id	sid	tid	token	lemma	pos
cril-239	1	1	automatic	automatic	ADJ
cril-239	1	2	opinion	opinion	NOUN
cril-239	1	3	polarity	polarity	NOUN
cril-239	1	4	classification	classification	NOUN
cril-239	1	5	of	of	ADP
cril-239	1	6	movie	movie	NOUN
cril-239	1	7	reviews	review	NOUN
cril-239	1	8	colorado	colorado	NOUN
cril-239	1	9	research	research	NOUN
cril-239	1	10	in	in	ADP
cril-239	1	11	linguistics	linguistic	NOUN
cril-239	1	12	.	.	PUNCT
cril-239	2	1	june	june	PROPN
cril-239	2	2	2004	2004	NUM
cril-239	2	3	.	.	PUNCT
cril-239	3	1	volume	volume	NOUN
cril-239	3	2	17	17	NUM
cril-239	3	3	,	,	PUNCT
cril-239	3	4	issue	issue	NOUN
cril-239	3	5	1	1	NUM
cril-239	3	6	.	.	PUNCT
cril-239	3	7	boulder	boulder	NOUN
cril-239	3	8	:	:	PUNCT
cril-239	3	9	university	university	NOUN
cril-239	3	10	of	of	ADP
cril-239	3	11	colorado	colorado	PROPN
cril-239	3	12	.	.	PUNCT
cril-239	4	1	©	©	PROPN
cril-239	4	2	2004	2004	NUM
cril-239	4	3	by	by	ADP
cril-239	4	4	franco	franco	PROPN
cril-239	4	5	salvetti	salvetti	PROPN
cril-239	4	6	,	,	PUNCT
cril-239	4	7	stephen	stephen	PROPN
cril-239	4	8	lewis	lewis	PROPN
cril-239	4	9	,	,	PUNCT
cril-239	4	10	christoph	christoph	PROPN
cril-239	4	11	reichenbach	reichenbach	PROPN
cril-239	4	12	.	.	PUNCT
cril-239	5	1	automatic	automatic	ADJ
cril-239	5	2	opinion	opinion	NOUN
cril-239	5	3	polarity	polarity	NOUN
cril-239	5	4	classification	classification	NOUN
cril-239	5	5	of	of	ADP
cril-239	5	6	movie	movie	NOUN
cril-239	5	7	reviews	review	NOUN
cril-239	5	8	franco	franco	PROPN
cril-239	5	9	salvetti	salvetti	PROPN
cril-239	5	10	department	department	PROPN
cril-239	5	11	of	of	ADP
cril-239	5	12	computer	computer	NOUN
cril-239	5	13	science	science	NOUN
cril-239	5	14	,	,	PUNCT
cril-239	5	15	university	university	NOUN
cril-239	5	16	of	of	ADP
cril-239	5	17	colorado	colorado	PROPN
cril-239	5	18	at	at	ADP
cril-239	5	19	boulder	boulder	PROPN
cril-239	5	20	stephen	stephen	PROPN
cril-239	5	21	lewis	lewis	PROPN
cril-239	5	22	department	department	PROPN
cril-239	5	23	of	of	ADP
cril-239	5	24	linguistics	linguistics	PROPN
cril-239	5	25	,	,	PUNCT
cril-239	5	26	university	university	NOUN
cril-239	5	27	of	of	ADP
cril-239	5	28	colorado	colorado	PROPN
cril-239	5	29	at	at	ADP
cril-239	5	30	boulder	boulder	PROPN
cril-239	5	31	christoph	christoph	PROPN
cril-239	5	32	reichenbach	reichenbach	PROPN
cril-239	5	33	department	department	PROPN
cril-239	5	34	of	of	ADP
cril-239	5	35	computer	computer	NOUN
cril-239	5	36	science	science	NOUN
cril-239	5	37	,	,	PUNCT
cril-239	5	38	university	university	NOUN
cril-239	5	39	of	of	ADP
cril-239	5	40	colorado	colorado	PROPN
cril-239	5	41	at	at	ADP
cril-239	5	42	boulder	boulder	NOUN
cril-239	5	43	one	one	NUM
cril-239	5	44	approach	approach	NOUN
cril-239	5	45	to	to	ADP
cril-239	5	46	assessing	assess	VERB
cril-239	5	47	overall	overall	ADJ
cril-239	5	48	opinion	opinion	NOUN
cril-239	5	49	polarity	polarity	NOUN
cril-239	5	50	(	(	PUNCT
cril-239	5	51	ovop	ovop	ADJ
cril-239	5	52	)	)	PUNCT
cril-239	5	53	of	of	ADP
cril-239	5	54	reviews	review	NOUN
cril-239	5	55	,	,	PUNCT
cril-239	5	56	a	a	DET
cril-239	5	57	concept	concept	NOUN
cril-239	5	58	defined	define	VERB
cril-239	5	59	in	in	ADP
cril-239	5	60	this	this	DET
cril-239	5	61	paper	paper	NOUN
cril-239	5	62	,	,	PUNCT
cril-239	5	63	is	be	AUX
cril-239	5	64	the	the	DET
cril-239	5	65	use	use	NOUN
cril-239	5	66	of	of	ADP
cril-239	5	67	supervised	supervised	ADJ
cril-239	5	68	machine	machine	NOUN
cril-239	5	69	learning	learn	VERB
cril-239	5	70	mechanisms	mechanism	NOUN
cril-239	5	71	.	.	PUNCT
cril-239	6	1	in	in	ADP
cril-239	6	2	this	this	DET
cril-239	6	3	paper	paper	NOUN
cril-239	6	4	,	,	PUNCT
cril-239	6	5	the	the	DET
cril-239	6	6	impact	impact	NOUN
cril-239	6	7	of	of	ADP
cril-239	6	8	lexical	lexical	ADJ
cril-239	6	9	filtering	filtering	NOUN
cril-239	6	10	,	,	PUNCT
cril-239	6	11	applied	apply	VERB
cril-239	6	12	to	to	ADP
cril-239	6	13	reviews	review	NOUN
cril-239	6	14	,	,	PUNCT
cril-239	6	15	on	on	ADP
cril-239	6	16	the	the	DET
cril-239	6	17	accuracy	accuracy	NOUN
cril-239	6	18	of	of	ADP
cril-239	6	19	two	two	NUM
cril-239	6	20	statistical	statistical	ADJ
cril-239	6	21	classifiers	classifier	NOUN
cril-239	6	22	(	(	PUNCT
cril-239	6	23	naive	naive	ADJ
cril-239	6	24	bayes	bayes	NOUN
cril-239	6	25	and	and	CCONJ
cril-239	6	26	markov	markov	NOUN
cril-239	6	27	model	model	NOUN
cril-239	6	28	)	)	PUNCT
cril-239	6	29	with	with	ADP
cril-239	6	30	respect	respect	NOUN
cril-239	6	31	to	to	ADP
cril-239	6	32	ovop	ovop	ADJ
cril-239	6	33	identification	identification	NOUN
cril-239	6	34	is	be	AUX
cril-239	6	35	observed	observe	VERB
cril-239	6	36	.	.	PUNCT
cril-239	7	1	two	two	NUM
cril-239	7	2	kinds	kind	NOUN
cril-239	7	3	of	of	ADP
cril-239	7	4	lexical	lexical	ADJ
cril-239	7	5	filters	filter	NOUN
cril-239	7	6	,	,	PUNCT
cril-239	7	7	one	one	NUM
cril-239	7	8	based	base	VERB
cril-239	7	9	on	on	ADP
cril-239	7	10	hypernymy	hypernymy	NOUN
cril-239	7	11	as	as	SCONJ
cril-239	7	12	provided	provide	VERB
cril-239	7	13	by	by	ADP
cril-239	7	14	wordnet	wordnet	NOUN
cril-239	7	15	(	(	PUNCT
cril-239	7	16	fellbaum	fellbaum	NOUN
cril-239	7	17	,	,	PUNCT
cril-239	7	18	1998	1998	NUM
cril-239	7	19	)	)	PUNCT
cril-239	7	20	,	,	PUNCT
cril-239	7	21	and	and	CCONJ
cril-239	7	22	one	one	NUM
cril-239	7	23	hand	hand	NOUN
cril-239	7	24	-	-	PUNCT
cril-239	7	25	crafted	craft	VERB
cril-239	7	26	filter	filter	NOUN
cril-239	7	27	based	base	VERB
cril-239	7	28	on	on	ADP
cril-239	7	29	part	part	NOUN
cril-239	7	30	-	-	PUNCT
cril-239	7	31	of	of	ADP
cril-239	7	32	-	-	PUNCT
cril-239	7	33	speech	speech	NOUN
cril-239	7	34	(	(	PUNCT
cril-239	7	35	pos	pos	NOUN
cril-239	7	36	)	)	PUNCT
cril-239	7	37	tags	tag	NOUN
cril-239	7	38	,	,	PUNCT
cril-239	7	39	are	be	AUX
cril-239	7	40	evaluated	evaluate	VERB
cril-239	7	41	.	.	PUNCT
cril-239	8	1	a	a	DET
cril-239	8	2	ranking	ranking	ADJ
cril-239	8	3	criterion	criterion	NOUN
cril-239	8	4	based	base	VERB
cril-239	8	5	on	on	ADP
cril-239	8	6	a	a	DET
cril-239	8	7	function	function	NOUN
cril-239	8	8	of	of	ADP
cril-239	8	9	the	the	DET
cril-239	8	10	probability	probability	NOUN
cril-239	8	11	of	of	ADP
cril-239	8	12	having	have	VERB
cril-239	8	13	positive	positive	ADJ
cril-239	8	14	or	or	CCONJ
cril-239	8	15	negative	negative	ADJ
cril-239	8	16	polarity	polarity	NOUN
cril-239	8	17	is	be	AUX
cril-239	8	18	introduced	introduce	VERB
cril-239	8	19	and	and	CCONJ
cril-239	8	20	verified	verify	VERB
cril-239	8	21	as	as	ADP
cril-239	8	22	being	be	AUX
cril-239	8	23	capable	capable	ADJ
cril-239	8	24	of	of	ADP
cril-239	8	25	achieving	achieve	VERB
cril-239	8	26	100	100	NUM
cril-239	8	27	%	%	NOUN
cril-239	8	28	accuracy	accuracy	NOUN
cril-239	8	29	with	with	ADP
cril-239	8	30	10	10	NUM
cril-239	8	31	%	%	NOUN
cril-239	8	32	recall	recall	NOUN
cril-239	8	33	.	.	PUNCT
cril-239	9	1	movie	movie	NOUN
cril-239	9	2	reviews	review	NOUN
cril-239	9	3	are	be	AUX
cril-239	9	4	used	use	VERB
cril-239	9	5	for	for	ADP
cril-239	9	6	training	training	NOUN
cril-239	9	7	and	and	CCONJ
cril-239	9	8	evaluation	evaluation	NOUN
cril-239	9	9	of	of	ADP
cril-239	9	10	each	each	DET
cril-239	9	11	statistical	statistical	ADJ
cril-239	9	12	classifier	classifier	NOUN
cril-239	9	13	,	,	PUNCT
cril-239	9	14	achieving	achieve	VERB
cril-239	9	15	80	80	NUM
cril-239	9	16	%	%	NOUN
cril-239	9	17	accuracy	accuracy	NOUN
cril-239	9	18	.	.	PUNCT
cril-239	10	1	1	1	X
cril-239	10	2	.	.	X
cril-239	10	3	introduction	introduction	NOUN
cril-239	10	4	the	the	DET
cril-239	10	5	dramatic	dramatic	ADJ
cril-239	10	6	increase	increase	NOUN
cril-239	10	7	in	in	ADP
cril-239	10	8	use	use	NOUN
cril-239	10	9	of	of	ADP
cril-239	10	10	the	the	DET
cril-239	10	11	internet	internet	NOUN
cril-239	10	12	as	as	ADP
cril-239	10	13	a	a	DET
cril-239	10	14	means	means	NOUN
cril-239	10	15	of	of	ADP
cril-239	10	16	communication	communication	NOUN
cril-239	10	17	has	have	AUX
cril-239	10	18	been	be	AUX
cril-239	10	19	accompanied	accompany	VERB
cril-239	10	20	by	by	ADP
cril-239	10	21	an	an	DET
cril-239	10	22	increase	increase	NOUN
cril-239	10	23	in	in	ADP
cril-239	10	24	freely	freely	ADV
cril-239	10	25	available	available	ADJ
cril-239	10	26	online	online	ADJ
cril-239	10	27	reviews	review	NOUN
cril-239	10	28	of	of	ADP
cril-239	10	29	products	product	NOUN
cril-239	10	30	and	and	CCONJ
cril-239	10	31	services	service	NOUN
cril-239	10	32	.	.	PUNCT
cril-239	11	1	although	although	SCONJ
cril-239	11	2	such	such	ADJ
cril-239	11	3	reviews	review	NOUN
cril-239	11	4	are	be	AUX
cril-239	11	5	a	a	DET
cril-239	11	6	valuable	valuable	ADJ
cril-239	11	7	resource	resource	NOUN
cril-239	11	8	to	to	ADP
cril-239	11	9	customers	customer	NOUN
cril-239	11	10	who	who	PRON
cril-239	11	11	want	want	VERB
cril-239	11	12	to	to	PART
cril-239	11	13	make	make	VERB
cril-239	11	14	wellinformed	wellinforme	VERB
cril-239	11	15	shopping	shopping	NOUN
cril-239	11	16	decisions	decision	NOUN
cril-239	11	17	,	,	PUNCT
cril-239	11	18	their	their	PRON
cril-239	11	19	abundance	abundance	NOUN
cril-239	11	20	and	and	CCONJ
cril-239	11	21	the	the	DET
cril-239	11	22	fact	fact	NOUN
cril-239	11	23	that	that	SCONJ
cril-239	11	24	they	they	PRON
cril-239	11	25	are	be	AUX
cril-239	11	26	mixed	mix	VERB
cril-239	11	27	in	in	ADP
cril-239	11	28	terms	term	NOUN
cril-239	11	29	of	of	ADP
cril-239	11	30	positive	positive	ADJ
cril-239	11	31	and	and	CCONJ
cril-239	11	32	negative	negative	ADJ
cril-239	11	33	overall	overall	ADJ
cril-239	11	34	opinion	opinion	NOUN
cril-239	11	35	polarity	polarity	NOUN
cril-239	11	36	are	be	AUX
cril-239	11	37	often	often	ADV
cril-239	11	38	obstacles	obstacle	NOUN
cril-239	11	39	.	.	PUNCT
cril-239	12	1	for	for	ADP
cril-239	12	2	instance	instance	NOUN
cril-239	12	3	,	,	PUNCT
cril-239	12	4	a	a	DET
cril-239	12	5	customer	customer	NOUN
cril-239	12	6	that	that	PRON
cril-239	12	7	is	be	AUX
cril-239	12	8	already	already	ADV
cril-239	12	9	interested	interested	ADJ
cril-239	12	10	in	in	ADP
cril-239	12	11	a	a	DET
cril-239	12	12	certain	certain	ADJ
cril-239	12	13	product	product	NOUN
cril-239	12	14	may	may	AUX
cril-239	12	15	want	want	VERB
cril-239	12	16	to	to	PART
cril-239	12	17	read	read	VERB
cril-239	12	18	some	some	DET
cril-239	12	19	negative	negative	ADJ
cril-239	12	20	reviews	review	NOUN
cril-239	12	21	just	just	ADV
cril-239	12	22	to	to	PART
cril-239	12	23	pinpoint	pinpoint	VERB
cril-239	12	24	possible	possible	ADJ
cril-239	12	25	drawbacks	drawback	NOUN
cril-239	12	26	,	,	PUNCT
cril-239	12	27	but	but	CCONJ
cril-239	12	28	has	have	VERB
cril-239	12	29	no	no	DET
cril-239	12	30	interest	interest	NOUN
cril-239	12	31	in	in	ADP
cril-239	12	32	spending	spending	NOUN
cril-239	12	33	time	time	NOUN
cril-239	12	34	reading	read	VERB
cril-239	12	35	positive	positive	ADJ
cril-239	12	36	reviews	review	NOUN
cril-239	12	37	.	.	PUNCT
cril-239	13	1	in	in	ADP
cril-239	13	2	contrast	contrast	NOUN
cril-239	13	3	,	,	PUNCT
cril-239	13	4	customers	customer	NOUN
cril-239	13	5	interested	interested	ADJ
cril-239	13	6	in	in	ADP
cril-239	13	7	watching	watch	VERB
cril-239	13	8	a	a	DET
cril-239	13	9	good	good	ADJ
cril-239	13	10	movie	movie	NOUN
cril-239	13	11	may	may	AUX
cril-239	13	12	want	want	VERB
cril-239	13	13	to	to	PART
cril-239	13	14	read	read	VERB
cril-239	13	15	reviews	review	NOUN
cril-239	13	16	that	that	PRON
cril-239	13	17	express	express	VERB
cril-239	13	18	a	a	DET
cril-239	13	19	positive	positive	ADJ
cril-239	13	20	overall	overall	ADJ
cril-239	13	21	opinion	opinion	NOUN
cril-239	13	22	polarity	polarity	NOUN
cril-239	13	23	.	.	PUNCT
cril-239	14	1	the	the	DET
cril-239	14	2	overall	overall	ADJ
cril-239	14	3	opinion	opinion	NOUN
cril-239	14	4	polarity	polarity	NOUN
cril-239	14	5	of	of	ADP
cril-239	14	6	a	a	DET
cril-239	14	7	review	review	NOUN
cril-239	14	8	,	,	PUNCT
cril-239	14	9	with	with	ADP
cril-239	14	10	values	value	NOUN
cril-239	14	11	expressed	express	VERB
cril-239	14	12	as	as	ADP
cril-239	14	13	positive	positive	ADJ
cril-239	14	14	or	or	CCONJ
cril-239	14	15	negative	negative	ADJ
cril-239	14	16	,	,	PUNCT
cril-239	14	17	can	can	AUX
cril-239	14	18	be	be	AUX
cril-239	14	19	represented	represent	VERB
cril-239	14	20	through	through	ADP
cril-239	14	21	the	the	DET
cril-239	14	22	classification	classification	NOUN
cril-239	14	23	that	that	SCONJ
cril-239	14	24	the	the	DET
cril-239	14	25	author	author	NOUN
cril-239	14	26	of	of	ADP
cril-239	14	27	a	a	DET
cril-239	14	28	review	review	NOUN
cril-239	14	29	would	would	AUX
cril-239	14	30	assign	assign	VERB
cril-239	14	31	to	to	ADP
cril-239	14	32	it	it	PRON
cril-239	14	33	,	,	PUNCT
cril-239	14	34	if	if	SCONJ
cril-239	14	35	requested	request	VERB
cril-239	14	36	.	.	PUNCT
cril-239	15	1	such	such	DET
cril-239	15	2	a	a	DET
cril-239	15	3	classification	classification	NOUN
cril-239	15	4	is	be	AUX
cril-239	15	5	here	here	ADV
cril-239	15	6	defined	define	VERB
cril-239	15	7	as	as	ADP
cril-239	15	8	the	the	DET
cril-239	15	9	overall	overall	ADJ
cril-239	15	10	opinion	opinion	NOUN
cril-239	15	11	polarity	polarity	NOUN
cril-239	15	12	(	(	PUNCT
cril-239	15	13	ovop	ovop	ADJ
cril-239	15	14	)	)	PUNCT
cril-239	15	15	of	of	ADP
cril-239	15	16	a	a	DET
cril-239	15	17	review	review	NOUN
cril-239	15	18	,	,	PUNCT
cril-239	15	19	or	or	CCONJ
cril-239	15	20	simply	simply	ADV
cril-239	15	21	the	the	DET
cril-239	15	22	polarity	polarity	NOUN
cril-239	15	23	.	.	PUNCT
cril-239	16	1	the	the	DET
cril-239	16	2	process	process	NOUN
cril-239	16	3	of	of	ADP
cril-239	16	4	identifying	identify	VERB
cril-239	16	5	ovop	ovop	NOUN
cril-239	16	6	of	of	ADP
cril-239	16	7	a	a	DET
cril-239	16	8	review	review	NOUN
cril-239	16	9	will	will	AUX
cril-239	16	10	be	be	AUX
cril-239	16	11	referred	refer	VERB
cril-239	16	12	to	to	ADP
cril-239	16	13	as	as	ADP
cril-239	16	14	overall	overall	ADJ
cril-239	16	15	opinion	opinion	NOUN
cril-239	16	16	polarity	polarity	NOUN
cril-239	16	17	identification	identification	NOUN
cril-239	16	18	(	(	PUNCT
cril-239	16	19	ovopi	ovopi	PROPN
cril-239	16	20	)	)	PUNCT
cril-239	16	21	.	.	PUNCT
cril-239	17	1	a	a	DET
cril-239	17	2	system	system	NOUN
cril-239	17	3	that	that	PRON
cril-239	17	4	is	be	AUX
cril-239	17	5	capable	capable	ADJ
cril-239	17	6	of	of	ADP
cril-239	17	7	labeling	label	VERB
cril-239	17	8	a	a	DET
cril-239	17	9	review	review	NOUN
cril-239	17	10	with	with	ADP
cril-239	17	11	its	its	PRON
cril-239	17	12	polarity	polarity	NOUN
cril-239	17	13	is	be	AUX
cril-239	17	14	valuable	valuable	ADJ
cril-239	17	15	for	for	ADP
cril-239	17	16	at	at	ADV
cril-239	17	17	least	least	ADV
cril-239	17	18	two	two	NUM
cril-239	17	19	reasons	reason	NOUN
cril-239	17	20	.	.	PUNCT
cril-239	18	1	first	first	ADV
cril-239	18	2	,	,	PUNCT
cril-239	18	3	it	it	PRON
cril-239	18	4	allows	allow	VERB
cril-239	18	5	the	the	DET
cril-239	18	6	reader	reader	NOUN
cril-239	18	7	interested	interest	VERB
cril-239	18	8	exclusively	exclusively	ADV
cril-239	18	9	in	in	ADP
cril-239	18	10	positive	positive	ADJ
cril-239	18	11	(	(	PUNCT
cril-239	18	12	or	or	CCONJ
cril-239	18	13	negative	negative	ADJ
cril-239	18	14	)	)	PUNCT
cril-239	18	15	reviews	review	NOUN
cril-239	18	16	to	to	PART
cril-239	18	17	save	save	VERB
cril-239	18	18	time	time	NOUN
cril-239	18	19	by	by	ADP
cril-239	18	20	reducing	reduce	VERB
cril-239	18	21	the	the	DET
cril-239	18	22	number	number	NOUN
cril-239	18	23	of	of	ADP
cril-239	18	24	reviews	review	NOUN
cril-239	18	25	to	to	PART
cril-239	18	26	be	be	AUX
cril-239	18	27	read	read	VERB
cril-239	18	28	.	.	PUNCT
cril-239	19	1	second	second	ADJ
cril-239	19	2	,	,	PUNCT
cril-239	19	3	since	since	SCONJ
cril-239	19	4	it	it	PRON
cril-239	19	5	is	be	AUX
cril-239	19	6	not	not	PART
cril-239	19	7	uncommon	uncommon	ADJ
cril-239	19	8	for	for	ADP
cril-239	19	9	a	a	DET
cril-239	19	10	review	review	NOUN
cril-239	19	11	that	that	PRON
cril-239	19	12	starts	start	VERB
cril-239	19	13	with	with	ADP
cril-239	19	14	positive	positive	ADJ
cril-239	19	15	polarity	polarity	NOUN
cril-239	19	16	to	to	PART
cril-239	19	17	turn	turn	VERB
cril-239	19	18	out	out	ADP
cril-239	19	19	to	to	PART
cril-239	19	20	be	be	AUX
cril-239	19	21	negative	negative	ADJ
cril-239	19	22	,	,	PUNCT
cril-239	19	23	or	or	CCONJ
cril-239	19	24	vice	vice	NOUN
cril-239	19	25	versa	versa	ADV
cril-239	19	26	,	,	PUNCT
cril-239	19	27	it	it	PRON
cril-239	19	28	avoids	avoid	VERB
cril-239	19	29	the	the	DET
cril-239	19	30	risk	risk	NOUN
cril-239	19	31	of	of	ADP
cril-239	19	32	a	a	DET
cril-239	19	33	reader	reader	NOUN
cril-239	19	34	erroneously	erroneously	ADV
cril-239	19	35	discarding	discard	VERB
cril-239	19	36	a	a	DET
cril-239	19	37	review	review	NOUN
cril-239	19	38	just	just	ADV
cril-239	19	39	because	because	SCONJ
cril-239	19	40	it	it	PRON
cril-239	19	41	first	first	ADV
cril-239	19	42	appears	appear	VERB
cril-239	19	43	to	to	PART
cril-239	19	44	have	have	VERB
cril-239	19	45	the	the	DET
cril-239	19	46	wrong	wrong	ADJ
cril-239	19	47	polarity	polarity	NOUN
cril-239	19	48	.	.	PUNCT
cril-239	20	1	in	in	ADP
cril-239	20	2	this	this	DET
cril-239	20	3	paper	paper	NOUN
cril-239	20	4	we	we	PRON
cril-239	20	5	frame	frame	VERB
cril-239	20	6	a	a	DET
cril-239	20	7	solution	solution	NOUN
cril-239	20	8	to	to	PART
cril-239	20	9	ovopi	ovopi	PROPN
cril-239	20	10	based	base	VERB
cril-239	20	11	on	on	ADP
cril-239	20	12	a	a	DET
cril-239	20	13	supervised	supervised	ADJ
cril-239	20	14	machine	machine	NOUN
cril-239	20	15	learning	learning	NOUN
cril-239	20	16	approach	approach	NOUN
cril-239	20	17	.	.	PUNCT
cril-239	21	1	in	in	ADP
cril-239	21	2	such	such	DET
cril-239	21	3	a	a	DET
cril-239	21	4	framework	framework	NOUN
cril-239	21	5	we	we	PRON
cril-239	21	6	observe	observe	VERB
cril-239	21	7	the	the	DET
cril-239	21	8	effects	effect	NOUN
cril-239	21	9	of	of	ADP
cril-239	21	10	lexical	lexical	ADJ
cril-239	21	11	filtering	filtering	NOUN
cril-239	21	12	,	,	PUNCT
cril-239	21	13	applied	apply	VERB
cril-239	21	14	to	to	ADP
cril-239	21	15	1	1	NUM
cril-239	21	16	salvetti	salvetti	VERB
cril-239	21	17	et	et	PROPN
cril-239	21	18	al	al	PROPN
cril-239	21	19	.	.	PROPN
cril-239	21	20	:	:	PUNCT
cril-239	22	1	automatic	automatic	ADJ
cril-239	22	2	opinion	opinion	NOUN
cril-239	22	3	polarity	polarity	NOUN
cril-239	22	4	classification	classification	NOUN
cril-239	22	5	of	of	ADP
cril-239	22	6	movie	movie	NOUN
cril-239	22	7	reviews	review	NOUN
cril-239	22	8	published	publish	VERB
cril-239	22	9	by	by	ADP
cril-239	22	10	cu	cu	PROPN
cril-239	22	11	scholar	scholar	NOUN
cril-239	22	12	,	,	PUNCT
cril-239	22	13	2004	2004	NUM
cril-239	22	14	colorado	colorado	PROPN
cril-239	22	15	research	research	NOUN
cril-239	22	16	in	in	ADP
cril-239	22	17	linguistics	linguistic	NOUN
cril-239	22	18	,	,	PUNCT
cril-239	22	19	volume	volume	NOUN
cril-239	22	20	17	17	NUM
cril-239	22	21	(	(	PUNCT
cril-239	22	22	2004	2004	NUM
cril-239	22	23	)	)	PUNCT
cril-239	22	24	2	2	NUM
cril-239	22	25	reviews	review	NOUN
cril-239	22	26	,	,	PUNCT
cril-239	22	27	on	on	ADP
cril-239	22	28	the	the	DET
cril-239	22	29	accuracy	accuracy	NOUN
cril-239	22	30	of	of	ADP
cril-239	22	31	two	two	NUM
cril-239	22	32	statistical	statistical	ADJ
cril-239	22	33	classifiers	classifier	NOUN
cril-239	22	34	trained	train	VERB
cril-239	22	35	on	on	ADP
cril-239	22	36	such	such	ADJ
cril-239	22	37	filtered	filter	VERB
cril-239	22	38	data	datum	NOUN
cril-239	22	39	.	.	PUNCT
cril-239	23	1	we	we	PRON
cril-239	23	2	have	have	AUX
cril-239	23	3	implemented	implement	VERB
cril-239	23	4	two	two	NUM
cril-239	23	5	different	different	ADJ
cril-239	23	6	kinds	kind	NOUN
cril-239	23	7	of	of	ADP
cril-239	23	8	lexical	lexical	ADJ
cril-239	23	9	filters	filter	NOUN
cril-239	23	10	,	,	PUNCT
cril-239	23	11	one	one	NUM
cril-239	23	12	based	base	VERB
cril-239	23	13	on	on	ADP
cril-239	23	14	hypernymy	hypernymy	NOUN
cril-239	23	15	as	as	SCONJ
cril-239	23	16	provided	provide	VERB
cril-239	23	17	by	by	ADP
cril-239	23	18	wordnet	wordnet	NOUN
cril-239	23	19	(	(	PUNCT
cril-239	23	20	fellbaum	fellbaum	NOUN
cril-239	23	21	,	,	PUNCT
cril-239	23	22	1998	1998	NUM
cril-239	23	23	)	)	PUNCT
cril-239	23	24	,	,	PUNCT
cril-239	23	25	and	and	CCONJ
cril-239	23	26	one	one	NUM
cril-239	23	27	based	base	VERB
cril-239	23	28	on	on	ADP
cril-239	23	29	part	part	NOUN
cril-239	23	30	-	-	PUNCT
cril-239	23	31	of	of	ADP
cril-239	23	32	-	-	PUNCT
cril-239	23	33	speech	speech	NOUN
cril-239	23	34	(	(	PUNCT
cril-239	23	35	pos	pos	NOUN
cril-239	23	36	)	)	PUNCT
cril-239	23	37	tags	tag	NOUN
cril-239	23	38	.	.	PUNCT
cril-239	24	1	the	the	DET
cril-239	24	2	results	result	NOUN
cril-239	24	3	obtained	obtain	VERB
cril-239	24	4	by	by	ADP
cril-239	24	5	experiments	experiment	NOUN
cril-239	24	6	based	base	VERB
cril-239	24	7	on	on	ADP
cril-239	24	8	movie	movie	NOUN
cril-239	24	9	reviews	review	NOUN
cril-239	24	10	revealed	reveal	VERB
cril-239	24	11	that	that	SCONJ
cril-239	24	12	wordnet	wordnet	NOUN
cril-239	24	13	filters	filter	NOUN
cril-239	24	14	produce	produce	VERB
cril-239	24	15	less	less	ADJ
cril-239	24	16	improvement	improvement	NOUN
cril-239	24	17	than	than	SCONJ
cril-239	24	18	do	do	VERB
cril-239	24	19	pos	pos	NOUN
cril-239	24	20	filters	filter	NOUN
cril-239	24	21	,	,	PUNCT
cril-239	24	22	and	and	CCONJ
cril-239	24	23	that	that	SCONJ
cril-239	24	24	for	for	ADP
cril-239	24	25	neither	neither	CCONJ
cril-239	24	26	is	be	AUX
cril-239	24	27	there	there	PRON
cril-239	24	28	evidence	evidence	NOUN
cril-239	24	29	of	of	ADP
cril-239	24	30	significantly	significantly	ADV
cril-239	24	31	improved	improve	VERB
cril-239	24	32	performance	performance	NOUN
cril-239	24	33	over	over	ADP
cril-239	24	34	the	the	DET
cril-239	24	35	system	system	NOUN
cril-239	24	36	without	without	ADP
cril-239	24	37	filters	filter	NOUN
cril-239	24	38	,	,	PUNCT
cril-239	24	39	although	although	SCONJ
cril-239	24	40	the	the	DET
cril-239	24	41	overall	overall	ADJ
cril-239	24	42	performance	performance	NOUN
cril-239	24	43	of	of	ADP
cril-239	24	44	our	our	PRON
cril-239	24	45	system	system	NOUN
cril-239	24	46	is	be	AUX
cril-239	24	47	comparable	comparable	ADJ
cril-239	24	48	to	to	ADP
cril-239	24	49	systems	system	NOUN
cril-239	24	50	in	in	ADP
cril-239	24	51	current	current	ADJ
cril-239	24	52	research	research	NOUN
cril-239	24	53	,	,	PUNCT
cril-239	24	54	achieving	achieve	VERB
cril-239	24	55	an	an	DET
cril-239	24	56	accuracy	accuracy	NOUN
cril-239	24	57	of	of	ADP
cril-239	24	58	81	81	NUM
cril-239	24	59	%	%	NOUN
cril-239	24	60	.	.	PUNCT
cril-239	25	1	in	in	ADP
cril-239	25	2	the	the	DET
cril-239	25	3	domain	domain	NOUN
cril-239	25	4	of	of	ADP
cril-239	25	5	ovopi	ovopi	NOUN
cril-239	25	6	of	of	ADP
cril-239	25	7	reviews	review	NOUN
cril-239	25	8	it	it	PRON
cril-239	25	9	is	be	AUX
cril-239	25	10	often	often	ADV
cril-239	25	11	acceptable	acceptable	ADJ
cril-239	25	12	to	to	PART
cril-239	25	13	sacrifice	sacrifice	VERB
cril-239	25	14	recall	recall	NOUN
cril-239	25	15	for	for	ADP
cril-239	25	16	accuracy	accuracy	NOUN
cril-239	25	17	.	.	PUNCT
cril-239	26	1	here	here	ADV
cril-239	26	2	we	we	PRON
cril-239	26	3	also	also	ADV
cril-239	26	4	present	present	VERB
cril-239	26	5	a	a	DET
cril-239	26	6	system	system	NOUN
cril-239	26	7	whereby	whereby	SCONJ
cril-239	26	8	the	the	DET
cril-239	26	9	reviews	review	NOUN
cril-239	26	10	are	be	AUX
cril-239	26	11	ranked	rank	VERB
cril-239	26	12	based	base	VERB
cril-239	26	13	on	on	ADP
cril-239	26	14	a	a	DET
cril-239	26	15	function	function	NOUN
cril-239	26	16	of	of	ADP
cril-239	26	17	the	the	DET
cril-239	26	18	probability	probability	NOUN
cril-239	26	19	of	of	ADP
cril-239	26	20	being	be	AUX
cril-239	26	21	positive	positive	ADJ
cril-239	26	22	/	/	SYM
cril-239	26	23	negative	negative	ADJ
cril-239	26	24	.	.	PUNCT
cril-239	27	1	using	use	VERB
cril-239	27	2	this	this	DET
cril-239	27	3	ranking	ranking	NOUN
cril-239	27	4	method	method	NOUN
cril-239	27	5	we	we	PRON
cril-239	27	6	achieve	achieve	VERB
cril-239	27	7	100	100	NUM
cril-239	27	8	%	%	NOUN
cril-239	27	9	accuracy	accuracy	NOUN
cril-239	27	10	when	when	SCONJ
cril-239	27	11	we	we	PRON
cril-239	27	12	accept	accept	VERB
cril-239	27	13	a	a	DET
cril-239	27	14	recall	recall	NOUN
cril-239	27	15	of	of	ADP
cril-239	27	16	10	10	NUM
cril-239	27	17	%	%	NOUN
cril-239	27	18	.	.	PUNCT
cril-239	28	1	this	this	DET
cril-239	28	2	result	result	NOUN
cril-239	28	3	is	be	AUX
cril-239	28	4	particularly	particularly	ADV
cril-239	28	5	interesting	interesting	ADJ
cril-239	28	6	for	for	ADP
cril-239	28	7	applications	application	NOUN
cril-239	28	8	that	that	PRON
cril-239	28	9	rely	rely	VERB
cril-239	28	10	on	on	ADP
cril-239	28	11	web	web	NOUN
cril-239	28	12	data	datum	NOUN
cril-239	28	13	,	,	PUNCT
cril-239	28	14	because	because	SCONJ
cril-239	28	15	the	the	DET
cril-239	28	16	customer	customer	NOUN
cril-239	28	17	is	be	AUX
cril-239	28	18	not	not	PART
cril-239	28	19	always	always	ADV
cril-239	28	20	interested	interested	ADJ
cril-239	28	21	in	in	ADP
cril-239	28	22	having	have	VERB
cril-239	28	23	all	all	DET
cril-239	28	24	the	the	DET
cril-239	28	25	possible	possible	ADJ
cril-239	28	26	reviews	review	NOUN
cril-239	28	27	,	,	PUNCT
cril-239	28	28	but	but	CCONJ
cril-239	28	29	many	many	ADJ
cril-239	28	30	times	time	NOUN
cril-239	28	31	is	be	AUX
cril-239	28	32	interested	interested	ADJ
cril-239	28	33	in	in	ADP
cril-239	28	34	having	have	VERB
cril-239	28	35	just	just	ADV
cril-239	28	36	a	a	DET
cril-239	28	37	few	few	ADJ
cril-239	28	38	positive	positive	ADJ
cril-239	28	39	and	and	CCONJ
cril-239	28	40	a	a	DET
cril-239	28	41	few	few	ADJ
cril-239	28	42	negative	negative	NOUN
cril-239	28	43	.	.	PUNCT
cril-239	29	1	from	from	ADP
cril-239	29	2	this	this	DET
cril-239	29	3	perspective	perspective	ADJ
cril-239	29	4	accuracy	accuracy	NOUN
cril-239	29	5	is	be	AUX
cril-239	29	6	more	more	ADV
cril-239	29	7	important	important	ADJ
cril-239	29	8	than	than	ADP
cril-239	29	9	recall	recall	NOUN
cril-239	29	10	.	.	PUNCT
cril-239	30	1	2	2	X
cril-239	30	2	.	.	X
cril-239	30	3	related	relate	VERB
cril-239	30	4	research	research	NOUN
cril-239	30	5	research	research	NOUN
cril-239	30	6	has	have	AUX
cril-239	30	7	demonstrated	demonstrate	VERB
cril-239	30	8	that	that	SCONJ
cril-239	30	9	there	there	PRON
cril-239	30	10	is	be	VERB
cril-239	30	11	a	a	DET
cril-239	30	12	strong	strong	ADJ
cril-239	30	13	positive	positive	ADJ
cril-239	30	14	correlation	correlation	NOUN
cril-239	30	15	between	between	ADP
cril-239	30	16	the	the	DET
cril-239	30	17	presence	presence	NOUN
cril-239	30	18	of	of	ADP
cril-239	30	19	adjectives	adjective	NOUN
cril-239	30	20	in	in	ADP
cril-239	30	21	a	a	DET
cril-239	30	22	sentence	sentence	NOUN
cril-239	30	23	and	and	CCONJ
cril-239	30	24	the	the	DET
cril-239	30	25	presence	presence	NOUN
cril-239	30	26	of	of	ADP
cril-239	30	27	opinion	opinion	NOUN
cril-239	30	28	(	(	PUNCT
cril-239	30	29	wiebe	wiebe	PROPN
cril-239	30	30	et	et	PROPN
cril-239	30	31	al	al	PROPN
cril-239	30	32	,	,	PUNCT
cril-239	30	33	1999	1999	NUM
cril-239	30	34	)	)	PUNCT
cril-239	30	35	.	.	PUNCT
cril-239	31	1	hatzivassiloglou	hatzivassiloglou	PROPN
cril-239	31	2	et	et	PROPN
cril-239	31	3	al	al	PROPN
cril-239	31	4	.	.	PROPN
cril-239	31	5	combined	combine	VERB
cril-239	31	6	a	a	DET
cril-239	31	7	log	log	NOUN
cril-239	31	8	-	-	PUNCT
cril-239	31	9	linear	linear	NOUN
cril-239	31	10	statistical	statistical	ADJ
cril-239	31	11	model	model	NOUN
cril-239	31	12	that	that	PRON
cril-239	31	13	examined	examine	VERB
cril-239	31	14	the	the	DET
cril-239	31	15	conjunctions	conjunction	NOUN
cril-239	31	16	between	between	ADP
cril-239	31	17	adjectives	adjective	NOUN
cril-239	31	18	,	,	PUNCT
cril-239	31	19	(	(	PUNCT
cril-239	31	20	such	such	ADJ
cril-239	31	21	as	as	ADP
cril-239	31	22	"	"	PUNCT
cril-239	31	23	and	and	CCONJ
cril-239	31	24	"	"	PUNCT
cril-239	31	25	,	,	PUNCT
cril-239	31	26	"	"	PUNCT
cril-239	31	27	but	but	CCONJ
cril-239	31	28	"	"	PUNCT
cril-239	31	29	,	,	PUNCT
cril-239	31	30	"	"	PUNCT
cril-239	31	31	or	or	CCONJ
cril-239	31	32	"	"	PUNCT
cril-239	31	33	)	)	PUNCT
cril-239	31	34	,	,	PUNCT
cril-239	31	35	with	with	ADP
cril-239	31	36	a	a	DET
cril-239	31	37	clustering	clustering	ADJ
cril-239	31	38	algorithm	algorithm	NOUN
cril-239	31	39	that	that	PRON
cril-239	31	40	grouped	group	VERB
cril-239	31	41	the	the	DET
cril-239	31	42	adjectives	adjective	NOUN
cril-239	31	43	into	into	ADP
cril-239	31	44	two	two	NUM
cril-239	31	45	sets	set	NOUN
cril-239	31	46	which	which	PRON
cril-239	31	47	were	be	AUX
cril-239	31	48	then	then	ADV
cril-239	31	49	labeled	label	VERB
cril-239	31	50	positive	positive	ADJ
cril-239	31	51	and	and	CCONJ
cril-239	31	52	negative	negative	ADJ
cril-239	31	53	(	(	PUNCT
cril-239	31	54	hatzivassiloglou	hatzivassiloglou	PROPN
cril-239	31	55	et	et	PROPN
cril-239	31	56	al	al	PROPN
cril-239	31	57	,	,	PUNCT
cril-239	31	58	1997	1997	NUM
cril-239	31	59	)	)	PUNCT
cril-239	31	60	.	.	PUNCT
cril-239	32	1	their	their	PRON
cril-239	32	2	model	model	NOUN
cril-239	32	3	predicted	predict	VERB
cril-239	32	4	whether	whether	SCONJ
cril-239	32	5	adjectives	adjective	NOUN
cril-239	32	6	carried	carry	VERB
cril-239	32	7	positive	positive	ADJ
cril-239	32	8	or	or	CCONJ
cril-239	32	9	negative	negative	ADJ
cril-239	32	10	polarity	polarity	NOUN
cril-239	32	11	with	with	ADP
cril-239	32	12	82	82	NUM
cril-239	32	13	%	%	NOUN
cril-239	32	14	accuracy	accuracy	NOUN
cril-239	32	15	.	.	PUNCT
cril-239	33	1	however	however	ADV
cril-239	33	2	,	,	PUNCT
cril-239	33	3	because	because	SCONJ
cril-239	33	4	the	the	DET
cril-239	33	5	model	model	NOUN
cril-239	33	6	was	be	AUX
cril-239	33	7	unsupervised	unsupervised	ADJ
cril-239	33	8	it	it	PRON
cril-239	33	9	required	require	VERB
cril-239	33	10	an	an	DET
cril-239	33	11	immense	immense	ADJ
cril-239	33	12	,	,	PUNCT
cril-239	33	13	21	21	NUM
cril-239	33	14	million	million	NUM
cril-239	33	15	word	word	NOUN
cril-239	33	16	corpus	corpus	NOUN
cril-239	33	17	to	to	PART
cril-239	33	18	function	function	VERB
cril-239	33	19	.	.	PUNCT
cril-239	34	1	turney	turney	PROPN
cril-239	34	2	extracted	extract	VERB
cril-239	34	3	n	n	CCONJ
cril-239	34	4	-	-	PUNCT
cril-239	34	5	grams	gram	NOUN
cril-239	34	6	based	base	VERB
cril-239	34	7	on	on	ADP
cril-239	34	8	adjectives	adjective	NOUN
cril-239	34	9	(	(	PUNCT
cril-239	34	10	turney	turney	NOUN
cril-239	34	11	,	,	PUNCT
cril-239	34	12	2002	2002	NUM
cril-239	34	13	)	)	PUNCT
cril-239	34	14	.	.	PUNCT
cril-239	35	1	in	in	ADP
cril-239	35	2	order	order	NOUN
cril-239	35	3	to	to	PART
cril-239	35	4	determine	determine	VERB
cril-239	35	5	if	if	SCONJ
cril-239	35	6	an	an	DET
cril-239	35	7	adjective	adjective	ADJ
cril-239	35	8	had	have	VERB
cril-239	35	9	a	a	DET
cril-239	35	10	positive	positive	ADJ
cril-239	35	11	/	/	SYM
cril-239	35	12	negative	negative	ADJ
cril-239	35	13	polarity	polarity	NOUN
cril-239	35	14	he	he	PRON
cril-239	35	15	used	use	VERB
cril-239	35	16	altavista	altavista	PROPN
cril-239	35	17	and	and	CCONJ
cril-239	35	18	its	its	PRON
cril-239	35	19	function	function	NOUN
cril-239	35	20	near	near	ADV
cril-239	35	21	.	.	PUNCT
cril-239	36	1	he	he	PRON
cril-239	36	2	combined	combine	VERB
cril-239	36	3	the	the	DET
cril-239	36	4	number	number	NOUN
cril-239	36	5	of	of	ADP
cril-239	36	6	co	co	NOUN
cril-239	36	7	-	-	NOUN
cril-239	36	8	occurrences	occurrence	NOUN
cril-239	36	9	of	of	ADP
cril-239	36	10	the	the	DET
cril-239	36	11	adjective	adjective	ADJ
cril-239	36	12	under	under	ADP
cril-239	36	13	investigation	investigation	NOUN
cril-239	36	14	near	near	ADP
cril-239	36	15	the	the	DET
cril-239	36	16	adjective	adjective	ADJ
cril-239	36	17	'	'	PUNCT
cril-239	36	18	excellent	excellent	ADJ
cril-239	36	19	'	'	PUNCT
cril-239	36	20	and	and	CCONJ
cril-239	36	21	near	near	ADP
cril-239	36	22	the	the	DET
cril-239	36	23	adjective	adjective	ADJ
cril-239	36	24	'	'	PUNCT
cril-239	36	25	poor	poor	ADJ
cril-239	36	26	'	'	PUNCT
cril-239	36	27	thinking	thinking	NOUN
cril-239	36	28	that	that	SCONJ
cril-239	36	29	high	high	ADJ
cril-239	36	30	occurrence	occurrence	NOUN
cril-239	36	31	near	near	ADP
cril-239	36	32	'	'	PUNCT
cril-239	36	33	poor	poor	ADJ
cril-239	36	34	'	'	PUNCT
cril-239	36	35	implies	imply	VERB
cril-239	36	36	negative	negative	ADJ
cril-239	36	37	polarity	polarity	NOUN
cril-239	36	38	and	and	CCONJ
cril-239	36	39	high	high	ADJ
cril-239	36	40	occurrence	occurrence	NOUN
cril-239	36	41	near	near	ADP
cril-239	36	42	'	'	PUNCT
cril-239	36	43	excellent	excellent	ADJ
cril-239	36	44	'	'	PUNCT
cril-239	36	45	implies	imply	VERB
cril-239	36	46	positive	positive	ADJ
cril-239	36	47	polarity	polarity	NOUN
cril-239	36	48	.	.	PUNCT
cril-239	37	1	turney	turney	PROPN
cril-239	37	2	achieved	achieve	VERB
cril-239	37	3	an	an	DET
cril-239	37	4	average	average	NOUN
cril-239	37	5	of	of	ADP
cril-239	37	6	74	74	NUM
cril-239	37	7	%	%	NOUN
cril-239	37	8	accuracy	accuracy	NOUN
cril-239	37	9	in	in	ADP
cril-239	37	10	ovopi	ovopi	NOUN
cril-239	37	11	across	across	ADP
cril-239	37	12	all	all	DET
cril-239	37	13	domains	domain	NOUN
cril-239	37	14	.	.	PUNCT
cril-239	38	1	the	the	DET
cril-239	38	2	performance	performance	NOUN
cril-239	38	3	on	on	ADP
cril-239	38	4	movie	movie	NOUN
cril-239	38	5	reviews	review	NOUN
cril-239	38	6	,	,	PUNCT
cril-239	38	7	however	however	ADV
cril-239	38	8	,	,	PUNCT
cril-239	38	9	was	be	AUX
cril-239	38	10	especially	especially	ADV
cril-239	38	11	poor	poor	ADJ
cril-239	38	12	at	at	ADP
cril-239	38	13	only	only	ADV
cril-239	38	14	65.8	65.8	NUM
cril-239	38	15	%	%	NOUN
cril-239	38	16	,	,	PUNCT
cril-239	38	17	indicating	indicate	VERB
cril-239	38	18	that	that	SCONJ
cril-239	38	19	ovopi	ovopi	NOUN
cril-239	38	20	for	for	ADP
cril-239	38	21	movie	movie	NOUN
cril-239	38	22	reviews	review	NOUN
cril-239	38	23	is	be	AUX
cril-239	38	24	a	a	DET
cril-239	38	25	more	more	ADV
cril-239	38	26	difficult	difficult	ADJ
cril-239	38	27	task	task	NOUN
cril-239	38	28	than	than	ADP
cril-239	38	29	for	for	ADP
cril-239	38	30	other	other	ADJ
cril-239	38	31	product	product	NOUN
cril-239	38	32	reviews	review	NOUN
cril-239	38	33	.	.	PUNCT
cril-239	39	1	pang	pang	NOUN
cril-239	39	2	et	et	PROPN
cril-239	39	3	al	al	PROPN
cril-239	39	4	.	.	PROPN
cril-239	39	5	concluded	conclude	VERB
cril-239	39	6	that	that	SCONJ
cril-239	39	7	the	the	DET
cril-239	39	8	task	task	NOUN
cril-239	39	9	of	of	ADP
cril-239	39	10	polarity	polarity	NOUN
cril-239	39	11	classification	classification	NOUN
cril-239	39	12	was	be	AUX
cril-239	39	13	not	not	PART
cril-239	39	14	the	the	DET
cril-239	39	15	same	same	ADJ
cril-239	39	16	as	as	ADP
cril-239	39	17	topic	topic	NOUN
cril-239	39	18	classification	classification	NOUN
cril-239	39	19	(	(	PUNCT
cril-239	39	20	pang	pang	NOUN
cril-239	39	21	et	et	PROPN
cril-239	39	22	al	al	PROPN
cril-239	39	23	,	,	PUNCT
cril-239	39	24	2002	2002	NUM
cril-239	39	25	)	)	PUNCT
cril-239	39	26	.	.	PUNCT
cril-239	40	1	they	they	PRON
cril-239	40	2	applied	apply	VERB
cril-239	40	3	naïve	naïve	ADJ
cril-239	40	4	bayes	bayes	NOUN
cril-239	40	5	,	,	PUNCT
cril-239	40	6	maximum	maximum	ADJ
cril-239	40	7	entropy	entropy	NOUN
cril-239	40	8	and	and	CCONJ
cril-239	40	9	support	support	VERB
cril-239	40	10	vector	vector	NOUN
cril-239	40	11	machine	machine	NOUN
cril-239	40	12	classification	classification	NOUN
cril-239	40	13	techniques	technique	NOUN
cril-239	40	14	to	to	ADP
cril-239	40	15	the	the	DET
cril-239	40	16	identification	identification	NOUN
cril-239	40	17	of	of	ADP
cril-239	40	18	the	the	DET
cril-239	40	19	polarity	polarity	NOUN
cril-239	40	20	of	of	ADP
cril-239	40	21	movie	movie	NOUN
cril-239	40	22	reviews	review	NOUN
cril-239	40	23	.	.	PUNCT
cril-239	41	1	they	they	PRON
cril-239	41	2	reported	report	VERB
cril-239	41	3	that	that	SCONJ
cril-239	41	4	the	the	DET
cril-239	41	5	naïve	naïve	ADJ
cril-239	41	6	bayes	bayes	NOUN
cril-239	41	7	method	method	NOUN
cril-239	41	8	returned	return	VERB
cril-239	41	9	a	a	DET
cril-239	41	10	77.3	77.3	NUM
cril-239	41	11	%	%	NOUN
cril-239	41	12	accuracy	accuracy	NOUN
cril-239	41	13	using	use	VERB
cril-239	41	14	bigrams	bigram	NOUN
cril-239	41	15	.	.	PUNCT
cril-239	42	1	their	their	PRON
cril-239	42	2	best	good	ADJ
cril-239	42	3	results	result	NOUN
cril-239	42	4	came	come	VERB
cril-239	42	5	using	use	VERB
cril-239	42	6	unigrams	unigram	NOUN
cril-239	42	7	,	,	PUNCT
cril-239	42	8	calculated	calculate	VERB
cril-239	42	9	by	by	ADP
cril-239	42	10	the	the	DET
cril-239	42	11	support	support	NOUN
cril-239	42	12	vector	vector	NOUN
cril-239	42	13	machine	machine	NOUN
cril-239	42	14	at	at	ADP
cril-239	42	15	82.9	82.9	NUM
cril-239	42	16	%	%	NOUN
cril-239	42	17	accuracy	accuracy	NOUN
cril-239	42	18	.	.	PUNCT
cril-239	43	1	maximum	maximum	ADJ
cril-239	43	2	entropy	entropy	PROPN
cril-239	43	3	performed	perform	VERB
cril-239	43	4	best	well	ADV
cril-239	43	5	using	use	VERB
cril-239	43	6	both	both	PRON
cril-239	43	7	unigrams	unigram	NOUN
cril-239	43	8	and	and	CCONJ
cril-239	43	9	bigrams	bigram	NOUN
cril-239	43	10	at	at	ADP
cril-239	43	11	80.8	80.8	NUM
cril-239	43	12	%	%	NOUN
cril-239	43	13	accuracy	accuracy	NOUN
cril-239	43	14	,	,	PUNCT
cril-239	43	15	and	and	CCONJ
cril-239	43	16	naïve	naïve	ADJ
cril-239	43	17	bayes	baye	NOUN
cril-239	43	18	performed	perform	VERB
cril-239	43	19	best	well	ADV
cril-239	43	20	at	at	ADP
cril-239	43	21	81.5	81.5	NUM
cril-239	43	22	%	%	NOUN
cril-239	43	23	using	use	VERB
cril-239	43	24	unigrams	unigram	NOUN
cril-239	43	25	with	with	ADP
cril-239	43	26	pos	pos	NOUN
cril-239	43	27	tags	tag	NOUN
cril-239	43	28	.	.	PUNCT
cril-239	44	1	2	2	NUM
cril-239	44	2	colorado	colorado	NOUN
cril-239	44	3	research	research	NOUN
cril-239	44	4	in	in	ADP
cril-239	44	5	linguistics	linguistic	NOUN
cril-239	44	6	,	,	PUNCT
cril-239	44	7	vol	vol	NOUN
cril-239	44	8	.	.	PROPN
cril-239	44	9	17	17	NUM
cril-239	45	1	[	[	X
cril-239	45	2	2004	2004	NUM
cril-239	45	3	]	]	X
cril-239	45	4	https://scholar.colorado.edu/cril/vol17/iss1/1	https://scholar.colorado.edu/cril/vol17/iss1/1	PROPN
cril-239	45	5	doi	doi	NOUN
cril-239	45	6	:	:	PUNCT
cril-239	45	7	https://doi.org/10.25810/atv1-v819	https://doi.org/10.25810/atv1-v819	VERB
cril-239	45	8	automatic	automatic	ADJ
cril-239	45	9	opinion	opinion	NOUN
cril-239	45	10	polarity	polarity	NOUN
cril-239	45	11	classification	classification	NOUN
cril-239	45	12	of	of	ADP
cril-239	45	13	movie	movie	NOUN
cril-239	45	14	reviews	review	NOUN
cril-239	45	15	3	3	NUM
cril-239	45	16	3	3	NUM
cril-239	45	17	.	.	PUNCT
cril-239	45	18	statistical	statistical	ADJ
cril-239	45	19	approaches	approach	NOUN
cril-239	45	20	to	to	ADP
cril-239	45	21	polarity	polarity	NOUN
cril-239	45	22	identification	identification	NOUN
cril-239	45	23	there	there	PRON
cril-239	45	24	are	be	VERB
cril-239	45	25	many	many	ADJ
cril-239	45	26	possible	possible	ADJ
cril-239	45	27	approaches	approach	NOUN
cril-239	45	28	to	to	ADP
cril-239	45	29	identifying	identify	VERB
cril-239	45	30	the	the	DET
cril-239	45	31	actual	actual	ADJ
cril-239	45	32	polarity	polarity	NOUN
cril-239	45	33	of	of	ADP
cril-239	45	34	a	a	DET
cril-239	45	35	document	document	NOUN
cril-239	45	36	.	.	PUNCT
cril-239	46	1	our	our	PRON
cril-239	46	2	analysis	analysis	NOUN
cril-239	46	3	uses	use	VERB
cril-239	46	4	statistical	statistical	ADJ
cril-239	46	5	methods	method	NOUN
cril-239	46	6	,	,	PUNCT
cril-239	46	7	namely	namely	ADV
cril-239	46	8	supervised	supervised	ADJ
cril-239	46	9	machine	machine	NOUN
cril-239	46	10	learning	learning	NOUN
cril-239	46	11	,	,	PUNCT
cril-239	46	12	to	to	PART
cril-239	46	13	identify	identify	VERB
cril-239	46	14	the	the	DET
cril-239	46	15	likelihood	likelihood	NOUN
cril-239	46	16	of	of	ADP
cril-239	46	17	reviews	review	NOUN
cril-239	46	18	having	have	VERB
cril-239	46	19	"	"	PUNCT
cril-239	46	20	positive	positive	ADJ
cril-239	46	21	"	"	PUNCT
cril-239	46	22	or	or	CCONJ
cril-239	46	23	"	"	PUNCT
cril-239	46	24	negative	negative	ADJ
cril-239	46	25	"	"	PUNCT
cril-239	46	26	polarity	polarity	NOUN
cril-239	46	27	with	with	ADP
cril-239	46	28	respect	respect	NOUN
cril-239	46	29	to	to	ADP
cril-239	46	30	previously	previously	ADV
cril-239	46	31	hand	hand	NOUN
cril-239	46	32	-	-	PUNCT
cril-239	46	33	classified	classify	VERB
cril-239	46	34	training	training	NOUN
cril-239	46	35	data	datum	NOUN
cril-239	46	36	.	.	PUNCT
cril-239	47	1	these	these	DET
cril-239	47	2	methods	method	NOUN
cril-239	47	3	are	be	AUX
cril-239	47	4	fairly	fairly	ADV
cril-239	47	5	standard	standard	ADJ
cril-239	47	6	and	and	CCONJ
cril-239	47	7	well	well	ADV
cril-239	47	8	understood	understand	VERB
cril-239	47	9	;	;	PUNCT
cril-239	47	10	we	we	PRON
cril-239	47	11	list	list	VERB
cril-239	47	12	them	they	PRON
cril-239	47	13	below	below	ADV
cril-239	47	14	for	for	ADP
cril-239	47	15	the	the	DET
cril-239	47	16	sake	sake	NOUN
cril-239	47	17	of	of	ADP
cril-239	47	18	completeness	completeness	NOUN
cril-239	47	19	.	.	PUNCT
cril-239	48	1	3.1	3.1	NUM
cril-239	48	2	.	.	PUNCT
cril-239	48	3	naïve	naïve	ADJ
cril-239	48	4	bayes	bayes	PROPN
cril-239	48	5	classifier	classifier	AUX
cril-239	48	6	the	the	DET
cril-239	48	7	naïve	naïve	ADJ
cril-239	48	8	bayes	bayes	NOUN
cril-239	48	9	classifier	classifier	NOUN
cril-239	48	10	is	be	AUX
cril-239	48	11	a	a	DET
cril-239	48	12	well	well	ADV
cril-239	48	13	-	-	PUNCT
cril-239	48	14	known	know	VERB
cril-239	48	15	supervised	supervised	ADJ
cril-239	48	16	machine	machine	NOUN
cril-239	48	17	learning	learning	NOUN
cril-239	48	18	approach	approach	NOUN
cril-239	48	19	.	.	PUNCT
cril-239	49	1	in	in	ADP
cril-239	49	2	this	this	DET
cril-239	49	3	paper	paper	NOUN
cril-239	49	4	the	the	DET
cril-239	49	5	"	"	PUNCT
cril-239	49	6	features	feature	NOUN
cril-239	49	7	"	"	PUNCT
cril-239	49	8	used	use	VERB
cril-239	49	9	to	to	PART
cril-239	49	10	develop	develop	VERB
cril-239	49	11	naïve	naïve	ADJ
cril-239	49	12	bayes	baye	NOUN
cril-239	49	13	are	be	AUX
cril-239	49	14	referred	refer	VERB
cril-239	49	15	to	to	ADP
cril-239	49	16	as	as	ADP
cril-239	49	17	"	"	PUNCT
cril-239	49	18	attributes	attribute	NOUN
cril-239	49	19	"	"	PUNCT
cril-239	49	20	to	to	PART
cril-239	49	21	avoid	avoid	VERB
cril-239	49	22	confusion	confusion	NOUN
cril-239	49	23	with	with	ADP
cril-239	49	24	text	text	NOUN
cril-239	49	25	"	"	PUNCT
cril-239	49	26	features	feature	NOUN
cril-239	49	27	.	.	PUNCT
cril-239	49	28	"	"	PUNCT
cril-239	50	1	in	in	ADP
cril-239	50	2	our	our	PRON
cril-239	50	3	approach	approach	NOUN
cril-239	50	4	,	,	PUNCT
cril-239	50	5	all	all	DET
cril-239	50	6	word	word	NOUN
cril-239	50	7	/	/	SYM
cril-239	50	8	pos	pos	NOUN
cril-239	50	9	-	-	PUNCT
cril-239	50	10	tag	tag	NOUN
cril-239	50	11	pairs	pair	NOUN
cril-239	50	12	that	that	PRON
cril-239	50	13	appear	appear	VERB
cril-239	50	14	in	in	ADP
cril-239	50	15	the	the	DET
cril-239	50	16	training	training	NOUN
cril-239	50	17	data	datum	NOUN
cril-239	50	18	are	be	AUX
cril-239	50	19	collected	collect	VERB
cril-239	50	20	and	and	CCONJ
cril-239	50	21	used	use	VERB
cril-239	50	22	as	as	ADP
cril-239	50	23	attributes	attribute	NOUN
cril-239	50	24	.	.	PUNCT
cril-239	51	1	the	the	DET
cril-239	51	2	formula	formula	NOUN
cril-239	51	3	of	of	ADP
cril-239	51	4	our	our	PRON
cril-239	51	5	naïve	naïve	ADJ
cril-239	51	6	bayes	bayes	NOUN
cril-239	51	7	classifier	classifier	NOUN
cril-239	51	8	is	be	AUX
cril-239	51	9	defined	define	VERB
cril-239	51	10	as	as	ADP
cril-239	51	11	where	where	SCONJ
cril-239	51	12	•	•	NOUN
cril-239	51	13	rv	rv	PROPN
cril-239	51	14	is	be	AUX
cril-239	51	15	the	the	DET
cril-239	51	16	review	review	NOUN
cril-239	51	17	under	under	ADP
cril-239	51	18	consideration	consideration	NOUN
cril-239	51	19	,	,	PUNCT
cril-239	51	20	•	•	NUM
cril-239	51	21	w	w	NOUN
cril-239	51	22	is	be	AUX
cril-239	51	23	a	a	DET
cril-239	51	24	word	word	NOUN
cril-239	51	25	/	/	SYM
cril-239	51	26	pos	pos	NOUN
cril-239	51	27	-	-	PUNCT
cril-239	51	28	tag	tag	NOUN
cril-239	51	29	pair	pair	NOUN
cril-239	51	30	that	that	PRON
cril-239	51	31	appears	appear	VERB
cril-239	51	32	in	in	ADP
cril-239	51	33	the	the	DET
cril-239	51	34	given	give	VERB
cril-239	51	35	document	document	NOUN
cril-239	51	36	,	,	PUNCT
cril-239	51	37	•	•	NUM
cril-239	51	38	pr(appw|class	pr(appw|class	NUM
cril-239	51	39	)	)	PUNCT
cril-239	51	40	is	be	AUX
cril-239	51	41	the	the	DET
cril-239	51	42	probability	probability	NOUN
cril-239	51	43	that	that	SCONJ
cril-239	51	44	a	a	DET
cril-239	51	45	word	word	NOUN
cril-239	51	46	/	/	SYM
cril-239	51	47	pos	pos	NOUN
cril-239	51	48	-	-	PUNCT
cril-239	51	49	tag	tag	NOUN
cril-239	51	50	pair	pair	NOUN
cril-239	51	51	appears	appear	VERB
cril-239	51	52	in	in	ADP
cril-239	51	53	a	a	DET
cril-239	51	54	document	document	NOUN
cril-239	51	55	of	of	ADP
cril-239	51	56	the	the	DET
cril-239	51	57	given	give	VERB
cril-239	51	58	class	class	NOUN
cril-239	51	59	in	in	ADP
cril-239	51	60	training	training	NOUN
cril-239	51	61	data	datum	NOUN
cril-239	51	62	,	,	PUNCT
cril-239	51	63	and	and	CCONJ
cril-239	51	64	•	•	NUM
cril-239	51	65	bc	bc	PROPN
cril-239	51	66	is	be	AUX
cril-239	51	67	an	an	DET
cril-239	51	68	estimated	estimate	VERB
cril-239	51	69	class	class	NOUN
cril-239	51	70	.	.	PUNCT
cril-239	52	1	one	one	NUM
cril-239	52	2	interesting	interesting	ADJ
cril-239	52	3	aspect	aspect	NOUN
cril-239	52	4	of	of	ADP
cril-239	52	5	this	this	DET
cril-239	52	6	particular	particular	ADJ
cril-239	52	7	application	application	NOUN
cril-239	52	8	of	of	ADP
cril-239	52	9	naïve	naïve	ADJ
cril-239	52	10	bayes	bayes	NOUN
cril-239	52	11	is	be	AUX
cril-239	52	12	that	that	SCONJ
cril-239	52	13	most	most	ADJ
cril-239	52	14	attributes	attribute	NOUN
cril-239	52	15	do	do	AUX
cril-239	52	16	not	not	PART
cril-239	52	17	appear	appear	VERB
cril-239	52	18	in	in	ADP
cril-239	52	19	a	a	DET
cril-239	52	20	test	test	NOUN
cril-239	52	21	review	review	NOUN
cril-239	52	22	,	,	PUNCT
cril-239	52	23	which	which	PRON
cril-239	52	24	means	mean	VERB
cril-239	52	25	most	most	ADJ
cril-239	52	26	factors	factor	NOUN
cril-239	52	27	in	in	ADP
cril-239	52	28	the	the	DET
cril-239	52	29	product	product	NOUN
cril-239	52	30	probability	probability	NOUN
cril-239	52	31	are	be	AUX
cril-239	52	32	based	base	VERB
cril-239	52	33	on	on	ADP
cril-239	52	34	what	what	PRON
cril-239	52	35	is	be	AUX
cril-239	52	36	not	not	PART
cril-239	52	37	written	write	VERB
cril-239	52	38	in	in	ADP
cril-239	52	39	a	a	DET
cril-239	52	40	review	review	NOUN
cril-239	52	41	.	.	PUNCT
cril-239	53	1	this	this	PRON
cril-239	53	2	is	be	AUX
cril-239	53	3	one	one	NUM
cril-239	53	4	major	major	ADJ
cril-239	53	5	difference	difference	NOUN
cril-239	53	6	from	from	ADP
cril-239	53	7	the	the	DET
cril-239	53	8	markov	markov	NOUN
cril-239	53	9	model	model	NOUN
cril-239	53	10	classifier	classifier	PROPN
cril-239	53	11	described	describe	VERB
cril-239	53	12	in	in	ADP
cril-239	53	13	the	the	DET
cril-239	53	14	next	next	ADJ
cril-239	53	15	section	section	NOUN
cril-239	53	16	.	.	PUNCT
cril-239	54	1	3.2	3.2	NUM
cril-239	54	2	.	.	PUNCT
cril-239	54	3	classifier	classifier	NOUN
cril-239	54	4	based	base	VERB
cril-239	54	5	on	on	ADP
cril-239	54	6	markov	markov	NOUN
cril-239	54	7	models	model	NOUN
cril-239	54	8	because	because	SCONJ
cril-239	54	9	the	the	DET
cril-239	54	10	naïve	naïve	ADJ
cril-239	54	11	bayes	bayes	NOUN
cril-239	54	12	classifier	classifier	NOUN
cril-239	54	13	defined	define	VERB
cril-239	54	14	in	in	ADP
cril-239	54	15	the	the	DET
cril-239	54	16	previous	previous	ADJ
cril-239	54	17	section	section	NOUN
cril-239	54	18	builds	build	VERB
cril-239	54	19	probabilistic	probabilistic	ADJ
cril-239	54	20	models	model	NOUN
cril-239	54	21	based	base	VERB
cril-239	54	22	on	on	ADP
cril-239	54	23	individual	individual	ADJ
cril-239	54	24	occurrences	occurrence	NOUN
cril-239	54	25	of	of	ADP
cril-239	54	26	words	word	NOUN
cril-239	54	27	,	,	PUNCT
cril-239	54	28	it	it	PRON
cril-239	54	29	is	be	AUX
cril-239	54	30	provided	provide	VERB
cril-239	54	31	with	with	ADP
cril-239	54	32	relatively	relatively	ADV
cril-239	54	33	little	little	ADJ
cril-239	54	34	information	information	NOUN
cril-239	54	35	regarding	regard	VERB
cril-239	54	36	the	the	DET
cril-239	54	37	phrasal	phrasal	ADJ
cril-239	54	38	structure	structure	NOUN
cril-239	54	39	.	.	PUNCT
cril-239	55	1	markov	markov	NOUN
cril-239	55	2	model	model	NOUN
cril-239	55	3	is	be	AUX
cril-239	55	4	a	a	DET
cril-239	55	5	widely	widely	ADV
cril-239	55	6	used	use	VERB
cril-239	55	7	probabilistic	probabilistic	ADJ
cril-239	55	8	model	model	NOUN
cril-239	55	9	that	that	PRON
cril-239	55	10	does	do	AUX
cril-239	55	11	capture	capture	VERB
cril-239	55	12	connectivity	connectivity	NOUN
cril-239	55	13	among	among	ADP
cril-239	55	14	words	word	NOUN
cril-239	55	15	.	.	PUNCT
cril-239	56	1	this	this	DET
cril-239	56	2	markov	markov	NOUN
cril-239	56	3	model	model	NOUN
cril-239	56	4	classifier	classifier	NOUN
cril-239	56	5	develops	develop	VERB
cril-239	56	6	two	two	NUM
cril-239	56	7	language	language	NOUN
cril-239	56	8	models	model	NOUN
cril-239	56	9	:	:	PUNCT
cril-239	56	10	one	one	NUM
cril-239	56	11	on	on	ADP
cril-239	56	12	positive	positive	ADJ
cril-239	56	13	reviews	review	NOUN
cril-239	56	14	and	and	CCONJ
cril-239	56	15	another	another	PRON
cril-239	56	16	on	on	ADP
cril-239	56	17	negative	negative	ADJ
cril-239	56	18	reviews	review	NOUN
cril-239	56	19	.	.	PUNCT
cril-239	57	1	the	the	DET
cril-239	57	2	classifier	classifier	NOUN
cril-239	57	3	then	then	ADV
cril-239	57	4	generates	generate	VERB
cril-239	57	5	two	two	NUM
cril-239	57	6	probabilities	probability	NOUN
cril-239	57	7	for	for	ADP
cril-239	57	8	an	an	DET
cril-239	57	9	unseen	unseen	ADJ
cril-239	57	10	review	review	NOUN
cril-239	57	11	,	,	PUNCT
cril-239	57	12	one	one	NUM
cril-239	57	13	from	from	ADP
cril-239	57	14	the	the	DET
cril-239	57	15	positive	positive	ADJ
cril-239	57	16	model	model	NOUN
cril-239	57	17	and	and	CCONJ
cril-239	57	18	the	the	DET
cril-239	57	19	other	other	ADJ
cril-239	57	20	from	from	ADP
cril-239	57	21	the	the	DET
cril-239	57	22	negative	negative	ADJ
cril-239	57	23	one	one	NUM
cril-239	57	24	.	.	PUNCT
cril-239	58	1	it	it	PRON
cril-239	58	2	compares	compare	VERB
cril-239	58	3	the	the	DET
cril-239	58	4	two	two	NUM
cril-239	58	5	probabilities	probability	NOUN
cril-239	58	6	and	and	CCONJ
cril-239	58	7	determines	determine	VERB
cril-239	58	8	the	the	DET
cril-239	58	9	classification	classification	NOUN
cril-239	58	10	.	.	PUNCT
cril-239	59	1	the	the	DET
cril-239	59	2	following	follow	VERB
cril-239	59	3	formula	formula	NOUN
cril-239	59	4	is	be	AUX
cril-239	59	5	the	the	DET
cril-239	59	6	one	one	NOUN
cril-239	59	7	used	use	VERB
cril-239	59	8	to	to	PART
cril-239	59	9	compute	compute	VERB
cril-239	59	10	the	the	DET
cril-239	59	11	probability	probability	NOUN
cril-239	59	12	that	that	SCONJ
cril-239	59	13	a	a	DET
cril-239	59	14	document	document	NOUN
cril-239	59	15	could	could	AUX
cril-239	59	16	be	be	AUX
cril-239	59	17	generated	generate	VERB
cril-239	59	18	using	use	VERB
cril-239	59	19	each	each	DET
cril-239	59	20	language	language	NOUN
cril-239	59	21	model	model	NOUN
cril-239	59	22	.	.	PUNCT
cril-239	60	1	3	3	NUM
cril-239	60	2	salvetti	salvetti	VERB
cril-239	60	3	et	et	PROPN
cril-239	60	4	al	al	PROPN
cril-239	60	5	.	.	PROPN
cril-239	60	6	:	:	PUNCT
cril-239	61	1	automatic	automatic	ADJ
cril-239	61	2	opinion	opinion	NOUN
cril-239	61	3	polarity	polarity	NOUN
cril-239	61	4	classification	classification	NOUN
cril-239	61	5	of	of	ADP
cril-239	61	6	movie	movie	NOUN
cril-239	61	7	reviews	review	NOUN
cril-239	61	8	published	publish	VERB
cril-239	61	9	by	by	ADP
cril-239	61	10	cu	cu	PROPN
cril-239	61	11	scholar	scholar	NOUN
cril-239	61	12	,	,	PUNCT
cril-239	61	13	2004	2004	NUM
cril-239	61	14	colorado	colorado	PROPN
cril-239	61	15	research	research	NOUN
cril-239	61	16	in	in	ADP
cril-239	61	17	linguistics	linguistic	NOUN
cril-239	61	18	,	,	PUNCT
cril-239	61	19	volume	volume	NOUN
cril-239	61	20	17	17	NUM
cril-239	61	21	(	(	PUNCT
cril-239	61	22	2004	2004	NUM
cril-239	61	23	)	)	PUNCT
cril-239	61	24	4	4	NUM
cril-239	61	25	where	where	SCONJ
cril-239	61	26	•	•	NUM
cril-239	61	27	rv	rv	PROPN
cril-239	61	28	is	be	AUX
cril-239	61	29	the	the	DET
cril-239	61	30	review	review	NOUN
cril-239	61	31	under	under	ADP
cril-239	61	32	consideration	consideration	NOUN
cril-239	61	33	,	,	PUNCT
cril-239	61	34	•	•	NUM
cril-239	61	35	sn	sn	PROPN
cril-239	61	36	is	be	AUX
cril-239	61	37	a	a	DET
cril-239	61	38	sentence	sentence	NOUN
cril-239	61	39	,	,	PUNCT
cril-239	61	40	•	•	NUM
cril-239	61	41	〈	〈	NOUN
cril-239	61	42	s	s	PRON
cril-239	61	43	〉	〉	NOUN
cril-239	61	44	is	be	AUX
cril-239	61	45	the	the	DET
cril-239	61	46	start	start	NOUN
cril-239	61	47	of	of	ADP
cril-239	61	48	a	a	DET
cril-239	61	49	sentence	sentence	NOUN
cril-239	61	50	,	,	PUNCT
cril-239	61	51	•	•	NUM
cril-239	61	52	〈	〈	NOUN
cril-239	61	53	/s	/	NOUN
cril-239	61	54	〉	〉	NOUN
cril-239	61	55	is	be	AUX
cril-239	61	56	the	the	DET
cril-239	61	57	end	end	NOUN
cril-239	61	58	of	of	ADP
cril-239	61	59	a	a	DET
cril-239	61	60	sentence	sentence	NOUN
cril-239	61	61	.	.	PUNCT
cril-239	62	1	4	4	X
cril-239	62	2	.	.	X
cril-239	62	3	features	feature	NOUN
cril-239	62	4	for	for	ADP
cril-239	62	5	analysis	analysis	NOUN
cril-239	62	6	statistical	statistical	ADJ
cril-239	62	7	analysis	analysis	NOUN
cril-239	62	8	depends	depend	VERB
cril-239	62	9	on	on	ADP
cril-239	62	10	a	a	DET
cril-239	62	11	sequence	sequence	NOUN
cril-239	62	12	of	of	ADP
cril-239	62	13	tokens	token	NOUN
cril-239	62	14	it	it	PRON
cril-239	62	15	uses	use	VERB
cril-239	62	16	as	as	ADP
cril-239	62	17	characteristic	characteristic	ADJ
cril-239	62	18	features	feature	NOUN
cril-239	62	19	of	of	ADP
cril-239	62	20	the	the	DET
cril-239	62	21	objects	object	NOUN
cril-239	62	22	it	it	PRON
cril-239	62	23	attempts	attempt	VERB
cril-239	62	24	to	to	PART
cril-239	62	25	analyze	analyze	VERB
cril-239	62	26	;	;	PUNCT
cril-239	62	27	the	the	DET
cril-239	62	28	only	only	ADJ
cril-239	62	29	necessary	necessary	ADJ
cril-239	62	30	property	property	NOUN
cril-239	62	31	of	of	ADP
cril-239	62	32	these	these	DET
cril-239	62	33	features	feature	NOUN
cril-239	62	34	is	be	AUX
cril-239	62	35	that	that	SCONJ
cril-239	62	36	it	it	PRON
cril-239	62	37	must	must	AUX
cril-239	62	38	be	be	AUX
cril-239	62	39	possible	possible	ADJ
cril-239	62	40	to	to	PART
cril-239	62	41	identify	identify	VERB
cril-239	62	42	whether	whether	SCONJ
cril-239	62	43	two	two	NUM
cril-239	62	44	features	feature	NOUN
cril-239	62	45	are	be	AUX
cril-239	62	46	equal	equal	ADJ
cril-239	62	47	.	.	PUNCT
cril-239	63	1	the	the	DET
cril-239	63	2	most	most	ADV
cril-239	63	3	straightforward	straightforward	ADJ
cril-239	63	4	way	way	NOUN
cril-239	63	5	of	of	ADP
cril-239	63	6	dealing	deal	VERB
cril-239	63	7	with	with	ADP
cril-239	63	8	the	the	DET
cril-239	63	9	information	information	NOUN
cril-239	63	10	we	we	PRON
cril-239	63	11	find	find	VERB
cril-239	63	12	within	within	ADP
cril-239	63	13	reviews	review	NOUN
cril-239	63	14	would	would	AUX
cril-239	63	15	be	be	AUX
cril-239	63	16	to	to	PART
cril-239	63	17	use	use	VERB
cril-239	63	18	individual	individual	ADJ
cril-239	63	19	words	word	NOUN
cril-239	63	20	from	from	ADP
cril-239	63	21	the	the	DET
cril-239	63	22	review	review	NOUN
cril-239	63	23	data	datum	NOUN
cril-239	63	24	as	as	ADP
cril-239	63	25	tokens	token	NOUN
cril-239	63	26	.	.	PUNCT
cril-239	64	1	however	however	ADV
cril-239	64	2	,	,	PUNCT
cril-239	64	3	just	just	ADV
cril-239	64	4	using	use	VERB
cril-239	64	5	the	the	DET
cril-239	64	6	words	word	NOUN
cril-239	64	7	discards	discard	VERB
cril-239	64	8	semantic	semantic	ADJ
cril-239	64	9	information	information	NOUN
cril-239	64	10	about	about	ADP
cril-239	64	11	the	the	DET
cril-239	64	12	remainder	remainder	NOUN
cril-239	64	13	of	of	ADP
cril-239	64	14	the	the	DET
cril-239	64	15	sentence	sentence	NOUN
cril-239	64	16	;	;	PUNCT
cril-239	64	17	as	as	ADP
cril-239	64	18	such	such	ADJ
cril-239	64	19	,	,	PUNCT
cril-239	64	20	it	it	PRON
cril-239	64	21	may	may	AUX
cril-239	64	22	be	be	AUX
cril-239	64	23	desirable	desirable	ADJ
cril-239	64	24	to	to	PART
cril-239	64	25	first	first	ADV
cril-239	64	26	perform	perform	VERB
cril-239	64	27	some	some	DET
cril-239	64	28	sort	sort	NOUN
cril-239	64	29	of	of	ADP
cril-239	64	30	semantic	semantic	ADJ
cril-239	64	31	analysis	analysis	NOUN
cril-239	64	32	to	to	PART
cril-239	64	33	enrich	enrich	VERB
cril-239	64	34	the	the	DET
cril-239	64	35	tokens	token	NOUN
cril-239	64	36	with	with	ADP
cril-239	64	37	useful	useful	ADJ
cril-239	64	38	information	information	NOUN
cril-239	64	39	,	,	PUNCT
cril-239	64	40	or	or	CCONJ
cril-239	64	41	even	even	ADV
cril-239	64	42	discard	discard	VERB
cril-239	64	43	misleading	misleading	ADJ
cril-239	64	44	or	or	CCONJ
cril-239	64	45	irrelevant	irrelevant	ADJ
cril-239	64	46	information	information	NOUN
cril-239	64	47	(	(	PUNCT
cril-239	64	48	noise	noise	NOUN
cril-239	64	49	)	)	PUNCT
cril-239	64	50	,	,	PUNCT
cril-239	64	51	in	in	ADP
cril-239	64	52	order	order	NOUN
cril-239	64	53	to	to	PART
cril-239	64	54	increase	increase	VERB
cril-239	64	55	accuracy	accuracy	NOUN
cril-239	64	56	.	.	PUNCT
cril-239	65	1	three	three	NUM
cril-239	65	2	basic	basic	ADJ
cril-239	65	3	approaches	approach	NOUN
cril-239	65	4	for	for	ADP
cril-239	65	5	handling	handle	VERB
cril-239	65	6	this	this	DET
cril-239	65	7	kind	kind	NOUN
cril-239	65	8	of	of	ADP
cril-239	65	9	data	data	NOUN
cril-239	65	10	preprocessing	preprocesse	VERB
cril-239	65	11	come	come	VERB
cril-239	65	12	to	to	ADP
cril-239	65	13	mind	mind	NOUN
cril-239	65	14	:	:	PUNCT
cril-239	65	15	•	•	NOUN
cril-239	65	16	leave	leave	VERB
cril-239	65	17	the	the	DET
cril-239	65	18	data	datum	NOUN
cril-239	65	19	as	as	SCONJ
cril-239	65	20	-	-	PUNCT
cril-239	65	21	is	be	AUX
cril-239	65	22	:	:	PUNCT
cril-239	65	23	each	each	DET
cril-239	65	24	word	word	NOUN
cril-239	65	25	will	will	AUX
cril-239	65	26	be	be	AUX
cril-239	65	27	represented	represent	VERB
cril-239	65	28	by	by	ADP
cril-239	65	29	itself	itself	PRON
cril-239	65	30	•	•	NOUN
cril-239	65	31	parts	part	NOUN
cril-239	65	32	-	-	PUNCT
cril-239	65	33	of	of	ADP
cril-239	65	34	-	-	PUNCT
cril-239	65	35	speech	speech	NOUN
cril-239	65	36	tagging	tagging	NOUN
cril-239	65	37	:	:	PUNCT
cril-239	65	38	each	each	DET
cril-239	65	39	word	word	NOUN
cril-239	65	40	is	be	AUX
cril-239	65	41	enriched	enrich	VERB
cril-239	65	42	by	by	ADP
cril-239	65	43	a	a	DET
cril-239	65	44	pos	pos	NOUN
cril-239	65	45	tag	tag	NOUN
cril-239	65	46	,	,	PUNCT
cril-239	65	47	as	as	SCONJ
cril-239	65	48	determined	determine	VERB
cril-239	65	49	by	by	ADP
cril-239	65	50	a	a	DET
cril-239	65	51	standard	standard	ADJ
cril-239	65	52	tagging	tagging	NOUN
cril-239	65	53	technique	technique	NOUN
cril-239	65	54	(	(	PUNCT
cril-239	65	55	such	such	ADJ
cril-239	65	56	as	as	ADP
cril-239	65	57	the	the	DET
cril-239	65	58	brill	brill	NOUN
cril-239	65	59	tagger	tagger	NOUN
cril-239	65	60	(	(	PUNCT
cril-239	65	61	brill	brill	PROPN
cril-239	65	62	,	,	PUNCT
cril-239	65	63	1995	1995	NUM
cril-239	65	64	)	)	PUNCT
cril-239	65	65	)	)	PUNCT
cril-239	65	66	•	•	NUM
cril-239	65	67	perform	perform	VERB
cril-239	65	68	pos	pos	NOUN
cril-239	65	69	tagging	tagging	NOUN
cril-239	65	70	and	and	CCONJ
cril-239	65	71	parse	parse	NOUN
cril-239	65	72	(	(	PUNCT
cril-239	65	73	using	use	VERB
cril-239	65	74	e.g.	e.g.	ADV
cril-239	65	75	the	the	DET
cril-239	65	76	penn	penn	PROPN
cril-239	65	77	treebank	treebank	PROPN
cril-239	65	78	(	(	PUNCT
cril-239	65	79	marcus	marcus	PROPN
cril-239	65	80	et	et	PROPN
cril-239	65	81	al	al	PROPN
cril-239	65	82	,	,	PUNCT
cril-239	65	83	1994	1994	NUM
cril-239	65	84	)	)	PUNCT
cril-239	65	85	)	)	PUNCT
cril-239	66	1	unfortunately	unfortunately	ADV
cril-239	66	2	,	,	PUNCT
cril-239	66	3	the	the	DET
cril-239	66	4	third	third	ADJ
cril-239	66	5	approach	approach	NOUN
cril-239	66	6	not	not	PART
cril-239	66	7	only	only	ADV
cril-239	66	8	had	have	VERB
cril-239	66	9	severe	severe	ADJ
cril-239	66	10	performance	performance	NOUN
cril-239	66	11	issues	issue	NOUN
cril-239	66	12	during	during	ADP
cril-239	66	13	our	our	PRON
cril-239	66	14	early	early	ADJ
cril-239	66	15	experiments	experiment	NOUN
cril-239	66	16	,	,	PUNCT
cril-239	66	17	but	but	CCONJ
cril-239	66	18	also	also	ADV
cril-239	66	19	raises	raise	VERB
cril-239	66	20	conceptional	conceptional	ADJ
cril-239	66	21	questions	question	NOUN
cril-239	66	22	of	of	ADP
cril-239	66	23	how	how	SCONJ
cril-239	66	24	such	such	ADJ
cril-239	66	25	data	datum	NOUN
cril-239	66	26	would	would	AUX
cril-239	66	27	be	be	AUX
cril-239	66	28	incorporated	incorporate	VERB
cril-239	66	29	into	into	ADP
cril-239	66	30	a	a	DET
cril-239	66	31	statistical	statistical	ADJ
cril-239	66	32	analysis	analysis	NOUN
cril-239	66	33	.	.	PUNCT
cril-239	67	1	we	we	PRON
cril-239	67	2	thus	thus	ADV
cril-239	67	3	focus	focus	VERB
cril-239	67	4	our	our	PRON
cril-239	67	5	analysis	analysis	NOUN
cril-239	67	6	in	in	ADP
cril-239	67	7	this	this	DET
cril-239	67	8	paper	paper	NOUN
cril-239	67	9	on	on	ADP
cril-239	67	10	postagged	postagge	VERB
cril-239	67	11	data	datum	NOUN
cril-239	67	12	(	(	PUNCT
cril-239	67	13	sentences	sentence	NOUN
cril-239	67	14	consisting	consist	VERB
cril-239	67	15	of	of	ADP
cril-239	67	16	words	word	NOUN
cril-239	67	17	enriched	enrich	VERB
cril-239	67	18	with	with	ADP
cril-239	67	19	information	information	NOUN
cril-239	67	20	about	about	ADP
cril-239	67	21	their	their	PRON
cril-239	67	22	parts	part	NOUN
cril-239	67	23	of	of	ADP
cril-239	67	24	speech	speech	NOUN
cril-239	67	25	)	)	PUNCT
cril-239	67	26	,	,	PUNCT
cril-239	67	27	which	which	PRON
cril-239	67	28	seems	seem	VERB
cril-239	67	29	to	to	PART
cril-239	67	30	be	be	AUX
cril-239	67	31	a	a	DET
cril-239	67	32	good	good	ADJ
cril-239	67	33	candidate	candidate	NOUN
cril-239	67	34	for	for	ADP
cril-239	67	35	a	a	DET
cril-239	67	36	worthwhile	worthwhile	ADJ
cril-239	67	37	source	source	NOUN
cril-239	67	38	of	of	ADP
cril-239	67	39	information	information	NOUN
cril-239	67	40	,	,	PUNCT
cril-239	67	41	for	for	ADP
cril-239	67	42	the	the	DET
cril-239	67	43	following	follow	VERB
cril-239	67	44	reasons	reason	NOUN
cril-239	67	45	:	:	PUNCT
cril-239	67	46	1	1	X
cril-239	67	47	.	.	X
cril-239	67	48	as	as	SCONJ
cril-239	67	49	discussed	discuss	VERB
cril-239	67	50	by	by	ADP
cril-239	67	51	losee	losee	NOUN
cril-239	67	52	(	(	PUNCT
cril-239	67	53	losee	losee	NOUN
cril-239	67	54	,	,	PUNCT
cril-239	67	55	2001	2001	NUM
cril-239	67	56	)	)	PUNCT
cril-239	67	57	,	,	PUNCT
cril-239	67	58	information	information	NOUN
cril-239	67	59	retrieval	retrieval	NOUN
cril-239	67	60	with	with	ADP
cril-239	67	61	pos	pos	NOUN
cril-239	67	62	-	-	PUNCT
cril-239	67	63	tagged	tag	VERB
cril-239	67	64	data	datum	NOUN
cril-239	67	65	improves	improve	VERB
cril-239	67	66	the	the	DET
cril-239	67	67	quality	quality	NOUN
cril-239	67	68	of	of	ADP
cril-239	67	69	an	an	DET
cril-239	67	70	analysis	analysis	NOUN
cril-239	67	71	in	in	ADP
cril-239	67	72	many	many	ADJ
cril-239	67	73	cases	case	NOUN
cril-239	67	74	,	,	PUNCT
cril-239	67	75	2	2	X
cril-239	67	76	.	.	X
cril-239	68	1	it	it	PRON
cril-239	68	2	is	be	AUX
cril-239	68	3	a	a	DET
cril-239	68	4	computationally	computationally	ADV
cril-239	68	5	inexpensive	inexpensive	ADJ
cril-239	68	6	way	way	NOUN
cril-239	68	7	of	of	ADP
cril-239	68	8	increasing	increase	VERB
cril-239	68	9	the	the	DET
cril-239	68	10	amount	amount	NOUN
cril-239	68	11	of	of	ADP
cril-239	68	12	(	(	PUNCT
cril-239	68	13	potentially	potentially	ADV
cril-239	68	14	)	)	PUNCT
cril-239	68	15	relevant	relevant	ADJ
cril-239	68	16	information	information	NOUN
cril-239	68	17	,	,	PUNCT
cril-239	68	18	3	3	X
cril-239	68	19	.	.	X
cril-239	69	1	it	it	PRON
cril-239	69	2	gives	give	VERB
cril-239	69	3	rise	rise	NOUN
cril-239	69	4	to	to	ADP
cril-239	69	5	pos	pos	NOUN
cril-239	69	6	-	-	PUNCT
cril-239	69	7	based	base	VERB
cril-239	69	8	filtering	filtering	NOUN
cril-239	69	9	techniques	technique	NOUN
cril-239	69	10	for	for	ADP
cril-239	69	11	further	further	ADJ
cril-239	69	12	refinement	refinement	NOUN
cril-239	69	13	,	,	PUNCT
cril-239	69	14	as	as	SCONJ
cril-239	69	15	we	we	PRON
cril-239	69	16	discuss	discuss	VERB
cril-239	69	17	below	below	ADV
cril-239	69	18	.	.	PUNCT
cril-239	70	1	we	we	PRON
cril-239	70	2	thus	thus	ADV
cril-239	70	3	make	make	VERB
cril-239	70	4	the	the	DET
cril-239	70	5	following	follow	VERB
cril-239	70	6	assumptions	assumption	NOUN
cril-239	70	7	about	about	ADP
cril-239	70	8	our	our	PRON
cril-239	70	9	test	test	NOUN
cril-239	70	10	and	and	CCONJ
cril-239	70	11	training	training	NOUN
cril-239	70	12	data	datum	NOUN
cril-239	70	13	:	:	PUNCT
cril-239	70	14	4	4	NUM
cril-239	70	15	colorado	colorado	NOUN
cril-239	70	16	research	research	NOUN
cril-239	70	17	in	in	ADP
cril-239	70	18	linguistics	linguistic	NOUN
cril-239	70	19	,	,	PUNCT
cril-239	70	20	vol	vol	NOUN
cril-239	70	21	.	.	PROPN
cril-239	70	22	17	17	NUM
cril-239	71	1	[	[	X
cril-239	71	2	2004	2004	NUM
cril-239	71	3	]	]	X
cril-239	71	4	https://scholar.colorado.edu/cril/vol17/iss1/1	https://scholar.colorado.edu/cril/vol17/iss1/1	PROPN
cril-239	71	5	doi	doi	NOUN
cril-239	71	6	:	:	PUNCT
cril-239	71	7	https://doi.org/10.25810/atv1-v819	https://doi.org/10.25810/atv1-v819	VERB
cril-239	71	8	automatic	automatic	ADJ
cril-239	71	9	opinion	opinion	NOUN
cril-239	71	10	polarity	polarity	NOUN
cril-239	71	11	classification	classification	NOUN
cril-239	71	12	of	of	ADP
cril-239	71	13	movie	movie	NOUN
cril-239	71	14	reviews	review	NOUN
cril-239	71	15	5	5	NUM
cril-239	71	16	1	1	NUM
cril-239	71	17	.	.	PUNCT
cril-239	72	1	all	all	DET
cril-239	72	2	words	word	NOUN
cril-239	72	3	are	be	AUX
cril-239	72	4	transformed	transform	VERB
cril-239	72	5	into	into	ADP
cril-239	72	6	upper	upper	ADJ
cril-239	72	7	case	case	NOUN
cril-239	72	8	,	,	PUNCT
cril-239	72	9	2	2	X
cril-239	72	10	.	.	X
cril-239	73	1	all	all	DET
cril-239	73	2	words	word	NOUN
cril-239	73	3	are	be	AUX
cril-239	73	4	stemmed	stem	VERB
cril-239	73	5	,	,	PUNCT
cril-239	73	6	3	3	X
cril-239	73	7	.	.	X
cril-239	74	1	all	all	DET
cril-239	74	2	words	word	NOUN
cril-239	74	3	are	be	AUX
cril-239	74	4	transformed	transform	VERB
cril-239	74	5	into	into	ADP
cril-239	74	6	(	(	PUNCT
cril-239	74	7	word	word	NOUN
cril-239	74	8	,	,	PUNCT
cril-239	74	9	pos	pos	NOUN
cril-239	74	10	)	)	PUNCT
cril-239	74	11	tuples	tuple	NOUN
cril-239	74	12	by	by	ADP
cril-239	74	13	pos	pos	NOUN
cril-239	74	14	tagging	tagging	NOUN
cril-239	74	15	(	(	PUNCT
cril-239	74	16	notation	notation	NOUN
cril-239	74	17	word	word	NOUN
cril-239	74	18	/	/	SYM
cril-239	74	19	pos	pos	NOUN
cril-239	74	20	)	)	PUNCT
cril-239	74	21	.	.	PUNCT
cril-239	75	1	all	all	PRON
cril-239	75	2	of	of	ADP
cril-239	75	3	these	these	PRON
cril-239	75	4	are	be	AUX
cril-239	75	5	computationally	computationally	ADV
cril-239	75	6	easy	easy	ADJ
cril-239	75	7	to	to	PART
cril-239	75	8	achieve	achieve	VERB
cril-239	75	9	(	(	PUNCT
cril-239	75	10	with	with	ADP
cril-239	75	11	a	a	DET
cril-239	75	12	reasonable	reasonable	ADJ
cril-239	75	13	amount	amount	NOUN
cril-239	75	14	of	of	ADP
cril-239	75	15	accuracy	accuracy	NOUN
cril-239	75	16	)	)	PUNCT
cril-239	75	17	using	use	VERB
cril-239	75	18	the	the	DET
cril-239	75	19	brill	brill	NOUN
cril-239	75	20	tagger	tagger	NOUN
cril-239	75	21	.	.	PUNCT
cril-239	76	1	5	5	X
cril-239	76	2	.	.	X
cril-239	76	3	part	part	NOUN
cril-239	76	4	of	of	ADP
cril-239	76	5	speech	speech	NOUN
cril-239	76	6	filters	filter	VERB
cril-239	76	7	careful	careful	ADJ
cril-239	76	8	analysis	analysis	NOUN
cril-239	76	9	of	of	ADP
cril-239	76	10	movie	movie	NOUN
cril-239	76	11	reviews	review	NOUN
cril-239	76	12	has	have	AUX
cril-239	76	13	made	make	VERB
cril-239	76	14	it	it	PRON
cril-239	76	15	clear	clear	ADJ
cril-239	76	16	that	that	SCONJ
cril-239	76	17	even	even	ADV
cril-239	76	18	the	the	DET
cril-239	76	19	most	most	ADV
cril-239	76	20	positive	positive	ADJ
cril-239	76	21	reviews	review	NOUN
cril-239	76	22	have	have	VERB
cril-239	76	23	portions	portion	NOUN
cril-239	76	24	with	with	ADP
cril-239	76	25	negative	negative	ADJ
cril-239	76	26	polarity	polarity	NOUN
cril-239	76	27	or	or	CCONJ
cril-239	76	28	no	no	DET
cril-239	76	29	clear	clear	ADJ
cril-239	76	30	polarity	polarity	NOUN
cril-239	76	31	at	at	ADV
cril-239	76	32	all	all	ADV
cril-239	76	33	.	.	PUNCT
cril-239	77	1	since	since	SCONJ
cril-239	77	2	the	the	DET
cril-239	77	3	training	training	NOUN
cril-239	77	4	data	datum	NOUN
cril-239	77	5	used	use	VERB
cril-239	77	6	here	here	ADV
cril-239	77	7	consists	consist	VERB
cril-239	77	8	of	of	ADP
cril-239	77	9	complete	complete	ADJ
cril-239	77	10	classified	classified	ADJ
cril-239	77	11	reviews	review	NOUN
cril-239	77	12	,	,	PUNCT
cril-239	77	13	the	the	DET
cril-239	77	14	presence	presence	NOUN
cril-239	77	15	of	of	ADP
cril-239	77	16	parts	part	NOUN
cril-239	77	17	with	with	ADP
cril-239	77	18	conflicting	conflicting	ADJ
cril-239	77	19	polarities	polarity	NOUN
cril-239	77	20	or	or	CCONJ
cril-239	77	21	lack	lack	NOUN
cril-239	77	22	of	of	ADP
cril-239	77	23	polarity	polarity	NOUN
cril-239	77	24	within	within	ADP
cril-239	77	25	a	a	DET
cril-239	77	26	review	review	NOUN
cril-239	77	27	presents	present	VERB
cril-239	77	28	a	a	DET
cril-239	77	29	major	major	ADJ
cril-239	77	30	obstacle	obstacle	NOUN
cril-239	77	31	for	for	ADP
cril-239	77	32	accurate	accurate	ADJ
cril-239	77	33	ovopi	ovopi	PROPN
cril-239	77	34	.	.	PUNCT
cril-239	78	1	as	as	ADP
cril-239	78	2	illustration	illustration	NOUN
cril-239	78	3	of	of	ADP
cril-239	78	4	this	this	DET
cril-239	78	5	inconsistent	inconsistent	ADJ
cril-239	78	6	polarity	polarity	NOUN
cril-239	78	7	,	,	PUNCT
cril-239	78	8	the	the	DET
cril-239	78	9	following	follow	VERB
cril-239	78	10	were	be	AUX
cril-239	78	11	all	all	PRON
cril-239	78	12	taken	take	VERB
cril-239	78	13	from	from	ADP
cril-239	78	14	a	a	DET
cril-239	78	15	single	single	ADJ
cril-239	78	16	review1	review1	NOUN
cril-239	78	17	.	.	PUNCT
cril-239	79	1	"	"	PUNCT
cril-239	79	2	special	special	ADJ
cril-239	79	3	effects	effect	NOUN
cril-239	79	4	are	be	AUX
cril-239	79	5	first	first	ADJ
cril-239	79	6	-	-	PUNCT
cril-239	79	7	rate	rate	NOUN
cril-239	79	8	"	"	PUNCT
cril-239	79	9	(	(	PUNCT
cril-239	79	10	positive	positive	ADJ
cril-239	79	11	polarity	polarity	NOUN
cril-239	79	12	)	)	PUNCT
cril-239	79	13	"	"	PUNCT
cril-239	79	14	the	the	DET
cril-239	79	15	character	character	NOUN
cril-239	79	16	is	be	AUX
cril-239	79	17	written	write	VERB
cril-239	79	18	thinly	thinly	ADV
cril-239	79	19	"	"	PUNCT
cril-239	79	20	(	(	PUNCT
cril-239	79	21	negative	negative	ADJ
cril-239	79	22	polarity	polarity	NOUN
cril-239	79	23	)	)	PUNCT
cril-239	79	24	"	"	PUNCT
cril-239	79	25	the	the	DET
cril-239	79	26	scenes	scene	NOUN
cril-239	79	27	were	be	AUX
cril-239	79	28	shot	shoot	VERB
cril-239	79	29	in	in	ADP
cril-239	79	30	short	short	ADJ
cril-239	79	31	segments	segment	NOUN
cril-239	79	32	"	"	PUNCT
cril-239	79	33	(	(	PUNCT
cril-239	79	34	no	no	DET
cril-239	79	35	clear	clear	ADJ
cril-239	79	36	polarity	polarity	NOUN
cril-239	79	37	)	)	PUNCT
cril-239	79	38	this	this	DET
cril-239	79	39	observation	observation	NOUN
cril-239	79	40	can	can	AUX
cril-239	79	41	be	be	AUX
cril-239	79	42	taken	take	VERB
cril-239	79	43	to	to	ADP
cril-239	79	44	lower	low	ADJ
cril-239	79	45	levels	level	NOUN
cril-239	79	46	as	as	ADV
cril-239	79	47	well	well	ADV
cril-239	79	48	.	.	PUNCT
cril-239	80	1	individual	individual	ADJ
cril-239	80	2	phrases	phrase	NOUN
cril-239	80	3	and	and	CCONJ
cril-239	80	4	words	word	NOUN
cril-239	80	5	vary	vary	VERB
cril-239	80	6	in	in	ADP
cril-239	80	7	their	their	PRON
cril-239	80	8	contribution	contribution	NOUN
cril-239	80	9	to	to	ADP
cril-239	80	10	opinion	opinion	NOUN
cril-239	80	11	polarity	polarity	NOUN
cril-239	80	12	.	.	PUNCT
cril-239	81	1	it	it	PRON
cril-239	81	2	may	may	AUX
cril-239	81	3	even	even	ADV
cril-239	81	4	be	be	AUX
cril-239	81	5	said	say	VERB
cril-239	81	6	that	that	SCONJ
cril-239	81	7	only	only	ADV
cril-239	81	8	some	some	DET
cril-239	81	9	part	part	NOUN
cril-239	81	10	of	of	ADP
cril-239	81	11	the	the	DET
cril-239	81	12	meaning	meaning	NOUN
cril-239	81	13	of	of	ADP
cril-239	81	14	a	a	DET
cril-239	81	15	word	word	NOUN
cril-239	81	16	contributes	contribute	VERB
cril-239	81	17	to	to	ADP
cril-239	81	18	opinion	opinion	NOUN
cril-239	81	19	polarity	polarity	NOUN
cril-239	81	20	(	(	PUNCT
cril-239	81	21	see	see	VERB
cril-239	81	22	wordnet	wordnet	NOUN
cril-239	81	23	filter	filter	NOUN
cril-239	81	24	section	section	NOUN
cril-239	81	25	below	below	ADP
cril-239	81	26	)	)	PUNCT
cril-239	81	27	.	.	PUNCT
cril-239	82	1	any	any	DET
cril-239	82	2	portion	portion	NOUN
cril-239	82	3	that	that	PRON
cril-239	82	4	does	do	AUX
cril-239	82	5	not	not	PART
cril-239	82	6	contribute	contribute	VERB
cril-239	82	7	to	to	ADP
cril-239	82	8	the	the	DET
cril-239	82	9	ovop	ovop	NOUN
cril-239	82	10	is	be	AUX
cril-239	82	11	noise	noise	NOUN
cril-239	82	12	.	.	PUNCT
cril-239	83	1	to	to	PART
cril-239	83	2	reduce	reduce	VERB
cril-239	83	3	noise	noise	NOUN
cril-239	83	4	,	,	PUNCT
cril-239	83	5	filters	filter	NOUN
cril-239	83	6	were	be	AUX
cril-239	83	7	developed	develop	VERB
cril-239	83	8	that	that	PRON
cril-239	83	9	use	use	VERB
cril-239	83	10	pos	pos	NOUN
cril-239	83	11	tags	tag	NOUN
cril-239	83	12	to	to	PART
cril-239	83	13	do	do	VERB
cril-239	83	14	the	the	DET
cril-239	83	15	following	following	NOUN
cril-239	83	16	.	.	PUNCT
cril-239	84	1	1	1	X
cril-239	84	2	.	.	X
cril-239	84	3	introduce	introduce	VERB
cril-239	84	4	custom	custom	NOUN
cril-239	84	5	parts	part	NOUN
cril-239	84	6	of	of	ADP
cril-239	84	7	speech	speech	NOUN
cril-239	84	8	when	when	SCONJ
cril-239	84	9	the	the	DET
cril-239	84	10	tagger	tagger	NOUN
cril-239	84	11	does	do	AUX
cril-239	84	12	not	not	PART
cril-239	84	13	provide	provide	VERB
cril-239	84	14	desired	desire	VERB
cril-239	84	15	specificity	specificity	NOUN
cril-239	84	16	(	(	PUNCT
cril-239	84	17	negation	negation	NOUN
cril-239	84	18	and	and	CCONJ
cril-239	84	19	copula	copula	NOUN
cril-239	84	20	)	)	PUNCT
cril-239	84	21	.	.	PUNCT
cril-239	85	1	2	2	X
cril-239	85	2	.	.	X
cril-239	85	3	remove	remove	VERB
cril-239	85	4	the	the	DET
cril-239	85	5	words	word	NOUN
cril-239	85	6	that	that	PRON
cril-239	85	7	are	be	AUX
cril-239	85	8	least	least	ADV
cril-239	85	9	likely	likely	ADJ
cril-239	85	10	to	to	PART
cril-239	85	11	contribute	contribute	VERB
cril-239	85	12	to	to	ADP
cril-239	85	13	the	the	DET
cril-239	85	14	polarity	polarity	NOUN
cril-239	85	15	of	of	ADP
cril-239	85	16	a	a	DET
cril-239	85	17	review	review	NOUN
cril-239	85	18	(	(	PUNCT
cril-239	85	19	determiner	determiner	NOUN
cril-239	85	20	,	,	PUNCT
cril-239	85	21	preposition	preposition	NOUN
cril-239	85	22	,	,	PUNCT
cril-239	85	23	etc	etc	X
cril-239	85	24	.	.	X
cril-239	85	25	)	)	PUNCT
cril-239	86	1	3	3	X
cril-239	86	2	.	.	X
cril-239	86	3	reduce	reduce	VERB
cril-239	86	4	parts	part	NOUN
cril-239	86	5	of	of	ADP
cril-239	86	6	speech	speech	NOUN
cril-239	86	7	that	that	PRON
cril-239	86	8	introduce	introduce	VERB
cril-239	86	9	unnecessary	unnecessary	ADJ
cril-239	86	10	variance	variance	NOUN
cril-239	86	11	to	to	PART
cril-239	86	12	pos	pos	VERB
cril-239	86	13	only	only	ADV
cril-239	86	14	.	.	PUNCT
cril-239	87	1	it	it	PRON
cril-239	87	2	may	may	AUX
cril-239	87	3	be	be	AUX
cril-239	87	4	useful	useful	ADJ
cril-239	87	5	,	,	PUNCT
cril-239	87	6	for	for	ADP
cril-239	87	7	instance	instance	NOUN
cril-239	87	8	,	,	PUNCT
cril-239	87	9	for	for	SCONJ
cril-239	87	10	the	the	DET
cril-239	87	11	classifier	classifier	NOUN
cril-239	87	12	to	to	PART
cril-239	87	13	record	record	VERB
cril-239	87	14	the	the	DET
cril-239	87	15	presence	presence	NOUN
cril-239	87	16	of	of	ADP
cril-239	87	17	a	a	DET
cril-239	87	18	proper	proper	ADJ
cril-239	87	19	noun	noun	NOUN
cril-239	87	20	.	.	PUNCT
cril-239	88	1	however	however	ADV
cril-239	88	2	,	,	PUNCT
cril-239	88	3	to	to	PART
cril-239	88	4	include	include	VERB
cril-239	88	5	individual	individual	ADJ
cril-239	88	6	proper	proper	ADJ
cril-239	88	7	nouns	noun	NOUN
cril-239	88	8	would	would	AUX
cril-239	88	9	unnecessarily	unnecessarily	ADV
cril-239	88	10	decrease	decrease	VERB
cril-239	88	11	the	the	DET
cril-239	88	12	probability	probability	NOUN
cril-239	88	13	of	of	ADP
cril-239	88	14	finding	find	VERB
cril-239	88	15	the	the	DET
cril-239	88	16	same	same	ADJ
cril-239	88	17	n	n	CCONJ
cril-239	88	18	-	-	PUNCT
cril-239	88	19	grams	gram	NOUN
cril-239	88	20	in	in	ADP
cril-239	88	21	the	the	DET
cril-239	88	22	test	test	NOUN
cril-239	88	23	data	datum	NOUN
cril-239	88	24	.	.	PUNCT
cril-239	89	1	experimentation	experimentation	NOUN
cril-239	89	2	involved	involve	VERB
cril-239	89	3	multiple	multiple	ADJ
cril-239	89	4	combinations	combination	NOUN
cril-239	89	5	of	of	ADP
cril-239	89	6	such	such	ADJ
cril-239	89	7	filter	filter	NOUN
cril-239	89	8	rules	rule	NOUN
cril-239	89	9	,	,	PUNCT
cril-239	89	10	yielding	yield	VERB
cril-239	89	11	several	several	ADJ
cril-239	89	12	separate	separate	ADJ
cril-239	89	13	filters	filter	NOUN
cril-239	89	14	.	.	PUNCT
cril-239	90	1	an	an	DET
cril-239	90	2	example	example	NOUN
cril-239	90	3	of	of	ADP
cril-239	90	4	a	a	DET
cril-239	90	5	specification	specification	NOUN
cril-239	90	6	of	of	ADP
cril-239	90	7	pos	pos	NOUN
cril-239	90	8	filter	filter	NOUN
cril-239	90	9	rules	rule	NOUN
cril-239	90	10	is	be	AUX
cril-239	90	11	shown	show	VERB
cril-239	90	12	in	in	ADP
cril-239	90	13	figure	figure	NOUN
cril-239	90	14	1	1	NUM
cril-239	90	15	.	.	SYM
cril-239	90	16	1	1	NUM
cril-239	90	17	apollo	apollo	PROPN
cril-239	90	18	13	13	NUM
cril-239	90	19	,	,	PUNCT
cril-239	90	20	a	a	DET
cril-239	90	21	film	film	NOUN
cril-239	90	22	review	review	NOUN
cril-239	90	23	by	by	ADP
cril-239	90	24	mark	mark	PROPN
cril-239	90	25	r.	r.	PROPN
cril-239	90	26	leeper	leeper	PROPN
cril-239	90	27	,	,	PUNCT
cril-239	90	28	copyright	copyright	NOUN
cril-239	90	29	©	©	PROPN
cril-239	90	30	1995	1995	NUM
cril-239	90	31	mark	mark	PROPN
cril-239	90	32	r.	r.	PROPN
cril-239	90	33	leeper	leeper	PROPN
cril-239	90	34	5	5	NUM
cril-239	90	35	salvetti	salvetti	VERB
cril-239	90	36	et	et	PROPN
cril-239	90	37	al	al	PROPN
cril-239	90	38	.	.	PROPN
cril-239	90	39	:	:	PUNCT
cril-239	91	1	automatic	automatic	ADJ
cril-239	91	2	opinion	opinion	NOUN
cril-239	91	3	polarity	polarity	NOUN
cril-239	91	4	classification	classification	NOUN
cril-239	91	5	of	of	ADP
cril-239	91	6	movie	movie	NOUN
cril-239	91	7	reviews	review	NOUN
cril-239	91	8	published	publish	VERB
cril-239	91	9	by	by	ADP
cril-239	91	10	cu	cu	PROPN
cril-239	91	11	scholar	scholar	NOUN
cril-239	91	12	,	,	PUNCT
cril-239	91	13	2004	2004	NUM
cril-239	91	14	colorado	colorado	PROPN
cril-239	91	15	research	research	NOUN
cril-239	91	16	in	in	ADP
cril-239	91	17	linguistics	linguistic	NOUN
cril-239	91	18	,	,	PUNCT
cril-239	91	19	volume	volume	NOUN
cril-239	91	20	17	17	NUM
cril-239	91	21	(	(	PUNCT
cril-239	91	22	2004	2004	NUM
cril-239	91	23	)	)	PUNCT
cril-239	91	24	6	6	NUM
cril-239	91	25	(	(	PUNCT
cril-239	91	26	1a	1a	NUM
cril-239	91	27	)	)	PUNCT
cril-239	91	28	copula	copula	NOUN
cril-239	91	29	conversion	conversion	NOUN
cril-239	91	30	:	:	PUNCT
cril-239	91	31	is/	is/	PROPN
cril-239	91	32	*	*	PUNCT
cril-239	92	1	*	*	PUNCT
cril-239	92	2	/cop	/cop	X
cril-239	92	3	(	(	PUNCT
cril-239	92	4	1b	1b	NUM
cril-239	92	5	)	)	PUNCT
cril-239	92	6	negation	negation	NOUN
cril-239	92	7	conversion	conversion	NOUN
cril-239	92	8	:	:	PUNCT
cril-239	92	9	not/	not/	PROPN
cril-239	92	10	*	*	X
cril-239	92	11	/neg	/neg	X
cril-239	92	12	(	(	PUNCT
cril-239	92	13	2	2	X
cril-239	92	14	)	)	PUNCT
cril-239	92	15	noun	noun	NOUN
cril-239	92	16	generalization	generalization	NOUN
cril-239	92	17	:	:	PUNCT
cril-239	92	18	*	*	PUNCT
cril-239	92	19	/nn	/nn	PUNCT
cril-239	92	20	/nn	/nn	PUNCT
cril-239	92	21	(	(	PUNCT
cril-239	92	22	3	3	X
cril-239	92	23	)	)	PUNCT
cril-239	92	24	pos	pos	NOUN
cril-239	92	25	tossing	tossing	NOUN
cril-239	92	26	:	:	PUNCT
cril-239	93	1	*	*	PUNCT
cril-239	93	2	/cc	/cc	NOUN
cril-239	94	1	ø	ø	NOUN
cril-239	94	2	figure	figure	NOUN
cril-239	94	3	1	1	NUM
cril-239	94	4	:	:	PUNCT
cril-239	94	5	abbreviated	abbreviate	VERB
cril-239	94	6	filter	filter	NOUN
cril-239	94	7	rule	rule	NOUN
cril-239	94	8	specification	specification	NOUN
cril-239	94	9	(	(	PUNCT
cril-239	94	10	illustrative	illustrative	ADJ
cril-239	94	11	details	detail	NOUN
cril-239	94	12	only	only	ADV
cril-239	94	13	)	)	PUNCT
cril-239	94	14	the	the	DET
cril-239	94	15	pos	pos	NOUN
cril-239	94	16	filters	filter	NOUN
cril-239	94	17	are	be	AUX
cril-239	94	18	not	not	PART
cril-239	94	19	designed	design	VERB
cril-239	94	20	to	to	PART
cril-239	94	21	reduce	reduce	VERB
cril-239	94	22	the	the	DET
cril-239	94	23	effects	effect	NOUN
cril-239	94	24	of	of	ADP
cril-239	94	25	conflicting	conflicting	ADJ
cril-239	94	26	polarity	polarity	NOUN
cril-239	94	27	.	.	PUNCT
cril-239	95	1	they	they	PRON
cril-239	95	2	are	be	AUX
cril-239	95	3	only	only	ADV
cril-239	95	4	designed	design	VERB
cril-239	95	5	to	to	PART
cril-239	95	6	reduce	reduce	VERB
cril-239	95	7	the	the	DET
cril-239	95	8	effect	effect	NOUN
cril-239	95	9	of	of	ADP
cril-239	95	10	lack	lack	NOUN
cril-239	95	11	of	of	ADP
cril-239	95	12	polarity	polarity	NOUN
cril-239	95	13	.	.	PUNCT
cril-239	96	1	the	the	DET
cril-239	96	2	effects	effect	NOUN
cril-239	96	3	of	of	ADP
cril-239	96	4	conflicting	conflicting	ADJ
cril-239	96	5	polarity	polarity	NOUN
cril-239	96	6	have	have	AUX
cril-239	96	7	instead	instead	ADV
cril-239	96	8	been	be	AUX
cril-239	96	9	addressed	address	VERB
cril-239	96	10	by	by	ADP
cril-239	96	11	careful	careful	ADJ
cril-239	96	12	preparation	preparation	NOUN
cril-239	96	13	of	of	ADP
cril-239	96	14	the	the	DET
cril-239	96	15	training	training	NOUN
cril-239	96	16	data	datum	NOUN
cril-239	96	17	as	as	SCONJ
cril-239	96	18	will	will	AUX
cril-239	96	19	be	be	AUX
cril-239	96	20	seen	see	VERB
cril-239	96	21	in	in	ADP
cril-239	96	22	the	the	DET
cril-239	96	23	following	follow	VERB
cril-239	96	24	section	section	NOUN
cril-239	96	25	.	.	PUNCT
cril-239	97	1	6	6	X
cril-239	97	2	.	.	PUNCT
cril-239	97	3	experiments	experiment	NOUN
cril-239	97	4	6.1	6.1	NUM
cril-239	97	5	.	.	PUNCT
cril-239	98	1	settings	setting	NOUN
cril-239	98	2	•	•	NUM
cril-239	98	3	data	datum	NOUN
cril-239	98	4	:	:	PUNCT
cril-239	98	5	taken	take	VERB
cril-239	98	6	from	from	ADP
cril-239	98	7	cornell	cornell	PROPN
cril-239	98	8	data	datum	NOUN
cril-239	98	9	(	(	PUNCT
cril-239	98	10	pang	pang	NOUN
cril-239	98	11	et	et	PROPN
cril-239	98	12	al	al	PROPN
cril-239	98	13	,	,	PUNCT
cril-239	98	14	2002	2002	NUM
cril-239	98	15	)	)	PUNCT
cril-239	98	16	•	•	NUM
cril-239	98	17	part	part	NOUN
cril-239	98	18	-	-	PUNCT
cril-239	98	19	of	of	ADP
cril-239	98	20	-	-	PUNCT
cril-239	98	21	speech	speech	NOUN
cril-239	98	22	tagger	tagger	NOUN
cril-239	98	23	:	:	PUNCT
cril-239	98	24	brill	brill	PROPN
cril-239	98	25	tagger	tagger	NOUN
cril-239	98	26	(	(	PUNCT
cril-239	98	27	brill	brill	PROPN
cril-239	98	28	,	,	PUNCT
cril-239	98	29	1995	1995	NUM
cril-239	98	30	)	)	PUNCT
cril-239	98	31	•	•	NUM
cril-239	98	32	wordnet	wordnet	NOUN
cril-239	98	33	:	:	PUNCT
cril-239	98	34	wordnet	wordnet	NOUN
cril-239	98	35	version	version	PROPN
cril-239	98	36	1.7.13	1.7.13	NUM
cril-239	98	37	(	(	PUNCT
cril-239	98	38	fellbaum	fellbaum	NOUN
cril-239	98	39	,	,	PUNCT
cril-239	98	40	1998	1998	NUM
cril-239	98	41	)	)	PUNCT
cril-239	98	42	movie	movie	NOUN
cril-239	98	43	reviews	review	NOUN
cril-239	98	44	are	be	AUX
cril-239	98	45	used	use	VERB
cril-239	98	46	for	for	ADP
cril-239	98	47	training	training	NOUN
cril-239	98	48	and	and	CCONJ
cril-239	98	49	evaluation	evaluation	NOUN
cril-239	98	50	of	of	ADP
cril-239	98	51	each	each	DET
cril-239	98	52	statistical	statistical	ADJ
cril-239	98	53	classifier	classifier	NOUN
cril-239	98	54	.	.	PUNCT
cril-239	99	1	the	the	DET
cril-239	99	2	decision	decision	NOUN
cril-239	99	3	to	to	PART
cril-239	99	4	use	use	VERB
cril-239	99	5	only	only	ADJ
cril-239	99	6	movie	movie	NOUN
cril-239	99	7	reviews	review	NOUN
cril-239	99	8	for	for	ADP
cril-239	99	9	training	training	NOUN
cril-239	99	10	and	and	CCONJ
cril-239	99	11	test	test	NOUN
cril-239	99	12	data	datum	NOUN
cril-239	99	13	was	be	AUX
cril-239	99	14	based	base	VERB
cril-239	99	15	on	on	ADP
cril-239	99	16	the	the	DET
cril-239	99	17	fact	fact	NOUN
cril-239	99	18	that	that	SCONJ
cril-239	99	19	ovopi	ovopi	NOUN
cril-239	99	20	of	of	ADP
cril-239	99	21	movie	movie	NOUN
cril-239	99	22	reviews	review	NOUN
cril-239	99	23	is	be	AUX
cril-239	99	24	particularly	particularly	ADV
cril-239	99	25	challenging	challenging	ADJ
cril-239	99	26	as	as	SCONJ
cril-239	99	27	shown	show	VERB
cril-239	99	28	by	by	ADP
cril-239	99	29	turney	turney	PROPN
cril-239	99	30	(	(	PUNCT
cril-239	99	31	turney	turney	PROPN
cril-239	99	32	,	,	PUNCT
cril-239	99	33	2002	2002	NUM
cril-239	99	34	)	)	PUNCT
cril-239	99	35	,	,	PUNCT
cril-239	99	36	and	and	CCONJ
cril-239	99	37	therefore	therefore	ADV
cril-239	99	38	can	can	AUX
cril-239	99	39	be	be	AUX
cril-239	99	40	considered	consider	VERB
cril-239	99	41	a	a	DET
cril-239	99	42	good	good	ADJ
cril-239	99	43	environment	environment	NOUN
cril-239	99	44	for	for	ADP
cril-239	99	45	testing	test	VERB
cril-239	99	46	any	any	DET
cril-239	99	47	system	system	NOUN
cril-239	99	48	designed	design	VERB
cril-239	99	49	for	for	ADP
cril-239	99	50	ovopi	ovopi	PROPN
cril-239	99	51	.	.	PUNCT
cril-239	100	1	the	the	DET
cril-239	100	2	other	other	ADJ
cril-239	100	3	reason	reason	NOUN
cril-239	100	4	for	for	ADP
cril-239	100	5	using	use	VERB
cril-239	100	6	movie	movie	NOUN
cril-239	100	7	reviews	review	NOUN
cril-239	100	8	is	be	AUX
cril-239	100	9	the	the	DET
cril-239	100	10	availability	availability	NOUN
cril-239	100	11	of	of	ADP
cril-239	100	12	large	large	ADJ
cril-239	100	13	bodies	body	NOUN
cril-239	100	14	of	of	ADP
cril-239	100	15	free	free	ADJ
cril-239	100	16	data	datum	NOUN
cril-239	100	17	on	on	ADP
cril-239	100	18	the	the	DET
cril-239	100	19	web	web	NOUN
cril-239	100	20	.	.	PUNCT
cril-239	101	1	specifically	specifically	ADV
cril-239	101	2	we	we	PRON
cril-239	101	3	used	use	VERB
cril-239	101	4	the	the	DET
cril-239	101	5	data	datum	NOUN
cril-239	101	6	available	available	ADJ
cril-239	101	7	through	through	ADP
cril-239	101	8	cornell	cornell	PROPN
cril-239	101	9	university	university	PROPN
cril-239	101	10	from	from	ADP
cril-239	101	11	the	the	DET
cril-239	101	12	internet	internet	NOUN
cril-239	101	13	movie	movie	NOUN
cril-239	101	14	database	database	NOUN
cril-239	101	15	.	.	PUNCT
cril-239	102	1	the	the	DET
cril-239	102	2	cornell	cornell	PROPN
cril-239	102	3	data	data	NOUN
cril-239	102	4	consists	consist	VERB
cril-239	102	5	of	of	ADP
cril-239	102	6	27,000	27,000	NUM
cril-239	102	7	movie	movie	NOUN
cril-239	102	8	reviews	review	NOUN
cril-239	102	9	in	in	ADP
cril-239	102	10	html	html	NOUN
cril-239	102	11	form	form	NOUN
cril-239	102	12	,	,	PUNCT
cril-239	102	13	using	use	VERB
cril-239	102	14	35	35	NUM
cril-239	102	15	different	different	ADJ
cril-239	102	16	rating	rating	NOUN
cril-239	102	17	scales	scale	NOUN
cril-239	102	18	such	such	ADJ
cril-239	102	19	as	as	ADP
cril-239	102	20	a.	a.	NOUN
cril-239	102	21	.	.	PUNCT
cril-239	102	22	.	.	PUNCT
cril-239	103	1	f	f	PROPN
cril-239	103	2	or	or	CCONJ
cril-239	103	3	1	1	NUM
cril-239	103	4	.	.	PUNCT
cril-239	103	5	.	.	PUNCT
cril-239	103	6	.	.	PUNCT
cril-239	104	1	10	10	NUM
cril-239	105	1	in	in	ADP
cril-239	105	2	addition	addition	NOUN
cril-239	105	3	to	to	ADP
cril-239	105	4	the	the	DET
cril-239	105	5	common	common	ADJ
cril-239	105	6	5	5	NUM
cril-239	105	7	star	star	NOUN
cril-239	105	8	system	system	NOUN
cril-239	105	9	.	.	PUNCT
cril-239	106	1	we	we	PRON
cril-239	106	2	divided	divide	VERB
cril-239	106	3	them	they	PRON
cril-239	106	4	into	into	ADP
cril-239	106	5	two	two	NUM
cril-239	106	6	classes	class	NOUN
cril-239	106	7	(	(	PUNCT
cril-239	106	8	positive	positive	ADJ
cril-239	106	9	and	and	CCONJ
cril-239	106	10	negative	negative	ADJ
cril-239	106	11	)	)	PUNCT
cril-239	106	12	and	and	CCONJ
cril-239	106	13	took	take	VERB
cril-239	106	14	100	100	NUM
cril-239	106	15	reviews	review	NOUN
cril-239	106	16	from	from	ADP
cril-239	106	17	each	each	DET
cril-239	106	18	class	class	NOUN
cril-239	106	19	as	as	ADP
cril-239	106	20	the	the	DET
cril-239	106	21	test	test	NOUN
cril-239	106	22	set	set	VERB
cril-239	106	23	.	.	PUNCT
cril-239	107	1	for	for	ADP
cril-239	107	2	training	training	NOUN
cril-239	107	3	sets	set	NOUN
cril-239	107	4	,	,	PUNCT
cril-239	107	5	we	we	PRON
cril-239	107	6	first	first	ADV
cril-239	107	7	identified	identify	VERB
cril-239	107	8	the	the	DET
cril-239	107	9	reviews	review	NOUN
cril-239	107	10	most	most	ADV
cril-239	107	11	likely	likely	ADJ
cril-239	107	12	to	to	PART
cril-239	107	13	be	be	AUX
cril-239	107	14	positive	positive	ADJ
cril-239	107	15	or	or	CCONJ
cril-239	107	16	negative	negative	ADJ
cril-239	107	17	.	.	PUNCT
cril-239	108	1	for	for	ADP
cril-239	108	2	instance	instance	NOUN
cril-239	108	3	,	,	PUNCT
cril-239	108	4	when	when	SCONJ
cril-239	108	5	reviews	review	NOUN
cril-239	108	6	contained	contain	VERB
cril-239	108	7	letter	letter	NOUN
cril-239	108	8	grade	grade	NOUN
cril-239	108	9	ratings	rating	NOUN
cril-239	108	10	,	,	PUNCT
cril-239	108	11	only	only	ADV
cril-239	108	12	the	the	DET
cril-239	108	13	a	a	PROPN
cril-239	108	14	and	and	CCONJ
cril-239	108	15	f	f	PROPN
cril-239	108	16	reviews	review	NOUN
cril-239	108	17	were	be	AUX
cril-239	108	18	selected	select	VERB
cril-239	108	19	.	.	PUNCT
cril-239	109	1	this	this	PRON
cril-239	109	2	was	be	AUX
cril-239	109	3	done	do	VERB
cril-239	109	4	in	in	ADP
cril-239	109	5	an	an	DET
cril-239	109	6	attempt	attempt	NOUN
cril-239	109	7	to	to	PART
cril-239	109	8	minimize	minimize	VERB
cril-239	109	9	the	the	DET
cril-239	109	10	effects	effect	NOUN
cril-239	109	11	of	of	ADP
cril-239	109	12	conflicting	conflicting	ADJ
cril-239	109	13	polarities	polarity	NOUN
cril-239	109	14	and	and	CCONJ
cril-239	109	15	to	to	PART
cril-239	109	16	maximize	maximize	VERB
cril-239	109	17	the	the	DET
cril-239	109	18	likelihood	likelihood	NOUN
cril-239	109	19	that	that	SCONJ
cril-239	109	20	our	our	PRON
cril-239	109	21	positive	positive	ADJ
cril-239	109	22	and	and	CCONJ
cril-239	109	23	negative	negative	ADJ
cril-239	109	24	labels	label	NOUN
cril-239	109	25	match	match	VERB
cril-239	109	26	those	those	PRON
cril-239	109	27	that	that	PRON
cril-239	109	28	the	the	DET
cril-239	109	29	authors	author	NOUN
cril-239	109	30	would	would	AUX
cril-239	109	31	have	have	AUX
cril-239	109	32	assigned	assign	VERB
cril-239	109	33	to	to	ADP
cril-239	109	34	the	the	DET
cril-239	109	35	reviews	review	NOUN
cril-239	109	36	.	.	PUNCT
cril-239	110	1	from	from	ADP
cril-239	110	2	these	these	DET
cril-239	110	3	reviews	review	NOUN
cril-239	110	4	,	,	PUNCT
cril-239	110	5	we	we	PRON
cril-239	110	6	took	take	VERB
cril-239	110	7	random	random	ADJ
cril-239	110	8	samples	sample	NOUN
cril-239	110	9	from	from	ADP
cril-239	110	10	each	each	DET
cril-239	110	11	class	class	NOUN
cril-239	110	12	in	in	ADP
cril-239	110	13	set	set	ADJ
cril-239	110	14	sizes	size	NOUN
cril-239	110	15	ranging	range	VERB
cril-239	110	16	from	from	ADP
cril-239	110	17	50	50	NUM
cril-239	110	18	to	to	PART
cril-239	110	19	750	750	NUM
cril-239	110	20	reviews	review	NOUN
cril-239	110	21	(	(	PUNCT
cril-239	110	22	in	in	ADP
cril-239	110	23	increments	increment	NOUN
cril-239	110	24	of	of	ADP
cril-239	110	25	50	50	NUM
cril-239	110	26	)	)	PUNCT
cril-239	110	27	.	.	PUNCT
cril-239	111	1	these	these	DET
cril-239	111	2	sets	set	NOUN
cril-239	111	3	consisted	consist	VERB
cril-239	111	4	of	of	ADP
cril-239	111	5	the	the	DET
cril-239	111	6	reviews	review	NOUN
cril-239	111	7	that	that	PRON
cril-239	111	8	remained	remain	VERB
cril-239	111	9	after	after	SCONJ
cril-239	111	10	the	the	DET
cril-239	111	11	test	test	NOUN
cril-239	111	12	sets	set	NOUN
cril-239	111	13	had	have	AUX
cril-239	111	14	been	be	AUX
cril-239	111	15	removed	remove	VERB
cril-239	111	16	.	.	PUNCT
cril-239	112	1	this	this	PRON
cril-239	112	2	resulted	result	VERB
cril-239	112	3	in	in	ADP
cril-239	112	4	training	training	NOUN
cril-239	112	5	set	set	VERB
cril-239	112	6	sizes	size	NOUN
cril-239	112	7	of	of	ADP
cril-239	112	8	100	100	NUM
cril-239	112	9	,	,	PUNCT
cril-239	112	10	200	200	NUM
cril-239	112	11	,	,	PUNCT
cril-239	112	12	...	...	PUNCT
cril-239	112	13	,	,	PUNCT
cril-239	112	14	1500	1500	NUM
cril-239	112	15	(	(	PUNCT
cril-239	112	16	in	in	ADP
cril-239	112	17	increments	increment	NOUN
cril-239	112	18	of	of	ADP
cril-239	112	19	100	100	NUM
cril-239	112	20	)	)	PUNCT
cril-239	112	21	.	.	PUNCT
cril-239	113	1	html	html	NOUN
cril-239	113	2	documents	document	NOUN
cril-239	113	3	were	be	AUX
cril-239	113	4	converted	convert	VERB
cril-239	113	5	to	to	ADP
cril-239	113	6	plain	plain	ADJ
cril-239	113	7	text	text	NOUN
cril-239	113	8	,	,	PUNCT
cril-239	113	9	6	6	NUM
cril-239	113	10	colorado	colorado	NOUN
cril-239	113	11	research	research	NOUN
cril-239	113	12	in	in	ADP
cril-239	113	13	linguistics	linguistic	NOUN
cril-239	113	14	,	,	PUNCT
cril-239	113	15	vol	vol	NOUN
cril-239	113	16	.	.	PROPN
cril-239	113	17	17	17	NUM
cril-239	114	1	[	[	X
cril-239	114	2	2004	2004	NUM
cril-239	114	3	]	]	X
cril-239	114	4	https://scholar.colorado.edu/cril/vol17/iss1/1	https://scholar.colorado.edu/cril/vol17/iss1/1	PROPN
cril-239	114	5	doi	doi	NOUN
cril-239	114	6	:	:	PUNCT
cril-239	114	7	https://doi.org/10.25810/atv1-v819	https://doi.org/10.25810/atv1-v819	VERB
cril-239	114	8	automatic	automatic	ADJ
cril-239	114	9	opinion	opinion	NOUN
cril-239	114	10	polarity	polarity	NOUN
cril-239	114	11	classification	classification	NOUN
cril-239	114	12	of	of	ADP
cril-239	114	13	movie	movie	NOUN
cril-239	114	14	reviews	review	NOUN
cril-239	114	15	7	7	NUM
cril-239	114	16	tagged	tag	VERB
cril-239	114	17	using	use	VERB
cril-239	114	18	the	the	DET
cril-239	114	19	brill	brill	NOUN
cril-239	114	20	tagger	tagger	NOUN
cril-239	114	21	,	,	PUNCT
cril-239	114	22	and	and	CCONJ
cril-239	114	23	fed	feed	VERB
cril-239	114	24	into	into	ADP
cril-239	114	25	filters	filter	NOUN
cril-239	114	26	and	and	CCONJ
cril-239	114	27	classifiers	classifier	NOUN
cril-239	114	28	.	.	PUNCT
cril-239	115	1	the	the	DET
cril-239	115	2	particular	particular	ADJ
cril-239	115	3	combinations	combination	NOUN
cril-239	115	4	of	of	ADP
cril-239	115	5	filters	filter	NOUN
cril-239	115	6	and	and	CCONJ
cril-239	115	7	classifiers	classifier	NOUN
cril-239	115	8	and	and	CCONJ
cril-239	115	9	their	their	PRON
cril-239	115	10	results	result	NOUN
cril-239	115	11	are	be	AUX
cril-239	115	12	described	describe	VERB
cril-239	115	13	in	in	ADP
cril-239	115	14	the	the	DET
cril-239	115	15	following	follow	VERB
cril-239	115	16	sections	section	NOUN
cril-239	115	17	.	.	PUNCT
cril-239	116	1	the	the	DET
cril-239	116	2	fact	fact	NOUN
cril-239	116	3	that	that	SCONJ
cril-239	116	4	as	as	ADP
cril-239	116	5	a	a	DET
cril-239	116	6	training	training	NOUN
cril-239	116	7	set	set	NOUN
cril-239	116	8	we	we	PRON
cril-239	116	9	used	use	VERB
cril-239	116	10	data	datum	NOUN
cril-239	116	11	labeled	label	VERB
cril-239	116	12	by	by	ADP
cril-239	116	13	a	a	DET
cril-239	116	14	reader	reader	NOUN
cril-239	116	15	and	and	CCONJ
cril-239	116	16	not	not	PART
cril-239	116	17	directly	directly	ADV
cril-239	116	18	by	by	ADP
cril-239	116	19	the	the	DET
cril-239	116	20	writer	writer	NOUN
cril-239	116	21	poses	pose	VERB
cril-239	116	22	a	a	DET
cril-239	116	23	potential	potential	ADJ
cril-239	116	24	problem	problem	NOUN
cril-239	116	25	.	.	PUNCT
cril-239	117	1	we	we	PRON
cril-239	117	2	are	be	AUX
cril-239	117	3	learning	learn	VERB
cril-239	117	4	a	a	DET
cril-239	117	5	function	function	NOUN
cril-239	117	6	that	that	PRON
cril-239	117	7	has	have	VERB
cril-239	117	8	to	to	PART
cril-239	117	9	mimic	mimic	VERB
cril-239	117	10	the	the	DET
cril-239	117	11	label	label	NOUN
cril-239	117	12	identified	identify	VERB
cril-239	117	13	by	by	ADP
cril-239	117	14	the	the	DET
cril-239	117	15	writer	writer	NOUN
cril-239	117	16	,	,	PUNCT
cril-239	117	17	but	but	CCONJ
cril-239	117	18	we	we	PRON
cril-239	117	19	are	be	AUX
cril-239	117	20	using	use	VERB
cril-239	117	21	data	datum	NOUN
cril-239	117	22	labeled	label	VERB
cril-239	117	23	by	by	ADP
cril-239	117	24	the	the	DET
cril-239	117	25	reader	reader	NOUN
cril-239	117	26	.	.	PUNCT
cril-239	118	1	we	we	PRON
cril-239	118	2	assume	assume	VERB
cril-239	118	3	that	that	SCONJ
cril-239	118	4	this	this	PRON
cril-239	118	5	is	be	AUX
cril-239	118	6	an	an	DET
cril-239	118	7	acceptable	acceptable	ADJ
cril-239	118	8	approximation	approximation	NOUN
cril-239	118	9	because	because	SCONJ
cril-239	118	10	there	there	PRON
cril-239	118	11	is	be	VERB
cril-239	118	12	a	a	DET
cril-239	118	13	strong	strong	ADJ
cril-239	118	14	practical	practical	ADJ
cril-239	118	15	relation	relation	NOUN
cril-239	118	16	between	between	ADP
cril-239	118	17	the	the	DET
cril-239	118	18	label	label	NOUN
cril-239	118	19	identified	identify	VERB
cril-239	118	20	by	by	ADP
cril-239	118	21	the	the	DET
cril-239	118	22	original	original	ADJ
cril-239	118	23	writer	writer	NOUN
cril-239	118	24	and	and	CCONJ
cril-239	118	25	the	the	DET
cril-239	118	26	reader	reader	NOUN
cril-239	118	27	.	.	PUNCT
cril-239	119	1	the	the	DET
cril-239	119	2	authors	author	NOUN
cril-239	119	3	themselves	themselves	PRON
cril-239	119	4	may	may	AUX
cril-239	119	5	not	not	PART
cril-239	119	6	have	have	AUX
cril-239	119	7	made	make	VERB
cril-239	119	8	the	the	DET
cril-239	119	9	polarity	polarity	NOUN
cril-239	119	10	classifications	classification	NOUN
cril-239	119	11	,	,	PUNCT
cril-239	119	12	but	but	CCONJ
cril-239	119	13	we	we	PRON
cril-239	119	14	assume	assume	VERB
cril-239	119	15	that	that	SCONJ
cril-239	119	16	language	language	NOUN
cril-239	119	17	is	be	AUX
cril-239	119	18	an	an	DET
cril-239	119	19	efficient	efficient	ADJ
cril-239	119	20	form	form	NOUN
cril-239	119	21	of	of	ADP
cril-239	119	22	communication	communication	NOUN
cril-239	119	23	.	.	PUNCT
cril-239	120	1	as	as	ADP
cril-239	120	2	such	such	ADJ
cril-239	120	3	,	,	PUNCT
cril-239	120	4	variances	variance	NOUN
cril-239	120	5	between	between	ADP
cril-239	120	6	author	author	NOUN
cril-239	120	7	and	and	CCONJ
cril-239	120	8	reader	reader	NOUN
cril-239	120	9	classification	classification	NOUN
cril-239	120	10	should	should	AUX
cril-239	120	11	be	be	AUX
cril-239	120	12	minimal	minimal	ADJ
cril-239	120	13	.	.	PUNCT
cril-239	121	1	6.2	6.2	NUM
cril-239	121	2	.	.	PUNCT
cril-239	122	1	naïve	naïve	ADJ
cril-239	122	2	bayes	bayes	PROPN
cril-239	122	3	according	accord	VERB
cril-239	122	4	to	to	ADP
cril-239	122	5	linguistic	linguistic	ADJ
cril-239	122	6	research	research	NOUN
cril-239	122	7	,	,	PUNCT
cril-239	122	8	adjectives	adjective	NOUN
cril-239	122	9	alone	alone	ADV
cril-239	122	10	are	be	AUX
cril-239	122	11	good	good	ADJ
cril-239	122	12	indicators	indicator	NOUN
cril-239	122	13	of	of	ADP
cril-239	122	14	subjective	subjective	ADJ
cril-239	122	15	expressions	expression	NOUN
cril-239	122	16	(	(	PUNCT
cril-239	122	17	wiebe	wiebe	PROPN
cril-239	122	18	,	,	PUNCT
cril-239	122	19	2000	2000	NUM
cril-239	122	20	)	)	PUNCT
cril-239	122	21	.	.	PUNCT
cril-239	123	1	therefore	therefore	ADV
cril-239	123	2	,	,	PUNCT
cril-239	123	3	determining	determine	VERB
cril-239	123	4	opinion	opinion	NOUN
cril-239	123	5	polarity	polarity	NOUN
cril-239	123	6	by	by	ADP
cril-239	123	7	analyzing	analyze	VERB
cril-239	123	8	occurrences	occurrence	NOUN
cril-239	123	9	of	of	ADP
cril-239	123	10	individual	individual	ADJ
cril-239	123	11	adjectives	adjective	NOUN
cril-239	123	12	in	in	ADP
cril-239	123	13	a	a	DET
cril-239	123	14	text	text	NOUN
cril-239	123	15	should	should	AUX
cril-239	123	16	be	be	AUX
cril-239	123	17	an	an	DET
cril-239	123	18	effective	effective	ADJ
cril-239	123	19	method	method	NOUN
cril-239	123	20	.	.	PUNCT
cril-239	124	1	to	to	PART
cril-239	124	2	identify	identify	VERB
cril-239	124	3	the	the	DET
cril-239	124	4	opinion	opinion	NOUN
cril-239	124	5	polarity	polarity	NOUN
cril-239	124	6	of	of	ADP
cril-239	124	7	movie	movie	NOUN
cril-239	124	8	reviews	review	NOUN
cril-239	124	9	,	,	PUNCT
cril-239	124	10	a	a	DET
cril-239	124	11	naïve	naïve	ADJ
cril-239	124	12	bayes	bayes	NOUN
cril-239	124	13	classifier	classifier	NOUN
cril-239	124	14	using	use	VERB
cril-239	124	15	adjectives	adjective	NOUN
cril-239	124	16	is	be	AUX
cril-239	124	17	a	a	DET
cril-239	124	18	promising	promising	ADJ
cril-239	124	19	model	model	NOUN
cril-239	124	20	.	.	PUNCT
cril-239	125	1	the	the	DET
cril-239	125	2	effectiveness	effectiveness	NOUN
cril-239	125	3	of	of	ADP
cril-239	125	4	adjectives	adjective	NOUN
cril-239	125	5	compared	compare	VERB
cril-239	125	6	to	to	ADP
cril-239	125	7	other	other	ADJ
cril-239	125	8	parts	part	NOUN
cril-239	125	9	-	-	PUNCT
cril-239	125	10	of	of	ADP
cril-239	125	11	-	-	PUNCT
cril-239	125	12	speech	speech	NOUN
cril-239	125	13	is	be	AUX
cril-239	125	14	evaluated	evaluate	VERB
cril-239	125	15	by	by	ADP
cril-239	125	16	applying	apply	VERB
cril-239	125	17	and	and	CCONJ
cril-239	125	18	comparing	compare	VERB
cril-239	125	19	the	the	DET
cril-239	125	20	results	result	NOUN
cril-239	125	21	on	on	ADP
cril-239	125	22	data	datum	NOUN
cril-239	125	23	with	with	ADP
cril-239	125	24	only	only	ADJ
cril-239	125	25	adjectives	adjective	NOUN
cril-239	125	26	against	against	ADP
cril-239	125	27	data	datum	NOUN
cril-239	125	28	with	with	ADP
cril-239	125	29	all	all	DET
cril-239	125	30	parts	part	NOUN
cril-239	125	31	-	-	PUNCT
cril-239	125	32	of	of	ADP
cril-239	125	33	-	-	PUNCT
cril-239	125	34	speech	speech	NOUN
cril-239	125	35	.	.	PUNCT
cril-239	126	1	the	the	DET
cril-239	126	2	impact	impact	NOUN
cril-239	126	3	of	of	ADP
cril-239	126	4	at	at	ADP
cril-239	126	5	-	-	PUNCT
cril-239	126	6	level	level	NOUN
cril-239	126	7	generalization	generalization	NOUN
cril-239	126	8	from	from	ADP
cril-239	126	9	adjectives	adjective	NOUN
cril-239	126	10	to	to	ADP
cril-239	126	11	synsets	synset	NOUN
cril-239	126	12	(	(	PUNCT
cril-239	126	13	or	or	CCONJ
cril-239	126	14	"	"	PUNCT
cril-239	126	15	sets	set	NOUN
cril-239	126	16	of	of	ADP
cril-239	126	17	synonyms	synonym	NOUN
cril-239	126	18	"	"	PUNCT
cril-239	126	19	;	;	PUNCT
cril-239	126	20	see	see	VERB
cril-239	126	21	"	"	PUNCT
cril-239	126	22	wordnet	wordnet	NOUN
cril-239	126	23	filtering	filtering	NOUN
cril-239	126	24	"	"	PUNCT
cril-239	126	25	,	,	PUNCT
cril-239	126	26	below	below	ADV
cril-239	126	27	)	)	PUNCT
cril-239	126	28	is	be	AUX
cril-239	126	29	also	also	ADV
cril-239	126	30	measured	measure	VERB
cril-239	126	31	.	.	PUNCT
cril-239	127	1	the	the	DET
cril-239	127	2	naïve	naïve	ADJ
cril-239	127	3	bayes	bayes	PROPN
cril-239	127	4	classifier	classifier	NOUN
cril-239	127	5	described	describe	VERB
cril-239	127	6	above	above	ADV
cril-239	127	7	was	be	AUX
cril-239	127	8	applied	apply	VERB
cril-239	127	9	to	to	ADP
cril-239	127	10	:	:	PUNCT
cril-239	127	11	1	1	X
cril-239	127	12	.	.	PUNCT
cril-239	127	13	tagged	tag	VERB
cril-239	127	14	data	datum	NOUN
cril-239	127	15	2	2	NUM
cril-239	127	16	.	.	PUNCT
cril-239	128	1	data	datum	NOUN
cril-239	128	2	containing	contain	VERB
cril-239	128	3	only	only	ADV
cril-239	128	4	the	the	DET
cril-239	128	5	adjectives	adjective	NOUN
cril-239	128	6	3	3	X
cril-239	128	7	.	.	PUNCT
cril-239	128	8	data	datum	NOUN
cril-239	128	9	containing	contain	VERB
cril-239	128	10	only	only	ADV
cril-239	128	11	the	the	DET
cril-239	128	12	synsets	synset	NOUN
cril-239	128	13	of	of	ADP
cril-239	128	14	the	the	DET
cril-239	128	15	adjectives	adjective	NOUN
cril-239	128	16	the	the	DET
cril-239	128	17	adjectives	adjective	NOUN
cril-239	128	18	in	in	ADP
cril-239	128	19	3	3	NUM
cril-239	128	20	were	be	AUX
cril-239	128	21	generalized	generalize	VERB
cril-239	128	22	to	to	ADP
cril-239	128	23	at	at	ADP
cril-239	128	24	-	-	PUNCT
cril-239	128	25	level	level	NOUN
cril-239	128	26	synsets	synset	NOUN
cril-239	128	27	using	use	VERB
cril-239	128	28	a	a	DET
cril-239	128	29	combination	combination	NOUN
cril-239	128	30	of	of	ADP
cril-239	128	31	the	the	DET
cril-239	128	32	pos	pos	NOUN
cril-239	128	33	filter	filter	NOUN
cril-239	128	34	module	module	NOUN
cril-239	128	35	and	and	CCONJ
cril-239	128	36	the	the	DET
cril-239	128	37	generalization	generalization	NOUN
cril-239	128	38	filter	filter	NOUN
cril-239	128	39	module	module	NOUN
cril-239	128	40	.	.	PUNCT
cril-239	129	1	for	for	ADP
cril-239	129	2	each	each	DET
cril-239	129	3	training	training	NOUN
cril-239	129	4	data	datum	NOUN
cril-239	129	5	set	set	VERB
cril-239	129	6	,	,	PUNCT
cril-239	129	7	addone	addone	NOUN
cril-239	129	8	smoothing	smoothing	NOUN
cril-239	129	9	was	be	AUX
cril-239	129	10	applied	apply	VERB
cril-239	129	11	to	to	ADP
cril-239	129	12	the	the	DET
cril-239	129	13	naïve	naïve	ADJ
cril-239	129	14	bayes	bayes	PROPN
cril-239	129	15	classifier	classifier	NOUN
cril-239	129	16	.	.	PUNCT
cril-239	130	1	table	table	NOUN
cril-239	130	2	1	1	NUM
cril-239	130	3	shows	show	VERB
cril-239	130	4	the	the	DET
cril-239	130	5	resulting	result	VERB
cril-239	130	6	accuracies	accuracy	NOUN
cril-239	130	7	of	of	ADP
cril-239	130	8	each	each	DET
cril-239	130	9	data	datum	NOUN
cril-239	130	10	set	set	VERB
cril-239	130	11	type	type	NOUN
cril-239	130	12	and	and	CCONJ
cril-239	130	13	size	size	NOUN
cril-239	130	14	.	.	PUNCT
cril-239	131	1	7	7	NUM
cril-239	131	2	salvetti	salvetti	VERB
cril-239	131	3	et	et	PROPN
cril-239	131	4	al	al	PROPN
cril-239	131	5	.	.	PROPN
cril-239	131	6	:	:	PUNCT
cril-239	132	1	automatic	automatic	ADJ
cril-239	132	2	opinion	opinion	NOUN
cril-239	132	3	polarity	polarity	NOUN
cril-239	132	4	classification	classification	NOUN
cril-239	132	5	of	of	ADP
cril-239	132	6	movie	movie	NOUN
cril-239	132	7	reviews	review	NOUN
cril-239	132	8	published	publish	VERB
cril-239	132	9	by	by	ADP
cril-239	132	10	cu	cu	PROPN
cril-239	132	11	scholar	scholar	NOUN
cril-239	132	12	,	,	PUNCT
cril-239	132	13	2004	2004	NUM
cril-239	132	14	colorado	colorado	PROPN
cril-239	132	15	research	research	NOUN
cril-239	132	16	in	in	ADP
cril-239	132	17	linguistics	linguistic	NOUN
cril-239	132	18	,	,	PUNCT
cril-239	132	19	volume	volume	NOUN
cril-239	132	20	17	17	NUM
cril-239	132	21	(	(	PUNCT
cril-239	132	22	2004	2004	NUM
cril-239	132	23	)	)	PUNCT
cril-239	132	24	8	8	NUM
cril-239	132	25	table	table	NOUN
cril-239	132	26	1	1	NUM
cril-239	132	27	:	:	PUNCT
cril-239	132	28	accuracies	accuracy	NOUN
cril-239	132	29	of	of	ADP
cril-239	132	30	naïve	naïve	ADJ
cril-239	132	31	bayes	bayes	NOUN
cril-239	132	32	classifier	classifier	NOUN
cril-239	132	33	.	.	PUNCT
cril-239	133	1	jj	jj	PROPN
cril-239	133	2	means	mean	VERB
cril-239	133	3	"	"	PUNCT
cril-239	133	4	adjectives	adjective	NOUN
cril-239	133	5	only	only	ADV
cril-239	133	6	"	"	PUNCT
cril-239	133	7	,	,	PUNCT
cril-239	133	8	wn	wn	PROPN
cril-239	133	9	indicates	indicate	VERB
cril-239	133	10	synset	synset	ADJ
cril-239	133	11	mapping	mapping	NOUN
cril-239	133	12	using	use	VERB
cril-239	133	13	the	the	DET
cril-239	133	14	generalization	generalization	NOUN
cril-239	133	15	filter	filter	NOUN
cril-239	133	16	.	.	PUNCT
cril-239	134	1	the	the	DET
cril-239	134	2	results	result	NOUN
cril-239	134	3	indicate	indicate	VERB
cril-239	134	4	that	that	SCONJ
cril-239	134	5	at	at	ADP
cril-239	134	6	-	-	PUNCT
cril-239	134	7	level	level	NOUN
cril-239	134	8	generalization	generalization	NOUN
cril-239	134	9	of	of	ADP
cril-239	134	10	adjectives	adjective	NOUN
cril-239	134	11	is	be	AUX
cril-239	134	12	not	not	PART
cril-239	134	13	effective	effective	ADJ
cril-239	134	14	and	and	CCONJ
cril-239	134	15	that	that	SCONJ
cril-239	134	16	extracting	extract	VERB
cril-239	134	17	only	only	ADV
cril-239	134	18	adjectives	adjective	NOUN
cril-239	134	19	degrades	degrade	VERB
cril-239	134	20	the	the	DET
cril-239	134	21	classifier	classifier	NOUN
cril-239	134	22	.	.	PUNCT
cril-239	135	1	however	however	ADV
cril-239	135	2	,	,	PUNCT
cril-239	135	3	this	this	PRON
cril-239	135	4	does	do	AUX
cril-239	135	5	not	not	PART
cril-239	135	6	imply	imply	VERB
cril-239	135	7	that	that	SCONJ
cril-239	135	8	filtering	filtering	NOUN
cril-239	135	9	does	do	AUX
cril-239	135	10	not	not	PART
cril-239	135	11	work	work	VERB
cril-239	135	12	.	.	PUNCT
cril-239	136	1	adjectives	adjective	NOUN
cril-239	136	2	constitute	constitute	VERB
cril-239	136	3	7.5	7.5	NUM
cril-239	136	4	%	%	NOUN
cril-239	136	5	of	of	ADP
cril-239	136	6	the	the	DET
cril-239	136	7	text	text	NOUN
cril-239	136	8	in	in	ADP
cril-239	136	9	the	the	DET
cril-239	136	10	data	datum	NOUN
cril-239	136	11	.	.	PUNCT
cril-239	137	1	the	the	DET
cril-239	137	2	accuracy	accuracy	NOUN
cril-239	137	3	achieved	achieve	VERB
cril-239	137	4	on	on	ADP
cril-239	137	5	such	such	DET
cril-239	137	6	a	a	DET
cril-239	137	7	small	small	ADJ
cril-239	137	8	portion	portion	NOUN
cril-239	137	9	of	of	ADP
cril-239	137	10	the	the	DET
cril-239	137	11	data	datum	NOUN
cril-239	137	12	indicates	indicate	VERB
cril-239	137	13	that	that	SCONJ
cril-239	137	14	a	a	DET
cril-239	137	15	significant	significant	ADJ
cril-239	137	16	portion	portion	NOUN
cril-239	137	17	of	of	ADP
cril-239	137	18	the	the	DET
cril-239	137	19	opinion	opinion	NOUN
cril-239	137	20	polarity	polarity	NOUN
cril-239	137	21	information	information	NOUN
cril-239	137	22	is	be	AUX
cril-239	137	23	carried	carry	VERB
cril-239	137	24	in	in	ADP
cril-239	137	25	the	the	DET
cril-239	137	26	adjectives	adjective	NOUN
cril-239	137	27	alone	alone	ADV
cril-239	137	28	.	.	PUNCT
cril-239	138	1	although	although	SCONJ
cril-239	138	2	the	the	DET
cril-239	138	3	resulting	result	VERB
cril-239	138	4	accuracies	accuracy	NOUN
cril-239	138	5	are	be	AUX
cril-239	138	6	better	well	ADJ
cril-239	138	7	in	in	ADP
cril-239	138	8	all	all	DET
cril-239	138	9	-	-	PUNCT
cril-239	138	10	pos	pos	NOUN
cril-239	138	11	data	datum	NOUN
cril-239	138	12	,	,	PUNCT
cril-239	138	13	adjectives	adjective	NOUN
cril-239	138	14	can	can	AUX
cril-239	138	15	still	still	ADV
cril-239	138	16	be	be	AUX
cril-239	138	17	considered	consider	VERB
cril-239	138	18	good	good	ADJ
cril-239	138	19	clues	clue	NOUN
cril-239	138	20	of	of	ADP
cril-239	138	21	opinion	opinion	NOUN
cril-239	138	22	polarity	polarity	NOUN
cril-239	138	23	.	.	PUNCT
cril-239	139	1	6.3	6.3	NUM
cril-239	139	2	.	.	PUNCT
cril-239	140	1	markov	markov	NOUN
cril-239	140	2	model	model	NOUN
cril-239	140	3	three	three	NUM
cril-239	140	4	types	type	NOUN
cril-239	140	5	of	of	ADP
cril-239	140	6	data	datum	NOUN
cril-239	140	7	are	be	AUX
cril-239	140	8	applied	apply	VERB
cril-239	140	9	to	to	ADP
cril-239	140	10	the	the	DET
cril-239	140	11	markov	markov	NOUN
cril-239	140	12	model	model	NOUN
cril-239	140	13	classifiers	classifier	NOUN
cril-239	140	14	described	describe	VERB
cril-239	140	15	previously	previously	ADV
cril-239	140	16	:	:	PUNCT
cril-239	140	17	1	1	X
cril-239	140	18	.	.	X
cril-239	140	19	tagged	tag	VERB
cril-239	140	20	data	datum	NOUN
cril-239	140	21	without	without	ADP
cril-239	140	22	any	any	DET
cril-239	140	23	filtering	filtering	NOUN
cril-239	140	24	,	,	PUNCT
cril-239	140	25	2	2	NUM
cril-239	140	26	.	.	PUNCT
cril-239	140	27	tagged	tag	VERB
cril-239	140	28	data	datum	NOUN
cril-239	140	29	with	with	ADP
cril-239	140	30	pos	pos	NOUN
cril-239	140	31	filters	filter	NOUN
cril-239	140	32	,	,	PUNCT
cril-239	140	33	3	3	X
cril-239	140	34	.	.	PUNCT
cril-239	140	35	tagged	tag	VERB
cril-239	140	36	data	datum	NOUN
cril-239	140	37	with	with	ADP
cril-239	140	38	both	both	DET
cril-239	140	39	pos	pos	NOUN
cril-239	140	40	filters	filter	NOUN
cril-239	140	41	and	and	CCONJ
cril-239	140	42	generalization	generalization	NOUN
cril-239	140	43	filters	filter	NOUN
cril-239	140	44	.	.	PUNCT
cril-239	141	1	witten	witten	PROPN
cril-239	141	2	-	-	PUNCT
cril-239	141	3	bell	bell	NOUN
cril-239	141	4	smoothing	smoothing	NOUN
cril-239	141	5	is	be	AUX
cril-239	141	6	applied	apply	VERB
cril-239	141	7	to	to	ADP
cril-239	141	8	this	this	DET
cril-239	141	9	classifier	classifier	NOUN
cril-239	141	10	.	.	PUNCT
cril-239	142	1	6.3.1	6.3.1	NUM
cril-239	142	2	.	.	PUNCT
cril-239	143	1	pos	pos	NOUN
cril-239	143	2	filtering	filter	VERB
cril-239	143	3	one	one	NUM
cril-239	143	4	design	design	NOUN
cril-239	143	5	principle	principle	NOUN
cril-239	143	6	of	of	ADP
cril-239	143	7	the	the	DET
cril-239	143	8	filter	filter	NOUN
cril-239	143	9	rules	rule	NOUN
cril-239	143	10	is	be	AUX
cril-239	143	11	that	that	SCONJ
cril-239	143	12	they	they	PRON
cril-239	143	13	filter	filter	VERB
cril-239	143	14	out	out	ADP
cril-239	143	15	parts	part	NOUN
cril-239	143	16	of	of	ADP
cril-239	143	17	speech	speech	NOUN
cril-239	143	18	that	that	PRON
cril-239	143	19	do	do	AUX
cril-239	143	20	not	not	PART
cril-239	143	21	contribute	contribute	VERB
cril-239	143	22	to	to	ADP
cril-239	143	23	the	the	DET
cril-239	143	24	opinion	opinion	NOUN
cril-239	143	25	polarity	polarity	NOUN
cril-239	143	26	and	and	CCONJ
cril-239	143	27	keep	keep	VERB
cril-239	143	28	the	the	DET
cril-239	143	29	parts	part	NOUN
cril-239	143	30	of	of	ADP
cril-239	143	31	speech	speech	NOUN
cril-239	143	32	that	that	PRON
cril-239	143	33	do	do	AUX
cril-239	143	34	contribute	contribute	VERB
cril-239	143	35	such	such	ADJ
cril-239	143	36	meaning	meaning	NOUN
cril-239	143	37	.	.	PUNCT
cril-239	144	1	based	base	VERB
cril-239	144	2	on	on	ADP
cril-239	144	3	analysis	analysis	NOUN
cril-239	144	4	of	of	ADP
cril-239	144	5	movie	movie	NOUN
cril-239	144	6	review	review	NOUN
cril-239	144	7	texts	text	NOUN
cril-239	144	8	,	,	PUNCT
cril-239	144	9	we	we	PRON
cril-239	144	10	devised	devise	VERB
cril-239	144	11	"	"	PUNCT
cril-239	144	12	filter	filter	NOUN
cril-239	144	13	rules	rule	NOUN
cril-239	144	14	"	"	PUNCT
cril-239	144	15	that	that	PRON
cril-239	144	16	take	take	VERB
cril-239	144	17	brill	brill	NOUN
cril-239	144	18	-	-	PUNCT
cril-239	144	19	tagged	tag	VERB
cril-239	144	20	text	text	NOUN
cril-239	144	21	as	as	ADP
cril-239	144	22	input	input	NOUN
cril-239	144	23	and	and	CCONJ
cril-239	144	24	return	return	VERB
cril-239	144	25	less	less	ADV
cril-239	144	26	noisy	noisy	ADJ
cril-239	144	27	,	,	PUNCT
cril-239	144	28	more	more	ADV
cril-239	144	29	concentrated	concentrated	ADJ
cril-239	144	30	sentences	sentence	NOUN
cril-239	144	31	that	that	PRON
cril-239	144	32	have	have	VERB
cril-239	144	33	a	a	DET
cril-239	144	34	combination	combination	NOUN
cril-239	144	35	of	of	ADP
cril-239	144	36	words	word	NOUN
cril-239	144	37	and	and	CCONJ
cril-239	144	38	word	word	NOUN
cril-239	144	39	/	/	SYM
cril-239	144	40	pos	pos	NOUN
cril-239	144	41	tag	tag	NOUN
cril-239	144	42	pairs	pair	NOUN
cril-239	144	43	removed	remove	VERB
cril-239	144	44	from	from	ADP
cril-239	144	45	the	the	DET
cril-239	144	46	original	original	NOUN
cril-239	144	47	.	.	PUNCT
cril-239	145	1	a	a	DET
cril-239	145	2	summary	summary	NOUN
cril-239	145	3	of	of	ADP
cril-239	145	4	the	the	DET
cril-239	145	5	filter	filter	NOUN
cril-239	145	6	rules	rule	NOUN
cril-239	145	7	defined	define	VERB
cril-239	145	8	in	in	ADP
cril-239	145	9	this	this	DET
cril-239	145	10	experiment	experiment	NOUN
cril-239	145	11	is	be	AUX
cril-239	145	12	shown	show	VERB
cril-239	145	13	in	in	ADP
cril-239	145	14	table	table	NOUN
cril-239	145	15	2	2	NUM
cril-239	145	16	.	.	SYM
cril-239	145	17	8	8	NUM
cril-239	145	18	colorado	colorado	NOUN
cril-239	145	19	research	research	NOUN
cril-239	145	20	in	in	ADP
cril-239	145	21	linguistics	linguistic	NOUN
cril-239	145	22	,	,	PUNCT
cril-239	145	23	vol	vol	NOUN
cril-239	145	24	.	.	PROPN
cril-239	145	25	17	17	NUM
cril-239	146	1	[	[	X
cril-239	146	2	2004	2004	NUM
cril-239	146	3	]	]	X
cril-239	146	4	https://scholar.colorado.edu/cril/vol17/iss1/1	https://scholar.colorado.edu/cril/vol17/iss1/1	PROPN
cril-239	146	5	doi	doi	NOUN
cril-239	146	6	:	:	PUNCT
cril-239	146	7	https://doi.org/10.25810/atv1-v819	https://doi.org/10.25810/atv1-v819	VERB
cril-239	146	8	automatic	automatic	ADJ
cril-239	146	9	opinion	opinion	NOUN
cril-239	146	10	polarity	polarity	NOUN
cril-239	146	11	classification	classification	NOUN
cril-239	146	12	of	of	ADP
cril-239	146	13	movie	movie	NOUN
cril-239	146	14	reviews	review	NOUN
cril-239	146	15	9	9	NUM
cril-239	146	16	table	table	NOUN
cril-239	146	17	2	2	NUM
cril-239	146	18	:	:	PUNCT
cril-239	146	19	summary	summary	NOUN
cril-239	146	20	of	of	ADP
cril-239	146	21	pos	pos	NOUN
cril-239	146	22	filter	filter	NOUN
cril-239	146	23	rules	rule	VERB
cril-239	146	24	wiebe	wiebe	PROPN
cril-239	146	25	et	et	PROPN
cril-239	146	26	al	al	PROPN
cril-239	146	27	.	.	PROPN
cril-239	146	28	,	,	PUNCT
cril-239	146	29	as	as	ADV
cril-239	146	30	well	well	ADV
cril-239	146	31	as	as	ADP
cril-239	146	32	other	other	ADJ
cril-239	146	33	researchers	researcher	NOUN
cril-239	146	34	,	,	PUNCT
cril-239	146	35	showed	show	VERB
cril-239	146	36	that	that	SCONJ
cril-239	146	37	subjectivity	subjectivity	NOUN
cril-239	146	38	is	be	AUX
cril-239	146	39	especially	especially	ADV
cril-239	146	40	concentrated	concentrate	VERB
cril-239	146	41	in	in	ADP
cril-239	146	42	adjectives	adjective	NOUN
cril-239	146	43	(	(	PUNCT
cril-239	146	44	wiebe	wiebe	PROPN
cril-239	146	45	et	et	PROPN
cril-239	146	46	al	al	PROPN
cril-239	146	47	,	,	PUNCT
cril-239	146	48	1999	1999	NUM
cril-239	146	49	;	;	PUNCT
cril-239	146	50	hatzivassiloglou	hatzivassiloglou	PROPN
cril-239	146	51	,	,	PUNCT
cril-239	146	52	2000	2000	NUM
cril-239	146	53	;	;	PUNCT
cril-239	146	54	turney	turney	PROPN
cril-239	146	55	et	et	PROPN
cril-239	146	56	al	al	PROPN
cril-239	146	57	,	,	PUNCT
cril-239	146	58	2003	2003	NUM
cril-239	146	59	)	)	PUNCT
cril-239	146	60	.	.	PUNCT
cril-239	147	1	therefore	therefore	ADV
cril-239	147	2	,	,	PUNCT
cril-239	147	3	no	no	DET
cril-239	147	4	adjectives	adjective	NOUN
cril-239	147	5	or	or	CCONJ
cril-239	147	6	their	their	PRON
cril-239	147	7	tags	tag	NOUN
cril-239	147	8	were	be	AUX
cril-239	147	9	removed	remove	VERB
cril-239	147	10	,	,	PUNCT
cril-239	147	11	nor	nor	CCONJ
cril-239	147	12	were	be	AUX
cril-239	147	13	copula	copula	ADJ
cril-239	147	14	verbs	verb	NOUN
cril-239	147	15	or	or	CCONJ
cril-239	147	16	negative	negative	ADJ
cril-239	147	17	markers	marker	NOUN
cril-239	147	18	.	.	PUNCT
cril-239	148	1	however	however	ADV
cril-239	148	2	,	,	PUNCT
cril-239	148	3	noisy	noisy	ADJ
cril-239	148	4	information	information	NOUN
cril-239	148	5	such	such	ADJ
cril-239	148	6	as	as	ADP
cril-239	148	7	determiners	determiner	NOUN
cril-239	148	8	,	,	PUNCT
cril-239	148	9	foreign	foreign	ADJ
cril-239	148	10	words	word	NOUN
cril-239	148	11	,	,	PUNCT
cril-239	148	12	prepositions	preposition	NOUN
cril-239	148	13	,	,	PUNCT
cril-239	148	14	modal	modal	ADJ
cril-239	148	15	verbs	verb	NOUN
cril-239	148	16	,	,	PUNCT
cril-239	148	17	possessives	possessive	NOUN
cril-239	148	18	,	,	PUNCT
cril-239	148	19	particles	particle	NOUN
cril-239	148	20	,	,	PUNCT
cril-239	148	21	interjections	interjection	NOUN
cril-239	148	22	,	,	PUNCT
cril-239	148	23	etc	etc	X
cril-239	148	24	.	.	X
cril-239	148	25	were	be	AUX
cril-239	148	26	removed	remove	VERB
cril-239	148	27	from	from	ADP
cril-239	148	28	the	the	DET
cril-239	148	29	text	text	NOUN
cril-239	148	30	stream	stream	NOUN
cril-239	148	31	.	.	PUNCT
cril-239	149	1	other	other	ADJ
cril-239	149	2	parts	part	NOUN
cril-239	149	3	of	of	ADP
cril-239	149	4	speech	speech	NOUN
cril-239	149	5	,	,	PUNCT
cril-239	149	6	such	such	ADJ
cril-239	149	7	as	as	ADP
cril-239	149	8	nouns	noun	NOUN
cril-239	149	9	and	and	CCONJ
cril-239	149	10	verbs	verb	NOUN
cril-239	149	11	,	,	PUNCT
cril-239	149	12	were	be	AUX
cril-239	149	13	removed	remove	VERB
cril-239	149	14	but	but	CCONJ
cril-239	149	15	their	their	PRON
cril-239	149	16	pos	pos	NOUN
cril-239	149	17	-	-	PUNCT
cril-239	149	18	tags	tag	NOUN
cril-239	149	19	were	be	AUX
cril-239	149	20	retained	retain	VERB
cril-239	149	21	.	.	PUNCT
cril-239	150	1	the	the	DET
cril-239	150	2	output	output	NOUN
cril-239	150	3	returned	return	VERB
cril-239	150	4	from	from	ADP
cril-239	150	5	the	the	DET
cril-239	150	6	filter	filter	NOUN
cril-239	150	7	did	do	AUX
cril-239	150	8	not	not	PART
cril-239	150	9	keep	keep	VERB
cril-239	150	10	the	the	DET
cril-239	150	11	original	original	ADJ
cril-239	150	12	sentence	sentence	NOUN
cril-239	150	13	structure	structure	NOUN
cril-239	150	14	.	.	PUNCT
cril-239	151	1	the	the	DET
cril-239	151	2	concrete	concrete	ADJ
cril-239	151	3	pos	pos	NOUN
cril-239	151	4	filtering	filtering	NOUN
cril-239	151	5	rules	rule	NOUN
cril-239	151	6	applied	apply	VERB
cril-239	151	7	in	in	ADP
cril-239	151	8	this	this	DET
cril-239	151	9	experiment	experiment	NOUN
cril-239	151	10	are	be	AUX
cril-239	151	11	shown	show	VERB
cril-239	151	12	in	in	ADP
cril-239	151	13	table	table	NOUN
cril-239	151	14	2	2	NUM
cril-239	151	15	.	.	PUNCT
cril-239	152	1	the	the	DET
cril-239	152	2	following	follow	VERB
cril-239	152	3	is	be	AUX
cril-239	152	4	an	an	DET
cril-239	152	5	example	example	NOUN
cril-239	152	6	of	of	ADP
cril-239	152	7	the	the	DET
cril-239	152	8	sentence	sentence	NOUN
cril-239	152	9	preprocessing	preprocessing	NOUN
cril-239	152	10	:	:	PUNCT
cril-239	152	11	•	•	ADP
cril-239	152	12	all	all	DET
cril-239	152	13	steve	steve	PROPN
cril-239	152	14	martin	martin	PROPN
cril-239	152	15	fans	fan	NOUN
cril-239	152	16	should	should	AUX
cril-239	152	17	be	be	AUX
cril-239	152	18	impressed	impressed	ADJ
cril-239	152	19	with	with	ADP
cril-239	152	20	this	this	DET
cril-239	152	21	wonderful	wonderful	ADJ
cril-239	152	22	new	new	ADJ
cril-239	152	23	comedy	comedy	NOUN
cril-239	152	24	•	•	INTJ
cril-239	152	25	/nnp	/nnp	PUNCT
cril-239	152	26	/nnp	/nnp	PUNCT
cril-239	152	27	/nn	/nn	PUNCT
cril-239	152	28	be	be	AUX
cril-239	152	29	/	/	SYM
cril-239	152	30	cop	cop	NOUN
cril-239	152	31	/vbn	/vbn	PROPN
cril-239	152	32	wonderful	wonderful	ADJ
cril-239	152	33	/	/	SYM
cril-239	152	34	jj	jj	PROPN
cril-239	152	35	new	new	PROPN
cril-239	152	36	/	/	SYM
cril-239	152	37	jj	jj	PROPN
cril-239	152	38	/nn	/nn	PUNCT
cril-239	153	1	the	the	DET
cril-239	153	2	resulting	result	VERB
cril-239	153	3	accuracies	accuracy	NOUN
cril-239	153	4	on	on	ADP
cril-239	153	5	pos	pos	NOUN
cril-239	153	6	filter	filter	NOUN
cril-239	153	7	rules	rule	NOUN
cril-239	153	8	and	and	CCONJ
cril-239	153	9	different	different	ADJ
cril-239	153	10	sizes	size	NOUN
cril-239	153	11	of	of	ADP
cril-239	153	12	data	datum	NOUN
cril-239	153	13	sets	set	NOUN
cril-239	153	14	are	be	AUX
cril-239	153	15	listed	list	VERB
cril-239	153	16	in	in	ADP
cril-239	153	17	table	table	NOUN
cril-239	153	18	3	3	NUM
cril-239	153	19	.	.	NOUN
cril-239	153	20	9	9	NUM
cril-239	153	21	salvetti	salvetti	VERB
cril-239	153	22	et	et	PROPN
cril-239	153	23	al	al	PROPN
cril-239	153	24	.	.	PROPN
cril-239	153	25	:	:	PUNCT
cril-239	154	1	automatic	automatic	ADJ
cril-239	154	2	opinion	opinion	NOUN
cril-239	154	3	polarity	polarity	NOUN
cril-239	154	4	classification	classification	NOUN
cril-239	154	5	of	of	ADP
cril-239	154	6	movie	movie	NOUN
cril-239	154	7	reviews	review	NOUN
cril-239	154	8	published	publish	VERB
cril-239	154	9	by	by	ADP
cril-239	154	10	cu	cu	PROPN
cril-239	154	11	scholar	scholar	NOUN
cril-239	154	12	,	,	PUNCT
cril-239	154	13	2004	2004	NUM
cril-239	154	14	colorado	colorado	PROPN
cril-239	154	15	research	research	NOUN
cril-239	154	16	in	in	ADP
cril-239	154	17	linguistics	linguistic	NOUN
cril-239	154	18	,	,	PUNCT
cril-239	154	19	volume	volume	NOUN
cril-239	154	20	17	17	NUM
cril-239	154	21	(	(	PUNCT
cril-239	154	22	2004	2004	NUM
cril-239	154	23	)	)	PUNCT
cril-239	154	24	10	10	NUM
cril-239	154	25	table	table	NOUN
cril-239	154	26	3	3	NUM
cril-239	154	27	:	:	PUNCT
cril-239	154	28	accuracies	accuracy	NOUN
cril-239	154	29	on	on	ADP
cril-239	154	30	pos	pos	NOUN
cril-239	154	31	filtering	filter	VERB
cril-239	154	32	7	7	NUM
cril-239	154	33	.	.	PUNCT
cril-239	155	1	wordnet	wordnet	NOUN
cril-239	155	2	filtering	filtering	NOUN
cril-239	155	3	in	in	ADP
cril-239	155	4	non	non	ADJ
cril-239	155	5	-	-	ADJ
cril-239	155	6	technical	technical	ADJ
cril-239	155	7	written	write	VERB
cril-239	155	8	text	text	NOUN
cril-239	155	9	it	it	PRON
cril-239	155	10	is	be	AUX
cril-239	155	11	uncommon	uncommon	ADJ
cril-239	155	12	to	to	PART
cril-239	155	13	encounter	encounter	VERB
cril-239	155	14	repetitions	repetition	NOUN
cril-239	155	15	of	of	ADP
cril-239	155	16	identical	identical	ADJ
cril-239	155	17	words	word	NOUN
cril-239	155	18	;	;	PUNCT
cril-239	155	19	this	this	PRON
cril-239	155	20	is	be	AUX
cril-239	155	21	generally	generally	ADV
cril-239	155	22	considered	consider	VERB
cril-239	155	23	"	"	PUNCT
cril-239	155	24	bad	bad	ADJ
cril-239	155	25	style	style	NOUN
cril-239	155	26	"	"	PUNCT
cril-239	155	27	.	.	PUNCT
cril-239	156	1	as	as	ADP
cril-239	156	2	such	such	ADJ
cril-239	156	3	,	,	PUNCT
cril-239	156	4	many	many	ADJ
cril-239	156	5	authors	author	NOUN
cril-239	156	6	attempt	attempt	VERB
cril-239	156	7	to	to	PART
cril-239	156	8	use	use	VERB
cril-239	156	9	synonyms	synonym	NOUN
cril-239	156	10	for	for	ADP
cril-239	156	11	words	word	NOUN
cril-239	156	12	whose	whose	DET
cril-239	156	13	meanings	meaning	NOUN
cril-239	156	14	they	they	PRON
cril-239	156	15	need	need	VERB
cril-239	156	16	often	often	ADV
cril-239	156	17	,	,	PUNCT
cril-239	156	18	propositions	proposition	NOUN
cril-239	156	19	,	,	PUNCT
cril-239	156	20	or	or	CCONJ
cril-239	156	21	even	even	ADV
cril-239	156	22	generalizations	generalization	NOUN
cril-239	156	23	.	.	PUNCT
cril-239	157	1	we	we	PRON
cril-239	157	2	attempted	attempt	VERB
cril-239	157	3	to	to	PART
cril-239	157	4	address	address	VERB
cril-239	157	5	two	two	NUM
cril-239	157	6	of	of	ADP
cril-239	157	7	these	these	DET
cril-239	157	8	perceived	perceive	VERB
cril-239	157	9	issues	issue	NOUN
cril-239	157	10	by	by	ADP
cril-239	157	11	identifying	identify	VERB
cril-239	157	12	words	word	NOUN
cril-239	157	13	with	with	ADP
cril-239	157	14	a	a	DET
cril-239	157	15	set	set	NOUN
cril-239	157	16	of	of	ADP
cril-239	157	17	likely	likely	ADJ
cril-239	157	18	synonyms	synonym	NOUN
cril-239	157	19	,	,	PUNCT
cril-239	157	20	and	and	CCONJ
cril-239	157	21	by	by	ADP
cril-239	157	22	hypernymy	hypernymy	NOUN
cril-239	157	23	generalization	generalization	NOUN
cril-239	157	24	.	.	PUNCT
cril-239	158	1	for	for	ADP
cril-239	158	2	the	the	DET
cril-239	158	3	implementation	implementation	NOUN
cril-239	158	4	of	of	ADP
cril-239	158	5	these	these	DET
cril-239	158	6	techniques	technique	NOUN
cril-239	158	7	,	,	PUNCT
cril-239	158	8	we	we	PRON
cril-239	158	9	took	take	VERB
cril-239	158	10	advantage	advantage	NOUN
cril-239	158	11	of	of	ADP
cril-239	158	12	the	the	DET
cril-239	158	13	wordnet	wordnet	NOUN
cril-239	158	14	(	(	PUNCT
cril-239	158	15	fellbaum	fellbaum	NOUN
cril-239	158	16	,	,	PUNCT
cril-239	158	17	1998	1998	NUM
cril-239	158	18	)	)	PUNCT
cril-239	158	19	system	system	NOUN
cril-239	158	20	,	,	PUNCT
cril-239	158	21	which	which	PRON
cril-239	158	22	provides	provide	VERB
cril-239	158	23	the	the	DET
cril-239	158	24	former	former	ADJ
cril-239	158	25	by	by	ADP
cril-239	158	26	means	mean	NOUN
cril-239	158	27	of	of	ADP
cril-239	158	28	synsets	synset	NOUN
cril-239	158	29	for	for	ADP
cril-239	158	30	four	four	NUM
cril-239	158	31	separate	separate	ADJ
cril-239	158	32	classes	class	NOUN
cril-239	158	33	of	of	ADP
cril-239	158	34	words	word	NOUN
cril-239	158	35	(	(	PUNCT
cril-239	158	36	verbs	verb	NOUN
cril-239	158	37	,	,	PUNCT
cril-239	158	38	nouns	noun	NOUN
cril-239	158	39	,	,	PUNCT
cril-239	158	40	adjectives	adjective	NOUN
cril-239	158	41	and	and	CCONJ
cril-239	158	42	adverbs	adverb	NOUN
cril-239	158	43	)	)	PUNCT
cril-239	158	44	,	,	PUNCT
cril-239	158	45	and	and	CCONJ
cril-239	158	46	the	the	DET
cril-239	158	47	latter	latter	ADJ
cril-239	158	48	through	through	ADP
cril-239	158	49	hypernymy	hypernymy	PROPN
cril-239	158	50	relations	relation	NOUN
cril-239	158	51	between	between	ADP
cril-239	158	52	synsets	synset	NOUN
cril-239	158	53	of	of	ADP
cril-239	158	54	the	the	DET
cril-239	158	55	same	same	ADJ
cril-239	158	56	class	class	NOUN
cril-239	158	57	.	.	PUNCT
cril-239	159	1	7.1	7.1	NUM
cril-239	159	2	.	.	PUNCT
cril-239	160	1	synonyms	synonyms	PROPN
cril-239	160	2	wordnet	wordnet	NOUN
cril-239	160	3	maps	map	VERB
cril-239	160	4	each	each	PRON
cril-239	160	5	of	of	ADP
cril-239	160	6	the	the	DET
cril-239	160	7	words	word	NOUN
cril-239	160	8	it	it	PRON
cril-239	160	9	supports	support	VERB
cril-239	160	10	into	into	ADP
cril-239	160	11	a	a	DET
cril-239	160	12	synset	synset	NOUN
cril-239	160	13	,	,	PUNCT
cril-239	160	14	which	which	PRON
cril-239	160	15	is	be	AUX
cril-239	160	16	an	an	DET
cril-239	160	17	abstract	abstract	ADJ
cril-239	160	18	entity	entity	NOUN
cril-239	160	19	encompassing	encompass	VERB
cril-239	160	20	all	all	DET
cril-239	160	21	words	word	NOUN
cril-239	160	22	with	with	ADP
cril-239	160	23	a	a	DET
cril-239	160	24	"	"	PUNCT
cril-239	160	25	reasonably	reasonably	ADJ
cril-239	160	26	"	"	PUNCT
cril-239	160	27	similar	similar	ADJ
cril-239	160	28	meaning	meaning	NOUN
cril-239	160	29	.	.	PUNCT
cril-239	161	1	in	in	ADP
cril-239	161	2	the	the	DET
cril-239	161	3	case	case	NOUN
cril-239	161	4	of	of	ADP
cril-239	161	5	ambiguous	ambiguous	ADJ
cril-239	161	6	words	word	NOUN
cril-239	161	7	,	,	PUNCT
cril-239	161	8	multiple	multiple	ADJ
cril-239	161	9	synsets	synset	NOUN
cril-239	161	10	may	may	AUX
cril-239	161	11	exist	exist	VERB
cril-239	161	12	for	for	ADP
cril-239	161	13	a	a	DET
cril-239	161	14	word	word	NOUN
cril-239	161	15	;	;	PUNCT
cril-239	161	16	in	in	ADP
cril-239	161	17	these	these	DET
cril-239	161	18	instances	instance	NOUN
cril-239	161	19	,	,	PUNCT
cril-239	161	20	we	we	PRON
cril-239	161	21	picked	pick	VERB
cril-239	161	22	the	the	DET
cril-239	161	23	first	first	ADJ
cril-239	161	24	one	one	NUM
cril-239	161	25	.	.	PUNCT
cril-239	162	1	note	note	VERB
cril-239	162	2	that	that	SCONJ
cril-239	162	3	synonyms	synonym	NOUN
cril-239	162	4	(	(	PUNCT
cril-239	162	5	and	and	CCONJ
cril-239	162	6	general	general	ADJ
cril-239	162	7	wordnet	wordnet	NOUN
cril-239	162	8	processing	processing	NOUN
cril-239	162	9	)	)	PUNCT
cril-239	162	10	are	be	AUX
cril-239	162	11	only	only	ADV
cril-239	162	12	available	available	ADJ
cril-239	162	13	in	in	ADP
cril-239	162	14	instances	instance	NOUN
cril-239	162	15	where	where	SCONJ
cril-239	162	16	the	the	DET
cril-239	162	17	word	word	NOUN
cril-239	162	18	under	under	ADP
cril-239	162	19	consideration	consideration	NOUN
cril-239	162	20	falls	fall	VERB
cril-239	162	21	in	in	ADP
cril-239	162	22	one	one	NUM
cril-239	162	23	of	of	ADP
cril-239	162	24	the	the	DET
cril-239	162	25	four	four	NUM
cril-239	162	26	classes	class	NOUN
cril-239	162	27	of	of	ADP
cril-239	162	28	words	word	NOUN
cril-239	162	29	we	we	PRON
cril-239	162	30	outlined	outline	VERB
cril-239	162	31	above	above	ADV
cril-239	162	32	.	.	PUNCT
cril-239	163	1	we	we	PRON
cril-239	163	2	determined	determine	VERB
cril-239	163	3	the	the	DET
cril-239	163	4	appropriate	appropriate	ADJ
cril-239	163	5	category	category	NOUN
cril-239	163	6	for	for	ADP
cril-239	163	7	each	each	DET
cril-239	163	8	word	word	NOUN
cril-239	163	9	by	by	ADP
cril-239	163	10	examining	examine	VERB
cril-239	163	11	the	the	DET
cril-239	163	12	tag	tag	NOUN
cril-239	163	13	it	it	PRON
cril-239	163	14	was	be	AUX
cril-239	163	15	assigned	assign	VERB
cril-239	163	16	by	by	ADP
cril-239	163	17	the	the	DET
cril-239	163	18	brill	brill	PROPN
cril-239	163	19	tagger	tagger	NOUN
cril-239	163	20	,	,	PUNCT
cril-239	163	21	not	not	PART
cril-239	163	22	touching	touch	VERB
cril-239	163	23	words	word	NOUN
cril-239	163	24	which	which	PRON
cril-239	163	25	fell	fall	VERB
cril-239	163	26	outside	outside	ADV
cril-239	163	27	of	of	ADP
cril-239	163	28	these	these	DET
cril-239	163	29	classes	class	NOUN
cril-239	163	30	.	.	PUNCT
cril-239	164	1	7.2	7.2	NUM
cril-239	164	2	.	.	PUNCT
cril-239	164	3	hypernyms	hypernym	NOUN
cril-239	164	4	for	for	ADP
cril-239	164	5	verbs	verb	NOUN
cril-239	164	6	and	and	CCONJ
cril-239	164	7	nouns	noun	NOUN
cril-239	164	8	,	,	PUNCT
cril-239	164	9	wordnet	wordnet	NOUN
cril-239	164	10	provides	provide	VERB
cril-239	164	11	a	a	DET
cril-239	164	12	hypernymy	hypernymy	NOUN
cril-239	164	13	relation	relation	NOUN
cril-239	164	14	,	,	PUNCT
cril-239	164	15	which	which	PRON
cril-239	164	16	can	can	AUX
cril-239	164	17	be	be	AUX
cril-239	164	18	informally	informally	ADV
cril-239	164	19	described	describe	VERB
cril-239	164	20	as	as	SCONJ
cril-239	164	21	follows	follow	VERB
cril-239	164	22	:	:	PUNCT
cril-239	164	23	let	let	VERB
cril-239	164	24	s1	s1	NOUN
cril-239	164	25	,	,	PUNCT
cril-239	164	26	s2	s2	NOUN
cril-239	164	27	be	be	VERB
cril-239	164	28	synsets	synset	NOUN
cril-239	164	29	.	.	PUNCT
cril-239	165	1	then	then	ADV
cril-239	165	2	s1	s1	PROPN
cril-239	165	3	is	be	AUX
cril-239	165	4	hypernym	hypernym	NOUN
cril-239	165	5	of	of	ADP
cril-239	165	6	s2	s2	PROPN
cril-239	165	7	,	,	PUNCT
cril-239	165	8	notation	notation	NOUN
cril-239	165	9	s1	s1	PROPN
cril-239	165	10	�	�	PROPN
cril-239	165	11	s2	s2	PROPN
cril-239	165	12	,	,	PUNCT
cril-239	165	13	if	if	SCONJ
cril-239	165	14	and	and	CCONJ
cril-239	165	15	only	only	ADV
cril-239	165	16	if	if	SCONJ
cril-239	165	17	anything	anything	PRON
cril-239	165	18	that	that	PRON
cril-239	165	19	can	can	AUX
cril-239	165	20	be	be	AUX
cril-239	165	21	described	describe	VERB
cril-239	165	22	by	by	ADP
cril-239	165	23	a	a	DET
cril-239	165	24	word	word	NOUN
cril-239	165	25	in	in	ADP
cril-239	165	26	s2	s2	NOUN
cril-239	165	27	can	can	AUX
cril-239	165	28	also	also	ADV
cril-239	165	29	be	be	AUX
cril-239	165	30	described	describe	VERB
cril-239	165	31	10	10	NUM
cril-239	165	32	colorado	colorado	NOUN
cril-239	165	33	research	research	NOUN
cril-239	165	34	in	in	ADP
cril-239	165	35	linguistics	linguistic	NOUN
cril-239	165	36	,	,	PUNCT
cril-239	165	37	vol	vol	NOUN
cril-239	165	38	.	.	PROPN
cril-239	165	39	17	17	NUM
cril-239	166	1	[	[	X
cril-239	166	2	2004	2004	NUM
cril-239	166	3	]	]	X
cril-239	166	4	https://scholar.colorado.edu/cril/vol17/iss1/1	https://scholar.colorado.edu/cril/vol17/iss1/1	PROPN
cril-239	166	5	doi	doi	NOUN
cril-239	166	6	:	:	PUNCT
cril-239	166	7	https://doi.org/10.25810/atv1-v819	https://doi.org/10.25810/atv1-v819	VERB
cril-239	166	8	automatic	automatic	ADJ
cril-239	166	9	opinion	opinion	NOUN
cril-239	166	10	polarity	polarity	NOUN
cril-239	166	11	classification	classification	NOUN
cril-239	166	12	of	of	ADP
cril-239	166	13	movie	movie	NOUN
cril-239	166	14	reviews	review	NOUN
cril-239	166	15	11	11	NUM
cril-239	166	16	by	by	ADP
cril-239	166	17	a	a	DET
cril-239	166	18	word	word	NOUN
cril-239	166	19	in	in	ADP
cril-239	166	20	s1	s1	NOUN
cril-239	166	21	,	,	PUNCT
cril-239	166	22	and	and	CCONJ
cril-239	166	23	s1	s1	PROPN
cril-239	166	24	≠	≠	PROPN
cril-239	166	25	s2	s2	PROPN
cril-239	166	26	.	.	PUNCT
cril-239	167	1	for	for	ADP
cril-239	167	2	each	each	PRON
cril-239	167	3	of	of	ADP
cril-239	167	4	the	the	DET
cril-239	167	5	hypernym	hypernym	NOUN
cril-239	167	6	categories	category	NOUN
cril-239	167	7	,	,	PUNCT
cril-239	167	8	we	we	PRON
cril-239	167	9	determine	determine	VERB
cril-239	167	10	a	a	DET
cril-239	167	11	set	set	NOUN
cril-239	167	12	of	of	ADP
cril-239	167	13	abstract	abstract	ADJ
cril-239	167	14	synsets	synset	NOUN
cril-239	167	15	a	a	DET
cril-239	167	16	such	such	ADJ
cril-239	167	17	that	that	PRON
cril-239	167	18	,	,	PUNCT
cril-239	167	19	for	for	ADP
cril-239	167	20	any	any	DET
cril-239	167	21	a	a	DET
cril-239	167	22	∈	∈	PROPN
cril-239	167	23	a	a	PRON
cril-239	167	24	,	,	PUNCT
cril-239	167	25	there	there	PRON
cril-239	167	26	does	do	AUX
cril-239	167	27	not	not	PART
cril-239	167	28	exist	exist	VERB
cril-239	167	29	any	any	DET
cril-239	167	30	s	s	NOUN
cril-239	167	31	such	such	ADJ
cril-239	167	32	that	that	DET
cril-239	167	33	s	s	PROPN
cril-239	167	34	�	�	PROPN
cril-239	167	35	a.	a.	NOUN
cril-239	167	36	we	we	PRON
cril-239	167	37	say	say	VERB
cril-239	167	38	that	that	SCONJ
cril-239	167	39	a	a	DET
cril-239	167	40	synset	synset	ADJ
cril-239	167	41	h	h	NOUN
cril-239	167	42	is	be	AUX
cril-239	167	43	a	a	DET
cril-239	167	44	level	level	NOUN
cril-239	167	45	n	n	DET
cril-239	167	46	hypernym	hypernym	NOUN
cril-239	167	47	of	of	ADP
cril-239	167	48	a	a	DET
cril-239	167	49	synset	synset	NOUN
cril-239	167	50	s	s	X
cril-239	167	51	if	if	SCONJ
cril-239	167	52	and	and	CCONJ
cril-239	167	53	only	only	ADV
cril-239	168	1	if	if	SCONJ
cril-239	168	2	h	h	PROPN
cril-239	168	3	�	�	PROPN
cril-239	168	4	*	*	PUNCT
cril-239	168	5	s	s	X
cril-239	168	6	and	and	CCONJ
cril-239	168	7	one	one	NUM
cril-239	168	8	of	of	ADP
cril-239	168	9	the	the	DET
cril-239	168	10	following	following	NOUN
cril-239	168	11	holds	hold	VERB
cril-239	168	12	for	for	ADP
cril-239	168	13	some	some	DET
cril-239	168	14	a	a	DET
cril-239	168	15	∈	∈	PROPN
cril-239	168	16	a	a	DET
cril-239	168	17	:	:	PUNCT
cril-239	168	18	1	1	NUM
cril-239	168	19	.	.	X
cril-239	168	20	a	a	DET
cril-239	168	21	�	�	PROPN
cril-239	168	22	n	n	CCONJ
cril-239	168	23	h	h	NOUN
cril-239	168	24	2	2	NUM
cril-239	168	25	.	.	PUNCT
cril-239	168	26	s	s	PART
cril-239	168	27	=	=	NOUN
cril-239	168	28	h	h	PROPN
cril-239	168	29	and	and	CCONJ
cril-239	168	30	a	a	DET
cril-239	168	31	�	�	PROPN
cril-239	168	32	l	l	PROPN
cril-239	168	33	s	s	PROPN
cril-239	168	34	,	,	PUNCT
cril-239	168	35	with	with	ADP
cril-239	168	36	l	l	X
cril-239	168	37	<	<	X
cril-239	168	38	n	n	X
cril-239	168	39	for	for	ADP
cril-239	168	40	example	example	NOUN
cril-239	168	41	,	,	PUNCT
cril-239	168	42	given	give	VERB
cril-239	168	43	the	the	DET
cril-239	168	44	wordnet	wordnet	NOUN
cril-239	168	45	database	database	NOUN
cril-239	168	46	,	,	PUNCT
cril-239	168	47	a	a	DET
cril-239	168	48	hypernym	hypernym	ADJ
cril-239	168	49	generalization	generalization	NOUN
cril-239	168	50	of	of	ADP
cril-239	168	51	level	level	NOUN
cril-239	168	52	4	4	NUM
cril-239	168	53	for	for	ADP
cril-239	168	54	the	the	DET
cril-239	168	55	nouns	noun	NOUN
cril-239	168	56	"	"	PUNCT
cril-239	168	57	movie	movie	NOUN
cril-239	168	58	"	"	PUNCT
cril-239	168	59	and	and	CCONJ
cril-239	168	60	"	"	PUNCT
cril-239	168	61	performance	performance	NOUN
cril-239	168	62	"	"	PUNCT
cril-239	168	63	will	will	AUX
cril-239	168	64	generalize	generalize	VERB
cril-239	168	65	both	both	PRON
cril-239	168	66	of	of	ADP
cril-239	168	67	them	they	PRON
cril-239	168	68	to	to	ADP
cril-239	168	69	one	one	NUM
cril-239	168	70	common	common	ADJ
cril-239	168	71	synset	synset	NOUN
cril-239	168	72	which	which	PRON
cril-239	168	73	can	can	AUX
cril-239	168	74	be	be	AUX
cril-239	168	75	characterized	characterize	VERB
cril-239	168	76	by	by	ADP
cril-239	168	77	the	the	DET
cril-239	168	78	word	word	NOUN
cril-239	168	79	"	"	PUNCT
cril-239	168	80	communication	communication	NOUN
cril-239	168	81	.	.	PUNCT
cril-239	168	82	"	"	PUNCT
cril-239	169	1	7.3	7.3	NUM
cril-239	169	2	.	.	PUNCT
cril-239	170	1	analysis	analysis	NOUN
cril-239	170	2	in	in	ADP
cril-239	170	3	order	order	NOUN
cril-239	170	4	to	to	PART
cril-239	170	5	determine	determine	VERB
cril-239	170	6	the	the	DET
cril-239	170	7	effects	effect	NOUN
cril-239	170	8	of	of	ADP
cril-239	170	9	translating	translate	VERB
cril-239	170	10	words	word	NOUN
cril-239	170	11	to	to	ADP
cril-239	170	12	synsets	synset	NOUN
cril-239	170	13	and	and	CCONJ
cril-239	170	14	performing	perform	VERB
cril-239	170	15	hypernymization	hypernymization	NOUN
cril-239	170	16	on	on	ADP
cril-239	170	17	them	they	PRON
cril-239	170	18	,	,	PUNCT
cril-239	170	19	we	we	PRON
cril-239	170	20	ran	run	VERB
cril-239	170	21	a	a	DET
cril-239	170	22	series	series	NOUN
cril-239	170	23	of	of	ADP
cril-239	170	24	tests	test	NOUN
cril-239	170	25	which	which	PRON
cril-239	170	26	quickly	quickly	ADV
cril-239	170	27	determined	determine	VERB
cril-239	170	28	that	that	SCONJ
cril-239	170	29	the	the	DET
cril-239	170	30	effects	effect	NOUN
cril-239	170	31	of	of	ADP
cril-239	170	32	pure	pure	ADJ
cril-239	170	33	synset	synset	ADJ
cril-239	170	34	translation	translation	NOUN
cril-239	170	35	were	be	AUX
cril-239	170	36	negligible	negligible	ADJ
cril-239	170	37	.	.	PUNCT
cril-239	171	1	we	we	PRON
cril-239	171	2	thus	thus	ADV
cril-239	171	3	experimented	experiment	VERB
cril-239	171	4	with	with	ADP
cril-239	171	5	the	the	DET
cril-239	171	6	computation	computation	NOUN
cril-239	171	7	of	of	ADP
cril-239	171	8	level	level	NOUN
cril-239	171	9	n	n	NOUN
cril-239	171	10	hypernyms	hypernyms	VERB
cril-239	171	11	with	with	ADP
cril-239	171	12	n	n	PRON
cril-239	171	13	∈	∈	PROPN
cril-239	171	14	{	{	PUNCT
cril-239	171	15	0	0	NUM
cril-239	171	16	.	.	PUNCT
cril-239	171	17	.	.	PUNCT
cril-239	171	18	.	.	PUNCT
cril-239	172	1	10	10	NUM
cril-239	172	2	}	}	PUNCT
cril-239	172	3	,	,	PUNCT
cril-239	172	4	separately	separately	ADV
cril-239	172	5	for	for	ADP
cril-239	172	6	nouns	noun	NOUN
cril-239	172	7	and	and	CCONJ
cril-239	172	8	verbs	verbs	PROPN
cril-239	172	9	.	.	PROPN
cril-239	172	10	figure	figure	NOUN
cril-239	172	11	2	2	NUM
cril-239	172	12	:	:	PUNCT
cril-239	172	13	hypernym	hypernym	NOUN
cril-239	172	14	generalization	generalization	NOUN
cril-239	172	15	with	with	ADP
cril-239	172	16	1500	1500	NUM
cril-239	172	17	reviews	review	NOUN
cril-239	172	18	from	from	ADP
cril-239	172	19	each	each	DET
cril-239	172	20	class	class	NOUN
cril-239	172	21	.	.	PUNCT
cril-239	173	1	the	the	DET
cril-239	173	2	x	x	NOUN
cril-239	173	3	and	and	CCONJ
cril-239	173	4	y	y	PROPN
cril-239	173	5	axis	axis	NOUN
cril-239	173	6	describe	describe	VERB
cril-239	173	7	the	the	DET
cril-239	173	8	level	level	NOUN
cril-239	173	9	of	of	ADP
cril-239	173	10	hypernym	hypernym	NOUN
cril-239	173	11	generalization	generalization	NOUN
cril-239	173	12	for	for	ADP
cril-239	173	13	nouns	noun	NOUN
cril-239	173	14	and	and	CCONJ
cril-239	173	15	verbs	verb	NOUN
cril-239	173	16	,	,	PUNCT
cril-239	173	17	z	z	NOUN
cril-239	173	18	the	the	DET
cril-239	173	19	accuracy	accuracy	NOUN
cril-239	173	20	we	we	PRON
cril-239	173	21	achieved	achieve	VERB
cril-239	173	22	.	.	PUNCT
cril-239	174	1	maximum	maximum	ADJ
cril-239	174	2	concreteness	concreteness	NOUN
cril-239	174	3	at	at	ADP
cril-239	174	4	level	level	NOUN
cril-239	174	5	10	10	NUM
cril-239	174	6	indicates	indicate	VERB
cril-239	174	7	no	no	DET
cril-239	174	8	generalization	generalization	NOUN
cril-239	174	9	.	.	PUNCT
cril-239	175	1	as	as	SCONJ
cril-239	175	2	we	we	PRON
cril-239	175	3	can	can	AUX
cril-239	175	4	see	see	VERB
cril-239	175	5	from	from	ADP
cril-239	175	6	figure	figure	NOUN
cril-239	175	7	2	2	NUM
cril-239	175	8	,	,	PUNCT
cril-239	175	9	applying	apply	VERB
cril-239	175	10	hypernym	hypernym	ADJ
cril-239	175	11	generalization	generalization	NOUN
cril-239	175	12	to	to	ADP
cril-239	175	13	information	information	NOUN
cril-239	175	14	gathered	gather	VERB
cril-239	175	15	from	from	ADP
cril-239	175	16	large	large	ADJ
cril-239	175	17	data	datum	NOUN
cril-239	175	18	sets	set	NOUN
cril-239	175	19	yielded	yield	VERB
cril-239	175	20	little	little	ADJ
cril-239	175	21	improvement	improvement	NOUN
cril-239	175	22	;	;	PUNCT
cril-239	175	23	instead	instead	ADV
cril-239	175	24	,	,	PUNCT
cril-239	175	25	we	we	PRON
cril-239	175	26	observed	observe	VERB
cril-239	175	27	a	a	DET
cril-239	175	28	degradation	degradation	NOUN
cril-239	175	29	in	in	ADP
cril-239	175	30	the	the	DET
cril-239	175	31	quality	quality	NOUN
cril-239	175	32	of	of	ADP
cril-239	175	33	our	our	PRON
cril-239	175	34	classification	classification	NOUN
cril-239	175	35	caused	cause	VERB
cril-239	175	36	by	by	ADP
cril-239	175	37	the	the	DET
cril-239	175	38	loss	loss	NOUN
cril-239	175	39	of	of	ADP
cril-239	175	40	information	information	NOUN
cril-239	175	41	.	.	PUNCT
cril-239	176	1	we	we	PRON
cril-239	176	2	assume	assume	VERB
cril-239	176	3	that	that	SCONJ
cril-239	176	4	for	for	SCONJ
cril-239	176	5	larger	large	ADJ
cril-239	176	6	data	datum	NOUN
cril-239	176	7	sets	set	VERB
cril-239	176	8	bigram	bigram	PROPN
cril-239	176	9	classification	classification	NOUN
cril-239	176	10	is	be	AUX
cril-239	176	11	already	already	ADV
cril-239	176	12	able	able	ADJ
cril-239	176	13	to	to	PART
cril-239	176	14	make	make	VERB
cril-239	176	15	use	use	NOUN
cril-239	176	16	of	of	ADP
cril-239	176	17	the	the	DET
cril-239	176	18	more	more	ADV
cril-239	176	19	fine	fine	ADV
cril-239	176	20	-	-	PUNCT
cril-239	176	21	grained	grain	VERB
cril-239	176	22	data	datum	NOUN
cril-239	176	23	present	present	ADJ
cril-239	176	24	.	.	PUNCT
cril-239	177	1	shrinking	shrink	VERB
cril-239	177	2	the	the	DET
cril-239	177	3	size	size	NOUN
cril-239	177	4	of	of	ADP
cril-239	177	5	our	our	PRON
cril-239	177	6	training	training	NOUN
cril-239	177	7	data	datum	NOUN
cril-239	177	8	,	,	PUNCT
cril-239	177	9	however	however	ADV
cril-239	177	10	,	,	PUNCT
cril-239	177	11	increased	increase	VERB
cril-239	177	12	the	the	DET
cril-239	177	13	impact	impact	NOUN
cril-239	177	14	of	of	ADP
cril-239	177	15	wordnet	wordnet	NOUN
cril-239	177	16	simplification	simplification	NOUN
cril-239	177	17	;	;	PUNCT
cril-239	177	18	for	for	ADP
cril-239	177	19	very	very	ADV
cril-239	177	20	small	small	ADJ
cril-239	177	21	data	data	NOUN
cril-239	177	22	sets	set	NOUN
cril-239	177	23	(	(	PUNCT
cril-239	177	24	50	50	NUM
cril-239	177	25	reviews	review	NOUN
cril-239	177	26	and	and	CCONJ
cril-239	177	27	less	less	ADJ
cril-239	177	28	,	,	PUNCT
cril-239	177	29	not	not	PART
cril-239	177	30	shown	show	VERB
cril-239	177	31	here	here	ADV
cril-239	177	32	)	)	PUNCT
cril-239	177	33	we	we	PRON
cril-239	177	34	observed	observe	VERB
cril-239	177	35	an	an	DET
cril-239	177	36	improvement	improvement	NOUN
cril-239	177	37	of	of	ADP
cril-239	177	38	2.5	2.5	NUM
cril-239	177	39	%	%	NOUN
cril-239	177	40	(	(	PUNCT
cril-239	177	41	absolute	absolute	ADJ
cril-239	177	42	)	)	PUNCT
cril-239	177	43	in	in	ADP
cril-239	177	44	comparison	comparison	NOUN
cril-239	177	45	to	to	ADP
cril-239	177	46	both	both	DET
cril-239	177	47	full	full	ADJ
cril-239	177	48	generalization	generalization	NOUN
cril-239	177	49	and	and	CCONJ
cril-239	177	50	no	no	DET
cril-239	177	51	generalization	generalization	NOUN
cril-239	177	52	at	at	ADV
cril-239	177	53	all	all	ADV
cril-239	177	54	.	.	PUNCT
cril-239	178	1	increasing	increase	VERB
cril-239	178	2	the	the	DET
cril-239	178	3	size	size	NOUN
cril-239	178	4	of	of	ADP
cril-239	178	5	the	the	DET
cril-239	178	6	set	set	NOUN
cril-239	178	7	of	of	ADP
cril-239	178	8	observable	observable	ADJ
cril-239	178	9	events	event	NOUN
cril-239	178	10	by	by	ADP
cril-239	178	11	using	use	VERB
cril-239	178	12	trigram	trigram	NOUN
cril-239	178	13	models	model	NOUN
cril-239	178	14	resulted	result	VERB
cril-239	178	15	in	in	ADP
cril-239	178	16	a	a	DET
cril-239	178	17	small	small	ADJ
cril-239	178	18	gain	gain	NOUN
cril-239	178	19	(	(	PUNCT
cril-239	178	20	around	around	ADV
cril-239	178	21	1	1	NUM
cril-239	178	22	%	%	NOUN
cril-239	178	23	)	)	PUNCT
cril-239	178	24	.	.	PUNCT
cril-239	179	1	interestingly	interestingly	ADV
cril-239	179	2	,	,	PUNCT
cril-239	179	3	the	the	DET
cril-239	179	4	effect	effect	NOUN
cril-239	179	5	of	of	ADP
cril-239	179	6	verb	verb	NOUN
cril-239	179	7	generalization	generalization	NOUN
cril-239	179	8	was	be	AUX
cril-239	179	9	relatively	relatively	ADV
cril-239	179	10	small	small	ADJ
cril-239	179	11	in	in	ADP
cril-239	179	12	comparison	comparison	NOUN
cril-239	179	13	to	to	ADP
cril-239	179	14	noun	noun	NOUN
cril-239	179	15	generalization	generalization	NOUN
cril-239	179	16	for	for	ADP
cril-239	179	17	similar	similar	ADJ
cril-239	179	18	hypernymy	hypernymy	NOUN
cril-239	179	19	levels	level	NOUN
cril-239	179	20	.	.	PUNCT
cril-239	180	1	11	11	NUM
cril-239	180	2	salvetti	salvetti	VERB
cril-239	180	3	et	et	PROPN
cril-239	180	4	al	al	PROPN
cril-239	180	5	.	.	PROPN
cril-239	180	6	:	:	PUNCT
cril-239	181	1	automatic	automatic	ADJ
cril-239	181	2	opinion	opinion	NOUN
cril-239	181	3	polarity	polarity	NOUN
cril-239	181	4	classification	classification	NOUN
cril-239	181	5	of	of	ADP
cril-239	181	6	movie	movie	NOUN
cril-239	181	7	reviews	review	NOUN
cril-239	181	8	published	publish	VERB
cril-239	181	9	by	by	ADP
cril-239	181	10	cu	cu	PROPN
cril-239	181	11	scholar	scholar	NOUN
cril-239	181	12	,	,	PUNCT
cril-239	181	13	2004	2004	NUM
cril-239	181	14	colorado	colorado	PROPN
cril-239	181	15	research	research	NOUN
cril-239	181	16	in	in	ADP
cril-239	181	17	linguistics	linguistic	NOUN
cril-239	181	18	,	,	PUNCT
cril-239	181	19	volume	volume	NOUN
cril-239	181	20	17	17	NUM
cril-239	181	21	(	(	PUNCT
cril-239	181	22	2004	2004	NUM
cril-239	181	23	)	)	PUNCT
cril-239	181	24	12	12	NUM
cril-239	181	25	7.4	7.4	NUM
cril-239	181	26	.	.	PUNCT
cril-239	182	1	discussion	discussion	NOUN
cril-239	182	2	our	our	PRON
cril-239	182	3	results	result	NOUN
cril-239	182	4	indicate	indicate	VERB
cril-239	182	5	that	that	SCONJ
cril-239	182	6	,	,	PUNCT
cril-239	182	7	except	except	SCONJ
cril-239	182	8	for	for	ADP
cril-239	182	9	very	very	ADV
cril-239	182	10	small	small	ADJ
cril-239	182	11	data	data	NOUN
cril-239	182	12	sets	set	NOUN
cril-239	182	13	,	,	PUNCT
cril-239	182	14	the	the	DET
cril-239	182	15	use	use	NOUN
cril-239	182	16	of	of	ADP
cril-239	182	17	word	word	NOUN
cril-239	182	18	-	-	PUNCT
cril-239	182	19	net	net	NOUN
cril-239	182	20	hypernymy	hypernymy	NOUN
cril-239	182	21	generalization	generalization	NOUN
cril-239	182	22	is	be	AUX
cril-239	182	23	not	not	PART
cril-239	182	24	significantly	significantly	ADV
cril-239	182	25	beneficial	beneficial	ADJ
cril-239	182	26	to	to	ADP
cril-239	182	27	the	the	DET
cril-239	182	28	classification	classification	NOUN
cril-239	182	29	process	process	NOUN
cril-239	182	30	.	.	PUNCT
cril-239	183	1	we	we	PRON
cril-239	183	2	assume	assume	VERB
cril-239	183	3	that	that	SCONJ
cril-239	183	4	this	this	PRON
cril-239	183	5	is	be	AUX
cril-239	183	6	due	due	ADJ
cril-239	183	7	to	to	ADP
cril-239	183	8	at	at	ADP
cril-239	183	9	least	least	ADJ
cril-239	183	10	the	the	DET
cril-239	183	11	following	following	ADJ
cril-239	183	12	reasons	reason	NOUN
cril-239	183	13	:	:	PUNCT
cril-239	183	14	•	•	NUM
cril-239	183	15	wordnet	wordnet	NOUN
cril-239	183	16	is	be	AUX
cril-239	183	17	too	too	ADV
cril-239	183	18	general	general	ADJ
cril-239	183	19	for	for	ADP
cril-239	183	20	our	our	PRON
cril-239	183	21	purposes	purpose	NOUN
cril-239	183	22	:	:	PUNCT
cril-239	183	23	it	it	PRON
cril-239	183	24	considers	consider	VERB
cril-239	183	25	many	many	ADJ
cril-239	183	26	meanings	meaning	NOUN
cril-239	183	27	and	and	CCONJ
cril-239	183	28	hypernymy	hypernymy	VERB
cril-239	183	29	relations	relation	NOUN
cril-239	183	30	which	which	PRON
cril-239	183	31	are	be	AUX
cril-239	183	32	rarely	rarely	ADV
cril-239	183	33	relevant	relevant	ADJ
cril-239	183	34	to	to	ADP
cril-239	183	35	the	the	DET
cril-239	183	36	field	field	NOUN
cril-239	183	37	of	of	ADP
cril-239	183	38	movie	movie	NOUN
cril-239	183	39	reviews	review	NOUN
cril-239	183	40	,	,	PUNCT
cril-239	183	41	but	but	CCONJ
cril-239	183	42	which	which	PRON
cril-239	183	43	potentially	potentially	ADV
cril-239	183	44	take	take	VERB
cril-239	183	45	precedence	precedence	NOUN
cril-239	183	46	over	over	ADP
cril-239	183	47	other	other	ADJ
cril-239	183	48	relations	relation	NOUN
cril-239	183	49	which	which	PRON
cril-239	183	50	might	might	AUX
cril-239	183	51	be	be	AUX
cril-239	183	52	more	more	ADV
cril-239	183	53	appropriate	appropriate	ADJ
cril-239	183	54	here	here	ADV
cril-239	183	55	.	.	PUNCT
cril-239	184	1	•	•	NOUN
cril-239	184	2	choosing	choose	VERB
cril-239	184	3	the	the	DET
cril-239	184	4	first	first	ADJ
cril-239	184	5	synset	synset	VERB
cril-239	184	6	out	out	ADP
cril-239	184	7	of	of	ADP
cril-239	184	8	the	the	DET
cril-239	184	9	set	set	NOUN
cril-239	184	10	of	of	ADP
cril-239	184	11	choices	choice	NOUN
cril-239	184	12	is	be	AUX
cril-239	184	13	unlikely	unlikely	ADJ
cril-239	184	14	to	to	PART
cril-239	184	15	yield	yield	VERB
cril-239	184	16	the	the	DET
cril-239	184	17	correct	correct	ADJ
cril-239	184	18	result	result	NOUN
cril-239	184	19	,	,	PUNCT
cril-239	184	20	given	give	VERB
cril-239	184	21	the	the	DET
cril-239	184	22	lack	lack	NOUN
cril-239	184	23	of	of	ADP
cril-239	184	24	wordnet	wordnet	NOUN
cril-239	184	25	's	's	PART
cril-239	184	26	specialization	specialization	NOUN
cril-239	184	27	on	on	ADP
cril-239	184	28	our	our	PRON
cril-239	184	29	domain	domain	NOUN
cril-239	184	30	of	of	ADP
cril-239	184	31	focus	focus	NOUN
cril-239	184	32	.	.	PUNCT
cril-239	185	1	•	•	NOUN
cril-239	185	2	for	for	ADP
cril-239	185	3	reasonably	reasonably	ADV
cril-239	185	4	large	large	ADJ
cril-239	185	5	data	datum	NOUN
cril-239	185	6	sets	set	NOUN
cril-239	185	7	,	,	PUNCT
cril-239	185	8	supervised	supervised	ADJ
cril-239	185	9	learning	learning	NOUN
cril-239	185	10	mechanisms	mechanism	NOUN
cril-239	185	11	gain	gain	VERB
cril-239	185	12	sufficient	sufficient	ADJ
cril-239	185	13	confidence	confidence	NOUN
cril-239	185	14	with	with	ADP
cril-239	185	15	related	related	ADJ
cril-239	185	16	words	word	NOUN
cril-239	185	17	to	to	PART
cril-239	185	18	make	make	VERB
cril-239	185	19	this	this	DET
cril-239	185	20	particular	particular	ADJ
cril-239	185	21	auxiliary	auxiliary	ADJ
cril-239	185	22	technique	technique	NOUN
cril-239	185	23	less	less	ADV
cril-239	185	24	useful	useful	ADJ
cril-239	185	25	.	.	PUNCT
cril-239	186	1	considering	consider	VERB
cril-239	186	2	this	this	PRON
cril-239	186	3	,	,	PUNCT
cril-239	186	4	the	the	DET
cril-239	186	5	use	use	NOUN
cril-239	186	6	of	of	ADP
cril-239	186	7	a	a	DET
cril-239	186	8	domain	domain	NOUN
cril-239	186	9	-	-	PUNCT
cril-239	186	10	specific	specific	ADJ
cril-239	186	11	database	database	NOUN
cril-239	186	12	seems	seem	VERB
cril-239	186	13	to	to	PART
cril-239	186	14	be	be	AUX
cril-239	186	15	a	a	DET
cril-239	186	16	promising	promising	ADJ
cril-239	186	17	approach	approach	NOUN
cril-239	186	18	to	to	ADP
cril-239	186	19	improving	improve	VERB
cril-239	186	20	our	our	PRON
cril-239	186	21	performance	performance	NOUN
cril-239	186	22	for	for	ADP
cril-239	186	23	this	this	DET
cril-239	186	24	technique	technique	NOUN
cril-239	186	25	.	.	PUNCT
cril-239	187	1	8	8	X
cril-239	187	2	.	.	X
cril-239	187	3	selection	selection	NOUN
cril-239	187	4	by	by	ADP
cril-239	187	5	ranking	rank	VERB
cril-239	187	6	the	the	DET
cril-239	187	7	probabilistic	probabilistic	ADJ
cril-239	187	8	models	model	NOUN
cril-239	187	9	computed	compute	VERB
cril-239	187	10	by	by	ADP
cril-239	187	11	the	the	DET
cril-239	187	12	naïve	naïve	ADJ
cril-239	187	13	bayes	bayes	PROPN
cril-239	187	14	classifiers	classifier	NOUN
cril-239	187	15	were	be	AUX
cril-239	187	16	sorted	sort	VERB
cril-239	187	17	by	by	ADP
cril-239	187	18	log	log	NOUN
cril-239	187	19	posterior	posterior	ADJ
cril-239	187	20	odds	odd	NOUN
cril-239	187	21	on	on	ADP
cril-239	187	22	positive	positive	ADJ
cril-239	187	23	and	and	CCONJ
cril-239	187	24	negative	negative	ADJ
cril-239	187	25	orientations	orientation	NOUN
cril-239	187	26	for	for	ADP
cril-239	187	27	the	the	DET
cril-239	187	28	purpose	purpose	NOUN
cril-239	187	29	of	of	ADP
cril-239	187	30	ranking	rank	VERB
cril-239	187	31	,	,	PUNCT
cril-239	187	32	i.e.	i.e.	X
cril-239	187	33	by	by	ADP
cril-239	187	34	a	a	DET
cril-239	187	35	"	"	PUNCT
cril-239	187	36	score	score	NOUN
cril-239	187	37	"	"	PUNCT
cril-239	187	38	computed	compute	VERB
cril-239	187	39	as	as	SCONJ
cril-239	187	40	follows	follow	VERB
cril-239	187	41	:	:	PUNCT
cril-239	187	42	score	score	NOUN
cril-239	187	43	=	=	PUNCT
cril-239	187	44	log	log	VERB
cril-239	187	45	pr(+|rv	pr(+|rv	PROPN
cril-239	187	46	)	)	PUNCT
cril-239	187	47	log	log	NOUN
cril-239	187	48	pr(-|rv	pr(-|rv	NOUN
cril-239	187	49	)	)	PUNCT
cril-239	187	50	where	where	SCONJ
cril-239	187	51	•	•	NOUN
cril-239	187	52	rv	rv	PROPN
cril-239	187	53	is	be	AUX
cril-239	187	54	the	the	DET
cril-239	187	55	review	review	NOUN
cril-239	187	56	under	under	ADP
cril-239	187	57	consideration	consideration	NOUN
cril-239	187	58	,	,	PUNCT
cril-239	187	59	•	•	ADP
cril-239	187	60	pr(+|rv	pr(+|rv	NOUN
cril-239	187	61	)	)	PUNCT
cril-239	187	62	is	be	AUX
cril-239	187	63	the	the	DET
cril-239	187	64	probability	probability	NOUN
cril-239	187	65	of	of	ADP
cril-239	187	66	rv	rv	PROPN
cril-239	187	67	being	be	AUX
cril-239	187	68	a	a	DET
cril-239	187	69	review	review	NOUN
cril-239	187	70	of	of	ADP
cril-239	187	71	positive	positive	ADJ
cril-239	187	72	polarity	polarity	NOUN
cril-239	187	73	,	,	PUNCT
cril-239	187	74	•	•	NUM
cril-239	187	75	pr(-|rv	pr(-|rv	NOUN
cril-239	187	76	)	)	PUNCT
cril-239	187	77	analogously	analogously	ADV
cril-239	187	78	is	be	AUX
cril-239	187	79	the	the	DET
cril-239	187	80	probability	probability	NOUN
cril-239	187	81	of	of	ADP
cril-239	187	82	the	the	DET
cril-239	187	83	review	review	NOUN
cril-239	187	84	being	be	AUX
cril-239	187	85	of	of	ADP
cril-239	187	86	negative	negative	ADJ
cril-239	187	87	polarity	polarity	NOUN
cril-239	187	88	.	.	PUNCT
cril-239	188	1	we	we	PRON
cril-239	188	2	modified	modify	VERB
cril-239	188	3	the	the	DET
cril-239	188	4	classifier	classifier	NOUN
cril-239	188	5	so	so	SCONJ
cril-239	188	6	that	that	SCONJ
cril-239	188	7	it	it	PRON
cril-239	188	8	:	:	PUNCT
cril-239	188	9	1	1	X
cril-239	188	10	.	.	X
cril-239	188	11	sorts	sort	VERB
cril-239	188	12	the	the	DET
cril-239	188	13	reviews	review	NOUN
cril-239	188	14	in	in	ADP
cril-239	188	15	the	the	DET
cril-239	188	16	test	test	NOUN
cril-239	188	17	data	datum	NOUN
cril-239	188	18	by	by	ADP
cril-239	188	19	log	log	NOUN
cril-239	188	20	posterior	posterior	ADJ
cril-239	188	21	odds	odd	NOUN
cril-239	188	22	2	2	NUM
cril-239	188	23	.	.	PUNCT
cril-239	188	24	returns	return	VERB
cril-239	188	25	the	the	DET
cril-239	188	26	first	first	ADJ
cril-239	188	27	n	n	NOUN
cril-239	188	28	reviews	review	NOUN
cril-239	188	29	from	from	ADP
cril-239	188	30	the	the	DET
cril-239	188	31	sorted	sorted	ADJ
cril-239	188	32	list	list	NOUN
cril-239	188	33	as	as	ADP
cril-239	188	34	positive	positive	ADJ
cril-239	188	35	reviews	review	NOUN
cril-239	188	36	3	3	NUM
cril-239	188	37	.	.	PUNCT
cril-239	189	1	returns	return	NOUN
cril-239	189	2	the	the	DET
cril-239	189	3	last	last	ADJ
cril-239	189	4	n	n	NOUN
cril-239	189	5	reviews	review	NOUN
cril-239	189	6	from	from	ADP
cril-239	189	7	the	the	DET
cril-239	189	8	sorted	sorted	ADJ
cril-239	189	9	list	list	NOUN
cril-239	189	10	as	as	ADP
cril-239	189	11	negative	negative	ADJ
cril-239	189	12	reviews	review	NOUN
cril-239	189	13	the	the	DET
cril-239	189	14	resulting	result	VERB
cril-239	189	15	accuracies	accuracy	NOUN
cril-239	189	16	and	and	CCONJ
cril-239	189	17	recalls	recall	VERB
cril-239	189	18	on	on	ADP
cril-239	189	19	different	different	ADJ
cril-239	189	20	n	n	NOUN
cril-239	189	21	are	be	AUX
cril-239	189	22	summarized	summarize	VERB
cril-239	189	23	in	in	ADP
cril-239	189	24	table	table	NOUN
cril-239	189	25	4	4	NUM
cril-239	189	26	.	.	NOUN
cril-239	189	27	12	12	NUM
cril-239	189	28	colorado	colorado	NOUN
cril-239	189	29	research	research	NOUN
cril-239	189	30	in	in	ADP
cril-239	189	31	linguistics	linguistic	NOUN
cril-239	189	32	,	,	PUNCT
cril-239	189	33	vol	vol	NOUN
cril-239	189	34	.	.	PROPN
cril-239	189	35	17	17	NUM
cril-239	190	1	[	[	X
cril-239	190	2	2004	2004	NUM
cril-239	190	3	]	]	X
cril-239	190	4	https://scholar.colorado.edu/cril/vol17/iss1/1	https://scholar.colorado.edu/cril/vol17/iss1/1	PROPN
cril-239	190	5	doi	doi	NOUN
cril-239	190	6	:	:	PUNCT
cril-239	190	7	https://doi.org/10.25810/atv1-v819	https://doi.org/10.25810/atv1-v819	VERB
cril-239	190	8	automatic	automatic	ADJ
cril-239	190	9	opinion	opinion	NOUN
cril-239	190	10	polarity	polarity	NOUN
cril-239	190	11	classification	classification	NOUN
cril-239	190	12	of	of	ADP
cril-239	190	13	movie	movie	NOUN
cril-239	190	14	reviews	review	NOUN
cril-239	190	15	13	13	NUM
cril-239	190	16	table	table	NOUN
cril-239	190	17	4	4	NUM
cril-239	190	18	:	:	PUNCT
cril-239	190	19	precisions	precision	NOUN
cril-239	190	20	and	and	CCONJ
cril-239	190	21	recalls	recall	VERB
cril-239	190	22	by	by	ADP
cril-239	190	23	number	number	NOUN
cril-239	190	24	of	of	ADP
cril-239	190	25	inputs	input	NOUN
cril-239	190	26	the	the	DET
cril-239	190	27	classifier	classifier	NOUN
cril-239	190	28	was	be	AUX
cril-239	190	29	trained	train	VERB
cril-239	190	30	on	on	ADP
cril-239	190	31	the	the	DET
cril-239	190	32	same	same	ADJ
cril-239	190	33	1500	1500	NUM
cril-239	190	34	review	review	NOUN
cril-239	190	35	data	datum	NOUN
cril-239	190	36	set	set	VERB
cril-239	190	37	and	and	CCONJ
cril-239	190	38	was	be	AUX
cril-239	190	39	used	use	VERB
cril-239	190	40	with	with	ADP
cril-239	190	41	ranking	rank	VERB
cril-239	190	42	on	on	ADP
cril-239	190	43	a	a	DET
cril-239	190	44	repository	repository	NOUN
cril-239	190	45	of	of	ADP
cril-239	190	46	200	200	NUM
cril-239	190	47	reviews	review	NOUN
cril-239	190	48	which	which	PRON
cril-239	190	49	were	be	AUX
cril-239	190	50	identical	identical	ADJ
cril-239	190	51	to	to	ADP
cril-239	190	52	the	the	DET
cril-239	190	53	test	test	NOUN
cril-239	190	54	data	datum	NOUN
cril-239	190	55	set	set	VERB
cril-239	190	56	.	.	PUNCT
cril-239	191	1	the	the	DET
cril-239	191	2	result	result	NOUN
cril-239	191	3	is	be	AUX
cril-239	191	4	very	very	ADV
cril-239	191	5	positive	positive	ADJ
cril-239	191	6	and	and	CCONJ
cril-239	191	7	indicates	indicate	VERB
cril-239	191	8	that	that	SCONJ
cril-239	191	9	adjectives	adjective	NOUN
cril-239	191	10	provide	provide	VERB
cril-239	191	11	enough	enough	ADJ
cril-239	191	12	sentiment	sentiment	NOUN
cril-239	191	13	to	to	PART
cril-239	191	14	detect	detect	VERB
cril-239	191	15	extremely	extremely	ADV
cril-239	191	16	positive	positive	ADJ
cril-239	191	17	or	or	CCONJ
cril-239	191	18	negative	negative	ADJ
cril-239	191	19	reviews	review	NOUN
cril-239	191	20	with	with	ADP
cril-239	191	21	good	good	ADJ
cril-239	191	22	accuracy	accuracy	NOUN
cril-239	191	23	.	.	PUNCT
cril-239	192	1	while	while	SCONJ
cril-239	192	2	the	the	DET
cril-239	192	3	number	number	NOUN
cril-239	192	4	of	of	ADP
cril-239	192	5	reviews	review	NOUN
cril-239	192	6	returned	return	VERB
cril-239	192	7	is	be	AUX
cril-239	192	8	specified	specify	VERB
cril-239	192	9	in	in	ADP
cril-239	192	10	this	this	DET
cril-239	192	11	particular	particular	ADJ
cril-239	192	12	example	example	NOUN
cril-239	192	13	,	,	PUNCT
cril-239	192	14	it	it	PRON
cril-239	192	15	is	be	AUX
cril-239	192	16	also	also	ADV
cril-239	192	17	possible	possible	ADJ
cril-239	192	18	to	to	PART
cril-239	192	19	use	use	VERB
cril-239	192	20	assurance	assurance	NOUN
cril-239	192	21	as	as	ADP
cril-239	192	22	the	the	DET
cril-239	192	23	cutoff	cutoff	NOUN
cril-239	192	24	criterion	criterion	NOUN
cril-239	192	25	by	by	ADP
cril-239	192	26	giving	give	VERB
cril-239	192	27	log	log	NOUN
cril-239	192	28	posterior	posterior	ADJ
cril-239	192	29	odds	odd	NOUN
cril-239	192	30	.	.	PUNCT
cril-239	193	1	9	9	X
cril-239	193	2	.	.	X
cril-239	193	3	discussion	discussion	NOUN
cril-239	193	4	taking	take	VERB
cril-239	193	5	all	all	DET
cril-239	193	6	results	result	NOUN
cril-239	193	7	into	into	ADP
cril-239	193	8	consideration	consideration	NOUN
cril-239	193	9	,	,	PUNCT
cril-239	193	10	both	both	CCONJ
cril-239	193	11	the	the	DET
cril-239	193	12	naïve	naïve	ADJ
cril-239	193	13	bayes	bayes	NOUN
cril-239	193	14	classifier	classifier	NOUN
cril-239	193	15	and	and	CCONJ
cril-239	193	16	bigram	bigram	PROPN
cril-239	193	17	markov	markov	PROPN
cril-239	193	18	model	model	NOUN
cril-239	193	19	classifier	classifier	NOUN
cril-239	193	20	performed	perform	VERB
cril-239	193	21	best	well	ADV
cril-239	193	22	when	when	SCONJ
cril-239	193	23	trained	train	VERB
cril-239	193	24	on	on	ADP
cril-239	193	25	sufficiently	sufficiently	ADV
cril-239	193	26	large	large	ADJ
cril-239	193	27	data	data	NOUN
cril-239	193	28	sets	set	NOUN
cril-239	193	29	without	without	ADP
cril-239	193	30	filtering	filter	VERB
cril-239	193	31	.	.	PUNCT
cril-239	194	1	for	for	ADP
cril-239	194	2	both	both	CCONJ
cril-239	194	3	bigram	bigram	PROPN
cril-239	194	4	and	and	CCONJ
cril-239	194	5	trigram	trigram	NOUN
cril-239	194	6	markov	markov	NOUN
cril-239	194	7	models	model	NOUN
cril-239	194	8	,	,	PUNCT
cril-239	194	9	we	we	PRON
cril-239	194	10	observed	observe	VERB
cril-239	194	11	a	a	DET
cril-239	194	12	noticeable	noticeable	ADJ
cril-239	194	13	improvement	improvement	NOUN
cril-239	194	14	with	with	ADP
cril-239	194	15	our	our	PRON
cril-239	194	16	generalization	generalization	NOUN
cril-239	194	17	filter	filter	NOUN
cril-239	194	18	when	when	SCONJ
cril-239	194	19	training	train	VERB
cril-239	194	20	on	on	ADP
cril-239	194	21	very	very	ADV
cril-239	194	22	small	small	ADJ
cril-239	194	23	data	data	NOUN
cril-239	194	24	sets	set	NOUN
cril-239	194	25	;	;	PUNCT
cril-239	194	26	for	for	ADP
cril-239	194	27	trigram	trigram	NOUN
cril-239	194	28	models	model	NOUN
cril-239	194	29	,	,	PUNCT
cril-239	194	30	this	this	DET
cril-239	194	31	improvement	improvement	NOUN
cril-239	194	32	even	even	ADV
cril-239	194	33	extended	extend	VERB
cril-239	194	34	to	to	ADP
cril-239	194	35	fairly	fairly	ADV
cril-239	194	36	large	large	ADJ
cril-239	194	37	data	data	NOUN
cril-239	194	38	sets	set	NOUN
cril-239	194	39	(	(	PUNCT
cril-239	194	40	1500	1500	NUM
cril-239	194	41	reviews	review	NOUN
cril-239	194	42	)	)	PUNCT
cril-239	194	43	.	.	PUNCT
cril-239	195	1	one	one	NUM
cril-239	195	2	explanation	explanation	NOUN
cril-239	195	3	for	for	ADP
cril-239	195	4	this	this	DET
cril-239	195	5	result	result	NOUN
cril-239	195	6	is	be	AUX
cril-239	195	7	that	that	SCONJ
cril-239	195	8	the	the	DET
cril-239	195	9	filters	filter	NOUN
cril-239	195	10	are	be	AUX
cril-239	195	11	unable	unable	ADJ
cril-239	195	12	to	to	PART
cril-239	195	13	make	make	VERB
cril-239	195	14	use	use	NOUN
cril-239	195	15	of	of	ADP
cril-239	195	16	the	the	DET
cril-239	195	17	more	more	ADV
cril-239	195	18	fine	fine	ADV
cril-239	195	19	-	-	PUNCT
cril-239	195	20	grained	grain	VERB
cril-239	195	21	information	information	NOUN
cril-239	195	22	provided	provide	VERB
cril-239	195	23	to	to	ADP
cril-239	195	24	them	they	PRON
cril-239	195	25	.	.	PUNCT
cril-239	196	1	a	a	DET
cril-239	196	2	likely	likely	ADJ
cril-239	196	3	reason	reason	NOUN
cril-239	196	4	for	for	ADP
cril-239	196	5	this	this	PRON
cril-239	196	6	is	be	AUX
cril-239	196	7	that	that	SCONJ
cril-239	196	8	the	the	DET
cril-239	196	9	ratio	ratio	NOUN
cril-239	196	10	between	between	ADP
cril-239	196	11	the	the	DET
cril-239	196	12	size	size	NOUN
cril-239	196	13	of	of	ADP
cril-239	196	14	the	the	DET
cril-239	196	15	set	set	NOUN
cril-239	196	16	of	of	ADP
cril-239	196	17	observable	observable	ADJ
cril-239	196	18	events	event	NOUN
cril-239	196	19	and	and	CCONJ
cril-239	196	20	the	the	DET
cril-239	196	21	size	size	NOUN
cril-239	196	22	of	of	ADP
cril-239	196	23	the	the	DET
cril-239	196	24	training	training	NOUN
cril-239	196	25	data	datum	NOUN
cril-239	196	26	set	set	VERB
cril-239	196	27	is	be	AUX
cril-239	196	28	comparatively	comparatively	ADV
cril-239	196	29	large	large	ADJ
cril-239	196	30	in	in	ADP
cril-239	196	31	both	both	DET
cril-239	196	32	cases	case	NOUN
cril-239	196	33	.	.	PUNCT
cril-239	197	1	however	however	ADV
cril-239	197	2	,	,	PUNCT
cril-239	197	3	further	further	ADJ
cril-239	197	4	research	research	NOUN
cril-239	197	5	and	and	CCONJ
cril-239	197	6	testing	testing	NOUN
cril-239	197	7	will	will	AUX
cril-239	197	8	be	be	AUX
cril-239	197	9	required	require	VERB
cril-239	197	10	in	in	ADP
cril-239	197	11	order	order	NOUN
cril-239	197	12	to	to	PART
cril-239	197	13	establish	establish	VERB
cril-239	197	14	a	a	DET
cril-239	197	15	more	more	ADV
cril-239	197	16	concrete	concrete	ADJ
cril-239	197	17	understanding	understanding	NOUN
cril-239	197	18	of	of	ADP
cril-239	197	19	the	the	DET
cril-239	197	20	usefulness	usefulness	NOUN
cril-239	197	21	of	of	ADP
cril-239	197	22	this	this	DET
cril-239	197	23	technique	technique	NOUN
cril-239	197	24	.	.	PUNCT
cril-239	198	1	the	the	DET
cril-239	198	2	learning	learn	VERB
cril-239	198	3	curve	curve	NOUN
cril-239	198	4	of	of	ADP
cril-239	198	5	classifiers	classifier	NOUN
cril-239	198	6	with	with	ADP
cril-239	198	7	the	the	DET
cril-239	198	8	pos	pos	NOUN
cril-239	198	9	filter	filter	NOUN
cril-239	198	10	and/or	and/or	CCONJ
cril-239	198	11	the	the	DET
cril-239	198	12	generalization	generalization	NOUN
cril-239	198	13	filter	filter	NOUN
cril-239	198	14	climbs	climb	VERB
cril-239	198	15	at	at	ADP
cril-239	198	16	higher	high	ADJ
cril-239	198	17	rates	rate	NOUN
cril-239	198	18	than	than	ADP
cril-239	198	19	those	those	PRON
cril-239	198	20	without	without	ADP
cril-239	198	21	the	the	DET
cril-239	198	22	filters	filter	NOUN
cril-239	198	23	and	and	CCONJ
cril-239	198	24	results	result	NOUN
cril-239	198	25	in	in	ADP
cril-239	198	26	lower	low	ADJ
cril-239	198	27	accuracy	accuracy	NOUN
cril-239	198	28	with	with	ADP
cril-239	198	29	larger	large	ADJ
cril-239	198	30	data	datum	NOUN
cril-239	198	31	sets	set	NOUN
cril-239	198	32	.	.	PUNCT
cril-239	199	1	one	one	NUM
cril-239	199	2	possible	possible	ADJ
cril-239	199	3	explanation	explanation	NOUN
cril-239	199	4	of	of	ADP
cril-239	199	5	the	the	DET
cril-239	199	6	higher	high	ADJ
cril-239	199	7	climbing	climbing	NOUN
cril-239	199	8	rates	rate	NOUN
cril-239	199	9	is	be	AUX
cril-239	199	10	that	that	SCONJ
cril-239	199	11	the	the	DET
cril-239	199	12	pos	pos	NOUN
cril-239	199	13	filter	filter	NOUN
cril-239	199	14	and	and	CCONJ
cril-239	199	15	the	the	DET
cril-239	199	16	generalization	generalization	NOUN
cril-239	199	17	filter	filter	NOUN
cril-239	199	18	compact	compact	VERB
cril-239	199	19	the	the	DET
cril-239	199	20	possible	possible	ADJ
cril-239	199	21	events	event	NOUN
cril-239	199	22	in	in	ADP
cril-239	199	23	language	language	NOUN
cril-239	199	24	models	model	NOUN
cril-239	199	25	while	while	SCONJ
cril-239	199	26	respecting	respect	VERB
cril-239	199	27	the	the	DET
cril-239	199	28	underlying	underlie	VERB
cril-239	199	29	model	model	NOUN
cril-239	199	30	by	by	ADP
cril-239	199	31	reducing	reduce	VERB
cril-239	199	32	vocabulary	vocabulary	NOUN
cril-239	199	33	.	.	PUNCT
cril-239	200	1	this	this	PRON
cril-239	200	2	also	also	ADV
cril-239	200	3	explains	explain	VERB
cril-239	200	4	why	why	SCONJ
cril-239	200	5	the	the	DET
cril-239	200	6	plateau	plateau	NOUN
cril-239	200	7	effect	effect	NOUN
cril-239	200	8	is	be	AUX
cril-239	200	9	observed	observe	VERB
cril-239	200	10	with	with	ADP
cril-239	200	11	smaller	small	ADJ
cril-239	200	12	data	datum	NOUN
cril-239	200	13	set	set	VERB
cril-239	200	14	sizes	size	NOUN
cril-239	200	15	.	.	PUNCT
cril-239	201	1	the	the	DET
cril-239	201	2	degraded	degraded	ADJ
cril-239	201	3	results	result	NOUN
cril-239	201	4	with	with	ADP
cril-239	201	5	filters	filter	NOUN
cril-239	201	6	also	also	ADV
cril-239	201	7	indicate	indicate	VERB
cril-239	201	8	that	that	SCONJ
cril-239	201	9	by	by	ADP
cril-239	201	10	removing	remove	VERB
cril-239	201	11	information	information	NOUN
cril-239	201	12	from	from	ADP
cril-239	201	13	training	training	NOUN
cril-239	201	14	and	and	CCONJ
cril-239	201	15	test	test	NOUN
cril-239	201	16	data	datum	NOUN
cril-239	201	17	,	,	PUNCT
cril-239	201	18	the	the	DET
cril-239	201	19	compacted	compact	VERB
cril-239	201	20	language	language	NOUN
cril-239	201	21	model	model	NOUN
cril-239	201	22	loses	lose	VERB
cril-239	201	23	resolution	resolution	NOUN
cril-239	201	24	.	.	PUNCT
cril-239	202	1	13	13	NUM
cril-239	202	2	salvetti	salvetti	NOUN
cril-239	202	3	et	et	PROPN
cril-239	202	4	al	al	PROPN
cril-239	202	5	.	.	PROPN
cril-239	202	6	:	:	PUNCT
cril-239	203	1	automatic	automatic	ADJ
cril-239	203	2	opinion	opinion	NOUN
cril-239	203	3	polarity	polarity	NOUN
cril-239	203	4	classification	classification	NOUN
cril-239	203	5	of	of	ADP
cril-239	203	6	movie	movie	NOUN
cril-239	203	7	reviews	review	NOUN
cril-239	203	8	published	publish	VERB
cril-239	203	9	by	by	ADP
cril-239	203	10	cu	cu	PROPN
cril-239	203	11	scholar	scholar	NOUN
cril-239	203	12	,	,	PUNCT
cril-239	203	13	2004	2004	NUM
cril-239	203	14	colorado	colorado	PROPN
cril-239	203	15	research	research	NOUN
cril-239	203	16	in	in	ADP
cril-239	203	17	linguistics	linguistic	NOUN
cril-239	203	18	,	,	PUNCT
cril-239	203	19	volume	volume	NOUN
cril-239	203	20	17	17	NUM
cril-239	203	21	(	(	PUNCT
cril-239	203	22	2004	2004	NUM
cril-239	203	23	)	)	PUNCT
cril-239	203	24	14	14	NUM
cril-239	203	25	10	10	NUM
cril-239	203	26	.	.	PUNCT
cril-239	204	1	conclusion	conclusion	VERB
cril-239	204	2	a	a	DET
cril-239	204	3	framework	framework	NOUN
cril-239	204	4	of	of	ADP
cril-239	204	5	two	two	NUM
cril-239	204	6	-	-	PUNCT
cril-239	204	7	phased	phase	VERB
cril-239	204	8	classification	classification	NOUN
cril-239	204	9	mechanism	mechanism	NOUN
cril-239	204	10	is	be	AUX
cril-239	204	11	introduced	introduce	VERB
cril-239	204	12	and	and	CCONJ
cril-239	204	13	implemented	implement	VERB
cril-239	204	14	with	with	ADP
cril-239	204	15	a	a	DET
cril-239	204	16	pos	pos	NOUN
cril-239	204	17	filter	filter	NOUN
cril-239	204	18	,	,	PUNCT
cril-239	204	19	a	a	DET
cril-239	204	20	generalization	generalization	NOUN
cril-239	204	21	filter	filter	NOUN
cril-239	204	22	,	,	PUNCT
cril-239	204	23	a	a	DET
cril-239	204	24	naïve	naïve	ADJ
cril-239	204	25	bayes	bayes	NOUN
cril-239	204	26	classifier	classifier	NOUN
cril-239	204	27	and	and	CCONJ
cril-239	204	28	a	a	DET
cril-239	204	29	markov	markov	NOUN
cril-239	204	30	model	model	NOUN
cril-239	204	31	classifier	classifier	NOUN
cril-239	204	32	.	.	PUNCT
cril-239	205	1	accuracies	accuracy	NOUN
cril-239	205	2	of	of	ADP
cril-239	205	3	combinations	combination	NOUN
cril-239	205	4	of	of	ADP
cril-239	205	5	filters	filter	NOUN
cril-239	205	6	and	and	CCONJ
cril-239	205	7	classifiers	classifier	NOUN
cril-239	205	8	are	be	AUX
cril-239	205	9	evaluated	evaluate	VERB
cril-239	205	10	by	by	ADP
cril-239	205	11	experiments	experiment	NOUN
cril-239	205	12	.	.	PUNCT
cril-239	206	1	although	although	SCONJ
cril-239	206	2	the	the	DET
cril-239	206	3	results	result	NOUN
cril-239	206	4	from	from	ADP
cril-239	206	5	classifications	classification	NOUN
cril-239	206	6	without	without	ADP
cril-239	206	7	filters	filter	NOUN
cril-239	206	8	are	be	AUX
cril-239	206	9	better	well	ADJ
cril-239	206	10	than	than	ADP
cril-239	206	11	those	those	PRON
cril-239	206	12	with	with	ADP
cril-239	206	13	filters	filter	NOUN
cril-239	206	14	,	,	PUNCT
cril-239	206	15	the	the	DET
cril-239	206	16	pos	pos	NOUN
cril-239	206	17	filters	filter	NOUN
cril-239	206	18	and	and	CCONJ
cril-239	206	19	generalization	generalization	NOUN
cril-239	206	20	filters	filter	NOUN
cril-239	206	21	are	be	AUX
cril-239	206	22	observed	observe	VERB
cril-239	206	23	to	to	PART
cril-239	206	24	still	still	ADV
cril-239	206	25	have	have	VERB
cril-239	206	26	potential	potential	NOUN
cril-239	206	27	to	to	PART
cril-239	206	28	improve	improve	VERB
cril-239	206	29	overall	overall	ADJ
cril-239	206	30	opinion	opinion	NOUN
cril-239	206	31	polarity	polarity	NOUN
cril-239	206	32	identification	identification	NOUN
cril-239	206	33	.	.	PUNCT
cril-239	207	1	generalization	generalization	NOUN
cril-239	207	2	filtering	filtering	NOUN
cril-239	207	3	using	use	VERB
cril-239	207	4	wordnet	wordnet	NOUN
cril-239	207	5	shows	show	VERB
cril-239	207	6	good	good	ADJ
cril-239	207	7	accuracy	accuracy	NOUN
cril-239	207	8	for	for	ADP
cril-239	207	9	small	small	ADJ
cril-239	207	10	data	datum	NOUN
cril-239	207	11	sets	set	NOUN
cril-239	207	12	and	and	CCONJ
cril-239	207	13	warrants	warrant	NOUN
cril-239	207	14	further	further	ADJ
cril-239	207	15	research	research	NOUN
cril-239	207	16	.	.	PUNCT
cril-239	208	1	using	use	VERB
cril-239	208	2	the	the	DET
cril-239	208	3	naïve	naïve	ADJ
cril-239	208	4	bayes	bayes	NOUN
cril-239	208	5	classifier	classifier	NOUN
cril-239	208	6	with	with	ADP
cril-239	208	7	ranking	rank	VERB
cril-239	208	8	on	on	ADP
cril-239	208	9	adjectives	adjective	NOUN
cril-239	208	10	has	have	AUX
cril-239	208	11	confirmed	confirm	VERB
cril-239	208	12	that	that	SCONJ
cril-239	208	13	desired	desire	VERB
cril-239	208	14	precision	precision	NOUN
cril-239	208	15	can	can	AUX
cril-239	208	16	be	be	AUX
cril-239	208	17	achieved	achieve	VERB
cril-239	208	18	by	by	ADP
cril-239	208	19	dropping	dropping	NOUN
cril-239	208	20	recalls	recall	VERB
cril-239	208	21	.	.	PUNCT
cril-239	209	1	for	for	ADP
cril-239	209	2	the	the	DET
cril-239	209	3	task	task	NOUN
cril-239	209	4	of	of	ADP
cril-239	209	5	finding	find	VERB
cril-239	209	6	reviews	review	NOUN
cril-239	209	7	of	of	ADP
cril-239	209	8	strong	strong	ADJ
cril-239	209	9	positive	positive	ADJ
cril-239	209	10	or	or	CCONJ
cril-239	209	11	negative	negative	ADJ
cril-239	209	12	polarity	polarity	NOUN
cril-239	209	13	within	within	ADP
cril-239	209	14	a	a	DET
cril-239	209	15	given	give	VERB
cril-239	209	16	data	datum	NOUN
cril-239	209	17	set	set	VERB
cril-239	209	18	,	,	PUNCT
cril-239	209	19	very	very	ADV
cril-239	209	20	high	high	ADJ
cril-239	209	21	precision	precision	NOUN
cril-239	209	22	was	be	AUX
cril-239	209	23	observed	observe	VERB
cril-239	209	24	for	for	ADP
cril-239	209	25	adequate	adequate	ADJ
cril-239	209	26	recall	recall	NOUN
cril-239	209	27	.	.	PUNCT
cril-239	210	1	acknowledgements	acknowledgement	NOUN
cril-239	210	2	the	the	DET
cril-239	210	3	authors	author	NOUN
cril-239	210	4	would	would	AUX
cril-239	210	5	like	like	VERB
cril-239	210	6	to	to	PART
cril-239	210	7	thank	thank	VERB
cril-239	210	8	tomohiro	tomohiro	PROPN
cril-239	210	9	oda	oda	PROPN
cril-239	210	10	for	for	ADP
cril-239	210	11	his	his	PRON
cril-239	210	12	extensive	extensive	ADJ
cril-239	210	13	help	help	NOUN
cril-239	210	14	and	and	CCONJ
cril-239	210	15	support	support	NOUN
cril-239	210	16	during	during	ADP
cril-239	210	17	the	the	DET
cril-239	210	18	course	course	NOUN
cril-239	210	19	of	of	ADP
cril-239	210	20	all	all	DET
cril-239	210	21	stages	stage	NOUN
cril-239	210	22	of	of	ADP
cril-239	210	23	the	the	DET
cril-239	210	24	project	project	NOUN
cril-239	210	25	.	.	PUNCT
cril-239	211	1	further	further	ADJ
cril-239	211	2	acknowledgements	acknowledgement	NOUN
cril-239	211	3	go	go	VERB
cril-239	211	4	to	to	ADP
cril-239	211	5	larry	larry	PROPN
cril-239	211	6	d.	d.	PROPN
cril-239	211	7	blair	blair	PROPN
cril-239	211	8	,	,	PUNCT
cril-239	211	9	assad	assad	PROPN
cril-239	211	10	jaharria	jaharria	PROPN
cril-239	211	11	,	,	PUNCT
cril-239	211	12	helen	helen	PROPN
cril-239	211	13	johnson	johnson	PROPN
cril-239	211	14	,	,	PUNCT
cril-239	211	15	jim	jim	PROPN
cril-239	211	16	martin	martin	PROPN
cril-239	211	17	,	,	PUNCT
cril-239	211	18	jeff	jeff	PROPN
cril-239	211	19	rueppel	rueppel	PROPN
cril-239	211	20	and	and	CCONJ
cril-239	211	21	philipp	philipp	PROPN
cril-239	211	22	wetzler	wetzler	NOUN
cril-239	211	23	for	for	ADP
cril-239	211	24	their	their	PRON
cril-239	211	25	valuable	valuable	ADJ
cril-239	211	26	contributions	contribution	NOUN
cril-239	211	27	.	.	PUNCT
cril-239	212	1	references	reference	NOUN
cril-239	212	2	eric	eric	PROPN
cril-239	212	3	brill	brill	PROPN
cril-239	212	4	.	.	PUNCT
cril-239	213	1	"	"	PUNCT
cril-239	213	2	transformation	transformation	NOUN
cril-239	213	3	-	-	PUNCT
cril-239	213	4	based	base	VERB
cril-239	213	5	error	error	NOUN
cril-239	213	6	-	-	PUNCT
cril-239	213	7	driven	drive	VERB
cril-239	213	8	learning	learning	NOUN
cril-239	213	9	and	and	CCONJ
cril-239	213	10	natural	natural	ADJ
cril-239	213	11	language	language	NOUN
cril-239	213	12	processing	processing	NOUN
cril-239	213	13	:	:	PUNCT
cril-239	213	14	a	a	DET
cril-239	213	15	case	case	NOUN
cril-239	213	16	study	study	NOUN
cril-239	213	17	in	in	ADP
cril-239	213	18	part	part	NOUN
cril-239	213	19	-	-	PUNCT
cril-239	213	20	of	of	ADP
cril-239	213	21	-	-	PUNCT
cril-239	213	22	speech	speech	NOUN
cril-239	213	23	tagging	tagging	NOUN
cril-239	213	24	"	"	PUNCT
cril-239	213	25	.	.	PUNCT
cril-239	214	1	computational	computational	ADJ
cril-239	214	2	linguistics	linguistic	NOUN
cril-239	214	3	,	,	PUNCT
cril-239	214	4	21(4):543–565	21(4):543–565	NOUN
cril-239	214	5	,	,	PUNCT
cril-239	214	6	1995	1995	NUM
cril-239	214	7	.	.	PUNCT
cril-239	215	1	tagger	tagger	NOUN
cril-239	215	2	available	available	ADJ
cril-239	215	3	from	from	ADP
cril-239	215	4	http://www.cs.jhu.edu/˜brill/rbt1	http://www.cs.jhu.edu/˜brill/rbt1	PROPN
cril-239	215	5	14.tar.z	14.tar.z	PROPN
cril-239	215	6	.	.	PUNCT
cril-239	216	1	christiane	christiane	PROPN
cril-239	216	2	fellbaum	fellbaum	PROPN
cril-239	216	3	.	.	PUNCT
cril-239	217	1	wordnet	wordnet	NOUN
cril-239	217	2	:	:	PUNCT
cril-239	217	3	an	an	DET
cril-239	217	4	electronic	electronic	ADJ
cril-239	217	5	lexical	lexical	ADJ
cril-239	217	6	database	database	NOUN
cril-239	217	7	,	,	PUNCT
cril-239	217	8	1998	1998	NUM
cril-239	217	9	.	.	PUNCT
cril-239	218	1	wordnet	wordnet	NOUN
cril-239	218	2	is	be	AUX
cril-239	218	3	available	available	ADJ
cril-239	218	4	from	from	ADP
cril-239	218	5	http://www.cogsci.princeton.edu/˜wn/.	http://www.cogsci.princeton.edu/˜wn/.	ADJ
cril-239	218	6	vasileios	vasileio	NOUN
cril-239	218	7	hatzivassiloglou	hatzivassiloglou	ADJ
cril-239	218	8	.	.	PUNCT
cril-239	219	1	effects	effect	NOUN
cril-239	219	2	of	of	ADP
cril-239	219	3	adjective	adjective	ADJ
cril-239	219	4	orientation	orientation	NOUN
cril-239	219	5	and	and	CCONJ
cril-239	219	6	gradability	gradability	NOUN
cril-239	219	7	on	on	ADP
cril-239	219	8	sentence	sentence	NOUN
cril-239	219	9	subjectivity	subjectivity	NOUN
cril-239	219	10	,	,	PUNCT
cril-239	219	11	2000	2000	NUM
cril-239	219	12	.	.	PUNCT
cril-239	220	1	vasileios	vasileio	NOUN
cril-239	220	2	hatzivassiloglou	hatzivassiloglou	PROPN
cril-239	220	3	and	and	CCONJ
cril-239	220	4	kathleen	kathleen	PROPN
cril-239	220	5	r.	r.	PROPN
cril-239	220	6	mckeown	mckeown	PROPN
cril-239	220	7	.	.	PUNCT
cril-239	221	1	"	"	PUNCT
cril-239	221	2	predicting	predict	VERB
cril-239	221	3	the	the	DET
cril-239	221	4	semantic	semantic	ADJ
cril-239	221	5	orientation	orientation	NOUN
cril-239	221	6	of	of	ADP
cril-239	221	7	adjectives	adjective	NOUN
cril-239	221	8	"	"	PUNCT
cril-239	221	9	.	.	PUNCT
cril-239	222	1	in	in	ADP
cril-239	222	2	philip	philip	PROPN
cril-239	222	3	r.	r.	PROPN
cril-239	222	4	cohen	cohen	PROPN
cril-239	222	5	and	and	CCONJ
cril-239	222	6	wolfgang	wolfgang	PROPN
cril-239	222	7	wahlster	wahlster	PROPN
cril-239	222	8	,	,	PUNCT
cril-239	222	9	editors	editor	NOUN
cril-239	222	10	,	,	PUNCT
cril-239	222	11	proceedings	proceeding	NOUN
cril-239	222	12	of	of	ADP
cril-239	222	13	the	the	DET
cril-239	222	14	thirty	thirty	NUM
cril-239	222	15	-	-	PUNCT
cril-239	222	16	fifth	fifth	ADJ
cril-239	222	17	annual	annual	ADJ
cril-239	222	18	meeting	meeting	NOUN
cril-239	222	19	of	of	ADP
cril-239	222	20	the	the	DET
cril-239	222	21	association	association	NOUN
cril-239	222	22	for	for	ADP
cril-239	222	23	computational	computational	ADJ
cril-239	222	24	linguistics	linguistic	NOUN
cril-239	222	25	and	and	CCONJ
cril-239	222	26	eighth	eighth	ADJ
cril-239	222	27	conference	conference	NOUN
cril-239	222	28	of	of	ADP
cril-239	222	29	the	the	DET
cril-239	222	30	european	european	ADJ
cril-239	222	31	chapter	chapter	NOUN
cril-239	222	32	of	of	ADP
cril-239	222	33	the	the	DET
cril-239	222	34	association	association	NOUN
cril-239	222	35	for	for	ADP
cril-239	222	36	computational	computational	ADJ
cril-239	222	37	linguistics	linguistic	NOUN
cril-239	222	38	,	,	PUNCT
cril-239	222	39	pages	page	NOUN
cril-239	222	40	174–181	174–181	NUM
cril-239	222	41	,	,	PUNCT
cril-239	222	42	somerset	somerset	PROPN
cril-239	222	43	,	,	PUNCT
cril-239	222	44	new	new	PROPN
cril-239	222	45	jersey	jersey	PROPN
cril-239	222	46	,	,	PUNCT
cril-239	222	47	1997	1997	NUM
cril-239	222	48	.	.	PUNCT
cril-239	223	1	association	association	NOUN
cril-239	223	2	for	for	ADP
cril-239	223	3	computational	computational	ADJ
cril-239	223	4	linguistics	linguistic	NOUN
cril-239	223	5	.	.	PUNCT
cril-239	224	1	robert	robert	PROPN
cril-239	224	2	m.	m.	PROPN
cril-239	224	3	losee	losee	PROPN
cril-239	224	4	.	.	PUNCT
cril-239	225	1	"	"	PUNCT
cril-239	225	2	natural	natural	ADJ
cril-239	225	3	language	language	NOUN
cril-239	225	4	processing	processing	NOUN
cril-239	225	5	in	in	ADP
cril-239	225	6	support	support	NOUN
cril-239	225	7	of	of	ADP
cril-239	225	8	decisionmaking	decisionmake	VERB
cril-239	225	9	:	:	PUNCT
cril-239	225	10	phrases	phrase	NOUN
cril-239	225	11	and	and	CCONJ
cril-239	225	12	part	part	NOUN
cril-239	225	13	-	-	PUNCT
cril-239	225	14	of	of	ADP
cril-239	225	15	-	-	PUNCT
cril-239	225	16	speech	speech	NOUN
cril-239	225	17	tagging	tagging	NOUN
cril-239	225	18	"	"	PUNCT
cril-239	225	19	.	.	PUNCT
cril-239	226	1	information	information	NOUN
cril-239	226	2	processing	processing	NOUN
cril-239	226	3	and	and	CCONJ
cril-239	226	4	management	management	NOUN
cril-239	226	5	,	,	PUNCT
cril-239	226	6	37(6):769–787	37(6):769–787	PROPN
cril-239	226	7	,	,	PUNCT
cril-239	226	8	2001	2001	NUM
cril-239	226	9	.	.	PUNCT
cril-239	227	1	mitchell	mitchell	PROPN
cril-239	227	2	p.	p.	PROPN
cril-239	227	3	marcus	marcus	PROPN
cril-239	227	4	,	,	PUNCT
cril-239	227	5	beatrice	beatrice	PROPN
cril-239	227	6	santorini	santorini	PROPN
cril-239	227	7	,	,	PUNCT
cril-239	227	8	and	and	CCONJ
cril-239	227	9	mary	mary	PROPN
cril-239	227	10	ann	ann	PROPN
cril-239	227	11	marcinkiewicz	marcinkiewicz	PROPN
cril-239	227	12	.	.	PUNCT
cril-239	228	1	"	"	PUNCT
cril-239	228	2	building	build	VERB
cril-239	228	3	a	a	DET
cril-239	228	4	large	large	ADJ
cril-239	228	5	annotated	annotated	ADJ
cril-239	228	6	corpus	corpus	NOUN
cril-239	228	7	of	of	ADP
cril-239	228	8	english	english	PROPN
cril-239	228	9	:	:	PUNCT
cril-239	228	10	the	the	DET
cril-239	228	11	penn	penn	PROPN
cril-239	228	12	treebank	treebank	PROPN
cril-239	228	13	"	"	PUNCT
cril-239	228	14	.	.	PUNCT
cril-239	229	1	computational	computational	ADJ
cril-239	229	2	linguistics	linguistic	NOUN
cril-239	229	3	,	,	PUNCT
cril-239	229	4	19(2):313–330	19(2):313–330	NUM
cril-239	229	5	,	,	PUNCT
cril-239	229	6	1994	1994	NUM
cril-239	229	7	.	.	PUNCT
cril-239	230	1	bo	bo	PROPN
cril-239	230	2	pang	pang	PROPN
cril-239	230	3	,	,	PUNCT
cril-239	230	4	lillian	lillian	PROPN
cril-239	230	5	lee	lee	PROPN
cril-239	230	6	,	,	PUNCT
cril-239	230	7	and	and	CCONJ
cril-239	230	8	shivakumar	shivakumar	PROPN
cril-239	230	9	vaithyanathan	vaithyanathan	PROPN
cril-239	230	10	.	.	PUNCT
cril-239	231	1	"	"	PUNCT
cril-239	231	2	thumbs	thumb	VERB
cril-239	231	3	up	up	ADP
cril-239	231	4	?	?	PUNCT
cril-239	232	1	sentiment	sentiment	NOUN
cril-239	232	2	classification	classification	NOUN
cril-239	232	3	using	use	VERB
cril-239	232	4	machine	machine	NOUN
cril-239	232	5	learning	learn	VERB
cril-239	232	6	techniques	technique	NOUN
cril-239	232	7	"	"	PUNCT
cril-239	232	8	.	.	PUNCT
cril-239	233	1	in	in	ADP
cril-239	233	2	proceedings	proceeding	NOUN
cril-239	233	3	of	of	ADP
cril-239	233	4	the	the	DET
cril-239	233	5	2002	2002	NUM
cril-239	233	6	conference	conference	NOUN
cril-239	233	7	on	on	ADP
cril-239	233	8	empirical	empirical	ADJ
cril-239	233	9	methods	method	NOUN
cril-239	233	10	in	in	ADP
cril-239	233	11	natural	natural	ADJ
cril-239	233	12	language	language	NOUN
cril-239	233	13	processing	processing	NOUN
cril-239	233	14	(	(	PUNCT
cril-239	233	15	emnlp	emnlp	ADJ
cril-239	233	16	)	)	PUNCT
cril-239	233	17	,	,	PUNCT
cril-239	233	18	2002	2002	NUM
cril-239	233	19	.	.	PUNCT
cril-239	234	1	movie	movie	NOUN
cril-239	234	2	review	review	NOUN
cril-239	234	3	14	14	NUM
cril-239	234	4	colorado	colorado	NOUN
cril-239	234	5	research	research	NOUN
cril-239	234	6	in	in	ADP
cril-239	234	7	linguistics	linguistic	NOUN
cril-239	234	8	,	,	PUNCT
cril-239	234	9	vol	vol	NOUN
cril-239	234	10	.	.	PROPN
cril-239	234	11	17	17	NUM
cril-239	235	1	[	[	X
cril-239	235	2	2004	2004	NUM
cril-239	235	3	]	]	X
cril-239	235	4	https://scholar.colorado.edu/cril/vol17/iss1/1	https://scholar.colorado.edu/cril/vol17/iss1/1	PROPN
cril-239	235	5	doi	doi	NOUN
cril-239	235	6	:	:	PUNCT
cril-239	235	7	https://doi.org/10.25810/atv1-v819	https://doi.org/10.25810/atv1-v819	VERB
cril-239	235	8	automatic	automatic	ADJ
cril-239	235	9	opinion	opinion	NOUN
cril-239	235	10	polarity	polarity	NOUN
cril-239	235	11	classification	classification	NOUN
cril-239	235	12	of	of	ADP
cril-239	235	13	movie	movie	NOUN
cril-239	235	14	reviews	review	NOUN
cril-239	235	15	15	15	NUM
cril-239	235	16	data	datum	NOUN
cril-239	235	17	available	available	ADJ
cril-239	235	18	at	at	ADP
cril-239	235	19	http://www.cs.cornell.edu/people/pabo/movie-reviewdata/movie.zip	http://www.cs.cornell.edu/people/pabo/movie-reviewdata/movie.zip	NOUN
cril-239	235	20	.	.	PUNCT
cril-239	236	1	peter	peter	PROPN
cril-239	236	2	turney	turney	PROPN
cril-239	236	3	and	and	CCONJ
cril-239	236	4	michael	michael	PROPN
cril-239	236	5	littman	littman	PROPN
cril-239	236	6	.	.	PUNCT
cril-239	237	1	"	"	PUNCT
cril-239	237	2	measuring	measure	VERB
cril-239	237	3	praise	praise	NOUN
cril-239	237	4	and	and	CCONJ
cril-239	237	5	criticism	criticism	NOUN
cril-239	237	6	:	:	PUNCT
cril-239	237	7	inference	inference	NOUN
cril-239	237	8	of	of	ADP
cril-239	237	9	semantic	semantic	ADJ
cril-239	237	10	orientation	orientation	NOUN
cril-239	237	11	from	from	ADP
cril-239	237	12	association	association	NOUN
cril-239	237	13	"	"	PUNCT
cril-239	237	14	.	.	PUNCT
cril-239	238	1	acm	acm	PROPN
cril-239	238	2	transactions	transaction	NOUN
cril-239	238	3	on	on	ADP
cril-239	238	4	information	information	NOUN
cril-239	238	5	systems	system	NOUN
cril-239	238	6	(	(	PUNCT
cril-239	238	7	tois	tois	NOUN
cril-239	238	8	)	)	PUNCT
cril-239	238	9	,	,	PUNCT
cril-239	238	10	21(4):315–346	21(4):315–346	NUM
cril-239	238	11	,	,	PUNCT
cril-239	238	12	2003	2003	NUM
cril-239	238	13	.	.	PUNCT
cril-239	239	1	peter	peter	PROPN
cril-239	239	2	turney	turney	PROPN
cril-239	239	3	.	.	PUNCT
cril-239	240	1	"	"	PUNCT
cril-239	240	2	thumbs	thumb	VERB
cril-239	240	3	up	up	ADP
cril-239	240	4	or	or	CCONJ
cril-239	240	5	thumbs	thumb	VERB
cril-239	240	6	down	down	ADP
cril-239	240	7	?	?	PUNCT
cril-239	241	1	semantic	semantic	ADJ
cril-239	241	2	orientation	orientation	NOUN
cril-239	241	3	applied	apply	VERB
cril-239	241	4	to	to	ADP
cril-239	241	5	unsupervised	unsupervised	ADJ
cril-239	241	6	classification	classification	NOUN
cril-239	241	7	of	of	ADP
cril-239	241	8	reviews	review	NOUN
cril-239	241	9	"	"	PUNCT
cril-239	241	10	.	.	PUNCT
cril-239	242	1	in	in	ADP
cril-239	242	2	proceedings	proceeding	NOUN
cril-239	242	3	of	of	ADP
cril-239	242	4	the	the	DET
cril-239	242	5	40th	40th	ADJ
cril-239	242	6	annual	annual	ADJ
cril-239	242	7	meeting	meeting	NOUN
cril-239	242	8	of	of	ADP
cril-239	242	9	the	the	DET
cril-239	242	10	association	association	NOUN
cril-239	242	11	for	for	ADP
cril-239	242	12	computational	computational	ADJ
cril-239	242	13	linguistics	linguistic	NOUN
cril-239	242	14	(	(	PUNCT
cril-239	242	15	acl'02	acl'02	NOUN
cril-239	242	16	)	)	PUNCT
cril-239	242	17	,	,	PUNCT
cril-239	242	18	pages	page	NOUN
cril-239	242	19	417–424	417–424	NUM
cril-239	242	20	,	,	PUNCT
cril-239	242	21	2002	2002	NUM
cril-239	242	22	.	.	PUNCT
cril-239	243	1	janyce	janyce	PROPN
cril-239	243	2	wiebe	wiebe	PROPN
cril-239	243	3	,	,	PUNCT
cril-239	243	4	rebecca	rebecca	PROPN
cril-239	243	5	f.	f.	PROPN
cril-239	243	6	bruce	bruce	PROPN
cril-239	243	7	,	,	PUNCT
cril-239	243	8	and	and	CCONJ
cril-239	243	9	thomas	thomas	PROPN
cril-239	243	10	o'hara	o'hara	PROPN
cril-239	243	11	.	.	PUNCT
cril-239	244	1	"	"	PUNCT
cril-239	244	2	development	development	NOUN
cril-239	244	3	and	and	CCONJ
cril-239	244	4	use	use	NOUN
cril-239	244	5	of	of	ADP
cril-239	244	6	a	a	DET
cril-239	244	7	goldstandard	goldstandard	NOUN
cril-239	244	8	data	datum	NOUN
cril-239	244	9	set	set	VERB
cril-239	244	10	for	for	ADP
cril-239	244	11	subjectivity	subjectivity	NOUN
cril-239	244	12	classifications	classification	NOUN
cril-239	244	13	"	"	PUNCT
cril-239	244	14	.	.	PUNCT
cril-239	245	1	in	in	ADP
cril-239	245	2	proc	proc	NOUN
cril-239	245	3	.	.	PUNCT
cril-239	246	1	37th	37th	ADJ
cril-239	246	2	annual	annual	ADJ
cril-239	246	3	meeting	meeting	NOUN
cril-239	246	4	of	of	ADP
cril-239	246	5	the	the	DET
cril-239	246	6	assoc	assoc	NOUN
cril-239	246	7	.	.	PUNCT
cril-239	247	1	for	for	ADP
cril-239	247	2	computational	computational	ADJ
cril-239	247	3	linguistics	linguistic	NOUN
cril-239	247	4	(	(	PUNCT
cril-239	247	5	acl-99	acl-99	NOUN
cril-239	247	6	)	)	PUNCT
cril-239	247	7	,	,	PUNCT
cril-239	247	8	1999	1999	NUM
cril-239	247	9	.	.	PUNCT
cril-239	248	1	janyce	janyce	PROPN
cril-239	248	2	wiebe	wiebe	PROPN
cril-239	248	3	.	.	PUNCT
cril-239	249	1	"	"	PUNCT
cril-239	249	2	learning	learn	VERB
cril-239	249	3	subjective	subjective	ADJ
cril-239	249	4	adjectives	adjective	NOUN
cril-239	249	5	from	from	ADP
cril-239	249	6	corpora	corpora	PROPN
cril-239	249	7	"	"	PUNCT
cril-239	249	8	.	.	PUNCT
cril-239	250	1	in	in	ADP
cril-239	250	2	aaai	aaai	PROPN
cril-239	250	3	/	/	SYM
cril-239	250	4	iaai	iaai	PROPN
cril-239	250	5	,	,	PUNCT
cril-239	250	6	2000	2000	NUM
cril-239	250	7	.	.	PUNCT
cril-239	251	1	15	15	NUM
cril-239	251	2	salvetti	salvetti	VERB
cril-239	251	3	et	et	PROPN
cril-239	251	4	al	al	PROPN
cril-239	251	5	.	.	PROPN
cril-239	251	6	:	:	PUNCT
cril-239	252	1	automatic	automatic	ADJ
cril-239	252	2	opinion	opinion	NOUN
cril-239	252	3	polarity	polarity	NOUN
cril-239	252	4	classification	classification	NOUN
cril-239	252	5	of	of	ADP
cril-239	252	6	movie	movie	NOUN
cril-239	252	7	reviews	review	NOUN
cril-239	252	8	published	publish	VERB
cril-239	252	9	by	by	ADP
cril-239	252	10	cu	cu	PROPN
cril-239	252	11	scholar	scholar	NOUN
cril-239	252	12	,	,	PUNCT
cril-239	252	13	2004	2004	NUM
cril-239	252	14	colorado	colorado	PROPN
cril-239	252	15	research	research	NOUN
cril-239	252	16	in	in	ADP
cril-239	252	17	linguistics	linguistic	NOUN
cril-239	252	18	6	6	NUM
cril-239	252	19	-	-	SYM
cril-239	252	20	2004	2004	NUM
cril-239	252	21	automatic	automatic	ADJ
cril-239	252	22	opinion	opinion	NOUN
cril-239	252	23	polarity	polarity	NOUN
cril-239	252	24	classification	classification	NOUN
cril-239	252	25	of	of	ADP
cril-239	252	26	movie	movie	NOUN
cril-239	252	27	reviews	review	NOUN
cril-239	252	28	franco	franco	PROPN
cril-239	252	29	salvetti	salvetti	PROPN
cril-239	252	30	stephen	stephen	PROPN
cril-239	252	31	lewis	lewis	PROPN
cril-239	252	32	christoph	christoph	PROPN
cril-239	252	33	reichenbach	reichenbach	PROPN
cril-239	252	34	recommended	recommend	VERB
cril-239	252	35	citation	citation	NOUN
cril-239	252	36	microsoft	microsoft	PROPN
cril-239	252	37	word	word	PROPN
cril-239	252	38	paper_salvetti_lewis_reichenbach.doc	paper_salvetti_lewis_reichenbach.doc	NOUN