
fcis-10302	139	11	deep	deep	ADJ
fcis-10302	139	12	model	model	NOUN
fcis-10302	139	13	:	:	PUNCT
fcis-10302	139	14	the	the	DET
fcis-10302	139	15	basic	basic	ADJ
fcis-10302	139	16	structure	structure	NOUN
fcis-10302	139	17	includes	include	VERB
fcis-10302	139	18	an	an	DET
fcis-10302	139	19	input	input	NOUN
fcis-10302	139	20	layer	layer	NOUN
fcis-10302	139	21	,	,	PUNCT
fcis-10302	139	22	a	a	DET
fcis-10302	139	23	conv1d	conv1d	ADJ
fcis-10302	139	24	layer	layer	NOUN
fcis-10302	139	25	,	,	PUNCT
fcis-10302	139	26	a	a	DET
fcis-10302	139	27	maxpool1d	maxpool1d	ADJ
fcis-10302	139	28	layer	layer	NOUN
fcis-10302	139	29	,	,	PUNCT
fcis-10302	139	30	batch	batch	VERB
fcis-10302	139	31	normalization	normalization	NOUN
fcis-10302	139	32	(	(	PUNCT
fcis-10302	139	33	bn	bn	NOUN
fcis-10302	139	34	)	)	PUNCT
fcis-10302	139	35	layer	layer	NOUN
fcis-10302	139	36	,	,	PUNCT
fcis-10302	139	37	a	a	DET
fcis-10302	139	38	gated	gate	VERB
fcis-10302	139	39	recurrent	recurrent	ADJ
fcis-10302	139	40	unit	unit	NOUN
fcis-10302	139	41	(	(	PUNCT
fcis-10302	139	42	gru	gru	NOUN
fcis-10302	139	43	)	)	PUNCT
fcis-10302	139	44	layer	layer	NOUN
fcis-10302	139	45	,	,	PUNCT
fcis-10302	139	46	a	a	DET
fcis-10302	139	47	flatten	flatten	ADJ
fcis-10302	139	48	layer	layer	NOUN
fcis-10302	139	49	,	,	PUNCT
fcis-10302	139	50	and	and	CCONJ
fcis-10302	139	51	a	a	DET
fcis-10302	139	52	dense	dense	ADJ
fcis-10302	139	53	layer	layer	NOUN
fcis-10302	139	54	.	.	PUNCT
fcis-10302	140	1	输入层	输入层	VERB
fcis-10302	140	2	x	x	X
fcis-10302	140	3	1dcnn	1dcnn	NUM
fcis-10302	140	4	layer	layer	NOUN
fcis-10302	140	5	maxpooling	maxpoole	VERB
fcis-10302	140	6	layer	layer	NOUN
fcis-10302	140	7	1dcnn	1dcnn	NUM
fcis-10302	140	8	layer	layer	NOUN
fcis-10302	140	9	batchnormalization	batchnormalization	NOUN
fcis-10302	140	10	layer	layer	NOUN
fcis-10302	140	11	maxpooling	maxpoole	VERB
fcis-10302	140	12	layer	layer	NOUN
fcis-10302	140	13	gru	gru	NOUN
fcis-10302	140	14	layer	layer	NOUN
fcis-10302	140	15	fatten	fatten	NOUN
fcis-10302	140	16	layer	layer	NOUN
fcis-10302	140	17	dense	dense	ADJ
fcis-10302	140	18	layer	layer	NOUN
fcis-10302	140	19	分类结果	分类结果	ADJ
fcis-10302	140	20	figure	figure	NOUN
fcis-10302	140	21	2	2	NUM
fcis-10302	140	22	.	.	X
fcis-10302	140	23	cnn	cnn	PROPN
fcis-10302	140	24	-	-	PUNCT
fcis-10302	140	25	gru	gru	PROPN
fcis-10302	140	26	model	model	NOUN
fcis-10302	140	27	firstly	firstly	ADV
fcis-10302	140	28	,	,	PUNCT
fcis-10302	140	29	each	each	DET
fcis-10302	140	30	selected	select	VERB
fcis-10302	140	31	traffic	traffic	NOUN
fcis-10302	140	32	feature	feature	NOUN
fcis-10302	140	33	is	be	AUX
fcis-10302	140	34	transformed	transform	VERB
fcis-10302	140	35	into	into	ADP
fcis-10302	140	36	a	a	DET
fcis-10302	140	37	(	(	PUNCT
fcis-10302	140	38	30,1	30,1	NOUN
fcis-10302	140	39	)	)	PUNCT
fcis-10302	140	40	tensor	tensor	NOUN
fcis-10302	140	41	and	and	CCONJ
fcis-10302	140	42	input	input	NOUN
fcis-10302	140	43	to	to	ADP
fcis-10302	140	44	the	the	DET
fcis-10302	140	45	input	input	NOUN
fcis-10302	140	46	layer	layer	NOUN
fcis-10302	140	47	.	.	PUNCT
fcis-10302	141	1	two	two	NUM
fcis-10302	141	2	layers	layer	NOUN
fcis-10302	141	3	of	of	ADP
fcis-10302	141	4	1d	1d	NUM
fcis-10302	141	5	convolutions	convolution	NOUN
fcis-10302	141	6	are	be	AUX
fcis-10302	141	7	applied	apply	VERB
fcis-10302	141	8	to	to	PART
fcis-10302	141	9	extract	extract	VERB
fcis-10302	141	10	features	feature	NOUN
fcis-10302	141	11	from	from	ADP
fcis-10302	141	12	iot	iot	ADJ
fcis-10302	141	13	traffic	traffic	NOUN
fcis-10302	141	14	,	,	PUNCT
fcis-10302	141	15	with	with	ADP
fcis-10302	141	16	multiple	multiple	ADJ
fcis-10302	141	17	channels	channel	NOUN
fcis-10302	141	18	set	set	VERB
fcis-10302	141	19	to	to	PART
fcis-10302	141	20	capture	capture	VERB
fcis-10302	141	21	multi	multi	ADJ
fcis-10302	141	22	-	-	ADJ
fcis-10302	141	23	dimensional	dimensional	ADJ
fcis-10302	141	24	features	feature	NOUN
fcis-10302	141	25	.	.	PUNCT
fcis-10302	142	1	the	the	DET
fcis-10302	142	2	trained	train	VERB
fcis-10302	142	3	features	feature	NOUN
fcis-10302	142	4	are	be	AUX
fcis-10302	142	5	then	then	ADV
fcis-10302	142	6	passed	pass	VERB
fcis-10302	142	7	to	to	ADP
fcis-10302	142	8	the	the	DET
fcis-10302	142	9	gru	gru	NOUN
fcis-10302	142	10	layer	layer	NOUN
fcis-10302	142	11	,	,	PUNCT
fcis-10302	142	12	which	which	PRON
fcis-10302	142	13	uses	use	VERB
fcis-10302	142	14	the	the	DET
fcis-10302	142	15	tanh	tanh	PROPN
fcis-10302	142	16	activation	activation	NOUN
fcis-10302	142	17	function	function	NOUN
fcis-10302	142	18	with	with	ADP
fcis-10302	142	19	64	64	NUM
fcis-10302	142	20	neurons	neuron	NOUN
fcis-10302	142	21	.	.	PUNCT
fcis-10302	143	1	finally	finally	ADV
fcis-10302	143	2	,	,	PUNCT
fcis-10302	143	3	the	the	DET
fcis-10302	143	4	flatten	flatten	ADJ
fcis-10302	143	5	layer	layer	NOUN
fcis-10302	143	6	flattens	flatten	VERB
fcis-10302	143	7	the	the	DET
fcis-10302	143	8	2d	2d	NUM
fcis-10302	143	9	feature	feature	NOUN
fcis-10302	143	10	maps	map	NOUN
fcis-10302	143	11	,	,	PUNCT
fcis-10302	143	12	and	and	CCONJ
fcis-10302	143	13	the	the	DET
fcis-10302	143	14	features	feature	NOUN
fcis-10302	143	15	are	be	AUX
fcis-10302	143	16	fed	feed	VERB
fcis-10302	143	17	into	into	ADP
fcis-10302	143	18	the	the	DET
fcis-10302	143	19	dense	dense	ADJ
fcis-10302	143	20	layer	layer	NOUN
fcis-10302	143	21	for	for	ADP
fcis-10302	143	22	classification	classification	NOUN
fcis-10302	143	23	.	.	PUNCT
fcis-10302	144	1	93	93	NUM
fcis-10302	144	2	the	the	DET
fcis-10302	144	3	commonly	commonly	ADV
fcis-10302	144	4	used	use	VERB
fcis-10302	144	5	activation	activation	NOUN
fcis-10302	144	6	functions	function	NOUN
fcis-10302	144	7	for	for	ADP
fcis-10302	144	8	classification	classification	NOUN
fcis-10302	144	9	problems	problem	NOUN
fcis-10302	144	10	are	be	AUX
fcis-10302	144	11	sigmoid	sigmoid	NOUN
fcis-10302	144	12	and	and	CCONJ
fcis-10302	144	13	softmax	softmax	NOUN
fcis-10302	144	14	.	.	PUNCT
fcis-10302	145	1	in	in	ADP
fcis-10302	145	2	this	this	DET
fcis-10302	145	3	model	model	NOUN
fcis-10302	145	4	,	,	PUNCT
fcis-10302	145	5	only	only	ADV
fcis-10302	145	6	the	the	DET
fcis-10302	145	7	softmax	softmax	NOUN
fcis-10302	145	8	activation	activation	NOUN
fcis-10302	145	9	function	function	NOUN
fcis-10302	145	10	is	be	AUX
fcis-10302	145	11	used	use	VERB
fcis-10302	145	12	,	,	PUNCT
fcis-10302	145	13	with	with	ADP
fcis-10302	145	14	the	the	DET
fcis-10302	145	15	number	number	NOUN
fcis-10302	145	16	of	of	ADP
fcis-10302	145	17	neurons	neuron	NOUN
fcis-10302	145	18	in	in	ADP
fcis-10302	145	19	this	this	DET
fcis-10302	145	20	layer	layer	NOUN
fcis-10302	145	21	set	set	VERB
fcis-10302	145	22	to	to	ADP
fcis-10302	145	23	10	10	NUM
fcis-10302	145	24	.	.	PUNCT
fcis-10302	146	1	to	to	PART
fcis-10302	146	2	improve	improve	VERB
fcis-10302	146	3	the	the	DET
fcis-10302	146	4	accuracy	accuracy	NOUN
fcis-10302	146	5	of	of	ADP
fcis-10302	146	6	the	the	DET
fcis-10302	146	7	model	model	NOUN
fcis-10302	146	8	,	,	PUNCT
fcis-10302	146	9	various	various	ADJ
fcis-10302	146	10	parameters	parameter	NOUN
fcis-10302	146	11	such	such	ADJ
fcis-10302	146	12	as	as	ADP
fcis-10302	146	13	learning	learn	VERB
fcis-10302	146	14	rate	rate	NOUN
fcis-10302	146	15	and	and	CCONJ
fcis-10302	146	16	weights	weight	NOUN
fcis-10302	146	17	need	need	VERB
fcis-10302	146	18	to	to	PART
fcis-10302	146	19	be	be	AUX
fcis-10302	146	20	adjusted	adjust	VERB
fcis-10302	146	21	.	.	PUNCT
fcis-10302	147	1	techniques	technique	NOUN
fcis-10302	147	2	such	such	ADJ
fcis-10302	147	3	as	as	ADP
fcis-10302	147	4	l1	l1	PROPN
fcis-10302	147	5	regularization	regularization	NOUN
fcis-10302	147	6	,	,	PUNCT
fcis-10302	147	7	l2	l2	NOUN
fcis-10302	147	8	regularization	regularization	NOUN
fcis-10302	147	9	,	,	PUNCT
fcis-10302	147	10	and	and	CCONJ
fcis-10302	147	11	batch	batch	NOUN
fcis-10302	147	12	normalization	normalization	NOUN
fcis-10302	147	13	can	can	AUX
fcis-10302	147	14	be	be	AUX
fcis-10302	147	15	used	use	VERB
fcis-10302	147	16	to	to	PART
fcis-10302	147	17	prevent	prevent	VERB
fcis-10302	147	18	overfitting	overfitting	NOUN
fcis-10302	147	19	.	.	PUNCT
fcis-10302	148	1	deep	deep	ADJ
fcis-10302	148	2	neural	neural	ADJ
fcis-10302	148	3	networks	network	NOUN
fcis-10302	148	4	have	have	VERB
fcis-10302	148	5	a	a	DET
fcis-10302	148	6	large	large	ADJ
fcis-10302	148	7	number	number	NOUN
fcis-10302	148	8	of	of	ADP
fcis-10302	148	9	parameters	parameter	NOUN
fcis-10302	148	10	,	,	PUNCT
fcis-10302	148	11	making	make	VERB
fcis-10302	148	12	the	the	DET
fcis-10302	148	13	model	model	NOUN
fcis-10302	148	14	complex	complex	NOUN
fcis-10302	148	15	and	and	CCONJ
fcis-10302	148	16	prone	prone	ADJ
fcis-10302	148	17	to	to	ADP
fcis-10302	148	18	overfitting	overfitte	VERB
fcis-10302	148	19	.	.	PUNCT
fcis-10302	149	1	the	the	DET
fcis-10302	149	2	use	use	NOUN
fcis-10302	149	3	of	of	ADP
fcis-10302	149	4	batch	batch	NOUN
fcis-10302	149	5	normalization	normalization	NOUN
fcis-10302	149	6	helps	help	VERB
fcis-10302	149	7	to	to	PART
fcis-10302	149	8	mitigate	mitigate	VERB
fcis-10302	149	9	the	the	DET
fcis-10302	149	10	occurrence	occurrence	NOUN
fcis-10302	149	11	of	of	ADP
fcis-10302	149	12	internal	internal	ADJ
fcis-10302	149	13	covariate	covariate	ADJ
fcis-10302	149	14	shift	shift	NOUN
fcis-10302	149	15	.	.	PUNCT
fcis-10302	150	1	by	by	ADP
fcis-10302	150	2	integrating	integrate	VERB
fcis-10302	150	3	the	the	DET
fcis-10302	150	4	convolutional	convolutional	ADJ
fcis-10302	150	5	features	feature	NOUN
fcis-10302	150	6	with	with	ADP
fcis-10302	150	7	the	the	DET
fcis-10302	150	8	batch	batch	NOUN
fcis-10302	150	9	normalization	normalization	NOUN
fcis-10302	150	10	layer	layer	NOUN
fcis-10302	150	11	,	,	PUNCT
fcis-10302	150	12	overfitting	overfitte	VERB
fcis-10302	150	13	of	of	ADP
fcis-10302	150	14	the	the	DET
fcis-10302	150	15	data	datum	NOUN
fcis-10302	150	16	is	be	AUX
fcis-10302	150	17	avoided	avoid	VERB
fcis-10302	150	18	.	.	PUNCT
fcis-10302	151	1	4	4	X
fcis-10302	151	2	.	.	X
fcis-10302	151	3	experimental	experimental	ADJ
fcis-10302	151	4	results	result	NOUN
fcis-10302	151	5	4.1	4.1	NUM
fcis-10302	151	6	.	.	PUNCT
fcis-10302	152	1	dataset	dataset	NOUN
fcis-10302	152	2	and	and	CCONJ
fcis-10302	152	3	data	datum	NOUN
fcis-10302	152	4	preprocessing	preprocesse	VERB
fcis-10302	152	5	the	the	DET
fcis-10302	152	6	experiment	experiment	NOUN
fcis-10302	152	7	utilizes	utilize	VERB
fcis-10302	152	8	the	the	DET
fcis-10302	152	9	n	n	NUM
fcis-10302	152	10	-	-	PUNCT
fcis-10302	152	11	baiot	baiot	NOUN
fcis-10302	152	12	dataset	dataset	NOUN
fcis-10302	152	13	created	create	VERB
fcis-10302	152	14	by	by	ADP
fcis-10302	152	15	meidan	meidan	PROPN
fcis-10302	152	16	et	et	PROPN
fcis-10302	152	17	al	al	PROPN
fcis-10302	152	18	.	.	PUNCT
fcis-10302	153	1	[	[	X
fcis-10302	153	2	14	14	NUM
fcis-10302	153	3	]	]	PUNCT
fcis-10302	153	4	.	.	PUNCT
fcis-10302	154	1	the	the	DET
fcis-10302	154	2	dataset	dataset	NOUN
fcis-10302	154	3	captures	capture	VERB
fcis-10302	154	4	traffic	traffic	NOUN
fcis-10302	154	5	by	by	ADP
fcis-10302	154	6	monitoring	monitor	VERB
fcis-10302	154	7	devices	device	NOUN
fcis-10302	154	8	during	during	ADP
fcis-10302	154	9	bashlite	bashlite	NOUN
fcis-10302	154	10	and	and	CCONJ
fcis-10302	154	11	mirai	mirai	PROPN
fcis-10302	154	12	botnet	botnet	NOUN
fcis-10302	154	13	attacks	attack	NOUN
fcis-10302	154	14	,	,	PUNCT
fcis-10302	154	15	with	with	ADP
fcis-10302	154	16	traffic	traffic	NOUN
fcis-10302	154	17	packets	packet	NOUN
fcis-10302	154	18	captured	capture	VERB
fcis-10302	154	19	using	use	VERB
fcis-10302	154	20	wireshark	wireshark	NOUN
fcis-10302	154	21	.	.	PUNCT
fcis-10302	155	1	a	a	DET
fcis-10302	155	2	total	total	NOUN
fcis-10302	155	3	of	of	ADP
fcis-10302	155	4	115	115	NUM
fcis-10302	155	5	features	feature	NOUN
fcis-10302	155	6	were	be	AUX
fcis-10302	155	7	extracted	extract	VERB
fcis-10302	155	8	from	from	ADP
fcis-10302	155	9	each	each	DET
fcis-10302	155	10	device	device	NOUN
fcis-10302	155	11	,	,	PUNCT
fcis-10302	155	12	collected	collect	VERB
fcis-10302	155	13	considering	consider	VERB
fcis-10302	155	14	fivetime	fivetime	ADJ
fcis-10302	155	15	intervals	interval	NOUN
fcis-10302	155	16	:	:	PUNCT
fcis-10302	155	17	1	1	NUM
fcis-10302	155	18	minute	minute	NOUN
fcis-10302	155	19	,	,	PUNCT
fcis-10302	155	20	10	10	NUM
fcis-10302	155	21	seconds	second	NOUN
fcis-10302	155	22	,	,	PUNCT
fcis-10302	155	23	1.5	1.5	NUM
fcis-10302	155	24	seconds	second	NOUN
fcis-10302	155	25	,	,	PUNCT
fcis-10302	155	26	500	500	NUM
fcis-10302	155	27	milliseconds	millisecond	NOUN
fcis-10302	155	28	,	,	PUNCT
fcis-10302	155	29	and	and	CCONJ
fcis-10302	155	30	100	100	NUM
fcis-10302	155	31	milliseconds	millisecond	NOUN
fcis-10302	155	32	(	(	PUNCT
fcis-10302	155	33	referred	refer	VERB
fcis-10302	155	34	to	to	ADP
fcis-10302	155	35	as	as	ADP
fcis-10302	155	36	l0.01	l0.01	PROPN
fcis-10302	155	37	,	,	PUNCT
fcis-10302	155	38	l0.1	l0.1	PROPN
fcis-10302	155	39	,	,	PUNCT
fcis-10302	155	40	l1	l1	PROPN
fcis-10302	155	41	,	,	PUNCT
fcis-10302	155	42	l3	l3	PROPN
fcis-10302	155	43	,	,	PUNCT
fcis-10302	155	44	and	and	CCONJ
fcis-10302	155	45	l5	l5	PROPN
fcis-10302	155	46	,	,	PUNCT
fcis-10302	155	47	respectively	respectively	ADV
fcis-10302	155	48	)	)	PUNCT
fcis-10302	155	49	.	.	PUNCT
fcis-10302	156	1	for	for	ADP
fcis-10302	156	2	example	example	NOUN
fcis-10302	156	3	,	,	PUNCT
fcis-10302	156	4	the	the	DET
fcis-10302	156	5	feature	feature	NOUN
fcis-10302	156	6	value	value	NOUN
fcis-10302	156	7	h_l5_mean	h_l5_mean	VERB
fcis-10302	156	8	represents	represent	VERB
fcis-10302	156	9	the	the	DET
fcis-10302	156	10	average	average	ADJ
fcis-10302	156	11	statistical	statistical	ADJ
fcis-10302	156	12	feature	feature	NOUN
fcis-10302	156	13	obtained	obtain	VERB
fcis-10302	156	14	within	within	ADP
fcis-10302	156	15	a	a	DET
fcis-10302	156	16	100	100	NUM
fcis-10302	156	17	-	-	PUNCT
fcis-10302	156	18	millisecond	millisecond	NOUN
fcis-10302	156	19	timeframe	timeframe	NOUN
fcis-10302	156	20	of	of	ADP
fcis-10302	156	21	host	host	NOUN
fcis-10302	156	22	ip	ip	NOUN
fcis-10302	156	23	traffic	traffic	NOUN
fcis-10302	156	24	data	datum	NOUN
fcis-10302	156	25	.	.	PUNCT
fcis-10302	157	1	for	for	ADP
fcis-10302	157	2	the	the	DET
fcis-10302	157	3	collected	collect	VERB
fcis-10302	157	4	iot	iot	NOUN
fcis-10302	157	5	traffic	traffic	NOUN
fcis-10302	157	6	,	,	PUNCT
fcis-10302	157	7	categorical	categorical	ADJ
fcis-10302	157	8	features	feature	NOUN
fcis-10302	157	9	are	be	AUX
fcis-10302	157	10	numerically	numerically	ADV
fcis-10302	157	11	encoded	encode	VERB
fcis-10302	157	12	using	use	VERB
fcis-10302	157	13	one	one	NUM
fcis-10302	157	14	-	-	PUNCT
fcis-10302	157	15	hot	hot	ADJ
fcis-10302	157	16	encoding	encoding	NOUN
fcis-10302	157	17	,	,	PUNCT
fcis-10302	157	18	and	and	CCONJ
fcis-10302	157	19	the	the	DET
fcis-10302	157	20	feature	feature	NOUN
fcis-10302	157	21	data	datum	NOUN
fcis-10302	157	22	is	be	AUX
fcis-10302	157	23	standardized	standardize	VERB
fcis-10302	157	24	using	use	VERB
fcis-10302	157	25	minmax	minmax	PROPN
fcis-10302	157	26	to	to	PART
fcis-10302	157	27	eliminate	eliminate	VERB
fcis-10302	157	28	the	the	DET
fcis-10302	157	29	scale	scale	NOUN
fcis-10302	157	30	differences	difference	NOUN
fcis-10302	157	31	between	between	ADP
fcis-10302	157	32	features	feature	NOUN
fcis-10302	157	33	,	,	PUNCT
fcis-10302	157	34	resulting	result	VERB
fcis-10302	157	35	in	in	ADP
fcis-10302	157	36	a	a	DET
fcis-10302	157	37	high	high	ADV
fcis-10302	157	38	-	-	PUNCT
fcis-10302	157	39	dimensional	dimensional	ADJ
fcis-10302	157	40	and	and	CCONJ
fcis-10302	157	41	standardized	standardized	ADJ
fcis-10302	157	42	dataset	dataset	NOUN
fcis-10302	157	43	.	.	PUNCT
fcis-10302	158	1	table	table	NOUN
fcis-10302	158	2	1	1	NUM
fcis-10302	158	3	.	.	PUNCT
fcis-10302	158	4	description	description	NOUN
fcis-10302	158	5	of	of	ADP
fcis-10302	158	6	specific	specific	ADJ
fcis-10302	158	7	statistical	statistical	ADJ
fcis-10302	158	8	features	feature	NOUN
fcis-10302	158	9	on	on	ADP
fcis-10302	158	10	the	the	DET
fcis-10302	158	11	nbaiot	nbaiot	ADJ
fcis-10302	158	12	dataset	dataset	NOUN
fcis-10302	158	13	feature	feature	NOUN
fcis-10302	158	14	type	type	NOUN
fcis-10302	158	15	abbreviation	abbreviation	NOUN
fcis-10302	158	16	details	detail	NOUN
fcis-10302	158	17	host	host	NOUN
fcis-10302	158	18	-	-	PUNCT
fcis-10302	158	19	ip	ip	NOUN
fcis-10302	158	20	h	h	NOUN
fcis-10302	158	21	traffic	traffic	NOUN
fcis-10302	158	22	data	datum	NOUN
fcis-10302	158	23	from	from	ADP
fcis-10302	158	24	specific	specific	ADJ
fcis-10302	158	25	internet	internet	NOUN
fcis-10302	158	26	protocol	protocol	NOUN
fcis-10302	158	27	addresses	address	NOUN
fcis-10302	158	28	hostmac&ip	hostmac&ip	NUM
fcis-10302	158	29	mi	mi	NOUN
fcis-10302	158	30	traffic	traffic	NOUN
fcis-10302	158	31	data	datum	NOUN
fcis-10302	158	32	from	from	ADP
fcis-10302	158	33	specific	specific	ADJ
fcis-10302	158	34	internet	internet	NOUN
fcis-10302	158	35	protocol	protocol	NOUN
fcis-10302	158	36	and	and	CCONJ
fcis-10302	158	37	media	medium	NOUN
fcis-10302	158	38	access	access	NOUN
fcis-10302	158	39	control	control	NOUN
fcis-10302	158	40	(	(	PUNCT
fcis-10302	158	41	mac	mac	NOUN
fcis-10302	158	42	)	)	PUNCT
fcis-10302	158	43	addresses	address	NOUN
fcis-10302	158	44	channel	channel	VERB
fcis-10302	158	45	hh	hh	PROPN
fcis-10302	158	46	traffic	traffic	NOUN
fcis-10302	158	47	collected	collect	VERB
fcis-10302	158	48	between	between	ADP
fcis-10302	158	49	specific	specific	ADJ
fcis-10302	158	50	hosts	host	NOUN
fcis-10302	158	51	socket	socket	VERB
fcis-10302	158	52	hphp	hphp	ADJ
fcis-10302	158	53	traffic	traffic	NOUN
fcis-10302	158	54	collected	collect	VERB
fcis-10302	158	55	between	between	ADP
fcis-10302	158	56	specific	specific	ADJ
fcis-10302	158	57	hosts	host	NOUN
fcis-10302	158	58	with	with	ADP
fcis-10302	158	59	port	port	NOUN
fcis-10302	158	60	information	information	NOUN
fcis-10302	158	61	network	network	NOUN
fcis-10302	158	62	jitter	jitter	NOUN
fcis-10302	158	63	hh_jit	hh_jit	PROPN
fcis-10302	158	64	statistical	statistical	ADJ
fcis-10302	158	65	values	value	NOUN
fcis-10302	158	66	of	of	ADP
fcis-10302	158	67	features	feature	NOUN
fcis-10302	158	68	related	relate	VERB
fcis-10302	158	69	to	to	ADP
fcis-10302	158	70	time	time	NOUN
fcis-10302	158	71	intervals	interval	NOUN
fcis-10302	158	72	between	between	ADP
fcis-10302	158	73	received	receive	VERB
fcis-10302	158	74	data	data	NOUN
fcis-10302	158	75	packets	packet	NOUN
fcis-10302	158	76	in	in	ADP
fcis-10302	158	77	the	the	DET
fcis-10302	158	78	channel	channel	NOUN
fcis-10302	158	79	category	category	NOUN
fcis-10302	158	80	4.2	4.2	NUM
fcis-10302	158	81	.	.	PUNCT
fcis-10302	159	1	feature	feature	NOUN
fcis-10302	159	2	selection	selection	NOUN
fcis-10302	159	3	the	the	DET
fcis-10302	159	4	xgboost	xgboost	PROPN
fcis-10302	159	5	algorithm	algorithm	PROPN
fcis-10302	159	6	is	be	AUX
fcis-10302	159	7	tested	test	VERB
fcis-10302	159	8	through	through	ADP
fcis-10302	159	9	extensive	extensive	ADJ
fcis-10302	159	10	experiments	experiment	NOUN
fcis-10302	159	11	,	,	PUNCT
fcis-10302	159	12	which	which	PRON
fcis-10302	159	13	include	include	VERB
fcis-10302	159	14	multiple	multiple	ADJ
fcis-10302	159	15	thresholds	threshold	NOUN
fcis-10302	159	16	for	for	ADP
fcis-10302	159	17	feature	feature	NOUN
fcis-10302	159	18	selection	selection	NOUN
fcis-10302	159	19	based	base	VERB
fcis-10302	159	20	on	on	ADP
fcis-10302	159	21	feature	feature	NOUN
fcis-10302	159	22	importance	importance	NOUN
fcis-10302	159	23	scores	score	NOUN
fcis-10302	159	24	.	.	PUNCT
fcis-10302	160	1	for	for	ADP
fcis-10302	160	2	each	each	DET
fcis-10302	160	3	input	input	NOUN
fcis-10302	160	4	variable	variable	NOUN
fcis-10302	160	5	,	,	PUNCT
fcis-10302	160	6	the	the	DET
fcis-10302	160	7	feature	feature	NOUN
fcis-10302	160	8	importance	importance	NOUN
fcis-10302	160	9	scores	score	NOUN
fcis-10302	160	10	allow	allow	VERB
fcis-10302	160	11	testing	testing	NOUN
fcis-10302	160	12	of	of	ADP
fcis-10302	160	13	subsets	subset	NOUN
fcis-10302	160	14	of	of	ADP
fcis-10302	160	15	features	feature	NOUN
fcis-10302	160	16	based	base	VERB
fcis-10302	160	17	on	on	ADP
fcis-10302	160	18	their	their	PRON
fcis-10302	160	19	importance	importance	NOUN
fcis-10302	160	20	.	.	PUNCT
fcis-10302	161	1	initially	initially	ADV
fcis-10302	161	2	,	,	PUNCT
fcis-10302	161	3	all	all	DET
fcis-10302	161	4	115	115	NUM
fcis-10302	161	5	features	feature	NOUN
fcis-10302	161	6	are	be	AUX
fcis-10302	161	7	used	use	VERB
fcis-10302	161	8	,	,	PUNCT
fcis-10302	161	9	and	and	CCONJ
fcis-10302	161	10	subsets	subset	NOUN
fcis-10302	161	11	containing	contain	VERB
fcis-10302	161	12	the	the	DET
fcis-10302	161	13	essential	essential	ADJ
fcis-10302	161	14	features	feature	NOUN
fcis-10302	161	15	are	be	AUX
fcis-10302	161	16	obtained	obtain	VERB
fcis-10302	161	17	.	.	PUNCT
fcis-10302	162	1	the	the	DET
fcis-10302	162	2	model	model	NOUN
fcis-10302	162	3	's	's	PART
fcis-10302	162	4	performance	performance	NOUN
fcis-10302	162	5	tends	tend	VERB
fcis-10302	162	6	to	to	PART
fcis-10302	162	7	overfit	overfit	VERB
fcis-10302	162	8	and	and	CCONJ
fcis-10302	162	9	results	result	NOUN
fcis-10302	162	10	in	in	ADP
fcis-10302	162	11	decreased	decrease	VERB
fcis-10302	162	12	prediction	prediction	NOUN
fcis-10302	162	13	accuracy	accuracy	NOUN
fcis-10302	162	14	as	as	ADP
fcis-10302	162	15	the	the	DET
fcis-10302	162	16	number	number	NOUN
fcis-10302	162	17	of	of	ADP
fcis-10302	162	18	selected	select	VERB
fcis-10302	162	19	features	feature	NOUN
fcis-10302	162	20	increases	increase	NOUN
fcis-10302	162	21	,	,	PUNCT
fcis-10302	162	22	creating	create	VERB
fcis-10302	162	23	a	a	DET
fcis-10302	162	24	trade	trade	NOUN
fcis-10302	162	25	-	-	PUNCT
fcis-10302	162	26	off	off	NOUN
fcis-10302	162	27	between	between	ADP
fcis-10302	162	28	the	the	DET
fcis-10302	162	29	number	number	NOUN
fcis-10302	162	30	of	of	ADP
fcis-10302	162	31	features	feature	NOUN
fcis-10302	162	32	and	and	CCONJ
fcis-10302	162	33	test	test	NOUN
fcis-10302	162	34	accuracy	accuracy	NOUN
fcis-10302	162	35	.	.	PUNCT
fcis-10302	163	1	therefore	therefore	ADV
fcis-10302	163	2	,	,	PUNCT
fcis-10302	163	3	the	the	DET
fcis-10302	163	4	optimal	optimal	ADJ
fcis-10302	163	5	combination	combination	NOUN
fcis-10302	163	6	of	of	ADP
fcis-10302	163	7	30	30	NUM
fcis-10302	163	8	features	feature	NOUN
fcis-10302	163	9	is	be	AUX
fcis-10302	163	10	selected	select	VERB
fcis-10302	163	11	from	from	ADP
fcis-10302	163	12	the	the	DET
fcis-10302	163	13	initial	initial	ADJ
fcis-10302	163	14	set	set	NOUN
fcis-10302	163	15	of	of	ADP
fcis-10302	163	16	115	115	NUM
fcis-10302	163	17	features	feature	NOUN
fcis-10302	163	18	.	.	PUNCT
fcis-10302	164	1	the	the	DET
fcis-10302	164	2	detailed	detailed	ADJ
fcis-10302	164	3	information	information	NOUN
fcis-10302	164	4	of	of	ADP
fcis-10302	164	5	the	the	DET
fcis-10302	164	6	selected	select	VERB
fcis-10302	164	7	features	feature	NOUN
fcis-10302	164	8	is	be	AUX
fcis-10302	164	9	provided	provide	VERB
fcis-10302	164	10	in	in	ADP
fcis-10302	164	11	table	table	NOUN
fcis-10302	164	12	2	2	NUM
fcis-10302	164	13	.	.	PUNCT
fcis-10302	165	1	these	these	DET
fcis-10302	165	2	selected	select	VERB
fcis-10302	165	3	features	feature	NOUN
fcis-10302	165	4	are	be	AUX
fcis-10302	165	5	then	then	ADV
fcis-10302	165	6	fed	feed	VERB
fcis-10302	165	7	into	into	ADP
fcis-10302	165	8	the	the	DET
fcis-10302	165	9	cnn	cnn	PROPN
fcis-10302	165	10	-	-	PUNCT
fcis-10302	165	11	gru	gru	PROPN
fcis-10302	165	12	model	model	NOUN
fcis-10302	165	13	for	for	ADP
fcis-10302	165	14	the	the	DET
fcis-10302	165	15	classification	classification	NOUN
fcis-10302	165	16	of	of	ADP
fcis-10302	165	17	iot	iot	ADJ
fcis-10302	165	18	malicious	malicious	ADJ
fcis-10302	165	19	traffic	traffic	NOUN
fcis-10302	165	20	.	.	PUNCT
fcis-10302	166	1	table	table	NOUN
fcis-10302	166	2	2	2	NUM
fcis-10302	166	3	.	.	X
fcis-10302	167	1	top	top	ADJ
fcis-10302	167	2	30	30	NUM
fcis-10302	167	3	features	feature	NOUN
fcis-10302	167	4	selected	select	VERB
fcis-10302	167	5	from	from	ADP
fcis-10302	167	6	the	the	DET
fcis-10302	167	7	n	n	NUM
fcis-10302	167	8	-	-	PUNCT
fcis-10302	167	9	baiot	baiot	NOUN
fcis-10302	167	10	dataset	dataset	NOUN
fcis-10302	167	11	of	of	ADP
fcis-10302	167	12	baby	baby	NOUN
fcis-10302	167	13	monitor	monitor	NOUN
fcis-10302	167	14	devices	device	NOUN
fcis-10302	167	15	no	no	PROPN
fcis-10302	167	16	.	.	PUNCT
fcis-10302	167	17	feature	feature	PROPN
fcis-10302	167	18	index	index	NOUN
fcis-10302	167	19	no	no	PROPN
fcis-10302	167	20	.	.	PUNCT
fcis-10302	168	1	feature	feature	PROPN
fcis-10302	168	2	index	index	NOUN
fcis-10302	168	3	1	1	NUM
fcis-10302	168	4	hphp_l0.01_weight	hphp_l0.01_weight	VERB
fcis-10302	168	5	89	89	NUM
fcis-10302	168	6	16	16	NUM
fcis-10302	168	7	mi_dir_l3_mean	mi_dir_l3_mean	ADJ
fcis-10302	168	8	4	4	NUM
fcis-10302	168	9	2	2	NUM
fcis-10302	168	10	hh_l5_magnitude	hh_l5_magnitude	NOUN
fcis-10302	168	11	21	21	NUM
fcis-10302	168	12	17	17	NUM
fcis-10302	168	13	mi_dir_lo.1_variance	mi_dir_lo.1_variance	NOUN
fcis-10302	168	14	11	11	NUM
fcis-10302	168	15	3	3	NUM
fcis-10302	168	16	mi_dir_l5_variance	mi_dir_l5_variance	NOUN
fcis-10302	168	17	2	2	NUM
fcis-10302	168	18	18	18	NUM
fcis-10302	168	19	hh_l1_magnitude	hh_l1_magnitude	NOUN
fcis-10302	168	20	35	35	NUM
fcis-10302	168	21	4	4	NUM
fcis-10302	168	22	hh_l0.1_covariance	hh_l0.1_covariance	NOUN
fcis-10302	168	23	44	44	NUM
fcis-10302	168	24	19	19	NUM
fcis-10302	168	25	mi_dir_l3_weight	mi_dir_l3_weight	NOUN
fcis-10302	168	26	3	3	NUM
fcis-10302	168	27	5	5	NUM
fcis-10302	168	28	mi_dir_l3_variance	mi_dir_l3_variance	NOUN
fcis-10302	168	29	5	5	NUM
fcis-10302	168	30	20	20	NUM
fcis-10302	168	31	mi_dir_l1_weight	mi_dir_l1_weight	NOUN
fcis-10302	168	32	6	6	NUM
fcis-10302	168	33	6	6	NUM
fcis-10302	168	34	mi_dir_l0.01_variance	mi_dir_l0.01_variance	NOUN
fcis-10302	168	35	14	14	NUM
fcis-10302	168	36	21	21	NUM
fcis-10302	168	37	hh_l0.01_covariance	hh_l0.01_covariance	PROPN
fcis-10302	168	38	51	51	NUM
fcis-10302	168	39	7	7	NUM
fcis-10302	168	40	mi_dir_l0.01_mean	mi_dir_l0.01_mean	NOUN
fcis-10302	168	41	13	13	NUM
fcis-10302	168	42	22	22	NUM
fcis-10302	169	1	hh_jit_l1_variance	hh_jit_l1_variance	NOUN
fcis-10302	169	2	58	58	NUM
fcis-10302	169	3	8	8	NUM
fcis-10302	169	4	mi_dir_l0.1_weight	mi_dir_l0.1_weight	NOUN
fcis-10302	169	5	9	9	NUM
fcis-10302	169	6	23	23	NUM
fcis-10302	169	7	mi_dir_l5_weight	mi_dir_l5_weight	NOUN
fcis-10302	169	8	0	0	NUM
fcis-10302	169	9	9	9	NUM
fcis-10302	169	10	mi_dir_l1_mean	mi_dir_l1_mean	NOUN
fcis-10302	169	11	7	7	NUM
fcis-10302	169	12	24	24	NUM
fcis-10302	169	13	hh_l0.01_magnitude	hh_l0.01_magnitude	NUM
fcis-10302	169	14	49	49	NUM
fcis-10302	169	15	10	10	NUM
fcis-10302	169	16	hh_l0.1_magnitude	hh_l0.1_magnitude	NOUN
fcis-10302	169	17	42	42	NUM
fcis-10302	169	18	25	25	NUM
fcis-10302	169	19	hh_l0.01_pcc	hh_l0.01_pcc	PROPN
fcis-10302	169	20	52	52	NUM
fcis-10302	169	21	11	11	NUM
fcis-10302	169	22	mi_dir_l0.1_mean	mi_dir_l0.1_mean	NOUN
fcis-10302	169	23	10	10	NUM
fcis-10302	169	24	26	26	NUM
fcis-10302	169	25	hphp_l3_std	hphp_l3_std	NOUN
fcis-10302	169	26	72	72	NUM
fcis-10302	169	27	12	12	NUM
fcis-10302	169	28	hh_l3_magnitude	hh_l3_magnitude	PROPN
fcis-10302	169	29	28	28	NUM
fcis-10302	169	30	27	27	NUM
fcis-10302	169	31	hh_l1_weight	hh_l1_weight	ADJ
fcis-10302	169	32	32	32	NUM
fcis-10302	169	33	13	13	NUM
fcis-10302	169	34	hh_jit_l0.01_mean	hh_jit_l0.01_mean	NOUN
fcis-10302	169	35	61	61	NUM
fcis-10302	169	36	28	28	NUM
fcis-10302	169	37	hh_l5_mean	hh_l5_mean	PROPN
fcis-10302	169	38	19	19	NUM
fcis-10302	169	39	14	14	NUM
fcis-10302	169	40	hh_l3_covariance	hh_l3_covariance	NOUN
fcis-10302	169	41	30	30	NUM
fcis-10302	169	42	29	29	NUM
fcis-10302	169	43	hh_jit_l0.1_variance	hh_jit_l0.1_variance	NOUN
fcis-10302	169	44	60	60	NUM
fcis-10302	169	45	15	15	NUM
fcis-10302	169	46	hh_jit_l0.1_mean	hh_jit_l0.1_mean	NOUN
fcis-10302	169	47	59	59	NUM
fcis-10302	169	48	30	30	NUM
fcis-10302	169	49	hh_jit_l5_mean	hh_jit_l5_mean	NOUN
fcis-10302	169	50	53	53	NUM
fcis-10302	169	51	4.3	4.3	NUM
fcis-10302	169	52	.	.	PUNCT
fcis-10302	170	1	performance	performance	NOUN
fcis-10302	170	2	evaluation	evaluation	NOUN
fcis-10302	170	3	metrics	metric	NOUN
fcis-10302	170	4	we	we	PRON
fcis-10302	170	5	use	use	VERB
fcis-10302	170	6	standard	standard	ADJ
fcis-10302	170	7	evaluation	evaluation	NOUN
fcis-10302	170	8	metrics	metric	NOUN
fcis-10302	170	9	such	such	ADJ
fcis-10302	170	10	as	as	ADP
fcis-10302	170	11	accuracy	accuracy	NOUN
fcis-10302	170	12	,	,	PUNCT
fcis-10302	170	13	precision	precision	NOUN
fcis-10302	170	14	,	,	PUNCT
fcis-10302	170	15	recall	recall	NOUN
fcis-10302	170	16	,	,	PUNCT
fcis-10302	170	17	and	and	CCONJ
fcis-10302	170	18	f1	f1	NOUN
fcis-10302	170	19	score	score	NOUN
fcis-10302	170	20	to	to	PART
fcis-10302	170	21	assess	assess	VERB
fcis-10302	170	22	the	the	DET
fcis-10302	170	23	performance	performance	NOUN
fcis-10302	170	24	of	of	ADP
fcis-10302	170	25	the	the	DET
fcis-10302	170	26	proposed	propose	VERB
fcis-10302	170	27	architecture	architecture	NOUN
fcis-10302	170	28	.	.	PUNCT
fcis-10302	171	1	to	to	PART
fcis-10302	171	2	compute	compute	VERB
fcis-10302	171	3	these	these	DET
fcis-10302	171	4	metrics	metric	NOUN
fcis-10302	171	5	,	,	PUNCT
fcis-10302	171	6	we	we	PRON
fcis-10302	171	7	need	need	VERB
fcis-10302	171	8	to	to	PART
fcis-10302	171	9	calculate	calculate	VERB
fcis-10302	171	10	true	true	ADJ
fcis-10302	171	11	positives	positive	NOUN
fcis-10302	171	12	(	(	PUNCT
fcis-10302	171	13	tp	tp	NOUN
fcis-10302	171	14	)	)	PUNCT
fcis-10302	171	15	,	,	PUNCT
fcis-10302	171	16	false	false	ADJ
fcis-10302	171	17	positives	positive	NOUN
fcis-10302	171	18	(	(	PUNCT
fcis-10302	171	19	fp	fp	NOUN
fcis-10302	171	20	)	)	PUNCT
fcis-10302	171	21	,	,	PUNCT
fcis-10302	171	22	true	true	ADJ
fcis-10302	171	23	negatives	negative	NOUN
fcis-10302	171	24	(	(	PUNCT
fcis-10302	171	25	tn	tn	NOUN
fcis-10302	171	26	)	)	PUNCT
fcis-10302	171	27	,	,	PUNCT
fcis-10302	171	28	and	and	CCONJ
fcis-10302	171	29	false	false	ADJ
fcis-10302	171	30	negatives	negative	NOUN
fcis-10302	171	31	(	(	PUNCT
fcis-10302	171	32	fn	fn	NOUN
fcis-10302	171	33	)	)	PUNCT
fcis-10302	171	34	.	.	PUNCT
fcis-10302	172	1	accuracy	accuracy	NOUN
fcis-10302	172	2	:	:	PUNCT
fcis-10302	172	3	94	94	NUM
fcis-10302	172	4	a	a	DET
fcis-10302	172	5	uracy	uracy	NOUN
fcis-10302	172	6	tp	tp	ADP
fcis-10302	172	7	tn	tn	PROPN
fcis-10302	172	8	cc	cc	PROPN
fcis-10302	172	9	tp	tp	ADP
fcis-10302	172	10	tn	tn	PROPN
fcis-10302	173	1	fp	fp	PROPN
fcis-10302	173	2	fn	fn	PROPN
fcis-10302	173	3			PROPN
fcis-10302	173	4			PROPN
fcis-10302	173	5			PROPN
fcis-10302	173	6			PUNCT
fcis-10302	173	7			X
fcis-10302	173	8	(	(	PUNCT
fcis-10302	173	9	15	15	NUM
fcis-10302	173	10	)	)	PUNCT
fcis-10302	173	11	precision	precision	NOUN
fcis-10302	173	12	:	:	PUNCT
fcis-10302	173	13	tp	tp	PART
fcis-10302	173	14	precision	precision	VERB
fcis-10302	173	15	tp	tp	ADP
fcis-10302	173	16	fp	fp	PROPN
fcis-10302	173	17			PROPN
fcis-10302	173	18			PROPN
fcis-10302	173	19	(	(	PUNCT
fcis-10302	173	20	16	16	NUM
fcis-10302	173	21	)	)	PUNCT
fcis-10302	173	22	recall	recall	NOUN
fcis-10302	173	23	:	:	PUNCT
fcis-10302	173	24	tp	tp	PART
fcis-10302	173	25	recall	recall	VERB
fcis-10302	173	26	tp	tp	ADP
fcis-10302	173	27	fn	fn	PROPN
fcis-10302	173	28			PROPN
fcis-10302	173	29			X
fcis-10302	173	30	(	(	PUNCT
fcis-10302	173	31	17	17	NUM
fcis-10302	173	32	)	)	PUNCT
fcis-10302	173	33	f1	f1	NOUN
fcis-10302	173	34	-	-	PUNCT
fcis-10302	173	35	score	score	NOUN
fcis-10302	173	36	:	:	PUNCT
fcis-10302	173	37	2	2	NUM
fcis-10302	173	38	1	1	NUM
fcis-10302	173	39	recall	recall	NOUN
fcis-10302	173	40	precision	precision	NOUN
fcis-10302	173	41	recall	recall	NOUN
fcis-10302	173	42	f	f	PROPN
fcis-10302	173	43	score	score	NOUN
fcis-10302	173	44	precision	precision	PROPN
fcis-10302	173	45			PROPN
fcis-10302	173	46			PRON
fcis-10302	173	47			ADV
fcis-10302	173	48			NUM
fcis-10302	173	49			NUM
fcis-10302	173	50	(	(	PUNCT
fcis-10302	173	51	18	18	NUM
fcis-10302	173	52	)	)	PUNCT
fcis-10302	173	53	4.4	4.4	NUM
fcis-10302	173	54	.	.	PUNCT
fcis-10302	174	1	experimental	experimental	ADJ
fcis-10302	174	2	results	result	NOUN
fcis-10302	174	3	the	the	DET
fcis-10302	174	4	experiments	experiment	NOUN
fcis-10302	174	5	were	be	AUX
fcis-10302	174	6	conducted	conduct	VERB
fcis-10302	174	7	in	in	ADP
fcis-10302	174	8	the	the	DET
fcis-10302	174	9	following	follow	VERB
fcis-10302	174	10	environment	environment	NOUN
fcis-10302	174	11	:	:	PUNCT
fcis-10302	174	12	cpu	cpu	VERB
fcis-10302	174	13	i7	i7	NOUN
fcis-10302	174	14	-	-	PUNCT
fcis-10302	174	15	11800h	11800h	NUM
fcis-10302	174	16	,	,	PUNCT
fcis-10302	174	17	16	16	NUM
fcis-10302	174	18	gb	gb	NOUN
fcis-10302	174	19	ram	ram	NOUN
fcis-10302	174	20	,	,	PUNCT
fcis-10302	174	21	operating	operating	NOUN
fcis-10302	174	22	system	system	NOUN
fcis-10302	174	23	win10	win10	PROPN
fcis-10302	174	24	,	,	PUNCT
fcis-10302	174	25	development	development	NOUN
fcis-10302	174	26	environment	environment	PROPN
fcis-10302	174	27	python	python	PROPN
fcis-10302	174	28	3.7.12	3.7.12	NUM
fcis-10302	174	29	,	,	PUNCT
fcis-10302	174	30	keras	keras	PROPN
fcis-10302	174	31	2.4.3	2.4.3	NUM
fcis-10302	174	32	,	,	PUNCT
fcis-10302	174	33	and	and	CCONJ
fcis-10302	174	34	tensorflow	tensorflow	NOUN
fcis-10302	174	35	2.3.0	2.3.0	NUM
fcis-10302	174	36	.	.	PUNCT
fcis-10302	175	1	the	the	DET
fcis-10302	175	2	dataset	dataset	NOUN
fcis-10302	175	3	used	use	VERB
fcis-10302	175	4	is	be	AUX
fcis-10302	175	5	the	the	DET
fcis-10302	175	6	nbaiot	nbaiot	ADJ
fcis-10302	175	7	dataset	dataset	NOUN
fcis-10302	175	8	,	,	PUNCT
fcis-10302	175	9	specifically	specifically	ADV
fcis-10302	175	10	the	the	DET
fcis-10302	175	11	traffic	traffic	NOUN
fcis-10302	175	12	data	datum	NOUN
fcis-10302	175	13	from	from	ADP
fcis-10302	175	14	the	the	DET
fcis-10302	175	15	philips	philip	NOUN
fcis-10302	175	16	b120n/10	b120n/10	VERB
fcis-10302	175	17	baby	baby	NOUN
fcis-10302	175	18	monitor	monitor	NOUN
fcis-10302	175	19	device	device	NOUN
fcis-10302	175	20	.	.	PUNCT
fcis-10302	176	1	to	to	PART
fcis-10302	176	2	evaluate	evaluate	VERB
fcis-10302	176	3	and	and	CCONJ
fcis-10302	176	4	examine	examine	VERB
fcis-10302	176	5	the	the	DET
fcis-10302	176	6	proposed	propose	VERB
fcis-10302	176	7	cnn	cnn	PROPN
fcis-10302	176	8	-	-	PUNCT
fcis-10302	176	9	gru	gru	PROPN
fcis-10302	176	10	model	model	NOUN
fcis-10302	176	11	for	for	ADP
fcis-10302	176	12	detecting	detect	VERB
fcis-10302	176	13	malicious	malicious	ADJ
fcis-10302	176	14	iot	iot	NOUN
fcis-10302	176	15	traffic	traffic	NOUN
fcis-10302	176	16	,	,	PUNCT
fcis-10302	176	17	the	the	DET
fcis-10302	176	18	model	model	NOUN
fcis-10302	176	19	uses	use	VERB
fcis-10302	176	20	the	the	DET
fcis-10302	176	21	adam	adam	PROPN
fcis-10302	176	22	optimizer	optimizer	NOUN
fcis-10302	176	23	with	with	ADP
fcis-10302	176	24	a	a	DET
fcis-10302	176	25	learning	learn	VERB
fcis-10302	176	26	rate	rate	NOUN
fcis-10302	176	27	of	of	ADP
fcis-10302	176	28	0.001	0.001	NUM
fcis-10302	176	29	.	.	PUNCT
fcis-10302	177	1	the	the	DET
fcis-10302	177	2	hyperparameters	hyperparameter	NOUN
fcis-10302	177	3	beta_1	beta_1	ADV
fcis-10302	177	4	and	and	CCONJ
fcis-10302	177	5	beta_2	beta_2	PRON
fcis-10302	177	6	is	be	AUX
fcis-10302	177	7	set	set	VERB
fcis-10302	177	8	to	to	ADP
fcis-10302	177	9	0.9	0.9	NUM
fcis-10302	177	10	and	and	CCONJ
fcis-10302	177	11	0.999	0.999	NUM
fcis-10302	177	12	,	,	PUNCT
fcis-10302	177	13	respectively	respectively	ADV
fcis-10302	177	14	.	.	PUNCT
fcis-10302	178	1	the	the	DET
fcis-10302	178	2	model	model	NOUN
fcis-10302	178	3	is	be	AUX
fcis-10302	178	4	trained	train	VERB
fcis-10302	178	5	for	for	ADP
fcis-10302	178	6	20	20	NUM
fcis-10302	178	7	epochs	epoch	NOUN
fcis-10302	178	8	with	with	ADP
fcis-10302	178	9	a	a	DET
fcis-10302	178	10	batch	batch	NOUN
fcis-10302	178	11	size	size	NOUN
fcis-10302	178	12	of	of	ADP
fcis-10302	178	13	256	256	NUM
fcis-10302	178	14	.	.	PUNCT
fcis-10302	179	1	the	the	DET
fcis-10302	179	2	training	training	NOUN
fcis-10302	179	3	and	and	CCONJ
fcis-10302	179	4	validation	validation	NOUN
fcis-10302	179	5	accuracy	accuracy	NOUN
fcis-10302	179	6	and	and	CCONJ
fcis-10302	179	7	loss	loss	NOUN
fcis-10302	179	8	are	be	AUX
fcis-10302	179	9	illustrated	illustrate	VERB
fcis-10302	179	10	in	in	ADP
fcis-10302	179	11	figure	figure	NOUN
fcis-10302	179	12	3	3	NUM
fcis-10302	179	13	.	.	PUNCT
fcis-10302	180	1	(	(	PUNCT
fcis-10302	180	2	a	a	X
fcis-10302	180	3	)	)	PUNCT
fcis-10302	180	4	accuracy	accuracy	NOUN
fcis-10302	180	5	(	(	PUNCT
fcis-10302	180	6	b	b	NOUN
fcis-10302	180	7	)	)	PUNCT
fcis-10302	180	8	loss	loss	NOUN
fcis-10302	180	9	figure	figure	NOUN
fcis-10302	180	10	3	3	NUM
fcis-10302	180	11	.	.	PUNCT
fcis-10302	180	12	training	training	NOUN
fcis-10302	180	13	and	and	CCONJ
fcis-10302	180	14	validation	validation	NOUN
fcis-10302	180	15	accuracy	accuracy	NOUN
fcis-10302	180	16	and	and	CCONJ
fcis-10302	180	17	loss	loss	NOUN
fcis-10302	180	18	to	to	PART
fcis-10302	180	19	highlight	highlight	VERB
fcis-10302	180	20	the	the	DET
fcis-10302	180	21	effectiveness	effectiveness	NOUN
fcis-10302	180	22	of	of	ADP
fcis-10302	180	23	the	the	DET
fcis-10302	180	24	proposed	propose	VERB
fcis-10302	180	25	approach	approach	NOUN
fcis-10302	180	26	,	,	PUNCT
fcis-10302	180	27	we	we	PRON
fcis-10302	180	28	compare	compare	VERB
fcis-10302	180	29	it	it	PRON
fcis-10302	180	30	with	with	ADP
fcis-10302	180	31	another	another	DET
fcis-10302	180	32	popular	popular	ADJ
fcis-10302	180	33	hybrid	hybrid	ADJ
fcis-10302	180	34	deep	deep	ADJ
fcis-10302	180	35	model	model	NOUN
fcis-10302	180	36	,	,	PUNCT
fcis-10302	180	37	cnnlstm	cnnlstm	NOUN
fcis-10302	180	38	.	.	PUNCT
fcis-10302	181	1	for	for	ADP
fcis-10302	181	2	evaluation	evaluation	NOUN
fcis-10302	181	3	,	,	PUNCT
fcis-10302	181	4	we	we	PRON
fcis-10302	181	5	use	use	VERB
fcis-10302	181	6	the	the	DET
fcis-10302	181	7	same	same	ADJ
fcis-10302	181	8	metrics	metric	NOUN
fcis-10302	181	9	for	for	ADP
fcis-10302	181	10	both	both	DET
fcis-10302	181	11	models	model	NOUN
fcis-10302	181	12	,	,	PUNCT
fcis-10302	181	13	and	and	CCONJ
fcis-10302	181	14	they	they	PRON
fcis-10302	181	15	are	be	AUX
fcis-10302	181	16	tested	test	VERB
fcis-10302	181	17	and	and	CCONJ
fcis-10302	181	18	trained	train	VERB
fcis-10302	181	19	on	on	ADP
fcis-10302	181	20	the	the	DET
fcis-10302	181	21	same	same	ADJ
fcis-10302	181	22	dataset	dataset	NOUN
fcis-10302	181	23	.	.	PUNCT
fcis-10302	182	1	the	the	DET
fcis-10302	182	2	cnn	cnn	PROPN
fcis-10302	182	3	-	-	PUNCT
fcis-10302	182	4	gru	gru	PROPN
fcis-10302	182	5	model	model	NOUN
fcis-10302	182	6	achieves	achieve	VERB
fcis-10302	182	7	an	an	DET
fcis-10302	182	8	overall	overall	ADJ
fcis-10302	182	9	accuracy	accuracy	NOUN
fcis-10302	182	10	of	of	ADP
fcis-10302	182	11	99.39	99.39	NUM
fcis-10302	182	12	%	%	NOUN
fcis-10302	182	13	,	,	PUNCT
fcis-10302	182	14	precision	precision	NOUN
fcis-10302	182	15	of	of	ADP
fcis-10302	182	16	99.78	99.78	NUM
fcis-10302	182	17	%	%	NOUN
fcis-10302	182	18	,	,	PUNCT
fcis-10302	182	19	recall	recall	NOUN
fcis-10302	182	20	of	of	ADP
fcis-10302	182	21	98.68	98.68	NUM
fcis-10302	182	22	%	%	NOUN
fcis-10302	182	23	,	,	PUNCT
fcis-10302	182	24	and	and	CCONJ
fcis-10302	182	25	f1	f1	NOUN
fcis-10302	182	26	-	-	PUNCT
fcis-10302	182	27	score	score	NOUN
fcis-10302	182	28	of	of	ADP
fcis-10302	182	29	99.59	99.59	NUM
fcis-10302	182	30	%	%	NOUN
fcis-10302	182	31	,	,	PUNCT
fcis-10302	182	32	which	which	PRON
fcis-10302	182	33	slightly	slightly	ADV
fcis-10302	182	34	outperforms	outperform	VERB
fcis-10302	182	35	the	the	DET
fcis-10302	182	36	cnn	cnn	PROPN
fcis-10302	182	37	-	-	PUNCT
fcis-10302	182	38	lstm	lstm	PROPN
fcis-10302	182	39	model	model	NOUN
fcis-10302	182	40	in	in	ADP
fcis-10302	182	41	the	the	DET
fcis-10302	182	42	accuracy	accuracy	NOUN
fcis-10302	182	43	of	of	ADP
fcis-10302	182	44	traffic	traffic	NOUN
fcis-10302	182	45	detection	detection	NOUN
fcis-10302	182	46	for	for	ADP
fcis-10302	182	47	each	each	DET
fcis-10302	182	48	category	category	NOUN
fcis-10302	182	49	.	.	PUNCT
fcis-10302	183	1	the	the	DET
fcis-10302	183	2	details	detail	NOUN
fcis-10302	183	3	of	of	ADP
fcis-10302	183	4	the	the	DET
fcis-10302	183	5	models	model	NOUN
fcis-10302	183	6	are	be	AUX
fcis-10302	183	7	provided	provide	VERB
fcis-10302	183	8	in	in	ADP
fcis-10302	183	9	figure	figure	NOUN
fcis-10302	183	10	4	4	NUM
fcis-10302	183	11	.	.	PUNCT
fcis-10302	183	12	training	training	NOUN
fcis-10302	183	13	time	time	NOUN
fcis-10302	183	14	is	be	AUX
fcis-10302	183	15	an	an	DET
fcis-10302	183	16	important	important	ADJ
fcis-10302	183	17	metric	metric	NOUN
fcis-10302	183	18	for	for	ADP
fcis-10302	183	19	evaluating	evaluate	VERB
fcis-10302	183	20	model	model	NOUN
fcis-10302	183	21	performance	performance	NOUN
fcis-10302	183	22	.	.	PUNCT
fcis-10302	184	1	the	the	DET
fcis-10302	184	2	training	training	NOUN
fcis-10302	184	3	time	time	NOUN
fcis-10302	184	4	for	for	ADP
fcis-10302	184	5	cnn	cnn	PROPN
fcis-10302	184	6	-	-	PUNCT
fcis-10302	184	7	gru	gru	PROPN
fcis-10302	184	8	is	be	AUX
fcis-10302	184	9	8	8	NUM
fcis-10302	184	10	milliseconds	millisecond	NOUN
fcis-10302	184	11	,	,	PUNCT
fcis-10302	184	12	which	which	PRON
fcis-10302	184	13	is	be	AUX
fcis-10302	184	14	significantly	significantly	ADV
fcis-10302	184	15	faster	fast	ADJ
fcis-10302	184	16	than	than	ADP
fcis-10302	184	17	the	the	DET
fcis-10302	184	18	training	training	NOUN
fcis-10302	184	19	time	time	NOUN
fcis-10302	184	20	of	of	ADP
fcis-10302	184	21	28	28	NUM
fcis-10302	184	22	milliseconds	millisecond	NOUN
fcis-10302	184	23	for	for	ADP
fcis-10302	184	24	cnn	cnn	PROPN
fcis-10302	184	25	-	-	PUNCT
fcis-10302	184	26	lstm	lstm	PROPN
fcis-10302	184	27	,	,	PUNCT
fcis-10302	184	28	indicating	indicate	VERB
fcis-10302	184	29	that	that	SCONJ
fcis-10302	184	30	the	the	DET
fcis-10302	184	31	proposed	propose	VERB
fcis-10302	184	32	model	model	NOUN
fcis-10302	184	33	ensures	ensure	VERB
fcis-10302	184	34	efficient	efficient	ADJ
fcis-10302	184	35	intrusion	intrusion	NOUN
fcis-10302	184	36	detection	detection	NOUN
fcis-10302	184	37	in	in	ADP
fcis-10302	184	38	iot	iot	NOUN
fcis-10302	184	39	while	while	SCONJ
fcis-10302	184	40	providing	provide	VERB
fcis-10302	184	41	faster	fast	ADJ
fcis-10302	184	42	response	response	NOUN
fcis-10302	184	43	time	time	NOUN
fcis-10302	184	44	,	,	PUNCT
fcis-10302	184	45	as	as	SCONJ
fcis-10302	184	46	shown	show	VERB
fcis-10302	184	47	in	in	ADP
fcis-10302	184	48	table	table	NOUN
fcis-10302	184	49	3	3	NUM
fcis-10302	184	50	.	.	PUNCT
fcis-10302	184	51	figure	figure	VERB
fcis-10302	184	52	4	4	NUM
fcis-10302	184	53	.	.	PUNCT
fcis-10302	184	54	comparison	comparison	NOUN
fcis-10302	184	55	of	of	ADP
fcis-10302	184	56	cnn	cnn	PROPN
fcis-10302	184	57	-	-	PUNCT
fcis-10302	184	58	lstm	lstm	PROPN
fcis-10302	184	59	and	and	CCONJ
fcis-10302	184	60	cnn	cnn	PROPN
fcis-10302	184	61	-	-	PUNCT
fcis-10302	184	62	gru	gru	PROPN
fcis-10302	184	63	experiments	experiment	NOUN
fcis-10302	184	64	table	table	VERB
fcis-10302	184	65	3	3	NUM
fcis-10302	184	66	.	.	PUNCT
fcis-10302	184	67	training	training	NOUN
fcis-10302	184	68	time	time	NOUN
fcis-10302	184	69	for	for	ADP
fcis-10302	184	70	cnn	cnn	PROPN
fcis-10302	184	71	-	-	PUNCT
fcis-10302	184	72	lstm	lstm	PROPN
fcis-10302	184	73	and	and	CCONJ
fcis-10302	184	74	cnn	cnn	PROPN
fcis-10302	184	75	-	-	PUNCT
fcis-10302	184	76	gru	gru	PROPN
fcis-10302	184	77	model	model	NOUN
fcis-10302	184	78	training	training	NOUN
fcis-10302	184	79	time	time	NOUN
fcis-10302	184	80	cnn	cnn	PROPN
fcis-10302	184	81	-	-	PUNCT
fcis-10302	184	82	lstm	lstm	PROPN
fcis-10302	184	83	28ms	28ms	NUM
fcis-10302	184	84	cnn	cnn	PROPN
fcis-10302	184	85	-	-	PUNCT
fcis-10302	184	86	gru	gru	PROPN
fcis-10302	184	87	8ms	8ms	ADJ
fcis-10302	184	88	furthermore	furthermore	ADV
fcis-10302	184	89	,	,	PUNCT
fcis-10302	184	90	to	to	PART
fcis-10302	184	91	demonstrate	demonstrate	VERB
fcis-10302	184	92	the	the	DET
fcis-10302	184	93	effectiveness	effectiveness	NOUN
fcis-10302	184	94	of	of	ADP
fcis-10302	184	95	the	the	DET
fcis-10302	184	96	model	model	NOUN
fcis-10302	184	97	,	,	PUNCT
fcis-10302	184	98	table	table	NOUN
fcis-10302	184	99	4	4	NUM
fcis-10302	184	100	presents	present	VERB
fcis-10302	184	101	a	a	DET
fcis-10302	184	102	comparison	comparison	NOUN
fcis-10302	184	103	with	with	ADP
fcis-10302	184	104	some	some	DET
fcis-10302	184	105	classical	classical	ADJ
fcis-10302	184	106	classifiers	classifier	NOUN
fcis-10302	184	107	from	from	ADP
fcis-10302	184	108	domestic	domestic	ADJ
fcis-10302	184	109	literature	literature	NOUN
fcis-10302	184	110	.	.	PUNCT
fcis-10302	185	1	it	it	PRON
fcis-10302	185	2	can	can	AUX
fcis-10302	185	3	be	be	AUX
fcis-10302	185	4	observed	observe	VERB
fcis-10302	185	5	that	that	SCONJ
fcis-10302	185	6	the	the	DET
fcis-10302	185	7	proposed	propose	VERB
fcis-10302	185	8	cnn	cnn	PROPN
fcis-10302	185	9	-	-	PUNCT
fcis-10302	185	10	gru	gru	PROPN
fcis-10302	185	11	model	model	NOUN
fcis-10302	185	12	outperforms	outperform	VERB
fcis-10302	185	13	other	other	ADJ
fcis-10302	185	14	models	model	NOUN
fcis-10302	185	15	in	in	ADP
fcis-10302	185	16	terms	term	NOUN
fcis-10302	185	17	of	of	ADP
fcis-10302	185	18	accuracy	accuracy	NOUN
fcis-10302	185	19	,	,	PUNCT
fcis-10302	185	20	f1	f1	NOUN
fcis-10302	185	21	-	-	PUNCT
fcis-10302	185	22	score	score	NOUN
fcis-10302	185	23	,	,	PUNCT
fcis-10302	185	24	precision	precision	NOUN
fcis-10302	185	25	,	,	PUNCT
fcis-10302	185	26	and	and	CCONJ
fcis-10302	185	27	recall	recall	NOUN
fcis-10302	185	28	,	,	PUNCT
fcis-10302	185	29	achieving	achieve	VERB
fcis-10302	185	30	99.78	99.78	NUM
fcis-10302	185	31	%	%	NOUN
fcis-10302	185	32	,	,	PUNCT
fcis-10302	185	33	99.59	99.59	NUM
fcis-10302	185	34	%	%	NOUN
fcis-10302	185	35	,	,	PUNCT
fcis-10302	185	36	99.52	99.52	NUM
fcis-10302	185	37	%	%	NOUN
fcis-10302	185	38	,	,	PUNCT
fcis-10302	185	39	and	and	CCONJ
fcis-10302	185	40	99.68	99.68	NUM
fcis-10302	185	41	%	%	NOUN
fcis-10302	185	42	respectively	respectively	ADV
fcis-10302	185	43	.	.	PUNCT
fcis-10302	186	1	table	table	NOUN
fcis-10302	186	2	4	4	NUM
fcis-10302	186	3	.	.	PUNCT
fcis-10302	187	1	experimental	experimental	ADJ
fcis-10302	187	2	result	result	PROPN
fcis-10302	187	3	comparison	comparison	NOUN
fcis-10302	187	4	(	(	PUNCT
fcis-10302	187	5	%	%	INTJ
fcis-10302	187	6	)	)	PUNCT
fcis-10302	187	7	model	model	NOUN
fcis-10302	187	8	dataset	dataset	NOUN
fcis-10302	187	9	accuracy	accuracy	NOUN
fcis-10302	187	10	precision	precision	NOUN
fcis-10302	187	11	recall	recall	VERB
fcis-10302	187	12	f1	f1	NOUN
fcis-10302	187	13	-	-	PUNCT
fcis-10302	187	14	score	score	NOUN
fcis-10302	187	15	lr	lr	NOUN
fcis-10302	187	16	-	-	NOUN
fcis-10302	187	17	ann	ann	PROPN
fcis-10302	187	18	[	[	X
fcis-10302	187	19	13	13	NUM
fcis-10302	187	20	]	]	SYM
fcis-10302	187	21	n	n	NUM
fcis-10302	187	22	-	-	PUNCT
fcis-10302	187	23	baiot	baiot	PROPN
fcis-10302	187	24	90.99	90.99	NUM
fcis-10302	187	25	%	%	NOUN
fcis-10302	187	26	90.09	90.09	NUM
fcis-10302	187	27	%	%	NOUN
fcis-10302	187	28	90.93	90.93	NUM
fcis-10302	187	29	%	%	NOUN
fcis-10302	187	30	svm	svm	NOUN
fcis-10302	188	1	[	[	X
fcis-10302	188	2	14	14	NUM
fcis-10302	188	3	]	]	SYM
fcis-10302	188	4	n	n	NUM
fcis-10302	188	5	-	-	PUNCT
fcis-10302	188	6	baiot	baiot	PROPN
fcis-10302	188	7	94.94	94.94	NUM
fcis-10302	188	8	%	%	NOUN
fcis-10302	188	9	90.84	90.84	NUM
fcis-10302	188	10	%	%	NOUN
fcis-10302	188	11	ae	ae	PROPN
fcis-10302	189	1	[	[	X
fcis-10302	189	2	15	15	NUM
fcis-10302	189	3	]	]	SYM
fcis-10302	189	4	n	n	X
fcis-10302	189	5	-	-	PUNCT
fcis-10302	189	6	baiot	baiot	NOUN
fcis-10302	189	7	98.85	98.85	NUM
fcis-10302	189	8	%	%	NOUN
fcis-10302	189	9	98.88	98.88	NUM
fcis-10302	189	10	%	%	NOUN
fcis-10302	189	11	99.10	99.10	NUM
fcis-10302	189	12	%	%	NOUN
fcis-10302	189	13	98.95	98.95	NUM
fcis-10302	189	14	%	%	NOUN
fcis-10302	189	15	my	my	PRON
fcis-10302	189	16	model	model	NOUN
fcis-10302	189	17	n	n	X
fcis-10302	189	18	-	-	PUNCT
fcis-10302	189	19	baiot	baiot	PROPN
fcis-10302	189	20	99.78	99.78	NUM
fcis-10302	189	21	%	%	NOUN
fcis-10302	189	22	99.52	99.52	NUM
fcis-10302	189	23	%	%	NOUN
fcis-10302	189	24	99.68	99.68	NUM
fcis-10302	189	25	%	%	NOUN
fcis-10302	189	26	99.59	99.59	NUM
fcis-10302	189	27	%	%	NOUN
fcis-10302	189	28	5	5	NUM
fcis-10302	189	29	.	.	PUNCT
fcis-10302	189	30	conclusion	conclusion	NOUN
fcis-10302	189	31	in	in	ADP
fcis-10302	189	32	this	this	DET
fcis-10302	189	33	paper	paper	NOUN
fcis-10302	189	34	,	,	PUNCT
fcis-10302	189	35	we	we	PRON
fcis-10302	189	36	proposed	propose	VERB
fcis-10302	189	37	an	an	DET
fcis-10302	189	38	intrusion	intrusion	NOUN
fcis-10302	189	39	detection	detection	NOUN
fcis-10302	189	40	method	method	NOUN
fcis-10302	189	41	for	for	ADP
fcis-10302	189	42	the	the	DET
fcis-10302	189	43	internet	internet	NOUN
fcis-10302	189	44	of	of	ADP
fcis-10302	189	45	things	thing	NOUN
fcis-10302	189	46	(	(	PUNCT
fcis-10302	189	47	iot	iot	NOUN
fcis-10302	189	48	)	)	PUNCT
fcis-10302	189	49	and	and	CCONJ
fcis-10302	189	50	developed	develop	VERB
fcis-10302	189	51	a	a	DET
fcis-10302	189	52	hybrid	hybrid	ADJ
fcis-10302	189	53	deep	deep	ADJ
fcis-10302	189	54	learning	learning	NOUN
fcis-10302	189	55	model	model	NOUN
fcis-10302	189	56	to	to	PART
fcis-10302	189	57	address	address	VERB
fcis-10302	189	58	the	the	DET
fcis-10302	189	59	diverse	diverse	ADJ
fcis-10302	189	60	potential	potential	ADJ
fcis-10302	189	61	security	security	NOUN
fcis-10302	189	62	threats	threat	NOUN
fcis-10302	189	63	in	in	ADP
fcis-10302	189	64	iot	iot	ADJ
fcis-10302	189	65	devices	device	NOUN
fcis-10302	189	66	.	.	PUNCT
fcis-10302	190	1	we	we	PRON
fcis-10302	190	2	utilized	utilize	VERB
fcis-10302	190	3	the	the	DET
fcis-10302	190	4	n	n	NUM
fcis-10302	190	5	-	-	PUNCT
fcis-10302	190	6	baiot	baiot	NOUN
fcis-10302	190	7	dataset	dataset	NOUN
fcis-10302	190	8	and	and	CCONJ
fcis-10302	190	9	applied	apply	VERB
fcis-10302	190	10	xgboost	xgboost	ADP
fcis-10302	190	11	feature	feature	NOUN
fcis-10302	190	12	selection	selection	NOUN
fcis-10302	190	13	to	to	PART
fcis-10302	190	14	reduce	reduce	VERB
fcis-10302	190	15	feature	feature	NOUN
fcis-10302	190	16	redundancy	redundancy	NOUN
fcis-10302	190	17	in	in	ADP
fcis-10302	190	18	the	the	DET
fcis-10302	190	19	iot	iot	NOUN
fcis-10302	190	20	dataset	dataset	NOUN
fcis-10302	190	21	.	.	PUNCT
fcis-10302	191	1	by	by	ADP
fcis-10302	191	2	integrating	integrate	VERB
fcis-10302	191	3	cnn	cnn	PROPN
fcis-10302	191	4	and	and	CCONJ
fcis-10302	191	5	gru	gru	PROPN
fcis-10302	191	6	,	,	PUNCT
fcis-10302	191	7	we	we	PRON
fcis-10302	191	8	achieved	achieve	VERB
fcis-10302	191	9	comprehensive	comprehensive	ADJ
fcis-10302	191	10	and	and	CCONJ
fcis-10302	191	11	effective	effective	ADJ
fcis-10302	191	12	feature	feature	NOUN
fcis-10302	191	13	learning	learning	NOUN
fcis-10302	191	14	.	.	PUNCT
fcis-10302	192	1	furthermore	furthermore	ADV
fcis-10302	192	2	,	,	PUNCT
fcis-10302	192	3	we	we	PRON
fcis-10302	192	4	conducted	conduct	VERB
fcis-10302	192	5	a	a	DET
fcis-10302	192	6	comprehensive	comprehensive	ADJ
fcis-10302	192	7	performance	performance	NOUN
fcis-10302	192	8	analysis	analysis	NOUN
fcis-10302	192	9	by	by	ADP
fcis-10302	192	10	comparing	compare	VERB
fcis-10302	192	11	our	our	PRON
fcis-10302	192	12	proposed	propose	VERB
fcis-10302	192	13	model	model	NOUN
fcis-10302	192	14	with	with	ADP
fcis-10302	192	15	state	state	NOUN
fcis-10302	192	16	-	-	PUNCT
fcis-10302	192	17	of	of	ADP
fcis-10302	192	18	-	-	PUNCT
fcis-10302	192	19	theart	theart	NOUN
fcis-10302	192	20	classification	classification	NOUN
fcis-10302	192	21	models	model	NOUN
fcis-10302	192	22	.	.	PUNCT
fcis-10302	193	1	the	the	DET
fcis-10302	193	2	experimental	experimental	ADJ
fcis-10302	193	3	results	result	NOUN
fcis-10302	193	4	demonstrate	demonstrate	VERB
fcis-10302	193	5	that	that	SCONJ
fcis-10302	193	6	our	our	PRON
fcis-10302	193	7	proposed	propose	VERB
fcis-10302	193	8	iot	iot	PROPN
fcis-10302	193	9	intrusion	intrusion	PROPN
fcis-10302	193	10	detection	detection	NOUN
fcis-10302	193	11	model	model	NOUN
fcis-10302	193	12	performs	perform	VERB
fcis-10302	193	13	well	well	ADV
fcis-10302	193	14	in	in	ADP
fcis-10302	193	15	terms	term	NOUN
fcis-10302	193	16	of	of	ADP
fcis-10302	193	17	accuracy	accuracy	NOUN
fcis-10302	193	18	,	,	PUNCT
fcis-10302	193	19	precision	precision	NOUN
fcis-10302	193	20	,	,	PUNCT
fcis-10302	193	21	recall	recall	NOUN
fcis-10302	193	22	,	,	PUNCT
fcis-10302	193	23	and	and	CCONJ
fcis-10302	193	24	f1score	f1score	NOUN
fcis-10302	193	25	,	,	PUNCT
fcis-10302	193	26	while	while	SCONJ
fcis-10302	193	27	maintaining	maintain	VERB
fcis-10302	193	28	a	a	DET
fcis-10302	193	29	shorter	short	ADJ
fcis-10302	193	30	training	training	NOUN
fcis-10302	193	31	time	time	NOUN
fcis-10302	193	32	.	.	PUNCT
fcis-10302	194	1	in	in	ADP
fcis-10302	194	2	future	future	ADJ
fcis-10302	194	3	work	work	NOUN
fcis-10302	194	4	,	,	PUNCT
fcis-10302	194	5	our	our	PRON
fcis-10302	194	6	goal	goal	NOUN
fcis-10302	194	7	is	be	AUX
fcis-10302	194	8	to	to	PART
fcis-10302	194	9	further	far	ADV
fcis-10302	194	10	enhance	enhance	VERB
fcis-10302	194	11	the	the	DET
fcis-10302	194	12	model	model	NOUN
fcis-10302	194	13	's	's	PART
fcis-10302	194	14	detection	detection	NOUN
fcis-10302	194	15	adaptability	adaptability	NOUN
fcis-10302	194	16	by	by	ADP
fcis-10302	194	17	training	train	VERB
fcis-10302	194	18	it	it	PRON
fcis-10302	194	19	on	on	ADP
fcis-10302	194	20	a	a	DET
fcis-10302	194	21	larger	large	ADJ
fcis-10302	194	22	variety	variety	NOUN
fcis-10302	194	23	of	of	ADP
fcis-10302	194	24	iot	iot	ADJ
fcis-10302	194	25	device	device	NOUN
fcis-10302	194	26	data	datum	NOUN
fcis-10302	194	27	,	,	PUNCT
fcis-10302	194	28	ensuring	ensure	VERB
fcis-10302	194	29	efficient	efficient	ADJ
fcis-10302	194	30	detection	detection	NOUN
fcis-10302	194	31	performance	performance	NOUN
fcis-10302	194	32	.	.	PUNCT
fcis-10302	195	1	references	reference	NOUN
fcis-10302	195	2	[	[	X
fcis-10302	195	3	1	1	NUM
fcis-10302	195	4	]	]	X
fcis-10302	195	5	gupta	gupta	PROPN
fcis-10302	195	6	b	b	PROPN
fcis-10302	195	7	b	b	PROPN
fcis-10302	195	8	,	,	PUNCT
fcis-10302	195	9	dahiya	dahiya	NOUN
fcis-10302	195	10	a.	a.	NOUN
fcis-10302	195	11	distributed	distribute	VERB
fcis-10302	195	12	denial	denial	NOUN
fcis-10302	195	13	of	of	ADP
fcis-10302	195	14	service	service	NOUN
fcis-10302	195	15	(	(	PUNCT
fcis-10302	195	16	ddos	ddos	NOUN
fcis-10302	195	17	)	)	PUNCT
fcis-10302	195	18	attacks	attack	NOUN
fcis-10302	195	19	:	:	PUNCT
fcis-10302	195	20	classification	classification	NOUN
fcis-10302	195	21	,	,	PUNCT
fcis-10302	195	22	attacks	attack	NOUN
fcis-10302	195	23	,	,	PUNCT
fcis-10302	195	24	challenges	challenge	NOUN
fcis-10302	195	25	and	and	CCONJ
fcis-10302	195	26	countermeasures	countermeasure	NOUN
fcis-10302	196	1	[	[	X
fcis-10302	197	1	m	m	X
fcis-10302	197	2	]	]	X
fcis-10302	197	3	.	.	PUNCT
fcis-10302	198	1	crc	crc	PROPN
fcis-10302	198	2	press	press	PROPN
fcis-10302	198	3	,	,	PUNCT
fcis-10302	198	4	2021	2021	NUM
fcis-10302	198	5	.	.	PUNCT
fcis-10302	199	1	97.50	97.50	NUM
fcis-10302	199	2	%	%	NOUN
fcis-10302	199	3	98.00	98.00	NUM
fcis-10302	199	4	%	%	NOUN
fcis-10302	199	5	98.50	98.50	NUM
fcis-10302	199	6	%	%	NOUN
fcis-10302	199	7	99.00	99.00	NUM
fcis-10302	199	8	%	%	NOUN
fcis-10302	199	9	99.50	99.50	NUM
fcis-10302	199	10	%	%	NOUN
fcis-10302	199	11	100.00	100.00	NUM
fcis-10302	199	12	%	%	NOUN
fcis-10302	199	13	accuracy	accuracy	NOUN
fcis-10302	199	14	precision	precision	NOUN
fcis-10302	199	15	recall	recall	VERB
fcis-10302	199	16	f1	f1	NOUN
fcis-10302	199	17	-	-	PUNCT
fcis-10302	199	18	score	score	NOUN
fcis-10302	199	19	cnn	cnn	PROPN
fcis-10302	199	20	-	-	PUNCT
fcis-10302	199	21	lstm	lstm	PROPN
fcis-10302	199	22	cnn	cnn	PROPN
fcis-10302	199	23	-	-	PUNCT
fcis-10302	199	24	gru	gru	PROPN
fcis-10302	199	25	95	95	NUM
fcis-10302	200	1	[	[	X
fcis-10302	200	2	2	2	NUM
fcis-10302	200	3	]	]	X
fcis-10302	200	4	chaabouni	chaabouni	PROPN
fcis-10302	200	5	n	n	CCONJ
fcis-10302	200	6	,	,	PUNCT
fcis-10302	200	7	mosbah	mosbah	PROPN
fcis-10302	200	8	m	m	PROPN
fcis-10302	200	9	,	,	PUNCT
fcis-10302	200	10	zemmari	zemmari	PROPN
fcis-10302	200	11	a	a	X
fcis-10302	200	12	,	,	PUNCT
fcis-10302	200	13	et	et	PROPN
fcis-10302	200	14	al	al	PROPN
fcis-10302	200	15	.	.	PUNCT
fcis-10302	200	16	network	network	PROPN
fcis-10302	200	17	intrusion	intrusion	PROPN
fcis-10302	200	18	detection	detection	NOUN
fcis-10302	200	19	for	for	ADP
fcis-10302	200	20	iot	iot	ADJ
fcis-10302	200	21	security	security	NOUN
fcis-10302	200	22	based	base	VERB
fcis-10302	200	23	on	on	ADP
fcis-10302	200	24	learning	learn	VERB
fcis-10302	200	25	techniques	technique	NOUN
fcis-10302	200	26	[	[	X
fcis-10302	200	27	j	j	X
fcis-10302	200	28	]	]	X
fcis-10302	200	29	.	.	PUNCT
fcis-10302	201	1	ieee	ieee	NOUN
fcis-10302	201	2	communications	communication	NOUN
fcis-10302	201	3	surveys	survey	NOUN
fcis-10302	201	4	&	&	CCONJ
fcis-10302	201	5	tutorials	tutorial	NOUN
fcis-10302	201	6	,	,	PUNCT
fcis-10302	201	7	2019	2019	NUM
fcis-10302	201	8	,	,	PUNCT
fcis-10302	201	9	21(3	21(3	NUM
fcis-10302	201	10	):	):	PUNCT
fcis-10302	201	11	2671	2671	NUM
fcis-10302	201	12	-	-	SYM
fcis-10302	201	13	2701	2701	NUM
fcis-10302	201	14	.	.	PUNCT
fcis-10302	202	1	[	[	X
fcis-10302	202	2	3	3	X
fcis-10302	202	3	]	]	X
fcis-10302	202	4	makhdoom	makhdoom	NOUN
fcis-10302	202	5	i	i	PROPN
fcis-10302	202	6	,	,	PUNCT
fcis-10302	202	7	abolhasan	abolhasan	PROPN
fcis-10302	202	8	m	m	PROPN
fcis-10302	202	9	,	,	PUNCT
fcis-10302	202	10	lipman	lipman	PROPN
fcis-10302	202	11	j	j	PROPN
fcis-10302	202	12	,	,	PUNCT
fcis-10302	202	13	et	et	PROPN
fcis-10302	202	14	al	al	PROPN
fcis-10302	202	15	.	.	PROPN
fcis-10302	202	16	anatomy	anatomy	PROPN
fcis-10302	202	17	of	of	ADP
fcis-10302	202	18	threats	threat	NOUN
fcis-10302	202	19	to	to	ADP
fcis-10302	202	20	the	the	DET
fcis-10302	202	21	internet	internet	NOUN
fcis-10302	202	22	of	of	ADP
fcis-10302	202	23	things[j	things[j	NOUN
fcis-10302	202	24	]	]	PUNCT
fcis-10302	202	25	.	.	PUNCT
fcis-10302	203	1	ieee	ieee	NOUN
fcis-10302	203	2	communications	communication	NOUN
fcis-10302	203	3	surveys	survey	NOUN
fcis-10302	203	4	&	&	CCONJ
fcis-10302	203	5	tutorials	tutorial	NOUN
fcis-10302	203	6	,	,	PUNCT
fcis-10302	203	7	2018	2018	NUM
fcis-10302	203	8	,	,	PUNCT
fcis-10302	203	9	21(2	21(2	NUM
fcis-10302	203	10	):	):	PUNCT
fcis-10302	203	11	1636	1636	NUM
fcis-10302	203	12	-	-	SYM
fcis-10302	203	13	1675	1675	NUM
fcis-10302	203	14	.	.	PUNCT
fcis-10302	204	1	[	[	X
fcis-10302	204	2	4	4	X
fcis-10302	204	3	]	]	PUNCT
fcis-10302	204	4	angrishi	angrishi	PROPN
fcis-10302	204	5	k.	k.	PROPN
fcis-10302	204	6	turning	turn	VERB
fcis-10302	204	7	internet	internet	NOUN
fcis-10302	204	8	of	of	ADP
fcis-10302	204	9	things	thing	NOUN
fcis-10302	204	10	(	(	PUNCT
fcis-10302	204	11	iot	iot	NOUN
fcis-10302	204	12	)	)	PUNCT
fcis-10302	204	13	into	into	ADP
fcis-10302	204	14	internet	internet	NOUN
fcis-10302	204	15	of	of	ADP
fcis-10302	204	16	vulnerabilities	vulnerability	NOUN
fcis-10302	204	17	(	(	PUNCT
fcis-10302	204	18	iov	iov	PROPN
fcis-10302	204	19	):	):	PUNCT
fcis-10302	204	20	iot	iot	PROPN
fcis-10302	204	21	botnets[j	botnets[j	PROPN
fcis-10302	204	22	]	]	PUNCT
fcis-10302	204	23	.	.	PUNCT
fcis-10302	205	1	arxiv	arxiv	PROPN
fcis-10302	205	2	preprint	preprint	PROPN
fcis-10302	205	3	arxiv	arxiv	PROPN
fcis-10302	205	4	:	:	PUNCT
fcis-10302	205	5	1702	1702	NUM
fcis-10302	205	6	.	.	PUNCT
fcis-10302	205	7	03681	03681	NUM
fcis-10302	205	8	,	,	PUNCT
fcis-10302	205	9	2017	2017	NUM
fcis-10302	205	10	.	.	PUNCT
fcis-10302	206	1	[	[	X
fcis-10302	206	2	5	5	NUM
fcis-10302	206	3	]	]	SYM
fcis-10302	206	4	bhunia	bhunia	NOUN
fcis-10302	206	5	s	s	PROPN
fcis-10302	206	6	s	s	PROPN
fcis-10302	206	7	,	,	PUNCT
fcis-10302	206	8	gurusamy	gurusamy	NOUN
fcis-10302	206	9	m.	m.	NOUN
fcis-10302	206	10	dynamic	dynamic	ADJ
fcis-10302	206	11	attack	attack	NOUN
fcis-10302	206	12	detection	detection	NOUN
fcis-10302	206	13	and	and	CCONJ
fcis-10302	206	14	mitigation	mitigation	NOUN
fcis-10302	206	15	in	in	ADP
fcis-10302	206	16	iot	iot	NOUN
fcis-10302	206	17	using	use	VERB
fcis-10302	206	18	sdn[c]//2017	sdn[c]//2017	ADP
fcis-10302	206	19	27th	27th	ADJ
fcis-10302	206	20	international	international	ADJ
fcis-10302	206	21	telecommunication	telecommunication	NOUN
fcis-10302	206	22	networks	network	NOUN
fcis-10302	206	23	and	and	CCONJ
fcis-10302	206	24	applications	application	NOUN
fcis-10302	206	25	conference	conference	NOUN
fcis-10302	206	26	(	(	PUNCT
fcis-10302	206	27	itnac	itnac	PROPN
fcis-10302	206	28	)	)	PUNCT
fcis-10302	206	29	.	.	PUNCT
fcis-10302	207	1	ieee	ieee	PROPN
fcis-10302	207	2	,	,	PUNCT
fcis-10302	207	3	2017	2017	NUM
fcis-10302	207	4	:	:	SYM
fcis-10302	207	5	1	1	NUM
fcis-10302	207	6	-	-	SYM
fcis-10302	207	7	6	6	NUM
fcis-10302	207	8	.	.	PUNCT
fcis-10302	208	1	[	[	X
fcis-10302	208	2	6	6	NUM
fcis-10302	208	3	]	]	SYM
fcis-10302	208	4	hodo	hodo	NOUN
fcis-10302	208	5	e	e	NOUN
fcis-10302	208	6	,	,	PUNCT
fcis-10302	208	7	bellekens	belleken	VERB
fcis-10302	208	8	x	x	SYM
fcis-10302	208	9	,	,	PUNCT
fcis-10302	208	10	hamilton	hamilton	PROPN
fcis-10302	208	11	a	a	PROPN
fcis-10302	208	12	,	,	PUNCT
fcis-10302	208	13	et	et	PROPN
fcis-10302	208	14	al	al	PROPN
fcis-10302	208	15	.	.	PROPN
fcis-10302	209	1	threat	threat	NOUN
fcis-10302	209	2	analysis	analysis	NOUN
fcis-10302	209	3	of	of	ADP
fcis-10302	209	4	iot	iot	ADJ
fcis-10302	209	5	networks	network	NOUN
fcis-10302	209	6	using	use	VERB
fcis-10302	209	7	artificial	artificial	ADJ
fcis-10302	209	8	neural	neural	ADJ
fcis-10302	209	9	network	network	NOUN
fcis-10302	209	10	intrusion	intrusion	NOUN
fcis-10302	209	11	detection	detection	NOUN
fcis-10302	209	12	system	system	NOUN
fcis-10302	209	13	[	[	X
fcis-10302	209	14	c]//2016	c]//2016	PROPN
fcis-10302	209	15	international	international	ADJ
fcis-10302	209	16	symposium	symposium	NOUN
fcis-10302	209	17	on	on	ADP
fcis-10302	209	18	networks	network	NOUN
fcis-10302	209	19	,	,	PUNCT
fcis-10302	209	20	computers	computer	NOUN
fcis-10302	209	21	and	and	CCONJ
fcis-10302	209	22	communications	communication	NOUN
fcis-10302	209	23	(	(	PUNCT
fcis-10302	209	24	isncc	isncc	ADJ
fcis-10302	209	25	)	)	PUNCT
fcis-10302	209	26	.	.	PUNCT
fcis-10302	210	1	ieee	ieee	PROPN
fcis-10302	210	2	,	,	PUNCT
fcis-10302	210	3	2016	2016	NUM
fcis-10302	210	4	:	:	PUNCT
fcis-10302	210	5	1	1	NUM
fcis-10302	210	6	-	-	SYM
fcis-10302	210	7	6	6	NUM
fcis-10302	210	8	.	.	PUNCT
fcis-10302	211	1	[	[	X
fcis-10302	211	2	7	7	X
fcis-10302	211	3	]	]	X
fcis-10302	211	4	chakraborty	chakraborty	NOUN
fcis-10302	211	5	d	d	PROPN
fcis-10302	211	6	,	,	PUNCT
fcis-10302	211	7	narayanan	narayanan	PROPN
fcis-10302	211	8	v	v	PROPN
fcis-10302	211	9	,	,	PUNCT
fcis-10302	211	10	ghosh	ghosh	PROPN
fcis-10302	211	11	a.	a.	NOUN
fcis-10302	211	12	integration	integration	NOUN
fcis-10302	211	13	of	of	ADP
fcis-10302	211	14	deep	deep	ADJ
fcis-10302	211	15	feature	feature	NOUN
fcis-10302	211	16	extraction	extraction	NOUN
fcis-10302	211	17	and	and	CCONJ
fcis-10302	211	18	ensemble	ensemble	ADJ
fcis-10302	211	19	learning	learning	NOUN
fcis-10302	211	20	for	for	ADP
fcis-10302	211	21	outlier	outlier	ADJ
fcis-10302	211	22	detection	detection	NOUN
fcis-10302	212	1	[	[	X
fcis-10302	212	2	j	j	X
fcis-10302	212	3	]	]	X
fcis-10302	212	4	.	.	PUNCT
fcis-10302	213	1	pattern	pattern	NOUN
fcis-10302	213	2	recognition	recognition	NOUN
fcis-10302	213	3	,	,	PUNCT
fcis-10302	213	4	2019	2019	NUM
fcis-10302	213	5	,	,	PUNCT
fcis-10302	213	6	89	89	NUM
fcis-10302	213	7	:	:	SYM
fcis-10302	213	8	161	161	NUM
fcis-10302	213	9	-	-	SYM
fcis-10302	213	10	171	171	NUM
fcis-10302	213	11	.	.	PUNCT
fcis-10302	214	1	[	[	X
fcis-10302	214	2	8	8	NUM
fcis-10302	214	3	]	]	X
fcis-10302	214	4	wang	wang	PROPN
fcis-10302	214	5	yun	yun	PROPN
fcis-10302	214	6	,	,	PUNCT
fcis-10302	214	7	yu	yu	PROPN
fcis-10302	214	8	yao	yao	PROPN
fcis-10302	214	9	,	,	PUNCT
fcis-10302	214	10	zhao	zhao	PROPN
fcis-10302	214	11	yujia	yujia	PROPN
fcis-10302	214	12	,	,	PUNCT
fcis-10302	214	13	et	et	PROPN
fcis-10302	214	14	al	al	PROPN
fcis-10302	214	15	.	.	PROPN
fcis-10302	214	16	intrusion	intrusion	NOUN
fcis-10302	214	17	detection	detection	NOUN
fcis-10302	214	18	and	and	CCONJ
fcis-10302	214	19	defense	defense	NOUN
fcis-10302	214	20	mechanism	mechanism	NOUN
fcis-10302	214	21	based	base	VERB
fcis-10302	214	22	on	on	ADP
fcis-10302	214	23	sdn	sdn	PROPN
fcis-10302	214	24	in	in	ADP
fcis-10302	214	25	home	home	NOUN
fcis-10302	214	26	internet	internet	NOUN
fcis-10302	214	27	of	of	ADP
fcis-10302	214	28	things	thing	NOUN
fcis-10302	214	29	[	[	X
fcis-10302	214	30	j	j	X
fcis-10302	214	31	]	]	X
fcis-10302	214	32	.	.	PUNCT
fcis-10302	215	1	control	control	PROPN
fcis-10302	215	2	engineering	engineering	PROPN
fcis-10302	215	3	,	,	PUNCT
fcis-10302	215	4	2021	2021	NUM
fcis-10302	215	5	,	,	PUNCT
fcis-10302	215	6	28(05	28(05	PROPN
fcis-10302	215	7	):	):	PUNCT
fcis-10302	215	8	1027	1027	NUM
fcis-10302	215	9	-	-	SYM
fcis-10302	215	10	1032	1032	NUM
fcis-10302	215	11	.	.	PUNCT
fcis-10302	216	1	doi	doi	NOUN
fcis-10302	216	2	:	:	PUNCT
fcis-10302	216	3	10	10	NUM
fcis-10302	216	4	.	.	PUNCT
fcis-10302	217	1	14107/	14107/	NUM
fcis-10302	217	2	j.cnki.kzgc.20200900	j.cnki.kzgc.20200900	NOUN
fcis-10302	217	3	.	.	PUNCT
fcis-10302	218	1	[	[	X
fcis-10302	218	2	9	9	NUM
fcis-10302	218	3	]	]	X
fcis-10302	218	4	alkahtani	alkahtani	PROPN
fcis-10302	218	5	h	h	PROPN
fcis-10302	218	6	,	,	PUNCT
fcis-10302	218	7	aldhyani	aldhyani	PROPN
fcis-10302	218	8	t	t	PROPN
fcis-10302	218	9	h	h	PROPN
fcis-10302	218	10	h.	h.	PROPN
fcis-10302	218	11	botnet	botnet	PROPN
fcis-10302	218	12	attack	attack	NOUN
fcis-10302	218	13	detection	detection	NOUN
fcis-10302	218	14	by	by	ADP
fcis-10302	218	15	using	use	VERB
fcis-10302	218	16	cnn	cnn	PROPN
fcis-10302	218	17	-	-	PUNCT
fcis-10302	218	18	lstm	lstm	PROPN
fcis-10302	218	19	model	model	NOUN
fcis-10302	218	20	for	for	ADP
fcis-10302	218	21	internet	internet	NOUN
fcis-10302	218	22	of	of	ADP
fcis-10302	218	23	things	thing	NOUN
fcis-10302	218	24	applications[j	applications[j	PROPN
fcis-10302	218	25	]	]	PUNCT
fcis-10302	218	26	.	.	PUNCT
fcis-10302	218	27	security	security	NOUN
fcis-10302	218	28	and	and	CCONJ
fcis-10302	218	29	communication	communication	NOUN
fcis-10302	218	30	networks	network	NOUN
fcis-10302	218	31	,	,	PUNCT
fcis-10302	218	32	2021	2021	NUM
fcis-10302	218	33	,	,	PUNCT
fcis-10302	218	34	2021	2021	NUM
fcis-10302	218	35	:	:	PUNCT
fcis-10302	218	36	1	1	NUM
fcis-10302	218	37	-	-	SYM
fcis-10302	218	38	23	23	NUM
fcis-10302	218	39	.	.	PUNCT
fcis-10302	219	1	[	[	X
fcis-10302	219	2	10	10	NUM
fcis-10302	219	3	]	]	X
fcis-10302	219	4	alkahtani	alkahtani	PROPN
fcis-10302	219	5	h	h	PROPN
fcis-10302	219	6	,	,	PUNCT
fcis-10302	219	7	aldhyani	aldhyani	PROPN
fcis-10302	219	8	t	t	PROPN
fcis-10302	219	9	h	h	PROPN
fcis-10302	219	10	h.	h.	PROPN
fcis-10302	219	11	botnet	botnet	PROPN
fcis-10302	219	12	attack	attack	NOUN
fcis-10302	219	13	detection	detection	NOUN
fcis-10302	219	14	by	by	ADP
fcis-10302	219	15	using	use	VERB
fcis-10302	219	16	cnn	cnn	PROPN
fcis-10302	219	17	-	-	PUNCT
fcis-10302	219	18	lstm	lstm	PROPN
fcis-10302	219	19	model	model	NOUN
fcis-10302	219	20	for	for	ADP
fcis-10302	219	21	internet	internet	NOUN
fcis-10302	219	22	of	of	ADP
fcis-10302	219	23	things	thing	NOUN
fcis-10302	219	24	applications[j	applications[j	PROPN
fcis-10302	219	25	]	]	PUNCT
fcis-10302	219	26	.	.	PUNCT
fcis-10302	220	1	security	security	NOUN
fcis-10302	220	2	and	and	CCONJ
fcis-10302	220	3	communication	communication	NOUN
fcis-10302	220	4	networks	network	NOUN
fcis-10302	220	5	,	,	PUNCT
fcis-10302	220	6	2021	2021	NUM
fcis-10302	220	7	,	,	PUNCT
fcis-10302	220	8	2021	2021	NUM
fcis-10302	220	9	:	:	PUNCT
fcis-10302	220	10	1	1	NUM
fcis-10302	220	11	-	-	SYM
fcis-10302	220	12	23	23	NUM
fcis-10302	220	13	.	.	PUNCT
fcis-10302	221	1	[	[	X
fcis-10302	221	2	11	11	NUM
fcis-10302	221	3	]	]	X
fcis-10302	221	4	li	li	PROPN
fcis-10302	221	5	xiaojia	xiaojia	PROPN
fcis-10302	221	6	,	,	PUNCT
fcis-10302	221	7	zhao	zhao	PROPN
fcis-10302	221	8	guosheng	guosheng	PROPN
fcis-10302	221	9	,	,	PUNCT
fcis-10302	221	10	wang	wang	PROPN
fcis-10302	221	11	yang	yang	PROPN
fcis-10302	221	12	,	,	PUNCT
fcis-10302	221	13	ning	ning	PROPN
fcis-10302	221	14	ke	ke	PROPN
fcis-10302	221	15	.	.	PUNCT
fcis-10302	222	1	iot	iot	PROPN
fcis-10302	222	2	intrusion	intrusion	PROPN
fcis-10302	222	3	detection	detection	NOUN
fcis-10302	222	4	model	model	NOUN
fcis-10302	222	5	for	for	ADP
fcis-10302	222	6	cnn	cnn	PROPN
fcis-10302	222	7	and	and	CCONJ
fcis-10302	222	8	rnn	rnn	VERB
fcis-10302	222	9	improvement	improvement	NOUN
fcis-10302	223	1	[	[	X
fcis-10302	223	2	j	j	X
fcis-10302	223	3	/	/	SYM
fcis-10302	223	4	ol	ol	PROPN
fcis-10302	223	5	]	]	PUNCT
fcis-10302	223	6	.	.	PUNCT
fcis-10302	224	1	computer	computer	NOUN
fcis-10302	224	2	engineering	engineering	NOUN
fcis-10302	224	3	and	and	CCONJ
fcis-10302	224	4	applications	application	NOUN
fcis-10302	224	5	:	:	PUNCT
fcis-10302	224	6	1	1	NUM
fcis-10302	224	7	-	-	SYM
fcis-10302	224	8	10	10	NUM
fcis-10302	225	1	[	[	X
fcis-10302	225	2	202304	202304	NUM
fcis-10302	225	3	-	-	SYM
fcis-10302	225	4	21	21	NUM
fcis-10302	225	5	]	]	PUNCT
fcis-10302	225	6	.	.	PUNCT
fcis-10302	226	1	http://kns.cnki	http://kns.cnki	PROPN
fcis-10302	226	2	.	.	PUNCT
fcis-10302	227	1	net	net	ADJ
fcis-10302	227	2	/	/	SYM
fcis-10302	227	3	kcms/	kcms/	NUM
fcis-10302	227	4	detail/11	detail/11	NOUN
fcis-10302	227	5	.	.	PUNCT
fcis-10302	228	1	2127	2127	NUM
fcis-10302	228	2	.	.	PUNCT
fcis-10302	229	1	tp	tp	X
fcis-10302	229	2	.	.	PUNCT
fcis-10302	229	3	2023	2023	NUM
fcis-10302	229	4	0302	0302	NUM
fcis-10302	229	5	.	.	PUNCT
fcis-10302	230	1	1415.004.html	1415.004.html	X
fcis-10302	230	2	.	.	PUNCT
fcis-10302	231	1	[	[	X
fcis-10302	231	2	12	12	NUM
fcis-10302	231	3	]	]	X
fcis-10302	231	4	liu	liu	PROPN
fcis-10302	231	5	j	j	PROPN
fcis-10302	231	6	,	,	PUNCT
fcis-10302	231	7	sun	sun	PROPN
fcis-10302	231	8	x	x	PRON
fcis-10302	231	9	,	,	PUNCT
fcis-10302	231	10	jin	jin	PROPN
fcis-10302	231	11	j.	j.	PROPN
fcis-10302	231	12	intrusion	intrusion	PROPN
fcis-10302	231	13	detection	detection	PROPN
fcis-10302	231	14	model	model	NOUN
fcis-10302	231	15	based	base	VERB
fcis-10302	231	16	on	on	ADP
fcis-10302	231	17	principal	principal	ADJ
fcis-10302	231	18	component	component	NOUN
fcis-10302	231	19	analysis	analysis	NOUN
fcis-10302	231	20	and	and	CCONJ
fcis-10302	231	21	recurrent	recurrent	ADJ
fcis-10302	231	22	neural	neural	ADJ
fcis-10302	231	23	network[j	network[j	NOUN
fcis-10302	231	24	]	]	PUNCT
fcis-10302	231	25	.	.	PUNCT
fcis-10302	232	1	journal	journal	PROPN
fcis-10302	232	2	of	of	ADP
fcis-10302	232	3	chinese	chinese	ADJ
fcis-10302	232	4	information	information	NOUN
fcis-10302	232	5	processing	processing	NOUN
fcis-10302	232	6	,	,	PUNCT
fcis-10302	232	7	2020	2020	NUM
fcis-10302	232	8	,	,	PUNCT
fcis-10302	232	9	34(10	34(10	NUM
fcis-10302	232	10	):	):	PUNCT
fcis-10302	232	11	105	105	NUM
fcis-10302	232	12	-	-	SYM
fcis-10302	232	13	112	112	NUM
fcis-10302	232	14	.	.	PUNCT
fcis-10302	233	1	[	[	X
fcis-10302	233	2	13	13	NUM
fcis-10302	233	3	]	]	X
fcis-10302	233	4	zhou	zhou	PROPN
fcis-10302	233	5	jieying	jieying	PROPN
fcis-10302	233	6	,	,	PUNCT
fcis-10302	233	7	he	he	PRON
fcis-10302	233	8	pengfei	pengfei	VERB
fcis-10302	233	9	,	,	PUNCT
fcis-10302	233	10	qiu	qiu	PROPN
fcis-10302	233	11	rongfa	rongfa	PROPN
fcis-10302	233	12	,	,	PUNCT
fcis-10302	233	13	et	et	PROPN
fcis-10302	233	14	al	al	PROPN
fcis-10302	233	15	.	.	PROPN
fcis-10302	233	16	intrusion	intrusion	PROPN
fcis-10302	233	17	detection	detection	PROPN
fcis-10302	233	18	research	research	NOUN
fcis-10302	233	19	based	base	VERB
fcis-10302	233	20	on	on	ADP
fcis-10302	233	21	the	the	DET
fcis-10302	233	22	fusion	fusion	NOUN
fcis-10302	233	23	of	of	ADP
fcis-10302	233	24	random	random	ADJ
fcis-10302	233	25	forest	forest	NOUN
fcis-10302	233	26	and	and	CCONJ
fcis-10302	233	27	gradient	gradient	NOUN
fcis-10302	233	28	boosting	boost	VERB
fcis-10302	233	29	tree[j	tree[j	NOUN
fcis-10302	233	30	]	]	PUNCT
fcis-10302	233	31	.	.	PUNCT
fcis-10302	234	1	journal	journal	PROPN
fcis-10302	234	2	of	of	ADP
fcis-10302	234	3	software	software	NOUN
fcis-10302	234	4	,	,	PUNCT
fcis-10302	234	5	2021	2021	NUM
fcis-10302	234	6	,	,	PUNCT
fcis-10302	234	7	32(10	32(10	NUM
fcis-10302	234	8	):	):	PUNCT
fcis-10302	234	9	3254	3254	NUM
fcis-10302	234	10	-	-	SYM
fcis-10302	234	11	3265	3265	NUM
fcis-10302	234	12	.	.	PUNCT
fcis-10302	235	1	doi	doi	NOUN
fcis-10302	235	2	:	:	PUNCT
fcis-10302	235	3	10.13328	10.13328	NUM
fcis-10302	235	4	/	/	SYM
fcis-10302	235	5	j.cnki.jos.006062	j.cnki.jos.006062	NOUN
fcis-10302	235	6	.	.	PUNCT
fcis-10302	236	1	[	[	X
fcis-10302	236	2	14	14	NUM
fcis-10302	236	3	]	]	X
fcis-10302	236	4	abbasi	abbasi	PROPN
fcis-10302	236	5	f	f	PROPN
fcis-10302	236	6	,	,	PUNCT
fcis-10302	236	7	naderan	naderan	NOUN
fcis-10302	236	8	m	m	PROPN
fcis-10302	236	9	,	,	PUNCT
fcis-10302	236	10	alavi	alavi	PROPN
fcis-10302	236	11	s	s	PROPN
fcis-10302	236	12	e.	e.	PROPN
fcis-10302	236	13	intrusion	intrusion	PROPN
fcis-10302	236	14	detection	detection	NOUN
fcis-10302	236	15	in	in	ADP
fcis-10302	236	16	iot	iot	NOUN
fcis-10302	236	17	with	with	ADP
fcis-10302	236	18	logistic	logistic	ADJ
fcis-10302	236	19	regression	regression	NOUN
fcis-10302	236	20	and	and	CCONJ
fcis-10302	236	21	artificial	artificial	ADJ
fcis-10302	236	22	neural	neural	ADJ
fcis-10302	236	23	network	network	NOUN
fcis-10302	236	24	:	:	PUNCT
fcis-10302	236	25	further	further	ADJ
fcis-10302	236	26	investigations	investigation	NOUN
fcis-10302	236	27	on	on	ADP
fcis-10302	236	28	n	n	CCONJ
fcis-10302	236	29	-	-	PUNCT
fcis-10302	236	30	baiot	baiot	NOUN
fcis-10302	236	31	dataset	dataset	ADJ
fcis-10302	236	32	devices[j	devices[j	PROPN
fcis-10302	236	33	]	]	PUNCT
fcis-10302	236	34	.	.	PUNCT
fcis-10302	237	1	journal	journal	PROPN
fcis-10302	237	2	of	of	ADP
fcis-10302	237	3	computing	computing	NOUN
fcis-10302	237	4	and	and	CCONJ
fcis-10302	237	5	security	security	NOUN
fcis-10302	237	6	,	,	PUNCT
fcis-10302	237	7	2021	2021	NUM
fcis-10302	237	8	,	,	PUNCT
fcis-10302	237	9	8(2	8(2	NUM
fcis-10302	237	10	):	):	PUNCT
fcis-10302	237	11	27	27	NUM
fcis-10302	237	12	-	-	SYM
fcis-10302	237	13	42	42	NUM
fcis-10302	237	14	.	.	PUNCT
fcis-10302	238	1	[	[	X
fcis-10302	238	2	15	15	NUM
fcis-10302	238	3	]	]	X
fcis-10302	238	4	nõmm	nõmm	NOUN
fcis-10302	238	5	s	s	PROPN
fcis-10302	238	6	,	,	PUNCT
fcis-10302	238	7	bahşi	bahşi	PROPN
fcis-10302	238	8	h.	h.	PROPN
fcis-10302	238	9	unsupervised	unsupervised	PROPN
fcis-10302	238	10	anomaly	anomaly	NOUN
fcis-10302	238	11	based	base	VERB
fcis-10302	238	12	botnet	botnet	NOUN
fcis-10302	238	13	detection	detection	NOUN
fcis-10302	238	14	in	in	ADP
fcis-10302	238	15	iot	iot	ADJ
fcis-10302	238	16	networks	network	NOUN
fcis-10302	238	17	[	[	X
fcis-10302	238	18	c]//2018	c]//2018	PROPN
fcis-10302	238	19	17th	17th	ADJ
fcis-10302	238	20	ieee	ieee	PROPN
fcis-10302	238	21	international	international	ADJ
fcis-10302	238	22	conference	conference	NOUN
fcis-10302	238	23	on	on	ADP
fcis-10302	238	24	machine	machine	NOUN
fcis-10302	238	25	learning	learning	NOUN
fcis-10302	238	26	and	and	CCONJ
fcis-10302	238	27	applications	application	NOUN
fcis-10302	238	28	(	(	PUNCT
fcis-10302	238	29	icmla	icmla	NOUN
fcis-10302	238	30	)	)	PUNCT
fcis-10302	238	31	.	.	PUNCT
fcis-10302	239	1	ieee	ieee	NOUN
fcis-10302	239	2	,	,	PUNCT
fcis-10302	239	3	2018	2018	NUM
fcis-10302	239	4	:	:	PUNCT
fcis-10302	239	5	1048	1048	NUM
fcis-10302	239	6	-	-	SYM
fcis-10302	239	7	1053	1053	NUM
fcis-10302	239	8	.	.	PUNCT
fcis-10302	240	1	[	[	X
fcis-10302	240	2	16	16	NUM
fcis-10302	240	3	]	]	AUX
fcis-10302	240	4	hezam	hezam	VERB
fcis-10302	240	5	a	a	DET
fcis-10302	240	6	a	a	NOUN
fcis-10302	240	7	,	,	PUNCT
fcis-10302	240	8	mostafa	mostafa	PROPN
fcis-10302	240	9	s	s	PROPN
fcis-10302	240	10	a	a	PRON
fcis-10302	240	11	,	,	PUNCT
fcis-10302	240	12	baharum	baharum	PROPN
fcis-10302	240	13	z	z	PROPN
fcis-10302	240	14	,	,	PUNCT
fcis-10302	240	15	et	et	PROPN
fcis-10302	240	16	al	al	PROPN
fcis-10302	240	17	.	.	PUNCT
fcis-10302	241	1	combining	combine	VERB
fcis-10302	241	2	deep	deep	ADJ
fcis-10302	241	3	learning	learning	NOUN
fcis-10302	241	4	models	model	NOUN
fcis-10302	241	5	for	for	ADP
fcis-10302	241	6	enhancing	enhance	VERB
fcis-10302	241	7	the	the	DET
fcis-10302	241	8	detection	detection	NOUN
fcis-10302	241	9	of	of	ADP
fcis-10302	241	10	botnet	botnet	NOUN
fcis-10302	241	11	attacks	attack	NOUN
fcis-10302	241	12	in	in	ADP
fcis-10302	241	13	multiple	multiple	ADJ
fcis-10302	241	14	sensors	sensor	NOUN
fcis-10302	241	15	internet	internet	NOUN
fcis-10302	241	16	of	of	ADP
fcis-10302	241	17	things	thing	NOUN
fcis-10302	241	18	networks[j	networks[j	PROPN
fcis-10302	241	19	]	]	PUNCT
fcis-10302	241	20	.	.	PUNCT
fcis-10302	242	1	joiv	joiv	PROPN
fcis-10302	242	2	:	:	PUNCT
fcis-10302	242	3	international	international	ADJ
fcis-10302	242	4	journal	journal	NOUN
fcis-10302	242	5	on	on	ADP
fcis-10302	242	6	informatics	informatic	NOUN
fcis-10302	242	7	visualization	visualization	NOUN
fcis-10302	242	8	,	,	PUNCT
fcis-10302	242	9	2021	2021	NUM
fcis-10302	242	10	,	,	PUNCT
fcis-10302	242	11	5(4	5(4	NUM
fcis-10302	242	12	):	):	PUNCT
fcis-10302	242	13	380	380	NUM
fcis-10302	242	14	-	-	SYM
fcis-10302	242	15	387	387	NUM
fcis-10302	242	16	.	.	PUNCT