id	sid	tid	token	lemma	pos
ajst-2553	1	1	academic	academic	ADJ
ajst-2553	1	2	journal	journal	NOUN
ajst-2553	1	3	of	of	ADP
ajst-2553	1	4	science	science	NOUN
ajst-2553	1	5	and	and	CCONJ
ajst-2553	1	6	technology	technology	NOUN
ajst-2553	1	7	issn	issn	NOUN
ajst-2553	1	8	:	:	PUNCT
ajst-2553	1	9	2771	2771	NUM
ajst-2553	1	10	-	-	SYM
ajst-2553	1	11	3032	3032	NUM
ajst-2553	1	12	|	|	NOUN
ajst-2553	1	13	vol	vol	NOUN
ajst-2553	1	14	.	.	PROPN
ajst-2553	2	1	3	3	NUM
ajst-2553	2	2	,	,	PUNCT
ajst-2553	2	3	no	no	INTJ
ajst-2553	2	4	.	.	NOUN
ajst-2553	2	5	3	3	NUM
ajst-2553	2	6	,	,	PUNCT
ajst-2553	2	7	2022	2022	NUM
ajst-2553	2	8	73	73	NUM
ajst-2553	2	9	research	research	NOUN
ajst-2553	2	10	on	on	ADP
ajst-2553	2	11	network	network	NOUN
ajst-2553	2	12	intrusion	intrusion	NOUN
ajst-2553	2	13	detection	detection	NOUN
ajst-2553	2	14	based	base	VERB
ajst-2553	2	15	on	on	ADP
ajst-2553	2	16	an	an	DET
ajst-2553	2	17	improved	improved	ADJ
ajst-2553	2	18	deep	deep	ADJ
ajst-2553	2	19	learning	learning	NOUN
ajst-2553	2	20	method	method	NOUN
ajst-2553	2	21	le	le	PROPN
ajst-2553	2	22	yang	yang	PROPN
ajst-2553	2	23	1	1	NUM
ajst-2553	2	24	,	,	PUNCT
ajst-2553	2	25	hua	hua	PROPN
ajst-2553	2	26	chen2	chen2	PROPN
ajst-2553	2	27	,	,	PUNCT
ajst-2553	2	28	*	*	PROPN
ajst-2553	2	29	1	1	NUM
ajst-2553	2	30	foshan	foshan	PROPN
ajst-2553	2	31	network	network	NOUN
ajst-2553	2	32	security	security	NOUN
ajst-2553	2	33	emergency	emergency	NOUN
ajst-2553	2	34	command	command	NOUN
ajst-2553	2	35	center	center	NOUN
ajst-2553	2	36	,	,	PUNCT
ajst-2553	2	37	foshan	foshan	PROPN
ajst-2553	2	38	,	,	PUNCT
ajst-2553	2	39	528000	528000	NUM
ajst-2553	2	40	,	,	PUNCT
ajst-2553	2	41	china	china	PROPN
ajst-2553	2	42	2	2	NUM
ajst-2553	2	43	foshan	foshan	PROPN
ajst-2553	2	44	human	human	PROPN
ajst-2553	2	45	resources	resource	NOUN
ajst-2553	2	46	public	public	ADJ
ajst-2553	2	47	service	service	NOUN
ajst-2553	2	48	center	center	NOUN
ajst-2553	2	49	,	,	PUNCT
ajst-2553	2	50	foshan	foshan	PROPN
ajst-2553	2	51	,	,	PUNCT
ajst-2553	2	52	528000	528000	NUM
ajst-2553	2	53	,	,	PUNCT
ajst-2553	2	54	china	china	PROPN
ajst-2553	2	55	*	*	PUNCT
ajst-2553	2	56	corresponding	correspond	VERB
ajst-2553	2	57	author	author	NOUN
ajst-2553	2	58	:	:	PUNCT
ajst-2553	3	1	hua	hua	PROPN
ajst-2553	3	2	chen	chen	PROPN
ajst-2553	3	3	(	(	PUNCT
ajst-2553	3	4	email	email	NOUN
ajst-2553	3	5	:	:	PUNCT
ajst-2553	3	6	chenh393@mail2.sysu.edu.cn	chenh393@mail2.sysu.edu.cn	NOUN
ajst-2553	3	7	)	)	PUNCT
ajst-2553	3	8	abstract	abstract	NOUN
ajst-2553	3	9	:	:	PUNCT
ajst-2553	3	10	network	network	NOUN
ajst-2553	3	11	intrusion	intrusion	NOUN
ajst-2553	3	12	detection	detection	NOUN
ajst-2553	3	13	is	be	AUX
ajst-2553	3	14	an	an	DET
ajst-2553	3	15	important	important	ADJ
ajst-2553	3	16	research	research	NOUN
ajst-2553	3	17	direction	direction	NOUN
ajst-2553	3	18	in	in	ADP
ajst-2553	3	19	the	the	DET
ajst-2553	3	20	field	field	NOUN
ajst-2553	3	21	of	of	ADP
ajst-2553	3	22	network	network	NOUN
ajst-2553	3	23	security	security	NOUN
ajst-2553	3	24	.	.	PUNCT
ajst-2553	4	1	the	the	DET
ajst-2553	4	2	traditional	traditional	ADJ
ajst-2553	4	3	detection	detection	NOUN
ajst-2553	4	4	algorithm	algorithm	NOUN
ajst-2553	4	5	is	be	AUX
ajst-2553	4	6	based	base	VERB
ajst-2553	4	7	on	on	ADP
ajst-2553	4	8	feature	feature	NOUN
ajst-2553	4	9	extraction	extraction	NOUN
ajst-2553	4	10	and	and	CCONJ
ajst-2553	4	11	feature	feature	NOUN
ajst-2553	4	12	separation	separation	NOUN
ajst-2553	4	13	,	,	PUNCT
ajst-2553	4	14	which	which	PRON
ajst-2553	4	15	has	have	VERB
ajst-2553	4	16	the	the	DET
ajst-2553	4	17	problems	problem	NOUN
ajst-2553	4	18	of	of	ADP
ajst-2553	4	19	low	low	ADJ
ajst-2553	4	20	detection	detection	NOUN
ajst-2553	4	21	accuracy	accuracy	NOUN
ajst-2553	4	22	and	and	CCONJ
ajst-2553	4	23	high	high	ADJ
ajst-2553	4	24	false	false	ADJ
ajst-2553	4	25	alarm	alarm	NOUN
ajst-2553	4	26	rate	rate	NOUN
ajst-2553	4	27	.	.	PUNCT
ajst-2553	5	1	in	in	ADP
ajst-2553	5	2	order	order	NOUN
ajst-2553	5	3	to	to	PART
ajst-2553	5	4	improve	improve	VERB
ajst-2553	5	5	the	the	DET
ajst-2553	5	6	accuracy	accuracy	NOUN
ajst-2553	5	7	of	of	ADP
ajst-2553	5	8	network	network	NOUN
ajst-2553	5	9	intrusion	intrusion	NOUN
ajst-2553	5	10	detection	detection	NOUN
ajst-2553	5	11	,	,	PUNCT
ajst-2553	5	12	this	this	DET
ajst-2553	5	13	paper	paper	NOUN
ajst-2553	5	14	proposes	propose	VERB
ajst-2553	5	15	an	an	DET
ajst-2553	5	16	intrusion	intrusion	NOUN
ajst-2553	5	17	detection	detection	NOUN
ajst-2553	5	18	model	model	NOUN
ajst-2553	5	19	based	base	VERB
ajst-2553	5	20	on	on	ADP
ajst-2553	5	21	deep	deep	ADJ
ajst-2553	5	22	asymmetric	asymmetric	ADJ
ajst-2553	5	23	convolutional	convolutional	ADJ
ajst-2553	5	24	encoder	encoder	NOUN
ajst-2553	5	25	and	and	CCONJ
ajst-2553	5	26	random	random	ADJ
ajst-2553	5	27	forest(rf	forest(rf	NOUN
ajst-2553	5	28	)	)	PUNCT
ajst-2553	5	29	.	.	PUNCT
ajst-2553	6	1	first	first	ADV
ajst-2553	6	2	,	,	PUNCT
ajst-2553	6	3	use	use	VERB
ajst-2553	6	4	dacae	dacae	NOUN
ajst-2553	6	5	to	to	PART
ajst-2553	6	6	extract	extract	VERB
ajst-2553	6	7	features	feature	NOUN
ajst-2553	6	8	from	from	ADP
ajst-2553	6	9	the	the	DET
ajst-2553	6	10	preprocessed	preprocesse	VERB
ajst-2553	6	11	data	datum	NOUN
ajst-2553	6	12	,	,	PUNCT
ajst-2553	6	13	and	and	CCONJ
ajst-2553	6	14	then	then	ADV
ajst-2553	6	15	use	use	VERB
ajst-2553	6	16	the	the	DET
ajst-2553	6	17	random	random	ADJ
ajst-2553	6	18	forest	forest	NOUN
ajst-2553	6	19	algorithm	algorithm	NOUN
ajst-2553	6	20	to	to	PART
ajst-2553	6	21	divide	divide	VERB
ajst-2553	6	22	the	the	DET
ajst-2553	6	23	network	network	NOUN
ajst-2553	6	24	traffic	traffic	NOUN
ajst-2553	6	25	data	datum	NOUN
ajst-2553	6	26	into	into	ADP
ajst-2553	6	27	normal	normal	ADJ
ajst-2553	6	28	and	and	CCONJ
ajst-2553	6	29	abnormal	abnormal	ADJ
ajst-2553	6	30	classes	class	NOUN
ajst-2553	6	31	,	,	PUNCT
ajst-2553	6	32	and	and	CCONJ
ajst-2553	6	33	finally	finally	ADV
ajst-2553	6	34	achieve	achieve	VERB
ajst-2553	6	35	the	the	DET
ajst-2553	6	36	purpose	purpose	NOUN
ajst-2553	6	37	of	of	ADP
ajst-2553	6	38	network	network	NOUN
ajst-2553	6	39	intrusion	intrusion	NOUN
ajst-2553	6	40	detection	detection	NOUN
ajst-2553	6	41	.	.	PUNCT
ajst-2553	7	1	it	it	PRON
ajst-2553	7	2	is	be	AUX
ajst-2553	7	3	tested	test	VERB
ajst-2553	7	4	on	on	ADP
ajst-2553	7	5	three	three	NUM
ajst-2553	7	6	public	public	ADJ
ajst-2553	7	7	benchmark	benchmark	NOUN
ajst-2553	7	8	datasets	dataset	NOUN
ajst-2553	7	9	of	of	ADP
ajst-2553	7	10	network	network	NOUN
ajst-2553	7	11	intrusion	intrusion	NOUN
ajst-2553	7	12	detection	detection	PROPN
ajst-2553	7	13	nsl	nsl	PROPN
ajst-2553	7	14	-	-	PUNCT
ajst-2553	7	15	kdd	kdd	PROPN
ajst-2553	7	16	and	and	CCONJ
ajst-2553	7	17	kdd99	kdd99	PROPN
ajst-2553	7	18	datasets	dataset	NOUN
ajst-2553	7	19	.	.	PUNCT
ajst-2553	8	1	the	the	DET
ajst-2553	8	2	experimental	experimental	ADJ
ajst-2553	8	3	results	result	NOUN
ajst-2553	8	4	show	show	VERB
ajst-2553	8	5	that	that	SCONJ
ajst-2553	8	6	the	the	DET
ajst-2553	8	7	accuracy	accuracy	NOUN
ajst-2553	8	8	and	and	CCONJ
ajst-2553	8	9	false	false	ADJ
ajst-2553	8	10	alarm	alarm	NOUN
ajst-2553	8	11	rate	rate	NOUN
ajst-2553	8	12	of	of	ADP
ajst-2553	8	13	the	the	DET
ajst-2553	8	14	improved	improve	VERB
ajst-2553	8	15	method	method	NOUN
ajst-2553	8	16	are	be	AUX
ajst-2553	8	17	better	well	ADJ
ajst-2553	8	18	than	than	ADP
ajst-2553	8	19	the	the	DET
ajst-2553	8	20	comparative	comparative	ADJ
ajst-2553	8	21	method	method	NOUN
ajst-2553	8	22	.	.	PUNCT
ajst-2553	9	1	keywords	keyword	NOUN
ajst-2553	9	2	:	:	PUNCT
ajst-2553	9	3	network	network	NOUN
ajst-2553	9	4	security	security	NOUN
ajst-2553	9	5	,	,	PUNCT
ajst-2553	9	6	intrusion	intrusion	NOUN
ajst-2553	9	7	detection	detection	NOUN
ajst-2553	9	8	,	,	PUNCT
ajst-2553	9	9	deep	deep	ADJ
ajst-2553	9	10	learning	learning	NOUN
ajst-2553	9	11	,	,	PUNCT
ajst-2553	9	12	random	random	ADJ
ajst-2553	9	13	forest	forest	NOUN
ajst-2553	9	14	.	.	PUNCT
ajst-2553	10	1	1	1	X
ajst-2553	10	2	.	.	X
ajst-2553	10	3	introduction	introduction	NOUN
ajst-2553	10	4	with	with	ADP
ajst-2553	10	5	the	the	DET
ajst-2553	10	6	development	development	NOUN
ajst-2553	10	7	of	of	ADP
ajst-2553	10	8	internet	internet	NOUN
ajst-2553	10	9	technology	technology	NOUN
ajst-2553	10	10	and	and	CCONJ
ajst-2553	10	11	application	application	NOUN
ajst-2553	10	12	,	,	PUNCT
ajst-2553	10	13	dos	do	NOUN
ajst-2553	10	14	attacks	attack	NOUN
ajst-2553	10	15	,	,	PUNCT
ajst-2553	10	16	ransomware	ransomware	NOUN
ajst-2553	10	17	and	and	CCONJ
ajst-2553	10	18	other	other	ADJ
ajst-2553	10	19	attacks	attack	NOUN
ajst-2553	10	20	are	be	AUX
ajst-2553	10	21	increasing	increase	VERB
ajst-2553	10	22	.	.	PUNCT
ajst-2553	11	1	in	in	ADP
ajst-2553	11	2	recent	recent	ADJ
ajst-2553	11	3	years	year	NOUN
ajst-2553	11	4	,	,	PUNCT
ajst-2553	11	5	with	with	SCONJ
ajst-2553	11	6	the	the	DET
ajst-2553	11	7	emergence	emergence	NOUN
ajst-2553	11	8	of	of	ADP
ajst-2553	11	9	big	big	ADJ
ajst-2553	11	10	data	datum	NOUN
ajst-2553	11	11	technology	technology	NOUN
ajst-2553	11	12	,	,	PUNCT
ajst-2553	11	13	deep	deep	ADJ
ajst-2553	11	14	learning	learning	NOUN
ajst-2553	11	15	,	,	PUNCT
ajst-2553	11	16	reinforcement	reinforcement	NOUN
ajst-2553	11	17	learning	learning	NOUN
ajst-2553	11	18	,	,	PUNCT
ajst-2553	11	19	visualization	visualization	NOUN
ajst-2553	11	20	and	and	CCONJ
ajst-2553	11	21	other	other	ADJ
ajst-2553	11	22	technologies	technology	NOUN
ajst-2553	11	23	have	have	AUX
ajst-2553	11	24	been	be	AUX
ajst-2553	11	25	widely	widely	ADV
ajst-2553	11	26	used	use	VERB
ajst-2553	11	27	,	,	PUNCT
ajst-2553	11	28	and	and	CCONJ
ajst-2553	11	29	have	have	AUX
ajst-2553	11	30	achieved	achieve	VERB
ajst-2553	11	31	great	great	ADJ
ajst-2553	11	32	success	success	NOUN
ajst-2553	11	33	in	in	ADP
ajst-2553	11	34	natural	natural	ADJ
ajst-2553	11	35	language	language	NOUN
ajst-2553	11	36	processing	processing	NOUN
ajst-2553	11	37	,	,	PUNCT
ajst-2553	11	38	image	image	NOUN
ajst-2553	11	39	recognition	recognition	NOUN
ajst-2553	11	40	,	,	PUNCT
ajst-2553	11	41	video	video	NOUN
ajst-2553	11	42	detection	detection	NOUN
ajst-2553	11	43	and	and	CCONJ
ajst-2553	11	44	other	other	ADJ
ajst-2553	11	45	fields	field	NOUN
ajst-2553	11	46	.	.	PUNCT
ajst-2553	12	1	at	at	ADP
ajst-2553	12	2	the	the	DET
ajst-2553	12	3	same	same	ADJ
ajst-2553	12	4	time	time	NOUN
ajst-2553	12	5	,	,	PUNCT
ajst-2553	12	6	in	in	ADP
ajst-2553	12	7	the	the	DET
ajst-2553	12	8	field	field	NOUN
ajst-2553	12	9	of	of	ADP
ajst-2553	12	10	network	network	NOUN
ajst-2553	12	11	security	security	NOUN
ajst-2553	12	12	,	,	PUNCT
ajst-2553	12	13	a	a	DET
ajst-2553	12	14	lot	lot	NOUN
ajst-2553	12	15	of	of	ADP
ajst-2553	12	16	research	research	NOUN
ajst-2553	12	17	work	work	NOUN
ajst-2553	12	18	has	have	AUX
ajst-2553	12	19	been	be	AUX
ajst-2553	12	20	done	do	VERB
ajst-2553	12	21	to	to	PART
ajst-2553	12	22	use	use	VERB
ajst-2553	12	23	these	these	DET
ajst-2553	12	24	technologies	technology	NOUN
ajst-2553	12	25	for	for	ADP
ajst-2553	12	26	network	network	NOUN
ajst-2553	12	27	intrusion	intrusion	NOUN
ajst-2553	12	28	detection	detection	NOUN
ajst-2553	12	29	,	,	PUNCT
ajst-2553	12	30	and	and	CCONJ
ajst-2553	12	31	some	some	DET
ajst-2553	12	32	results	result	NOUN
ajst-2553	12	33	have	have	AUX
ajst-2553	12	34	been	be	AUX
ajst-2553	12	35	achieved	achieve	VERB
ajst-2553	12	36	.	.	PUNCT
ajst-2553	13	1	julisch	julisch	NOUN
ajst-2553	14	1	[	[	X
ajst-2553	14	2	1	1	X
ajst-2553	14	3	]	]	PUNCT
ajst-2553	14	4	proposed	propose	VERB
ajst-2553	14	5	a	a	DET
ajst-2553	14	6	new	new	ADJ
ajst-2553	14	7	method	method	NOUN
ajst-2553	14	8	for	for	ADP
ajst-2553	14	9	detecting	detect	VERB
ajst-2553	14	10	intrusion	intrusion	NOUN
ajst-2553	14	11	alerts	alert	NOUN
ajst-2553	14	12	.	.	PUNCT
ajst-2553	15	1	the	the	DET
ajst-2553	15	2	data	data	NOUN
ajst-2553	15	3	is	be	AUX
ajst-2553	15	4	processed	process	VERB
ajst-2553	15	5	by	by	ADP
ajst-2553	15	6	eliminating	eliminate	VERB
ajst-2553	15	7	the	the	DET
ajst-2553	15	8	root	root	NOUN
ajst-2553	15	9	cause	cause	NOUN
ajst-2553	15	10	of	of	ADP
ajst-2553	15	11	the	the	DET
ajst-2553	15	12	event	event	NOUN
ajst-2553	15	13	,	,	PUNCT
ajst-2553	15	14	and	and	CCONJ
ajst-2553	15	15	the	the	DET
ajst-2553	15	16	new	new	ADJ
ajst-2553	15	17	alarm	alarm	NOUN
ajst-2553	15	18	clustering	cluster	VERB
ajst-2553	15	19	method	method	NOUN
ajst-2553	15	20	is	be	AUX
ajst-2553	15	21	used	use	VERB
ajst-2553	15	22	to	to	PART
ajst-2553	15	23	help	help	VERB
ajst-2553	15	24	researchers	researcher	NOUN
ajst-2553	15	25	identify	identify	VERB
ajst-2553	15	26	the	the	DET
ajst-2553	15	27	network	network	NOUN
ajst-2553	15	28	anomaly	anomaly	NOUN
ajst-2553	15	29	detection	detection	NOUN
ajst-2553	15	30	type	type	NOUN
ajst-2553	15	31	.	.	PUNCT
ajst-2553	16	1	zhang	zhang	X
ajst-2553	17	1	[	[	X
ajst-2553	17	2	2	2	NUM
ajst-2553	17	3	]	]	PUNCT
ajst-2553	17	4	et	et	NOUN
ajst-2553	17	5	al	al	PROPN
ajst-2553	17	6	applied	apply	VERB
ajst-2553	17	7	random	random	ADJ
ajst-2553	17	8	forest	forest	NOUN
ajst-2553	17	9	algorithm	algorithm	NOUN
ajst-2553	17	10	to	to	ADP
ajst-2553	17	11	network	network	NOUN
ajst-2553	17	12	intrusion	intrusion	NOUN
ajst-2553	17	13	detection	detection	NOUN
ajst-2553	17	14	system	system	NOUN
ajst-2553	17	15	.	.	PUNCT
ajst-2553	18	1	its	its	PRON
ajst-2553	18	2	network	network	NOUN
ajst-2553	18	3	service	service	NOUN
ajst-2553	18	4	mode	mode	NOUN
ajst-2553	18	5	is	be	AUX
ajst-2553	18	6	constructed	construct	VERB
ajst-2553	18	7	and	and	CCONJ
ajst-2553	18	8	implemented	implement	VERB
ajst-2553	18	9	on	on	ADP
ajst-2553	18	10	the	the	DET
ajst-2553	18	11	network	network	NOUN
ajst-2553	18	12	data	datum	NOUN
ajst-2553	18	13	flow	flow	VERB
ajst-2553	18	14	through	through	ADP
ajst-2553	18	15	the	the	DET
ajst-2553	18	16	random	random	ADJ
ajst-2553	18	17	forest	forest	NOUN
ajst-2553	18	18	algorithm	algorithm	NOUN
ajst-2553	18	19	.	.	PUNCT
ajst-2553	19	1	the	the	DET
ajst-2553	19	2	algorithm	algorithm	NOUN
ajst-2553	19	3	is	be	AUX
ajst-2553	19	4	based	base	VERB
ajst-2553	19	5	on	on	ADP
ajst-2553	19	6	unsupervised	unsupervised	ADJ
ajst-2553	19	7	learning	learning	NOUN
ajst-2553	19	8	to	to	PART
ajst-2553	19	9	overcome	overcome	VERB
ajst-2553	19	10	the	the	DET
ajst-2553	19	11	label	label	NOUN
ajst-2553	19	12	dependence	dependence	NOUN
ajst-2553	19	13	problem	problem	NOUN
ajst-2553	19	14	in	in	ADP
ajst-2553	19	15	supervised	supervised	ADJ
ajst-2553	19	16	learning	learning	NOUN
ajst-2553	19	17	.	.	PUNCT
ajst-2553	20	1	chen	chen	PROPN
ajst-2553	21	1	[	[	X
ajst-2553	21	2	3	3	X
ajst-2553	21	3	]	]	PUNCT
ajst-2553	21	4	et	et	PROPN
ajst-2553	21	5	al	al	PROPN
ajst-2553	21	6	.	.	PROPN
ajst-2553	21	7	used	use	VERB
ajst-2553	21	8	two	two	NUM
ajst-2553	21	9	data	datum	NOUN
ajst-2553	21	10	mining	mining	NOUN
ajst-2553	21	11	methods	method	NOUN
ajst-2553	21	12	,	,	PUNCT
ajst-2553	21	13	artificial	artificial	ADJ
ajst-2553	21	14	neural	neural	ADJ
ajst-2553	21	15	network	network	NOUN
ajst-2553	21	16	(	(	PUNCT
ajst-2553	21	17	ann	ann	PROPN
ajst-2553	21	18	)	)	PUNCT
ajst-2553	21	19	and	and	CCONJ
ajst-2553	21	20	support	support	VERB
ajst-2553	21	21	vector	vector	NOUN
ajst-2553	21	22	machine	machine	NOUN
ajst-2553	21	23	(	(	PUNCT
ajst-2553	21	24	svm	svm	PROPN
ajst-2553	21	25	)	)	PUNCT
ajst-2553	21	26	to	to	PART
ajst-2553	21	27	detect	detect	VERB
ajst-2553	21	28	potential	potential	ADJ
ajst-2553	21	29	system	system	NOUN
ajst-2553	21	30	intrusions	intrusion	NOUN
ajst-2553	21	31	.	.	PUNCT
ajst-2553	22	1	gao	gao	PROPN
ajst-2553	23	1	[	[	X
ajst-2553	23	2	4	4	X
ajst-2553	23	3	]	]	PUNCT
ajst-2553	23	4	et	et	PROPN
ajst-2553	23	5	al	al	PROPN
ajst-2553	23	6	.	.	PROPN
ajst-2553	23	7	began	begin	VERB
ajst-2553	23	8	to	to	PART
ajst-2553	23	9	introduce	introduce	VERB
ajst-2553	23	10	deep	deep	ADJ
ajst-2553	23	11	belief	belief	NOUN
ajst-2553	23	12	networks	network	NOUN
ajst-2553	23	13	into	into	ADP
ajst-2553	23	14	the	the	DET
ajst-2553	23	15	field	field	NOUN
ajst-2553	23	16	of	of	ADP
ajst-2553	23	17	anomaly	anomaly	NOUN
ajst-2553	23	18	detection	detection	NOUN
ajst-2553	23	19	,	,	PUNCT
ajst-2553	23	20	and	and	CCONJ
ajst-2553	23	21	composed	compose	VERB
ajst-2553	23	22	multiple	multiple	ADJ
ajst-2553	23	23	layers	layer	NOUN
ajst-2553	23	24	of	of	ADP
ajst-2553	23	25	restricted	restrict	VERB
ajst-2553	23	26	boltzmann	boltzmann	PROPN
ajst-2553	23	27	machines	machine	NOUN
ajst-2553	23	28	into	into	ADP
ajst-2553	23	29	neural	neural	ADJ
ajst-2553	23	30	network	network	NOUN
ajst-2553	23	31	classifiers	classifier	NOUN
ajst-2553	23	32	to	to	PART
ajst-2553	23	33	obtain	obtain	VERB
ajst-2553	23	34	better	well	ADJ
ajst-2553	23	35	performance	performance	NOUN
ajst-2553	23	36	.	.	PUNCT
ajst-2553	24	1	vinayaku	vinayaku	PROPN
ajst-2553	24	2	mar	mar	PROPN
ajst-2553	25	1	[	[	X
ajst-2553	25	2	5	5	NUM
ajst-2553	25	3	]	]	PUNCT
ajst-2553	25	4	used	use	VERB
ajst-2553	25	5	convolutional	convolutional	ADJ
ajst-2553	25	6	neural	neural	ADJ
ajst-2553	25	7	network	network	NOUN
ajst-2553	25	8	(	(	PUNCT
ajst-2553	25	9	cnn	cnn	PROPN
ajst-2553	25	10	)	)	PUNCT
ajst-2553	25	11	for	for	ADP
ajst-2553	25	12	network	network	NOUN
ajst-2553	25	13	intrusion	intrusion	NOUN
ajst-2553	25	14	detection	detection	NOUN
ajst-2553	25	15	.	.	PUNCT
ajst-2553	26	1	the	the	DET
ajst-2553	26	2	network	network	NOUN
ajst-2553	26	3	traffic	traffic	NOUN
ajst-2553	26	4	is	be	AUX
ajst-2553	26	5	modeled	model	VERB
ajst-2553	26	6	as	as	ADP
ajst-2553	26	7	a	a	DET
ajst-2553	26	8	time	time	NOUN
ajst-2553	26	9	series	series	NOUN
ajst-2553	26	10	,	,	PUNCT
ajst-2553	26	11	and	and	CCONJ
ajst-2553	26	12	the	the	DET
ajst-2553	26	13	data	data	NOUN
ajst-2553	26	14	packets	packet	NOUN
ajst-2553	26	15	of	of	ADP
ajst-2553	26	16	tcp	tcp	PROPN
ajst-2553	26	17	/	/	SYM
ajst-2553	26	18	ip	ip	NOUN
ajst-2553	26	19	protocol	protocol	NOUN
ajst-2553	26	20	are	be	AUX
ajst-2553	26	21	modeled	model	VERB
ajst-2553	26	22	using	use	VERB
ajst-2553	26	23	supervised	supervised	ADJ
ajst-2553	26	24	learning	learning	NOUN
ajst-2553	26	25	method	method	NOUN
ajst-2553	26	26	within	within	ADP
ajst-2553	26	27	a	a	DET
ajst-2553	26	28	predefined	predefine	VERB
ajst-2553	26	29	time	time	NOUN
ajst-2553	26	30	range	range	NOUN
ajst-2553	26	31	.	.	PUNCT
ajst-2553	27	1	the	the	DET
ajst-2553	27	2	validity	validity	NOUN
ajst-2553	27	3	of	of	ADP
ajst-2553	27	4	the	the	DET
ajst-2553	27	5	network	network	NOUN
ajst-2553	27	6	structure	structure	NOUN
ajst-2553	27	7	in	in	ADP
ajst-2553	27	8	intrusion	intrusion	NOUN
ajst-2553	27	9	detection	detection	NOUN
ajst-2553	27	10	is	be	AUX
ajst-2553	27	11	also	also	ADV
ajst-2553	27	12	proved	prove	VERB
ajst-2553	27	13	on	on	ADP
ajst-2553	27	14	kdd99	kdd99	PROPN
ajst-2553	27	15	data	datum	NOUN
ajst-2553	27	16	set	set	VERB
ajst-2553	27	17	.	.	PUNCT
ajst-2553	28	1	the	the	DET
ajst-2553	28	2	traditional	traditional	ADJ
ajst-2553	28	3	machine	machine	NOUN
ajst-2553	28	4	learning	learning	NOUN
ajst-2553	28	5	methods	method	NOUN
ajst-2553	28	6	are	be	AUX
ajst-2553	28	7	still	still	ADV
ajst-2553	28	8	based	base	VERB
ajst-2553	28	9	on	on	ADP
ajst-2553	28	10	the	the	DET
ajst-2553	28	11	prior	prior	ADJ
ajst-2553	28	12	knowledge	knowledge	NOUN
ajst-2553	28	13	of	of	ADP
ajst-2553	28	14	experts	expert	NOUN
ajst-2553	28	15	,	,	PUNCT
ajst-2553	28	16	and	and	CCONJ
ajst-2553	28	17	still	still	ADV
ajst-2553	28	18	rely	rely	VERB
ajst-2553	28	19	on	on	ADP
ajst-2553	28	20	the	the	DET
ajst-2553	28	21	label	label	NOUN
ajst-2553	28	22	of	of	ADP
ajst-2553	28	23	data	datum	NOUN
ajst-2553	28	24	sets	set	NOUN
ajst-2553	28	25	.	.	PUNCT
ajst-2553	29	1	an	an	DET
ajst-2553	29	2	excellent	excellent	ADJ
ajst-2553	29	3	intrusion	intrusion	NOUN
ajst-2553	29	4	detection	detection	NOUN
ajst-2553	29	5	system	system	NOUN
ajst-2553	29	6	should	should	AUX
ajst-2553	29	7	not	not	PART
ajst-2553	29	8	only	only	ADV
ajst-2553	29	9	detect	detect	VERB
ajst-2553	29	10	anomalies	anomaly	NOUN
ajst-2553	29	11	from	from	ADP
ajst-2553	29	12	known	know	VERB
ajst-2553	29	13	network	network	NOUN
ajst-2553	29	14	traffic	traffic	NOUN
ajst-2553	29	15	,	,	PUNCT
ajst-2553	29	16	but	but	CCONJ
ajst-2553	29	17	also	also	ADV
ajst-2553	29	18	detect	detect	VERB
ajst-2553	29	19	new	new	ADJ
ajst-2553	29	20	unknown	unknown	ADJ
ajst-2553	29	21	traffic	traffic	NOUN
ajst-2553	29	22	from	from	ADP
ajst-2553	29	23	massive	massive	ADJ
ajst-2553	29	24	and	and	CCONJ
ajst-2553	29	25	highdimensional	highdimensional	ADJ
ajst-2553	29	26	data	datum	NOUN
ajst-2553	29	27	flows	flow	VERB
ajst-2553	29	28	.	.	PUNCT
ajst-2553	30	1	2	2	X
ajst-2553	30	2	.	.	NUM
ajst-2553	30	3	improved	improve	VERB
ajst-2553	30	4	deep	deep	ADJ
ajst-2553	30	5	learning	learning	NOUN
ajst-2553	30	6	model	model	NOUN
ajst-2553	30	7	2.1	2.1	NUM
ajst-2553	30	8	.	.	PUNCT
ajst-2553	31	1	autoencoder	autoencoder	NOUN
ajst-2553	31	2	based	base	VERB
ajst-2553	31	3	on	on	ADP
ajst-2553	31	4	the	the	DET
ajst-2553	31	5	back	back	ADJ
ajst-2553	31	6	-	-	PUNCT
ajst-2553	31	7	propagation	propagation	NOUN
ajst-2553	31	8	algorithm	algorithm	NOUN
ajst-2553	31	9	and	and	CCONJ
ajst-2553	31	10	optimization	optimization	NOUN
ajst-2553	31	11	method	method	NOUN
ajst-2553	31	12	,	,	PUNCT
ajst-2553	31	13	the	the	DET
ajst-2553	31	14	auto	auto	NOUN
ajst-2553	31	15	-	-	PUNCT
ajst-2553	31	16	encoder	encoder	NOUN
ajst-2553	31	17	uses	use	VERB
ajst-2553	31	18	the	the	DET
ajst-2553	31	19	input	input	NOUN
ajst-2553	31	20	data	datum	NOUN
ajst-2553	31	21	x	x	X
ajst-2553	31	22	itself	itself	PRON
ajst-2553	31	23	as	as	ADP
ajst-2553	31	24	a	a	DET
ajst-2553	31	25	supervision	supervision	NOUN
ajst-2553	31	26	to	to	PART
ajst-2553	31	27	guide	guide	VERB
ajst-2553	31	28	the	the	DET
ajst-2553	31	29	neural	neural	ADJ
ajst-2553	31	30	network	network	NOUN
ajst-2553	31	31	to	to	PART
ajst-2553	31	32	try	try	VERB
ajst-2553	31	33	to	to	PART
ajst-2553	31	34	learn	learn	VERB
ajst-2553	31	35	a	a	DET
ajst-2553	31	36	mapping	mapping	NOUN
ajst-2553	31	37	relationship	relationship	NOUN
ajst-2553	31	38	,	,	PUNCT
ajst-2553	31	39	thereby	thereby	ADV
ajst-2553	31	40	obtaining	obtain	VERB
ajst-2553	31	41	a	a	DET
ajst-2553	31	42	reconstructed	reconstructed	ADJ
ajst-2553	31	43	output	output	NOUN
ajst-2553	31	44	xr	xr	NOUN
ajst-2553	31	45	.	.	PUNCT
ajst-2553	32	1	in	in	ADP
ajst-2553	32	2	the	the	DET
ajst-2553	32	3	time	time	NOUN
ajst-2553	32	4	series	series	PROPN
ajst-2553	32	5	anomaly	anomaly	PROPN
ajst-2553	32	6	detection	detection	PROPN
ajst-2553	32	7	scenario	scenario	NOUN
ajst-2553	32	8	,	,	PUNCT
ajst-2553	32	9	anomalies	anomaly	NOUN
ajst-2553	32	10	are	be	AUX
ajst-2553	32	11	rare	rare	ADJ
ajst-2553	32	12	for	for	ADP
ajst-2553	32	13	normal	normal	ADJ
ajst-2553	32	14	,	,	PUNCT
ajst-2553	32	15	so	so	CCONJ
ajst-2553	32	16	if	if	SCONJ
ajst-2553	32	17	the	the	DET
ajst-2553	32	18	difference	difference	NOUN
ajst-2553	32	19	between	between	ADP
ajst-2553	32	20	the	the	DET
ajst-2553	32	21	output	output	NOUN
ajst-2553	32	22	xr	xr	PROPN
ajst-2553	32	23	reconstructed	reconstruct	VERB
ajst-2553	32	24	by	by	ADP
ajst-2553	32	25	the	the	DET
ajst-2553	32	26	auto	auto	NOUN
ajst-2553	32	27	-	-	PUNCT
ajst-2553	32	28	encoder	encoder	NOUN
ajst-2553	32	29	and	and	CCONJ
ajst-2553	32	30	the	the	DET
ajst-2553	32	31	original	original	ADJ
ajst-2553	32	32	input	input	NOUN
ajst-2553	32	33	exceeds	exceed	VERB
ajst-2553	32	34	a	a	DET
ajst-2553	32	35	certain	certain	ADJ
ajst-2553	32	36	threshold	threshold	NOUN
ajst-2553	32	37	(	(	PUNCT
ajst-2553	32	38	threshold	threshold	NOUN
ajst-2553	32	39	)	)	PUNCT
ajst-2553	32	40	,	,	PUNCT
ajst-2553	32	41	the	the	DET
ajst-2553	32	42	original	original	ADJ
ajst-2553	32	43	time	time	NOUN
ajst-2553	32	44	series	series	NOUN
ajst-2553	32	45	is	be	AUX
ajst-2553	32	46	abnormal	abnormal	ADJ
ajst-2553	32	47	.	.	PUNCT
ajst-2553	33	1	the	the	DET
ajst-2553	33	2	function	function	NOUN
ajst-2553	33	3	of	of	ADP
ajst-2553	33	4	the	the	DET
ajst-2553	33	5	encoder	encoder	NOUN
ajst-2553	33	6	is	be	AUX
ajst-2553	33	7	to	to	PART
ajst-2553	33	8	encode	encode	VERB
ajst-2553	33	9	the	the	DET
ajst-2553	33	10	highdimensional	highdimensional	ADJ
ajst-2553	33	11	input	input	NOUN
ajst-2553	33	12	x	x	X
ajst-2553	33	13	into	into	ADP
ajst-2553	33	14	the	the	DET
ajst-2553	33	15	low	low	ADV
ajst-2553	33	16	-	-	PUNCT
ajst-2553	33	17	dimensional	dimensional	ADJ
ajst-2553	33	18	hidden	hidden	ADJ
ajst-2553	33	19	variable	variable	ADJ
ajst-2553	33	20	h	h	NOUN
ajst-2553	33	21	so	so	SCONJ
ajst-2553	33	22	as	as	SCONJ
ajst-2553	33	23	to	to	PART
ajst-2553	33	24	force	force	VERB
ajst-2553	33	25	the	the	DET
ajst-2553	33	26	neural	neural	ADJ
ajst-2553	33	27	network	network	NOUN
ajst-2553	33	28	to	to	PART
ajst-2553	33	29	learn	learn	VERB
ajst-2553	33	30	the	the	DET
ajst-2553	33	31	most	most	ADV
ajst-2553	33	32	informative	informative	ADJ
ajst-2553	33	33	features	feature	NOUN
ajst-2553	33	34	;	;	PUNCT
ajst-2553	33	35	the	the	DET
ajst-2553	33	36	function	function	NOUN
ajst-2553	33	37	of	of	ADP
ajst-2553	33	38	the	the	DET
ajst-2553	33	39	decoder	decoder	NOUN
ajst-2553	33	40	is	be	AUX
ajst-2553	33	41	to	to	PART
ajst-2553	33	42	restore	restore	VERB
ajst-2553	33	43	the	the	DET
ajst-2553	33	44	hidden	hide	VERB
ajst-2553	33	45	variable	variable	ADJ
ajst-2553	33	46	h	h	NOUN
ajst-2553	33	47	of	of	ADP
ajst-2553	33	48	the	the	DET
ajst-2553	33	49	hidden	hide	VERB
ajst-2553	33	50	layer	layer	NOUN
ajst-2553	33	51	to	to	ADP
ajst-2553	33	52	the	the	DET
ajst-2553	33	53	initial	initial	ADJ
ajst-2553	33	54	dimension	dimension	NOUN
ajst-2553	33	55	.	.	PUNCT
ajst-2553	34	1	the	the	DET
ajst-2553	34	2	best	good	ADJ
ajst-2553	34	3	state	state	NOUN
ajst-2553	34	4	is	be	AUX
ajst-2553	34	5	that	that	SCONJ
ajst-2553	34	6	the	the	DET
ajst-2553	34	7	output	output	NOUN
ajst-2553	34	8	of	of	ADP
ajst-2553	34	9	the	the	DET
ajst-2553	34	10	decoder	decoder	NOUN
ajst-2553	34	11	can	can	AUX
ajst-2553	34	12	perfectly	perfectly	ADV
ajst-2553	34	13	or	or	CCONJ
ajst-2553	34	14	approximately	approximately	ADV
ajst-2553	34	15	recover	recover	VERB
ajst-2553	34	16	the	the	DET
ajst-2553	34	17	original	original	ADJ
ajst-2553	34	18	input	input	NOUN
ajst-2553	34	19	,	,	PUNCT
ajst-2553	34	20	that	that	ADV
ajst-2553	34	21	is	is	ADV
ajst-2553	34	22	,	,	PUNCT
ajst-2553	34	23	x	x	PRON
ajst-2553	34	24	'	'	PUNCT
ajst-2553	34	25	≈	≈	PROPN
ajst-2553	34	26	x.	x.	NOUN
ajst-2553	34	27	figure	figure	NOUN
ajst-2553	34	28	1	1	NUM
ajst-2553	34	29	.	.	PUNCT
ajst-2553	35	1	the	the	DET
ajst-2553	35	2	antoencoder	antoencoder	NOUN
ajst-2553	35	3	as	as	SCONJ
ajst-2553	35	4	shown	show	VERB
ajst-2553	35	5	in	in	ADP
ajst-2553	35	6	figure	figure	NOUN
ajst-2553	35	7	1	1	NUM
ajst-2553	35	8	:	:	PUNCT
ajst-2553	35	9	the	the	DET
ajst-2553	35	10	encoding	encoding	NOUN
ajst-2553	35	11	process	process	NOUN
ajst-2553	35	12	of	of	ADP
ajst-2553	35	13	the	the	DET
ajst-2553	35	14	original	original	ADJ
ajst-2553	35	15	data	datum	NOUN
ajst-2553	35	16	x	x	PUNCT
ajst-2553	35	17	from	from	ADP
ajst-2553	35	18	the	the	DET
ajst-2553	35	19	input	input	NOUN
ajst-2553	35	20	74	74	NUM
ajst-2553	35	21	layer	layer	NOUN
ajst-2553	35	22	to	to	ADP
ajst-2553	35	23	the	the	DET
ajst-2553	35	24	hidden	hide	VERB
ajst-2553	35	25	layer	layer	NOUN
ajst-2553	35	26	:	:	PUNCT
ajst-2553	35	27			NOUN
ajst-2553	35	28			SYM
ajst-2553	35	29			NOUN
ajst-2553	35	30	1bx1wx	1bx1wx	NOUN
ajst-2553	35	31	1	1	NUM
ajst-2553	35	32	h	h	NOUN
ajst-2553	35	33			NUM
ajst-2553	35	34	g	g	NUM
ajst-2553	35	35	(	(	PUNCT
ajst-2553	35	36	1	1	NUM
ajst-2553	35	37	)	)	PUNCT
ajst-2553	35	38	the	the	DET
ajst-2553	35	39	decoding	decode	VERB
ajst-2553	35	40	process	process	NOUN
ajst-2553	35	41	from	from	ADP
ajst-2553	35	42	the	the	DET
ajst-2553	35	43	hidden	hide	VERB
ajst-2553	35	44	layer	layer	NOUN
ajst-2553	35	45	to	to	ADP
ajst-2553	35	46	the	the	DET
ajst-2553	35	47	output	output	NOUN
ajst-2553	35	48	layer	layer	NOUN
ajst-2553	35	49	:	:	PUNCT
ajst-2553	35	50			NOUN
ajst-2553	35	51			PROPN
ajst-2553	35	52			NOUN
ajst-2553	35	53	22	22	PROPN
ajst-2553	35	54	bhhx	bhhx	NOUN
ajst-2553	35	55	2	2	NUM
ajst-2553	35	56			ADV
ajst-2553	35	57			PROPN
ajst-2553	35	58			NUM
ajst-2553	35	59	wg	wg	PROPN
ajst-2553	35	60			PROPN
ajst-2553	35	61	(	(	PUNCT
ajst-2553	35	62	2	2	NUM
ajst-2553	35	63	)	)	PUNCT
ajst-2553	35	64	then	then	ADV
ajst-2553	35	65	the	the	DET
ajst-2553	35	66	optimization	optimization	NOUN
ajst-2553	35	67	objective	objective	ADJ
ajst-2553	35	68	function	function	NOUN
ajst-2553	35	69	of	of	ADP
ajst-2553	35	70	the	the	DET
ajst-2553	35	71	algorithm	algorithm	NOUN
ajst-2553	35	72	is	be	AUX
ajst-2553	35	73	:	:	PUNCT
ajst-2553	35	74	)	)	PUNCT
ajst-2553	35	75	,	,	PUNCT
ajst-2553	35	76	(	(	PUNCT
ajst-2553	35	77	rxxdistssminimizelo	rxxdistssminimizelo	VERB
ajst-2553	35	78			NUM
ajst-2553	35	79	(	(	PUNCT
ajst-2553	35	80	3	3	NUM
ajst-2553	35	81	)	)	PUNCT
ajst-2553	35	82	where	where	SCONJ
ajst-2553	35	83	dist	dist	NOUN
ajst-2553	35	84	is	be	AUX
ajst-2553	35	85	the	the	DET
ajst-2553	35	86	distance	distance	NOUN
ajst-2553	35	87	measurement	measurement	NOUN
ajst-2553	35	88	function	function	NOUN
ajst-2553	35	89	,	,	PUNCT
ajst-2553	35	90	usually	usually	ADV
ajst-2553	35	91	using	use	VERB
ajst-2553	35	92	(	(	PUNCT
ajst-2553	35	93	mean	mean	ADJ
ajst-2553	35	94	square	square	ADJ
ajst-2553	35	95	difference	difference	NOUN
ajst-2553	35	96	)	)	PUNCT
ajst-2553	35	97	mse	mse	NOUN
ajst-2553	35	98	.	.	PUNCT
ajst-2553	36	1	2.2	2.2	NUM
ajst-2553	36	2	.	.	PUNCT
ajst-2553	37	1	convolutional	convolutional	ADJ
ajst-2553	37	2	autoencoder	autoencoder	NOUN
ajst-2553	37	3	the	the	DET
ajst-2553	37	4	training	training	NOUN
ajst-2553	37	5	method	method	NOUN
ajst-2553	37	6	of	of	ADP
ajst-2553	37	7	the	the	DET
ajst-2553	37	8	convolutional	convolutional	ADJ
ajst-2553	37	9	autoencoder	autoencoder	NOUN
ajst-2553	37	10	is	be	AUX
ajst-2553	37	11	similar	similar	ADJ
ajst-2553	37	12	to	to	ADP
ajst-2553	37	13	that	that	PRON
ajst-2553	37	14	of	of	ADP
ajst-2553	37	15	the	the	DET
ajst-2553	37	16	autoencoder	autoencoder	NOUN
ajst-2553	37	17	.	.	PUNCT
ajst-2553	38	1	in	in	ADP
ajst-2553	38	2	each	each	DET
ajst-2553	38	3	round	round	NOUN
ajst-2553	38	4	of	of	ADP
ajst-2553	38	5	training	training	NOUN
ajst-2553	38	6	operation	operation	NOUN
ajst-2553	38	7	of	of	ADP
ajst-2553	38	8	the	the	DET
ajst-2553	38	9	convolutional	convolutional	ADJ
ajst-2553	38	10	layer	layer	NOUN
ajst-2553	38	11	,	,	PUNCT
ajst-2553	38	12	k	k	PROPN
ajst-2553	38	13	convolution	convolution	NOUN
ajst-2553	38	14	sums	sum	NOUN
ajst-2553	38	15	are	be	AUX
ajst-2553	38	16	initialized	initialize	VERB
ajst-2553	38	17	,	,	PUNCT
ajst-2553	38	18	and	and	CCONJ
ajst-2553	38	19	each	each	DET
ajst-2553	38	20	convolution	convolution	NOUN
ajst-2553	38	21	has	have	VERB
ajst-2553	38	22	a	a	DET
ajst-2553	38	23	weight	weight	NOUN
ajst-2553	38	24	w	w	NOUN
ajst-2553	38	25	and	and	CCONJ
ajst-2553	38	26	a	a	DET
ajst-2553	38	27	bias	bias	NOUN
ajst-2553	38	28	b.	b.	PROPN
ajst-2553	39	1	the	the	DET
ajst-2553	39	2	convolution	convolution	NOUN
ajst-2553	39	3	kernel	kernel	PROPN
ajst-2553	39	4	produces	produce	VERB
ajst-2553	39	5	k	k	PROPN
ajst-2553	39	6	feature	feature	NOUN
ajst-2553	39	7	maps	map	NOUN
ajst-2553	39	8	after	after	ADP
ajst-2553	39	9	the	the	DET
ajst-2553	39	10	convolution	convolution	NOUN
ajst-2553	39	11	operation	operation	NOUN
ajst-2553	39	12	on	on	ADP
ajst-2553	39	13	the	the	DET
ajst-2553	39	14	input	input	NOUN
ajst-2553	39	15	vector	vector	NOUN
ajst-2553	39	16	x.	x.	NOUN
ajst-2553	40	1	the	the	DET
ajst-2553	40	2	formula	formula	NOUN
ajst-2553	40	3	is	be	AUX
ajst-2553	40	4	shown	show	VERB
ajst-2553	40	5	in	in	ADP
ajst-2553	40	6	(	(	PUNCT
ajst-2553	40	7	4.3	4.3	NUM
ajst-2553	40	8	)	)	PUNCT
ajst-2553	40	9	,	,	PUNCT
ajst-2553	40	10	σ	σ	PROPN
ajst-2553	40	11	represents	represent	VERB
ajst-2553	40	12	the	the	DET
ajst-2553	40	13	activation	activation	NOUN
ajst-2553	40	14	function	function	NOUN
ajst-2553	40	15	,	,	PUNCT
ajst-2553	40	16	and	and	CCONJ
ajst-2553	40	17	the	the	DET
ajst-2553	40	18	symbol	symbol	NOUN
ajst-2553	40	19	*	*	PUNCT
ajst-2553	40	20	is	be	AUX
ajst-2553	40	21	the	the	DET
ajst-2553	40	22	convolution	convolution	NOUN
ajst-2553	40	23	operation	operation	NOUN
ajst-2553	40	24	.	.	PUNCT
ajst-2553	41	1			NOUN
ajst-2553	41	2	kkk	kkk	PROPN
ajst-2553	41	3	bwxh	bwxh	NOUN
ajst-2553	41	4			PROPN
ajst-2553	41	5	(	(	PUNCT
ajst-2553	41	6	4	4	X
ajst-2553	41	7	)	)	PUNCT
ajst-2553	41	8	figure	figure	NOUN
ajst-2553	41	9	2	2	NUM
ajst-2553	41	10	.	.	PUNCT
ajst-2553	41	11	convolutional	convolutional	ADJ
ajst-2553	41	12	autoencoder	autoencoder	NOUN
ajst-2553	41	13	reconstruct	reconstruct	VERB
ajst-2553	41	14	the	the	DET
ajst-2553	41	15	convolutional	convolutional	ADJ
ajst-2553	41	16	layer	layer	NOUN
ajst-2553	41	17	obtained	obtain	VERB
ajst-2553	41	18	from	from	ADP
ajst-2553	41	19	the	the	DET
ajst-2553	41	20	previous	previous	ADJ
ajst-2553	41	21	layer	layer	NOUN
ajst-2553	41	22	to	to	PART
ajst-2553	41	23	obtain	obtain	VERB
ajst-2553	41	24	the	the	DET
ajst-2553	41	25	following	follow	VERB
ajst-2553	41	26	equation	equation	NOUN
ajst-2553	41	27	(	(	PUNCT
ajst-2553	41	28	5	5	NUM
ajst-2553	41	29	)	)	PUNCT
ajst-2553	41	30	)	)	PUNCT
ajst-2553	42	1	(	(	PUNCT
ajst-2553	42	2			PROPN
ajst-2553	42	3			NUM
ajst-2553	42	4	k	k	PROPN
ajst-2553	42	5	kckwkhy	kckwkhy	PROPN
ajst-2553	42	6			PROPN
ajst-2553	42	7	(	(	PUNCT
ajst-2553	42	8	5	5	NUM
ajst-2553	42	9	)	)	PUNCT
ajst-2553	42	10	in	in	ADP
ajst-2553	42	11	equation	equation	NOUN
ajst-2553	42	12	(	(	PUNCT
ajst-2553	42	13	5	5	NUM
ajst-2553	42	14	)	)	PUNCT
ajst-2553	42	15	,	,	PUNCT
ajst-2553	42	16	y	y	PROPN
ajst-2553	42	17	is	be	AUX
ajst-2553	42	18	the	the	DET
ajst-2553	42	19	feature	feature	NOUN
ajst-2553	42	20	after	after	ADP
ajst-2553	42	21	reconstruction	reconstruction	NOUN
ajst-2553	42	22	of	of	ADP
ajst-2553	42	23	the	the	DET
ajst-2553	42	24	convolution	convolution	NOUN
ajst-2553	42	25	output	output	NOUN
ajst-2553	42	26	,	,	PUNCT
ajst-2553	42	27	w	w	PROPN
ajst-2553	42	28	represents	represent	VERB
ajst-2553	42	29	the	the	DET
ajst-2553	42	30	transpose	transpose	NOUN
ajst-2553	42	31	of	of	ADP
ajst-2553	42	32	the	the	DET
ajst-2553	42	33	weight	weight	NOUN
ajst-2553	42	34	w	w	PROPN
ajst-2553	42	35	of	of	ADP
ajst-2553	42	36	the	the	DET
ajst-2553	42	37	k	k	NOUN
ajst-2553	42	38	-	-	PUNCT
ajst-2553	42	39	th	th	VERB
ajst-2553	42	40	feature	feature	NOUN
ajst-2553	42	41	,	,	PUNCT
ajst-2553	42	42	and	and	CCONJ
ajst-2553	42	43	ck	ck	PROPN
ajst-2553	42	44	represents	represent	VERB
ajst-2553	42	45	the	the	DET
ajst-2553	42	46	bias	bias	NOUN
ajst-2553	42	47	.	.	PUNCT
ajst-2553	43	1	the	the	DET
ajst-2553	43	2	convolutional	convolutional	ADJ
ajst-2553	43	3	autoencoder	autoencoder	NOUN
ajst-2553	43	4	compares	compare	VERB
ajst-2553	43	5	the	the	DET
ajst-2553	43	6	input	input	NOUN
ajst-2553	43	7	data	datum	NOUN
ajst-2553	43	8	with	with	ADP
ajst-2553	43	9	the	the	DET
ajst-2553	43	10	output	output	NOUN
ajst-2553	43	11	data	datum	NOUN
ajst-2553	43	12	results	result	NOUN
ajst-2553	43	13	,	,	PUNCT
ajst-2553	43	14	and	and	CCONJ
ajst-2553	43	15	trains	train	NOUN
ajst-2553	43	16	and	and	CCONJ
ajst-2553	43	17	adjusts	adjust	VERB
ajst-2553	43	18	the	the	DET
ajst-2553	43	19	parameters	parameter	NOUN
ajst-2553	43	20	for	for	ADP
ajst-2553	43	21	optimization	optimization	NOUN
ajst-2553	43	22	.	.	PUNCT
ajst-2553	44	1	2.3	2.3	NUM
ajst-2553	44	2	.	.	PUNCT
ajst-2553	45	1	deep	deep	ADJ
ajst-2553	45	2	asymmetric	asymmetric	ADJ
ajst-2553	45	3	convolutional	convolutional	ADJ
ajst-2553	45	4	autoencoder	autoencoder	NOUN
ajst-2553	45	5	the	the	DET
ajst-2553	45	6	input	input	NOUN
ajst-2553	45	7	vector	vector	NOUN
ajst-2553	45	8	of	of	ADP
ajst-2553	45	9	the	the	DET
ajst-2553	45	10	encoder	encoder	NOUN
ajst-2553	45	11	in	in	ADP
ajst-2553	45	12	the	the	DET
ajst-2553	45	13	dacae	dacae	PROPN
ajst-2553	45	14	model	model	NOUN
ajst-2553	45	15	is	be	AUX
ajst-2553	45	16	set	set	VERB
ajst-2553	45	17	to	to	ADP
ajst-2553	45	18	lrx	lrx	PROPN
ajst-2553	45	19	.	.	PUNCT
ajst-2553	46	1	after	after	ADP
ajst-2553	46	2	learning	learn	VERB
ajst-2553	46	3	the	the	DET
ajst-2553	46	4	input	input	NOUN
ajst-2553	46	5	of	of	ADP
ajst-2553	46	6	each	each	DET
ajst-2553	46	7	layer	layer	NOUN
ajst-2553	46	8	,	,	PUNCT
ajst-2553	46	9	the	the	DET
ajst-2553	46	10	hidden	hide	VERB
ajst-2553	46	11	layer	layer	NOUN
ajst-2553	46	12	is	be	AUX
ajst-2553	46	13	coded	code	VERB
ajst-2553	46	14	and	and	CCONJ
ajst-2553	46	15	mapped	map	VERB
ajst-2553	46	16	to	to	ADP
ajst-2553	46	17	l	l	NOUN
ajst-2553	46	18	i	i	PRON
ajst-2553	46	19	rx	rx	VERB
ajst-2553	46	20			NOUN
ajst-2553	46	21	,	,	PUNCT
ajst-2553	46	22	and	and	CCONJ
ajst-2553	46	23	l	l	NOUN
ajst-2553	46	24	represents	represent	VERB
ajst-2553	46	25	the	the	DET
ajst-2553	46	26	dimension	dimension	NOUN
ajst-2553	46	27	of	of	ADP
ajst-2553	46	28	the	the	DET
ajst-2553	46	29	vector	vector	NOUN
ajst-2553	46	30	.	.	PUNCT
ajst-2553	47	1	the	the	DET
ajst-2553	47	2	coding	code	VERB
ajst-2553	47	3	function	function	NOUN
ajst-2553	47	4	can	can	AUX
ajst-2553	47	5	be	be	AUX
ajst-2553	47	6	determined	determine	VERB
ajst-2553	47	7	as	as	ADP
ajst-2553	47	8	:	:	PUNCT
ajst-2553	47	9	nibhwh	nibhwh	NOUN
ajst-2553	47	10	iiii	iiii	NOUN
ajst-2553	47	11	...	...	PUNCT
ajst-2553	47	12	,	,	PUNCT
ajst-2553	47	13	2,1	2,1	NUM
ajst-2553	47	14	)	)	PUNCT
ajst-2553	47	15	,	,	PUNCT
ajst-2553	47	16	(	(	PUNCT
ajst-2553	47	17			X
ajst-2553	47	18	(	(	PUNCT
ajst-2553	47	19	6	6	NUM
ajst-2553	47	20	)	)	PUNCT
ajst-2553	47	21	the	the	DET
ajst-2553	47	22	dac	dac	PROPN
ajst-2553	47	23	ae	ae	PROPN
ajst-2553	47	24	model	model	PROPN
ajst-2553	47	25	proposed	propose	VERB
ajst-2553	47	26	in	in	ADP
ajst-2553	47	27	this	this	DET
ajst-2553	47	28	paper	paper	NOUN
ajst-2553	47	29	does	do	AUX
ajst-2553	47	30	not	not	PART
ajst-2553	47	31	have	have	VERB
ajst-2553	47	32	a	a	DET
ajst-2553	47	33	decoder	decoder	NOUN
ajst-2553	47	34	.	.	PUNCT
ajst-2553	48	1	after	after	SCONJ
ajst-2553	48	2	the	the	DET
ajst-2553	48	3	conversion	conversion	NOUN
ajst-2553	48	4	between	between	ADP
ajst-2553	48	5	the	the	DET
ajst-2553	48	6	n	n	CCONJ
ajst-2553	48	7	-	-	PUNCT
ajst-2553	48	8	layer	layer	NOUN
ajst-2553	48	9	encoder	encoder	NOUN
ajst-2553	48	10	hidden	hide	VERB
ajst-2553	48	11	layer	layer	NOUN
ajst-2553	48	12	and	and	CCONJ
ajst-2553	48	13	the	the	DET
ajst-2553	48	14	sigmoid	sigmoid	NOUN
ajst-2553	48	15	activation	activation	NOUN
ajst-2553	48	16	function	function	NOUN
ajst-2553	48	17	,	,	PUNCT
ajst-2553	48	18	the	the	DET
ajst-2553	48	19	output	output	NOUN
ajst-2553	48	20	data	datum	NOUN
ajst-2553	48	21	can	can	AUX
ajst-2553	48	22	be	be	AUX
ajst-2553	48	23	expressed	express	VERB
ajst-2553	48	24	as	as	ADP
ajst-2553	48	25	equation	equation	NOUN
ajst-2553	48	26	(	(	PUNCT
ajst-2553	48	27	4.6	4.6	NUM
ajst-2553	48	28	)	)	PUNCT
ajst-2553	48	29	nibhwh	nibhwh	NOUN
ajst-2553	48	30	ninnn	ninnn	ADV
ajst-2553	48	31	...	...	PUNCT
ajst-2553	48	32	,	,	PUNCT
ajst-2553	48	33	2,1	2,1	NUM
ajst-2553	48	34	)	)	PUNCT
ajst-2553	48	35	,	,	PUNCT
ajst-2553	48	36	(	(	PUNCT
ajst-2553	48	37			X
ajst-2553	48	38			NOUN
ajst-2553	48	39	(	(	PUNCT
ajst-2553	48	40	7	7	NUM
ajst-2553	48	41	)	)	PUNCT
ajst-2553	48	42	in	in	ADP
ajst-2553	48	43	the	the	DET
ajst-2553	48	44	unsupervised	unsupervised	ADJ
ajst-2553	48	45	learning	learning	NOUN
ajst-2553	48	46	process	process	NOUN
ajst-2553	48	47	,	,	PUNCT
ajst-2553	48	48	back	back	ADJ
ajst-2553	48	49	propagation	propagation	NOUN
ajst-2553	48	50	is	be	AUX
ajst-2553	48	51	used	use	VERB
ajst-2553	48	52	for	for	ADP
ajst-2553	48	53	error	error	NOUN
ajst-2553	48	54	adjustment	adjustment	NOUN
ajst-2553	48	55	.	.	PUNCT
ajst-2553	49	1	finally	finally	ADV
ajst-2553	49	2	,	,	PUNCT
ajst-2553	49	3	the	the	DET
ajst-2553	49	4	reconstruction	reconstruction	NOUN
ajst-2553	49	5	error	error	NOUN
ajst-2553	49	6	generated	generate	VERB
ajst-2553	49	7	by	by	ADP
ajst-2553	49	8	dacae	dacae	PROPN
ajst-2553	49	9	can	can	AUX
ajst-2553	49	10	be	be	AUX
ajst-2553	49	11	expressed	express	VERB
ajst-2553	49	12	as	as	ADP
ajst-2553	49	13	equation	equation	NOUN
ajst-2553	49	14	(	(	PUNCT
ajst-2553	49	15	7	7	NUM
ajst-2553	49	16	):	):	PUNCT
ajst-2553	49	17			NOUN
ajst-2553	49	18			SYM
ajst-2553	49	19			NOUN
ajst-2553	49	20			PROPN
ajst-2553	49	21			NUM
ajst-2553	49	22			PROPN
ajst-2553	49	23	m	m	VERB
ajst-2553	49	24	i	i	NOUN
ajst-2553	49	25	ii	ii	NOUN
ajst-2553	49	26	yx	yx	INTJ
ajst-2553	49	27	m	m	PROPN
ajst-2553	49	28	e	e	NOUN
ajst-2553	49	29	1	1	NUM
ajst-2553	49	30	2	2	NUM
ajst-2553	49	31	2	2	NUM
ajst-2553	49	32	1	1	NOUN
ajst-2553	49	33	(	(	PUNCT
ajst-2553	49	34	8)	8)	NUM
ajst-2553	49	35	2.4	2.4	NUM
ajst-2553	49	36	.	.	PUNCT
ajst-2553	50	1	random	random	ADJ
ajst-2553	50	2	forest	forest	NOUN
ajst-2553	50	3	random	random	ADJ
ajst-2553	50	4	forest	forest	NOUN
ajst-2553	50	5	is	be	AUX
ajst-2553	50	6	a	a	DET
ajst-2553	50	7	classifier	classifier	NOUN
ajst-2553	50	8	that	that	PRON
ajst-2553	50	9	uses	use	VERB
ajst-2553	50	10	multiple	multiple	ADJ
ajst-2553	50	11	decision	decision	NOUN
ajst-2553	50	12	trees	tree	NOUN
ajst-2553	50	13	to	to	PART
ajst-2553	50	14	train	train	VERB
ajst-2553	50	15	and	and	CCONJ
ajst-2553	50	16	predict	predict	VERB
ajst-2553	50	17	samples	sample	NOUN
ajst-2553	50	18	.	.	PUNCT
ajst-2553	51	1	it	it	PRON
ajst-2553	51	2	contains	contain	VERB
ajst-2553	51	3	multiple	multiple	ADJ
ajst-2553	51	4	decision	decision	NOUN
ajst-2553	51	5	tree	tree	NOUN
ajst-2553	51	6	classifiers	classifier	NOUN
ajst-2553	51	7	,	,	PUNCT
ajst-2553	51	8	and	and	CCONJ
ajst-2553	51	9	the	the	DET
ajst-2553	51	10	output	output	NOUN
ajst-2553	51	11	categories	category	NOUN
ajst-2553	51	12	are	be	AUX
ajst-2553	51	13	determined	determine	VERB
ajst-2553	51	14	by	by	ADP
ajst-2553	51	15	the	the	DET
ajst-2553	51	16	mode	mode	NOUN
ajst-2553	51	17	of	of	ADP
ajst-2553	51	18	the	the	DET
ajst-2553	51	19	categories	category	NOUN
ajst-2553	51	20	output	output	NOUN
ajst-2553	51	21	by	by	ADP
ajst-2553	51	22	individual	individual	ADJ
ajst-2553	51	23	trees	tree	NOUN
ajst-2553	51	24	.	.	PUNCT
ajst-2553	52	1	random	random	ADJ
ajst-2553	52	2	forest	forest	NOUN
ajst-2553	52	3	is	be	AUX
ajst-2553	52	4	a	a	PRON
ajst-2553	52	5	flexible	flexible	ADJ
ajst-2553	52	6	and	and	CCONJ
ajst-2553	52	7	easy	easy	ADJ
ajst-2553	52	8	to	to	PART
ajst-2553	52	9	use	use	VERB
ajst-2553	52	10	machine	machine	NOUN
ajst-2553	52	11	learning	learning	NOUN
ajst-2553	52	12	algorithm	algorithm	NOUN
ajst-2553	52	13	.	.	PUNCT
ajst-2553	53	1	even	even	ADV
ajst-2553	53	2	without	without	ADP
ajst-2553	53	3	super	super	ADJ
ajst-2553	53	4	parameter	parameter	NOUN
ajst-2553	53	5	tuning	tuning	NOUN
ajst-2553	53	6	,	,	PUNCT
ajst-2553	53	7	it	it	PRON
ajst-2553	53	8	can	can	AUX
ajst-2553	53	9	get	get	VERB
ajst-2553	53	10	good	good	ADJ
ajst-2553	53	11	results	result	NOUN
ajst-2553	53	12	in	in	ADP
ajst-2553	53	13	most	most	ADJ
ajst-2553	53	14	cases	case	NOUN
ajst-2553	53	15	.	.	PUNCT
ajst-2553	54	1	random	random	ADJ
ajst-2553	54	2	forest	forest	NOUN
ajst-2553	54	3	is	be	AUX
ajst-2553	54	4	also	also	ADV
ajst-2553	54	5	one	one	NUM
ajst-2553	54	6	of	of	ADP
ajst-2553	54	7	the	the	DET
ajst-2553	54	8	most	most	ADV
ajst-2553	54	9	commonly	commonly	ADV
ajst-2553	54	10	used	use	VERB
ajst-2553	54	11	algorithms	algorithm	NOUN
ajst-2553	54	12	,	,	PUNCT
ajst-2553	54	13	because	because	SCONJ
ajst-2553	54	14	it	it	PRON
ajst-2553	54	15	is	be	AUX
ajst-2553	54	16	simple	simple	ADJ
ajst-2553	54	17	and	and	CCONJ
ajst-2553	54	18	can	can	AUX
ajst-2553	54	19	be	be	AUX
ajst-2553	54	20	used	use	VERB
ajst-2553	54	21	for	for	ADP
ajst-2553	54	22	both	both	CCONJ
ajst-2553	54	23	classification	classification	NOUN
ajst-2553	54	24	and	and	CCONJ
ajst-2553	54	25	regression	regression	NOUN
ajst-2553	54	26	.	.	PUNCT
ajst-2553	55	1	random	random	ADJ
ajst-2553	55	2	forest	forest	NOUN
ajst-2553	55	3	integrates	integrate	NOUN
ajst-2553	55	4	all	all	DET
ajst-2553	55	5	classified	classified	ADJ
ajst-2553	55	6	voting	voting	NOUN
ajst-2553	55	7	results	result	NOUN
ajst-2553	55	8	,	,	PUNCT
ajst-2553	55	9	and	and	CCONJ
ajst-2553	55	10	specifies	specify	VERB
ajst-2553	55	11	the	the	DET
ajst-2553	55	12	category	category	NOUN
ajst-2553	55	13	with	with	ADP
ajst-2553	55	14	the	the	DET
ajst-2553	55	15	most	most	ADJ
ajst-2553	55	16	votes	vote	NOUN
ajst-2553	55	17	as	as	ADP
ajst-2553	55	18	the	the	DET
ajst-2553	55	19	final	final	ADJ
ajst-2553	55	20	output	output	NOUN
ajst-2553	55	21	,	,	PUNCT
ajst-2553	55	22	which	which	PRON
ajst-2553	55	23	is	be	AUX
ajst-2553	55	24	the	the	DET
ajst-2553	55	25	simplest	simple	ADJ
ajst-2553	55	26	bagging	bagging	NOUN
ajst-2553	55	27	idea	idea	NOUN
ajst-2553	55	28	.	.	PUNCT
ajst-2553	56	1	rf	rf	VERB
ajst-2553	56	2	basic	basic	ADJ
ajst-2553	56	3	construction	construction	NOUN
ajst-2553	56	4	algorithm	algorithm	NOUN
ajst-2553	56	5	process	process	NOUN
ajst-2553	56	6	:	:	PUNCT
ajst-2553	56	7	1	1	NUM
ajst-2553	56	8	)	)	PUNCT
ajst-2553	56	9	.	.	PUNCT
ajst-2553	57	1	n	n	PROPN
ajst-2553	57	2	represents	represent	VERB
ajst-2553	57	3	the	the	DET
ajst-2553	57	4	number	number	NOUN
ajst-2553	57	5	of	of	ADP
ajst-2553	57	6	training	training	NOUN
ajst-2553	57	7	cases	case	NOUN
ajst-2553	57	8	(	(	PUNCT
ajst-2553	57	9	samples	sample	NOUN
ajst-2553	57	10	)	)	PUNCT
ajst-2553	57	11	,	,	PUNCT
ajst-2553	57	12	and	and	CCONJ
ajst-2553	57	13	m	m	VERB
ajst-2553	57	14	represents	represent	VERB
ajst-2553	57	15	the	the	DET
ajst-2553	57	16	number	number	NOUN
ajst-2553	57	17	of	of	ADP
ajst-2553	57	18	features	feature	NOUN
ajst-2553	57	19	.	.	PUNCT
ajst-2553	58	1	2	2	NUM
ajst-2553	58	2	)	)	PUNCT
ajst-2553	58	3	.	.	PUNCT
ajst-2553	59	1	input	input	VERB
ajst-2553	59	2	the	the	DET
ajst-2553	59	3	feature	feature	NOUN
ajst-2553	59	4	number	number	NOUN
ajst-2553	59	5	m	m	VERB
ajst-2553	59	6	to	to	PART
ajst-2553	59	7	determine	determine	VERB
ajst-2553	59	8	the	the	DET
ajst-2553	59	9	decision	decision	NOUN
ajst-2553	59	10	result	result	NOUN
ajst-2553	59	11	of	of	ADP
ajst-2553	59	12	a	a	DET
ajst-2553	59	13	node	node	NOUN
ajst-2553	59	14	on	on	ADP
ajst-2553	59	15	the	the	DET
ajst-2553	59	16	decision	decision	NOUN
ajst-2553	59	17	tree	tree	NOUN
ajst-2553	59	18	;	;	PUNCT
ajst-2553	59	19	where	where	SCONJ
ajst-2553	59	20	m	m	PROPN
ajst-2553	59	21	should	should	AUX
ajst-2553	59	22	be	be	AUX
ajst-2553	59	23	much	much	ADV
ajst-2553	59	24	less	less	ADJ
ajst-2553	59	25	than	than	ADP
ajst-2553	59	26	m.	m.	NOUN
ajst-2553	59	27	3	3	NUM
ajst-2553	59	28	)	)	PUNCT
ajst-2553	59	29	.	.	PUNCT
ajst-2553	60	1	take	take	VERB
ajst-2553	60	2	n	n	PRON
ajst-2553	60	3	samples	sample	NOUN
ajst-2553	60	4	from	from	ADP
ajst-2553	60	5	n	n	DET
ajst-2553	60	6	training	training	NOUN
ajst-2553	60	7	cases	case	NOUN
ajst-2553	60	8	(	(	PUNCT
ajst-2553	60	9	samples	sample	NOUN
ajst-2553	60	10	)	)	PUNCT
ajst-2553	60	11	in	in	ADP
ajst-2553	60	12	the	the	DET
ajst-2553	60	13	way	way	NOUN
ajst-2553	60	14	of	of	ADP
ajst-2553	60	15	put	put	NOUN
ajst-2553	60	16	back	back	ADV
ajst-2553	60	17	sampling	sample	VERB
ajst-2553	60	18	to	to	PART
ajst-2553	60	19	form	form	VERB
ajst-2553	60	20	a	a	DET
ajst-2553	60	21	training	training	NOUN
ajst-2553	60	22	set	set	NOUN
ajst-2553	60	23	(	(	PUNCT
ajst-2553	60	24	bootstrap	bootstrap	NOUN
ajst-2553	60	25	sampling	sampling	NOUN
ajst-2553	60	26	)	)	PUNCT
ajst-2553	60	27	,	,	PUNCT
ajst-2553	60	28	and	and	CCONJ
ajst-2553	60	29	use	use	VERB
ajst-2553	60	30	the	the	DET
ajst-2553	60	31	unselected	unselected	ADJ
ajst-2553	60	32	cases	case	NOUN
ajst-2553	60	33	(	(	PUNCT
ajst-2553	60	34	samples	sample	NOUN
ajst-2553	60	35	)	)	PUNCT
ajst-2553	60	36	for	for	ADP
ajst-2553	60	37	prediction	prediction	NOUN
ajst-2553	60	38	to	to	PART
ajst-2553	60	39	evaluate	evaluate	VERB
ajst-2553	60	40	the	the	DET
ajst-2553	60	41	error	error	NOUN
ajst-2553	60	42	.	.	PUNCT
ajst-2553	60	43	4	4	NUM
ajst-2553	60	44	)	)	PUNCT
ajst-2553	60	45	.	.	PUNCT
ajst-2553	61	1	for	for	ADP
ajst-2553	61	2	each	each	DET
ajst-2553	61	3	node	node	NOUN
ajst-2553	61	4	,	,	PUNCT
ajst-2553	61	5	m	m	VERB
ajst-2553	61	6	features	feature	NOUN
ajst-2553	61	7	are	be	AUX
ajst-2553	61	8	randomly	randomly	ADV
ajst-2553	61	9	selected	select	VERB
ajst-2553	61	10	,	,	PUNCT
ajst-2553	61	11	and	and	CCONJ
ajst-2553	61	12	the	the	DET
ajst-2553	61	13	decision	decision	NOUN
ajst-2553	61	14	of	of	ADP
ajst-2553	61	15	each	each	DET
ajst-2553	61	16	node	node	NOUN
ajst-2553	61	17	on	on	ADP
ajst-2553	61	18	the	the	DET
ajst-2553	61	19	decision	decision	NOUN
ajst-2553	61	20	tree	tree	NOUN
ajst-2553	61	21	is	be	AUX
ajst-2553	61	22	determined	determine	VERB
ajst-2553	61	23	based	base	VERB
ajst-2553	61	24	on	on	ADP
ajst-2553	61	25	these	these	DET
ajst-2553	61	26	features	feature	NOUN
ajst-2553	61	27	.	.	PUNCT
ajst-2553	62	1	according	accord	VERB
ajst-2553	62	2	to	to	ADP
ajst-2553	62	3	the	the	DET
ajst-2553	62	4	m	m	NOUN
ajst-2553	62	5	features	feature	NOUN
ajst-2553	62	6	,	,	PUNCT
ajst-2553	62	7	calculate	calculate	VERB
ajst-2553	62	8	the	the	DET
ajst-2553	62	9	optimal	optimal	ADJ
ajst-2553	62	10	splitting	splitting	NOUN
ajst-2553	62	11	method	method	NOUN
ajst-2553	62	12	.	.	PUNCT
ajst-2553	63	1	5	5	NUM
ajst-2553	63	2	)	)	PUNCT
ajst-2553	63	3	.	.	PUNCT
ajst-2553	64	1	each	each	DET
ajst-2553	64	2	tree	tree	NOUN
ajst-2553	64	3	grows	grow	VERB
ajst-2553	64	4	completely	completely	ADV
ajst-2553	64	5	without	without	ADP
ajst-2553	64	6	pruning	prune	VERB
ajst-2553	64	7	,	,	PUNCT
ajst-2553	64	8	which	which	PRON
ajst-2553	64	9	may	may	AUX
ajst-2553	64	10	be	be	AUX
ajst-2553	64	11	used	use	VERB
ajst-2553	64	12	after	after	ADP
ajst-2553	64	13	building	build	VERB
ajst-2553	64	14	a	a	DET
ajst-2553	64	15	normal	normal	ADJ
ajst-2553	64	16	tree	tree	NOUN
ajst-2553	64	17	classifier	classifier	NOUN
ajst-2553	64	18	)	)	PUNCT
ajst-2553	64	19	.	.	PUNCT
ajst-2553	65	1	2.5	2.5	NUM
ajst-2553	65	2	.	.	PUNCT
ajst-2553	66	1	establishment	establishment	NOUN
ajst-2553	66	2	of	of	ADP
ajst-2553	66	3	improved	improved	ADJ
ajst-2553	66	4	deep	deep	ADJ
ajst-2553	66	5	learning	learning	NOUN
ajst-2553	66	6	model	model	NOUN
ajst-2553	66	7	the	the	DET
ajst-2553	66	8	combination	combination	NOUN
ajst-2553	66	9	of	of	ADP
ajst-2553	66	10	deep	deep	ADJ
ajst-2553	66	11	learning	learning	NOUN
ajst-2553	66	12	model	model	NOUN
ajst-2553	66	13	and	and	CCONJ
ajst-2553	66	14	shallow	shallow	ADJ
ajst-2553	66	15	classification	classification	NOUN
ajst-2553	66	16	learning	learning	NOUN
ajst-2553	66	17	model	model	NOUN
ajst-2553	66	18	can	can	AUX
ajst-2553	66	19	better	well	ADV
ajst-2553	66	20	improve	improve	VERB
ajst-2553	66	21	the	the	DET
ajst-2553	66	22	accuracy	accuracy	NOUN
ajst-2553	66	23	of	of	ADP
ajst-2553	66	24	classification	classification	NOUN
ajst-2553	66	25	detection	detection	NOUN
ajst-2553	66	26	.	.	PUNCT
ajst-2553	67	1	the	the	DET
ajst-2553	67	2	overall	overall	ADJ
ajst-2553	67	3	framework	framework	NOUN
ajst-2553	67	4	of	of	ADP
ajst-2553	67	5	the	the	DET
ajst-2553	67	6	dacae	dacae	NOUN
ajst-2553	67	7	-	-	PUNCT
ajst-2553	67	8	rf	rf	NOUN
ajst-2553	67	9	based	base	VERB
ajst-2553	67	10	detection	detection	NOUN
ajst-2553	67	11	model	model	NOUN
ajst-2553	67	12	includes	include	VERB
ajst-2553	67	13	three	three	NUM
ajst-2553	67	14	modules	module	NOUN
ajst-2553	67	15	:	:	PUNCT
ajst-2553	67	16	data	datum	NOUN
ajst-2553	67	17	preprocessing	preprocessing	NOUN
ajst-2553	67	18	module	module	NOUN
ajst-2553	67	19	,	,	PUNCT
ajst-2553	67	20	depth	depth	NOUN
ajst-2553	67	21	asymmetric	asymmetric	ADJ
ajst-2553	67	22	convolutional	convolutional	ADJ
ajst-2553	67	23	encoder	encoder	NOUN
ajst-2553	67	24	training	training	NOUN
ajst-2553	67	25	module	module	NOUN
ajst-2553	67	26	,	,	PUNCT
ajst-2553	67	27	and	and	CCONJ
ajst-2553	67	28	intrusion	intrusion	NOUN
ajst-2553	67	29	detection	detection	NOUN
ajst-2553	67	30	module	module	NOUN
ajst-2553	67	31	.	.	PUNCT
ajst-2553	68	1	the	the	DET
ajst-2553	68	2	proposed	propose	VERB
ajst-2553	68	3	model	model	NOUN
ajst-2553	68	4	is	be	AUX
ajst-2553	68	5	shown	show	VERB
ajst-2553	68	6	in	in	ADP
ajst-2553	68	7	figure	figure	NOUN
ajst-2553	68	8	4.3	4.3	NUM
ajst-2553	68	9	.	.	PUNCT
ajst-2553	69	1	it	it	PRON
ajst-2553	69	2	uses	use	VERB
ajst-2553	69	3	dacae	dacae	NOUN
ajst-2553	69	4	to	to	PART
ajst-2553	69	5	extract	extract	VERB
ajst-2553	69	6	features	feature	NOUN
ajst-2553	69	7	from	from	ADP
ajst-2553	69	8	the	the	DET
ajst-2553	69	9	preprocessed	preprocesse	VERB
ajst-2553	69	10	data	datum	NOUN
ajst-2553	69	11	,	,	PUNCT
ajst-2553	69	12	and	and	CCONJ
ajst-2553	69	13	uses	use	VERB
ajst-2553	69	14	the	the	DET
ajst-2553	69	15	random	random	ADJ
ajst-2553	69	16	forest	forest	NOUN
ajst-2553	69	17	algorithm	algorithm	NOUN
ajst-2553	69	18	to	to	PART
ajst-2553	69	19	divide	divide	VERB
ajst-2553	69	20	the	the	DET
ajst-2553	69	21	network	network	NOUN
ajst-2553	69	22	traffic	traffic	NOUN
ajst-2553	69	23	data	datum	NOUN
ajst-2553	69	24	into	into	ADP
ajst-2553	69	25	normal	normal	ADJ
ajst-2553	69	26	classes	class	NOUN
ajst-2553	69	27	and	and	CCONJ
ajst-2553	69	28	abnormal	abnormal	ADJ
ajst-2553	69	29	classes	class	NOUN
ajst-2553	69	30	by	by	ADP
ajst-2553	69	31	generating	generate	VERB
ajst-2553	69	32	new	new	ADJ
ajst-2553	69	33	abstract	abstract	ADJ
ajst-2553	69	34	features	feature	NOUN
ajst-2553	69	35	.	.	PUNCT
ajst-2553	70	1	for	for	ADP
ajst-2553	70	2	the	the	DET
ajst-2553	70	3	current	current	ADJ
ajst-2553	70	4	data	datum	NOUN
ajst-2553	70	5	and	and	CCONJ
ajst-2553	70	6	models	model	NOUN
ajst-2553	70	7	,	,	PUNCT
ajst-2553	70	8	there	there	PRON
ajst-2553	70	9	are	be	VERB
ajst-2553	70	10	some	some	DET
ajst-2553	70	11	unavoidable	unavoidable	ADJ
ajst-2553	70	12	problems	problem	NOUN
ajst-2553	70	13	,	,	PUNCT
ajst-2553	70	14	such	such	ADJ
ajst-2553	70	15	as	as	ADP
ajst-2553	70	16	the	the	DET
ajst-2553	70	17	scarcity	scarcity	NOUN
ajst-2553	70	18	of	of	ADP
ajst-2553	70	19	labeled	label	VERB
ajst-2553	70	20	data	datum	NOUN
ajst-2553	70	21	resources	resource	NOUN
ajst-2553	70	22	,	,	PUNCT
ajst-2553	70	23	and	and	CCONJ
ajst-2553	70	24	the	the	DET
ajst-2553	70	25	problem	problem	NOUN
ajst-2553	70	26	of	of	ADP
ajst-2553	70	27	gradient	gradient	ADJ
ajst-2553	70	28	dispersion	dispersion	NOUN
ajst-2553	70	29	is	be	AUX
ajst-2553	70	30	prone	prone	ADJ
ajst-2553	70	31	to	to	PART
ajst-2553	70	32	occur	occur	VERB
ajst-2553	70	33	in	in	ADP
ajst-2553	70	34	deep	deep	ADJ
ajst-2553	70	35	neural	neural	ADJ
ajst-2553	70	36	networks	network	NOUN
ajst-2553	70	37	.	.	PUNCT
ajst-2553	71	1	therefore	therefore	ADV
ajst-2553	71	2	,	,	PUNCT
ajst-2553	71	3	using	use	VERB
ajst-2553	71	4	an	an	DET
ajst-2553	71	5	75	75	NUM
ajst-2553	71	6	unsupervised	unsupervised	ADJ
ajst-2553	71	7	asymmetric	asymmetric	ADJ
ajst-2553	71	8	convolutional	convolutional	ADJ
ajst-2553	71	9	encoder	encoder	NOUN
ajst-2553	71	10	can	can	AUX
ajst-2553	71	11	avoid	avoid	VERB
ajst-2553	71	12	this	this	DET
ajst-2553	71	13	problem	problem	NOUN
ajst-2553	71	14	.	.	PUNCT
ajst-2553	72	1	(	(	PUNCT
ajst-2553	72	2	1	1	X
ajst-2553	72	3	)	)	PUNCT
ajst-2553	72	4	in	in	ADP
ajst-2553	72	5	the	the	DET
ajst-2553	72	6	data	datum	NOUN
ajst-2553	72	7	preprocessing	preprocessing	NOUN
ajst-2553	72	8	module	module	NOUN
ajst-2553	72	9	,	,	PUNCT
ajst-2553	72	10	the	the	DET
ajst-2553	72	11	intrusion	intrusion	NOUN
ajst-2553	72	12	detection	detection	NOUN
ajst-2553	72	13	model	model	NOUN
ajst-2553	72	14	will	will	AUX
ajst-2553	72	15	be	be	AUX
ajst-2553	72	16	tested	test	VERB
ajst-2553	72	17	with	with	ADP
ajst-2553	72	18	two	two	NUM
ajst-2553	72	19	datasets	dataset	NOUN
ajst-2553	72	20	,	,	PUNCT
ajst-2553	72	21	the	the	DET
ajst-2553	72	22	kd99based	kd99base	VERB
ajst-2553	72	23	nsl	nsl	PROPN
ajst-2553	72	24	-	-	PUNCT
ajst-2553	72	25	kdd	kdd	PROPN
ajst-2553	72	26	dataset	dataset	NOUN
ajst-2553	72	27	and	and	CCONJ
ajst-2553	72	28	the	the	DET
ajst-2553	72	29	nsl	nsl	PROPN
ajst-2553	72	30	-	-	PUNCT
ajst-2553	72	31	kdd	kdd	PROPN
ajst-2553	72	32	dataset	dataset	NOUN
ajst-2553	72	33	.	.	PUNCT
ajst-2553	73	1	the	the	DET
ajst-2553	73	2	preprocessing	preprocessing	NOUN
ajst-2553	73	3	of	of	ADP
ajst-2553	73	4	the	the	DET
ajst-2553	73	5	dataset	dataset	NOUN
ajst-2553	73	6	includes	include	VERB
ajst-2553	73	7	removing	remove	VERB
ajst-2553	73	8	socket	socket	NOUN
ajst-2553	73	9	information	information	NOUN
ajst-2553	73	10	,	,	PUNCT
ajst-2553	73	11	label	label	NOUN
ajst-2553	73	12	encoding	encoding	NOUN
ajst-2553	73	13	and	and	CCONJ
ajst-2553	73	14	data	datum	NOUN
ajst-2553	73	15	normalization	normalization	NOUN
ajst-2553	73	16	,	,	PUNCT
ajst-2553	73	17	etc	etc	X
ajst-2553	73	18	.	.	X
ajst-2553	73	19	among	among	ADP
ajst-2553	73	20	them	they	PRON
ajst-2553	73	21	,	,	PUNCT
ajst-2553	73	22	the	the	DET
ajst-2553	73	23	1d	1d	NUM
ajst-2553	73	24	data	datum	NOUN
ajst-2553	73	25	of	of	ADP
ajst-2553	73	26	41	41	NUM
ajst-2553	73	27	features	feature	NOUN
ajst-2553	73	28	is	be	AUX
ajst-2553	73	29	filled	fill	VERB
ajst-2553	73	30	with	with	ADP
ajst-2553	73	31	0	0	NUM
ajst-2553	73	32	to	to	ADP
ajst-2553	73	33	64	64	NUM
ajst-2553	73	34	features	feature	NOUN
ajst-2553	73	35	,	,	PUNCT
ajst-2553	73	36	and	and	CCONJ
ajst-2553	73	37	then	then	ADV
ajst-2553	73	38	converted	convert	VERB
ajst-2553	73	39	into	into	ADP
ajst-2553	73	40	8×8	8×8	NUM
ajst-2553	73	41	2d	2d	NUM
ajst-2553	73	42	data	datum	NOUN
ajst-2553	73	43	.	.	PUNCT
ajst-2553	74	1	(	(	PUNCT
ajst-2553	74	2	2	2	X
ajst-2553	74	3	)	)	PUNCT
ajst-2553	74	4	in	in	ADP
ajst-2553	74	5	the	the	DET
ajst-2553	74	6	deep	deep	ADJ
ajst-2553	74	7	asymmetric	asymmetric	ADJ
ajst-2553	74	8	convolutional	convolutional	ADJ
ajst-2553	74	9	encoder	encoder	NOUN
ajst-2553	74	10	module	module	NOUN
ajst-2553	74	11	,	,	PUNCT
ajst-2553	74	12	the	the	DET
ajst-2553	74	13	corresponding	correspond	VERB
ajst-2553	74	14	encoder	encoder	NOUN
ajst-2553	74	15	structure	structure	NOUN
ajst-2553	74	16	is	be	AUX
ajst-2553	74	17	established	establish	VERB
ajst-2553	74	18	through	through	ADP
ajst-2553	74	19	multi	multi	ADJ
ajst-2553	74	20	-	-	ADJ
ajst-2553	74	21	layer	layer	ADJ
ajst-2553	74	22	convolutional	convolutional	ADJ
ajst-2553	74	23	layers	layer	NOUN
ajst-2553	74	24	and	and	CCONJ
ajst-2553	74	25	pooling	pool	VERB
ajst-2553	74	26	layers	layer	NOUN
ajst-2553	74	27	,	,	PUNCT
ajst-2553	74	28	and	and	CCONJ
ajst-2553	74	29	unsupervised	unsupervised	ADJ
ajst-2553	74	30	pre	pre	ADJ
ajst-2553	74	31	-	-	NOUN
ajst-2553	74	32	training	training	NOUN
ajst-2553	74	33	is	be	AUX
ajst-2553	74	34	performed	perform	VERB
ajst-2553	74	35	for	for	ADP
ajst-2553	74	36	weight	weight	NOUN
ajst-2553	74	37	adjustment	adjustment	NOUN
ajst-2553	74	38	.	.	PUNCT
ajst-2553	75	1	(	(	PUNCT
ajst-2553	75	2	3	3	X
ajst-2553	75	3	)	)	PUNCT
ajst-2553	75	4	after	after	SCONJ
ajst-2553	75	5	the	the	DET
ajst-2553	75	6	first	first	ADJ
ajst-2553	75	7	two	two	NUM
ajst-2553	75	8	modules	module	NOUN
ajst-2553	75	9	preprocess	preprocess	VERB
ajst-2553	75	10	the	the	DET
ajst-2553	75	11	data	datum	NOUN
ajst-2553	75	12	and	and	CCONJ
ajst-2553	75	13	perform	perform	VERB
ajst-2553	75	14	feature	feature	NOUN
ajst-2553	75	15	extraction	extraction	NOUN
ajst-2553	75	16	on	on	ADP
ajst-2553	75	17	the	the	DET
ajst-2553	75	18	neural	neural	ADJ
ajst-2553	75	19	network	network	NOUN
ajst-2553	75	20	,	,	PUNCT
ajst-2553	75	21	new	new	ADJ
ajst-2553	75	22	abstract	abstract	ADJ
ajst-2553	75	23	feature	feature	NOUN
ajst-2553	75	24	data	datum	NOUN
ajst-2553	75	25	is	be	AUX
ajst-2553	75	26	obtained	obtain	VERB
ajst-2553	75	27	and	and	CCONJ
ajst-2553	75	28	the	the	DET
ajst-2553	75	29	new	new	ADJ
ajst-2553	75	30	data	datum	NOUN
ajst-2553	75	31	is	be	AUX
ajst-2553	75	32	used	use	VERB
ajst-2553	75	33	as	as	ADP
ajst-2553	75	34	the	the	DET
ajst-2553	75	35	training	training	NOUN
ajst-2553	75	36	data	datum	NOUN
ajst-2553	75	37	of	of	ADP
ajst-2553	75	38	the	the	DET
ajst-2553	75	39	random	random	ADJ
ajst-2553	75	40	forest	forest	NOUN
ajst-2553	75	41	.	.	PUNCT
ajst-2553	76	1	random	random	ADJ
ajst-2553	76	2	forest	forest	NOUN
ajst-2553	76	3	makes	make	VERB
ajst-2553	76	4	judgments	judgment	NOUN
ajst-2553	76	5	and	and	CCONJ
ajst-2553	76	6	identifies	identify	VERB
ajst-2553	76	7	the	the	DET
ajst-2553	76	8	types	type	NOUN
ajst-2553	76	9	of	of	ADP
ajst-2553	76	10	intrusion	intrusion	NOUN
ajst-2553	76	11	data	datum	NOUN
ajst-2553	76	12	to	to	PART
ajst-2553	76	13	achieve	achieve	VERB
ajst-2553	76	14	the	the	DET
ajst-2553	76	15	purpose	purpose	NOUN
ajst-2553	76	16	of	of	ADP
ajst-2553	76	17	intrusion	intrusion	NOUN
ajst-2553	76	18	detection	detection	NOUN
ajst-2553	76	19	.	.	PUNCT
ajst-2553	77	1	figure	figure	NOUN
ajst-2553	77	2	3	3	NUM
ajst-2553	77	3	.	.	PUNCT
ajst-2553	78	1	using	use	VERB
ajst-2553	78	2	dacae	dacae	NOUN
ajst-2553	78	3	-	-	PUNCT
ajst-2553	78	4	rf	rf	NOUN
ajst-2553	78	5	to	to	PART
ajst-2553	78	6	detect	detect	VERB
ajst-2553	78	7	network	network	NOUN
ajst-2553	78	8	traffic	traffic	NOUN
ajst-2553	78	9	intrusion	intrusion	NOUN
ajst-2553	78	10	3	3	NUM
ajst-2553	78	11	.	.	PUNCT
ajst-2553	78	12	experiments	experiment	NOUN
ajst-2553	78	13	in	in	ADP
ajst-2553	78	14	the	the	DET
ajst-2553	78	15	experimental	experimental	ADJ
ajst-2553	78	16	part	part	NOUN
ajst-2553	78	17	,	,	PUNCT
ajst-2553	78	18	the	the	DET
ajst-2553	78	19	dacae	dacae	NOUN
ajst-2553	78	20	-	-	PUNCT
ajst-2553	78	21	rf	rf	NOUN
ajst-2553	78	22	detection	detection	NOUN
ajst-2553	78	23	model	model	NOUN
ajst-2553	78	24	is	be	AUX
ajst-2553	78	25	compared	compare	VERB
ajst-2553	78	26	with	with	ADP
ajst-2553	78	27	the	the	DET
ajst-2553	78	28	model	model	NOUN
ajst-2553	78	29	in	in	ADP
ajst-2553	78	30	research	research	NOUN
ajst-2553	78	31	[	[	X
ajst-2553	78	32	6	6	NUM
ajst-2553	78	33	]	]	PUNCT
ajst-2553	78	34	.	.	PUNCT
ajst-2553	79	1	the	the	DET
ajst-2553	79	2	experimental	experimental	ADJ
ajst-2553	79	3	results	result	NOUN
ajst-2553	79	4	were	be	AUX
ajst-2553	79	5	compared	compare	VERB
ajst-2553	79	6	between	between	ADP
ajst-2553	79	7	nsl	nsl	NOUN
ajst-2553	79	8	-	-	PUNCT
ajst-2553	79	9	kdd	kdd	PROPN
ajst-2553	79	10	and	and	CCONJ
ajst-2553	79	11	kdd99	kdd99	PROPN
ajst-2553	79	12	datasets	dataset	NOUN
ajst-2553	79	13	.	.	PUNCT
ajst-2553	80	1	the	the	DET
ajst-2553	80	2	evaluation	evaluation	NOUN
ajst-2553	80	3	criteria	criterion	NOUN
ajst-2553	80	4	include	include	VERB
ajst-2553	80	5	accuracy	accuracy	NOUN
ajst-2553	80	6	,	,	PUNCT
ajst-2553	80	7	precision	precision	NOUN
ajst-2553	80	8	,	,	PUNCT
ajst-2553	80	9	recall	recall	NOUN
ajst-2553	80	10	and	and	CCONJ
ajst-2553	80	11	1	1	NUM
ajst-2553	80	12	-	-	PUNCT
ajst-2553	80	13	score	score	NOUN
ajst-2553	80	14	.	.	PUNCT
ajst-2553	81	1	the	the	DET
ajst-2553	81	2	proposed	propose	VERB
ajst-2553	81	3	model	model	NOUN
ajst-2553	81	4	and	and	CCONJ
ajst-2553	81	5	the	the	DET
ajst-2553	81	6	detection	detection	NOUN
ajst-2553	81	7	model	model	NOUN
ajst-2553	81	8	mentioned	mention	VERB
ajst-2553	81	9	in	in	ADP
ajst-2553	81	10	the	the	DET
ajst-2553	81	11	study	study	NOUN
ajst-2553	81	12	[	[	X
ajst-2553	81	13	6	6	NUM
ajst-2553	81	14	]	]	PUNCT
ajst-2553	81	15	were	be	AUX
ajst-2553	81	16	compared	compare	VERB
ajst-2553	81	17	and	and	CCONJ
ajst-2553	81	18	analyzed	analyze	VERB
ajst-2553	81	19	,	,	PUNCT
ajst-2553	81	20	as	as	SCONJ
ajst-2553	81	21	shown	show	VERB
ajst-2553	81	22	in	in	ADP
ajst-2553	81	23	table1	table1	PROPN
ajst-2553	81	24	and	and	CCONJ
ajst-2553	81	25	table	table	NOUN
ajst-2553	81	26	2	2	NUM
ajst-2553	81	27	.	.	PUNCT
ajst-2553	81	28	table	table	NOUN
ajst-2553	81	29	1	1	NUM
ajst-2553	81	30	.	.	PUNCT
ajst-2553	82	1	performance	performance	NOUN
ajst-2553	82	2	of	of	ADP
ajst-2553	82	3	dacae	dacae	NOUN
ajst-2553	82	4	-	-	PUNCT
ajst-2553	82	5	rf	rf	NOUN
ajst-2553	82	6	on	on	ADP
ajst-2553	82	7	kdd99	kdd99	PROPN
ajst-2553	82	8	dataset	dataset	NOUN
ajst-2553	82	9	it	it	PRON
ajst-2553	82	10	can	can	AUX
ajst-2553	82	11	be	be	AUX
ajst-2553	82	12	seen	see	VERB
ajst-2553	82	13	from	from	ADP
ajst-2553	82	14	the	the	DET
ajst-2553	82	15	table	table	NOUN
ajst-2553	82	16	1	1	NUM
ajst-2553	82	17	that	that	SCONJ
ajst-2553	82	18	the	the	DET
ajst-2553	82	19	dacae	dacae	NOUN
ajst-2553	82	20	-	-	PUNCT
ajst-2553	82	21	rf	rf	NOUN
ajst-2553	82	22	detection	detection	NOUN
ajst-2553	82	23	model	model	NOUN
ajst-2553	82	24	proposed	propose	VERB
ajst-2553	82	25	in	in	ADP
ajst-2553	82	26	this	this	DET
ajst-2553	82	27	paper	paper	NOUN
ajst-2553	82	28	is	be	AUX
ajst-2553	82	29	better	well	ADJ
ajst-2553	82	30	than	than	ADP
ajst-2553	82	31	s	s	NOUN
ajst-2553	82	32	-	-	NOUN
ajst-2553	82	33	ndae	ndae	NOUN
ajst-2553	82	34	in	in	ADP
ajst-2553	82	35	accuracy	accuracy	NOUN
ajst-2553	82	36	,	,	PUNCT
ajst-2553	82	37	precision	precision	NOUN
ajst-2553	82	38	,	,	PUNCT
ajst-2553	82	39	recall	recall	NOUN
ajst-2553	82	40	rate	rate	NOUN
ajst-2553	82	41	and	and	CCONJ
ajst-2553	82	42	1	1	NUM
ajst-2553	82	43	-	-	PUNCT
ajst-2553	82	44	score	score	NOUN
ajst-2553	82	45	in	in	ADP
ajst-2553	82	46	the	the	DET
ajst-2553	82	47	classification	classification	NOUN
ajst-2553	82	48	of	of	ADP
ajst-2553	82	49	large	large	ADJ
ajst-2553	82	50	-	-	PUNCT
ajst-2553	82	51	scale	scale	NOUN
ajst-2553	82	52	samples	sample	NOUN
ajst-2553	82	53	,	,	PUNCT
ajst-2553	82	54	such	such	ADJ
ajst-2553	82	55	as	as	ADP
ajst-2553	82	56	dos	do	NOUN
ajst-2553	82	57	normal	normal	ADJ
ajst-2553	82	58	,	,	PUNCT
ajst-2553	82	59	probe	probe	NOUN
ajst-2553	82	60	and	and	CCONJ
ajst-2553	82	61	other	other	ADJ
ajst-2553	82	62	attack	attack	NOUN
ajst-2553	82	63	types	type	NOUN
ajst-2553	82	64	.	.	PUNCT
ajst-2553	83	1	s	s	X
ajst-2553	83	2	-	-	PUNCT
ajst-2553	83	3	ndae	ndae	ADJ
ajst-2553	83	4	and	and	CCONJ
ajst-2553	83	5	dacae	dacae	NOUN
ajst-2553	83	6	-	-	PUNCT
ajst-2553	83	7	rf	rf	NOUN
ajst-2553	83	8	also	also	ADV
ajst-2553	83	9	use	use	VERB
ajst-2553	83	10	the	the	DET
ajst-2553	83	11	encoder	encoder	NOUN
ajst-2553	83	12	structure	structure	NOUN
ajst-2553	83	13	to	to	PART
ajst-2553	83	14	extract	extract	VERB
ajst-2553	83	15	the	the	DET
ajst-2553	83	16	features	feature	NOUN
ajst-2553	83	17	of	of	ADP
ajst-2553	83	18	the	the	DET
ajst-2553	83	19	data	datum	NOUN
ajst-2553	83	20	,	,	PUNCT
ajst-2553	83	21	and	and	CCONJ
ajst-2553	83	22	use	use	VERB
ajst-2553	83	23	the	the	DET
ajst-2553	83	24	random	random	ADJ
ajst-2553	83	25	forest	forest	NOUN
ajst-2553	83	26	to	to	PART
ajst-2553	83	27	classify	classify	VERB
ajst-2553	83	28	the	the	DET
ajst-2553	83	29	types	type	NOUN
ajst-2553	83	30	of	of	ADP
ajst-2553	83	31	network	network	NOUN
ajst-2553	83	32	intrusion	intrusion	NOUN
ajst-2553	83	33	.	.	PUNCT
ajst-2553	84	1	the	the	DET
ajst-2553	84	2	experimental	experimental	ADJ
ajst-2553	84	3	results	result	NOUN
ajst-2553	84	4	show	show	VERB
ajst-2553	84	5	that	that	SCONJ
ajst-2553	84	6	the	the	DET
ajst-2553	84	7	model	model	NOUN
ajst-2553	84	8	proposed	propose	VERB
ajst-2553	84	9	in	in	ADP
ajst-2553	84	10	this	this	DET
ajst-2553	84	11	paper	paper	NOUN
ajst-2553	84	12	has	have	VERB
ajst-2553	84	13	better	well	ADJ
ajst-2553	84	14	detection	detection	NOUN
ajst-2553	84	15	results	result	NOUN
ajst-2553	84	16	on	on	ADP
ajst-2553	84	17	anomaly	anomaly	NOUN
ajst-2553	84	18	types	type	NOUN
ajst-2553	84	19	than	than	ADP
ajst-2553	84	20	s	s	NOUN
ajst-2553	84	21	-	-	NOUN
ajst-2553	84	22	ndae	ndae	NOUN
ajst-2553	84	23	.	.	PUNCT
ajst-2553	85	1	the	the	DET
ajst-2553	85	2	experiment	experiment	NOUN
ajst-2553	85	3	also	also	ADV
ajst-2553	85	4	evaluated	evaluate	VERB
ajst-2553	85	5	the	the	DET
ajst-2553	85	6	dacae	dacae	NOUN
ajst-2553	85	7	-	-	PUNCT
ajst-2553	85	8	rf	rf	NOUN
ajst-2553	85	9	model	model	NOUN
ajst-2553	85	10	on	on	ADP
ajst-2553	85	11	the	the	DET
ajst-2553	85	12	nsl	nsl	PROPN
ajst-2553	85	13	-	-	PUNCT
ajst-2553	85	14	kdd	kdd	PROPN
ajst-2553	85	15	data	datum	NOUN
ajst-2553	85	16	set	set	VERB
ajst-2553	85	17	.	.	PUNCT
ajst-2553	86	1	table	table	NOUN
ajst-2553	86	2	2	2	NUM
ajst-2553	86	3	is	be	AUX
ajst-2553	86	4	different	different	ADJ
ajst-2553	86	5	results	result	NOUN
ajst-2553	86	6	of	of	ADP
ajst-2553	86	7	performance	performance	NOUN
ajst-2553	86	8	indicators	indicator	NOUN
ajst-2553	86	9	of	of	ADP
ajst-2553	86	10	dacae	dacae	NOUN
ajst-2553	86	11	-	-	PUNCT
ajst-2553	86	12	rf	rf	NUM
ajst-2553	86	13	intrusion	intrusion	NOUN
ajst-2553	86	14	detection	detection	NOUN
ajst-2553	86	15	classification	classification	NOUN
ajst-2553	86	16	.	.	PUNCT
ajst-2553	87	1	table	table	NOUN
ajst-2553	87	2	2	2	NUM
ajst-2553	87	3	.	.	PUNCT
ajst-2553	87	4	performance	performance	NOUN
ajst-2553	87	5	of	of	ADP
ajst-2553	87	6	dacae	dacae	NOUN
ajst-2553	87	7	-	-	PUNCT
ajst-2553	87	8	rf	rf	NOUN
ajst-2553	87	9	on	on	ADP
ajst-2553	87	10	kdd99	kdd99	NOUN
ajst-2553	87	11	dataset	dataset	VERB
ajst-2553	87	12	in	in	ADP
ajst-2553	87	13	the	the	DET
ajst-2553	87	14	nsl	nsl	NOUN
ajst-2553	87	15	-	-	PUNCT
ajst-2553	87	16	kdd	kdd	PROPN
ajst-2553	87	17	dataset	dataset	NOUN
ajst-2553	88	1	,	,	PUNCT
ajst-2553	88	2	the	the	DET
ajst-2553	88	3	total	total	ADJ
ajst-2553	88	4	accuracy	accuracy	NOUN
ajst-2553	88	5	,	,	PUNCT
ajst-2553	88	6	precision	precision	NOUN
ajst-2553	88	7	,	,	PUNCT
ajst-2553	88	8	recall	recall	VERB
ajst-2553	88	9	and1	and1	PROPN
ajst-2553	88	10	-	-	PUNCT
ajst-2553	88	11	score	score	NOUN
ajst-2553	88	12	has	have	VERB
ajst-2553	88	13	some	some	DET
ajst-2553	88	14	improvement	improvement	NOUN
ajst-2553	88	15	compared	compare	VERB
ajst-2553	88	16	with	with	ADP
ajst-2553	88	17	the	the	DET
ajst-2553	88	18	model	model	NOUN
ajst-2553	88	19	proposed	propose	VERB
ajst-2553	88	20	in	in	ADP
ajst-2553	88	21	other	other	ADJ
ajst-2553	88	22	studies	study	NOUN
ajst-2553	88	23	.	.	PUNCT
ajst-2553	89	1	in	in	ADP
ajst-2553	89	2	the	the	DET
ajst-2553	89	3	classification	classification	NOUN
ajst-2553	89	4	of	of	ADP
ajst-2553	89	5	probe	probe	NOUN
ajst-2553	89	6	intrusion	intrusion	NOUN
ajst-2553	89	7	network	network	NOUN
ajst-2553	89	8	types	type	NOUN
ajst-2553	89	9	,	,	PUNCT
ajst-2553	89	10	the	the	DET
ajst-2553	89	11	accuracy	accuracy	NOUN
ajst-2553	89	12	of	of	ADP
ajst-2553	89	13	dacae	dacae	NOUN
ajst-2553	89	14	-	-	PUNCT
ajst-2553	89	15	rf	rf	NOUN
ajst-2553	89	16	is	be	AUX
ajst-2553	89	17	improved	improve	VERB
ajst-2553	89	18	from	from	ADP
ajst-2553	89	19	94.67	94.67	NUM
ajst-2553	89	20	%	%	NOUN
ajst-2553	89	21	to	to	ADP
ajst-2553	89	22	96.29	96.29	NUM
ajst-2553	89	23	%	%	NOUN
ajst-2553	89	24	compared	compare	VERB
ajst-2553	89	25	with	with	ADP
ajst-2553	89	26	s	s	NOUN
ajst-2553	89	27	-	-	NOUN
ajst-2553	89	28	ndae	ndae	NOUN
ajst-2553	89	29	.	.	PUNCT
ajst-2553	90	1	in	in	ADP
ajst-2553	90	2	particular	particular	ADJ
ajst-2553	90	3	,	,	PUNCT
ajst-2553	90	4	in	in	ADP
ajst-2553	90	5	the	the	DET
ajst-2553	90	6	classification	classification	NOUN
ajst-2553	90	7	of	of	ADP
ajst-2553	90	8	r2l	r2l	NOUN
ajst-2553	90	9	intrusion	intrusion	NOUN
ajst-2553	90	10	events	event	NOUN
ajst-2553	90	11	,	,	PUNCT
ajst-2553	90	12	the	the	DET
ajst-2553	90	13	scores	score	NOUN
ajst-2553	90	14	of	of	ADP
ajst-2553	90	15	various	various	ADJ
ajst-2553	90	16	indicators	indicator	NOUN
ajst-2553	90	17	of	of	ADP
ajst-2553	90	18	dacae	dacae	NOUN
ajst-2553	90	19	-	-	PUNCT
ajst-2553	90	20	rf	rf	NOUN
ajst-2553	90	21	have	have	AUX
ajst-2553	90	22	been	be	AUX
ajst-2553	90	23	significantly	significantly	ADV
ajst-2553	90	24	improved	improve	VERB
ajst-2553	90	25	.	.	PUNCT
ajst-2553	91	1	in	in	ADP
ajst-2553	91	2	particular	particular	ADJ
ajst-2553	91	3	,	,	PUNCT
ajst-2553	91	4	the	the	DET
ajst-2553	91	5	recall	recall	NOUN
ajst-2553	91	6	rate	rate	NOUN
ajst-2553	91	7	and	and	CCONJ
ajst-2553	91	8	"	"	PUNCT
ajst-2553	91	9	f1score	f1score	NOUN
ajst-2553	91	10	"	"	PUNCT
ajst-2553	91	11	reached	reach	VERB
ajst-2553	91	12	15.27	15.27	NUM
ajst-2553	91	13	%	%	NOUN
ajst-2553	91	14	and	and	CCONJ
ajst-2553	91	15	26.49	26.49	NUM
ajst-2553	91	16	%	%	NOUN
ajst-2553	91	17	respectively	respectively	ADV
ajst-2553	91	18	.	.	PUNCT
ajst-2553	92	1	this	this	DET
ajst-2553	92	2	result	result	NOUN
ajst-2553	92	3	also	also	ADV
ajst-2553	92	4	shows	show	VERB
ajst-2553	92	5	that	that	SCONJ
ajst-2553	92	6	under	under	ADP
ajst-2553	92	7	different	different	ADJ
ajst-2553	92	8	data	data	NOUN
ajst-2553	92	9	sets	set	NOUN
ajst-2553	92	10	,	,	PUNCT
ajst-2553	92	11	the	the	DET
ajst-2553	92	12	dacae	dacae	NOUN
ajst-2553	92	13	-	-	PUNCT
ajst-2553	92	14	rf	rf	NOUN
ajst-2553	92	15	detection	detection	NOUN
ajst-2553	92	16	model	model	NOUN
ajst-2553	92	17	can	can	AUX
ajst-2553	92	18	effectively	effectively	ADV
ajst-2553	92	19	detect	detect	VERB
ajst-2553	92	20	intrusion	intrusion	NOUN
ajst-2553	92	21	anomaly	anomaly	NOUN
ajst-2553	92	22	types	type	NOUN
ajst-2553	92	23	.	.	PUNCT
ajst-2553	93	1	in	in	ADP
ajst-2553	93	2	the	the	DET
ajst-2553	93	3	last	last	ADJ
ajst-2553	93	4	evaluation	evaluation	NOUN
ajst-2553	93	5	,	,	PUNCT
ajst-2553	93	6	the	the	DET
ajst-2553	93	7	performance	performance	NOUN
ajst-2553	93	8	of	of	ADP
ajst-2553	93	9	multi	multi	ADJ
ajst-2553	93	10	classification	classification	NOUN
ajst-2553	93	11	of	of	ADP
ajst-2553	93	12	dacae	dacae	NOUN
ajst-2553	93	13	-	-	PUNCT
ajst-2553	93	14	rf	rf	NOUN
ajst-2553	93	15	model	model	NOUN
ajst-2553	93	16	is	be	AUX
ajst-2553	93	17	tested	test	VERB
ajst-2553	93	18	.	.	PUNCT
ajst-2553	94	1	we	we	PRON
ajst-2553	94	2	processed	process	VERB
ajst-2553	94	3	the	the	DET
ajst-2553	94	4	nsl	nsl	NOUN
ajst-2553	94	5	-	-	PUNCT
ajst-2553	94	6	kdd	kdd	PROPN
ajst-2553	94	7	training	training	NOUN
ajst-2553	94	8	set	set	NOUN
ajst-2553	94	9	and	and	CCONJ
ajst-2553	94	10	reserved	reserve	VERB
ajst-2553	94	11	more	more	ADJ
ajst-2553	94	12	than	than	ADP
ajst-2553	94	13	100	100	NUM
ajst-2553	94	14	intrusion	intrusion	NOUN
ajst-2553	94	15	event	event	NOUN
ajst-2553	94	16	categories	category	NOUN
ajst-2553	94	17	.	.	PUNCT
ajst-2553	95	1	the	the	DET
ajst-2553	95	2	purpose	purpose	NOUN
ajst-2553	95	3	of	of	ADP
ajst-2553	95	4	this	this	DET
ajst-2553	95	5	experiment	experiment	NOUN
ajst-2553	95	6	is	be	AUX
ajst-2553	95	7	to	to	PART
ajst-2553	95	8	test	test	VERB
ajst-2553	95	9	the	the	DET
ajst-2553	95	10	stability	stability	NOUN
ajst-2553	95	11	and	and	CCONJ
ajst-2553	95	12	effectiveness	effectiveness	NOUN
ajst-2553	95	13	of	of	ADP
ajst-2553	95	14	dacae	dacae	NOUN
ajst-2553	95	15	-	-	PUNCT
ajst-2553	95	16	rf	rf	NOUN
ajst-2553	95	17	when	when	SCONJ
ajst-2553	95	18	the	the	DET
ajst-2553	95	19	number	number	NOUN
ajst-2553	95	20	of	of	ADP
ajst-2553	95	21	intrusion	intrusion	NOUN
ajst-2553	95	22	attack	attack	NOUN
ajst-2553	95	23	types	type	NOUN
ajst-2553	95	24	increases	increase	VERB
ajst-2553	95	25	.	.	PUNCT
ajst-2553	96	1	it	it	PRON
ajst-2553	96	2	is	be	AUX
ajst-2553	96	3	a	a	DET
ajst-2553	96	4	standard	standard	NOUN
ajst-2553	96	5	to	to	PART
ajst-2553	96	6	measure	measure	VERB
ajst-2553	96	7	the	the	DET
ajst-2553	96	8	quality	quality	NOUN
ajst-2553	96	9	of	of	ADP
ajst-2553	96	10	intrusion	intrusion	NOUN
ajst-2553	96	11	detection	detection	NOUN
ajst-2553	96	12	system	system	NOUN
ajst-2553	96	13	that	that	PRON
ajst-2553	96	14	can	can	AUX
ajst-2553	96	15	effectively	effectively	ADV
ajst-2553	96	16	classify	classify	VERB
ajst-2553	96	17	a	a	DET
ajst-2553	96	18	variety	variety	NOUN
ajst-2553	96	19	of	of	ADP
ajst-2553	96	20	different	different	ADJ
ajst-2553	96	21	types	type	NOUN
ajst-2553	96	22	of	of	ADP
ajst-2553	96	23	intrusion	intrusion	NOUN
ajst-2553	96	24	events	event	NOUN
ajst-2553	96	25	.	.	PUNCT
ajst-2553	97	1	as	as	SCONJ
ajst-2553	97	2	shown	show	VERB
ajst-2553	97	3	in	in	ADP
ajst-2553	97	4	table	table	NOUN
ajst-2553	97	5	3	3	NUM
ajst-2553	97	6	,	,	PUNCT
ajst-2553	97	7	dacae	dacae	NOUN
ajst-2553	97	8	-	-	PUNCT
ajst-2553	97	9	rf	rf	NOUN
ajst-2553	97	10	can	can	AUX
ajst-2553	97	11	effectively	effectively	ADV
ajst-2553	97	12	detect	detect	VERB
ajst-2553	97	13	the	the	DET
ajst-2553	97	14	types	type	NOUN
ajst-2553	97	15	of	of	ADP
ajst-2553	97	16	intrusion	intrusion	NOUN
ajst-2553	97	17	events	event	NOUN
ajst-2553	97	18	in	in	ADP
ajst-2553	97	19	the	the	DET
ajst-2553	97	20	case	case	NOUN
ajst-2553	97	21	of	of	ADP
ajst-2553	97	22	multiple	multiple	ADJ
ajst-2553	97	23	classifications	classification	NOUN
ajst-2553	97	24	with	with	ADP
ajst-2553	97	25	increased	increase	VERB
ajst-2553	97	26	detection	detection	NOUN
ajst-2553	97	27	categories	category	NOUN
ajst-2553	97	28	.	.	PUNCT
ajst-2553	98	1	this	this	PRON
ajst-2553	98	2	shows	show	VERB
ajst-2553	98	3	that	that	SCONJ
ajst-2553	98	4	the	the	DET
ajst-2553	98	5	dacae	dacae	NOUN
ajst-2553	98	6	-	-	PUNCT
ajst-2553	98	7	rf	rf	NOUN
ajst-2553	98	8	model	model	NOUN
ajst-2553	98	9	can	can	AUX
ajst-2553	98	10	still	still	ADV
ajst-2553	98	11	maintain	maintain	VERB
ajst-2553	98	12	a	a	DET
ajst-2553	98	13	high	high	ADJ
ajst-2553	98	14	level	level	NOUN
ajst-2553	98	15	of	of	ADP
ajst-2553	98	16	detection	detection	NOUN
ajst-2553	98	17	in	in	ADP
ajst-2553	98	18	the	the	DET
ajst-2553	98	19	case	case	NOUN
ajst-2553	98	20	of	of	ADP
ajst-2553	98	21	multi	multi	ADJ
ajst-2553	98	22	classification	classification	NOUN
ajst-2553	98	23	.	.	PUNCT
ajst-2553	99	1	table	table	NOUN
ajst-2553	99	2	3	3	NUM
ajst-2553	99	3	.	.	PUNCT
ajst-2553	99	4	dacae	dacae	NOUN
ajst-2553	99	5	-	-	PUNCT
ajst-2553	99	6	rf	rf	NOUN
ajst-2553	99	7	’s	’s	PART
ajst-2553	99	8	multi	multi	NOUN
ajst-2553	99	9	-	-	NOUN
ajst-2553	99	10	classification	classification	NOUN
ajst-2553	99	11	on	on	ADP
ajst-2553	99	12	nsl	nsl	PROPN
ajst-2553	99	13	-	-	PUNCT
ajst-2553	99	14	kdd	kdd	PROPN
ajst-2553	99	15	dateset	dateset	ADJ
ajst-2553	99	16	label	label	NOUN
ajst-2553	99	17	precision	precision	NOUN
ajst-2553	99	18	(	(	PUNCT
ajst-2553	99	19	%	%	INTJ
ajst-2553	99	20	)	)	PUNCT
ajst-2553	99	21	accuracy	accuracy	NOUN
ajst-2553	99	22	(	(	PUNCT
ajst-2553	99	23	%	%	INTJ
ajst-2553	99	24	)	)	PUNCT
ajst-2553	99	25	recall	recall	NOUN
ajst-2553	99	26	(	(	PUNCT
ajst-2553	99	27	%	%	NOUN
ajst-2553	99	28	)	)	PUNCT
ajst-2553	99	29	f1	f1	NOUN
ajst-2553	99	30	-	-	PUNCT
ajst-2553	99	31	score	score	NOUN
ajst-2553	99	32	(	(	PUNCT
ajst-2553	99	33	%	%	INTJ
ajst-2553	99	34	)	)	PUNCT
ajst-2553	99	35	back	back	ADV
ajst-2553	99	36	97.11	97.11	NUM
ajst-2553	99	37	97.71	97.71	NUM
ajst-2553	99	38	97.08	97.08	NUM
ajst-2553	99	39	97.39	97.39	NUM
ajst-2553	99	40	neptune	neptune	NOUN
ajst-2553	99	41	97.23	97.23	NUM
ajst-2553	99	42	97.95	97.95	NUM
ajst-2553	99	43	99.29	99.29	NUM
ajst-2553	99	44	98.62	98.62	NUM
ajst-2553	99	45	pod	pod	NOUN
ajst-2553	99	46	100	100	NUM
ajst-2553	99	47	100	100	NUM
ajst-2553	99	48	100	100	NUM
ajst-2553	99	49	100	100	NUM
ajst-2553	99	50	smurf	smurf	NOUN
ajst-2553	99	51	97.98	97.98	NUM
ajst-2553	99	52	98.96	98.96	NUM
ajst-2553	99	53	99.22	99.22	NUM
ajst-2553	99	54	99.09	99.09	NUM
ajst-2553	99	55	teardrop	teardrop	NOUN
ajst-2553	99	56	99.04	99.04	NUM
ajst-2553	99	57	99.16	99.16	NUM
ajst-2553	99	58	97.52	97.52	NUM
ajst-2553	99	59	98.33	98.33	NUM
ajst-2553	99	60	ipsweep	ipsweep	NOUN
ajst-2553	99	61	91.32	91.32	NUM
ajst-2553	99	62	85.41	85.41	NUM
ajst-2553	99	63	92.90	92.90	NUM
ajst-2553	99	64	89.00	89.00	NUM
ajst-2553	99	65	nmap	nmap	NUM
ajst-2553	99	66	81.33	81.33	NUM
ajst-2553	100	1	80.17	80.17	NUM
ajst-2553	100	2	60.60	60.60	NUM
ajst-2553	100	3	69.02	69.02	NUM
ajst-2553	100	4	portsweep	portsweep	NOUN
ajst-2553	100	5	92.27	92.27	NUM
ajst-2553	100	6	93.87	93.87	NUM
ajst-2553	100	7	86.83	86.83	NUM
ajst-2553	100	8	90.21	90.21	NUM
ajst-2553	100	9	satan	satan	PROPN
ajst-2553	100	10	89.34	89.34	NUM
ajst-2553	100	11	88.47	88.47	NUM
ajst-2553	100	12	79.17	79.17	NUM
ajst-2553	100	13	83.56	83.56	NUM
ajst-2553	100	14	warezclient	warezclient	NOUN
ajst-2553	100	15	94.11	94.11	NUM
ajst-2553	100	16	90.27	90.27	NUM
ajst-2553	100	17	93.4	93.4	NUM
ajst-2553	100	18	91.81	91.81	NUM
ajst-2553	100	19	normal	normal	ADJ
ajst-2553	100	20	98.39	98.39	NUM
ajst-2553	100	21	99.27	99.27	NUM
ajst-2553	100	22	99.53	99.53	NUM
ajst-2553	100	23	99.40	99.40	NUM
ajst-2553	100	24	total	total	NOUN
ajst-2553	100	25	98.72	98.72	NUM
ajst-2553	100	26	98.66	98.66	NUM
ajst-2553	100	27	98.72	98.72	NUM
ajst-2553	100	28	98.69	98.69	NUM
ajst-2553	100	29	76	76	NUM
ajst-2553	100	30	4	4	NUM
ajst-2553	100	31	.	.	PUNCT
ajst-2553	100	32	conclusion	conclusion	NOUN
ajst-2553	100	33	there	there	PRON
ajst-2553	100	34	are	be	VERB
ajst-2553	100	35	a	a	DET
ajst-2553	100	36	large	large	ADJ
ajst-2553	100	37	number	number	NOUN
ajst-2553	100	38	of	of	ADP
ajst-2553	100	39	unlabeled	unlabeled	ADJ
ajst-2553	100	40	network	network	NOUN
ajst-2553	100	41	data	datum	NOUN
ajst-2553	100	42	in	in	ADP
ajst-2553	100	43	the	the	DET
ajst-2553	100	44	internet	internet	NOUN
ajst-2553	100	45	environment	environment	NOUN
ajst-2553	100	46	,	,	PUNCT
ajst-2553	100	47	and	and	CCONJ
ajst-2553	100	48	intrusion	intrusion	NOUN
ajst-2553	100	49	detection	detection	NOUN
ajst-2553	100	50	systems	system	NOUN
ajst-2553	100	51	need	need	VERB
ajst-2553	100	52	to	to	PART
ajst-2553	100	53	detect	detect	VERB
ajst-2553	100	54	and	and	CCONJ
ajst-2553	100	55	classify	classify	VERB
ajst-2553	100	56	under	under	ADP
ajst-2553	100	57	a	a	DET
ajst-2553	100	58	variety	variety	NOUN
ajst-2553	100	59	of	of	ADP
ajst-2553	100	60	intrusion	intrusion	NOUN
ajst-2553	100	61	events	event	NOUN
ajst-2553	100	62	.	.	PUNCT
ajst-2553	101	1	this	this	DET
ajst-2553	101	2	paper	paper	NOUN
ajst-2553	101	3	combines	combine	VERB
ajst-2553	101	4	the	the	DET
ajst-2553	101	5	advantages	advantage	NOUN
ajst-2553	101	6	of	of	ADP
ajst-2553	101	7	traditional	traditional	ADJ
ajst-2553	101	8	self	self	NOUN
ajst-2553	101	9	-	-	PUNCT
ajst-2553	101	10	encoder	encoder	NOUN
ajst-2553	101	11	and	and	CCONJ
ajst-2553	101	12	convolutional	convolutional	ADJ
ajst-2553	101	13	self	self	NOUN
ajst-2553	101	14	-	-	PUNCT
ajst-2553	101	15	encoder	encoder	NOUN
ajst-2553	101	16	to	to	PART
ajst-2553	101	17	propose	propose	VERB
ajst-2553	101	18	a	a	DET
ajst-2553	101	19	depth	depth	NOUN
ajst-2553	101	20	asymmetric	asymmetric	ADJ
ajst-2553	101	21	convolutional	convolutional	ADJ
ajst-2553	101	22	encoder	encoder	NOUN
ajst-2553	101	23	.	.	PUNCT
ajst-2553	102	1	the	the	DET
ajst-2553	102	2	model	model	NOUN
ajst-2553	102	3	can	can	AUX
ajst-2553	102	4	make	make	VERB
ajst-2553	102	5	good	good	ADJ
ajst-2553	102	6	use	use	NOUN
ajst-2553	102	7	of	of	ADP
ajst-2553	102	8	the	the	DET
ajst-2553	102	9	ability	ability	NOUN
ajst-2553	102	10	of	of	ADP
ajst-2553	102	11	convolution	convolution	NOUN
ajst-2553	102	12	kernel	kernel	NOUN
ajst-2553	102	13	to	to	PART
ajst-2553	102	14	extract	extract	VERB
ajst-2553	102	15	local	local	ADJ
ajst-2553	102	16	optimal	optimal	ADJ
ajst-2553	102	17	features	feature	NOUN
ajst-2553	102	18	,	,	PUNCT
ajst-2553	102	19	extract	extract	VERB
ajst-2553	102	20	new	new	ADJ
ajst-2553	102	21	abstract	abstract	ADJ
ajst-2553	102	22	features	feature	NOUN
ajst-2553	102	23	through	through	ADP
ajst-2553	102	24	stacked	stack	VERB
ajst-2553	102	25	convolution	convolution	NOUN
ajst-2553	102	26	encoders	encoder	NOUN
ajst-2553	102	27	,	,	PUNCT
ajst-2553	102	28	and	and	CCONJ
ajst-2553	102	29	use	use	VERB
ajst-2553	102	30	random	random	ADJ
ajst-2553	102	31	forest	forest	NOUN
ajst-2553	102	32	algorithm	algorithm	NOUN
ajst-2553	102	33	to	to	PART
ajst-2553	102	34	detect	detect	VERB
ajst-2553	102	35	and	and	CCONJ
ajst-2553	102	36	classify	classify	VERB
ajst-2553	102	37	the	the	DET
ajst-2553	102	38	generated	generate	VERB
ajst-2553	102	39	features	feature	NOUN
ajst-2553	102	40	.	.	PUNCT
ajst-2553	103	1	from	from	ADP
ajst-2553	103	2	the	the	DET
ajst-2553	103	3	experimental	experimental	ADJ
ajst-2553	103	4	results	result	NOUN
ajst-2553	103	5	,	,	PUNCT
ajst-2553	103	6	the	the	DET
ajst-2553	103	7	dacae	dacae	NOUN
ajst-2553	103	8	-	-	PUNCT
ajst-2553	103	9	rf	rf	NOUN
ajst-2553	103	10	model	model	NOUN
ajst-2553	103	11	has	have	VERB
ajst-2553	103	12	better	well	ADJ
ajst-2553	103	13	intrusion	intrusion	NOUN
ajst-2553	103	14	detection	detection	NOUN
ajst-2553	103	15	performance	performance	NOUN
ajst-2553	103	16	than	than	ADP
ajst-2553	103	17	the	the	DET
ajst-2553	103	18	traditional	traditional	ADJ
ajst-2553	103	19	model	model	NOUN
ajst-2553	103	20	.	.	PUNCT
ajst-2553	104	1	especially	especially	ADV
ajst-2553	104	2	in	in	ADP
ajst-2553	104	3	terms	term	NOUN
ajst-2553	104	4	of	of	ADP
ajst-2553	104	5	the	the	DET
ajst-2553	104	6	ability	ability	NOUN
ajst-2553	104	7	to	to	PART
ajst-2553	104	8	detect	detect	VERB
ajst-2553	104	9	small	small	ADJ
ajst-2553	104	10	sample	sample	NOUN
ajst-2553	104	11	data	datum	NOUN
ajst-2553	104	12	,	,	PUNCT
ajst-2553	104	13	the	the	DET
ajst-2553	104	14	dacae	dacae	NOUN
ajst-2553	104	15	-	-	PUNCT
ajst-2553	104	16	rf	rf	NOUN
ajst-2553	104	17	model	model	NOUN
ajst-2553	104	18	has	have	VERB
ajst-2553	104	19	better	well	ADJ
ajst-2553	104	20	detection	detection	NOUN
ajst-2553	104	21	ability	ability	NOUN
ajst-2553	104	22	.	.	PUNCT
ajst-2553	105	1	in	in	ADP
ajst-2553	105	2	terms	term	NOUN
ajst-2553	105	3	of	of	ADP
ajst-2553	105	4	multi	multi	PROPN
ajst-2553	105	5	sample	sample	PROPN
ajst-2553	105	6	intrusion	intrusion	NOUN
ajst-2553	105	7	detection	detection	NOUN
ajst-2553	105	8	,	,	PUNCT
ajst-2553	105	9	the	the	DET
ajst-2553	105	10	model	model	NOUN
ajst-2553	105	11	can	can	AUX
ajst-2553	105	12	maintain	maintain	VERB
ajst-2553	105	13	a	a	DET
ajst-2553	105	14	good	good	ADJ
ajst-2553	105	15	detection	detection	NOUN
ajst-2553	105	16	level	level	NOUN
ajst-2553	105	17	.	.	PUNCT
ajst-2553	106	1	references	reference	NOUN
ajst-2553	106	2	[	[	X
ajst-2553	106	3	1	1	NUM
ajst-2553	106	4	]	]	PUNCT
ajst-2553	106	5	julisch	julisch	NOUN
ajst-2553	106	6	,	,	PUNCT
ajst-2553	106	7	klaus	klaus	NOUN
ajst-2553	106	8	.	.	PUNCT
ajst-2553	107	1	using	use	VERB
ajst-2553	107	2	root	root	NOUN
ajst-2553	107	3	cause	cause	VERB
ajst-2553	107	4	analysis	analysis	NOUN
ajst-2553	107	5	to	to	PART
ajst-2553	107	6	handle	handle	VERB
ajst-2553	107	7	intrusion	intrusion	NOUN
ajst-2553	107	8	detection	detection	NOUN
ajst-2553	107	9	alarms	alarm	NOUN
ajst-2553	107	10	.	.	PUNCT
ajst-2553	108	1	diss	diss	PROPN
ajst-2553	108	2	.	.	PUNCT
ajst-2553	109	1	universität	universität	PROPN
ajst-2553	109	2	dortmund	dortmund	PROPN
ajst-2553	109	3	,	,	PUNCT
ajst-2553	109	4	2003	2003	NUM
ajst-2553	109	5	.	.	PUNCT
ajst-2553	110	1	[	[	X
ajst-2553	110	2	2	2	NUM
ajst-2553	110	3	]	]	X
ajst-2553	110	4	zhang	zhang	PROPN
ajst-2553	110	5	,	,	PUNCT
ajst-2553	110	6	jiong	jiong	PROPN
ajst-2553	110	7	,	,	PUNCT
ajst-2553	110	8	and	and	CCONJ
ajst-2553	110	9	mohammad	mohammad	PROPN
ajst-2553	110	10	zulkernine	zulkernine	PROPN
ajst-2553	110	11	.	.	PUNCT
ajst-2553	111	1	"	"	PUNCT
ajst-2553	111	2	anomaly	anomaly	NOUN
ajst-2553	111	3	based	base	VERB
ajst-2553	111	4	network	network	NOUN
ajst-2553	111	5	intrusion	intrusion	NOUN
ajst-2553	111	6	detection	detection	NOUN
ajst-2553	111	7	with	with	ADP
ajst-2553	111	8	unsupervised	unsupervised	ADJ
ajst-2553	111	9	outlier	outlier	NOUN
ajst-2553	111	10	detection	detection	NOUN
ajst-2553	111	11	.	.	PUNCT
ajst-2553	111	12	"	"	PUNCT
ajst-2553	112	1	2006	2006	NUM
ajst-2553	112	2	ieee	ieee	PROPN
ajst-2553	112	3	international	international	ADJ
ajst-2553	112	4	conference	conference	NOUN
ajst-2553	112	5	on	on	ADP
ajst-2553	112	6	communications	communication	NOUN
ajst-2553	112	7	.	.	PUNCT
ajst-2553	113	1	vol	vol	NOUN
ajst-2553	113	2	.	.	PROPN
ajst-2553	114	1	5	5	NUM
ajst-2553	114	2	.	.	X
ajst-2553	114	3	ieee	ieee	NOUN
ajst-2553	114	4	,	,	PUNCT
ajst-2553	114	5	2006	2006	NUM
ajst-2553	114	6	.	.	PUNCT
ajst-2553	115	1	[	[	X
ajst-2553	115	2	3	3	NUM
ajst-2553	115	3	]	]	X
ajst-2553	115	4	chen	chen	PROPN
ajst-2553	115	5	,	,	PUNCT
ajst-2553	115	6	wun	wun	PROPN
ajst-2553	115	7	-	-	PROPN
ajst-2553	115	8	hwa	hwa	PROPN
ajst-2553	115	9	,	,	PUNCT
ajst-2553	115	10	sheng	sheng	PROPN
ajst-2553	115	11	-	-	PUNCT
ajst-2553	115	12	hsun	hsun	PROPN
ajst-2553	115	13	hsu	hsu	PROPN
ajst-2553	115	14	,	,	PUNCT
ajst-2553	115	15	and	and	CCONJ
ajst-2553	115	16	hwang	hwang	PROPN
ajst-2553	115	17	-	-	PUNCT
ajst-2553	115	18	pin	pin	PROPN
ajst-2553	115	19	shen	shen	NOUN
ajst-2553	115	20	.	.	PUNCT
ajst-2553	116	1	"	"	PUNCT
ajst-2553	116	2	application	application	NOUN
ajst-2553	116	3	of	of	ADP
ajst-2553	116	4	svm	svm	PROPN
ajst-2553	116	5	and	and	CCONJ
ajst-2553	116	6	ann	ann	PROPN
ajst-2553	116	7	for	for	ADP
ajst-2553	116	8	intrusion	intrusion	NOUN
ajst-2553	116	9	detection	detection	NOUN
ajst-2553	116	10	.	.	PUNCT
ajst-2553	116	11	"	"	PUNCT
ajst-2553	117	1	computers	computer	NOUN
ajst-2553	117	2	&	&	CCONJ
ajst-2553	117	3	operations	operation	NOUN
ajst-2553	117	4	research	research	VERB
ajst-2553	117	5	32.10	32.10	NUM
ajst-2553	117	6	(	(	PUNCT
ajst-2553	117	7	2005	2005	NUM
ajst-2553	117	8	):	):	PUNCT
ajst-2553	117	9	2617	2617	NUM
ajst-2553	117	10	-	-	SYM
ajst-2553	117	11	2634	2634	NUM
ajst-2553	117	12	.	.	PUNCT
ajst-2553	118	1	[	[	X
ajst-2553	118	2	4	4	NUM
ajst-2553	118	3	]	]	X
ajst-2553	118	4	gao	gao	PROPN
ajst-2553	118	5	,	,	PUNCT
ajst-2553	118	6	ni	ni	PROPN
ajst-2553	118	7	,	,	PUNCT
ajst-2553	118	8	et	et	PROPN
ajst-2553	118	9	al	al	PROPN
ajst-2553	118	10	.	.	PUNCT
ajst-2553	119	1	"	"	PUNCT
ajst-2553	119	2	an	an	DET
ajst-2553	119	3	intrusion	intrusion	NOUN
ajst-2553	119	4	detection	detection	NOUN
ajst-2553	119	5	model	model	NOUN
ajst-2553	119	6	based	base	VERB
ajst-2553	119	7	on	on	ADP
ajst-2553	119	8	deep	deep	ADJ
ajst-2553	119	9	belief	belief	NOUN
ajst-2553	119	10	networks	network	NOUN
ajst-2553	119	11	.	.	PUNCT
ajst-2553	119	12	"	"	PUNCT
ajst-2553	120	1	2014	2014	NUM
ajst-2553	120	2	second	second	ADJ
ajst-2553	120	3	international	international	ADJ
ajst-2553	120	4	conference	conference	NOUN
ajst-2553	120	5	on	on	ADP
ajst-2553	120	6	advanced	advanced	ADJ
ajst-2553	120	7	cloud	cloud	NOUN
ajst-2553	120	8	and	and	CCONJ
ajst-2553	120	9	big	big	ADJ
ajst-2553	120	10	data	datum	NOUN
ajst-2553	120	11	.	.	PUNCT
ajst-2553	121	1	ieee	ieee	PROPN
ajst-2553	121	2	,	,	PUNCT
ajst-2553	121	3	2014	2014	NUM
ajst-2553	121	4	.	.	PUNCT
ajst-2553	122	1	[	[	X
ajst-2553	122	2	5	5	NUM
ajst-2553	122	3	]	]	X
ajst-2553	122	4	r.vinayakumar	r.vinayakumar	PROPN
ajst-2553	122	5	,	,	PUNCT
ajst-2553	122	6	k.	k.	PROPN
ajst-2553	122	7	p.	p.	PROPN
ajst-2553	122	8	soman	soman	NOUN
ajst-2553	122	9	and	and	CCONJ
ajst-2553	122	10	p.	p.	PROPN
ajst-2553	122	11	poornachandran	poornachandran	NOUN
ajst-2553	122	12	,	,	PUNCT
ajst-2553	122	13	"	"	PUNCT
ajst-2553	122	14	applying	apply	VERB
ajst-2553	122	15	convolutional	convolutional	ADJ
ajst-2553	122	16	neural	neural	ADJ
ajst-2553	122	17	network	network	NOUN
ajst-2553	122	18	for	for	ADP
ajst-2553	122	19	network	network	NOUN
ajst-2553	122	20	intrusion	intrusion	NOUN
ajst-2553	122	21	detection	detection	NOUN
ajst-2553	122	22	,	,	PUNCT
ajst-2553	122	23	"	"	PUNCT
ajst-2553	122	24	2017	2017	NUM
ajst-2553	122	25	international	international	ADJ
ajst-2553	122	26	conference	conference	NOUN
ajst-2553	122	27	on	on	ADP
ajst-2553	122	28	advances	advance	NOUN
ajst-2553	122	29	in	in	ADP
ajst-2553	122	30	computing	computing	NOUN
ajst-2553	122	31	,	,	PUNCT
ajst-2553	122	32	communications	communication	NOUN
ajst-2553	122	33	and	and	CCONJ
ajst-2553	122	34	informatics	informatic	NOUN
ajst-2553	122	35	(	(	PUNCT
ajst-2553	122	36	icacci	icacci	PROPN
ajst-2553	122	37	)	)	PUNCT
ajst-2553	122	38	,	,	PUNCT
ajst-2553	122	39	2017	2017	NUM
ajst-2553	122	40	,	,	PUNCT
ajst-2553	122	41	pp	pp	ADJ
ajst-2553	122	42	.	.	PUNCT
ajst-2553	122	43	1222	1222	NUM
ajst-2553	122	44	-	-	SYM
ajst-2553	122	45	1228	1228	NUM
ajst-2553	122	46	.	.	PUNCT
ajst-2553	123	1	[	[	X
ajst-2553	123	2	6	6	NUM
ajst-2553	123	3	]	]	PUNCT
ajst-2553	123	4	shone	shone	PROPN
ajst-2553	123	5	,	,	PUNCT
ajst-2553	123	6	nathan	nathan	PROPN
ajst-2553	123	7	,	,	PUNCT
ajst-2553	123	8	et	et	PROPN
ajst-2553	123	9	al	al	PROPN
ajst-2553	123	10	.	.	PUNCT
ajst-2553	124	1	"	"	PUNCT
ajst-2553	124	2	a	a	DET
ajst-2553	124	3	deep	deep	ADJ
ajst-2553	124	4	learning	learning	NOUN
ajst-2553	124	5	approach	approach	NOUN
ajst-2553	124	6	to	to	ADP
ajst-2553	124	7	network	network	NOUN
ajst-2553	124	8	intrusion	intrusion	NOUN
ajst-2553	124	9	detection	detection	NOUN
ajst-2553	124	10	.	.	PUNCT
ajst-2553	124	11	"	"	PUNCT
ajst-2553	125	1	ieee	ieee	NOUN
ajst-2553	125	2	transactions	transaction	NOUN
ajst-2553	125	3	on	on	ADP
ajst-2553	125	4	emerging	emerge	VERB
ajst-2553	125	5	topics	topic	NOUN
ajst-2553	125	6	in	in	ADP
ajst-2553	125	7	computational	computational	ADJ
ajst-2553	125	8	intelligence	intelligence	NOUN
ajst-2553	125	9	2.1	2.1	NUM
ajst-2553	125	10	(	(	PUNCT
ajst-2553	125	11	2018	2018	NUM
ajst-2553	125	12	):	):	PUNCT
ajst-2553	125	13	41	41	NUM
ajst-2553	125	14	-	-	SYM
ajst-2553	125	15	50	50	NUM
ajst-2553	125	16	.	.	PUNCT