id	sid	tid	token	lemma	pos
fcis-10735	1	1	frontiers	frontier	NOUN
fcis-10735	1	2	in	in	ADP
fcis-10735	1	3	computing	computing	NOUN
fcis-10735	1	4	and	and	CCONJ
fcis-10735	1	5	intelligent	intelligent	ADJ
fcis-10735	1	6	systems	system	NOUN
fcis-10735	1	7	issn	issn	VERB
fcis-10735	1	8	:	:	PUNCT
fcis-10735	1	9	2832	2832	NUM
fcis-10735	1	10	-	-	SYM
fcis-10735	1	11	6024	6024	NUM
fcis-10735	1	12	|	|	NOUN
fcis-10735	1	13	vol	vol	NOUN
fcis-10735	1	14	.	.	PROPN
fcis-10735	2	1	4	4	NUM
fcis-10735	2	2	,	,	PUNCT
fcis-10735	2	3	no	no	INTJ
fcis-10735	2	4	.	.	NOUN
fcis-10735	2	5	3	3	NUM
fcis-10735	2	6	,	,	PUNCT
fcis-10735	2	7	2023	2023	NUM
fcis-10735	2	8	7	7	NUM
fcis-10735	2	9	real	real	ADJ
fcis-10735	2	10	-	-	PUNCT
fcis-10735	2	11	time	time	NOUN
fcis-10735	2	12	classifier	classifier	NOUN
fcis-10735	2	13	of	of	ADP
fcis-10735	2	14	multilingual	multilingual	ADJ
fcis-10735	2	15	font	font	NOUN
fcis-10735	2	16	styles	style	NOUN
fcis-10735	2	17	based	base	VERB
fcis-10735	2	18	on	on	ADP
fcis-10735	2	19	resnet	resnet	NOUN
fcis-10735	2	20	,	,	PUNCT
fcis-10735	2	21	swordnet	swordnet	NOUN
fcis-10735	2	22	,	,	PUNCT
fcis-10735	2	23	logistic	logistic	ADJ
fcis-10735	2	24	regression	regression	NOUN
fcis-10735	2	25	and	and	CCONJ
fcis-10735	2	26	random	random	ADJ
fcis-10735	2	27	forest	forest	NOUN
fcis-10735	2	28	algorithms	algorithms	NOUN
fcis-10735	2	29	yue	yue	PROPN
fcis-10735	2	30	wu	wu	PROPN
fcis-10735	3	1	*	*	PUNCT
fcis-10735	3	2	school	school	NOUN
fcis-10735	3	3	of	of	ADP
fcis-10735	3	4	ai	ai	PROPN
fcis-10735	3	5	and	and	CCONJ
fcis-10735	3	6	advanced	advanced	ADJ
fcis-10735	3	7	computing	computing	NOUN
fcis-10735	3	8	,	,	PUNCT
fcis-10735	3	9	xi'an	xi'an	PROPN
fcis-10735	3	10	jiaotong	jiaotong	PROPN
fcis-10735	3	11	-	-	PUNCT
fcis-10735	3	12	liverpool	liverpool	PROPN
fcis-10735	3	13	university	university	PROPN
fcis-10735	3	14	,	,	PUNCT
fcis-10735	3	15	jiangsu	jiangsu	PROPN
fcis-10735	3	16	215123	215123	NUM
fcis-10735	3	17	,	,	PUNCT
fcis-10735	3	18	china	china	PROPN
fcis-10735	3	19	*	*	PUNCT
fcis-10735	3	20	corresponding	correspond	VERB
fcis-10735	3	21	author	author	NOUN
fcis-10735	3	22	email	email	NOUN
fcis-10735	3	23	:	:	PUNCT
fcis-10735	3	24	wuyue20022021@163.com	wuyue20022021@163.com	X
fcis-10735	3	25	abstract	abstract	NOUN
fcis-10735	3	26	:	:	PUNCT
fcis-10735	3	27	different	different	ADJ
fcis-10735	3	28	languages	language	NOUN
fcis-10735	3	29	have	have	VERB
fcis-10735	3	30	different	different	ADJ
fcis-10735	3	31	characters	character	NOUN
fcis-10735	3	32	.	.	PUNCT
fcis-10735	4	1	at	at	ADP
fcis-10735	4	2	the	the	DET
fcis-10735	4	3	same	same	ADJ
fcis-10735	4	4	time	time	NOUN
fcis-10735	4	5	,	,	PUNCT
fcis-10735	4	6	each	each	DET
fcis-10735	4	7	character	character	NOUN
fcis-10735	4	8	has	have	VERB
fcis-10735	4	9	a	a	DET
fcis-10735	4	10	lot	lot	NOUN
fcis-10735	4	11	of	of	ADP
fcis-10735	4	12	font	font	NOUN
fcis-10735	4	13	styles	style	NOUN
fcis-10735	4	14	.	.	PUNCT
fcis-10735	5	1	this	this	PRON
fcis-10735	5	2	makes	make	VERB
fcis-10735	5	3	it	it	PRON
fcis-10735	5	4	difficult	difficult	ADJ
fcis-10735	5	5	for	for	SCONJ
fcis-10735	5	6	humans	human	NOUN
fcis-10735	5	7	to	to	PART
fcis-10735	5	8	recognize	recognize	VERB
fcis-10735	5	9	different	different	ADJ
fcis-10735	5	10	font	font	NOUN
fcis-10735	5	11	styles	style	NOUN
fcis-10735	5	12	for	for	ADP
fcis-10735	5	13	different	different	ADJ
fcis-10735	5	14	characters	character	NOUN
fcis-10735	5	15	.	.	PUNCT
fcis-10735	6	1	however	however	ADV
fcis-10735	6	2	,	,	PUNCT
fcis-10735	6	3	being	be	AUX
fcis-10735	6	4	able	able	ADJ
fcis-10735	6	5	to	to	PART
fcis-10735	6	6	detect	detect	VERB
fcis-10735	6	7	and	and	CCONJ
fcis-10735	6	8	identify	identify	VERB
fcis-10735	6	9	these	these	DET
fcis-10735	6	10	font	font	NOUN
fcis-10735	6	11	styles	style	NOUN
fcis-10735	6	12	quickly	quickly	ADV
fcis-10735	6	13	and	and	CCONJ
fcis-10735	6	14	accurately	accurately	ADV
fcis-10735	6	15	has	have	VERB
fcis-10735	6	16	many	many	ADJ
fcis-10735	6	17	important	important	ADJ
fcis-10735	6	18	application	application	NOUN
fcis-10735	6	19	use	use	NOUN
fcis-10735	6	20	cases	case	NOUN
fcis-10735	6	21	in	in	ADP
fcis-10735	6	22	different	different	ADJ
fcis-10735	6	23	fields	field	NOUN
fcis-10735	6	24	.	.	PUNCT
fcis-10735	7	1	at	at	ADP
fcis-10735	7	2	the	the	DET
fcis-10735	7	3	same	same	ADJ
fcis-10735	7	4	time	time	NOUN
fcis-10735	7	5	,	,	PUNCT
fcis-10735	7	6	a	a	DET
fcis-10735	7	7	large	large	ADJ
fcis-10735	7	8	number	number	NOUN
fcis-10735	7	9	of	of	ADP
fcis-10735	7	10	internet	internet	NOUN
fcis-10735	7	11	users	user	NOUN
fcis-10735	7	12	use	use	VERB
fcis-10735	7	13	web	web	NOUN
fcis-10735	7	14	pages	page	NOUN
fcis-10735	7	15	to	to	PART
fcis-10735	7	16	query	query	VERB
fcis-10735	7	17	font	font	NOUN
fcis-10735	7	18	styles	style	NOUN
fcis-10735	7	19	.	.	PUNCT
fcis-10735	8	1	therefore	therefore	ADV
fcis-10735	8	2	,	,	PUNCT
fcis-10735	8	3	i	i	PRON
fcis-10735	8	4	choose	choose	VERB
fcis-10735	8	5	to	to	PART
fcis-10735	8	6	make	make	VERB
fcis-10735	8	7	this	this	DET
fcis-10735	8	8	real	real	ADJ
fcis-10735	8	9	-	-	PUNCT
fcis-10735	8	10	time	time	NOUN
fcis-10735	8	11	multilingual	multilingual	ADJ
fcis-10735	8	12	font	font	NOUN
fcis-10735	8	13	style	style	NOUN
fcis-10735	8	14	recognition	recognition	NOUN
fcis-10735	8	15	algorithm	algorithm	NOUN
fcis-10735	8	16	.	.	PUNCT
fcis-10735	9	1	in	in	ADP
fcis-10735	9	2	this	this	DET
fcis-10735	9	3	paper	paper	NOUN
fcis-10735	9	4	,	,	PUNCT
fcis-10735	9	5	i	i	PRON
fcis-10735	9	6	propose	propose	VERB
fcis-10735	9	7	an	an	DET
fcis-10735	9	8	algorithm	algorithm	NOUN
fcis-10735	9	9	that	that	PRON
fcis-10735	9	10	recognizes	recognize	VERB
fcis-10735	9	11	the	the	DET
fcis-10735	9	12	input	input	NOUN
fcis-10735	9	13	text	text	NOUN
fcis-10735	9	14	and	and	CCONJ
fcis-10735	9	15	pictures	picture	NOUN
fcis-10735	9	16	in	in	ADP
fcis-10735	9	17	real	real	ADJ
fcis-10735	9	18	time	time	NOUN
fcis-10735	9	19	to	to	PART
fcis-10735	9	20	judge	judge	VERB
fcis-10735	9	21	the	the	DET
fcis-10735	9	22	language	language	NOUN
fcis-10735	9	23	and	and	CCONJ
fcis-10735	9	24	style	style	NOUN
fcis-10735	9	25	of	of	ADP
fcis-10735	9	26	the	the	DET
fcis-10735	9	27	text	text	NOUN
fcis-10735	9	28	.	.	PUNCT
fcis-10735	10	1	it	it	PRON
fcis-10735	10	2	includes	include	VERB
fcis-10735	10	3	resnet	resnet	NOUN
fcis-10735	10	4	,	,	PUNCT
fcis-10735	10	5	swordnet	swordnet	NOUN
fcis-10735	10	6	,	,	PUNCT
fcis-10735	10	7	logistic	logistic	ADJ
fcis-10735	10	8	regression	regression	NOUN
fcis-10735	10	9	and	and	CCONJ
fcis-10735	10	10	random	random	ADJ
fcis-10735	10	11	forest	forest	NOUN
fcis-10735	10	12	algorithms	algorithm	NOUN
fcis-10735	10	13	.	.	PUNCT
fcis-10735	11	1	the	the	DET
fcis-10735	11	2	whole	whole	ADJ
fcis-10735	11	3	algorithm	algorithm	NOUN
fcis-10735	11	4	also	also	ADV
fcis-10735	11	5	calls	call	VERB
fcis-10735	11	6	pytesseract	pytesseract	NOUN
fcis-10735	11	7	and	and	CCONJ
fcis-10735	11	8	google	google	PROPN
fcis-10735	11	9	tesseract	tesseract	NOUN
fcis-10735	11	10	to	to	PART
fcis-10735	11	11	realize	realize	VERB
fcis-10735	11	12	text	text	NOUN
fcis-10735	11	13	recognition	recognition	NOUN
fcis-10735	11	14	and	and	CCONJ
fcis-10735	11	15	text	text	NOUN
fcis-10735	11	16	positioning	positioning	NOUN
fcis-10735	11	17	.	.	PUNCT
fcis-10735	12	1	i	i	PRON
fcis-10735	12	2	used	use	VERB
fcis-10735	12	3	font	font	NOUN
fcis-10735	12	4	datasets	dataset	NOUN
fcis-10735	12	5	used	use	VERB
fcis-10735	12	6	in	in	ADP
fcis-10735	12	7	"	"	PUNCT
fcis-10735	12	8	font	font	NOUN
fcis-10735	12	9	and	and	CCONJ
fcis-10735	12	10	calligraphy	calligraphy	NOUN
fcis-10735	12	11	style	style	NOUN
fcis-10735	12	12	recognition	recognition	NOUN
fcis-10735	12	13	using	use	VERB
fcis-10735	12	14	complex	complex	ADJ
fcis-10735	12	15	wavelet	wavelet	NOUN
fcis-10735	12	16	transform	transform	NOUN
fcis-10735	12	17	"	"	PUNCT
fcis-10735	12	18	for	for	ADP
fcis-10735	12	19	training	training	NOUN
fcis-10735	12	20	.	.	PUNCT
fcis-10735	13	1	at	at	ADP
fcis-10735	13	2	the	the	DET
fcis-10735	13	3	same	same	ADJ
fcis-10735	13	4	time	time	NOUN
fcis-10735	13	5	,	,	PUNCT
fcis-10735	13	6	i	i	PRON
fcis-10735	13	7	also	also	ADV
fcis-10735	13	8	built	build	VERB
fcis-10735	13	9	an	an	DET
fcis-10735	13	10	image	image	NOUN
fcis-10735	13	11	text	text	NOUN
fcis-10735	13	12	recognition	recognition	NOUN
fcis-10735	13	13	algorithm	algorithm	NOUN
fcis-10735	13	14	and	and	CCONJ
fcis-10735	13	15	generated	generate	VERB
fcis-10735	13	16	various	various	ADJ
fcis-10735	13	17	font	font	NOUN
fcis-10735	13	18	styles	style	NOUN
fcis-10735	13	19	as	as	ADP
fcis-10735	13	20	a	a	DET
fcis-10735	13	21	data	data	NOUN
fcis-10735	13	22	source	source	NOUN
fcis-10735	13	23	.	.	PUNCT
fcis-10735	14	1	based	base	VERB
fcis-10735	14	2	on	on	ADP
fcis-10735	14	3	this	this	DET
fcis-10735	14	4	data	datum	NOUN
fcis-10735	14	5	,	,	PUNCT
fcis-10735	14	6	we	we	PRON
fcis-10735	14	7	adjusted	adjust	VERB
fcis-10735	14	8	the	the	DET
fcis-10735	14	9	parameters	parameter	NOUN
fcis-10735	14	10	and	and	CCONJ
fcis-10735	14	11	finally	finally	ADV
fcis-10735	14	12	achieved	achieve	VERB
fcis-10735	14	13	an	an	DET
fcis-10735	14	14	accuracy	accuracy	NOUN
fcis-10735	14	15	rate	rate	NOUN
fcis-10735	14	16	higher	high	ADJ
fcis-10735	14	17	than	than	ADP
fcis-10735	14	18	90	90	NUM
fcis-10735	14	19	%	%	NOUN
fcis-10735	14	20	.	.	PUNCT
fcis-10735	15	1	keywords	keyword	NOUN
fcis-10735	15	2	:	:	PUNCT
fcis-10735	15	3	font	font	VERB
fcis-10735	15	4	style	style	NOUN
fcis-10735	15	5	recognition	recognition	NOUN
fcis-10735	15	6	;	;	PUNCT
fcis-10735	15	7	real	real	ADJ
fcis-10735	15	8	-	-	PUNCT
fcis-10735	15	9	time	time	NOUN
fcis-10735	15	10	font	font	NOUN
fcis-10735	15	11	identification	identification	NOUN
fcis-10735	15	12	;	;	PUNCT
fcis-10735	15	13	image	image	NOUN
fcis-10735	15	14	text	text	NOUN
fcis-10735	15	15	recognition	recognition	NOUN
fcis-10735	15	16	;	;	PUNCT
fcis-10735	15	17	language	language	NOUN
fcis-10735	15	18	identification	identification	NOUN
fcis-10735	15	19	;	;	PUNCT
fcis-10735	15	20	resnet	resnet	NOUN
fcis-10735	15	21	;	;	PUNCT
fcis-10735	15	22	random	random	ADJ
fcis-10735	15	23	forest	forest	NOUN
fcis-10735	15	24	;	;	PUNCT
fcis-10735	15	25	logistic	logistic	ADJ
fcis-10735	15	26	regression	regression	NOUN
fcis-10735	15	27	;	;	PUNCT
fcis-10735	15	28	pytesseract	pytesseract	NOUN
fcis-10735	15	29	.	.	PUNCT
fcis-10735	16	1	1	1	X
fcis-10735	16	2	.	.	X
fcis-10735	16	3	dataset	dataset	VERB
fcis-10735	16	4	1.1	1.1	NUM
fcis-10735	16	5	.	.	PUNCT
fcis-10735	17	1	real	real	ADJ
fcis-10735	17	2	-	-	PUNCT
fcis-10735	17	3	time	time	NOUN
fcis-10735	17	4	font	font	NOUN
fcis-10735	17	5	style	style	NOUN
fcis-10735	17	6	dataset	dataset	VERB
fcis-10735	17	7	this	this	DET
fcis-10735	17	8	dataset	dataset	NOUN
fcis-10735	17	9	was	be	AUX
fcis-10735	17	10	created	create	VERB
fcis-10735	17	11	by	by	ADP
fcis-10735	17	12	me	i	PRON
fcis-10735	17	13	.	.	PUNCT
fcis-10735	18	1	the	the	DET
fcis-10735	18	2	data	datum	NOUN
fcis-10735	18	3	set	set	VERB
fcis-10735	18	4	is	be	AUX
fcis-10735	18	5	generated	generate	VERB
fcis-10735	18	6	in	in	ADP
fcis-10735	18	7	real	real	ADJ
fcis-10735	18	8	-	-	PUNCT
fcis-10735	18	9	time	time	NOUN
fcis-10735	18	10	,	,	PUNCT
fcis-10735	18	11	and	and	CCONJ
fcis-10735	18	12	its	its	PRON
fcis-10735	18	13	content	content	NOUN
fcis-10735	18	14	is	be	AUX
fcis-10735	18	15	very	very	ADV
fcis-10735	18	16	varied	varied	ADJ
fcis-10735	18	17	.	.	PUNCT
fcis-10735	19	1	it	it	PRON
fcis-10735	19	2	mainly	mainly	ADV
fcis-10735	19	3	uses	use	VERB
fcis-10735	19	4	google	google	PROPN
fcis-10735	19	5	's	's	PART
fcis-10735	19	6	tesseract	tesseract	NOUN
fcis-10735	19	7	and	and	CCONJ
fcis-10735	19	8	various	various	ADJ
fcis-10735	19	9	font	font	NOUN
fcis-10735	19	10	packages	package	NOUN
fcis-10735	19	11	for	for	ADP
fcis-10735	19	12	generation	generation	NOUN
fcis-10735	19	13	.	.	PUNCT
fcis-10735	20	1	i	i	PRON
fcis-10735	20	2	will	will	AUX
fcis-10735	20	3	describe	describe	VERB
fcis-10735	20	4	how	how	SCONJ
fcis-10735	20	5	the	the	DET
fcis-10735	20	6	entire	entire	ADJ
fcis-10735	20	7	database	database	NOUN
fcis-10735	20	8	is	be	AUX
fcis-10735	20	9	generated	generate	VERB
fcis-10735	20	10	and	and	CCONJ
fcis-10735	20	11	used	use	VERB
fcis-10735	20	12	in	in	ADP
fcis-10735	20	13	real	real	ADJ
fcis-10735	20	14	time	time	NOUN
fcis-10735	20	15	.	.	PUNCT
fcis-10735	21	1	1.1.1	1.1.1	X
fcis-10735	21	2	.	.	PUNCT
fcis-10735	21	3	data	datum	NOUN
fcis-10735	21	4	establishment	establishment	NOUN
fcis-10735	21	5	process	process	NOUN
fcis-10735	21	6	first	first	ADV
fcis-10735	21	7	,	,	PUNCT
fcis-10735	21	8	the	the	DET
fcis-10735	21	9	user	user	NOUN
fcis-10735	21	10	needs	need	VERB
fcis-10735	21	11	to	to	PART
fcis-10735	21	12	find	find	VERB
fcis-10735	21	13	a	a	DET
fcis-10735	21	14	picture	picture	NOUN
fcis-10735	21	15	of	of	ADP
fcis-10735	21	16	some	some	DET
fcis-10735	21	17	text	text	NOUN
fcis-10735	21	18	as	as	SCONJ
fcis-10735	21	19	shown	show	VERB
fcis-10735	21	20	in	in	ADP
fcis-10735	21	21	figure	figure	NOUN
fcis-10735	21	22	1	1	NUM
fcis-10735	21	23	.	.	PUNCT
fcis-10735	22	1	the	the	DET
fcis-10735	22	2	selected	select	VERB
fcis-10735	22	3	picture	picture	NOUN
fcis-10735	22	4	should	should	AUX
fcis-10735	22	5	be	be	AUX
fcis-10735	22	6	clear	clear	ADJ
fcis-10735	22	7	and	and	CCONJ
fcis-10735	22	8	the	the	DET
fcis-10735	22	9	text	text	NOUN
fcis-10735	22	10	in	in	ADP
fcis-10735	22	11	the	the	DET
fcis-10735	22	12	picture	picture	NOUN
fcis-10735	22	13	should	should	AUX
fcis-10735	22	14	be	be	AUX
fcis-10735	22	15	parallel	parallel	ADJ
fcis-10735	22	16	to	to	ADP
fcis-10735	22	17	the	the	DET
fcis-10735	22	18	picture	picture	NOUN
fcis-10735	22	19	.	.	PUNCT
fcis-10735	23	1	text	text	NOUN
fcis-10735	23	2	and	and	CCONJ
fcis-10735	23	3	pictures	picture	NOUN
fcis-10735	23	4	in	in	ADP
fcis-10735	23	5	multiple	multiple	ADJ
fcis-10735	23	6	languages	language	NOUN
fcis-10735	23	7	can	can	AUX
fcis-10735	23	8	be	be	AUX
fcis-10735	23	9	input	input	NOUN
fcis-10735	23	10	and	and	CCONJ
fcis-10735	23	11	the	the	DET
fcis-10735	23	12	input	input	NOUN
fcis-10735	23	13	picture	picture	NOUN
fcis-10735	23	14	should	should	AUX
fcis-10735	23	15	be	be	AUX
fcis-10735	23	16	placed	place	VERB
fcis-10735	23	17	in	in	ADP
fcis-10735	23	18	the	the	DET
fcis-10735	23	19	'	'	PUNCT
fcis-10735	23	20	test	test	NOUN
fcis-10735	23	21	'	'	PUNCT
fcis-10735	23	22	folder	folder	NOUN
fcis-10735	23	23	.	.	PUNCT
fcis-10735	24	1	fig	fig	PROPN
fcis-10735	24	2	1	1	NUM
fcis-10735	24	3	.	.	PUNCT
fcis-10735	24	4	sentence	sentence	NOUN
fcis-10735	24	5	image	image	NOUN
fcis-10735	24	6	second	second	ADJ
fcis-10735	24	7	,	,	PUNCT
fcis-10735	24	8	i	i	PRON
fcis-10735	24	9	call	call	VERB
fcis-10735	24	10	pytesseract	pytesseract	NOUN
fcis-10735	24	11	and	and	CCONJ
fcis-10735	24	12	google	google	PROPN
fcis-10735	24	13	tesseract	tesseract	NOUN
fcis-10735	24	14	to	to	PART
fcis-10735	24	15	perform	perform	VERB
fcis-10735	24	16	recognition	recognition	NOUN
fcis-10735	24	17	on	on	ADP
fcis-10735	24	18	the	the	DET
fcis-10735	24	19	input	input	NOUN
fcis-10735	24	20	image	image	NOUN
fcis-10735	24	21	.	.	PUNCT
fcis-10735	25	1	in	in	ADP
fcis-10735	25	2	this	this	DET
fcis-10735	25	3	way	way	NOUN
fcis-10735	25	4	,	,	PUNCT
fcis-10735	25	5	i	i	PRON
fcis-10735	25	6	can	can	AUX
fcis-10735	25	7	identify	identify	VERB
fcis-10735	25	8	the	the	DET
fcis-10735	25	9	text	text	NOUN
fcis-10735	25	10	content	content	NOUN
fcis-10735	25	11	and	and	CCONJ
fcis-10735	25	12	text	text	NOUN
fcis-10735	25	13	position	position	NOUN
fcis-10735	25	14	in	in	ADP
fcis-10735	25	15	the	the	DET
fcis-10735	25	16	image	image	NOUN
fcis-10735	25	17	.	.	PUNCT
fcis-10735	26	1	based	base	VERB
fcis-10735	26	2	on	on	ADP
fcis-10735	26	3	the	the	DET
fcis-10735	26	4	position	position	NOUN
fcis-10735	26	5	and	and	CCONJ
fcis-10735	26	6	content	content	NOUN
fcis-10735	26	7	of	of	ADP
fcis-10735	26	8	the	the	DET
fcis-10735	26	9	image	image	NOUN
fcis-10735	26	10	,	,	PUNCT
fcis-10735	26	11	it	it	PRON
fcis-10735	26	12	allows	allow	VERB
fcis-10735	26	13	the	the	DET
fcis-10735	26	14	output	output	NOUN
fcis-10735	26	15	of	of	ADP
fcis-10735	26	16	the	the	DET
fcis-10735	26	17	box	box	NOUN
fcis-10735	26	18	-	-	PUNCT
fcis-10735	26	19	selected	select	VERB
fcis-10735	26	20	image	image	NOUN
fcis-10735	26	21	as	as	SCONJ
fcis-10735	26	22	shown	show	VERB
fcis-10735	26	23	in	in	ADP
fcis-10735	26	24	figure	figure	NOUN
fcis-10735	26	25	2	2	NUM
fcis-10735	26	26	.	.	PUNCT
fcis-10735	27	1	the	the	DET
fcis-10735	27	2	output	output	NOUN
fcis-10735	27	3	image	image	NOUN
fcis-10735	27	4	size	size	NOUN
fcis-10735	27	5	is	be	AUX
fcis-10735	27	6	unified	unify	VERB
fcis-10735	27	7	to	to	ADP
fcis-10735	27	8	220	220	NUM
fcis-10735	27	9	*	*	SYM
fcis-10735	27	10	100	100	NUM
fcis-10735	28	1	.	.	PUNCT
fcis-10735	28	2	fig	fig	NOUN
fcis-10735	28	3	2	2	NUM
fcis-10735	28	4	.	.	PUNCT
fcis-10735	29	1	picture	picture	NOUN
fcis-10735	29	2	of	of	ADP
fcis-10735	29	3	single	single	ADJ
fcis-10735	29	4	font	font	NOUN
fcis-10735	29	5	third	third	ADV
fcis-10735	29	6	,	,	PUNCT
fcis-10735	29	7	the	the	DET
fcis-10735	29	8	algorithm	algorithm	NOUN
fcis-10735	29	9	calls	call	VERB
fcis-10735	29	10	the	the	DET
fcis-10735	29	11	font	font	NOUN
fcis-10735	29	12	package	package	NOUN
fcis-10735	29	13	to	to	PART
fcis-10735	29	14	generate	generate	VERB
fcis-10735	29	15	images	image	NOUN
fcis-10735	29	16	with	with	ADP
fcis-10735	29	17	different	different	ADJ
fcis-10735	29	18	font	font	NOUN
fcis-10735	29	19	styles	style	NOUN
fcis-10735	29	20	of	of	ADP
fcis-10735	29	21	the	the	DET
fcis-10735	29	22	same	same	ADJ
fcis-10735	29	23	text	text	NOUN
fcis-10735	29	24	as	as	SCONJ
fcis-10735	29	25	shown	show	VERB
fcis-10735	29	26	in	in	ADP
fcis-10735	29	27	figure	figure	NOUN
fcis-10735	29	28	3	3	NUM
fcis-10735	29	29	.	.	PUNCT
fcis-10735	29	30	fonts	font	NOUN
fcis-10735	29	31	are	be	AUX
fcis-10735	29	32	all	all	PRON
fcis-10735	29	33	stored	store	VERB
fcis-10735	29	34	in	in	ADP
fcis-10735	29	35	the	the	DET
fcis-10735	29	36	'	'	PUNCT
fcis-10735	29	37	fonts	font	NOUN
fcis-10735	29	38	'	'	PART
fcis-10735	29	39	folder	folder	NOUN
fcis-10735	29	40	and	and	CCONJ
fcis-10735	29	41	the	the	DET
fcis-10735	29	42	number	number	NOUN
fcis-10735	29	43	of	of	ADP
fcis-10735	29	44	datasets	dataset	NOUN
fcis-10735	29	45	generated	generate	VERB
fcis-10735	29	46	is	be	AUX
fcis-10735	29	47	affected	affect	VERB
fcis-10735	29	48	by	by	ADP
fcis-10735	29	49	the	the	DET
fcis-10735	29	50	number	number	NOUN
fcis-10735	29	51	of	of	ADP
fcis-10735	29	52	fonts	font	NOUN
fcis-10735	29	53	.	.	PUNCT
fcis-10735	30	1	the	the	DET
fcis-10735	30	2	output	output	NOUN
fcis-10735	30	3	image	image	NOUN
fcis-10735	30	4	size	size	NOUN
fcis-10735	30	5	is	be	AUX
fcis-10735	30	6	also	also	ADV
fcis-10735	30	7	unified	unified	ADJ
fcis-10735	30	8	to	to	ADP
fcis-10735	30	9	220	220	NUM
fcis-10735	30	10	*	*	SYM
fcis-10735	30	11	100	100	NUM
fcis-10735	30	12	.	.	PUNCT
fcis-10735	31	1	all	all	DET
fcis-10735	31	2	generated	generate	VERB
fcis-10735	31	3	text	text	NOUN
fcis-10735	31	4	images	image	NOUN
fcis-10735	31	5	will	will	AUX
fcis-10735	31	6	be	be	AUX
fcis-10735	31	7	stored	store	VERB
fcis-10735	31	8	in	in	ADP
fcis-10735	31	9	the	the	DET
fcis-10735	31	10	'	'	PUNCT
fcis-10735	31	11	train	train	NOUN
fcis-10735	31	12	'	'	PUNCT
fcis-10735	31	13	folder	folder	NOUN
fcis-10735	31	14	.	.	PUNCT
fcis-10735	32	1	fig	fig	PROPN
fcis-10735	32	2	3	3	NUM
fcis-10735	32	3	.	.	PUNCT
fcis-10735	33	1	pictures	picture	NOUN
fcis-10735	33	2	of	of	ADP
fcis-10735	33	3	fonts	font	NOUN
fcis-10735	33	4	1.1.2	1.1.2	NUM
fcis-10735	33	5	.	.	PUNCT
fcis-10735	34	1	data	datum	NOUN
fcis-10735	34	2	description	description	NOUN
fcis-10735	34	3	and	and	CCONJ
fcis-10735	34	4	feature	feature	VERB
fcis-10735	34	5	analysis	analysis	NOUN
fcis-10735	34	6	the	the	DET
fcis-10735	34	7	text	text	NOUN
fcis-10735	34	8	in	in	ADP
fcis-10735	34	9	the	the	DET
fcis-10735	34	10	images	image	NOUN
fcis-10735	34	11	is	be	AUX
fcis-10735	34	12	black	black	ADJ
fcis-10735	34	13	and	and	CCONJ
fcis-10735	34	14	the	the	DET
fcis-10735	34	15	background	background	NOUN
fcis-10735	34	16	is	be	AUX
fcis-10735	34	17	white	white	ADJ
fcis-10735	34	18	.	.	PUNCT
fcis-10735	35	1	the	the	DET
fcis-10735	35	2	images	image	NOUN
fcis-10735	35	3	are	be	AUX
fcis-10735	35	4	arranged	arrange	VERB
fcis-10735	35	5	in	in	ADP
fcis-10735	35	6	bgr	bgr	PROPN
fcis-10735	35	7	and	and	CCONJ
fcis-10735	35	8	they	they	PRON
fcis-10735	35	9	are	be	AUX
fcis-10735	35	10	of	of	ADP
fcis-10735	35	11	the	the	DET
fcis-10735	35	12	same	same	ADJ
fcis-10735	35	13	size	size	NOUN
fcis-10735	35	14	.	.	PUNCT
fcis-10735	36	1	all	all	DET
fcis-10735	36	2	images	image	NOUN
fcis-10735	36	3	are	be	AUX
fcis-10735	36	4	of	of	ADP
fcis-10735	36	5	size	size	NOUN
fcis-10735	36	6	220	220	NUM
fcis-10735	36	7	*	*	SYM
fcis-10735	36	8	100	100	NUM
fcis-10735	36	9	.	.	PUNCT
fcis-10735	37	1	the	the	DET
fcis-10735	37	2	data	datum	NOUN
fcis-10735	37	3	is	be	AUX
fcis-10735	37	4	displayed	display	VERB
fcis-10735	37	5	as	as	ADP
fcis-10735	37	6	a	a	DET
fcis-10735	37	7	picture	picture	NOUN
fcis-10735	37	8	and	and	CCONJ
fcis-10735	37	9	stored	store	VERB
fcis-10735	37	10	in	in	ADP
fcis-10735	37	11	'	'	PUNCT
fcis-10735	37	12	.png	.png	ADJ
fcis-10735	37	13	'	'	PUNCT
fcis-10735	37	14	format	format	NOUN
fcis-10735	37	15	.	.	PUNCT
fcis-10735	38	1	the	the	DET
fcis-10735	38	2	image	image	NOUN
fcis-10735	38	3	name	name	NOUN
fcis-10735	38	4	is	be	AUX
fcis-10735	38	5	the	the	DET
fcis-10735	38	6	label	label	NOUN
fcis-10735	38	7	of	of	ADP
fcis-10735	38	8	the	the	DET
fcis-10735	38	9	image	image	NOUN
fcis-10735	38	10	.	.	PUNCT
fcis-10735	39	1	the	the	DET
fcis-10735	39	2	format	format	NOUN
fcis-10735	39	3	is	be	AUX
fcis-10735	39	4	'	'	PUNCT
fcis-10735	39	5	text_font	text_font	PRON
fcis-10735	39	6	style.png	style.png	NOUN
fcis-10735	39	7	'	'	NUM
fcis-10735	39	8	.	.	PUNCT
fcis-10735	40	1	all	all	DET
fcis-10735	40	2	fonts	font	NOUN
fcis-10735	40	3	are	be	AUX
fcis-10735	40	4	fully	fully	ADV
fcis-10735	40	5	displayed	display	VERB
fcis-10735	40	6	and	and	CCONJ
fcis-10735	40	7	marked	mark	VERB
fcis-10735	40	8	.	.	PUNCT
fcis-10735	41	1	8	8	NUM
fcis-10735	41	2	fig	fig	NOUN
fcis-10735	41	3	4	4	NUM
fcis-10735	41	4	.	.	PUNCT
fcis-10735	42	1	pictures	picture	NOUN
fcis-10735	42	2	of	of	ADP
fcis-10735	42	3	fonts	font	NOUN
fcis-10735	42	4	fig	fig	NOUN
fcis-10735	42	5	5	5	NUM
fcis-10735	42	6	.	.	PUNCT
fcis-10735	43	1	pictures	picture	NOUN
fcis-10735	43	2	information	information	PROPN
fcis-10735	43	3	fig	fig	PROPN
fcis-10735	43	4	6	6	NUM
fcis-10735	43	5	.	.	PUNCT
fcis-10735	44	1	pictures	picture	NOUN
fcis-10735	44	2	of	of	ADP
fcis-10735	44	3	single	single	ADJ
fcis-10735	44	4	font	font	NOUN
fcis-10735	44	5	fig	fig	NOUN
fcis-10735	44	6	7	7	NUM
fcis-10735	44	7	.	.	PUNCT
fcis-10735	45	1	pictures	picture	NOUN
fcis-10735	45	2	of	of	ADP
fcis-10735	45	3	fonts	font	NOUN
fcis-10735	45	4	pros	pro	NOUN
fcis-10735	45	5	:	:	PUNCT
fcis-10735	45	6	the	the	DET
fcis-10735	45	7	entire	entire	ADJ
fcis-10735	45	8	dataset	dataset	NOUN
fcis-10735	45	9	is	be	AUX
fcis-10735	45	10	generated	generate	VERB
fcis-10735	45	11	in	in	ADP
fcis-10735	45	12	real	real	ADJ
fcis-10735	45	13	-	-	PUNCT
fcis-10735	45	14	time	time	NOUN
fcis-10735	45	15	,	,	PUNCT
fcis-10735	45	16	which	which	PRON
fcis-10735	45	17	means	mean	VERB
fcis-10735	45	18	that	that	SCONJ
fcis-10735	45	19	the	the	DET
fcis-10735	45	20	training	training	NOUN
fcis-10735	45	21	data	datum	NOUN
fcis-10735	45	22	can	can	AUX
fcis-10735	45	23	be	be	AUX
fcis-10735	45	24	modified	modify	VERB
fcis-10735	45	25	.	.	PUNCT
fcis-10735	46	1	so	so	ADV
fcis-10735	46	2	,	,	PUNCT
fcis-10735	46	3	the	the	DET
fcis-10735	46	4	user	user	NOUN
fcis-10735	46	5	can	can	AUX
fcis-10735	46	6	generate	generate	VERB
fcis-10735	46	7	training	training	NOUN
fcis-10735	46	8	data	datum	NOUN
fcis-10735	46	9	based	base	VERB
fcis-10735	46	10	on	on	ADP
fcis-10735	46	11	the	the	DET
fcis-10735	46	12	text	text	NOUN
fcis-10735	46	13	he	he	PRON
fcis-10735	46	14	wants	want	VERB
fcis-10735	46	15	to	to	PART
fcis-10735	46	16	query	query	VERB
fcis-10735	46	17	.	.	PUNCT
fcis-10735	47	1	this	this	PRON
fcis-10735	47	2	allows	allow	VERB
fcis-10735	47	3	the	the	DET
fcis-10735	47	4	user	user	NOUN
fcis-10735	47	5	to	to	PART
fcis-10735	47	6	choose	choose	VERB
fcis-10735	47	7	the	the	DET
fcis-10735	47	8	language	language	NOUN
fcis-10735	47	9	they	they	PRON
fcis-10735	47	10	want	want	VERB
fcis-10735	47	11	to	to	PART
fcis-10735	47	12	use	use	VERB
fcis-10735	47	13	and	and	CCONJ
fcis-10735	47	14	the	the	DET
fcis-10735	47	15	font	font	NOUN
fcis-10735	47	16	package	package	NOUN
fcis-10735	47	17	they	they	PRON
fcis-10735	47	18	own	own	VERB
fcis-10735	47	19	to	to	PART
fcis-10735	47	20	generate	generate	VERB
fcis-10735	47	21	the	the	DET
fcis-10735	47	22	data	datum	NOUN
fcis-10735	47	23	.	.	PUNCT
fcis-10735	48	1	it	it	PRON
fcis-10735	48	2	also	also	ADV
fcis-10735	48	3	means	mean	VERB
fcis-10735	48	4	that	that	SCONJ
fcis-10735	48	5	we	we	PRON
fcis-10735	48	6	can	can	AUX
fcis-10735	48	7	select	select	VERB
fcis-10735	48	8	the	the	DET
fcis-10735	48	9	same	same	ADJ
fcis-10735	48	10	text	text	NOUN
fcis-10735	48	11	as	as	SCONJ
fcis-10735	48	12	the	the	DET
fcis-10735	48	13	query	query	NOUN
fcis-10735	48	14	text	text	NOUN
fcis-10735	48	15	during	during	ADP
fcis-10735	48	16	the	the	DET
fcis-10735	48	17	model	model	NOUN
fcis-10735	48	18	learning	learning	NOUN
fcis-10735	48	19	process	process	NOUN
fcis-10735	48	20	to	to	PART
fcis-10735	48	21	improve	improve	VERB
fcis-10735	48	22	the	the	DET
fcis-10735	48	23	accuracy	accuracy	NOUN
fcis-10735	48	24	and	and	CCONJ
fcis-10735	48	25	reduce	reduce	VERB
fcis-10735	48	26	the	the	DET
fcis-10735	48	27	model	model	NOUN
fcis-10735	48	28	learning	learning	NOUN
fcis-10735	48	29	time	time	NOUN
fcis-10735	48	30	.	.	PUNCT
fcis-10735	49	1	cons	con	NOUN
fcis-10735	49	2	:	:	PUNCT
fcis-10735	49	3	the	the	DET
fcis-10735	49	4	amount	amount	NOUN
fcis-10735	49	5	of	of	ADP
fcis-10735	49	6	data	datum	NOUN
fcis-10735	49	7	generated	generate	VERB
fcis-10735	49	8	is	be	AUX
fcis-10735	49	9	relatively	relatively	ADV
fcis-10735	49	10	small	small	ADJ
fcis-10735	49	11	.	.	PUNCT
fcis-10735	50	1	it	it	PRON
fcis-10735	50	2	takes	take	VERB
fcis-10735	50	3	more	more	ADJ
fcis-10735	50	4	time	time	NOUN
fcis-10735	50	5	to	to	PART
fcis-10735	50	6	generate	generate	VERB
fcis-10735	50	7	a	a	DET
fcis-10735	50	8	large	large	ADJ
fcis-10735	50	9	amount	amount	NOUN
fcis-10735	50	10	of	of	ADP
fcis-10735	50	11	data	datum	NOUN
fcis-10735	50	12	.	.	PUNCT
fcis-10735	51	1	the	the	DET
fcis-10735	51	2	results	result	NOUN
fcis-10735	51	3	of	of	ADP
fcis-10735	51	4	training	training	NOUN
fcis-10735	51	5	using	use	VERB
fcis-10735	51	6	this	this	DET
fcis-10735	51	7	dataset	dataset	NOUN
fcis-10735	51	8	are	be	AUX
fcis-10735	51	9	for	for	ADP
fcis-10735	51	10	specific	specific	ADJ
fcis-10735	51	11	texts	text	NOUN
fcis-10735	51	12	and	and	CCONJ
fcis-10735	51	13	have	have	VERB
fcis-10735	51	14	no	no	DET
fcis-10735	51	15	great	great	ADJ
fcis-10735	51	16	generality	generality	NOUN
fcis-10735	51	17	.	.	PUNCT
fcis-10735	52	1	1.1.3	1.1.3	NUM
fcis-10735	52	2	.	.	PUNCT
fcis-10735	53	1	data	datum	NOUN
fcis-10735	53	2	pre	pre	ADJ
fcis-10735	53	3	-	-	ADJ
fcis-10735	53	4	processing	processing	ADJ
fcis-10735	53	5	1.1.3.1	1.1.3.1	NUM
fcis-10735	53	6	pictures	picture	NOUN
fcis-10735	53	7	imported	import	VERB
fcis-10735	53	8	and	and	CCONJ
fcis-10735	53	9	converted	convert	VERB
fcis-10735	53	10	to	to	ADP
fcis-10735	53	11	black	black	ADJ
fcis-10735	53	12	and	and	CCONJ
fcis-10735	53	13	white	white	ADJ
fcis-10735	53	14	images	image	NOUN
fcis-10735	53	15	i	i	PRON
fcis-10735	53	16	use	use	VERB
fcis-10735	53	17	the	the	DET
fcis-10735	53	18	cv2	cv2	PROPN
fcis-10735	53	19	package	package	NOUN
fcis-10735	53	20	to	to	PART
fcis-10735	53	21	import	import	VERB
fcis-10735	53	22	the	the	DET
fcis-10735	53	23	images	image	NOUN
fcis-10735	53	24	from	from	ADP
fcis-10735	53	25	the	the	DET
fcis-10735	53	26	folder	folder	NOUN
fcis-10735	53	27	and	and	CCONJ
fcis-10735	53	28	convert	convert	VERB
fcis-10735	53	29	the	the	DET
fcis-10735	53	30	images	image	NOUN
fcis-10735	53	31	to	to	ADP
fcis-10735	53	32	black	black	ADJ
fcis-10735	53	33	and	and	CCONJ
fcis-10735	53	34	white	white	ADJ
fcis-10735	53	35	.	.	PUNCT
fcis-10735	54	1	at	at	ADP
fcis-10735	54	2	the	the	DET
fcis-10735	54	3	same	same	ADJ
fcis-10735	54	4	time	time	NOUN
fcis-10735	54	5	,	,	PUNCT
fcis-10735	54	6	the	the	DET
fcis-10735	54	7	algorithm	algorithm	NOUN
fcis-10735	54	8	also	also	ADV
fcis-10735	54	9	reads	read	VERB
fcis-10735	54	10	the	the	DET
fcis-10735	54	11	image	image	NOUN
fcis-10735	54	12	name	name	NOUN
fcis-10735	54	13	and	and	CCONJ
fcis-10735	54	14	enters	enter	VERB
fcis-10735	54	15	it	it	PRON
fcis-10735	54	16	into	into	ADP
fcis-10735	54	17	'	'	PUNCT
fcis-10735	54	18	words	word	NOUN
fcis-10735	54	19	'	'	PUNCT
fcis-10735	54	20	and	and	CCONJ
fcis-10735	54	21	'	'	PUNCT
fcis-10735	54	22	labels	label	NOUN
fcis-10735	54	23	'	'	PUNCT
fcis-10735	54	24	.	.	PUNCT
fcis-10735	55	1	1.1.3.2	1.1.3.2	NUM
fcis-10735	55	2	image	image	NOUN
fcis-10735	55	3	feature	feature	NOUN
fcis-10735	55	4	transformation	transformation	NOUN
fcis-10735	55	5	the	the	DET
fcis-10735	55	6	algorithm	algorithm	NOUN
fcis-10735	55	7	calculates	calculate	VERB
fcis-10735	55	8	the	the	DET
fcis-10735	55	9	histogram	histogram	NOUN
fcis-10735	55	10	after	after	ADP
fcis-10735	55	11	dividing	divide	VERB
fcis-10735	55	12	each	each	DET
fcis-10735	55	13	channel	channel	NOUN
fcis-10735	55	14	into	into	ADP
fcis-10735	55	15	8	8	NUM
fcis-10735	55	16	groups	group	NOUN
fcis-10735	55	17	using	use	VERB
fcis-10735	55	18	the	the	DET
fcis-10735	55	19	feature	feature	NOUN
fcis-10735	55	20	transformation	transformation	NOUN
fcis-10735	55	21	function	function	NOUN
fcis-10735	55	22	.	.	PUNCT
fcis-10735	56	1	finally	finally	ADV
fcis-10735	56	2	,	,	PUNCT
fcis-10735	56	3	it	it	PRON
fcis-10735	56	4	flattens	flatten	VERB
fcis-10735	56	5	the	the	DET
fcis-10735	56	6	8x8x8	8x8x8	NUM
fcis-10735	56	7	multidimensional	multidimensional	ADJ
fcis-10735	56	8	array	array	NOUN
fcis-10735	56	9	.	.	PUNCT
fcis-10735	57	1	in	in	ADP
fcis-10735	57	2	this	this	DET
fcis-10735	57	3	way	way	NOUN
fcis-10735	57	4	,	,	PUNCT
fcis-10735	57	5	i	i	PRON
fcis-10735	57	6	get	get	VERB
fcis-10735	57	7	all	all	DET
fcis-10735	57	8	the	the	DET
fcis-10735	57	9	image	image	NOUN
fcis-10735	57	10	features	feature	VERB
fcis-10735	57	11	.	.	PUNCT
fcis-10735	58	1	1.2	1.2	NUM
fcis-10735	58	2	.	.	PUNCT
fcis-10735	58	3	font	font	NOUN
fcis-10735	58	4	datasets	dataset	NOUN
fcis-10735	58	5	used	use	VERB
fcis-10735	58	6	in	in	ADP
fcis-10735	58	7	"	"	PUNCT
fcis-10735	58	8	font	font	NOUN
fcis-10735	58	9	and	and	CCONJ
fcis-10735	58	10	calligraphy	calligraphy	NOUN
fcis-10735	58	11	style	style	NOUN
fcis-10735	58	12	recognition	recognition	NOUN
fcis-10735	58	13	using	use	VERB
fcis-10735	58	14	complex	complex	ADJ
fcis-10735	58	15	wavelet	wavelet	NOUN
fcis-10735	58	16	transform	transform	NOUN
fcis-10735	58	17	"	"	PUNCT
fcis-10735	58	18	1.2.1	1.2.1	NUM
fcis-10735	58	19	.	.	PUNCT
fcis-10735	59	1	data	datum	NOUN
fcis-10735	59	2	description	description	NOUN
fcis-10735	59	3	the	the	DET
fcis-10735	59	4	database	database	NOUN
fcis-10735	59	5	is	be	AUX
fcis-10735	59	6	from	from	ADP
fcis-10735	59	7	font	font	NOUN
fcis-10735	59	8	datasets	dataset	NOUN
fcis-10735	59	9	used	use	VERB
fcis-10735	59	10	in	in	ADP
fcis-10735	59	11	"	"	PUNCT
fcis-10735	59	12	font	font	NOUN
fcis-10735	59	13	and	and	CCONJ
fcis-10735	59	14	calligraphy	calligraphy	NOUN
fcis-10735	59	15	style	style	NOUN
fcis-10735	59	16	recognition	recognition	NOUN
fcis-10735	59	17	using	use	VERB
fcis-10735	59	18	complex	complex	ADJ
fcis-10735	59	19	wavelet	wavelet	NOUN
fcis-10735	59	20	transform	transform	NOUN
fcis-10735	59	21	"	"	PUNCT
fcis-10735	59	22	shared	share	VERB
fcis-10735	59	23	by	by	ADP
fcis-10735	59	24	alican	alican	ADJ
fcis-10735	59	25	bozkurt	bozkurt	PROPN
fcis-10735	60	1	[	[	X
fcis-10735	60	2	1	1	NUM
fcis-10735	60	3	]	]	PUNCT
fcis-10735	60	4	.	.	PUNCT
fcis-10735	61	1	it	it	PRON
fcis-10735	61	2	includes	include	VERB
fcis-10735	61	3	multiple	multiple	ADJ
fcis-10735	61	4	languages	language	NOUN
fcis-10735	61	5	,	,	PUNCT
fcis-10735	61	6	such	such	ADJ
fcis-10735	61	7	as	as	ADP
fcis-10735	61	8	arabic	arabic	ADJ
fcis-10735	61	9	,	,	PUNCT
fcis-10735	61	10	chinese	chinese	PROPN
fcis-10735	61	11	,	,	PUNCT
fcis-10735	61	12	and	and	CCONJ
fcis-10735	61	13	latin	latin	NOUN
fcis-10735	61	14	-	-	PUNCT
fcis-10735	61	15	based	base	VERB
fcis-10735	61	16	languages	language	NOUN
fcis-10735	61	17	.	.	PUNCT
fcis-10735	62	1	it	it	PRON
fcis-10735	62	2	has	have	VERB
fcis-10735	62	3	a	a	DET
fcis-10735	62	4	total	total	NOUN
fcis-10735	62	5	of	of	ADP
fcis-10735	62	6	158	158	NUM
fcis-10735	62	7	fonts	font	NOUN
fcis-10735	62	8	and	and	CCONJ
fcis-10735	62	9	each	each	DET
fcis-10735	62	10	font	font	NOUN
fcis-10735	62	11	supplies	supply	VERB
fcis-10735	62	12	rich	rich	ADJ
fcis-10735	62	13	font	font	NOUN
fcis-10735	62	14	images	image	NOUN
fcis-10735	62	15	.	.	PUNCT
fcis-10735	63	1	all	all	DET
fcis-10735	63	2	labels	label	NOUN
fcis-10735	63	3	are	be	AUX
fcis-10735	63	4	stored	store	VERB
fcis-10735	63	5	in	in	ADP
fcis-10735	63	6	the	the	DET
fcis-10735	63	7	'	'	PUNCT
fcis-10735	63	8	font	font	NOUN
fcis-10735	63	9	datasets.csv	datasets.csv	NOUN
fcis-10735	63	10	'	'	PART
fcis-10735	63	11	file	file	NOUN
fcis-10735	63	12	and	and	CCONJ
fcis-10735	63	13	hold	hold	VERB
fcis-10735	63	14	'	'	PUNCT
fcis-10735	63	15	i	i	PRON
fcis-10735	63	16	d	d	NOUN
fcis-10735	63	17	'	'	PROPN
fcis-10735	63	18	,	,	PUNCT
fcis-10735	63	19	'	'	PUNCT
fcis-10735	63	20	alphabet	alphabet	NOUN
fcis-10735	63	21	'	'	PUNCT
fcis-10735	63	22	,	,	PUNCT
fcis-10735	63	23	'	'	PUNCT
fcis-10735	63	24	font	font	NOUN
fcis-10735	63	25	'	'	PUNCT
fcis-10735	63	26	,	,	PUNCT
fcis-10735	63	27	'	'	PUNCT
fcis-10735	63	28	emphasis	emphasis	NOUN
fcis-10735	63	29	'	'	PUNCT
fcis-10735	63	30	and	and	CCONJ
fcis-10735	63	31	'	'	PUNCT
fcis-10735	63	32	noise	noise	NOUN
fcis-10735	63	33	level	level	NOUN
fcis-10735	63	34	'	'	PUNCT
fcis-10735	63	35	as	as	SCONJ
fcis-10735	63	36	shown	show	VERB
fcis-10735	63	37	in	in	ADP
fcis-10735	63	38	figure	figure	NOUN
fcis-10735	63	39	9	9	NUM
fcis-10735	63	40	.	.	PUNCT
fcis-10735	63	41	fig	fig	NOUN
fcis-10735	63	42	8	8	NUM
fcis-10735	63	43	.	.	PUNCT
fcis-10735	64	1	different	different	ADJ
fcis-10735	64	2	languages	language	NOUN
fcis-10735	64	3	of	of	ADP
fcis-10735	64	4	fonts	font	NOUN
fcis-10735	64	5	fig	fig	NOUN
fcis-10735	64	6	9	9	NUM
fcis-10735	64	7	.	.	PUNCT
fcis-10735	65	1	fonts	font	NOUN
fcis-10735	65	2	information	information	NOUN
fcis-10735	65	3	1.2.2	1.2.2	NUM
fcis-10735	65	4	.	.	PUNCT
fcis-10735	66	1	feature	feature	NOUN
fcis-10735	66	2	analysis	analysis	NOUN
fcis-10735	66	3	pros	pro	NOUN
fcis-10735	66	4	:	:	PUNCT
fcis-10735	66	5	this	this	DET
fcis-10735	66	6	dataset	dataset	NOUN
fcis-10735	66	7	has	have	AUX
fcis-10735	66	8	font	font	VERB
fcis-10735	66	9	data	datum	NOUN
fcis-10735	66	10	in	in	ADP
fcis-10735	66	11	multiple	multiple	ADJ
fcis-10735	66	12	languages	language	NOUN
fcis-10735	66	13	and	and	CCONJ
fcis-10735	66	14	the	the	DET
fcis-10735	66	15	data	datum	NOUN
fcis-10735	66	16	is	be	AUX
fcis-10735	66	17	categorized	categorize	VERB
fcis-10735	66	18	in	in	ADP
fcis-10735	66	19	detail	detail	NOUN
fcis-10735	66	20	.	.	PUNCT
fcis-10735	67	1	the	the	DET
fcis-10735	67	2	number	number	NOUN
fcis-10735	67	3	of	of	ADP
fcis-10735	67	4	samples	sample	NOUN
fcis-10735	67	5	of	of	ADP
fcis-10735	67	6	each	each	DET
fcis-10735	67	7	font	font	NOUN
fcis-10735	67	8	is	be	AUX
fcis-10735	67	9	large	large	ADJ
fcis-10735	67	10	,	,	PUNCT
fcis-10735	67	11	which	which	PRON
fcis-10735	67	12	is	be	AUX
fcis-10735	67	13	conducive	conducive	ADJ
fcis-10735	67	14	to	to	ADP
fcis-10735	67	15	model	model	NOUN
fcis-10735	67	16	learning	learning	NOUN
fcis-10735	67	17	and	and	CCONJ
fcis-10735	67	18	generalization	generalization	NOUN
fcis-10735	67	19	using	use	VERB
fcis-10735	67	20	the	the	DET
fcis-10735	67	21	model	model	NOUN
fcis-10735	67	22	results	result	NOUN
fcis-10735	67	23	.	.	PUNCT
fcis-10735	68	1	the	the	DET
fcis-10735	68	2	dataset	dataset	NOUN
fcis-10735	68	3	annotates	annotate	VERB
fcis-10735	68	4	font	font	VERB
fcis-10735	68	5	details	detail	NOUN
fcis-10735	68	6	such	such	ADJ
fcis-10735	68	7	as	as	ADP
fcis-10735	68	8	italics	italic	NOUN
fcis-10735	68	9	and	and	CCONJ
fcis-10735	68	10	bold	bold	ADJ
fcis-10735	68	11	.	.	PUNCT
fcis-10735	69	1	cons	con	NOUN
fcis-10735	69	2	:	:	PUNCT
fcis-10735	69	3	font	font	VERB
fcis-10735	69	4	styles	style	NOUN
fcis-10735	69	5	other	other	ADJ
fcis-10735	69	6	than	than	ADP
fcis-10735	69	7	those	those	PRON
fcis-10735	69	8	in	in	ADP
fcis-10735	69	9	the	the	DET
fcis-10735	69	10	dataset	dataset	NOUN
fcis-10735	69	11	are	be	AUX
fcis-10735	69	12	not	not	PART
fcis-10735	69	13	recognized	recognize	VERB
fcis-10735	69	14	.	.	PUNCT
fcis-10735	70	1	at	at	ADP
fcis-10735	70	2	the	the	DET
fcis-10735	70	3	same	same	ADJ
fcis-10735	70	4	time	time	NOUN
fcis-10735	70	5	,	,	PUNCT
fcis-10735	70	6	the	the	DET
fcis-10735	70	7	text	text	NOUN
fcis-10735	70	8	that	that	PRON
fcis-10735	70	9	is	be	AUX
fcis-10735	70	10	not	not	PART
fcis-10735	70	11	in	in	ADP
fcis-10735	70	12	the	the	DET
fcis-10735	70	13	dataset	dataset	NOUN
fcis-10735	70	14	may	may	AUX
fcis-10735	70	15	not	not	PART
fcis-10735	70	16	be	be	AUX
fcis-10735	70	17	recognized	recognize	VERB
fcis-10735	70	18	.	.	PUNCT
fcis-10735	71	1	all	all	DET
fcis-10735	71	2	text	text	NOUN
fcis-10735	71	3	is	be	AUX
fcis-10735	71	4	concentrated	concentrate	VERB
fcis-10735	71	5	on	on	ADP
fcis-10735	71	6	a	a	DET
fcis-10735	71	7	single	single	ADJ
fcis-10735	71	8	image	image	NOUN
fcis-10735	71	9	.	.	PUNCT
fcis-10735	72	1	this	this	PRON
fcis-10735	72	2	requires	require	VERB
fcis-10735	72	3	splitting	split	VERB
fcis-10735	72	4	the	the	DET
fcis-10735	72	5	text	text	NOUN
fcis-10735	72	6	on	on	ADP
fcis-10735	72	7	the	the	DET
fcis-10735	72	8	image	image	NOUN
fcis-10735	72	9	.	.	PUNCT
fcis-10735	73	1	1.2.3	1.2.3	X
fcis-10735	73	2	.	.	PUNCT
fcis-10735	73	3	data	data	PROPN
fcis-10735	73	4	pre	pre	ADJ
fcis-10735	73	5	-	-	ADJ
fcis-10735	73	6	processing	process	VERB
fcis-10735	73	7	1.2.3.1	1.2.3.1	NUM
fcis-10735	73	8	character	character	NOUN
fcis-10735	73	9	recognition	recognition	NOUN
fcis-10735	73	10	and	and	CCONJ
fcis-10735	73	11	location	location	NOUN
fcis-10735	73	12	recognition	recognition	NOUN
fcis-10735	73	13	i	i	PRON
fcis-10735	73	14	use	use	VERB
fcis-10735	73	15	the	the	DET
fcis-10735	73	16	chinese	chinese	ADJ
fcis-10735	73	17	traditional	traditional	ADJ
fcis-10735	73	18	recognition	recognition	NOUN
fcis-10735	73	19	module	module	NOUN
fcis-10735	73	20	of	of	ADP
fcis-10735	73	21	google	google	PROPN
fcis-10735	73	22	tesseract	tesseract	NOUN
fcis-10735	73	23	to	to	PART
fcis-10735	73	24	recognize	recognize	VERB
fcis-10735	73	25	the	the	DET
fcis-10735	73	26	dataset	dataset	NOUN
fcis-10735	73	27	images	image	NOUN
fcis-10735	73	28	.	.	PUNCT
fcis-10735	74	1	the	the	DET
fcis-10735	74	2	algorithm	algorithm	NOUN
fcis-10735	74	3	will	will	AUX
fcis-10735	74	4	locate	locate	VERB
fcis-10735	74	5	all	all	DET
fcis-10735	74	6	the	the	DET
fcis-10735	74	7	recognized	recognize	VERB
fcis-10735	74	8	text	text	NOUN
fcis-10735	74	9	on	on	ADP
fcis-10735	74	10	the	the	DET
fcis-10735	74	11	entire	entire	ADJ
fcis-10735	74	12	image	image	NOUN
fcis-10735	74	13	.	.	PUNCT
fcis-10735	75	1	at	at	ADP
fcis-10735	75	2	the	the	DET
fcis-10735	75	3	same	same	ADJ
fcis-10735	75	4	time	time	NOUN
fcis-10735	75	5	,	,	PUNCT
fcis-10735	75	6	it	it	PRON
fcis-10735	75	7	will	will	AUX
fcis-10735	75	8	also	also	ADV
fcis-10735	75	9	provide	provide	VERB
fcis-10735	75	10	the	the	DET
fcis-10735	75	11	height	height	NOUN
fcis-10735	75	12	,	,	PUNCT
fcis-10735	75	13	width	width	VERB
fcis-10735	75	14	and	and	CCONJ
fcis-10735	75	15	text	text	NOUN
fcis-10735	75	16	content	content	NOUN
fcis-10735	75	17	information	information	NOUN
fcis-10735	75	18	of	of	ADP
fcis-10735	75	19	the	the	DET
fcis-10735	75	20	text	text	NOUN
fcis-10735	75	21	.	.	PUNCT
fcis-10735	76	1	1.2.3.2	1.2.3.2	NUM
fcis-10735	76	2	text	text	NOUN
fcis-10735	76	3	image	image	NOUN
fcis-10735	76	4	segmentation	segmentation	NOUN
fcis-10735	76	5	the	the	DET
fcis-10735	76	6	algorithm	algorithm	NOUN
fcis-10735	76	7	uses	use	VERB
fcis-10735	76	8	the	the	DET
fcis-10735	76	9	location	location	NOUN
fcis-10735	76	10	information	information	NOUN
fcis-10735	76	11	of	of	ADP
fcis-10735	76	12	the	the	DET
fcis-10735	76	13	text	text	NOUN
fcis-10735	76	14	in	in	ADP
fcis-10735	76	15	9	9	NUM
fcis-10735	76	16	the	the	DET
fcis-10735	76	17	image	image	NOUN
fcis-10735	76	18	to	to	PART
fcis-10735	76	19	segment	segment	VERB
fcis-10735	76	20	the	the	DET
fcis-10735	76	21	image	image	NOUN
fcis-10735	76	22	.	.	PUNCT
fcis-10735	77	1	it	it	PRON
fcis-10735	77	2	segments	segment	VERB
fcis-10735	77	3	the	the	DET
fcis-10735	77	4	image	image	NOUN
fcis-10735	77	5	and	and	CCONJ
fcis-10735	77	6	sets	set	VERB
fcis-10735	77	7	the	the	DET
fcis-10735	77	8	size	size	NOUN
fcis-10735	77	9	to	to	ADP
fcis-10735	77	10	64	64	NUM
fcis-10735	77	11	*	*	SYM
fcis-10735	77	12	64	64	NUM
fcis-10735	77	13	.	.	PUNCT
fcis-10735	78	1	finally	finally	ADV
fcis-10735	78	2	,	,	PUNCT
fcis-10735	78	3	the	the	DET
fcis-10735	78	4	algorithm	algorithm	NOUN
fcis-10735	78	5	stores	store	VERB
fcis-10735	78	6	all	all	DET
fcis-10735	78	7	the	the	DET
fcis-10735	78	8	generated	generate	VERB
fcis-10735	78	9	images	image	NOUN
fcis-10735	78	10	in	in	ADP
fcis-10735	78	11	a	a	DET
fcis-10735	78	12	folder	folder	NOUN
fcis-10735	78	13	.	.	PUNCT
fcis-10735	79	1	2	2	X
fcis-10735	79	2	.	.	X
fcis-10735	79	3	model	model	PROPN
fcis-10735	79	4	fig	fig	PROPN
fcis-10735	79	5	10	10	NUM
fcis-10735	79	6	.	.	PUNCT
fcis-10735	80	1	workflow	workflow	NOUN
fcis-10735	80	2	2.1	2.1	NUM
fcis-10735	80	3	.	.	PUNCT
fcis-10735	81	1	random	random	ADJ
fcis-10735	81	2	forest	forest	NOUN
fcis-10735	81	3	algorithm	algorithm	NOUN
fcis-10735	81	4	2.1.1	2.1.1	NUM
fcis-10735	81	5	.	.	PUNCT
fcis-10735	82	1	workflow	workflow	VERB
fcis-10735	82	2	the	the	DET
fcis-10735	82	3	algorithm	algorithm	NOUN
fcis-10735	82	4	flow	flow	NOUN
fcis-10735	82	5	of	of	ADP
fcis-10735	82	6	random	random	ADJ
fcis-10735	82	7	forest	forest	NOUN
fcis-10735	82	8	:	:	PUNCT
fcis-10735	82	9	(	(	PUNCT
fcis-10735	82	10	1	1	X
fcis-10735	82	11	)	)	PUNCT
fcis-10735	82	12	the	the	DET
fcis-10735	82	13	algorithm	algorithm	NOUN
fcis-10735	82	14	uses	use	VERB
fcis-10735	82	15	bootstrap	bootstrap	NOUN
fcis-10735	82	16	to	to	PART
fcis-10735	82	17	extract	extract	VERB
fcis-10735	82	18	n	n	DET
fcis-10735	82	19	samples	sample	NOUN
fcis-10735	82	20	with	with	ADP
fcis-10735	82	21	replacement	replacement	NOUN
fcis-10735	82	22	from	from	ADP
fcis-10735	82	23	the	the	DET
fcis-10735	82	24	training	training	NOUN
fcis-10735	82	25	set	set	NOUN
fcis-10735	82	26	r	r	NOUN
fcis-10735	82	27	and	and	CCONJ
fcis-10735	82	28	rk	rk	NOUN
fcis-10735	82	29	is	be	AUX
fcis-10735	82	30	a	a	DET
fcis-10735	82	31	training	training	NOUN
fcis-10735	82	32	subset	subset	NOUN
fcis-10735	82	33	of	of	ADP
fcis-10735	82	34	r.	r.	PROPN
fcis-10735	82	35	fig	fig	PROPN
fcis-10735	82	36	11	11	NUM
fcis-10735	82	37	.	.	PUNCT
fcis-10735	83	1	workflow	workflow	NOUN
fcis-10735	83	2	of	of	ADP
fcis-10735	83	3	random	random	ADJ
fcis-10735	83	4	forest	forest	NOUN
fcis-10735	83	5	algorithm	algorithm	NOUN
fcis-10735	83	6	[	[	X
fcis-10735	83	7	2	2	NUM
fcis-10735	83	8	]	]	PUNCT
fcis-10735	83	9	(	(	PUNCT
fcis-10735	83	10	2	2	NUM
fcis-10735	83	11	)	)	PUNCT
fcis-10735	83	12	for	for	ADP
fcis-10735	83	13	the	the	DET
fcis-10735	83	14	training	training	NOUN
fcis-10735	83	15	subset	subset	VERB
fcis-10735	83	16	rk	rk	NOUN
fcis-10735	83	17	,	,	PUNCT
fcis-10735	83	18	the	the	DET
fcis-10735	83	19	algorithm	algorithm	NOUN
fcis-10735	83	20	randomly	randomly	ADV
fcis-10735	83	21	extracts	extract	NOUN
fcis-10735	83	22	m	m	VERB
fcis-10735	83	23	features	feature	NOUN
fcis-10735	83	24	from	from	ADP
fcis-10735	83	25	the	the	DET
fcis-10735	83	26	feature	feature	NOUN
fcis-10735	83	27	set	set	VERB
fcis-10735	83	28	f	f	PROPN
fcis-10735	83	29	without	without	ADP
fcis-10735	83	30	replacement	replacement	NOUN
fcis-10735	83	31	.	.	PUNCT
fcis-10735	84	1	among	among	ADP
fcis-10735	84	2	them	they	PRON
fcis-10735	84	3	,	,	PUNCT
fcis-10735	84	4	m	m	NOUN
fcis-10735	84	5	=	=	NOUN
fcis-10735	84	6	log2^m	log2^m	NOUN
fcis-10735	84	7	.	.	PUNCT
fcis-10735	85	1	it	it	PRON
fcis-10735	85	2	is	be	AUX
fcis-10735	85	3	used	use	VERB
fcis-10735	85	4	as	as	ADP
fcis-10735	85	5	the	the	DET
fcis-10735	85	6	basis	basis	NOUN
fcis-10735	85	7	for	for	ADP
fcis-10735	85	8	the	the	DET
fcis-10735	85	9	split	split	NOUN
fcis-10735	85	10	of	of	ADP
fcis-10735	85	11	each	each	DET
fcis-10735	85	12	node	node	NOUN
fcis-10735	85	13	on	on	ADP
fcis-10735	85	14	the	the	DET
fcis-10735	85	15	decision	decision	NOUN
fcis-10735	85	16	tree	tree	NOUN
fcis-10735	85	17	to	to	PART
fcis-10735	85	18	generate	generate	VERB
fcis-10735	85	19	a	a	DET
fcis-10735	85	20	complete	complete	ADJ
fcis-10735	85	21	decision	decision	NOUN
fcis-10735	85	22	tree	tree	NOUN
fcis-10735	85	23	sk	sk	VERB
fcis-10735	85	24	from	from	ADP
fcis-10735	85	25	the	the	DET
fcis-10735	85	26	root	root	NOUN
fcis-10735	85	27	node	node	NOUN
fcis-10735	85	28	from	from	ADP
fcis-10735	85	29	top	top	NOUN
fcis-10735	85	30	to	to	ADP
fcis-10735	85	31	bottom	bottom	NOUN
fcis-10735	85	32	without	without	ADP
fcis-10735	85	33	pruning	prune	VERB
fcis-10735	85	34	.	.	PUNCT
fcis-10735	86	1	(	(	PUNCT
fcis-10735	86	2	3	3	X
fcis-10735	86	3	)	)	PUNCT
fcis-10735	86	4	the	the	DET
fcis-10735	86	5	algorithm	algorithm	NOUN
fcis-10735	86	6	repeats	repeat	VERB
fcis-10735	86	7	steps	step	NOUN
fcis-10735	86	8	1	1	NUM
fcis-10735	86	9	and	and	CCONJ
fcis-10735	86	10	2	2	NUM
fcis-10735	86	11	n	n	NUM
fcis-10735	86	12	times	time	NOUN
fcis-10735	86	13	to	to	PART
fcis-10735	86	14	get	get	VERB
fcis-10735	86	15	n	n	PRON
fcis-10735	86	16	training	train	VERB
fcis-10735	86	17	subsets	subset	NOUN
fcis-10735	86	18	r1	r1	NOUN
fcis-10735	86	19	,	,	PUNCT
fcis-10735	86	20	r2	r2	PROPN
fcis-10735	86	21	…	…	SYM
fcis-10735	86	22	rn	rn	NOUN
fcis-10735	86	23	.	.	PROPN
fcis-10735	87	1	at	at	ADP
fcis-10735	87	2	the	the	DET
fcis-10735	87	3	same	same	ADJ
fcis-10735	87	4	time	time	NOUN
fcis-10735	87	5	,	,	PUNCT
fcis-10735	87	6	it	it	PRON
fcis-10735	87	7	generates	generate	VERB
fcis-10735	87	8	decision	decision	NOUN
fcis-10735	87	9	trees	tree	NOUN
fcis-10735	87	10	s1	s1	NOUN
fcis-10735	87	11	,	,	PUNCT
fcis-10735	87	12	s2	s2	PROPN
fcis-10735	87	13	…	…	SYM
fcis-10735	87	14	sn	sn	PROPN
fcis-10735	87	15	and	and	CCONJ
fcis-10735	87	16	combines	combine	VERB
fcis-10735	87	17	n	n	PRON
fcis-10735	87	18	decision	decision	NOUN
fcis-10735	87	19	trees	tree	NOUN
fcis-10735	87	20	to	to	PART
fcis-10735	87	21	form	form	VERB
fcis-10735	87	22	a	a	DET
fcis-10735	87	23	random	random	ADJ
fcis-10735	87	24	forest	forest	NOUN
fcis-10735	87	25	.	.	PUNCT
fcis-10735	88	1	(	(	PUNCT
fcis-10735	88	2	4	4	X
fcis-10735	88	3	)	)	PUNCT
fcis-10735	88	4	the	the	DET
fcis-10735	88	5	algorithm	algorithm	NOUN
fcis-10735	88	6	inputs	input	VERB
fcis-10735	88	7	the	the	DET
fcis-10735	88	8	sample	sample	NOUN
fcis-10735	88	9	dμ	dμ	VERB
fcis-10735	88	10	of	of	ADP
fcis-10735	88	11	the	the	DET
fcis-10735	88	12	test	test	NOUN
fcis-10735	88	13	set	set	VERB
fcis-10735	88	14	d	d	NOUN
fcis-10735	88	15	into	into	ADP
fcis-10735	88	16	the	the	DET
fcis-10735	88	17	random	random	ADJ
fcis-10735	88	18	forest	forest	NOUN
fcis-10735	88	19	and	and	CCONJ
fcis-10735	88	20	lets	let	VERB
fcis-10735	88	21	each	each	DET
fcis-10735	88	22	decision	decision	NOUN
fcis-10735	88	23	tree	tree	NOUN
fcis-10735	88	24	make	make	VERB
fcis-10735	88	25	a	a	DET
fcis-10735	88	26	decision	decision	NOUN
fcis-10735	88	27	on	on	ADP
fcis-10735	88	28	dμ	dμ	NOUN
fcis-10735	88	29	.	.	PUNCT
fcis-10735	89	1	then	then	ADV
fcis-10735	89	2	,	,	PUNCT
fcis-10735	89	3	it	it	PRON
fcis-10735	89	4	uses	use	VERB
fcis-10735	89	5	the	the	DET
fcis-10735	89	6	majority	majority	NOUN
fcis-10735	89	7	voting	voting	NOUN
fcis-10735	89	8	method	method	NOUN
fcis-10735	89	9	to	to	PART
fcis-10735	89	10	vote	vote	VERB
fcis-10735	89	11	on	on	ADP
fcis-10735	89	12	the	the	DET
fcis-10735	89	13	decision	decision	NOUN
fcis-10735	89	14	result	result	NOUN
fcis-10735	89	15	and	and	CCONJ
fcis-10735	89	16	finally	finally	ADV
fcis-10735	89	17	decides	decide	VERB
fcis-10735	89	18	the	the	DET
fcis-10735	89	19	classification	classification	NOUN
fcis-10735	89	20	of	of	ADP
fcis-10735	89	21	dμ	dμ	PROPN
fcis-10735	89	22	.	.	PUNCT
fcis-10735	90	1	(	(	PUNCT
fcis-10735	90	2	5	5	X
fcis-10735	90	3	)	)	PUNCT
fcis-10735	90	4	the	the	DET
fcis-10735	90	5	algorithm	algorithm	NOUN
fcis-10735	90	6	repeats	repeat	VERB
fcis-10735	90	7	step	step	VERB
fcis-10735	90	8	4	4	NUM
fcis-10735	90	9	for	for	ADP
fcis-10735	90	10	λ	λ	PROPN
fcis-10735	90	11	times	time	NOUN
fcis-10735	90	12	until	until	ADP
fcis-10735	90	13	the	the	DET
fcis-10735	90	14	classification	classification	NOUN
fcis-10735	90	15	of	of	ADP
fcis-10735	90	16	test	test	NOUN
fcis-10735	90	17	set	set	VERB
fcis-10735	90	18	d	d	NOUN
fcis-10735	90	19	is	be	AUX
fcis-10735	90	20	completed	complete	VERB
fcis-10735	90	21	.	.	PUNCT
fcis-10735	91	1	2.1.2	2.1.2	X
fcis-10735	91	2	.	.	PUNCT
fcis-10735	91	3	hardware	hardware	NOUN
fcis-10735	91	4	usage	usage	NOUN
fcis-10735	91	5	the	the	DET
fcis-10735	91	6	random	random	ADJ
fcis-10735	91	7	forest	forest	NOUN
fcis-10735	91	8	algorithm	algorithm	NOUN
fcis-10735	91	9	uses	use	VERB
fcis-10735	91	10	cpu	cpu	NOUN
fcis-10735	91	11	to	to	PART
fcis-10735	91	12	perform	perform	VERB
fcis-10735	91	13	operations	operation	NOUN
fcis-10735	91	14	.	.	PUNCT
fcis-10735	92	1	there	there	PRON
fcis-10735	92	2	is	be	VERB
fcis-10735	92	3	no	no	DET
fcis-10735	92	4	error	error	NOUN
fcis-10735	92	5	during	during	ADP
fcis-10735	92	6	the	the	DET
fcis-10735	92	7	whole	whole	ADJ
fcis-10735	92	8	running	running	NOUN
fcis-10735	92	9	process	process	NOUN
fcis-10735	92	10	and	and	CCONJ
fcis-10735	92	11	the	the	DET
fcis-10735	92	12	operation	operation	NOUN
fcis-10735	92	13	speed	speed	NOUN
fcis-10735	92	14	is	be	AUX
fcis-10735	92	15	fast	fast	ADJ
fcis-10735	92	16	.	.	PUNCT
fcis-10735	93	1	hardware	hardware	NOUN
fcis-10735	93	2	list	list	NOUN
fcis-10735	93	3	:	:	PUNCT
fcis-10735	93	4	device	device	NOUN
fcis-10735	93	5	name	name	NOUN
fcis-10735	93	6	:	:	PUNCT
fcis-10735	93	7	laptop-9lij8tiu	laptop-9lij8tiu	NOUN
fcis-10735	93	8	processor	processor	NOUN
fcis-10735	93	9	:	:	PUNCT
fcis-10735	93	10	11th	11th	ADJ
fcis-10735	93	11	gen	gen	PROPN
fcis-10735	93	12	intel(r	intel(r	PROPN
fcis-10735	93	13	)	)	PUNCT
fcis-10735	93	14	core	core	NOUN
fcis-10735	93	15	(	(	PUNCT
fcis-10735	93	16	tm	tm	NOUN
fcis-10735	93	17	)	)	PUNCT
fcis-10735	93	18	i9	i9	NOUN
fcis-10735	93	19	-	-	PUNCT
fcis-10735	93	20	11900h	11900h	NUM
fcis-10735	93	21	@	@	ADP
fcis-10735	93	22	2.50ghz	2.50ghz	NUM
fcis-10735	93	23	2.50	2.50	NUM
fcis-10735	93	24	ghz	ghz	NOUN
fcis-10735	93	25	on	on	ADP
fcis-10735	93	26	-	-	PUNCT
fcis-10735	93	27	board	board	NOUN
fcis-10735	93	28	ram	ram	NOUN
fcis-10735	93	29	:	:	PUNCT
fcis-10735	93	30	16.0	16.0	NUM
fcis-10735	93	31	gb	gb	NOUN
fcis-10735	93	32	(	(	PUNCT
fcis-10735	93	33	15.7	15.7	NUM
fcis-10735	93	34	gb	gb	ADV
fcis-10735	93	35	available	available	ADJ
fcis-10735	93	36	)	)	PUNCT
fcis-10735	93	37	system	system	NOUN
fcis-10735	93	38	type	type	NOUN
fcis-10735	93	39	:	:	PUNCT
fcis-10735	93	40	64	64	NUM
fcis-10735	93	41	-	-	PUNCT
fcis-10735	93	42	bit	bit	NOUN
fcis-10735	93	43	operating	operate	VERB
fcis-10735	93	44	system	system	NOUN
fcis-10735	93	45	,	,	PUNCT
fcis-10735	93	46	x64	x64	NOUN
fcis-10735	93	47	-	-	PUNCT
fcis-10735	93	48	based	base	VERB
fcis-10735	93	49	processor	processor	NOUN
fcis-10735	93	50	2.1.3	2.1.3	NUM
fcis-10735	93	51	.	.	PUNCT
fcis-10735	94	1	theory	theory	NOUN
fcis-10735	94	2	an	an	DET
fcis-10735	94	3	ensemble	ensemble	ADJ
fcis-10735	94	4	model	model	NOUN
fcis-10735	94	5	like	like	ADP
fcis-10735	94	6	random	random	ADJ
fcis-10735	94	7	forest	forest	NOUN
fcis-10735	94	8	is	be	AUX
fcis-10735	94	9	composed	compose	VERB
fcis-10735	94	10	of	of	ADP
fcis-10735	94	11	a	a	DET
fcis-10735	94	12	bunch	bunch	NOUN
fcis-10735	94	13	of	of	ADP
fcis-10735	94	14	decision	decision	NOUN
fcis-10735	94	15	trees	tree	NOUN
fcis-10735	94	16	.	.	PUNCT
fcis-10735	95	1	there	there	PRON
fcis-10735	95	2	are	be	VERB
fcis-10735	95	3	multiple	multiple	ADJ
fcis-10735	95	4	algorithms	algorithm	NOUN
fcis-10735	95	5	for	for	ADP
fcis-10735	95	6	decision	decision	NOUN
fcis-10735	95	7	trees	tree	NOUN
fcis-10735	95	8	,	,	PUNCT
fcis-10735	95	9	the	the	DET
fcis-10735	95	10	most	most	ADV
fcis-10735	95	11	commonly	commonly	ADV
fcis-10735	95	12	used	use	VERB
fcis-10735	95	13	ones	one	NOUN
fcis-10735	95	14	are	be	AUX
fcis-10735	95	15	id3	id3	NOUN
fcis-10735	95	16	and	and	CCONJ
fcis-10735	95	17	cart	cart	NOUN
fcis-10735	95	18	.	.	PUNCT
fcis-10735	96	1	id3	id3	NOUN
fcis-10735	96	2	trees	tree	NOUN
fcis-10735	96	3	use	use	VERB
fcis-10735	96	4	information	information	NOUN
fcis-10735	96	5	gain	gain	NOUN
fcis-10735	96	6	,	,	PUNCT
fcis-10735	96	7	while	while	SCONJ
fcis-10735	96	8	cart	cart	NOUN
fcis-10735	96	9	trees	tree	NOUN
fcis-10735	96	10	use	use	VERB
fcis-10735	96	11	the	the	DET
fcis-10735	96	12	gini	gini	PROPN
fcis-10735	96	13	index	index	NOUN
fcis-10735	96	14	to	to	PART
fcis-10735	96	15	decide	decide	VERB
fcis-10735	96	16	when	when	SCONJ
fcis-10735	96	17	to	to	PART
fcis-10735	96	18	split	split	VERB
fcis-10735	96	19	.	.	PUNCT
fcis-10735	97	1	random	random	ADJ
fcis-10735	97	2	forest	forest	NOUN
fcis-10735	97	3	is	be	AUX
fcis-10735	97	4	designed	design	VERB
fcis-10735	97	5	to	to	PART
fcis-10735	97	6	reduce	reduce	VERB
fcis-10735	97	7	overfitting	overfitting	NOUN
fcis-10735	97	8	and	and	CCONJ
fcis-10735	97	9	variance	variance	NOUN
fcis-10735	97	10	by	by	ADP
fcis-10735	97	11	using	use	VERB
fcis-10735	97	12	a	a	DET
fcis-10735	97	13	bagging	bagging	NOUN
fcis-10735	97	14	algorithm	algorithm	NOUN
fcis-10735	97	15	.	.	PUNCT
fcis-10735	98	1	2.1.3.1	2.1.3.1	NUM
fcis-10735	98	2	id3	id3	NOUN
fcis-10735	98	3	tree	tree	NOUN
fcis-10735	98	4	and	and	CCONJ
fcis-10735	98	5	information	information	NOUN
fcis-10735	98	6	gain	gain	VERB
fcis-10735	98	7	the	the	DET
fcis-10735	98	8	whole	whole	ADJ
fcis-10735	98	9	point	point	NOUN
fcis-10735	98	10	of	of	ADP
fcis-10735	98	11	an	an	DET
fcis-10735	98	12	id3	id3	NOUN
fcis-10735	98	13	tree	tree	NOUN
fcis-10735	98	14	is	be	AUX
fcis-10735	98	15	to	to	PART
fcis-10735	98	16	maximize	maximize	VERB
fcis-10735	98	17	information	information	NOUN
fcis-10735	98	18	gain	gain	NOUN
fcis-10735	98	19	and	and	CCONJ
fcis-10735	98	20	information	information	NOUN
fcis-10735	98	21	gain	gain	NOUN
fcis-10735	98	22	is	be	AUX
fcis-10735	98	23	a	a	DET
fcis-10735	98	24	measure	measure	NOUN
fcis-10735	98	25	of	of	ADP
fcis-10735	98	26	impurity	impurity	NOUN
fcis-10735	98	27	using	use	VERB
fcis-10735	98	28	entropy	entropy	PROPN
fcis-10735	98	29	.	.	PUNCT
fcis-10735	99	1	entropy	entropy	PROPN
fcis-10735	99	2	is	be	AUX
fcis-10735	99	3	a	a	DET
fcis-10735	99	4	measure	measure	NOUN
fcis-10735	99	5	of	of	ADP
fcis-10735	99	6	(	(	PUNCT
fcis-10735	99	7	dis)order	dis)order	PROPN
fcis-10735	99	8	,	,	PUNCT
fcis-10735	99	9	which	which	PRON
fcis-10735	99	10	is	be	AUX
fcis-10735	99	11	able	able	ADJ
fcis-10735	99	12	to	to	PART
fcis-10735	99	13	represent	represent	VERB
fcis-10735	99	14	the	the	DET
fcis-10735	99	15	missing	miss	VERB
fcis-10735	99	16	flow	flow	NOUN
fcis-10735	99	17	of	of	ADP
fcis-10735	99	18	information	information	NOUN
fcis-10735	99	19	,	,	PUNCT
fcis-10735	99	20	or	or	CCONJ
fcis-10735	99	21	the	the	DET
fcis-10735	99	22	degree	degree	NOUN
fcis-10735	99	23	of	of	ADP
fcis-10735	99	24	confusion	confusion	NOUN
fcis-10735	99	25	in	in	ADP
fcis-10735	99	26	the	the	DET
fcis-10735	99	27	data	datum	NOUN
fcis-10735	99	28	.	.	PUNCT
fcis-10735	100	1	entry	entry	NOUN
fcis-10735	100	2	publicity	publicity	NOUN
fcis-10735	100	3	:	:	PUNCT
fcis-10735	100	4	2.1.3.2	2.1.3.2	NUM
fcis-10735	100	5	cart	cart	NOUN
fcis-10735	100	6	tree	tree	NOUN
fcis-10735	100	7	and	and	CCONJ
fcis-10735	100	8	gini	gini	PROPN
fcis-10735	100	9	index	index	NOUN
fcis-10735	100	10	the	the	DET
fcis-10735	100	11	decision	decision	NOUN
fcis-10735	100	12	tree	tree	NOUN
fcis-10735	100	13	of	of	ADP
fcis-10735	100	14	the	the	DET
fcis-10735	100	15	cart	cart	NOUN
fcis-10735	100	16	algorithm	algorithm	NOUN
fcis-10735	100	17	is	be	AUX
fcis-10735	100	18	designed	design	VERB
fcis-10735	100	19	to	to	PART
fcis-10735	100	20	minimize	minimize	VERB
fcis-10735	100	21	the	the	DET
fcis-10735	100	22	gini	gini	PROPN
fcis-10735	100	23	index	index	NOUN
fcis-10735	100	24	.	.	PUNCT
fcis-10735	101	1	the	the	DET
fcis-10735	101	2	gini	gini	PROPN
fcis-10735	101	3	index	index	NOUN
fcis-10735	101	4	can	can	AUX
fcis-10735	101	5	represent	represent	VERB
fcis-10735	101	6	how	how	SCONJ
fcis-10735	101	7	often	often	ADV
fcis-10735	101	8	randomly	randomly	ADV
fcis-10735	101	9	selected	select	VERB
fcis-10735	101	10	data	datum	NOUN
fcis-10735	101	11	points	point	NOUN
fcis-10735	101	12	in	in	ADP
fcis-10735	101	13	a	a	DET
fcis-10735	101	14	dataset	dataset	NOUN
fcis-10735	101	15	may	may	AUX
fcis-10735	101	16	be	be	AUX
fcis-10735	101	17	misclassified	misclassifie	VERB
fcis-10735	101	18	.	.	PUNCT
fcis-10735	102	1	the	the	DET
fcis-10735	102	2	gini	gini	PROPN
fcis-10735	102	3	index	index	NOUN
fcis-10735	102	4	is	be	AUX
fcis-10735	102	5	calculated	calculate	VERB
fcis-10735	102	6	as	as	ADP
fcis-10735	102	7	:	:	PUNCT
fcis-10735	102	8	among	among	ADP
fcis-10735	102	9	them	they	PRON
fcis-10735	102	10	,	,	PUNCT
fcis-10735	102	11	pk	pk	NOUN
fcis-10735	102	12	represents	represent	VERB
fcis-10735	102	13	the	the	DET
fcis-10735	102	14	probability	probability	NOUN
fcis-10735	102	15	that	that	SCONJ
fcis-10735	102	16	the	the	DET
fcis-10735	102	17	sample	sample	NOUN
fcis-10735	102	18	belongs	belong	VERB
fcis-10735	102	19	to	to	ADP
fcis-10735	102	20	the	the	DET
fcis-10735	102	21	kth	kth	PROPN
fcis-10735	102	22	category	category	NOUN
fcis-10735	102	23	.	.	PUNCT
fcis-10735	103	1	2.1.3.3	2.1.3.3	NUM
fcis-10735	103	2	bagging	bag	VERB
fcis-10735	103	3	algorithm	algorithm	NOUN
fcis-10735	103	4	the	the	DET
fcis-10735	103	5	bagging	bagging	NOUN
fcis-10735	103	6	algorithm	algorithm	NOUN
fcis-10735	103	7	is	be	AUX
fcis-10735	103	8	mainly	mainly	ADV
fcis-10735	103	9	used	use	VERB
fcis-10735	103	10	to	to	PART
fcis-10735	103	11	deal	deal	VERB
fcis-10735	103	12	with	with	ADP
fcis-10735	103	13	weak	weak	ADJ
fcis-10735	103	14	learners	learner	NOUN
fcis-10735	103	15	.	.	PUNCT
fcis-10735	104	1	the	the	DET
fcis-10735	104	2	accuracy	accuracy	NOUN
fcis-10735	104	3	of	of	ADP
fcis-10735	104	4	the	the	DET
fcis-10735	104	5	weak	weak	ADJ
fcis-10735	104	6	learning	learning	NOUN
fcis-10735	104	7	algorithm	algorithm	NOUN
fcis-10735	104	8	is	be	AUX
fcis-10735	104	9	not	not	PART
fcis-10735	104	10	high	high	ADJ
fcis-10735	104	11	,	,	PUNCT
fcis-10735	104	12	but	but	CCONJ
fcis-10735	104	13	the	the	DET
fcis-10735	104	14	accuracy	accuracy	NOUN
fcis-10735	104	15	of	of	ADP
fcis-10735	104	16	the	the	DET
fcis-10735	104	17	prediction	prediction	NOUN
fcis-10735	104	18	function	function	NOUN
fcis-10735	104	19	sequence	sequence	NOUN
fcis-10735	104	20	is	be	AUX
fcis-10735	104	21	improved	improve	VERB
fcis-10735	104	22	after	after	ADP
fcis-10735	104	23	repeated	repeat	VERB
fcis-10735	104	24	use	use	NOUN
fcis-10735	104	25	[	[	X
fcis-10735	104	26	3	3	NUM
fcis-10735	104	27	]	]	PUNCT
fcis-10735	104	28	.	.	PUNCT
fcis-10735	105	1	the	the	DET
fcis-10735	105	2	random	random	ADJ
fcis-10735	105	3	forest	forest	NOUN
fcis-10735	105	4	uses	use	VERB
fcis-10735	105	5	the	the	DET
fcis-10735	105	6	bagging	bagging	NOUN
fcis-10735	105	7	algorithm	algorithm	NOUN
fcis-10735	105	8	to	to	PART
fcis-10735	105	9	avoid	avoid	VERB
fcis-10735	105	10	the	the	DET
fcis-10735	105	11	defects	defect	NOUN
fcis-10735	105	12	of	of	ADP
fcis-10735	105	13	high	high	ADJ
fcis-10735	105	14	variance	variance	NOUN
fcis-10735	105	15	and	and	CCONJ
fcis-10735	105	16	10	10	NUM
fcis-10735	105	17	overfitting	overfitting	NOUN
fcis-10735	105	18	.	.	PUNCT
fcis-10735	106	1	fig	fig	NOUN
fcis-10735	106	2	12	12	NUM
fcis-10735	106	3	.	.	PUNCT
fcis-10735	107	1	bootstrapping	bootstrappe	VERB
fcis-10735	107	2	[	[	X
fcis-10735	107	3	3	3	NUM
fcis-10735	107	4	]	]	SYM
fcis-10735	107	5	2.1.4	2.1.4	NUM
fcis-10735	107	6	.	.	PUNCT
fcis-10735	108	1	implementation	implementation	NOUN
fcis-10735	108	2	2.1.4.1	2.1.4.1	NUM
fcis-10735	108	3	import	import	NOUN
fcis-10735	108	4	all	all	DET
fcis-10735	108	5	required	require	VERB
fcis-10735	108	6	libraries	library	NOUN
fcis-10735	108	7	the	the	DET
fcis-10735	108	8	random	random	ADJ
fcis-10735	108	9	forest	forest	NOUN
fcis-10735	108	10	algorithm	algorithm	NOUN
fcis-10735	108	11	mainly	mainly	ADV
fcis-10735	108	12	uses	use	VERB
fcis-10735	108	13	the	the	DET
fcis-10735	108	14	sklearn	sklearn	ADJ
fcis-10735	108	15	platform	platform	NOUN
fcis-10735	108	16	for	for	ADP
fcis-10735	108	17	model	model	NOUN
fcis-10735	108	18	building	building	NOUN
fcis-10735	108	19	and	and	CCONJ
fcis-10735	108	20	training	training	NOUN
fcis-10735	108	21	.	.	PUNCT
fcis-10735	109	1	the	the	DET
fcis-10735	109	2	algorithm	algorithm	NOUN
fcis-10735	109	3	also	also	ADV
fcis-10735	109	4	calls	call	VERB
fcis-10735	109	5	pandas	panda	NOUN
fcis-10735	109	6	for	for	ADP
fcis-10735	109	7	reading	read	VERB
fcis-10735	109	8	files	file	NOUN
fcis-10735	109	9	and	and	CCONJ
fcis-10735	109	10	cv2	cv2	PROPN
fcis-10735	109	11	for	for	ADP
fcis-10735	109	12	image	image	NOUN
fcis-10735	109	13	processing	processing	NOUN
fcis-10735	109	14	.	.	PUNCT
fcis-10735	110	1	some	some	DET
fcis-10735	110	2	basic	basic	ADJ
fcis-10735	110	3	data	datum	NOUN
fcis-10735	110	4	processing	processing	NOUN
fcis-10735	110	5	packages	package	NOUN
fcis-10735	110	6	are	be	AUX
fcis-10735	110	7	also	also	ADV
fcis-10735	110	8	called	call	VERB
fcis-10735	110	9	.	.	PUNCT
fcis-10735	111	1	2.1.4.2	2.1.4.2	NUM
fcis-10735	111	2	modeling	model	VERB
fcis-10735	111	3	2.1.4.2.1	2.1.4.2.1	PRON
fcis-10735	111	4	random	random	ADJ
fcis-10735	111	5	forest	forest	NOUN
fcis-10735	111	6	model	model	NOUN
fcis-10735	111	7	parameter	parameter	NOUN
fcis-10735	111	8	optimization	optimization	NOUN
fcis-10735	111	9	in	in	ADP
fcis-10735	111	10	the	the	DET
fcis-10735	111	11	whole	whole	ADJ
fcis-10735	111	12	modeling	modeling	NOUN
fcis-10735	111	13	process	process	NOUN
fcis-10735	111	14	,	,	PUNCT
fcis-10735	111	15	the	the	DET
fcis-10735	111	16	model	model	NOUN
fcis-10735	111	17	parameters	parameter	NOUN
fcis-10735	111	18	are	be	AUX
fcis-10735	111	19	tested	test	VERB
fcis-10735	111	20	to	to	PART
fcis-10735	111	21	determine	determine	VERB
fcis-10735	111	22	the	the	DET
fcis-10735	111	23	optimal	optimal	ADJ
fcis-10735	111	24	parameters	parameter	NOUN
fcis-10735	111	25	.	.	PUNCT
fcis-10735	112	1	n_estimators	n_estimator	NOUN
fcis-10735	112	2	:	:	PUNCT
fcis-10735	112	3	the	the	DET
fcis-10735	112	4	number	number	NOUN
fcis-10735	112	5	of	of	ADP
fcis-10735	112	6	decision	decision	NOUN
fcis-10735	112	7	tree	tree	NOUN
fcis-10735	112	8	models	model	NOUN
fcis-10735	112	9	included	include	VERB
fcis-10735	112	10	in	in	ADP
fcis-10735	112	11	the	the	DET
fcis-10735	112	12	random	random	ADJ
fcis-10735	112	13	forest	forest	NOUN
fcis-10735	112	14	model	model	NOUN
fcis-10735	112	15	.	.	PUNCT
fcis-10735	113	1	fig	fig	PROPN
fcis-10735	113	2	13	13	NUM
fcis-10735	113	3	.	.	NOUN
fcis-10735	113	4	0	0	NUM
fcis-10735	113	5	-	-	SYM
fcis-10735	113	6	200	200	NUM
fcis-10735	113	7	estimators	estimator	NOUN
fcis-10735	113	8	’	'	PUNCT
fcis-10735	113	9	result	result	VERB
fcis-10735	113	10	fig	fig	NOUN
fcis-10735	113	11	14	14	NUM
fcis-10735	113	12	.	.	PUNCT
fcis-10735	114	1	146	146	NUM
fcis-10735	114	2	-	-	SYM
fcis-10735	114	3	156	156	NUM
fcis-10735	114	4	estimators	estimator	NOUN
fcis-10735	114	5	’	'	PUNCT
fcis-10735	114	6	result	result	VERB
fcis-10735	114	7	max_depth	max_depth	NOUN
fcis-10735	114	8	:	:	PUNCT
fcis-10735	114	9	the	the	DET
fcis-10735	114	10	maximum	maximum	ADJ
fcis-10735	114	11	depth	depth	NOUN
fcis-10735	114	12	of	of	ADP
fcis-10735	114	13	the	the	DET
fcis-10735	114	14	decision	decision	NOUN
fcis-10735	114	15	tree	tree	NOUN
fcis-10735	114	16	model	model	NOUN
fcis-10735	114	17	.	.	PUNCT
fcis-10735	115	1	max_features	max_feature	NOUN
fcis-10735	115	2	:	:	PUNCT
fcis-10735	115	3	the	the	DET
fcis-10735	115	4	maximum	maximum	ADJ
fcis-10735	115	5	number	number	NOUN
fcis-10735	115	6	of	of	ADP
fcis-10735	115	7	features	feature	NOUN
fcis-10735	115	8	to	to	PART
fcis-10735	115	9	select	select	VERB
fcis-10735	115	10	when	when	SCONJ
fcis-10735	115	11	building	build	VERB
fcis-10735	115	12	a	a	DET
fcis-10735	115	13	decision	decision	NOUN
fcis-10735	115	14	tree	tree	NOUN
fcis-10735	115	15	.	.	PUNCT
fcis-10735	116	1	min_samples_leaf	min_samples_leaf	ADV
fcis-10735	116	2	:	:	PUNCT
fcis-10735	116	3	the	the	DET
fcis-10735	116	4	minimum	minimum	ADJ
fcis-10735	116	5	number	number	NOUN
fcis-10735	116	6	of	of	ADP
fcis-10735	116	7	samples	sample	NOUN
fcis-10735	116	8	for	for	ADP
fcis-10735	116	9	leaf	leaf	NOUN
fcis-10735	116	10	nodes	node	NOUN
fcis-10735	116	11	.	.	PUNCT
fcis-10735	117	1	min_samples_split	min_samples_split	ADJ
fcis-10735	117	2	:	:	PUNCT
fcis-10735	117	3	the	the	DET
fcis-10735	117	4	minimum	minimum	ADJ
fcis-10735	117	5	number	number	NOUN
fcis-10735	117	6	of	of	ADP
fcis-10735	117	7	samples	sample	NOUN
fcis-10735	117	8	allowed	allow	VERB
fcis-10735	117	9	to	to	PART
fcis-10735	117	10	split	split	VERB
fcis-10735	117	11	the	the	DET
fcis-10735	117	12	current	current	ADJ
fcis-10735	117	13	node	node	NOUN
fcis-10735	117	14	.	.	PUNCT
fcis-10735	118	1	criterion	criterion	NOUN
fcis-10735	118	2	:	:	PUNCT
fcis-10735	118	3	node	node	ADJ
fcis-10735	118	4	splitting	splitting	NOUN
fcis-10735	118	5	basis	basis	NOUN
fcis-10735	118	6	.	.	PUNCT
fcis-10735	119	1	2.1.4.2.2	2.1.4.2.2	NUM
fcis-10735	119	2	the	the	DET
fcis-10735	119	3	final	final	ADJ
fcis-10735	119	4	model	model	NOUN
fcis-10735	119	5	after	after	ADP
fcis-10735	119	6	parameter	parameter	NOUN
fcis-10735	119	7	optimization	optimization	NOUN
fcis-10735	119	8	the	the	DET
fcis-10735	119	9	final	final	ADJ
fcis-10735	119	10	model	model	NOUN
fcis-10735	119	11	is	be	AUX
fcis-10735	119	12	modified	modify	VERB
fcis-10735	119	13	according	accord	VERB
fcis-10735	119	14	to	to	ADP
fcis-10735	119	15	the	the	DET
fcis-10735	119	16	parameter	parameter	NOUN
fcis-10735	119	17	tuning	tune	VERB
fcis-10735	119	18	test	test	NOUN
fcis-10735	119	19	and	and	CCONJ
fcis-10735	119	20	the	the	DET
fcis-10735	119	21	final	final	ADJ
fcis-10735	119	22	model	model	NOUN
fcis-10735	119	23	parameters	parameter	NOUN
fcis-10735	119	24	are	be	AUX
fcis-10735	119	25	obtained	obtain	VERB
fcis-10735	119	26	.	.	PUNCT
fcis-10735	120	1	2.1.4.3	2.1.4.3	NUM
fcis-10735	120	2	training	training	NOUN
fcis-10735	120	3	and	and	CCONJ
fcis-10735	120	4	results	result	VERB
fcis-10735	120	5	the	the	DET
fcis-10735	120	6	training	training	NOUN
fcis-10735	120	7	model	model	NOUN
fcis-10735	120	8	is	be	AUX
fcis-10735	120	9	performed	perform	VERB
fcis-10735	120	10	using	use	VERB
fcis-10735	120	11	the	the	DET
fcis-10735	120	12	input	input	NOUN
fcis-10735	121	1	x	x	NOUN
fcis-10735	121	2	values	value	NOUN
fcis-10735	121	3	(	(	PUNCT
fcis-10735	121	4	font	font	NOUN
fcis-10735	121	5	images	image	NOUN
fcis-10735	121	6	)	)	PUNCT
fcis-10735	121	7	and	and	CCONJ
fcis-10735	121	8	y	y	PROPN
fcis-10735	121	9	values	value	NOUN
fcis-10735	121	10	(	(	PUNCT
fcis-10735	121	11	labels	label	NOUN
fcis-10735	121	12	)	)	PUNCT
fcis-10735	121	13	.	.	PUNCT
fcis-10735	122	1	the	the	DET
fcis-10735	122	2	whole	whole	ADJ
fcis-10735	122	3	training	training	NOUN
fcis-10735	122	4	process	process	NOUN
fcis-10735	122	5	takes	take	VERB
fcis-10735	122	6	a	a	DET
fcis-10735	122	7	short	short	ADJ
fcis-10735	122	8	time	time	NOUN
fcis-10735	122	9	and	and	CCONJ
fcis-10735	122	10	generates	generate	VERB
fcis-10735	122	11	the	the	DET
fcis-10735	122	12	training	training	NOUN
fcis-10735	122	13	result	result	NOUN
fcis-10735	122	14	model	model	NOUN
fcis-10735	122	15	.	.	PUNCT
fcis-10735	123	1	fig	fig	NOUN
fcis-10735	123	2	15	15	NUM
fcis-10735	123	3	.	.	PUNCT
fcis-10735	124	1	random	random	ADJ
fcis-10735	124	2	forest	forest	NOUN
fcis-10735	124	3	classifier	classifier	NOUN
fcis-10735	124	4	the	the	DET
fcis-10735	124	5	font	font	NOUN
fcis-10735	124	6	recognition	recognition	NOUN
fcis-10735	124	7	results	result	NOUN
fcis-10735	124	8	of	of	ADP
fcis-10735	124	9	the	the	DET
fcis-10735	124	10	random	random	ADJ
fcis-10735	124	11	forest	forest	NOUN
fcis-10735	124	12	algorithm	algorithm	NOUN
fcis-10735	124	13	for	for	ADP
fcis-10735	124	14	the	the	DET
fcis-10735	124	15	test	test	NOUN
fcis-10735	124	16	image	image	NOUN
fcis-10735	124	17	.	.	PUNCT
fcis-10735	125	1	it	it	PRON
fcis-10735	125	2	runs	run	VERB
fcis-10735	125	3	fast	fast	ADV
fcis-10735	125	4	and	and	CCONJ
fcis-10735	125	5	gives	give	VERB
fcis-10735	125	6	no	no	DET
fcis-10735	125	7	error	error	NOUN
fcis-10735	125	8	messages	message	NOUN
fcis-10735	125	9	.	.	PUNCT
fcis-10735	126	1	2.2	2.2	NUM
fcis-10735	126	2	.	.	PUNCT
fcis-10735	127	1	logistic	logistic	ADJ
fcis-10735	127	2	regression	regression	NOUN
fcis-10735	127	3	2.2.1	2.2.1	NUM
fcis-10735	127	4	.	.	PUNCT
fcis-10735	128	1	workflow	workflow	NOUN
fcis-10735	128	2	fig	fig	NOUN
fcis-10735	128	3	16	16	NUM
fcis-10735	128	4	.	.	PUNCT
fcis-10735	129	1	workflow	workflow	NOUN
fcis-10735	129	2	of	of	ADP
fcis-10735	129	3	logistic	logistic	ADJ
fcis-10735	129	4	regression	regression	NOUN
fcis-10735	129	5	2.2.2	2.2.2	NUM
fcis-10735	129	6	.	.	PUNCT
fcis-10735	129	7	hardware	hardware	NOUN
fcis-10735	129	8	usage	usage	NOUN
fcis-10735	129	9	the	the	DET
fcis-10735	129	10	logistic	logistic	ADJ
fcis-10735	129	11	regression	regression	NOUN
fcis-10735	129	12	algorithm	algorithm	NOUN
fcis-10735	129	13	uses	use	VERB
fcis-10735	129	14	cpu	cpu	NOUN
fcis-10735	129	15	to	to	PART
fcis-10735	129	16	perform	perform	VERB
fcis-10735	129	17	operations	operation	NOUN
fcis-10735	129	18	.	.	PUNCT
fcis-10735	130	1	there	there	PRON
fcis-10735	130	2	is	be	VERB
fcis-10735	130	3	no	no	DET
fcis-10735	130	4	error	error	NOUN
fcis-10735	130	5	during	during	ADP
fcis-10735	130	6	the	the	DET
fcis-10735	130	7	whole	whole	ADJ
fcis-10735	130	8	running	running	NOUN
fcis-10735	130	9	process	process	NOUN
fcis-10735	130	10	and	and	CCONJ
fcis-10735	130	11	the	the	DET
fcis-10735	130	12	operation	operation	NOUN
fcis-10735	130	13	speed	speed	NOUN
fcis-10735	130	14	is	be	AUX
fcis-10735	130	15	fast	fast	ADJ
fcis-10735	130	16	.	.	PUNCT
fcis-10735	131	1	hardware	hardware	NOUN
fcis-10735	131	2	list	list	NOUN
fcis-10735	131	3	:	:	PUNCT
fcis-10735	131	4	device	device	NOUN
fcis-10735	131	5	name	name	NOUN
fcis-10735	131	6	:	:	PUNCT
fcis-10735	131	7	laptop-9lij8tiu	laptop-9lij8tiu	NOUN
fcis-10735	131	8	processor	processor	NOUN
fcis-10735	131	9	:	:	PUNCT
fcis-10735	131	10	11th	11th	ADJ
fcis-10735	131	11	gen	gen	PROPN
fcis-10735	131	12	intel(r	intel(r	PROPN
fcis-10735	131	13	)	)	PUNCT
fcis-10735	131	14	core	core	NOUN
fcis-10735	131	15	(	(	PUNCT
fcis-10735	131	16	tm	tm	NOUN
fcis-10735	131	17	)	)	PUNCT
fcis-10735	131	18	i9	i9	NOUN
fcis-10735	131	19	-	-	PUNCT
fcis-10735	131	20	11900h	11900h	NUM
fcis-10735	131	21	@	@	ADP
fcis-10735	131	22	2.50ghz	2.50ghz	NUM
fcis-10735	131	23	2.50	2.50	NUM
fcis-10735	131	24	ghz	ghz	NOUN
fcis-10735	131	25	on	on	ADP
fcis-10735	131	26	-	-	PUNCT
fcis-10735	131	27	board	board	NOUN
fcis-10735	131	28	ram	ram	NOUN
fcis-10735	131	29	:	:	PUNCT
fcis-10735	131	30	16.0	16.0	NUM
fcis-10735	131	31	gb	gb	NOUN
fcis-10735	131	32	(	(	PUNCT
fcis-10735	131	33	15.7	15.7	NUM
fcis-10735	131	34	gb	gb	ADV
fcis-10735	131	35	available	available	ADJ
fcis-10735	131	36	)	)	PUNCT
fcis-10735	131	37	system	system	NOUN
fcis-10735	131	38	type	type	NOUN
fcis-10735	131	39	:	:	PUNCT
fcis-10735	131	40	64	64	NUM
fcis-10735	131	41	-	-	PUNCT
fcis-10735	131	42	bit	bit	NOUN
fcis-10735	131	43	operating	operate	VERB
fcis-10735	131	44	system	system	NOUN
fcis-10735	131	45	,	,	PUNCT
fcis-10735	131	46	x64	x64	NOUN
fcis-10735	131	47	-	-	PUNCT
fcis-10735	131	48	based	base	VERB
fcis-10735	131	49	processor	processor	NOUN
fcis-10735	131	50	2.2.3	2.2.3	NUM
fcis-10735	131	51	.	.	PUNCT
fcis-10735	132	1	theory	theory	NOUN
fcis-10735	132	2	the	the	DET
fcis-10735	132	3	principle	principle	NOUN
fcis-10735	132	4	of	of	ADP
fcis-10735	132	5	logistic	logistic	ADJ
fcis-10735	132	6	regression	regression	NOUN
fcis-10735	132	7	is	be	AUX
fcis-10735	132	8	to	to	PART
fcis-10735	132	9	use	use	VERB
fcis-10735	132	10	the	the	DET
fcis-10735	132	11	logistic	logistic	ADJ
fcis-10735	132	12	function	function	NOUN
fcis-10735	132	13	to	to	PART
fcis-10735	132	14	map	map	VERB
fcis-10735	132	15	the	the	DET
fcis-10735	132	16	results	result	NOUN
fcis-10735	132	17	of	of	ADP
fcis-10735	132	18	linear	linear	ADJ
fcis-10735	132	19	regression	regression	NOUN
fcis-10735	132	20	(	(	PUNCT
fcis-10735	132	21	-∞	-∞	PROPN
fcis-10735	132	22	,	,	PUNCT
fcis-10735	132	23	∞	∞	PROPN
fcis-10735	132	24	)	)	PUNCT
fcis-10735	132	25	to	to	ADP
fcis-10735	132	26	(	(	PUNCT
fcis-10735	132	27	0,1	0,1	NUM
fcis-10735	132	28	)	)	PUNCT
fcis-10735	132	29	.	.	PUNCT
fcis-10735	133	1	logistic	logistic	ADJ
fcis-10735	133	2	regression	regression	NOUN
fcis-10735	133	3	assumes	assume	VERB
fcis-10735	133	4	that	that	SCONJ
fcis-10735	133	5	the	the	DET
fcis-10735	133	6	data	datum	NOUN
fcis-10735	133	7	follow	follow	VERB
fcis-10735	133	8	this	this	DET
fcis-10735	133	9	distribution	distribution	NOUN
fcis-10735	133	10	,	,	PUNCT
fcis-10735	133	11	and	and	CCONJ
fcis-10735	133	12	then	then	ADV
fcis-10735	133	13	uses	use	VERB
fcis-10735	133	14	maximum	maximum	ADJ
fcis-10735	133	15	likelihood	likelihood	NOUN
fcis-10735	133	16	estimation	estimation	NOUN
fcis-10735	133	17	to	to	PART
fcis-10735	133	18	estimate	estimate	VERB
fcis-10735	133	19	parameters	parameter	NOUN
fcis-10735	133	20	.	.	PUNCT
fcis-10735	134	1	2.2.3.1	2.2.3.1	NUM
fcis-10735	134	2	logistic	logistic	ADJ
fcis-10735	134	3	distribution	distribution	NOUN
fcis-10735	134	4	logistic	logistic	ADJ
fcis-10735	134	5	distribution	distribution	NOUN
fcis-10735	134	6	is	be	AUX
fcis-10735	134	7	a	a	DET
fcis-10735	134	8	continuous	continuous	ADJ
fcis-10735	134	9	probability	probability	NOUN
fcis-10735	134	10	distribution	distribution	NOUN
fcis-10735	134	11	,	,	PUNCT
fcis-10735	134	12	and	and	CCONJ
fcis-10735	134	13	its	its	PRON
fcis-10735	134	14	distribution	distribution	NOUN
fcis-10735	134	15	function	function	NOUN
fcis-10735	134	16	and	and	CCONJ
fcis-10735	134	17	density	density	NOUN
fcis-10735	134	18	function	function	NOUN
fcis-10735	134	19	are	be	AUX
fcis-10735	134	20	:	:	PUNCT
fcis-10735	134	21	among	among	ADP
fcis-10735	134	22	them	they	PRON
fcis-10735	134	23	,	,	PUNCT
fcis-10735	134	24	μ	μ	PROPN
fcis-10735	134	25	represents	represent	VERB
fcis-10735	134	26	the	the	DET
fcis-10735	134	27	position	position	NOUN
fcis-10735	134	28	parameter	parameter	NOUN
fcis-10735	134	29	,	,	PUNCT
fcis-10735	134	30	and	and	CCONJ
fcis-10735	134	31	y>0	y>0	NOUN
fcis-10735	134	32	is	be	AUX
fcis-10735	134	33	the	the	DET
fcis-10735	134	34	shape	shape	NOUN
fcis-10735	134	35	parameter	parameter	NOUN
fcis-10735	134	36	[	[	X
fcis-10735	134	37	4	4	NUM
fcis-10735	134	38	]	]	PUNCT
fcis-10735	134	39	.	.	PUNCT
fcis-10735	135	1	2.2.3.2	2.2.3.2	NUM
fcis-10735	135	2	logistic	logistic	ADJ
fcis-10735	135	3	regression	regression	NOUN
fcis-10735	135	4	mathematical	mathematical	ADJ
fcis-10735	135	5	expression	expression	NOUN
fcis-10735	135	6	of	of	ADP
fcis-10735	135	7	linear	linear	PROPN
fcis-10735	135	8	regression	regression	NOUN
fcis-10735	135	9	function	function	NOUN
fcis-10735	135	10	:	:	PUNCT
fcis-10735	135	11	11	11	NUM
fcis-10735	135	12	where	where	SCONJ
fcis-10735	135	13	xi	xi	PROPN
fcis-10735	135	14	is	be	AUX
fcis-10735	135	15	the	the	DET
fcis-10735	135	16	independent	independent	ADJ
fcis-10735	135	17	variable	variable	NOUN
fcis-10735	135	18	,	,	PUNCT
fcis-10735	135	19	y	y	PROPN
fcis-10735	135	20	is	be	AUX
fcis-10735	135	21	the	the	DET
fcis-10735	135	22	dependent	dependent	ADJ
fcis-10735	135	23	variable	variable	NOUN
fcis-10735	135	24	,	,	PUNCT
fcis-10735	135	25	the	the	DET
fcis-10735	135	26	range	range	NOUN
fcis-10735	135	27	of	of	ADP
fcis-10735	135	28	y	y	PROPN
fcis-10735	135	29	is	be	AUX
fcis-10735	135	30	(	(	PUNCT
fcis-10735	135	31	-∞	-∞	INTJ
fcis-10735	135	32	,	,	PUNCT
fcis-10735	135	33	∞	∞	PROPN
fcis-10735	135	34	)	)	PUNCT
fcis-10735	135	35	,	,	PUNCT
fcis-10735	135	36	θ0	θ0	PROPN
fcis-10735	135	37	is	be	AUX
fcis-10735	135	38	a	a	DET
fcis-10735	135	39	constant	constant	ADJ
fcis-10735	135	40	term	term	NOUN
fcis-10735	135	41	,	,	PUNCT
fcis-10735	135	42	θi(i=1,2,	θi(i=1,2,	NOUN
fcis-10735	135	43	...	...	PUNCT
fcis-10735	135	44	,n	,n	PUNCT
fcis-10735	135	45	)	)	PUNCT
fcis-10735	135	46	is	be	AUX
fcis-10735	135	47	the	the	DET
fcis-10735	135	48	coefficient	coefficient	NOUN
fcis-10735	135	49	to	to	PART
fcis-10735	135	50	be	be	AUX
fcis-10735	135	51	calculated	calculate	VERB
fcis-10735	135	52	,	,	PUNCT
fcis-10735	135	53	different	different	ADJ
fcis-10735	135	54	weights	weight	NOUN
fcis-10735	135	55	θi	θi	PROPN
fcis-10735	135	56	reflects	reflect	VERB
fcis-10735	135	57	the	the	DET
fcis-10735	135	58	different	different	ADJ
fcis-10735	135	59	contribution	contribution	NOUN
fcis-10735	135	60	of	of	ADP
fcis-10735	135	61	the	the	DET
fcis-10735	135	62	independent	independent	ADJ
fcis-10735	135	63	variable	variable	NOUN
fcis-10735	135	64	to	to	ADP
fcis-10735	135	65	the	the	DET
fcis-10735	135	66	dependent	dependent	ADJ
fcis-10735	135	67	variable	variable	NOUN
fcis-10735	135	68	[	[	X
fcis-10735	135	69	4	4	NUM
fcis-10735	135	70	]	]	PUNCT
fcis-10735	135	71	.	.	PUNCT
fcis-10735	136	1	2.2.3.3	2.2.3.3	NUM
fcis-10735	136	2	cost	cost	NOUN
fcis-10735	136	3	function	function	NOUN
fcis-10735	136	4	likelihood	likelihood	NOUN
fcis-10735	136	5	function	function	NOUN
fcis-10735	136	6	:	:	PUNCT
fcis-10735	136	7	in	in	ADP
fcis-10735	136	8	machine	machine	NOUN
fcis-10735	136	9	learning	learn	VERB
fcis-10735	136	10	we	we	PRON
fcis-10735	136	11	have	have	VERB
fcis-10735	136	12	the	the	DET
fcis-10735	136	13	concept	concept	NOUN
fcis-10735	136	14	of	of	ADP
fcis-10735	136	15	a	a	DET
fcis-10735	136	16	loss	loss	NOUN
fcis-10735	136	17	function	function	NOUN
fcis-10735	136	18	,	,	PUNCT
fcis-10735	136	19	which	which	PRON
fcis-10735	136	20	measures	measure	VERB
fcis-10735	136	21	how	how	SCONJ
fcis-10735	136	22	wrong	wrong	ADJ
fcis-10735	136	23	a	a	DET
fcis-10735	136	24	model	model	NOUN
fcis-10735	136	25	's	's	PART
fcis-10735	136	26	predictions	prediction	NOUN
fcis-10735	136	27	are	be	AUX
fcis-10735	136	28	.	.	PUNCT
fcis-10735	137	1	if	if	SCONJ
fcis-10735	137	2	we	we	PRON
fcis-10735	137	3	take	take	VERB
fcis-10735	137	4	the	the	DET
fcis-10735	137	5	average	average	ADJ
fcis-10735	137	6	log	log	NOUN
fcis-10735	137	7	-	-	PUNCT
fcis-10735	137	8	likelihood	likelihood	NOUN
fcis-10735	137	9	loss	loss	NOUN
fcis-10735	137	10	over	over	ADP
fcis-10735	137	11	the	the	DET
fcis-10735	137	12	entire	entire	ADJ
fcis-10735	137	13	dataset	dataset	NOUN
fcis-10735	137	14	,	,	PUNCT
fcis-10735	137	15	we	we	PRON
fcis-10735	137	16	can	can	AUX
fcis-10735	137	17	get	get	VERB
fcis-10735	137	18	:	:	PUNCT
fcis-10735	137	19	2.2.3.4	2.2.3.4	NUM
fcis-10735	137	20	gradient	gradient	ADJ
fcis-10735	137	21	descent	descent	NOUN
fcis-10735	137	22	and	and	CCONJ
fcis-10735	137	23	newton	newton	PROPN
fcis-10735	137	24	's	's	PART
fcis-10735	137	25	method	method	NOUN
fcis-10735	137	26	stochastic	stochastic	ADJ
fcis-10735	137	27	gradient	gradient	ADJ
fcis-10735	137	28	descent	descent	NOUN
fcis-10735	137	29	gradient	gradient	NOUN
fcis-10735	137	30	descent	descent	NOUN
fcis-10735	137	31	is	be	AUX
fcis-10735	137	32	to	to	PART
fcis-10735	137	33	find	find	VERB
fcis-10735	137	34	the	the	DET
fcis-10735	137	35	descending	descend	VERB
fcis-10735	137	36	direction	direction	NOUN
fcis-10735	137	37	through	through	ADP
fcis-10735	137	38	the	the	DET
fcis-10735	137	39	first	first	ADJ
fcis-10735	137	40	derivative	derivative	NOUN
fcis-10735	137	41	of	of	ADP
fcis-10735	137	42	j(w	j(w	PROPN
fcis-10735	137	43	)	)	PUNCT
fcis-10735	137	44	to	to	PART
fcis-10735	137	45	w	w	NOUN
fcis-10735	137	46	and	and	CCONJ
fcis-10735	137	47	update	update	VERB
fcis-10735	137	48	the	the	DET
fcis-10735	137	49	parameters	parameter	NOUN
fcis-10735	137	50	in	in	ADP
fcis-10735	137	51	an	an	DET
fcis-10735	137	52	iterative	iterative	NOUN
fcis-10735	137	53	manner	manner	NOUN
fcis-10735	137	54	.	.	PUNCT
fcis-10735	138	1	the	the	DET
fcis-10735	138	2	update	update	NOUN
fcis-10735	138	3	method	method	NOUN
fcis-10735	138	4	is	be	AUX
fcis-10735	138	5	:	:	PUNCT
fcis-10735	138	6	where	where	SCONJ
fcis-10735	138	7	k	k	PROPN
fcis-10735	138	8	is	be	AUX
fcis-10735	138	9	the	the	DET
fcis-10735	138	10	number	number	NOUN
fcis-10735	138	11	of	of	ADP
fcis-10735	138	12	iterations	iteration	NOUN
fcis-10735	138	13	.	.	PUNCT
fcis-10735	139	1	after	after	ADP
fcis-10735	139	2	each	each	DET
fcis-10735	139	3	parameter	parameter	NOUN
fcis-10735	139	4	update	update	NOUN
fcis-10735	139	5	,	,	PUNCT
fcis-10735	139	6	you	you	PRON
fcis-10735	139	7	can	can	AUX
fcis-10735	139	8	compare	compare	VERB
fcis-10735	139	9	the	the	PRON
fcis-10735	139	10	to	to	PART
fcis-10735	139	11	stop	stop	VERB
fcis-10735	139	12	the	the	DET
fcis-10735	139	13	iteration	iteration	NOUN
fcis-10735	139	14	when	when	SCONJ
fcis-10735	139	15	it	it	PRON
fcis-10735	139	16	is	be	AUX
fcis-10735	139	17	less	less	ADJ
fcis-10735	139	18	than	than	ADP
fcis-10735	139	19	the	the	DET
fcis-10735	139	20	threshold	threshold	NOUN
fcis-10735	139	21	or	or	CCONJ
fcis-10735	139	22	when	when	SCONJ
fcis-10735	139	23	the	the	DET
fcis-10735	139	24	maximum	maximum	ADJ
fcis-10735	139	25	number	number	NOUN
fcis-10735	139	26	of	of	ADP
fcis-10735	139	27	iterations	iteration	NOUN
fcis-10735	139	28	is	be	AUX
fcis-10735	139	29	reached	reach	VERB
fcis-10735	139	30	.	.	PUNCT
fcis-10735	140	1	the	the	DET
fcis-10735	140	2	idea	idea	NOUN
fcis-10735	140	3	of	of	ADP
fcis-10735	140	4	newton	newton	PROPN
fcis-10735	140	5	's	's	PART
fcis-10735	140	6	method	method	NOUN
fcis-10735	140	7	is	be	AUX
fcis-10735	140	8	to	to	PART
fcis-10735	140	9	perform	perform	VERB
fcis-10735	140	10	a	a	DET
fcis-10735	140	11	second	second	ADJ
fcis-10735	140	12	-	-	PUNCT
fcis-10735	140	13	order	order	NOUN
fcis-10735	140	14	taylor	taylor	NOUN
fcis-10735	140	15	expansion	expansion	NOUN
fcis-10735	140	16	of	of	ADP
fcis-10735	140	17	f(x	f(x	PROPN
fcis-10735	140	18	)	)	PUNCT
fcis-10735	140	19	in	in	ADP
fcis-10735	140	20	the	the	DET
fcis-10735	140	21	vicinity	vicinity	NOUN
fcis-10735	140	22	of	of	ADP
fcis-10735	140	23	the	the	DET
fcis-10735	140	24	existing	exist	VERB
fcis-10735	140	25	minimum	minimum	NOUN
fcis-10735	140	26	point	point	NOUN
fcis-10735	140	27	estimate	estimate	NOUN
fcis-10735	140	28	,	,	PUNCT
fcis-10735	140	29	and	and	CCONJ
fcis-10735	140	30	then	then	ADV
fcis-10735	140	31	find	find	VERB
fcis-10735	140	32	the	the	DET
fcis-10735	140	33	next	next	ADJ
fcis-10735	140	34	estimate	estimate	NOUN
fcis-10735	140	35	of	of	ADP
fcis-10735	140	36	the	the	DET
fcis-10735	140	37	minimum	minimum	ADJ
fcis-10735	140	38	point	point	NOUN
fcis-10735	140	39	.	.	PUNCT
fcis-10735	141	1	assuming	assume	VERB
fcis-10735	141	2	w^k	w^k	PROPN
fcis-10735	141	3	is	be	AUX
fcis-10735	141	4	the	the	DET
fcis-10735	141	5	current	current	ADJ
fcis-10735	141	6	minimum	minimum	ADJ
fcis-10735	141	7	estimate	estimate	NOUN
fcis-10735	141	8	,	,	PUNCT
fcis-10735	141	9	then	then	ADV
fcis-10735	141	10	there	there	PRON
fcis-10735	141	11	are	be	VERB
fcis-10735	141	12	:	:	PUNCT
fcis-10735	141	13	2.2.3.5	2.2.3.5	NUM
fcis-10735	141	14	regularization	regularization	NOUN
fcis-10735	141	15	objective	objective	ADJ
fcis-10735	141	16	function	function	NOUN
fcis-10735	141	17	:	:	PUNCT
fcis-10735	141	18	2.2.4	2.2.4	NUM
fcis-10735	141	19	.	.	PUNCT
fcis-10735	142	1	implementation	implementation	NOUN
fcis-10735	142	2	2.2.4.1	2.2.4.1	NUM
fcis-10735	142	3	import	import	NOUN
fcis-10735	142	4	all	all	DET
fcis-10735	142	5	required	require	VERB
fcis-10735	142	6	libraries	library	NOUN
fcis-10735	142	7	the	the	DET
fcis-10735	142	8	logistic	logistic	ADJ
fcis-10735	142	9	regression	regression	NOUN
fcis-10735	142	10	algorithm	algorithm	NOUN
fcis-10735	142	11	calls	call	VERB
fcis-10735	142	12	logistic	logistic	ADJ
fcis-10735	142	13	regression	regression	NOUN
fcis-10735	142	14	in	in	ADP
fcis-10735	142	15	sklearn	sklearn	NOUN
fcis-10735	142	16	to	to	PART
fcis-10735	142	17	build	build	VERB
fcis-10735	142	18	the	the	DET
fcis-10735	142	19	model	model	NOUN
fcis-10735	142	20	2.2.4.2	2.2.4.2	NUM
fcis-10735	142	21	modeling	model	VERB
fcis-10735	142	22	2.2.4.2.1	2.2.4.2.1	NUM
fcis-10735	142	23	parameter	parameter	NOUN
fcis-10735	142	24	tuning	tune	VERB
fcis-10735	142	25	fig	fig	NOUN
fcis-10735	142	26	17	17	NUM
fcis-10735	142	27	.	.	PUNCT
fcis-10735	143	1	result	result	PROPN
fcis-10735	143	2	chart	chart	NOUN
fcis-10735	143	3	i	i	PRON
fcis-10735	143	4	use	use	VERB
fcis-10735	143	5	gridsearchcv	gridsearchcv	NOUN
fcis-10735	143	6	in	in	ADP
fcis-10735	143	7	sklearn	sklearn	PROPN
fcis-10735	143	8	to	to	PART
fcis-10735	143	9	optimize	optimize	VERB
fcis-10735	143	10	the	the	DET
fcis-10735	143	11	parameters	parameter	NOUN
fcis-10735	143	12	of	of	ADP
fcis-10735	143	13	the	the	DET
fcis-10735	143	14	model	model	NOUN
fcis-10735	143	15	.	.	PUNCT
fcis-10735	144	1	the	the	DET
fcis-10735	144	2	function	function	NOUN
fcis-10735	144	3	of	of	ADP
fcis-10735	144	4	gridsearchcv	gridsearchcv	NOUN
fcis-10735	144	5	is	be	AUX
fcis-10735	144	6	to	to	PART
fcis-10735	144	7	automatically	automatically	ADV
fcis-10735	144	8	adjust	adjust	VERB
fcis-10735	144	9	parameters	parameter	NOUN
fcis-10735	144	10	.	.	PUNCT
fcis-10735	145	1	the	the	DET
fcis-10735	145	2	user	user	NOUN
fcis-10735	145	3	only	only	ADV
fcis-10735	145	4	needs	need	VERB
fcis-10735	145	5	to	to	PART
fcis-10735	145	6	input	input	VERB
fcis-10735	145	7	the	the	DET
fcis-10735	145	8	parameters	parameter	NOUN
fcis-10735	145	9	,	,	PUNCT
fcis-10735	145	10	and	and	CCONJ
fcis-10735	145	11	the	the	DET
fcis-10735	145	12	optimized	optimize	VERB
fcis-10735	145	13	results	result	NOUN
fcis-10735	145	14	and	and	CCONJ
fcis-10735	145	15	parameters	parameter	NOUN
fcis-10735	145	16	can	can	AUX
fcis-10735	145	17	be	be	AUX
fcis-10735	145	18	given	give	VERB
fcis-10735	145	19	.	.	PUNCT
fcis-10735	146	1	2.2.4.2.2	2.2.4.2.2	NUM
fcis-10735	146	2	final	final	ADJ
fcis-10735	146	3	model	model	NOUN
fcis-10735	146	4	i	i	PRON
fcis-10735	146	5	build	build	VERB
fcis-10735	146	6	a	a	DET
fcis-10735	146	7	model	model	NOUN
fcis-10735	146	8	based	base	VERB
fcis-10735	146	9	on	on	ADP
fcis-10735	146	10	the	the	DET
fcis-10735	146	11	results	result	NOUN
fcis-10735	146	12	of	of	ADP
fcis-10735	146	13	parameter	parameter	NOUN
fcis-10735	146	14	optimization	optimization	NOUN
fcis-10735	146	15	.	.	PUNCT
fcis-10735	147	1	2.2.4.3	2.2.4.3	NUM
fcis-10735	147	2	training	training	NOUN
fcis-10735	147	3	and	and	CCONJ
fcis-10735	147	4	results	result	VERB
fcis-10735	147	5	the	the	DET
fcis-10735	147	6	training	training	NOUN
fcis-10735	147	7	model	model	NOUN
fcis-10735	147	8	is	be	AUX
fcis-10735	147	9	performed	perform	VERB
fcis-10735	147	10	using	use	VERB
fcis-10735	147	11	the	the	DET
fcis-10735	147	12	input	input	NOUN
fcis-10735	148	1	x	x	NOUN
fcis-10735	148	2	values	value	NOUN
fcis-10735	148	3	(	(	PUNCT
fcis-10735	148	4	font	font	NOUN
fcis-10735	148	5	images	image	NOUN
fcis-10735	148	6	)	)	PUNCT
fcis-10735	148	7	and	and	CCONJ
fcis-10735	148	8	y	y	PROPN
fcis-10735	148	9	values	value	NOUN
fcis-10735	148	10	(	(	PUNCT
fcis-10735	148	11	labels	label	NOUN
fcis-10735	148	12	)	)	PUNCT
fcis-10735	148	13	.	.	PUNCT
fcis-10735	149	1	the	the	DET
fcis-10735	149	2	whole	whole	ADJ
fcis-10735	149	3	training	training	NOUN
fcis-10735	149	4	process	process	NOUN
fcis-10735	149	5	takes	take	VERB
fcis-10735	149	6	a	a	DET
fcis-10735	149	7	short	short	ADJ
fcis-10735	149	8	time	time	NOUN
fcis-10735	149	9	and	and	CCONJ
fcis-10735	149	10	generates	generate	VERB
fcis-10735	149	11	the	the	DET
fcis-10735	149	12	training	training	NOUN
fcis-10735	149	13	result	result	NOUN
fcis-10735	149	14	model	model	NOUN
fcis-10735	149	15	.	.	PUNCT
fcis-10735	150	1	the	the	DET
fcis-10735	150	2	font	font	NOUN
fcis-10735	150	3	recognition	recognition	NOUN
fcis-10735	150	4	results	result	NOUN
fcis-10735	150	5	of	of	ADP
fcis-10735	150	6	the	the	DET
fcis-10735	150	7	random	random	ADJ
fcis-10735	150	8	forest	forest	NOUN
fcis-10735	150	9	algorithm	algorithm	NOUN
fcis-10735	150	10	for	for	ADP
fcis-10735	150	11	the	the	DET
fcis-10735	150	12	test	test	NOUN
fcis-10735	150	13	image	image	NOUN
fcis-10735	150	14	are	be	AUX
fcis-10735	150	15	shown	show	VERB
fcis-10735	150	16	in	in	ADP
fcis-10735	150	17	the	the	DET
fcis-10735	150	18	figure	figure	NOUN
fcis-10735	150	19	.	.	PUNCT
fcis-10735	151	1	it	it	PRON
fcis-10735	151	2	runs	run	VERB
fcis-10735	151	3	fast	fast	ADV
fcis-10735	151	4	and	and	CCONJ
fcis-10735	151	5	gives	give	VERB
fcis-10735	151	6	no	no	DET
fcis-10735	151	7	error	error	NOUN
fcis-10735	151	8	messages	message	NOUN
fcis-10735	151	9	.	.	PUNCT
fcis-10735	152	1	fig	fig	NOUN
fcis-10735	152	2	18	18	NUM
fcis-10735	152	3	.	.	PUNCT
fcis-10735	153	1	logistic	logistic	ADJ
fcis-10735	153	2	regression	regression	NOUN
fcis-10735	153	3	2.3	2.3	NUM
fcis-10735	153	4	.	.	PUNCT
fcis-10735	154	1	resnet	resnet	NOUN
fcis-10735	154	2	and	and	CCONJ
fcis-10735	154	3	swordnet	swordnet	PROPN
fcis-10735	154	4	2.3.1	2.3.1	NUM
fcis-10735	154	5	.	.	PUNCT
fcis-10735	155	1	workflow	workflow	NOUN
fcis-10735	155	2	resnet	resnet	NOUN
fcis-10735	155	3	:	:	PUNCT
fcis-10735	155	4	fig	fig	NOUN
fcis-10735	155	5	19	19	NUM
fcis-10735	155	6	.	.	PUNCT
fcis-10735	156	1	workflow	workflow	NOUN
fcis-10735	156	2	of	of	ADP
fcis-10735	156	3	resnet	resnet	NOUN
fcis-10735	156	4	[	[	X
fcis-10735	156	5	5	5	NUM
fcis-10735	156	6	]	]	PUNCT
fcis-10735	156	7	swordnet	swordnet	NOUN
fcis-10735	156	8	:	:	PUNCT
fcis-10735	156	9	fig	fig	NOUN
fcis-10735	156	10	20	20	NUM
fcis-10735	156	11	.	.	PUNCT
fcis-10735	157	1	swordnet	swordnet	NOUN
fcis-10735	157	2	structure	structure	NOUN
fcis-10735	157	3	diagram	diagram	NOUN
fcis-10735	157	4	[	[	X
fcis-10735	157	5	6	6	NUM
fcis-10735	157	6	]	]	SYM
fcis-10735	157	7	2.3.2	2.3.2	NUM
fcis-10735	157	8	.	.	PUNCT
fcis-10735	157	9	hardware	hardware	NOUN
fcis-10735	157	10	usage	usage	NOUN
fcis-10735	157	11	the	the	DET
fcis-10735	157	12	neural	neural	ADJ
fcis-10735	157	13	network	network	NOUN
fcis-10735	157	14	algorithm	algorithm	NOUN
fcis-10735	157	15	can	can	AUX
fcis-10735	157	16	use	use	VERB
fcis-10735	157	17	gpu	gpu	NOUN
fcis-10735	157	18	to	to	PART
fcis-10735	157	19	accelerate	accelerate	VERB
fcis-10735	157	20	the	the	DET
fcis-10735	157	21	model	model	NOUN
fcis-10735	157	22	operation	operation	NOUN
fcis-10735	157	23	process	process	NOUN
fcis-10735	157	24	.	.	PUNCT
fcis-10735	158	1	therefore	therefore	ADV
fcis-10735	158	2	,	,	PUNCT
fcis-10735	158	3	i	i	PRON
fcis-10735	158	4	choose	choose	VERB
fcis-10735	158	5	gpu1	gpu1	NOUN
fcis-10735	158	6	(	(	PUNCT
fcis-10735	158	7	rtx3060	rtx3060	VERB
fcis-10735	158	8	90w	90w	NUM
fcis-10735	158	9	)	)	PUNCT
fcis-10735	158	10	to	to	PART
fcis-10735	158	11	build	build	VERB
fcis-10735	158	12	and	and	CCONJ
fcis-10735	158	13	operate	operate	VERB
fcis-10735	158	14	the	the	DET
fcis-10735	158	15	entire	entire	ADJ
fcis-10735	158	16	neural	neural	ADJ
fcis-10735	158	17	network	network	NOUN
fcis-10735	158	18	model	model	NOUN
fcis-10735	158	19	.	.	PUNCT
fcis-10735	159	1	hardware	hardware	NOUN
fcis-10735	159	2	list	list	NOUN
fcis-10735	159	3	:	:	PUNCT
fcis-10735	159	4	device	device	NOUN
fcis-10735	159	5	name	name	NOUN
fcis-10735	159	6	:	:	PUNCT
fcis-10735	159	7	laptop-9lij8tiu	laptop-9lij8tiu	NOUN
fcis-10735	159	8	processor	processor	NOUN
fcis-10735	159	9	:	:	PUNCT
fcis-10735	159	10	11th	11th	ADJ
fcis-10735	159	11	gen	gen	PROPN
fcis-10735	159	12	intel(r	intel(r	PROPN
fcis-10735	159	13	)	)	PUNCT
fcis-10735	159	14	core	core	NOUN
fcis-10735	159	15	(	(	PUNCT
fcis-10735	159	16	tm	tm	NOUN
fcis-10735	159	17	)	)	PUNCT
fcis-10735	159	18	i9	i9	NOUN
fcis-10735	159	19	-	-	PUNCT
fcis-10735	159	20	11900h	11900h	NUM
fcis-10735	159	21	@	@	ADP
fcis-10735	159	22	2.50ghz	2.50ghz	NUM
fcis-10735	159	23	2.50	2.50	NUM
fcis-10735	159	24	ghz	ghz	NOUN
fcis-10735	159	25	12	12	NUM
fcis-10735	159	26	on	on	ADP
fcis-10735	159	27	-	-	PUNCT
fcis-10735	159	28	board	board	NOUN
fcis-10735	159	29	ram	ram	NOUN
fcis-10735	159	30	:	:	PUNCT
fcis-10735	159	31	16.0	16.0	NUM
fcis-10735	159	32	gb	gb	NOUN
fcis-10735	159	33	(	(	PUNCT
fcis-10735	159	34	15.7	15.7	NUM
fcis-10735	159	35	gb	gb	ADV
fcis-10735	159	36	available	available	ADJ
fcis-10735	159	37	)	)	PUNCT
fcis-10735	159	38	system	system	NOUN
fcis-10735	159	39	type	type	NOUN
fcis-10735	159	40	:	:	PUNCT
fcis-10735	159	41	64	64	NUM
fcis-10735	159	42	-	-	PUNCT
fcis-10735	159	43	bit	bit	NOUN
fcis-10735	159	44	operating	operate	VERB
fcis-10735	159	45	system	system	NOUN
fcis-10735	159	46	,	,	PUNCT
fcis-10735	159	47	x64	x64	NOUN
fcis-10735	159	48	-	-	PUNCT
fcis-10735	159	49	based	base	VERB
fcis-10735	159	50	processor	processor	NOUN
fcis-10735	159	51	2.3.3	2.3.3	NUM
fcis-10735	159	52	.	.	PUNCT
fcis-10735	160	1	theory	theory	NOUN
fcis-10735	160	2	2.3.3.1	2.3.3.1	NUM
fcis-10735	160	3	resnet	resnet	VERB
fcis-10735	160	4	2.3.3.1.1	2.3.3.1.1	PROPN
fcis-10735	160	5	residual	residual	ADJ
fcis-10735	160	6	learning	learning	NOUN
fcis-10735	160	7	resnet	resnet	NOUN
fcis-10735	160	8	has	have	VERB
fcis-10735	160	9	many	many	ADJ
fcis-10735	160	10	bypass	bypass	NOUN
fcis-10735	160	11	branches	branch	NOUN
fcis-10735	160	12	to	to	PART
fcis-10735	160	13	directly	directly	ADV
fcis-10735	160	14	connect	connect	VERB
fcis-10735	160	15	the	the	DET
fcis-10735	160	16	input	input	NOUN
fcis-10735	160	17	to	to	ADP
fcis-10735	160	18	the	the	DET
fcis-10735	160	19	following	follow	VERB
fcis-10735	160	20	layers	layer	NOUN
fcis-10735	160	21	,	,	PUNCT
fcis-10735	160	22	so	so	SCONJ
fcis-10735	160	23	that	that	SCONJ
fcis-10735	160	24	the	the	DET
fcis-10735	160	25	latter	latter	ADJ
fcis-10735	160	26	layers	layer	NOUN
fcis-10735	160	27	can	can	AUX
fcis-10735	160	28	directly	directly	ADV
fcis-10735	160	29	learn	learn	VERB
fcis-10735	160	30	the	the	DET
fcis-10735	160	31	residuals	residual	NOUN
fcis-10735	160	32	.	.	PUNCT
fcis-10735	161	1	this	this	DET
fcis-10735	161	2	structure	structure	NOUN
fcis-10735	161	3	is	be	AUX
fcis-10735	161	4	also	also	ADV
fcis-10735	161	5	called	call	VERB
fcis-10735	161	6	shortcut	shortcut	NOUN
fcis-10735	161	7	or	or	CCONJ
fcis-10735	161	8	skip	skip	ADJ
fcis-10735	161	9	connections	connection	NOUN
fcis-10735	161	10	.	.	PUNCT
fcis-10735	162	1	2.3.3.1.2	2.3.3.1.2	NUM
fcis-10735	162	2	equivalent	equivalent	ADJ
fcis-10735	162	3	mapping	mapping	NOUN
fcis-10735	162	4	via	via	ADP
fcis-10735	162	5	shortcut	shortcut	NOUN
fcis-10735	162	6	when	when	SCONJ
fcis-10735	162	7	the	the	DET
fcis-10735	162	8	input	input	NOUN
fcis-10735	162	9	and	and	CCONJ
fcis-10735	162	10	output	output	NOUN
fcis-10735	162	11	are	be	AUX
fcis-10735	162	12	the	the	DET
fcis-10735	162	13	same	same	ADJ
fcis-10735	162	14	,	,	PUNCT
fcis-10735	162	15	the	the	DET
fcis-10735	162	16	identity	identity	NOUN
fcis-10735	162	17	shortcut	shortcut	NOUN
fcis-10735	162	18	connection	connection	NOUN
fcis-10735	162	19	is	be	AUX
fcis-10735	162	20	used	use	VERB
fcis-10735	162	21	:	:	PUNCT
fcis-10735	162	22	when	when	SCONJ
fcis-10735	162	23	the	the	DET
fcis-10735	162	24	input	input	NOUN
fcis-10735	162	25	and	and	CCONJ
fcis-10735	162	26	output	output	NOUN
fcis-10735	162	27	dimensions	dimension	NOUN
fcis-10735	162	28	are	be	AUX
fcis-10735	162	29	different	different	ADJ
fcis-10735	162	30	,	,	PUNCT
fcis-10735	162	31	a	a	DET
fcis-10735	162	32	projection	projection	NOUN
fcis-10735	162	33	shortcut	shortcut	NOUN
fcis-10735	162	34	connection	connection	NOUN
fcis-10735	162	35	and	and	CCONJ
fcis-10735	162	36	padding	padding	NOUN
fcis-10735	162	37	with	with	ADP
fcis-10735	162	38	zeros	zero	NOUN
fcis-10735	162	39	are	be	AUX
fcis-10735	162	40	used	use	VERB
fcis-10735	162	41	:	:	PUNCT
fcis-10735	162	42	2.3.3.1.3	2.3.3.1.3	NUM
fcis-10735	162	43	network	network	NOUN
fcis-10735	162	44	structure	structure	NOUN
fcis-10735	162	45	resnet	resnet	NOUN
fcis-10735	162	46	mainly	mainly	ADV
fcis-10735	162	47	has	have	VERB
fcis-10735	162	48	five	five	NUM
fcis-10735	162	49	main	main	ADJ
fcis-10735	162	50	forms	form	NOUN
fcis-10735	162	51	:	:	PUNCT
fcis-10735	162	52	res18	res18	ADJ
fcis-10735	162	53	,	,	PUNCT
fcis-10735	162	54	res34	res34	ADV
fcis-10735	162	55	,	,	PUNCT
fcis-10735	162	56	res50	res50	NOUN
fcis-10735	162	57	,	,	PUNCT
fcis-10735	162	58	res101	res101	PROPN
fcis-10735	162	59	,	,	PUNCT
fcis-10735	162	60	res152	res152	PROPN
fcis-10735	162	61	.	.	PUNCT
fcis-10735	163	1	each	each	DET
fcis-10735	163	2	network	network	NOUN
fcis-10735	163	3	includes	include	VERB
fcis-10735	163	4	three	three	NUM
fcis-10735	163	5	main	main	ADJ
fcis-10735	163	6	parts	part	NOUN
fcis-10735	163	7	:	:	PUNCT
fcis-10735	163	8	input	input	NOUN
fcis-10735	163	9	part	part	NOUN
fcis-10735	163	10	,	,	PUNCT
fcis-10735	163	11	output	output	NOUN
fcis-10735	163	12	part	part	NOUN
fcis-10735	163	13	and	and	CCONJ
fcis-10735	163	14	intermediate	intermediate	ADJ
fcis-10735	163	15	convolution	convolution	NOUN
fcis-10735	163	16	part	part	NOUN
fcis-10735	163	17	2.3.3.2	2.3.3.2	NUM
fcis-10735	163	18	swordnet	swordnet	NOUN
fcis-10735	163	19	swordnet	swordnet	NOUN
fcis-10735	163	20	consists	consist	VERB
fcis-10735	163	21	of	of	ADP
fcis-10735	163	22	15	15	NUM
fcis-10735	163	23	convolutional	convolutional	ADJ
fcis-10735	163	24	layers	layer	NOUN
fcis-10735	163	25	.	.	PUNCT
fcis-10735	164	1	the	the	DET
fcis-10735	164	2	kernel	kernel	PROPN
fcis-10735	164	3	size	size	NOUN
fcis-10735	164	4	of	of	ADP
fcis-10735	164	5	each	each	DET
fcis-10735	164	6	layer	layer	NOUN
fcis-10735	164	7	is	be	AUX
fcis-10735	164	8	3	3	NUM
fcis-10735	164	9	*	*	SYM
fcis-10735	164	10	3	3	NUM
fcis-10735	164	11	and	and	CCONJ
fcis-10735	164	12	the	the	DET
fcis-10735	164	13	stride	stride	NOUN
fcis-10735	164	14	is	be	AUX
fcis-10735	164	15	2	2	NUM
fcis-10735	164	16	.	.	PUNCT
fcis-10735	165	1	2.3.3.2.1	2.3.3.2.1	NUM
fcis-10735	165	2	maxpooling	maxpooling	NOUN
fcis-10735	165	3	layer	layer	NOUN
fcis-10735	165	4	the	the	DET
fcis-10735	165	5	maximum	maximum	ADJ
fcis-10735	165	6	subsampling	subsample	VERB
fcis-10735	165	7	function	function	NOUN
fcis-10735	165	8	takes	take	VERB
fcis-10735	165	9	the	the	DET
fcis-10735	165	10	maximum	maximum	ADJ
fcis-10735	165	11	value	value	NOUN
fcis-10735	165	12	of	of	ADP
fcis-10735	165	13	all	all	DET
fcis-10735	165	14	neurons	neuron	NOUN
fcis-10735	165	15	in	in	ADP
fcis-10735	165	16	the	the	DET
fcis-10735	165	17	area	area	NOUN
fcis-10735	165	18	(	(	PUNCT
fcis-10735	165	19	max	max	NOUN
fcis-10735	165	20	-	-	PUNCT
fcis-10735	165	21	pooling	pooling	NOUN
fcis-10735	165	22	)	)	PUNCT
fcis-10735	165	23	.	.	PUNCT
fcis-10735	166	1	the	the	DET
fcis-10735	166	2	algorithm	algorithm	NOUN
fcis-10735	166	3	not	not	PART
fcis-10735	166	4	only	only	ADV
fcis-10735	166	5	calculates	calculate	VERB
fcis-10735	166	6	the	the	DET
fcis-10735	166	7	maximum	maximum	ADJ
fcis-10735	166	8	value	value	NOUN
fcis-10735	166	9	in	in	ADP
fcis-10735	166	10	the	the	DET
fcis-10735	166	11	pool	pool	NOUN
fcis-10735	166	12	area	area	NOUN
fcis-10735	166	13	,	,	PUNCT
fcis-10735	166	14	but	but	CCONJ
fcis-10735	166	15	also	also	ADV
fcis-10735	166	16	records	record	VERB
fcis-10735	166	17	the	the	DET
fcis-10735	166	18	position	position	NOUN
fcis-10735	166	19	of	of	ADP
fcis-10735	166	20	the	the	DET
fcis-10735	166	21	maximum	maximum	ADJ
fcis-10735	166	22	value	value	NOUN
fcis-10735	166	23	in	in	ADP
fcis-10735	166	24	the	the	DET
fcis-10735	166	25	input	input	NOUN
fcis-10735	166	26	data	data	PROPN
fcis-10735	166	27	.	.	PUNCT
fcis-10735	167	1	fig	fig	PROPN
fcis-10735	167	2	21	21	NUM
fcis-10735	167	3	.	.	PUNCT
fcis-10735	168	1	maxpooling	maxpoole	VERB
fcis-10735	168	2	layer	layer	NOUN
fcis-10735	168	3	[	[	X
fcis-10735	168	4	7	7	X
fcis-10735	168	5	]	]	SYM
fcis-10735	168	6	2.3.3.2.2	2.3.3.2.2	NUM
fcis-10735	168	7	skip	skip	ADJ
fcis-10735	168	8	connection	connection	NOUN
fcis-10735	168	9	fig	fig	NOUN
fcis-10735	168	10	22	22	NUM
fcis-10735	168	11	.	.	PUNCT
fcis-10735	169	1	skip	skip	ADJ
fcis-10735	169	2	connection	connection	NOUN
fcis-10735	169	3	[	[	X
fcis-10735	169	4	8	8	NUM
fcis-10735	169	5	]	]	PUNCT
fcis-10735	169	6	make	make	VERB
fcis-10735	169	7	a	a	DET
fcis-10735	169	8	residual	residual	ADJ
fcis-10735	169	9	block	block	NOUN
fcis-10735	169	10	(	(	PUNCT
fcis-10735	169	11	rb	rb	NOUN
fcis-10735	169	12	):	):	PUNCT
fcis-10735	169	13	this	this	DET
fcis-10735	169	14	block	block	NOUN
fcis-10735	169	15	provides	provide	VERB
fcis-10735	169	16	shortcuts	shortcut	NOUN
fcis-10735	169	17	in	in	ADP
fcis-10735	169	18	the	the	DET
fcis-10735	169	19	form	form	NOUN
fcis-10735	169	20	of	of	ADP
fcis-10735	169	21	an	an	DET
fcis-10735	169	22	"	"	PUNCT
fcis-10735	169	23	identity	identity	NOUN
fcis-10735	169	24	function	function	NOUN
fcis-10735	169	25	"	"	PUNCT
fcis-10735	169	26	.	.	PUNCT
fcis-10735	170	1	2.3.3.2.3	2.3.3.2.3	NUM
fcis-10735	170	2	global	global	ADJ
fcis-10735	170	3	average	average	ADJ
fcis-10735	170	4	pool	pool	NOUN
fcis-10735	170	5	the	the	DET
fcis-10735	170	6	network	network	NOUN
fcis-10735	170	7	in	in	ADP
fcis-10735	170	8	network	network	NOUN
fcis-10735	170	9	work	work	NOUN
fcis-10735	170	10	uses	use	VERB
fcis-10735	170	11	gap	gap	NOUN
fcis-10735	170	12	to	to	PART
fcis-10735	170	13	replace	replace	VERB
fcis-10735	170	14	the	the	DET
fcis-10735	170	15	last	last	ADJ
fcis-10735	170	16	fully	fully	ADV
fcis-10735	170	17	connected	connect	VERB
fcis-10735	170	18	layer	layer	NOUN
fcis-10735	170	19	,	,	PUNCT
fcis-10735	170	20	which	which	PRON
fcis-10735	170	21	directly	directly	ADV
fcis-10735	170	22	realizes	realize	VERB
fcis-10735	170	23	dimensionality	dimensionality	NOUN
fcis-10735	170	24	reduction	reduction	NOUN
fcis-10735	170	25	,	,	PUNCT
fcis-10735	170	26	and	and	CCONJ
fcis-10735	170	27	more	more	ADV
fcis-10735	170	28	importantly	importantly	ADV
fcis-10735	170	29	,	,	PUNCT
fcis-10735	170	30	greatly	greatly	ADV
fcis-10735	170	31	reduces	reduce	VERB
fcis-10735	170	32	the	the	DET
fcis-10735	170	33	parameters	parameter	NOUN
fcis-10735	170	34	of	of	ADP
fcis-10735	170	35	the	the	DET
fcis-10735	170	36	network	network	NOUN
fcis-10735	170	37	.	.	PUNCT
fcis-10735	171	1	fig	fig	NOUN
fcis-10735	171	2	23	23	NUM
fcis-10735	171	3	.	.	PUNCT
fcis-10735	172	1	global	global	ADJ
fcis-10735	172	2	average	average	ADJ
fcis-10735	172	3	pool	pool	NOUN
fcis-10735	172	4	[	[	X
fcis-10735	172	5	9	9	NUM
fcis-10735	172	6	]	]	SYM
fcis-10735	172	7	2.3.4	2.3.4	NUM
fcis-10735	172	8	.	.	PUNCT
fcis-10735	173	1	implementation	implementation	NOUN
fcis-10735	173	2	2.3.4.1	2.3.4.1	NUM
fcis-10735	173	3	import	import	NOUN
fcis-10735	173	4	all	all	DET
fcis-10735	173	5	required	require	VERB
fcis-10735	173	6	libraries	library	NOUN
fcis-10735	173	7	resnet	resnet	NOUN
fcis-10735	173	8	and	and	CCONJ
fcis-10735	173	9	swordnet	swordnet	PROPN
fcis-10735	173	10	mainly	mainly	ADV
fcis-10735	173	11	call	call	VERB
fcis-10735	173	12	the	the	DET
fcis-10735	173	13	ptorch	ptorch	NOUN
fcis-10735	173	14	package	package	NOUN
fcis-10735	173	15	for	for	ADP
fcis-10735	173	16	model	model	NOUN
fcis-10735	173	17	building	building	NOUN
fcis-10735	173	18	.	.	PUNCT
fcis-10735	174	1	it	it	PRON
fcis-10735	174	2	includes	include	VERB
fcis-10735	174	3	the	the	DET
fcis-10735	174	4	necessary	necessary	ADJ
fcis-10735	174	5	models	model	NOUN
fcis-10735	174	6	for	for	ADP
fcis-10735	174	7	neural	neural	ADJ
fcis-10735	174	8	network	network	NOUN
fcis-10735	174	9	construction	construction	NOUN
fcis-10735	174	10	.	.	PUNCT
fcis-10735	175	1	2.3.4.2	2.3.4.2	NUM
fcis-10735	175	2	modeling	model	VERB
fcis-10735	175	3	resnet	resnet	NOUN
fcis-10735	175	4	:	:	PUNCT
fcis-10735	175	5	fig	fig	NOUN
fcis-10735	175	6	24	24	NUM
fcis-10735	175	7	.	.	PUNCT
fcis-10735	176	1	resnet	resnet	ADJ
fcis-10735	176	2	detail	detail	NOUN
fcis-10735	176	3	[	[	X
fcis-10735	176	4	6	6	NUM
fcis-10735	176	5	]	]	X
fcis-10735	176	6	different	different	ADJ
fcis-10735	176	7	resnet	resnet	NOUN
fcis-10735	176	8	architectures	architecture	NOUN
fcis-10735	176	9	are	be	AUX
fcis-10735	176	10	unified	unify	VERB
fcis-10735	176	11	with	with	ADP
fcis-10735	176	12	one	one	NUM
fcis-10735	176	13	layer	layer	NOUN
fcis-10735	176	14	of	of	ADP
fcis-10735	176	15	feature	feature	NOUN
fcis-10735	176	16	extraction	extraction	NOUN
fcis-10735	176	17	and	and	CCONJ
fcis-10735	176	18	four	four	NUM
fcis-10735	176	19	layers	layer	NOUN
fcis-10735	176	20	of	of	ADP
fcis-10735	176	21	residuals	residual	NOUN
fcis-10735	176	22	,	,	PUNCT
fcis-10735	176	23	and	and	CCONJ
fcis-10735	176	24	the	the	DET
fcis-10735	176	25	difference	difference	NOUN
fcis-10735	176	26	lies	lie	VERB
fcis-10735	176	27	in	in	ADP
fcis-10735	176	28	the	the	DET
fcis-10735	176	29	depth	depth	NOUN
fcis-10735	176	30	of	of	ADP
fcis-10735	176	31	each	each	DET
fcis-10735	176	32	layer	layer	NOUN
fcis-10735	176	33	of	of	ADP
fcis-10735	176	34	residuals	residual	NOUN
fcis-10735	176	35	.	.	PUNCT
fcis-10735	177	1	for	for	ADP
fcis-10735	177	2	images	image	NOUN
fcis-10735	177	3	,	,	PUNCT
fcis-10735	177	4	the	the	DET
fcis-10735	177	5	feature	feature	NOUN
fcis-10735	177	6	map	map	NOUN
fcis-10735	177	7	size	size	NOUN
fcis-10735	177	8	changes	change	NOUN
fcis-10735	177	9	as	as	SCONJ
fcis-10735	177	10	follows	follow	VERB
fcis-10735	177	11	:	:	PUNCT
fcis-10735	177	12	(	(	PUNCT
fcis-10735	177	13	32	32	NUM
fcis-10735	177	14	,	,	PUNCT
fcis-10735	177	15	32	32	NUM
fcis-10735	177	16	,	,	PUNCT
fcis-10735	177	17	3	3	X
fcis-10735	177	18	)	)	PUNCT
fcis-10735	177	19	-	-	PUNCT
fcis-10735	177	20	>	>	X
fcis-10735	178	1	[	[	X
fcis-10735	178	2	conv2d	conv2d	X
fcis-10735	178	3	]	]	X
fcis-10735	178	4	-	-	PUNCT
fcis-10735	178	5	>	>	X
fcis-10735	178	6	(	(	PUNCT
fcis-10735	178	7	32	32	NUM
fcis-10735	178	8	,	,	PUNCT
fcis-10735	178	9	32	32	NUM
fcis-10735	178	10	,	,	PUNCT
fcis-10735	178	11	64	64	NUM
fcis-10735	178	12	)	)	PUNCT
fcis-10735	178	13	-	-	PUNCT
fcis-10735	178	14	>	>	X
fcis-10735	179	1	[	[	X
fcis-10735	179	2	res1	res1	NOUN
fcis-10735	179	3	]	]	PUNCT
fcis-10735	179	4	-	-	PUNCT
fcis-10735	179	5	>	>	X
fcis-10735	179	6	(	(	PUNCT
fcis-10735	179	7	32	32	NUM
fcis-10735	179	8	,	,	PUNCT
fcis-10735	179	9	32	32	NUM
fcis-10735	179	10	,	,	PUNCT
fcis-10735	179	11	64	64	NUM
fcis-10735	179	12	)	)	PUNCT
fcis-10735	179	13	-	-	PUNCT
fcis-10735	179	14	>	>	X
fcis-10735	180	1	[	[	X
fcis-10735	180	2	res2	res2	NOUN
fcis-10735	180	3	]	]	X
fcis-10735	180	4	-	-	PUNCT
fcis-10735	180	5	>	>	X
fcis-10735	180	6	(	(	PUNCT
fcis-10735	180	7	16	16	NUM
fcis-10735	180	8	,	,	PUNCT
fcis-10735	180	9	16	16	NUM
fcis-10735	180	10	,	,	PUNCT
fcis-10735	180	11	128	128	NUM
fcis-10735	180	12	)	)	PUNCT
fcis-10735	180	13	-	-	PUNCT
fcis-10735	180	14	>	>	X
fcis-10735	181	1	[	[	X
fcis-10735	181	2	res3	res3	X
fcis-10735	181	3	]	]	X
fcis-10735	181	4	-	-	PUNCT
fcis-10735	181	5	>	>	X
fcis-10735	181	6	(	(	PUNCT
fcis-10735	181	7	8	8	NUM
fcis-10735	181	8	,	,	PUNCT
fcis-10735	181	9	8	8	NUM
fcis-10735	181	10	,	,	PUNCT
fcis-10735	181	11	256	256	NUM
fcis-10735	181	12	)	)	PUNCT
fcis-10735	181	13	->[res4	->[res4	NOUN
fcis-10735	181	14	]	]	X
fcis-10735	181	15	-	-	PUNCT
fcis-10735	181	16	>	>	X
fcis-10735	181	17	(	(	PUNCT
fcis-10735	181	18	4	4	NUM
fcis-10735	181	19	,	,	PUNCT
fcis-10735	181	20	4	4	NUM
fcis-10735	181	21	,	,	PUNCT
fcis-10735	181	22	512	512	NUM
fcis-10735	181	23	)	)	PUNCT
fcis-10735	181	24	-	-	PUNCT
fcis-10735	181	25	>	>	X
fcis-10735	182	1	[	[	X
fcis-10735	182	2	avgpool	avgpool	X
fcis-10735	182	3	]	]	PUNCT
fcis-10735	182	4	-	-	PUNCT
fcis-10735	182	5	>	>	X
fcis-10735	182	6	(	(	PUNCT
fcis-10735	182	7	1	1	NUM
fcis-10735	182	8	,	,	PUNCT
fcis-10735	182	9	1	1	NUM
fcis-10735	182	10	,	,	PUNCT
fcis-10735	182	11	512	512	NUM
fcis-10735	182	12	)	)	PUNCT
fcis-10735	182	13	-	-	PUNCT
fcis-10735	182	14	>	>	X
fcis-10735	183	1	[	[	X
fcis-10735	183	2	reshape	reshape	NOUN
fcis-10735	183	3	]	]	PUNCT
fcis-10735	183	4	-	-	PUNCT
fcis-10735	183	5	>	>	X
fcis-10735	183	6	(	(	PUNCT
fcis-10735	183	7	512	512	NUM
fcis-10735	183	8	)	)	PUNCT
fcis-10735	183	9	-	-	PUNCT
fcis-10735	183	10	>	>	X
fcis-10735	184	1	[	[	X
fcis-10735	184	2	linear	linear	X
fcis-10735	184	3	]	]	X
fcis-10735	184	4	-	-	PUNCT
fcis-10735	184	5	>	>	X
fcis-10735	184	6	(	(	PUNCT
fcis-10735	184	7	10	10	NUM
fcis-10735	184	8	)	)	PUNCT
fcis-10735	184	9	.	.	PUNCT
fcis-10735	185	1	swordnet	swordnet	NOUN
fcis-10735	185	2	:	:	PUNCT
fcis-10735	185	3	the	the	DET
fcis-10735	185	4	input	input	NOUN
fcis-10735	185	5	image	image	NOUN
fcis-10735	185	6	of	of	ADP
fcis-10735	185	7	swordnet	swordnet	NOUN
fcis-10735	185	8	is	be	AUX
fcis-10735	185	9	in	in	ADP
fcis-10735	185	10	96	96	NUM
fcis-10735	185	11	*	*	SYM
fcis-10735	185	12	96	96	NUM
fcis-10735	185	13	*	*	SYM
fcis-10735	185	14	3	3	NUM
fcis-10735	185	15	format	format	NOUN
fcis-10735	185	16	.	.	PUNCT
fcis-10735	186	1	the	the	DET
fcis-10735	186	2	algorithm	algorithm	NOUN
fcis-10735	186	3	has	have	VERB
fcis-10735	186	4	a	a	DET
fcis-10735	186	5	total	total	NOUN
fcis-10735	186	6	of	of	ADP
fcis-10735	186	7	15	15	NUM
fcis-10735	186	8	convolutional	convolutional	ADJ
fcis-10735	186	9	blocks	block	NOUN
fcis-10735	186	10	.	.	PUNCT
fcis-10735	187	1	each	each	DET
fcis-10735	187	2	block	block	NOUN
fcis-10735	187	3	consists	consist	VERB
fcis-10735	187	4	of	of	ADP
fcis-10735	187	5	convolutional	convolutional	ADJ
fcis-10735	187	6	layer	layer	NOUN
fcis-10735	187	7	,	,	PUNCT
fcis-10735	187	8	bn	bn	NOUN
fcis-10735	187	9	layer	layer	NOUN
fcis-10735	187	10	,	,	PUNCT
fcis-10735	187	11	relu	relu	NOUN
fcis-10735	187	12	activation	activation	NOUN
fcis-10735	187	13	function	function	NOUN
fcis-10735	187	14	layer	layer	NOUN
fcis-10735	187	15	and	and	CCONJ
fcis-10735	187	16	maxpooling	maxpoole	VERB
fcis-10735	187	17	layer	layer	NOUN
fcis-10735	187	18	.	.	PUNCT
fcis-10735	188	1	swordnet	swordnet	PROPN
fcis-10735	188	2	also	also	ADV
fcis-10735	188	3	includes	include	VERB
fcis-10735	188	4	3	3	NUM
fcis-10735	188	5	skip	skip	ADJ
fcis-10735	188	6	connections	connection	NOUN
fcis-10735	188	7	.	.	PUNCT
fcis-10735	189	1	at	at	ADP
fcis-10735	189	2	the	the	DET
fcis-10735	189	3	end	end	NOUN
fcis-10735	189	4	of	of	ADP
fcis-10735	189	5	swordnet	swordnet	NOUN
fcis-10735	189	6	gap	gap	NOUN
fcis-10735	189	7	and	and	CCONJ
fcis-10735	189	8	13	13	NUM
fcis-10735	189	9	swordnet	swordnet	NOUN
fcis-10735	189	10	will	will	AUX
fcis-10735	189	11	output	output	VERB
fcis-10735	189	12	the	the	DET
fcis-10735	189	13	predicted	predict	VERB
fcis-10735	189	14	font	font	NOUN
fcis-10735	189	15	style	style	NOUN
fcis-10735	189	16	.	.	PUNCT
fcis-10735	190	1	fig	fig	NOUN
fcis-10735	190	2	25	25	NUM
fcis-10735	190	3	.	.	PUNCT
fcis-10735	191	1	swordnet	swordnet	NOUN
fcis-10735	191	2	detail	detail	NOUN
fcis-10735	191	3	[	[	X
fcis-10735	191	4	6	6	NUM
fcis-10735	191	5	]	]	PUNCT
fcis-10735	191	6	2.3.4.3	2.3.4.3	NUM
fcis-10735	191	7	training	training	NOUN
fcis-10735	191	8	and	and	CCONJ
fcis-10735	191	9	results	result	NOUN
fcis-10735	191	10	both	both	DET
fcis-10735	191	11	resnet	resnet	NOUN
fcis-10735	191	12	and	and	CCONJ
fcis-10735	191	13	swordnet	swordnet	NOUN
fcis-10735	191	14	have	have	VERB
fcis-10735	191	15	good	good	ADJ
fcis-10735	191	16	accuracy	accuracy	NOUN
fcis-10735	191	17	after	after	ADP
fcis-10735	191	18	training	training	NOUN
fcis-10735	191	19	.	.	PUNCT
fcis-10735	192	1	and	and	CCONJ
fcis-10735	192	2	according	accord	VERB
fcis-10735	192	3	to	to	ADP
fcis-10735	192	4	my	my	PRON
fcis-10735	192	5	speculation	speculation	NOUN
fcis-10735	192	6	,	,	PUNCT
fcis-10735	192	7	the	the	DET
fcis-10735	192	8	accuracy	accuracy	NOUN
fcis-10735	192	9	rate	rate	NOUN
fcis-10735	192	10	of	of	ADP
fcis-10735	192	11	resnet	resnet	NOUN
fcis-10735	192	12	and	and	CCONJ
fcis-10735	192	13	swordnet	swordnet	NOUN
fcis-10735	192	14	will	will	AUX
fcis-10735	192	15	reach	reach	VERB
fcis-10735	192	16	more	more	ADJ
fcis-10735	192	17	than	than	ADP
fcis-10735	192	18	0.90	0.90	NUM
fcis-10735	192	19	.	.	PUNCT
fcis-10735	193	1	swordnet	swordnet	PROPN
fcis-10735	193	2	will	will	AUX
fcis-10735	193	3	have	have	VERB
fcis-10735	193	4	better	well	ADJ
fcis-10735	193	5	prediction	prediction	NOUN
fcis-10735	193	6	effect	effect	NOUN
fcis-10735	193	7	than	than	ADP
fcis-10735	193	8	resnet	resnet	NOUN
fcis-10735	193	9	.	.	PUNCT
fcis-10735	194	1	3	3	X
fcis-10735	194	2	.	.	X
fcis-10735	194	3	evaluation	evaluation	NOUN
fcis-10735	194	4	3.1	3.1	NUM
fcis-10735	194	5	.	.	PUNCT
fcis-10735	194	6	random	random	ADJ
fcis-10735	194	7	forest	forest	NOUN
fcis-10735	194	8	algorithm	algorithm	NOUN
fcis-10735	194	9	3.1.1	3.1.1	NUM
fcis-10735	194	10	.	.	PUNCT
fcis-10735	195	1	results	result	NOUN
fcis-10735	195	2	accuracy	accuracy	NOUN
fcis-10735	195	3	:	:	PUNCT
fcis-10735	195	4	recall	recall	NOUN
fcis-10735	195	5	:	:	PUNCT
fcis-10735	195	6	confusion	confusion	NOUN
fcis-10735	195	7	matrix	matrix	NOUN
fcis-10735	195	8	:	:	PUNCT
fcis-10735	195	9	fig	fig	NOUN
fcis-10735	195	10	26	26	NUM
fcis-10735	195	11	.	.	PUNCT
fcis-10735	196	1	confusion	confusion	NOUN
fcis-10735	196	2	matrix	matrix	NOUN
fcis-10735	196	3	3.1.2	3.1.2	NUM
fcis-10735	196	4	.	.	PUNCT
fcis-10735	197	1	cross	cross	PROPN
fcis-10735	197	2	validation	validation	PROPN
fcis-10735	197	3	fig	fig	PROPN
fcis-10735	197	4	27	27	NUM
fcis-10735	197	5	.	.	PUNCT
fcis-10735	198	1	cross	cross	VERB
fcis-10735	198	2	validation	validation	PROPN
fcis-10735	198	3	3.1.3	3.1.3	NUM
fcis-10735	198	4	.	.	PUNCT
fcis-10735	199	1	analysis	analysis	NOUN
fcis-10735	199	2	i	i	PRON
fcis-10735	199	3	get	get	VERB
fcis-10735	199	4	the	the	DET
fcis-10735	199	5	accuracy	accuracy	NOUN
fcis-10735	199	6	of	of	ADP
fcis-10735	199	7	random	random	ADJ
fcis-10735	199	8	forest	forest	NOUN
fcis-10735	199	9	algorithm	algorithm	NOUN
fcis-10735	199	10	on	on	ADP
fcis-10735	199	11	the	the	DET
fcis-10735	199	12	training	training	NOUN
fcis-10735	199	13	set	set	NOUN
fcis-10735	199	14	is	be	AUX
fcis-10735	199	15	0.994	0.994	NUM
fcis-10735	199	16	and	and	CCONJ
fcis-10735	199	17	the	the	DET
fcis-10735	199	18	recall	recall	NOUN
fcis-10735	199	19	is	be	AUX
fcis-10735	199	20	0.994	0.994	NUM
fcis-10735	199	21	.	.	PUNCT
fcis-10735	200	1	this	this	PRON
fcis-10735	200	2	means	mean	VERB
fcis-10735	200	3	that	that	SCONJ
fcis-10735	200	4	14	14	NUM
fcis-10735	200	5	random	random	ADJ
fcis-10735	200	6	forest	forest	NOUN
fcis-10735	200	7	algorithm	algorithm	NOUN
fcis-10735	200	8	can	can	AUX
fcis-10735	200	9	get	get	VERB
fcis-10735	200	10	good	good	ADJ
fcis-10735	200	11	prediction	prediction	NOUN
fcis-10735	200	12	performance	performance	NOUN
fcis-10735	200	13	with	with	ADP
fcis-10735	200	14	parameter	parameter	NOUN
fcis-10735	200	15	optimization	optimization	NOUN
fcis-10735	200	16	.	.	PUNCT
fcis-10735	201	1	we	we	PRON
fcis-10735	201	2	can	can	AUX
fcis-10735	201	3	see	see	VERB
fcis-10735	201	4	that	that	PRON
fcis-10735	201	5	random	random	ADJ
fcis-10735	201	6	forest	forest	NOUN
fcis-10735	201	7	predicts	predict	VERB
fcis-10735	201	8	the	the	DET
fcis-10735	201	9	wrong	wrong	ADJ
fcis-10735	201	10	font	font	NOUN
fcis-10735	201	11	style	style	NOUN
fcis-10735	201	12	as	as	SCONJ
fcis-10735	201	13	shown	show	VERB
fcis-10735	201	14	in	in	ADP
fcis-10735	201	15	figure	figure	NOUN
fcis-10735	201	16	.	.	PUNCT
fcis-10735	202	1	from	from	ADP
fcis-10735	202	2	visual	visual	ADJ
fcis-10735	202	3	observation	observation	NOUN
fcis-10735	202	4	,	,	PUNCT
fcis-10735	202	5	we	we	PRON
fcis-10735	202	6	can	can	AUX
fcis-10735	202	7	clearly	clearly	ADV
fcis-10735	202	8	see	see	VERB
fcis-10735	202	9	that	that	SCONJ
fcis-10735	202	10	these	these	DET
fcis-10735	202	11	misidentified	misidentifie	VERB
fcis-10735	202	12	fonts	font	NOUN
fcis-10735	202	13	have	have	VERB
fcis-10735	202	14	some	some	DET
fcis-10735	202	15	similarities	similarity	NOUN
fcis-10735	202	16	.	.	PUNCT
fcis-10735	203	1	it	it	PRON
fcis-10735	203	2	may	may	AUX
fcis-10735	203	3	be	be	AUX
fcis-10735	203	4	an	an	DET
fcis-10735	203	5	error	error	NOUN
fcis-10735	203	6	due	due	ADP
fcis-10735	203	7	to	to	ADP
fcis-10735	203	8	high	high	ADJ
fcis-10735	203	9	font	font	NOUN
fcis-10735	203	10	similarity	similarity	NOUN
fcis-10735	203	11	.	.	PUNCT
fcis-10735	204	1	the	the	DET
fcis-10735	204	2	accuracy	accuracy	NOUN
fcis-10735	204	3	rate	rate	NOUN
fcis-10735	204	4	,	,	PUNCT
fcis-10735	204	5	recall	recall	NOUN
fcis-10735	204	6	rate	rate	NOUN
fcis-10735	204	7	and	and	CCONJ
fcis-10735	204	8	f1	f1	NOUN
fcis-10735	204	9	value	value	NOUN
fcis-10735	204	10	of	of	ADP
fcis-10735	204	11	the	the	DET
fcis-10735	204	12	entire	entire	ADJ
fcis-10735	204	13	model	model	NOUN
fcis-10735	204	14	are	be	AUX
fcis-10735	204	15	all	all	ADV
fcis-10735	204	16	high	high	ADJ
fcis-10735	204	17	.	.	PUNCT
fcis-10735	205	1	this	this	PRON
fcis-10735	205	2	means	mean	VERB
fcis-10735	205	3	that	that	SCONJ
fcis-10735	205	4	the	the	DET
fcis-10735	205	5	model	model	NOUN
fcis-10735	205	6	is	be	AUX
fcis-10735	205	7	successfully	successfully	ADV
fcis-10735	205	8	established	establish	VERB
fcis-10735	205	9	and	and	CCONJ
fcis-10735	205	10	the	the	DET
fcis-10735	205	11	parameter	parameter	NOUN
fcis-10735	205	12	optimization	optimization	NOUN
fcis-10735	205	13	effect	effect	NOUN
fcis-10735	205	14	is	be	AUX
fcis-10735	205	15	good	good	ADJ
fcis-10735	205	16	.	.	PUNCT
fcis-10735	206	1	fig	fig	NOUN
fcis-10735	206	2	28	28	NUM
fcis-10735	206	3	.	.	PUNCT
fcis-10735	207	1	the	the	DET
fcis-10735	207	2	result	result	NOUN
fcis-10735	207	3	fonts	font	VERB
fcis-10735	207	4	3.2	3.2	NUM
fcis-10735	207	5	.	.	PUNCT
fcis-10735	208	1	logistic	logistic	ADJ
fcis-10735	208	2	regression	regression	NOUN
fcis-10735	208	3	3.2.1	3.2.1	NUM
fcis-10735	208	4	.	.	PUNCT
fcis-10735	209	1	results	result	NOUN
fcis-10735	209	2	accuracy	accuracy	NOUN
fcis-10735	209	3	:	:	PUNCT
fcis-10735	209	4	recall	recall	NOUN
fcis-10735	209	5	:	:	PUNCT
fcis-10735	209	6	confusion	confusion	NOUN
fcis-10735	209	7	matrix	matrix	NOUN
fcis-10735	209	8	:	:	PUNCT
fcis-10735	209	9	fig	fig	NOUN
fcis-10735	209	10	29	29	NUM
fcis-10735	209	11	.	.	PUNCT
fcis-10735	210	1	confusion	confusion	NOUN
fcis-10735	210	2	matrix	matrix	NOUN
fcis-10735	210	3	3.2.2	3.2.2	NUM
fcis-10735	210	4	.	.	PUNCT
fcis-10735	211	1	cross	cross	NOUN
fcis-10735	211	2	validation	validation	PROPN
fcis-10735	211	3	fig	fig	NOUN
fcis-10735	211	4	30	30	NUM
fcis-10735	211	5	.	.	PUNCT
fcis-10735	212	1	cross	cross	VERB
fcis-10735	212	2	validation	validation	PROPN
fcis-10735	212	3	3.2.3	3.2.3	NUM
fcis-10735	212	4	.	.	PUNCT
fcis-10735	213	1	analysis	analysis	NOUN
fcis-10735	213	2	the	the	DET
fcis-10735	213	3	final	final	ADJ
fcis-10735	213	4	prediction	prediction	NOUN
fcis-10735	213	5	result	result	NOUN
fcis-10735	213	6	of	of	ADP
fcis-10735	213	7	the	the	DET
fcis-10735	213	8	logistic	logistic	ADJ
fcis-10735	213	9	regression	regression	NOUN
fcis-10735	213	10	model	model	NOUN
fcis-10735	213	11	has	have	VERB
fcis-10735	213	12	an	an	DET
fcis-10735	213	13	accuracy	accuracy	NOUN
fcis-10735	213	14	of	of	ADP
fcis-10735	213	15	0.71	0.71	NUM
fcis-10735	213	16	and	and	CCONJ
fcis-10735	213	17	a	a	DET
fcis-10735	213	18	recall	recall	NOUN
fcis-10735	213	19	of	of	ADP
fcis-10735	213	20	0.70	0.70	NUM
fcis-10735	213	21	.	.	PUNCT
fcis-10735	214	1	the	the	DET
fcis-10735	214	2	prediction	prediction	NOUN
fcis-10735	214	3	accuracy	accuracy	NOUN
fcis-10735	214	4	of	of	ADP
fcis-10735	214	5	this	this	DET
fcis-10735	214	6	model	model	NOUN
fcis-10735	214	7	exceeds	exceed	VERB
fcis-10735	214	8	0.6	0.6	NUM
fcis-10735	214	9	,	,	PUNCT
fcis-10735	214	10	which	which	PRON
fcis-10735	214	11	means	mean	VERB
fcis-10735	214	12	that	that	SCONJ
fcis-10735	214	13	the	the	DET
fcis-10735	214	14	entire	entire	ADJ
fcis-10735	214	15	model	model	NOUN
fcis-10735	214	16	can	can	AUX
fcis-10735	214	17	make	make	VERB
fcis-10735	214	18	accurate	accurate	ADJ
fcis-10735	214	19	predictions	prediction	NOUN
fcis-10735	214	20	but	but	CCONJ
fcis-10735	214	21	the	the	DET
fcis-10735	214	22	final	final	ADJ
fcis-10735	214	23	result	result	NOUN
fcis-10735	214	24	may	may	AUX
fcis-10735	214	25	still	still	ADV
fcis-10735	214	26	have	have	VERB
fcis-10735	214	27	errors	error	NOUN
fcis-10735	214	28	.	.	PUNCT
fcis-10735	215	1	according	accord	VERB
fcis-10735	215	2	to	to	ADP
fcis-10735	215	3	the	the	DET
fcis-10735	215	4	results	result	NOUN
fcis-10735	215	5	of	of	ADP
fcis-10735	215	6	cross	cross	NOUN
fcis-10735	215	7	validation	validation	NOUN
fcis-10735	215	8	and	and	CCONJ
fcis-10735	215	9	confusion	confusion	NOUN
fcis-10735	215	10	matrix	matrix	NOUN
fcis-10735	215	11	i	i	PRON
fcis-10735	215	12	can	can	AUX
fcis-10735	215	13	conclude	conclude	VERB
fcis-10735	215	14	that	that	SCONJ
fcis-10735	215	15	this	this	DET
fcis-10735	215	16	model	model	NOUN
fcis-10735	215	17	has	have	VERB
fcis-10735	215	18	the	the	DET
fcis-10735	215	19	lowest	low	ADJ
fcis-10735	215	20	prediction	prediction	NOUN
fcis-10735	215	21	accuracy	accuracy	NOUN
fcis-10735	215	22	for	for	ADP
fcis-10735	215	23	'	'	PUNCT
fcis-10735	215	24	antquabi	antquabi	NOUN
fcis-10735	215	25	'	'	PART
fcis-10735	215	26	font	font	NOUN
fcis-10735	215	27	it	it	PRON
fcis-10735	215	28	is	be	AUX
fcis-10735	215	29	only	only	ADV
fcis-10735	215	30	0.428	0.428	NUM
fcis-10735	215	31	.	.	PUNCT
fcis-10735	216	1	at	at	ADP
fcis-10735	216	2	the	the	DET
fcis-10735	216	3	same	same	ADJ
fcis-10735	216	4	time	time	NOUN
fcis-10735	216	5	,	,	PUNCT
fcis-10735	216	6	this	this	DET
fcis-10735	216	7	model	model	NOUN
fcis-10735	216	8	has	have	VERB
fcis-10735	216	9	the	the	DET
fcis-10735	216	10	lowest	low	ADJ
fcis-10735	216	11	recall	recall	NOUN
fcis-10735	216	12	rate	rate	NOUN
fcis-10735	216	13	for	for	ADP
fcis-10735	216	14	'	'	PUNCT
fcis-10735	216	15	harngton	harngton	NOUN
fcis-10735	216	16	'	'	PUNCT
fcis-10735	216	17	font	font	NOUN
fcis-10735	216	18	which	which	PRON
fcis-10735	216	19	is	be	AUX
fcis-10735	216	20	only	only	ADV
fcis-10735	216	21	0.388	0.388	NUM
fcis-10735	216	22	.	.	PUNCT
fcis-10735	217	1	this	this	PRON
fcis-10735	217	2	means	mean	VERB
fcis-10735	217	3	that	that	SCONJ
fcis-10735	217	4	the	the	DET
fcis-10735	217	5	model	model	NOUN
fcis-10735	217	6	's	's	PART
fcis-10735	217	7	learning	learning	NOUN
fcis-10735	217	8	of	of	ADP
fcis-10735	217	9	some	some	DET
fcis-10735	217	10	font	font	NOUN
fcis-10735	217	11	images	image	NOUN
fcis-10735	217	12	is	be	AUX
fcis-10735	217	13	still	still	ADV
fcis-10735	217	14	incomplete	incomplete	ADJ
fcis-10735	217	15	and	and	CCONJ
fcis-10735	217	16	the	the	DET
fcis-10735	217	17	prediction	prediction	NOUN
fcis-10735	217	18	results	result	NOUN
fcis-10735	217	19	may	may	AUX
fcis-10735	217	20	be	be	AUX
fcis-10735	217	21	reduced	reduce	VERB
fcis-10735	217	22	due	due	ADP
fcis-10735	217	23	to	to	ADP
fcis-10735	217	24	the	the	DET
fcis-10735	217	25	high	high	ADJ
fcis-10735	217	26	similarity	similarity	NOUN
fcis-10735	217	27	of	of	ADP
fcis-10735	217	28	the	the	DET
fcis-10735	217	29	overall	overall	ADJ
fcis-10735	217	30	image	image	NOUN
fcis-10735	217	31	set	set	VERB
fcis-10735	217	32	.	.	PUNCT
fcis-10735	218	1	3.3	3.3	NUM
fcis-10735	218	2	.	.	PUNCT
fcis-10735	218	3	resnet	resnet	NOUN
fcis-10735	218	4	and	and	CCONJ
fcis-10735	218	5	swordnet	swordnet	PROPN
fcis-10735	218	6	3.3.1	3.3.1	PROPN
fcis-10735	218	7	.	.	PUNCT
fcis-10735	218	8	results	result	NOUN
fcis-10735	218	9	according	accord	VERB
fcis-10735	218	10	to	to	ADP
fcis-10735	218	11	the	the	DET
fcis-10735	218	12	figure	figure	NOUN
fcis-10735	218	13	32	32	NUM
fcis-10735	218	14	,	,	PUNCT
fcis-10735	218	15	it	it	PRON
fcis-10735	218	16	can	can	AUX
fcis-10735	218	17	be	be	AUX
fcis-10735	218	18	concluded	conclude	VERB
fcis-10735	218	19	that	that	SCONJ
fcis-10735	218	20	the	the	DET
fcis-10735	218	21	prediction	prediction	NOUN
fcis-10735	218	22	results	result	NOUN
fcis-10735	218	23	of	of	ADP
fcis-10735	218	24	resnet	resnet	NOUN
fcis-10735	218	25	and	and	CCONJ
fcis-10735	218	26	swordnet	swordnet	NOUN
fcis-10735	218	27	are	be	AUX
fcis-10735	218	28	higher	high	ADJ
fcis-10735	218	29	than	than	ADP
fcis-10735	218	30	15	15	NUM
fcis-10735	218	31	other	other	ADJ
fcis-10735	218	32	models	model	NOUN
fcis-10735	218	33	.	.	PUNCT
fcis-10735	219	1	at	at	ADP
fcis-10735	219	2	the	the	DET
fcis-10735	219	3	same	same	ADJ
fcis-10735	219	4	time	time	NOUN
fcis-10735	219	5	,	,	PUNCT
fcis-10735	219	6	swordnet	swordnet	NOUN
fcis-10735	219	7	has	have	VERB
fcis-10735	219	8	better	well	ADJ
fcis-10735	219	9	prediction	prediction	NOUN
fcis-10735	219	10	accuracy	accuracy	NOUN
fcis-10735	219	11	than	than	ADP
fcis-10735	219	12	resnet	resnet	NOUN
fcis-10735	219	13	.	.	PUNCT
fcis-10735	220	1	fig	fig	NOUN
fcis-10735	220	2	31	31	NUM
fcis-10735	220	3	.	.	PUNCT
fcis-10735	221	1	accuracy	accuracy	NOUN
fcis-10735	221	2	results	result	NOUN
fcis-10735	222	1	[	[	X
fcis-10735	222	2	6	6	NUM
fcis-10735	222	3	]	]	SYM
fcis-10735	222	4	3.3.2	3.3.2	NUM
fcis-10735	222	5	.	.	PUNCT
fcis-10735	223	1	analysis	analysis	NOUN
fcis-10735	223	2	the	the	DET
fcis-10735	223	3	prediction	prediction	NOUN
fcis-10735	223	4	accuracy	accuracy	NOUN
fcis-10735	223	5	of	of	ADP
fcis-10735	223	6	resnet	resnet	NOUN
fcis-10735	223	7	and	and	CCONJ
fcis-10735	223	8	swordnet	swordnet	NOUN
fcis-10735	223	9	is	be	AUX
fcis-10735	223	10	very	very	ADV
fcis-10735	223	11	high	high	ADJ
fcis-10735	223	12	.	.	PUNCT
fcis-10735	224	1	the	the	DET
fcis-10735	224	2	model	model	NOUN
fcis-10735	224	3	can	can	AUX
fcis-10735	224	4	also	also	ADV
fcis-10735	224	5	be	be	AUX
fcis-10735	224	6	generalized	generalize	VERB
fcis-10735	224	7	.	.	PUNCT
fcis-10735	225	1	however	however	ADV
fcis-10735	225	2	,	,	PUNCT
fcis-10735	225	3	the	the	DET
fcis-10735	225	4	training	training	NOUN
fcis-10735	225	5	of	of	ADP
fcis-10735	225	6	the	the	DET
fcis-10735	225	7	entire	entire	ADJ
fcis-10735	225	8	model	model	NOUN
fcis-10735	225	9	requires	require	VERB
fcis-10735	225	10	a	a	DET
fcis-10735	225	11	long	long	ADJ
fcis-10735	225	12	time	time	NOUN
fcis-10735	225	13	and	and	CCONJ
fcis-10735	225	14	a	a	DET
fcis-10735	225	15	large	large	ADJ
fcis-10735	225	16	amount	amount	NOUN
fcis-10735	225	17	of	of	ADP
fcis-10735	225	18	data	datum	NOUN
fcis-10735	225	19	.	.	PUNCT
fcis-10735	226	1	4	4	X
fcis-10735	226	2	.	.	X
fcis-10735	226	3	discussion	discussion	NOUN
fcis-10735	226	4	4.1	4.1	NUM
fcis-10735	226	5	.	.	PUNCT
fcis-10735	227	1	random	random	ADJ
fcis-10735	227	2	forest	forest	NOUN
fcis-10735	227	3	algorithm	algorithm	NOUN
fcis-10735	227	4	pros	pro	NOUN
fcis-10735	227	5	:	:	PUNCT
fcis-10735	227	6	random	random	ADJ
fcis-10735	227	7	forest	forest	NOUN
fcis-10735	227	8	can	can	AUX
fcis-10735	227	9	produce	produce	VERB
fcis-10735	227	10	very	very	ADV
fcis-10735	227	11	high	high	ADJ
fcis-10735	227	12	dimensional	dimensional	ADJ
fcis-10735	227	13	data	datum	NOUN
fcis-10735	227	14	.	.	PUNCT
fcis-10735	228	1	it	it	PRON
fcis-10735	228	2	does	do	AUX
fcis-10735	228	3	not	not	PART
fcis-10735	228	4	need	need	VERB
fcis-10735	228	5	to	to	PART
fcis-10735	228	6	reduce	reduce	VERB
fcis-10735	228	7	dimensionality	dimensionality	NOUN
fcis-10735	228	8	and	and	CCONJ
fcis-10735	228	9	it	it	PRON
fcis-10735	228	10	does	do	AUX
fcis-10735	228	11	not	not	PART
fcis-10735	228	12	need	need	VERB
fcis-10735	228	13	to	to	PART
fcis-10735	228	14	do	do	VERB
fcis-10735	228	15	feature	feature	NOUN
fcis-10735	228	16	selection	selection	NOUN
fcis-10735	228	17	.	.	PUNCT
fcis-10735	229	1	therefore	therefore	ADV
fcis-10735	229	2	,	,	PUNCT
fcis-10735	229	3	random	random	ADJ
fcis-10735	229	4	forest	forest	NOUN
fcis-10735	229	5	algorithm	algorithm	NOUN
fcis-10735	229	6	can	can	AUX
fcis-10735	229	7	also	also	ADV
fcis-10735	229	8	be	be	AUX
fcis-10735	229	9	used	use	VERB
fcis-10735	229	10	for	for	ADP
fcis-10735	229	11	image	image	NOUN
fcis-10735	229	12	classification	classification	NOUN
fcis-10735	229	13	.	.	PUNCT
fcis-10735	230	1	random	random	ADJ
fcis-10735	230	2	forest	forest	NOUN
fcis-10735	230	3	is	be	AUX
fcis-10735	230	4	not	not	PART
fcis-10735	230	5	easy	easy	ADJ
fcis-10735	230	6	to	to	PART
fcis-10735	230	7	overfit	overfit	VERB
fcis-10735	230	8	.	.	PUNCT
fcis-10735	231	1	therefore	therefore	ADV
fcis-10735	231	2	,	,	PUNCT
fcis-10735	231	3	users	user	NOUN
fcis-10735	231	4	can	can	AUX
fcis-10735	231	5	continuously	continuously	ADV
fcis-10735	231	6	optimize	optimize	VERB
fcis-10735	231	7	the	the	DET
fcis-10735	231	8	parameters	parameter	NOUN
fcis-10735	231	9	of	of	ADP
fcis-10735	231	10	random	random	ADJ
fcis-10735	231	11	forest	forest	NOUN
fcis-10735	231	12	.	.	PUNCT
fcis-10735	232	1	if	if	SCONJ
fcis-10735	232	2	a	a	DET
fcis-10735	232	3	large	large	ADJ
fcis-10735	232	4	part	part	NOUN
fcis-10735	232	5	of	of	ADP
fcis-10735	232	6	the	the	DET
fcis-10735	232	7	features	feature	NOUN
fcis-10735	232	8	is	be	AUX
fcis-10735	232	9	missing	miss	VERB
fcis-10735	232	10	,	,	PUNCT
fcis-10735	232	11	the	the	DET
fcis-10735	232	12	random	random	ADJ
fcis-10735	232	13	forest	forest	NOUN
fcis-10735	232	14	can	can	AUX
fcis-10735	232	15	still	still	ADV
fcis-10735	232	16	maintain	maintain	VERB
fcis-10735	232	17	the	the	DET
fcis-10735	232	18	accuracy	accuracy	NOUN
fcis-10735	232	19	.	.	PUNCT
fcis-10735	233	1	at	at	ADP
fcis-10735	233	2	the	the	DET
fcis-10735	233	3	same	same	ADJ
fcis-10735	233	4	time	time	NOUN
fcis-10735	233	5	,	,	PUNCT
fcis-10735	233	6	the	the	DET
fcis-10735	233	7	random	random	ADJ
fcis-10735	233	8	forest	forest	NOUN
fcis-10735	233	9	algorithm	algorithm	NOUN
fcis-10735	233	10	can	can	AUX
fcis-10735	233	11	balance	balance	VERB
fcis-10735	233	12	the	the	DET
fcis-10735	233	13	error	error	NOUN
fcis-10735	233	14	for	for	ADP
fcis-10735	233	15	the	the	DET
fcis-10735	233	16	data	datum	NOUN
fcis-10735	233	17	imbalance	imbalance	NOUN
fcis-10735	233	18	.	.	PUNCT
fcis-10735	234	1	random	random	ADJ
fcis-10735	234	2	forest	forest	NOUN
fcis-10735	234	3	is	be	AUX
fcis-10735	234	4	simple	simple	ADJ
fcis-10735	234	5	and	and	CCONJ
fcis-10735	234	6	easy	easy	ADJ
fcis-10735	234	7	to	to	PART
fcis-10735	234	8	use	use	VERB
fcis-10735	234	9	.	.	PUNCT
fcis-10735	235	1	it	it	PRON
fcis-10735	235	2	does	do	AUX
fcis-10735	235	3	n't	not	PART
fcis-10735	235	4	require	require	VERB
fcis-10735	235	5	building	build	VERB
fcis-10735	235	6	a	a	DET
fcis-10735	235	7	model	model	NOUN
fcis-10735	235	8	to	to	PART
fcis-10735	235	9	run	run	VERB
fcis-10735	235	10	and	and	CCONJ
fcis-10735	235	11	it	it	PRON
fcis-10735	235	12	trains	train	VERB
fcis-10735	235	13	well	well	ADV
fcis-10735	235	14	.	.	PUNCT
fcis-10735	236	1	cons	con	NOUN
fcis-10735	236	2	:	:	PUNCT
fcis-10735	236	3	random	random	ADJ
fcis-10735	236	4	forest	forest	NOUN
fcis-10735	236	5	will	will	AUX
fcis-10735	236	6	overfit	overfit	VERB
fcis-10735	236	7	on	on	ADP
fcis-10735	236	8	some	some	DET
fcis-10735	236	9	noisy	noisy	ADJ
fcis-10735	236	10	classification	classification	NOUN
fcis-10735	236	11	or	or	CCONJ
fcis-10735	236	12	regression	regression	NOUN
fcis-10735	236	13	problems	problem	NOUN
fcis-10735	236	14	.	.	PUNCT
fcis-10735	237	1	therefore	therefore	ADV
fcis-10735	237	2	,	,	PUNCT
fcis-10735	237	3	when	when	SCONJ
fcis-10735	237	4	inputting	inputte	VERB
fcis-10735	237	5	images	image	NOUN
fcis-10735	237	6	,	,	PUNCT
fcis-10735	237	7	it	it	PRON
fcis-10735	237	8	is	be	AUX
fcis-10735	237	9	necessary	necessary	ADJ
fcis-10735	237	10	to	to	PART
fcis-10735	237	11	preprocess	preprocess	VERB
fcis-10735	237	12	the	the	DET
fcis-10735	237	13	images	image	NOUN
fcis-10735	237	14	and	and	CCONJ
fcis-10735	237	15	analyze	analyze	VERB
fcis-10735	237	16	the	the	DET
fcis-10735	237	17	input	input	NOUN
fcis-10735	237	18	data	datum	NOUN
fcis-10735	237	19	.	.	PUNCT
fcis-10735	238	1	for	for	ADP
fcis-10735	238	2	datasets	dataset	NOUN
fcis-10735	238	3	with	with	ADP
fcis-10735	238	4	a	a	DET
fcis-10735	238	5	large	large	ADJ
fcis-10735	238	6	amount	amount	NOUN
fcis-10735	238	7	of	of	ADP
fcis-10735	238	8	data	datum	NOUN
fcis-10735	238	9	,	,	PUNCT
fcis-10735	238	10	the	the	DET
fcis-10735	238	11	random	random	ADJ
fcis-10735	238	12	forest	forest	NOUN
fcis-10735	238	13	algorithm	algorithm	NOUN
fcis-10735	238	14	takes	take	VERB
fcis-10735	238	15	a	a	DET
fcis-10735	238	16	long	long	ADJ
fcis-10735	238	17	time	time	NOUN
fcis-10735	238	18	for	for	ADP
fcis-10735	238	19	model	model	NOUN
fcis-10735	238	20	learning	learning	NOUN
fcis-10735	238	21	and	and	CCONJ
fcis-10735	238	22	may	may	AUX
fcis-10735	238	23	not	not	PART
fcis-10735	238	24	output	output	VERB
fcis-10735	238	25	results	result	NOUN
fcis-10735	238	26	for	for	ADP
fcis-10735	238	27	datasets	dataset	NOUN
fcis-10735	238	28	larger	large	ADJ
fcis-10735	238	29	than	than	SCONJ
fcis-10735	238	30	expected	expect	VERB
fcis-10735	238	31	.	.	PUNCT
fcis-10735	239	1	4.2	4.2	NUM
fcis-10735	239	2	.	.	PUNCT
fcis-10735	240	1	logistic	logistic	ADJ
fcis-10735	240	2	regression	regression	NOUN
fcis-10735	240	3	pros	pro	NOUN
fcis-10735	240	4	:	:	PUNCT
fcis-10735	240	5	logistic	logistic	ADJ
fcis-10735	240	6	regression	regression	NOUN
fcis-10735	240	7	classification	classification	NOUN
fcis-10735	240	8	is	be	AUX
fcis-10735	240	9	very	very	ADV
fcis-10735	240	10	computationally	computationally	ADV
fcis-10735	240	11	small	small	ADJ
fcis-10735	240	12	and	and	CCONJ
fcis-10735	240	13	fast	fast	ADJ
fcis-10735	240	14	.	.	PUNCT
fcis-10735	241	1	it	it	PRON
fcis-10735	241	2	uses	use	VERB
fcis-10735	241	3	less	less	ADJ
fcis-10735	241	4	storage	storage	NOUN
fcis-10735	241	5	resources	resource	NOUN
fcis-10735	241	6	.	.	PUNCT
fcis-10735	242	1	logistic	logistic	ADJ
fcis-10735	242	2	regression	regression	NOUN
fcis-10735	242	3	is	be	AUX
fcis-10735	242	4	computationally	computationally	ADV
fcis-10735	242	5	inexpensive	inexpensive	ADJ
fcis-10735	242	6	and	and	CCONJ
fcis-10735	242	7	easy	easy	ADJ
fcis-10735	242	8	to	to	PART
fcis-10735	242	9	implement	implement	VERB
fcis-10735	242	10	.	.	PUNCT
fcis-10735	243	1	it	it	PRON
fcis-10735	243	2	also	also	ADV
fcis-10735	243	3	has	have	VERB
fcis-10735	243	4	a	a	DET
fcis-10735	243	5	handy	handy	ADJ
fcis-10735	243	6	observation	observation	NOUN
fcis-10735	243	7	sample	sample	NOUN
fcis-10735	243	8	probability	probability	NOUN
fcis-10735	243	9	score	score	NOUN
fcis-10735	243	10	.	.	PUNCT
fcis-10735	244	1	logistic	logistic	ADJ
fcis-10735	244	2	regression	regression	NOUN
fcis-10735	244	3	is	be	AUX
fcis-10735	244	4	simple	simple	ADJ
fcis-10735	244	5	to	to	PART
fcis-10735	244	6	implement	implement	VERB
fcis-10735	244	7	and	and	CCONJ
fcis-10735	244	8	widely	widely	ADV
fcis-10735	244	9	used	use	VERB
fcis-10735	244	10	.	.	PUNCT
fcis-10735	245	1	it	it	PRON
fcis-10735	245	2	has	have	VERB
fcis-10735	245	3	numerous	numerous	ADJ
fcis-10735	245	4	improvements	improvement	NOUN
fcis-10735	245	5	.	.	PUNCT
fcis-10735	246	1	cons	con	NOUN
fcis-10735	246	2	:	:	PUNCT
fcis-10735	246	3	logistic	logistic	ADJ
fcis-10735	246	4	regression	regression	NOUN
fcis-10735	246	5	is	be	AUX
fcis-10735	246	6	prone	prone	ADJ
fcis-10735	246	7	to	to	ADP
fcis-10735	246	8	underfitting	underfitte	VERB
fcis-10735	246	9	,	,	PUNCT
fcis-10735	246	10	and	and	CCONJ
fcis-10735	246	11	the	the	DET
fcis-10735	246	12	accuracy	accuracy	NOUN
fcis-10735	246	13	is	be	AUX
fcis-10735	246	14	generally	generally	ADV
fcis-10735	246	15	not	not	PART
fcis-10735	246	16	very	very	ADV
fcis-10735	246	17	high	high	ADJ
fcis-10735	246	18	.	.	PUNCT
fcis-10735	247	1	it	it	PRON
fcis-10735	247	2	may	may	AUX
fcis-10735	247	3	not	not	PART
fcis-10735	247	4	perform	perform	VERB
fcis-10735	247	5	well	well	ADV
fcis-10735	247	6	in	in	ADP
fcis-10735	247	7	using	use	VERB
fcis-10735	247	8	model	model	NOUN
fcis-10735	247	9	predictions	prediction	NOUN
fcis-10735	247	10	.	.	PUNCT
fcis-10735	248	1	logistic	logistic	ADJ
fcis-10735	248	2	regression	regression	NOUN
fcis-10735	248	3	can	can	AUX
fcis-10735	248	4	not	not	PART
fcis-10735	248	5	handle	handle	VERB
fcis-10735	248	6	a	a	DET
fcis-10735	248	7	large	large	ADJ
fcis-10735	248	8	number	number	NOUN
fcis-10735	248	9	of	of	ADP
fcis-10735	248	10	multiclass	multiclass	ADJ
fcis-10735	248	11	features	feature	NOUN
fcis-10735	248	12	or	or	CCONJ
fcis-10735	248	13	variables	variable	NOUN
fcis-10735	248	14	very	very	ADV
fcis-10735	248	15	well	well	ADV
fcis-10735	248	16	.	.	PUNCT
fcis-10735	249	1	for	for	ADP
fcis-10735	249	2	multi	multi	ADJ
fcis-10735	249	3	-	-	ADJ
fcis-10735	249	4	classification	classification	ADJ
fcis-10735	249	5	problems	problem	NOUN
fcis-10735	249	6	,	,	PUNCT
fcis-10735	249	7	softmax	softmax	PROPN
fcis-10735	249	8	is	be	AUX
fcis-10735	249	9	required	require	VERB
fcis-10735	249	10	.	.	PUNCT
fcis-10735	250	1	when	when	SCONJ
fcis-10735	250	2	using	use	VERB
fcis-10735	250	3	logistic	logistic	ADJ
fcis-10735	250	4	regression	regression	NOUN
fcis-10735	250	5	,	,	PUNCT
fcis-10735	250	6	a	a	DET
fcis-10735	250	7	transformation	transformation	NOUN
fcis-10735	250	8	is	be	AUX
fcis-10735	250	9	required	require	VERB
fcis-10735	250	10	for	for	ADP
fcis-10735	250	11	nonlinear	nonlinear	ADJ
fcis-10735	250	12	features	feature	NOUN
fcis-10735	250	13	.	.	PUNCT
fcis-10735	251	1	4.3	4.3	NUM
fcis-10735	251	2	.	.	X
fcis-10735	251	3	resnet	resnet	NOUN
fcis-10735	251	4	and	and	CCONJ
fcis-10735	251	5	swordnet	swordnet	PROPN
fcis-10735	251	6	4.3.1	4.3.1	NUM
fcis-10735	251	7	.	.	PUNCT
fcis-10735	251	8	resnet	resnet	PROPN
fcis-10735	251	9	pros	pro	NOUN
fcis-10735	251	10	:	:	PUNCT
fcis-10735	251	11	even	even	ADV
fcis-10735	251	12	in	in	ADP
fcis-10735	251	13	the	the	DET
fcis-10735	251	14	most	most	ADV
fcis-10735	251	15	extreme	extreme	ADJ
fcis-10735	251	16	case	case	NOUN
fcis-10735	251	17	(	(	PUNCT
fcis-10735	251	18	where	where	SCONJ
fcis-10735	251	19	the	the	DET
fcis-10735	251	20	optimal	optimal	ADJ
fcis-10735	251	21	function	function	NOUN
fcis-10735	251	22	is	be	AUX
fcis-10735	251	23	equal	equal	ADJ
fcis-10735	251	24	to	to	ADP
fcis-10735	251	25	the	the	DET
fcis-10735	251	26	identity	identity	NOUN
fcis-10735	251	27	function	function	NOUN
fcis-10735	251	28	)	)	PUNCT
fcis-10735	251	29	,	,	PUNCT
fcis-10735	251	30	the	the	DET
fcis-10735	251	31	residual	residual	ADJ
fcis-10735	251	32	function	function	NOUN
fcis-10735	251	33	does	do	AUX
fcis-10735	251	34	not	not	PART
fcis-10735	251	35	produce	produce	VERB
fcis-10735	251	36	a	a	DET
fcis-10735	251	37	"	"	PUNCT
fcis-10735	251	38	visible	visible	ADJ
fcis-10735	251	39	"	"	PUNCT
fcis-10735	251	40	degradation	degradation	NOUN
fcis-10735	251	41	.	.	PUNCT
fcis-10735	252	1	since	since	SCONJ
fcis-10735	252	2	the	the	DET
fcis-10735	252	3	weights	weight	NOUN
fcis-10735	252	4	of	of	ADP
fcis-10735	252	5	the	the	DET
fcis-10735	252	6	residual	residual	ADJ
fcis-10735	252	7	function	function	NOUN
fcis-10735	252	8	part	part	NOUN
fcis-10735	252	9	are	be	AUX
fcis-10735	252	10	initialized	initialize	VERB
fcis-10735	252	11	close	close	ADV
fcis-10735	252	12	to	to	ADP
fcis-10735	252	13	zero	zero	NUM
fcis-10735	252	14	,	,	PUNCT
fcis-10735	252	15	they	they	PRON
fcis-10735	252	16	do	do	AUX
fcis-10735	252	17	not	not	PART
fcis-10735	252	18	get	get	VERB
fcis-10735	252	19	a	a	DET
fcis-10735	252	20	large	large	ADJ
fcis-10735	252	21	gradient	gradient	NOUN
fcis-10735	252	22	(	(	PUNCT
fcis-10735	252	23	compared	compare	VERB
fcis-10735	252	24	to	to	ADP
fcis-10735	252	25	1	1	NUM
fcis-10735	252	26	)	)	PUNCT
fcis-10735	252	27	.	.	PUNCT
fcis-10735	253	1	the	the	DET
fcis-10735	253	2	residual	residual	ADJ
fcis-10735	253	3	function	function	NOUN
fcis-10735	253	4	will	will	AUX
fcis-10735	253	5	only	only	ADV
fcis-10735	253	6	undergo	undergo	VERB
fcis-10735	253	7	extremely	extremely	ADV
fcis-10735	253	8	limited	limited	ADJ
fcis-10735	253	9	changes	change	NOUN
fcis-10735	253	10	when	when	SCONJ
fcis-10735	253	11	it	it	PRON
fcis-10735	253	12	is	be	AUX
fcis-10735	253	13	updated	update	VERB
fcis-10735	253	14	,	,	PUNCT
fcis-10735	253	15	and	and	CCONJ
fcis-10735	253	16	the	the	DET
fcis-10735	253	17	impact	impact	NOUN
fcis-10735	253	18	on	on	ADP
fcis-10735	253	19	the	the	DET
fcis-10735	253	20	identity	identity	NOUN
fcis-10735	253	21	function	function	NOUN
fcis-10735	253	22	is	be	AUX
fcis-10735	253	23	almost	almost	ADV
fcis-10735	253	24	negligible	negligible	ADJ
fcis-10735	253	25	.	.	PUNCT
fcis-10735	254	1	with	with	ADP
fcis-10735	254	2	the	the	DET
fcis-10735	254	3	shortcut	shortcut	NOUN
fcis-10735	254	4	connection	connection	NOUN
fcis-10735	254	5	,	,	PUNCT
fcis-10735	254	6	the	the	DET
fcis-10735	254	7	signal	signal	NOUN
fcis-10735	254	8	can	can	AUX
fcis-10735	254	9	be	be	AUX
fcis-10735	254	10	easily	easily	ADV
fcis-10735	254	11	propagated	propagate	VERB
fcis-10735	254	12	throughout	throughout	ADP
fcis-10735	254	13	the	the	DET
fcis-10735	254	14	network	network	NOUN
fcis-10735	254	15	,	,	PUNCT
fcis-10735	254	16	thus	thus	ADV
fcis-10735	254	17	avoiding	avoid	VERB
fcis-10735	254	18	"	"	PUNCT
fcis-10735	254	19	vanishing	vanish	VERB
fcis-10735	254	20	gradient	gradient	NOUN
fcis-10735	254	21	"	"	PUNCT
fcis-10735	254	22	.	.	PUNCT
fcis-10735	255	1	global	global	ADJ
fcis-10735	255	2	average	average	ADJ
fcis-10735	255	3	pooling	pooling	NOUN
fcis-10735	255	4	of	of	ADP
fcis-10735	255	5	resnet	resnet	NOUN
fcis-10735	255	6	can	can	AUX
fcis-10735	255	7	suppress	suppress	VERB
fcis-10735	255	8	overfitting	overfitting	NOUN
fcis-10735	255	9	and	and	CCONJ
fcis-10735	255	10	make	make	VERB
fcis-10735	255	11	input	input	NOUN
fcis-10735	255	12	size	size	NOUN
fcis-10735	255	13	more	more	ADV
fcis-10735	255	14	flexible	flexible	ADJ
fcis-10735	255	15	.	.	PUNCT
fcis-10735	256	1	resnet	resnet	PROPN
fcis-10735	256	2	solves	solve	VERB
fcis-10735	256	3	the	the	DET
fcis-10735	256	4	gradient	gradient	ADJ
fcis-10735	256	5	problem	problem	NOUN
fcis-10735	256	6	to	to	PART
fcis-10735	256	7	build	build	VERB
fcis-10735	256	8	a	a	DET
fcis-10735	256	9	deeper	deep	ADJ
fcis-10735	256	10	neural	neural	ADJ
fcis-10735	256	11	network	network	NOUN
fcis-10735	256	12	model	model	NOUN
fcis-10735	256	13	.	.	PUNCT
fcis-10735	257	1	therefore	therefore	ADV
fcis-10735	257	2	,	,	PUNCT
fcis-10735	257	3	the	the	DET
fcis-10735	257	4	model	model	NOUN
fcis-10735	257	5	can	can	AUX
fcis-10735	257	6	learn	learn	VERB
fcis-10735	257	7	more	more	ADJ
fcis-10735	257	8	abstract	abstract	ADJ
fcis-10735	257	9	features	feature	NOUN
fcis-10735	257	10	,	,	PUNCT
fcis-10735	257	11	and	and	CCONJ
fcis-10735	257	12	the	the	DET
fcis-10735	257	13	prediction	prediction	NOUN
fcis-10735	257	14	results	result	NOUN
fcis-10735	257	15	will	will	AUX
fcis-10735	257	16	be	be	AUX
fcis-10735	257	17	better	well	ADJ
fcis-10735	257	18	.	.	PUNCT
fcis-10735	258	1	cons	con	NOUN
fcis-10735	258	2	:	:	PUNCT
fcis-10735	258	3	resnet	resnet	NOUN
fcis-10735	258	4	training	training	NOUN
fcis-10735	258	5	is	be	AUX
fcis-10735	258	6	slow	slow	ADJ
fcis-10735	258	7	and	and	CCONJ
fcis-10735	258	8	training	training	NOUN
fcis-10735	258	9	time	time	NOUN
fcis-10735	258	10	is	be	AUX
fcis-10735	258	11	long	long	ADJ
fcis-10735	258	12	with	with	ADP
fcis-10735	258	13	the	the	DET
fcis-10735	258	14	same	same	ADJ
fcis-10735	258	15	number	number	NOUN
fcis-10735	258	16	of	of	ADP
fcis-10735	258	17	layers	layer	NOUN
fcis-10735	258	18	,	,	PUNCT
fcis-10735	258	19	resnet	resnet	NOUN
fcis-10735	258	20	training	training	NOUN
fcis-10735	258	21	requires	require	VERB
fcis-10735	258	22	more	more	ADJ
fcis-10735	258	23	computing	computing	NOUN
fcis-10735	258	24	power	power	NOUN
fcis-10735	258	25	than	than	ADP
fcis-10735	258	26	other	other	ADJ
fcis-10735	258	27	models	model	NOUN
fcis-10735	258	28	.	.	PUNCT
fcis-10735	259	1	resnet	resnet	NOUN
fcis-10735	259	2	helps	help	VERB
fcis-10735	259	3	us	we	PRON
fcis-10735	259	4	avoid	avoid	VERB
fcis-10735	259	5	the	the	DET
fcis-10735	259	6	problem	problem	NOUN
fcis-10735	259	7	that	that	SCONJ
fcis-10735	259	8	deep	deep	ADJ
fcis-10735	259	9	models	model	NOUN
fcis-10735	259	10	are	be	AUX
fcis-10735	259	11	difficult	difficult	ADJ
fcis-10735	259	12	to	to	PART
fcis-10735	259	13	optimize	optimize	VERB
fcis-10735	259	14	without	without	ADP
fcis-10735	259	15	really	really	ADV
fcis-10735	259	16	solving	solve	VERB
fcis-10735	259	17	it	it	PRON
fcis-10735	259	18	.	.	PUNCT
fcis-10735	260	1	although	although	SCONJ
fcis-10735	260	2	resnet	resnet	NOUN
fcis-10735	260	3	stacks	stack	VERB
fcis-10735	260	4	many	many	ADJ
fcis-10735	260	5	network	network	NOUN
fcis-10735	260	6	layers	layer	NOUN
fcis-10735	260	7	,	,	PUNCT
fcis-10735	260	8	the	the	DET
fcis-10735	260	9	effective	effective	ADJ
fcis-10735	260	10	depth	depth	NOUN
fcis-10735	260	11	is	be	AUX
fcis-10735	260	12	small	small	ADJ
fcis-10735	260	13	.	.	PUNCT
fcis-10735	261	1	4.3.2	4.3.2	X
fcis-10735	261	2	.	.	PUNCT
fcis-10735	262	1	swordnet	swordnet	PROPN
fcis-10735	262	2	pros	pros	PROPN
fcis-10735	262	3	:	:	PUNCT
fcis-10735	262	4	global	global	ADJ
fcis-10735	262	5	average	average	ADJ
fcis-10735	262	6	pooling	pooling	NOUN
fcis-10735	262	7	of	of	ADP
fcis-10735	262	8	swordnet	swordnet	NOUN
fcis-10735	262	9	can	can	AUX
fcis-10735	262	10	suppress	suppress	VERB
fcis-10735	262	11	overfitting	overfitting	NOUN
fcis-10735	262	12	and	and	CCONJ
fcis-10735	262	13	make	make	VERB
fcis-10735	262	14	input	input	NOUN
fcis-10735	262	15	size	size	NOUN
fcis-10735	262	16	more	more	ADV
fcis-10735	262	17	flexible	flexible	ADJ
fcis-10735	262	18	.	.	PUNCT
fcis-10735	263	1	its	its	PRON
fcis-10735	263	2	model	model	NOUN
fcis-10735	263	3	accuracy	accuracy	NOUN
fcis-10735	263	4	is	be	AUX
fcis-10735	263	5	high	high	ADJ
fcis-10735	263	6	and	and	CCONJ
fcis-10735	263	7	higher	high	ADJ
fcis-10735	263	8	than	than	ADP
fcis-10735	263	9	that	that	PRON
fcis-10735	263	10	of	of	ADP
fcis-10735	263	11	similar	similar	ADJ
fcis-10735	263	12	algorithms	algorithm	NOUN
fcis-10735	263	13	.	.	PUNCT
fcis-10735	264	1	cons	con	NOUN
fcis-10735	264	2	:	:	PUNCT
fcis-10735	264	3	swordnet	swordnet	NOUN
fcis-10735	264	4	model	model	NOUN
fcis-10735	264	5	training	training	NOUN
fcis-10735	264	6	takes	take	VERB
fcis-10735	264	7	a	a	DET
fcis-10735	264	8	long	long	ADJ
fcis-10735	264	9	time	time	NOUN
fcis-10735	264	10	.	.	PUNCT
fcis-10735	265	1	the	the	DET
fcis-10735	265	2	entire	entire	ADJ
fcis-10735	265	3	swordnet	swordnet	NOUN
fcis-10735	265	4	model	model	NOUN
fcis-10735	265	5	has	have	VERB
fcis-10735	265	6	more	more	ADJ
fcis-10735	265	7	layers	layer	NOUN
fcis-10735	265	8	,	,	PUNCT
fcis-10735	265	9	which	which	PRON
fcis-10735	265	10	means	mean	VERB
fcis-10735	265	11	more	more	ADJ
fcis-10735	265	12	time	time	NOUN
fcis-10735	265	13	for	for	ADP
fcis-10735	265	14	model	model	NOUN
fcis-10735	265	15	training	training	NOUN
fcis-10735	265	16	.	.	PUNCT
fcis-10735	266	1	4.4	4.4	NUM
fcis-10735	266	2	.	.	PUNCT
fcis-10735	267	1	pros&cons	pros&con	NOUN
fcis-10735	267	2	of	of	ADP
fcis-10735	267	3	the	the	DET
fcis-10735	267	4	whole	whole	ADJ
fcis-10735	267	5	algorithm	algorithm	NOUN
fcis-10735	267	6	pros	pro	NOUN
fcis-10735	267	7	:	:	PUNCT
fcis-10735	267	8	the	the	DET
fcis-10735	267	9	whole	whole	ADJ
fcis-10735	267	10	algorithm	algorithm	NOUN
fcis-10735	267	11	has	have	VERB
fcis-10735	267	12	a	a	DET
fcis-10735	267	13	fast	fast	ADJ
fcis-10735	267	14	operation	operation	NOUN
fcis-10735	267	15	speed	speed	NOUN
fcis-10735	267	16	and	and	CCONJ
fcis-10735	267	17	a	a	DET
fcis-10735	267	18	high	high	ADJ
fcis-10735	267	19	prediction	prediction	NOUN
fcis-10735	267	20	accuracy	accuracy	NOUN
fcis-10735	267	21	.	.	PUNCT
fcis-10735	268	1	it	it	PRON
fcis-10735	268	2	can	can	AUX
fcis-10735	268	3	perform	perform	VERB
fcis-10735	268	4	real	real	ADJ
fcis-10735	268	5	-	-	PUNCT
fcis-10735	268	6	time	time	NOUN
fcis-10735	268	7	text	text	NOUN
fcis-10735	268	8	style	style	NOUN
fcis-10735	268	9	inference	inference	NOUN
fcis-10735	268	10	on	on	ADP
fcis-10735	268	11	the	the	DET
fcis-10735	268	12	text	text	NOUN
fcis-10735	268	13	in	in	ADP
fcis-10735	268	14	the	the	DET
fcis-10735	268	15	input	input	NOUN
fcis-10735	268	16	image	image	NOUN
fcis-10735	268	17	.	.	PUNCT
fcis-10735	269	1	the	the	DET
fcis-10735	269	2	amount	amount	NOUN
fcis-10735	269	3	of	of	ADP
fcis-10735	269	4	data	datum	NOUN
fcis-10735	269	5	required	require	VERB
fcis-10735	269	6	by	by	ADP
fcis-10735	269	7	the	the	DET
fcis-10735	269	8	algorithm	algorithm	NOUN
fcis-10735	269	9	is	be	AUX
fcis-10735	269	10	small	small	ADJ
fcis-10735	269	11	.	.	PUNCT
fcis-10735	270	1	this	this	PRON
fcis-10735	270	2	means	mean	VERB
fcis-10735	270	3	that	that	SCONJ
fcis-10735	270	4	the	the	DET
fcis-10735	270	5	algorithm	algorithm	NOUN
fcis-10735	270	6	is	be	AUX
fcis-10735	270	7	also	also	ADV
fcis-10735	270	8	very	very	ADV
fcis-10735	270	9	adaptable	adaptable	ADJ
fcis-10735	270	10	to	to	ADP
fcis-10735	270	11	the	the	DET
fcis-10735	270	12	lack	lack	NOUN
fcis-10735	270	13	of	of	ADP
fcis-10735	270	14	data	datum	NOUN
fcis-10735	270	15	.	.	PUNCT
fcis-10735	271	1	the	the	DET
fcis-10735	271	2	algorithm	algorithm	NOUN
fcis-10735	271	3	can	can	AUX
fcis-10735	271	4	not	not	PART
fcis-10735	271	5	only	only	ADV
fcis-10735	271	6	recognize	recognize	VERB
fcis-10735	271	7	the	the	DET
fcis-10735	271	8	font	font	NOUN
fcis-10735	271	9	style	style	NOUN
fcis-10735	271	10	but	but	CCONJ
fcis-10735	271	11	also	also	ADV
fcis-10735	271	12	recognize	recognize	VERB
fcis-10735	271	13	the	the	DET
fcis-10735	271	14	text	text	NOUN
fcis-10735	271	15	content	content	NOUN
fcis-10735	271	16	.	.	PUNCT
fcis-10735	272	1	cons	con	NOUN
fcis-10735	272	2	:	:	PUNCT
fcis-10735	272	3	the	the	DET
fcis-10735	272	4	algorithm	algorithm	NOUN
fcis-10735	272	5	may	may	AUX
fcis-10735	272	6	not	not	PART
fcis-10735	272	7	have	have	VERB
fcis-10735	272	8	good	good	ADJ
fcis-10735	272	9	prediction	prediction	NOUN
fcis-10735	272	10	results	result	NOUN
fcis-10735	272	11	for	for	ADP
fcis-10735	272	12	some	some	DET
fcis-10735	272	13	languages	language	NOUN
fcis-10735	272	14	such	such	ADJ
fcis-10735	272	15	as	as	ADP
fcis-10735	272	16	traditional	traditional	ADJ
fcis-10735	272	17	chinese	chinese	PROPN
fcis-10735	272	18	.	.	PUNCT
fcis-10735	273	1	the	the	DET
fcis-10735	273	2	whole	whole	ADJ
fcis-10735	273	3	algorithm	algorithm	NOUN
fcis-10735	273	4	uses	use	VERB
fcis-10735	273	5	three	three	NUM
fcis-10735	273	6	different	different	ADJ
fcis-10735	273	7	methods	method	NOUN
fcis-10735	273	8	which	which	PRON
fcis-10735	273	9	may	may	AUX
fcis-10735	273	10	result	result	VERB
fcis-10735	273	11	in	in	ADP
fcis-10735	273	12	more	more	ADJ
fcis-10735	273	13	models	model	NOUN
fcis-10735	273	14	.	.	PUNCT
fcis-10735	274	1	4.5	4.5	NUM
fcis-10735	274	2	.	.	PUNCT
fcis-10735	275	1	problems	problem	NOUN
fcis-10735	275	2	4.5.1	4.5.1	NUM
fcis-10735	275	3	.	.	PUNCT
fcis-10735	276	1	data	data	NOUN
fcis-10735	276	2	processing	processing	NOUN
fcis-10735	276	3	issues	issue	NOUN
fcis-10735	276	4	improperly	improperly	ADV
fcis-10735	276	5	formatted	format	VERB
fcis-10735	276	6	input	input	NOUN
fcis-10735	276	7	of	of	ADP
fcis-10735	276	8	data	datum	NOUN
fcis-10735	276	9	in	in	ADP
fcis-10735	276	10	data	datum	NOUN
fcis-10735	276	11	processing	processing	NOUN
fcis-10735	276	12	may	may	AUX
fcis-10735	276	13	cause	cause	VERB
fcis-10735	276	14	training	training	NOUN
fcis-10735	276	15	results	result	NOUN
fcis-10735	276	16	not	not	PART
fcis-10735	276	17	to	to	PART
fcis-10735	276	18	be	be	AUX
fcis-10735	276	19	as	as	SCONJ
fcis-10735	276	20	expected	expect	VERB
fcis-10735	276	21	.	.	PUNCT
fcis-10735	277	1	4.5.2	4.5.2	NUM
fcis-10735	277	2	.	.	PUNCT
fcis-10735	277	3	modification	modification	NOUN
fcis-10735	277	4	and	and	CCONJ
fcis-10735	277	5	optimization	optimization	NOUN
fcis-10735	277	6	of	of	ADP
fcis-10735	277	7	model	model	NOUN
fcis-10735	277	8	parameters	parameter	NOUN
fcis-10735	277	9	during	during	ADP
fcis-10735	277	10	parameter	parameter	NOUN
fcis-10735	277	11	tuning	tuning	NOUN
fcis-10735	277	12	,	,	PUNCT
fcis-10735	277	13	each	each	DET
fcis-10735	277	14	model	model	NOUN
fcis-10735	277	15	has	have	VERB
fcis-10735	277	16	many	many	ADJ
fcis-10735	277	17	parameters	parameter	NOUN
fcis-10735	277	18	.	.	PUNCT
fcis-10735	278	1	optimizing	optimize	VERB
fcis-10735	278	2	parameters	parameter	NOUN
fcis-10735	278	3	requires	require	VERB
fcis-10735	278	4	code	code	NOUN
fcis-10735	278	5	to	to	PART
fcis-10735	278	6	traverse	traverse	VERB
fcis-10735	278	7	the	the	DET
fcis-10735	278	8	possible	possible	ADJ
fcis-10735	278	9	situations	situation	NOUN
fcis-10735	278	10	in	in	ADP
fcis-10735	278	11	the	the	DET
fcis-10735	278	12	parametric	parametric	ADJ
fcis-10735	278	13	model	model	NOUN
fcis-10735	278	14	.	.	PUNCT
fcis-10735	279	1	this	this	PRON
fcis-10735	279	2	will	will	AUX
fcis-10735	279	3	consume	consume	VERB
fcis-10735	279	4	a	a	DET
fcis-10735	279	5	lot	lot	NOUN
fcis-10735	279	6	of	of	ADP
fcis-10735	279	7	16	16	NUM
fcis-10735	279	8	time	time	NOUN
fcis-10735	279	9	,	,	PUNCT
fcis-10735	279	10	and	and	CCONJ
fcis-10735	279	11	some	some	DET
fcis-10735	279	12	parameters	parameter	NOUN
fcis-10735	279	13	have	have	VERB
fcis-10735	279	14	less	less	ADJ
fcis-10735	279	15	influence	influence	NOUN
fcis-10735	279	16	on	on	ADP
fcis-10735	279	17	model	model	NOUN
fcis-10735	279	18	optimization	optimization	NOUN
fcis-10735	279	19	.	.	PUNCT
fcis-10735	280	1	5	5	X
fcis-10735	280	2	.	.	X
fcis-10735	280	3	conclusion	conclusion	NOUN
fcis-10735	280	4	i	i	PRON
fcis-10735	280	5	built	build	VERB
fcis-10735	280	6	this	this	DET
fcis-10735	280	7	real	real	ADJ
fcis-10735	280	8	-	-	PUNCT
fcis-10735	280	9	time	time	NOUN
fcis-10735	280	10	font	font	NOUN
fcis-10735	280	11	style	style	NOUN
fcis-10735	280	12	recognition	recognition	NOUN
fcis-10735	280	13	algorithm	algorithm	NOUN
fcis-10735	280	14	based	base	VERB
fcis-10735	280	15	on	on	ADP
fcis-10735	280	16	resnet	resnet	NOUN
fcis-10735	280	17	,	,	PUNCT
fcis-10735	280	18	swordnet	swordnet	NOUN
fcis-10735	280	19	,	,	PUNCT
fcis-10735	280	20	logistic	logistic	ADJ
fcis-10735	280	21	regression	regression	NOUN
fcis-10735	280	22	and	and	CCONJ
fcis-10735	280	23	random	random	ADJ
fcis-10735	280	24	forest	forest	NOUN
fcis-10735	280	25	.	.	PUNCT
fcis-10735	281	1	users	user	NOUN
fcis-10735	281	2	can	can	AUX
fcis-10735	281	3	input	input	VERB
fcis-10735	281	4	multiple	multiple	ADJ
fcis-10735	281	5	images	image	NOUN
fcis-10735	281	6	of	of	ADP
fcis-10735	281	7	any	any	DET
fcis-10735	281	8	size	size	NOUN
fcis-10735	281	9	and	and	CCONJ
fcis-10735	281	10	get	get	VERB
fcis-10735	281	11	a	a	DET
fcis-10735	281	12	final	final	ADJ
fcis-10735	281	13	font	font	NOUN
fcis-10735	281	14	style	style	NOUN
fcis-10735	281	15	prediction	prediction	NOUN
fcis-10735	281	16	result	result	NOUN
fcis-10735	281	17	.	.	PUNCT
fcis-10735	282	1	in	in	ADP
fcis-10735	282	2	the	the	DET
fcis-10735	282	3	end	end	NOUN
fcis-10735	282	4	,	,	PUNCT
fcis-10735	282	5	the	the	DET
fcis-10735	282	6	accuracy	accuracy	NOUN
fcis-10735	282	7	of	of	ADP
fcis-10735	282	8	the	the	DET
fcis-10735	282	9	prediction	prediction	NOUN
fcis-10735	282	10	results	result	NOUN
fcis-10735	282	11	of	of	ADP
fcis-10735	282	12	resnet	resnet	NOUN
fcis-10735	282	13	and	and	CCONJ
fcis-10735	282	14	swordnet	swordnet	NOUN
fcis-10735	282	15	is	be	AUX
fcis-10735	282	16	above	above	ADP
fcis-10735	282	17	0.93	0.93	NUM
fcis-10735	282	18	.	.	PUNCT
fcis-10735	283	1	the	the	DET
fcis-10735	283	2	accuracy	accuracy	NOUN
fcis-10735	283	3	of	of	ADP
fcis-10735	283	4	the	the	DET
fcis-10735	283	5	prediction	prediction	NOUN
fcis-10735	283	6	results	result	NOUN
fcis-10735	283	7	of	of	ADP
fcis-10735	283	8	logistic	logistic	ADJ
fcis-10735	283	9	regression	regression	NOUN
fcis-10735	283	10	is	be	AUX
fcis-10735	283	11	0.71	0.71	NUM
fcis-10735	283	12	and	and	CCONJ
fcis-10735	283	13	the	the	DET
fcis-10735	283	14	prediction	prediction	NOUN
fcis-10735	283	15	result	result	NOUN
fcis-10735	283	16	of	of	ADP
fcis-10735	283	17	random	random	ADJ
fcis-10735	283	18	forestd	forestd	NOUN
fcis-10735	283	19	is	be	AUX
fcis-10735	283	20	0.99	0.99	NUM
fcis-10735	283	21	.	.	PUNCT
fcis-10735	284	1	finally	finally	ADV
fcis-10735	284	2	,	,	PUNCT
fcis-10735	284	3	the	the	DET
fcis-10735	284	4	accuracy	accuracy	NOUN
fcis-10735	284	5	of	of	ADP
fcis-10735	284	6	the	the	DET
fcis-10735	284	7	prediction	prediction	NOUN
fcis-10735	284	8	results	result	NOUN
fcis-10735	284	9	of	of	ADP
fcis-10735	284	10	the	the	DET
fcis-10735	284	11	entire	entire	ADJ
fcis-10735	284	12	model	model	NOUN
fcis-10735	284	13	is	be	AUX
fcis-10735	284	14	above	above	ADP
fcis-10735	284	15	0.90	0.90	NUM
fcis-10735	284	16	.	.	PUNCT
fcis-10735	285	1	however	however	ADV
fcis-10735	285	2	,	,	PUNCT
fcis-10735	285	3	this	this	DET
fcis-10735	285	4	font	font	NOUN
fcis-10735	285	5	style	style	NOUN
fcis-10735	285	6	recognizer	recognizer	NOUN
fcis-10735	285	7	is	be	AUX
fcis-10735	285	8	not	not	PART
fcis-10735	285	9	finalized	finalize	VERB
fcis-10735	285	10	.	.	PUNCT
fcis-10735	286	1	i	i	PRON
fcis-10735	286	2	hope	hope	VERB
fcis-10735	286	3	to	to	PART
fcis-10735	286	4	build	build	VERB
fcis-10735	286	5	a	a	DET
fcis-10735	286	6	model	model	NOUN
fcis-10735	286	7	that	that	PRON
fcis-10735	286	8	can	can	AUX
fcis-10735	286	9	classify	classify	VERB
fcis-10735	286	10	all	all	DET
fcis-10735	286	11	font	font	NOUN
fcis-10735	286	12	styles	style	NOUN
fcis-10735	286	13	.	.	PUNCT
fcis-10735	287	1	at	at	ADP
fcis-10735	287	2	the	the	DET
fcis-10735	287	3	same	same	ADJ
fcis-10735	287	4	time	time	NOUN
fcis-10735	287	5	,	,	PUNCT
fcis-10735	287	6	although	although	SCONJ
fcis-10735	287	7	resnet	resnet	NOUN
fcis-10735	287	8	and	and	CCONJ
fcis-10735	287	9	swordnet	swordnet	NOUN
fcis-10735	287	10	have	have	AUX
fcis-10735	287	11	completed	complete	VERB
fcis-10735	287	12	the	the	DET
fcis-10735	287	13	deep	deep	ADJ
fcis-10735	287	14	neural	neural	ADJ
fcis-10735	287	15	network	network	NOUN
fcis-10735	287	16	,	,	PUNCT
fcis-10735	287	17	their	their	PRON
fcis-10735	287	18	learning	learning	NOUN
fcis-10735	287	19	ability	ability	NOUN
fcis-10735	287	20	is	be	AUX
fcis-10735	287	21	not	not	PART
fcis-10735	287	22	high	high	ADJ
fcis-10735	287	23	.	.	PUNCT
fcis-10735	288	1	i	i	PRON
fcis-10735	288	2	hope	hope	VERB
fcis-10735	288	3	that	that	SCONJ
fcis-10735	288	4	the	the	DET
fcis-10735	288	5	entire	entire	ADJ
fcis-10735	288	6	neural	neural	ADJ
fcis-10735	288	7	network	network	NOUN
fcis-10735	288	8	can	can	AUX
fcis-10735	288	9	be	be	AUX
fcis-10735	288	10	further	far	ADV
fcis-10735	288	11	optimized	optimize	VERB
fcis-10735	288	12	to	to	PART
fcis-10735	288	13	find	find	VERB
fcis-10735	288	14	a	a	DET
fcis-10735	288	15	neural	neural	ADJ
fcis-10735	288	16	network	network	NOUN
fcis-10735	288	17	model	model	NOUN
fcis-10735	288	18	that	that	PRON
fcis-10735	288	19	can	can	AUX
fcis-10735	288	20	solve	solve	VERB
fcis-10735	288	21	the	the	DET
fcis-10735	288	22	gradient	gradient	ADJ
fcis-10735	288	23	problem	problem	NOUN
fcis-10735	288	24	.	.	PUNCT
fcis-10735	289	1	for	for	ADP
fcis-10735	289	2	the	the	DET
fcis-10735	289	3	field	field	NOUN
fcis-10735	289	4	of	of	ADP
fcis-10735	289	5	font	font	NOUN
fcis-10735	289	6	style	style	NOUN
fcis-10735	289	7	recognition	recognition	NOUN
fcis-10735	289	8	,	,	PUNCT
fcis-10735	289	9	i	i	PRON
fcis-10735	289	10	hope	hope	VERB
fcis-10735	289	11	that	that	SCONJ
fcis-10735	289	12	scientists	scientist	NOUN
fcis-10735	289	13	it	it	PRON
fcis-10735	289	14	is	be	AUX
fcis-10735	289	15	possible	possible	ADJ
fcis-10735	289	16	to	to	PART
fcis-10735	289	17	continue	continue	VERB
fcis-10735	289	18	to	to	PART
fcis-10735	289	19	develop	develop	VERB
fcis-10735	289	20	font	font	NOUN
fcis-10735	289	21	style	style	NOUN
fcis-10735	289	22	recognition	recognition	NOUN
fcis-10735	289	23	technology	technology	NOUN
fcis-10735	289	24	and	and	CCONJ
fcis-10735	289	25	continue	continue	VERB
fcis-10735	289	26	to	to	PART
fcis-10735	289	27	study	study	VERB
fcis-10735	289	28	font	font	NOUN
fcis-10735	289	29	recognition	recognition	NOUN
fcis-10735	289	30	models	model	NOUN
fcis-10735	289	31	.	.	PUNCT
fcis-10735	290	1	in	in	ADP
fcis-10735	290	2	this	this	DET
fcis-10735	290	3	way	way	NOUN
fcis-10735	290	4	,	,	PUNCT
fcis-10735	290	5	people	people	NOUN
fcis-10735	290	6	can	can	AUX
fcis-10735	290	7	recognize	recognize	VERB
fcis-10735	290	8	ancient	ancient	ADJ
fcis-10735	290	9	texts	text	NOUN
fcis-10735	290	10	and	and	CCONJ
fcis-10735	290	11	have	have	VERB
fcis-10735	290	12	an	an	DET
fcis-10735	290	13	impact	impact	NOUN
fcis-10735	290	14	on	on	ADP
fcis-10735	290	15	future	future	ADJ
fcis-10735	290	16	text	text	NOUN
fcis-10735	290	17	development	development	NOUN
fcis-10735	290	18	.	.	PUNCT
fcis-10735	291	1	references	reference	NOUN
fcis-10735	291	2	[	[	X
fcis-10735	291	3	1	1	NUM
fcis-10735	291	4	]	]	PUNCT
fcis-10735	291	5	a.	a.	NOUN
fcis-10735	291	6	bozkurt	bozkurt	PROPN
fcis-10735	291	7	,	,	PUNCT
fcis-10735	291	8	(	(	PUNCT
fcis-10735	291	9	2013	2013	NUM
fcis-10735	291	10	,	,	PUNCT
fcis-10735	291	11	nov	nov	PROPN
fcis-10735	291	12	.	.	PROPN
fcis-10735	291	13	19	19	NUM
fcis-10735	291	14	)	)	PUNCT
fcis-10735	291	15	.	.	PUNCT
fcis-10735	292	1	font	font	NOUN
fcis-10735	292	2	datasets	dataset	NOUN
fcis-10735	292	3	used	use	VERB
fcis-10735	292	4	in	in	ADP
fcis-10735	292	5	"	"	PUNCT
fcis-10735	292	6	font	font	NOUN
fcis-10735	292	7	and	and	CCONJ
fcis-10735	292	8	calligraphy	calligraphy	NOUN
fcis-10735	292	9	style	style	NOUN
fcis-10735	292	10	recognition	recognition	NOUN
fcis-10735	292	11	using	use	VERB
fcis-10735	292	12	complex	complex	ADJ
fcis-10735	292	13	wavelet	wavelet	NOUN
fcis-10735	292	14	transform	transform	NOUN
fcis-10735	292	15	"	"	PUNCT
fcis-10735	293	1	[	[	X
fcis-10735	293	2	online	online	X
fcis-10735	293	3	]	]	X
fcis-10735	293	4	.	.	PUNCT
fcis-10735	294	1	available	available	ADJ
fcis-10735	294	2	:	:	PUNCT
fcis-10735	294	3	https://	https://	PROPN
fcis-10735	294	4	github	github	NOUN
fcis-10735	294	5	.	.	PUNCT
fcis-10735	295	1	com/	com/	NOUN
fcis-10735	295	2	alicanb/	alicanb/	NUM
fcis-10735	295	3	fonts	font	NOUN
fcis-10735	295	4	.	.	PUNCT
fcis-10735	296	1	[	[	X
fcis-10735	296	2	2	2	NUM
fcis-10735	296	3	]	]	X
fcis-10735	296	4	acron	acron	NOUN
fcis-10735	296	5	(	(	PUNCT
fcis-10735	296	6	2018	2018	NUM
fcis-10735	296	7	)	)	PUNCT
fcis-10735	296	8	australian	australian	ADJ
fcis-10735	296	9	cybercrime	cybercrime	NOUN
fcis-10735	296	10	online	online	ADJ
fcis-10735	296	11	reporting	reporting	NOUN
fcis-10735	296	12	network	network	NOUN
fcis-10735	296	13	(	(	PUNCT
fcis-10735	296	14	acorn	acorn	NOUN
fcis-10735	296	15	)	)	PUNCT
fcis-10735	297	1	[	[	X
fcis-10735	297	2	online	online	X
fcis-10735	297	3	]	]	X
fcis-10735	297	4	.	.	PUNCT
fcis-10735	298	1	available	available	ADJ
fcis-10735	298	2	:	:	PUNCT
fcis-10735	299	1	http://www	http://www	PROPN
fcis-10735	299	2	.	.	PROPN
fcis-10735	299	3	acorn	acorn	PROPN
fcis-10735	299	4	.	.	PUNCT
fcis-10735	300	1	gov	gov	PROPN
fcis-10735	300	2	.	.	PROPN
fcis-10735	300	3	au	au	PROPN
fcis-10735	300	4	/	/	SYM
fcis-10735	300	5	learn	learn	VERB
fcis-10735	300	6	-	-	PUNCT
fcis-10735	300	7	about	about	ADP
fcis-10735	300	8	-	-	PUNCT
fcis-10735	300	9	cybercrime	cybercrime	NOUN
fcis-10735	300	10	.	.	PUNCT
fcis-10735	301	1	[	[	X
fcis-10735	301	2	3	3	NUM
fcis-10735	301	3	]	]	PUNCT
fcis-10735	301	4	a.	a.	NOUN
fcis-10735	301	5	biswal	biswal	NOUN
fcis-10735	301	6	,	,	PUNCT
fcis-10735	301	7	(	(	PUNCT
fcis-10735	301	8	2021	2021	NUM
fcis-10735	301	9	,	,	PUNCT
fcis-10735	301	10	sep	sep	PROPN
fcis-10735	301	11	.	.	PROPN
fcis-10735	301	12	14	14	NUM
fcis-10735	301	13	)	)	PUNCT
fcis-10735	301	14	.	.	PUNCT
fcis-10735	302	1	bagging	bag	VERB
fcis-10735	302	2	in	in	ADP
fcis-10735	302	3	machine	machine	NOUN
fcis-10735	302	4	learning	learning	NOUN
fcis-10735	302	5	:	:	PUNCT
fcis-10735	302	6	step	step	VERB
fcis-10735	302	7	to	to	PART
fcis-10735	302	8	perform	perform	VERB
fcis-10735	302	9	and	and	CCONJ
fcis-10735	302	10	its	its	PRON
fcis-10735	302	11	advantages	advantage	NOUN
fcis-10735	302	12	[	[	X
fcis-10735	302	13	online	online	X
fcis-10735	302	14	]	]	X
fcis-10735	302	15	.	.	PUNCT
fcis-10735	303	1	available	available	ADJ
fcis-10735	303	2	:	:	PUNCT
fcis-10735	303	3	https://	https://	PROPN
fcis-10735	303	4	www.simplilearn.com/tutorials/machine-learning-tutorial/	www.simplilearn.com/tutorials/machine-learning-tutorial/	PROPN
fcis-10735	303	5	bagging	bagging	NOUN
fcis-10735	303	6	-	-	PUNCT
fcis-10735	303	7	in	in	ADP
fcis-10735	303	8	-	-	PUNCT
fcis-10735	303	9	machine	machine	NOUN
fcis-10735	303	10	-	-	PUNCT
fcis-10735	303	11	learning	learning	NOUN
fcis-10735	303	12	.	.	PUNCT
fcis-10735	304	1	[	[	X
fcis-10735	304	2	4	4	NUM
fcis-10735	304	3	]	]	PUNCT
fcis-10735	304	4	a.	a.	NOUN
fcis-10735	304	5	ze	ze	PROPN
fcis-10735	304	6	,	,	PUNCT
fcis-10735	304	7	(	(	PUNCT
fcis-10735	304	8	2021	2021	NUM
fcis-10735	304	9	,	,	PUNCT
fcis-10735	304	10	jan	jan	PROPN
fcis-10735	304	11	.	.	PROPN
fcis-10735	304	12	20	20	NUM
fcis-10735	304	13	)	)	PUNCT
fcis-10735	304	14	.	.	PUNCT
fcis-10735	305	1	[	[	X
fcis-10735	305	2	machine	machine	NOUN
fcis-10735	305	3	learning	learning	NOUN
fcis-10735	305	4	]	]	PUNCT
fcis-10735	305	5	logistic	logistic	ADJ
fcis-10735	305	6	regression	regression	NOUN
fcis-10735	305	7	(	(	PUNCT
fcis-10735	305	8	very	very	ADV
fcis-10735	305	9	detailed	detailed	ADJ
fcis-10735	305	10	)	)	PUNCT
fcis-10735	306	1	[	[	X
fcis-10735	306	2	online	online	X
fcis-10735	306	3	]	]	X
fcis-10735	306	4	.	.	PUNCT
fcis-10735	307	1	available	available	ADJ
fcis-10735	307	2	:	:	PUNCT
fcis-10735	308	1	https://zhuanlan	https://zhuanlan	PROPN
fcis-10735	308	2	.	.	PUNCT
fcis-10735	308	3	zhihu	zhihu	PROPN
fcis-10735	308	4	.	.	PUNCT
fcis-10735	309	1	com/	com/	PROPN
fcis-10735	309	2	p/74874291	p/74874291	PROPN
fcis-10735	309	3	.	.	PUNCT
fcis-10735	310	1	[	[	X
fcis-10735	310	2	5	5	NUM
fcis-10735	310	3	]	]	X
fcis-10735	310	4	s.h	s.h	PROPN
fcis-10735	310	5	.	.	PROPN
fcis-10735	310	6	tsang	tsang	PROPN
fcis-10735	310	7	,	,	PUNCT
fcis-10735	310	8	(	(	PUNCT
fcis-10735	310	9	2018	2018	NUM
fcis-10735	310	10	,	,	PUNCT
fcis-10735	310	11	sep	sep	PROPN
fcis-10735	310	12	.	.	PROPN
fcis-10735	310	13	16	16	NUM
fcis-10735	310	14	)	)	PUNCT
fcis-10735	310	15	.	.	PUNCT
fcis-10735	311	1	review	review	NOUN
fcis-10735	311	2	:	:	PUNCT
fcis-10735	311	3	resnet	resnet	PROPN
fcis-10735	311	4	--winner	--winner	PUNCT
fcis-10735	311	5	of	of	ADP
fcis-10735	311	6	ilsvrc	ilsvrc	PROPN
fcis-10735	311	7	2015	2015	NUM
fcis-10735	311	8	(	(	PUNCT
fcis-10735	311	9	image	image	NOUN
fcis-10735	311	10	classification	classification	NOUN
fcis-10735	311	11	,	,	PUNCT
fcis-10735	311	12	localization	localization	NOUN
fcis-10735	311	13	,	,	PUNCT
fcis-10735	311	14	detection	detection	NOUN
fcis-10735	311	15	)	)	PUNCT
fcis-10735	312	1	[	[	X
fcis-10735	312	2	online	online	X
fcis-10735	312	3	]	]	X
fcis-10735	312	4	.	.	PUNCT
fcis-10735	313	1	available	available	ADJ
fcis-10735	313	2	:	:	PUNCT
fcis-10735	313	3	https://towardsdatascience.com/reviewresnet-winner-of-ilsvrc-2015-image-classification-localization	https://towardsdatascience.com/reviewresnet-winner-of-ilsvrc-2015-image-classification-localization	NOUN
fcis-10735	313	4	-detection	-detection	NOUN
fcis-10735	313	5	-	-	PUNCT
fcis-10735	313	6	e39402bfa5d8	e39402bfa5d8	NUM
fcis-10735	313	7	.	.	PUNCT
fcis-10735	314	1	[	[	X
fcis-10735	314	2	6	6	NUM
fcis-10735	314	3	]	]	PUNCT
fcis-10735	314	4	x.	x.	NOUN
fcis-10735	314	5	li	li	PROPN
fcis-10735	314	6	,	,	PUNCT
fcis-10735	314	7	et.al	et.al	PROPN
fcis-10735	314	8	,	,	PUNCT
fcis-10735	314	9	“	"	PUNCT
fcis-10735	314	10	swordnet	swordnet	NOUN
fcis-10735	314	11	:	:	PUNCT
fcis-10735	314	12	chinese	chinese	ADJ
fcis-10735	314	13	character	character	NOUN
fcis-10735	314	14	font	font	VERB
fcis-10735	314	15	style	style	NOUN
fcis-10735	314	16	recognition	recognition	NOUN
fcis-10735	314	17	network	network	NOUN
fcis-10735	314	18	”	"	PUNCT
fcis-10735	314	19	tianjin	tianjin	NOUN
fcis-10735	314	20	,	,	PUNCT
fcis-10735	314	21	2016	2016	NUM
fcis-10735	314	22	,	,	PUNCT
fcis-10735	314	23	pp.1	pp.1	NOUN
fcis-10735	314	24	-	-	PUNCT
fcis-10735	314	25	11	11	NUM
fcis-10735	314	26	.	.	PUNCT
fcis-10735	315	1	[	[	X
fcis-10735	315	2	7	7	NUM
fcis-10735	315	3	]	]	X
fcis-10735	315	4	wangcy	wangcy	NOUN
fcis-10735	315	5	,	,	PUNCT
fcis-10735	315	6	(	(	PUNCT
fcis-10735	315	7	2016	2016	NUM
fcis-10735	315	8	,	,	PUNCT
fcis-10735	315	9	dec	dec	PROPN
fcis-10735	315	10	.	.	PROPN
fcis-10735	315	11	14	14	NUM
fcis-10735	315	12	)	)	PUNCT
fcis-10735	315	13	.	.	PUNCT
fcis-10735	316	1	maxpool	maxpool	NOUN
fcis-10735	316	2	layer	layer	NOUN
fcis-10735	317	1	[	[	X
fcis-10735	317	2	online	online	X
fcis-10735	317	3	]	]	X
fcis-10735	317	4	.	.	PUNCT
fcis-10735	318	1	available	available	ADJ
fcis-10735	318	2	:	:	PUNCT
fcis-10735	318	3	https	https	NOUN
fcis-10735	318	4	:	:	PUNCT
fcis-10735	318	5	//	//	SYM
fcis-10735	318	6	www.jianshu.com/p/6928203bf75b	www.jianshu.com/p/6928203bf75b	PROPN
fcis-10735	318	7	.	.	PUNCT
fcis-10735	319	1	[	[	X
fcis-10735	319	2	8	8	NUM
fcis-10735	319	3	]	]	X
fcis-10735	319	4	l.	l.	PROPN
fcis-10735	319	5	han	han	PROPN
fcis-10735	319	6	,	,	PUNCT
fcis-10735	319	7	et.al	et.al	PROPN
fcis-10735	319	8	,	,	PUNCT
fcis-10735	319	9	“	"	PUNCT
fcis-10735	319	10	a	a	DET
fcis-10735	319	11	new	new	ADJ
fcis-10735	319	12	method	method	NOUN
fcis-10735	319	13	of	of	ADP
fcis-10735	319	14	mixed	mixed	ADJ
fcis-10735	319	15	gas	gas	NOUN
fcis-10735	319	16	identification	identification	NOUN
fcis-10735	319	17	based	base	VERB
fcis-10735	319	18	on	on	ADP
fcis-10735	319	19	a	a	DET
fcis-10735	319	20	convolutional	convolutional	ADJ
fcis-10735	319	21	neural	neural	ADJ
fcis-10735	319	22	network	network	NOUN
fcis-10735	319	23	for	for	ADP
fcis-10735	319	24	time	time	NOUN
fcis-10735	319	25	series	series	PROPN
fcis-10735	319	26	classification	classification	PROPN
fcis-10735	319	27	”	"	PUNCT
fcis-10735	319	28	beijing	beijing	PROPN
fcis-10735	319	29	,	,	PUNCT
fcis-10735	319	30	2019	2019	NUM
fcis-10735	319	31	,	,	PUNCT
fcis-10735	319	32	pp.9	pp.9	NOUN
fcis-10735	319	33	-	-	SYM
fcis-10735	319	34	9	9	NUM
fcis-10735	319	35	.	.	PUNCT
fcis-10735	320	1	[	[	X
fcis-10735	320	2	9	9	NUM
fcis-10735	320	3	]	]	X
fcis-10735	320	4	yalesaleng	yalesaleng	NOUN
fcis-10735	320	5	,	,	PUNCT
fcis-10735	320	6	(	(	PUNCT
fcis-10735	320	7	2018	2018	NUM
fcis-10735	320	8	,	,	PUNCT
fcis-10735	320	9	jul	jul	PROPN
fcis-10735	320	10	.	.	PROPN
fcis-10735	320	11	13	13	NUM
fcis-10735	320	12	)	)	PUNCT
fcis-10735	320	13	.	.	PUNCT
fcis-10735	321	1	global	global	ADJ
fcis-10735	321	2	average	average	ADJ
fcis-10735	321	3	pooling	pooling	NOUN
fcis-10735	321	4	(	(	PUNCT
fcis-10735	321	5	gap	gap	NOUN
fcis-10735	321	6	)	)	PUNCT
fcis-10735	322	1	[	[	X
fcis-10735	322	2	online	online	X
fcis-10735	322	3	]	]	X
fcis-10735	322	4	.	.	PUNCT
fcis-10735	323	1	available	available	ADJ
fcis-10735	323	2	:	:	PUNCT
fcis-10735	323	3	https://www.jianshu.com/p/04f7771f4da2	https://www.jianshu.com/p/04f7771f4da2	PROPN
fcis-10735	323	4	.	.	PUNCT