id	sid	tid	token	lemma	pos
dem-1058	1	1	dynamic	dynamic	ADJ
dem-1058	1	2	econometric	econometric	ADJ
dem-1058	1	3	models	model	NOUN
dem-1058	1	4	vol	vol	NOUN
dem-1058	1	5	.	.	PROPN
dem-1058	1	6	11	11	NUM
dem-1058	1	7	–	–	PUNCT
dem-1058	1	8	nicolaus	nicolaus	PROPN
dem-1058	1	9	copernicus	copernicus	PROPN
dem-1058	1	10	university	university	PROPN
dem-1058	1	11	–	–	PUNCT
dem-1058	1	12	toruń	toruń	NOUN
dem-1058	1	13	–	–	PUNCT
dem-1058	1	14	2011	2011	NUM
dem-1058	1	15	mariola	mariola	PROPN
dem-1058	1	16	piłatowska	piłatowska	PROPN
dem-1058	1	17	nicolaus	nicolaus	PROPN
dem-1058	1	18	copernicus	copernicus	PROPN
dem-1058	1	19	university	university	PROPN
dem-1058	1	20	in	in	ADP
dem-1058	1	21	toruń	toruń	PROPN
dem-1058	1	22	information	information	NOUN
dem-1058	1	23	and	and	CCONJ
dem-1058	1	24	prediction	prediction	NOUN
dem-1058	1	25	criteria	criterion	NOUN
dem-1058	1	26	in	in	ADP
dem-1058	1	27	selecting	select	VERB
dem-1058	1	28	the	the	DET
dem-1058	1	29	forecasting	forecasting	NOUN
dem-1058	1	30	model	model	NOUN
dem-1058	1	31	a	a	DET
dem-1058	1	32	b	b	PROPN
dem-1058	1	33	s	s	ADP
dem-1058	1	34	t	t	PROPN
dem-1058	1	35	r	r	NOUN
dem-1058	1	36	a	a	DET
dem-1058	1	37	c	c	NOUN
dem-1058	1	38	t.	t.	NOUN
dem-1058	1	39	the	the	DET
dem-1058	1	40	purpose	purpose	NOUN
dem-1058	1	41	of	of	ADP
dem-1058	1	42	the	the	DET
dem-1058	1	43	paper	paper	NOUN
dem-1058	1	44	it	it	PRON
dem-1058	1	45	to	to	PART
dem-1058	1	46	compare	compare	VERB
dem-1058	1	47	the	the	DET
dem-1058	1	48	performance	performance	NOUN
dem-1058	1	49	of	of	ADP
dem-1058	1	50	both	both	DET
dem-1058	1	51	information	information	NOUN
dem-1058	1	52	and	and	CCONJ
dem-1058	1	53	prediction	prediction	NOUN
dem-1058	1	54	criteria	criterion	NOUN
dem-1058	1	55	in	in	ADP
dem-1058	1	56	selecting	select	VERB
dem-1058	1	57	the	the	DET
dem-1058	1	58	forecasting	forecasting	NOUN
dem-1058	1	59	model	model	NOUN
dem-1058	1	60	on	on	ADP
dem-1058	1	61	empirical	empirical	ADJ
dem-1058	1	62	data	datum	NOUN
dem-1058	1	63	for	for	ADP
dem-1058	1	64	poland	poland	PROPN
dem-1058	1	65	when	when	SCONJ
dem-1058	1	66	the	the	DET
dem-1058	1	67	data	data	NOUN
dem-1058	1	68	generating	generating	NOUN
dem-1058	1	69	model	model	NOUN
dem-1058	1	70	is	be	AUX
dem-1058	1	71	unknown	unknown	ADJ
dem-1058	1	72	.	.	PUNCT
dem-1058	2	1	the	the	DET
dem-1058	2	2	attention	attention	NOUN
dem-1058	2	3	will	will	AUX
dem-1058	2	4	especially	especially	ADV
dem-1058	2	5	focus	focus	VERB
dem-1058	2	6	on	on	ADP
dem-1058	2	7	the	the	DET
dem-1058	2	8	evolution	evolution	NOUN
dem-1058	2	9	of	of	ADP
dem-1058	2	10	information	information	NOUN
dem-1058	2	11	criteria	criterion	NOUN
dem-1058	2	12	(	(	PUNCT
dem-1058	2	13	aic	aic	PROPN
dem-1058	2	14	,	,	PUNCT
dem-1058	2	15	bic	bic	PROPN
dem-1058	2	16	)	)	PUNCT
dem-1058	2	17	and	and	CCONJ
dem-1058	2	18	accumulated	accumulate	VERB
dem-1058	2	19	prediction	prediction	NOUN
dem-1058	2	20	error	error	NOUN
dem-1058	2	21	(	(	PUNCT
dem-1058	2	22	ape	ape	NOUN
dem-1058	2	23	)	)	PUNCT
dem-1058	2	24	for	for	ADP
dem-1058	2	25	increasing	increase	VERB
dem-1058	2	26	sample	sample	NOUN
dem-1058	2	27	sizes	size	NOUN
dem-1058	2	28	and	and	CCONJ
dem-1058	2	29	rolling	roll	VERB
dem-1058	2	30	windows	window	NOUN
dem-1058	2	31	of	of	ADP
dem-1058	2	32	different	different	ADJ
dem-1058	2	33	size	size	NOUN
dem-1058	2	34	,	,	PUNCT
dem-1058	2	35	and	and	CCONJ
dem-1058	2	36	also	also	ADV
dem-1058	2	37	the	the	DET
dem-1058	2	38	impact	impact	NOUN
dem-1058	2	39	of	of	ADP
dem-1058	2	40	initial	initial	ADJ
dem-1058	2	41	sample	sample	NOUN
dem-1058	2	42	and	and	CCONJ
dem-1058	2	43	rolling	rolling	ADJ
dem-1058	2	44	window	window	NOUN
dem-1058	2	45	sizes	size	NOUN
dem-1058	2	46	on	on	ADP
dem-1058	2	47	the	the	DET
dem-1058	2	48	selection	selection	NOUN
dem-1058	2	49	of	of	ADP
dem-1058	2	50	forecasting	forecasting	NOUN
dem-1058	2	51	model	model	NOUN
dem-1058	2	52	.	.	PUNCT
dem-1058	3	1	the	the	DET
dem-1058	3	2	best	good	ADJ
dem-1058	3	3	forecasting	forecasting	NOUN
dem-1058	3	4	model	model	NOUN
dem-1058	3	5	will	will	AUX
dem-1058	3	6	be	be	AUX
dem-1058	3	7	chosen	choose	VERB
dem-1058	3	8	from	from	ADP
dem-1058	3	9	the	the	DET
dem-1058	3	10	set	set	NOUN
dem-1058	3	11	including	include	VERB
dem-1058	3	12	three	three	NUM
dem-1058	3	13	models	model	NOUN
dem-1058	3	14	:	:	PUNCT
dem-1058	3	15	autoregressive	autoregressive	ADJ
dem-1058	3	16	model	model	NOUN
dem-1058	3	17	,	,	PUNCT
dem-1058	3	18	ar	ar	PROPN
dem-1058	3	19	(	(	PUNCT
dem-1058	3	20	with	with	ADP
dem-1058	3	21	or	or	CCONJ
dem-1058	3	22	without	without	ADP
dem-1058	3	23	a	a	DET
dem-1058	3	24	deterministic	deterministic	ADJ
dem-1058	3	25	trend	trend	NOUN
dem-1058	3	26	)	)	PUNCT
dem-1058	3	27	,	,	PUNCT
dem-1058	3	28	arima	arima	NOUN
dem-1058	3	29	model	model	NOUN
dem-1058	3	30	and	and	CCONJ
dem-1058	3	31	random	random	ADJ
dem-1058	3	32	walk	walk	NOUN
dem-1058	3	33	(	(	PUNCT
dem-1058	3	34	rw	rw	NOUN
dem-1058	3	35	)	)	PUNCT
dem-1058	3	36	model	model	NOUN
dem-1058	3	37	.	.	PUNCT
dem-1058	4	1	k	k	PROPN
dem-1058	4	2	e	e	PROPN
dem-1058	4	3	y	y	PROPN
dem-1058	4	4	w	w	NOUN
dem-1058	4	5	o	o	NOUN
dem-1058	4	6	r	r	NOUN
dem-1058	4	7	d	d	X
dem-1058	4	8	s	s	X
dem-1058	4	9	:	:	PUNCT
dem-1058	4	10	information	information	NOUN
dem-1058	4	11	and	and	CCONJ
dem-1058	4	12	prediction	prediction	NOUN
dem-1058	4	13	criteria	criterion	NOUN
dem-1058	4	14	,	,	PUNCT
dem-1058	4	15	accumulated	accumulate	VERB
dem-1058	4	16	prediction	prediction	NOUN
dem-1058	4	17	error	error	NOUN
dem-1058	4	18	,	,	PUNCT
dem-1058	4	19	model	model	NOUN
dem-1058	4	20	selection	selection	NOUN
dem-1058	4	21	.	.	PUNCT
dem-1058	5	1	introduction	introduction	NOUN
dem-1058	5	2	the	the	DET
dem-1058	5	3	model	model	NOUN
dem-1058	5	4	selection	selection	NOUN
dem-1058	5	5	literature	literature	NOUN
dem-1058	5	6	has	have	AUX
dem-1058	5	7	recently	recently	ADV
dem-1058	5	8	emphasized	emphasize	VERB
dem-1058	5	9	the	the	DET
dem-1058	5	10	necessity	necessity	NOUN
dem-1058	5	11	of	of	ADP
dem-1058	5	12	considering	consider	VERB
dem-1058	5	13	the	the	DET
dem-1058	5	14	choice	choice	NOUN
dem-1058	5	15	of	of	ADP
dem-1058	5	16	model	model	NOUN
dem-1058	5	17	depending	depend	VERB
dem-1058	5	18	on	on	ADP
dem-1058	5	19	the	the	DET
dem-1058	5	20	purpose	purpose	NOUN
dem-1058	5	21	of	of	ADP
dem-1058	5	22	econometric	econometric	ADJ
dem-1058	5	23	modeling	modeling	NOUN
dem-1058	5	24	.	.	PUNCT
dem-1058	6	1	in	in	ADP
dem-1058	6	2	modeling	model	VERB
dem-1058	6	3	approach	approach	NOUN
dem-1058	6	4	the	the	DET
dem-1058	6	5	two	two	NUM
dem-1058	6	6	aims	aim	NOUN
dem-1058	6	7	are	be	AUX
dem-1058	6	8	mentioned	mention	VERB
dem-1058	6	9	the	the	DET
dem-1058	6	10	most	most	ADV
dem-1058	6	11	frequently	frequently	ADV
dem-1058	6	12	,	,	PUNCT
dem-1058	6	13	namely	namely	ADV
dem-1058	6	14	searching	search	VERB
dem-1058	6	15	'	'	PUNCT
dem-1058	6	16	true	true	ADJ
dem-1058	6	17	'	'	PUNCT
dem-1058	6	18	model	model	NOUN
dem-1058	6	19	and	and	CCONJ
dem-1058	6	20	selecting	select	VERB
dem-1058	6	21	the	the	DET
dem-1058	6	22	best	good	ADJ
dem-1058	6	23	forecasting	forecasting	NOUN
dem-1058	6	24	model	model	NOUN
dem-1058	6	25	(	(	PUNCT
dem-1058	6	26	optimizing	optimize	VERB
dem-1058	6	27	prediction	prediction	NOUN
dem-1058	6	28	)	)	PUNCT
dem-1058	6	29	.	.	PUNCT
dem-1058	7	1	that	that	DET
dem-1058	7	2	first	first	ADJ
dem-1058	7	3	aim	aim	NOUN
dem-1058	7	4	of	of	ADP
dem-1058	7	5	modeling	modeling	NOUN
dem-1058	7	6	is	be	AUX
dem-1058	7	7	hard	hard	ADJ
dem-1058	7	8	to	to	PART
dem-1058	7	9	realize	realize	VERB
dem-1058	7	10	because	because	SCONJ
dem-1058	7	11	the	the	DET
dem-1058	7	12	economic	economic	ADJ
dem-1058	7	13	reality	reality	NOUN
dem-1058	7	14	is	be	AUX
dem-1058	7	15	seen	see	VERB
dem-1058	7	16	as	as	ADP
dem-1058	7	17	a	a	DET
dem-1058	7	18	complex	complex	ADJ
dem-1058	7	19	and	and	CCONJ
dem-1058	7	20	dynamically	dynamically	ADV
dem-1058	7	21	evolving	evolve	VERB
dem-1058	7	22	structure	structure	NOUN
dem-1058	7	23	whose	whose	DET
dem-1058	7	24	mechanism	mechanism	NOUN
dem-1058	7	25	is	be	AUX
dem-1058	7	26	hidden	hide	VERB
dem-1058	7	27	and	and	CCONJ
dem-1058	7	28	almost	almost	ADV
dem-1058	7	29	impossible	impossible	ADJ
dem-1058	7	30	to	to	PART
dem-1058	7	31	uncover	uncover	VERB
dem-1058	7	32	.	.	PUNCT
dem-1058	8	1	therefore	therefore	ADV
dem-1058	8	2	the	the	DET
dem-1058	8	3	model	model	NOUN
dem-1058	8	4	is	be	AUX
dem-1058	8	5	an	an	DET
dem-1058	8	6	approximation	approximation	NOUN
dem-1058	8	7	(	(	PUNCT
dem-1058	8	8	or	or	CCONJ
dem-1058	8	9	simplification	simplification	NOUN
dem-1058	8	10	)	)	PUNCT
dem-1058	8	11	of	of	ADP
dem-1058	8	12	reality	reality	NOUN
dem-1058	8	13	which	which	PRON
dem-1058	8	14	represents	represent	VERB
dem-1058	8	15	the	the	DET
dem-1058	8	16	relevance	relevance	NOUN
dem-1058	8	17	of	of	ADP
dem-1058	8	18	a	a	DET
dem-1058	8	19	particular	particular	ADJ
dem-1058	8	20	phenomenon	phenomenon	NOUN
dem-1058	8	21	.	.	PUNCT
dem-1058	9	1	it	it	PRON
dem-1058	9	2	is	be	AUX
dem-1058	9	3	advocated	advocate	VERB
dem-1058	9	4	to	to	PART
dem-1058	9	5	assume	assume	VERB
dem-1058	9	6	that	that	SCONJ
dem-1058	9	7	each	each	DET
dem-1058	9	8	model	model	NOUN
dem-1058	9	9	is	be	AUX
dem-1058	9	10	not	not	PART
dem-1058	9	11	true	true	ADJ
dem-1058	9	12	by	by	ADP
dem-1058	9	13	definition	definition	NOUN
dem-1058	9	14	(	(	PUNCT
dem-1058	9	15	taub	taub	PROPN
dem-1058	9	16	,	,	PUNCT
dem-1058	9	17	1993	1993	NUM
dem-1058	9	18	;	;	PUNCT
dem-1058	9	19	deleew	deleew	VERB
dem-1058	9	20	,	,	PUNCT
dem-1058	9	21	1998	1998	NUM
dem-1058	9	22	)	)	PUNCT
dem-1058	9	23	or	or	CCONJ
dem-1058	9	24	that	that	SCONJ
dem-1058	9	25	"	"	PUNCT
dem-1058	9	26	all	all	DET
dem-1058	9	27	models	model	NOUN
dem-1058	9	28	are	be	AUX
dem-1058	9	29	wrong	wrong	ADJ
dem-1058	9	30	,	,	PUNCT
dem-1058	9	31	but	but	CCONJ
dem-1058	9	32	some	some	PRON
dem-1058	9	33	are	be	AUX
dem-1058	9	34	useful	useful	ADJ
dem-1058	9	35	"	"	PUNCT
dem-1058	9	36	(	(	PUNCT
dem-1058	9	37	box	box	NOUN
dem-1058	9	38	,	,	PUNCT
dem-1058	9	39	1976	1976	NUM
dem-1058	9	40	)	)	PUNCT
dem-1058	9	41	.	.	PUNCT
dem-1058	10	1	having	have	VERB
dem-1058	10	2	in	in	ADP
dem-1058	10	3	mind	mind	NOUN
dem-1058	10	4	that	that	SCONJ
dem-1058	10	5	none	none	NOUN
dem-1058	10	6	of	of	ADP
dem-1058	10	7	models	model	NOUN
dem-1058	10	8	can	can	AUX
dem-1058	10	9	not	not	PART
dem-1058	10	10	reflect	reflect	VERB
dem-1058	10	11	all	all	PRON
dem-1058	10	12	of	of	ADP
dem-1058	10	13	reality	reality	NOUN
dem-1058	10	14	,	,	PUNCT
dem-1058	10	15	the	the	DET
dem-1058	10	16	debate	debate	NOUN
dem-1058	10	17	concerning	concern	VERB
dem-1058	10	18	true	true	ADJ
dem-1058	10	19	models	model	NOUN
dem-1058	10	20	should	should	AUX
dem-1058	10	21	be	be	AUX
dem-1058	10	22	completed	complete	VERB
dem-1058	10	23	because	because	SCONJ
dem-1058	10	24	it	it	PRON
dem-1058	10	25	seems	seem	VERB
dem-1058	10	26	to	to	PART
dem-1058	10	27	be	be	AUX
dem-1058	10	28	unproductive	unproductive	ADJ
dem-1058	10	29	.	.	PUNCT
dem-1058	11	1	hence	hence	ADV
dem-1058	11	2	,	,	PUNCT
dem-1058	11	3	the	the	DET
dem-1058	11	4	second	second	ADJ
dem-1058	11	5	aim	aim	NOUN
dem-1058	11	6	of	of	ADP
dem-1058	11	7	modeling	modeling	NOUN
dem-1058	11	8	,	,	PUNCT
dem-1058	11	9	i.e.	i.e.	X
dem-1058	11	10	selecting	select	VERB
dem-1058	11	11	the	the	DET
dem-1058	11	12	best	good	ADJ
dem-1058	11	13	forecasting	forecasting	NOUN
dem-1058	11	14	model	model	NOUN
dem-1058	11	15	acquires	acquire	VERB
dem-1058	11	16	relevance	relevance	NOUN
dem-1058	11	17	from	from	ADP
dem-1058	11	18	the	the	DET
dem-1058	11	19	practical	practical	ADJ
dem-1058	11	20	point	point	NOUN
dem-1058	11	21	of	of	ADP
dem-1058	11	22	view	view	NOUN
dem-1058	11	23	.	.	PUNCT
dem-1058	12	1	mariola	mariola	PROPN
dem-1058	12	2	piłatowska	piłatowska	PROPN
dem-1058	12	3	22	22	NUM
dem-1058	12	4	in	in	ADP
dem-1058	12	5	predictive	predictive	ADJ
dem-1058	12	6	approach	approach	NOUN
dem-1058	12	7	the	the	DET
dem-1058	12	8	goal	goal	NOUN
dem-1058	12	9	of	of	ADP
dem-1058	12	10	selecting	select	VERB
dem-1058	12	11	the	the	DET
dem-1058	12	12	true	true	ADJ
dem-1058	12	13	model	model	NOUN
dem-1058	12	14	is	be	AUX
dem-1058	12	15	abandoned	abandon	VERB
dem-1058	12	16	and	and	CCONJ
dem-1058	12	17	the	the	DET
dem-1058	12	18	attention	attention	NOUN
dem-1058	12	19	focuses	focus	VERB
dem-1058	12	20	on	on	ADP
dem-1058	12	21	seeking	seek	VERB
dem-1058	12	22	a	a	DET
dem-1058	12	23	model	model	NOUN
dem-1058	12	24	with	with	ADP
dem-1058	12	25	as	as	ADP
dem-1058	12	26	small	small	ADJ
dem-1058	12	27	predictive	predictive	ADJ
dem-1058	12	28	errors	error	NOUN
dem-1058	12	29	as	as	ADP
dem-1058	12	30	possible	possible	ADJ
dem-1058	12	31	.	.	PUNCT
dem-1058	13	1	it	it	PRON
dem-1058	13	2	should	should	AUX
dem-1058	13	3	be	be	AUX
dem-1058	13	4	emphasized	emphasize	VERB
dem-1058	13	5	that	that	SCONJ
dem-1058	13	6	in	in	ADP
dem-1058	13	7	forecasting	forecast	VERB
dem-1058	13	8	situation	situation	NOUN
dem-1058	13	9	the	the	DET
dem-1058	13	10	misspecified	misspecifie	VERB
dem-1058	13	11	model	model	NOUN
dem-1058	13	12	are	be	AUX
dem-1058	13	13	allowed	allow	VERB
dem-1058	13	14	because	because	SCONJ
dem-1058	13	15	such	such	DET
dem-1058	13	16	a	a	DET
dem-1058	13	17	model	model	NOUN
dem-1058	13	18	may	may	AUX
dem-1058	13	19	yield	yield	VERB
dem-1058	13	20	excellent	excellent	ADJ
dem-1058	13	21	forecasts	forecast	NOUN
dem-1058	13	22	.	.	PUNCT
dem-1058	14	1	on	on	ADP
dem-1058	14	2	the	the	DET
dem-1058	14	3	other	other	ADJ
dem-1058	14	4	hand	hand	NOUN
dem-1058	14	5	good	good	ADJ
dem-1058	14	6	prediction	prediction	NOUN
dem-1058	14	7	is	be	AUX
dem-1058	14	8	treated	treat	VERB
dem-1058	14	9	as	as	ADP
dem-1058	14	10	a	a	DET
dem-1058	14	11	test	test	NOUN
dem-1058	14	12	of	of	ADP
dem-1058	14	13	any	any	DET
dem-1058	14	14	subsidiary	subsidiary	NOUN
dem-1058	14	15	aim	aim	NOUN
dem-1058	14	16	,	,	PUNCT
dem-1058	14	17	i.e.	i.e.	X
dem-1058	14	18	if	if	SCONJ
dem-1058	14	19	the	the	DET
dem-1058	14	20	purpose	purpose	NOUN
dem-1058	14	21	of	of	ADP
dem-1058	14	22	an	an	DET
dem-1058	14	23	analysis	analysis	NOUN
dem-1058	14	24	is	be	AUX
dem-1058	14	25	to	to	PART
dem-1058	14	26	estimate	estimate	VERB
dem-1058	14	27	parameters	parameter	NOUN
dem-1058	14	28	,	,	PUNCT
dem-1058	14	29	then	then	ADV
dem-1058	14	30	the	the	DET
dem-1058	14	31	best	well	ADV
dem-1058	14	32	estimated	estimate	VERB
dem-1058	14	33	model	model	NOUN
dem-1058	14	34	should	should	AUX
dem-1058	14	35	give	give	VERB
dem-1058	14	36	the	the	DET
dem-1058	14	37	best	good	ADJ
dem-1058	14	38	prediction	prediction	NOUN
dem-1058	14	39	;	;	PUNCT
dem-1058	14	40	if	if	SCONJ
dem-1058	14	41	the	the	DET
dem-1058	14	42	purpose	purpose	NOUN
dem-1058	14	43	of	of	ADP
dem-1058	14	44	an	an	DET
dem-1058	14	45	analysis	analysis	NOUN
dem-1058	14	46	is	be	AUX
dem-1058	14	47	hypothesis	hypothesis	NOUN
dem-1058	14	48	testing	testing	NOUN
dem-1058	14	49	,	,	PUNCT
dem-1058	14	50	then	then	ADV
dem-1058	14	51	any	any	DET
dem-1058	14	52	rejected	reject	VERB
dem-1058	14	53	model	model	NOUN
dem-1058	14	54	should	should	AUX
dem-1058	14	55	give	give	VERB
dem-1058	14	56	worse	bad	ADJ
dem-1058	14	57	forecasts	forecast	NOUN
dem-1058	14	58	than	than	ADP
dem-1058	14	59	any	any	DET
dem-1058	14	60	accepted	accepted	ADJ
dem-1058	14	61	model	model	NOUN
dem-1058	14	62	(	(	PUNCT
dem-1058	14	63	clarke	clarke	PROPN
dem-1058	14	64	,	,	PUNCT
dem-1058	14	65	2001	2001	NUM
dem-1058	14	66	;	;	PUNCT
dem-1058	14	67	de	de	PROPN
dem-1058	14	68	luna	luna	PROPN
dem-1058	14	69	,	,	PUNCT
dem-1058	14	70	skouras	skoura	NOUN
dem-1058	14	71	,	,	PUNCT
dem-1058	14	72	2003	2003	NUM
dem-1058	14	73	;	;	PUNCT
dem-1058	14	74	kunst	kunst	PROPN
dem-1058	14	75	,	,	PUNCT
dem-1058	14	76	2003	2003	NUM
dem-1058	14	77	)	)	PUNCT
dem-1058	14	78	.	.	PUNCT
dem-1058	15	1	to	to	PART
dem-1058	15	2	select	select	VERB
dem-1058	15	3	the	the	DET
dem-1058	15	4	forecasting	forecasting	NOUN
dem-1058	15	5	model	model	NOUN
dem-1058	15	6	the	the	DET
dem-1058	15	7	different	different	ADJ
dem-1058	15	8	model	model	NOUN
dem-1058	15	9	selection	selection	NOUN
dem-1058	15	10	methods	method	NOUN
dem-1058	15	11	can	can	AUX
dem-1058	15	12	be	be	AUX
dem-1058	15	13	used	use	VERB
dem-1058	15	14	,	,	PUNCT
dem-1058	15	15	for	for	ADP
dem-1058	15	16	instance	instance	NOUN
dem-1058	15	17	information	information	NOUN
dem-1058	15	18	and	and	CCONJ
dem-1058	15	19	prediction	prediction	NOUN
dem-1058	15	20	criteria	criterion	NOUN
dem-1058	15	21	.	.	PUNCT
dem-1058	16	1	however	however	ADV
dem-1058	16	2	,	,	PUNCT
dem-1058	16	3	question	question	NOUN
dem-1058	16	4	may	may	AUX
dem-1058	16	5	arise	arise	VERB
dem-1058	16	6	whether	whether	SCONJ
dem-1058	16	7	the	the	DET
dem-1058	16	8	performance	performance	NOUN
dem-1058	16	9	of	of	ADP
dem-1058	16	10	both	both	DET
dem-1058	16	11	criteria	criterion	NOUN
dem-1058	16	12	is	be	AUX
dem-1058	16	13	the	the	DET
dem-1058	16	14	same	same	ADJ
dem-1058	16	15	with	with	ADP
dem-1058	16	16	regard	regard	NOUN
dem-1058	16	17	to	to	ADP
dem-1058	16	18	the	the	DET
dem-1058	16	19	choice	choice	NOUN
dem-1058	16	20	of	of	ADP
dem-1058	16	21	model	model	NOUN
dem-1058	16	22	.	.	PUNCT
dem-1058	17	1	it	it	PRON
dem-1058	17	2	has	have	AUX
dem-1058	17	3	been	be	AUX
dem-1058	17	4	shown	show	VERB
dem-1058	17	5	on	on	ADP
dem-1058	17	6	simulated	simulated	ADJ
dem-1058	17	7	data1	data1	PROPN
dem-1058	17	8	(	(	PUNCT
dem-1058	17	9	kunst	kunst	PROPN
dem-1058	17	10	,	,	PUNCT
dem-1058	17	11	2003	2003	NUM
dem-1058	17	12	)	)	PUNCT
dem-1058	17	13	that	that	SCONJ
dem-1058	17	14	information	information	NOUN
dem-1058	17	15	criteria	criterion	NOUN
dem-1058	17	16	should	should	AUX
dem-1058	17	17	be	be	AUX
dem-1058	17	18	rather	rather	ADV
dem-1058	17	19	used	use	VERB
dem-1058	17	20	if	if	SCONJ
dem-1058	17	21	one	one	PRON
dem-1058	17	22	is	be	AUX
dem-1058	17	23	interested	interested	ADJ
dem-1058	17	24	in	in	ADP
dem-1058	17	25	finding	find	VERB
dem-1058	17	26	'	'	PUNCT
dem-1058	17	27	true	true	ADJ
dem-1058	17	28	'	'	PUNCT
dem-1058	17	29	model	model	NOUN
dem-1058	17	30	.	.	PUNCT
dem-1058	18	1	if	if	SCONJ
dem-1058	18	2	the	the	DET
dem-1058	18	3	purpose	purpose	NOUN
dem-1058	18	4	of	of	ADP
dem-1058	18	5	analysis	analysis	NOUN
dem-1058	18	6	is	be	AUX
dem-1058	18	7	to	to	PART
dem-1058	18	8	choose	choose	VERB
dem-1058	18	9	a	a	DET
dem-1058	18	10	forecasting	forecasting	NOUN
dem-1058	18	11	model	model	NOUN
dem-1058	18	12	,	,	PUNCT
dem-1058	18	13	the	the	DET
dem-1058	18	14	prediction	prediction	NOUN
dem-1058	18	15	criteria	criterion	NOUN
dem-1058	18	16	are	be	AUX
dem-1058	18	17	preferred	prefer	VERB
dem-1058	18	18	because	because	SCONJ
dem-1058	18	19	they	they	PRON
dem-1058	18	20	select	select	VERB
dem-1058	18	21	the	the	DET
dem-1058	18	22	model	model	NOUN
dem-1058	18	23	yielding	yield	VERB
dem-1058	18	24	the	the	DET
dem-1058	18	25	smallest	small	ADJ
dem-1058	18	26	prediction	prediction	NOUN
dem-1058	18	27	error	error	NOUN
dem-1058	18	28	,	,	PUNCT
dem-1058	18	29	although	although	SCONJ
dem-1058	18	30	sometimes	sometimes	ADV
dem-1058	18	31	it	it	PRON
dem-1058	18	32	may	may	AUX
dem-1058	18	33	be	be	AUX
dem-1058	18	34	an	an	DET
dem-1058	18	35	incorrect	incorrect	ADJ
dem-1058	18	36	choice	choice	NOUN
dem-1058	18	37	(	(	PUNCT
dem-1058	18	38	not	not	PART
dem-1058	18	39	true	true	ADJ
dem-1058	18	40	model	model	NOUN
dem-1058	18	41	)	)	PUNCT
dem-1058	18	42	.	.	PUNCT
dem-1058	19	1	however	however	ADV
dem-1058	19	2	,	,	PUNCT
dem-1058	19	3	in	in	ADP
dem-1058	19	4	economic	economic	ADJ
dem-1058	19	5	reality	reality	NOUN
dem-1058	19	6	the	the	DET
dem-1058	19	7	true	true	ADJ
dem-1058	19	8	model	model	NOUN
dem-1058	19	9	is	be	AUX
dem-1058	19	10	unknown	unknown	ADJ
dem-1058	19	11	,	,	PUNCT
dem-1058	19	12	therefore	therefore	ADV
dem-1058	19	13	it	it	PRON
dem-1058	19	14	is	be	AUX
dem-1058	19	15	worth	worth	ADJ
dem-1058	19	16	checking	check	VERB
dem-1058	19	17	the	the	DET
dem-1058	19	18	performance	performance	NOUN
dem-1058	19	19	of	of	ADP
dem-1058	19	20	information	information	NOUN
dem-1058	19	21	and	and	CCONJ
dem-1058	19	22	predictiion	predictiion	NOUN
dem-1058	19	23	criteria	criterion	NOUN
dem-1058	19	24	in	in	ADP
dem-1058	19	25	practical	practical	ADJ
dem-1058	19	26	context	context	NOUN
dem-1058	19	27	(	(	PUNCT
dem-1058	19	28	empirical	empirical	ADJ
dem-1058	19	29	data	datum	NOUN
dem-1058	19	30	)	)	PUNCT
dem-1058	19	31	.	.	PUNCT
dem-1058	20	1	the	the	DET
dem-1058	20	2	purpose	purpose	NOUN
dem-1058	20	3	of	of	ADP
dem-1058	20	4	the	the	DET
dem-1058	20	5	paper	paper	NOUN
dem-1058	20	6	it	it	PRON
dem-1058	20	7	to	to	PART
dem-1058	20	8	compare	compare	VERB
dem-1058	20	9	the	the	DET
dem-1058	20	10	performance	performance	NOUN
dem-1058	20	11	of	of	ADP
dem-1058	20	12	both	both	DET
dem-1058	20	13	information	information	NOUN
dem-1058	20	14	and	and	CCONJ
dem-1058	20	15	prediction	prediction	NOUN
dem-1058	20	16	criteria	criterion	NOUN
dem-1058	20	17	in	in	ADP
dem-1058	20	18	selecting	select	VERB
dem-1058	20	19	the	the	DET
dem-1058	20	20	forecasting	forecasting	NOUN
dem-1058	20	21	model	model	NOUN
dem-1058	20	22	on	on	ADP
dem-1058	20	23	empirical	empirical	ADJ
dem-1058	20	24	data	datum	NOUN
dem-1058	20	25	when	when	SCONJ
dem-1058	20	26	the	the	DET
dem-1058	20	27	data	data	NOUN
dem-1058	20	28	generating	generating	NOUN
dem-1058	20	29	model	model	NOUN
dem-1058	20	30	is	be	AUX
dem-1058	20	31	unknown	unknown	ADJ
dem-1058	20	32	.	.	PUNCT
dem-1058	21	1	the	the	DET
dem-1058	21	2	attention	attention	NOUN
dem-1058	21	3	will	will	AUX
dem-1058	21	4	especially	especially	ADV
dem-1058	21	5	focus	focus	VERB
dem-1058	21	6	on	on	ADP
dem-1058	21	7	the	the	DET
dem-1058	21	8	evolution	evolution	NOUN
dem-1058	21	9	of	of	ADP
dem-1058	21	10	information	information	NOUN
dem-1058	21	11	criteria	criterion	NOUN
dem-1058	21	12	(	(	PUNCT
dem-1058	21	13	aic	aic	PROPN
dem-1058	21	14	,	,	PUNCT
dem-1058	21	15	bic	bic	PROPN
dem-1058	21	16	)	)	PUNCT
dem-1058	21	17	and	and	CCONJ
dem-1058	21	18	accumulated	accumulate	VERB
dem-1058	21	19	prediction	prediction	NOUN
dem-1058	21	20	error	error	NOUN
dem-1058	21	21	(	(	PUNCT
dem-1058	21	22	ape	ape	NOUN
dem-1058	21	23	)	)	PUNCT
dem-1058	21	24	for	for	ADP
dem-1058	21	25	increasing	increase	VERB
dem-1058	21	26	sample	sample	NOUN
dem-1058	21	27	sizes	size	NOUN
dem-1058	21	28	and	and	CCONJ
dem-1058	21	29	rolling	roll	VERB
dem-1058	21	30	windows	window	NOUN
dem-1058	21	31	of	of	ADP
dem-1058	21	32	different	different	ADJ
dem-1058	21	33	size	size	NOUN
dem-1058	21	34	,	,	PUNCT
dem-1058	21	35	and	and	CCONJ
dem-1058	21	36	also	also	ADV
dem-1058	21	37	the	the	DET
dem-1058	21	38	impact	impact	NOUN
dem-1058	21	39	of	of	ADP
dem-1058	21	40	sample	sample	NOUN
dem-1058	21	41	and	and	CCONJ
dem-1058	21	42	rolling	rolling	ADJ
dem-1058	21	43	window	window	NOUN
dem-1058	21	44	sizes	size	NOUN
dem-1058	21	45	on	on	ADP
dem-1058	21	46	the	the	DET
dem-1058	21	47	selection	selection	NOUN
dem-1058	21	48	of	of	ADP
dem-1058	21	49	forecasting	forecasting	NOUN
dem-1058	21	50	model	model	NOUN
dem-1058	21	51	.	.	PUNCT
dem-1058	22	1	the	the	DET
dem-1058	22	2	best	good	ADJ
dem-1058	22	3	forecasting	forecasting	NOUN
dem-1058	22	4	model	model	NOUN
dem-1058	22	5	will	will	AUX
dem-1058	22	6	be	be	AUX
dem-1058	22	7	chosen	choose	VERB
dem-1058	22	8	from	from	ADP
dem-1058	22	9	the	the	DET
dem-1058	22	10	set	set	NOUN
dem-1058	22	11	including	include	VERB
dem-1058	22	12	three	three	NUM
dem-1058	22	13	models	model	NOUN
dem-1058	22	14	:	:	PUNCT
dem-1058	22	15	autoregressive	autoregressive	ADJ
dem-1058	22	16	model	model	NOUN
dem-1058	22	17	,	,	PUNCT
dem-1058	22	18	ar	ar	PROPN
dem-1058	22	19	(	(	PUNCT
dem-1058	22	20	with	with	ADP
dem-1058	22	21	or	or	CCONJ
dem-1058	22	22	without	without	ADP
dem-1058	22	23	a	a	DET
dem-1058	22	24	deterministic	deterministic	ADJ
dem-1058	22	25	trend	trend	NOUN
dem-1058	22	26	)	)	PUNCT
dem-1058	22	27	,	,	PUNCT
dem-1058	22	28	arima	arima	NOUN
dem-1058	22	29	model	model	NOUN
dem-1058	22	30	and	and	CCONJ
dem-1058	22	31	random	random	ADJ
dem-1058	22	32	walk	walk	NOUN
dem-1058	22	33	(	(	PUNCT
dem-1058	22	34	rw	rw	NOUN
dem-1058	22	35	)	)	PUNCT
dem-1058	22	36	model	model	NOUN
dem-1058	22	37	on	on	ADP
dem-1058	22	38	the	the	DET
dem-1058	22	39	basis	basis	NOUN
dem-1058	22	40	of	of	ADP
dem-1058	22	41	empirical	empirical	ADJ
dem-1058	22	42	data	datum	NOUN
dem-1058	22	43	for	for	ADP
dem-1058	22	44	poland	poland	PROPN
dem-1058	22	45	.	.	PUNCT
dem-1058	23	1	the	the	DET
dem-1058	23	2	choice	choice	NOUN
dem-1058	23	3	of	of	ADP
dem-1058	23	4	model	model	NOUN
dem-1058	23	5	will	will	AUX
dem-1058	23	6	be	be	AUX
dem-1058	23	7	carried	carry	VERB
dem-1058	23	8	out	out	ADP
dem-1058	23	9	using	use	VERB
dem-1058	23	10	information	information	NOUN
dem-1058	23	11	(	(	PUNCT
dem-1058	23	12	aic	aic	PROPN
dem-1058	23	13	,	,	PUNCT
dem-1058	23	14	bic	bic	PROPN
dem-1058	23	15	)	)	PUNCT
dem-1058	23	16	and	and	CCONJ
dem-1058	23	17	prediction	prediction	NOUN
dem-1058	23	18	(	(	PUNCT
dem-1058	23	19	ape	ape	NOUN
dem-1058	23	20	and	and	CCONJ
dem-1058	23	21	mse	mse	NOUN
dem-1058	23	22	,	,	PUNCT
dem-1058	23	23	mape	mape	NOUN
dem-1058	23	24	,	,	PUNCT
dem-1058	23	25	u	u	NOUN
dem-1058	23	26	)	)	PUNCT
dem-1058	23	27	criteria	criterion	NOUN
dem-1058	23	28	.	.	PUNCT
dem-1058	24	1	the	the	DET
dem-1058	24	2	decision	decision	NOUN
dem-1058	24	3	of	of	ADP
dem-1058	24	4	selecting	select	VERB
dem-1058	24	5	a	a	DET
dem-1058	24	6	model	model	NOUN
dem-1058	24	7	by	by	ADP
dem-1058	24	8	information	information	NOUN
dem-1058	24	9	and	and	CCONJ
dem-1058	24	10	prediction	prediction	NOUN
dem-1058	24	11	criteria	criterion	NOUN
dem-1058	24	12	is	be	AUX
dem-1058	24	13	checked	check	VERB
dem-1058	24	14	in	in	ADP
dem-1058	24	15	out	out	ADP
dem-1058	24	16	-	-	PUNCT
dem-1058	24	17	of	of	ADP
dem-1058	24	18	-	-	PUNCT
dem-1058	24	19	sample	sample	NOUN
dem-1058	24	20	forecasting	forecasting	NOUN
dem-1058	24	21	by	by	ADP
dem-1058	24	22	comparing	compare	VERB
dem-1058	24	23	accuracy	accuracy	NOUN
dem-1058	24	24	measures	measure	NOUN
dem-1058	24	25	for	for	ADP
dem-1058	24	26	given	give	VERB
dem-1058	24	27	forecast	forecast	NOUN
dem-1058	24	28	models	model	NOUN
dem-1058	24	29	.	.	PUNCT
dem-1058	25	1	1	1	NUM
dem-1058	25	2	the	the	DET
dem-1058	25	3	true	true	ADJ
dem-1058	25	4	data	datum	NOUN
dem-1058	25	5	were	be	AUX
dem-1058	25	6	generated	generate	VERB
dem-1058	25	7	from	from	ADP
dem-1058	25	8	arma(1	arma(1	ADJ
dem-1058	25	9	,	,	PUNCT
dem-1058	25	10	1	1	X
dem-1058	25	11	)	)	PUNCT
dem-1058	25	12	models	model	NOUN
dem-1058	25	13	with	with	ADP
dem-1058	25	14	100	100	NUM
dem-1058	25	15	+	+	SYM
dem-1058	25	16	100	100	NUM
dem-1058	25	17	+	+	SYM
dem-1058	25	18	10	10	NUM
dem-1058	25	19	observations	observation	NOUN
dem-1058	25	20	(	(	PUNCT
dem-1058	25	21	first	first	ADJ
dem-1058	25	22	100	100	NUM
dem-1058	25	23	observations	observation	NOUN
dem-1058	25	24	were	be	AUX
dem-1058	25	25	discarded	discard	VERB
dem-1058	25	26	;	;	PUNCT
dem-1058	25	27	1000	1000	NUM
dem-1058	25	28	replication	replication	NOUN
dem-1058	25	29	were	be	AUX
dem-1058	25	30	conducted	conduct	VERB
dem-1058	25	31	)	)	PUNCT
dem-1058	25	32	.	.	PUNCT
dem-1058	26	1	the	the	DET
dem-1058	26	2	set	set	NOUN
dem-1058	26	3	of	of	ADP
dem-1058	26	4	models	model	NOUN
dem-1058	26	5	included	include	VERB
dem-1058	26	6	:	:	PUNCT
dem-1058	26	7	arma(1	arma(1	NOUN
dem-1058	26	8	,	,	PUNCT
dem-1058	26	9	1	1	NUM
dem-1058	26	10	)	)	PUNCT
dem-1058	26	11	,	,	PUNCT
dem-1058	26	12	ar(1	ar(1	PROPN
dem-1058	26	13	)	)	PUNCT
dem-1058	26	14	and	and	CCONJ
dem-1058	26	15	ma(1	ma(1	NOUN
dem-1058	26	16	)	)	PUNCT
dem-1058	26	17	models	model	NOUN
dem-1058	26	18	.	.	PUNCT
dem-1058	27	1	to	to	PART
dem-1058	27	2	select	select	VERB
dem-1058	27	3	a	a	DET
dem-1058	27	4	model	model	NOUN
dem-1058	27	5	the	the	DET
dem-1058	27	6	aic	aic	PROPN
dem-1058	27	7	information	information	NOUN
dem-1058	27	8	criterion	criterion	NOUN
dem-1058	27	9	and	and	CCONJ
dem-1058	27	10	the	the	DET
dem-1058	27	11	mean	mean	ADJ
dem-1058	27	12	squared	square	VERB
dem-1058	27	13	error	error	NOUN
dem-1058	27	14	(	(	PUNCT
dem-1058	27	15	mse	mse	NOUN
dem-1058	27	16	)	)	PUNCT
dem-1058	27	17	based	base	VERB
dem-1058	27	18	on	on	ADP
dem-1058	27	19	prediction	prediction	NOUN
dem-1058	27	20	error	error	NOUN
dem-1058	27	21	from	from	ADP
dem-1058	27	22	10	10	NUM
dem-1058	27	23	one	one	NUM
dem-1058	27	24	-	-	PUNCT
dem-1058	27	25	step	step	NOUN
dem-1058	27	26	-	-	PUNCT
dem-1058	27	27	ahead	ahead	NOUN
dem-1058	27	28	forecasts	forecast	NOUN
dem-1058	27	29	were	be	AUX
dem-1058	27	30	applied	apply	VERB
dem-1058	27	31	(	(	PUNCT
dem-1058	27	32	see	see	VERB
dem-1058	27	33	kunst	kunst	PROPN
dem-1058	27	34	,	,	PUNCT
dem-1058	27	35	2003	2003	NUM
dem-1058	27	36	)	)	PUNCT
dem-1058	27	37	.	.	PUNCT
dem-1058	28	1	information	information	NOUN
dem-1058	28	2	and	and	CCONJ
dem-1058	28	3	prediction	prediction	NOUN
dem-1058	28	4	criteria	criterion	NOUN
dem-1058	28	5	in	in	ADP
dem-1058	28	6	selecting	select	VERB
dem-1058	28	7	the	the	DET
dem-1058	28	8	forecasting	forecasting	NOUN
dem-1058	28	9	model	model	NOUN
dem-1058	28	10	23	23	NUM
dem-1058	28	11	1	1	NUM
dem-1058	28	12	.	.	PUNCT
dem-1058	29	1	information	information	NOUN
dem-1058	29	2	and	and	CCONJ
dem-1058	29	3	prediction	prediction	NOUN
dem-1058	29	4	criteria	criterion	NOUN
dem-1058	29	5	generally	generally	ADV
dem-1058	29	6	,	,	PUNCT
dem-1058	29	7	information	information	NOUN
dem-1058	29	8	criterion	criterion	NOUN
dem-1058	29	9	takes	take	VERB
dem-1058	29	10	the	the	DET
dem-1058	29	11	form	form	NOUN
dem-1058	29	12	:	:	PUNCT
dem-1058	29	13	,	,	PUNCT
dem-1058	29	14	)	)	PUNCT
dem-1058	29	15	ˆ(ln2	ˆ(ln2	ADJ
dem-1058	29	16	qlic	qlic	NOUN
dem-1058	29	17			NUM
dem-1058	29	18			X
dem-1058	29	19	where	where	SCONJ
dem-1058	29	20	)	)	PUNCT
dem-1058	29	21	ˆ(l	ˆ(l	PRON
dem-1058	29	22	is	be	AUX
dem-1058	29	23	the	the	DET
dem-1058	29	24	likelihood	likelihood	NOUN
dem-1058	29	25	function	function	NOUN
dem-1058	29	26	,	,	PUNCT
dem-1058	29	27	and	and	CCONJ
dem-1058	29	28	q	q	NOUN
dem-1058	29	29	is	be	AUX
dem-1058	29	30	a	a	DET
dem-1058	29	31	penalty	penalty	NOUN
dem-1058	29	32	term	term	NOUN
dem-1058	29	33	that	that	PRON
dem-1058	29	34	is	be	AUX
dem-1058	29	35	a	a	DET
dem-1058	29	36	function	function	NOUN
dem-1058	29	37	of	of	ADP
dem-1058	29	38	the	the	DET
dem-1058	29	39	number	number	NOUN
dem-1058	29	40	of	of	ADP
dem-1058	29	41	parameters	parameter	NOUN
dem-1058	29	42	k	k	PROPN
dem-1058	29	43	and	and	CCONJ
dem-1058	29	44	the	the	DET
dem-1058	29	45	number	number	NOUN
dem-1058	29	46	of	of	ADP
dem-1058	29	47	observations	observation	NOUN
dem-1058	29	48	n	n	CCONJ
dem-1058	29	49	;	;	PUNCT
dem-1058	29	50	this	this	DET
dem-1058	29	51	penalty	penalty	NOUN
dem-1058	29	52	guards	guard	NOUN
dem-1058	29	53	against	against	ADP
dem-1058	29	54	overfitting	overfitte	VERB
dem-1058	29	55	,	,	PUNCT
dem-1058	29	56	i.e.	i.e.	X
dem-1058	29	57	using	use	VERB
dem-1058	29	58	too	too	ADV
dem-1058	29	59	many	many	ADJ
dem-1058	29	60	parameters	parameter	NOUN
dem-1058	29	61	.	.	PUNCT
dem-1058	30	1	for	for	ADP
dem-1058	30	2	akaike	akaike	NOUN
dem-1058	30	3	's	's	PART
dem-1058	30	4	information	information	NOUN
dem-1058	30	5	criterion	criterion	NOUN
dem-1058	30	6	(	(	PUNCT
dem-1058	30	7	aic	aic	PROPN
dem-1058	30	8	)	)	PUNCT
dem-1058	30	9	the	the	DET
dem-1058	30	10	penalty	penalty	NOUN
dem-1058	30	11	is	be	AUX
dem-1058	30	12	equal	equal	ADJ
dem-1058	30	13	to	to	ADP
dem-1058	30	14	,	,	PUNCT
dem-1058	30	15	2kq	2kq	ADJ
dem-1058	30	16			NOUN
dem-1058	30	17	for	for	ADP
dem-1058	30	18	schwartz	schwartz	PROPN
dem-1058	30	19	(	(	PUNCT
dem-1058	30	20	bayes	bayes	PROPN
dem-1058	30	21	)	)	PUNCT
dem-1058	30	22	information	information	NOUN
dem-1058	30	23	criterion	criterion	NOUN
dem-1058	30	24	−	−	PROPN
dem-1058	30	25	)	)	PUNCT
dem-1058	30	26	,	,	PUNCT
dem-1058	30	27	ln(nkq	ln(nkq	PROPN
dem-1058	30	28			NUM
dem-1058	30	29	for	for	ADP
dem-1058	30	30	hannan	hannan	PROPN
dem-1058	30	31	-	-	PUNCT
dem-1058	30	32	quinn	quinn	PROPN
dem-1058	30	33	information	information	NOUN
dem-1058	30	34	criterion	criterion	NOUN
dem-1058	30	35	−	−	PROPN
dem-1058	30	36	)	)	PUNCT
dem-1058	30	37	.ln(ln2	.ln(ln2	PUNCT
dem-1058	31	1	nkq	nkq	DET
dem-1058	31	2			NOUN
dem-1058	31	3	for	for	ADP
dem-1058	31	4	small	small	ADJ
dem-1058	31	5	samples	sample	NOUN
dem-1058	31	6	the	the	DET
dem-1058	31	7	aic	aic	PROPN
dem-1058	31	8	criterion	criterion	NOUN
dem-1058	31	9	is	be	AUX
dem-1058	31	10	biased	bias	VERB
dem-1058	31	11	and	and	CCONJ
dem-1058	31	12	may	may	AUX
dem-1058	31	13	suggest	suggest	VERB
dem-1058	31	14	a	a	DET
dem-1058	31	15	model	model	NOUN
dem-1058	31	16	with	with	ADP
dem-1058	31	17	a	a	DET
dem-1058	31	18	high	high	ADJ
dem-1058	31	19	number	number	NOUN
dem-1058	31	20	of	of	ADP
dem-1058	31	21	parameters	parameter	NOUN
dem-1058	31	22	compared	compare	VERB
dem-1058	31	23	with	with	ADP
dem-1058	31	24	the	the	DET
dem-1058	31	25	number	number	NOUN
dem-1058	31	26	of	of	ADP
dem-1058	31	27	observations	observation	NOUN
dem-1058	31	28	)	)	PUNCT
dem-1058	32	1	.40/	.40/	PROPN
dem-1058	32	2	(	(	PUNCT
dem-1058	32	3	kn	kn	NOUN
dem-1058	32	4	thus	thus	ADV
dem-1058	32	5	a	a	DET
dem-1058	32	6	bias	bias	NOUN
dem-1058	32	7	-	-	PUNCT
dem-1058	32	8	corrected	correct	VERB
dem-1058	32	9	version	version	NOUN
dem-1058	32	10	,	,	PUNCT
dem-1058	32	11	aicc	aicc	NOUN
dem-1058	32	12	,	,	PUNCT
dem-1058	32	13	is	be	AUX
dem-1058	32	14	increasingly	increasingly	ADV
dem-1058	32	15	used	use	VERB
dem-1058	32	16	.	.	PUNCT
dem-1058	33	1	the	the	DET
dem-1058	33	2	latter	latter	ADJ
dem-1058	33	3	is	be	AUX
dem-1058	33	4	given	give	VERB
dem-1058	33	5	by	by	ADP
dem-1058	33	6	adding	add	VERB
dem-1058	33	7	the	the	DET
dem-1058	33	8	quantity	quantity	NOUN
dem-1058	33	9	)	)	PUNCT
dem-1058	34	1	1/()1(2	1/()1(2	NUM
dem-1058	34	2			PROPN
dem-1058	34	3	knkk	knkk	VERB
dem-1058	34	4	to	to	ADP
dem-1058	34	5	the	the	DET
dem-1058	34	6	ordinary	ordinary	ADJ
dem-1058	34	7	aic	aic	PROPN
dem-1058	34	8	.	.	PUNCT
dem-1058	35	1	the	the	DET
dem-1058	35	2	bic	bic	PROPN
dem-1058	35	3	(	(	PUNCT
dem-1058	35	4	like	like	ADP
dem-1058	35	5	the	the	DET
dem-1058	35	6	aicc	aicc	NOUN
dem-1058	35	7	criterion	criterion	NOUN
dem-1058	35	8	)	)	PUNCT
dem-1058	35	9	penalizes	penalize	VERB
dem-1058	35	10	the	the	DET
dem-1058	35	11	addition	addition	NOUN
dem-1058	35	12	of	of	ADP
dem-1058	35	13	extra	extra	ADJ
dem-1058	35	14	parameters	parameter	NOUN
dem-1058	35	15	more	more	ADV
dem-1058	35	16	severely	severely	ADV
dem-1058	35	17	than	than	ADP
dem-1058	35	18	the	the	DET
dem-1058	35	19	aic	aic	PROPN
dem-1058	35	20	,	,	PUNCT
dem-1058	35	21	and	and	CCONJ
dem-1058	35	22	should	should	AUX
dem-1058	35	23	be	be	AUX
dem-1058	35	24	preferred	prefer	VERB
dem-1058	35	25	to	to	ADP
dem-1058	35	26	the	the	DET
dem-1058	35	27	ordinary	ordinary	ADJ
dem-1058	35	28	aic	aic	PROPN
dem-1058	35	29	in	in	ADP
dem-1058	35	30	timeseries	timeserie	NOUN
dem-1058	35	31	analysis	analysis	NOUN
dem-1058	35	32	especially	especially	ADV
dem-1058	35	33	when	when	SCONJ
dem-1058	35	34	the	the	DET
dem-1058	35	35	number	number	NOUN
dem-1058	35	36	of	of	ADP
dem-1058	35	37	parameters	parameter	NOUN
dem-1058	35	38	is	be	AUX
dem-1058	35	39	high	high	ADJ
dem-1058	35	40	compared	compare	VERB
dem-1058	35	41	with	with	ADP
dem-1058	35	42	the	the	DET
dem-1058	35	43	number	number	NOUN
dem-1058	35	44	of	of	ADP
dem-1058	35	45	observations	observation	NOUN
dem-1058	35	46	.	.	PUNCT
dem-1058	36	1	applying	apply	VERB
dem-1058	36	2	information	information	NOUN
dem-1058	36	3	criteria	criterion	NOUN
dem-1058	36	4	in	in	ADP
dem-1058	36	5	model	model	NOUN
dem-1058	36	6	selection	selection	NOUN
dem-1058	36	7	the	the	DET
dem-1058	36	8	model	model	NOUN
dem-1058	36	9	with	with	ADP
dem-1058	36	10	the	the	DET
dem-1058	36	11	minimum	minimum	NOUN
dem-1058	36	12	of	of	ADP
dem-1058	36	13	a	a	DET
dem-1058	36	14	given	give	VERB
dem-1058	36	15	information	information	NOUN
dem-1058	36	16	criterion	criterion	NOUN
dem-1058	36	17	is	be	AUX
dem-1058	36	18	chosen	choose	VERB
dem-1058	36	19	.	.	PUNCT
dem-1058	37	1	to	to	ADP
dem-1058	37	2	traditional	traditional	ADJ
dem-1058	37	3	prediction	prediction	NOUN
dem-1058	37	4	criteria	criterion	NOUN
dem-1058	37	5	,	,	PUNCT
dem-1058	37	6	used	use	VERB
dem-1058	37	7	both	both	PRON
dem-1058	37	8	in	in	ADP
dem-1058	37	9	the	the	DET
dem-1058	37	10	accuracy	accuracy	NOUN
dem-1058	37	11	evaluation	evaluation	NOUN
dem-1058	37	12	and	and	CCONJ
dem-1058	37	13	selection	selection	NOUN
dem-1058	37	14	of	of	ADP
dem-1058	37	15	forecasting	forecasting	NOUN
dem-1058	37	16	model	model	NOUN
dem-1058	37	17	,	,	PUNCT
dem-1058	37	18	belong	belong	ADJ
dem-1058	37	19	:	:	PUNCT
dem-1058	37	20	mean	mean	ADJ
dem-1058	37	21	absolute	absolute	ADJ
dem-1058	37	22	error	error	NOUN
dem-1058	37	23	,	,	PUNCT
dem-1058	37	24	||	||	NOUN
dem-1058	37	25	1	1	NUM
dem-1058	37	26	t	t	NOUN
dem-1058	37	27	e	e	PROPN
dem-1058	37	28	mae	mae	PROPN
dem-1058	37	29	t	t	PROPN
dem-1058	37	30	t	t	PROPN
dem-1058	37	31	t	t	PROPN
dem-1058	37	32			NUM
dem-1058	37	33	mean	mean	VERB
dem-1058	37	34	square	square	NOUN
dem-1058	37	35	error	error	NOUN
dem-1058	37	36	,	,	PUNCT
dem-1058	37	37	1	1	NUM
dem-1058	37	38	2	2	NUM
dem-1058	37	39	t	t	NOUN
dem-1058	37	40	e	e	PROPN
dem-1058	37	41	mse	mse	PROPN
dem-1058	37	42	t	t	PROPN
dem-1058	37	43	t	t	PROPN
dem-1058	37	44	t	t	PROPN
dem-1058	37	45			NUM
dem-1058	37	46	root	root	NOUN
dem-1058	37	47	mean	mean	VERB
dem-1058	37	48	square	square	NOUN
dem-1058	37	49	error	error	NOUN
dem-1058	37	50	,	,	PUNCT
dem-1058	38	1	msermse	msermse	ADJ
dem-1058	38	2			PROPN
dem-1058	38	3	mean	mean	VERB
dem-1058	38	4	absolute	absolute	ADJ
dem-1058	38	5	percentage	percentage	NOUN
dem-1058	38	6	error	error	NOUN
dem-1058	38	7	%	%	NOUN
dem-1058	38	8	,	,	PUNCT
dem-1058	38	9	100	100	NUM
dem-1058	38	10	|/|	|/|	NOUN
dem-1058	38	11	1	1	NUM
dem-1058	38	12	t	t	NOUN
dem-1058	38	13	ye	ye	NUM
dem-1058	38	14	mape	mape	PROPN
dem-1058	38	15	t	t	PROPN
dem-1058	38	16	t	t	PROPN
dem-1058	38	17	t	t	PROPN
dem-1058	38	18	t	t	PROPN
dem-1058	38	19			NUM
dem-1058	38	20	theil	theil	PROPN
dem-1058	38	21	's	's	PART
dem-1058	38	22	inequality	inequality	NOUN
dem-1058	38	23	coefficient	coefficient	NOUN
dem-1058	38	24	,	,	PUNCT
dem-1058	38	25	model)benchmark	model)benchmark	NOUN
dem-1058	38	26	''	''	PUNCT
dem-1058	38	27	(	(	PUNCT
dem-1058	38	28	)	)	PUNCT
dem-1058	38	29	modelnew	modelnew	NOUN
dem-1058	38	30	''	''	PUNCT
dem-1058	39	1	(	(	PUNCT
dem-1058	39	2	rmse	rmse	PROPN
dem-1058	39	3	rmse	rmse	PROPN
dem-1058	39	4	u	u	PROPN
dem-1058	39	5			PRON
dem-1058	39	6	where	where	SCONJ
dem-1058	39	7	te	te	PROPN
dem-1058	39	8	denotes	denote	VERB
dem-1058	39	9	prediction	prediction	NOUN
dem-1058	39	10	error	error	NOUN
dem-1058	39	11	,	,	PUNCT
dem-1058	39	12	,	,	PUNCT
dem-1058	39	13	ˆttt	ˆttt	PROPN
dem-1058	39	14	yye	yye	PROPN
dem-1058	39	15			PROPN
dem-1058	40	1	ty	ty	ADP
dem-1058	40	2	−	−	PROPN
dem-1058	40	3	realization	realization	NOUN
dem-1058	40	4	of	of	ADP
dem-1058	40	5	y	y	PROPN
dem-1058	40	6	in	in	ADP
dem-1058	40	7	period	period	NOUN
dem-1058	40	8	t	t	PROPN
dem-1058	40	9	,	,	PUNCT
dem-1058	40	10	tŷ	tŷ	PROPN
dem-1058	40	11	−	−	PROPN
dem-1058	40	12	forecast	forecast	NOUN
dem-1058	40	13	of	of	ADP
dem-1058	40	14	y	y	PROPN
dem-1058	40	15	for	for	ADP
dem-1058	40	16	period	period	NOUN
dem-1058	40	17	t.	t.	NOUN
dem-1058	40	18	applying	apply	VERB
dem-1058	40	19	usual	usual	ADJ
dem-1058	40	20	accuracy	accuracy	NOUN
dem-1058	40	21	measures	measure	VERB
dem-1058	40	22	the	the	DET
dem-1058	40	23	model	model	NOUN
dem-1058	40	24	with	with	ADP
dem-1058	40	25	the	the	DET
dem-1058	40	26	smallest	small	ADJ
dem-1058	40	27	value	value	NOUN
dem-1058	40	28	of	of	ADP
dem-1058	40	29	given	give	VERB
dem-1058	40	30	measure	measure	NOUN
dem-1058	40	31	is	be	AUX
dem-1058	40	32	selected	select	VERB
dem-1058	40	33	what	what	PRON
dem-1058	40	34	corresponds	correspond	VERB
dem-1058	40	35	to	to	ADP
dem-1058	40	36	the	the	DET
dem-1058	40	37	smallest	small	ADJ
dem-1058	40	38	prediction	prediction	NOUN
dem-1058	40	39	error	error	NOUN
dem-1058	40	40	.	.	PUNCT
dem-1058	41	1	mariola	mariola	PROPN
dem-1058	41	2	piłatowska	piłatowska	PROPN
dem-1058	41	3	24	24	NUM
dem-1058	41	4	the	the	DET
dem-1058	41	5	theil	theil	NOUN
dem-1058	41	6	's	's	PART
dem-1058	41	7	inequality	inequality	NOUN
dem-1058	41	8	coefficient	coefficient	NOUN
dem-1058	41	9	u	u	NOUN
dem-1058	41	10	indicates	indicate	VERB
dem-1058	41	11	whether	whether	SCONJ
dem-1058	41	12	a	a	DET
dem-1058	41	13	given	give	VERB
dem-1058	41	14	model	model	NOUN
dem-1058	41	15	is	be	AUX
dem-1058	41	16	worse	bad	ADJ
dem-1058	41	17	(	(	PUNCT
dem-1058	41	18	u	u	NOUN
dem-1058	41	19	>	>	X
dem-1058	41	20	1	1	NUM
dem-1058	41	21	)	)	PUNCT
dem-1058	41	22	or	or	CCONJ
dem-1058	41	23	better	well	ADJ
dem-1058	41	24	(	(	PUNCT
dem-1058	41	25	u	u	NOUN
dem-1058	41	26	<	<	X
dem-1058	41	27	1	1	NUM
dem-1058	41	28	)	)	PUNCT
dem-1058	41	29	than	than	ADP
dem-1058	41	30	the	the	DET
dem-1058	41	31	random	random	ADJ
dem-1058	41	32	walk	walk	NOUN
dem-1058	41	33	model	model	NOUN
dem-1058	41	34	(	(	PUNCT
dem-1058	41	35	tt	tt	PROPN
dem-1058	41	36	yy	yy	PROPN
dem-1058	41	37	1ˆ	1ˆ	PRON
dem-1058	41	38	)	)	PUNCT
dem-1058	41	39	considered	consider	VERB
dem-1058	41	40	as	as	ADP
dem-1058	41	41	a	a	DET
dem-1058	41	42	benchmark	benchmark	NOUN
dem-1058	41	43	model	model	NOUN
dem-1058	41	44	.	.	PUNCT
dem-1058	42	1	it	it	PRON
dem-1058	42	2	is	be	AUX
dem-1058	42	3	worth	worth	ADJ
dem-1058	42	4	highlighting	highlight	VERB
dem-1058	42	5	that	that	SCONJ
dem-1058	42	6	the	the	DET
dem-1058	42	7	choice	choice	NOUN
dem-1058	42	8	of	of	ADP
dem-1058	42	9	accuracy	accuracy	NOUN
dem-1058	42	10	measure	measure	NOUN
dem-1058	42	11	can	can	AUX
dem-1058	42	12	affect	affect	VERB
dem-1058	42	13	the	the	DET
dem-1058	42	14	ranking	ranking	NOUN
dem-1058	42	15	of	of	ADP
dem-1058	42	16	forecasting	forecasting	NOUN
dem-1058	42	17	methods	method	NOUN
dem-1058	42	18	,	,	PUNCT
dem-1058	42	19	and	and	CCONJ
dem-1058	42	20	also	also	ADV
dem-1058	42	21	models	model	NOUN
dem-1058	42	22	(	(	PUNCT
dem-1058	42	23	armstrong	armstrong	PROPN
dem-1058	42	24	,	,	PUNCT
dem-1058	42	25	2001	2001	NUM
dem-1058	42	26	;	;	PUNCT
dem-1058	42	27	armstrong	armstrong	PROPN
dem-1058	42	28	,	,	PUNCT
dem-1058	42	29	fildes	filde	NOUN
dem-1058	42	30	,	,	PUNCT
dem-1058	42	31	1995	1995	NUM
dem-1058	42	32	)	)	PUNCT
dem-1058	42	33	.	.	PUNCT
dem-1058	43	1	for	for	ADP
dem-1058	43	2	instance	instance	NOUN
dem-1058	43	3	,	,	PUNCT
dem-1058	43	4	the	the	DET
dem-1058	43	5	mse	mse	NOUN
dem-1058	43	6	depends	depend	VERB
dem-1058	43	7	on	on	ADP
dem-1058	43	8	the	the	DET
dem-1058	43	9	scale	scale	NOUN
dem-1058	43	10	in	in	ADP
dem-1058	43	11	which	which	PRON
dem-1058	43	12	the	the	DET
dem-1058	43	13	variable	variable	NOUN
dem-1058	43	14	is	be	AUX
dem-1058	43	15	measured	measure	VERB
dem-1058	43	16	.	.	PUNCT
dem-1058	44	1	this	this	PRON
dem-1058	44	2	means	mean	VERB
dem-1058	44	3	that	that	SCONJ
dem-1058	44	4	the	the	DET
dem-1058	44	5	mse	mse	NOUN
dem-1058	44	6	is	be	AUX
dem-1058	44	7	appropriate	appropriate	ADJ
dem-1058	44	8	only	only	ADV
dem-1058	44	9	for	for	ADP
dem-1058	44	10	assessing	assess	VERB
dem-1058	44	11	the	the	DET
dem-1058	44	12	results	result	NOUN
dem-1058	44	13	for	for	ADP
dem-1058	44	14	a	a	DET
dem-1058	44	15	single	single	ADJ
dem-1058	44	16	time	time	NOUN
dem-1058	44	17	series	series	NOUN
dem-1058	44	18	,	,	PUNCT
dem-1058	44	19	and	and	CCONJ
dem-1058	44	20	should	should	AUX
dem-1058	44	21	be	be	AUX
dem-1058	44	22	avoided	avoid	VERB
dem-1058	44	23	to	to	PART
dem-1058	44	24	assess	assess	VERB
dem-1058	44	25	accuracy	accuracy	NOUN
dem-1058	44	26	across	across	ADP
dem-1058	44	27	(	(	PUNCT
dem-1058	44	28	many	many	ADJ
dem-1058	44	29	)	)	PUNCT
dem-1058	44	30	different	different	ADJ
dem-1058	44	31	series	series	NOUN
dem-1058	44	32	.	.	PUNCT
dem-1058	45	1	in	in	ADP
dem-1058	45	2	that	that	DET
dem-1058	45	3	case	case	NOUN
dem-1058	45	4	the	the	DET
dem-1058	45	5	scale	scale	NOUN
dem-1058	45	6	-	-	PUNCT
dem-1058	45	7	independent	independent	ADJ
dem-1058	45	8	measures	measure	NOUN
dem-1058	45	9	are	be	AUX
dem-1058	45	10	required2	required2	NOUN
dem-1058	45	11	,	,	PUNCT
dem-1058	45	12	e.g.	e.g.	ADV
dem-1058	45	13	mape	mape	NOUN
dem-1058	45	14	or	or	CCONJ
dem-1058	45	15	theil	theil	PROPN
dem-1058	45	16	's	's	PART
dem-1058	45	17	inequality	inequality	NOUN
dem-1058	45	18	coefficient	coefficient	NOUN
dem-1058	45	19	(	(	PUNCT
dem-1058	45	20	u	u	NOUN
dem-1058	45	21	)	)	PUNCT
dem-1058	45	22	.	.	PUNCT
dem-1058	46	1	besides	besides	SCONJ
dem-1058	46	2	,	,	PUNCT
dem-1058	46	3	for	for	ADP
dem-1058	46	4	the	the	DET
dem-1058	46	5	reason	reason	NOUN
dem-1058	46	6	that	that	SCONJ
dem-1058	46	7	the	the	DET
dem-1058	46	8	loss	loss	NOUN
dem-1058	46	9	function	function	NOUN
dem-1058	46	10	may	may	AUX
dem-1058	46	11	be	be	AUX
dem-1058	46	12	asymmetric	asymmetric	ADJ
dem-1058	46	13	(	(	PUNCT
dem-1058	46	14	e.g.	e.g.	ADV
dem-1058	46	15	underforecasting	underforecasting	NOUN
dem-1058	46	16	is	be	AUX
dem-1058	46	17	worse	bad	ADJ
dem-1058	46	18	than	than	ADP
dem-1058	46	19	overforecasting	overforecaste	VERB
dem-1058	46	20	)	)	PUNCT
dem-1058	46	21	,	,	PUNCT
dem-1058	46	22	it	it	PRON
dem-1058	46	23	may	may	AUX
dem-1058	46	24	be	be	AUX
dem-1058	46	25	important	important	ADJ
dem-1058	46	26	to	to	PART
dem-1058	46	27	forecast	forecast	VERB
dem-1058	46	28	the	the	DET
dem-1058	46	29	direction	direction	NOUN
dem-1058	46	30	of	of	ADP
dem-1058	46	31	movement	movement	NOUN
dem-1058	46	32	or	or	CCONJ
dem-1058	46	33	to	to	PART
dem-1058	46	34	predict	predict	VERB
dem-1058	46	35	large	large	ADJ
dem-1058	46	36	movements	movement	NOUN
dem-1058	46	37	(	(	PUNCT
dem-1058	46	38	chatfield	chatfield	PROPN
dem-1058	46	39	,	,	PUNCT
dem-1058	46	40	2000	2000	NUM
dem-1058	46	41	)	)	PUNCT
dem-1058	46	42	.	.	PUNCT
dem-1058	47	1	however	however	ADV
dem-1058	47	2	,	,	PUNCT
dem-1058	47	3	it	it	PRON
dem-1058	47	4	is	be	AUX
dem-1058	47	5	not	not	PART
dem-1058	47	6	possible	possible	ADJ
dem-1058	47	7	to	to	PART
dem-1058	47	8	select	select	VERB
dem-1058	47	9	a	a	DET
dem-1058	47	10	measure	measure	NOUN
dem-1058	47	11	of	of	ADP
dem-1058	47	12	forecast	forecast	NOUN
dem-1058	47	13	accuracy	accuracy	NOUN
dem-1058	47	14	that	that	PRON
dem-1058	47	15	is	be	AUX
dem-1058	47	16	scale	scale	NOUN
dem-1058	47	17	-	-	PUNCT
dem-1058	47	18	independent	independent	ADJ
dem-1058	47	19	and	and	CCONJ
dem-1058	47	20	yet	yet	ADV
dem-1058	47	21	satisfies	satisfy	VERB
dem-1058	47	22	the	the	DET
dem-1058	47	23	demands	demand	NOUN
dem-1058	47	24	of	of	ADP
dem-1058	47	25	the	the	DET
dem-1058	47	26	appropriate	appropriate	ADJ
dem-1058	47	27	loss	loss	NOUN
dem-1058	47	28	function	function	NOUN
dem-1058	47	29	.	.	PUNCT
dem-1058	48	1	in	in	ADP
dem-1058	48	2	conclusion	conclusion	NOUN
dem-1058	48	3	,	,	PUNCT
dem-1058	48	4	there	there	PRON
dem-1058	48	5	is	be	VERB
dem-1058	48	6	no	no	DET
dem-1058	48	7	measure	measure	NOUN
dem-1058	48	8	suitable	suitable	ADJ
dem-1058	48	9	for	for	ADP
dem-1058	48	10	all	all	DET
dem-1058	48	11	types	type	NOUN
dem-1058	48	12	of	of	ADP
dem-1058	48	13	data	datum	NOUN
dem-1058	48	14	and	and	CCONJ
dem-1058	48	15	all	all	DET
dem-1058	48	16	contexts	context	NOUN
dem-1058	48	17	.	.	PUNCT
dem-1058	49	1	there	there	PRON
dem-1058	49	2	are	be	VERB
dem-1058	49	3	many	many	ADJ
dem-1058	49	4	empirical	empirical	ADJ
dem-1058	49	5	evidence	evidence	NOUN
dem-1058	49	6	that	that	SCONJ
dem-1058	49	7	a	a	DET
dem-1058	49	8	method	method	NOUN
dem-1058	49	9	which	which	PRON
dem-1058	49	10	is	be	AUX
dem-1058	49	11	'	'	PUNCT
dem-1058	49	12	best	good	ADJ
dem-1058	49	13	'	'	PUNCT
dem-1058	49	14	under	under	ADP
dem-1058	49	15	one	one	NUM
dem-1058	49	16	criterion	criterion	NOUN
dem-1058	49	17	need	need	AUX
dem-1058	49	18	not	not	PART
dem-1058	49	19	be	be	AUX
dem-1058	49	20	'	'	PUNCT
dem-1058	49	21	best	good	ADJ
dem-1058	49	22	'	'	PUNCT
dem-1058	49	23	under	under	ADP
dem-1058	49	24	alternative	alternative	ADJ
dem-1058	49	25	criteria	criterion	NOUN
dem-1058	49	26	(	(	PUNCT
dem-1058	49	27	e.g.	e.g.	ADV
dem-1058	49	28	swanson	swanson	NOUN
dem-1058	49	29	and	and	CCONJ
dem-1058	49	30	white	white	ADJ
dem-1058	49	31	,	,	PUNCT
dem-1058	49	32	1997	1997	NUM
dem-1058	49	33	)	)	PUNCT
dem-1058	49	34	.	.	PUNCT
dem-1058	50	1	the	the	DET
dem-1058	50	2	choice	choice	NOUN
dem-1058	50	3	of	of	ADP
dem-1058	50	4	forecasting	forecasting	NOUN
dem-1058	50	5	model	model	NOUN
dem-1058	50	6	may	may	AUX
dem-1058	50	7	also	also	ADV
dem-1058	50	8	be	be	AUX
dem-1058	50	9	carried	carry	VERB
dem-1058	50	10	out	out	ADP
dem-1058	50	11	by	by	ADP
dem-1058	50	12	the	the	DET
dem-1058	50	13	accumulated	accumulate	VERB
dem-1058	50	14	prediction	prediction	NOUN
dem-1058	50	15	error	error	NOUN
dem-1058	50	16	,	,	PUNCT
dem-1058	50	17	ape	ape	NOUN
dem-1058	50	18	,	,	PUNCT
dem-1058	50	19	(	(	PUNCT
dem-1058	50	20	rissanen	rissanen	NOUN
dem-1058	50	21	,	,	PUNCT
dem-1058	50	22	1986	1986	NUM
dem-1058	50	23	)	)	PUNCT
dem-1058	50	24	.	.	PUNCT
dem-1058	51	1	according	accord	VERB
dem-1058	51	2	to	to	ADP
dem-1058	51	3	the	the	DET
dem-1058	51	4	ape	ape	NOUN
dem-1058	51	5	the	the	DET
dem-1058	51	6	most	most	ADV
dem-1058	51	7	useful	useful	ADJ
dem-1058	51	8	model	model	NOUN
dem-1058	51	9	is	be	AUX
dem-1058	51	10	the	the	DET
dem-1058	51	11	model	model	NOUN
dem-1058	51	12	with	with	ADP
dem-1058	51	13	the	the	DET
dem-1058	51	14	smallest	small	ADJ
dem-1058	51	15	out	out	ADP
dem-1058	51	16	-	-	PUNCT
dem-1058	51	17	of	of	ADP
dem-1058	51	18	-	-	PUNCT
dem-1058	51	19	sample	sample	NOUN
dem-1058	51	20	one	one	NUM
dem-1058	51	21	-	-	PUNCT
dem-1058	51	22	step	step	NOUN
dem-1058	51	23	-	-	PUNCT
dem-1058	51	24	ahead	ahead	NOUN
dem-1058	51	25	prediction	prediction	NOUN
dem-1058	51	26	error	error	NOUN
dem-1058	51	27	.	.	PUNCT
dem-1058	52	1	the	the	DET
dem-1058	52	2	ape	ape	NOUN
dem-1058	52	3	method	method	NOUN
dem-1058	52	4	proceeds	proceed	NOUN
dem-1058	52	5	by	by	ADP
dem-1058	52	6	calculating	calculate	VERB
dem-1058	52	7	sequential	sequential	ADJ
dem-1058	52	8	one	one	NUM
dem-1058	52	9	-	-	PUNCT
dem-1058	52	10	step	step	NOUN
dem-1058	52	11	-	-	PUNCT
dem-1058	52	12	ahead	ahead	NOUN
dem-1058	52	13	forecasts	forecast	NOUN
dem-1058	52	14	based	base	VERB
dem-1058	52	15	on	on	ADP
dem-1058	52	16	gradually	gradually	ADV
dem-1058	52	17	increasing	increase	VERB
dem-1058	52	18	sample	sample	NOUN
dem-1058	52	19	.	.	PUNCT
dem-1058	53	1	for	for	ADP
dem-1058	53	2	model	model	NOUN
dem-1058	53	3	mj	mj	PROPN
dem-1058	53	4	the	the	DET
dem-1058	53	5	ape	ape	NOUN
dem-1058	53	6	is	be	AUX
dem-1058	53	7	calculated	calculate	VERB
dem-1058	53	8	as	as	ADP
dem-1058	53	9	follows	follow	VERB
dem-1058	53	10	(	(	PUNCT
dem-1058	53	11	wagenmaker	wagenmaker	NOUN
dem-1058	53	12	,	,	PUNCT
dem-1058	53	13	grünwald	grünwald	NOUN
dem-1058	53	14	,	,	PUNCT
dem-1058	53	15	steyvers	steyver	NOUN
dem-1058	53	16	,	,	PUNCT
dem-1058	53	17	2006	2006	NUM
dem-1058	53	18	):	):	PUNCT
dem-1058	53	19	1	1	NUM
dem-1058	53	20	.	.	X
dem-1058	53	21	determine	determine	VERB
dem-1058	53	22	the	the	DET
dem-1058	53	23	smallest	small	ADJ
dem-1058	53	24	number	number	NOUN
dem-1058	53	25	s	s	VERB
dem-1058	53	26	of	of	ADP
dem-1058	53	27	observations	observation	NOUN
dem-1058	53	28	that	that	PRON
dem-1058	53	29	makes	make	VERB
dem-1058	53	30	the	the	DET
dem-1058	53	31	model	model	NOUN
dem-1058	53	32	identifiable	identifiable	ADJ
dem-1058	53	33	.	.	PUNCT
dem-1058	54	1	set	set	VERB
dem-1058	54	2	,	,	PUNCT
dem-1058	54	3	1	1	NUM
dem-1058	54	4	si	si	NOUN
dem-1058	55	1	so	so	SCONJ
dem-1058	55	2	that	that	SCONJ
dem-1058	55	3	.1	.1	NUM
dem-1058	55	4	si	si	PROPN
dem-1058	55	5			PROPN
dem-1058	55	6	2	2	NUM
dem-1058	55	7	.	.	PUNCT
dem-1058	55	8	based	base	VERB
dem-1058	55	9	on	on	ADP
dem-1058	55	10	the	the	DET
dem-1058	55	11	first	first	ADJ
dem-1058	55	12	1i	1i	NUM
dem-1058	55	13	observations	observation	NOUN
dem-1058	55	14	,	,	PUNCT
dem-1058	55	15	calculate	calculate	VERB
dem-1058	55	16	a	a	DET
dem-1058	55	17	prediction	prediction	NOUN
dem-1058	55	18	ip̂	ip̂	VERB
dem-1058	55	19	for	for	ADP
dem-1058	55	20	the	the	DET
dem-1058	55	21	next	next	ADJ
dem-1058	55	22	observation	observation	NOUN
dem-1058	55	23	.i	.i	NOUN
dem-1058	56	1	3	3	X
dem-1058	56	2	.	.	X
dem-1058	56	3	calculate	calculate	VERB
dem-1058	56	4	the	the	DET
dem-1058	56	5	prediction	prediction	NOUN
dem-1058	56	6	error	error	NOUN
dem-1058	56	7	for	for	ADP
dem-1058	56	8	observation	observation	NOUN
dem-1058	56	9	i	i	PRON
dem-1058	56	10	,	,	PUNCT
dem-1058	56	11	e.g.	e.g.	ADV
dem-1058	56	12	squared	square	VERB
dem-1058	56	13	difference	difference	NOUN
dem-1058	56	14	between	between	ADP
dem-1058	56	15	the	the	DET
dem-1058	56	16	predicted	predict	VERB
dem-1058	56	17	value	value	NOUN
dem-1058	56	18	ip̂	ip̂	VERB
dem-1058	56	19	and	and	CCONJ
dem-1058	56	20	the	the	DET
dem-1058	56	21	observed	observed	ADJ
dem-1058	56	22	value	value	NOUN
dem-1058	56	23	.ix	.ix	PUNCT
dem-1058	57	1	4	4	X
dem-1058	57	2	.	.	X
dem-1058	57	3	increase	increase	VERB
dem-1058	57	4	i	i	PRON
dem-1058	57	5	by	by	ADP
dem-1058	57	6	1	1	NUM
dem-1058	57	7	and	and	CCONJ
dem-1058	57	8	repeat	repeat	VERB
dem-1058	57	9	steps	step	NOUN
dem-1058	57	10	2	2	NUM
dem-1058	57	11	and	and	CCONJ
dem-1058	57	12	3	3	NUM
dem-1058	57	13	until	until	ADP
dem-1058	57	14	.ni	.ni	PUNCT
dem-1058	58	1			NUM
dem-1058	58	2	5	5	X
dem-1058	58	3	.	.	X
dem-1058	58	4	sum	sum	VERB
dem-1058	58	5	all	all	PRON
dem-1058	58	6	of	of	ADP
dem-1058	58	7	the	the	DET
dem-1058	58	8	one	one	NUM
dem-1058	58	9	-	-	PUNCT
dem-1058	58	10	step	step	NOUN
dem-1058	58	11	-	-	PUNCT
dem-1058	58	12	ahead	ahead	NOUN
dem-1058	58	13	prediction	prediction	NOUN
dem-1058	58	14	errors	error	NOUN
dem-1058	58	15	as	as	SCONJ
dem-1058	58	16	calculated	calculate	VERB
dem-1058	58	17	in	in	ADP
dem-1058	58	18	step	step	NOUN
dem-1058	58	19	3	3	NUM
dem-1058	58	20	.	.	PUNCT
dem-1058	59	1	the	the	DET
dem-1058	59	2	result	result	NOUN
dem-1058	59	3	is	be	AUX
dem-1058	59	4	the	the	DET
dem-1058	59	5	ape	ape	NOUN
dem-1058	59	6	.	.	PUNCT
dem-1058	60	1	for	for	ADP
dem-1058	60	2	model	model	NOUN
dem-1058	60	3	jm	jm	PROPN
dem-1058	60	4	the	the	DET
dem-1058	60	5	accumulated	accumulate	VERB
dem-1058	60	6	prediction	prediction	NOUN
dem-1058	60	7	error	error	NOUN
dem-1058	60	8	is	be	AUX
dem-1058	60	9	given	give	VERB
dem-1058	60	10	by	by	ADP
dem-1058	60	11	:	:	PUNCT
dem-1058	60	12	)	)	PUNCT
dem-1058	61	1	]	]	PUNCT
dem-1058	61	2	,	,	PUNCT
dem-1058	61	3	ˆ	ˆ	PROPN
dem-1058	61	4	(	(	PUNCT
dem-1058	61	5	,	,	PUNCT
dem-1058	61	6	[	[	X
dem-1058	61	7	)	)	PUNCT
dem-1058	61	8	(	(	PUNCT
dem-1058	61	9	11	11	NUM
dem-1058	61	10			NOUN
dem-1058	61	11	ii	ii	NOUN
dem-1058	61	12	n	n	NOUN
dem-1058	61	13	si	si	X
dem-1058	61	14	ij	ij	NUM
dem-1058	61	15	xpxdmape	xpxdmape	PROPN
dem-1058	61	16	2	2	NUM
dem-1058	61	17	mentzer	mentzer	PROPN
dem-1058	61	18	and	and	CCONJ
dem-1058	61	19	kahn	kahn	PROPN
dem-1058	61	20	(	(	PUNCT
dem-1058	61	21	1995	1995	NUM
dem-1058	61	22	)	)	PUNCT
dem-1058	61	23	found	find	VERB
dem-1058	61	24	in	in	ADP
dem-1058	61	25	a	a	DET
dem-1058	61	26	survey	survey	NOUN
dem-1058	61	27	of	of	ADP
dem-1058	61	28	207	207	NUM
dem-1058	61	29	forecasting	forecasting	NOUN
dem-1058	61	30	executives	executive	NOUN
dem-1058	61	31	in	in	ADP
dem-1058	61	32	us	we	PRON
dem-1058	61	33	that	that	DET
dem-1058	61	34	mape	mape	NOUN
dem-1058	61	35	was	be	AUX
dem-1058	61	36	the	the	DET
dem-1058	61	37	most	most	ADV
dem-1058	61	38	commonly	commonly	ADV
dem-1058	61	39	used	use	VERB
dem-1058	61	40	measure	measure	NOUN
dem-1058	61	41	(	(	PUNCT
dem-1058	61	42	52	52	NUM
dem-1058	61	43	%	%	NOUN
dem-1058	61	44	)	)	PUNCT
dem-1058	61	45	while	while	SCONJ
dem-1058	61	46	only	only	ADV
dem-1058	61	47	10	10	NUM
dem-1058	61	48	%	%	NOUN
dem-1058	61	49	used	use	VERB
dem-1058	61	50	mse	mse	PROPN
dem-1058	61	51	.	.	PUNCT
dem-1058	62	1	information	information	NOUN
dem-1058	62	2	and	and	CCONJ
dem-1058	62	3	prediction	prediction	NOUN
dem-1058	62	4	criteria	criterion	NOUN
dem-1058	62	5	in	in	ADP
dem-1058	62	6	selecting	select	VERB
dem-1058	62	7	the	the	DET
dem-1058	62	8	forecasting	forecasting	NOUN
dem-1058	62	9	model	model	NOUN
dem-1058	62	10	25	25	NUM
dem-1058	62	11	where	where	SCONJ
dem-1058	62	12	d	d	NOUN
dem-1058	62	13	indicates	indicate	VERB
dem-1058	62	14	the	the	DET
dem-1058	62	15	specific	specific	ADJ
dem-1058	62	16	loss	loss	NOUN
dem-1058	62	17	function	function	NOUN
dem-1058	62	18	that	that	PRON
dem-1058	62	19	quantifies	quantify	VERB
dem-1058	62	20	the	the	DET
dem-1058	62	21	discrepancy	discrepancy	NOUN
dem-1058	62	22	between	between	ADP
dem-1058	62	23	observed	observed	ADJ
dem-1058	62	24	and	and	CCONJ
dem-1058	62	25	predicted	predict	VERB
dem-1058	62	26	values	value	NOUN
dem-1058	62	27	.	.	PUNCT
dem-1058	63	1	in	in	ADP
dem-1058	63	2	the	the	DET
dem-1058	63	3	case	case	NOUN
dem-1058	63	4	of	of	ADP
dem-1058	63	5	point	point	NOUN
dem-1058	63	6	predictions	prediction	NOUN
dem-1058	63	7	one	one	NUM
dem-1058	63	8	typically	typically	ADV
dem-1058	63	9	uses	use	VERB
dem-1058	63	10	the	the	DET
dem-1058	63	11	squared	square	VERB
dem-1058	63	12	error	error	NOUN
dem-1058	63	13	2)ˆ	2)ˆ	NOUN
dem-1058	63	14	(	(	PUNCT
dem-1058	63	15	ii	ii	NOUN
dem-1058	63	16	px	px	NOUN
dem-1058	63	17			PROPN
dem-1058	63	18	,	,	PUNCT
dem-1058	63	19	but	but	CCONJ
dem-1058	63	20	another	another	DET
dem-1058	63	21	choice	choice	NOUN
dem-1058	63	22	would	would	AUX
dem-1058	63	23	be	be	AUX
dem-1058	63	24	to	to	PART
dem-1058	63	25	compute	compute	VERB
dem-1058	63	26	the	the	DET
dem-1058	63	27	absolute	absolute	ADJ
dem-1058	63	28	value	value	NOUN
dem-1058	63	29	loss	loss	NOUN
dem-1058	63	30	ii	ii	NOUN
dem-1058	63	31	px	px	NOUN
dem-1058	63	32	ˆ	ˆ	NOUN
dem-1058	63	33	,	,	PUNCT
dem-1058	63	34	or	or	CCONJ
dem-1058	63	35	more	more	ADV
dem-1058	63	36	generally	generally	ADV
dem-1058	63	37	,	,	PUNCT
dem-1058	63	38	an	an	DET
dem-1058	63	39	α	α	NOUN
dem-1058	63	40	-	-	PUNCT
dem-1058	63	41	loss	loss	NOUN
dem-1058	63	42	function	function	NOUN
dem-1058	63	43	,	,	PUNCT
dem-1058	63	44			X
dem-1058	63	45	ii	ii	NOUN
dem-1058	63	46	px	px	PROPN
dem-1058	63	47	ˆ	ˆ	NOUN
dem-1058	63	48	,	,	PUNCT
dem-1058	63	49	]	]	PUNCT
dem-1058	63	50	2,1[	2,1[	NUM
dem-1058	63	51	(	(	PUNCT
dem-1058	63	52	rissanen	rissanen	NOUN
dem-1058	63	53	,	,	PUNCT
dem-1058	63	54	2003	2003	NUM
dem-1058	63	55	)	)	PUNCT
dem-1058	63	56	.	.	PUNCT
dem-1058	64	1	2	2	X
dem-1058	64	2	.	.	X
dem-1058	64	3	selection	selection	NOUN
dem-1058	64	4	of	of	ADP
dem-1058	64	5	forecasting	forecasting	NOUN
dem-1058	64	6	model	model	NOUN
dem-1058	64	7	–	–	PUNCT
dem-1058	64	8	empirical	empirical	ADJ
dem-1058	64	9	examples	example	NOUN
dem-1058	64	10	to	to	PART
dem-1058	64	11	compare	compare	VERB
dem-1058	64	12	the	the	DET
dem-1058	64	13	performance	performance	NOUN
dem-1058	64	14	of	of	ADP
dem-1058	64	15	information	information	NOUN
dem-1058	64	16	and	and	CCONJ
dem-1058	64	17	prediction	prediction	NOUN
dem-1058	64	18	criteria	criterion	NOUN
dem-1058	64	19	in	in	ADP
dem-1058	64	20	selecting	select	VERB
dem-1058	64	21	the	the	DET
dem-1058	64	22	best	good	ADJ
dem-1058	64	23	forecasting	forecasting	NOUN
dem-1058	64	24	model	model	NOUN
dem-1058	64	25	the	the	DET
dem-1058	64	26	monthly	monthly	ADJ
dem-1058	64	27	data	datum	NOUN
dem-1058	64	28	on	on	ADP
dem-1058	64	29	consumer	consumer	NOUN
dem-1058	64	30	price	price	NOUN
dem-1058	64	31	index	index	NOUN
dem-1058	64	32	cpi	cpi	NOUN
dem-1058	64	33	(	(	PUNCT
dem-1058	64	34	corresponding	correspond	VERB
dem-1058	64	35	period	period	NOUN
dem-1058	64	36	of	of	ADP
dem-1058	64	37	previous	previous	ADJ
dem-1058	64	38	year	year	NOUN
dem-1058	64	39	=	=	NOUN
dem-1058	64	40	100	100	NUM
dem-1058	64	41	)	)	PUNCT
dem-1058	64	42	and	and	CCONJ
dem-1058	64	43	industry	industry	NOUN
dem-1058	64	44	production	production	NOUN
dem-1058	64	45	ip	ip	NOUN
dem-1058	64	46	(	(	PUNCT
dem-1058	64	47	in	in	ADP
dem-1058	64	48	billion	billion	NUM
dem-1058	64	49	pln	pln	PROPN
dem-1058	64	50	zl	zl	PROPN
dem-1058	64	51	)	)	PUNCT
dem-1058	64	52	in	in	ADP
dem-1058	64	53	poland	poland	PROPN
dem-1058	64	54	were	be	AUX
dem-1058	64	55	used3	used3	PROPN
dem-1058	64	56	(	(	PUNCT
dem-1058	64	57	in	in	ADP
dem-1058	64	58	the	the	DET
dem-1058	64	59	period	period	NOUN
dem-1058	64	60	2002:01−2010:12	2002:01−2010:12	NUM
dem-1058	64	61	,	,	PUNCT
dem-1058	64	62	108	108	NUM
dem-1058	64	63	observations	observation	NOUN
dem-1058	64	64	,	,	PUNCT
dem-1058	64	65	data	datum	NOUN
dem-1058	64	66	are	be	AUX
dem-1058	64	67	seasonally	seasonally	ADV
dem-1058	64	68	adjusted	adjust	VERB
dem-1058	64	69	)	)	PUNCT
dem-1058	64	70	.	.	PUNCT
dem-1058	65	1	the	the	DET
dem-1058	65	2	reason	reason	NOUN
dem-1058	65	3	for	for	ADP
dem-1058	65	4	such	such	ADJ
dem-1058	65	5	selection	selection	NOUN
dem-1058	65	6	of	of	ADP
dem-1058	65	7	time	time	NOUN
dem-1058	65	8	series	series	NOUN
dem-1058	65	9	is	be	AUX
dem-1058	65	10	the	the	DET
dem-1058	65	11	desire	desire	NOUN
dem-1058	65	12	to	to	PART
dem-1058	65	13	check	check	VERB
dem-1058	65	14	the	the	DET
dem-1058	65	15	performance	performance	NOUN
dem-1058	65	16	of	of	ADP
dem-1058	65	17	selection	selection	NOUN
dem-1058	65	18	criteria	criterion	NOUN
dem-1058	65	19	with	with	ADP
dem-1058	65	20	regard	regard	NOUN
dem-1058	65	21	to	to	ADP
dem-1058	65	22	time	time	NOUN
dem-1058	65	23	series	series	NOUN
dem-1058	65	24	of	of	ADP
dem-1058	65	25	different	different	ADJ
dem-1058	65	26	properties	property	NOUN
dem-1058	65	27	,	,	PUNCT
dem-1058	65	28	i.e.	i.e.	X
dem-1058	65	29	in	in	ADP
dem-1058	65	30	the	the	DET
dem-1058	65	31	above	above	ADJ
dem-1058	65	32	case	case	NOUN
dem-1058	65	33	cpi	cpi	PROPN
dem-1058	65	34	is	be	AUX
dem-1058	65	35	expected	expect	VERB
dem-1058	65	36	to	to	PART
dem-1058	65	37	be	be	AUX
dem-1058	65	38	rather	rather	ADV
dem-1058	65	39	an	an	DET
dem-1058	65	40	integrated	integrate	VERB
dem-1058	65	41	process	process	NOUN
dem-1058	65	42	,	,	PUNCT
dem-1058	65	43	and	and	CCONJ
dem-1058	65	44	ip	ip	VERB
dem-1058	65	45	−	−	PROPN
dem-1058	65	46	rather	rather	ADV
dem-1058	65	47	a	a	DET
dem-1058	65	48	stationary	stationary	ADJ
dem-1058	65	49	process	process	NOUN
dem-1058	65	50	around	around	ADP
dem-1058	65	51	deterministic	deterministic	ADJ
dem-1058	65	52	trend4	trend4	NOUN
dem-1058	65	53	.	.	PUNCT
dem-1058	66	1	the	the	DET
dem-1058	66	2	set	set	NOUN
dem-1058	66	3	of	of	ADP
dem-1058	66	4	candidate	candidate	NOUN
dem-1058	66	5	models	model	NOUN
dem-1058	66	6	for	for	ADP
dem-1058	66	7	cpi	cpi	PROPN
dem-1058	66	8	includes	include	VERB
dem-1058	66	9	:	:	PUNCT
dem-1058	66	10	autoregressive	autoregressive	ADJ
dem-1058	66	11	model	model	NOUN
dem-1058	66	12	,	,	PUNCT
dem-1058	66	13	ar(12	ar(12	PROPN
dem-1058	66	14	)	)	PUNCT
dem-1058	66	15	,	,	PUNCT
dem-1058	66	16	arima(12,1,0	arima(12,1,0	ADJ
dem-1058	66	17	)	)	PUNCT
dem-1058	66	18	model	model	NOUN
dem-1058	66	19	and	and	CCONJ
dem-1058	66	20	random	random	ADJ
dem-1058	66	21	walk	walk	NOUN
dem-1058	66	22	model	model	NOUN
dem-1058	66	23	(	(	PUNCT
dem-1058	66	24	rw	rw	NOUN
dem-1058	66	25	)	)	PUNCT
dem-1058	66	26	as	as	ADP
dem-1058	66	27	a	a	DET
dem-1058	66	28	benchmark	benchmark	NOUN
dem-1058	66	29	model	model	NOUN
dem-1058	66	30	;	;	PUNCT
dem-1058	66	31	for	for	ADP
dem-1058	66	32	industry	industry	NOUN
dem-1058	66	33	production	production	NOUN
dem-1058	66	34	ip	ip	ADP
dem-1058	66	35	this	this	DET
dem-1058	66	36	set	set	NOUN
dem-1058	66	37	was	be	AUX
dem-1058	66	38	as	as	SCONJ
dem-1058	66	39	follows	follow	VERB
dem-1058	66	40	:	:	PUNCT
dem-1058	66	41	linear	linear	ADJ
dem-1058	66	42	trend	trend	NOUN
dem-1058	66	43	model	model	NOUN
dem-1058	66	44	with	with	ADP
dem-1058	66	45	autoregression	autoregression	NOUN
dem-1058	66	46	of	of	ADP
dem-1058	66	47	twelfth	twelfth	ADJ
dem-1058	66	48	order	order	NOUN
dem-1058	66	49	(	(	PUNCT
dem-1058	66	50	further	far	ADV
dem-1058	66	51	denoted	denote	VERB
dem-1058	66	52	as	as	ADP
dem-1058	66	53	ar	ar	NOUN
dem-1058	66	54	for	for	ADP
dem-1058	66	55	the	the	DET
dem-1058	66	56	convenience	convenience	NOUN
dem-1058	66	57	of	of	ADP
dem-1058	66	58	presentation	presentation	NOUN
dem-1058	66	59	)	)	PUNCT
dem-1058	66	60	,	,	PUNCT
dem-1058	66	61	arima(12,1,0	arima(12,1,0	ADJ
dem-1058	66	62	)	)	PUNCT
dem-1058	66	63	and	and	CCONJ
dem-1058	66	64	random	random	ADJ
dem-1058	66	65	walk	walk	NOUN
dem-1058	66	66	model	model	NOUN
dem-1058	66	67	as	as	ADP
dem-1058	66	68	a	a	DET
dem-1058	66	69	benchmark	benchmark	NOUN
dem-1058	66	70	model5	model5	NOUN
dem-1058	66	71	.	.	PUNCT
dem-1058	67	1	two	two	NUM
dem-1058	67	2	versions	version	NOUN
dem-1058	67	3	of	of	ADP
dem-1058	67	4	estimation	estimation	NOUN
dem-1058	67	5	procedures	procedure	NOUN
dem-1058	67	6	are	be	AUX
dem-1058	67	7	considered	consider	VERB
dem-1058	67	8	:	:	PUNCT
dem-1058	67	9	version	version	NOUN
dem-1058	68	1	i	i	PRON
dem-1058	68	2	:	:	PUNCT
dem-1058	68	3	the	the	DET
dem-1058	68	4	models	model	NOUN
dem-1058	68	5	are	be	AUX
dem-1058	68	6	iteratively	iteratively	ADV
dem-1058	68	7	estimated	estimate	VERB
dem-1058	68	8	beginning	begin	VERB
dem-1058	68	9	with	with	ADP
dem-1058	68	10	the	the	DET
dem-1058	68	11	initial	initial	ADJ
dem-1058	68	12	sample	sample	NOUN
dem-1058	68	13	size	size	NOUN
dem-1058	68	14	n(s	n(s	PROPN
dem-1058	68	15	)	)	PUNCT
dem-1058	68	16	which	which	PRON
dem-1058	68	17	is	be	AUX
dem-1058	68	18	being	be	AUX
dem-1058	68	19	increased	increase	VERB
dem-1058	68	20	by	by	ADP
dem-1058	68	21	one	one	NUM
dem-1058	68	22	until	until	ADP
dem-1058	68	23	n	n	NOUN
dem-1058	68	24	=	=	SYM
dem-1058	68	25	108	108	NUM
dem-1058	68	26	(	(	PUNCT
dem-1058	68	27	until	until	ADP
dem-1058	68	28	2010:12	2010:12	NUM
dem-1058	68	29	)	)	PUNCT
dem-1058	68	30	;	;	PUNCT
dem-1058	68	31	three	three	NUM
dem-1058	68	32	sizes	size	NOUN
dem-1058	68	33	of	of	ADP
dem-1058	68	34	initial	initial	ADJ
dem-1058	68	35	sample	sample	NOUN
dem-1058	68	36	n(s	n(s	PROPN
dem-1058	68	37	)	)	PUNCT
dem-1058	68	38	are	be	AUX
dem-1058	68	39	taken	take	VERB
dem-1058	68	40	:	:	PUNCT
dem-1058	68	41	40	40	NUM
dem-1058	68	42	,	,	PUNCT
dem-1058	68	43	60	60	NUM
dem-1058	68	44	,	,	PUNCT
dem-1058	68	45	80	80	NUM
dem-1058	68	46	;	;	PUNCT
dem-1058	68	47	version	version	PROPN
dem-1058	68	48	ii	ii	PROPN
dem-1058	68	49	:	:	PUNCT
dem-1058	68	50	the	the	DET
dem-1058	68	51	models	model	NOUN
dem-1058	68	52	are	be	AUX
dem-1058	68	53	iteratively	iteratively	ADV
dem-1058	68	54	estimated	estimate	VERB
dem-1058	68	55	for	for	ADP
dem-1058	68	56	rolling	rolling	ADJ
dem-1058	68	57	window	window	NOUN
dem-1058	68	58	size	size	NOUN
dem-1058	68	59	of	of	ADP
dem-1058	68	60	40	40	NUM
dem-1058	68	61	,	,	PUNCT
dem-1058	68	62	60	60	NUM
dem-1058	68	63	and	and	CCONJ
dem-1058	68	64	80	80	NUM
dem-1058	68	65	observations	observation	NOUN
dem-1058	68	66	.	.	PUNCT
dem-1058	69	1	the	the	DET
dem-1058	69	2	question	question	NOUN
dem-1058	69	3	is	be	AUX
dem-1058	69	4	to	to	PART
dem-1058	69	5	what	what	DET
dem-1058	69	6	extent	extent	NOUN
dem-1058	69	7	the	the	DET
dem-1058	69	8	size	size	NOUN
dem-1058	69	9	of	of	ADP
dem-1058	69	10	initial	initial	ADJ
dem-1058	69	11	sample	sample	NOUN
dem-1058	69	12	and	and	CCONJ
dem-1058	69	13	rolling	rolling	ADJ
dem-1058	69	14	window	window	NOUN
dem-1058	69	15	have	have	VERB
dem-1058	69	16	the	the	DET
dem-1058	69	17	impact	impact	NOUN
dem-1058	69	18	on	on	ADP
dem-1058	69	19	the	the	DET
dem-1058	69	20	choice	choice	NOUN
dem-1058	69	21	of	of	ADP
dem-1058	69	22	forecast	forecast	NOUN
dem-1058	69	23	model	model	NOUN
dem-1058	69	24	when	when	SCONJ
dem-1058	69	25	the	the	DET
dem-1058	69	26	information	information	NOUN
dem-1058	69	27	criteria	criteria	VERB
dem-1058	69	28	3	3	NUM
dem-1058	69	29	data	datum	NOUN
dem-1058	69	30	have	have	AUX
dem-1058	69	31	been	be	AUX
dem-1058	69	32	taken	take	VERB
dem-1058	69	33	from	from	ADP
dem-1058	69	34	the	the	DET
dem-1058	69	35	statistical	statistical	ADJ
dem-1058	69	36	bulletin	bulletin	NOUN
dem-1058	69	37	of	of	ADP
dem-1058	69	38	the	the	DET
dem-1058	69	39	central	central	ADJ
dem-1058	69	40	statistical	statistical	ADJ
dem-1058	69	41	office	office	NOUN
dem-1058	69	42	in	in	ADP
dem-1058	69	43	poland	poland	PROPN
dem-1058	69	44	.	.	PUNCT
dem-1058	70	1	4	4	NUM
dem-1058	70	2	many	many	ADJ
dem-1058	70	3	empirical	empirical	ADJ
dem-1058	70	4	studies	study	NOUN
dem-1058	70	5	conclude	conclude	VERB
dem-1058	70	6	that	that	SCONJ
dem-1058	70	7	the	the	DET
dem-1058	70	8	consumer	consumer	NOUN
dem-1058	70	9	price	price	NOUN
dem-1058	70	10	index	index	NOUN
dem-1058	70	11	(	(	PUNCT
dem-1058	70	12	cpi	cpi	PROPN
dem-1058	70	13	)	)	PUNCT
dem-1058	70	14	as	as	ADP
dem-1058	70	15	a	a	DET
dem-1058	70	16	financial	financial	ADJ
dem-1058	70	17	series	series	NOUN
dem-1058	70	18	is	be	AUX
dem-1058	70	19	rather	rather	ADV
dem-1058	70	20	integrated	integrated	ADJ
dem-1058	70	21	and	and	CCONJ
dem-1058	70	22	then	then	ADV
dem-1058	70	23	the	the	DET
dem-1058	70	24	arima	arima	PROPN
dem-1058	70	25	model	model	NOUN
dem-1058	70	26	is	be	AUX
dem-1058	70	27	more	more	ADV
dem-1058	70	28	appropriate	appropriate	ADJ
dem-1058	70	29	than	than	ADP
dem-1058	70	30	the	the	DET
dem-1058	70	31	ar	ar	NOUN
dem-1058	70	32	model	model	NOUN
dem-1058	70	33	(	(	PUNCT
dem-1058	70	34	with	with	ADP
dem-1058	70	35	or	or	CCONJ
dem-1058	70	36	without	without	ADP
dem-1058	70	37	deterministic	deterministic	ADJ
dem-1058	70	38	trend	trend	NOUN
dem-1058	70	39	)	)	PUNCT
dem-1058	70	40	.	.	PUNCT
dem-1058	71	1	whereas	whereas	SCONJ
dem-1058	71	2	,	,	PUNCT
dem-1058	71	3	the	the	DET
dem-1058	71	4	industry	industry	NOUN
dem-1058	71	5	production	production	NOUN
dem-1058	71	6	(	(	PUNCT
dem-1058	71	7	ip	ip	NOUN
dem-1058	71	8	)	)	PUNCT
dem-1058	71	9	is	be	AUX
dem-1058	71	10	treated	treat	VERB
dem-1058	71	11	rather	rather	ADV
dem-1058	71	12	as	as	ADP
dem-1058	71	13	a	a	DET
dem-1058	71	14	stationary	stationary	ADJ
dem-1058	71	15	process	process	NOUN
dem-1058	71	16	around	around	ADP
dem-1058	71	17	deterministic	deterministic	ADJ
dem-1058	71	18	trend	trend	NOUN
dem-1058	71	19	and	and	CCONJ
dem-1058	71	20	then	then	ADV
dem-1058	71	21	the	the	DET
dem-1058	71	22	ar	ar	PROPN
dem-1058	71	23	model	model	NOUN
dem-1058	71	24	(	(	PUNCT
dem-1058	71	25	with	with	ADP
dem-1058	71	26	or	or	CCONJ
dem-1058	71	27	without	without	ADP
dem-1058	71	28	deterministic	deterministic	ADJ
dem-1058	71	29	trend	trend	NOUN
dem-1058	71	30	)	)	PUNCT
dem-1058	71	31	is	be	AUX
dem-1058	71	32	often	often	ADV
dem-1058	71	33	taken	take	VERB
dem-1058	71	34	as	as	ADP
dem-1058	71	35	a	a	DET
dem-1058	71	36	more	more	ADV
dem-1058	71	37	correct	correct	ADJ
dem-1058	71	38	model	model	NOUN
dem-1058	71	39	.	.	PUNCT
dem-1058	72	1	however	however	ADV
dem-1058	72	2	,	,	PUNCT
dem-1058	72	3	it	it	PRON
dem-1058	72	4	may	may	AUX
dem-1058	72	5	occur	occur	VERB
dem-1058	72	6	in	in	ADP
dem-1058	72	7	forecasting	forecast	VERB
dem-1058	72	8	that	that	SCONJ
dem-1058	72	9	an	an	DET
dem-1058	72	10	inappropriate	inappropriate	ADJ
dem-1058	72	11	model	model	NOUN
dem-1058	72	12	will	will	AUX
dem-1058	72	13	yield	yield	VERB
dem-1058	72	14	better	well	ADJ
dem-1058	72	15	forecasts	forecast	NOUN
dem-1058	72	16	.	.	PUNCT
dem-1058	73	1	to	to	PART
dem-1058	73	2	take	take	VERB
dem-1058	73	3	into	into	ADP
dem-1058	73	4	account	account	NOUN
dem-1058	73	5	this	this	DET
dem-1058	73	6	possibility	possibility	NOUN
dem-1058	73	7	,	,	PUNCT
dem-1058	73	8	both	both	DET
dem-1058	73	9	models	model	NOUN
dem-1058	73	10	are	be	AUX
dem-1058	73	11	used	use	VERB
dem-1058	73	12	in	in	ADP
dem-1058	73	13	empirical	empirical	ADJ
dem-1058	73	14	study	study	NOUN
dem-1058	73	15	.	.	PUNCT
dem-1058	74	1	5	5	NUM
dem-1058	74	2	the	the	DET
dem-1058	74	3	order	order	NOUN
dem-1058	74	4	of	of	ADP
dem-1058	74	5	autoregression	autoregression	NOUN
dem-1058	74	6	was	be	AUX
dem-1058	74	7	fixed	fix	VERB
dem-1058	74	8	at	at	ADP
dem-1058	74	9	12	12	NUM
dem-1058	74	10	as	as	ADP
dem-1058	74	11	the	the	DET
dem-1058	74	12	potentially	potentially	ADV
dem-1058	74	13	highest	high	ADJ
dem-1058	74	14	order	order	NOUN
dem-1058	74	15	reflecting	reflect	VERB
dem-1058	74	16	monthly	monthly	ADJ
dem-1058	74	17	frequency	frequency	NOUN
dem-1058	74	18	of	of	ADP
dem-1058	74	19	data	data	PROPN
dem-1058	74	20	.	.	PUNCT
dem-1058	75	1	mariola	mariola	PROPN
dem-1058	75	2	piłatowska	piłatowska	PROPN
dem-1058	75	3	26	26	NUM
dem-1058	75	4	(	(	PUNCT
dem-1058	75	5	aic	aic	PROPN
dem-1058	75	6	,	,	PUNCT
dem-1058	75	7	bic	bic	PROPN
dem-1058	75	8	)	)	PUNCT
dem-1058	75	9	and	and	CCONJ
dem-1058	75	10	prediction	prediction	NOUN
dem-1058	75	11	criteria	criterion	NOUN
dem-1058	75	12	(	(	PUNCT
dem-1058	75	13	ape_se	ape_se	PROPN
dem-1058	75	14	,	,	PUNCT
dem-1058	75	15	ape_ae	ape_ae	NUM
dem-1058	75	16	)	)	PUNCT
dem-1058	75	17	are	be	AUX
dem-1058	75	18	used	use	VERB
dem-1058	75	19	as	as	ADP
dem-1058	75	20	selection	selection	NOUN
dem-1058	75	21	criteria	criterion	NOUN
dem-1058	75	22	.	.	PUNCT
dem-1058	76	1	notations	notation	NOUN
dem-1058	76	2	ape_se	ape_se	X
dem-1058	76	3	and	and	CCONJ
dem-1058	76	4	ape_ae	ape_ae	PUNCT
dem-1058	76	5	denote	denote	VERB
dem-1058	76	6	the	the	DET
dem-1058	76	7	accumulated	accumulate	VERB
dem-1058	76	8	prediction	prediction	NOUN
dem-1058	76	9	error	error	NOUN
dem-1058	76	10	(	(	PUNCT
dem-1058	76	11	ape	ape	NOUN
dem-1058	76	12	)	)	PUNCT
dem-1058	76	13	that	that	PRON
dem-1058	76	14	uses	use	VERB
dem-1058	76	15	the	the	DET
dem-1058	76	16	squared	square	VERB
dem-1058	76	17	error	error	NOUN
dem-1058	76	18	and	and	CCONJ
dem-1058	76	19	absolute	absolute	ADJ
dem-1058	76	20	error	error	NOUN
dem-1058	76	21	respectively	respectively	ADV
dem-1058	76	22	as	as	ADP
dem-1058	76	23	a	a	DET
dem-1058	76	24	loss	loss	NOUN
dem-1058	76	25	function	function	NOUN
dem-1058	76	26	.	.	PUNCT
dem-1058	77	1	to	to	PART
dem-1058	77	2	make	make	VERB
dem-1058	77	3	a	a	DET
dem-1058	77	4	choice	choice	NOUN
dem-1058	77	5	of	of	ADP
dem-1058	77	6	best	good	ADJ
dem-1058	77	7	forecasting	forecasting	NOUN
dem-1058	77	8	model	model	NOUN
dem-1058	77	9	the	the	DET
dem-1058	77	10	information	information	NOUN
dem-1058	77	11	criteria	criterion	NOUN
dem-1058	77	12	(	(	PUNCT
dem-1058	77	13	aic	aic	PROPN
dem-1058	77	14	and	and	CCONJ
dem-1058	77	15	bic	bic	PROPN
dem-1058	77	16	)	)	PUNCT
dem-1058	77	17	and	and	CCONJ
dem-1058	77	18	prediction	prediction	NOUN
dem-1058	77	19	criteria	criterion	NOUN
dem-1058	77	20	(	(	PUNCT
dem-1058	77	21	ape_se	ape_se	X
dem-1058	77	22	and	and	CCONJ
dem-1058	77	23	ape_ae	ape_ae	NUM
dem-1058	77	24	)	)	PUNCT
dem-1058	77	25	were	be	AUX
dem-1058	77	26	calculated	calculate	VERB
dem-1058	77	27	at	at	ADP
dem-1058	77	28	each	each	DET
dem-1058	77	29	iteration	iteration	NOUN
dem-1058	77	30	.	.	PUNCT
dem-1058	78	1	the	the	DET
dem-1058	78	2	results	result	NOUN
dem-1058	78	3	are	be	AUX
dem-1058	78	4	presented	present	VERB
dem-1058	78	5	in	in	ADP
dem-1058	78	6	figures	figure	NOUN
dem-1058	78	7	1−8	1−8	NUM
dem-1058	78	8	as	as	ADP
dem-1058	78	9	difference	difference	NOUN
dem-1058	78	10	in	in	ADP
dem-1058	78	11	a	a	DET
dem-1058	78	12	given	give	VERB
dem-1058	78	13	criterion	criterion	NOUN
dem-1058	78	14	for	for	ADP
dem-1058	78	15	pairs	pair	NOUN
dem-1058	78	16	of	of	ADP
dem-1058	78	17	models	model	NOUN
dem-1058	78	18	,	,	PUNCT
dem-1058	78	19	i.e.	i.e.	X
dem-1058	78	20	aic(mi)−aic(mj	aic(mi)−aic(mj	X
dem-1058	78	21	)	)	PUNCT
dem-1058	78	22	,	,	PUNCT
dem-1058	78	23	bic(mi)−bic(mj	bic(mi)−bic(mj	NOUN
dem-1058	78	24	)	)	PUNCT
dem-1058	78	25	,	,	PUNCT
dem-1058	78	26	ape_se(mi)−ape_se(mj	ape_se(mi)−ape_se(mj	X
dem-1058	78	27	)	)	PUNCT
dem-1058	78	28	,	,	PUNCT
dem-1058	78	29	ape_ae(mi)−ape_ae(mj	ape_ae(mi)−ape_ae(mj	NOUN
dem-1058	78	30	)	)	PUNCT
dem-1058	78	31	,	,	PUNCT
dem-1058	78	32	and	and	CCONJ
dem-1058	78	33	also	also	ADV
dem-1058	78	34	in	in	ADP
dem-1058	78	35	tables	table	NOUN
dem-1058	78	36	presenting	present	VERB
dem-1058	78	37	the	the	DET
dem-1058	78	38	choices	choice	NOUN
dem-1058	78	39	of	of	ADP
dem-1058	78	40	forecasting	forecasting	NOUN
dem-1058	78	41	model	model	NOUN
dem-1058	78	42	for	for	ADP
dem-1058	78	43	all	all	DET
dem-1058	78	44	pairs	pair	NOUN
dem-1058	78	45	of	of	ADP
dem-1058	78	46	models	model	NOUN
dem-1058	78	47	.	.	PUNCT
dem-1058	79	1	these	these	DET
dem-1058	79	2	differences	difference	NOUN
dem-1058	79	3	in	in	ADP
dem-1058	79	4	selection	selection	NOUN
dem-1058	79	5	criteria	criterion	NOUN
dem-1058	79	6	are	be	AUX
dem-1058	79	7	interpreted	interpret	VERB
dem-1058	79	8	as	as	SCONJ
dem-1058	79	9	follows	follow	VERB
dem-1058	79	10	:	:	PUNCT
dem-1058	79	11	the	the	DET
dem-1058	79	12	positive	positive	ADJ
dem-1058	79	13	differences	difference	NOUN
dem-1058	79	14	favor	favor	VERB
dem-1058	79	15	the	the	DET
dem-1058	79	16	second	second	ADJ
dem-1058	79	17	model	model	NOUN
dem-1058	79	18	over	over	ADP
dem-1058	79	19	the	the	DET
dem-1058	79	20	first	first	ADJ
dem-1058	79	21	one	one	NUM
dem-1058	79	22	in	in	ADP
dem-1058	79	23	a	a	DET
dem-1058	79	24	pair	pair	NOUN
dem-1058	79	25	of	of	ADP
dem-1058	79	26	models	model	NOUN
dem-1058	79	27	(	(	PUNCT
dem-1058	79	28	this	this	PRON
dem-1058	79	29	means	mean	VERB
dem-1058	79	30	either	either	CCONJ
dem-1058	79	31	a	a	DET
dem-1058	79	32	lower	low	ADJ
dem-1058	79	33	value	value	NOUN
dem-1058	79	34	of	of	ADP
dem-1058	79	35	information	information	NOUN
dem-1058	79	36	criterion	criterion	NOUN
dem-1058	79	37	or	or	CCONJ
dem-1058	79	38	smaller	small	ADJ
dem-1058	79	39	prediction	prediction	NOUN
dem-1058	79	40	error	error	NOUN
dem-1058	79	41	for	for	ADP
dem-1058	79	42	the	the	DET
dem-1058	79	43	second	second	ADJ
dem-1058	79	44	model	model	NOUN
dem-1058	79	45	)	)	PUNCT
dem-1058	79	46	,	,	PUNCT
dem-1058	79	47	and	and	CCONJ
dem-1058	79	48	the	the	DET
dem-1058	79	49	negative	negative	ADJ
dem-1058	79	50	differences	difference	NOUN
dem-1058	79	51	indicate	indicate	VERB
dem-1058	79	52	that	that	SCONJ
dem-1058	79	53	the	the	DET
dem-1058	79	54	first	first	ADJ
dem-1058	79	55	model	model	NOUN
dem-1058	79	56	outperforms	outperform	VERB
dem-1058	79	57	the	the	DET
dem-1058	79	58	second	second	ADJ
dem-1058	79	59	one	one	NUM
dem-1058	79	60	.	.	PUNCT
dem-1058	80	1	the	the	DET
dem-1058	80	2	sign	sign	NOUN
dem-1058	80	3	of	of	ADP
dem-1058	80	4	differences	difference	NOUN
dem-1058	80	5	in	in	ADP
dem-1058	80	6	criteria	criterion	NOUN
dem-1058	80	7	may	may	AUX
dem-1058	80	8	change	change	VERB
dem-1058	80	9	in	in	ADP
dem-1058	80	10	time	time	NOUN
dem-1058	80	11	which	which	PRON
dem-1058	80	12	indicates	indicate	VERB
dem-1058	80	13	that	that	SCONJ
dem-1058	80	14	one	one	NUM
dem-1058	80	15	model	model	NOUN
dem-1058	80	16	has	have	AUX
dem-1058	80	17	become	become	VERB
dem-1058	80	18	outdated	outdated	ADJ
dem-1058	80	19	.	.	PUNCT
dem-1058	81	1	however	however	ADV
dem-1058	81	2	,	,	PUNCT
dem-1058	81	3	from	from	ADP
dem-1058	81	4	the	the	DET
dem-1058	81	5	forecasting	forecasting	NOUN
dem-1058	81	6	point	point	NOUN
dem-1058	81	7	of	of	ADP
dem-1058	81	8	view	view	NOUN
dem-1058	81	9	the	the	DET
dem-1058	81	10	most	most	ADV
dem-1058	81	11	important	important	ADJ
dem-1058	81	12	is	be	AUX
dem-1058	81	13	the	the	DET
dem-1058	81	14	sign	sign	NOUN
dem-1058	81	15	of	of	ADP
dem-1058	81	16	differences	difference	NOUN
dem-1058	81	17	in	in	ADP
dem-1058	81	18	criteria	criterion	NOUN
dem-1058	81	19	at	at	ADP
dem-1058	81	20	the	the	DET
dem-1058	81	21	end	end	NOUN
dem-1058	81	22	of	of	ADP
dem-1058	81	23	studied	study	VERB
dem-1058	81	24	period	period	NOUN
dem-1058	81	25	,	,	PUNCT
dem-1058	81	26	therefore	therefore	ADV
dem-1058	81	27	the	the	DET
dem-1058	81	28	choice	choice	NOUN
dem-1058	81	29	of	of	ADP
dem-1058	81	30	the	the	DET
dem-1058	81	31	best	good	ADJ
dem-1058	81	32	forecasting	forecasting	NOUN
dem-1058	81	33	model	model	NOUN
dem-1058	81	34	has	have	AUX
dem-1058	81	35	been	be	AUX
dem-1058	81	36	made	make	VERB
dem-1058	81	37	basing	base	VERB
dem-1058	81	38	on	on	ADP
dem-1058	81	39	the	the	DET
dem-1058	81	40	performance	performance	NOUN
dem-1058	81	41	of	of	ADP
dem-1058	81	42	differences	difference	NOUN
dem-1058	81	43	in	in	ADP
dem-1058	81	44	criteria	criterion	NOUN
dem-1058	81	45	at	at	ADP
dem-1058	81	46	the	the	DET
dem-1058	81	47	end	end	NOUN
dem-1058	81	48	of	of	ADP
dem-1058	81	49	sample	sample	NOUN
dem-1058	81	50	(	(	PUNCT
dem-1058	81	51	at	at	ADP
dem-1058	81	52	least	least	ADV
dem-1058	81	53	three	three	NUM
dem-1058	81	54	observations	observation	NOUN
dem-1058	81	55	with	with	ADP
dem-1058	81	56	the	the	DET
dem-1058	81	57	same	same	ADJ
dem-1058	81	58	sign	sign	NOUN
dem-1058	81	59	of	of	ADP
dem-1058	81	60	difference	difference	NOUN
dem-1058	81	61	in	in	ADP
dem-1058	81	62	a	a	DET
dem-1058	81	63	given	give	VERB
dem-1058	81	64	criterion	criterion	NOUN
dem-1058	81	65	)	)	PUNCT
dem-1058	81	66	.	.	PUNCT
dem-1058	82	1	figure	figure	NOUN
dem-1058	82	2	1	1	NUM
dem-1058	82	3	(	(	PUNCT
dem-1058	82	4	row	row	NOUN
dem-1058	82	5	1	1	NUM
dem-1058	82	6	)	)	PUNCT
dem-1058	82	7	demonstrates	demonstrate	VERB
dem-1058	82	8	that	that	SCONJ
dem-1058	82	9	independently	independently	ADV
dem-1058	82	10	of	of	ADP
dem-1058	82	11	initial	initial	ADJ
dem-1058	82	12	sample	sample	NOUN
dem-1058	82	13	size	size	NOUN
dem-1058	82	14	the	the	DET
dem-1058	82	15	performance	performance	NOUN
dem-1058	82	16	of	of	ADP
dem-1058	82	17	differences	difference	NOUN
dem-1058	82	18	in	in	ADP
dem-1058	82	19	aic	aic	PROPN
dem-1058	82	20	criterion	criterion	NOUN
dem-1058	82	21	for	for	ADP
dem-1058	82	22	pairs	pair	NOUN
dem-1058	82	23	of	of	ADP
dem-1058	82	24	models	model	NOUN
dem-1058	82	25	is	be	AUX
dem-1058	82	26	similar	similar	ADJ
dem-1058	82	27	,	,	PUNCT
dem-1058	82	28	i.e.	i.e.	X
dem-1058	82	29	the	the	DET
dem-1058	82	30	aic	aic	PROPN
dem-1058	82	31	criterion	criterion	NOUN
dem-1058	82	32	favors	favor	VERB
dem-1058	82	33	the	the	DET
dem-1058	82	34	ar	ar	PROPN
dem-1058	82	35	model	model	NOUN
dem-1058	82	36	over	over	ADP
dem-1058	82	37	the	the	DET
dem-1058	82	38	arima	arima	PROPN
dem-1058	82	39	and	and	CCONJ
dem-1058	82	40	rw	rw	NOUN
dem-1058	82	41	models	model	NOUN
dem-1058	82	42	,	,	PUNCT
dem-1058	82	43	and	and	CCONJ
dem-1058	82	44	the	the	DET
dem-1058	82	45	arima	arima	PROPN
dem-1058	82	46	model	model	NOUN
dem-1058	82	47	is	be	AUX
dem-1058	82	48	better	well	ADJ
dem-1058	82	49	in	in	ADP
dem-1058	82	50	sense	sense	NOUN
dem-1058	82	51	of	of	ADP
dem-1058	82	52	aic	aic	PROPN
dem-1058	82	53	criterion	criterion	NOUN
dem-1058	82	54	than	than	ADP
dem-1058	82	55	the	the	DET
dem-1058	82	56	rw	rw	PROPN
dem-1058	82	57	model	model	NOUN
dem-1058	82	58	.	.	PUNCT
dem-1058	83	1	however	however	ADV
dem-1058	83	2	,	,	PUNCT
dem-1058	83	3	the	the	DET
dem-1058	83	4	results	result	NOUN
dem-1058	83	5	for	for	ADP
dem-1058	83	6	different	different	ADJ
dem-1058	83	7	rolling	rolling	ADJ
dem-1058	83	8	window	window	NOUN
dem-1058	83	9	sizes	size	NOUN
dem-1058	83	10	are	be	AUX
dem-1058	83	11	different	different	ADJ
dem-1058	83	12	(	(	PUNCT
dem-1058	83	13	fig	fig	NOUN
dem-1058	83	14	.	.	NOUN
dem-1058	83	15	1	1	NUM
dem-1058	83	16	,	,	PUNCT
dem-1058	83	17	row	row	NOUN
dem-1058	83	18	2	2	NUM
dem-1058	83	19	)	)	PUNCT
dem-1058	83	20	.	.	PUNCT
dem-1058	84	1	while	while	SCONJ
dem-1058	84	2	the	the	DET
dem-1058	84	3	dominance	dominance	NOUN
dem-1058	84	4	of	of	ADP
dem-1058	84	5	the	the	DET
dem-1058	84	6	ar	ar	PROPN
dem-1058	84	7	model	model	NOUN
dem-1058	84	8	over	over	ADP
dem-1058	84	9	arima	arima	PROPN
dem-1058	84	10	model	model	NOUN
dem-1058	84	11	is	be	AUX
dem-1058	84	12	maintained	maintain	VERB
dem-1058	84	13	,	,	PUNCT
dem-1058	84	14	the	the	DET
dem-1058	84	15	aic	aic	PROPN
dem-1058	84	16	criterion	criterion	NOUN
dem-1058	84	17	prefers	prefer	VERB
dem-1058	84	18	the	the	DET
dem-1058	84	19	rw	rw	PROPN
dem-1058	84	20	model	model	NOUN
dem-1058	84	21	over	over	ADP
dem-1058	84	22	ar	ar	PROPN
dem-1058	84	23	and	and	CCONJ
dem-1058	84	24	arima	arima	PROPN
dem-1058	84	25	model	model	NOUN
dem-1058	84	26	(	(	PUNCT
dem-1058	84	27	differences	difference	NOUN
dem-1058	84	28	in	in	ADP
dem-1058	84	29	aic	aic	PROPN
dem-1058	84	30	for	for	ADP
dem-1058	84	31	pairs	pair	NOUN
dem-1058	84	32	of	of	ADP
dem-1058	84	33	models	model	NOUN
dem-1058	84	34	,	,	PUNCT
dem-1058	84	35	aic(arima)−aic(rw	aic(arima)−aic(rw	NOUN
dem-1058	84	36	)	)	PUNCT
dem-1058	84	37	and	and	CCONJ
dem-1058	84	38	aic(ar)−aic(rw	aic(ar)−aic(rw	ADJ
dem-1058	84	39	)	)	PUNCT
dem-1058	84	40	,	,	PUNCT
dem-1058	84	41	are	be	AUX
dem-1058	84	42	positive	positive	ADJ
dem-1058	84	43	)	)	PUNCT
dem-1058	84	44	what	what	PRON
dem-1058	84	45	is	be	AUX
dem-1058	84	46	opposite	opposite	ADJ
dem-1058	84	47	to	to	ADP
dem-1058	84	48	results	result	NOUN
dem-1058	84	49	obtained	obtain	VERB
dem-1058	84	50	for	for	ADP
dem-1058	84	51	increasing	increase	VERB
dem-1058	84	52	by	by	ADP
dem-1058	84	53	one	one	NUM
dem-1058	84	54	sample	sample	NOUN
dem-1058	84	55	size	size	NOUN
dem-1058	84	56	(	(	PUNCT
dem-1058	84	57	fig	fig	NOUN
dem-1058	84	58	.	.	PUNCT
dem-1058	85	1	1	1	NUM
dem-1058	85	2	,	,	PUNCT
dem-1058	85	3	row	row	NOUN
dem-1058	85	4	1	1	NUM
dem-1058	85	5	)	)	PUNCT
dem-1058	85	6	.	.	PUNCT
dem-1058	86	1	the	the	DET
dem-1058	86	2	different	different	ADJ
dem-1058	86	3	results	result	NOUN
dem-1058	86	4	for	for	ADP
dem-1058	86	5	window	window	NOUN
dem-1058	86	6	size	size	NOUN
dem-1058	86	7	of	of	ADP
dem-1058	86	8	80	80	NUM
dem-1058	86	9	observations	observation	NOUN
dem-1058	86	10	in	in	ADP
dem-1058	86	11	comparison	comparison	NOUN
dem-1058	86	12	with	with	ADP
dem-1058	86	13	those	those	PRON
dem-1058	86	14	obtained	obtain	VERB
dem-1058	86	15	for	for	ADP
dem-1058	86	16	window	window	NOUN
dem-1058	86	17	size	size	NOUN
dem-1058	86	18	of	of	ADP
dem-1058	86	19	40	40	NUM
dem-1058	86	20	and	and	CCONJ
dem-1058	86	21	60	60	NUM
dem-1058	86	22	observations	observation	NOUN
dem-1058	86	23	could	could	AUX
dem-1058	86	24	suggest	suggest	VERB
dem-1058	86	25	the	the	DET
dem-1058	86	26	influence	influence	NOUN
dem-1058	86	27	of	of	ADP
dem-1058	86	28	window	window	NOUN
dem-1058	86	29	size	size	NOUN
dem-1058	86	30	on	on	ADP
dem-1058	86	31	a	a	DET
dem-1058	86	32	choice	choice	NOUN
dem-1058	86	33	of	of	ADP
dem-1058	86	34	model	model	NOUN
dem-1058	86	35	by	by	ADP
dem-1058	86	36	the	the	DET
dem-1058	86	37	aic	aic	PROPN
dem-1058	86	38	criterion	criterion	NOUN
dem-1058	86	39	,	,	PUNCT
dem-1058	86	40	but	but	CCONJ
dem-1058	86	41	in	in	ADP
dem-1058	86	42	that	that	DET
dem-1058	86	43	case	case	NOUN
dem-1058	86	44	it	it	PRON
dem-1058	86	45	is	be	AUX
dem-1058	86	46	rather	rather	ADV
dem-1058	86	47	the	the	DET
dem-1058	86	48	problem	problem	NOUN
dem-1058	86	49	of	of	ADP
dem-1058	86	50	too	too	ADV
dem-1058	86	51	large	large	ADJ
dem-1058	86	52	window	window	NOUN
dem-1058	86	53	size	size	NOUN
dem-1058	86	54	with	with	ADP
dem-1058	86	55	regard	regard	NOUN
dem-1058	86	56	to	to	ADP
dem-1058	86	57	the	the	DET
dem-1058	86	58	number	number	NOUN
dem-1058	86	59	of	of	ADP
dem-1058	86	60	observations	observation	NOUN
dem-1058	86	61	.	.	PUNCT
dem-1058	87	1	the	the	DET
dem-1058	87	2	results	result	NOUN
dem-1058	87	3	of	of	ADP
dem-1058	87	4	model	model	NOUN
dem-1058	87	5	selection	selection	NOUN
dem-1058	87	6	for	for	ADP
dem-1058	87	7	consumer	consumer	NOUN
dem-1058	87	8	price	price	NOUN
dem-1058	87	9	index	index	NOUN
dem-1058	87	10	(	(	PUNCT
dem-1058	87	11	cpi	cpi	PROPN
dem-1058	87	12	)	)	PUNCT
dem-1058	87	13	by	by	ADP
dem-1058	87	14	the	the	DET
dem-1058	87	15	bic	bic	PROPN
dem-1058	87	16	criterion	criterion	NOUN
dem-1058	87	17	seem	seem	VERB
dem-1058	87	18	to	to	PART
dem-1058	87	19	be	be	AUX
dem-1058	87	20	more	more	ADV
dem-1058	87	21	stable	stable	ADJ
dem-1058	87	22	and	and	CCONJ
dem-1058	87	23	insensitive	insensitive	ADJ
dem-1058	87	24	as	as	ADV
dem-1058	87	25	well	well	ADV
dem-1058	87	26	to	to	ADP
dem-1058	87	27	the	the	DET
dem-1058	87	28	initial	initial	ADJ
dem-1058	87	29	sample	sample	NOUN
dem-1058	87	30	size	size	NOUN
dem-1058	87	31	(	(	PUNCT
dem-1058	87	32	fig	fig	NOUN
dem-1058	87	33	.	.	PUNCT
dem-1058	88	1	2	2	NUM
dem-1058	88	2	,	,	PUNCT
dem-1058	88	3	row	row	NOUN
dem-1058	88	4	1	1	NUM
dem-1058	88	5	)	)	PUNCT
dem-1058	88	6	as	as	ADP
dem-1058	88	7	rolling	roll	VERB
dem-1058	88	8	window	window	NOUN
dem-1058	88	9	size	size	NOUN
dem-1058	88	10	(	(	PUNCT
dem-1058	88	11	fig	fig	NOUN
dem-1058	88	12	.	.	PUNCT
dem-1058	89	1	2	2	NUM
dem-1058	89	2	,	,	PUNCT
dem-1058	89	3	row	row	NOUN
dem-1058	89	4	2	2	NUM
dem-1058	89	5	)	)	PUNCT
dem-1058	89	6	in	in	ADP
dem-1058	89	7	comparison	comparison	NOUN
dem-1058	89	8	with	with	ADP
dem-1058	89	9	the	the	DET
dem-1058	89	10	selection	selection	NOUN
dem-1058	89	11	by	by	ADP
dem-1058	89	12	the	the	DET
dem-1058	89	13	aic	aic	PROPN
dem-1058	89	14	criterion	criterion	NOUN
dem-1058	89	15	.	.	PUNCT
dem-1058	90	1	the	the	DET
dem-1058	90	2	bic	bic	PROPN
dem-1058	90	3	criterion	criterion	NOUN
dem-1058	90	4	favors	favor	VERB
dem-1058	90	5	the	the	DET
dem-1058	90	6	rw	rw	PROPN
dem-1058	90	7	model	model	NOUN
dem-1058	90	8	over	over	ADP
dem-1058	90	9	the	the	DET
dem-1058	90	10	ar	ar	PROPN
dem-1058	90	11	and	and	CCONJ
dem-1058	90	12	arima	arima	PROPN
dem-1058	90	13	models	model	NOUN
dem-1058	90	14	.	.	PUNCT
dem-1058	91	1	as	as	SCONJ
dem-1058	91	2	previously	previously	ADV
dem-1058	91	3	the	the	DET
dem-1058	91	4	ar	ar	NOUN
dem-1058	91	5	is	be	AUX
dem-1058	91	6	preferred	prefer	VERB
dem-1058	91	7	over	over	ADP
dem-1058	91	8	the	the	DET
dem-1058	91	9	arima	arima	PROPN
dem-1058	91	10	model	model	NOUN
dem-1058	91	11	.	.	PUNCT
dem-1058	92	1	summing	sum	VERB
dem-1058	92	2	up	up	ADP
dem-1058	92	3	,	,	PUNCT
dem-1058	92	4	the	the	DET
dem-1058	92	5	choices	choice	NOUN
dem-1058	92	6	by	by	ADP
dem-1058	92	7	information	information	NOUN
dem-1058	92	8	criterion	criterion	NOUN
dem-1058	92	9	differ	differ	VERB
dem-1058	92	10	,	,	PUNCT
dem-1058	92	11	i.e.	i.e.	X
dem-1058	92	12	the	the	DET
dem-1058	92	13	aic	aic	PROPN
dem-1058	92	14	criterion	criterion	NOUN
dem-1058	92	15	prefers	prefer	VERB
dem-1058	92	16	the	the	DET
dem-1058	92	17	ar	ar	NOUN
dem-1058	92	18	model	model	NOUN
dem-1058	92	19	for	for	ADP
dem-1058	92	20	cpi	cpi	PROPN
dem-1058	92	21	in	in	ADP
dem-1058	92	22	the	the	DET
dem-1058	92	23	case	case	NOUN
dem-1058	92	24	of	of	ADP
dem-1058	92	25	increasing	increase	VERB
dem-1058	92	26	sample	sample	NOUN
dem-1058	92	27	size	size	NOUN
dem-1058	92	28	(	(	PUNCT
dem-1058	92	29	version	version	NOUN
dem-1058	92	30	information	information	NOUN
dem-1058	92	31	and	and	CCONJ
dem-1058	92	32	prediction	prediction	NOUN
dem-1058	92	33	criteria	criterion	NOUN
dem-1058	92	34	in	in	ADP
dem-1058	92	35	selecting	select	VERB
dem-1058	92	36	the	the	DET
dem-1058	92	37	forecasting	forecasting	NOUN
dem-1058	92	38	model	model	NOUN
dem-1058	92	39	27	27	NUM
dem-1058	92	40	i	i	NOUN
dem-1058	92	41	)	)	PUNCT
dem-1058	92	42	,	,	PUNCT
dem-1058	92	43	and	and	CCONJ
dem-1058	92	44	model	model	NOUN
dem-1058	92	45	rw	rw	PROPN
dem-1058	92	46	−	−	PROPN
dem-1058	92	47	in	in	ADP
dem-1058	92	48	the	the	DET
dem-1058	92	49	case	case	NOUN
dem-1058	92	50	of	of	ADP
dem-1058	92	51	rolling	rolling	ADJ
dem-1058	92	52	window	window	NOUN
dem-1058	92	53	size	size	NOUN
dem-1058	92	54	,	,	PUNCT
dem-1058	92	55	version	version	NOUN
dem-1058	92	56	ii	ii	PROPN
dem-1058	92	57	(	(	PUNCT
dem-1058	92	58	except	except	SCONJ
dem-1058	92	59	the	the	DET
dem-1058	92	60	window	window	NOUN
dem-1058	92	61	size	size	NOUN
dem-1058	92	62	of	of	ADP
dem-1058	92	63	80	80	NUM
dem-1058	92	64	observations	observation	NOUN
dem-1058	92	65	)	)	PUNCT
dem-1058	92	66	,	,	PUNCT
dem-1058	92	67	while	while	SCONJ
dem-1058	92	68	the	the	DET
dem-1058	92	69	bic	bic	PROPN
dem-1058	92	70	criterion	criterion	NOUN
dem-1058	92	71	prefers	prefer	VERB
dem-1058	92	72	the	the	DET
dem-1058	92	73	rw	rw	PROPN
dem-1058	92	74	model	model	NOUN
dem-1058	92	75	in	in	ADP
dem-1058	92	76	both	both	DET
dem-1058	92	77	version	version	NOUN
dem-1058	92	78	.	.	PUNCT
dem-1058	93	1	figures	figure	NOUN
dem-1058	93	2	3	3	NUM
dem-1058	93	3	and	and	CCONJ
dem-1058	93	4	4	4	NUM
dem-1058	93	5	demonstrate	demonstrate	VERB
dem-1058	93	6	the	the	DET
dem-1058	93	7	differences	difference	NOUN
dem-1058	93	8	in	in	ADP
dem-1058	93	9	prediction	prediction	NOUN
dem-1058	93	10	criteria	criterion	NOUN
dem-1058	93	11	ape_se	ape_se	X
dem-1058	93	12	and	and	CCONJ
dem-1058	93	13	ape_ae	ape_ae	NUM
dem-1058	93	14	respectively	respectively	ADV
dem-1058	93	15	.	.	PUNCT
dem-1058	94	1	the	the	DET
dem-1058	94	2	ape_se	ape_se	NOUN
dem-1058	94	3	criterion	criterion	NOUN
dem-1058	94	4	favors	favor	VERB
dem-1058	94	5	the	the	DET
dem-1058	94	6	rw	rw	PROPN
dem-1058	94	7	model	model	NOUN
dem-1058	94	8	over	over	ADP
dem-1058	94	9	the	the	DET
dem-1058	94	10	ar	ar	PROPN
dem-1058	94	11	and	and	CCONJ
dem-1058	94	12	arima	arima	PROPN
dem-1058	94	13	model	model	NOUN
dem-1058	94	14	(	(	PUNCT
dem-1058	94	15	fig	fig	NOUN
dem-1058	94	16	.	.	PUNCT
dem-1058	95	1	3	3	NUM
dem-1058	95	2	,	,	PUNCT
dem-1058	95	3	row	row	NOUN
dem-1058	95	4	1	1	NUM
dem-1058	95	5	and	and	CCONJ
dem-1058	95	6	2	2	NUM
dem-1058	95	7	)	)	PUNCT
dem-1058	95	8	except	except	SCONJ
dem-1058	95	9	the	the	DET
dem-1058	95	10	initial	initial	ADJ
dem-1058	95	11	sample	sample	NOUN
dem-1058	95	12	size	size	NOUN
dem-1058	95	13	and	and	CCONJ
dem-1058	95	14	rolling	rolling	ADJ
dem-1058	95	15	window	window	NOUN
dem-1058	95	16	size	size	NOUN
dem-1058	95	17	of	of	ADP
dem-1058	95	18	80	80	NUM
dem-1058	95	19	observations	observation	NOUN
dem-1058	95	20	when	when	SCONJ
dem-1058	95	21	the	the	DET
dem-1058	95	22	arima	arima	NOUN
dem-1058	95	23	model	model	NOUN
dem-1058	95	24	is	be	AUX
dem-1058	95	25	preferred	prefer	VERB
dem-1058	95	26	over	over	ADP
dem-1058	95	27	the	the	DET
dem-1058	95	28	rw	rw	NOUN
dem-1058	95	29	and	and	CCONJ
dem-1058	95	30	ar	ar	NOUN
dem-1058	95	31	models	model	NOUN
dem-1058	95	32	.	.	PUNCT
dem-1058	96	1	the	the	DET
dem-1058	96	2	lack	lack	NOUN
dem-1058	96	3	of	of	ADP
dem-1058	96	4	support	support	NOUN
dem-1058	96	5	for	for	ADP
dem-1058	96	6	the	the	DET
dem-1058	96	7	rw	rw	PROPN
dem-1058	96	8	model	model	NOUN
dem-1058	96	9	in	in	ADP
dem-1058	96	10	sample	sample	NOUN
dem-1058	96	11	of	of	ADP
dem-1058	96	12	80	80	NUM
dem-1058	96	13	observations	observation	NOUN
dem-1058	96	14	shows	show	VERB
dem-1058	96	15	rather	rather	ADV
dem-1058	96	16	the	the	DET
dem-1058	96	17	influence	influence	NOUN
dem-1058	96	18	of	of	ADP
dem-1058	96	19	initial	initial	ADJ
dem-1058	96	20	sample	sample	NOUN
dem-1058	96	21	size	size	NOUN
dem-1058	96	22	and	and	CCONJ
dem-1058	96	23	rolling	rolling	ADJ
dem-1058	96	24	window	window	NOUN
dem-1058	96	25	size	size	NOUN
dem-1058	96	26	on	on	ADP
dem-1058	96	27	selecting	select	VERB
dem-1058	96	28	the	the	DET
dem-1058	96	29	model	model	NOUN
dem-1058	96	30	.	.	PUNCT
dem-1058	97	1	hence	hence	ADV
dem-1058	97	2	,	,	PUNCT
dem-1058	97	3	the	the	DET
dem-1058	97	4	size	size	NOUN
dem-1058	97	5	of	of	ADP
dem-1058	97	6	initial	initial	ADJ
dem-1058	97	7	sample	sample	NOUN
dem-1058	97	8	or	or	CCONJ
dem-1058	97	9	rolling	rolling	ADJ
dem-1058	97	10	window	window	NOUN
dem-1058	97	11	should	should	AUX
dem-1058	97	12	not	not	PART
dem-1058	97	13	be	be	AUX
dem-1058	97	14	to	to	PART
dem-1058	97	15	large	large	VERB
dem-1058	97	16	with	with	ADP
dem-1058	97	17	regard	regard	NOUN
dem-1058	97	18	to	to	ADP
dem-1058	97	19	total	total	ADJ
dem-1058	97	20	number	number	NOUN
dem-1058	97	21	of	of	ADP
dem-1058	97	22	observations	observation	NOUN
dem-1058	97	23	.	.	PUNCT
dem-1058	98	1	observing	observe	VERB
dem-1058	98	2	the	the	DET
dem-1058	98	3	differences	difference	NOUN
dem-1058	98	4	in	in	ADP
dem-1058	98	5	ape_ae	ape_ae	ADP
dem-1058	98	6	the	the	DET
dem-1058	98	7	influence	influence	NOUN
dem-1058	98	8	of	of	ADP
dem-1058	98	9	the	the	DET
dem-1058	98	10	size	size	NOUN
dem-1058	98	11	of	of	ADP
dem-1058	98	12	initial	initial	ADJ
dem-1058	98	13	sample	sample	NOUN
dem-1058	98	14	and	and	CCONJ
dem-1058	98	15	rolling	rolling	ADJ
dem-1058	98	16	window	window	NOUN
dem-1058	98	17	is	be	AUX
dem-1058	98	18	much	much	ADV
dem-1058	98	19	more	more	ADV
dem-1058	98	20	distinct	distinct	ADJ
dem-1058	98	21	(	(	PUNCT
dem-1058	98	22	fig	fig	NOUN
dem-1058	98	23	.	.	PUNCT
dem-1058	99	1	4	4	NUM
dem-1058	99	2	,	,	PUNCT
dem-1058	99	3	row	row	NOUN
dem-1058	99	4	1	1	NUM
dem-1058	99	5	and	and	CCONJ
dem-1058	99	6	2	2	NUM
dem-1058	99	7	)	)	PUNCT
dem-1058	99	8	.	.	PUNCT
dem-1058	100	1	for	for	ADP
dem-1058	100	2	initial	initial	ADJ
dem-1058	100	3	sample	sample	NOUN
dem-1058	100	4	size	size	NOUN
dem-1058	100	5	and	and	CCONJ
dem-1058	100	6	window	window	NOUN
dem-1058	100	7	size	size	NOUN
dem-1058	100	8	of	of	ADP
dem-1058	100	9	40	40	NUM
dem-1058	100	10	observations	observation	NOUN
dem-1058	100	11	the	the	DET
dem-1058	100	12	ape_ae	ape_ae	NOUN
dem-1058	100	13	criterion	criterion	NOUN
dem-1058	100	14	prefers	prefer	VERB
dem-1058	100	15	the	the	DET
dem-1058	100	16	rw	rw	PROPN
dem-1058	100	17	model	model	NOUN
dem-1058	100	18	over	over	ADP
dem-1058	100	19	the	the	DET
dem-1058	100	20	arima	arima	PROPN
dem-1058	100	21	and	and	CCONJ
dem-1058	100	22	ar	ar	NOUN
dem-1058	100	23	models	model	NOUN
dem-1058	100	24	.	.	PUNCT
dem-1058	101	1	however	however	ADV
dem-1058	101	2	,	,	PUNCT
dem-1058	101	3	for	for	ADP
dem-1058	101	4	sample	sample	NOUN
dem-1058	101	5	(	(	PUNCT
dem-1058	101	6	or	or	CCONJ
dem-1058	101	7	window	window	NOUN
dem-1058	101	8	)	)	PUNCT
dem-1058	101	9	of	of	ADP
dem-1058	101	10	60	60	NUM
dem-1058	101	11	observations	observation	NOUN
dem-1058	101	12	this	this	DET
dem-1058	101	13	criterion	criterion	NOUN
dem-1058	101	14	favors	favor	VERB
dem-1058	101	15	the	the	DET
dem-1058	101	16	arima	arima	PROPN
dem-1058	101	17	model	model	NOUN
dem-1058	101	18	over	over	ADP
dem-1058	101	19	the	the	DET
dem-1058	101	20	ar	ar	PROPN
dem-1058	101	21	and	and	CCONJ
dem-1058	101	22	rw	rw	NOUN
dem-1058	101	23	models	model	NOUN
dem-1058	101	24	,	,	PUNCT
dem-1058	101	25	and	and	CCONJ
dem-1058	101	26	for	for	ADP
dem-1058	101	27	sample	sample	NOUN
dem-1058	101	28	(	(	PUNCT
dem-1058	101	29	or	or	CCONJ
dem-1058	101	30	window	window	NOUN
dem-1058	101	31	)	)	PUNCT
dem-1058	101	32	of	of	ADP
dem-1058	101	33	80	80	NUM
dem-1058	101	34	observations	observation	NOUN
dem-1058	101	35	–	–	PUNCT
dem-1058	101	36	the	the	DET
dem-1058	101	37	ar	ar	NOUN
dem-1058	101	38	model	model	NOUN
dem-1058	101	39	over	over	ADP
dem-1058	101	40	the	the	DET
dem-1058	101	41	arima	arima	PROPN
dem-1058	101	42	and	and	CCONJ
dem-1058	101	43	rw	rw	NOUN
dem-1058	101	44	models	model	NOUN
dem-1058	101	45	.	.	PUNCT
dem-1058	102	1	table	table	NOUN
dem-1058	102	2	1	1	NUM
dem-1058	102	3	.	.	PUNCT
dem-1058	103	1	results	result	NOUN
dem-1058	103	2	of	of	ADP
dem-1058	103	3	model	model	NOUN
dem-1058	103	4	selection	selection	NOUN
dem-1058	103	5	for	for	ADP
dem-1058	103	6	cpi	cpi	PROPN
dem-1058	103	7	in	in	ADP
dem-1058	103	8	poland	poland	PROPN
dem-1058	103	9	using	use	VERB
dem-1058	103	10	information	information	NOUN
dem-1058	103	11	(	(	PUNCT
dem-1058	103	12	aic	aic	PROPN
dem-1058	103	13	,	,	PUNCT
dem-1058	103	14	bic	bic	PROPN
dem-1058	103	15	)	)	PUNCT
dem-1058	103	16	and	and	CCONJ
dem-1058	103	17	prediction	prediction	NOUN
dem-1058	103	18	criteria	criterion	NOUN
dem-1058	103	19	(	(	PUNCT
dem-1058	103	20	ape_se	ape_se	PROPN
dem-1058	103	21	,	,	PUNCT
dem-1058	103	22	ape_ae	ape_ae	NUM
dem-1058	103	23	)	)	PUNCT
dem-1058	103	24	pairs	pair	NOUN
dem-1058	103	25	of	of	ADP
dem-1058	103	26	models	model	NOUN
dem-1058	103	27	selection	selection	NOUN
dem-1058	103	28	criteria	criterion	NOUN
dem-1058	103	29	version	version	NOUN
dem-1058	104	1	i	i	PRON
dem-1058	104	2	version	version	VERB
dem-1058	104	3	ii	ii	PROPN
dem-1058	104	4	n=40	n=40	PROPN
dem-1058	104	5	n=60	n=60	X
dem-1058	104	6	n=80	n=80	X
dem-1058	104	7	n=40	n=40	PROPN
dem-1058	104	8	n=60	n=60	X
dem-1058	104	9	n=80	n=80	PROPN
dem-1058	104	10	arima	arima	PROPN
dem-1058	104	11	vs.	vs.	ADP
dem-1058	104	12	ar	ar	PROPN
dem-1058	104	13	aic	aic	PROPN
dem-1058	104	14	ar	ar	PROPN
dem-1058	104	15	ar	ar	PROPN
dem-1058	104	16	ar	ar	PROPN
dem-1058	104	17	ar	ar	PROPN
dem-1058	104	18	ar	ar	PROPN
dem-1058	104	19	ar	ar	PROPN
dem-1058	104	20	bic	bic	PROPN
dem-1058	104	21	ar	ar	PROPN
dem-1058	104	22	ar	ar	PROPN
dem-1058	104	23	ar	ar	PROPN
dem-1058	104	24	ar	ar	PROPN
dem-1058	104	25	ar	ar	PROPN
dem-1058	104	26	ar	ar	PROPN
dem-1058	104	27	ape_se	ape_se	PROPN
dem-1058	105	1	arima	arima	PROPN
dem-1058	105	2	arima	arima	PROPN
dem-1058	105	3	arima	arima	PROPN
dem-1058	105	4	ar	ar	PROPN
dem-1058	105	5	arima	arima	PROPN
dem-1058	105	6	arima	arima	PROPN
dem-1058	105	7	ape_ae	ape_ae	PROPN
dem-1058	105	8	ar	ar	PROPN
dem-1058	105	9	arima	arima	PROPN
dem-1058	105	10	ar	ar	PROPN
dem-1058	105	11	ar	ar	PROPN
dem-1058	105	12	arima	arima	PROPN
dem-1058	105	13	ar	ar	PROPN
dem-1058	105	14	arima	arima	PROPN
dem-1058	105	15	vs.	vs.	ADP
dem-1058	105	16	rw	rw	NOUN
dem-1058	105	17	aic	aic	PROPN
dem-1058	105	18	arima	arima	PROPN
dem-1058	105	19	arima	arima	PROPN
dem-1058	105	20	arima	arima	PROPN
dem-1058	105	21	rw	rw	PROPN
dem-1058	105	22	rw	rw	PROPN
dem-1058	105	23	arima	arima	PROPN
dem-1058	105	24	bic	bic	PROPN
dem-1058	105	25	rw	rw	PROPN
dem-1058	105	26	rw	rw	PROPN
dem-1058	105	27	rw	rw	PROPN
dem-1058	105	28	rw	rw	PROPN
dem-1058	105	29	rw	rw	PROPN
dem-1058	105	30	rw	rw	PROPN
dem-1058	105	31	ape_se	ape_se	PROPN
dem-1058	105	32	rw	rw	PROPN
dem-1058	105	33	rw	rw	PROPN
dem-1058	105	34	arima	arima	PROPN
dem-1058	105	35	rw	rw	PROPN
dem-1058	105	36	rw	rw	PROPN
dem-1058	105	37	arima	arima	PROPN
dem-1058	105	38	ape_ae	ape_ae	X
dem-1058	105	39	rw	rw	PROPN
dem-1058	105	40	arima	arima	PROPN
dem-1058	105	41	arima	arima	PROPN
dem-1058	105	42	rw	rw	PROPN
dem-1058	105	43	arima	arima	PROPN
dem-1058	105	44	arima	arima	PROPN
dem-1058	105	45	ar	ar	PROPN
dem-1058	105	46	vs.	vs.	PROPN
dem-1058	105	47	rw	rw	PROPN
dem-1058	105	48	aic	aic	PROPN
dem-1058	105	49	ar	ar	PROPN
dem-1058	105	50	ar	ar	PROPN
dem-1058	105	51	ar	ar	PROPN
dem-1058	105	52	rw	rw	PROPN
dem-1058	105	53	rw	rw	PROPN
dem-1058	105	54	ar	ar	PROPN
dem-1058	105	55	bic	bic	PROPN
dem-1058	105	56	rw	rw	PROPN
dem-1058	105	57	rw	rw	PROPN
dem-1058	105	58	rw	rw	PROPN
dem-1058	105	59	rw	rw	PROPN
dem-1058	105	60	rw	rw	PROPN
dem-1058	105	61	rw	rw	PROPN
dem-1058	106	1	ape_se	ape_se	PROPN
dem-1058	106	2	rw	rw	PROPN
dem-1058	106	3	rw	rw	PROPN
dem-1058	106	4	ar	ar	PROPN
dem-1058	106	5	rw	rw	PROPN
dem-1058	106	6	rw	rw	PROPN
dem-1058	106	7	ar	ar	PROPN
dem-1058	106	8	ape_ae	ape_ae	PROPN
dem-1058	106	9	rw	rw	PROPN
dem-1058	106	10	ar	ar	PROPN
dem-1058	106	11	ar	ar	PROPN
dem-1058	106	12	rw	rw	PROPN
dem-1058	106	13	ar	ar	PROPN
dem-1058	106	14	ar	ar	PROPN
dem-1058	106	15	mariola	mariola	PROPN
dem-1058	106	16	piłatowska	piłatowska	PROPN
dem-1058	106	17	28	28	NUM
dem-1058	106	18	table	table	NOUN
dem-1058	106	19	2	2	NUM
dem-1058	106	20	.	.	PUNCT
dem-1058	106	21	accuracy	accuracy	NOUN
dem-1058	106	22	measures	measure	NOUN
dem-1058	106	23	for	for	ADP
dem-1058	106	24	one	one	NUM
dem-1058	106	25	-	-	PUNCT
dem-1058	106	26	step	step	NOUN
dem-1058	106	27	-	-	PUNCT
dem-1058	106	28	ahead	ahead	NOUN
dem-1058	106	29	forecasts	forecast	NOUN
dem-1058	106	30	of	of	ADP
dem-1058	106	31	cpi	cpi	NOUN
dem-1058	106	32	from	from	ADP
dem-1058	106	33	different	different	ADJ
dem-1058	106	34	models	model	NOUN
dem-1058	106	35	in	in	ADP
dem-1058	106	36	the	the	DET
dem-1058	106	37	period	period	NOUN
dem-1058	106	38	2011:01−20011:06	2011:01−20011:06	NUM
dem-1058	106	39	−	−	PROPN
dem-1058	106	40	version	version	NOUN
dem-1058	106	41	i	i	PRON
dem-1058	106	42	accuracy	accuracy	NOUN
dem-1058	106	43	measures	measure	NOUN
dem-1058	106	44	models	model	NOUN
dem-1058	106	45	arima	arima	PROPN
dem-1058	106	46	ar	ar	PROPN
dem-1058	106	47	rw	rw	PROPN
dem-1058	106	48	mse	mse	PROPN
dem-1058	106	49	0.34	0.34	NUM
dem-1058	106	50	0.35	0.35	NUM
dem-1058	106	51	0.26	0.26	NUM
dem-1058	106	52	rmse	rmse	NOUN
dem-1058	106	53	0.58	0.58	NUM
dem-1058	106	54	0.59	0.59	NUM
dem-1058	106	55	0.51	0.51	NUM
dem-1058	106	56	u	u	NOUN
dem-1058	106	57	1.31	1.31	NUM
dem-1058	106	58	1.36	1.36	NUM
dem-1058	106	59	1.00	1.00	NUM
dem-1058	106	60	mape	mape	NOUN
dem-1058	106	61	(	(	PUNCT
dem-1058	106	62	%	%	NOUN
dem-1058	106	63	)	)	PUNCT
dem-1058	106	64	0.48	0.48	NUM
dem-1058	106	65	0.46	0.46	NUM
dem-1058	106	66	0.41	0.41	NUM
dem-1058	106	67	table	table	NOUN
dem-1058	106	68	3	3	NUM
dem-1058	106	69	.	.	PUNCT
dem-1058	106	70	accuracy	accuracy	NOUN
dem-1058	106	71	measures	measure	NOUN
dem-1058	106	72	for	for	ADP
dem-1058	106	73	one	one	NUM
dem-1058	106	74	-	-	PUNCT
dem-1058	106	75	step	step	NOUN
dem-1058	106	76	-	-	PUNCT
dem-1058	106	77	ahead	ahead	NOUN
dem-1058	106	78	forecasts	forecast	NOUN
dem-1058	106	79	of	of	ADP
dem-1058	106	80	cpi	cpi	NOUN
dem-1058	106	81	from	from	ADP
dem-1058	106	82	different	different	ADJ
dem-1058	106	83	models	model	NOUN
dem-1058	106	84	in	in	ADP
dem-1058	106	85	the	the	DET
dem-1058	106	86	period	period	NOUN
dem-1058	106	87	2011:01−2011:06	2011:01−2011:06	NUM
dem-1058	106	88	−	−	PROPN
dem-1058	106	89	version	version	PROPN
dem-1058	106	90	ii	ii	PROPN
dem-1058	106	91	accuracy	accuracy	NOUN
dem-1058	106	92	measures	measure	NOUN
dem-1058	106	93	n=40	n=40	ADJ
dem-1058	106	94	n=60	n=60	X
dem-1058	106	95	n=80	n=80	PROPN
dem-1058	106	96	arima	arima	PROPN
dem-1058	106	97	ar	ar	PROPN
dem-1058	106	98	rw	rw	PROPN
dem-1058	106	99	arima	arima	PROPN
dem-1058	106	100	ar	ar	PROPN
dem-1058	106	101	rw	rw	PROPN
dem-1058	106	102	arima	arima	PROPN
dem-1058	106	103	ar	ar	PROPN
dem-1058	106	104	rw	rw	PROPN
dem-1058	106	105	mse	mse	PROPN
dem-1058	106	106	0.314	0.314	NUM
dem-1058	106	107	0.292	0.292	NUM
dem-1058	106	108	0.265	0.265	NUM
dem-1058	106	109	0.305	0.305	NUM
dem-1058	106	110	0.365	0.365	NUM
dem-1058	106	111	0.276	0.276	NUM
dem-1058	106	112	0.297	0.297	NUM
dem-1058	106	113	0.342	0.342	NUM
dem-1058	106	114	0.279	0.279	NUM
dem-1058	106	115	rmse	rmse	NOUN
dem-1058	106	116	0.560	0.560	NUM
dem-1058	106	117	0.540	0.540	NUM
dem-1058	106	118	0.515	0.515	NUM
dem-1058	106	119	0.552	0.552	NUM
dem-1058	106	120	0.604	0.604	NUM
dem-1058	106	121	0.526	0.526	NUM
dem-1058	106	122	0.545	0.545	NUM
dem-1058	106	123	0.585	0.585	NUM
dem-1058	106	124	0.528	0.528	NUM
dem-1058	106	125	u	u	NOUN
dem-1058	106	126	1.183	1.183	NUM
dem-1058	106	127	1.099	1.099	NUM
dem-1058	106	128	1.000	1.000	NUM
dem-1058	106	129	1.104	1.104	NUM
dem-1058	106	130	1.322	1.322	NUM
dem-1058	106	131	1.000	1.000	NUM
dem-1058	106	132	1.063	1.063	NUM
dem-1058	106	133	1.224	1.224	NUM
dem-1058	106	134	1.000	1.000	NUM
dem-1058	106	135	mape	mape	NOUN
dem-1058	106	136	0.443	0.443	NUM
dem-1058	106	137	%	%	NOUN
dem-1058	106	138	0.434	0.434	NUM
dem-1058	106	139	%	%	NOUN
dem-1058	106	140	0.423	0.423	NUM
dem-1058	106	141	%	%	NOUN
dem-1058	106	142	0.453	0.453	NUM
dem-1058	106	143	%	%	NOUN
dem-1058	106	144	0.470	0.470	NUM
dem-1058	106	145	%	%	NOUN
dem-1058	106	146	0.421	0.421	NUM
dem-1058	106	147	%	%	NOUN
dem-1058	106	148	0.444	0.444	NUM
dem-1058	106	149	%	%	NOUN
dem-1058	106	150	0.450	0.450	NUM
dem-1058	106	151	%	%	NOUN
dem-1058	106	152	0.432	0.432	NUM
dem-1058	106	153	%	%	NOUN
dem-1058	106	154	generally	generally	ADV
dem-1058	106	155	,	,	PUNCT
dem-1058	106	156	the	the	DET
dem-1058	106	157	rw	rw	NOUN
dem-1058	106	158	model	model	NOUN
dem-1058	106	159	should	should	AUX
dem-1058	106	160	be	be	AUX
dem-1058	106	161	chosen	choose	VERB
dem-1058	106	162	as	as	ADP
dem-1058	106	163	the	the	DET
dem-1058	106	164	best	good	ADJ
dem-1058	106	165	model	model	NOUN
dem-1058	106	166	for	for	ADP
dem-1058	106	167	cpi	cpi	NOUN
dem-1058	106	168	because	because	SCONJ
dem-1058	106	169	it	it	PRON
dem-1058	106	170	is	be	AUX
dem-1058	106	171	preferred	prefer	VERB
dem-1058	106	172	by	by	ADP
dem-1058	106	173	all	all	DET
dem-1058	106	174	criteria	criterion	NOUN
dem-1058	106	175	in	in	ADP
dem-1058	106	176	the	the	DET
dem-1058	106	177	case	case	NOUN
dem-1058	106	178	of	of	ADP
dem-1058	106	179	rolling	rolling	ADJ
dem-1058	106	180	windows	window	NOUN
dem-1058	106	181	(	(	PUNCT
dem-1058	106	182	except	except	SCONJ
dem-1058	106	183	window	window	NOUN
dem-1058	106	184	of	of	ADP
dem-1058	106	185	80	80	NUM
dem-1058	106	186	observations	observation	NOUN
dem-1058	106	187	for	for	ADP
dem-1058	106	188	ape_se	ape_se	PRON
dem-1058	106	189	and	and	CCONJ
dem-1058	106	190	ape_ae	ape_ae	NUM
dem-1058	106	191	and	and	CCONJ
dem-1058	106	192	window	window	NOUN
dem-1058	106	193	of	of	ADP
dem-1058	106	194	60	60	NUM
dem-1058	106	195	observations	observation	NOUN
dem-1058	106	196	for	for	ADP
dem-1058	106	197	ape_ae	ape_ae	NUM
dem-1058	106	198	which	which	PRON
dem-1058	106	199	seem	seem	VERB
dem-1058	106	200	rather	rather	ADV
dem-1058	106	201	too	too	ADV
dem-1058	106	202	large	large	ADJ
dem-1058	106	203	with	with	ADP
dem-1058	106	204	regard	regard	NOUN
dem-1058	106	205	to	to	ADP
dem-1058	106	206	number	number	NOUN
dem-1058	106	207	of	of	ADP
dem-1058	106	208	observations	observation	NOUN
dem-1058	106	209	)	)	PUNCT
dem-1058	106	210	and	and	CCONJ
dem-1058	106	211	also	also	ADV
dem-1058	106	212	in	in	ADP
dem-1058	106	213	the	the	DET
dem-1058	106	214	case	case	NOUN
dem-1058	106	215	of	of	ADP
dem-1058	106	216	increasing	increase	VERB
dem-1058	106	217	sample	sample	NOUN
dem-1058	106	218	size	size	NOUN
dem-1058	106	219	except	except	SCONJ
dem-1058	106	220	the	the	DET
dem-1058	106	221	aic	aic	PROPN
dem-1058	106	222	criterion	criterion	NOUN
dem-1058	106	223	which	which	PRON
dem-1058	106	224	favored	favor	VERB
dem-1058	106	225	the	the	DET
dem-1058	106	226	ar	ar	NOUN
dem-1058	106	227	model	model	NOUN
dem-1058	106	228	and	and	CCONJ
dem-1058	106	229	prediction	prediction	NOUN
dem-1058	106	230	criteria	criterion	NOUN
dem-1058	106	231	for	for	ADP
dem-1058	106	232	initial	initial	ADJ
dem-1058	106	233	sample	sample	NOUN
dem-1058	106	234	size	size	NOUN
dem-1058	106	235	of	of	ADP
dem-1058	106	236	80	80	NUM
dem-1058	106	237	observations	observation	NOUN
dem-1058	106	238	(	(	PUNCT
dem-1058	106	239	for	for	ADP
dem-1058	106	240	both	both	DET
dem-1058	106	241	criteria	criterion	NOUN
dem-1058	106	242	ape	ape	NOUN
dem-1058	106	243	)	)	PUNCT
dem-1058	106	244	and	and	CCONJ
dem-1058	106	245	of	of	ADP
dem-1058	106	246	60	60	NUM
dem-1058	106	247	observations	observation	NOUN
dem-1058	106	248	(	(	PUNCT
dem-1058	106	249	for	for	ADP
dem-1058	106	250	ape_ae	ape_ae	NUM
dem-1058	106	251	)	)	PUNCT
dem-1058	106	252	−	−	PROPN
dem-1058	106	253	see	see	VERB
dem-1058	106	254	figure	figure	NOUN
dem-1058	106	255	1−4	1−4	NUM
dem-1058	106	256	and	and	CCONJ
dem-1058	106	257	table	table	NOUN
dem-1058	106	258	1	1	NUM
dem-1058	106	259	.	.	PUNCT
dem-1058	107	1	if	if	SCONJ
dem-1058	107	2	the	the	DET
dem-1058	107	3	rw	rw	NOUN
dem-1058	107	4	model	model	NOUN
dem-1058	107	5	is	be	AUX
dem-1058	107	6	really	really	ADV
dem-1058	107	7	the	the	DET
dem-1058	107	8	best	good	ADJ
dem-1058	107	9	one	one	NUM
dem-1058	107	10	,	,	PUNCT
dem-1058	107	11	its	its	PRON
dem-1058	107	12	predictive	predictive	ADJ
dem-1058	107	13	performance	performance	NOUN
dem-1058	107	14	should	should	AUX
dem-1058	107	15	be	be	AUX
dem-1058	107	16	also	also	ADV
dem-1058	107	17	confirmed	confirm	VERB
dem-1058	107	18	in	in	ADP
dem-1058	107	19	out	out	ADP
dem-1058	107	20	-	-	PUNCT
dem-1058	107	21	of	of	ADP
dem-1058	107	22	-	-	PUNCT
dem-1058	107	23	sample	sample	NOUN
dem-1058	107	24	evaluation	evaluation	NOUN
dem-1058	107	25	.	.	PUNCT
dem-1058	108	1	out	out	ADP
dem-1058	108	2	-	-	PUNCT
dem-1058	108	3	of	of	ADP
dem-1058	108	4	-	-	PUNCT
dem-1058	108	5	sample	sample	NOUN
dem-1058	108	6	forecast	forecast	NOUN
dem-1058	108	7	evaluation	evaluation	NOUN
dem-1058	108	8	(	(	PUNCT
dem-1058	108	9	i.e.	i.e.	X
dem-1058	108	10	in	in	ADP
dem-1058	108	11	the	the	DET
dem-1058	108	12	period	period	NOUN
dem-1058	108	13	2011:01−2011:06	2011:01−2011:06	NUM
dem-1058	108	14	)	)	PUNCT
dem-1058	108	15	has	have	AUX
dem-1058	108	16	been	be	AUX
dem-1058	108	17	realized	realize	VERB
dem-1058	108	18	by	by	ADP
dem-1058	108	19	the	the	DET
dem-1058	108	20	measures	measure	NOUN
dem-1058	108	21	of	of	ADP
dem-1058	108	22	accuracy	accuracy	NOUN
dem-1058	108	23	(	(	PUNCT
dem-1058	108	24	mse	mse	NOUN
dem-1058	108	25	,	,	PUNCT
dem-1058	108	26	rmse	rmse	NOUN
dem-1058	108	27	,	,	PUNCT
dem-1058	108	28	u	u	NOUN
dem-1058	108	29	,	,	PUNCT
dem-1058	108	30	mape	mape	NOUN
dem-1058	108	31	)	)	PUNCT
dem-1058	108	32	−	−	NOUN
dem-1058	108	33	see	see	VERB
dem-1058	108	34	table	table	NOUN
dem-1058	108	35	2	2	NUM
dem-1058	108	36	and	and	CCONJ
dem-1058	108	37	3	3	NUM
dem-1058	108	38	.	.	PUNCT
dem-1058	109	1	the	the	DET
dem-1058	109	2	results	result	NOUN
dem-1058	109	3	in	in	ADP
dem-1058	109	4	table	table	NOUN
dem-1058	109	5	2	2	NUM
dem-1058	109	6	and	and	CCONJ
dem-1058	109	7	3	3	NUM
dem-1058	109	8	indicate	indicate	VERB
dem-1058	109	9	that	that	SCONJ
dem-1058	109	10	one	one	NUM
dem-1058	109	11	-	-	PUNCT
dem-1058	109	12	step	step	NOUN
dem-1058	109	13	-	-	PUNCT
dem-1058	109	14	ahead	ahead	NOUN
dem-1058	109	15	forecasts	forecast	NOUN
dem-1058	109	16	of	of	ADP
dem-1058	109	17	cpi	cpi	PROPN
dem-1058	109	18	made	make	VERB
dem-1058	109	19	from	from	ADP
dem-1058	109	20	the	the	DET
dem-1058	109	21	rw	rw	PROPN
dem-1058	109	22	model	model	NOUN
dem-1058	109	23	have	have	VERB
dem-1058	109	24	the	the	DET
dem-1058	109	25	smallest	small	ADJ
dem-1058	109	26	prediction	prediction	NOUN
dem-1058	109	27	errors	error	NOUN
dem-1058	109	28	independently	independently	ADV
dem-1058	109	29	of	of	ADP
dem-1058	109	30	the	the	DET
dem-1058	109	31	versions	version	NOUN
dem-1058	109	32	(	(	PUNCT
dem-1058	109	33	iteratively	iteratively	ADV
dem-1058	109	34	increasing	increase	VERB
dem-1058	109	35	sample	sample	NOUN
dem-1058	109	36	size	size	NOUN
dem-1058	109	37	by	by	ADP
dem-1058	109	38	one	one	NUM
dem-1058	109	39	−	−	PROPN
dem-1058	109	40	version	version	NOUN
dem-1058	109	41	i	i	PRON
dem-1058	109	42	,	,	PUNCT
dem-1058	109	43	or	or	CCONJ
dem-1058	109	44	rolling	rolling	ADJ
dem-1058	109	45	window	window	NOUN
dem-1058	109	46	of	of	ADP
dem-1058	109	47	given	give	VERB
dem-1058	109	48	size	size	NOUN
dem-1058	109	49	−	−	PROPN
dem-1058	109	50	version	version	PROPN
dem-1058	109	51	ii	ii	PROPN
dem-1058	109	52	)	)	PUNCT
dem-1058	109	53	.	.	PUNCT
dem-1058	110	1	the	the	DET
dem-1058	110	2	out	out	ADV
dem-1058	110	3	-	-	PUNCT
dem-1058	110	4	of	of	ADP
dem-1058	110	5	-	-	PUNCT
dem-1058	110	6	sample	sample	NOUN
dem-1058	110	7	performance	performance	NOUN
dem-1058	110	8	of	of	ADP
dem-1058	110	9	rw	rw	NOUN
dem-1058	110	10	model	model	NOUN
dem-1058	110	11	is	be	AUX
dem-1058	110	12	in	in	ADP
dem-1058	110	13	rough	rough	ADJ
dem-1058	110	14	agreement	agreement	NOUN
dem-1058	110	15	with	with	ADP
dem-1058	110	16	choices	choice	NOUN
dem-1058	110	17	received	receive	VERB
dem-1058	110	18	by	by	ADP
dem-1058	110	19	ape	ape	NOUN
dem-1058	110	20	criteria	criterion	NOUN
dem-1058	110	21	but	but	CCONJ
dem-1058	110	22	also	also	ADV
dem-1058	110	23	bic	bic	PROPN
dem-1058	110	24	criterion	criterion	NOUN
dem-1058	110	25	,	,	PUNCT
dem-1058	110	26	thus	thus	ADV
dem-1058	110	27	providing	provide	VERB
dem-1058	110	28	evidence	evidence	NOUN
dem-1058	110	29	of	of	ADP
dem-1058	110	30	correct	correct	ADJ
dem-1058	110	31	model	model	NOUN
dem-1058	110	32	selection	selection	NOUN
dem-1058	110	33	by	by	ADP
dem-1058	110	34	these	these	DET
dem-1058	110	35	criteria	criterion	NOUN
dem-1058	110	36	.	.	PUNCT
dem-1058	111	1	the	the	DET
dem-1058	111	2	results	result	NOUN
dem-1058	111	3	of	of	ADP
dem-1058	111	4	selecting	select	VERB
dem-1058	111	5	the	the	DET
dem-1058	111	6	best	good	ADJ
dem-1058	111	7	model	model	NOUN
dem-1058	111	8	for	for	ADP
dem-1058	111	9	industry	industry	NOUN
dem-1058	111	10	production	production	NOUN
dem-1058	111	11	(	(	PUNCT
dem-1058	111	12	ip	ip	NOUN
dem-1058	111	13	)	)	PUNCT
dem-1058	111	14	in	in	ADP
dem-1058	111	15	poland	poland	PROPN
dem-1058	111	16	are	be	AUX
dem-1058	111	17	presented	present	VERB
dem-1058	111	18	in	in	ADP
dem-1058	111	19	figures	figure	NOUN
dem-1058	111	20	5−8	5−8	NUM
dem-1058	111	21	.	.	PUNCT
dem-1058	112	1	figure	figure	NOUN
dem-1058	112	2	5	5	NUM
dem-1058	112	3	shows	show	VERB
dem-1058	112	4	that	that	SCONJ
dem-1058	112	5	the	the	DET
dem-1058	112	6	aic	aic	PROPN
dem-1058	112	7	criterion	criterion	NOUN
dem-1058	112	8	favors	favor	VERB
dem-1058	112	9	the	the	DET
dem-1058	112	10	ar	ar	PROPN
dem-1058	112	11	model	model	NOUN
dem-1058	112	12	over	over	ADP
dem-1058	112	13	the	the	DET
dem-1058	112	14	arima	arima	PROPN
dem-1058	112	15	and	and	CCONJ
dem-1058	112	16	rw	rw	NOUN
dem-1058	112	17	models	model	NOUN
dem-1058	112	18	independently	independently	ADV
dem-1058	112	19	of	of	ADP
dem-1058	112	20	initial	initial	ADJ
dem-1058	112	21	sample	sample	NOUN
dem-1058	112	22	size	size	NOUN
dem-1058	112	23	and	and	CCONJ
dem-1058	112	24	rolling	rolling	ADJ
dem-1058	112	25	windows	window	NOUN
dem-1058	112	26	.	.	PUNCT
dem-1058	113	1	hence	hence	ADV
dem-1058	113	2	,	,	PUNCT
dem-1058	113	3	in	in	ADP
dem-1058	113	4	that	that	DET
dem-1058	113	5	case	case	NOUN
dem-1058	113	6	the	the	DET
dem-1058	113	7	selection	selection	NOUN
dem-1058	113	8	of	of	ADP
dem-1058	113	9	a	a	DET
dem-1058	113	10	model	model	NOUN
dem-1058	113	11	by	by	ADP
dem-1058	113	12	the	the	DET
dem-1058	113	13	aic	aic	PROPN
dem-1058	113	14	criterion	criterion	NOUN
dem-1058	113	15	seems	seem	VERB
dem-1058	113	16	to	to	PART
dem-1058	113	17	be	be	AUX
dem-1058	113	18	insensitive	insensitive	ADJ
dem-1058	113	19	to	to	ADP
dem-1058	113	20	the	the	DET
dem-1058	113	21	size	size	NOUN
dem-1058	113	22	of	of	ADP
dem-1058	113	23	initial	initial	ADJ
dem-1058	113	24	sample	sample	NOUN
dem-1058	113	25	and	and	CCONJ
dem-1058	113	26	rolling	roll	VERB
dem-1058	113	27	information	information	NOUN
dem-1058	113	28	and	and	CCONJ
dem-1058	113	29	prediction	prediction	NOUN
dem-1058	113	30	criteria	criterion	NOUN
dem-1058	113	31	in	in	ADP
dem-1058	113	32	selecting	select	VERB
dem-1058	113	33	the	the	DET
dem-1058	113	34	forecasting	forecasting	NOUN
dem-1058	113	35	model	model	NOUN
dem-1058	113	36	29	29	NUM
dem-1058	113	37	windows	window	NOUN
dem-1058	113	38	.	.	PUNCT
dem-1058	114	1	when	when	SCONJ
dem-1058	114	2	the	the	DET
dem-1058	114	3	bic	bic	PROPN
dem-1058	114	4	criterion	criterion	NOUN
dem-1058	114	5	has	have	AUX
dem-1058	114	6	been	be	AUX
dem-1058	114	7	used	use	VERB
dem-1058	114	8	the	the	DET
dem-1058	114	9	results	result	NOUN
dem-1058	114	10	were	be	AUX
dem-1058	114	11	similar	similar	ADJ
dem-1058	114	12	but	but	CCONJ
dem-1058	114	13	only	only	ADV
dem-1058	114	14	in	in	ADP
dem-1058	114	15	version	version	NOUN
dem-1058	114	16	i	i	PRON
dem-1058	114	17	(	(	PUNCT
dem-1058	114	18	initial	initial	ADJ
dem-1058	114	19	sample	sample	NOUN
dem-1058	114	20	size	size	NOUN
dem-1058	114	21	increased	increase	VERB
dem-1058	114	22	by	by	ADP
dem-1058	114	23	one	one	NUM
dem-1058	114	24	)	)	PUNCT
dem-1058	114	25	−	−	NOUN
dem-1058	115	1	see	see	VERB
dem-1058	115	2	figure	figure	NOUN
dem-1058	115	3	6	6	NUM
dem-1058	115	4	,	,	PUNCT
dem-1058	115	5	row	row	NOUN
dem-1058	115	6	1	1	NUM
dem-1058	115	7	and	and	CCONJ
dem-1058	115	8	for	for	ADP
dem-1058	115	9	window	window	NOUN
dem-1058	115	10	size	size	NOUN
dem-1058	115	11	of	of	ADP
dem-1058	115	12	80	80	NUM
dem-1058	115	13	observations	observation	NOUN
dem-1058	115	14	(	(	PUNCT
dem-1058	115	15	version	version	PROPN
dem-1058	115	16	ii	ii	PROPN
dem-1058	115	17	,	,	PUNCT
dem-1058	115	18	see	see	VERB
dem-1058	115	19	figure	figure	NOUN
dem-1058	115	20	6	6	NUM
dem-1058	115	21	,	,	PUNCT
dem-1058	115	22	row	row	NOUN
dem-1058	115	23	2	2	NUM
dem-1058	115	24	)	)	PUNCT
dem-1058	115	25	.	.	PUNCT
dem-1058	116	1	for	for	ADP
dem-1058	116	2	window	window	NOUN
dem-1058	116	3	sizes	size	NOUN
dem-1058	116	4	of	of	ADP
dem-1058	116	5	40	40	NUM
dem-1058	116	6	and	and	CCONJ
dem-1058	116	7	60	60	NUM
dem-1058	116	8	observations	observation	NOUN
dem-1058	116	9	the	the	DET
dem-1058	116	10	bic	bic	PROPN
dem-1058	116	11	criterion	criterion	NOUN
dem-1058	116	12	prefers	prefer	VERB
dem-1058	116	13	the	the	DET
dem-1058	116	14	rw	rw	PROPN
dem-1058	116	15	model	model	NOUN
dem-1058	116	16	over	over	ADP
dem-1058	116	17	the	the	DET
dem-1058	116	18	arima	arima	PROPN
dem-1058	116	19	and	and	CCONJ
dem-1058	116	20	ar	ar	NOUN
dem-1058	116	21	model	model	NOUN
dem-1058	116	22	(	(	PUNCT
dem-1058	116	23	figure	figure	NOUN
dem-1058	116	24	6	6	NUM
dem-1058	116	25	,	,	PUNCT
dem-1058	116	26	row	row	NOUN
dem-1058	116	27	2	2	NUM
dem-1058	116	28	)	)	PUNCT
dem-1058	116	29	.	.	PUNCT
dem-1058	117	1	the	the	DET
dem-1058	117	2	choice	choice	NOUN
dem-1058	117	3	of	of	ADP
dem-1058	117	4	model	model	NOUN
dem-1058	117	5	of	of	ADP
dem-1058	117	6	industry	industry	NOUN
dem-1058	117	7	production	production	NOUN
dem-1058	117	8	(	(	PUNCT
dem-1058	117	9	ip	ip	NOUN
dem-1058	117	10	)	)	PUNCT
dem-1058	117	11	in	in	ADP
dem-1058	117	12	poland	poland	PROPN
dem-1058	117	13	by	by	ADP
dem-1058	117	14	prediction	prediction	NOUN
dem-1058	117	15	criteria	criterion	NOUN
dem-1058	117	16	is	be	AUX
dem-1058	117	17	different	different	ADJ
dem-1058	117	18	than	than	ADP
dem-1058	117	19	by	by	ADP
dem-1058	117	20	information	information	NOUN
dem-1058	117	21	criteria	criterion	NOUN
dem-1058	117	22	.	.	PUNCT
dem-1058	118	1	namely	namely	ADV
dem-1058	118	2	,	,	PUNCT
dem-1058	118	3	the	the	DET
dem-1058	118	4	ape_se	ape_se	NOUN
dem-1058	118	5	criterion	criterion	NOUN
dem-1058	118	6	favors	favor	VERB
dem-1058	118	7	the	the	DET
dem-1058	118	8	arima	arima	PROPN
dem-1058	118	9	model	model	NOUN
dem-1058	118	10	over	over	ADP
dem-1058	118	11	the	the	DET
dem-1058	118	12	ar	ar	PROPN
dem-1058	118	13	and	and	CCONJ
dem-1058	118	14	rw	rw	PROPN
dem-1058	118	15	model	model	NOUN
dem-1058	118	16	except	except	SCONJ
dem-1058	118	17	the	the	DET
dem-1058	118	18	initial	initial	ADJ
dem-1058	118	19	sample	sample	NOUN
dem-1058	118	20	of	of	ADP
dem-1058	118	21	80	80	NUM
dem-1058	118	22	observations	observation	NOUN
dem-1058	118	23	(	(	PUNCT
dem-1058	118	24	version	version	NOUN
dem-1058	118	25	i	i	NOUN
dem-1058	118	26	)	)	PUNCT
dem-1058	118	27	and	and	CCONJ
dem-1058	118	28	window	window	NOUN
dem-1058	118	29	size	size	NOUN
dem-1058	118	30	of	of	ADP
dem-1058	118	31	80	80	NUM
dem-1058	118	32	observations	observation	NOUN
dem-1058	118	33	(	(	PUNCT
dem-1058	118	34	version	version	NOUN
dem-1058	118	35	ii	ii	PROPN
dem-1058	118	36	)	)	PUNCT
dem-1058	118	37	−	−	NOUN
dem-1058	118	38	figure	figure	NOUN
dem-1058	118	39	7	7	NUM
dem-1058	118	40	.	.	PUNCT
dem-1058	119	1	this	this	DET
dem-1058	119	2	dominance	dominance	NOUN
dem-1058	119	3	of	of	ADP
dem-1058	119	4	model	model	NOUN
dem-1058	119	5	arima	arima	PROPN
dem-1058	119	6	is	be	AUX
dem-1058	119	7	maintained	maintain	VERB
dem-1058	119	8	when	when	SCONJ
dem-1058	119	9	the	the	DET
dem-1058	119	10	ape_ae	ape_ae	ADJ
dem-1058	119	11	criterion	criterion	NOUN
dem-1058	119	12	is	be	AUX
dem-1058	119	13	used	use	VERB
dem-1058	119	14	to	to	PART
dem-1058	119	15	select	select	VERB
dem-1058	119	16	the	the	DET
dem-1058	119	17	best	good	ADJ
dem-1058	119	18	model	model	NOUN
dem-1058	119	19	(	(	PUNCT
dem-1058	119	20	except	except	SCONJ
dem-1058	119	21	the	the	DET
dem-1058	119	22	initial	initial	ADJ
dem-1058	119	23	sample	sample	NOUN
dem-1058	119	24	size	size	NOUN
dem-1058	119	25	of	of	ADP
dem-1058	119	26	60	60	NUM
dem-1058	119	27	observations	observation	NOUN
dem-1058	119	28	,	,	PUNCT
dem-1058	119	29	version	version	NOUN
dem-1058	119	30	i	i	NOUN
dem-1058	119	31	)	)	PUNCT
dem-1058	119	32	−	−	PROPN
dem-1058	119	33	figure	figure	NOUN
dem-1058	119	34	8	8	NUM
dem-1058	119	35	.	.	PUNCT
dem-1058	119	36	summing	sum	VERB
dem-1058	119	37	up	up	ADP
dem-1058	119	38	,	,	PUNCT
dem-1058	119	39	according	accord	VERB
dem-1058	119	40	to	to	ADP
dem-1058	119	41	information	information	NOUN
dem-1058	119	42	criteria	criterion	NOUN
dem-1058	119	43	(	(	PUNCT
dem-1058	119	44	aic	aic	PROPN
dem-1058	119	45	,	,	PUNCT
dem-1058	119	46	bic	bic	PROPN
dem-1058	119	47	)	)	PUNCT
dem-1058	119	48	the	the	DET
dem-1058	119	49	ar	ar	PROPN
dem-1058	119	50	model	model	NOUN
dem-1058	119	51	should	should	AUX
dem-1058	119	52	be	be	AUX
dem-1058	119	53	chosen	choose	VERB
dem-1058	119	54	as	as	ADP
dem-1058	119	55	the	the	DET
dem-1058	119	56	best	good	ADJ
dem-1058	119	57	model	model	NOUN
dem-1058	119	58	for	for	ADP
dem-1058	119	59	industry	industry	NOUN
dem-1058	119	60	production	production	NOUN
dem-1058	119	61	,	,	PUNCT
dem-1058	119	62	ip	ip	NOUN
dem-1058	119	63	,	,	PUNCT
dem-1058	119	64	(	(	PUNCT
dem-1058	119	65	except	except	SCONJ
dem-1058	119	66	the	the	DET
dem-1058	119	67	choice	choice	NOUN
dem-1058	119	68	of	of	ADP
dem-1058	119	69	bic	bic	ADJ
dem-1058	119	70	criterion	criterion	NOUN
dem-1058	119	71	for	for	ADP
dem-1058	119	72	rolling	rolling	ADJ
dem-1058	119	73	window	window	NOUN
dem-1058	119	74	of	of	ADP
dem-1058	119	75	40	40	NUM
dem-1058	119	76	and	and	CCONJ
dem-1058	119	77	60	60	NUM
dem-1058	119	78	observations	observation	NOUN
dem-1058	119	79	when	when	SCONJ
dem-1058	119	80	the	the	DET
dem-1058	119	81	rw	rw	NOUN
dem-1058	119	82	model	model	NOUN
dem-1058	119	83	is	be	AUX
dem-1058	119	84	preferred	prefer	VERB
dem-1058	119	85	)	)	PUNCT
dem-1058	120	1	−	−	PROPN
dem-1058	120	2	see	see	VERB
dem-1058	120	3	figure	figure	NOUN
dem-1058	120	4	5−6	5−6	NUM
dem-1058	120	5	and	and	CCONJ
dem-1058	120	6	table	table	NOUN
dem-1058	120	7	4	4	NUM
dem-1058	120	8	.	.	PUNCT
dem-1058	121	1	whereas	whereas	SCONJ
dem-1058	121	2	the	the	DET
dem-1058	121	3	choice	choice	NOUN
dem-1058	121	4	of	of	ADP
dem-1058	121	5	prediction	prediction	NOUN
dem-1058	121	6	criteria	criterion	NOUN
dem-1058	121	7	(	(	PUNCT
dem-1058	121	8	ape_se	ape_se	PROPN
dem-1058	121	9	,	,	PUNCT
dem-1058	121	10	ape_ae	ape_ae	NUM
dem-1058	121	11	)	)	PUNCT
dem-1058	121	12	is	be	AUX
dem-1058	121	13	the	the	DET
dem-1058	121	14	arima	arima	PROPN
dem-1058	121	15	model	model	NOUN
dem-1058	121	16	(	(	PUNCT
dem-1058	121	17	except	except	SCONJ
dem-1058	121	18	initial	initial	ADJ
dem-1058	121	19	sample	sample	NOUN
dem-1058	121	20	of	of	ADP
dem-1058	121	21	60	60	NUM
dem-1058	121	22	observations	observation	NOUN
dem-1058	121	23	)	)	PUNCT
dem-1058	121	24	−	−	NOUN
dem-1058	122	1	see	see	VERB
dem-1058	122	2	figure	figure	NOUN
dem-1058	122	3	7	7	NUM
dem-1058	122	4	-	-	SYM
dem-1058	122	5	8	8	NUM
dem-1058	122	6	and	and	CCONJ
dem-1058	122	7	table	table	NOUN
dem-1058	122	8	4	4	NUM
dem-1058	122	9	.	.	PUNCT
dem-1058	122	10	table	table	NOUN
dem-1058	122	11	4	4	NUM
dem-1058	122	12	.	.	PUNCT
dem-1058	122	13	results	result	NOUN
dem-1058	122	14	of	of	ADP
dem-1058	122	15	model	model	NOUN
dem-1058	122	16	selection	selection	NOUN
dem-1058	122	17	for	for	ADP
dem-1058	122	18	ip	ip	NOUN
dem-1058	122	19	in	in	ADP
dem-1058	122	20	poland	poland	NOUN
dem-1058	122	21	using	use	VERB
dem-1058	122	22	information	information	NOUN
dem-1058	122	23	(	(	PUNCT
dem-1058	122	24	aic	aic	PROPN
dem-1058	122	25	,	,	PUNCT
dem-1058	122	26	bic	bic	PROPN
dem-1058	122	27	)	)	PUNCT
dem-1058	122	28	and	and	CCONJ
dem-1058	122	29	prediction	prediction	NOUN
dem-1058	122	30	criteria	criterion	NOUN
dem-1058	122	31	(	(	PUNCT
dem-1058	122	32	ape_se	ape_se	PROPN
dem-1058	122	33	,	,	PUNCT
dem-1058	122	34	ape_ae	ape_ae	NUM
dem-1058	122	35	)	)	PUNCT
dem-1058	122	36	pairs	pair	NOUN
dem-1058	122	37	of	of	ADP
dem-1058	122	38	models	model	NOUN
dem-1058	122	39	selection	selection	NOUN
dem-1058	122	40	criteria	criterion	NOUN
dem-1058	122	41	version	version	NOUN
dem-1058	123	1	i	i	PRON
dem-1058	123	2	version	version	VERB
dem-1058	123	3	ii	ii	PROPN
dem-1058	123	4	n=40	n=40	PROPN
dem-1058	123	5	n=60	n=60	X
dem-1058	123	6	n=80	n=80	X
dem-1058	123	7	n=40	n=40	PROPN
dem-1058	123	8	n=60	n=60	X
dem-1058	123	9	n=80	n=80	PROPN
dem-1058	123	10	arima	arima	PROPN
dem-1058	123	11	vs.	vs.	ADP
dem-1058	123	12	ar	ar	PROPN
dem-1058	123	13	aic	aic	PROPN
dem-1058	123	14	ar	ar	PROPN
dem-1058	123	15	ar	ar	PROPN
dem-1058	123	16	ar	ar	PROPN
dem-1058	123	17	ar	ar	PROPN
dem-1058	123	18	ar	ar	PROPN
dem-1058	123	19	ar	ar	PROPN
dem-1058	123	20	bic	bic	PROPN
dem-1058	123	21	ar	ar	PROPN
dem-1058	123	22	ar	ar	PROPN
dem-1058	123	23	ar	ar	PROPN
dem-1058	123	24	ar	ar	PROPN
dem-1058	123	25	ar	ar	PROPN
dem-1058	123	26	ar	ar	PROPN
dem-1058	123	27	ape_se	ape_se	PROPN
dem-1058	124	1	arima	arima	PROPN
dem-1058	124	2	arima	arima	PROPN
dem-1058	124	3	ar	ar	PROPN
dem-1058	124	4	arima	arima	PROPN
dem-1058	124	5	arima	arima	PROPN
dem-1058	124	6	ar	ar	PROPN
dem-1058	124	7	ape_ae	ape_ae	PROPN
dem-1058	124	8	arima	arima	PROPN
dem-1058	124	9	ar	ar	PROPN
dem-1058	124	10	arima	arima	PROPN
dem-1058	124	11	arima	arima	PROPN
dem-1058	124	12	arima	arima	PROPN
dem-1058	124	13	ar	ar	PROPN
dem-1058	124	14	arima	arima	PROPN
dem-1058	124	15	vs.	vs.	ADP
dem-1058	124	16	rw	rw	NOUN
dem-1058	124	17	aic	aic	PROPN
dem-1058	124	18	arima	arima	PROPN
dem-1058	124	19	arima	arima	PROPN
dem-1058	124	20	arima	arima	PROPN
dem-1058	124	21	arima	arima	PROPN
dem-1058	124	22	arima	arima	PROPN
dem-1058	124	23	arima	arima	PROPN
dem-1058	124	24	bic	bic	PROPN
dem-1058	124	25	arima	arima	PROPN
dem-1058	124	26	arima	arima	PROPN
dem-1058	124	27	arima	arima	PROPN
dem-1058	124	28	rw	rw	PROPN
dem-1058	124	29	rw	rw	PROPN
dem-1058	124	30	arima	arima	PROPN
dem-1058	124	31	ape_se	ape_se	PROPN
dem-1058	125	1	arima	arima	PROPN
dem-1058	125	2	arima	arima	PROPN
dem-1058	125	3	arima	arima	PROPN
dem-1058	125	4	arima	arima	PROPN
dem-1058	125	5	arima	arima	PROPN
dem-1058	125	6	arima	arima	PROPN
dem-1058	125	7	ape_ae	ape_ae	X
dem-1058	125	8	arima	arima	PROPN
dem-1058	125	9	arima	arima	PROPN
dem-1058	125	10	arima	arima	PROPN
dem-1058	125	11	arima	arima	PROPN
dem-1058	125	12	arima	arima	PROPN
dem-1058	125	13	arima	arima	PROPN
dem-1058	125	14	ar	ar	PROPN
dem-1058	125	15	vs.	vs.	PROPN
dem-1058	125	16	rw	rw	PROPN
dem-1058	125	17	aic	aic	PROPN
dem-1058	125	18	ar	ar	PROPN
dem-1058	125	19	ar	ar	PROPN
dem-1058	125	20	ar	ar	PROPN
dem-1058	125	21	ar	ar	PROPN
dem-1058	125	22	ar	ar	PROPN
dem-1058	125	23	ar	ar	PROPN
dem-1058	125	24	bic	bic	PROPN
dem-1058	125	25	ar	ar	PROPN
dem-1058	125	26	ar	ar	PROPN
dem-1058	125	27	ar	ar	PROPN
dem-1058	125	28	rw	rw	PROPN
dem-1058	125	29	rw	rw	PROPN
dem-1058	125	30	ar	ar	PROPN
dem-1058	125	31	ape_se	ape_se	PROPN
dem-1058	125	32	ar	ar	PROPN
dem-1058	125	33	ar	ar	PROPN
dem-1058	125	34	ar	ar	PROPN
dem-1058	125	35	rw	rw	PROPN
dem-1058	125	36	ar	ar	PROPN
dem-1058	125	37	ar	ar	PROPN
dem-1058	125	38	ape_ae	ape_ae	PROPN
dem-1058	125	39	ar	ar	PROPN
dem-1058	125	40	ar	ar	PROPN
dem-1058	125	41	ar	ar	PROPN
dem-1058	125	42	ar	ar	PROPN
dem-1058	125	43	ar	ar	PROPN
dem-1058	125	44	ar	ar	PROPN
dem-1058	125	45	out	out	ADP
dem-1058	125	46	-	-	PUNCT
dem-1058	125	47	of	of	ADP
dem-1058	125	48	-	-	PUNCT
dem-1058	125	49	sample	sample	NOUN
dem-1058	125	50	evaluation	evaluation	NOUN
dem-1058	125	51	of	of	ADP
dem-1058	125	52	ip	ip	PRON
dem-1058	125	53	forecasts	forecast	NOUN
dem-1058	125	54	(	(	PUNCT
dem-1058	125	55	i.e.	i.e.	X
dem-1058	125	56	in	in	ADP
dem-1058	125	57	the	the	DET
dem-1058	125	58	period	period	NOUN
dem-1058	125	59	2011:01−2011:06	2011:01−2011:06	NUM
dem-1058	125	60	)	)	PUNCT
dem-1058	125	61	has	have	AUX
dem-1058	125	62	been	be	AUX
dem-1058	125	63	realized	realize	VERB
dem-1058	125	64	by	by	ADP
dem-1058	125	65	the	the	DET
dem-1058	125	66	measures	measure	NOUN
dem-1058	125	67	of	of	ADP
dem-1058	125	68	accuracy	accuracy	NOUN
dem-1058	125	69	(	(	PUNCT
dem-1058	125	70	mse	mse	NOUN
dem-1058	125	71	,	,	PUNCT
dem-1058	125	72	rmse	rmse	NOUN
dem-1058	125	73	,	,	PUNCT
dem-1058	125	74	u	u	NOUN
dem-1058	125	75	,	,	PUNCT
dem-1058	125	76	mape	mape	NOUN
dem-1058	125	77	)	)	PUNCT
dem-1058	125	78	−	−	NOUN
dem-1058	125	79	see	see	VERB
dem-1058	125	80	table	table	NOUN
dem-1058	125	81	5	5	NUM
dem-1058	125	82	and	and	CCONJ
dem-1058	125	83	6	6	NUM
dem-1058	125	84	.	.	PUNCT
dem-1058	126	1	the	the	DET
dem-1058	126	2	results	result	NOUN
dem-1058	126	3	in	in	ADP
dem-1058	126	4	table	table	NOUN
dem-1058	126	5	5	5	NUM
dem-1058	126	6	and	and	CCONJ
dem-1058	126	7	6	6	NUM
dem-1058	126	8	show	show	VERB
dem-1058	126	9	that	that	SCONJ
dem-1058	126	10	both	both	CCONJ
dem-1058	126	11	the	the	DET
dem-1058	126	12	arima	arima	PROPN
dem-1058	126	13	model	model	NOUN
dem-1058	126	14	and	and	CCONJ
dem-1058	126	15	ar	ar	PROPN
dem-1058	126	16	model	model	NOUN
dem-1058	126	17	have	have	VERB
dem-1058	126	18	similar	similar	ADJ
dem-1058	126	19	predictive	predictive	ADJ
dem-1058	126	20	value	value	NOUN
dem-1058	126	21	because	because	SCONJ
dem-1058	126	22	the	the	DET
dem-1058	126	23	accuracy	accuracy	NOUN
dem-1058	126	24	measures	measure	NOUN
dem-1058	126	25	of	of	ADP
dem-1058	126	26	one	one	NUM
dem-1058	126	27	-	-	PUNCT
dem-1058	126	28	stepmariola	stepmariola	NOUN
dem-1058	126	29	piłatowska	piłatowska	NOUN
dem-1058	126	30	30	30	NUM
dem-1058	126	31	ahead	ahead	ADV
dem-1058	126	32	forecasts	forecast	NOUN
dem-1058	126	33	of	of	ADP
dem-1058	126	34	ip	ip	NOUN
dem-1058	126	35	do	do	AUX
dem-1058	126	36	not	not	PART
dem-1058	126	37	differ	differ	VERB
dem-1058	126	38	much	much	ADJ
dem-1058	126	39	,	,	PUNCT
dem-1058	126	40	so	so	ADV
dem-1058	126	41	these	these	DET
dem-1058	126	42	model	model	NOUN
dem-1058	126	43	may	may	AUX
dem-1058	126	44	complete	complete	VERB
dem-1058	126	45	with	with	ADP
dem-1058	126	46	each	each	DET
dem-1058	126	47	other	other	ADJ
dem-1058	126	48	.	.	PUNCT
dem-1058	127	1	in	in	ADP
dem-1058	127	2	case	case	NOUN
dem-1058	127	3	of	of	ADP
dem-1058	127	4	version	version	NOUN
dem-1058	127	5	i	i	PRON
dem-1058	127	6	(	(	PUNCT
dem-1058	127	7	iteratively	iteratively	ADV
dem-1058	127	8	increasing	increase	VERB
dem-1058	127	9	sample	sample	NOUN
dem-1058	127	10	size	size	NOUN
dem-1058	127	11	by	by	ADP
dem-1058	127	12	one	one	NUM
dem-1058	127	13	)	)	PUNCT
dem-1058	127	14	the	the	DET
dem-1058	127	15	rmse	rmse	NOUN
dem-1058	127	16	and	and	CCONJ
dem-1058	127	17	u	u	PRON
dem-1058	127	18	measure	measure	NOUN
dem-1058	127	19	prefer	prefer	VERB
dem-1058	127	20	the	the	DET
dem-1058	127	21	arima	arima	PROPN
dem-1058	127	22	model	model	NOUN
dem-1058	127	23	as	as	ADP
dem-1058	127	24	a	a	DET
dem-1058	127	25	model	model	NOUN
dem-1058	127	26	with	with	ADP
dem-1058	127	27	the	the	DET
dem-1058	127	28	smallest	small	ADJ
dem-1058	127	29	prediction	prediction	NOUN
dem-1058	127	30	error	error	NOUN
dem-1058	127	31	but	but	CCONJ
dem-1058	127	32	the	the	DET
dem-1058	127	33	mape	mape	NOUN
dem-1058	127	34	indicates	indicate	VERB
dem-1058	127	35	the	the	DET
dem-1058	127	36	ar	ar	NOUN
dem-1058	127	37	model	model	NOUN
dem-1058	127	38	(	(	PUNCT
dem-1058	127	39	table	table	NOUN
dem-1058	127	40	5	5	NUM
dem-1058	127	41	)	)	PUNCT
dem-1058	127	42	.	.	PUNCT
dem-1058	128	1	the	the	DET
dem-1058	128	2	opposite	opposite	ADJ
dem-1058	128	3	result	result	NOUN
dem-1058	128	4	is	be	AUX
dem-1058	128	5	obtained	obtain	VERB
dem-1058	128	6	in	in	ADP
dem-1058	128	7	version	version	PROPN
dem-1058	128	8	ii	ii	PROPN
dem-1058	128	9	for	for	ADP
dem-1058	128	10	rolling	rolling	ADJ
dem-1058	128	11	window	window	NOUN
dem-1058	128	12	of	of	ADP
dem-1058	128	13	40	40	NUM
dem-1058	128	14	observations	observation	NOUN
dem-1058	128	15	(	(	PUNCT
dem-1058	128	16	table	table	NOUN
dem-1058	128	17	6	6	NUM
dem-1058	128	18	)	)	PUNCT
dem-1058	128	19	.	.	PUNCT
dem-1058	129	1	table	table	NOUN
dem-1058	129	2	5	5	NUM
dem-1058	129	3	.	.	PUNCT
dem-1058	130	1	accuracy	accuracy	NOUN
dem-1058	130	2	measures	measure	NOUN
dem-1058	130	3	for	for	ADP
dem-1058	130	4	one	one	NUM
dem-1058	130	5	-	-	PUNCT
dem-1058	130	6	step	step	NOUN
dem-1058	130	7	-	-	PUNCT
dem-1058	130	8	ahead	ahead	NOUN
dem-1058	130	9	forecasts	forecast	NOUN
dem-1058	130	10	of	of	ADP
dem-1058	130	11	ip	ip	NOUN
dem-1058	130	12	from	from	ADP
dem-1058	130	13	different	different	ADJ
dem-1058	130	14	models	model	NOUN
dem-1058	130	15	in	in	ADP
dem-1058	130	16	the	the	DET
dem-1058	130	17	period	period	NOUN
dem-1058	130	18	2011:01−20011:06	2011:01−20011:06	NUM
dem-1058	130	19	−	−	PROPN
dem-1058	130	20	version	version	NOUN
dem-1058	130	21	i	i	PRON
dem-1058	130	22	accuracy	accuracy	NOUN
dem-1058	130	23	measures	measure	NOUN
dem-1058	130	24	models	model	NOUN
dem-1058	130	25	arima	arima	PROPN
dem-1058	130	26	ar	ar	PROPN
dem-1058	130	27	rw	rw	PROPN
dem-1058	130	28	rmse	rmse	PROPN
dem-1058	130	29	2673.30	2673.30	NUM
dem-1058	130	30	2679.50	2679.50	NUM
dem-1058	130	31	3390.70	3390.70	NUM
dem-1058	130	32	u	u	NOUN
dem-1058	130	33	0.622	0.622	NUM
dem-1058	130	34	0.624	0.624	NUM
dem-1058	130	35	1.00	1.00	NUM
dem-1058	130	36	mape	mape	NOUN
dem-1058	130	37	(	(	PUNCT
dem-1058	130	38	%	%	NOUN
dem-1058	130	39	)	)	PUNCT
dem-1058	130	40	2.12	2.12	NUM
dem-1058	130	41	2.03	2.03	NUM
dem-1058	130	42	3.52	3.52	NUM
dem-1058	130	43	table	table	NOUN
dem-1058	130	44	6	6	NUM
dem-1058	130	45	.	.	PUNCT
dem-1058	131	1	accuracy	accuracy	NOUN
dem-1058	131	2	measures	measure	NOUN
dem-1058	131	3	for	for	ADP
dem-1058	131	4	one	one	NUM
dem-1058	131	5	-	-	PUNCT
dem-1058	131	6	step	step	NOUN
dem-1058	131	7	-	-	PUNCT
dem-1058	131	8	ahead	ahead	NOUN
dem-1058	131	9	forecasts	forecast	NOUN
dem-1058	131	10	of	of	ADP
dem-1058	131	11	ip	ip	NOUN
dem-1058	131	12	from	from	ADP
dem-1058	131	13	different	different	ADJ
dem-1058	131	14	models	model	NOUN
dem-1058	131	15	in	in	ADP
dem-1058	131	16	the	the	DET
dem-1058	131	17	period	period	NOUN
dem-1058	131	18	2011:01−2011:06	2011:01−2011:06	NUM
dem-1058	131	19	−	−	PROPN
dem-1058	131	20	version	version	PROPN
dem-1058	131	21	ii	ii	PROPN
dem-1058	131	22	accuracy	accuracy	NOUN
dem-1058	131	23	measures	measure	NOUN
dem-1058	131	24	n=40	n=40	ADJ
dem-1058	131	25	n=60	n=60	X
dem-1058	131	26	n=80	n=80	PROPN
dem-1058	131	27	arima	arima	PROPN
dem-1058	131	28	ar	ar	PROPN
dem-1058	131	29	rw	rw	PROPN
dem-1058	131	30	arima	arima	PROPN
dem-1058	131	31	ar	ar	PROPN
dem-1058	131	32	rw	rw	PROPN
dem-1058	131	33	arima	arima	PROPN
dem-1058	131	34	ar	ar	PROPN
dem-1058	131	35	rw	rw	PROPN
dem-1058	131	36	rmse	rmse	PROPN
dem-1058	131	37	2916.6	2916.6	NUM
dem-1058	131	38	2908.4	2908.4	NUM
dem-1058	131	39	3411.8	3411.8	NUM
dem-1058	131	40	2716.42	2716.42	NUM
dem-1058	131	41	2863.65	2863.65	NUM
dem-1058	131	42	3386.73	3386.73	NUM
dem-1058	131	43	2787.6	2787.6	NUM
dem-1058	131	44	2697.1	2697.1	NUM
dem-1058	131	45	3395.6	3395.6	NUM
dem-1058	131	46	u	u	NOUN
dem-1058	131	47	0.731	0.731	NUM
dem-1058	131	48	0.727	0.727	NUM
dem-1058	131	49	1.000	1.000	NUM
dem-1058	131	50	0.643	0.643	NUM
dem-1058	131	51	0.715	0.715	NUM
dem-1058	131	52	1.000	1.000	NUM
dem-1058	131	53	0.674	0.674	NUM
dem-1058	131	54	0.631	0.631	NUM
dem-1058	131	55	1.00	1.00	NUM
dem-1058	131	56	mape	mape	NOUN
dem-1058	131	57	(	(	PUNCT
dem-1058	131	58	%	%	NOUN
dem-1058	131	59	)	)	PUNCT
dem-1058	131	60	2.390	2.390	NUM
dem-1058	131	61	2.460	2.460	NUM
dem-1058	131	62	3.510	3.510	NUM
dem-1058	131	63	2.12	2.12	NUM
dem-1058	131	64	%	%	NOUN
dem-1058	131	65	2.210	2.210	NUM
dem-1058	131	66	3.500	3.500	NUM
dem-1058	131	67	2.220	2.220	NUM
dem-1058	131	68	1.990	1.990	NUM
dem-1058	131	69	3.520	3.520	NUM
dem-1058	131	70	for	for	ADP
dem-1058	131	71	the	the	DET
dem-1058	131	72	window	window	NOUN
dem-1058	131	73	of	of	ADP
dem-1058	131	74	60	60	NUM
dem-1058	131	75	and	and	CCONJ
dem-1058	131	76	80	80	NUM
dem-1058	131	77	observations	observation	NOUN
dem-1058	131	78	the	the	DET
dem-1058	131	79	arima	arima	PROPN
dem-1058	131	80	model	model	NOUN
dem-1058	131	81	and	and	CCONJ
dem-1058	131	82	ar	ar	PROPN
dem-1058	131	83	model	model	NOUN
dem-1058	131	84	give	give	VERB
dem-1058	131	85	the	the	DET
dem-1058	131	86	smallest	small	ADJ
dem-1058	131	87	prediction	prediction	NOUN
dem-1058	131	88	errors	error	NOUN
dem-1058	131	89	.	.	PUNCT
dem-1058	132	1	it	it	PRON
dem-1058	132	2	is	be	AUX
dem-1058	132	3	worth	worth	ADJ
dem-1058	132	4	noting	note	VERB
dem-1058	132	5	that	that	SCONJ
dem-1058	132	6	the	the	DET
dem-1058	132	7	arima	arima	PROPN
dem-1058	132	8	and	and	CCONJ
dem-1058	132	9	ar	ar	NOUN
dem-1058	132	10	models	model	NOUN
dem-1058	132	11	substantially	substantially	ADV
dem-1058	132	12	outperform	outperform	VERB
dem-1058	132	13	the	the	DET
dem-1058	132	14	rw	rw	PROPN
dem-1058	132	15	model	model	NOUN
dem-1058	132	16	.	.	PUNCT
dem-1058	133	1	on	on	ADP
dem-1058	133	2	the	the	DET
dem-1058	133	3	whole	whole	NOUN
dem-1058	133	4	,	,	PUNCT
dem-1058	133	5	the	the	DET
dem-1058	133	6	predictive	predictive	ADJ
dem-1058	133	7	performance	performance	NOUN
dem-1058	133	8	of	of	ADP
dem-1058	133	9	arima	arima	PROPN
dem-1058	133	10	and	and	CCONJ
dem-1058	133	11	ar	ar	NOUN
dem-1058	133	12	models	model	NOUN
dem-1058	133	13	is	be	AUX
dem-1058	133	14	in	in	ADP
dem-1058	133	15	agreement	agreement	NOUN
dem-1058	133	16	with	with	ADP
dem-1058	133	17	choices	choice	NOUN
dem-1058	133	18	obtained	obtain	VERB
dem-1058	133	19	by	by	ADP
dem-1058	133	20	ape	ape	NOUN
dem-1058	133	21	criteria	criterion	NOUN
dem-1058	133	22	and	and	CCONJ
dem-1058	133	23	aic	aic	PROPN
dem-1058	133	24	criterion	criterion	NOUN
dem-1058	133	25	.	.	PUNCT
dem-1058	134	1	conclusions	conclusion	NOUN
dem-1058	134	2	the	the	DET
dem-1058	134	3	results	result	NOUN
dem-1058	134	4	of	of	ADP
dem-1058	134	5	choosing	choose	VERB
dem-1058	134	6	the	the	DET
dem-1058	134	7	forecasting	forecasting	NOUN
dem-1058	134	8	model	model	NOUN
dem-1058	134	9	on	on	ADP
dem-1058	134	10	empirical	empirical	ADJ
dem-1058	134	11	data	datum	NOUN
dem-1058	134	12	for	for	ADP
dem-1058	134	13	poland	poland	PROPN
dem-1058	134	14	showed	show	VERB
dem-1058	134	15	that	that	SCONJ
dem-1058	134	16	the	the	DET
dem-1058	134	17	size	size	NOUN
dem-1058	134	18	of	of	ADP
dem-1058	134	19	initial	initial	ADJ
dem-1058	134	20	sample	sample	NOUN
dem-1058	134	21	(	(	PUNCT
dem-1058	134	22	version	version	NOUN
dem-1058	134	23	i	i	PROPN
dem-1058	134	24	)	)	PUNCT
dem-1058	134	25	and	and	CCONJ
dem-1058	134	26	rolling	roll	VERB
dem-1058	134	27	windows	window	NOUN
dem-1058	134	28	(	(	PUNCT
dem-1058	134	29	version	version	NOUN
dem-1058	134	30	ii	ii	NOUN
dem-1058	134	31	)	)	PUNCT
dem-1058	134	32	do	do	AUX
dem-1058	134	33	have	have	VERB
dem-1058	134	34	an	an	DET
dem-1058	134	35	impact	impact	NOUN
dem-1058	134	36	on	on	ADP
dem-1058	134	37	the	the	DET
dem-1058	134	38	choice	choice	NOUN
dem-1058	134	39	of	of	ADP
dem-1058	134	40	forecasting	forecasting	NOUN
dem-1058	134	41	model	model	NOUN
dem-1058	134	42	(	(	PUNCT
dem-1058	134	43	within	within	ADP
dem-1058	134	44	a	a	DET
dem-1058	134	45	given	give	VERB
dem-1058	134	46	version	version	NOUN
dem-1058	134	47	)	)	PUNCT
dem-1058	134	48	,	,	PUNCT
dem-1058	134	49	especially	especially	ADV
dem-1058	134	50	it	it	PRON
dem-1058	134	51	concerns	concern	VERB
dem-1058	134	52	the	the	DET
dem-1058	134	53	ape	ape	NOUN
dem-1058	134	54	.	.	PUNCT
dem-1058	135	1	the	the	DET
dem-1058	135	2	size	size	NOUN
dem-1058	135	3	of	of	ADP
dem-1058	135	4	initial	initial	ADJ
dem-1058	135	5	sample	sample	NOUN
dem-1058	135	6	and	and	CCONJ
dem-1058	135	7	rolling	rolling	ADJ
dem-1058	135	8	windows	window	NOUN
dem-1058	135	9	should	should	AUX
dem-1058	135	10	be	be	AUX
dem-1058	135	11	relatively	relatively	ADV
dem-1058	135	12	small	small	ADJ
dem-1058	135	13	with	with	ADP
dem-1058	135	14	regard	regard	NOUN
dem-1058	135	15	to	to	ADP
dem-1058	135	16	sample	sample	NOUN
dem-1058	135	17	size	size	NOUN
dem-1058	135	18	because	because	SCONJ
dem-1058	135	19	too	too	ADV
dem-1058	135	20	large	large	ADJ
dem-1058	135	21	initial	initial	ADJ
dem-1058	135	22	sample	sample	NOUN
dem-1058	135	23	(	(	PUNCT
dem-1058	135	24	or	or	CCONJ
dem-1058	135	25	window	window	NOUN
dem-1058	135	26	)	)	PUNCT
dem-1058	135	27	enables	enable	VERB
dem-1058	135	28	to	to	PART
dem-1058	135	29	follow	follow	VERB
dem-1058	135	30	the	the	DET
dem-1058	135	31	evolution	evolution	NOUN
dem-1058	135	32	of	of	ADP
dem-1058	135	33	ape	ape	NOUN
dem-1058	135	34	for	for	ADP
dem-1058	135	35	sufficient	sufficient	ADJ
dem-1058	135	36	number	number	NOUN
dem-1058	135	37	of	of	ADP
dem-1058	135	38	periods	period	NOUN
dem-1058	135	39	.	.	PUNCT
dem-1058	136	1	there	there	PRON
dem-1058	136	2	are	be	VERB
dem-1058	136	3	no	no	DET
dem-1058	136	4	relevant	relevant	ADJ
dem-1058	136	5	differences	difference	NOUN
dem-1058	136	6	between	between	ADP
dem-1058	136	7	the	the	DET
dem-1058	136	8	selection	selection	NOUN
dem-1058	136	9	of	of	ADP
dem-1058	136	10	forecasting	forecasting	NOUN
dem-1058	136	11	model	model	NOUN
dem-1058	136	12	for	for	ADP
dem-1058	136	13	the	the	DET
dem-1058	136	14	initial	initial	ADJ
dem-1058	136	15	sample	sample	NOUN
dem-1058	136	16	and	and	CCONJ
dem-1058	136	17	rolling	rolling	ADJ
dem-1058	136	18	windows	window	NOUN
dem-1058	136	19	(	(	PUNCT
dem-1058	136	20	within	within	ADP
dem-1058	136	21	the	the	DET
dem-1058	136	22	comparative	comparative	ADJ
dem-1058	136	23	number	number	NOUN
dem-1058	136	24	of	of	ADP
dem-1058	136	25	observations	observation	NOUN
dem-1058	136	26	)	)	PUNCT
dem-1058	136	27	.	.	PUNCT
dem-1058	137	1	therefore	therefore	ADV
dem-1058	137	2	it	it	PRON
dem-1058	137	3	seems	seem	VERB
dem-1058	137	4	sufficient	sufficient	ADJ
dem-1058	137	5	to	to	PART
dem-1058	137	6	calculate	calculate	VERB
dem-1058	137	7	the	the	DET
dem-1058	137	8	ape	ape	NOUN
dem-1058	137	9	only	only	ADV
dem-1058	137	10	for	for	ADP
dem-1058	137	11	some	some	DET
dem-1058	137	12	small	small	ADJ
dem-1058	137	13	initial	initial	ADJ
dem-1058	137	14	sample	sample	NOUN
dem-1058	137	15	size	size	NOUN
dem-1058	137	16	increasing	increase	VERB
dem-1058	137	17	iteratively	iteratively	ADV
dem-1058	137	18	by	by	ADP
dem-1058	137	19	one	one	NUM
dem-1058	137	20	observation	observation	NOUN
dem-1058	137	21	.	.	PUNCT
dem-1058	138	1	both	both	PRON
dem-1058	138	2	prediction	prediction	NOUN
dem-1058	138	3	and	and	CCONJ
dem-1058	138	4	information	information	NOUN
dem-1058	138	5	criteria	criterion	NOUN
dem-1058	138	6	are	be	AUX
dem-1058	138	7	useful	useful	ADJ
dem-1058	138	8	in	in	ADP
dem-1058	138	9	selecting	select	VERB
dem-1058	138	10	the	the	DET
dem-1058	138	11	forecasting	forecasting	NOUN
dem-1058	138	12	model	model	NOUN
dem-1058	138	13	,	,	PUNCT
dem-1058	138	14	however	however	ADV
dem-1058	138	15	the	the	DET
dem-1058	138	16	choice	choice	NOUN
dem-1058	138	17	of	of	ADP
dem-1058	138	18	models	model	NOUN
dem-1058	138	19	by	by	ADP
dem-1058	138	20	prediction	prediction	NOUN
dem-1058	138	21	criteria	criterion	NOUN
dem-1058	138	22	is	be	AUX
dem-1058	138	23	supposed	suppose	VERB
dem-1058	138	24	to	to	PART
dem-1058	138	25	be	be	AUX
dem-1058	138	26	much	much	ADV
dem-1058	138	27	more	more	ADJ
dem-1058	138	28	in	in	ADP
dem-1058	138	29	agreement	agreement	NOUN
dem-1058	138	30	with	with	ADP
dem-1058	138	31	their	their	PRON
dem-1058	138	32	best	good	ADJ
dem-1058	138	33	out	out	ADP
dem-1058	138	34	-	-	PUNCT
dem-1058	138	35	of	of	ADP
dem-1058	138	36	-	-	PUNCT
dem-1058	138	37	sample	sample	NOUN
dem-1058	138	38	performance	performance	NOUN
dem-1058	138	39	.	.	PUNCT
dem-1058	139	1	information	information	NOUN
dem-1058	139	2	and	and	CCONJ
dem-1058	139	3	prediction	prediction	NOUN
dem-1058	139	4	criteria	criterion	NOUN
dem-1058	139	5	in	in	ADP
dem-1058	139	6	selecting	select	VERB
dem-1058	139	7	the	the	DET
dem-1058	139	8	forecasting	forecasting	NOUN
dem-1058	139	9	model	model	NOUN
dem-1058	139	10	31	31	NUM
dem-1058	139	11	references	reference	NOUN
dem-1058	139	12	armstrong	armstrong	PROPN
dem-1058	139	13	,	,	PUNCT
dem-1058	139	14	j.	j.	PROPN
dem-1058	139	15	s.	s.	PROPN
dem-1058	139	16	(	(	PUNCT
dem-1058	139	17	2001	2001	NUM
dem-1058	139	18	)	)	PUNCT
dem-1058	139	19	,	,	PUNCT
dem-1058	139	20	principles	principle	NOUN
dem-1058	139	21	of	of	ADP
dem-1058	139	22	forecasting	forecasting	NOUN
dem-1058	139	23	,	,	PUNCT
dem-1058	139	24	springer	springer	NOUN
dem-1058	139	25	,	,	PUNCT
dem-1058	139	26	new	new	PROPN
dem-1058	139	27	york	york	PROPN
dem-1058	139	28	.	.	PUNCT
dem-1058	140	1	armstrong	armstrong	PROPN
dem-1058	140	2	,	,	PUNCT
dem-1058	140	3	j.	j.	PROPN
dem-1058	140	4	s.	s.	PROPN
dem-1058	140	5	,	,	PUNCT
dem-1058	140	6	fildes	fildes	PROPN
dem-1058	140	7	,	,	PUNCT
dem-1058	140	8	r.	r.	PROPN
dem-1058	140	9	(	(	PUNCT
dem-1058	140	10	1995	1995	NUM
dem-1058	140	11	)	)	PUNCT
dem-1058	140	12	,	,	PUNCT
dem-1058	140	13	on	on	ADP
dem-1058	140	14	the	the	DET
dem-1058	140	15	selection	selection	NOUN
dem-1058	140	16	of	of	ADP
dem-1058	140	17	error	error	NOUN
dem-1058	140	18	measures	measure	NOUN
dem-1058	140	19	for	for	ADP
dem-1058	140	20	comparisons	comparison	NOUN
dem-1058	140	21	among	among	ADP
dem-1058	140	22	forecasting	forecasting	NOUN
dem-1058	140	23	methods	method	NOUN
dem-1058	140	24	,	,	PUNCT
dem-1058	140	25	journal	journal	NOUN
dem-1058	140	26	of	of	ADP
dem-1058	140	27	forecasting	forecasting	NOUN
dem-1058	140	28	,	,	PUNCT
dem-1058	140	29	14	14	NUM
dem-1058	140	30	,	,	PUNCT
dem-1058	140	31	67−71	67−71	NUM
dem-1058	140	32	.	.	PUNCT
dem-1058	141	1	box	box	PROPN
dem-1058	141	2	,	,	PUNCT
dem-1058	141	3	g.	g.	PROPN
dem-1058	141	4	e.	e.	PROPN
dem-1058	141	5	p.	p.	PROPN
dem-1058	141	6	(	(	PUNCT
dem-1058	141	7	1976	1976	NUM
dem-1058	141	8	)	)	PUNCT
dem-1058	141	9	,	,	PUNCT
dem-1058	141	10	science	science	NOUN
dem-1058	141	11	and	and	CCONJ
dem-1058	141	12	statistics	statistic	NOUN
dem-1058	141	13	,	,	PUNCT
dem-1058	141	14	journal	journal	NOUN
dem-1058	141	15	of	of	ADP
dem-1058	141	16	the	the	DET
dem-1058	141	17	american	american	PROPN
dem-1058	141	18	association	association	PROPN
dem-1058	141	19	,	,	PUNCT
dem-1058	141	20	71	71	NUM
dem-1058	141	21	,	,	PUNCT
dem-1058	141	22	791−799	791−799	NUM
dem-1058	141	23	.	.	X
dem-1058	141	24	burnham	burnham	PROPN
dem-1058	141	25	,	,	PUNCT
dem-1058	141	26	k.	k.	PROPN
dem-1058	142	1	p.	p.	PROPN
dem-1058	142	2	,	,	PUNCT
dem-1058	142	3	anderson	anderson	PROPN
dem-1058	142	4	,	,	PUNCT
dem-1058	142	5	d.	d.	PROPN
dem-1058	142	6	r.	r.	PROPN
dem-1058	142	7	(	(	PUNCT
dem-1058	142	8	2002	2002	NUM
dem-1058	142	9	)	)	PUNCT
dem-1058	142	10	,	,	PUNCT
dem-1058	142	11	model	model	NOUN
dem-1058	142	12	selection	selection	NOUN
dem-1058	142	13	and	and	CCONJ
dem-1058	142	14	multimodel	multimodel	PROPN
dem-1058	142	15	inference	inference	PROPN
dem-1058	142	16	,	,	PUNCT
dem-1058	142	17	springer	springer	NOUN
dem-1058	142	18	.	.	PUNCT
dem-1058	143	1	chatfield	chatfield	PROPN
dem-1058	143	2	,	,	PUNCT
dem-1058	143	3	c.	c.	PROPN
dem-1058	143	4	(	(	PUNCT
dem-1058	143	5	2000	2000	NUM
dem-1058	143	6	)	)	PUNCT
dem-1058	143	7	,	,	PUNCT
dem-1058	143	8	time	time	NOUN
dem-1058	143	9	-	-	PUNCT
dem-1058	143	10	series	series	NOUN
dem-1058	143	11	forecasting	forecasting	NOUN
dem-1058	143	12	,	,	PUNCT
dem-1058	143	13	chapman	chapman	PROPN
dem-1058	143	14	&	&	CCONJ
dem-1058	143	15	hall	hall	PROPN
dem-1058	143	16	/	/	SYM
dem-1058	143	17	crc	crc	PROPN
dem-1058	143	18	,	,	PUNCT
dem-1058	143	19	boca	boca	PROPN
dem-1058	143	20	raton	raton	PROPN
dem-1058	143	21	-	-	PUNCT
dem-1058	143	22	london	london	PROPN
dem-1058	143	23	-	-	PUNCT
dem-1058	143	24	new	new	PROPN
dem-1058	143	25	york	york	PROPN
dem-1058	143	26	-	-	PUNCT
dem-1058	143	27	washington	washington	PROPN
dem-1058	143	28	.	.	PUNCT
dem-1058	144	1	clarke	clarke	PROPN
dem-1058	144	2	,	,	PUNCT
dem-1058	144	3	b.	b.	PROPN
dem-1058	144	4	(	(	PUNCT
dem-1058	144	5	2001	2001	NUM
dem-1058	144	6	)	)	PUNCT
dem-1058	144	7	,	,	PUNCT
dem-1058	144	8	combining	combine	VERB
dem-1058	144	9	model	model	NOUN
dem-1058	144	10	selection	selection	NOUN
dem-1058	144	11	procedures	procedure	NOUN
dem-1058	144	12	for	for	ADP
dem-1058	144	13	online	online	ADJ
dem-1058	144	14	prediction	prediction	NOUN
dem-1058	144	15	,	,	PUNCT
dem-1058	144	16	sankhya	sankhya	NOUN
dem-1058	144	17	:	:	PUNCT
dem-1058	144	18	the	the	DET
dem-1058	144	19	indian	indian	PROPN
dem-1058	144	20	journal	journal	PROPN
dem-1058	144	21	of	of	ADP
dem-1058	144	22	statistics	statistic	NOUN
dem-1058	144	23	,	,	PUNCT
dem-1058	144	24	63	63	NUM
dem-1058	144	25	,	,	PUNCT
dem-1058	144	26	series	series	NOUN
dem-1058	144	27	a	a	NOUN
dem-1058	144	28	,	,	PUNCT
dem-1058	144	29	229−249	229−249	NOUN
dem-1058	144	30	.	.	PUNCT
dem-1058	144	31	de	de	PROPN
dem-1058	144	32	leeuw	leeuw	PROPN
dem-1058	144	33	,	,	PUNCT
dem-1058	144	34	j.	j.	PROPN
dem-1058	144	35	(	(	PUNCT
dem-1058	144	36	1988	1988	NUM
dem-1058	144	37	)	)	PUNCT
dem-1058	144	38	,	,	PUNCT
dem-1058	144	39	model	model	NOUN
dem-1058	144	40	selection	selection	NOUN
dem-1058	144	41	in	in	ADP
dem-1058	144	42	multinomial	multinomial	ADJ
dem-1058	144	43	experiments	experiment	NOUN
dem-1058	144	44	,	,	PUNCT
dem-1058	144	45	in	in	ADP
dem-1058	144	46	:	:	PUNCT
dem-1058	144	47	t.	t.	PROPN
dem-1058	144	48	k.	k.	PROPN
dem-1058	144	49	dijkstra	dijkstra	PROPN
dem-1058	144	50	(	(	PUNCT
dem-1058	144	51	ed	ed	NOUN
dem-1058	144	52	.	.	PUNCT
dem-1058	144	53	)	)	PUNCT
dem-1058	144	54	,	,	PUNCT
dem-1058	144	55	on	on	ADP
dem-1058	144	56	model	model	NOUN
dem-1058	144	57	uncertainty	uncertainty	NOUN
dem-1058	144	58	and	and	CCONJ
dem-1058	144	59	its	its	PRON
dem-1058	144	60	statistical	statistical	ADJ
dem-1058	144	61	implication	implication	NOUN
dem-1058	144	62	,	,	PUNCT
dem-1058	144	63	lecture	lecture	NOUN
dem-1058	144	64	notes	note	NOUN
dem-1058	144	65	in	in	ADP
dem-1058	144	66	economics	economic	NOUN
dem-1058	144	67	and	and	CCONJ
dem-1058	144	68	mathematical	mathematical	ADJ
dem-1058	144	69	systems	system	NOUN
dem-1058	144	70	,	,	PUNCT
dem-1058	144	71	springer	springer	NOUN
dem-1058	144	72	-	-	PUNCT
dem-1058	144	73	verlag	verlag	PROPN
dem-1058	144	74	,	,	PUNCT
dem-1058	144	75	new	new	PROPN
dem-1058	144	76	york	york	PROPN
dem-1058	144	77	.	.	PUNCT
dem-1058	145	1	dawid	dawid	PROPN
dem-1058	145	2	,	,	PUNCT
dem-1058	145	3	a.	a.	NOUN
dem-1058	145	4	p.	p.	NOUN
dem-1058	145	5	(	(	PUNCT
dem-1058	145	6	1984	1984	NUM
dem-1058	145	7	)	)	PUNCT
dem-1058	145	8	,	,	PUNCT
dem-1058	145	9	statistical	statistical	ADJ
dem-1058	145	10	theory	theory	NOUN
dem-1058	145	11	:	:	PUNCT
dem-1058	145	12	the	the	DET
dem-1058	145	13	prequential	prequential	ADJ
dem-1058	145	14	approach	approach	NOUN
dem-1058	145	15	,	,	PUNCT
dem-1058	145	16	journal	journal	NOUN
dem-1058	145	17	of	of	ADP
dem-1058	145	18	royal	royal	ADJ
dem-1058	145	19	statistical	statistical	ADJ
dem-1058	145	20	society	society	NOUN
dem-1058	145	21	series	series	PROPN
dem-1058	145	22	b	b	PROPN
dem-1058	145	23	,	,	PUNCT
dem-1058	145	24	147	147	NUM
dem-1058	145	25	,	,	PUNCT
dem-1058	145	26	278−292	278−292	NOUN
dem-1058	145	27	.	.	NOUN
dem-1058	145	28	de	de	X
dem-1058	145	29	luna	luna	PROPN
dem-1058	145	30	,	,	PUNCT
dem-1058	145	31	x.	x.	PROPN
dem-1058	145	32	,	,	PUNCT
dem-1058	145	33	skouras	skouras	PROPN
dem-1058	145	34	,	,	PUNCT
dem-1058	145	35	k.	k.	PROPN
dem-1058	145	36	(	(	PUNCT
dem-1058	145	37	2003	2003	NUM
dem-1058	145	38	)	)	PUNCT
dem-1058	145	39	,	,	PUNCT
dem-1058	145	40	choosing	choose	VERB
dem-1058	145	41	a	a	DET
dem-1058	145	42	model	model	NOUN
dem-1058	145	43	selection	selection	NOUN
dem-1058	145	44	strategy	strategy	NOUN
dem-1058	145	45	,	,	PUNCT
dem-1058	145	46	scandinavian	scandinavian	ADJ
dem-1058	145	47	journal	journal	NOUN
dem-1058	145	48	of	of	ADP
dem-1058	145	49	statistics	statistic	NOUN
dem-1058	145	50	,	,	PUNCT
dem-1058	145	51	30	30	NUM
dem-1058	145	52	,	,	PUNCT
dem-1058	145	53	113−128	113−128	NUM
dem-1058	145	54	.	.	PUNCT
dem-1058	146	1	kunst	kunst	PROPN
dem-1058	146	2	,	,	PUNCT
dem-1058	146	3	r.	r.	PROPN
dem-1058	146	4	m.	m.	PROPN
dem-1058	146	5	(	(	PUNCT
dem-1058	146	6	2003	2003	NUM
dem-1058	146	7	)	)	PUNCT
dem-1058	146	8	,	,	PUNCT
dem-1058	146	9	testing	test	VERB
dem-1058	146	10	for	for	ADP
dem-1058	146	11	relative	relative	ADJ
dem-1058	146	12	predictive	predictive	ADJ
dem-1058	146	13	accuracy	accuracy	NOUN
dem-1058	146	14	:	:	PUNCT
dem-1058	146	15	a	a	DET
dem-1058	146	16	critical	critical	ADJ
dem-1058	146	17	viewpoint	viewpoint	NOUN
dem-1058	146	18	,	,	PUNCT
dem-1058	146	19	economics	economics	NOUN
dem-1058	146	20	series	series	NOUN
dem-1058	146	21	,	,	PUNCT
dem-1058	146	22	130	130	NUM
dem-1058	146	23	,	,	PUNCT
dem-1058	146	24	1−45	1−45	NUM
dem-1058	146	25	,	,	PUNCT
dem-1058	146	26	department	department	NOUN
dem-1058	146	27	of	of	ADP
dem-1058	146	28	economics	economic	NOUN
dem-1058	146	29	and	and	CCONJ
dem-1058	146	30	finance	finance	NOUN
dem-1058	146	31	,	,	PUNCT
dem-1058	146	32	vienna	vienna	PROPN
dem-1058	146	33	.	.	PUNCT
dem-1058	147	1	mentzer	mentzer	PROPN
dem-1058	147	2	,	,	PUNCT
dem-1058	147	3	j.	j.	PROPN
dem-1058	147	4	t.	t.	PROPN
dem-1058	147	5	,	,	PUNCT
dem-1058	147	6	kahn	kahn	PROPN
dem-1058	147	7	,	,	PUNCT
dem-1058	147	8	k.	k.	PROPN
dem-1058	147	9	b.	b.	PROPN
dem-1058	147	10	(	(	PUNCT
dem-1058	147	11	1995	1995	NUM
dem-1058	147	12	)	)	PUNCT
dem-1058	147	13	,	,	PUNCT
dem-1058	147	14	forecasting	forecasting	NOUN
dem-1058	147	15	technique	technique	NOUN
dem-1058	147	16	familiarity	familiarity	NOUN
dem-1058	147	17	,	,	PUNCT
dem-1058	147	18	satisfaction	satisfaction	NOUN
dem-1058	147	19	,	,	PUNCT
dem-1058	147	20	usage	usage	NOUN
dem-1058	147	21	,	,	PUNCT
dem-1058	147	22	and	and	CCONJ
dem-1058	147	23	application	application	NOUN
dem-1058	147	24	,	,	PUNCT
dem-1058	147	25	journal	journal	NOUN
dem-1058	147	26	of	of	ADP
dem-1058	147	27	forecasting	forecasting	NOUN
dem-1058	147	28	,	,	PUNCT
dem-1058	147	29	14	14	NUM
dem-1058	147	30	,	,	PUNCT
dem-1058	147	31	465−476	465−476	PROPN
dem-1058	147	32	.	.	PUNCT
dem-1058	147	33	rissanen	rissanen	PROPN
dem-1058	147	34	,	,	PUNCT
dem-1058	147	35	j.	j.	PROPN
dem-1058	147	36	(	(	PUNCT
dem-1058	147	37	1986	1986	NUM
dem-1058	147	38	)	)	PUNCT
dem-1058	147	39	,	,	PUNCT
dem-1058	147	40	order	order	NOUN
dem-1058	147	41	estimation	estimation	NOUN
dem-1058	147	42	by	by	ADP
dem-1058	147	43	accumulated	accumulate	VERB
dem-1058	147	44	prediction	prediction	NOUN
dem-1058	147	45	errors	error	NOUN
dem-1058	147	46	,	,	PUNCT
dem-1058	147	47	journal	journal	NOUN
dem-1058	147	48	of	of	ADP
dem-1058	147	49	applied	apply	VERB
dem-1058	147	50	probability	probability	NOUN
dem-1058	147	51	,	,	PUNCT
dem-1058	147	52	23a	23a	NUM
dem-1058	147	53	,	,	PUNCT
dem-1058	147	54	55	55	NUM
dem-1058	147	55	-	-	SYM
dem-1058	147	56	61	61	NUM
dem-1058	147	57	.	.	PUNCT
dem-1058	147	58	skouras	skoura	NOUN
dem-1058	147	59	,	,	PUNCT
dem-1058	147	60	k.	k.	PROPN
dem-1058	147	61	,	,	PUNCT
dem-1058	147	62	dawid	dawid	PROPN
dem-1058	147	63	,	,	PUNCT
dem-1058	147	64	a.	a.	NOUN
dem-1058	147	65	p.	p.	NOUN
dem-1058	147	66	(	(	PUNCT
dem-1058	147	67	1998	1998	NUM
dem-1058	147	68	)	)	PUNCT
dem-1058	147	69	,	,	PUNCT
dem-1058	147	70	on	on	ADP
dem-1058	147	71	efficient	efficient	ADJ
dem-1058	147	72	point	point	NOUN
dem-1058	147	73	prediction	prediction	NOUN
dem-1058	147	74	systems	system	NOUN
dem-1058	147	75	,	,	PUNCT
dem-1058	147	76	journal	journal	NOUN
dem-1058	147	77	of	of	ADP
dem-1058	147	78	royal	royal	ADJ
dem-1058	147	79	statistical	statistical	ADJ
dem-1058	147	80	society	society	NOUN
dem-1058	147	81	b	b	PROPN
dem-1058	147	82	,	,	PUNCT
dem-1058	147	83	60	60	NUM
dem-1058	147	84	,	,	PUNCT
dem-1058	147	85	765−780	765−780	PROPN
dem-1058	147	86	.	.	PUNCT
dem-1058	147	87	swanson	swanson	PROPN
dem-1058	147	88	,	,	PUNCT
dem-1058	147	89	n.	n.	PROPN
dem-1058	147	90	r.	r.	PROPN
dem-1058	147	91	,	,	PUNCT
dem-1058	147	92	white	white	PROPN
dem-1058	147	93	,	,	PUNCT
dem-1058	147	94	h.	h.	PROPN
dem-1058	147	95	(	(	PUNCT
dem-1058	147	96	1997	1997	NUM
dem-1058	147	97	)	)	PUNCT
dem-1058	147	98	,	,	PUNCT
dem-1058	147	99	forecasting	forecast	VERB
dem-1058	147	100	economic	economic	ADJ
dem-1058	147	101	time	time	NOUN
dem-1058	147	102	series	series	PROPN
dem-1058	147	103	using	use	VERB
dem-1058	147	104	flexible	flexible	ADJ
dem-1058	147	105	versus	versus	ADP
dem-1058	147	106	fixed	fix	VERB
dem-1058	147	107	specification	specification	NOUN
dem-1058	147	108	and	and	CCONJ
dem-1058	147	109	linear	linear	PROPN
dem-1058	147	110	versus	versus	ADP
dem-1058	147	111	nonlinear	nonlinear	ADJ
dem-1058	147	112	econometric	econometric	ADJ
dem-1058	147	113	models	model	NOUN
dem-1058	147	114	,	,	PUNCT
dem-1058	147	115	journal	journal	NOUN
dem-1058	147	116	of	of	ADP
dem-1058	147	117	forecasting	forecasting	NOUN
dem-1058	147	118	,	,	PUNCT
dem-1058	147	119	13	13	NUM
dem-1058	147	120	,	,	PUNCT
dem-1058	147	121	439−461	439−461	NUM
dem-1058	147	122	.	.	PUNCT
dem-1058	148	1	taub	taub	PROPN
dem-1058	148	2	,	,	PUNCT
dem-1058	148	3	f.	f.	PROPN
dem-1058	148	4	b.	b.	PROPN
dem-1058	148	5	(	(	PUNCT
dem-1058	148	6	1993	1993	NUM
dem-1058	148	7	)	)	PUNCT
dem-1058	148	8	,	,	PUNCT
dem-1058	148	9	book	book	NOUN
dem-1058	148	10	review	review	NOUN
dem-1058	148	11	:	:	PUNCT
dem-1058	148	12	estimating	estimate	VERB
dem-1058	148	13	ecological	ecological	ADJ
dem-1058	148	14	risks	risk	NOUN
dem-1058	148	15	,	,	PUNCT
dem-1058	148	16	ecology	ecology	NOUN
dem-1058	148	17	,	,	PUNCT
dem-1058	148	18	74	74	NUM
dem-1058	148	19	,	,	PUNCT
dem-1058	148	20	1290−1291	1290−1291	NUM
dem-1058	148	21	.	.	PUNCT
dem-1058	148	22	wagenmaker	wagenmaker	PROPN
dem-1058	148	23	,	,	PUNCT
dem-1058	148	24	e	e	PROPN
dem-1058	148	25	-	-	PROPN
dem-1058	148	26	j.	j.	PROPN
dem-1058	148	27	,	,	PUNCT
dem-1058	148	28	grünwald	grünwald	NOUN
dem-1058	148	29	,	,	PUNCT
dem-1058	148	30	p.	p.	NOUN
dem-1058	148	31	,	,	PUNCT
dem-1058	148	32	steyvers	steyver	NOUN
dem-1058	148	33	,	,	PUNCT
dem-1058	148	34	m.	m.	NOUN
dem-1058	148	35	(	(	PUNCT
dem-1058	148	36	2006	2006	NUM
dem-1058	148	37	)	)	PUNCT
dem-1058	148	38	,	,	PUNCT
dem-1058	148	39	accumulative	accumulative	ADJ
dem-1058	148	40	prediction	prediction	NOUN
dem-1058	148	41	error	error	NOUN
dem-1058	148	42	and	and	CCONJ
dem-1058	148	43	the	the	DET
dem-1058	148	44	selection	selection	NOUN
dem-1058	148	45	of	of	ADP
dem-1058	148	46	time	time	NOUN
dem-1058	148	47	series	series	PROPN
dem-1058	148	48	models	model	NOUN
dem-1058	148	49	,	,	PUNCT
dem-1058	148	50	journal	journal	NOUN
dem-1058	148	51	of	of	ADP
dem-1058	148	52	mathematical	mathematical	ADJ
dem-1058	148	53	psychology	psychology	NOUN
dem-1058	148	54	,	,	PUNCT
dem-1058	148	55	50	50	NUM
dem-1058	148	56	,	,	PUNCT
dem-1058	148	57	149−166	149−166	PROPN
dem-1058	148	58	.	.	PUNCT
dem-1058	148	59	kryteria	kryteria	PROPN
dem-1058	148	60	informacyjne	informacyjne	PROPN
dem-1058	148	61	i	i	PRON
dem-1058	148	62	predykcyjne	predykcyjne	VERB
dem-1058	148	63	w	w	PRON
dem-1058	148	64	wyborze	wyborze	PROPN
dem-1058	148	65	modelu	modelu	PROPN
dem-1058	148	66	prognostycznego	prognostycznego	PROPN
dem-1058	148	67	z	z	PROPN
dem-1058	148	68	a	a	DET
dem-1058	148	69	r	r	NOUN
dem-1058	148	70	y	y	PROPN
dem-1058	148	71	s	s	PROPN
dem-1058	148	72	t	t	NOUN
dem-1058	148	73	r	r	NOUN
dem-1058	148	74	e	e	PROPN
dem-1058	148	75	ś	ś	PROPN
dem-1058	148	76	c	c	PROPN
dem-1058	148	77	i.	i.	PROPN
dem-1058	148	78	celem	celem	PROPN
dem-1058	149	1	artykułu	artykułu	PROPN
dem-1058	149	2	jest	jest	PROPN
dem-1058	149	3	porównanie	porównanie	PROPN
dem-1058	149	4	zachowania	zachowania	PROPN
dem-1058	149	5	się	się	PROPN
dem-1058	149	6	kryteriów	kryteriów	PROPN
dem-1058	149	7	informacyjnych	informacyjnych	PROPN
dem-1058	149	8	i	i	PRON
dem-1058	149	9	predykcyjnych	predykcyjnych	VERB
dem-1058	149	10	w	w	PROPN
dem-1058	149	11	wyborze	wyborze	PROPN
dem-1058	149	12	modelu	modelu	PROPN
dem-1058	149	13	prognostycznego	prognostycznego	NOUN
dem-1058	149	14	na	na	PROPN
dem-1058	149	15	podstawie	podstawie	NOUN
dem-1058	149	16	danych	danych	PROPN
dem-1058	149	17	empirycznych	empirycznych	PROPN
dem-1058	149	18	dla	dla	PROPN
dem-1058	149	19	polski	polski	PROPN
dem-1058	149	20	,	,	PUNCT
dem-1058	149	21	przy	przy	PROPN
dem-1058	149	22	założeniu	założeniu	PROPN
dem-1058	149	23	nieznajomości	nieznajomości	PROPN
dem-1058	149	24	modelu	modelu	PROPN
dem-1058	149	25	generującego	generującego	PROPN
dem-1058	149	26	dane	dane	PROPN
dem-1058	149	27	.	.	PUNCT
dem-1058	150	1	uwaga	uwaga	PROPN
dem-1058	150	2	będzie	będzie	PROPN
dem-1058	150	3	poświęcona	poświęcona	PROPN
dem-1058	150	4	śledzeniu	śledzeniu	PROPN
dem-1058	150	5	zmian	zmian	PROPN
dem-1058	150	6	kryteriów	kryteriów	PROPN
dem-1058	150	7	informacyjnych	informacyjnych	PROPN
dem-1058	150	8	(	(	PUNCT
dem-1058	150	9	aic	aic	PROPN
dem-1058	150	10	,	,	PUNCT
dem-1058	150	11	bic	bic	PROPN
dem-1058	150	12	)	)	PUNCT
dem-1058	150	13	oraz	oraz	PROPN
dem-1058	150	14	skumulowanego	skumulowanego	PROPN
dem-1058	150	15	błędu	błędu	PROPN
dem-1058	150	16	prognoz	prognoz	PROPN
dem-1058	150	17	(	(	PUNCT
dem-1058	150	18	ape	ape	NOUN
dem-1058	150	19	)	)	PUNCT
dem-1058	150	20	dla	dla	NOUN
dem-1058	150	21	próby	próby	NOUN
dem-1058	150	22	powiększanej	powiększanej	NOUN
dem-1058	150	23	iteracyjnie	iteracyjnie	NOUN
dem-1058	150	24	o	o	PROPN
dem-1058	151	1	jedną	jedną	ADV
dem-1058	151	2	obserwację	obserwację	PROPN
dem-1058	152	1	i	i	PRON
dem-1058	152	2	ruchowych	ruchowych	VERB
dem-1058	152	3	okien	okien	NOUN
dem-1058	152	4	(	(	PUNCT
dem-1058	152	5	o	o	PROPN
dem-1058	152	6	różnej	różnej	NOUN
dem-1058	152	7	wielkości	wielkości	NOUN
dem-1058	152	8	)	)	PUNCT
dem-1058	152	9	,	,	PUNCT
dem-1058	152	10	a	a	DET
dem-1058	152	11	także	także	NOUN
dem-1058	152	12	ocenie	ocenie	NOUN
dem-1058	152	13	wpływu	wpływu	PROPN
dem-1058	152	14	wielkości	wielkości	PROPN
dem-1058	152	15	próby	próby	PROPN
dem-1058	152	16	(	(	PUNCT
dem-1058	152	17	startowej	startowej	PROPN
dem-1058	152	18	)	)	PUNCT
dem-1058	153	1	i	i	PRON
dem-1058	153	2	ruchomego	ruchomego	VERB
dem-1058	153	3	okna	okna	NOUN
dem-1058	153	4	na	na	ADP
dem-1058	153	5	wybór	wybór	PROPN
dem-1058	153	6	modelu	modelu	NOUN
dem-1058	153	7	prognostycznego	prognostycznego	NOUN
dem-1058	153	8	.	.	PUNCT
dem-1058	154	1	wybór	wybór	ADJ
dem-1058	154	2	najlepszego	najlepszego	PROPN
dem-1058	154	3	modelu	modelu	PROPN
dem-1058	154	4	prognostycznego	prognostycznego	PROPN
dem-1058	154	5	jest	jest	PROPN
dem-1058	154	6	dokonywany	dokonywany	PROPN
dem-1058	154	7	spośród	spośród	PROPN
dem-1058	154	8	następującego	następującego	PROPN
dem-1058	154	9	zestawu	zestawu	PROPN
dem-1058	154	10	modeli	modeli	ADJ
dem-1058	154	11	:	:	PUNCT
dem-1058	154	12	model	model	PROPN
dem-1058	154	13	autoregresyjny	autoregresyjny	PROPN
dem-1058	154	14	(	(	PUNCT
dem-1058	154	15	ar	ar	PROPN
dem-1058	154	16	,	,	PUNCT
dem-1058	154	17	z	z	NOUN
dem-1058	154	18	trendem	trendem	NOUN
dem-1058	155	1	i	i	PRON
dem-1058	155	2	bez	bez	VERB
dem-1058	155	3	trendu	trendu	NOUN
dem-1058	155	4	deterministycznego	deterministycznego	NOUN
dem-1058	155	5	)	)	PUNCT
dem-1058	155	6	,	,	PUNCT
dem-1058	155	7	model	model	NOUN
dem-1058	155	8	arima	arima	PROPN
dem-1058	155	9	,	,	PUNCT
dem-1058	155	10	model	model	NOUN
dem-1058	155	11	błądzenia	błądzenia	PROPN
dem-1058	155	12	przypadkowego	przypadkowego	PROPN
dem-1058	155	13	(	(	PUNCT
dem-1058	155	14	rw	rw	NOUN
dem-1058	155	15	)	)	PUNCT
dem-1058	155	16	.	.	PUNCT
dem-1058	156	1	s	s	PART
dem-1058	157	1	ł	ł	PROPN
dem-1058	157	2	o	o	X
dem-1058	157	3	w	w	NOUN
dem-1058	157	4	a	a	DET
dem-1058	157	5	k	k	X
dem-1058	157	6	l	l	NOUN
dem-1058	157	7	u	u	NOUN
dem-1058	158	1	c	c	NOUN
dem-1058	158	2	z	z	NOUN
dem-1058	158	3	o	o	X
dem-1058	158	4	w	w	PROPN
dem-1058	159	1	e	e	NOUN
dem-1058	159	2	:	:	PUNCT
dem-1058	159	3	kryteria	kryteria	PROPN
dem-1058	159	4	informacyjne	informacyjne	PROPN
dem-1058	159	5	,	,	PUNCT
dem-1058	159	6	kryteria	kryteria	PROPN
dem-1058	159	7	predykcyjne	predykcyjne	PROPN
dem-1058	159	8	,	,	PUNCT
dem-1058	159	9	skumulowany	skumulowany	ADJ
dem-1058	159	10	błąd	błąd	NOUN
dem-1058	159	11	predykcji	predykcji	NOUN
dem-1058	159	12	,	,	PUNCT
dem-1058	159	13	wybór	wybór	PROPN
dem-1058	159	14	modelu	modelu	NOUN
dem-1058	159	15	.	.	PUNCT
dem-1058	160	1	figure	figure	NOUN
dem-1058	160	2	1	1	NUM
dem-1058	160	3	.	.	PUNCT
dem-1058	161	1	differences	difference	NOUN
dem-1058	161	2	in	in	ADP
dem-1058	161	3	aic	aic	PROPN
dem-1058	161	4	information	information	NOUN
dem-1058	161	5	criterion	criterion	NOUN
dem-1058	161	6	(	(	PUNCT
dem-1058	161	7	version	version	NOUN
dem-1058	161	8	i	i	PRON
dem-1058	161	9	row	row	VERB
dem-1058	161	10	1	1	NUM
dem-1058	161	11	,	,	PUNCT
dem-1058	161	12	version	version	NOUN
dem-1058	161	13	ii	ii	PROPN
dem-1058	161	14	row	row	NOUN
dem-1058	161	15	2	2	NUM
dem-1058	161	16	)	)	PUNCT
dem-1058	161	17	for	for	ADP
dem-1058	161	18	pairs	pair	NOUN
dem-1058	161	19	of	of	ADP
dem-1058	161	20	models	model	NOUN
dem-1058	161	21	(	(	PUNCT
dem-1058	161	22	arima	arima	PROPN
dem-1058	161	23	vs.	vs.	X
dem-1058	161	24	ar	ar	PROPN
dem-1058	161	25	,	,	PUNCT
dem-1058	161	26	arima	arima	PROPN
dem-1058	161	27	vs.	vs.	ADP
dem-1058	161	28	rw	rw	PROPN
dem-1058	161	29	,	,	PUNCT
dem-1058	161	30	ar	ar	PROPN
dem-1058	161	31	vs.	vs.	NOUN
dem-1058	161	32	rw	rw	NOUN
dem-1058	161	33	)	)	PUNCT
dem-1058	161	34	depending	depend	VERB
dem-1058	161	35	on	on	ADP
dem-1058	161	36	starting	start	VERB
dem-1058	161	37	sample	sample	NOUN
dem-1058	161	38	size	size	NOUN
dem-1058	161	39	and	and	CCONJ
dem-1058	161	40	size	size	NOUN
dem-1058	161	41	of	of	ADP
dem-1058	161	42	rolling	rolling	ADJ
dem-1058	161	43	window	window	NOUN
dem-1058	161	44	for	for	ADP
dem-1058	161	45	cpi	cpi	PROPN
dem-1058	161	46	in	in	ADP
dem-1058	161	47	poland	poland	PROPN
dem-1058	161	48	figure	figure	NOUN
dem-1058	161	49	2	2	NUM
dem-1058	161	50	.	.	PUNCT
dem-1058	161	51	differences	difference	NOUN
dem-1058	161	52	in	in	ADP
dem-1058	161	53	bic	bic	PROPN
dem-1058	161	54	information	information	NOUN
dem-1058	161	55	criterion	criterion	NOUN
dem-1058	161	56	(	(	PUNCT
dem-1058	161	57	version	version	NOUN
dem-1058	161	58	i	i	PRON
dem-1058	161	59	row	row	VERB
dem-1058	161	60	1	1	NUM
dem-1058	161	61	,	,	PUNCT
dem-1058	161	62	version	version	NOUN
dem-1058	161	63	ii	ii	PROPN
dem-1058	161	64	row	row	NOUN
dem-1058	161	65	2	2	NUM
dem-1058	161	66	)	)	PUNCT
dem-1058	161	67	for	for	ADP
dem-1058	161	68	pairs	pair	NOUN
dem-1058	161	69	of	of	ADP
dem-1058	161	70	models	model	NOUN
dem-1058	161	71	(	(	PUNCT
dem-1058	161	72	arima	arima	PROPN
dem-1058	161	73	vs.	vs.	X
dem-1058	161	74	ar	ar	PROPN
dem-1058	161	75	,	,	PUNCT
dem-1058	161	76	arima	arima	PROPN
dem-1058	161	77	vs.	vs.	ADP
dem-1058	161	78	rw	rw	PROPN
dem-1058	161	79	,	,	PUNCT
dem-1058	161	80	ar	ar	PROPN
dem-1058	161	81	vs.	vs.	NOUN
dem-1058	161	82	rw	rw	NOUN
dem-1058	161	83	)	)	PUNCT
dem-1058	161	84	depending	depend	VERB
dem-1058	161	85	on	on	ADP
dem-1058	161	86	starting	start	VERB
dem-1058	161	87	sample	sample	NOUN
dem-1058	161	88	size	size	NOUN
dem-1058	161	89	and	and	CCONJ
dem-1058	161	90	size	size	NOUN
dem-1058	161	91	of	of	ADP
dem-1058	161	92	rolling	rolling	ADJ
dem-1058	161	93	window	window	NOUN
dem-1058	161	94	for	for	ADP
dem-1058	161	95	cpi	cpi	PROPN
dem-1058	161	96	in	in	ADP
dem-1058	161	97	poland	poland	PROPN
dem-1058	161	98	figure	figure	NOUN
dem-1058	162	1	3	3	NUM
dem-1058	162	2	.	.	PUNCT
dem-1058	162	3	differences	difference	NOUN
dem-1058	162	4	in	in	ADP
dem-1058	162	5	prediction	prediction	NOUN
dem-1058	162	6	criterion	criterion	NOUN
dem-1058	162	7	ape_se	ape_se	PUNCT
dem-1058	163	1	(	(	PUNCT
dem-1058	163	2	version	version	NOUN
dem-1058	163	3	i	i	PRON
dem-1058	163	4	row	row	VERB
dem-1058	163	5	1	1	NUM
dem-1058	163	6	,	,	PUNCT
dem-1058	163	7	version	version	NOUN
dem-1058	163	8	ii	ii	PROPN
dem-1058	163	9	row	row	NOUN
dem-1058	163	10	2	2	NUM
dem-1058	163	11	)	)	PUNCT
dem-1058	163	12	for	for	ADP
dem-1058	163	13	pairs	pair	NOUN
dem-1058	163	14	of	of	ADP
dem-1058	163	15	models	model	NOUN
dem-1058	163	16	(	(	PUNCT
dem-1058	163	17	arima	arima	PROPN
dem-1058	163	18	vs.	vs.	X
dem-1058	163	19	ar	ar	PROPN
dem-1058	163	20	,	,	PUNCT
dem-1058	163	21	arima	arima	PROPN
dem-1058	163	22	vs.	vs.	ADP
dem-1058	163	23	rw	rw	PROPN
dem-1058	163	24	,	,	PUNCT
dem-1058	163	25	ar	ar	PROPN
dem-1058	163	26	vs.	vs.	NOUN
dem-1058	163	27	rw	rw	NOUN
dem-1058	163	28	)	)	PUNCT
dem-1058	163	29	depending	depend	VERB
dem-1058	163	30	on	on	ADP
dem-1058	163	31	starting	start	VERB
dem-1058	163	32	sample	sample	NOUN
dem-1058	163	33	size	size	NOUN
dem-1058	163	34	and	and	CCONJ
dem-1058	163	35	size	size	NOUN
dem-1058	163	36	of	of	ADP
dem-1058	163	37	rolling	rolling	ADJ
dem-1058	163	38	window	window	NOUN
dem-1058	163	39	for	for	ADP
dem-1058	163	40	cpi	cpi	PROPN
dem-1058	163	41	in	in	ADP
dem-1058	163	42	poland	poland	PROPN
dem-1058	163	43	figure	figure	NOUN
dem-1058	163	44	4	4	NUM
dem-1058	163	45	.	.	PUNCT
dem-1058	163	46	differences	difference	NOUN
dem-1058	163	47	in	in	ADP
dem-1058	163	48	prediction	prediction	NOUN
dem-1058	163	49	criterion	criterion	NOUN
dem-1058	163	50	ape_ae	ape_ae	X
dem-1058	164	1	(	(	PUNCT
dem-1058	164	2	version	version	NOUN
dem-1058	164	3	i	i	PRON
dem-1058	164	4	row	row	VERB
dem-1058	164	5	1	1	NUM
dem-1058	164	6	,	,	PUNCT
dem-1058	164	7	version	version	NOUN
dem-1058	164	8	ii	ii	PROPN
dem-1058	164	9	row	row	NOUN
dem-1058	164	10	2	2	NUM
dem-1058	164	11	)	)	PUNCT
dem-1058	164	12	for	for	ADP
dem-1058	164	13	pairs	pair	NOUN
dem-1058	164	14	of	of	ADP
dem-1058	164	15	models	model	NOUN
dem-1058	164	16	(	(	PUNCT
dem-1058	164	17	arima	arima	PROPN
dem-1058	164	18	vs.	vs.	X
dem-1058	164	19	ar	ar	PROPN
dem-1058	164	20	,	,	PUNCT
dem-1058	164	21	arima	arima	PROPN
dem-1058	164	22	vs.	vs.	ADP
dem-1058	164	23	rw	rw	PROPN
dem-1058	164	24	,	,	PUNCT
dem-1058	164	25	ar	ar	PROPN
dem-1058	164	26	vs.	vs.	NOUN
dem-1058	164	27	rw	rw	NOUN
dem-1058	164	28	)	)	PUNCT
dem-1058	164	29	depending	depend	VERB
dem-1058	164	30	on	on	ADP
dem-1058	164	31	starting	start	VERB
dem-1058	164	32	sample	sample	NOUN
dem-1058	164	33	size	size	NOUN
dem-1058	164	34	and	and	CCONJ
dem-1058	164	35	size	size	NOUN
dem-1058	164	36	of	of	ADP
dem-1058	164	37	rolling	rolling	ADJ
dem-1058	164	38	window	window	NOUN
dem-1058	164	39	for	for	ADP
dem-1058	164	40	cpi	cpi	PROPN
dem-1058	164	41	in	in	ADP
dem-1058	164	42	poland	poland	PROPN
dem-1058	164	43	figure	figure	NOUN
dem-1058	164	44	5	5	NUM
dem-1058	164	45	.	.	PUNCT
dem-1058	164	46	differences	difference	NOUN
dem-1058	164	47	in	in	ADP
dem-1058	164	48	aic	aic	PROPN
dem-1058	164	49	criterion	criterion	NOUN
dem-1058	164	50	(	(	PUNCT
dem-1058	164	51	version	version	NOUN
dem-1058	164	52	i	i	PRON
dem-1058	164	53	row	row	VERB
dem-1058	164	54	1	1	NUM
dem-1058	164	55	,	,	PUNCT
dem-1058	164	56	version	version	NOUN
dem-1058	164	57	ii	ii	PROPN
dem-1058	164	58	row	row	NOUN
dem-1058	164	59	2	2	NUM
dem-1058	164	60	)	)	PUNCT
dem-1058	164	61	for	for	ADP
dem-1058	164	62	pairs	pair	NOUN
dem-1058	164	63	of	of	ADP
dem-1058	164	64	models	model	NOUN
dem-1058	164	65	(	(	PUNCT
dem-1058	164	66	arima	arima	PROPN
dem-1058	164	67	vs.	vs.	X
dem-1058	164	68	ar	ar	PROPN
dem-1058	164	69	,	,	PUNCT
dem-1058	164	70	arima	arima	PROPN
dem-1058	164	71	vs.	vs.	ADP
dem-1058	164	72	rw	rw	PROPN
dem-1058	164	73	,	,	PUNCT
dem-1058	164	74	ar	ar	PROPN
dem-1058	164	75	vs.	vs.	NOUN
dem-1058	164	76	rw	rw	NOUN
dem-1058	164	77	)	)	PUNCT
dem-1058	164	78	depending	depend	VERB
dem-1058	164	79	on	on	ADP
dem-1058	164	80	starting	start	VERB
dem-1058	164	81	sample	sample	NOUN
dem-1058	164	82	size	size	NOUN
dem-1058	164	83	and	and	CCONJ
dem-1058	164	84	size	size	NOUN
dem-1058	164	85	of	of	ADP
dem-1058	164	86	rolling	rolling	ADJ
dem-1058	164	87	window	window	NOUN
dem-1058	164	88	for	for	ADP
dem-1058	164	89	ip	ip	NOUN
dem-1058	164	90	in	in	ADP
dem-1058	164	91	poland	poland	PROPN
dem-1058	164	92	figure	figure	NOUN
dem-1058	164	93	6	6	NUM
dem-1058	164	94	.	.	PUNCT
dem-1058	164	95	differences	difference	NOUN
dem-1058	164	96	in	in	ADP
dem-1058	164	97	bic	bic	PROPN
dem-1058	164	98	criterion	criterion	NOUN
dem-1058	164	99	(	(	PUNCT
dem-1058	164	100	version	version	NOUN
dem-1058	164	101	i	i	PRON
dem-1058	164	102	row	row	VERB
dem-1058	164	103	1	1	NUM
dem-1058	164	104	,	,	PUNCT
dem-1058	164	105	version	version	NOUN
dem-1058	164	106	ii	ii	PROPN
dem-1058	164	107	row	row	NOUN
dem-1058	164	108	2	2	NUM
dem-1058	164	109	)	)	PUNCT
dem-1058	164	110	for	for	ADP
dem-1058	164	111	pairs	pair	NOUN
dem-1058	164	112	of	of	ADP
dem-1058	164	113	models	model	NOUN
dem-1058	164	114	(	(	PUNCT
dem-1058	164	115	arima	arima	PROPN
dem-1058	164	116	vs.	vs.	X
dem-1058	164	117	ar	ar	PROPN
dem-1058	164	118	,	,	PUNCT
dem-1058	164	119	arima	arima	PROPN
dem-1058	164	120	vs.	vs.	ADP
dem-1058	164	121	rw	rw	PROPN
dem-1058	164	122	,	,	PUNCT
dem-1058	164	123	ar	ar	PROPN
dem-1058	164	124	vs.	vs.	NOUN
dem-1058	164	125	rw	rw	NOUN
dem-1058	164	126	)	)	PUNCT
dem-1058	164	127	depending	depend	VERB
dem-1058	164	128	on	on	ADP
dem-1058	164	129	starting	start	VERB
dem-1058	164	130	sample	sample	NOUN
dem-1058	164	131	size	size	NOUN
dem-1058	164	132	and	and	CCONJ
dem-1058	164	133	size	size	NOUN
dem-1058	164	134	of	of	ADP
dem-1058	164	135	rolling	rolling	ADJ
dem-1058	164	136	window	window	NOUN
dem-1058	164	137	for	for	ADP
dem-1058	164	138	ip	ip	NOUN
dem-1058	164	139	in	in	ADP
dem-1058	164	140	poland	poland	PROPN
dem-1058	164	141	figure	figure	NOUN
dem-1058	165	1	7	7	NUM
dem-1058	165	2	.	.	PUNCT
dem-1058	165	3	differences	difference	NOUN
dem-1058	165	4	in	in	ADP
dem-1058	165	5	prediction	prediction	NOUN
dem-1058	165	6	criterion	criterion	NOUN
dem-1058	165	7	ape_se	ape_se	PUNCT
dem-1058	166	1	(	(	PUNCT
dem-1058	166	2	version	version	NOUN
dem-1058	166	3	i	i	PRON
dem-1058	166	4	row	row	VERB
dem-1058	166	5	1	1	NUM
dem-1058	166	6	,	,	PUNCT
dem-1058	166	7	version	version	NOUN
dem-1058	166	8	ii	ii	PROPN
dem-1058	166	9	row	row	NOUN
dem-1058	166	10	2	2	NUM
dem-1058	166	11	)	)	PUNCT
dem-1058	166	12	for	for	ADP
dem-1058	166	13	pairs	pair	NOUN
dem-1058	166	14	of	of	ADP
dem-1058	166	15	models	model	NOUN
dem-1058	166	16	(	(	PUNCT
dem-1058	166	17	arima	arima	PROPN
dem-1058	166	18	vs.	vs.	X
dem-1058	166	19	ar	ar	PROPN
dem-1058	166	20	,	,	PUNCT
dem-1058	166	21	arima	arima	PROPN
dem-1058	166	22	vs.	vs.	ADP
dem-1058	166	23	rw	rw	PROPN
dem-1058	166	24	,	,	PUNCT
dem-1058	166	25	ar	ar	PROPN
dem-1058	166	26	vs.	vs.	NOUN
dem-1058	166	27	rw	rw	NOUN
dem-1058	166	28	)	)	PUNCT
dem-1058	166	29	depending	depend	VERB
dem-1058	166	30	on	on	ADP
dem-1058	166	31	starting	start	VERB
dem-1058	166	32	sample	sample	NOUN
dem-1058	166	33	size	size	NOUN
dem-1058	166	34	and	and	CCONJ
dem-1058	166	35	size	size	NOUN
dem-1058	166	36	of	of	ADP
dem-1058	166	37	rolling	rolling	ADJ
dem-1058	166	38	window	window	NOUN
dem-1058	166	39	for	for	ADP
dem-1058	166	40	ip	ip	NOUN
dem-1058	166	41	in	in	ADP
dem-1058	166	42	poland	poland	PROPN
dem-1058	166	43	figure	figure	NOUN
dem-1058	166	44	8	8	NUM
dem-1058	166	45	.	.	PUNCT
dem-1058	167	1	differences	difference	NOUN
dem-1058	167	2	in	in	ADP
dem-1058	167	3	prediction	prediction	NOUN
dem-1058	167	4	criterion	criterion	NOUN
dem-1058	167	5	ape_ae	ape_ae	X
dem-1058	168	1	(	(	PUNCT
dem-1058	168	2	version	version	NOUN
dem-1058	168	3	i	i	PRON
dem-1058	168	4	row	row	VERB
dem-1058	168	5	1	1	NUM
dem-1058	168	6	,	,	PUNCT
dem-1058	168	7	version	version	NOUN
dem-1058	168	8	ii	ii	PROPN
dem-1058	168	9	row	row	NOUN
dem-1058	168	10	2	2	NUM
dem-1058	168	11	)	)	PUNCT
dem-1058	168	12	for	for	ADP
dem-1058	168	13	pairs	pair	NOUN
dem-1058	168	14	of	of	ADP
dem-1058	168	15	models	model	NOUN
dem-1058	168	16	(	(	PUNCT
dem-1058	168	17	arima	arima	PROPN
dem-1058	168	18	vs.	vs.	X
dem-1058	168	19	ar	ar	PROPN
dem-1058	168	20	,	,	PUNCT
dem-1058	168	21	arima	arima	PROPN
dem-1058	168	22	vs.	vs.	ADP
dem-1058	168	23	rw	rw	PROPN
dem-1058	168	24	,	,	PUNCT
dem-1058	168	25	ar	ar	PROPN
dem-1058	168	26	vs.	vs.	NOUN
dem-1058	168	27	rw	rw	NOUN
dem-1058	168	28	)	)	PUNCT
dem-1058	168	29	depending	depend	VERB
dem-1058	168	30	on	on	ADP
dem-1058	168	31	starting	start	VERB
dem-1058	168	32	sample	sample	NOUN
dem-1058	168	33	size	size	NOUN
dem-1058	168	34	and	and	CCONJ
dem-1058	168	35	size	size	NOUN
dem-1058	168	36	of	of	ADP
dem-1058	168	37	rolling	rolling	ADJ
dem-1058	168	38	window	window	NOUN
dem-1058	168	39	for	for	ADP
dem-1058	168	40	ip	ip	NOUN
dem-1058	168	41	in	in	ADP
dem-1058	168	42	poland	poland	PROPN
dem-1058	168	43	02_piłatowska_m	02_piłatowska_m	NUM
dem-1058	168	44	piłatowska	piłatowska	PROPN
dem-1058	168	45	dem	dem	PROPN
dem-1058	168	46	2011	2011	NUM
dem-1058	168	47	wykresy	wykresy	NOUN