id	sid	tid	token	lemma	pos
ajst-32555	1	1	116	116	NUM
ajst-32555	1	2	machine	machine	NOUN
ajst-32555	1	3	learning	learning	NOUN
ajst-32555	1	4	-	-	PUNCT
ajst-32555	1	5	based	base	VERB
ajst-32555	1	6	prediction	prediction	NOUN
ajst-32555	1	7	of	of	ADP
ajst-32555	1	8	telecom	telecom	NOUN
ajst-32555	1	9	customer	customer	NOUN
ajst-32555	1	10	churn	churn	NOUN
ajst-32555	1	11	:	:	PUNCT
ajst-32555	1	12	comparative	comparative	ADJ
ajst-32555	1	13	model	model	NOUN
ajst-32555	1	14	analysis	analysis	NOUN
ajst-32555	1	15	yixiang	yixiang	PROPN
ajst-32555	1	16	zhang	zhang	PROPN
ajst-32555	1	17	university	university	PROPN
ajst-32555	1	18	of	of	ADP
ajst-32555	1	19	illinois	illinois	PROPN
ajst-32555	1	20	urbana	urbana	PROPN
ajst-32555	1	21	-	-	PUNCT
ajst-32555	1	22	champaign	champaign	PROPN
ajst-32555	1	23	,	,	PUNCT
ajst-32555	1	24	illinois	illinois	PROPN
ajst-32555	1	25	,	,	PUNCT
ajst-32555	1	26	the	the	DET
ajst-32555	1	27	united	united	PROPN
ajst-32555	1	28	states	states	PROPN
ajst-32555	1	29	dithosw@gmail.com	dithosw@gmail.com	X
ajst-32555	1	30	abstract	abstract	ADJ
ajst-32555	1	31	.	.	PUNCT
ajst-32555	2	1	in	in	ADP
ajst-32555	2	2	the	the	DET
ajst-32555	2	3	increasingly	increasingly	ADV
ajst-32555	2	4	competitive	competitive	ADJ
ajst-32555	2	5	telecommunications	telecommunication	NOUN
ajst-32555	2	6	market	market	NOUN
ajst-32555	2	7	,	,	PUNCT
ajst-32555	2	8	customer	customer	NOUN
ajst-32555	2	9	churn	churn	NOUN
ajst-32555	2	10	has	have	AUX
ajst-32555	2	11	become	become	VERB
ajst-32555	2	12	a	a	DET
ajst-32555	2	13	key	key	ADJ
ajst-32555	2	14	challenge	challenge	NOUN
ajst-32555	2	15	affecting	affect	VERB
ajst-32555	2	16	the	the	DET
ajst-32555	2	17	profitability	profitability	NOUN
ajst-32555	2	18	and	and	CCONJ
ajst-32555	2	19	sustainable	sustainable	ADJ
ajst-32555	2	20	growth	growth	NOUN
ajst-32555	2	21	of	of	ADP
ajst-32555	2	22	enterprises	enterprise	NOUN
ajst-32555	2	23	.	.	PUNCT
ajst-32555	3	1	this	this	DET
ajst-32555	3	2	study	study	NOUN
ajst-32555	3	3	aims	aim	VERB
ajst-32555	3	4	to	to	PART
ajst-32555	3	5	conduct	conduct	VERB
ajst-32555	3	6	an	an	DET
ajst-32555	3	7	empirical	empirical	ADJ
ajst-32555	3	8	analysis	analysis	NOUN
ajst-32555	3	9	to	to	PART
ajst-32555	3	10	compare	compare	VERB
ajst-32555	3	11	the	the	DET
ajst-32555	3	12	performance	performance	NOUN
ajst-32555	3	13	of	of	ADP
ajst-32555	3	14	various	various	ADJ
ajst-32555	3	15	machine	machine	NOUN
ajst-32555	3	16	learning	learning	NOUN
ajst-32555	3	17	models	model	NOUN
ajst-32555	3	18	in	in	ADP
ajst-32555	3	19	predicting	predict	VERB
ajst-32555	3	20	customer	customer	NOUN
ajst-32555	3	21	churn	churn	NOUN
ajst-32555	3	22	in	in	ADP
ajst-32555	3	23	the	the	DET
ajst-32555	3	24	telecommunications	telecommunications	NOUN
ajst-32555	3	25	industry	industry	NOUN
ajst-32555	3	26	.	.	PUNCT
ajst-32555	4	1	in	in	ADP
ajst-32555	4	2	this	this	DET
ajst-32555	4	3	study	study	NOUN
ajst-32555	4	4	,	,	PUNCT
ajst-32555	4	5	python	python	NOUN
ajst-32555	4	6	and	and	CCONJ
ajst-32555	4	7	related	relate	VERB
ajst-32555	4	8	machine	machine	NOUN
ajst-32555	4	9	learning	learning	NOUN
ajst-32555	4	10	libraries	library	NOUN
ajst-32555	4	11	were	be	AUX
ajst-32555	4	12	utilized	utilize	VERB
ajst-32555	4	13	to	to	PART
ajst-32555	4	14	build	build	VERB
ajst-32555	4	15	a	a	DET
ajst-32555	4	16	prediction	prediction	NOUN
ajst-32555	4	17	process	process	NOUN
ajst-32555	4	18	,	,	PUNCT
ajst-32555	4	19	and	and	CCONJ
ajst-32555	4	20	strict	strict	ADJ
ajst-32555	4	21	performance	performance	NOUN
ajst-32555	4	22	evaluations	evaluation	NOUN
ajst-32555	4	23	were	be	AUX
ajst-32555	4	24	conducted	conduct	VERB
ajst-32555	4	25	on	on	ADP
ajst-32555	4	26	three	three	NUM
ajst-32555	4	27	models	model	NOUN
ajst-32555	4	28	:	:	PUNCT
ajst-32555	4	29	logistic	logistic	ADJ
ajst-32555	4	30	regression	regression	NOUN
ajst-32555	4	31	,	,	PUNCT
ajst-32555	4	32	random	random	ADJ
ajst-32555	4	33	forest	forest	NOUN
ajst-32555	4	34	,	,	PUNCT
ajst-32555	4	35	and	and	CCONJ
ajst-32555	4	36	extreme	extreme	ADJ
ajst-32555	4	37	gradient	gradient	NOUN
ajst-32555	4	38	boost	boost	NOUN
ajst-32555	4	39	.	.	PUNCT
ajst-32555	5	1	the	the	DET
ajst-32555	5	2	evaluation	evaluation	NOUN
ajst-32555	5	3	indicators	indicator	NOUN
ajst-32555	5	4	include	include	VERB
ajst-32555	5	5	accuracy	accuracy	NOUN
ajst-32555	5	6	rate	rate	NOUN
ajst-32555	5	7	,	,	PUNCT
ajst-32555	5	8	precision	precision	NOUN
ajst-32555	5	9	rate	rate	NOUN
ajst-32555	5	10	,	,	PUNCT
ajst-32555	5	11	recall	recall	NOUN
ajst-32555	5	12	rate	rate	NOUN
ajst-32555	5	13	,	,	PUNCT
ajst-32555	5	14	f1	f1	NOUN
ajst-32555	5	15	score	score	NOUN
ajst-32555	5	16	,	,	PUNCT
ajst-32555	5	17	and	and	CCONJ
ajst-32555	5	18	the	the	DET
ajst-32555	5	19	area	area	NOUN
ajst-32555	5	20	under	under	ADP
ajst-32555	5	21	the	the	DET
ajst-32555	5	22	receiver	receiver	NOUN
ajst-32555	5	23	operating	operate	VERB
ajst-32555	5	24	characteristic	characteristic	ADJ
ajst-32555	5	25	curve	curve	NOUN
ajst-32555	5	26	(	(	PUNCT
ajst-32555	5	27	roc	roc	NOUN
ajst-32555	5	28	-	-	PUNCT
ajst-32555	5	29	auc	auc	NOUN
ajst-32555	5	30	)	)	PUNCT
ajst-32555	5	31	.	.	PUNCT
ajst-32555	6	1	the	the	DET
ajst-32555	6	2	research	research	NOUN
ajst-32555	6	3	results	result	NOUN
ajst-32555	6	4	show	show	VERB
ajst-32555	6	5	that	that	SCONJ
ajst-32555	6	6	the	the	DET
ajst-32555	6	7	optimized	optimize	VERB
ajst-32555	6	8	xgboost	xgboost	PROPN
ajst-32555	6	9	model	model	PROPN
ajst-32555	6	10	exhibits	exhibit	VERB
ajst-32555	6	11	the	the	DET
ajst-32555	6	12	best	good	ADJ
ajst-32555	6	13	performance	performance	NOUN
ajst-32555	6	14	in	in	ADP
ajst-32555	6	15	all	all	DET
ajst-32555	6	16	evaluation	evaluation	NOUN
ajst-32555	6	17	indicators	indicator	NOUN
ajst-32555	6	18	,	,	PUNCT
ajst-32555	6	19	demonstrating	demonstrate	VERB
ajst-32555	6	20	its	its	PRON
ajst-32555	6	21	outstanding	outstanding	ADJ
ajst-32555	6	22	ability	ability	NOUN
ajst-32555	6	23	to	to	PART
ajst-32555	6	24	handle	handle	VERB
ajst-32555	6	25	such	such	ADJ
ajst-32555	6	26	classification	classification	NOUN
ajst-32555	6	27	problems	problem	NOUN
ajst-32555	6	28	.	.	PUNCT
ajst-32555	7	1	in	in	ADP
ajst-32555	7	2	addition	addition	NOUN
ajst-32555	7	3	,	,	PUNCT
ajst-32555	7	4	by	by	ADP
ajst-32555	7	5	analyzing	analyze	VERB
ajst-32555	7	6	the	the	DET
ajst-32555	7	7	feature	feature	NOUN
ajst-32555	7	8	importance	importance	NOUN
ajst-32555	7	9	of	of	ADP
ajst-32555	7	10	the	the	DET
ajst-32555	7	11	xgboost	xgboost	PROPN
ajst-32555	7	12	model	model	PROPN
ajst-32555	7	13	,	,	PUNCT
ajst-32555	7	14	this	this	DET
ajst-32555	7	15	study	study	NOUN
ajst-32555	7	16	identified	identify	VERB
ajst-32555	7	17	the	the	DET
ajst-32555	7	18	key	key	ADJ
ajst-32555	7	19	drivers	driver	NOUN
ajst-32555	7	20	influencing	influence	VERB
ajst-32555	7	21	customer	customer	NOUN
ajst-32555	7	22	churn	churn	NOUN
ajst-32555	7	23	,	,	PUNCT
ajst-32555	7	24	among	among	ADP
ajst-32555	7	25	which	which	PRON
ajst-32555	7	26	contract	contract	NOUN
ajst-32555	7	27	type	type	NOUN
ajst-32555	7	28	,	,	PUNCT
ajst-32555	7	29	customer	customer	NOUN
ajst-32555	7	30	online	online	NOUN
ajst-32555	7	31	duration	duration	NOUN
ajst-32555	7	32	,	,	PUNCT
ajst-32555	7	33	and	and	CCONJ
ajst-32555	7	34	monthly	monthly	ADJ
ajst-32555	7	35	fee	fee	NOUN
ajst-32555	7	36	are	be	AUX
ajst-32555	7	37	the	the	DET
ajst-32555	7	38	most	most	ADV
ajst-32555	7	39	significant	significant	ADJ
ajst-32555	7	40	predictors	predictor	NOUN
ajst-32555	7	41	.	.	PUNCT
ajst-32555	8	1	these	these	DET
ajst-32555	8	2	findings	finding	NOUN
ajst-32555	8	3	not	not	PART
ajst-32555	8	4	only	only	ADV
ajst-32555	8	5	provide	provide	VERB
ajst-32555	8	6	telecom	telecom	NOUN
ajst-32555	8	7	operators	operator	NOUN
ajst-32555	8	8	with	with	ADP
ajst-32555	8	9	high	high	ADJ
ajst-32555	8	10	-	-	PUNCT
ajst-32555	8	11	precision	precision	NOUN
ajst-32555	8	12	churn	churn	NOUN
ajst-32555	8	13	warning	warning	NOUN
ajst-32555	8	14	tools	tool	NOUN
ajst-32555	8	15	but	but	CCONJ
ajst-32555	8	16	also	also	ADV
ajst-32555	8	17	offer	offer	VERB
ajst-32555	8	18	data	datum	NOUN
ajst-32555	8	19	-	-	PUNCT
ajst-32555	8	20	driven	drive	VERB
ajst-32555	8	21	decision	decision	NOUN
ajst-32555	8	22	support	support	NOUN
ajst-32555	8	23	for	for	SCONJ
ajst-32555	8	24	them	they	PRON
ajst-32555	8	25	to	to	PART
ajst-32555	8	26	formulate	formulate	VERB
ajst-32555	8	27	precise	precise	ADJ
ajst-32555	8	28	and	and	CCONJ
ajst-32555	8	29	effective	effective	ADJ
ajst-32555	8	30	customer	customer	NOUN
ajst-32555	8	31	retention	retention	NOUN
ajst-32555	8	32	strategies	strategy	NOUN
ajst-32555	8	33	.	.	PUNCT
ajst-32555	9	1	keywords	keyword	NOUN
ajst-32555	9	2	:	:	PUNCT
ajst-32555	9	3	customer	customer	NOUN
ajst-32555	9	4	churn	churn	NOUN
ajst-32555	9	5	;	;	PUNCT
ajst-32555	9	6	machine	machine	NOUN
ajst-32555	9	7	learning	learning	NOUN
ajst-32555	9	8	;	;	PUNCT
ajst-32555	9	9	python	python	NOUN
ajst-32555	9	10	;	;	PUNCT
ajst-32555	9	11	telecommunications	telecommunication	NOUN
ajst-32555	9	12	;	;	PUNCT
ajst-32555	9	13	xgboost	xgboost	X
ajst-32555	9	14	.	.	PUNCT
ajst-32555	10	1	1	1	X
ajst-32555	10	2	.	.	X
ajst-32555	10	3	introduction	introduction	NOUN
ajst-32555	10	4	in	in	ADP
ajst-32555	10	5	today	today	NOUN
ajst-32555	10	6	's	's	PART
ajst-32555	10	7	highly	highly	ADV
ajst-32555	10	8	saturated	saturated	ADJ
ajst-32555	10	9	and	and	CCONJ
ajst-32555	10	10	competitive	competitive	ADJ
ajst-32555	10	11	market	market	NOUN
ajst-32555	10	12	environment	environment	NOUN
ajst-32555	10	13	,	,	PUNCT
ajst-32555	10	14	customer	customer	NOUN
ajst-32555	10	15	relationship	relationship	NOUN
ajst-32555	10	16	management	management	NOUN
ajst-32555	10	17	(	(	PUNCT
ajst-32555	10	18	crm	crm	NOUN
ajst-32555	10	19	)	)	PUNCT
ajst-32555	10	20	has	have	AUX
ajst-32555	10	21	become	become	VERB
ajst-32555	10	22	a	a	DET
ajst-32555	10	23	core	core	NOUN
ajst-32555	10	24	strategy	strategy	NOUN
ajst-32555	10	25	for	for	ADP
ajst-32555	10	26	business	business	NOUN
ajst-32555	10	27	survival	survival	NOUN
ajst-32555	10	28	and	and	CCONJ
ajst-32555	10	29	growth	growth	NOUN
ajst-32555	10	30	.	.	PUNCT
ajst-32555	11	1	among	among	ADP
ajst-32555	11	2	all	all	DET
ajst-32555	11	3	customer	customer	NOUN
ajst-32555	11	4	-	-	PUNCT
ajst-32555	11	5	related	relate	VERB
ajst-32555	11	6	business	business	NOUN
ajst-32555	11	7	metrics	metric	NOUN
ajst-32555	11	8	,	,	PUNCT
ajst-32555	11	9	customer	customer	NOUN
ajst-32555	11	10	churn	churn	NOUN
ajst-32555	11	11	is	be	AUX
ajst-32555	11	12	particularly	particularly	ADV
ajst-32555	11	13	critical	critical	ADJ
ajst-32555	11	14	[	[	X
ajst-32555	11	15	1	1	NUM
ajst-32555	11	16	]	]	PUNCT
ajst-32555	11	17	.	.	PUNCT
ajst-32555	12	1	the	the	DET
ajst-32555	12	2	customer	customer	NOUN
ajst-32555	12	3	churn	churn	NOUN
ajst-32555	12	4	rate	rate	NOUN
ajst-32555	12	5	refers	refer	VERB
ajst-32555	12	6	to	to	ADP
ajst-32555	12	7	the	the	DET
ajst-32555	12	8	proportion	proportion	NOUN
ajst-32555	12	9	of	of	ADP
ajst-32555	12	10	customers	customer	NOUN
ajst-32555	12	11	who	who	PRON
ajst-32555	12	12	cease	cease	VERB
ajst-32555	12	13	using	use	VERB
ajst-32555	12	14	a	a	DET
ajst-32555	12	15	company	company	NOUN
ajst-32555	12	16	's	's	PART
ajst-32555	12	17	products	product	NOUN
ajst-32555	12	18	or	or	CCONJ
ajst-32555	12	19	services	service	NOUN
ajst-32555	12	20	within	within	ADP
ajst-32555	12	21	a	a	DET
ajst-32555	12	22	specific	specific	ADJ
ajst-32555	12	23	period	period	NOUN
ajst-32555	12	24	.	.	PUNCT
ajst-32555	13	1	this	this	DET
ajst-32555	13	2	metric	metric	ADJ
ajst-32555	13	3	directly	directly	ADV
ajst-32555	13	4	impacts	impact	VERB
ajst-32555	13	5	a	a	DET
ajst-32555	13	6	company	company	NOUN
ajst-32555	13	7	's	's	PART
ajst-32555	13	8	revenue	revenue	NOUN
ajst-32555	13	9	stability	stability	NOUN
ajst-32555	13	10	and	and	CCONJ
ajst-32555	13	11	long	long	ADJ
ajst-32555	13	12	-	-	PUNCT
ajst-32555	13	13	term	term	NOUN
ajst-32555	13	14	growth	growth	NOUN
ajst-32555	13	15	potential	potential	NOUN
ajst-32555	13	16	[	[	X
ajst-32555	13	17	2	2	NUM
ajst-32555	13	18	]	]	PUNCT
ajst-32555	13	19	.	.	PUNCT
ajst-32555	14	1	extensive	extensive	ADJ
ajst-32555	14	2	research	research	NOUN
ajst-32555	14	3	and	and	CCONJ
ajst-32555	14	4	business	business	NOUN
ajst-32555	14	5	practice	practice	NOUN
ajst-32555	14	6	demonstrate	demonstrate	VERB
ajst-32555	14	7	that	that	SCONJ
ajst-32555	14	8	retaining	retain	VERB
ajst-32555	14	9	an	an	DET
ajst-32555	14	10	existing	exist	VERB
ajst-32555	14	11	customer	customer	NOUN
ajst-32555	14	12	costs	cost	VERB
ajst-32555	14	13	significantly	significantly	ADV
ajst-32555	14	14	less	less	ADJ
ajst-32555	14	15	than	than	ADP
ajst-32555	14	16	acquiring	acquire	VERB
ajst-32555	14	17	a	a	DET
ajst-32555	14	18	new	new	ADJ
ajst-32555	14	19	one	one	NOUN
ajst-32555	14	20	.	.	PUNCT
ajst-32555	15	1	for	for	ADP
ajst-32555	15	2	instance	instance	NOUN
ajst-32555	15	3	,	,	PUNCT
ajst-32555	15	4	studies	study	NOUN
ajst-32555	15	5	indicate	indicate	VERB
ajst-32555	15	6	that	that	SCONJ
ajst-32555	15	7	reducing	reduce	VERB
ajst-32555	15	8	the	the	DET
ajst-32555	15	9	customer	customer	NOUN
ajst-32555	15	10	churn	churn	NOUN
ajst-32555	15	11	rate	rate	NOUN
ajst-32555	15	12	by	by	ADP
ajst-32555	15	13	5	5	NUM
ajst-32555	15	14	%	%	NOUN
ajst-32555	15	15	can	can	AUX
ajst-32555	15	16	increase	increase	VERB
ajst-32555	15	17	corporate	corporate	ADJ
ajst-32555	15	18	profits	profit	NOUN
ajst-32555	15	19	by	by	ADP
ajst-32555	15	20	25	25	NUM
ajst-32555	15	21	%	%	NOUN
ajst-32555	15	22	to	to	PART
ajst-32555	15	23	85	85	NUM
ajst-32555	15	24	%	%	NOUN
ajst-32555	15	25	[	[	X
ajst-32555	15	26	3	3	NUM
ajst-32555	15	27	]	]	PUNCT
ajst-32555	15	28	.	.	PUNCT
ajst-32555	16	1	for	for	ADP
ajst-32555	16	2	the	the	DET
ajst-32555	16	3	telecommunications	telecommunications	NOUN
ajst-32555	16	4	industry	industry	NOUN
ajst-32555	16	5	,	,	PUNCT
ajst-32555	16	6	customer	customer	NOUN
ajst-32555	16	7	churn	churn	NOUN
ajst-32555	16	8	poses	pose	VERB
ajst-32555	16	9	particularly	particularly	ADV
ajst-32555	16	10	severe	severe	ADJ
ajst-32555	16	11	challenges	challenge	NOUN
ajst-32555	16	12	.	.	PUNCT
ajst-32555	17	1	characterized	characterize	VERB
ajst-32555	17	2	by	by	ADP
ajst-32555	17	3	a	a	DET
ajst-32555	17	4	subscription	subscription	NOUN
ajst-32555	17	5	-	-	PUNCT
ajst-32555	17	6	based	base	VERB
ajst-32555	17	7	business	business	NOUN
ajst-32555	17	8	model	model	NOUN
ajst-32555	17	9	and	and	CCONJ
ajst-32555	17	10	near	near	ADV
ajst-32555	17	11	-	-	PUNCT
ajst-32555	17	12	saturated	saturate	VERB
ajst-32555	17	13	market	market	NOUN
ajst-32555	17	14	penetration	penetration	NOUN
ajst-32555	17	15	,	,	PUNCT
ajst-32555	17	16	competition	competition	NOUN
ajst-32555	17	17	among	among	ADP
ajst-32555	17	18	major	major	ADJ
ajst-32555	17	19	operators	operator	NOUN
ajst-32555	17	20	primarily	primarily	ADV
ajst-32555	17	21	centers	center	VERB
ajst-32555	17	22	on	on	ADP
ajst-32555	17	23	existing	exist	VERB
ajst-32555	17	24	customer	customer	NOUN
ajst-32555	17	25	bases	basis	NOUN
ajst-32555	17	26	[	[	X
ajst-32555	17	27	4	4	NUM
ajst-32555	17	28	]	]	PUNCT
ajst-32555	17	29	.	.	PUNCT
ajst-32555	18	1	statistics	statistic	NOUN
ajst-32555	18	2	indicate	indicate	VERB
ajst-32555	18	3	that	that	SCONJ
ajst-32555	18	4	the	the	DET
ajst-32555	18	5	telecom	telecom	NOUN
ajst-32555	18	6	industry	industry	NOUN
ajst-32555	18	7	's	's	PART
ajst-32555	18	8	average	average	ADJ
ajst-32555	18	9	annual	annual	ADJ
ajst-32555	18	10	customer	customer	NOUN
ajst-32555	18	11	churn	churn	NOUN
ajst-32555	18	12	rate	rate	NOUN
ajst-32555	18	13	ranks	rank	VERB
ajst-32555	18	14	among	among	ADP
ajst-32555	18	15	the	the	DET
ajst-32555	18	16	highest	high	ADJ
ajst-32555	18	17	across	across	ADP
ajst-32555	18	18	major	major	ADJ
ajst-32555	18	19	sectors	sector	NOUN
ajst-32555	18	20	,	,	PUNCT
ajst-32555	18	21	sometimes	sometimes	ADV
ajst-32555	18	22	exceeding	exceed	VERB
ajst-32555	18	23	30	30	NUM
ajst-32555	18	24	%	%	NOUN
ajst-32555	19	1	[	[	X
ajst-32555	19	2	5	5	NUM
ajst-32555	19	3	]	]	PUNCT
ajst-32555	19	4	.	.	PUNCT
ajst-32555	20	1	therefore	therefore	ADV
ajst-32555	20	2	,	,	PUNCT
ajst-32555	20	3	establishing	establish	VERB
ajst-32555	20	4	a	a	DET
ajst-32555	20	5	mechanism	mechanism	NOUN
ajst-32555	20	6	that	that	PRON
ajst-32555	20	7	accurately	accurately	ADV
ajst-32555	20	8	identifies	identify	VERB
ajst-32555	20	9	potential	potential	ADJ
ajst-32555	20	10	churn	churn	NOUN
ajst-32555	20	11	customers	customer	NOUN
ajst-32555	20	12	and	and	CCONJ
ajst-32555	20	13	understands	understand	VERB
ajst-32555	20	14	their	their	PRON
ajst-32555	20	15	reasons	reason	NOUN
ajst-32555	20	16	for	for	ADP
ajst-32555	20	17	leaving	leave	VERB
ajst-32555	20	18	is	be	AUX
ajst-32555	20	19	not	not	PART
ajst-32555	20	20	only	only	ADV
ajst-32555	20	21	a	a	DET
ajst-32555	20	22	necessity	necessity	NOUN
ajst-32555	20	23	for	for	SCONJ
ajst-32555	20	24	telecom	telecom	NOUN
ajst-32555	20	25	operators	operator	NOUN
ajst-32555	20	26	to	to	PART
ajst-32555	20	27	enhance	enhance	VERB
ajst-32555	20	28	profitability	profitability	NOUN
ajst-32555	20	29	but	but	CCONJ
ajst-32555	20	30	also	also	ADV
ajst-32555	20	31	a	a	DET
ajst-32555	20	32	strategic	strategic	ADJ
ajst-32555	20	33	imperative	imperative	NOUN
ajst-32555	20	34	to	to	PART
ajst-32555	20	35	maintain	maintain	VERB
ajst-32555	20	36	core	core	NOUN
ajst-32555	20	37	competitiveness	competitiveness	NOUN
ajst-32555	20	38	[	[	X
ajst-32555	20	39	6	6	NUM
ajst-32555	20	40	]	]	PUNCT
ajst-32555	20	41	.	.	PUNCT
ajst-32555	21	1	traditionally	traditionally	ADV
ajst-32555	21	2	,	,	PUNCT
ajst-32555	21	3	customer	customer	NOUN
ajst-32555	21	4	retention	retention	NOUN
ajst-32555	21	5	strategies	strategy	NOUN
ajst-32555	21	6	have	have	AUX
ajst-32555	21	7	been	be	AUX
ajst-32555	21	8	largely	largely	ADV
ajst-32555	21	9	reactive	reactive	ADJ
ajst-32555	21	10	,	,	PUNCT
ajst-32555	21	11	involving	involve	VERB
ajst-32555	21	12	remedial	remedial	ADJ
ajst-32555	21	13	actions	action	NOUN
ajst-32555	21	14	only	only	ADV
ajst-32555	21	15	after	after	SCONJ
ajst-32555	21	16	customers	customer	NOUN
ajst-32555	21	17	submit	submit	VERB
ajst-32555	21	18	cancellation	cancellation	NOUN
ajst-32555	21	19	requests	request	NOUN
ajst-32555	21	20	—	—	PUNCT
ajst-32555	21	21	often	often	ADV
ajst-32555	21	22	too	too	ADV
ajst-32555	21	23	late	late	ADJ
ajst-32555	21	24	.	.	PUNCT
ajst-32555	22	1	with	with	ADP
ajst-32555	22	2	the	the	DET
ajst-32555	22	3	advancement	advancement	NOUN
ajst-32555	22	4	of	of	ADP
ajst-32555	22	5	big	big	ADJ
ajst-32555	22	6	data	datum	NOUN
ajst-32555	22	7	technology	technology	NOUN
ajst-32555	22	8	and	and	CCONJ
ajst-32555	22	9	artificial	artificial	ADJ
ajst-32555	22	10	intelligence	intelligence	NOUN
ajst-32555	22	11	,	,	PUNCT
ajst-32555	22	12	machine	machine	NOUN
ajst-32555	22	13	learning	learning	NOUN
ajst-32555	22	14	has	have	AUX
ajst-32555	22	15	revolutionized	revolutionize	VERB
ajst-32555	22	16	customer	customer	NOUN
ajst-32555	22	17	churn	churn	NOUN
ajst-32555	22	18	management	management	NOUN
ajst-32555	22	19	,	,	PUNCT
ajst-32555	22	20	propelling	propel	VERB
ajst-32555	22	21	enterprises	enterprise	NOUN
ajst-32555	22	22	from	from	ADP
ajst-32555	22	23	passive	passive	ADJ
ajst-32555	22	24	response	response	NOUN
ajst-32555	22	25	to	to	ADP
ajst-32555	22	26	proactive	proactive	ADJ
ajst-32555	22	27	prediction	prediction	NOUN
ajst-32555	22	28	[	[	X
ajst-32555	22	29	7	7	NUM
ajst-32555	22	30	]	]	PUNCT
ajst-32555	22	31	.	.	PUNCT
ajst-32555	23	1	by	by	ADP
ajst-32555	23	2	analyzing	analyze	VERB
ajst-32555	23	3	vast	vast	ADJ
ajst-32555	23	4	amounts	amount	NOUN
ajst-32555	23	5	of	of	ADP
ajst-32555	23	6	historical	historical	ADJ
ajst-32555	23	7	customer	customer	NOUN
ajst-32555	23	8	data	datum	NOUN
ajst-32555	23	9	—	—	PUNCT
ajst-32555	23	10	encompassing	encompass	VERB
ajst-32555	23	11	demographics	demographic	NOUN
ajst-32555	23	12	,	,	PUNCT
ajst-32555	23	13	account	account	NOUN
ajst-32555	23	14	details	detail	NOUN
ajst-32555	23	15	,	,	PUNCT
ajst-32555	23	16	service	service	NOUN
ajst-32555	23	17	usage	usage	NOUN
ajst-32555	23	18	records	record	NOUN
ajst-32555	23	19	,	,	PUNCT
ajst-32555	23	20	and	and	CCONJ
ajst-32555	23	21	consumption	consumption	NOUN
ajst-32555	23	22	behavior	behavior	NOUN
ajst-32555	23	23	—	—	PUNCT
ajst-32555	23	24	machine	machine	NOUN
ajst-32555	23	25	learning	learning	NOUN
ajst-32555	23	26	models	model	NOUN
ajst-32555	23	27	can	can	AUX
ajst-32555	23	28	learn	learn	VERB
ajst-32555	23	29	and	and	CCONJ
ajst-32555	23	30	identify	identify	VERB
ajst-32555	23	31	complex	complex	ADJ
ajst-32555	23	32	,	,	PUNCT
ajst-32555	23	33	nonlinear	nonlinear	ADJ
ajst-32555	23	34	patterns	pattern	NOUN
ajst-32555	23	35	underlying	underlie	VERB
ajst-32555	23	36	customer	customer	NOUN
ajst-32555	23	37	churn	churn	NOUN
ajst-32555	23	38	[	[	X
ajst-32555	23	39	2	2	NUM
ajst-32555	23	40	]	]	PUNCT
ajst-32555	23	41	.	.	PUNCT
ajst-32555	24	1	this	this	DET
ajst-32555	24	2	predictive	predictive	ADJ
ajst-32555	24	3	capability	capability	NOUN
ajst-32555	24	4	enables	enable	VERB
ajst-32555	24	5	companies	company	NOUN
ajst-32555	24	6	to	to	PART
ajst-32555	24	7	calculate	calculate	VERB
ajst-32555	24	8	each	each	DET
ajst-32555	24	9	customer	customer	NOUN
ajst-32555	24	10	's	's	PART
ajst-32555	24	11	probability	probability	NOUN
ajst-32555	24	12	of	of	ADP
ajst-32555	24	13	future	future	ADJ
ajst-32555	24	14	churn	churn	NOUN
ajst-32555	24	15	.	.	PUNCT
ajst-32555	25	1	based	base	VERB
ajst-32555	25	2	on	on	ADP
ajst-32555	25	3	this	this	DET
ajst-32555	25	4	risk	risk	NOUN
ajst-32555	25	5	score	score	NOUN
ajst-32555	25	6	,	,	PUNCT
ajst-32555	25	7	businesses	business	NOUN
ajst-32555	25	8	can	can	AUX
ajst-32555	25	9	identify	identify	VERB
ajst-32555	25	10	high	high	ADJ
ajst-32555	25	11	-	-	PUNCT
ajst-32555	25	12	risk	risk	NOUN
ajst-32555	25	13	customer	customer	NOUN
ajst-32555	25	14	segments	segment	NOUN
ajst-32555	25	15	and	and	CCONJ
ajst-32555	25	16	implement	implement	VERB
ajst-32555	25	17	precise	precise	ADJ
ajst-32555	25	18	,	,	PUNCT
ajst-32555	25	19	personalized	personalized	ADJ
ajst-32555	25	20	retention	retention	NOUN
ajst-32555	25	21	interventions	intervention	NOUN
ajst-32555	25	22	before	before	SCONJ
ajst-32555	25	23	they	they	PRON
ajst-32555	25	24	decide	decide	VERB
ajst-32555	25	25	to	to	PART
ajst-32555	25	26	leave	leave	VERB
ajst-32555	25	27	—	—	PUNCT
ajst-32555	25	28	such	such	ADJ
ajst-32555	25	29	as	as	ADP
ajst-32555	25	30	offering	offer	VERB
ajst-32555	25	31	customized	customize	VERB
ajst-32555	25	32	package	package	NOUN
ajst-32555	25	33	deals	deal	NOUN
ajst-32555	25	34	,	,	PUNCT
ajst-32555	25	35	providing	provide	VERB
ajst-32555	25	36	value	value	NOUN
ajst-32555	25	37	-	-	PUNCT
ajst-32555	25	38	added	add	VERB
ajst-32555	25	39	services	service	NOUN
ajst-32555	25	40	,	,	PUNCT
ajst-32555	25	41	or	or	CCONJ
ajst-32555	25	42	117	117	NUM
ajst-32555	25	43	initiating	initiate	VERB
ajst-32555	25	44	proactive	proactive	ADJ
ajst-32555	25	45	customer	customer	NOUN
ajst-32555	25	46	care	care	NOUN
ajst-32555	25	47	communications	communication	NOUN
ajst-32555	25	48	[	[	X
ajst-32555	25	49	8	8	NUM
ajst-32555	25	50	]	]	PUNCT
ajst-32555	25	51	.	.	PUNCT
ajst-32555	26	1	this	this	DET
ajst-32555	26	2	data	data	NOUN
ajst-32555	26	3	-	-	PUNCT
ajst-32555	26	4	driven	drive	VERB
ajst-32555	26	5	proactive	proactive	ADJ
ajst-32555	26	6	retention	retention	NOUN
ajst-32555	26	7	strategy	strategy	NOUN
ajst-32555	26	8	not	not	PART
ajst-32555	26	9	only	only	ADV
ajst-32555	26	10	significantly	significantly	ADV
ajst-32555	26	11	improves	improve	VERB
ajst-32555	26	12	retention	retention	NOUN
ajst-32555	26	13	success	success	NOUN
ajst-32555	26	14	rates	rate	NOUN
ajst-32555	26	15	but	but	CCONJ
ajst-32555	26	16	also	also	ADV
ajst-32555	26	17	optimizes	optimize	VERB
ajst-32555	26	18	marketing	marketing	NOUN
ajst-32555	26	19	resource	resource	NOUN
ajst-32555	26	20	allocation	allocation	NOUN
ajst-32555	26	21	.	.	PUNCT
ajst-32555	27	1	it	it	PRON
ajst-32555	27	2	concentrates	concentrate	VERB
ajst-32555	27	3	limited	limited	ADJ
ajst-32555	27	4	budgets	budget	NOUN
ajst-32555	27	5	on	on	ADP
ajst-32555	27	6	customers	customer	NOUN
ajst-32555	27	7	most	most	ADV
ajst-32555	27	8	likely	likely	ADJ
ajst-32555	27	9	to	to	PART
ajst-32555	27	10	churn	churn	VERB
ajst-32555	27	11	and	and	CCONJ
ajst-32555	27	12	with	with	ADP
ajst-32555	27	13	the	the	DET
ajst-32555	27	14	highest	high	ADJ
ajst-32555	27	15	retention	retention	NOUN
ajst-32555	27	16	value	value	NOUN
ajst-32555	27	17	,	,	PUNCT
ajst-32555	27	18	thereby	thereby	ADV
ajst-32555	27	19	maximizing	maximize	VERB
ajst-32555	27	20	the	the	DET
ajst-32555	27	21	effectiveness	effectiveness	NOUN
ajst-32555	27	22	of	of	ADP
ajst-32555	27	23	customer	customer	NOUN
ajst-32555	27	24	relationship	relationship	NOUN
ajst-32555	27	25	management	management	NOUN
ajst-32555	28	1	[	[	X
ajst-32555	28	2	3	3	NUM
ajst-32555	28	3	]	]	PUNCT
ajst-32555	28	4	.	.	PUNCT
ajst-32555	29	1	existing	exist	VERB
ajst-32555	29	2	literature	literature	NOUN
ajst-32555	29	3	employs	employ	VERB
ajst-32555	29	4	various	various	ADJ
ajst-32555	29	5	techniques	technique	NOUN
ajst-32555	29	6	ranging	range	VERB
ajst-32555	29	7	from	from	ADP
ajst-32555	29	8	traditional	traditional	ADJ
ajst-32555	29	9	statistical	statistical	ADJ
ajst-32555	29	10	models	model	NOUN
ajst-32555	29	11	to	to	ADP
ajst-32555	29	12	advanced	advanced	ADJ
ajst-32555	29	13	deep	deep	ADJ
ajst-32555	29	14	learning	learning	NOUN
ajst-32555	29	15	methods	method	NOUN
ajst-32555	29	16	.	.	PUNCT
ajst-32555	30	1	for	for	ADP
ajst-32555	30	2	instance	instance	NOUN
ajst-32555	30	3	,	,	PUNCT
ajst-32555	30	4	ahmad	ahmad	PROPN
ajst-32555	30	5	,	,	PUNCT
ajst-32555	30	6	jafar	jafar	PROPN
ajst-32555	30	7	,	,	PUNCT
ajst-32555	30	8	&	&	CCONJ
ajst-32555	30	9	aljoumaa	aljoumaa	ADV
ajst-32555	30	10	achieved	achieve	VERB
ajst-32555	30	11	an	an	DET
ajst-32555	30	12	impressive	impressive	ADJ
ajst-32555	30	13	auc	auc	NOUN
ajst-32555	30	14	of	of	ADP
ajst-32555	30	15	93.3	93.3	NUM
ajst-32555	30	16	%	%	NOUN
ajst-32555	30	17	on	on	ADP
ajst-32555	30	18	a	a	DET
ajst-32555	30	19	large	large	ADJ
ajst-32555	30	20	telecom	telecom	NOUN
ajst-32555	30	21	company	company	NOUN
ajst-32555	30	22	's	's	PART
ajst-32555	30	23	real	real	ADJ
ajst-32555	30	24	-	-	PUNCT
ajst-32555	30	25	world	world	NOUN
ajst-32555	30	26	big	big	ADJ
ajst-32555	30	27	data	datum	NOUN
ajst-32555	30	28	platform	platform	NOUN
ajst-32555	30	29	[	[	X
ajst-32555	30	30	9	9	NUM
ajst-32555	30	31	]	]	PUNCT
ajst-32555	30	32	.	.	PUNCT
ajst-32555	31	1	this	this	DET
ajst-32555	31	2	result	result	NOUN
ajst-32555	31	3	was	be	AUX
ajst-32555	31	4	attained	attain	VERB
ajst-32555	31	5	by	by	ADP
ajst-32555	31	6	constructing	construct	VERB
ajst-32555	31	7	innovative	innovative	ADJ
ajst-32555	31	8	feature	feature	NOUN
ajst-32555	31	9	engineering	engineering	NOUN
ajst-32555	31	10	methods	method	NOUN
ajst-32555	31	11	(	(	PUNCT
ajst-32555	31	12	such	such	ADJ
ajst-32555	31	13	as	as	ADP
ajst-32555	31	14	social	social	ADJ
ajst-32555	31	15	network	network	NOUN
ajst-32555	31	16	analysis	analysis	NOUN
ajst-32555	31	17	features	feature	NOUN
ajst-32555	31	18	)	)	PUNCT
ajst-32555	31	19	and	and	CCONJ
ajst-32555	31	20	leveraging	leverage	VERB
ajst-32555	31	21	advanced	advanced	ADJ
ajst-32555	31	22	algorithms	algorithm	NOUN
ajst-32555	31	23	like	like	ADP
ajst-32555	31	24	xgboost	xgboost	ADV
ajst-32555	31	25	,	,	PUNCT
ajst-32555	31	26	demonstrating	demonstrate	VERB
ajst-32555	31	27	the	the	DET
ajst-32555	31	28	high	high	ADJ
ajst-32555	31	29	level	level	NOUN
ajst-32555	31	30	of	of	ADP
ajst-32555	31	31	churn	churn	NOUN
ajst-32555	31	32	prediction	prediction	NOUN
ajst-32555	31	33	achievable	achievable	ADJ
ajst-32555	31	34	with	with	ADP
ajst-32555	31	35	rich	rich	ADJ
ajst-32555	31	36	data	datum	NOUN
ajst-32555	31	37	sources	source	NOUN
ajst-32555	31	38	and	and	CCONJ
ajst-32555	31	39	robust	robust	ADJ
ajst-32555	31	40	computational	computational	ADJ
ajst-32555	31	41	platforms	platform	NOUN
ajst-32555	31	42	[	[	X
ajst-32555	31	43	9	9	NUM
ajst-32555	31	44	]	]	PUNCT
ajst-32555	31	45	.	.	PUNCT
ajst-32555	32	1	concurrently	concurrently	ADV
ajst-32555	32	2	,	,	PUNCT
ajst-32555	32	3	academia	academia	NOUN
ajst-32555	32	4	increasingly	increasingly	ADV
ajst-32555	32	5	emphasizes	emphasize	VERB
ajst-32555	32	6	model	model	NOUN
ajst-32555	32	7	interpretability	interpretability	NOUN
ajst-32555	32	8	.	.	PUNCT
ajst-32555	33	1	a	a	DET
ajst-32555	33	2	black	black	ADJ
ajst-32555	33	3	-	-	PUNCT
ajst-32555	33	4	box	box	NOUN
ajst-32555	33	5	model	model	NOUN
ajst-32555	33	6	with	with	ADP
ajst-32555	33	7	high	high	ADJ
ajst-32555	33	8	predictive	predictive	ADJ
ajst-32555	33	9	accuracy	accuracy	NOUN
ajst-32555	33	10	but	but	CCONJ
ajst-32555	33	11	no	no	DET
ajst-32555	33	12	decision	decision	NOUN
ajst-32555	33	13	logic	logic	NOUN
ajst-32555	33	14	explanation	explanation	NOUN
ajst-32555	33	15	holds	hold	VERB
ajst-32555	33	16	limited	limited	ADJ
ajst-32555	33	17	commercial	commercial	ADJ
ajst-32555	33	18	value	value	NOUN
ajst-32555	33	19	,	,	PUNCT
ajst-32555	33	20	as	as	SCONJ
ajst-32555	33	21	it	it	PRON
ajst-32555	33	22	fails	fail	VERB
ajst-32555	33	23	to	to	PART
ajst-32555	33	24	inform	inform	VERB
ajst-32555	33	25	decision	decision	NOUN
ajst-32555	33	26	-	-	PUNCT
ajst-32555	33	27	makers	maker	NOUN
ajst-32555	33	28	about	about	ADP
ajst-32555	33	29	why	why	SCONJ
ajst-32555	33	30	customers	customer	NOUN
ajst-32555	33	31	churn	churn	VERB
ajst-32555	33	32	,	,	PUNCT
ajst-32555	33	33	hindering	hinder	VERB
ajst-32555	33	34	effective	effective	ADJ
ajst-32555	33	35	intervention	intervention	NOUN
ajst-32555	33	36	strategies	strategy	NOUN
ajst-32555	33	37	.	.	PUNCT
ajst-32555	34	1	addressing	address	VERB
ajst-32555	34	2	this	this	DET
ajst-32555	34	3	challenge	challenge	NOUN
ajst-32555	34	4	,	,	PUNCT
ajst-32555	34	5	de	de	X
ajst-32555	34	6	caigny	caigny	PROPN
ajst-32555	34	7	,	,	PUNCT
ajst-32555	34	8	coussement	coussement	ADJ
ajst-32555	34	9	,	,	PUNCT
ajst-32555	34	10	&	&	CCONJ
ajst-32555	34	11	de	de	X
ajst-32555	34	12	bock	bock	NOUN
ajst-32555	34	13	proposed	propose	VERB
ajst-32555	34	14	the	the	DET
ajst-32555	34	15	logit	logit	PROPN
ajst-32555	34	16	leaf	leaf	NOUN
ajst-32555	34	17	model	model	NOUN
ajst-32555	34	18	(	(	PUNCT
ajst-32555	34	19	llm	llm	PROPN
ajst-32555	34	20	)	)	PUNCT
ajst-32555	34	21	,	,	PUNCT
ajst-32555	34	22	a	a	DET
ajst-32555	34	23	hybrid	hybrid	ADJ
ajst-32555	34	24	algorithm	algorithm	NOUN
ajst-32555	34	25	that	that	PRON
ajst-32555	34	26	ingeniously	ingeniously	ADV
ajst-32555	34	27	combines	combine	VERB
ajst-32555	34	28	the	the	DET
ajst-32555	34	29	strengths	strength	NOUN
ajst-32555	34	30	of	of	ADP
ajst-32555	34	31	decision	decision	NOUN
ajst-32555	34	32	trees	tree	NOUN
ajst-32555	34	33	(	(	PUNCT
ajst-32555	34	34	excelling	excel	VERB
ajst-32555	34	35	at	at	ADP
ajst-32555	34	36	capturing	capture	VERB
ajst-32555	34	37	variable	variable	ADJ
ajst-32555	34	38	interactions	interaction	NOUN
ajst-32555	34	39	)	)	PUNCT
ajst-32555	34	40	and	and	CCONJ
ajst-32555	34	41	logistic	logistic	ADJ
ajst-32555	34	42	regression	regression	NOUN
ajst-32555	34	43	(	(	PUNCT
ajst-32555	34	44	excelling	excel	VERB
ajst-32555	34	45	at	at	ADP
ajst-32555	34	46	handling	handle	VERB
ajst-32555	34	47	linear	linear	NOUN
ajst-32555	34	48	relationships	relationship	NOUN
ajst-32555	34	49	)	)	PUNCT
ajst-32555	35	1	[	[	X
ajst-32555	35	2	10	10	NUM
ajst-32555	35	3	]	]	PUNCT
ajst-32555	35	4	.	.	PUNCT
ajst-32555	36	1	this	this	DET
ajst-32555	36	2	approach	approach	NOUN
ajst-32555	36	3	aims	aim	VERB
ajst-32555	36	4	to	to	PART
ajst-32555	36	5	balance	balance	VERB
ajst-32555	36	6	predictive	predictive	ADJ
ajst-32555	36	7	performance	performance	NOUN
ajst-32555	36	8	with	with	ADP
ajst-32555	36	9	model	model	NOUN
ajst-32555	36	10	interpretability	interpretability	NOUN
ajst-32555	36	11	.	.	PUNCT
ajst-32555	37	1	the	the	DET
ajst-32555	37	2	structure	structure	NOUN
ajst-32555	37	3	of	of	ADP
ajst-32555	37	4	this	this	DET
ajst-32555	37	5	paper	paper	NOUN
ajst-32555	37	6	is	be	AUX
ajst-32555	37	7	as	as	SCONJ
ajst-32555	37	8	follows	follow	VERB
ajst-32555	37	9	:	:	PUNCT
ajst-32555	37	10	section	section	NOUN
ajst-32555	37	11	2	2	NUM
ajst-32555	37	12	details	detail	NOUN
ajst-32555	37	13	the	the	DET
ajst-32555	37	14	dataset	dataset	NOUN
ajst-32555	37	15	,	,	PUNCT
ajst-32555	37	16	data	datum	NOUN
ajst-32555	37	17	preprocessing	preprocessing	NOUN
ajst-32555	37	18	methods	method	NOUN
ajst-32555	37	19	,	,	PUNCT
ajst-32555	37	20	selected	select	VERB
ajst-32555	37	21	machine	machine	NOUN
ajst-32555	37	22	learning	learning	NOUN
ajst-32555	37	23	models	model	NOUN
ajst-32555	37	24	,	,	PUNCT
ajst-32555	37	25	and	and	CCONJ
ajst-32555	37	26	evaluation	evaluation	NOUN
ajst-32555	37	27	protocols	protocol	NOUN
ajst-32555	37	28	.	.	PUNCT
ajst-32555	38	1	section	section	NOUN
ajst-32555	38	2	3	3	NUM
ajst-32555	38	3	presents	present	VERB
ajst-32555	38	4	comprehensive	comprehensive	ADJ
ajst-32555	38	5	experimental	experimental	ADJ
ajst-32555	38	6	results	result	NOUN
ajst-32555	38	7	,	,	PUNCT
ajst-32555	38	8	including	include	VERB
ajst-32555	38	9	comparative	comparative	ADJ
ajst-32555	38	10	analysis	analysis	NOUN
ajst-32555	38	11	of	of	ADP
ajst-32555	38	12	model	model	NOUN
ajst-32555	38	13	performance	performance	NOUN
ajst-32555	38	14	,	,	PUNCT
ajst-32555	38	15	identification	identification	NOUN
ajst-32555	38	16	of	of	ADP
ajst-32555	38	17	key	key	ADJ
ajst-32555	38	18	features	feature	NOUN
ajst-32555	38	19	,	,	PUNCT
ajst-32555	38	20	and	and	CCONJ
ajst-32555	38	21	in	in	ADP
ajst-32555	38	22	-	-	PUNCT
ajst-32555	38	23	depth	depth	NOUN
ajst-32555	38	24	exploration	exploration	NOUN
ajst-32555	38	25	of	of	ADP
ajst-32555	38	26	their	their	PRON
ajst-32555	38	27	business	business	NOUN
ajst-32555	38	28	implications	implication	NOUN
ajst-32555	38	29	.	.	PUNCT
ajst-32555	39	1	section	section	NOUN
ajst-32555	39	2	4	4	NUM
ajst-32555	39	3	summarizes	summarize	NOUN
ajst-32555	39	4	the	the	DET
ajst-32555	39	5	findings	finding	NOUN
ajst-32555	39	6	,	,	PUNCT
ajst-32555	39	7	discusses	discuss	VERB
ajst-32555	39	8	limitations	limitation	NOUN
ajst-32555	39	9	,	,	PUNCT
ajst-32555	39	10	and	and	CCONJ
ajst-32555	39	11	outlines	outline	VERB
ajst-32555	39	12	future	future	ADJ
ajst-32555	39	13	research	research	NOUN
ajst-32555	39	14	directions	direction	NOUN
ajst-32555	39	15	.	.	PUNCT
ajst-32555	40	1	2	2	X
ajst-32555	40	2	.	.	X
ajst-32555	40	3	data	datum	NOUN
ajst-32555	40	4	&	&	CCONJ
ajst-32555	40	5	methods	method	NOUN
ajst-32555	40	6	2.1	2.1	NUM
ajst-32555	40	7	dataset	dataset	VERB
ajst-32555	40	8	and	and	CCONJ
ajst-32555	40	9	preprocessing	preprocesse	VERB
ajst-32555	40	10	the	the	DET
ajst-32555	40	11	data	data	NOUN
ajst-32555	40	12	source	source	NOUN
ajst-32555	40	13	employed	employ	VERB
ajst-32555	40	14	in	in	ADP
ajst-32555	40	15	this	this	DET
ajst-32555	40	16	study	study	NOUN
ajst-32555	40	17	is	be	AUX
ajst-32555	40	18	the	the	DET
ajst-32555	40	19	telco	telco	ADJ
ajst-32555	40	20	customer	customer	NOUN
ajst-32555	40	21	churn	churn	NOUN
ajst-32555	40	22	dataset	dataset	NOUN
ajst-32555	40	23	,	,	PUNCT
ajst-32555	40	24	originally	originally	ADV
ajst-32555	40	25	provided	provide	VERB
ajst-32555	40	26	by	by	ADP
ajst-32555	40	27	ibm	ibm	PROPN
ajst-32555	40	28	analytics	analytic	NOUN
ajst-32555	40	29	.	.	PUNCT
ajst-32555	41	1	the	the	DET
ajst-32555	41	2	dataset	dataset	NOUN
ajst-32555	41	3	features	feature	VERB
ajst-32555	41	4	a	a	DET
ajst-32555	41	5	clear	clear	ADJ
ajst-32555	41	6	structure	structure	NOUN
ajst-32555	41	7	,	,	PUNCT
ajst-32555	41	8	comprising	comprise	VERB
ajst-32555	41	9	7,043	7,043	NUM
ajst-32555	41	10	customer	customer	NOUN
ajst-32555	41	11	records	record	NOUN
ajst-32555	41	12	(	(	PUNCT
ajst-32555	41	13	rows	row	NOUN
ajst-32555	41	14	)	)	PUNCT
ajst-32555	41	15	and	and	CCONJ
ajst-32555	41	16	21	21	NUM
ajst-32555	41	17	features	feature	NOUN
ajst-32555	41	18	(	(	PUNCT
ajst-32555	41	19	columns	column	NOUN
ajst-32555	41	20	)	)	PUNCT
ajst-32555	41	21	describing	describe	VERB
ajst-32555	41	22	customer	customer	NOUN
ajst-32555	41	23	attributes	attribute	VERB
ajst-32555	41	24	[	[	PUNCT
ajst-32555	41	25	11	11	NUM
ajst-32555	41	26	]	]	PUNCT
ajst-32555	41	27	.	.	PUNCT
ajst-32555	42	1	the	the	DET
ajst-32555	42	2	target	target	NOUN
ajst-32555	42	3	variable	variable	NOUN
ajst-32555	42	4	,	,	PUNCT
ajst-32555	42	5	churn	churn	NOUN
ajst-32555	42	6	,	,	PUNCT
ajst-32555	42	7	is	be	AUX
ajst-32555	42	8	a	a	DET
ajst-32555	42	9	binary	binary	ADJ
ajst-32555	42	10	categorical	categorical	ADJ
ajst-32555	42	11	variable	variable	NOUN
ajst-32555	42	12	representing	represent	VERB
ajst-32555	42	13	whether	whether	SCONJ
ajst-32555	42	14	a	a	DET
ajst-32555	42	15	customer	customer	NOUN
ajst-32555	42	16	leaves	leave	VERB
ajst-32555	42	17	the	the	DET
ajst-32555	42	18	service	service	NOUN
ajst-32555	42	19	.	.	PUNCT
ajst-32555	43	1	the	the	DET
ajst-32555	43	2	preliminary	preliminary	ADJ
ajst-32555	43	3	analysis	analysis	NOUN
ajst-32555	43	4	highlighted	highlight	VERB
ajst-32555	43	5	class	class	NOUN
ajst-32555	43	6	imbalance	imbalance	NOUN
ajst-32555	43	7	,	,	PUNCT
ajst-32555	43	8	as	as	SCONJ
ajst-32555	43	9	about	about	ADP
ajst-32555	43	10	73.5	73.5	NUM
ajst-32555	43	11	%	%	NOUN
ajst-32555	43	12	of	of	ADP
ajst-32555	43	13	the	the	DET
ajst-32555	43	14	customers	customer	NOUN
ajst-32555	43	15	were	be	AUX
ajst-32555	43	16	classified	classify	VERB
ajst-32555	43	17	as	as	ADP
ajst-32555	43	18	non	non	ADJ
ajst-32555	43	19	-	-	ADJ
ajst-32555	43	20	churning	churning	ADJ
ajst-32555	43	21	customers	customer	NOUN
ajst-32555	43	22	and	and	CCONJ
ajst-32555	43	23	about	about	ADV
ajst-32555	43	24	26.5	26.5	NUM
ajst-32555	43	25	%	%	NOUN
ajst-32555	43	26	were	be	AUX
ajst-32555	43	27	classified	classify	VERB
ajst-32555	43	28	as	as	ADP
ajst-32555	43	29	customers	customer	NOUN
ajst-32555	43	30	who	who	PRON
ajst-32555	43	31	quit	quit	VERB
ajst-32555	43	32	[	[	PUNCT
ajst-32555	43	33	4	4	NUM
ajst-32555	43	34	]	]	PUNCT
ajst-32555	43	35	,	,	PUNCT
ajst-32555	43	36	prompting	prompt	VERB
ajst-32555	43	37	the	the	DET
ajst-32555	43	38	inclusion	inclusion	NOUN
ajst-32555	43	39	of	of	ADP
ajst-32555	43	40	multi	multi	ADJ
ajst-32555	43	41	-	-	ADJ
ajst-32555	43	42	dimensional	dimensional	ADJ
ajst-32555	43	43	evaluation	evaluation	NOUN
ajst-32555	43	44	metrics	metric	NOUN
ajst-32555	43	45	.	.	PUNCT
ajst-32555	44	1	table	table	NOUN
ajst-32555	44	2	1	1	NUM
ajst-32555	44	3	summarizes	summarize	NOUN
ajst-32555	44	4	several	several	ADJ
ajst-32555	44	5	important	important	ADJ
ajst-32555	44	6	features	feature	NOUN
ajst-32555	44	7	in	in	ADP
ajst-32555	44	8	the	the	DET
ajst-32555	44	9	dataset	dataset	NOUN
ajst-32555	44	10	,	,	PUNCT
ajst-32555	44	11	including	include	VERB
ajst-32555	44	12	data	datum	NOUN
ajst-32555	44	13	types	type	NOUN
ajst-32555	44	14	and	and	CCONJ
ajst-32555	44	15	example	example	NOUN
ajst-32555	44	16	values	value	NOUN
ajst-32555	44	17	.	.	PUNCT
ajst-32555	45	1	to	to	PART
ajst-32555	45	2	ensure	ensure	VERB
ajst-32555	45	3	data	datum	NOUN
ajst-32555	45	4	quality	quality	NOUN
ajst-32555	45	5	and	and	CCONJ
ajst-32555	45	6	meet	meet	VERB
ajst-32555	45	7	machine	machine	NOUN
ajst-32555	45	8	learning	learning	NOUN
ajst-32555	45	9	model	model	NOUN
ajst-32555	45	10	input	input	NOUN
ajst-32555	45	11	requirements	requirement	NOUN
ajst-32555	45	12	,	,	PUNCT
ajst-32555	45	13	all	all	DET
ajst-32555	45	14	operations	operation	NOUN
ajst-32555	45	15	were	be	AUX
ajst-32555	45	16	performed	perform	VERB
ajst-32555	45	17	in	in	ADP
ajst-32555	45	18	a	a	DET
ajst-32555	45	19	python	python	NOUN
ajst-32555	45	20	environment	environment	NOUN
ajst-32555	45	21	,	,	PUNCT
ajst-32555	45	22	primarily	primarily	ADV
ajst-32555	45	23	relying	rely	VERB
ajst-32555	45	24	on	on	ADP
ajst-32555	45	25	pandas	panda	NOUN
ajst-32555	45	26	and	and	CCONJ
ajst-32555	45	27	numpy	numpy	NOUN
ajst-32555	45	28	for	for	ADP
ajst-32555	45	29	data	data	NOUN
ajst-32555	45	30	manipulation	manipulation	NOUN
ajst-32555	45	31	,	,	PUNCT
ajst-32555	45	32	and	and	CCONJ
ajst-32555	45	33	utilizing	utilize	VERB
ajst-32555	45	34	the	the	DET
ajst-32555	45	35	scikit	scikit	NOUN
ajst-32555	45	36	-	-	PUNCT
ajst-32555	45	37	learn	learn	VERB
ajst-32555	45	38	library	library	NOUN
ajst-32555	45	39	for	for	ADP
ajst-32555	45	40	feature	feature	NOUN
ajst-32555	45	41	transformation	transformation	NOUN
ajst-32555	46	1	[	[	X
ajst-32555	46	2	2	2	NUM
ajst-32555	46	3	]	]	PUNCT
ajst-32555	46	4	.	.	PUNCT
ajst-32555	47	1	the	the	DET
ajst-32555	47	2	totalcharges	totalcharge	NOUN
ajst-32555	47	3	column	column	NOUN
ajst-32555	47	4	was	be	AUX
ajst-32555	47	5	incorrectly	incorrectly	ADV
ajst-32555	47	6	identified	identify	VERB
ajst-32555	47	7	as	as	ADP
ajst-32555	47	8	an	an	DET
ajst-32555	47	9	object	object	NOUN
ajst-32555	47	10	type	type	NOUN
ajst-32555	47	11	.	.	PUNCT
ajst-32555	48	1	this	this	PRON
ajst-32555	48	2	occurred	occur	VERB
ajst-32555	48	3	because	because	SCONJ
ajst-32555	48	4	the	the	DET
ajst-32555	48	5	field	field	NOUN
ajst-32555	48	6	contained	contain	VERB
ajst-32555	48	7	empty	empty	ADJ
ajst-32555	48	8	strings	string	NOUN
ajst-32555	48	9	for	for	ADP
ajst-32555	48	10	new	new	ADJ
ajst-32555	48	11	customers	customer	NOUN
ajst-32555	48	12	with	with	ADP
ajst-32555	48	13	a	a	DET
ajst-32555	48	14	tenure	tenure	NOUN
ajst-32555	48	15	of	of	ADP
ajst-32555	48	16	0	0	NUM
ajst-32555	48	17	.	.	PUNCT
ajst-32555	49	1	this	this	DET
ajst-32555	49	2	study	study	NOUN
ajst-32555	49	3	first	first	ADV
ajst-32555	49	4	used	use	VERB
ajst-32555	49	5	the	the	DET
ajst-32555	49	6	pd.to_numeric	pd.to_numeric	ADJ
ajst-32555	49	7	function	function	NOUN
ajst-32555	49	8	to	to	PART
ajst-32555	49	9	force	force	VERB
ajst-32555	49	10	the	the	DET
ajst-32555	49	11	column	column	NOUN
ajst-32555	49	12	into	into	ADP
ajst-32555	49	13	a	a	DET
ajst-32555	49	14	numeric	numeric	ADJ
ajst-32555	49	15	type	type	NOUN
ajst-32555	49	16	,	,	PUNCT
ajst-32555	49	17	setting	set	VERB
ajst-32555	49	18	unconvertible	unconvertible	ADJ
ajst-32555	49	19	empty	empty	ADJ
ajst-32555	49	20	strings	string	NOUN
ajst-32555	49	21	as	as	ADP
ajst-32555	49	22	missing	miss	VERB
ajst-32555	49	23	values	value	NOUN
ajst-32555	49	24	(	(	PUNCT
ajst-32555	49	25	nan	nan	PROPN
ajst-32555	49	26	)	)	PUNCT
ajst-32555	49	27	.	.	PUNCT
ajst-32555	50	1	analysis	analysis	NOUN
ajst-32555	50	2	revealed	reveal	VERB
ajst-32555	50	3	that	that	SCONJ
ajst-32555	50	4	11	11	NUM
ajst-32555	50	5	samples	sample	NOUN
ajst-32555	50	6	had	have	VERB
ajst-32555	50	7	their	their	PRON
ajst-32555	50	8	totalcharges	totalcharge	NOUN
ajst-32555	50	9	become	become	VERB
ajst-32555	50	10	missing	missing	ADJ
ajst-32555	50	11	,	,	PUNCT
ajst-32555	50	12	and	and	CCONJ
ajst-32555	50	13	all	all	DET
ajst-32555	50	14	these	these	DET
ajst-32555	50	15	samples	sample	NOUN
ajst-32555	50	16	had	have	VERB
ajst-32555	50	17	a	a	DET
ajst-32555	50	18	tenure	tenure	NOUN
ajst-32555	50	19	of	of	ADP
ajst-32555	50	20	0	0	NUM
ajst-32555	51	1	[	[	X
ajst-32555	51	2	8	8	NUM
ajst-32555	51	3	]	]	PUNCT
ajst-32555	51	4	.	.	PUNCT
ajst-32555	52	1	given	give	VERB
ajst-32555	52	2	that	that	SCONJ
ajst-32555	52	3	these	these	DET
ajst-32555	52	4	customers	customer	NOUN
ajst-32555	52	5	had	have	AUX
ajst-32555	52	6	not	not	PART
ajst-32555	52	7	yet	yet	ADV
ajst-32555	52	8	incurred	incur	VERB
ajst-32555	52	9	total	total	ADJ
ajst-32555	52	10	charges	charge	NOUN
ajst-32555	52	11	,	,	PUNCT
ajst-32555	52	12	filling	fill	VERB
ajst-32555	52	13	these	these	DET
ajst-32555	52	14	missing	miss	VERB
ajst-32555	52	15	values	value	NOUN
ajst-32555	52	16	with	with	ADP
ajst-32555	52	17	0	0	NUM
ajst-32555	52	18	was	be	AUX
ajst-32555	52	19	a	a	DET
ajst-32555	52	20	logical	logical	ADJ
ajst-32555	52	21	approach	approach	NOUN
ajst-32555	52	22	.	.	PUNCT
ajst-32555	53	1	beyond	beyond	ADP
ajst-32555	53	2	this	this	PRON
ajst-32555	53	3	,	,	PUNCT
ajst-32555	53	4	no	no	DET
ajst-32555	53	5	other	other	ADJ
ajst-32555	53	6	significant	significant	ADJ
ajst-32555	53	7	missing	miss	VERB
ajst-32555	53	8	values	value	NOUN
ajst-32555	53	9	required	require	VERB
ajst-32555	53	10	handling	handling	NOUN
ajst-32555	53	11	in	in	ADP
ajst-32555	53	12	the	the	DET
ajst-32555	53	13	dataset	dataset	NOUN
ajst-32555	53	14	[	[	X
ajst-32555	53	15	8	8	NUM
ajst-32555	53	16	]	]	PUNCT
ajst-32555	53	17	.	.	PUNCT
ajst-32555	54	1	binary	binary	PROPN
ajst-32555	54	2	categorical	categorical	ADJ
ajst-32555	54	3	features	feature	NOUN
ajst-32555	54	4	(	(	PUNCT
ajst-32555	54	5	e.g.	e.g.	ADV
ajst-32555	54	6	,	,	PUNCT
ajst-32555	54	7	yes	yes	INTJ
ajst-32555	54	8	/	/	SYM
ajst-32555	54	9	no	no	NOUN
ajst-32555	54	10	)	)	PUNCT
ajst-32555	54	11	were	be	AUX
ajst-32555	54	12	mapped	map	VERB
ajst-32555	54	13	to	to	ADP
ajst-32555	54	14	1	1	NUM
ajst-32555	54	15	and	and	CCONJ
ajst-32555	54	16	0	0	NUM
ajst-32555	54	17	.	.	PUNCT
ajst-32555	55	1	for	for	ADP
ajst-32555	55	2	nominal	nominal	ADJ
ajst-32555	55	3	variables	variable	NOUN
ajst-32555	55	4	with	with	ADP
ajst-32555	55	5	three	three	NUM
ajst-32555	55	6	or	or	CCONJ
ajst-32555	55	7	more	more	ADJ
ajst-32555	55	8	categories	category	NOUN
ajst-32555	55	9	(	(	PUNCT
ajst-32555	55	10	e.g.	e.g.	ADV
ajst-32555	55	11	,	,	PUNCT
ajst-32555	55	12	contract	contract	NOUN
ajst-32555	55	13	,	,	PUNCT
ajst-32555	55	14	internetservice	internetservice	NOUN
ajst-32555	55	15	,	,	PUNCT
ajst-32555	55	16	paymentmethod	paymentmethod	NOUN
ajst-32555	55	17	)	)	PUNCT
ajst-32555	55	18	,	,	PUNCT
ajst-32555	55	19	one	one	NUM
ajst-32555	55	20	-	-	PUNCT
ajst-32555	55	21	hot	hot	ADJ
ajst-32555	55	22	encoding	encoding	NOUN
ajst-32555	55	23	is	be	AUX
ajst-32555	55	24	applied	apply	VERB
ajst-32555	55	25	.	.	PUNCT
ajst-32555	56	1	this	this	DET
ajst-32555	56	2	method	method	NOUN
ajst-32555	56	3	converts	convert	VERB
ajst-32555	56	4	each	each	DET
ajst-32555	56	5	category	category	NOUN
ajst-32555	56	6	into	into	ADP
ajst-32555	56	7	a	a	DET
ajst-32555	56	8	new	new	ADJ
ajst-32555	56	9	binary	binary	ADJ
ajst-32555	56	10	feature	feature	NOUN
ajst-32555	56	11	column	column	NOUN
ajst-32555	56	12	,	,	PUNCT
ajst-32555	56	13	preventing	prevent	VERB
ajst-32555	56	14	the	the	DET
ajst-32555	56	15	model	model	NOUN
ajst-32555	56	16	from	from	ADP
ajst-32555	56	17	erroneously	erroneously	ADV
ajst-32555	56	18	assigning	assign	VERB
ajst-32555	56	19	ordinal	ordinal	ADJ
ajst-32555	56	20	relationships	relationship	NOUN
ajst-32555	56	21	between	between	ADP
ajst-32555	56	22	categories	category	NOUN
ajst-32555	56	23	.	.	PUNCT
ajst-32555	57	1	continuous	continuous	ADJ
ajst-32555	57	2	numerical	numerical	ADJ
ajst-32555	57	3	features	feature	NOUN
ajst-32555	57	4	like	like	ADP
ajst-32555	57	5	tenure	tenure	NOUN
ajst-32555	57	6	,	,	PUNCT
ajst-32555	57	7	monthlycharges	monthlycharge	NOUN
ajst-32555	57	8	,	,	PUNCT
ajst-32555	57	9	and	and	CCONJ
ajst-32555	57	10	totalcharges	totalcharge	VERB
ajst-32555	57	11	undergo	undergo	VERB
ajst-32555	57	12	standardization	standardization	NOUN
ajst-32555	57	13	.	.	PUNCT
ajst-32555	58	1	using	use	VERB
ajst-32555	58	2	scikit	scikit	NOUN
ajst-32555	58	3	-	-	PUNCT
ajst-32555	58	4	learn	learn	NOUN
ajst-32555	58	5	's	's	PART
ajst-32555	58	6	standardscaler	standardscaler	NOUN
ajst-32555	58	7	,	,	PUNCT
ajst-32555	58	8	each	each	DET
ajst-32555	58	9	feature	feature	NOUN
ajst-32555	58	10	's	's	PART
ajst-32555	58	11	data	datum	NOUN
ajst-32555	58	12	is	be	AUX
ajst-32555	58	13	transformed	transform	VERB
ajst-32555	58	14	into	into	ADP
ajst-32555	58	15	a	a	DET
ajst-32555	58	16	distribution	distribution	NOUN
ajst-32555	58	17	with	with	ADP
ajst-32555	58	18	a	a	DET
ajst-32555	58	19	mean	mean	NOUN
ajst-32555	58	20	of	of	ADP
ajst-32555	58	21	0	0	NUM
ajst-32555	58	22	and	and	CCONJ
ajst-32555	58	23	a	a	DET
ajst-32555	58	24	standard	standard	ADJ
ajst-32555	58	25	deviation	deviation	NOUN
ajst-32555	58	26	of	of	ADP
ajst-32555	58	27	1	1	NUM
ajst-32555	58	28	.	.	PUNCT
ajst-32555	59	1	this	this	DET
ajst-32555	59	2	step	step	NOUN
ajst-32555	59	3	is	be	AUX
ajst-32555	59	4	crucial	crucial	ADJ
ajst-32555	59	5	for	for	ADP
ajst-32555	59	6	models	model	NOUN
ajst-32555	59	7	sensitive	sensitive	ADJ
ajst-32555	59	8	to	to	PART
ajst-32555	59	9	feature	feature	VERB
ajst-32555	59	10	scaling	scaling	NOUN
ajst-32555	59	11	(	(	PUNCT
ajst-32555	59	12	e.g.	e.g.	ADV
ajst-32555	59	13	,	,	PUNCT
ajst-32555	59	14	logistic	logistic	ADJ
ajst-32555	59	15	regression	regression	NOUN
ajst-32555	59	16	and	and	CCONJ
ajst-32555	59	17	support	support	VERB
ajst-32555	59	18	vector	vector	NOUN
ajst-32555	59	19	machines	machine	NOUN
ajst-32555	59	20	)	)	PUNCT
ajst-32555	59	21	,	,	PUNCT
ajst-32555	59	22	eliminating	eliminate	VERB
ajst-32555	59	23	the	the	DET
ajst-32555	59	24	influence	influence	NOUN
ajst-32555	59	25	of	of	ADP
ajst-32555	59	26	units	unit	NOUN
ajst-32555	59	27	and	and	CCONJ
ajst-32555	59	28	ensuring	ensure	VERB
ajst-32555	59	29	all	all	DET
ajst-32555	59	30	features	feature	NOUN
ajst-32555	59	31	carry	carry	VERB
ajst-32555	59	32	equal	equal	ADJ
ajst-32555	59	33	118	118	NUM
ajst-32555	59	34	weight	weight	NOUN
ajst-32555	59	35	during	during	ADP
ajst-32555	59	36	model	model	NOUN
ajst-32555	59	37	training	training	NOUN
ajst-32555	59	38	[	[	X
ajst-32555	59	39	12	12	NUM
ajst-32555	59	40	]	]	PUNCT
ajst-32555	59	41	.	.	PUNCT
ajst-32555	60	1	to	to	PART
ajst-32555	60	2	objectively	objectively	ADV
ajst-32555	60	3	evaluate	evaluate	VERB
ajst-32555	60	4	model	model	NOUN
ajst-32555	60	5	generalization	generalization	NOUN
ajst-32555	60	6	,	,	PUNCT
ajst-32555	60	7	the	the	DET
ajst-32555	60	8	preprocessed	preprocesse	VERB
ajst-32555	60	9	dataset	dataset	NOUN
ajst-32555	60	10	was	be	AUX
ajst-32555	60	11	randomly	randomly	ADV
ajst-32555	60	12	split	split	VERB
ajst-32555	60	13	into	into	ADP
ajst-32555	60	14	training	training	NOUN
ajst-32555	60	15	and	and	CCONJ
ajst-32555	60	16	test	test	NOUN
ajst-32555	60	17	sets	set	NOUN
ajst-32555	60	18	at	at	ADP
ajst-32555	60	19	an	an	DET
ajst-32555	60	20	80:20	80:20	NUM
ajst-32555	60	21	ratio	ratio	NOUN
ajst-32555	60	22	.	.	PUNCT
ajst-32555	61	1	all	all	DET
ajst-32555	61	2	models	model	NOUN
ajst-32555	61	3	were	be	AUX
ajst-32555	61	4	trained	train	VERB
ajst-32555	61	5	exclusively	exclusively	ADV
ajst-32555	61	6	on	on	ADP
ajst-32555	61	7	the	the	DET
ajst-32555	61	8	training	training	NOUN
ajst-32555	61	9	set	set	NOUN
ajst-32555	61	10	,	,	PUNCT
ajst-32555	61	11	while	while	SCONJ
ajst-32555	61	12	final	final	ADJ
ajst-32555	61	13	performance	performance	NOUN
ajst-32555	61	14	evaluation	evaluation	NOUN
ajst-32555	61	15	was	be	AUX
ajst-32555	61	16	conducted	conduct	VERB
ajst-32555	61	17	on	on	ADP
ajst-32555	61	18	the	the	DET
ajst-32555	61	19	unseen	unseen	ADJ
ajst-32555	61	20	test	test	NOUN
ajst-32555	61	21	set	set	VERB
ajst-32555	61	22	to	to	PART
ajst-32555	61	23	ensure	ensure	VERB
ajst-32555	61	24	impartial	impartial	ADJ
ajst-32555	61	25	and	and	CCONJ
ajst-32555	61	26	reliable	reliable	ADJ
ajst-32555	61	27	assessment	assessment	NOUN
ajst-32555	62	1	[	[	X
ajst-32555	62	2	4	4	NUM
ajst-32555	62	3	]	]	PUNCT
ajst-32555	62	4	.	.	PUNCT
ajst-32555	63	1	table	table	NOUN
ajst-32555	63	2	1	1	NUM
ajst-32555	63	3	.	.	PUNCT
ajst-32555	63	4	dataset	dataset	ADJ
ajst-32555	63	5	feature	feature	NOUN
ajst-32555	63	6	description	description	NOUN
ajst-32555	63	7	.	.	PUNCT
ajst-32555	64	1	feature	feature	NOUN
ajst-32555	64	2	name	name	NOUN
ajst-32555	64	3	description	description	NOUN
ajst-32555	64	4	data	datum	NOUN
ajst-32555	64	5	type	type	NOUN
ajst-32555	64	6	example	example	NOUN
ajst-32555	64	7	values	value	NOUN
ajst-32555	64	8	gender	gender	NOUN
ajst-32555	64	9	customer	customer	NOUN
ajst-32555	64	10	gender	gender	NOUN
ajst-32555	64	11	categorical	categorical	ADJ
ajst-32555	64	12	male	male	NOUN
ajst-32555	64	13	,	,	PUNCT
ajst-32555	64	14	female	female	ADJ
ajst-32555	64	15	seniorcitizen	seniorcitizen	NOUN
ajst-32555	64	16	senior	senior	ADJ
ajst-32555	64	17	citizen	citizen	NOUN
ajst-32555	64	18	(	(	PUNCT
ajst-32555	64	19	65	65	NUM
ajst-32555	64	20	+	+	NOUN
ajst-32555	64	21	)	)	PUNCT
ajst-32555	64	22	binary	binary	NOUN
ajst-32555	64	23	1	1	NUM
ajst-32555	64	24	,	,	PUNCT
ajst-32555	64	25	0	0	NUM
ajst-32555	64	26	partner	partner	NOUN
ajst-32555	64	27	has	have	VERB
ajst-32555	64	28	a	a	DET
ajst-32555	64	29	partner	partner	NOUN
ajst-32555	64	30	?	?	PUNCT
ajst-32555	65	1	binary	binary	NOUN
ajst-32555	66	1	yes	yes	INTJ
ajst-32555	66	2	,	,	PUNCT
ajst-32555	66	3	no	no	DET
ajst-32555	66	4	dependents	dependent	NOUN
ajst-32555	66	5	has	have	VERB
ajst-32555	66	6	dependents	dependent	NOUN
ajst-32555	66	7	?	?	PUNCT
ajst-32555	67	1	binary	binary	NOUN
ajst-32555	68	1	yes	yes	INTJ
ajst-32555	68	2	,	,	PUNCT
ajst-32555	68	3	no	no	DET
ajst-32555	68	4	tenure	tenure	NOUN
ajst-32555	68	5	customer	customer	NOUN
ajst-32555	68	6	tenure	tenure	NOUN
ajst-32555	68	7	(	(	PUNCT
ajst-32555	68	8	months	month	NOUN
ajst-32555	68	9	)	)	PUNCT
ajst-32555	68	10	numeric	numeric	ADP
ajst-32555	68	11	1	1	NUM
ajst-32555	68	12	,	,	PUNCT
ajst-32555	68	13	34	34	NUM
ajst-32555	68	14	,	,	PUNCT
ajst-32555	68	15	72	72	NUM
ajst-32555	68	16	contract	contract	NOUN
ajst-32555	68	17	contract	contract	NOUN
ajst-32555	68	18	type	type	NOUN
ajst-32555	68	19	categorical	categorical	ADJ
ajst-32555	68	20	month	month	NOUN
ajst-32555	68	21	-	-	PUNCT
ajst-32555	68	22	to	to	ADP
ajst-32555	68	23	-	-	PUNCT
ajst-32555	68	24	month	month	NOUN
ajst-32555	68	25	,	,	PUNCT
ajst-32555	68	26	one	one	NUM
ajst-32555	68	27	year	year	NOUN
ajst-32555	68	28	,	,	PUNCT
ajst-32555	68	29	twoyear	twoyear	ADJ
ajst-32555	68	30	payment	payment	NOUN
ajst-32555	68	31	method	method	NOUN
ajst-32555	68	32	payment	payment	NOUN
ajst-32555	68	33	method	method	NOUN
ajst-32555	68	34	categorical	categorical	ADJ
ajst-32555	68	35	e.h	e.h	PROPN
ajst-32555	68	36	.	.	PROPN
ajst-32555	68	37	,	,	PUNCT
ajst-32555	68	38	electronic	electronic	ADJ
ajst-32555	68	39	check	check	NOUN
ajst-32555	68	40	,	,	PUNCT
ajst-32555	68	41	mailed	mail	VERB
ajst-32555	68	42	check	check	NOUN
ajst-32555	68	43	internet	internet	NOUN
ajst-32555	68	44	service	service	NOUN
ajst-32555	68	45	type	type	NOUN
ajst-32555	68	46	of	of	ADP
ajst-32555	68	47	internet	internet	NOUN
ajst-32555	68	48	service	service	NOUN
ajst-32555	68	49	categorical	categorical	ADJ
ajst-32555	68	50	dsl	dsl	PROPN
ajst-32555	68	51	,	,	PUNCT
ajst-32555	68	52	fiber	fiber	NOUN
ajst-32555	68	53	optic	optic	NOUN
ajst-32555	68	54	,	,	PUNCT
ajst-32555	68	55	no	no	DET
ajst-32555	68	56	online	online	ADJ
ajst-32555	68	57	security	security	NOUN
ajst-32555	68	58	subscribes	subscribe	NOUN
ajst-32555	68	59	to	to	ADP
ajst-32555	68	60	online	online	ADJ
ajst-32555	68	61	security	security	NOUN
ajst-32555	68	62	services	service	NOUN
ajst-32555	68	63	?	?	PUNCT
ajst-32555	69	1	binary	binary	NOUN
ajst-32555	69	2	yes	yes	INTJ
ajst-32555	69	3	,	,	PUNCT
ajst-32555	69	4	no	no	INTJ
ajst-32555	69	5	,	,	PUNCT
ajst-32555	69	6	no	no	DET
ajst-32555	69	7	internet	internet	NOUN
ajst-32555	69	8	service	service	NOUN
ajst-32555	69	9	techsupport	techsupport	NOUN
ajst-32555	69	10	subscribe	subscribe	VERB
ajst-32555	69	11	to	to	ADP
ajst-32555	69	12	technical	technical	ADJ
ajst-32555	69	13	support	support	NOUN
ajst-32555	69	14	services	service	NOUN
ajst-32555	69	15	?	?	PUNCT
ajst-32555	70	1	binary	binary	NOUN
ajst-32555	70	2	yes	yes	INTJ
ajst-32555	70	3	,	,	PUNCT
ajst-32555	70	4	no	no	INTJ
ajst-32555	70	5	,	,	PUNCT
ajst-32555	70	6	no	no	DET
ajst-32555	70	7	internet	internet	NOUN
ajst-32555	70	8	service	service	NOUN
ajst-32555	70	9	monthly	monthly	ADJ
ajst-32555	70	10	charges	charge	NOUN
ajst-32555	70	11	monthly	monthly	ADJ
ajst-32555	70	12	billing	billing	NOUN
ajst-32555	70	13	charges	charge	NOUN
ajst-32555	70	14	numeric	numeric	ADP
ajst-32555	70	15	29.85	29.85	NUM
ajst-32555	70	16	,	,	PUNCT
ajst-32555	70	17	56.95	56.95	NUM
ajst-32555	70	18	,	,	PUNCT
ajst-32555	70	19	108.15	108.15	NUM
ajst-32555	70	20	total	total	ADJ
ajst-32555	70	21	charges	charge	NOUN
ajst-32555	70	22	total	total	ADJ
ajst-32555	70	23	bill	bill	NOUN
ajst-32555	70	24	charges	charge	NOUN
ajst-32555	70	25	numeric	numeric	ADP
ajst-32555	70	26	29.85	29.85	NUM
ajst-32555	70	27	,	,	PUNCT
ajst-32555	70	28	1889.5	1889.5	NUM
ajst-32555	70	29	,	,	PUNCT
ajst-32555	70	30	8684.8	8684.8	NUM
ajst-32555	70	31	churn	churn	NOUN
ajst-32555	70	32	churn	churn	NOUN
ajst-32555	70	33	status	status	NOUN
ajst-32555	70	34	(	(	PUNCT
ajst-32555	70	35	target	target	NOUN
ajst-32555	70	36	variable	variable	NOUN
ajst-32555	70	37	)	)	PUNCT
ajst-32555	70	38	binary	binary	NOUN
ajst-32555	71	1	yes	yes	INTJ
ajst-32555	71	2	,	,	PUNCT
ajst-32555	71	3	no	no	DET
ajst-32555	71	4	2.2	2.2	NUM
ajst-32555	71	5	machine	machine	NOUN
ajst-32555	71	6	learning	learning	NOUN
ajst-32555	71	7	models	model	NOUN
ajst-32555	71	8	this	this	DET
ajst-32555	71	9	study	study	NOUN
ajst-32555	71	10	selected	select	VERB
ajst-32555	71	11	three	three	NUM
ajst-32555	71	12	classification	classification	NOUN
ajst-32555	71	13	algorithms	algorithm	NOUN
ajst-32555	71	14	that	that	PRON
ajst-32555	71	15	are	be	AUX
ajst-32555	71	16	widely	widely	ADV
ajst-32555	71	17	applied	apply	VERB
ajst-32555	71	18	and	and	CCONJ
ajst-32555	71	19	representative	representative	NOUN
ajst-32555	71	20	in	in	ADP
ajst-32555	71	21	both	both	CCONJ
ajst-32555	71	22	academia	academia	NOUN
ajst-32555	71	23	and	and	CCONJ
ajst-32555	71	24	industry	industry	NOUN
ajst-32555	71	25	.	.	PUNCT
ajst-32555	72	1	logistic	logistic	ADJ
ajst-32555	72	2	regression	regression	NOUN
ajst-32555	72	3	serves	serve	VERB
ajst-32555	72	4	as	as	ADP
ajst-32555	72	5	the	the	DET
ajst-32555	72	6	baseline	baseline	NOUN
ajst-32555	72	7	model	model	NOUN
ajst-32555	72	8	for	for	ADP
ajst-32555	72	9	this	this	DET
ajst-32555	72	10	comparative	comparative	ADJ
ajst-32555	72	11	study	study	NOUN
ajst-32555	72	12	.	.	PUNCT
ajst-32555	73	1	it	it	PRON
ajst-32555	73	2	is	be	AUX
ajst-32555	73	3	a	a	DET
ajst-32555	73	4	classical	classical	ADJ
ajst-32555	73	5	generalized	generalized	ADJ
ajst-32555	73	6	linear	linear	ADJ
ajst-32555	73	7	model	model	NOUN
ajst-32555	73	8	that	that	PRON
ajst-32555	73	9	maps	map	VERB
ajst-32555	73	10	the	the	DET
ajst-32555	73	11	output	output	NOUN
ajst-32555	73	12	of	of	ADP
ajst-32555	73	13	linear	linear	ADJ
ajst-32555	73	14	regression	regression	NOUN
ajst-32555	73	15	to	to	ADP
ajst-32555	73	16	the	the	DET
ajst-32555	73	17	interval	interval	NOUN
ajst-32555	73	18	(	(	PUNCT
ajst-32555	73	19	0	0	NUM
ajst-32555	73	20	,	,	PUNCT
ajst-32555	73	21	1	1	NUM
ajst-32555	73	22	)	)	PUNCT
ajst-32555	73	23	via	via	ADP
ajst-32555	73	24	the	the	DET
ajst-32555	73	25	sigmoid	sigmoid	NOUN
ajst-32555	73	26	function	function	NOUN
ajst-32555	73	27	,	,	PUNCT
ajst-32555	73	28	yielding	yield	VERB
ajst-32555	73	29	the	the	DET
ajst-32555	73	30	probability	probability	NOUN
ajst-32555	73	31	of	of	ADP
ajst-32555	73	32	a	a	DET
ajst-32555	73	33	sample	sample	NOUN
ajst-32555	73	34	belonging	belong	VERB
ajst-32555	73	35	to	to	ADP
ajst-32555	73	36	a	a	DET
ajst-32555	73	37	specific	specific	ADJ
ajst-32555	73	38	class	class	NOUN
ajst-32555	73	39	.	.	PUNCT
ajst-32555	74	1	its	its	PRON
ajst-32555	74	2	advantages	advantage	NOUN
ajst-32555	74	3	include	include	VERB
ajst-32555	74	4	simplicity	simplicity	NOUN
ajst-32555	74	5	,	,	PUNCT
ajst-32555	74	6	computational	computational	ADJ
ajst-32555	74	7	efficiency	efficiency	NOUN
ajst-32555	74	8	,	,	PUNCT
ajst-32555	74	9	fast	fast	ADJ
ajst-32555	74	10	training	training	NOUN
ajst-32555	74	11	,	,	PUNCT
ajst-32555	74	12	and	and	CCONJ
ajst-32555	74	13	highly	highly	ADV
ajst-32555	74	14	interpretable	interpretable	ADJ
ajst-32555	74	15	model	model	NOUN
ajst-32555	74	16	coefficients	coefficient	NOUN
ajst-32555	74	17	that	that	PRON
ajst-32555	74	18	intuitively	intuitively	ADV
ajst-32555	74	19	reflect	reflect	VERB
ajst-32555	74	20	the	the	DET
ajst-32555	74	21	direction	direction	NOUN
ajst-32555	74	22	and	and	CCONJ
ajst-32555	74	23	strength	strength	NOUN
ajst-32555	74	24	of	of	ADP
ajst-32555	74	25	each	each	DET
ajst-32555	74	26	feature	feature	NOUN
ajst-32555	74	27	's	's	PART
ajst-32555	74	28	influence	influence	NOUN
ajst-32555	74	29	on	on	ADP
ajst-32555	74	30	prediction	prediction	NOUN
ajst-32555	74	31	outcomes	outcome	NOUN
ajst-32555	74	32	[	[	X
ajst-32555	74	33	4	4	NUM
ajst-32555	74	34	]	]	PUNCT
ajst-32555	74	35	.	.	PUNCT
ajst-32555	75	1	random	random	ADJ
ajst-32555	75	2	forest	forest	NOUN
ajst-32555	75	3	is	be	AUX
ajst-32555	75	4	a	a	DET
ajst-32555	75	5	representative	representative	ADJ
ajst-32555	75	6	algorithm	algorithm	NOUN
ajst-32555	75	7	of	of	ADP
ajst-32555	75	8	the	the	DET
ajst-32555	75	9	bagging	bagging	NOUN
ajst-32555	75	10	concept	concept	NOUN
ajst-32555	75	11	in	in	ADP
ajst-32555	75	12	ensemble	ensemble	ADJ
ajst-32555	75	13	learning	learning	NOUN
ajst-32555	75	14	,	,	PUNCT
ajst-32555	75	15	renowned	renowne	VERB
ajst-32555	75	16	for	for	ADP
ajst-32555	75	17	its	its	PRON
ajst-32555	75	18	robust	robust	ADJ
ajst-32555	75	19	performance	performance	NOUN
ajst-32555	75	20	and	and	CCONJ
ajst-32555	75	21	resilience	resilience	NOUN
ajst-32555	75	22	against	against	ADP
ajst-32555	75	23	overfitting	overfitte	VERB
ajst-32555	75	24	[	[	X
ajst-32555	75	25	13	13	NUM
ajst-32555	75	26	]	]	PUNCT
ajst-32555	75	27	.	.	PUNCT
ajst-32555	76	1	it	it	PRON
ajst-32555	76	2	operates	operate	VERB
ajst-32555	76	3	by	by	ADP
ajst-32555	76	4	constructing	construct	VERB
ajst-32555	76	5	multiple	multiple	ADJ
ajst-32555	76	6	decision	decision	NOUN
ajst-32555	76	7	trees	tree	NOUN
ajst-32555	76	8	using	use	VERB
ajst-32555	76	9	bootstrap	bootstrap	NOUN
ajst-32555	76	10	aggregating	aggregating	NOUN
ajst-32555	76	11	(	(	PUNCT
ajst-32555	76	12	bootstrap	bootstrap	NOUN
ajst-32555	76	13	sampling	sampling	NOUN
ajst-32555	76	14	)	)	PUNCT
ajst-32555	76	15	.	.	PUNCT
ajst-32555	77	1	during	during	ADP
ajst-32555	77	2	each	each	DET
ajst-32555	77	3	tree	tree	NOUN
ajst-32555	77	4	's	's	PART
ajst-32555	77	5	growth	growth	NOUN
ajst-32555	77	6	,	,	PUNCT
ajst-32555	77	7	a	a	DET
ajst-32555	77	8	double	double	ADJ
ajst-32555	77	9	randomness	randomness	NOUN
ajst-32555	77	10	is	be	AUX
ajst-32555	77	11	introduced	introduce	VERB
ajst-32555	77	12	:	:	PUNCT
ajst-32555	77	13	not	not	PART
ajst-32555	77	14	only	only	ADV
ajst-32555	77	15	are	be	AUX
ajst-32555	77	16	the	the	DET
ajst-32555	77	17	training	training	NOUN
ajst-32555	77	18	samples	sample	NOUN
ajst-32555	77	19	randomly	randomly	ADV
ajst-32555	77	20	drawn	draw	VERB
ajst-32555	77	21	,	,	PUNCT
ajst-32555	77	22	but	but	CCONJ
ajst-32555	77	23	the	the	DET
ajst-32555	77	24	features	feature	NOUN
ajst-32555	77	25	selected	select	VERB
ajst-32555	77	26	for	for	ADP
ajst-32555	77	27	node	node	ADJ
ajst-32555	77	28	splitting	splitting	NOUN
ajst-32555	77	29	are	be	AUX
ajst-32555	77	30	also	also	ADV
ajst-32555	77	31	randomly	randomly	ADV
ajst-32555	77	32	sampled	sample	VERB
ajst-32555	77	33	from	from	ADP
ajst-32555	77	34	a	a	DET
ajst-32555	77	35	subset	subset	NOUN
ajst-32555	77	36	of	of	ADP
ajst-32555	77	37	all	all	DET
ajst-32555	77	38	features	feature	NOUN
ajst-32555	77	39	.	.	PUNCT
ajst-32555	78	1	the	the	DET
ajst-32555	78	2	final	final	ADJ
ajst-32555	78	3	prediction	prediction	NOUN
ajst-32555	78	4	is	be	AUX
ajst-32555	78	5	determined	determine	VERB
ajst-32555	78	6	by	by	ADP
ajst-32555	78	7	the	the	DET
ajst-32555	78	8	majority	majority	NOUN
ajst-32555	78	9	vote	vote	NOUN
ajst-32555	78	10	of	of	ADP
ajst-32555	78	11	all	all	DET
ajst-32555	78	12	decision	decision	NOUN
ajst-32555	78	13	trees	tree	NOUN
ajst-32555	78	14	.	.	PUNCT
ajst-32555	79	1	this	this	DET
ajst-32555	79	2	process	process	NOUN
ajst-32555	79	3	reduces	reduce	VERB
ajst-32555	79	4	the	the	DET
ajst-32555	79	5	correlation	correlation	NOUN
ajst-32555	79	6	between	between	ADP
ajst-32555	79	7	individual	individual	ADJ
ajst-32555	79	8	trees	tree	NOUN
ajst-32555	79	9	,	,	PUNCT
ajst-32555	79	10	effectively	effectively	ADV
ajst-32555	79	11	lowering	lower	VERB
ajst-32555	79	12	model	model	NOUN
ajst-32555	79	13	variance	variance	NOUN
ajst-32555	79	14	while	while	SCONJ
ajst-32555	79	15	enhancing	enhance	VERB
ajst-32555	79	16	overall	overall	ADV
ajst-32555	79	17	119	119	NUM
ajst-32555	79	18	stability	stability	NOUN
ajst-32555	79	19	and	and	CCONJ
ajst-32555	79	20	accuracy	accuracy	NOUN
ajst-32555	79	21	.	.	PUNCT
ajst-32555	80	1	additionally	additionally	ADV
ajst-32555	80	2	,	,	PUNCT
ajst-32555	80	3	random	random	ADJ
ajst-32555	80	4	forest	forest	NOUN
ajst-32555	80	5	has	have	VERB
ajst-32555	80	6	the	the	DET
ajst-32555	80	7	natural	natural	ADJ
ajst-32555	80	8	ability	ability	NOUN
ajst-32555	80	9	to	to	PART
ajst-32555	80	10	output	output	VERB
ajst-32555	80	11	feature	feature	NOUN
ajst-32555	80	12	importance	importance	NOUN
ajst-32555	80	13	rankings	ranking	NOUN
ajst-32555	80	14	.	.	PUNCT
ajst-32555	81	1	xgboost	xgboost	PROPN
ajst-32555	81	2	is	be	AUX
ajst-32555	81	3	an	an	DET
ajst-32555	81	4	efficient	efficient	ADJ
ajst-32555	81	5	,	,	PUNCT
ajst-32555	81	6	flexible	flexible	ADJ
ajst-32555	81	7	,	,	PUNCT
ajst-32555	81	8	and	and	CCONJ
ajst-32555	81	9	portable	portable	ADJ
ajst-32555	81	10	implementation	implementation	NOUN
ajst-32555	81	11	of	of	ADP
ajst-32555	81	12	gradient	gradient	ADJ
ajst-32555	81	13	boosting	boost	VERB
ajst-32555	81	14	algorithms	algorithm	NOUN
ajst-32555	81	15	,	,	PUNCT
ajst-32555	81	16	ranking	rank	VERB
ajst-32555	81	17	among	among	ADP
ajst-32555	81	18	the	the	DET
ajst-32555	81	19	most	most	ADV
ajst-32555	81	20	powerful	powerful	ADJ
ajst-32555	81	21	tools	tool	NOUN
ajst-32555	81	22	in	in	ADP
ajst-32555	81	23	contemporary	contemporary	ADJ
ajst-32555	81	24	machine	machine	NOUN
ajst-32555	81	25	learning	learning	NOUN
ajst-32555	81	26	.	.	PUNCT
ajst-32555	82	1	it	it	PRON
ajst-32555	82	2	consistently	consistently	ADV
ajst-32555	82	3	delivers	deliver	VERB
ajst-32555	82	4	outstanding	outstanding	ADJ
ajst-32555	82	5	performance	performance	NOUN
ajst-32555	82	6	in	in	ADP
ajst-32555	82	7	data	datum	NOUN
ajst-32555	82	8	science	science	NOUN
ajst-32555	82	9	competitions	competition	NOUN
ajst-32555	82	10	and	and	CCONJ
ajst-32555	82	11	real	real	ADJ
ajst-32555	82	12	-	-	PUNCT
ajst-32555	82	13	world	world	NOUN
ajst-32555	82	14	applications	application	NOUN
ajst-32555	82	15	.	.	PUNCT
ajst-32555	83	1	unlike	unlike	ADP
ajst-32555	83	2	random	random	ADJ
ajst-32555	83	3	forests	forest	NOUN
ajst-32555	83	4	that	that	PRON
ajst-32555	83	5	build	build	VERB
ajst-32555	83	6	trees	tree	NOUN
ajst-32555	83	7	in	in	ADP
ajst-32555	83	8	parallel	parallel	NOUN
ajst-32555	83	9	,	,	PUNCT
ajst-32555	83	10	xgboost	xgboost	ADV
ajst-32555	83	11	employs	employ	VERB
ajst-32555	83	12	a	a	DET
ajst-32555	83	13	sequential	sequential	ADJ
ajst-32555	83	14	approach	approach	NOUN
ajst-32555	83	15	,	,	PUNCT
ajst-32555	83	16	iteratively	iteratively	ADV
ajst-32555	83	17	constructing	construct	VERB
ajst-32555	83	18	a	a	DET
ajst-32555	83	19	series	series	NOUN
ajst-32555	83	20	of	of	ADP
ajst-32555	83	21	decision	decision	NOUN
ajst-32555	83	22	trees	tree	NOUN
ajst-32555	83	23	.	.	PUNCT
ajst-32555	84	1	each	each	DET
ajst-32555	84	2	new	new	ADJ
ajst-32555	84	3	tree	tree	NOUN
ajst-32555	84	4	is	be	AUX
ajst-32555	84	5	trained	train	VERB
ajst-32555	84	6	to	to	PART
ajst-32555	84	7	minimize	minimize	VERB
ajst-32555	84	8	the	the	DET
ajst-32555	84	9	residuals	residual	NOUN
ajst-32555	84	10	(	(	PUNCT
ajst-32555	84	11	i.e.	i.e.	X
ajst-32555	84	12	,	,	PUNCT
ajst-32555	84	13	errors	error	NOUN
ajst-32555	84	14	)	)	PUNCT
ajst-32555	84	15	from	from	ADP
ajst-32555	84	16	the	the	DET
ajst-32555	84	17	previous	previous	ADJ
ajst-32555	84	18	round	round	NOUN
ajst-32555	84	19	by	by	ADP
ajst-32555	84	20	optimizing	optimize	VERB
ajst-32555	84	21	along	along	ADP
ajst-32555	84	22	the	the	DET
ajst-32555	84	23	negative	negative	ADJ
ajst-32555	84	24	gradient	gradient	NOUN
ajst-32555	84	25	of	of	ADP
ajst-32555	84	26	the	the	DET
ajst-32555	84	27	loss	loss	NOUN
ajst-32555	84	28	function	function	NOUN
ajst-32555	84	29	,	,	PUNCT
ajst-32555	84	30	progressively	progressively	ADV
ajst-32555	84	31	correcting	correct	VERB
ajst-32555	84	32	errors	error	NOUN
ajst-32555	84	33	to	to	PART
ajst-32555	84	34	achieve	achieve	VERB
ajst-32555	84	35	high	high	ADJ
ajst-32555	84	36	prediction	prediction	NOUN
ajst-32555	84	37	accuracy	accuracy	NOUN
ajst-32555	84	38	.	.	PUNCT
ajst-32555	85	1	xgboost	xgboost	PROPN
ajst-32555	85	2	also	also	ADV
ajst-32555	85	3	incorporates	incorporate	VERB
ajst-32555	85	4	several	several	ADJ
ajst-32555	85	5	key	key	ADJ
ajst-32555	85	6	optimizations	optimization	NOUN
ajst-32555	85	7	,	,	PUNCT
ajst-32555	85	8	including	include	VERB
ajst-32555	85	9	regularization	regularization	NOUN
ajst-32555	85	10	terms	term	NOUN
ajst-32555	85	11	to	to	PART
ajst-32555	85	12	prevent	prevent	VERB
ajst-32555	85	13	overfitting	overfitting	NOUN
ajst-32555	85	14	and	and	CCONJ
ajst-32555	85	15	support	support	NOUN
ajst-32555	85	16	for	for	ADP
ajst-32555	85	17	parallel	parallel	ADJ
ajst-32555	85	18	computation	computation	NOUN
ajst-32555	85	19	to	to	PART
ajst-32555	85	20	enhance	enhance	VERB
ajst-32555	85	21	efficiency	efficiency	NOUN
ajst-32555	85	22	[	[	X
ajst-32555	85	23	14	14	NUM
ajst-32555	85	24	]	]	SYM
ajst-32555	85	25	.	.	PUNCT
ajst-32555	86	1	2.3	2.3	NUM
ajst-32555	86	2	evaluation	evaluation	NOUN
ajst-32555	86	3	metrics	metric	NOUN
ajst-32555	86	4	and	and	CCONJ
ajst-32555	86	5	protocol	protocol	NOUN
ajst-32555	86	6	given	give	VERB
ajst-32555	86	7	the	the	DET
ajst-32555	86	8	imbalanced	imbalanced	ADJ
ajst-32555	86	9	nature	nature	NOUN
ajst-32555	86	10	of	of	ADP
ajst-32555	86	11	the	the	DET
ajst-32555	86	12	dataset	dataset	NOUN
ajst-32555	86	13	,	,	PUNCT
ajst-32555	86	14	relying	rely	VERB
ajst-32555	86	15	solely	solely	ADV
ajst-32555	86	16	on	on	ADP
ajst-32555	86	17	accuracy	accuracy	NOUN
ajst-32555	86	18	metrics	metric	NOUN
ajst-32555	86	19	may	may	AUX
ajst-32555	86	20	be	be	AUX
ajst-32555	86	21	misleading	misleading	ADJ
ajst-32555	86	22	.	.	PUNCT
ajst-32555	87	1	for	for	ADP
ajst-32555	87	2	instance	instance	NOUN
ajst-32555	87	3	,	,	PUNCT
ajst-32555	87	4	a	a	DET
ajst-32555	87	5	model	model	NOUN
ajst-32555	87	6	predicting	predict	VERB
ajst-32555	87	7	all	all	DET
ajst-32555	87	8	customers	customer	NOUN
ajst-32555	87	9	as	as	ADP
ajst-32555	87	10	non	non	ADJ
ajst-32555	87	11	-	-	ADJ
ajst-32555	87	12	churn	churn	NOUN
ajst-32555	87	13	could	could	AUX
ajst-32555	87	14	achieve	achieve	VERB
ajst-32555	87	15	73.5	73.5	NUM
ajst-32555	87	16	%	%	NOUN
ajst-32555	87	17	accuracy	accuracy	NOUN
ajst-32555	87	18	,	,	PUNCT
ajst-32555	87	19	yet	yet	CCONJ
ajst-32555	87	20	remain	remain	VERB
ajst-32555	87	21	useless	useless	ADJ
ajst-32555	87	22	for	for	ADP
ajst-32555	87	23	identifying	identify	VERB
ajst-32555	87	24	churners	churner	NOUN
ajst-32555	87	25	.	.	PUNCT
ajst-32555	88	1	therefore	therefore	ADV
ajst-32555	88	2	,	,	PUNCT
ajst-32555	88	3	this	this	DET
ajst-32555	88	4	study	study	NOUN
ajst-32555	88	5	employs	employ	VERB
ajst-32555	88	6	a	a	DET
ajst-32555	88	7	comprehensive	comprehensive	ADJ
ajst-32555	88	8	set	set	NOUN
ajst-32555	88	9	of	of	ADP
ajst-32555	88	10	evaluation	evaluation	NOUN
ajst-32555	88	11	metrics	metric	NOUN
ajst-32555	88	12	that	that	PRON
ajst-32555	88	13	holistically	holistically	ADV
ajst-32555	88	14	reflect	reflect	VERB
ajst-32555	88	15	model	model	NOUN
ajst-32555	88	16	performance	performance	NOUN
ajst-32555	88	17	on	on	ADP
ajst-32555	88	18	imbalanced	imbalanced	ADJ
ajst-32555	88	19	data	datum	NOUN
ajst-32555	88	20	.	.	PUNCT
ajst-32555	89	1	confusion	confusion	NOUN
ajst-32555	89	2	matrix	matrix	NOUN
ajst-32555	89	3	:	:	PUNCT
ajst-32555	89	4	serves	serve	VERB
ajst-32555	89	5	as	as	ADP
ajst-32555	89	6	the	the	DET
ajst-32555	89	7	foundation	foundation	NOUN
ajst-32555	89	8	for	for	ADP
ajst-32555	89	9	all	all	DET
ajst-32555	89	10	classification	classification	NOUN
ajst-32555	89	11	metrics	metric	NOUN
ajst-32555	89	12	,	,	PUNCT
ajst-32555	89	13	comprising	comprise	VERB
ajst-32555	89	14	four	four	NUM
ajst-32555	89	15	core	core	NOUN
ajst-32555	89	16	components	component	NOUN
ajst-32555	89	17	:	:	PUNCT
ajst-32555	89	18	true	true	ADJ
ajst-32555	89	19	positives	positive	NOUN
ajst-32555	89	20	(	(	PUNCT
ajst-32555	89	21	tp	tp	NOUN
ajst-32555	89	22	)	)	PUNCT
ajst-32555	89	23	,	,	PUNCT
ajst-32555	89	24	true	true	ADJ
ajst-32555	89	25	negatives	negative	NOUN
ajst-32555	89	26	(	(	PUNCT
ajst-32555	89	27	tn	tn	NOUN
ajst-32555	89	28	)	)	PUNCT
ajst-32555	89	29	,	,	PUNCT
ajst-32555	89	30	false	false	ADJ
ajst-32555	89	31	positives	positive	NOUN
ajst-32555	89	32	(	(	PUNCT
ajst-32555	89	33	fp	fp	NOUN
ajst-32555	89	34	)	)	PUNCT
ajst-32555	89	35	,	,	PUNCT
ajst-32555	89	36	and	and	CCONJ
ajst-32555	89	37	false	false	ADJ
ajst-32555	89	38	negatives	negative	NOUN
ajst-32555	89	39	(	(	PUNCT
ajst-32555	89	40	fn	fn	NOUN
ajst-32555	89	41	)	)	PUNCT
ajst-32555	89	42	.	.	PUNCT
ajst-32555	90	1	accuracy	accuracy	NOUN
ajst-32555	90	2	:	:	PUNCT
ajst-32555	90	3	the	the	DET
ajst-32555	90	4	proportion	proportion	NOUN
ajst-32555	90	5	of	of	ADP
ajst-32555	90	6	correctly	correctly	ADV
ajst-32555	90	7	predicted	predict	VERB
ajst-32555	90	8	samples	sample	NOUN
ajst-32555	90	9	out	out	ADP
ajst-32555	90	10	of	of	ADP
ajst-32555	90	11	the	the	DET
ajst-32555	90	12	total	total	ADJ
ajst-32555	90	13	samples	sample	NOUN
ajst-32555	90	14	.	.	PUNCT
ajst-32555	91	1	accuracy=	accuracy=	NUM
ajst-32555	91	2	tp+tn	tp+tn	PRON
ajst-32555	91	3	tp+tn+fp+fn	tp+tn+fp+fn	PROPN
ajst-32555	91	4	precision	precision	NOUN
ajst-32555	91	5	:	:	PUNCT
ajst-32555	91	6	the	the	DET
ajst-32555	91	7	proportion	proportion	NOUN
ajst-32555	91	8	of	of	ADP
ajst-32555	91	9	customers	customer	NOUN
ajst-32555	91	10	who	who	PRON
ajst-32555	91	11	actually	actually	ADV
ajst-32555	91	12	churn	churn	VERB
ajst-32555	91	13	among	among	ADP
ajst-32555	91	14	all	all	DET
ajst-32555	91	15	customers	customer	NOUN
ajst-32555	91	16	predicted	predict	VERB
ajst-32555	91	17	by	by	ADP
ajst-32555	91	18	the	the	DET
ajst-32555	91	19	model	model	NOUN
ajst-32555	91	20	to	to	PART
ajst-32555	91	21	churn	churn	VERB
ajst-32555	91	22	.	.	PUNCT
ajst-32555	92	1	a	a	DET
ajst-32555	92	2	high	high	ADJ
ajst-32555	92	3	precision	precision	NOUN
ajst-32555	92	4	indicates	indicate	VERB
ajst-32555	92	5	less	less	ADJ
ajst-32555	92	6	waste	waste	NOUN
ajst-32555	92	7	in	in	ADP
ajst-32555	92	8	retention	retention	NOUN
ajst-32555	92	9	costs	cost	NOUN
ajst-32555	92	10	.	.	PUNCT
ajst-32555	93	1	precision=	precision=	NUM
ajst-32555	93	2	tp	tp	ADP
ajst-32555	93	3	tp+fp	tp+fp	PROPN
ajst-32555	93	4	recall	recall	PROPN
ajst-32555	93	5	:	:	PUNCT
ajst-32555	93	6	the	the	DET
ajst-32555	93	7	proportion	proportion	NOUN
ajst-32555	93	8	of	of	ADP
ajst-32555	93	9	all	all	PRON
ajst-32555	93	10	truly	truly	ADV
ajst-32555	93	11	churned	churn	VERB
ajst-32555	93	12	customers	customer	NOUN
ajst-32555	93	13	successfully	successfully	ADV
ajst-32555	93	14	identified	identify	VERB
ajst-32555	93	15	by	by	ADP
ajst-32555	93	16	the	the	DET
ajst-32555	93	17	model	model	NOUN
ajst-32555	93	18	.	.	PUNCT
ajst-32555	94	1	a	a	DET
ajst-32555	94	2	high	high	ADJ
ajst-32555	94	3	recall	recall	NOUN
ajst-32555	94	4	rate	rate	NOUN
ajst-32555	94	5	indicates	indicate	VERB
ajst-32555	94	6	coverage	coverage	NOUN
ajst-32555	94	7	of	of	ADP
ajst-32555	94	8	more	more	ADJ
ajst-32555	94	9	potential	potential	ADJ
ajst-32555	94	10	churn	churn	NOUN
ajst-32555	94	11	customers	customer	NOUN
ajst-32555	94	12	.	.	PUNCT
ajst-32555	95	1	recall=	recall=	NOUN
ajst-32555	95	2	tp	tp	NOUN
ajst-32555	95	3	tp+fn	tp+fn	NOUN
ajst-32555	95	4	f1	f1	NOUN
ajst-32555	95	5	score	score	NOUN
ajst-32555	95	6	:	:	PUNCT
ajst-32555	95	7	as	as	ADP
ajst-32555	95	8	the	the	DET
ajst-32555	95	9	harmonic	harmonic	ADJ
ajst-32555	95	10	mean	mean	NOUN
ajst-32555	95	11	of	of	ADP
ajst-32555	95	12	precision	precision	NOUN
ajst-32555	95	13	and	and	CCONJ
ajst-32555	95	14	recall	recall	NOUN
ajst-32555	95	15	,	,	PUNCT
ajst-32555	95	16	it	it	PRON
ajst-32555	95	17	provides	provide	VERB
ajst-32555	95	18	a	a	DET
ajst-32555	95	19	single	single	ADJ
ajst-32555	95	20	metric	metric	ADJ
ajst-32555	95	21	balancing	balancing	NOUN
ajst-32555	95	22	both	both	PRON
ajst-32555	95	23	.	.	PUNCT
ajst-32555	96	1	f1=2×	f1=2×	ADJ
ajst-32555	96	2	precision×recall	precision×recall	NOUN
ajst-32555	96	3	precision+recall	precision+recall	PROPN
ajst-32555	96	4	auc	auc	NOUN
ajst-32555	96	5	(	(	PUNCT
ajst-32555	96	6	area	area	NOUN
ajst-32555	96	7	under	under	ADP
ajst-32555	96	8	the	the	DET
ajst-32555	96	9	roc	roc	PROPN
ajst-32555	96	10	curve	curve	NOUN
ajst-32555	96	11	):	):	PUNCT
ajst-32555	96	12	assesses	assess	VERB
ajst-32555	96	13	the	the	DET
ajst-32555	96	14	model	model	NOUN
ajst-32555	96	15	's	's	PART
ajst-32555	96	16	skill	skill	NOUN
ajst-32555	96	17	in	in	ADP
ajst-32555	96	18	distinguishing	distinguish	VERB
ajst-32555	96	19	positive	positive	ADJ
ajst-32555	96	20	from	from	ADP
ajst-32555	96	21	negative	negative	ADJ
ajst-32555	96	22	samples	sample	NOUN
ajst-32555	96	23	across	across	ADP
ajst-32555	96	24	all	all	DET
ajst-32555	96	25	classification	classification	NOUN
ajst-32555	96	26	thresholds	threshold	NOUN
ajst-32555	96	27	.	.	PUNCT
ajst-32555	97	1	values	value	NOUN
ajst-32555	97	2	can	can	AUX
ajst-32555	97	3	range	range	VERB
ajst-32555	97	4	from	from	ADP
ajst-32555	97	5	0.5	0.5	NUM
ajst-32555	97	6	to	to	ADP
ajst-32555	97	7	1	1	NUM
ajst-32555	97	8	,	,	PUNCT
ajst-32555	97	9	with	with	ADP
ajst-32555	97	10	larger	large	ADJ
ajst-32555	97	11	values	value	NOUN
ajst-32555	97	12	indicating	indicate	VERB
ajst-32555	97	13	better	well	ADJ
ajst-32555	97	14	discrimination	discrimination	NOUN
ajst-32555	97	15	.	.	PUNCT
ajst-32555	98	1	auc	auc	NOUN
ajst-32555	98	2	is	be	AUX
ajst-32555	98	3	well	well	ADV
ajst-32555	98	4	-	-	PUNCT
ajst-32555	98	5	suited	suit	VERB
ajst-32555	98	6	to	to	PART
ajst-32555	98	7	evaluate	evaluate	VERB
ajst-32555	98	8	performance	performance	NOUN
ajst-32555	98	9	on	on	ADP
ajst-32555	98	10	imbalanced	imbalanced	ADJ
ajst-32555	98	11	datasets	dataset	NOUN
ajst-32555	98	12	.	.	PUNCT
ajst-32555	99	1	3	3	X
ajst-32555	99	2	.	.	NOUN
ajst-32555	99	3	results	result	VERB
ajst-32555	99	4	3.1	3.1	NUM
ajst-32555	99	5	comparative	comparative	ADJ
ajst-32555	99	6	performance	performance	NOUN
ajst-32555	99	7	of	of	ADP
ajst-32555	99	8	models	model	NOUN
ajst-32555	99	9	to	to	PART
ajst-32555	99	10	fairly	fairly	ADV
ajst-32555	99	11	evaluate	evaluate	VERB
ajst-32555	99	12	the	the	DET
ajst-32555	99	13	predictive	predictive	ADJ
ajst-32555	99	14	capabilities	capability	NOUN
ajst-32555	99	15	of	of	ADP
ajst-32555	99	16	logistic	logistic	ADJ
ajst-32555	99	17	regression	regression	NOUN
ajst-32555	99	18	,	,	PUNCT
ajst-32555	99	19	random	random	ADJ
ajst-32555	99	20	forest	forest	NOUN
ajst-32555	99	21	,	,	PUNCT
ajst-32555	99	22	and	and	CCONJ
ajst-32555	99	23	xgboost	xgboost	ADV
ajst-32555	99	24	,	,	PUNCT
ajst-32555	99	25	this	this	DET
ajst-32555	99	26	study	study	NOUN
ajst-32555	99	27	calculated	calculate	VERB
ajst-32555	99	28	the	the	DET
ajst-32555	99	29	aforementioned	aforementioned	ADJ
ajst-32555	99	30	performance	performance	NOUN
ajst-32555	99	31	metrics	metric	NOUN
ajst-32555	99	32	on	on	ADP
ajst-32555	99	33	the	the	DET
ajst-32555	99	34	reserved	reserved	ADJ
ajst-32555	99	35	20	20	NUM
ajst-32555	99	36	%	%	NOUN
ajst-32555	99	37	test	test	NOUN
ajst-32555	99	38	set	set	VERB
ajst-32555	99	39	.	.	PUNCT
ajst-32555	100	1	all	all	DET
ajst-32555	100	2	models	model	NOUN
ajst-32555	100	3	utilized	utilize	VERB
ajst-32555	100	4	their	their	PRON
ajst-32555	100	5	final	final	ADJ
ajst-32555	100	6	versions	version	NOUN
ajst-32555	100	7	trained	train	VERB
ajst-32555	100	8	on	on	ADP
ajst-32555	100	9	the	the	DET
ajst-32555	100	10	80	80	NUM
ajst-32555	100	11	%	%	NOUN
ajst-32555	100	12	training	training	NOUN
ajst-32555	100	13	set	set	NOUN
ajst-32555	100	14	.	.	PUNCT
ajst-32555	101	1	detailed	detailed	ADJ
ajst-32555	101	2	performance	performance	NOUN
ajst-32555	101	3	comparison	comparison	NOUN
ajst-32555	101	4	results	result	NOUN
ajst-32555	101	5	are	be	AUX
ajst-32555	101	6	shown	show	VERB
ajst-32555	101	7	in	in	ADP
ajst-32555	101	8	table	table	NOUN
ajst-32555	101	9	2	2	NUM
ajst-32555	101	10	.	.	PUNCT
ajst-32555	102	1	the	the	DET
ajst-32555	102	2	comparison	comparison	NOUN
ajst-32555	102	3	results	result	VERB
ajst-32555	102	4	in	in	ADP
ajst-32555	102	5	table	table	NOUN
ajst-32555	102	6	2	2	NUM
ajst-32555	102	7	reveal	reveal	VERB
ajst-32555	102	8	several	several	ADJ
ajst-32555	102	9	key	key	ADJ
ajst-32555	102	10	points	point	NOUN
ajst-32555	102	11	:	:	PUNCT
ajst-32555	102	12			ADJ
ajst-32555	102	13	benchmark	benchmark	ADJ
ajst-32555	102	14	model	model	NOUN
ajst-32555	102	15	performance	performance	NOUN
ajst-32555	102	16	:	:	PUNCT
ajst-32555	102	17	the	the	DET
ajst-32555	102	18	baseline	baseline	ADJ
ajst-32555	102	19	logistic	logistic	ADJ
ajst-32555	102	20	regression	regression	NOUN
ajst-32555	102	21	model	model	NOUN
ajst-32555	102	22	achieved	achieve	VERB
ajst-32555	102	23	an	an	DET
ajst-32555	102	24	accuracy	accuracy	NOUN
ajst-32555	102	25	of	of	ADP
ajst-32555	102	26	80.4	80.4	NUM
ajst-32555	102	27	%	%	NOUN
ajst-32555	102	28	and	and	CCONJ
ajst-32555	102	29	an	an	DET
ajst-32555	102	30	auc	auc	NOUN
ajst-32555	102	31	value	value	NOUN
ajst-32555	102	32	of	of	ADP
ajst-32555	102	33	0.843	0.843	NUM
ajst-32555	102	34	,	,	PUNCT
ajst-32555	102	35	indicating	indicate	VERB
ajst-32555	102	36	that	that	SCONJ
ajst-32555	102	37	even	even	ADV
ajst-32555	102	38	a	a	DET
ajst-32555	102	39	simple	simple	ADJ
ajst-32555	102	40	linear	linear	NOUN
ajst-32555	102	41	model	model	NOUN
ajst-32555	102	42	can	can	AUX
ajst-32555	102	43	capture	capture	VERB
ajst-32555	102	44	some	some	DET
ajst-32555	102	45	predictive	predictive	ADJ
ajst-32555	102	46	signals	signal	NOUN
ajst-32555	102	47	on	on	ADP
ajst-32555	102	48	this	this	DET
ajst-32555	102	49	dataset	dataset	NOUN
ajst-32555	102	50	.	.	PUNCT
ajst-32555	103	1	this	this	DET
ajst-32555	103	2	performance	performance	NOUN
ajst-32555	103	3	level	level	NOUN
ajst-32555	103	4	aligns	align	VERB
ajst-32555	103	5	with	with	ADP
ajst-32555	103	6	reports	report	NOUN
ajst-32555	103	7	of	of	ADP
ajst-32555	103	8	logistic	logistic	ADJ
ajst-32555	103	9	regression	regression	NOUN
ajst-32555	103	10	on	on	ADP
ajst-32555	103	11	this	this	DET
ajst-32555	103	12	dataset	dataset	NOUN
ajst-32555	103	13	.	.	PUNCT
ajst-32555	104	1	its	its	PRON
ajst-32555	104	2	recall	recall	NOUN
ajst-32555	104	3	(	(	PUNCT
ajst-32555	104	4	0.548	0.548	NUM
ajst-32555	104	5	)	)	PUNCT
ajst-32555	104	6	is	be	AUX
ajst-32555	104	7	relatively	relatively	ADV
ajst-32555	104	8	low	low	ADJ
ajst-32555	104	9	,	,	PUNCT
ajst-32555	104	10	meaning	mean	VERB
ajst-32555	104	11	the	the	DET
ajst-32555	104	12	model	model	NOUN
ajst-32555	104	13	missed	miss	VERB
ajst-32555	104	14	nearly	nearly	ADV
ajst-32555	104	15	half	half	NOUN
ajst-32555	104	16	of	of	ADP
ajst-32555	104	17	the	the	DET
ajst-32555	104	18	actual	actual	ADJ
ajst-32555	104	19	churn	churn	NOUN
ajst-32555	104	20	customers	customer	NOUN
ajst-32555	104	21	.	.	PUNCT
ajst-32555	105	1	120	120	NUM
ajst-32555	105	2			ADJ
ajst-32555	105	3	superiority	superiority	NOUN
ajst-32555	105	4	of	of	ADP
ajst-32555	105	5	ensemble	ensemble	ADJ
ajst-32555	105	6	models	model	NOUN
ajst-32555	105	7	:	:	PUNCT
ajst-32555	105	8	compared	compare	VERB
ajst-32555	105	9	to	to	ADP
ajst-32555	105	10	logistic	logistic	ADJ
ajst-32555	105	11	regression	regression	NOUN
ajst-32555	105	12	,	,	PUNCT
ajst-32555	105	13	two	two	NUM
ajst-32555	105	14	ensemble	ensemble	ADJ
ajst-32555	105	15	learning	learning	NOUN
ajst-32555	105	16	models	model	NOUN
ajst-32555	105	17	—	—	PUNCT
ajst-32555	105	18	random	random	ADJ
ajst-32555	105	19	forest	forest	NOUN
ajst-32555	105	20	and	and	CCONJ
ajst-32555	105	21	xgboost	xgboost	PROPN
ajst-32555	105	22	—	—	PUNCT
ajst-32555	105	23	demonstrated	demonstrate	VERB
ajst-32555	105	24	stronger	strong	ADJ
ajst-32555	105	25	performance	performance	NOUN
ajst-32555	105	26	across	across	ADP
ajst-32555	105	27	multiple	multiple	ADJ
ajst-32555	105	28	key	key	ADJ
ajst-32555	105	29	metrics	metric	NOUN
ajst-32555	105	30	,	,	PUNCT
ajst-32555	105	31	particularly	particularly	ADV
ajst-32555	105	32	in	in	ADP
ajst-32555	105	33	auc	auc	NOUN
ajst-32555	105	34	,	,	PUNCT
ajst-32555	105	35	a	a	DET
ajst-32555	105	36	core	core	NOUN
ajst-32555	105	37	indicator	indicator	NOUN
ajst-32555	105	38	measuring	measure	VERB
ajst-32555	105	39	overall	overall	ADJ
ajst-32555	105	40	model	model	NOUN
ajst-32555	105	41	discriminative	discriminative	NOUN
ajst-32555	105	42	capability	capability	NOUN
ajst-32555	105	43	.	.	PUNCT
ajst-32555	106	1	xgboost	xgboost	PROPN
ajst-32555	106	2	achieved	achieve	VERB
ajst-32555	106	3	the	the	DET
ajst-32555	106	4	highest	high	ADJ
ajst-32555	106	5	auc	auc	NOUN
ajst-32555	106	6	value	value	NOUN
ajst-32555	106	7	of	of	ADP
ajst-32555	106	8	0.857	0.857	NUM
ajst-32555	106	9	among	among	ADP
ajst-32555	106	10	the	the	DET
ajst-32555	106	11	three	three	NUM
ajst-32555	106	12	models	model	NOUN
ajst-32555	106	13	.	.	PUNCT
ajst-32555	107	1			PROPN
ajst-32555	107	2	xgboost	xgboost	PROPN
ajst-32555	107	3	's	's	PART
ajst-32555	107	4	leading	lead	VERB
ajst-32555	107	5	performance	performance	NOUN
ajst-32555	107	6	:	:	PUNCT
ajst-32555	107	7	between	between	ADP
ajst-32555	107	8	random	random	ADJ
ajst-32555	107	9	forest	forest	NOUN
ajst-32555	107	10	and	and	CCONJ
ajst-32555	107	11	xgboost	xgboost	PRON
ajst-32555	107	12	,	,	PUNCT
ajst-32555	107	13	xgboost	xgboost	PROPN
ajst-32555	107	14	outperformed	outperform	VERB
ajst-32555	107	15	in	in	ADP
ajst-32555	107	16	accuracy	accuracy	NOUN
ajst-32555	107	17	,	,	PUNCT
ajst-32555	107	18	precision	precision	NOUN
ajst-32555	107	19	,	,	PUNCT
ajst-32555	107	20	f1	f1	NOUN
ajst-32555	107	21	score	score	NOUN
ajst-32555	107	22	,	,	PUNCT
ajst-32555	107	23	and	and	CCONJ
ajst-32555	107	24	auc	auc	NOUN
ajst-32555	107	25	.	.	PUNCT
ajst-32555	108	1	this	this	PRON
ajst-32555	108	2	aligns	align	VERB
ajst-32555	108	3	with	with	ADP
ajst-32555	108	4	xgboost	xgboost	PROPN
ajst-32555	108	5	's	's	PART
ajst-32555	108	6	established	establish	VERB
ajst-32555	108	7	reputation	reputation	NOUN
ajst-32555	108	8	as	as	ADP
ajst-32555	108	9	an	an	DET
ajst-32555	108	10	advanced	advanced	ADJ
ajst-32555	108	11	implementation	implementation	NOUN
ajst-32555	108	12	of	of	ADP
ajst-32555	108	13	gradient	gradient	ADJ
ajst-32555	108	14	boosting	boost	VERB
ajst-32555	108	15	algorithms	algorithm	NOUN
ajst-32555	108	16	.	.	PUNCT
ajst-32555	109	1	while	while	SCONJ
ajst-32555	109	2	random	random	ADJ
ajst-32555	109	3	forest	forest	NOUN
ajst-32555	109	4	reduces	reduce	VERB
ajst-32555	109	5	variance	variance	NOUN
ajst-32555	109	6	through	through	ADP
ajst-32555	109	7	bagging	bagging	NOUN
ajst-32555	109	8	,	,	PUNCT
ajst-32555	109	9	xgboost	xgboost	X
ajst-32555	109	10	progressively	progressively	ADV
ajst-32555	109	11	minimizes	minimize	VERB
ajst-32555	109	12	bias	bias	NOUN
ajst-32555	109	13	via	via	ADP
ajst-32555	109	14	boosting	boost	VERB
ajst-32555	109	15	.	.	PUNCT
ajst-32555	110	1	its	its	PRON
ajst-32555	110	2	sequential	sequential	ADJ
ajst-32555	110	3	learning	learning	NOUN
ajst-32555	110	4	approach	approach	NOUN
ajst-32555	110	5	allows	allow	VERB
ajst-32555	110	6	it	it	PRON
ajst-32555	110	7	to	to	PART
ajst-32555	110	8	focus	focus	VERB
ajst-32555	110	9	more	more	ADV
ajst-32555	110	10	intently	intently	ADV
ajst-32555	110	11	on	on	ADP
ajst-32555	110	12	correcting	correct	VERB
ajst-32555	110	13	errors	error	NOUN
ajst-32555	110	14	from	from	ADP
ajst-32555	110	15	previous	previous	ADJ
ajst-32555	110	16	rounds	round	NOUN
ajst-32555	110	17	,	,	PUNCT
ajst-32555	110	18	achieving	achieve	VERB
ajst-32555	110	19	higher	high	ADJ
ajst-32555	110	20	precision	precision	NOUN
ajst-32555	110	21	across	across	ADP
ajst-32555	110	22	many	many	ADJ
ajst-32555	110	23	tasks	task	NOUN
ajst-32555	110	24	.	.	PUNCT
ajst-32555	111	1	table	table	NOUN
ajst-32555	111	2	2	2	NUM
ajst-32555	111	3	.	.	PUNCT
ajst-32555	111	4	machine	machine	NOUN
ajst-32555	111	5	learning	learning	NOUN
ajst-32555	111	6	model	model	NOUN
ajst-32555	111	7	performance	performance	NOUN
ajst-32555	111	8	comparison	comparison	NOUN
ajst-32555	111	9	.	.	PUNCT
ajst-32555	112	1	model	model	NOUN
ajst-32555	112	2	accuracy	accuracy	PROPN
ajst-32555	112	3	precision	precision	NOUN
ajst-32555	112	4	recall	recall	NOUN
ajst-32555	112	5	f1	f1	PROPN
ajst-32555	112	6	score	score	NOUN
ajst-32555	112	7	auc	auc	NOUN
ajst-32555	112	8	logistic	logistic	ADJ
ajst-32555	112	9	regression	regression	NOUN
ajst-32555	112	10	0.804	0.804	NUM
ajst-32555	112	11	0.651	0.651	NUM
ajst-32555	112	12	0.548	0.548	NUM
ajst-32555	112	13	0.595	0.595	NUM
ajst-32555	112	14	0.843	0.843	NUM
ajst-32555	112	15	random	random	ADJ
ajst-32555	112	16	forest	forest	NOUN
ajst-32555	112	17	0.795	0.795	NUM
ajst-32555	112	18	0.635	0.635	NUM
ajst-32555	112	19	0.579	0.579	NUM
ajst-32555	112	20	0.606	0.606	NUM
ajst-32555	112	21	0.849	0.849	NUM
ajst-32555	112	22	xgboost	xgboost	ADP
ajst-32555	112	23	0.812	0.812	NUM
ajst-32555	112	24	0.673	0.673	NUM
ajst-32555	112	25	0.585	0.585	NUM
ajst-32555	112	26	0.626	0.626	NUM
ajst-32555	112	27	0.857	0.857	NUM
ajst-32555	112	28	3.2	3.2	NUM
ajst-32555	112	29	feature	feature	NOUN
ajst-32555	112	30	importance	importance	NOUN
ajst-32555	112	31	analysis	analysis	NOUN
ajst-32555	112	32	of	of	ADP
ajst-32555	112	33	xgboost	xgboost	PRON
ajst-32555	112	34	to	to	PART
ajst-32555	112	35	investigate	investigate	VERB
ajst-32555	112	36	the	the	DET
ajst-32555	112	37	primary	primary	ADJ
ajst-32555	112	38	factors	factor	NOUN
ajst-32555	112	39	driving	drive	VERB
ajst-32555	112	40	customer	customer	NOUN
ajst-32555	112	41	churn	churn	NOUN
ajst-32555	112	42	,	,	PUNCT
ajst-32555	112	43	this	this	DET
ajst-32555	112	44	study	study	NOUN
ajst-32555	112	45	leveraged	leverage	VERB
ajst-32555	112	46	the	the	DET
ajst-32555	112	47	built	build	VERB
ajst-32555	112	48	-	-	PUNCT
ajst-32555	112	49	in	in	ADP
ajst-32555	112	50	feature	feature	NOUN
ajst-32555	112	51	importance	importance	NOUN
ajst-32555	112	52	evaluation	evaluation	NOUN
ajst-32555	112	53	function	function	NOUN
ajst-32555	112	54	within	within	ADP
ajst-32555	112	55	the	the	DET
ajst-32555	112	56	top	top	ADV
ajst-32555	112	57	-	-	PUNCT
ajst-32555	112	58	performing	perform	VERB
ajst-32555	112	59	xgboost	xgboost	ADJ
ajst-32555	112	60	model	model	NOUN
ajst-32555	112	61	.	.	PUNCT
ajst-32555	113	1	this	this	DET
ajst-32555	113	2	function	function	NOUN
ajst-32555	113	3	quantifies	quantify	VERB
ajst-32555	113	4	each	each	DET
ajst-32555	113	5	feature	feature	NOUN
ajst-32555	113	6	's	's	PART
ajst-32555	113	7	contribution	contribution	NOUN
ajst-32555	113	8	to	to	PART
ajst-32555	113	9	prediction	prediction	VERB
ajst-32555	113	10	outcomes	outcome	NOUN
ajst-32555	113	11	by	by	ADP
ajst-32555	113	12	calculating	calculate	VERB
ajst-32555	113	13	the	the	DET
ajst-32555	113	14	total	total	ADJ
ajst-32555	113	15	number	number	NOUN
ajst-32555	113	16	of	of	ADP
ajst-32555	113	17	times	time	NOUN
ajst-32555	113	18	it	it	PRON
ajst-32555	113	19	was	be	AUX
ajst-32555	113	20	used	use	VERB
ajst-32555	113	21	for	for	ADP
ajst-32555	113	22	node	node	NOUN
ajst-32555	113	23	splits	split	NOUN
ajst-32555	113	24	or	or	CCONJ
ajst-32555	113	25	the	the	DET
ajst-32555	113	26	total	total	ADJ
ajst-32555	113	27	gain	gain	NOUN
ajst-32555	113	28	across	across	ADP
ajst-32555	113	29	all	all	DET
ajst-32555	113	30	decision	decision	NOUN
ajst-32555	113	31	trees	tree	NOUN
ajst-32555	113	32	in	in	ADP
ajst-32555	113	33	the	the	DET
ajst-32555	113	34	model	model	NOUN
ajst-32555	113	35	.	.	PUNCT
ajst-32555	114	1	the	the	DET
ajst-32555	114	2	following	follow	VERB
ajst-32555	114	3	figure	figure	NOUN
ajst-32555	114	4	displays	display	VERB
ajst-32555	114	5	the	the	DET
ajst-32555	114	6	top	top	ADJ
ajst-32555	114	7	15	15	NUM
ajst-32555	114	8	influential	influential	ADJ
ajst-32555	114	9	features	feature	NOUN
ajst-32555	114	10	and	and	CCONJ
ajst-32555	114	11	their	their	PRON
ajst-32555	114	12	relative	relative	ADJ
ajst-32555	114	13	importance	importance	NOUN
ajst-32555	114	14	.	.	PUNCT
ajst-32555	115	1	figure	figure	NOUN
ajst-32555	115	2	1	1	NUM
ajst-32555	115	3	.	.	PUNCT
ajst-32555	115	4	top	top	ADJ
ajst-32555	115	5	15	15	NUM
ajst-32555	115	6	important	important	ADJ
ajst-32555	115	7	features	feature	NOUN
ajst-32555	115	8	output	output	NOUN
ajst-32555	115	9	by	by	ADP
ajst-32555	115	10	the	the	DET
ajst-32555	115	11	xgboost	xgboost	PROPN
ajst-32555	115	12	model	model	PROPN
ajst-32555	115	13	.	.	PUNCT
ajst-32555	116	1	the	the	DET
ajst-32555	116	2	analysis	analysis	NOUN
ajst-32555	116	3	reveals	reveal	VERB
ajst-32555	116	4	that	that	SCONJ
ajst-32555	116	5	the	the	DET
ajst-32555	116	6	most	most	ADV
ajst-32555	116	7	influential	influential	ADJ
ajst-32555	116	8	features	feature	NOUN
ajst-32555	116	9	for	for	ADP
ajst-32555	116	10	predicting	predict	VERB
ajst-32555	116	11	customer	customer	NOUN
ajst-32555	116	12	churn	churn	NOUN
ajst-32555	116	13	are	be	AUX
ajst-32555	116	14	highly	highly	ADV
ajst-32555	116	15	concentrated	concentrated	ADJ
ajst-32555	116	16	in	in	ADP
ajst-32555	116	17	contract	contract	NOUN
ajst-32555	116	18	details	detail	NOUN
ajst-32555	116	19	and	and	CCONJ
ajst-32555	116	20	service	service	NOUN
ajst-32555	116	21	selections	selection	NOUN
ajst-32555	116	22	,	,	PUNCT
ajst-32555	116	23	rather	rather	ADV
ajst-32555	116	24	than	than	ADP
ajst-32555	116	25	personal	personal	ADJ
ajst-32555	116	26	demographic	demographic	ADJ
ajst-32555	116	27	characteristics	characteristic	NOUN
ajst-32555	116	28	.	.	PUNCT
ajst-32555	117	1	the	the	DET
ajst-32555	117	2	top	top	ADV
ajst-32555	117	3	-	-	PUNCT
ajst-32555	117	4	ranked	rank	VERB
ajst-32555	117	5	key	key	ADJ
ajst-32555	117	6	features	feature	NOUN
ajst-32555	117	7	and	and	CCONJ
ajst-32555	117	8	their	their	PRON
ajst-32555	117	9	relative	relative	ADJ
ajst-32555	117	10	importance	importance	NOUN
ajst-32555	117	11	are	be	AUX
ajst-32555	117	12	as	as	SCONJ
ajst-32555	117	13	follows	follow	VERB
ajst-32555	117	14	:	:	PUNCT
ajst-32555	117	15	1	1	X
ajst-32555	117	16	.	.	X
ajst-32555	117	17	contract	contract	NOUN
ajst-32555	117	18	type	type	NOUN
ajst-32555	117	19	:	:	PUNCT
ajst-32555	117	20	a	a	DET
ajst-32555	117	21	two	two	NUM
ajst-32555	117	22	-	-	PUNCT
ajst-32555	117	23	year	year	NOUN
ajst-32555	117	24	contract	contract	NOUN
ajst-32555	117	25	(	(	PUNCT
ajst-32555	117	26	importance	importance	NOUN
ajst-32555	117	27	score	score	NOUN
ajst-32555	117	28	0.322	0.322	NUM
ajst-32555	117	29	)	)	PUNCT
ajst-32555	117	30	is	be	AUX
ajst-32555	117	31	the	the	DET
ajst-32555	117	32	most	most	ADV
ajst-32555	117	33	critical	critical	ADJ
ajst-32555	117	34	factor	factor	NOUN
ajst-32555	117	35	for	for	ADP
ajst-32555	117	36	stabilizing	stabilize	VERB
ajst-32555	117	37	customer	customer	NOUN
ajst-32555	117	38	relationships	relationship	NOUN
ajst-32555	117	39	.	.	PUNCT
ajst-32555	118	1	one	one	NUM
ajst-32555	118	2	-	-	PUNCT
ajst-32555	118	3	year	year	NOUN
ajst-32555	118	4	contract	contract	NOUN
ajst-32555	118	5	(	(	PUNCT
ajst-32555	118	6	0.133	0.133	NUM
ajst-32555	118	7	)	)	PUNCT
ajst-32555	118	8	.	.	PUNCT
ajst-32555	119	1	customers	customer	NOUN
ajst-32555	119	2	on	on	ADP
ajst-32555	119	3	monthly	monthly	ADJ
ajst-32555	119	4	payment	payment	NOUN
ajst-32555	119	5	plans	plan	NOUN
ajst-32555	119	6	constitute	constitute	VERB
ajst-32555	119	7	the	the	DET
ajst-32555	119	8	group	group	NOUN
ajst-32555	119	9	with	with	ADP
ajst-32555	119	10	the	the	DET
ajst-32555	119	11	highest	high	ADJ
ajst-32555	119	12	churn	churn	NOUN
ajst-32555	119	13	risk	risk	NOUN
ajst-32555	119	14	.	.	PUNCT
ajst-32555	120	1	2	2	X
ajst-32555	120	2	.	.	X
ajst-32555	120	3	internet	internet	NOUN
ajst-32555	120	4	service	service	NOUN
ajst-32555	120	5	type	type	NOUN
ajst-32555	120	6	:	:	PUNCT
ajst-32555	120	7	whether	whether	SCONJ
ajst-32555	120	8	using	use	VERB
ajst-32555	120	9	fiber	fiber	NOUN
ajst-32555	120	10	optic	optic	NOUN
ajst-32555	120	11	(	(	PUNCT
ajst-32555	120	12	internetservice_fiber	internetservice_fiber	NOUN
ajst-32555	120	13	optic	optic	NOUN
ajst-32555	120	14	)	)	PUNCT
ajst-32555	120	15	is	be	AUX
ajst-32555	120	16	the	the	DET
ajst-32555	120	17	second	second	ADJ
ajst-32555	120	18	most	most	ADV
ajst-32555	120	19	important	important	ADJ
ajst-32555	120	20	predictive	predictive	ADJ
ajst-32555	120	21	feature	feature	NOUN
ajst-32555	120	22	(	(	PUNCT
ajst-32555	120	23	0.216	0.216	NUM
ajst-32555	120	24	)	)	PUNCT
ajst-32555	120	25	.	.	PUNCT
ajst-32555	121	1	this	this	PRON
ajst-32555	121	2	typically	typically	ADV
ajst-32555	121	3	indicates	indicate	VERB
ajst-32555	121	4	that	that	SCONJ
ajst-32555	121	5	customers	customer	NOUN
ajst-32555	121	6	using	use	VERB
ajst-32555	121	7	fiber	fiber	NOUN
ajst-32555	121	8	121	121	NUM
ajst-32555	121	9	optic	optic	NOUN
ajst-32555	121	10	services	service	NOUN
ajst-32555	121	11	may	may	AUX
ajst-32555	121	12	exhibit	exhibit	VERB
ajst-32555	121	13	higher	high	ADJ
ajst-32555	121	14	churn	churn	NOUN
ajst-32555	121	15	tendencies	tendency	NOUN
ajst-32555	121	16	due	due	ADP
ajst-32555	121	17	to	to	ADP
ajst-32555	121	18	price	price	NOUN
ajst-32555	121	19	sensitivity	sensitivity	NOUN
ajst-32555	121	20	or	or	CCONJ
ajst-32555	121	21	service	service	NOUN
ajst-32555	121	22	stability	stability	NOUN
ajst-32555	121	23	concerns	concern	NOUN
ajst-32555	121	24	.	.	PUNCT
ajst-32555	122	1	3	3	X
ajst-32555	122	2	.	.	X
ajst-32555	122	3	tenure	tenure	NOUN
ajst-32555	122	4	:	:	PUNCT
ajst-32555	122	5	tenure	tenure	NOUN
ajst-32555	122	6	ranked	rank	VERB
ajst-32555	122	7	sixth	sixth	ADJ
ajst-32555	122	8	in	in	ADP
ajst-32555	122	9	importance	importance	NOUN
ajst-32555	122	10	(	(	PUNCT
ajst-32555	122	11	0.018	0.018	NUM
ajst-32555	122	12	)	)	PUNCT
ajst-32555	122	13	,	,	PUNCT
ajst-32555	122	14	indicating	indicate	VERB
ajst-32555	122	15	that	that	SCONJ
ajst-32555	122	16	relatively	relatively	ADV
ajst-32555	122	17	stable	stable	ADJ
ajst-32555	122	18	long	long	ADJ
ajst-32555	122	19	-	-	PUNCT
ajst-32555	122	20	term	term	NOUN
ajst-32555	122	21	customers	customer	NOUN
ajst-32555	122	22	have	have	VERB
ajst-32555	122	23	lower	low	ADJ
ajst-32555	122	24	churn	churn	NOUN
ajst-32555	122	25	probability	probability	NOUN
ajst-32555	122	26	,	,	PUNCT
ajst-32555	122	27	while	while	SCONJ
ajst-32555	122	28	newer	new	ADJ
ajst-32555	122	29	customers	customer	NOUN
ajst-32555	122	30	with	with	ADP
ajst-32555	122	31	shorter	short	ADJ
ajst-32555	122	32	tenure	tenure	NOUN
ajst-32555	122	33	are	be	AUX
ajst-32555	122	34	more	more	ADV
ajst-32555	122	35	prone	prone	ADJ
ajst-32555	122	36	to	to	PART
ajst-32555	122	37	churn	churn	VERB
ajst-32555	122	38	.	.	PUNCT
ajst-32555	123	1	4	4	X
ajst-32555	123	2	.	.	X
ajst-32555	123	3	payment	payment	NOUN
ajst-32555	123	4	method	method	NOUN
ajst-32555	123	5	:	:	PUNCT
ajst-32555	123	6	the	the	DET
ajst-32555	123	7	importance	importance	NOUN
ajst-32555	123	8	score	score	NOUN
ajst-32555	123	9	for	for	ADP
ajst-32555	123	10	electronic	electronic	ADJ
ajst-32555	123	11	check	check	NOUN
ajst-32555	123	12	(	(	PUNCT
ajst-32555	123	13	paymentmethod_electronic	paymentmethod_electronic	ADJ
ajst-32555	123	14	check	check	NOUN
ajst-32555	123	15	)	)	PUNCT
ajst-32555	123	16	is	be	AUX
ajst-32555	123	17	0.016	0.016	NUM
ajst-32555	123	18	,	,	PUNCT
ajst-32555	123	19	indicating	indicate	VERB
ajst-32555	123	20	that	that	SCONJ
ajst-32555	123	21	customers	customer	NOUN
ajst-32555	123	22	using	use	VERB
ajst-32555	123	23	this	this	DET
ajst-32555	123	24	payment	payment	NOUN
ajst-32555	123	25	method	method	NOUN
ajst-32555	123	26	face	face	VERB
ajst-32555	123	27	a	a	DET
ajst-32555	123	28	higher	high	ADJ
ajst-32555	123	29	churn	churn	NOUN
ajst-32555	123	30	risk	risk	NOUN
ajst-32555	123	31	.	.	PUNCT
ajst-32555	124	1	5	5	X
ajst-32555	124	2	.	.	X
ajst-32555	124	3	total	total	ADJ
ajst-32555	124	4	charges	charge	NOUN
ajst-32555	124	5	&	&	CCONJ
ajst-32555	124	6	monthly	monthly	ADJ
ajst-32555	124	7	charges	charge	NOUN
ajst-32555	124	8	:	:	PUNCT
ajst-32555	124	9	both	both	DET
ajst-32555	124	10	features	feature	NOUN
ajst-32555	124	11	scored	score	VERB
ajst-32555	124	12	0.012	0.012	NUM
ajst-32555	124	13	in	in	ADP
ajst-32555	124	14	importance	importance	NOUN
ajst-32555	124	15	.	.	PUNCT
ajst-32555	125	1	customers	customer	NOUN
ajst-32555	125	2	with	with	ADP
ajst-32555	125	3	lower	low	ADJ
ajst-32555	125	4	total	total	ADJ
ajst-32555	125	5	charges	charge	NOUN
ajst-32555	125	6	may	may	AUX
ajst-32555	125	7	be	be	AUX
ajst-32555	125	8	more	more	ADV
ajst-32555	125	9	prone	prone	ADJ
ajst-32555	125	10	to	to	PART
ajst-32555	125	11	churn	churn	VERB
ajst-32555	125	12	as	as	SCONJ
ajst-32555	125	13	they	they	PRON
ajst-32555	125	14	have	have	AUX
ajst-32555	125	15	n't	not	PART
ajst-32555	125	16	invested	invest	VERB
ajst-32555	125	17	heavily	heavily	ADV
ajst-32555	125	18	in	in	ADP
ajst-32555	125	19	the	the	DET
ajst-32555	125	20	service	service	NOUN
ajst-32555	125	21	yet	yet	ADV
ajst-32555	125	22	;	;	PUNCT
ajst-32555	125	23	similarly	similarly	ADV
ajst-32555	125	24	,	,	PUNCT
ajst-32555	125	25	customers	customer	NOUN
ajst-32555	125	26	with	with	ADP
ajst-32555	125	27	higher	high	ADJ
ajst-32555	125	28	monthly	monthly	ADJ
ajst-32555	125	29	charges	charge	NOUN
ajst-32555	125	30	also	also	ADV
ajst-32555	125	31	exhibit	exhibit	VERB
ajst-32555	125	32	greater	great	ADJ
ajst-32555	125	33	churn	churn	NOUN
ajst-32555	125	34	tendencies	tendency	NOUN
ajst-32555	125	35	.	.	PUNCT
ajst-32555	126	1	6	6	NUM
ajst-32555	126	2	.	.	X
ajst-32555	126	3	value	value	NOUN
ajst-32555	126	4	-	-	PUNCT
ajst-32555	126	5	added	add	VERB
ajst-32555	126	6	services	service	NOUN
ajst-32555	126	7	subscriptions	subscription	NOUN
ajst-32555	126	8	:	:	PUNCT
ajst-32555	126	9	whether	whether	SCONJ
ajst-32555	126	10	customers	customer	NOUN
ajst-32555	126	11	subscribe	subscribe	VERB
ajst-32555	126	12	to	to	ADP
ajst-32555	126	13	value	value	NOUN
ajst-32555	126	14	-	-	PUNCT
ajst-32555	126	15	added	add	VERB
ajst-32555	126	16	services	service	NOUN
ajst-32555	126	17	like	like	ADP
ajst-32555	126	18	onlinesecurity_yes	onlinesecurity_yes	PROPN
ajst-32555	126	19	(	(	PUNCT
ajst-32555	126	20	0.012	0.012	NUM
ajst-32555	126	21	)	)	PUNCT
ajst-32555	126	22	and	and	CCONJ
ajst-32555	126	23	techsupport_yes	techsupport_yes	NUM
ajst-32555	126	24	(	(	PUNCT
ajst-32555	126	25	0.012	0.012	NUM
ajst-32555	126	26	)	)	PUNCT
ajst-32555	126	27	also	also	ADV
ajst-32555	126	28	aids	aid	VERB
ajst-32555	126	29	in	in	ADP
ajst-32555	126	30	predicting	predict	VERB
ajst-32555	126	31	churn	churn	NOUN
ajst-32555	126	32	.	.	PUNCT
ajst-32555	127	1	customers	customer	NOUN
ajst-32555	127	2	not	not	PART
ajst-32555	127	3	subscribing	subscribe	VERB
ajst-32555	127	4	to	to	ADP
ajst-32555	127	5	these	these	DET
ajst-32555	127	6	services	service	NOUN
ajst-32555	127	7	typically	typically	ADV
ajst-32555	127	8	exhibit	exhibit	VERB
ajst-32555	127	9	higher	high	ADJ
ajst-32555	127	10	churn	churn	NOUN
ajst-32555	127	11	rates	rate	NOUN
ajst-32555	127	12	.	.	PUNCT
ajst-32555	128	1	these	these	DET
ajst-32555	128	2	findings	finding	NOUN
ajst-32555	128	3	align	align	VERB
ajst-32555	128	4	with	with	ADP
ajst-32555	128	5	previous	previous	ADJ
ajst-32555	128	6	exploratory	exploratory	ADJ
ajst-32555	128	7	data	datum	NOUN
ajst-32555	128	8	analysis	analysis	NOUN
ajst-32555	128	9	of	of	ADP
ajst-32555	128	10	this	this	DET
ajst-32555	128	11	dataset	dataset	NOUN
ajst-32555	128	12	.	.	PUNCT
ajst-32555	129	1	notably	notably	ADV
ajst-32555	129	2	,	,	PUNCT
ajst-32555	129	3	as	as	SCONJ
ajst-32555	129	4	shown	show	VERB
ajst-32555	129	5	in	in	ADP
ajst-32555	129	6	the	the	DET
ajst-32555	129	7	chart	chart	NOUN
ajst-32555	129	8	,	,	PUNCT
ajst-32555	129	9	demographic	demographic	ADJ
ajst-32555	129	10	characteristics	characteristic	NOUN
ajst-32555	129	11	like	like	ADP
ajst-32555	129	12	gender	gender	NOUN
ajst-32555	129	13	and	and	CCONJ
ajst-32555	129	14	relationship	relationship	NOUN
ajst-32555	129	15	status	status	NOUN
ajst-32555	129	16	did	do	AUX
ajst-32555	129	17	not	not	PART
ajst-32555	129	18	appear	appear	VERB
ajst-32555	129	19	among	among	ADP
ajst-32555	129	20	the	the	DET
ajst-32555	129	21	top	top	ADJ
ajst-32555	129	22	15	15	NUM
ajst-32555	129	23	most	most	ADV
ajst-32555	129	24	significant	significant	ADJ
ajst-32555	129	25	features	feature	NOUN
ajst-32555	129	26	.	.	PUNCT
ajst-32555	130	1	this	this	PRON
ajst-32555	130	2	strongly	strongly	ADV
ajst-32555	130	3	indicates	indicate	VERB
ajst-32555	130	4	that	that	SCONJ
ajst-32555	130	5	customer	customer	NOUN
ajst-32555	130	6	churn	churn	NOUN
ajst-32555	130	7	decisions	decision	NOUN
ajst-32555	130	8	are	be	AUX
ajst-32555	130	9	more	more	ADV
ajst-32555	130	10	influenced	influence	VERB
ajst-32555	130	11	by	by	ADP
ajst-32555	130	12	their	their	PRON
ajst-32555	130	13	service	service	NOUN
ajst-32555	130	14	contracts	contract	NOUN
ajst-32555	130	15	with	with	ADP
ajst-32555	130	16	the	the	DET
ajst-32555	130	17	company	company	NOUN
ajst-32555	130	18	,	,	PUNCT
ajst-32555	130	19	fees	fee	NOUN
ajst-32555	130	20	,	,	PUNCT
ajst-32555	130	21	and	and	CCONJ
ajst-32555	130	22	specific	specific	ADJ
ajst-32555	130	23	service	service	NOUN
ajst-32555	130	24	usage	usage	NOUN
ajst-32555	130	25	experiences	experience	NOUN
ajst-32555	130	26	than	than	ADP
ajst-32555	130	27	by	by	ADP
ajst-32555	130	28	their	their	PRON
ajst-32555	130	29	personal	personal	ADJ
ajst-32555	130	30	identity	identity	NOUN
ajst-32555	130	31	background	background	NOUN
ajst-32555	130	32	.	.	PUNCT
ajst-32555	131	1	4	4	X
ajst-32555	131	2	.	.	X
ajst-32555	131	3	discussion	discussion	VERB
ajst-32555	131	4	the	the	DET
ajst-32555	131	5	analysis	analysis	NOUN
ajst-32555	131	6	of	of	ADP
ajst-32555	131	7	feature	feature	NOUN
ajst-32555	131	8	importance	importance	NOUN
ajst-32555	131	9	provides	provide	VERB
ajst-32555	131	10	actionable	actionable	ADJ
ajst-32555	131	11	insights	insight	NOUN
ajst-32555	131	12	for	for	ADP
ajst-32555	131	13	telecom	telecom	NOUN
ajst-32555	131	14	operators	operator	NOUN
ajst-32555	131	15	seeking	seek	VERB
ajst-32555	131	16	to	to	PART
ajst-32555	131	17	reduce	reduce	VERB
ajst-32555	131	18	customer	customer	NOUN
ajst-32555	131	19	churn	churn	NOUN
ajst-32555	131	20	.	.	PUNCT
ajst-32555	132	1	this	this	DET
ajst-32555	132	2	section	section	NOUN
ajst-32555	132	3	translates	translate	VERB
ajst-32555	132	4	the	the	DET
ajst-32555	132	5	model	model	NOUN
ajst-32555	132	6	's	's	PART
ajst-32555	132	7	key	key	ADJ
ajst-32555	132	8	findings	finding	NOUN
ajst-32555	132	9	into	into	ADP
ajst-32555	132	10	data	data	NOUN
ajst-32555	132	11	-	-	PUNCT
ajst-32555	132	12	driven	drive	VERB
ajst-32555	132	13	business	business	NOUN
ajst-32555	132	14	strategies	strategy	NOUN
ajst-32555	132	15	.	.	PUNCT
ajst-32555	133	1	a	a	DET
ajst-32555	133	2	primary	primary	ADJ
ajst-32555	133	3	finding	finding	NOUN
ajst-32555	133	4	is	be	AUX
ajst-32555	133	5	that	that	SCONJ
ajst-32555	133	6	customers	customer	NOUN
ajst-32555	133	7	on	on	ADP
ajst-32555	133	8	month	month	NOUN
ajst-32555	133	9	-	-	PUNCT
ajst-32555	133	10	to	to	ADP
ajst-32555	133	11	-	-	PUNCT
ajst-32555	133	12	month	month	NOUN
ajst-32555	133	13	contracts	contract	NOUN
ajst-32555	133	14	are	be	AUX
ajst-32555	133	15	the	the	DET
ajst-32555	133	16	most	most	ADV
ajst-32555	133	17	prone	prone	ADJ
ajst-32555	133	18	to	to	PART
ajst-32555	133	19	churn	churn	VERB
ajst-32555	133	20	due	due	ADP
ajst-32555	133	21	to	to	ADP
ajst-32555	133	22	a	a	DET
ajst-32555	133	23	lack	lack	NOUN
ajst-32555	133	24	of	of	ADP
ajst-32555	133	25	long	long	ADJ
ajst-32555	133	26	-	-	PUNCT
ajst-32555	133	27	term	term	NOUN
ajst-32555	133	28	commitment	commitment	NOUN
ajst-32555	133	29	and	and	CCONJ
ajst-32555	133	30	low	low	ADJ
ajst-32555	133	31	switching	switching	NOUN
ajst-32555	133	32	costs	cost	NOUN
ajst-32555	133	33	.	.	PUNCT
ajst-32555	134	1	to	to	PART
ajst-32555	134	2	address	address	VERB
ajst-32555	134	3	this	this	PRON
ajst-32555	134	4	,	,	PUNCT
ajst-32555	134	5	operators	operator	NOUN
ajst-32555	134	6	should	should	AUX
ajst-32555	134	7	design	design	VERB
ajst-32555	134	8	targeted	target	VERB
ajst-32555	134	9	marketing	marketing	NOUN
ajst-32555	134	10	campaigns	campaign	NOUN
ajst-32555	134	11	that	that	PRON
ajst-32555	134	12	encourage	encourage	VERB
ajst-32555	134	13	these	these	DET
ajst-32555	134	14	customers	customer	NOUN
ajst-32555	134	15	to	to	PART
ajst-32555	134	16	transition	transition	VERB
ajst-32555	134	17	to	to	PART
ajst-32555	134	18	oneor	oneor	VERB
ajst-32555	134	19	two	two	NUM
ajst-32555	134	20	-	-	PUNCT
ajst-32555	134	21	year	year	NOUN
ajst-32555	134	22	contracts	contract	NOUN
ajst-32555	134	23	.	.	PUNCT
ajst-32555	135	1	effective	effective	ADJ
ajst-32555	135	2	incentives	incentive	NOUN
ajst-32555	135	3	could	could	AUX
ajst-32555	135	4	include	include	VERB
ajst-32555	135	5	first	first	ADJ
ajst-32555	135	6	-	-	PUNCT
ajst-32555	135	7	month	month	NOUN
ajst-32555	135	8	discounts	discount	NOUN
ajst-32555	135	9	,	,	PUNCT
ajst-32555	135	10	complimentary	complimentary	ADJ
ajst-32555	135	11	data	datum	NOUN
ajst-32555	135	12	allowances	allowance	NOUN
ajst-32555	135	13	,	,	PUNCT
ajst-32555	135	14	or	or	CCONJ
ajst-32555	135	15	service	service	NOUN
ajst-32555	135	16	upgrades	upgrade	NOUN
ajst-32555	135	17	to	to	PART
ajst-32555	135	18	lock	lock	VERB
ajst-32555	135	19	in	in	ADP
ajst-32555	135	20	customer	customer	NOUN
ajst-32555	135	21	relationships	relationship	NOUN
ajst-32555	135	22	and	and	CCONJ
ajst-32555	135	23	significantly	significantly	ADV
ajst-32555	135	24	reduce	reduce	VERB
ajst-32555	135	25	churn	churn	NOUN
ajst-32555	135	26	risk	risk	NOUN
ajst-32555	135	27	.	.	PUNCT
ajst-32555	136	1	similarly	similarly	ADV
ajst-32555	136	2	,	,	PUNCT
ajst-32555	136	3	new	new	ADJ
ajst-32555	136	4	customers	customer	NOUN
ajst-32555	136	5	with	with	ADP
ajst-32555	136	6	a	a	DET
ajst-32555	136	7	short	short	ADJ
ajst-32555	136	8	tenure	tenure	NOUN
ajst-32555	136	9	represent	represent	VERB
ajst-32555	136	10	another	another	DET
ajst-32555	136	11	high	high	ADJ
ajst-32555	136	12	-	-	PUNCT
ajst-32555	136	13	risk	risk	NOUN
ajst-32555	136	14	segment	segment	NOUN
ajst-32555	136	15	,	,	PUNCT
ajst-32555	136	16	as	as	SCONJ
ajst-32555	136	17	the	the	DET
ajst-32555	136	18	early	early	ADJ
ajst-32555	136	19	stage	stage	NOUN
ajst-32555	136	20	is	be	AUX
ajst-32555	136	21	a	a	DET
ajst-32555	136	22	critical	critical	ADJ
ajst-32555	136	23	period	period	NOUN
ajst-32555	136	24	for	for	ADP
ajst-32555	136	25	building	building	NOUN
ajst-32555	136	26	loyalty	loyalty	NOUN
ajst-32555	136	27	.	.	PUNCT
ajst-32555	137	1	it	it	PRON
ajst-32555	137	2	is	be	AUX
ajst-32555	137	3	recommended	recommend	VERB
ajst-32555	137	4	to	to	PART
ajst-32555	137	5	implement	implement	VERB
ajst-32555	137	6	a	a	DET
ajst-32555	137	7	robust	robust	ADJ
ajst-32555	137	8	new	new	ADJ
ajst-32555	137	9	customer	customer	NOUN
ajst-32555	137	10	onboarding	onboarding	NOUN
ajst-32555	137	11	program	program	NOUN
ajst-32555	137	12	.	.	PUNCT
ajst-32555	138	1	proactively	proactively	ADV
ajst-32555	138	2	engaging	engage	VERB
ajst-32555	138	3	with	with	ADP
ajst-32555	138	4	customers	customer	NOUN
ajst-32555	138	5	within	within	ADP
ajst-32555	138	6	the	the	DET
ajst-32555	138	7	first	first	ADJ
ajst-32555	138	8	three	three	NUM
ajst-32555	138	9	months	month	NOUN
ajst-32555	138	10	via	via	ADP
ajst-32555	138	11	sms	sms	PROPN
ajst-32555	138	12	,	,	PUNCT
ajst-32555	138	13	app	app	NOUN
ajst-32555	138	14	notifications	notification	NOUN
ajst-32555	138	15	,	,	PUNCT
ajst-32555	138	16	or	or	CCONJ
ajst-32555	138	17	phone	phone	NOUN
ajst-32555	138	18	calls	call	NOUN
ajst-32555	138	19	can	can	AUX
ajst-32555	138	20	ensure	ensure	VERB
ajst-32555	138	21	a	a	DET
ajst-32555	138	22	seamless	seamless	ADJ
ajst-32555	138	23	experience	experience	NOUN
ajst-32555	138	24	,	,	PUNCT
ajst-32555	138	25	address	address	VERB
ajst-32555	138	26	any	any	DET
ajst-32555	138	27	initial	initial	ADJ
ajst-32555	138	28	queries	query	NOUN
ajst-32555	138	29	,	,	PUNCT
ajst-32555	138	30	and	and	CCONJ
ajst-32555	138	31	reinforce	reinforce	VERB
ajst-32555	138	32	the	the	DET
ajst-32555	138	33	value	value	NOUN
ajst-32555	138	34	of	of	ADP
ajst-32555	138	35	their	their	PRON
ajst-32555	138	36	subscription	subscription	NOUN
ajst-32555	138	37	.	.	PUNCT
ajst-32555	139	1	the	the	DET
ajst-32555	139	2	balance	balance	NOUN
ajst-32555	139	3	between	between	ADP
ajst-32555	139	4	price	price	NOUN
ajst-32555	139	5	and	and	CCONJ
ajst-32555	139	6	perceived	perceive	VERB
ajst-32555	139	7	value	value	NOUN
ajst-32555	139	8	is	be	AUX
ajst-32555	139	9	also	also	ADV
ajst-32555	139	10	crucial	crucial	ADJ
ajst-32555	139	11	.	.	PUNCT
ajst-32555	140	1	churn	churn	NOUN
ajst-32555	140	2	rates	rate	NOUN
ajst-32555	140	3	are	be	AUX
ajst-32555	140	4	elevated	elevate	VERB
ajst-32555	140	5	among	among	ADP
ajst-32555	140	6	customers	customer	NOUN
ajst-32555	140	7	with	with	ADP
ajst-32555	140	8	high	high	ADJ
ajst-32555	140	9	-	-	PUNCT
ajst-32555	140	10	fee	fee	NOUN
ajst-32555	140	11	fiber	fiber	NOUN
ajst-32555	140	12	optic	optic	NOUN
ajst-32555	140	13	plans	plan	NOUN
ajst-32555	140	14	who	who	PRON
ajst-32555	140	15	do	do	AUX
ajst-32555	140	16	not	not	PART
ajst-32555	140	17	subscribe	subscribe	VERB
ajst-32555	140	18	to	to	ADP
ajst-32555	140	19	value	value	NOUN
ajst-32555	140	20	-	-	PUNCT
ajst-32555	140	21	added	add	VERB
ajst-32555	140	22	services	service	NOUN
ajst-32555	140	23	,	,	PUNCT
ajst-32555	140	24	suggesting	suggest	VERB
ajst-32555	140	25	they	they	PRON
ajst-32555	140	26	feel	feel	VERB
ajst-32555	140	27	the	the	DET
ajst-32555	140	28	service	service	NOUN
ajst-32555	140	29	value	value	NOUN
ajst-32555	140	30	does	do	AUX
ajst-32555	140	31	not	not	PART
ajst-32555	140	32	match	match	VERB
ajst-32555	140	33	the	the	DET
ajst-32555	140	34	premium	premium	NOUN
ajst-32555	140	35	price	price	NOUN
ajst-32555	140	36	.	.	PUNCT
ajst-32555	141	1	to	to	PART
ajst-32555	141	2	counter	counter	VERB
ajst-32555	141	3	this	this	PRON
ajst-32555	141	4	,	,	PUNCT
ajst-32555	141	5	operators	operator	NOUN
ajst-32555	141	6	should	should	AUX
ajst-32555	141	7	redesign	redesign	VERB
ajst-32555	141	8	their	their	PRON
ajst-32555	141	9	product	product	NOUN
ajst-32555	141	10	portfolio	portfolio	NOUN
ajst-32555	141	11	by	by	ADP
ajst-32555	141	12	bundling	bundle	VERB
ajst-32555	141	13	services	service	NOUN
ajst-32555	141	14	like	like	ADP
ajst-32555	141	15	online	online	ADJ
ajst-32555	141	16	security	security	NOUN
ajst-32555	141	17	and	and	CCONJ
ajst-32555	141	18	technical	technical	ADJ
ajst-32555	141	19	support	support	NOUN
ajst-32555	141	20	with	with	ADP
ajst-32555	141	21	these	these	DET
ajst-32555	141	22	premium	premium	NOUN
ajst-32555	141	23	plans	plan	NOUN
ajst-32555	141	24	.	.	PUNCT
ajst-32555	142	1	this	this	DET
ajst-32555	142	2	approach	approach	NOUN
ajst-32555	142	3	not	not	PART
ajst-32555	142	4	only	only	ADV
ajst-32555	142	5	increases	increase	VERB
ajst-32555	142	6	customer	customer	NOUN
ajst-32555	142	7	stickiness	stickiness	NOUN
ajst-32555	142	8	but	but	CCONJ
ajst-32555	142	9	also	also	ADV
ajst-32555	142	10	better	well	ADV
ajst-32555	142	11	justifies	justify	VERB
ajst-32555	142	12	the	the	DET
ajst-32555	142	13	higher	high	ADJ
ajst-32555	142	14	price	price	NOUN
ajst-32555	142	15	point	point	NOUN
ajst-32555	142	16	,	,	PUNCT
ajst-32555	142	17	enhancing	enhance	VERB
ajst-32555	142	18	overall	overall	ADJ
ajst-32555	142	19	perceived	perceive	VERB
ajst-32555	142	20	value	value	NOUN
ajst-32555	142	21	.	.	PUNCT
ajst-32555	143	1	finally	finally	ADV
ajst-32555	143	2	,	,	PUNCT
ajst-32555	143	3	the	the	DET
ajst-32555	143	4	payment	payment	NOUN
ajst-32555	143	5	method	method	NOUN
ajst-32555	143	6	significantly	significantly	ADV
ajst-32555	143	7	influences	influence	VERB
ajst-32555	143	8	retention	retention	NOUN
ajst-32555	143	9	,	,	PUNCT
ajst-32555	143	10	with	with	ADP
ajst-32555	143	11	electronic	electronic	ADJ
ajst-32555	143	12	check	check	NOUN
ajst-32555	143	13	users	user	NOUN
ajst-32555	143	14	exhibiting	exhibit	VERB
ajst-32555	143	15	a	a	DET
ajst-32555	143	16	higher	high	ADJ
ajst-32555	143	17	churn	churn	NOUN
ajst-32555	143	18	rate	rate	NOUN
ajst-32555	143	19	,	,	PUNCT
ajst-32555	143	20	likely	likely	ADJ
ajst-32555	143	21	due	due	ADP
ajst-32555	143	22	to	to	ADP
ajst-32555	143	23	payment	payment	NOUN
ajst-32555	143	24	failures	failure	NOUN
ajst-32555	143	25	or	or	CCONJ
ajst-32555	143	26	inconvenience	inconvenience	NOUN
ajst-32555	143	27	.	.	PUNCT
ajst-32555	144	1	operators	operator	NOUN
ajst-32555	144	2	should	should	AUX
ajst-32555	144	3	promote	promote	VERB
ajst-32555	144	4	more	more	ADV
ajst-32555	144	5	stable	stable	ADJ
ajst-32555	144	6	and	and	CCONJ
ajst-32555	144	7	convenient	convenient	ADJ
ajst-32555	144	8	automated	automate	VERB
ajst-32555	144	9	payment	payment	NOUN
ajst-32555	144	10	methods	method	NOUN
ajst-32555	144	11	,	,	PUNCT
ajst-32555	144	12	such	such	ADJ
ajst-32555	144	13	as	as	ADP
ajst-32555	144	14	credit	credit	NOUN
ajst-32555	144	15	cards	card	NOUN
ajst-32555	144	16	or	or	CCONJ
ajst-32555	144	17	bank	bank	NOUN
ajst-32555	144	18	auto	auto	NOUN
ajst-32555	144	19	-	-	PUNCT
ajst-32555	144	20	debit	debit	NOUN
ajst-32555	144	21	,	,	PUNCT
ajst-32555	144	22	by	by	ADP
ajst-32555	144	23	offering	offer	VERB
ajst-32555	144	24	small	small	ADJ
ajst-32555	144	25	bill	bill	NOUN
ajst-32555	144	26	discounts	discount	NOUN
ajst-32555	144	27	as	as	ADP
ajst-32555	144	28	an	an	DET
ajst-32555	144	29	incentive	incentive	NOUN
ajst-32555	144	30	.	.	PUNCT
ajst-32555	145	1	this	this	PRON
ajst-32555	145	2	can	can	AUX
ajst-32555	145	3	effectively	effectively	ADV
ajst-32555	145	4	reduce	reduce	VERB
ajst-32555	145	5	unintentional	unintentional	ADJ
ajst-32555	145	6	churn	churn	NOUN
ajst-32555	145	7	caused	cause	VERB
ajst-32555	145	8	by	by	ADP
ajst-32555	145	9	payment	payment	NOUN
ajst-32555	145	10	friction	friction	NOUN
ajst-32555	145	11	.	.	PUNCT
ajst-32555	146	1	in	in	ADP
ajst-32555	146	2	summary	summary	NOUN
ajst-32555	146	3	,	,	PUNCT
ajst-32555	146	4	the	the	DET
ajst-32555	146	5	model	model	NOUN
ajst-32555	146	6	not	not	PART
ajst-32555	146	7	only	only	ADV
ajst-32555	146	8	predicts	predict	NOUN
ajst-32555	146	9	who	who	PRON
ajst-32555	146	10	is	be	AUX
ajst-32555	146	11	likely	likely	ADJ
ajst-32555	146	12	to	to	PART
ajst-32555	146	13	churn	churn	VERB
ajst-32555	146	14	but	but	CCONJ
ajst-32555	146	15	also	also	ADV
ajst-32555	146	16	reveals	reveal	VERB
ajst-32555	146	17	the	the	DET
ajst-32555	146	18	underlying	underlie	VERB
ajst-32555	146	19	reasons	reason	NOUN
ajst-32555	146	20	through	through	ADP
ajst-32555	146	21	feature	feature	NOUN
ajst-32555	146	22	importance	importance	NOUN
ajst-32555	146	23	analysis	analysis	NOUN
ajst-32555	146	24	,	,	PUNCT
ajst-32555	146	25	providing	provide	VERB
ajst-32555	146	26	telecom	telecom	NOUN
ajst-32555	146	27	enterprises	enterprise	NOUN
ajst-32555	146	28	with	with	ADP
ajst-32555	146	29	a	a	DET
ajst-32555	146	30	clear	clear	ADJ
ajst-32555	146	31	roadmap	roadmap	NOUN
ajst-32555	146	32	for	for	ADP
ajst-32555	146	33	precise	precise	ADJ
ajst-32555	146	34	and	and	CCONJ
ajst-32555	146	35	effective	effective	ADJ
ajst-32555	146	36	interventions	intervention	NOUN
ajst-32555	146	37	.	.	PUNCT
ajst-32555	147	1	5	5	X
ajst-32555	147	2	.	.	X
ajst-32555	147	3	conclusion	conclusion	NOUN
ajst-32555	147	4	this	this	DET
ajst-32555	147	5	study	study	NOUN
ajst-32555	147	6	systematically	systematically	ADV
ajst-32555	147	7	compared	compare	VERB
ajst-32555	147	8	three	three	NUM
ajst-32555	147	9	machine	machine	NOUN
ajst-32555	147	10	learning	learning	NOUN
ajst-32555	147	11	models	model	NOUN
ajst-32555	147	12	for	for	ADP
ajst-32555	147	13	telecom	telecom	NOUN
ajst-32555	147	14	churn	churn	NOUN
ajst-32555	147	15	prediction	prediction	NOUN
ajst-32555	147	16	and	and	CCONJ
ajst-32555	147	17	validated	validate	VERB
ajst-32555	147	18	a	a	DET
ajst-32555	147	19	reproducible	reproducible	ADJ
ajst-32555	147	20	workflow	workflow	NOUN
ajst-32555	147	21	that	that	PRON
ajst-32555	147	22	transforms	transform	VERB
ajst-32555	147	23	model	model	NOUN
ajst-32555	147	24	outputs	output	NOUN
ajst-32555	147	25	into	into	ADP
ajst-32555	147	26	actionable	actionable	ADJ
ajst-32555	147	27	business	business	NOUN
ajst-32555	147	28	guidance	guidance	NOUN
ajst-32555	147	29	.	.	PUNCT
ajst-32555	148	1	the	the	DET
ajst-32555	148	2	findings	finding	NOUN
ajst-32555	148	3	confirm	confirm	VERB
ajst-32555	148	4	that	that	SCONJ
ajst-32555	148	5	ensemble	ensemble	ADJ
ajst-32555	148	6	learning	learning	NOUN
ajst-32555	148	7	models	model	NOUN
ajst-32555	148	8	significantly	significantly	ADV
ajst-32555	148	9	outperform	outperform	VERB
ajst-32555	148	10	traditional	traditional	ADJ
ajst-32555	148	11	122	122	NUM
ajst-32555	148	12	linear	linear	NOUN
ajst-32555	148	13	models	model	NOUN
ajst-32555	148	14	in	in	ADP
ajst-32555	148	15	this	this	DET
ajst-32555	148	16	context	context	NOUN
ajst-32555	148	17	.	.	PUNCT
ajst-32555	149	1	specifically	specifically	ADV
ajst-32555	149	2	,	,	PUNCT
ajst-32555	149	3	the	the	DET
ajst-32555	149	4	xgboost	xgboost	PROPN
ajst-32555	149	5	model	model	NOUN
ajst-32555	149	6	demonstrated	demonstrate	VERB
ajst-32555	149	7	the	the	DET
ajst-32555	149	8	most	most	ADV
ajst-32555	149	9	robust	robust	ADJ
ajst-32555	149	10	performance	performance	NOUN
ajst-32555	149	11	,	,	PUNCT
ajst-32555	149	12	achieving	achieve	VERB
ajst-32555	149	13	the	the	DET
ajst-32555	149	14	highest	high	ADJ
ajst-32555	149	15	scores	score	NOUN
ajst-32555	149	16	across	across	ADP
ajst-32555	149	17	all	all	DET
ajst-32555	149	18	key	key	ADJ
ajst-32555	149	19	metrics	metric	NOUN
ajst-32555	149	20	,	,	PUNCT
ajst-32555	149	21	including	include	VERB
ajst-32555	149	22	an	an	DET
ajst-32555	149	23	area	area	NOUN
ajst-32555	149	24	under	under	ADP
ajst-32555	149	25	the	the	DET
ajst-32555	149	26	curve	curve	NOUN
ajst-32555	149	27	(	(	PUNCT
ajst-32555	149	28	auc	auc	NOUN
ajst-32555	149	29	)	)	PUNCT
ajst-32555	149	30	of	of	ADP
ajst-32555	149	31	0.857	0.857	NUM
ajst-32555	149	32	.	.	PUNCT
ajst-32555	150	1	through	through	ADP
ajst-32555	150	2	feature	feature	NOUN
ajst-32555	150	3	importance	importance	NOUN
ajst-32555	150	4	analysis	analysis	NOUN
ajst-32555	150	5	of	of	ADP
ajst-32555	150	6	the	the	DET
ajst-32555	150	7	xgboost	xgboost	PROPN
ajst-32555	150	8	model	model	PROPN
ajst-32555	150	9	,	,	PUNCT
ajst-32555	150	10	the	the	DET
ajst-32555	150	11	study	study	NOUN
ajst-32555	150	12	identified	identify	VERB
ajst-32555	150	13	that	that	SCONJ
ajst-32555	150	14	the	the	DET
ajst-32555	150	15	most	most	ADV
ajst-32555	150	16	critical	critical	ADJ
ajst-32555	150	17	churn	churn	NOUN
ajst-32555	150	18	drivers	driver	NOUN
ajst-32555	150	19	relate	relate	VERB
ajst-32555	150	20	to	to	PART
ajst-32555	150	21	contract	contract	NOUN
ajst-32555	150	22	and	and	CCONJ
ajst-32555	150	23	financial	financial	ADJ
ajst-32555	150	24	status	status	NOUN
ajst-32555	150	25	—	—	PUNCT
ajst-32555	150	26	including	include	VERB
ajst-32555	150	27	contract	contract	NOUN
ajst-32555	150	28	type	type	NOUN
ajst-32555	150	29	,	,	PUNCT
ajst-32555	150	30	customer	customer	NOUN
ajst-32555	150	31	tenure	tenure	NOUN
ajst-32555	150	32	,	,	PUNCT
ajst-32555	150	33	and	and	CCONJ
ajst-32555	150	34	monthly	monthly	ADJ
ajst-32555	150	35	fees	fee	NOUN
ajst-32555	150	36	—	—	PUNCT
ajst-32555	150	37	rather	rather	ADV
ajst-32555	150	38	than	than	ADP
ajst-32555	150	39	demographic	demographic	ADJ
ajst-32555	150	40	characteristics	characteristic	NOUN
ajst-32555	150	41	.	.	PUNCT
ajst-32555	151	1	however	however	ADV
ajst-32555	151	2	,	,	PUNCT
ajst-32555	151	3	the	the	DET
ajst-32555	151	4	study	study	NOUN
ajst-32555	151	5	has	have	VERB
ajst-32555	151	6	certain	certain	ADJ
ajst-32555	151	7	limitations	limitation	NOUN
ajst-32555	151	8	that	that	PRON
ajst-32555	151	9	should	should	AUX
ajst-32555	151	10	be	be	AUX
ajst-32555	151	11	acknowledged	acknowledge	VERB
ajst-32555	151	12	.	.	PUNCT
ajst-32555	152	1	the	the	DET
ajst-32555	152	2	findings	finding	NOUN
ajst-32555	152	3	are	be	AUX
ajst-32555	152	4	based	base	VERB
ajst-32555	152	5	on	on	ADP
ajst-32555	152	6	a	a	DET
ajst-32555	152	7	single	single	ADJ
ajst-32555	152	8	,	,	PUNCT
ajst-32555	152	9	public	public	ADJ
ajst-32555	152	10	benchmark	benchmark	NOUN
ajst-32555	152	11	dataset	dataset	NOUN
ajst-32555	152	12	,	,	PUNCT
ajst-32555	152	13	which	which	PRON
ajst-32555	152	14	may	may	AUX
ajst-32555	152	15	limit	limit	VERB
ajst-32555	152	16	generalizability	generalizability	NOUN
ajst-32555	152	17	to	to	ADP
ajst-32555	152	18	specific	specific	ADJ
ajst-32555	152	19	commercial	commercial	ADJ
ajst-32555	152	20	environments	environment	NOUN
ajst-32555	152	21	.	.	PUNCT
ajst-32555	153	1	furthermore	furthermore	ADV
ajst-32555	153	2	,	,	PUNCT
ajst-32555	153	3	the	the	DET
ajst-32555	153	4	analysis	analysis	NOUN
ajst-32555	153	5	used	use	VERB
ajst-32555	153	6	a	a	DET
ajst-32555	153	7	static	static	ADJ
ajst-32555	153	8	data	datum	NOUN
ajst-32555	153	9	snapshot	snapshot	NOUN
ajst-32555	153	10	,	,	PUNCT
ajst-32555	153	11	failing	fail	VERB
ajst-32555	153	12	to	to	PART
ajst-32555	153	13	capture	capture	VERB
ajst-32555	153	14	the	the	DET
ajst-32555	153	15	dynamic	dynamic	ADJ
ajst-32555	153	16	,	,	PUNCT
ajst-32555	153	17	time	time	NOUN
ajst-32555	153	18	-	-	PUNCT
ajst-32555	153	19	series	series	NOUN
ajst-32555	153	20	nature	nature	NOUN
ajst-32555	153	21	of	of	ADP
ajst-32555	153	22	customer	customer	NOUN
ajst-32555	153	23	behavior	behavior	NOUN
ajst-32555	153	24	that	that	PRON
ajst-32555	153	25	may	may	AUX
ajst-32555	153	26	contain	contain	VERB
ajst-32555	153	27	critical	critical	ADJ
ajst-32555	153	28	churn	churn	NOUN
ajst-32555	153	29	signals	signal	NOUN
ajst-32555	153	30	.	.	PUNCT
ajst-32555	154	1	the	the	DET
ajst-32555	154	2	comparative	comparative	ADJ
ajst-32555	154	3	scope	scope	NOUN
ajst-32555	154	4	was	be	AUX
ajst-32555	154	5	limited	limit	VERB
ajst-32555	154	6	to	to	ADP
ajst-32555	154	7	three	three	NUM
ajst-32555	154	8	representative	representative	ADJ
ajst-32555	154	9	models	model	NOUN
ajst-32555	154	10	,	,	PUNCT
ajst-32555	154	11	and	and	CCONJ
ajst-32555	154	12	no	no	DET
ajst-32555	154	13	specialized	specialized	ADJ
ajst-32555	154	14	techniques	technique	NOUN
ajst-32555	154	15	like	like	ADP
ajst-32555	154	16	smote	smote	NOUN
ajst-32555	154	17	were	be	AUX
ajst-32555	154	18	applied	apply	VERB
ajst-32555	154	19	at	at	ADP
ajst-32555	154	20	the	the	DET
ajst-32555	154	21	data	data	NOUN
ajst-32555	154	22	level	level	NOUN
ajst-32555	154	23	to	to	PART
ajst-32555	154	24	address	address	VERB
ajst-32555	154	25	the	the	DET
ajst-32555	154	26	dataset	dataset	NOUN
ajst-32555	154	27	's	's	PART
ajst-32555	154	28	class	class	NOUN
ajst-32555	154	29	imbalance	imbalance	NOUN
ajst-32555	154	30	.	.	PUNCT
ajst-32555	155	1	future	future	ADJ
ajst-32555	155	2	research	research	NOUN
ajst-32555	155	3	should	should	AUX
ajst-32555	155	4	therefore	therefore	ADV
ajst-32555	155	5	aim	aim	VERB
ajst-32555	155	6	to	to	PART
ajst-32555	155	7	incorporate	incorporate	VERB
ajst-32555	155	8	time	time	NOUN
ajst-32555	155	9	-	-	PUNCT
ajst-32555	155	10	series	series	NOUN
ajst-32555	155	11	data	datum	NOUN
ajst-32555	155	12	to	to	PART
ajst-32555	155	13	capture	capture	VERB
ajst-32555	155	14	evolving	evolve	VERB
ajst-32555	155	15	customer	customer	NOUN
ajst-32555	155	16	behaviors	behavior	NOUN
ajst-32555	155	17	for	for	ADP
ajst-32555	155	18	more	more	ADV
ajst-32555	155	19	timely	timely	ADJ
ajst-32555	155	20	and	and	CCONJ
ajst-32555	155	21	accurate	accurate	ADJ
ajst-32555	155	22	predictions	prediction	NOUN
ajst-32555	155	23	.	.	PUNCT
ajst-32555	156	1	it	it	PRON
ajst-32555	156	2	is	be	AUX
ajst-32555	156	3	also	also	ADV
ajst-32555	156	4	recommended	recommend	VERB
ajst-32555	156	5	to	to	PART
ajst-32555	156	6	construct	construct	VERB
ajst-32555	156	7	a	a	DET
ajst-32555	156	8	business	business	NOUN
ajst-32555	156	9	-	-	PUNCT
ajst-32555	156	10	oriented	orient	VERB
ajst-32555	156	11	evaluation	evaluation	NOUN
ajst-32555	156	12	framework	framework	NOUN
ajst-32555	156	13	that	that	PRON
ajst-32555	156	14	weighs	weigh	VERB
ajst-32555	156	15	technical	technical	ADJ
ajst-32555	156	16	accuracy	accuracy	NOUN
ajst-32555	156	17	against	against	ADP
ajst-32555	156	18	factors	factor	NOUN
ajst-32555	156	19	like	like	ADP
ajst-32555	156	20	customer	customer	NOUN
ajst-32555	156	21	value	value	NOUN
ajst-32555	156	22	and	and	CCONJ
ajst-32555	156	23	retention	retention	NOUN
ajst-32555	156	24	costs	cost	NOUN
ajst-32555	156	25	to	to	PART
ajst-32555	156	26	identify	identify	VERB
ajst-32555	156	27	the	the	DET
ajst-32555	156	28	model	model	NOUN
ajst-32555	156	29	that	that	PRON
ajst-32555	156	30	delivers	deliver	VERB
ajst-32555	156	31	maximum	maximum	ADJ
ajst-32555	156	32	business	business	NOUN
ajst-32555	156	33	profit	profit	NOUN
ajst-32555	156	34	.	.	PUNCT
ajst-32555	157	1	finally	finally	ADV
ajst-32555	157	2	,	,	PUNCT
ajst-32555	157	3	enhancing	enhance	VERB
ajst-32555	157	4	the	the	DET
ajst-32555	157	5	interpretability	interpretability	NOUN
ajst-32555	157	6	of	of	ADP
ajst-32555	157	7	high	high	ADJ
ajst-32555	157	8	-	-	PUNCT
ajst-32555	157	9	performance	performance	NOUN
ajst-32555	157	10	models	model	NOUN
ajst-32555	157	11	like	like	ADP
ajst-32555	157	12	xgboost	xgboost	ADV
ajst-32555	157	13	is	be	AUX
ajst-32555	157	14	of	of	ADP
ajst-32555	157	15	vital	vital	ADJ
ajst-32555	157	16	importance	importance	NOUN
ajst-32555	157	17	.	.	PUNCT
ajst-32555	158	1	applying	apply	VERB
ajst-32555	158	2	advanced	advanced	ADJ
ajst-32555	158	3	interpretability	interpretability	NOUN
ajst-32555	158	4	tools	tool	NOUN
ajst-32555	158	5	can	can	AUX
ajst-32555	158	6	provide	provide	VERB
ajst-32555	158	7	customized	customized	ADJ
ajst-32555	158	8	churn	churn	NOUN
ajst-32555	158	9	reasons	reason	NOUN
ajst-32555	158	10	for	for	ADP
ajst-32555	158	11	each	each	DET
ajst-32555	158	12	high	high	ADJ
ajst-32555	158	13	-	-	PUNCT
ajst-32555	158	14	risk	risk	NOUN
ajst-32555	158	15	customer	customer	NOUN
ajst-32555	158	16	,	,	PUNCT
ajst-32555	158	17	allowing	allow	VERB
ajst-32555	158	18	for	for	ADP
ajst-32555	158	19	the	the	DET
ajst-32555	158	20	formulation	formulation	NOUN
ajst-32555	158	21	of	of	ADP
ajst-32555	158	22	more	more	ADV
ajst-32555	158	23	precise	precise	ADJ
ajst-32555	158	24	and	and	CCONJ
ajst-32555	158	25	targeted	target	VERB
ajst-32555	158	26	retention	retention	NOUN
ajst-32555	158	27	strategies	strategy	NOUN
ajst-32555	158	28	.	.	PUNCT
ajst-32555	159	1	references	reference	NOUN
ajst-32555	159	2	[	[	X
ajst-32555	159	3	1	1	NUM
ajst-32555	159	4	]	]	PUNCT
ajst-32555	159	5	kaggle	kaggle	NOUN
ajst-32555	159	6	.	.	PUNCT
ajst-32555	160	1	(	(	PUNCT
ajst-32555	160	2	2025	2025	NUM
ajst-32555	160	3	)	)	PUNCT
ajst-32555	160	4	.	.	PUNCT
ajst-32555	161	1	customer	customer	NOUN
ajst-32555	161	2	churn	churn	NOUN
ajst-32555	161	3	dataset	dataset	NOUN
ajst-32555	161	4	.	.	PUNCT
ajst-32555	162	1	https://	https://	PROPN
ajst-32555	162	2	www.kaggle.com/	www.kaggle.com/	VERB
ajst-32555	162	3	datasets/	datasets/	NOUN
ajst-32555	162	4	muhammadshahidazeem/	muhammadshahidazeem/	NUM
ajst-32555	162	5	customer	customer	NOUN
ajst-32555	162	6	-	-	PUNCT
ajst-32555	162	7	churn	churn	NOUN
ajst-32555	162	8	-	-	PUNCT
ajst-32555	162	9	dataset	dataset	NOUN
ajst-32555	162	10	.	.	PUNCT
ajst-32555	163	1	[	[	X
ajst-32555	163	2	2	2	NUM
ajst-32555	163	3	]	]	PUNCT
ajst-32555	163	4	google	google	PROPN
ajst-32555	163	5	dataset	dataset	NOUN
ajst-32555	163	6	search	search	NOUN
ajst-32555	163	7	.	.	PUNCT
ajst-32555	164	1	(	(	PUNCT
ajst-32555	164	2	2025	2025	NUM
ajst-32555	164	3	)	)	PUNCT
ajst-32555	164	4	.	.	PUNCT
ajst-32555	165	1	telco	telco	ADJ
ajst-32555	165	2	customer	customer	NOUN
ajst-32555	165	3	churn	churn	NOUN
ajst-32555	165	4	dataset	dataset	NOUN
ajst-32555	165	5	.	.	PUNCT
ajst-32555	166	1	https://toolbox	https://toolbox	PROPN
ajst-32555	166	2	.	.	PUNCT
ajst-32555	166	3	google	google	PROPN
ajst-32555	166	4	.	.	PUNCT
ajst-32555	167	1	com/	com/	NOUN
ajst-32555	167	2	datasetsearch/	datasetsearch/	NUM
ajst-32555	167	3	search	search	NOUN
ajst-32555	167	4	?	?	PUNCT
ajst-32555	168	1	query	query	NOUN
ajst-32555	168	2	=	=	SYM
ajst-32555	168	3	telco%20customer%20churn%20dataset%20	telco%20customer%20churn%20dataset%20	NOUN
ajst-32555	168	4	-	-	PUNCT
ajst-32555	168	5	site%3akaggle.com	site%3akaggle.com	NOUN
ajst-32555	168	6	.	.	PUNCT
ajst-32555	169	1	[	[	X
ajst-32555	169	2	3	3	NUM
ajst-32555	169	3	]	]	PUNCT
ajst-32555	169	4	xu	xu	PROPN
ajst-32555	169	5	,	,	PUNCT
ajst-32555	169	6	t.	t.	PROPN
ajst-32555	169	7	,	,	PUNCT
ajst-32555	169	8	ma	ma	PROPN
ajst-32555	169	9	,	,	PUNCT
ajst-32555	169	10	y.	y.	PROPN
ajst-32555	169	11	,	,	PUNCT
ajst-32555	169	12	&	&	CCONJ
ajst-32555	169	13	kim	kim	PROPN
ajst-32555	169	14	,	,	PUNCT
ajst-32555	169	15	k.	k.	PROPN
ajst-32555	169	16	(	(	PUNCT
ajst-32555	169	17	2021	2021	NUM
ajst-32555	169	18	)	)	PUNCT
ajst-32555	169	19	.	.	PUNCT
ajst-32555	170	1	telecom	telecom	NOUN
ajst-32555	170	2	churn	churn	NOUN
ajst-32555	170	3	prediction	prediction	NOUN
ajst-32555	170	4	system	system	NOUN
ajst-32555	170	5	based	base	VERB
ajst-32555	170	6	on	on	ADP
ajst-32555	170	7	ensemble	ensemble	ADJ
ajst-32555	170	8	learning	learning	NOUN
ajst-32555	170	9	using	use	VERB
ajst-32555	170	10	feature	feature	NOUN
ajst-32555	170	11	grouping	grouping	NOUN
ajst-32555	170	12	.	.	PUNCT
ajst-32555	171	1	applied	apply	VERB
ajst-32555	171	2	sciences	science	NOUN
ajst-32555	171	3	,	,	PUNCT
ajst-32555	171	4	11(11	11(11	NUM
ajst-32555	171	5	)	)	PUNCT
ajst-32555	171	6	,	,	PUNCT
ajst-32555	171	7	4742	4742	NUM
ajst-32555	171	8	.	.	PUNCT
ajst-32555	172	1	https://doi.org/10.3390/app11114742	https://doi.org/10.3390/app11114742	ADV
ajst-32555	172	2	.	.	PUNCT
ajst-32555	173	1	[	[	X
ajst-32555	173	2	4	4	NUM
ajst-32555	173	3	]	]	X
ajst-32555	173	4	wei	wei	PROPN
ajst-32555	173	5	,	,	PUNCT
ajst-32555	173	6	s.	s.	PROPN
ajst-32555	173	7	(	(	PUNCT
ajst-32555	173	8	2025	2025	NUM
ajst-32555	173	9	)	)	PUNCT
ajst-32555	173	10	.	.	PUNCT
ajst-32555	174	1	comparative	comparative	ADJ
ajst-32555	174	2	analysis	analysis	NOUN
ajst-32555	174	3	of	of	ADP
ajst-32555	174	4	machine	machine	NOUN
ajst-32555	174	5	learning	learning	NOUN
ajst-32555	174	6	models	model	NOUN
ajst-32555	174	7	for	for	ADP
ajst-32555	174	8	telecom	telecom	NOUN
ajst-32555	174	9	customer	customer	NOUN
ajst-32555	174	10	churn	churn	NOUN
ajst-32555	174	11	prediction	prediction	NOUN
ajst-32555	174	12	.	.	PUNCT
ajst-32555	175	1	theoretical	theoretical	ADJ
ajst-32555	175	2	and	and	CCONJ
ajst-32555	175	3	natural	natural	ADJ
ajst-32555	175	4	science	science	NOUN
ajst-32555	175	5	,	,	PUNCT
ajst-32555	175	6	134	134	NUM
ajst-32555	175	7	,	,	PUNCT
ajst-32555	175	8	24–30	24–30	NUM
ajst-32555	175	9	.	.	PUNCT
ajst-32555	176	1	https://www.ewadirect	https://www.ewadirect	ADJ
ajst-32555	176	2	.	.	PUNCT
ajst-32555	177	1	com/	com/	NOUN
ajst-32555	177	2	proceedings/	proceedings/	NUM
ajst-32555	177	3	tns/	tns/	NUM
ajst-32555	177	4	article/	article/	NUM
ajst-32555	177	5	view/26483	view/26483	NOUN
ajst-32555	177	6	.	.	PUNCT
ajst-32555	178	1	[	[	X
ajst-32555	178	2	5	5	NUM
ajst-32555	178	3	]	]	X
ajst-32555	178	4	chang	chang	PROPN
ajst-32555	178	5	,	,	PUNCT
ajst-32555	178	6	v.	v.	PROPN
ajst-32555	178	7	,	,	PUNCT
ajst-32555	178	8	hall	hall	PROPN
ajst-32555	178	9	,	,	PUNCT
ajst-32555	178	10	k.	k.	PROPN
ajst-32555	178	11	,	,	PUNCT
ajst-32555	178	12	xu	xu	PROPN
ajst-32555	178	13	,	,	PUNCT
ajst-32555	178	14	q.	q.	PROPN
ajst-32555	178	15	a.	a.	PROPN
ajst-32555	178	16	,	,	PUNCT
ajst-32555	178	17	amao	amao	ADJ
ajst-32555	178	18	,	,	PUNCT
ajst-32555	178	19	f.	f.	PROPN
ajst-32555	178	20	o.	o.	PROPN
ajst-32555	178	21	,	,	PUNCT
ajst-32555	178	22	ganatra	ganatra	PROPN
ajst-32555	178	23	,	,	PUNCT
ajst-32555	178	24	m.	m.	NOUN
ajst-32555	178	25	a.	a.	PROPN
ajst-32555	178	26	,	,	PUNCT
ajst-32555	178	27	&	&	CCONJ
ajst-32555	178	28	benson	benson	PROPN
ajst-32555	178	29	,	,	PUNCT
ajst-32555	178	30	v.	v.	PROPN
ajst-32555	178	31	(	(	PUNCT
ajst-32555	178	32	2024	2024	NUM
ajst-32555	178	33	)	)	PUNCT
ajst-32555	178	34	.	.	PUNCT
ajst-32555	179	1	prediction	prediction	NOUN
ajst-32555	179	2	of	of	ADP
ajst-32555	179	3	customer	customer	NOUN
ajst-32555	179	4	churn	churn	NOUN
ajst-32555	179	5	behavior	behavior	NOUN
ajst-32555	179	6	in	in	ADP
ajst-32555	179	7	the	the	DET
ajst-32555	179	8	telecommunication	telecommunication	NOUN
ajst-32555	179	9	industry	industry	NOUN
ajst-32555	179	10	using	use	VERB
ajst-32555	179	11	machine	machine	NOUN
ajst-32555	179	12	learning	learning	NOUN
ajst-32555	179	13	models	model	NOUN
ajst-32555	179	14	.	.	PUNCT
ajst-32555	180	1	algorithms	algorithm	NOUN
ajst-32555	180	2	,	,	PUNCT
ajst-32555	180	3	17(6	17(6	NUM
ajst-32555	180	4	)	)	PUNCT
ajst-32555	180	5	,	,	PUNCT
ajst-32555	180	6	231	231	NUM
ajst-32555	180	7	.	.	PUNCT
ajst-32555	181	1	https://doi.org/10.3390/a17060231	https://doi.org/10.3390/a17060231	NUM
ajst-32555	181	2	.	.	PUNCT
ajst-32555	182	1	[	[	X
ajst-32555	182	2	6	6	NUM
ajst-32555	182	3	]	]	PUNCT
ajst-32555	182	4	kaggle	kaggle	NOUN
ajst-32555	182	5	.	.	PUNCT
ajst-32555	183	1	(	(	PUNCT
ajst-32555	183	2	2025	2025	NUM
ajst-32555	183	3	)	)	PUNCT
ajst-32555	183	4	.	.	PUNCT
ajst-32555	184	1	telco	telco	NOUN
ajst-32555	184	2	customer	customer	NOUN
ajst-32555	184	3	churn	churn	NOUN
ajst-32555	184	4	:	:	PUNCT
ajst-32555	184	5	exploratory	exploratory	ADJ
ajst-32555	184	6	data	datum	NOUN
ajst-32555	184	7	analysis	analysis	NOUN
ajst-32555	184	8	.	.	PUNCT
ajst-32555	185	1	https://www.kaggle.com/	https://www.kaggle.com/	NOUN
ajst-32555	185	2	code/	code/	NUM
ajst-32555	185	3	supratimhaldar/	supratimhaldar/	NUM
ajst-32555	185	4	telco	telco	NOUN
ajst-32555	185	5	-	-	PUNCT
ajst-32555	185	6	customer	customer	NOUN
ajst-32555	185	7	-	-	PUNCT
ajst-32555	185	8	churn	churn	NOUN
ajst-32555	185	9	-	-	PUNCT
ajst-32555	185	10	exploratory	exploratory	ADJ
ajst-32555	185	11	-	-	PUNCT
ajst-32555	185	12	data	data	NOUN
ajst-32555	185	13	-	-	PUNCT
ajst-32555	185	14	analysis	analysis	NOUN
ajst-32555	185	15	.	.	PUNCT
ajst-32555	186	1	[	[	X
ajst-32555	186	2	7	7	NUM
ajst-32555	186	3	]	]	X
ajst-32555	186	4	medium	medium	NOUN
ajst-32555	186	5	.	.	PUNCT
ajst-32555	187	1	(	(	PUNCT
ajst-32555	187	2	2019	2019	NUM
ajst-32555	187	3	)	)	PUNCT
ajst-32555	187	4	.	.	PUNCT
ajst-32555	188	1	machine	machine	NOUN
ajst-32555	188	2	learning	learning	NOUN
ajst-32555	188	3	case	case	NOUN
ajst-32555	188	4	study	study	NOUN
ajst-32555	188	5	:	:	PUNCT
ajst-32555	188	6	telco	telco	ADJ
ajst-32555	188	7	customer	customer	NOUN
ajst-32555	188	8	churn	churn	NOUN
ajst-32555	188	9	prediction	prediction	NOUN
ajst-32555	188	10	.	.	PUNCT
ajst-32555	189	1	https://medium.com/	https://medium.com/	VERB
ajst-32555	189	2	@	@	ADP
ajst-32555	189	3	manureservations	manureservation	NOUN
ajst-32555	189	4	/	/	SYM
ajst-32555	189	5	machine	machine	NOUN
ajst-32555	189	6	-	-	PUNCT
ajst-32555	189	7	learning	learning	NOUN
ajst-32555	189	8	-	-	PUNCT
ajst-32555	189	9	case	case	NOUN
ajst-32555	189	10	-	-	PUNCT
ajst-32555	189	11	study	study	NOUN
ajst-32555	189	12	-	-	PUNCT
ajst-32555	189	13	telco	telco	NOUN
ajst-32555	189	14	-	-	PUNCT
ajst-32555	189	15	customer	customer	NOUN
ajst-32555	189	16	-	-	PUNCT
ajst-32555	189	17	churn	churn	NOUN
ajst-32555	189	18	-	-	PUNCT
ajst-32555	189	19	prediction	prediction	NOUN
ajst-32555	189	20	-	-	PUNCT
ajst-32555	189	21	a5f228364945	a5f228364945	NOUN
ajst-32555	189	22	.	.	PUNCT
ajst-32555	190	1	[	[	X
ajst-32555	190	2	8	8	NUM
ajst-32555	190	3	]	]	PUNCT
ajst-32555	190	4	kaggle	kaggle	NOUN
ajst-32555	190	5	.	.	PUNCT
ajst-32555	191	1	(	(	PUNCT
ajst-32555	191	2	2025	2025	NUM
ajst-32555	191	3	)	)	PUNCT
ajst-32555	191	4	.	.	PUNCT
ajst-32555	192	1	telco	telco	NOUN
ajst-32555	192	2	customer	customer	NOUN
ajst-32555	192	3	churn	churn	VERB
ajst-32555	192	4	eda	eda	PROPN
ajst-32555	192	5	&	&	CCONJ
ajst-32555	192	6	predictions	prediction	NOUN
ajst-32555	192	7	+	+	CCONJ
ajst-32555	192	8	shap	shap	NOUN
ajst-32555	192	9	.	.	PUNCT
ajst-32555	193	1	https://www.kaggle	https://www.kaggle	PROPN
ajst-32555	193	2	.	.	PUNCT
ajst-32555	194	1	com/	com/	NOUN
ajst-32555	194	2	code/	code/	NUM
ajst-32555	194	3	jonaspalucibarbosa	jonaspalucibarbosa	PROPN
ajst-32555	194	4	/	/	SYM
ajst-32555	194	5	telco	telco	NOUN
ajst-32555	194	6	-	-	PUNCT
ajst-32555	194	7	customer	customer	NOUN
ajst-32555	194	8	-	-	PUNCT
ajst-32555	194	9	churn	churn	NOUN
ajst-32555	194	10	-	-	PUNCT
ajst-32555	194	11	eda	eda	NOUN
ajst-32555	194	12	-	-	PUNCT
ajst-32555	194	13	predictions	prediction	NOUN
ajst-32555	194	14	-	-	PUNCT
ajst-32555	194	15	shap	shap	NOUN
ajst-32555	194	16	.	.	PUNCT
ajst-32555	195	1	[	[	X
ajst-32555	195	2	9	9	NUM
ajst-32555	195	3	]	]	X
ajst-32555	195	4	ahmad	ahmad	PROPN
ajst-32555	195	5	,	,	PUNCT
ajst-32555	195	6	j.	j.	PROPN
ajst-32555	195	7	,	,	PUNCT
ajst-32555	195	8	jafar	jafar	PROPN
ajst-32555	195	9	,	,	PUNCT
ajst-32555	195	10	r.	r.	PROPN
ajst-32555	195	11	,	,	PUNCT
ajst-32555	195	12	&	&	CCONJ
ajst-32555	195	13	aljoumaa	aljoumaa	PROPN
ajst-32555	195	14	,	,	PUNCT
ajst-32555	195	15	f.	f.	PROPN
ajst-32555	195	16	(	(	PUNCT
ajst-32555	195	17	2025	2025	NUM
ajst-32555	195	18	)	)	PUNCT
ajst-32555	195	19	.	.	PUNCT
ajst-32555	195	20	customer	customer	NOUN
ajst-32555	195	21	churn	churn	NOUN
ajst-32555	195	22	prediction	prediction	NOUN
ajst-32555	195	23	in	in	ADP
ajst-32555	195	24	telecom	telecom	NOUN
ajst-32555	195	25	using	use	VERB
ajst-32555	195	26	machine	machine	NOUN
ajst-32555	195	27	learning	learning	NOUN
ajst-32555	195	28	and	and	CCONJ
ajst-32555	195	29	social	social	ADJ
ajst-32555	195	30	network	network	NOUN
ajst-32555	195	31	analysis	analysis	NOUN
ajst-32555	195	32	in	in	ADP
ajst-32555	195	33	big	big	ADJ
ajst-32555	195	34	data	datum	NOUN
ajst-32555	195	35	platform	platform	NOUN
ajst-32555	195	36	.	.	PUNCT
ajst-32555	196	1	researchgate	researchgate	NOUN
ajst-32555	196	2	.	.	PUNCT
ajst-32555	197	1	https://www	https://www	PROPN
ajst-32555	197	2	.	.	PUNCT
ajst-32555	198	1	researchgate	researchgate	NOUN
ajst-32555	198	2	.	.	PUNCT
ajst-32555	199	1	net/	net/	PROPN
ajst-32555	199	2	publication/386615188_customer_churn_prediction_in_telecom_using_machine_learning_and_social	publication/386615188_customer_churn_prediction_in_telecom_using_machine_learning_and_social	PROPN
ajst-32555	199	3	_	_	PUNCT
ajst-32555	199	4	network_analysis_in_big_data_platform	network_analysis_in_big_data_platform	NOUN
ajst-32555	199	5	.	.	PUNCT
ajst-32555	200	1	[	[	X
ajst-32555	200	2	10	10	NUM
ajst-32555	200	3	]	]	X
ajst-32555	200	4	de	de	X
ajst-32555	200	5	caigny	caigny	PROPN
ajst-32555	200	6	,	,	PUNCT
ajst-32555	200	7	s.	s.	PROPN
ajst-32555	200	8	,	,	PUNCT
ajst-32555	200	9	coussement	coussement	ADJ
ajst-32555	200	10	,	,	PUNCT
ajst-32555	200	11	k.	k.	PROPN
ajst-32555	200	12	,	,	PUNCT
ajst-32555	200	13	&	&	CCONJ
ajst-32555	200	14	de	de	X
ajst-32555	200	15	bock	bock	PROPN
ajst-32555	200	16	,	,	PUNCT
ajst-32555	200	17	k.	k.	PROPN
ajst-32555	200	18	(	(	PUNCT
ajst-32555	200	19	2025	2025	NUM
ajst-32555	200	20	)	)	PUNCT
ajst-32555	200	21	.	.	PUNCT
ajst-32555	201	1	a	a	DET
ajst-32555	201	2	new	new	ADJ
ajst-32555	201	3	hybrid	hybrid	ADJ
ajst-32555	201	4	classification	classification	NOUN
ajst-32555	201	5	algorithm	algorithm	NOUN
ajst-32555	201	6	for	for	ADP
ajst-32555	201	7	customer	customer	NOUN
ajst-32555	201	8	churn	churn	NOUN
ajst-32555	201	9	prediction	prediction	NOUN
ajst-32555	201	10	based	base	VERB
ajst-32555	201	11	on	on	ADP
ajst-32555	201	12	logistic	logistic	ADJ
ajst-32555	201	13	regression	regression	NOUN
ajst-32555	201	14	and	and	CCONJ
ajst-32555	201	15	decision	decision	NOUN
ajst-32555	201	16	trees	tree	NOUN
ajst-32555	201	17	.	.	PUNCT
ajst-32555	202	1	researchgate	researchgate	NOUN
ajst-32555	202	2	.	.	PUNCT
ajst-32555	203	1	https://www.researchgate.net/publication/323130432_a_new_hybrid_classification_algorithm_for_c	https://www.researchgate.net/publication/323130432_a_new_hybrid_classification_algorithm_for_c	PROPN
ajst-32555	203	2	ustomer_churn_prediction_based_on_logistic_regression_and_decision_trees	ustomer_churn_prediction_based_on_logistic_regression_and_decision_tree	NOUN
ajst-32555	203	3	.	.	PUNCT
ajst-32555	204	1	[	[	X
ajst-32555	204	2	11	11	NUM
ajst-32555	204	3	]	]	PUNCT
ajst-32555	204	4	kaggle	kaggle	NOUN
ajst-32555	204	5	.	.	PUNCT
ajst-32555	205	1	(	(	PUNCT
ajst-32555	205	2	2025	2025	NUM
ajst-32555	205	3	)	)	PUNCT
ajst-32555	205	4	.	.	PUNCT
ajst-32555	206	1	telco	telco	NOUN
ajst-32555	206	2	customer	customer	NOUN
ajst-32555	206	3	churn	churn	NOUN
ajst-32555	206	4	.	.	PUNCT
ajst-32555	207	1	https://www.kaggle.com/datasets/blastchar/telco-customer-churn	https://www.kaggle.com/datasets/blastchar/telco-customer-churn	NOUN
ajst-32555	207	2	.	.	PUNCT
ajst-32555	208	1	[	[	X
ajst-32555	208	2	12	12	NUM
ajst-32555	208	3	]	]	SYM
ajst-32555	208	4	observable	observable	ADJ
ajst-32555	208	5	.	.	PUNCT
ajst-32555	209	1	(	(	PUNCT
ajst-32555	209	2	2023	2023	NUM
ajst-32555	209	3	)	)	PUNCT
ajst-32555	209	4	.	.	PUNCT
ajst-32555	210	1	end	end	NOUN
ajst-32555	210	2	-	-	PUNCT
ajst-32555	210	3	to	to	ADP
ajst-32555	210	4	-	-	PUNCT
ajst-32555	210	5	end	end	NOUN
ajst-32555	210	6	machine	machine	NOUN
ajst-32555	210	7	learning	learning	NOUN
ajst-32555	210	8	project	project	NOUN
ajst-32555	210	9	:	:	PUNCT
ajst-32555	210	10	telco	telco	ADJ
ajst-32555	210	11	customer	customer	NOUN
ajst-32555	210	12	churn	churn	NOUN
ajst-32555	210	13	/	/	PUNCT
ajst-32555	210	14	his	his	PRON
ajst-32555	210	15	codeship	codeship	NOUN
ajst-32555	210	16	.	.	PUNCT
ajst-32555	211	1	https://observablehq.com/@ealecho/end-to-end-machine-learning-project-telco-customer-churn	https://observablehq.com/@ealecho/end-to-end-machine-learning-project-telco-customer-churn	NOUN
ajst-32555	211	2	.	.	PUNCT
ajst-32555	212	1	123	123	NUM
ajst-32555	212	2	[	[	SYM
ajst-32555	212	3	13	13	NUM
ajst-32555	212	4	]	]	X
ajst-32555	212	5	medium	medium	NOUN
ajst-32555	212	6	.	.	PUNCT
ajst-32555	213	1	(	(	PUNCT
ajst-32555	213	2	2020	2020	NUM
ajst-32555	213	3	)	)	PUNCT
ajst-32555	213	4	.	.	PUNCT
ajst-32555	214	1	comprehensive	comprehensive	ADJ
ajst-32555	214	2	report	report	NOUN
ajst-32555	214	3	:	:	PUNCT
ajst-32555	214	4	telecom	telecom	PROPN
ajst-32555	214	5	customer	customer	NOUN
ajst-32555	214	6	churn	churn	NOUN
ajst-32555	214	7	analysis	analysis	NOUN
ajst-32555	214	8	and	and	CCONJ
ajst-32555	214	9	recommendations	recommendation	NOUN
ajst-32555	214	10	.	.	PUNCT
ajst-32555	215	1	https://medium.com/@kinghenrymorgansdiary/comprehensive-report-telecom-customer-churnanalysis-and-recommendations-398eedaf3466	https://medium.com/@kinghenrymorgansdiary/comprehensive-report-telecom-customer-churnanalysis-and-recommendations-398eedaf3466	NOUN
ajst-32555	215	2	.	.	PUNCT
ajst-32555	216	1	[	[	X
ajst-32555	216	2	14	14	NUM
ajst-32555	216	3	]	]	X
ajst-32555	216	4	ahmad	ahmad	PROPN
ajst-32555	216	5	,	,	PUNCT
ajst-32555	216	6	j.	j.	PROPN
ajst-32555	216	7	,	,	PUNCT
ajst-32555	216	8	jafar	jafar	PROPN
ajst-32555	216	9	,	,	PUNCT
ajst-32555	216	10	r.	r.	PROPN
ajst-32555	216	11	,	,	PUNCT
ajst-32555	216	12	&	&	CCONJ
ajst-32555	216	13	aljoumaa	aljoumaa	PROPN
ajst-32555	216	14	,	,	PUNCT
ajst-32555	216	15	f.	f.	PROPN
ajst-32555	216	16	(	(	PUNCT
ajst-32555	216	17	2024	2024	NUM
ajst-32555	216	18	)	)	PUNCT
ajst-32555	216	19	.	.	PUNCT
ajst-32555	216	20	customer	customer	NOUN
ajst-32555	216	21	churn	churn	NOUN
ajst-32555	216	22	prediction	prediction	NOUN
ajst-32555	216	23	in	in	ADP
ajst-32555	216	24	telecom	telecom	NOUN
ajst-32555	216	25	using	use	VERB
ajst-32555	216	26	machine	machine	NOUN
ajst-32555	216	27	learning	learning	NOUN
ajst-32555	216	28	and	and	CCONJ
ajst-32555	216	29	social	social	ADJ
ajst-32555	216	30	network	network	NOUN
ajst-32555	216	31	analysis	analysis	NOUN
ajst-32555	216	32	in	in	ADP
ajst-32555	216	33	big	big	ADJ
ajst-32555	216	34	data	datum	NOUN
ajst-32555	216	35	platform	platform	NOUN
ajst-32555	216	36	.	.	PUNCT
ajst-32555	217	1	researchgate	researchgate	NOUN
ajst-32555	217	2	.	.	PUNCT
ajst-32555	218	1	https://www	https://www	PROPN
ajst-32555	218	2	.	.	PUNCT
ajst-32555	219	1	researchgate	researchgate	NOUN
ajst-32555	219	2	.	.	PUNCT
ajst-32555	220	1	net/	net/	PROPN
ajst-32555	220	2	publication/386615188_customer_churn_prediction_in_telecom_using_machine_learning_and_social	publication/386615188_customer_churn_prediction_in_telecom_using_machine_learning_and_social	PROPN
ajst-32555	220	3	_	_	SYM
ajst-32555	220	4	network_analysis_in_big_data_platform	network_analysis_in_big_data_platform	NOUN
ajst-32555	220	5	.	.	PUNCT