id	sid	tid	token	lemma	pos
altman	1	1	Building	build	VBG
altman	1	2	a	a	DT
altman	1	3	Machine	Machine	NNP
altman	1	4	Learning	Learning	NNP
altman	1	5	Pipeline	Pipeline	NNP
altman	1	6	As	as	IN
altman	1	7	a	a	DT
altman	1	8	new	new	JJ
altman	1	9	machine	machine	NN
altman	1	10	learning	learning	NN
altman	1	11	(	(	-LRB-
altman	1	12	ML	ML	NNP
altman	1	13	)	)	-RRB-
altman	1	14	practitioner	practitioner	NN
altman	1	15	,	,	,
altman	1	16	it	-PRON-	PRP
altman	1	17	is	be	VBZ
altman	1	18	important	important	JJ
altman	1	19	to	to	TO
altman	1	20	develop	develop	VB
altman	1	21	a	a	DT
altman	1	22	mindful	mindful	JJ
altman	1	23	approach	approach	NN
altman	1	24	to	to	IN
altman	1	25	the	the	DT
altman	1	26	craft	craft	NN
altman	1	27	.	.	.
altman	2	1	By	by	IN
altman	2	2	mindful	mindful	JJ
altman	2	3	,	,	,
altman	2	4	I	-PRON-	PRP
altman	2	5	mean	mean	VBP
altman	2	6	possessing	possess	VBG
altman	2	7	the	the	DT
altman	2	8	ability	ability	NN
altman	2	9	to	to	TO
altman	2	10	think	think	VB
altman	2	11	clearly	clearly	RB
altman	2	12	about	about	IN
altman	2	13	each	each	DT
altman	2	14	individual	individual	JJ
altman	2	15	piece	piece	NN
altman	2	16	of	of	IN
altman	2	17	the	the	DT
altman	2	18	process	process	NN
altman	2	19	,	,	,
altman	2	20	and	and	CC
altman	2	21	understanding	understand	VBG
altman	2	22	how	how	WRB
altman	2	23	each	each	DT
altman	2	24	piece	piece	NN
altman	2	25	fits	fit	VBZ
altman	2	26	into	into	IN
altman	2	27	the	the	DT
altman	2	28	larger	large	JJR
altman	2	29	whole	whole	NN
altman	2	30	.	.	.
altman	3	1	In	in	IN
altman	3	2	my	-PRON-	PRP$
altman	3	3	experience	experience	NN
altman	3	4	,	,	,
altman	3	5	there	there	EX
altman	3	6	are	be	VBP
altman	3	7	many	many	JJ
altman	3	8	good	good	JJ
altman	3	9	tutorials	tutorial	NNS
altman	3	10	available	available	JJ
altman	3	11	that	that	WDT
altman	3	12	will	will	MD
altman	3	13	help	help	VB
altman	3	14	you	-PRON-	PRP
altman	3	15	work	work	VB
altman	3	16	with	with	IN
altman	3	17	an	an	DT
altman	3	18	individual	individual	JJ
altman	3	19	tool	tool	NN
altman	3	20	,	,	,
altman	3	21	deploy	deploy	VB
altman	3	22	a	a	DT
altman	3	23	specific	specific	JJ
altman	3	24	algorithm	algorithm	NN
altman	3	25	,	,	,
altman	3	26	or	or	CC
altman	3	27	complete	complete	VB
altman	3	28	a	a	DT
altman	3	29	single	single	JJ
altman	3	30	task	task	NN
altman	3	31	.	.	.
altman	4	1	It	-PRON-	PRP
altman	4	2	is	be	VBZ
altman	4	3	more	more	RBR
altman	4	4	difficult	difficult	JJ
altman	4	5	to	to	TO
altman	4	6	find	find	VB
altman	4	7	guidelines	guideline	NNS
altman	4	8	for	for	IN
altman	4	9	building	build	VBG
altman	4	10	a	a	DT
altman	4	11	holistic	holistic	JJ
altman	4	12	system	system	NN
altman	4	13	that	that	WDT
altman	4	14	supports	support	VBZ
altman	4	15	the	the	DT
altman	4	16	entire	entire	JJ
altman	4	17	ML	ML	NNP
altman	4	18	workflow	workflow	NN
altman	4	19	.	.	.
altman	5	1	My	-PRON-	PRP$
altman	5	2	aim	aim	NN
altman	5	3	is	be	VBZ
altman	5	4	to	to	TO
altman	5	5	help	help	VB
altman	5	6	you	-PRON-	PRP
altman	5	7	build	build	VB
altman	5	8	just	just	RB
altman	5	9	such	such	PDT
altman	5	10	a	a	DT
altman	5	11	system	system	NN
altman	5	12	,	,	,
altman	5	13	so	so	IN
altman	5	14	that	that	IN
altman	5	15	you	-PRON-	PRP
altman	5	16	are	be	VBP
altman	5	17	free	free	JJ
altman	5	18	to	to	TO
altman	5	19	focus	focus	VB
altman	5	20	on	on	IN
altman	5	21	inquiry	inquiry	NN
altman	5	22	and	and	CC
altman	5	23	discovery	discovery	NN
altman	5	24	rather	rather	RB
altman	5	25	than	than	IN
altman	5	26	struggling	struggle	VBG
altman	5	27	with	with	IN
altman	5	28	infrastructure	infrastructure	NN
altman	5	29	and	and	CC
altman	5	30	process	process	NN
altman	5	31	.	.	.
altman	6	1	I	-PRON-	PRP
altman	6	2	write	write	VBP
altman	6	3	this	this	DT
altman	6	4	as	as	IN
altman	6	5	a	a	DT
altman	6	6	software	software	NN
altman	6	7	developer	developer	NN
altman	6	8	who	who	WP
altman	6	9	has	have	VBZ
altman	6	10	,	,	,
altman	6	11	at	at	IN
altman	6	12	one	one	CD
altman	6	13	time	time	NN
altman	6	14	or	or	CC
altman	6	15	another	another	DT
altman	6	16	,	,	,
altman	6	17	been	be	VBN
altman	6	18	on	on	IN
altman	6	19	the	the	DT
altman	6	20	wrong	wrong	JJ
altman	6	21	end	end	NN
altman	6	22	of	of	IN
altman	6	23	all	all	PDT
altman	6	24	the	the	DT
altman	6	25	recommendations	recommendation	NNS
altman	6	26	presented	present	VBN
altman	6	27	here	here	RB
altman	6	28	,	,	,
altman	6	29	and	and	CC
altman	6	30	hopes	hope	VBZ
altman	6	31	to	to	TO
altman	6	32	save	save	VB
altman	6	33	you	-PRON-	PRP
altman	6	34	from	from	IN
altman	6	35	similar	similar	JJ
altman	6	36	headaches	headache	NNS
altman	6	37	.	.	.
altman	7	1	Many	many	JJ
altman	7	2	of	of	IN
altman	7	3	the	the	DT
altman	7	4	examples	example	NNS
altman	7	5	and	and	CC
altman	7	6	design	design	NN
altman	7	7	choices	choice	NNS
altman	7	8	are	be	VBP
altman	7	9	drawn	draw	VBN
altman	7	10	from	from	IN
altman	7	11	my	-PRON-	PRP$
altman	7	12	experiences	experience	NNS
altman	7	13	at	at	IN
altman	7	14	the	the	DT
altman	7	15	Digital	Digital	NNP
altman	7	16	Public	Public	NNP
altman	7	17	Library	Library	NNP
altman	7	18	of	of	IN
altman	7	19	America	America	NNP
altman	7	20	,	,	,
altman	7	21	where	where	WRB
altman	7	22	I	-PRON-	PRP
altman	7	23	have	have	VBP
altman	7	24	worked	work	VBN
altman	7	25	alongside	alongside	IN
altman	7	26	a	a	DT
altman	7	27	very	very	RB
altman	7	28	talented	talented	JJ
altman	7	29	team	team	NN
altman	7	30	of	of	IN
altman	7	31	developers	developer	NNS
altman	7	32	.	.	.
altman	8	1	This	this	DT
altman	8	2	is	be	VBZ
altman	8	3	by	by	IN
altman	8	4	no	no	DT
altman	8	5	means	means	NN
altman	8	6	an	an	DT
altman	8	7	exhaustive	exhaustive	JJ
altman	8	8	text	text	NN
altman	8	9	,	,	,
altman	8	10	but	but	CC
altman	8	11	rather	rather	RB
altman	8	12	a	a	DT
altman	8	13	bit	bit	NN
altman	8	14	of	of	IN
altman	8	15	pragmatic	pragmatic	JJ
altman	8	16	advice	advice	NN
altman	8	17	and	and	CC
altman	8	18	a	a	DT
altman	8	19	jumping	jumping	NN
altman	8	20	-	-	HYPH
altman	8	21	off	off	RP
altman	8	22	point	point	NN
altman	8	23	for	for	IN
altman	8	24	further	further	JJ
altman	8	25	research	research	NN
altman	8	26	,	,	,
altman	8	27	designed	design	VBN
altman	8	28	to	to	TO
altman	8	29	give	give	VB
altman	8	30	you	-PRON-	PRP
altman	8	31	a	a	DT
altman	8	32	clearer	clear	JJR
altman	8	33	idea	idea	NN
altman	8	34	of	of	IN
altman	8	35	which	which	WDT
altman	8	36	questions	question	NNS
altman	8	37	to	to	TO
altman	8	38	ask	ask	VB
altman	8	39	throughout	throughout	IN
altman	8	40	your	-PRON-	PRP$
altman	8	41	practice	practice	NN
altman	8	42	.	.	.
altman	9	1	This	this	DT
altman	9	2	article	article	NN
altman	9	3	reviews	review	VBZ
altman	9	4	the	the	DT
altman	9	5	basic	basic	JJ
altman	9	6	machine	machine	NN
altman	9	7	learning	learning	NN
altman	9	8	workflow	workflow	NN
altman	9	9	,	,	,
altman	9	10	discussing	discuss	VBG
altman	9	11	design	design	NN
altman	9	12	considerations	consideration	NNS
altman	9	13	along	along	IN
altman	9	14	the	the	DT
altman	9	15	way	way	NN
altman	9	16	.	.	.
altman	10	1	It	-PRON-	PRP
altman	10	2	offers	offer	VBZ
altman	10	3	recommendations	recommendation	NNS
altman	10	4	for	for	IN
altman	10	5	data	datum	NNS
altman	10	6	storage	storage	NN
altman	10	7	,	,	,
altman	10	8	guidelines	guideline	NNS
altman	10	9	on	on	IN
altman	10	10	selecting	select	VBG
altman	10	11	and	and	CC
altman	10	12	working	work	VBG
altman	10	13	with	with	IN
altman	10	14	ML	ML	NNP
altman	10	15	algorithms	algorithm	NNS
altman	10	16	,	,	,
altman	10	17	and	and	CC
altman	10	18	questions	question	NNS
altman	10	19	for	for	IN
altman	10	20	tool	tool	NN
altman	10	21	selection	selection	NN
altman	10	22	.	.	.
altman	11	1	Finally	finally	RB
altman	11	2	,	,	,
altman	11	3	it	-PRON-	PRP
altman	11	4	describes	describe	VBZ
altman	11	5	some	some	DT
altman	11	6	challenges	challenge	NNS
altman	11	7	with	with	IN
altman	11	8	scaling	scale	VBG
altman	11	9	up	up	RP
altman	11	10	.	.	.
altman	12	1	My	-PRON-	PRP$
altman	12	2	hope	hope	NN
altman	12	3	is	be	VBZ
altman	12	4	that	that	IN
altman	12	5	the	the	DT
altman	12	6	insight	insight	NN
altman	12	7	presented	present	VBD
altman	12	8	here	here	RB
altman	12	9	,	,	,
altman	12	10	combined	combine	VBN
altman	12	11	with	with	IN
altman	12	12	your	-PRON-	PRP$
altman	12	13	good	good	JJ
altman	12	14	judgement	judgement	NN
altman	12	15	,	,	,
altman	12	16	will	will	MD
altman	12	17	empower	empower	VB
altman	12	18	you	-PRON-	PRP
altman	12	19	to	to	TO
altman	12	20	get	get	VB
altman	12	21	started	start	VBN
altman	12	22	with	with	IN
altman	12	23	the	the	DT
altman	12	24	actual	actual	JJ
altman	12	25	practice	practice	NN
altman	12	26	of	of	IN
altman	12	27	designing	design	VBG
altman	12	28	and	and	CC
altman	12	29	executing	execute	VBG
altman	12	30	a	a	DT
altman	12	31	machine	machine	NN
altman	12	32	learning	learning	NN
altman	12	33	project	project	NN
altman	12	34	.	.	.
altman	13	1	The	the	DT
altman	13	2	machine	machine	NN
altman	13	3	learning	learning	NN
altman	13	4	pipeline	pipeline	NN
altman	13	5	The	the	DT
altman	13	6	metaphor	metaphor	NN
altman	13	7	of	of	IN
altman	13	8	a	a	DT
altman	13	9	pipeline	pipeline	NN
altman	13	10	is	be	VBZ
altman	13	11	often	often	RB
altman	13	12	used	use	VBN
altman	13	13	for	for	IN
altman	13	14	a	a	DT
altman	13	15	machine	machine	NN
altman	13	16	learning	learn	VBG
altman	13	17	workflow	workflow	NN
altman	13	18	.	.	.
altman	14	1	This	this	DT
altman	14	2	metaphor	metaphor	NN
altman	14	3	captures	capture	VBZ
altman	14	4	the	the	DT
altman	14	5	idea	idea	NN
altman	14	6	of	of	IN
altman	14	7	data	datum	NNS
altman	14	8	channeled	channel	VBN
altman	14	9	through	through	IN
altman	14	10	a	a	DT
altman	14	11	series	series	NN
altman	14	12	of	of	IN
altman	14	13	sequential	sequential	JJ
altman	14	14	transformations	transformation	NNS
altman	14	15	.	.	.
altman	15	1	However	however	RB
altman	15	2	,	,	,
altman	15	3	it	-PRON-	PRP
altman	15	4	is	be	VBZ
altman	15	5	important	important	JJ
altman	15	6	to	to	TO
altman	15	7	note	note	VB
altman	15	8	that	that	IN
altman	15	9	each	each	DT
altman	15	10	stage	stage	NN
altman	15	11	in	in	IN
altman	15	12	the	the	DT
altman	15	13	process	process	NN
altman	15	14	will	will	MD
altman	15	15	need	need	VB
altman	15	16	to	to	TO
altman	15	17	be	be	VB
altman	15	18	repeated	repeat	VBN
altman	15	19	and	and	CC
altman	15	20	honed	hone	VBN
altman	15	21	throughout	throughout	IN
altman	15	22	the	the	DT
altman	15	23	course	course	NN
altman	15	24	of	of	IN
altman	15	25	your	-PRON-	PRP$
altman	15	26	project	project	NN
altman	15	27	.	.	.
altman	16	1	Therefore	therefore	RB
altman	16	2	,	,	,
altman	16	3	do	do	VBP
altman	16	4	n’t	not	RB
altman	16	5	think	think	VB
altman	16	6	of	of	IN
altman	16	7	yourself	-PRON-	PRP
altman	16	8	as	as	IN
altman	16	9	building	build	VBG
altman	16	10	a	a	DT
altman	16	11	single	single	JJ
altman	16	12	intelligent	intelligent	JJ
altman	16	13	model	model	NN
altman	16	14	,	,	,
altman	16	15	such	such	JJ
altman	16	16	as	as	IN
altman	16	17	a	a	DT
altman	16	18	decision	decision	NN
altman	16	19	tree	tree	NN
altman	16	20	or	or	CC
altman	16	21	clustering	clustering	NN
altman	16	22	algorithm	algorithm	NN
altman	16	23	.	.	.
altman	17	1	Instead	instead	RB
altman	17	2	,	,	,
altman	17	3	build	build	VB
altman	17	4	a	a	DT
altman	17	5	pipeline	pipeline	NN
altman	17	6	with	with	IN
altman	17	7	pieces	piece	NNS
altman	17	8	that	that	WDT
altman	17	9	can	can	MD
altman	17	10	be	be	VB
altman	17	11	swapped	swap	VBN
altman	17	12	in	in	IN
altman	17	13	and	and	CC
altman	17	14	out	out	RB
altman	17	15	as	as	IN
altman	17	16	needed	need	VBN
altman	17	17	.	.	.
altman	18	1	Data	datum	NNS
altman	18	2	flows	flow	VBZ
altman	18	3	through	through	IN
altman	18	4	the	the	DT
altman	18	5	pipeline	pipeline	NN
altman	18	6	and	and	CC
altman	18	7	outputs	output	VBZ
altman	18	8	a	a	DT
altman	18	9	version	version	NN
altman	18	10	of	of	IN
altman	18	11	a	a	DT
altman	18	12	decision	decision	NN
altman	18	13	tree	tree	NN
altman	18	14	,	,	,
altman	18	15	clustering	cluster	VBG
altman	18	16	algorithm	algorithm	NN
altman	18	17	,	,	,
altman	18	18	or	or	CC
altman	18	19	other	other	JJ
altman	18	20	intelligent	intelligent	JJ
altman	18	21	model	model	NN
altman	18	22	.	.	.
altman	19	1	Throughout	throughout	IN
altman	19	2	your	-PRON-	PRP$
altman	19	3	process	process	NN
altman	19	4	,	,	,
altman	19	5	you	-PRON-	PRP
altman	19	6	will	will	MD
altman	19	7	tweak	tweak	VB
altman	19	8	your	-PRON-	PRP$
altman	19	9	pipeline	pipeline	NN
altman	19	10	,	,	,
altman	19	11	making	make	VBG
altman	19	12	many	many	JJ
altman	19	13	intelligent	intelligent	JJ
altman	19	14	models	model	NNS
altman	19	15	.	.	.
altman	20	1	Eventually	eventually	RB
altman	20	2	you	-PRON-	PRP
altman	20	3	will	will	MD
altman	20	4	select	select	VB
altman	20	5	the	the	DT
altman	20	6	best	good	JJS
altman	20	7	model	model	NN
altman	20	8	for	for	IN
altman	20	9	your	-PRON-	PRP$
altman	20	10	use	use	NN
altman	20	11	case	case	NN
altman	20	12	.	.	.
altman	21	1	To	to	TO
altman	21	2	use	use	VB
altman	21	3	another	another	DT
altman	21	4	metaphor	metaphor	NN
altman	21	5	,	,	,
altman	21	6	do	do	VB
altman	21	7	n’t	not	RB
altman	21	8	build	build	VB
altman	21	9	a	a	DT
altman	21	10	car	car	NN
altman	21	11	,	,	,
altman	21	12	build	build	VB
altman	21	13	an	an	DT
altman	21	14	assembly	assembly	NN
altman	21	15	line	line	NN
altman	21	16	for	for	IN
altman	21	17	making	make	VBG
altman	21	18	cars	car	NNS
altman	21	19	.	.	.
altman	22	1	While	while	IN
altman	22	2	the	the	DT
altman	22	3	final	final	JJ
altman	22	4	output	output	NN
altman	22	5	of	of	IN
altman	22	6	a	a	DT
altman	22	7	machine	machine	NN
altman	22	8	learning	learning	NN
altman	22	9	workflow	workflow	NN
altman	22	10	is	be	VBZ
altman	22	11	some	some	DT
altman	22	12	sort	sort	NN
altman	22	13	of	of	IN
altman	22	14	intelligent	intelligent	JJ
altman	22	15	model	model	NN
altman	22	16	,	,	,
altman	22	17	there	there	EX
altman	22	18	are	be	VBP
altman	22	19	many	many	JJ
altman	22	20	factors	factor	NNS
altman	22	21	that	that	WDT
altman	22	22	make	make	VBP
altman	22	23	repetition	repetition	NN
altman	22	24	and	and	CC
altman	22	25	iteration	iteration	NN
altman	22	26	necessary	necessary	JJ
altman	22	27	.	.	.
altman	23	1	ML	ML	NNP
altman	23	2	processes	process	NNS
altman	23	3	often	often	RB
altman	23	4	involve	involve	VBP
altman	23	5	subjective	subjective	JJ
altman	23	6	decisions	decision	NNS
altman	23	7	,	,	,
altman	23	8	such	such	JJ
altman	23	9	as	as	IN
altman	23	10	which	which	WDT
altman	23	11	data	data	NN
altman	23	12	points	point	NNS
altman	23	13	to	to	TO
altman	23	14	ignore	ignore	VB
altman	23	15	,	,	,
altman	23	16	or	or	CC
altman	23	17	which	which	WDT
altman	23	18	configurations	configuration	VBZ
altman	23	19	you	-PRON-	PRP
altman	23	20	select	select	VBP
altman	23	21	for	for	IN
altman	23	22	your	-PRON-	PRP$
altman	23	23	algorithm	algorithm	NN
altman	23	24	.	.	.
altman	24	1	You	-PRON-	PRP
altman	24	2	will	will	MD
altman	24	3	want	want	VB
altman	24	4	to	to	TO
altman	24	5	test	test	VB
altman	24	6	different	different	JJ
altman	24	7	possibilities	possibility	NNS
altman	24	8	to	to	TO
altman	24	9	see	see	VB
altman	24	10	what	what	WP
altman	24	11	works	work	VBZ
altman	24	12	best	well	RBS
altman	24	13	.	.	.
altman	25	1	As	as	IN
altman	25	2	you	-PRON-	PRP
altman	25	3	learn	learn	VBP
altman	25	4	more	more	JJR
altman	25	5	about	about	IN
altman	25	6	your	-PRON-	PRP$
altman	25	7	dataset	dataset	NN
altman	25	8	throughout	throughout	IN
altman	25	9	the	the	DT
altman	25	10	course	course	NN
altman	25	11	of	of	IN
altman	25	12	the	the	DT
altman	25	13	project	project	NN
altman	25	14	,	,	,
altman	25	15	you	-PRON-	PRP
altman	25	16	will	will	MD
altman	25	17	go	go	VB
altman	25	18	back	back	RB
altman	25	19	and	and	CC
altman	25	20	tweak	tweak	VB
altman	25	21	parts	part	NNS
altman	25	22	of	of	IN
altman	25	23	your	-PRON-	PRP$
altman	25	24	process	process	NN
altman	25	25	.	.	.
altman	26	1	You	-PRON-	PRP
altman	26	2	may	may	MD
altman	26	3	discover	discover	VB
altman	26	4	biases	bias	NNS
altman	26	5	in	in	IN
altman	26	6	your	-PRON-	PRP$
altman	26	7	data	datum	NNS
altman	26	8	or	or	CC
altman	26	9	algorithms	algorithm	NNS
altman	26	10	that	that	WDT
altman	26	11	need	need	VBP
altman	26	12	to	to	TO
altman	26	13	be	be	VB
altman	26	14	addressed	address	VBN
altman	26	15	.	.	.
altman	27	1	If	if	IN
altman	27	2	you	-PRON-	PRP
altman	27	3	are	be	VBP
altman	27	4	working	work	VBG
altman	27	5	collaboratively	collaboratively	RB
altman	27	6	,	,	,
altman	27	7	you	-PRON-	PRP
altman	27	8	will	will	MD
altman	27	9	be	be	VB
altman	27	10	incorporating	incorporate	VBG
altman	27	11	asynchronous	asynchronous	JJ
altman	27	12	feedback	feedback	NN
altman	27	13	from	from	IN
altman	27	14	members	member	NNS
altman	27	15	of	of	IN
altman	27	16	your	-PRON-	PRP$
altman	27	17	team	team	NN
altman	27	18	.	.	.
altman	28	1	At	at	IN
altman	28	2	some	some	DT
altman	28	3	point	point	NN
altman	28	4	,	,	,
altman	28	5	you	-PRON-	PRP
altman	28	6	may	may	MD
altman	28	7	need	need	VB
altman	28	8	to	to	TO
altman	28	9	introduce	introduce	VB
altman	28	10	new	new	JJ
altman	28	11	or	or	CC
altman	28	12	revised	revise	VBN
altman	28	13	data	datum	NNS
altman	28	14	,	,	,
altman	28	15	or	or	CC
altman	28	16	try	try	VB
altman	28	17	a	a	DT
altman	28	18	new	new	JJ
altman	28	19	tool	tool	NN
altman	28	20	or	or	CC
altman	28	21	algorithm	algorithm	NN
altman	28	22	.	.	.
altman	29	1	It	-PRON-	PRP
altman	29	2	is	be	VBZ
altman	29	3	also	also	RB
altman	29	4	prudent	prudent	JJ
altman	29	5	to	to	TO
altman	29	6	expect	expect	VB
altman	29	7	and	and	CC
altman	29	8	plan	plan	VB
altman	29	9	for	for	IN
altman	29	10	errors	error	NNS
altman	29	11	.	.	.
altman	30	1	Human	human	JJ
altman	30	2	errors	error	NNS
altman	30	3	are	be	VBP
altman	30	4	inevitable	inevitable	JJ
altman	30	5	,	,	,
altman	30	6	and	and	CC
altman	30	7	hardware	hardware	NN
altman	30	8	errors	error	NNS
altman	30	9	,	,	,
altman	30	10	such	such	JJ
altman	30	11	as	as	IN
altman	30	12	network	network	NN
altman	30	13	timeouts	timeout	NNS
altman	30	14	or	or	CC
altman	30	15	memory	memory	NN
altman	30	16	overloads	overload	NNS
altman	30	17	,	,	,
altman	30	18	are	be	VBP
altman	30	19	common	common	JJ
altman	30	20	.	.	.
altman	31	1	For	for	IN
altman	31	2	all	all	DT
altman	31	3	of	of	IN
altman	31	4	these	these	DT
altman	31	5	reasons	reason	NNS
altman	31	6	,	,	,
altman	31	7	you	-PRON-	PRP
altman	31	8	will	will	MD
altman	31	9	be	be	VB
altman	31	10	well	well	RB
altman	31	11	-	-	HYPH
altman	31	12	served	serve	VBN
altman	31	13	by	by	IN
altman	31	14	a	a	DT
altman	31	15	pipeline	pipeline	NN
altman	31	16	composed	compose	VBN
altman	31	17	of	of	IN
altman	31	18	modular	modular	JJ
altman	31	19	,	,	,
altman	31	20	repeatable	repeatable	JJ
altman	31	21	steps	step	NNS
altman	31	22	,	,	,
altman	31	23	each	each	DT
altman	31	24	with	with	IN
altman	31	25	discrete	discrete	JJ
altman	31	26	and	and	CC
altman	31	27	stable	stable	JJ
altman	31	28	output	output	NN
altman	31	29	.	.	.
altman	32	1	A	a	DT
altman	32	2	modular	modular	JJ
altman	32	3	pipeline	pipeline	NN
altman	32	4	supports	support	VBZ
altman	32	5	a	a	DT
altman	32	6	batch	batch	NN
altman	32	7	processing	processing	NN
altman	32	8	workflow	workflow	NN
altman	32	9	,	,	,
altman	32	10	in	in	IN
altman	32	11	which	which	WDT
altman	32	12	whole	whole	JJ
altman	32	13	datasets	dataset	NNS
altman	32	14	undergo	undergo	VBP
altman	32	15	a	a	DT
altman	32	16	series	series	NN
altman	32	17	of	of	IN
altman	32	18	transformations	transformation	NNS
altman	32	19	.	.	.
altman	33	1	During	during	IN
altman	33	2	each	each	DT
altman	33	3	step	step	NN
altman	33	4	of	of	IN
altman	33	5	the	the	DT
altman	33	6	process	process	NN
altman	33	7	,	,	,
altman	33	8	a	a	DT
altman	33	9	large	large	JJ
altman	33	10	amount	amount	NN
altman	33	11	of	of	IN
altman	33	12	data	datum	NNS
altman	33	13	(	(	-LRB-
altman	33	14	possibly	possibly	RB
altman	33	15	the	the	DT
altman	33	16	entire	entire	JJ
altman	33	17	dataset	dataset	NN
altman	33	18	)	)	-RRB-
altman	33	19	is	be	VBZ
altman	33	20	transformed	transform	VBN
altman	33	21	all	all	RB
altman	33	22	at	at	IN
altman	33	23	once	once	RB
altman	33	24	and	and	CC
altman	33	25	then	then	RB
altman	33	26	incrementally	incrementally	RB
altman	33	27	stored	store	VBD
altman	33	28	.	.	.
altman	34	1	This	this	DT
altman	34	2	can	can	MD
altman	34	3	be	be	VB
altman	34	4	contrasted	contrast	VBN
altman	34	5	with	with	IN
altman	34	6	a	a	DT
altman	34	7	real	real	JJ
altman	34	8	-	-	HYPH
altman	34	9	time	time	NN
altman	34	10	workflow	workflow	NN
altman	34	11	,	,	,
altman	34	12	in	in	IN
altman	34	13	which	which	WDT
altman	34	14	individual	individual	JJ
altman	34	15	records	record	NNS
altman	34	16	are	be	VBP
altman	34	17	transformed	transform	VBN
altman	34	18	instantaneously	instantaneously	RB
altman	34	19	(	(	-LRB-
altman	34	20	e.g.	e.g.	RB
altman	35	1	a	a	DT
altman	35	2	librarian	librarian	NN
altman	35	3	updates	update	VBZ
altman	35	4	a	a	DT
altman	35	5	single	single	JJ
altman	35	6	record	record	NN
altman	35	7	in	in	IN
altman	35	8	library	library	JJ
altman	35	9	catalog	catalog	NN
altman	35	10	)	)	-RRB-
altman	35	11	;	;	:
altman	35	12	or	or	CC
altman	35	13	a	a	DT
altman	35	14	streaming	streaming	NN
altman	35	15	workflow	workflow	NN
altman	35	16	,	,	,
altman	35	17	in	in	IN
altman	35	18	which	which	WDT
altman	35	19	a	a	DT
altman	35	20	continuous	continuous	JJ
altman	35	21	flow	flow	NN
altman	35	22	of	of	IN
altman	35	23	data	datum	NNS
altman	35	24	is	be	VBZ
altman	35	25	pushed	push	VBN
altman	35	26	through	through	IN
altman	35	27	an	an	DT
altman	35	28	entire	entire	JJ
altman	35	29	pipeline	pipeline	NN
altman	35	30	,	,	,
altman	35	31	often	often	RB
altman	35	32	without	without	IN
altman	35	33	incremental	incremental	JJ
altman	35	34	storage	storage	NN
altman	35	35	along	along	IN
altman	35	36	the	the	DT
altman	35	37	way	way	NN
altman	35	38	(	(	-LRB-
altman	35	39	e.g.	e.g.	RB
altman	36	1	performing	perform	VBG
altman	36	2	analysis	analysis	NN
altman	36	3	on	on	IN
altman	36	4	a	a	DT
altman	36	5	continuous	continuous	JJ
altman	36	6	stream	stream	NN
altman	36	7	of	of	IN
altman	36	8	new	new	JJ
altman	36	9	tweets	tweet	NNS
altman	36	10	)	)	-RRB-
altman	36	11	.	.	.
altman	37	1	Batch	batch	NN
altman	37	2	processing	processing	NN
altman	37	3	is	be	VBZ
altman	37	4	common	common	JJ
altman	37	5	in	in	IN
altman	37	6	the	the	DT
altman	37	7	research	research	NN
altman	37	8	and	and	CC
altman	37	9	development	development	NN
altman	37	10	phase	phase	NN
altman	37	11	of	of	IN
altman	37	12	an	an	DT
altman	37	13	ML	ML	NNP
altman	37	14	project	project	NN
altman	37	15	,	,	,
altman	37	16	and	and	CC
altman	37	17	may	may	MD
altman	37	18	also	also	RB
altman	37	19	be	be	VB
altman	37	20	a	a	DT
altman	37	21	good	good	JJ
altman	37	22	choice	choice	NN
altman	37	23	for	for	IN
altman	37	24	a	a	DT
altman	37	25	production	production	NN
altman	37	26	system	system	NN
altman	37	27	.	.	.
altman	38	1	When	when	WRB
altman	38	2	designing	design	VBG
altman	38	3	any	any	DT
altman	38	4	step	step	NN
altman	38	5	in	in	IN
altman	38	6	the	the	DT
altman	38	7	batch	batch	NN
altman	38	8	processing	processing	NN
altman	38	9	pipeline	pipeline	NN
altman	38	10	,	,	,
altman	38	11	assume	assume	VBP
altman	38	12	that	that	IN
altman	38	13	at	at	IN
altman	38	14	some	some	DT
altman	38	15	point	point	NN
altman	38	16	you	-PRON-	PRP
altman	38	17	will	will	MD
altman	38	18	need	need	VB
altman	38	19	to	to	TO
altman	38	20	repeat	repeat	VB
altman	38	21	it	-PRON-	PRP
altman	38	22	either	either	CC
altman	38	23	exactly	exactly	RB
altman	38	24	as	as	IN
altman	38	25	is	be	VBZ
altman	38	26	,	,	,
altman	38	27	or	or	CC
altman	38	28	with	with	IN
altman	38	29	modifications	modification	NNS
altman	38	30	.	.	.
altman	39	1	Documenting	document	VBG
altman	39	2	your	-PRON-	PRP$
altman	39	3	process	process	NN
altman	39	4	lets	let	VBZ
altman	39	5	you	-PRON-	PRP
altman	39	6	compare	compare	VB
altman	39	7	the	the	DT
altman	39	8	outputs	output	NNS
altman	39	9	of	of	IN
altman	39	10	different	different	JJ
altman	39	11	variations	variation	NNS
altman	39	12	and	and	CC
altman	39	13	communicate	communicate	VB
altman	39	14	the	the	DT
altman	39	15	ways	way	NNS
altman	39	16	in	in	IN
altman	39	17	which	which	WDT
altman	39	18	your	-PRON-	PRP$
altman	39	19	choices	choice	NNS
altman	39	20	impact	impact	VBP
altman	39	21	the	the	DT
altman	39	22	final	final	JJ
altman	39	23	results	result	NNS
altman	39	24	.	.	.
altman	40	1	If	if	IN
altman	40	2	you	-PRON-	PRP
altman	40	3	’re	be	VBP
altman	40	4	writing	write	VBG
altman	40	5	code	code	NN
altman	40	6	,	,	,
altman	40	7	version	version	NN
altman	40	8	control	control	NN
altman	40	9	software	software	NN
altman	40	10	can	can	MD
altman	40	11	help	help	VB
altman	40	12	.	.	.
altman	41	1	If	if	IN
altman	41	2	you	-PRON-	PRP
altman	41	3	’re	be	VBP
altman	41	4	doing	do	VBG
altman	41	5	more	more	JJR
altman	41	6	manual	manual	JJ
altman	41	7	data	data	NN
altman	41	8	manipulations	manipulation	NNS
altman	41	9	,	,	,
altman	41	10	such	such	JJ
altman	41	11	as	as	IN
altman	41	12	editing	editing	NN
altman	41	13	data	datum	NNS
altman	41	14	in	in	IN
altman	41	15	spreadsheets	spreadsheet	NNS
altman	41	16	,	,	,
altman	41	17	you	-PRON-	PRP
altman	41	18	will	will	MD
altman	41	19	need	need	VB
altman	41	20	an	an	DT
altman	41	21	intentional	intentional	JJ
altman	41	22	system	system	NN
altman	41	23	of	of	IN
altman	41	24	documenting	document	VBG
altman	41	25	exactly	exactly	RB
altman	41	26	which	which	WDT
altman	41	27	transformations	transformation	NNS
altman	41	28	you	-PRON-	PRP
altman	41	29	are	be	VBP
altman	41	30	applying	apply	VBG
altman	41	31	to	to	IN
altman	41	32	your	-PRON-	PRP$
altman	41	33	data	datum	NNS
altman	41	34	.	.	.
altman	42	1	It	-PRON-	PRP
altman	42	2	is	be	VBZ
altman	42	3	generally	generally	RB
altman	42	4	preferable	preferable	JJ
altman	42	5	to	to	TO
altman	42	6	automate	automate	VB
altman	42	7	processes	process	NNS
altman	42	8	wherever	wherever	WRB
altman	42	9	possible	possible	JJ
altman	42	10	so	so	IN
altman	42	11	that	that	IN
altman	42	12	you	-PRON-	PRP
altman	42	13	can	can	MD
altman	42	14	repeat	repeat	VB
altman	42	15	them	-PRON-	PRP
altman	42	16	with	with	IN
altman	42	17	ease	ease	NN
altman	42	18	and	and	CC
altman	42	19	consistency	consistency	NN
altman	42	20	.	.	.
altman	43	1	A	a	DT
altman	43	2	concrete	concrete	JJ
altman	43	3	example	example	NN
altman	43	4	from	from	IN
altman	43	5	my	-PRON-	PRP$
altman	43	6	own	own	JJ
altman	43	7	experience	experience	NN
altman	43	8	demonstrates	demonstrate	VBZ
altman	43	9	the	the	DT
altman	43	10	importance	importance	NN
altman	43	11	of	of	IN
altman	43	12	a	a	DT
altman	43	13	pipeline	pipeline	NN
altman	43	14	that	that	WDT
altman	43	15	supports	support	VBZ
altman	43	16	repetition	repetition	NN
altman	43	17	.	.	.
altman	44	1	In	in	IN
altman	44	2	my	-PRON-	PRP$
altman	44	3	first	first	JJ
altman	44	4	ever	ever	RB
altman	44	5	ML	ML	NNP
altman	44	6	project	project	NN
altman	44	7	,	,	,
altman	44	8	I	-PRON-	PRP
altman	44	9	worked	work	VBD
altman	44	10	with	with	IN
altman	44	11	a	a	DT
altman	44	12	set	set	NN
altman	44	13	of	of	IN
altman	44	14	XML	xml	NN
altman	44	15	library	library	NN
altman	44	16	data	datum	NNS
altman	44	17	converted	convert	VBN
altman	44	18	to	to	IN
altman	44	19	CSV	CSV	NNP
altman	44	20	.	.	.
altman	45	1	I	-PRON-	PRP
altman	45	2	did	do	VBD
altman	45	3	most	most	JJS
altman	45	4	of	of	IN
altman	45	5	my	-PRON-	PRP$
altman	45	6	data	data	NN
altman	45	7	cleanup	cleanup	NN
altman	45	8	by	by	IN
altman	45	9	hand	hand	NN
altman	45	10	using	use	VBG
altman	45	11	spreadsheet	spreadsheet	NN
altman	45	12	software	software	NN
altman	45	13	,	,	,
altman	45	14	and	and	CC
altman	45	15	was	be	VBD
altman	45	16	not	not	RB
altman	45	17	careful	careful	JJ
altman	45	18	about	about	IN
altman	45	19	preserving	preserve	VBG
altman	45	20	the	the	DT
altman	45	21	formulas	formula	NNS
altman	45	22	for	for	IN
altman	45	23	each	each	DT
altman	45	24	step	step	NN
altman	45	25	of	of	IN
altman	45	26	the	the	DT
altman	45	27	process	process	NN
altman	45	28	;	;	:
altman	45	29	instead	instead	RB
altman	45	30	,	,	,
altman	45	31	I	-PRON-	PRP
altman	45	32	deleted	delete	VBD
altman	45	33	and	and	CC
altman	45	34	wrote	write	VBD
altman	45	35	over	over	IN
altman	45	36	many	many	JJ
altman	45	37	important	important	JJ
altman	45	38	intermediate	intermediate	JJ
altman	45	39	computations	computation	NNS
altman	45	40	,	,	,
altman	45	41	saving	save	VBG
altman	45	42	only	only	RB
altman	45	43	the	the	DT
altman	45	44	final	final	JJ
altman	45	45	results	result	NNS
altman	45	46	.	.	.
altman	46	1	This	this	DT
altman	46	2	whole	whole	JJ
altman	46	3	process	process	NN
altman	46	4	took	take	VBD
altman	46	5	me	-PRON-	PRP
altman	46	6	countless	countless	JJ
altman	46	7	hours	hour	NNS
altman	46	8	,	,	,
altman	46	9	and	and	CC
altman	46	10	when	when	WRB
altman	46	11	an	an	DT
altman	46	12	updated	update	VBN
altman	46	13	dataset	dataset	NN
altman	46	14	became	become	VBD
altman	46	15	available	available	JJ
altman	46	16	,	,	,
altman	46	17	there	there	EX
altman	46	18	was	be	VBD
altman	46	19	no	no	DT
altman	46	20	way	way	NN
altman	46	21	to	to	TO
altman	46	22	reproduce	reproduce	VB
altman	46	23	my	-PRON-	PRP$
altman	46	24	painstaking	painstaking	JJ
altman	46	25	cleanup	cleanup	NN
altman	46	26	process	process	NN
altman	46	27	.	.	.
altman	47	1	I	-PRON-	PRP
altman	47	2	was	be	VBD
altman	47	3	stuck	stick	VBN
altman	47	4	with	with	IN
altman	47	5	outdated	outdated	JJ
altman	47	6	data	datum	NNS
altman	47	7	,	,	,
altman	47	8	and	and	CC
altman	47	9	my	-PRON-	PRP$
altman	47	10	final	final	JJ
altman	47	11	output	output	NN
altman	47	12	was	be	VBD
altman	47	13	doomed	doom	VBN
altman	47	14	to	to	TO
altman	47	15	grow	grow	VB
altman	47	16	more	more	RBR
altman	47	17	and	and	CC
altman	47	18	more	more	RBR
altman	47	19	irrelevant	irrelevant	JJ
altman	47	20	as	as	IN
altman	47	21	time	time	NN
altman	47	22	wore	wear	VBD
altman	47	23	on	on	RB
altman	47	24	.	.	.
altman	48	1	Since	since	IN
altman	48	2	then	then	RB
altman	48	3	,	,	,
altman	48	4	I	-PRON-	PRP
altman	48	5	have	have	VBP
altman	48	6	always	always	RB
altman	48	7	written	write	VBN
altman	48	8	repeatable	repeatable	JJ
altman	48	9	scripts	script	NNS
altman	48	10	for	for	IN
altman	48	11	all	all	DT
altman	48	12	my	-PRON-	PRP$
altman	48	13	data	data	NN
altman	48	14	cleanup	cleanup	NN
altman	48	15	tasks	task	NNS
altman	48	16	.	.	.
altman	49	1	Each	each	DT
altman	49	2	decision	decision	NN
altman	49	3	you	-PRON-	PRP
altman	49	4	make	make	VBP
altman	49	5	will	will	MD
altman	49	6	have	have	VB
altman	49	7	an	an	DT
altman	49	8	impact	impact	NN
altman	49	9	on	on	IN
altman	49	10	the	the	DT
altman	49	11	final	final	JJ
altman	49	12	results	result	NNS
altman	49	13	,	,	,
altman	49	14	so	so	CC
altman	49	15	it	-PRON-	PRP
altman	49	16	is	be	VBZ
altman	49	17	important	important	JJ
altman	49	18	to	to	TO
altman	49	19	keep	keep	VB
altman	49	20	clear	clear	JJ
altman	49	21	documentation	documentation	NN
altman	49	22	and	and	CC
altman	49	23	to	to	TO
altman	49	24	verify	verify	VB
altman	49	25	your	-PRON-	PRP$
altman	49	26	assumptions	assumption	NNS
altman	49	27	and	and	CC
altman	49	28	hypotheses	hypothesis	NNS
altman	49	29	wherever	wherever	WRB
altman	49	30	possible	possible	JJ
altman	49	31	.	.	.
altman	50	1	Sometimes	sometimes	RB
altman	50	2	there	there	EX
altman	50	3	will	will	MD
altman	50	4	be	be	VB
altman	50	5	explicit	explicit	JJ
altman	50	6	tests	test	NNS
altman	50	7	to	to	TO
altman	50	8	perform	perform	VB
altman	50	9	;	;	:
altman	50	10	at	at	IN
altman	50	11	other	other	JJ
altman	50	12	times	time	NNS
altman	50	13	,	,	,
altman	50	14	you	-PRON-	PRP
altman	50	15	may	may	MD
altman	50	16	just	just	RB
altman	50	17	need	need	VB
altman	50	18	to	to	TO
altman	50	19	look	look	VB
altman	50	20	at	at	IN
altman	50	21	data	datum	NNS
altman	50	22	—	—	:
altman	50	23	make	make	VB
altman	50	24	a	a	DT
altman	50	25	quick	quick	JJ
altman	50	26	visualization	visualization	NN
altman	50	27	,	,	,
altman	50	28	perform	perform	VB
altman	50	29	a	a	DT
altman	50	30	simple	simple	JJ
altman	50	31	calculation	calculation	NN
altman	50	32	,	,	,
altman	50	33	or	or	CC
altman	50	34	glance	glance	NN
altman	50	35	through	through	IN
altman	50	36	a	a	DT
altman	50	37	sample	sample	NN
altman	50	38	of	of	IN
altman	50	39	records	record	NNS
altman	50	40	.	.	.
altman	51	1	Be	be	VB
altman	51	2	cognizant	cognizant	JJ
altman	51	3	of	of	IN
altman	51	4	the	the	DT
altman	51	5	potential	potential	NN
altman	51	6	to	to	TO
altman	51	7	introduce	introduce	VB
altman	51	8	error	error	NN
altman	51	9	or	or	CC
altman	51	10	bias	bias	NN
altman	51	11	.	.	.
altman	52	1	For	for	IN
altman	52	2	example	example	NN
altman	52	3	,	,	,
altman	52	4	you	-PRON-	PRP
altman	52	5	could	could	MD
altman	52	6	remove	remove	VB
altman	52	7	a	a	DT
altman	52	8	field	field	NN
altman	52	9	that	that	WDT
altman	52	10	you	-PRON-	PRP
altman	52	11	do	do	VBP
altman	52	12	n’t	not	RB
altman	52	13	think	think	VB
altman	52	14	is	be	VBZ
altman	52	15	important	important	JJ
altman	52	16	,	,	,
altman	52	17	but	but	CC
altman	52	18	that	that	DT
altman	52	19	would	would	MD
altman	52	20	,	,	,
altman	52	21	in	in	IN
altman	52	22	fact	fact	NN
altman	52	23	,	,	,
altman	52	24	have	have	VBP
altman	52	25	a	a	DT
altman	52	26	meaningful	meaningful	JJ
altman	52	27	impact	impact	NN
altman	52	28	on	on	IN
altman	52	29	the	the	DT
altman	52	30	final	final	JJ
altman	52	31	result	result	NN
altman	52	32	.	.	.
altman	53	1	All	all	DT
altman	53	2	of	of	IN
altman	53	3	these	these	DT
altman	53	4	precautions	precaution	NNS
altman	53	5	will	will	MD
altman	53	6	strengthen	strengthen	VB
altman	53	7	confidence	confidence	NN
altman	53	8	in	in	IN
altman	53	9	your	-PRON-	PRP$
altman	53	10	final	final	JJ
altman	53	11	outcomes	outcome	NNS
altman	53	12	and	and	CC
altman	53	13	make	make	VB
altman	53	14	them	-PRON-	PRP
altman	53	15	intelligible	intelligible	JJ
altman	53	16	to	to	IN
altman	53	17	your	-PRON-	PRP$
altman	53	18	collaborators	collaborator	NNS
altman	53	19	and	and	CC
altman	53	20	other	other	JJ
altman	53	21	audiences	audience	NNS
altman	53	22	.	.	.
altman	54	1	The	the	DT
altman	54	2	pipeline	pipeline	NN
altman	54	3	for	for	IN
altman	54	4	a	a	DT
altman	54	5	machine	machine	NN
altman	54	6	learning	learning	NN
altman	54	7	project	project	NN
altman	54	8	generally	generally	RB
altman	54	9	comprises	comprise	VBZ
altman	54	10	five	five	CD
altman	54	11	stages	stage	NNS
altman	54	12	:	:	:
altman	54	13	data	datum	NNS
altman	54	14	acquisition	acquisition	NN
altman	54	15	,	,	,
altman	54	16	data	data	NN
altman	54	17	preparation	preparation	NN
altman	54	18	,	,	,
altman	54	19	model	model	NN
altman	54	20	training	training	NN
altman	54	21	and	and	CC
altman	54	22	testing	testing	NN
altman	54	23	,	,	,
altman	54	24	evaluation	evaluation	NN
altman	54	25	and	and	CC
altman	54	26	analysis	analysis	NN
altman	54	27	,	,	,
altman	54	28	and	and	CC
altman	54	29	application	application	NN
altman	54	30	of	of	IN
altman	54	31	results	result	NNS
altman	54	32	.	.	.
altman	55	1	Data	datum	NNS
altman	55	2	acquisition	acquisition	NN
altman	55	3	The	the	DT
altman	55	4	first	first	JJ
altman	55	5	step	step	NN
altman	55	6	is	be	VBZ
altman	55	7	to	to	TO
altman	55	8	acquire	acquire	VB
altman	55	9	the	the	DT
altman	55	10	data	datum	NNS
altman	55	11	that	that	WDT
altman	55	12	you	-PRON-	PRP
altman	55	13	will	will	MD
altman	55	14	be	be	VB
altman	55	15	using	use	VBG
altman	55	16	for	for	IN
altman	55	17	your	-PRON-	PRP$
altman	55	18	machine	machine	NN
altman	55	19	learning	learning	NN
altman	55	20	project	project	NN
altman	55	21	.	.	.
altman	56	1	You	-PRON-	PRP
altman	56	2	may	may	MD
altman	56	3	need	need	VB
altman	56	4	to	to	TO
altman	56	5	combine	combine	VB
altman	56	6	data	datum	NNS
altman	56	7	from	from	IN
altman	56	8	several	several	JJ
altman	56	9	different	different	JJ
altman	56	10	sources	source	NNS
altman	56	11	.	.	.
altman	57	1	There	there	EX
altman	57	2	are	be	VBP
altman	57	3	many	many	JJ
altman	57	4	ways	way	NNS
altman	57	5	to	to	TO
altman	57	6	acquire	acquire	VB
altman	57	7	data	datum	NNS
altman	57	8	,	,	,
altman	57	9	including	include	VBG
altman	57	10	downloading	download	VBG
altman	57	11	files	file	NNS
altman	57	12	,	,	,
altman	57	13	querying	query	VBG
altman	57	14	a	a	DT
altman	57	15	database	database	NN
altman	57	16	or	or	CC
altman	57	17	API	api	NN
altman	57	18	,	,	,
altman	57	19	or	or	CC
altman	57	20	scraping	scrape	VBG
altman	57	21	web	web	NN
altman	57	22	pages	page	NNS
altman	57	23	.	.	.
altman	58	1	Depending	depend	VBG
altman	58	2	on	on	IN
altman	58	3	the	the	DT
altman	58	4	size	size	NN
altman	58	5	of	of	IN
altman	58	6	the	the	DT
altman	58	7	source	source	NN
altman	58	8	data	datum	NNS
altman	58	9	and	and	CC
altman	58	10	how	how	WRB
altman	58	11	it	-PRON-	PRP
altman	58	12	is	be	VBZ
altman	58	13	made	make	VBN
altman	58	14	available	available	JJ
altman	58	15	,	,	,
altman	58	16	this	this	DT
altman	58	17	can	can	MD
altman	58	18	be	be	VB
altman	58	19	a	a	DT
altman	58	20	quick	quick	JJ
altman	58	21	and	and	CC
altman	58	22	simple	simple	JJ
altman	58	23	step	step	NN
altman	58	24	or	or	CC
altman	58	25	the	the	DT
altman	58	26	most	most	RBS
altman	58	27	challenging	challenging	JJ
altman	58	28	bottleneck	bottleneck	NN
altman	58	29	in	in	IN
altman	58	30	your	-PRON-	PRP$
altman	58	31	pipeline	pipeline	NN
altman	58	32	.	.	.
altman	59	1	However	however	RB
altman	59	2	you	-PRON-	PRP
altman	59	3	get	get	VBP
altman	59	4	your	-PRON-	PRP$
altman	59	5	initial	initial	JJ
altman	59	6	data	datum	NNS
altman	59	7	,	,	,
altman	59	8	it	-PRON-	PRP
altman	59	9	is	be	VBZ
altman	59	10	generally	generally	RB
altman	59	11	a	a	DT
altman	59	12	good	good	JJ
altman	59	13	idea	idea	NN
altman	59	14	to	to	TO
altman	59	15	save	save	VB
altman	59	16	a	a	DT
altman	59	17	copy	copy	NN
altman	59	18	in	in	IN
altman	59	19	the	the	DT
altman	59	20	rawest	raw	JJS
altman	59	21	possible	possible	JJ
altman	59	22	form	form	NN
altman	59	23	and	and	CC
altman	59	24	treat	treat	VB
altman	59	25	that	that	DT
altman	59	26	copy	copy	VBP
altman	59	27	as	as	IN
altman	59	28	immutable	immutable	JJ
altman	59	29	,	,	,
altman	59	30	at	at	IN
altman	59	31	least	least	JJS
altman	59	32	during	during	IN
altman	59	33	the	the	DT
altman	59	34	initial	initial	JJ
altman	59	35	phase	phase	NN
altman	59	36	of	of	IN
altman	59	37	testing	test	VBG
altman	59	38	different	different	JJ
altman	59	39	algorithms	algorithm	NNS
altman	59	40	or	or	CC
altman	59	41	configurations	configuration	NNS
altman	59	42	.	.	.
altman	60	1	Having	have	VBG
altman	60	2	a	a	DT
altman	60	3	raw	raw	JJ
altman	60	4	,	,	,
altman	60	5	immutable	immutable	JJ
altman	60	6	copy	copy	NN
altman	60	7	of	of	IN
altman	60	8	your	-PRON-	PRP$
altman	60	9	initial	initial	JJ
altman	60	10	dataset	dataset	NN
altman	60	11	(	(	-LRB-
altman	60	12	or	or	CC
altman	60	13	datasets	dataset	NNS
altman	60	14	)	)	-RRB-
altman	60	15	ensures	ensure	VBZ
altman	60	16	that	that	IN
altman	60	17	you	-PRON-	PRP
altman	60	18	can	can	MD
altman	60	19	always	always	RB
altman	60	20	go	go	VB
altman	60	21	back	back	RB
altman	60	22	to	to	IN
altman	60	23	the	the	DT
altman	60	24	beginning	beginning	NN
altman	60	25	of	of	IN
altman	60	26	your	-PRON-	PRP$
altman	60	27	ML	ML	NNP
altman	60	28	process	process	NN
altman	60	29	and	and	CC
altman	60	30	start	start	VB
altman	60	31	over	over	RP
altman	60	32	with	with	IN
altman	60	33	exactly	exactly	RB
altman	60	34	the	the	DT
altman	60	35	same	same	JJ
altman	60	36	input	input	NN
altman	60	37	.	.	.
altman	61	1	It	-PRON-	PRP
altman	61	2	will	will	MD
altman	61	3	also	also	RB
altman	61	4	save	save	VB
altman	61	5	you	-PRON-	PRP
altman	61	6	from	from	IN
altman	61	7	the	the	DT
altman	61	8	possibility	possibility	NN
altman	61	9	that	that	IN
altman	61	10	the	the	DT
altman	61	11	source	source	NN
altman	61	12	data	datum	NNS
altman	61	13	will	will	MD
altman	61	14	change	change	VB
altman	61	15	from	from	IN
altman	61	16	beneath	beneath	IN
altman	61	17	you	-PRON-	PRP
altman	61	18	,	,	,
altman	61	19	thereby	thereby	RB
altman	61	20	compromising	compromise	VBG
altman	61	21	your	-PRON-	PRP$
altman	61	22	ability	ability	NN
altman	61	23	to	to	TO
altman	61	24	compare	compare	VB
altman	61	25	the	the	DT
altman	61	26	outputs	output	NNS
altman	61	27	of	of	IN
altman	61	28	different	different	JJ
altman	61	29	operations	operation	NNS
altman	61	30	(	(	-LRB-
altman	61	31	for	for	IN
altman	61	32	more	more	JJR
altman	61	33	on	on	IN
altman	61	34	this	this	DT
altman	61	35	,	,	,
altman	61	36	see	see	VB
altman	61	37	the	the	DT
altman	61	38	section	section	NN
altman	61	39	on	on	IN
altman	61	40	data	datum	NNS
altman	61	41	storage	storage	NN
altman	61	42	)	)	-RRB-
altman	61	43	.	.	.
altman	62	1	If	if	IN
altman	62	2	possible	possible	JJ
altman	62	3	,	,	,
altman	62	4	it	-PRON-	PRP
altman	62	5	’s	’	VBZ
altman	62	6	often	often	RB
altman	62	7	worthwhile	worthwhile	JJ
altman	62	8	to	to	TO
altman	62	9	learn	learn	VB
altman	62	10	about	about	IN
altman	62	11	how	how	WRB
altman	62	12	the	the	DT
altman	62	13	original	original	JJ
altman	62	14	data	datum	NNS
altman	62	15	was	be	VBD
altman	62	16	created	create	VBN
altman	62	17	,	,	,
altman	62	18	especially	especially	RB
altman	62	19	if	if	IN
altman	62	20	you	-PRON-	PRP
altman	62	21	are	be	VBP
altman	62	22	getting	get	VBG
altman	62	23	data	datum	NNS
altman	62	24	from	from	IN
altman	62	25	multiple	multiple	JJ
altman	62	26	sources	source	NNS
altman	62	27	that	that	WDT
altman	62	28	differ	differ	VBP
altman	62	29	in	in	IN
altman	62	30	subtle	subtle	JJ
altman	62	31	ways	way	NNS
altman	62	32	.	.	.
altman	63	1	Data	datum	NNS
altman	63	2	preparation	preparation	NN
altman	63	3	Data	Data	NNP
altman	63	4	preparation	preparation	NN
altman	63	5	involves	involve	VBZ
altman	63	6	cleaning	cleaning	NN
altman	63	7	data	datum	NNS
altman	63	8	and	and	CC
altman	63	9	transforming	transform	VBG
altman	63	10	it	-PRON-	PRP
altman	63	11	into	into	IN
altman	63	12	an	an	DT
altman	63	13	appropriate	appropriate	JJ
altman	63	14	format	format	NN
altman	63	15	for	for	IN
altman	63	16	subsequent	subsequent	JJ
altman	63	17	machine	machine	NN
altman	63	18	learning	learning	NN
altman	63	19	tasks	task	NNS
altman	63	20	.	.	.
altman	64	1	This	this	DT
altman	64	2	is	be	VBZ
altman	64	3	often	often	RB
altman	64	4	the	the	DT
altman	64	5	part	part	NN
altman	64	6	of	of	IN
altman	64	7	the	the	DT
altman	64	8	process	process	NN
altman	64	9	that	that	WDT
altman	64	10	requires	require	VBZ
altman	64	11	the	the	DT
altman	64	12	most	most	JJS
altman	64	13	work	work	NN
altman	64	14	,	,	,
altman	64	15	and	and	CC
altman	64	16	you	-PRON-	PRP
altman	64	17	should	should	MD
altman	64	18	expect	expect	VB
altman	64	19	to	to	TO
altman	64	20	iterate	iterate	VB
altman	64	21	over	over	IN
altman	64	22	your	-PRON-	PRP$
altman	64	23	data	data	NN
altman	64	24	preparations	preparation	NNS
altman	64	25	many	many	JJ
altman	64	26	times	time	NNS
altman	64	27	,	,	,
altman	64	28	even	even	RB
altman	64	29	after	after	IN
altman	64	30	you	-PRON-	PRP
altman	64	31	’ve	have	VB
altman	64	32	started	start	VBN
altman	64	33	training	train	VBG
altman	64	34	and	and	CC
altman	64	35	testing	testing	NN
altman	64	36	models	model	NNS
altman	64	37	.	.	.
altman	65	1	The	the	DT
altman	65	2	first	first	JJ
altman	65	3	step	step	NN
altman	65	4	of	of	IN
altman	65	5	data	datum	NNS
altman	65	6	preparation	preparation	NN
altman	65	7	is	be	VBZ
altman	65	8	to	to	TO
altman	65	9	parse	parse	VB
altman	65	10	your	-PRON-	PRP$
altman	65	11	acquired	acquire	VBN
altman	65	12	data	datum	NNS
altman	65	13	and	and	CC
altman	65	14	transform	transform	VB
altman	65	15	it	-PRON-	PRP
altman	65	16	into	into	IN
altman	65	17	a	a	DT
altman	65	18	common	common	JJ
altman	65	19	,	,	,
altman	65	20	usable	usable	JJ
altman	65	21	schema	schema	NN
altman	65	22	.	.	.
altman	66	1	Acquired	acquire	VBN
altman	66	2	data	datum	NNS
altman	66	3	often	often	RB
altman	66	4	comes	come	VBZ
altman	66	5	in	in	IN
altman	66	6	file	file	NN
altman	66	7	formats	format	NNS
altman	66	8	that	that	WDT
altman	66	9	are	be	VBP
altman	66	10	good	good	JJ
altman	66	11	for	for	IN
altman	66	12	data	datum	NNS
altman	66	13	sharing	sharing	NN
altman	66	14	,	,	,
altman	66	15	such	such	JJ
altman	66	16	as	as	IN
altman	66	17	XML	xml	NN
altman	66	18	,	,	,
altman	66	19	JSON	JSON	NNP
altman	66	20	,	,	,
altman	66	21	or	or	CC
altman	66	22	CSV	CSV	NNP
altman	66	23	.	.	.
altman	67	1	You	-PRON-	PRP
altman	67	2	can	can	MD
altman	67	3	parse	parse	VB
altman	67	4	these	these	DT
altman	67	5	files	file	NNS
altman	67	6	into	into	IN
altman	67	7	whatever	whatever	WDT
altman	67	8	schema	schema	NN
altman	67	9	makes	make	VBZ
altman	67	10	sense	sense	NN
altman	67	11	to	to	TO
altman	67	12	manage	manage	VB
altman	67	13	the	the	DT
altman	67	14	various	various	JJ
altman	67	15	transformations	transformation	NNS
altman	67	16	you	-PRON-	PRP
altman	67	17	want	want	VBP
altman	67	18	to	to	TO
altman	67	19	perform	perform	VB
altman	67	20	,	,	,
altman	67	21	but	but	CC
altman	67	22	it	-PRON-	PRP
altman	67	23	can	can	MD
altman	67	24	help	help	VB
altman	67	25	to	to	TO
altman	67	26	have	have	VB
altman	67	27	a	a	DT
altman	67	28	sense	sense	NN
altman	67	29	of	of	IN
altman	67	30	where	where	WRB
altman	67	31	you	-PRON-	PRP
altman	67	32	are	be	VBP
altman	67	33	headed	head	VBN
altman	67	34	.	.	.
altman	68	1	Your	-PRON-	PRP$
altman	68	2	eventual	eventual	JJ
altman	68	3	choice	choice	NN
altman	68	4	of	of	IN
altman	68	5	data	datum	NNS
altman	68	6	format	format	NN
altman	68	7	will	will	MD
altman	68	8	likely	likely	RB
altman	68	9	be	be	VB
altman	68	10	dictated	dictate	VBN
altman	68	11	by	by	IN
altman	68	12	your	-PRON-	PRP$
altman	68	13	ML	ML	NNP
altman	68	14	algorithms	algorithm	NNS
altman	68	15	;	;	:
altman	68	16	likely	likely	JJ
altman	68	17	candidates	candidate	NNS
altman	68	18	include	include	VBP
altman	68	19	multidimensional	multidimensional	JJ
altman	68	20	arrays	array	NNS
altman	68	21	,	,	,
altman	68	22	tensors	tensor	NNS
altman	68	23	,	,	,
altman	68	24	matrices	matrix	NNS
altman	68	25	,	,	,
altman	68	26	and	and	CC
altman	68	27	DataFrames	DataFrames	NNP
altman	68	28	.	.	.
altman	69	1	Look	look	VB
altman	69	2	ahead	ahead	RB
altman	69	3	to	to	IN
altman	69	4	specific	specific	JJ
altman	69	5	functions	function	NNS
altman	69	6	in	in	IN
altman	69	7	the	the	DT
altman	69	8	specific	specific	JJ
altman	69	9	libraries	library	NNS
altman	69	10	you	-PRON-	PRP
altman	69	11	plan	plan	VBP
altman	69	12	to	to	TO
altman	69	13	use	use	VB
altman	69	14	,	,	,
altman	69	15	and	and	CC
altman	69	16	see	see	VB
altman	69	17	what	what	WDT
altman	69	18	type	type	NN
altman	69	19	of	of	IN
altman	69	20	input	input	NN
altman	69	21	data	datum	NNS
altman	69	22	is	be	VBZ
altman	69	23	required	require	VBN
altman	69	24	.	.	.
altman	70	1	You	-PRON-	PRP
altman	70	2	do	do	VBP
altman	70	3	n’t	not	RB
altman	70	4	have	have	VB
altman	70	5	to	to	TO
altman	70	6	use	use	VB
altman	70	7	these	these	DT
altman	70	8	same	same	JJ
altman	70	9	formats	format	NNS
altman	70	10	during	during	IN
altman	70	11	your	-PRON-	PRP$
altman	70	12	data	data	NN
altman	70	13	preparations	preparation	NNS
altman	70	14	,	,	,
altman	70	15	though	though	IN
altman	70	16	it	-PRON-	PRP
altman	70	17	can	can	MD
altman	70	18	simplify	simplify	VB
altman	70	19	the	the	DT
altman	70	20	process	process	NN
altman	70	21	.	.	.
altman	71	1	Data	datum	NNS
altman	71	2	cleanup	cleanup	NN
altman	71	3	and	and	CC
altman	71	4	transformation	transformation	NN
altman	71	5	is	be	VBZ
altman	71	6	an	an	DT
altman	71	7	art	art	NN
altman	71	8	.	.	.
altman	72	1	Data	datum	NNS
altman	72	2	is	be	VBZ
altman	72	3	messy	messy	JJ
altman	72	4	,	,	,
altman	72	5	and	and	CC
altman	72	6	the	the	DT
altman	72	7	messier	messy	JJR
altman	72	8	the	the	DT
altman	72	9	data	datum	NNS
altman	72	10	,	,	,
altman	72	11	the	the	DT
altman	72	12	harder	hard	JJR
altman	72	13	it	-PRON-	PRP
altman	72	14	is	be	VBZ
altman	72	15	to	to	TO
altman	72	16	analyze	analyze	VB
altman	72	17	and	and	CC
altman	72	18	uncover	uncover	VB
altman	72	19	underlying	underlying	JJ
altman	72	20	patterns	pattern	NNS
altman	72	21	.	.	.
altman	73	1	Yet	yet	CC
altman	73	2	,	,	,
altman	73	3	we	-PRON-	PRP
altman	73	4	are	be	VBP
altman	73	5	only	only	RB
altman	73	6	human	human	JJ
altman	73	7	,	,	,
altman	73	8	and	and	CC
altman	73	9	perfect	perfect	JJ
altman	73	10	data	data	NN
altman	73	11	is	be	VBZ
altman	73	12	far	far	RB
altman	73	13	beyond	beyond	IN
altman	73	14	our	-PRON-	PRP$
altman	73	15	reach	reach	NN
altman	73	16	.	.	.
altman	74	1	To	to	TO
altman	74	2	strike	strike	VB
altman	74	3	a	a	DT
altman	74	4	workable	workable	JJ
altman	74	5	balance	balance	NN
altman	74	6	,	,	,
altman	74	7	focus	focus	VBP
altman	74	8	on	on	IN
altman	74	9	those	those	DT
altman	74	10	cleanup	cleanup	NN
altman	74	11	tasks	task	NNS
altman	74	12	that	that	WDT
altman	74	13	you	-PRON-	PRP
altman	74	14	know	know	VBP
altman	74	15	(	(	-LRB-
altman	74	16	or	or	CC
altman	74	17	strongly	strongly	RB
altman	74	18	suspect	suspect	VB
altman	74	19	)	)	-RRB-
altman	74	20	will	will	MD
altman	74	21	have	have	VB
altman	74	22	a	a	DT
altman	74	23	significant	significant	JJ
altman	74	24	impact	impact	NN
altman	74	25	on	on	IN
altman	74	26	the	the	DT
altman	74	27	final	final	JJ
altman	74	28	product	product	NN
altman	74	29	.	.	.
altman	75	1	Cleanup	cleanup	NN
altman	75	2	and	and	CC
altman	75	3	transformation	transformation	NN
altman	75	4	operations	operation	NNS
altman	75	5	include	include	VBP
altman	75	6	removing	remove	VBG
altman	75	7	punctuation	punctuation	NN
altman	75	8	or	or	CC
altman	75	9	stopwords	stopword	NNS
altman	75	10	from	from	IN
altman	75	11	textual	textual	JJ
altman	75	12	data	datum	NNS
altman	75	13	,	,	,
altman	75	14	standardizing	standardize	VBG
altman	75	15	date	date	NN
altman	75	16	and	and	CC
altman	75	17	number	number	NN
altman	75	18	formats	format	NNS
altman	75	19	,	,	,
altman	75	20	replacing	replace	VBG
altman	75	21	missing	miss	VBG
altman	75	22	or	or	CC
altman	75	23	dummy	dummy	JJ
altman	75	24	values	value	NNS
altman	75	25	with	with	IN
altman	75	26	a	a	DT
altman	75	27	meaningful	meaningful	JJ
altman	75	28	default	default	NN
altman	75	29	,	,	,
altman	75	30	and	and	CC
altman	75	31	excluding	exclude	VBG
altman	75	32	data	datum	NNS
altman	75	33	that	that	WDT
altman	75	34	is	be	VBZ
altman	75	35	known	know	VBN
altman	75	36	to	to	TO
altman	75	37	be	be	VB
altman	75	38	erroneous	erroneous	JJ
altman	75	39	or	or	CC
altman	75	40	atypical	atypical	JJ
altman	75	41	.	.	.
altman	76	1	You	-PRON-	PRP
altman	76	2	will	will	MD
altman	76	3	select	select	VB
altman	76	4	relevant	relevant	JJ
altman	76	5	data	data	NN
altman	76	6	points	point	NNS
altman	76	7	,	,	,
altman	76	8	and	and	CC
altman	76	9	you	-PRON-	PRP
altman	76	10	may	may	MD
altman	76	11	need	need	VB
altman	76	12	to	to	TO
altman	76	13	represent	represent	VB
altman	76	14	them	-PRON-	PRP
altman	76	15	in	in	IN
altman	76	16	a	a	DT
altman	76	17	new	new	JJ
altman	76	18	way	way	NN
altman	76	19	:	:	:
altman	76	20	a	a	DT
altman	76	21	birth	birth	NN
altman	76	22	date	date	NN
altman	76	23	becomes	become	VBZ
altman	76	24	age	age	NN
altman	76	25	range	range	NN
altman	76	26	;	;	:
altman	76	27	a	a	DT
altman	76	28	place	place	NN
altman	76	29	name	name	NN
altman	76	30	becomes	become	VBZ
altman	76	31	geo	geo	NN
altman	76	32	-	-	HYPH
altman	76	33	coordinates	coordinate	NNS
altman	76	34	;	;	:
altman	76	35	a	a	DT
altman	76	36	text	text	NN
altman	76	37	document	document	NN
altman	76	38	becomes	become	VBZ
altman	76	39	a	a	DT
altman	76	40	word	word	NN
altman	76	41	density	density	NN
altman	76	42	vector	vector	NN
altman	76	43	.	.	.
altman	77	1	There	there	EX
altman	77	2	are	be	VBP
altman	77	3	many	many	JJ
altman	77	4	possible	possible	JJ
altman	77	5	normalizations	normalization	NNS
altman	77	6	to	to	TO
altman	77	7	perform	perform	VB
altman	77	8	,	,	,
altman	77	9	depending	depend	VBG
altman	77	10	on	on	IN
altman	77	11	your	-PRON-	PRP$
altman	77	12	dataset	dataset	NN
altman	77	13	and	and	CC
altman	77	14	which	which	WDT
altman	77	15	algorithm(s	algorithm(s	NNP
altman	77	16	)	)	-RRB-
altman	77	17	you	-PRON-	PRP
altman	77	18	plan	plan	VBP
altman	77	19	to	to	TO
altman	77	20	use	use	VB
altman	77	21	.	.	.
altman	78	1	It	-PRON-	PRP
altman	78	2	’s	’	VBZ
altman	78	3	not	not	RB
altman	78	4	a	a	DT
altman	78	5	bad	bad	JJ
altman	78	6	idea	idea	NN
altman	78	7	to	to	TO
altman	78	8	ensure	ensure	VB
altman	78	9	that	that	IN
altman	78	10	there	there	EX
altman	78	11	’s	’	VBZ
altman	78	12	a	a	DT
altman	78	13	genuinely	genuinely	RB
altman	78	14	unique	unique	JJ
altman	78	15	identifier	identifier	NN
altman	78	16	for	for	IN
altman	78	17	each	each	DT
altman	78	18	record	record	NN
altman	78	19	(	(	-LRB-
altman	78	20	even	even	RB
altman	78	21	if	if	IN
altman	78	22	you	-PRON-	PRP
altman	78	23	do	do	VBP
altman	78	24	n’t	not	RB
altman	78	25	see	see	VB
altman	78	26	an	an	DT
altman	78	27	immediate	immediate	JJ
altman	78	28	need	need	NN
altman	78	29	for	for	IN
altman	78	30	one	one	CD
altman	78	31	)	)	-RRB-
altman	78	32	.	.	.
altman	79	1	This	this	DT
altman	79	2	is	be	VBZ
altman	79	3	also	also	RB
altman	79	4	a	a	DT
altman	79	5	good	good	JJ
altman	79	6	time	time	NN
altman	79	7	to	to	TO
altman	79	8	reflect	reflect	VB
altman	79	9	on	on	IN
altman	79	10	any	any	DT
altman	79	11	biases	bias	NNS
altman	79	12	that	that	WDT
altman	79	13	might	may	MD
altman	79	14	be	be	VB
altman	79	15	inherent	inherent	JJ
altman	79	16	in	in	IN
altman	79	17	your	-PRON-	PRP$
altman	79	18	data	datum	NNS
altman	79	19	,	,	,
altman	79	20	and	and	CC
altman	79	21	whether	whether	IN
altman	79	22	or	or	CC
altman	79	23	not	not	RB
altman	79	24	you	-PRON-	PRP
altman	79	25	can	can	MD
altman	79	26	adjust	adjust	VB
altman	79	27	for	for	IN
altman	79	28	them	-PRON-	PRP
altman	79	29	;	;	:
altman	79	30	even	even	RB
altman	79	31	if	if	IN
altman	79	32	you	-PRON-	PRP
altman	79	33	can	can	MD
altman	79	34	not	not	RB
altman	79	35	,	,	,
altman	79	36	understanding	understand	VBG
altman	79	37	how	how	WRB
altman	79	38	they	-PRON-	PRP
altman	79	39	might	may	MD
altman	79	40	impact	impact	VB
altman	79	41	the	the	DT
altman	79	42	ML	ML	NNP
altman	79	43	process	process	NN
altman	79	44	will	will	MD
altman	79	45	help	help	VB
altman	79	46	you	-PRON-	PRP
altman	79	47	conduct	conduct	VB
altman	79	48	a	a	DT
altman	79	49	more	more	RBR
altman	79	50	nuanced	nuanced	JJ
altman	79	51	analysis	analysis	NN
altman	79	52	and	and	CC
altman	79	53	frame	frame	VB
altman	79	54	your	-PRON-	PRP$
altman	79	55	final	final	JJ
altman	79	56	results	result	NNS
altman	79	57	.	.	.
altman	80	1	At	at	IN
altman	80	2	the	the	DT
altman	80	3	very	very	RB
altman	80	4	least	least	JJS
altman	80	5	,	,	,
altman	80	6	you	-PRON-	PRP
altman	80	7	can	can	MD
altman	80	8	record	record	VB
altman	80	9	biases	bias	NNS
altman	80	10	in	in	IN
altman	80	11	the	the	DT
altman	80	12	documentation	documentation	NN
altman	80	13	so	so	IN
altman	80	14	that	that	IN
altman	80	15	future	future	JJ
altman	80	16	researchers	researcher	NNS
altman	80	17	will	will	MD
altman	80	18	be	be	VB
altman	80	19	aware	aware	JJ
altman	80	20	of	of	IN
altman	80	21	them	-PRON-	PRP
altman	80	22	and	and	CC
altman	80	23	react	react	VBP
altman	80	24	accordingly	accordingly	RB
altman	80	25	.	.	.
altman	81	1	As	as	IN
altman	81	2	you	-PRON-	PRP
altman	81	3	become	become	VBP
altman	81	4	more	more	RBR
altman	81	5	familiar	familiar	JJ
altman	81	6	with	with	IN
altman	81	7	the	the	DT
altman	81	8	data	datum	NNS
altman	81	9	,	,	,
altman	81	10	you	-PRON-	PRP
altman	81	11	will	will	MD
altman	81	12	likely	likely	RB
altman	81	13	hone	hone	VB
altman	81	14	your	-PRON-	PRP$
altman	81	15	cleanup	cleanup	NN
altman	81	16	process	process	NN
altman	81	17	and	and	CC
altman	81	18	iterate	iterate	VB
altman	81	19	through	through	IN
altman	81	20	the	the	DT
altman	81	21	steps	step	NNS
altman	81	22	multiple	multiple	JJ
altman	81	23	times	time	NNS
altman	81	24	.	.	.
altman	82	1	The	the	DT
altman	82	2	more	more	RBR
altman	82	3	you	-PRON-	PRP
altman	82	4	can	can	MD
altman	82	5	learn	learn	VB
altman	82	6	about	about	IN
altman	82	7	the	the	DT
altman	82	8	data	datum	NNS
altman	82	9	,	,	,
altman	82	10	the	the	DT
altman	82	11	better	well	JJR
altman	82	12	your	-PRON-	PRP$
altman	82	13	preparations	preparation	NNS
altman	82	14	will	will	MD
altman	82	15	be	be	VB
altman	82	16	.	.	.
altman	83	1	During	during	IN
altman	83	2	the	the	DT
altman	83	3	data	data	NN
altman	83	4	preparation	preparation	NN
altman	83	5	phase	phase	NN
altman	83	6	,	,	,
altman	83	7	practitioners	practitioner	NNS
altman	83	8	often	often	RB
altman	83	9	make	make	VBP
altman	83	10	use	use	NN
altman	83	11	of	of	IN
altman	83	12	visualizations	visualization	NNS
altman	83	13	and	and	CC
altman	83	14	query	query	NN
altman	83	15	frameworks	framework	VBZ
altman	83	16	to	to	TO
altman	83	17	picture	picture	VB
altman	83	18	their	-PRON-	PRP$
altman	83	19	data	datum	NNS
altman	83	20	holistically	holistically	RB
altman	83	21	,	,	,
altman	83	22	identify	identify	VB
altman	83	23	patterns	pattern	NNS
altman	83	24	,	,	,
altman	83	25	and	and	CC
altman	83	26	find	find	VB
altman	83	27	errors	error	NNS
altman	83	28	or	or	CC
altman	83	29	outliers	outlier	NNS
altman	83	30	.	.	.
altman	84	1	Some	some	DT
altman	84	2	ML	ML	NNP
altman	84	3	tools	tool	NNS
altman	84	4	support	support	VBP
altman	84	5	these	these	DT
altman	84	6	features	feature	NNS
altman	84	7	out	out	RP
altman	84	8	-	-	HYPH
altman	84	9	of	of	IN
altman	84	10	-	-	HYPH
altman	84	11	the	the	DT
altman	84	12	-	-	HYPH
altman	84	13	box	box	NN
altman	84	14	,	,	,
altman	84	15	or	or	CC
altman	84	16	are	be	VBP
altman	84	17	intentionally	intentionally	RB
altman	84	18	interoperable	interoperable	JJ
altman	84	19	with	with	IN
altman	84	20	external	external	JJ
altman	84	21	query	query	NN
altman	84	22	and	and	CC
altman	84	23	visualization	visualization	NN
altman	84	24	tools	tool	NNS
altman	84	25	.	.	.
altman	85	1	For	for	IN
altman	85	2	a	a	DT
altman	85	3	lightweight	lightweight	JJ
altman	85	4	tool	tool	NN
altman	85	5	,	,	,
altman	85	6	consider	consider	VB
altman	85	7	spreadsheet	spreadsheet	NN
altman	85	8	or	or	CC
altman	85	9	notebook	notebook	VBP
altman	85	10	software	software	NN
altman	85	11	.	.	.
altman	86	1	Depending	depend	VBG
altman	86	2	on	on	IN
altman	86	3	your	-PRON-	PRP$
altman	86	4	use	use	NN
altman	86	5	case	case	NN
altman	86	6	,	,	,
altman	86	7	it	-PRON-	PRP
altman	86	8	may	may	MD
altman	86	9	be	be	VB
altman	86	10	worthwhile	worthwhile	JJ
altman	86	11	to	to	TO
altman	86	12	put	put	VB
altman	86	13	your	-PRON-	PRP$
altman	86	14	data	datum	NNS
altman	86	15	into	into	IN
altman	86	16	a	a	DT
altman	86	17	temporary	temporary	JJ
altman	86	18	database	database	NN
altman	86	19	or	or	CC
altman	86	20	search	search	NN
altman	86	21	index	index	NN
altman	86	22	so	so	IN
altman	86	23	that	that	IN
altman	86	24	you	-PRON-	PRP
altman	86	25	can	can	MD
altman	86	26	make	make	VB
altman	86	27	use	use	NN
altman	86	28	of	of	IN
altman	86	29	a	a	DT
altman	86	30	more	more	RBR
altman	86	31	sophisticated	sophisticated	JJ
altman	86	32	query	query	NN
altman	86	33	interface	interface	NN
altman	86	34	.	.	.
altman	87	1	Model	model	NN
altman	87	2	testing	testing	NN
altman	87	3	and	and	CC
altman	87	4	training	training	NN
altman	87	5	During	during	IN
altman	87	6	the	the	DT
altman	87	7	testing	testing	NN
altman	87	8	and	and	CC
altman	87	9	training	training	NN
altman	87	10	phase	phase	NN
altman	87	11	,	,	,
altman	87	12	you	-PRON-	PRP
altman	87	13	will	will	MD
altman	87	14	build	build	VB
altman	87	15	multiple	multiple	JJ
altman	87	16	models	model	NNS
altman	87	17	and	and	CC
altman	87	18	determine	determine	VB
altman	87	19	which	which	WDT
altman	87	20	one	one	PRP
altman	87	21	gives	give	VBZ
altman	87	22	you	-PRON-	PRP
altman	87	23	the	the	DT
altman	87	24	best	good	JJS
altman	87	25	results	result	NNS
altman	87	26	.	.	.
altman	88	1	One	one	CD
altman	88	2	of	of	IN
altman	88	3	the	the	DT
altman	88	4	main	main	JJ
altman	88	5	ways	way	NNS
altman	88	6	you	-PRON-	PRP
altman	88	7	will	will	MD
altman	88	8	tune	tune	VB
altman	88	9	your	-PRON-	PRP$
altman	88	10	model	model	NN
altman	88	11	is	be	VBZ
altman	88	12	by	by	IN
altman	88	13	trying	try	VBG
altman	88	14	multiple	multiple	JJ
altman	88	15	combinations	combination	NNS
altman	88	16	of	of	IN
altman	88	17	hyperparameters	hyperparameter	NNS
altman	88	18	.	.	.
altman	89	1	A	a	DT
altman	89	2	hyperparameter	hyperparameter	NN
altman	89	3	is	be	VBZ
altman	89	4	a	a	DT
altman	89	5	value	value	NN
altman	89	6	that	that	WDT
altman	89	7	you	-PRON-	PRP
altman	89	8	set	set	VBP
altman	89	9	before	before	IN
altman	89	10	you	-PRON-	PRP
altman	89	11	run	run	VBP
altman	89	12	the	the	DT
altman	89	13	learning	learning	NN
altman	89	14	process	process	NN
altman	89	15	,	,	,
altman	89	16	which	which	WDT
altman	89	17	impacts	impact	VBZ
altman	89	18	how	how	WRB
altman	89	19	the	the	DT
altman	89	20	learning	learning	NN
altman	89	21	process	process	NN
altman	89	22	works	work	VBZ
altman	89	23	.	.	.
altman	90	1	Hyperparameters	hyperparameter	NNS
altman	90	2	control	control	VBP
altman	90	3	things	thing	NNS
altman	90	4	like	like	IN
altman	90	5	the	the	DT
altman	90	6	number	number	NN
altman	90	7	of	of	IN
altman	90	8	learning	learn	VBG
altman	90	9	cycles	cycle	NNS
altman	90	10	an	an	DT
altman	90	11	algorithm	algorithm	NN
altman	90	12	will	will	MD
altman	90	13	iterate	iterate	VB
altman	90	14	through	through	IN
altman	90	15	,	,	,
altman	90	16	the	the	DT
altman	90	17	number	number	NN
altman	90	18	of	of	IN
altman	90	19	layers	layer	NNS
altman	90	20	in	in	IN
altman	90	21	a	a	DT
altman	90	22	neural	neural	JJ
altman	90	23	net	net	NN
altman	90	24	,	,	,
altman	90	25	the	the	DT
altman	90	26	characteristics	characteristic	NNS
altman	90	27	of	of	IN
altman	90	28	a	a	DT
altman	90	29	cluster	cluster	NN
altman	90	30	,	,	,
altman	90	31	or	or	CC
altman	90	32	the	the	DT
altman	90	33	number	number	NN
altman	90	34	of	of	IN
altman	90	35	decision	decision	NN
altman	90	36	trees	tree	NNS
altman	90	37	in	in	IN
altman	90	38	a	a	DT
altman	90	39	forest	forest	NN
altman	90	40	.	.	.
altman	91	1	Often	often	RB
altman	91	2	,	,	,
altman	91	3	you	-PRON-	PRP
altman	91	4	will	will	MD
altman	91	5	also	also	RB
altman	91	6	want	want	VB
altman	91	7	to	to	TO
altman	91	8	circle	circle	VB
altman	91	9	back	back	RB
altman	91	10	to	to	IN
altman	91	11	your	-PRON-	PRP$
altman	91	12	data	data	NN
altman	91	13	preparation	preparation	NN
altman	91	14	steps	step	NNS
altman	91	15	to	to	TO
altman	91	16	try	try	VB
altman	91	17	different	different	JJ
altman	91	18	configurations	configuration	NNS
altman	91	19	,	,	,
altman	91	20	apply	apply	VB
altman	91	21	new	new	JJ
altman	91	22	enhancements	enhancement	NNS
altman	91	23	,	,	,
altman	91	24	or	or	CC
altman	91	25	address	address	VB
altman	91	26	new	new	JJ
altman	91	27	problems	problem	NNS
altman	91	28	and	and	CC
altman	91	29	particularities	particularity	NNS
altman	91	30	that	that	WDT
altman	91	31	you	-PRON-	PRP
altman	91	32	’ve	have	VB
altman	91	33	uncovered	uncover	VBN
altman	91	34	.	.	.
altman	92	1	The	the	DT
altman	92	2	process	process	NN
altman	92	3	is	be	VBZ
altman	92	4	deceptively	deceptively	RB
altman	92	5	simple	simple	JJ
altman	92	6	:	:	:
altman	92	7	try	try	VB
altman	92	8	out	out	RP
altman	92	9	different	different	JJ
altman	92	10	configurations	configuration	NNS
altman	92	11	until	until	IN
altman	92	12	you	-PRON-	PRP
altman	92	13	get	get	VBP
altman	92	14	a	a	DT
altman	92	15	good	good	JJ
altman	92	16	result	result	NN
altman	92	17	.	.	.
altman	93	1	The	the	DT
altman	93	2	challenge	challenge	NN
altman	93	3	comes	come	VBZ
altman	93	4	when	when	WRB
altman	93	5	you	-PRON-	PRP
altman	93	6	try	try	VBP
altman	93	7	to	to	TO
altman	93	8	define	define	VB
altman	93	9	what	what	WP
altman	93	10	constitutes	constitute	VBZ
altman	93	11	a	a	DT
altman	93	12	good	good	JJ
altman	93	13	(	(	-LRB-
altman	93	14	or	or	CC
altman	93	15	good	good	RB
altman	93	16	-	-	HYPH
altman	93	17	enough	enough	JJ
altman	93	18	)	)	-RRB-
altman	93	19	result	result	NN
altman	93	20	.	.	.
altman	94	1	Measuring	measure	VBG
altman	94	2	the	the	DT
altman	94	3	quality	quality	NN
altman	94	4	of	of	IN
altman	94	5	a	a	DT
altman	94	6	machine	machine	NN
altman	94	7	learning	learn	VBG
altman	94	8	model	model	NN
altman	94	9	takes	take	VBZ
altman	94	10	finesse	finesse	NN
altman	94	11	.	.	.
altman	95	1	Start	start	VB
altman	95	2	by	by	IN
altman	95	3	asking	ask	VBG
altman	95	4	:	:	:
altman	95	5	What	what	WP
altman	95	6	would	would	MD
altman	95	7	you	-PRON-	PRP
altman	95	8	expect	expect	VB
altman	95	9	to	to	TO
altman	95	10	see	see	VB
altman	95	11	if	if	IN
altman	95	12	the	the	DT
altman	95	13	model	model	NN
altman	95	14	learned	learn	VBD
altman	95	15	perfectly	perfectly	RB
altman	95	16	?	?	.
altman	96	1	Equally	equally	RB
altman	96	2	important	important	JJ
altman	96	3	,	,	,
altman	96	4	what	what	WP
altman	96	5	would	would	MD
altman	96	6	you	-PRON-	PRP
altman	96	7	expect	expect	VB
altman	96	8	to	to	TO
altman	96	9	see	see	VB
altman	96	10	if	if	IN
altman	96	11	the	the	DT
altman	96	12	model	model	NN
altman	96	13	did	do	VBD
altman	96	14	n’t	not	RB
altman	96	15	learn	learn	VB
altman	96	16	anything	anything	NN
altman	96	17	at	at	RB
altman	96	18	all	all	RB
altman	96	19	?	?	.
altman	97	1	You	-PRON-	PRP
altman	97	2	can	can	MD
altman	97	3	often	often	RB
altman	97	4	utilize	utilize	VB
altman	97	5	randomness	randomness	NN
altman	97	6	as	as	IN
altman	97	7	a	a	DT
altman	97	8	stand	stand	NN
altman	97	9	-	-	HYPH
altman	97	10	in	in	NN
altman	97	11	for	for	IN
altman	97	12	no	no	DT
altman	97	13	learning	learning	NN
altman	97	14	,	,	,
altman	97	15	e.g.	e.g.	RB
altman	98	1	“	"	``
altman	98	2	if	if	IN
altman	98	3	a	a	DT
altman	98	4	result	result	NN
altman	98	5	was	be	VBD
altman	98	6	selected	select	VBN
altman	98	7	at	at	IN
altman	98	8	random	random	JJ
altman	98	9	,	,	,
altman	98	10	the	the	DT
altman	98	11	probability	probability	NN
altman	98	12	of	of	IN
altman	98	13	the	the	DT
altman	98	14	desired	desire	VBN
altman	98	15	outcome	outcome	NN
altman	98	16	would	would	MD
altman	98	17	be	be	VB
altman	98	18	X	x	NN
altman	98	19	”	"	''
altman	98	20	.	.	.
altman	99	1	These	these	DT
altman	99	2	two	two	CD
altman	99	3	questions	question	NNS
altman	99	4	will	will	MD
altman	99	5	help	help	VB
altman	99	6	you	-PRON-	PRP
altman	99	7	to	to	TO
altman	99	8	set	set	VB
altman	99	9	benchmarks	benchmark	NNS
altman	99	10	at	at	IN
altman	99	11	both	both	DT
altman	99	12	extremes	extreme	NNS
altman	99	13	of	of	IN
altman	99	14	the	the	DT
altman	99	15	realm	realm	NN
altman	99	16	of	of	IN
altman	99	17	possible	possible	JJ
altman	99	18	outcomes	outcome	NNS
altman	99	19	.	.	.
altman	100	1	Perfection	perfection	NN
altman	100	2	is	be	VBZ
altman	100	3	illusive	illusive	JJ
altman	100	4	,	,	,
altman	100	5	and	and	CC
altman	100	6	the	the	DT
altman	100	7	return	return	NN
altman	100	8	on	on	IN
altman	100	9	investment	investment	NN
altman	100	10	dwindles	dwindle	NNS
altman	100	11	after	after	IN
altman	100	12	a	a	DT
altman	100	13	while	while	NN
altman	100	14	,	,	,
altman	100	15	so	so	RB
altman	100	16	be	be	VB
altman	100	17	prepared	prepared	JJ
altman	100	18	to	to	TO
altman	100	19	stop	stop	VB
altman	100	20	training	train	VBG
altman	100	21	once	once	IN
altman	100	22	you	-PRON-	PRP
altman	100	23	’ve	have	VB
altman	100	24	arrived	arrive	VBD
altman	100	25	at	at	IN
altman	100	26	an	an	DT
altman	100	27	acceptably	acceptably	RB
altman	100	28	good	good	JJ
altman	100	29	model	model	NN
altman	100	30	.	.	.
altman	101	1	In	in	IN
altman	101	2	a	a	DT
altman	101	3	supervised	supervise	VBN
altman	101	4	learning	learning	NN
altman	101	5	problem	problem	NN
altman	101	6	the	the	DT
altman	101	7	dataset	dataset	NN
altman	101	8	is	be	VBZ
altman	101	9	split	split	VBN
altman	101	10	into	into	IN
altman	101	11	training	training	NN
altman	101	12	and	and	CC
altman	101	13	testing	testing	NN
altman	101	14	datasets	dataset	NNS
altman	101	15	.	.	.
altman	102	1	The	the	DT
altman	102	2	algorithm	algorithm	NNP
altman	102	3	uses	use	VBZ
altman	102	4	the	the	DT
altman	102	5	training	training	NN
altman	102	6	data	datum	NNS
altman	102	7	to	to	TO
altman	102	8	“	"	``
altman	102	9	learn	learn	VB
altman	102	10	”	"	''
altman	102	11	a	a	DT
altman	102	12	set	set	NN
altman	102	13	of	of	IN
altman	102	14	rules	rule	NNS
altman	102	15	that	that	WDT
altman	102	16	it	-PRON-	PRP
altman	102	17	can	can	MD
altman	102	18	subsequently	subsequently	RB
altman	102	19	apply	apply	VB
altman	102	20	to	to	IN
altman	102	21	new	new	JJ
altman	102	22	,	,	,
altman	102	23	unseen	unseen	JJ
altman	102	24	data	datum	NNS
altman	102	25	to	to	TO
altman	102	26	predict	predict	VB
altman	102	27	the	the	DT
altman	102	28	outcome	outcome	NN
altman	102	29	.	.	.
altman	103	1	The	the	DT
altman	103	2	testing	testing	NN
altman	103	3	dataset	dataset	NN
altman	103	4	(	(	-LRB-
altman	103	5	also	also	RB
altman	103	6	called	call	VBN
altman	103	7	a	a	DT
altman	103	8	validation	validation	NN
altman	103	9	dataset	dataset	NN
altman	103	10	)	)	-RRB-
altman	103	11	is	be	VBZ
altman	103	12	used	use	VBN
altman	103	13	to	to	TO
altman	103	14	test	test	VB
altman	103	15	how	how	WRB
altman	103	16	well	well	RB
altman	103	17	the	the	DT
altman	103	18	model	model	NN
altman	103	19	performs	perform	VBZ
altman	103	20	.	.	.
altman	104	1	Often	often	RB
altman	104	2	,	,	,
altman	104	3	a	a	DT
altman	104	4	third	third	JJ
altman	104	5	dataset	dataset	NN
altman	104	6	is	be	VBZ
altman	104	7	held	hold	VBN
altman	104	8	out	out	RP
altman	104	9	as	as	RB
altman	104	10	well	well	RB
altman	104	11	,	,	,
altman	104	12	reserved	reserve	VBN
altman	104	13	for	for	IN
altman	104	14	final	final	JJ
altman	104	15	testing	testing	NN
altman	104	16	after	after	IN
altman	104	17	the	the	DT
altman	104	18	model	model	NN
altman	104	19	has	have	VBZ
altman	104	20	been	be	VBN
altman	104	21	trained	train	VBN
altman	104	22	.	.	.
altman	105	1	This	this	DT
altman	105	2	third	third	JJ
altman	105	3	dataset	dataset	NN
altman	105	4	provides	provide	VBZ
altman	105	5	an	an	DT
altman	105	6	additional	additional	JJ
altman	105	7	bulwark	bulwark	NN
altman	105	8	against	against	IN
altman	105	9	bias	bias	NN
altman	105	10	and	and	CC
altman	105	11	overfitting	overfitting	NN
altman	105	12	.	.	.
altman	106	1	Results	result	NNS
altman	106	2	are	be	VBP
altman	106	3	typically	typically	RB
altman	106	4	evaluated	evaluate	VBN
altman	106	5	based	base	VBN
altman	106	6	on	on	IN
altman	106	7	some	some	DT
altman	106	8	statistical	statistical	JJ
altman	106	9	measurement	measurement	NN
altman	106	10	that	that	WDT
altman	106	11	is	be	VBZ
altman	106	12	directly	directly	RB
altman	106	13	relevant	relevant	JJ
altman	106	14	to	to	IN
altman	106	15	your	-PRON-	PRP$
altman	106	16	research	research	NN
altman	106	17	question	question	NN
altman	106	18	.	.	.
altman	107	1	In	in	IN
altman	107	2	a	a	DT
altman	107	3	classification	classification	NN
altman	107	4	problem	problem	NN
altman	107	5	,	,	,
altman	107	6	you	-PRON-	PRP
altman	107	7	might	may	MD
altman	107	8	optimize	optimize	VB
altman	107	9	for	for	IN
altman	107	10	recall	recall	NN
altman	107	11	or	or	CC
altman	107	12	precision	precision	NN
altman	107	13	.	.	.
altman	108	1	In	in	IN
altman	108	2	a	a	DT
altman	108	3	regression	regression	NN
altman	108	4	problem	problem	NN
altman	108	5	,	,	,
altman	108	6	you	-PRON-	PRP
altman	108	7	can	can	MD
altman	108	8	use	use	VB
altman	108	9	formulas	formula	NNS
altman	108	10	such	such	JJ
altman	108	11	as	as	IN
altman	108	12	the	the	DT
altman	108	13	root	root	NN
altman	108	14	-	-	HYPH
altman	108	15	mean	mean	JJ
altman	108	16	square	square	JJ
altman	108	17	deviation	deviation	NN
altman	108	18	to	to	TO
altman	108	19	measure	measure	VB
altman	108	20	how	how	WRB
altman	108	21	well	well	RB
altman	108	22	the	the	DT
altman	108	23	regression	regression	NN
altman	108	24	line	line	NN
altman	108	25	matches	match	VBZ
altman	108	26	the	the	DT
altman	108	27	actual	actual	JJ
altman	108	28	data	data	NN
altman	108	29	points	point	NNS
altman	108	30	.	.	.
altman	109	1	How	how	WRB
altman	109	2	you	-PRON-	PRP
altman	109	3	choose	choose	VBP
altman	109	4	to	to	TO
altman	109	5	optimize	optimize	VB
altman	109	6	your	-PRON-	PRP$
altman	109	7	model	model	NN
altman	109	8	will	will	MD
altman	109	9	depend	depend	VB
altman	109	10	on	on	IN
altman	109	11	your	-PRON-	PRP$
altman	109	12	specific	specific	JJ
altman	109	13	context	context	NN
altman	109	14	and	and	CC
altman	109	15	priorities	priority	NNS
altman	109	16	.	.	.
altman	110	1	Testing	test	VBG
altman	110	2	an	an	DT
altman	110	3	unsupervised	unsupervised	JJ
altman	110	4	model	model	NN
altman	110	5	is	be	VBZ
altman	110	6	not	not	RB
altman	110	7	as	as	RB
altman	110	8	straightforward	straightforward	JJ
altman	110	9	,	,	,
altman	110	10	since	since	IN
altman	110	11	there	there	EX
altman	110	12	is	be	VBZ
altman	110	13	no	no	DT
altman	110	14	preconceived	preconceived	JJ
altman	110	15	notion	notion	NN
altman	110	16	of	of	IN
altman	110	17	correct	correct	JJ
altman	110	18	and	and	CC
altman	110	19	incorrect	incorrect	JJ
altman	110	20	categorization	categorization	NN
altman	110	21	.	.	.
altman	111	1	You	-PRON-	PRP
altman	111	2	can	can	MD
altman	111	3	sometimes	sometimes	RB
altman	111	4	rely	rely	VB
altman	111	5	on	on	IN
altman	111	6	a	a	DT
altman	111	7	known	know	VBN
altman	111	8	pattern	pattern	NN
altman	111	9	in	in	IN
altman	111	10	the	the	DT
altman	111	11	underlying	underlying	JJ
altman	111	12	dataset	dataset	NN
altman	111	13	that	that	IN
altman	111	14	you	-PRON-	PRP
altman	111	15	would	would	MD
altman	111	16	reasonably	reasonably	RB
altman	111	17	expect	expect	VB
altman	111	18	to	to	TO
altman	111	19	be	be	VB
altman	111	20	reflected	reflect	VBN
altman	111	21	in	in	IN
altman	111	22	a	a	DT
altman	111	23	successful	successful	JJ
altman	111	24	model	model	NN
altman	111	25	.	.	.
altman	112	1	There	there	EX
altman	112	2	may	may	MD
altman	112	3	also	also	RB
altman	112	4	be	be	VB
altman	112	5	characteristics	characteristic	NNS
altman	112	6	of	of	IN
altman	112	7	the	the	DT
altman	112	8	final	final	JJ
altman	112	9	model	model	NN
altman	112	10	that	that	WDT
altman	112	11	indicate	indicate	VBP
altman	112	12	success	success	NN
altman	112	13	.	.	.
altman	113	1	For	for	IN
altman	113	2	example	example	NN
altman	113	3	,	,	,
altman	113	4	if	if	IN
altman	113	5	you	-PRON-	PRP
altman	113	6	are	be	VBP
altman	113	7	working	work	VBG
altman	113	8	with	with	IN
altman	113	9	a	a	DT
altman	113	10	clustering	cluster	VBG
altman	113	11	algorithm	algorithm	NN
altman	113	12	,	,	,
altman	113	13	models	model	NNS
altman	113	14	with	with	IN
altman	113	15	dense	dense	JJ
altman	113	16	,	,	,
altman	113	17	well	well	RB
altman	113	18	-	-	HYPH
altman	113	19	defined	define	VBN
altman	113	20	clusters	cluster	NNS
altman	113	21	are	be	VBP
altman	113	22	probably	probably	RB
altman	113	23	better	well	JJR
altman	113	24	than	than	IN
altman	113	25	sparse	sparse	VB
altman	113	26	clusters	cluster	NNS
altman	113	27	with	with	IN
altman	113	28	vague	vague	JJ
altman	113	29	boundaries	boundary	NNS
altman	113	30	.	.	.
altman	114	1	In	in	IN
altman	114	2	unsupervised	unsupervised	JJ
altman	114	3	learning	learning	NN
altman	114	4	,	,	,
altman	114	5	you	-PRON-	PRP
altman	114	6	may	may	MD
altman	114	7	want	want	VB
altman	114	8	to	to	TO
altman	114	9	hold	hold	VB
altman	114	10	back	back	RP
altman	114	11	some	some	DT
altman	114	12	portion	portion	NN
altman	114	13	of	of	IN
altman	114	14	your	-PRON-	PRP$
altman	114	15	data	datum	NNS
altman	114	16	to	to	TO
altman	114	17	perform	perform	VB
altman	114	18	an	an	DT
altman	114	19	independent	independent	JJ
altman	114	20	validation	validation	NN
altman	114	21	of	of	IN
altman	114	22	your	-PRON-	PRP$
altman	114	23	results	result	NNS
altman	114	24	,	,	,
altman	114	25	or	or	CC
altman	114	26	you	-PRON-	PRP
altman	114	27	may	may	MD
altman	114	28	use	use	VB
altman	114	29	the	the	DT
altman	114	30	entire	entire	JJ
altman	114	31	dataset	dataset	NN
altman	114	32	to	to	TO
altman	114	33	build	build	VB
altman	114	34	the	the	DT
altman	114	35	model	model	NN
altman	114	36	—	—	:
altman	114	37	it	-PRON-	PRP
altman	114	38	depends	depend	VBZ
altman	114	39	on	on	IN
altman	114	40	what	what	WDT
altman	114	41	type	type	NN
altman	114	42	of	of	IN
altman	114	43	testing	testing	NN
altman	114	44	you	-PRON-	PRP
altman	114	45	want	want	VBP
altman	114	46	to	to	TO
altman	114	47	perform	perform	VB
altman	114	48	.	.	.
altman	115	1	Application	application	NN
altman	115	2	of	of	IN
altman	115	3	results	result	NNS
altman	115	4	As	as	IN
altman	115	5	the	the	DT
altman	115	6	final	final	JJ
altman	115	7	step	step	NN
altman	115	8	of	of	IN
altman	115	9	your	-PRON-	PRP$
altman	115	10	workflow	workflow	NN
altman	115	11	,	,	,
altman	115	12	you	-PRON-	PRP
altman	115	13	will	will	MD
altman	115	14	use	use	VB
altman	115	15	your	-PRON-	PRP$
altman	115	16	intelligent	intelligent	JJ
altman	115	17	model	model	NN
altman	115	18	to	to	TO
altman	115	19	perform	perform	VB
altman	115	20	some	some	DT
altman	115	21	task	task	NN
altman	115	22	.	.	.
altman	116	1	Perhaps	perhaps	RB
altman	116	2	you	-PRON-	PRP
altman	116	3	will	will	MD
altman	116	4	use	use	VB
altman	116	5	it	-PRON-	PRP
altman	116	6	for	for	IN
altman	116	7	scholarly	scholarly	JJ
altman	116	8	analysis	analysis	NN
altman	116	9	of	of	IN
altman	116	10	a	a	DT
altman	116	11	dataset	dataset	NN
altman	116	12	,	,	,
altman	116	13	or	or	CC
altman	116	14	perhaps	perhaps	RB
altman	116	15	you	-PRON-	PRP
altman	116	16	will	will	MD
altman	116	17	integrate	integrate	VB
altman	116	18	it	-PRON-	PRP
altman	116	19	into	into	IN
altman	116	20	a	a	DT
altman	116	21	software	software	NN
altman	116	22	project	project	NN
altman	116	23	.	.	.
altman	117	1	If	if	IN
altman	117	2	it	-PRON-	PRP
altman	117	3	is	be	VBZ
altman	117	4	the	the	DT
altman	117	5	former	former	JJ
altman	117	6	,	,	,
altman	117	7	consider	consider	VB
altman	117	8	how	how	WRB
altman	117	9	to	to	TO
altman	117	10	export	export	VB
altman	117	11	any	any	DT
altman	117	12	final	final	JJ
altman	117	13	data	datum	NNS
altman	117	14	and	and	CC
altman	117	15	preserve	preserve	VB
altman	117	16	the	the	DT
altman	117	17	artifacts	artifact	NNS
altman	117	18	of	of	IN
altman	117	19	your	-PRON-	PRP$
altman	117	20	project	project	NN
altman	117	21	.	.	.
altman	118	1	If	if	IN
altman	118	2	it	-PRON-	PRP
altman	118	3	is	be	VBZ
altman	118	4	the	the	DT
altman	118	5	latter	latter	JJ
altman	118	6	,	,	,
altman	118	7	consider	consider	VB
altman	118	8	how	how	WRB
altman	118	9	the	the	DT
altman	118	10	model	model	NN
altman	118	11	,	,	,
altman	118	12	its	-PRON-	PRP$
altman	118	13	outputs	output	NNS
altman	118	14	,	,	,
altman	118	15	and	and	CC
altman	118	16	its	-PRON-	PRP$
altman	118	17	continued	continued	JJ
altman	118	18	maintenance	maintenance	NN
altman	118	19	will	will	MD
altman	118	20	fit	fit	VB
altman	118	21	into	into	IN
altman	118	22	existing	exist	VBG
altman	118	23	systems	system	NNS
altman	118	24	and	and	CC
altman	118	25	workflows	workflow	NNS
altman	118	26	.	.	.
altman	119	1	Planning	plan	VBG
altman	119	2	for	for	IN
altman	119	3	interoperability	interoperability	NN
altman	119	4	may	may	MD
altman	119	5	influence	influence	VB
altman	119	6	decisions	decision	NNS
altman	119	7	from	from	IN
altman	119	8	tool	tool	NN
altman	119	9	selection	selection	NN
altman	119	10	to	to	IN
altman	119	11	data	data	NN
altman	119	12	formats	format	NNS
altman	119	13	and	and	CC
altman	119	14	storage	storage	NN
altman	119	15	.	.	.
altman	120	1	Immutable	immutable	JJ
altman	120	2	data	datum	NNS
altman	120	3	storage	storage	NN
altman	120	4	Immutable	immutable	JJ
altman	120	5	data	datum	NNS
altman	120	6	storage	storage	NN
altman	120	7	can	can	MD
altman	120	8	benefit	benefit	VB
altman	120	9	the	the	DT
altman	120	10	batch	batch	NN
altman	120	11	-	-	HYPH
altman	120	12	processing	process	VBG
altman	120	13	ML	ML	NNP
altman	120	14	pipeline	pipeline	NN
altman	120	15	,	,	,
altman	120	16	especially	especially	RB
altman	120	17	during	during	IN
altman	120	18	the	the	DT
altman	120	19	initial	initial	JJ
altman	120	20	research	research	NN
altman	120	21	and	and	CC
altman	120	22	development	development	NN
altman	120	23	phase	phase	NN
altman	120	24	.	.	.
altman	121	1	This	this	DT
altman	121	2	type	type	NN
altman	121	3	of	of	IN
altman	121	4	data	datum	NNS
altman	121	5	storage	storage	NN
altman	121	6	supports	support	VBZ
altman	121	7	iteration	iteration	NN
altman	121	8	and	and	CC
altman	121	9	allows	allow	VBZ
altman	121	10	you	-PRON-	PRP
altman	121	11	to	to	TO
altman	121	12	compare	compare	VB
altman	121	13	the	the	DT
altman	121	14	results	result	NNS
altman	121	15	of	of	IN
altman	121	16	many	many	JJ
altman	121	17	different	different	JJ
altman	121	18	experiments	experiment	NNS
altman	121	19	.	.	.
altman	122	1	Treating	treat	VBG
altman	122	2	data	datum	NNS
altman	122	3	as	as	IN
altman	122	4	immutable	immutable	JJ
altman	122	5	means	mean	VBZ
altman	122	6	that	that	IN
altman	122	7	after	after	IN
altman	122	8	each	each	DT
altman	122	9	significant	significant	JJ
altman	122	10	change	change	NN
altman	122	11	or	or	CC
altman	122	12	set	set	NN
altman	122	13	of	of	IN
altman	122	14	changes	change	NNS
altman	122	15	to	to	IN
altman	122	16	your	-PRON-	PRP$
altman	122	17	data	datum	NNS
altman	122	18	,	,	,
altman	122	19	you	-PRON-	PRP
altman	122	20	save	save	VBP
altman	122	21	a	a	DT
altman	122	22	new	new	JJ
altman	122	23	snapshot	snapshot	NN
altman	122	24	of	of	IN
altman	122	25	the	the	DT
altman	122	26	dataset	dataset	NN
altman	122	27	that	that	WDT
altman	122	28	is	be	VBZ
altman	122	29	never	never	RB
altman	122	30	edited	edit	VBN
altman	122	31	or	or	CC
altman	122	32	changed	change	VBN
altman	122	33	.	.	.
altman	123	1	It	-PRON-	PRP
altman	123	2	also	also	RB
altman	123	3	allows	allow	VBZ
altman	123	4	you	-PRON-	PRP
altman	123	5	to	to	TO
altman	123	6	be	be	VB
altman	123	7	flexible	flexible	JJ
altman	123	8	and	and	CC
altman	123	9	adaptive	adaptive	JJ
altman	123	10	with	with	IN
altman	123	11	your	-PRON-	PRP$
altman	123	12	data	data	NN
altman	123	13	model	model	NN
altman	123	14	.	.	.
altman	124	1	Immutable	immutable	JJ
altman	124	2	data	datum	NNS
altman	124	3	storage	storage	NN
altman	124	4	has	have	VBZ
altman	124	5	become	become	VBN
altman	124	6	a	a	DT
altman	124	7	popular	popular	JJ
altman	124	8	choice	choice	NN
altman	124	9	for	for	IN
altman	124	10	data	data	NN
altman	124	11	-	-	HYPH
altman	124	12	intensive	intensive	JJ
altman	124	13	or	or	CC
altman	124	14	“	"	``
altman	124	15	big	big	JJ
altman	124	16	data	datum	NNS
altman	124	17	”	"	''
altman	124	18	applications	application	NNS
altman	124	19	as	as	IN
altman	124	20	a	a	DT
altman	124	21	way	way	NN
altman	124	22	to	to	TO
altman	124	23	easily	easily	RB
altman	124	24	assemble	assemble	VB
altman	124	25	large	large	JJ
altman	124	26	quantities	quantity	NNS
altman	124	27	of	of	IN
altman	124	28	data	datum	NNS
altman	124	29	,	,	,
altman	124	30	often	often	RB
altman	124	31	from	from	IN
altman	124	32	multiple	multiple	JJ
altman	124	33	sources	source	NNS
altman	124	34	,	,	,
altman	124	35	without	without	IN
altman	124	36	having	have	VBG
altman	124	37	to	to	TO
altman	124	38	spend	spend	VB
altman	124	39	time	time	NN
altman	124	40	upfront	upfront	NN
altman	124	41	crafting	craft	VBG
altman	124	42	a	a	DT
altman	124	43	strict	strict	JJ
altman	124	44	data	data	NN
altman	124	45	model	model	NN
altman	124	46	.	.	.
altman	125	1	You	-PRON-	PRP
altman	125	2	may	may	MD
altman	125	3	have	have	VB
altman	125	4	heard	hear	VBN
altman	125	5	the	the	DT
altman	125	6	term	term	NN
altman	125	7	“	"	``
altman	125	8	data	data	NNP
altman	125	9	lake	lake	NN
altman	125	10	”	"	''
altman	125	11	to	to	TO
altman	125	12	refer	refer	VB
altman	125	13	to	to	IN
altman	125	14	such	such	JJ
altman	125	15	large	large	JJ
altman	125	16	,	,	,
altman	125	17	unstructured	unstructured	JJ
altman	125	18	collections	collection	NNS
altman	125	19	of	of	IN
altman	125	20	data	datum	NNS
altman	125	21	.	.	.
altman	126	1	This	this	DT
altman	126	2	can	can	MD
altman	126	3	be	be	VB
altman	126	4	contrasted	contrast	VBN
altman	126	5	with	with	IN
altman	126	6	a	a	DT
altman	126	7	“	"	``
altman	126	8	data	data	NN
altman	126	9	warehouse	warehouse	NN
altman	126	10	”	"	''
altman	126	11	,	,	,
altman	126	12	which	which	WDT
altman	126	13	usually	usually	RB
altman	126	14	indicates	indicate	VBZ
altman	126	15	a	a	DT
altman	126	16	highly	highly	RB
altman	126	17	structured	structured	JJ
altman	126	18	,	,	,
altman	126	19	centralized	centralize	VBD
altman	126	20	repository	repository	NN
altman	126	21	such	such	JJ
altman	126	22	as	as	IN
altman	126	23	a	a	DT
altman	126	24	relational	relational	JJ
altman	126	25	database	database	NN
altman	126	26	.	.	.
altman	127	1	To	to	TO
altman	127	2	demonstrate	demonstrate	VB
altman	127	3	how	how	WRB
altman	127	4	immutable	immutable	JJ
altman	127	5	supports	support	NNS
altman	127	6	iteration	iteration	NN
altman	127	7	and	and	CC
altman	127	8	experimentation	experimentation	NN
altman	127	9	,	,	,
altman	127	10	consider	consider	VB
altman	127	11	the	the	DT
altman	127	12	following	follow	VBG
altman	127	13	scenario	scenario	NN
altman	127	14	:	:	:
altman	127	15	You	-PRON-	PRP
altman	127	16	start	start	VBP
altman	127	17	with	with	IN
altman	127	18	an	an	DT
altman	127	19	input	input	NN
altman	127	20	file	file	NN
altman	127	21	my_data.csv	my_data.csv	NN
altman	127	22	,	,	,
altman	127	23	and	and	CC
altman	127	24	then	then	RB
altman	127	25	perform	perform	VB
altman	127	26	some	some	DT
altman	127	27	cleanup	cleanup	JJ
altman	127	28	operation	operation	NN
altman	127	29	over	over	IN
altman	127	30	the	the	DT
altman	127	31	data	datum	NNS
altman	127	32	,	,	,
altman	127	33	such	such	JJ
altman	127	34	as	as	IN
altman	127	35	converting	convert	VBG
altman	127	36	all	all	DT
altman	127	37	measurements	measurement	NNS
altman	127	38	in	in	IN
altman	127	39	miles	mile	NNS
altman	127	40	to	to	IN
altman	127	41	kilometers	kilometer	NNS
altman	127	42	,	,	,
altman	127	43	rounded	round	VBD
altman	127	44	to	to	IN
altman	127	45	the	the	DT
altman	127	46	nearest	near	JJS
altman	127	47	whole	whole	JJ
altman	127	48	number	number	NN
altman	127	49	.	.	.
altman	128	1	If	if	IN
altman	128	2	you	-PRON-	PRP
altman	128	3	were	be	VBD
altman	128	4	treating	treat	VBG
altman	128	5	your	-PRON-	PRP$
altman	128	6	data	datum	NNS
altman	128	7	as	as	IN
altman	128	8	mutable	mutable	JJ
altman	128	9	,	,	,
altman	128	10	you	-PRON-	PRP
altman	128	11	might	may	MD
altman	128	12	overwrite	overwrite	VB
altman	128	13	the	the	DT
altman	128	14	original	original	JJ
altman	128	15	contents	content	NNS
altman	128	16	of	of	IN
altman	128	17	my_data.csv	my_data.csv	NNP
altman	128	18	with	with	IN
altman	128	19	the	the	DT
altman	128	20	transformed	transform	VBN
altman	128	21	values	value	NNS
altman	128	22	.	.	.
altman	129	1	The	the	DT
altman	129	2	problem	problem	NN
altman	129	3	with	with	IN
altman	129	4	this	this	DT
altman	129	5	approach	approach	NN
altman	129	6	comes	come	VBZ
altman	129	7	if	if	IN
altman	129	8	you	-PRON-	PRP
altman	129	9	want	want	VBP
altman	129	10	to	to	TO
altman	129	11	test	test	VB
altman	129	12	some	some	DT
altman	129	13	alteration	alteration	NN
altman	129	14	of	of	IN
altman	129	15	your	-PRON-	PRP$
altman	129	16	cleanup	cleanup	NN
altman	129	17	operation	operation	NN
altman	129	18	.	.	.
altman	130	1	Say	say	VB
altman	130	2	,	,	,
altman	130	3	for	for	IN
altman	130	4	example	example	NN
altman	130	5	,	,	,
altman	130	6	you	-PRON-	PRP
altman	130	7	wanted	want	VBD
altman	130	8	to	to	TO
altman	130	9	round	round	VB
altman	130	10	all	all	DT
altman	130	11	your	-PRON-	PRP$
altman	130	12	conversions	conversion	NNS
altman	130	13	to	to	IN
altman	130	14	the	the	DT
altman	130	15	nearest	near	JJS
altman	130	16	tenth	tenth	NN
altman	130	17	instead	instead	RB
altman	130	18	.	.	.
altman	131	1	Since	since	IN
altman	131	2	you	-PRON-	PRP
altman	131	3	no	no	RB
altman	131	4	longer	longer	RB
altman	131	5	have	have	VBP
altman	131	6	your	-PRON-	PRP$
altman	131	7	original	original	JJ
altman	131	8	data	datum	NNS
altman	131	9	,	,	,
altman	131	10	you	-PRON-	PRP
altman	131	11	would	would	MD
altman	131	12	have	have	VB
altman	131	13	to	to	TO
altman	131	14	start	start	VB
altman	131	15	the	the	DT
altman	131	16	entire	entire	JJ
altman	131	17	ML	ML	NNP
altman	131	18	process	process	NN
altman	131	19	from	from	IN
altman	131	20	the	the	DT
altman	131	21	top	top	NN
altman	131	22	.	.	.
altman	132	1	If	if	IN
altman	132	2	you	-PRON-	PRP
altman	132	3	instead	instead	RB
altman	132	4	treated	treat	VBD
altman	132	5	your	-PRON-	PRP$
altman	132	6	data	datum	NNS
altman	132	7	as	as	IN
altman	132	8	immutable	immutable	JJ
altman	132	9	,	,	,
altman	132	10	you	-PRON-	PRP
altman	132	11	would	would	MD
altman	132	12	keep	keep	VB
altman	132	13	my_data.csv	my_data.csv	NNP
altman	132	14	in	in	IN
altman	132	15	its	-PRON-	PRP$
altman	132	16	original	original	JJ
altman	132	17	state	state	NN
altman	132	18	,	,	,
altman	132	19	and	and	CC
altman	132	20	save	save	VB
altman	132	21	the	the	DT
altman	132	22	output	output	NN
altman	132	23	of	of	IN
altman	132	24	your	-PRON-	PRP$
altman	132	25	cleanup	cleanup	NN
altman	132	26	operation	operation	NN
altman	132	27	in	in	IN
altman	132	28	a	a	DT
altman	132	29	new	new	JJ
altman	132	30	file	file	NN
altman	132	31	,	,	,
altman	132	32	say	say	VBP
altman	132	33	my_clean_data.csv	my_clean_data.csv	CD
altman	132	34	.	.	.
altman	133	1	That	that	DT
altman	133	2	way	way	NN
altman	133	3	,	,	,
altman	133	4	you	-PRON-	PRP
altman	133	5	could	could	MD
altman	133	6	return	return	VB
altman	133	7	to	to	IN
altman	133	8	my_data.csv	my_data.csv	NNP
altman	133	9	as	as	IN
altman	133	10	many	many	JJ
altman	133	11	times	time	NNS
altman	133	12	as	as	IN
altman	133	13	you	-PRON-	PRP
altman	133	14	wished	wish	VBD
altman	133	15	,	,	,
altman	133	16	try	try	VB
altman	133	17	different	different	JJ
altman	133	18	operations	operation	NNS
altman	133	19	on	on	IN
altman	133	20	this	this	DT
altman	133	21	data	data	NN
altman	133	22	,	,	,
altman	133	23	and	and	CC
altman	133	24	easily	easily	RB
altman	133	25	compare	compare	VB
altman	133	26	the	the	DT
altman	133	27	results	result	NNS
altman	133	28	of	of	IN
altman	133	29	these	these	DT
altman	133	30	operations	operation	NNS
altman	133	31	knowing	know	VBG
altman	133	32	the	the	DT
altman	133	33	source	source	NN
altman	133	34	data	datum	NNS
altman	133	35	was	be	VBD
altman	133	36	exactly	exactly	RB
altman	133	37	the	the	DT
altman	133	38	same	same	JJ
altman	133	39	for	for	IN
altman	133	40	each	each	DT
altman	133	41	one	one	NN
altman	133	42	.	.	.
altman	134	1	Think	think	VB
altman	134	2	of	of	IN
altman	134	3	each	each	DT
altman	134	4	immutable	immutable	JJ
altman	134	5	dataset	dataset	NN
altman	134	6	as	as	IN
altman	134	7	a	a	DT
altman	134	8	place	place	NN
altman	134	9	in	in	IN
altman	134	10	your	-PRON-	PRP$
altman	134	11	process	process	NN
altman	134	12	that	that	WDT
altman	134	13	you	-PRON-	PRP
altman	134	14	can	can	MD
altman	134	15	safely	safely	RB
altman	134	16	rewind	rewind	VB
altman	134	17	to	to	TO
altman	134	18	anytime	anytime	VB
altman	134	19	you	-PRON-	PRP
altman	134	20	want	want	VBP
altman	134	21	to	to	TO
altman	134	22	try	try	VB
altman	134	23	something	something	NN
altman	134	24	new	new	JJ
altman	134	25	or	or	CC
altman	134	26	correct	correct	JJ
altman	134	27	for	for	IN
altman	134	28	some	some	DT
altman	134	29	error	error	NN
altman	134	30	or	or	CC
altman	134	31	failure	failure	NN
altman	134	32	.	.	.
altman	135	1	To	to	TO
altman	135	2	illustrate	illustrate	VB
altman	135	3	the	the	DT
altman	135	4	benefits	benefit	NNS
altman	135	5	of	of	IN
altman	135	6	a	a	DT
altman	135	7	flexible	flexible	JJ
altman	135	8	data	data	NN
altman	135	9	model	model	NN
altman	135	10	,	,	,
altman	135	11	consider	consider	VB
altman	135	12	a	a	DT
altman	135	13	mutable	mutable	JJ
altman	135	14	data	data	NN
altman	135	15	store	store	NN
altman	135	16	,	,	,
altman	135	17	such	such	JJ
altman	135	18	as	as	IN
altman	135	19	a	a	DT
altman	135	20	relational	relational	JJ
altman	135	21	database	database	NN
altman	135	22	.	.	.
altman	136	1	Before	before	IN
altman	136	2	you	-PRON-	PRP
altman	136	3	put	put	VBP
altman	136	4	any	any	DT
altman	136	5	data	datum	NNS
altman	136	6	into	into	IN
altman	136	7	the	the	DT
altman	136	8	database	database	NN
altman	136	9	,	,	,
altman	136	10	you	-PRON-	PRP
altman	136	11	would	would	MD
altman	136	12	first	first	RB
altman	136	13	need	need	VB
altman	136	14	to	to	TO
altman	136	15	design	design	VB
altman	136	16	a	a	DT
altman	136	17	system	system	NN
altman	136	18	of	of	IN
altman	136	19	tables	table	NNS
altman	136	20	with	with	IN
altman	136	21	set	set	NN
altman	136	22	fields	field	NNS
altman	136	23	and	and	CC
altman	136	24	datatypes	datatype	NNS
altman	136	25	,	,	,
altman	136	26	and	and	CC
altman	136	27	the	the	DT
altman	136	28	relationships	relationship	NNS
altman	136	29	between	between	IN
altman	136	30	those	those	DT
altman	136	31	tables	table	NNS
altman	136	32	.	.	.
altman	137	1	This	this	DT
altman	137	2	can	can	MD
altman	137	3	feel	feel	VB
altman	137	4	like	like	IN
altman	137	5	putting	put	VBG
altman	137	6	the	the	DT
altman	137	7	cart	cart	NN
altman	137	8	before	before	IN
altman	137	9	the	the	DT
altman	137	10	horse	horse	NN
altman	137	11	,	,	,
altman	137	12	especially	especially	RB
altman	137	13	if	if	IN
altman	137	14	you	-PRON-	PRP
altman	137	15	are	be	VBP
altman	137	16	starting	start	VBG
altman	137	17	with	with	IN
altman	137	18	a	a	DT
altman	137	19	dataset	dataset	NN
altman	137	20	with	with	IN
altman	137	21	which	which	WDT
altman	137	22	you	-PRON-	PRP
altman	137	23	are	be	VBP
altman	137	24	not	not	RB
altman	137	25	yet	yet	RB
altman	137	26	intimately	intimately	RB
altman	137	27	familiar	familiar	JJ
altman	137	28	,	,	,
altman	137	29	and	and	CC
altman	137	30	you	-PRON-	PRP
altman	137	31	want	want	VBP
altman	137	32	the	the	DT
altman	137	33	ability	ability	NN
altman	137	34	to	to	TO
altman	137	35	experiment	experiment	VB
altman	137	36	with	with	IN
altman	137	37	different	different	JJ
altman	137	38	algorithms	algorithm	NNS
altman	137	39	,	,	,
altman	137	40	all	all	DT
altman	137	41	of	of	IN
altman	137	42	which	which	WDT
altman	137	43	might	may	MD
altman	137	44	require	require	VB
altman	137	45	slightly	slightly	RB
altman	137	46	different	different	JJ
altman	137	47	transformations	transformation	NNS
altman	137	48	on	on	IN
altman	137	49	the	the	DT
altman	137	50	original	original	JJ
altman	137	51	dataset	dataset	NN
altman	137	52	.	.	.
altman	138	1	Revisiting	revisit	VBG
altman	138	2	the	the	DT
altman	138	3	example	example	NN
altman	138	4	in	in	IN
altman	138	5	the	the	DT
altman	138	6	previous	previous	JJ
altman	138	7	paragraph	paragraph	NN
altman	138	8	,	,	,
altman	138	9	you	-PRON-	PRP
altman	138	10	might	may	MD
altman	138	11	initially	initially	RB
altman	138	12	have	have	VB
altman	138	13	defined	define	VBN
altman	138	14	your	-PRON-	PRP$
altman	138	15	distance	distance	NN
altman	138	16	datatype	datatype	NN
altman	138	17	as	as	IN
altman	138	18	an	an	DT
altman	138	19	integer	integer	NN
altman	138	20	(	(	-LRB-
altman	138	21	when	when	WRB
altman	138	22	you	-PRON-	PRP
altman	138	23	were	be	VBD
altman	138	24	rounding	round	VBG
altman	138	25	to	to	IN
altman	138	26	the	the	DT
altman	138	27	nearest	near	JJS
altman	138	28	whole	whole	JJ
altman	138	29	number	number	NN
altman	138	30	)	)	-RRB-
altman	138	31	,	,	,
altman	138	32	and	and	CC
altman	138	33	would	would	MD
altman	138	34	later	later	RB
altman	138	35	have	have	VB
altman	138	36	to	to	TO
altman	138	37	change	change	VB
altman	138	38	it	-PRON-	PRP
altman	138	39	to	to	IN
altman	138	40	a	a	DT
altman	138	41	floating	floating	JJ
altman	138	42	point	point	NN
altman	138	43	number	number	NN
altman	138	44	(	(	-LRB-
altman	138	45	when	when	WRB
altman	138	46	you	-PRON-	PRP
altman	138	47	were	be	VBD
altman	138	48	rounding	round	VBG
altman	138	49	to	to	IN
altman	138	50	the	the	DT
altman	138	51	nearest	near	JJS
altman	138	52	tenth	tenth	NN
altman	138	53	)	)	-RRB-
altman	138	54	.	.	.
altman	139	1	Making	make	VBG
altman	139	2	this	this	DT
altman	139	3	change	change	NN
altman	139	4	would	would	MD
altman	139	5	mean	mean	VB
altman	139	6	altering	alter	VBG
altman	139	7	the	the	DT
altman	139	8	database	database	NN
altman	139	9	schema	schema	NN
altman	139	10	and	and	CC
altman	139	11	migrating	migrate	VBG
altman	139	12	all	all	DT
altman	139	13	of	of	IN
altman	139	14	the	the	DT
altman	139	15	existing	exist	VBG
altman	139	16	data	datum	NNS
altman	139	17	to	to	IN
altman	139	18	the	the	DT
altman	139	19	new	new	JJ
altman	139	20	type	type	NN
altman	139	21	,	,	,
altman	139	22	which	which	WDT
altman	139	23	is	be	VBZ
altman	139	24	a	a	DT
altman	139	25	nontrivial	nontrivial	JJ
altman	139	26	task	task	NN
altman	139	27	—	—	:
altman	139	28	especially	especially	RB
altman	139	29	if	if	IN
altman	139	30	you	-PRON-	PRP
altman	139	31	later	later	RB
altman	139	32	decide	decide	VBP
altman	139	33	to	to	TO
altman	139	34	revert	revert	VB
altman	139	35	back	back	RB
altman	139	36	to	to	IN
altman	139	37	the	the	DT
altman	139	38	original	original	JJ
altman	139	39	type	type	NN
altman	139	40	.	.	.
altman	140	1	By	by	IN
altman	140	2	contrast	contrast	NN
altman	140	3	,	,	,
altman	140	4	if	if	IN
altman	140	5	you	-PRON-	PRP
altman	140	6	were	be	VBD
altman	140	7	working	work	VBG
altman	140	8	with	with	IN
altman	140	9	immutable	immutable	JJ
altman	140	10	CSV	csv	NN
altman	140	11	files	file	NNS
altman	140	12	,	,	,
altman	140	13	it	-PRON-	PRP
altman	140	14	would	would	MD
altman	140	15	be	be	VB
altman	140	16	much	much	RB
altman	140	17	easier	easy	JJR
altman	140	18	to	to	TO
altman	140	19	write	write	VB
altman	140	20	out	out	RP
altman	140	21	two	two	CD
altman	140	22	files	file	NNS
altman	140	23	,	,	,
altman	140	24	one	one	CD
altman	140	25	with	with	IN
altman	140	26	each	each	DT
altman	140	27	data	data	NN
altman	140	28	type	type	NN
altman	140	29	,	,	,
altman	140	30	and	and	CC
altman	140	31	keep	keep	VB
altman	140	32	whichever	whichever	WDT
altman	140	33	one	one	NN
altman	140	34	ultimately	ultimately	RB
altman	140	35	proved	prove	VBD
altman	140	36	most	most	RBS
altman	140	37	effective	effective	JJ
altman	140	38	.	.	.
altman	141	1	Throughout	throughout	IN
altman	141	2	your	-PRON-	PRP$
altman	141	3	ML	ML	NNP
altman	141	4	process	process	NN
altman	141	5	,	,	,
altman	141	6	you	-PRON-	PRP
altman	141	7	can	can	MD
altman	141	8	create	create	VB
altman	141	9	several	several	JJ
altman	141	10	incremental	incremental	JJ
altman	141	11	datasets	dataset	NNS
altman	141	12	that	that	WDT
altman	141	13	are	be	VBP
altman	141	14	essentially	essentially	RB
altman	141	15	read	read	VBN
altman	141	16	-	-	:
altman	141	17	only	only	RB
altman	141	18	.	.	.
altman	142	1	There	there	EX
altman	142	2	’s	’	VBZ
altman	142	3	no	no	DT
altman	142	4	one	one	CD
altman	142	5	correct	correct	JJ
altman	142	6	data	data	NN
altman	142	7	storage	storage	NN
altman	142	8	format	format	NN
altman	142	9	,	,	,
altman	142	10	but	but	CC
altman	142	11	ideally	ideally	RB
altman	142	12	you	-PRON-	PRP
altman	142	13	would	would	MD
altman	142	14	use	use	VB
altman	142	15	something	something	NN
altman	142	16	simple	simple	JJ
altman	142	17	and	and	CC
altman	142	18	space	space	NN
altman	142	19	-	-	HYPH
altman	142	20	efficient	efficient	JJ
altman	142	21	with	with	IN
altman	142	22	the	the	DT
altman	142	23	capacity	capacity	NN
altman	142	24	for	for	IN
altman	142	25	interoperability	interoperability	NN
altman	142	26	with	with	IN
altman	142	27	different	different	JJ
altman	142	28	tools	tool	NNS
altman	142	29	,	,	,
altman	142	30	such	such	JJ
altman	142	31	as	as	IN
altman	142	32	flat	flat	JJ
altman	142	33	files	file	NNS
altman	142	34	(	(	-LRB-
altman	142	35	plain	plain	JJ
altman	142	36	text	text	NN
altman	142	37	files	file	NNS
altman	142	38	without	without	IN
altman	142	39	extraneous	extraneous	JJ
altman	142	40	markup	markup	NN
altman	142	41	,	,	,
altman	142	42	such	such	JJ
altman	142	43	as	as	IN
altman	142	44	TXT	TXT	NNP
altman	142	45	,	,	,
altman	142	46	CSV	CSV	NNP
altman	142	47	,	,	,
altman	142	48	or	or	CC
altman	142	49	Parquet	Parquet	NNP
altman	142	50	)	)	-RRB-
altman	142	51	.	.	.
altman	143	1	Even	even	RB
altman	143	2	if	if	IN
altman	143	3	your	-PRON-	PRP$
altman	143	4	data	data	NN
altman	143	5	is	be	VBZ
altman	143	6	ultimately	ultimately	RB
altman	143	7	destined	destine	VBN
altman	143	8	for	for	IN
altman	143	9	a	a	DT
altman	143	10	different	different	JJ
altman	143	11	kind	kind	NN
altman	143	12	of	of	IN
altman	143	13	datastore	datastore	NN
altman	143	14	,	,	,
altman	143	15	such	such	JJ
altman	143	16	as	as	IN
altman	143	17	a	a	DT
altman	143	18	relational	relational	JJ
altman	143	19	database	database	NN
altman	143	20	or	or	CC
altman	143	21	triplestore	triplestore	NN
altman	143	22	,	,	,
altman	143	23	consider	consider	VB
altman	143	24	using	use	VBG
altman	143	25	simple	simple	JJ
altman	143	26	,	,	,
altman	143	27	immutable	immutable	JJ
altman	143	28	storage	storage	NN
altman	143	29	as	as	IN
altman	143	30	an	an	DT
altman	143	31	intermediary	intermediary	NN
altman	143	32	to	to	TO
altman	143	33	facilitate	facilitate	VB
altman	143	34	iteration	iteration	NN
altman	143	35	and	and	CC
altman	143	36	experimentation	experimentation	NN
altman	143	37	.	.	.
altman	144	1	If	if	IN
altman	144	2	you	-PRON-	PRP
altman	144	3	’re	be	VBP
altman	144	4	concerned	concerned	JJ
altman	144	5	about	about	IN
altman	144	6	overwhelming	overwhelm	VBG
altman	144	7	your	-PRON-	PRP$
altman	144	8	local	local	JJ
altman	144	9	drive	drive	NN
altman	144	10	,	,	,
altman	144	11	cloud	cloud	NN
altman	144	12	storage	storage	NN
altman	144	13	is	be	VBZ
altman	144	14	a	a	DT
altman	144	15	good	good	JJ
altman	144	16	option	option	NN
altman	144	17	,	,	,
altman	144	18	especially	especially	RB
altman	144	19	if	if	IN
altman	144	20	you	-PRON-	PRP
altman	144	21	can	can	MD
altman	144	22	read	read	VB
altman	144	23	and	and	CC
altman	144	24	write	write	VB
altman	144	25	directly	directly	RB
altman	144	26	from	from	IN
altman	144	27	your	-PRON-	PRP$
altman	144	28	programs	program	NNS
altman	144	29	or	or	CC
altman	144	30	software	software	NN
altman	144	31	services	service	NNS
altman	144	32	.	.	.
altman	145	1	One	one	CD
altman	145	2	final	final	JJ
altman	145	3	benefit	benefit	NN
altman	145	4	of	of	IN
altman	145	5	immutable	immutable	JJ
altman	145	6	storage	storage	NN
altman	145	7	relates	relate	VBZ
altman	145	8	to	to	TO
altman	145	9	scale	scale	VB
altman	145	10	.	.	.
altman	146	1	Batch	batch	NN
altman	146	2	processing	processing	NN
altman	146	3	workflows	workflow	NNS
altman	146	4	with	with	IN
altman	146	5	immutable	immutable	JJ
altman	146	6	data	datum	NNS
altman	146	7	are	be	VBP
altman	146	8	also	also	RB
altman	146	9	design	design	NN
altman	146	10	principles	principle	NNS
altman	146	11	for	for	IN
altman	146	12	distributed	distribute	VBN
altman	146	13	data	datum	NNS
altman	146	14	processing	processing	NN
altman	146	15	frameworks	framework	NNS
altman	146	16	,	,	,
altman	146	17	such	such	JJ
altman	146	18	as	as	IN
altman	146	19	MapReduce	MapReduce	NNP
altman	146	20	and	and	CC
altman	146	21	Spark	Spark	NNP
altman	146	22	.	.	.
altman	147	1	Therefore	therefore	RB
altman	147	2	,	,	,
altman	147	3	if	if	IN
altman	147	4	you	-PRON-	PRP
altman	147	5	need	need	VBP
altman	147	6	to	to	TO
altman	147	7	scale	scale	VB
altman	147	8	your	-PRON-	PRP$
altman	147	9	ML	ML	NNP
altman	147	10	project	project	NN
altman	147	11	using	use	VBG
altman	147	12	distributed	distribute	VBN
altman	147	13	processing	processing	NN
altman	147	14	,	,	,
altman	147	15	the	the	DT
altman	147	16	integration	integration	NN
altman	147	17	will	will	MD
altman	147	18	be	be	VB
altman	147	19	more	more	RBR
altman	147	20	seamless	seamless	NN
altman	147	21	(	(	-LRB-
altman	147	22	for	for	IN
altman	147	23	more	more	JJR
altman	147	24	,	,	,
altman	147	25	see	see	VB
altman	147	26	the	the	DT
altman	147	27	section	section	NN
altman	147	28	on	on	IN
altman	147	29	scaling	scale	VBG
altman	147	30	up	up	RP
altman	147	31	)	)	-RRB-
altman	147	32	.	.	.
altman	148	1	Comment	comment	NN
altman	148	2	by	by	IN
altman	148	3	Daniel	Daniel	NNP
altman	148	4	Johnson	Johnson	NNP
altman	148	5	:	:	:
altman	148	6	Am	be	VBP
altman	148	7	I	-PRON-	PRP
altman	148	8	parsing	parse	VBG
altman	148	9	this	this	DT
altman	148	10	right	right	NN
altman	148	11	?	?	.
altman	149	1	Organizing	organize	VBG
altman	149	2	Immutable	Immutable	NNP
altman	149	3	Data	datum	NNS
altman	149	4	One	one	CD
altman	149	5	of	of	IN
altman	149	6	the	the	DT
altman	149	7	challenges	challenge	NNS
altman	149	8	in	in	IN
altman	149	9	working	work	VBG
altman	149	10	with	with	IN
altman	149	11	immutable	immutable	JJ
altman	149	12	data	data	NN
altman	149	13	stores	store	NNS
altman	149	14	is	be	VBZ
altman	149	15	keeping	keep	VBG
altman	149	16	it	-PRON-	PRP
altman	149	17	all	all	DT
altman	149	18	organized	organize	VBN
altman	149	19	,	,	,
altman	149	20	especially	especially	RB
altman	149	21	with	with	IN
altman	149	22	multiple	multiple	JJ
altman	149	23	users	user	NNS
altman	149	24	.	.	.
altman	150	1	A	a	DT
altman	150	2	little	little	JJ
altman	150	3	planning	planning	NN
altman	150	4	can	can	MD
altman	150	5	save	save	VB
altman	150	6	you	-PRON-	PRP
altman	150	7	from	from	IN
altman	150	8	losing	lose	VBG
altman	150	9	track	track	NN
altman	150	10	of	of	IN
altman	150	11	your	-PRON-	PRP$
altman	150	12	experiments	experiment	NNS
altman	150	13	and	and	CC
altman	150	14	results	result	NNS
altman	150	15	.	.	.
altman	151	1	A	a	DT
altman	151	2	well	well	RB
altman	151	3	-	-	HYPH
altman	151	4	ordered	order	VBN
altman	151	5	directory	directory	NN
altman	151	6	structure	structure	NN
altman	151	7	,	,	,
altman	151	8	informative	informative	JJ
altman	151	9	and	and	CC
altman	151	10	consistent	consistent	JJ
altman	151	11	file	file	NN
altman	151	12	names	name	NNS
altman	151	13	,	,	,
altman	151	14	liberal	liberal	JJ
altman	151	15	use	use	NN
altman	151	16	of	of	IN
altman	151	17	timestamps	timestamp	NNS
altman	151	18	,	,	,
altman	151	19	and	and	CC
altman	151	20	disciplined	discipline	VBD
altman	151	21	note	note	NN
altman	151	22	-	-	HYPH
altman	151	23	taking	taking	NN
altman	151	24	are	be	VBP
altman	151	25	simple	simple	JJ
altman	151	26	but	but	CC
altman	151	27	effective	effective	JJ
altman	151	28	strategies	strategy	NNS
altman	151	29	.	.	.
altman	152	1	For	for	IN
altman	152	2	example	example	NN
altman	152	3	,	,	,
altman	152	4	say	say	VBP
altman	152	5	you	-PRON-	PRP
altman	152	6	were	be	VBD
altman	152	7	acquiring	acquire	VBG
altman	152	8	MARCXML	marcxml	NN
altman	152	9	records	record	NNS
altman	152	10	from	from	IN
altman	152	11	the	the	DT
altman	152	12	University	University	NNP
altman	152	13	Library	Library	NNP
altman	152	14	’s	’s	,
altman	152	15	API	api	NN
altman	152	16	feed	feed	NN
altman	152	17	,	,	,
altman	152	18	parsing	parse	VBG
altman	152	19	out	out	RP
altman	152	20	subject	subject	JJ
altman	152	21	terms	term	NNS
altman	152	22	,	,	,
altman	152	23	and	and	CC
altman	152	24	building	build	VBG
altman	152	25	a	a	DT
altman	152	26	clustering	cluster	VBG
altman	152	27	algorithm	algorithm	NN
altman	152	28	around	around	IN
altman	152	29	these	these	DT
altman	152	30	terms	term	NNS
altman	152	31	.	.	.
altman	153	1	Let	let	VB
altman	153	2	us	-PRON-	PRP
altman	153	3	explore	explore	VB
altman	153	4	one	one	CD
altman	153	5	possible	possible	JJ
altman	153	6	way	way	NN
altman	153	7	that	that	WDT
altman	153	8	you	-PRON-	PRP
altman	153	9	could	could	MD
altman	153	10	organize	organize	VB
altman	153	11	your	-PRON-	PRP$
altman	153	12	data	datum	NNS
altman	153	13	outputs	output	NNS
altman	153	14	through	through	IN
altman	153	15	each	each	DT
altman	153	16	step	step	NN
altman	153	17	of	of	IN
altman	153	18	the	the	DT
altman	153	19	machine	machine	NN
altman	153	20	learning	learning	NN
altman	153	21	pipeline	pipeline	NN
altman	153	22	.	.	.
altman	154	1	To	to	TO
altman	154	2	enforce	enforce	VB
altman	154	3	a	a	DT
altman	154	4	naming	name	VBG
altman	154	5	convention	convention	NN
altman	154	6	,	,	,
altman	154	7	create	create	VB
altman	154	8	a	a	DT
altman	154	9	helper	helper	NN
altman	154	10	method	method	NN
altman	154	11	that	that	WDT
altman	154	12	generates	generate	VBZ
altman	154	13	the	the	DT
altman	154	14	output	output	NN
altman	154	15	path	path	NN
altman	154	16	for	for	IN
altman	154	17	each	each	DT
altman	154	18	run	run	NN
altman	154	19	of	of	IN
altman	154	20	a	a	DT
altman	154	21	particular	particular	JJ
altman	154	22	data	data	NN
altman	154	23	process	process	NN
altman	154	24	.	.	.
altman	155	1	This	this	DT
altman	155	2	output	output	NN
altman	155	3	path	path	NN
altman	155	4	includes	include	VBZ
altman	155	5	the	the	DT
altman	155	6	date	date	NN
altman	155	7	and	and	CC
altman	155	8	timestamp	timestamp	NN
altman	155	9	of	of	IN
altman	155	10	the	the	DT
altman	155	11	run	run	NN
altman	155	12	—	—	:
altman	155	13	that	that	DT
altman	155	14	way	way	NN
altman	155	15	you	-PRON-	PRP
altman	155	16	wo	will	MD
altman	155	17	n’t	not	RB
altman	155	18	have	have	VB
altman	155	19	to	to	TO
altman	155	20	think	think	VB
altman	155	21	about	about	IN
altman	155	22	naming	name	VBG
altman	155	23	each	each	DT
altman	155	24	individual	individual	JJ
altman	155	25	file	file	NN
altman	155	26	,	,	,
altman	155	27	and	and	CC
altman	155	28	can	can	MD
altman	155	29	avoid	avoid	VB
altman	155	30	the	the	DT
altman	155	31	phenomenon	phenomenon	NN
altman	155	32	of	of	IN
altman	155	33	a	a	DT
altman	155	34	mess	mess	NN
altman	155	35	of	of	IN
altman	155	36	files	file	NNS
altman	155	37	called	call	VBD
altman	155	38	my_clean_data.csv	my_clean_data.csv	NNP
altman	155	39	,	,	,
altman	155	40	my_cleaner_data.csv	my_cleaner_data.csv	UH
altman	155	41	,	,	,
altman	155	42	my_final_cleanest_data.csv	my_final_cleanest_data.csv	CD
altman	155	43	,	,	,
altman	155	44	etc	etc	FW
altman	155	45	.	.	.
altman	156	1	Your	-PRON-	PRP$
altman	156	2	file	file	NN
altman	156	3	path	path	NN
altman	156	4	for	for	IN
altman	156	5	the	the	DT
altman	156	6	acquired	acquire	VBN
altman	156	7	data	datum	NNS
altman	156	8	might	may	MD
altman	156	9	be	be	VB
altman	156	10	in	in	IN
altman	156	11	the	the	DT
altman	156	12	format	format	NN
altman	156	13	:	:	:
altman	156	14	myProject	myProject	NNP
altman	156	15	/	/	SYM
altman	156	16	acquisitions	acquisition	NNS
altman	156	17	/	/	SYM
altman	156	18	marc_YYYYMMDD_HHMMSS.xml	marc_yyyymmdd_hhmmss.xml	NN
altman	156	19	In	in	IN
altman	156	20	this	this	DT
altman	156	21	case	case	NN
altman	156	22	,	,	,
altman	156	23	YYMMDD	YYMMDD	NNP
altman	156	24	represents	represent	VBZ
altman	156	25	the	the	DT
altman	156	26	date	date	NN
altman	156	27	and	and	CC
altman	156	28	HHMMSS	hhmmss	NN
altman	156	29	represents	represent	VBZ
altman	156	30	the	the	DT
altman	156	31	timestamp	timestamp	NN
altman	156	32	.	.	.
altman	157	1	Your	-PRON-	PRP$
altman	157	2	file	file	NN
altman	157	3	path	path	NN
altman	157	4	for	for	IN
altman	157	5	prepared	prepared	JJ
altman	157	6	and	and	CC
altman	157	7	cleaned	clean	VBN
altman	157	8	data	datum	NNS
altman	157	9	might	may	MD
altman	157	10	be	be	VB
altman	157	11	:	:	:
altman	157	12	myProject	myProject	NNP
altman	157	13	/	/	SYM
altman	157	14	clean_datasets	clean_datasets	NNPS
altman	157	15	/	/	SYM
altman	157	16	subjects_YYYYMMDD_HHMMSS.csv	subjects_YYYYMMDD_HHMMSS.csv	NNP
altman	157	17	Finally	finally	RB
altman	157	18	,	,	,
altman	157	19	each	each	DT
altman	157	20	clustering	cluster	VBG
altman	157	21	model	model	NN
altman	157	22	you	-PRON-	PRP
altman	157	23	build	build	VBP
altman	157	24	could	could	MD
altman	157	25	be	be	VB
altman	157	26	saved	save	VBN
altman	157	27	using	use	VBG
altman	157	28	the	the	DT
altman	157	29	file	file	NN
altman	157	30	path	path	NN
altman	157	31	pattern	pattern	NN
altman	157	32	:	:	:
altman	157	33	myProject	myproject	NN
altman	157	34	/	/	SYM
altman	157	35	models	model	NNS
altman	157	36	/	/	SYM
altman	157	37	cluster_YYYYMMDD_HHMMSS	cluster_YYYYMMDD_HHMMSS	NNP
altman	157	38	Following	follow	VBG
altman	157	39	this	this	DT
altman	157	40	general	general	JJ
altman	157	41	pattern	pattern	NN
altman	157	42	,	,	,
altman	157	43	you	-PRON-	PRP
altman	157	44	can	can	MD
altman	157	45	organize	organize	VB
altman	157	46	all	all	DT
altman	157	47	of	of	IN
altman	157	48	the	the	DT
altman	157	49	outputs	output	NNS
altman	157	50	for	for	IN
altman	157	51	your	-PRON-	PRP$
altman	157	52	entire	entire	JJ
altman	157	53	project	project	NN
altman	157	54	.	.	.
altman	158	1	Using	use	VBG
altman	158	2	date	date	NN
altman	158	3	and	and	CC
altman	158	4	timestamps	timestamp	NNS
altman	158	5	in	in	IN
altman	158	6	the	the	DT
altman	158	7	file	file	NN
altman	158	8	name	name	NN
altman	158	9	also	also	RB
altman	158	10	enables	enable	VBZ
altman	158	11	easy	easy	RB
altman	158	12	sorting	sorting	NN
altman	158	13	and	and	CC
altman	158	14	retrieval	retrieval	NN
altman	158	15	of	of	IN
altman	158	16	the	the	DT
altman	158	17	most	most	RBS
altman	158	18	recent	recent	JJ
altman	158	19	output	output	NN
altman	158	20	.	.	.
altman	159	1	For	for	IN
altman	159	2	each	each	DT
altman	159	3	data	data	NN
altman	159	4	output	output	NN
altman	159	5	,	,	,
altman	159	6	you	-PRON-	PRP
altman	159	7	will	will	MD
altman	159	8	want	want	VB
altman	159	9	to	to	TO
altman	159	10	maintain	maintain	VB
altman	159	11	a	a	DT
altman	159	12	record	record	NN
altman	159	13	of	of	IN
altman	159	14	the	the	DT
altman	159	15	exact	exact	JJ
altman	159	16	input	input	NN
altman	159	17	,	,	,
altman	159	18	any	any	DT
altman	159	19	special	special	JJ
altman	159	20	attributes	attribute	NNS
altman	159	21	of	of	IN
altman	159	22	the	the	DT
altman	159	23	process	process	NN
altman	159	24	(	(	-LRB-
altman	159	25	e.g.	e.g.	RB
altman	160	1	“	"	``
altman	160	2	this	this	DT
altman	160	3	time	time	NN
altman	160	4	I	-PRON-	PRP
altman	160	5	rounded	round	VBD
altman	160	6	decimals	decimal	NNS
altman	160	7	to	to	IN
altman	160	8	the	the	DT
altman	160	9	nearest	near	JJS
altman	160	10	hundredth	hundredth	JJ
altman	160	11	”	"	''
altman	160	12	)	)	-RRB-
altman	160	13	,	,	,
altman	160	14	and	and	CC
altman	160	15	metrics	metric	NNS
altman	160	16	that	that	WDT
altman	160	17	will	will	MD
altman	160	18	help	help	VB
altman	160	19	you	-PRON-	PRP
altman	160	20	determine	determine	VB
altman	160	21	success	success	NN
altman	160	22	or	or	CC
altman	160	23	failure	failure	NN
altman	160	24	of	of	IN
altman	160	25	the	the	DT
altman	160	26	process	process	NN
altman	160	27	.	.	.
altman	161	1	If	if	IN
altman	161	2	you	-PRON-	PRP
altman	161	3	can	can	MD
altman	161	4	generate	generate	VB
altman	161	5	this	this	DT
altman	161	6	information	information	NN
altman	161	7	automatically	automatically	RB
altman	161	8	for	for	IN
altman	161	9	each	each	DT
altman	161	10	process	process	NN
altman	161	11	,	,	,
altman	161	12	all	all	PDT
altman	161	13	the	the	DT
altman	161	14	better	well	JJR
altman	161	15	for	for	IN
altman	161	16	ensuring	ensure	VBG
altman	161	17	an	an	DT
altman	161	18	accurate	accurate	JJ
altman	161	19	record	record	NN
altman	161	20	.	.	.
altman	162	1	One	one	CD
altman	162	2	strategy	strategy	NN
altman	162	3	is	be	VBZ
altman	162	4	to	to	TO
altman	162	5	include	include	VB
altman	162	6	a	a	DT
altman	162	7	second	second	JJ
altman	162	8	helper	helper	NN
altman	162	9	method	method	NN
altman	162	10	in	in	IN
altman	162	11	your	-PRON-	PRP$
altman	162	12	program	program	NN
altman	162	13	that	that	WDT
altman	162	14	will	will	MD
altman	162	15	generate	generate	VB
altman	162	16	and	and	CC
altman	162	17	write	write	VB
altman	162	18	out	out	RP
altman	162	19	a	a	DT
altman	162	20	companion	companion	NN
altman	162	21	file	file	NN
altman	162	22	to	to	IN
altman	162	23	each	each	DT
altman	162	24	data	data	NN
altman	162	25	output	output	NN
altman	162	26	.	.	.
altman	163	1	The	the	DT
altman	163	2	companion	companion	NN
altman	163	3	file	file	NN
altman	163	4	contains	contain	VBZ
altman	163	5	information	information	NN
altman	163	6	that	that	WDT
altman	163	7	will	will	MD
altman	163	8	help	help	VB
altman	163	9	evaluate	evaluate	VB
altman	163	10	results	result	NNS
altman	163	11	,	,	,
altman	163	12	detect	detect	VB
altman	163	13	errors	error	NNS
altman	163	14	,	,	,
altman	163	15	perform	perform	VB
altman	163	16	optimizations	optimization	NNS
altman	163	17	,	,	,
altman	163	18	and	and	CC
altman	163	19	differentiate	differentiate	JJ
altman	163	20	between	between	IN
altman	163	21	any	any	DT
altman	163	22	two	two	CD
altman	163	23	data	datum	NNS
altman	163	24	outputs	output	NNS
altman	163	25	.	.	.
altman	164	1	In	in	IN
altman	164	2	the	the	DT
altman	164	3	example	example	NN
altman	164	4	project	project	NN
altman	164	5	,	,	,
altman	164	6	you	-PRON-	PRP
altman	164	7	could	could	MD
altman	164	8	accompany	accompany	VB
altman	164	9	the	the	DT
altman	164	10	acquisition	acquisition	NN
altman	164	11	output	output	NN
altman	164	12	with	with	IN
altman	164	13	a	a	DT
altman	164	14	text	text	NN
altman	164	15	file	file	NN
altman	164	16	detailing	detail	VBG
altman	164	17	the	the	DT
altman	164	18	exact	exact	JJ
altman	164	19	API	api	NN
altman	164	20	call	call	NN
altman	164	21	used	use	VBN
altman	164	22	to	to	TO
altman	164	23	fetch	fetch	VB
altman	164	24	the	the	DT
altman	164	25	data	datum	NNS
altman	164	26	,	,	,
altman	164	27	the	the	DT
altman	164	28	number	number	NN
altman	164	29	of	of	IN
altman	164	30	records	record	NNS
altman	164	31	acquired	acquire	VBN
altman	164	32	,	,	,
altman	164	33	and	and	CC
altman	164	34	the	the	DT
altman	164	35	runtime	runtime	NN
altman	164	36	for	for	IN
altman	164	37	the	the	DT
altman	164	38	process	process	NN
altman	164	39	.	.	.
altman	165	1	Keeping	keep	VBG
altman	165	2	companion	companion	NN
altman	165	3	files	file	NNS
altman	165	4	as	as	RB
altman	165	5	close	close	RB
altman	165	6	as	as	IN
altman	165	7	possible	possible	JJ
altman	165	8	to	to	IN
altman	165	9	their	-PRON-	PRP$
altman	165	10	outputs	output	NNS
altman	165	11	helps	help	VBZ
altman	165	12	prevent	prevent	VB
altman	165	13	accidental	accidental	JJ
altman	165	14	separation	separation	NN
altman	165	15	,	,	,
altman	165	16	so	so	RB
altman	165	17	save	save	VB
altman	165	18	it	-PRON-	PRP
altman	165	19	to	to	IN
altman	165	20	:	:	:
altman	165	21	myProject	myProject	NNP
altman	165	22	/	/	SYM
altman	165	23	acquisition	acquisition	NN
altman	165	24	/	/	SYM
altman	165	25	marc_YYYYMMDD_HHMMSS.txt	marc_yyyymmdd_hhmmss.txt	NN
altman	165	26	In	in	IN
altman	165	27	this	this	DT
altman	165	28	case	case	NN
altman	165	29	,	,	,
altman	165	30	the	the	DT
altman	165	31	date	date	NN
altman	165	32	and	and	CC
altman	165	33	timestamp	timestamp	NN
altman	165	34	should	should	MD
altman	165	35	exactly	exactly	RB
altman	165	36	match	match	VB
altman	165	37	that	that	DT
altman	165	38	of	of	IN
altman	165	39	its	-PRON-	PRP$
altman	165	40	companion	companion	NN
altman	165	41	xml	xml	NN
altman	165	42	file	file	NN
altman	165	43	.	.	.
altman	166	1	When	when	WRB
altman	166	2	running	running	NN
altman	166	3	processes	process	NNS
altman	166	4	that	that	WDT
altman	166	5	test	test	NN
altman	166	6	and	and	CC
altman	166	7	train	train	NN
altman	166	8	models	model	NNS
altman	166	9	,	,	,
altman	166	10	you	-PRON-	PRP
altman	166	11	can	can	MD
altman	166	12	include	include	VB
altman	166	13	information	information	NN
altman	166	14	in	in	IN
altman	166	15	your	-PRON-	PRP$
altman	166	16	companion	companion	NN
altman	166	17	file	file	NN
altman	166	18	about	about	IN
altman	166	19	hyperparameters	hyperparameter	NNS
altman	166	20	and	and	CC
altman	166	21	whatever	whatever	WDT
altman	166	22	metrics	metric	NNS
altman	166	23	you	-PRON-	PRP
altman	166	24	are	be	VBP
altman	166	25	using	use	VBG
altman	166	26	to	to	TO
altman	166	27	evaluate	evaluate	VB
altman	166	28	the	the	DT
altman	166	29	quality	quality	NN
altman	166	30	of	of	IN
altman	166	31	the	the	DT
altman	166	32	model	model	NN
altman	166	33	.	.	.
altman	167	1	In	in	IN
altman	167	2	our	-PRON-	PRP$
altman	167	3	example	example	NN
altman	167	4	,	,	,
altman	167	5	the	the	DT
altman	167	6	companion	companion	NN
altman	167	7	file	file	NN
altman	167	8	to	to	IN
altman	167	9	each	each	DT
altman	167	10	cluster	cluster	NN
altman	167	11	model	model	NN
altman	167	12	may	may	MD
altman	167	13	contain	contain	VB
altman	167	14	the	the	DT
altman	167	15	file	file	NN
altman	167	16	path	path	NN
altman	167	17	for	for	IN
altman	167	18	the	the	DT
altman	167	19	cleaned	clean	VBN
altman	167	20	input	input	NN
altman	167	21	data	datum	NNS
altman	167	22	,	,	,
altman	167	23	the	the	DT
altman	167	24	number	number	NN
altman	167	25	of	of	IN
altman	167	26	clusters	cluster	NNS
altman	167	27	,	,	,
altman	167	28	and	and	CC
altman	167	29	a	a	DT
altman	167	30	measure	measure	NN
altman	167	31	of	of	IN
altman	167	32	cluster	cluster	NN
altman	167	33	variance	variance	NN
altman	167	34	.	.	.
altman	168	1	Algorithm	Algorithm	NNP
altman	168	2	Selection	Selection	NNP
altman	168	3	As	as	IN
altman	168	4	you	-PRON-	PRP
altman	168	5	begin	begin	VBP
altman	168	6	ingesting	ingest	VBG
altman	168	7	and	and	CC
altman	168	8	preparing	prepare	VBG
altman	168	9	data	datum	NNS
altman	168	10	,	,	,
altman	168	11	you	-PRON-	PRP
altman	168	12	'll	will	MD
altman	168	13	want	want	VB
altman	168	14	to	to	TO
altman	168	15	explore	explore	VB
altman	168	16	possible	possible	JJ
altman	168	17	machine	machine	NN
altman	168	18	learning	learn	VBG
altman	168	19	algorithms	algorithm	NNS
altman	168	20	to	to	TO
altman	168	21	perform	perform	VB
altman	168	22	on	on	IN
altman	168	23	your	-PRON-	PRP$
altman	168	24	dataset	dataset	NN
altman	168	25	.	.	.
altman	169	1	Choose	choose	VB
altman	169	2	an	an	DT
altman	169	3	algorithm	algorithm	NN
altman	169	4	that	that	WDT
altman	169	5	fits	fit	VBZ
altman	169	6	your	-PRON-	PRP$
altman	169	7	research	research	NN
altman	169	8	question	question	NN
altman	169	9	and	and	CC
altman	169	10	data	datum	NNS
altman	169	11	.	.	.
altman	170	1	If	if	IN
altman	170	2	you	-PRON-	PRP
altman	170	3	’re	be	VBZ
altman	170	4	not	not	RB
altman	170	5	sure	sure	JJ
altman	170	6	which	which	WDT
altman	170	7	algorithm	algorithm	VBP
altman	170	8	to	to	TO
altman	170	9	choose	choose	VB
altman	170	10	and	and	CC
altman	170	11	not	not	RB
altman	170	12	constrained	constrain	VBN
altman	170	13	by	by	IN
altman	170	14	time	time	NN
altman	170	15	,	,	,
altman	170	16	experiment	experiment	NN
altman	170	17	with	with	IN
altman	170	18	several	several	JJ
altman	170	19	different	different	JJ
altman	170	20	options	option	NNS
altman	170	21	and	and	CC
altman	170	22	see	see	VB
altman	170	23	which	which	WDT
altman	170	24	one	one	NN
altman	170	25	yields	yield	VBZ
altman	170	26	the	the	DT
altman	170	27	best	good	JJS
altman	170	28	results	result	NNS
altman	170	29	.	.	.
altman	171	1	Start	start	VB
altman	171	2	by	by	IN
altman	171	3	determining	determine	VBG
altman	171	4	what	what	WP
altman	171	5	general	general	JJ
altman	171	6	type	type	NN
altman	171	7	of	of	IN
altman	171	8	learning	learning	NN
altman	171	9	algorithm	algorithm	NN
altman	171	10	you	-PRON-	PRP
altman	171	11	need	need	VBP
altman	171	12	,	,	,
altman	171	13	and	and	CC
altman	171	14	proceed	proceed	VB
altman	171	15	from	from	IN
altman	171	16	there	there	RB
altman	171	17	to	to	IN
altman	171	18	research	research	NN
altman	171	19	and	and	CC
altman	171	20	select	select	VB
altman	171	21	one	one	NN
altman	171	22	that	that	WDT
altman	171	23	specifically	specifically	RB
altman	171	24	addresses	address	VBZ
altman	171	25	your	-PRON-	PRP$
altman	171	26	research	research	NN
altman	171	27	question	question	NN
altman	171	28	.	.	.
altman	172	1	In	in	IN
altman	172	2	supervised	supervised	JJ
altman	172	3	learning	learning	NN
altman	172	4	,	,	,
altman	172	5	you	-PRON-	PRP
altman	172	6	train	train	VBP
altman	172	7	a	a	DT
altman	172	8	model	model	NN
altman	172	9	to	to	TO
altman	172	10	predict	predict	VB
altman	172	11	an	an	DT
altman	172	12	output	output	NN
altman	172	13	condition	condition	NN
altman	172	14	based	base	VBN
altman	172	15	on	on	IN
altman	172	16	given	give	VBN
altman	172	17	input	input	NN
altman	172	18	conditions	condition	NNS
altman	172	19	;	;	:
altman	172	20	for	for	IN
altman	172	21	example	example	NN
altman	172	22	,	,	,
altman	172	23	predicting	predict	VBG
altman	172	24	whether	whether	IN
altman	172	25	or	or	CC
altman	172	26	not	not	RB
altman	172	27	a	a	DT
altman	172	28	patient	patient	NN
altman	172	29	has	have	VBZ
altman	172	30	some	some	DT
altman	172	31	disease	disease	NN
altman	172	32	based	base	VBN
altman	172	33	on	on	IN
altman	172	34	their	-PRON-	PRP$
altman	172	35	symptoms	symptom	NNS
altman	172	36	,	,	,
altman	172	37	or	or	CC
altman	172	38	the	the	DT
altman	172	39	topic	topic	NN
altman	172	40	of	of	IN
altman	172	41	a	a	DT
altman	172	42	news	news	NN
altman	172	43	article	article	NN
altman	172	44	based	base	VBN
altman	172	45	on	on	IN
altman	172	46	keywords	keyword	NNS
altman	172	47	in	in	IN
altman	172	48	the	the	DT
altman	172	49	text	text	NN
altman	172	50	.	.	.
altman	173	1	In	in	IN
altman	173	2	order	order	NN
altman	173	3	for	for	IN
altman	173	4	supervised	supervise	VBN
altman	173	5	learning	learn	VBG
altman	173	6	to	to	IN
altman	173	7	work	work	NN
altman	173	8	,	,	,
altman	173	9	you	-PRON-	PRP
altman	173	10	need	need	VBP
altman	173	11	labeled	label	VBN
altman	173	12	training	training	NN
altman	173	13	data	datum	NNS
altman	173	14	,	,	,
altman	173	15	meaning	mean	VBG
altman	173	16	data	datum	NNS
altman	173	17	in	in	IN
altman	173	18	which	which	WDT
altman	173	19	the	the	DT
altman	173	20	outcome	outcome	NN
altman	173	21	is	be	VBZ
altman	173	22	already	already	RB
altman	173	23	known	know	VBN
altman	173	24	.	.	.
altman	174	1	Examples	example	NNS
altman	174	2	include	include	VBP
altman	174	3	records	record	NNS
altman	174	4	of	of	IN
altman	174	5	symptoms	symptom	NNS
altman	174	6	in	in	IN
altman	174	7	patients	patient	NNS
altman	174	8	who	who	WP
altman	174	9	were	be	VBD
altman	174	10	known	know	VBN
altman	174	11	to	to	TO
altman	174	12	have	have	VB
altman	174	13	the	the	DT
altman	174	14	disease	disease	NN
altman	174	15	(	(	-LRB-
altman	174	16	or	or	CC
altman	174	17	not	not	RB
altman	174	18	)	)	-RRB-
altman	174	19	,	,	,
altman	174	20	or	or	CC
altman	174	21	news	news	NN
altman	174	22	articles	article	NNS
altman	174	23	that	that	WDT
altman	174	24	have	have	VBP
altman	174	25	already	already	RB
altman	174	26	been	be	VBN
altman	174	27	assigned	assign	VBN
altman	174	28	topics	topic	NNS
altman	174	29	.	.	.
altman	175	1	Classification	classification	NN
altman	175	2	and	and	CC
altman	175	3	regression	regression	NN
altman	175	4	are	be	VBP
altman	175	5	both	both	DT
altman	175	6	types	type	NNS
altman	175	7	of	of	IN
altman	175	8	supervised	supervise	VBN
altman	175	9	learning	learning	NN
altman	175	10	.	.	.
altman	176	1	In	in	IN
altman	176	2	a	a	DT
altman	176	3	classification	classification	NN
altman	176	4	problem	problem	NN
altman	176	5	,	,	,
altman	176	6	you	-PRON-	PRP
altman	176	7	are	be	VBP
altman	176	8	predicting	predict	VBG
altman	176	9	a	a	DT
altman	176	10	discrete	discrete	JJ
altman	176	11	number	number	NN
altman	176	12	of	of	IN
altman	176	13	possible	possible	JJ
altman	176	14	outcomes	outcome	NNS
altman	176	15	.	.	.
altman	177	1	For	for	IN
altman	177	2	example	example	NN
altman	177	3	,	,	,
altman	177	4	“	"	``
altman	177	5	based	base	VBN
altman	177	6	on	on	IN
altman	177	7	what	what	WP
altman	177	8	I	-PRON-	PRP
altman	177	9	know	know	VBP
altman	177	10	about	about	IN
altman	177	11	this	this	DT
altman	177	12	book	book	NN
altman	177	13	,	,	,
altman	177	14	will	will	MD
altman	177	15	it	-PRON-	PRP
altman	177	16	make	make	VB
altman	177	17	the	the	DT
altman	177	18	New	New	NNP
altman	177	19	York	York	NNP
altman	177	20	Times	Times	NNP
altman	177	21	Best	Best	NNP
altman	177	22	Seller	Seller	NNP
altman	177	23	list	list	NN
altman	177	24	?	?	.
altman	177	25	”	"	''
altman	177	26	is	be	VBZ
altman	177	27	a	a	DT
altman	177	28	classification	classification	NN
altman	177	29	problem	problem	NN
altman	177	30	because	because	IN
altman	177	31	there	there	EX
altman	177	32	are	be	VBP
altman	177	33	two	two	CD
altman	177	34	discrete	discrete	JJ
altman	177	35	outcomes	outcome	NNS
altman	177	36	:	:	:
altman	177	37	yes	yes	UH
altman	177	38	or	or	CC
altman	177	39	no	no	UH
altman	177	40	.	.	.
altman	178	1	Classification	classification	NN
altman	178	2	algorithms	algorithm	NNS
altman	178	3	include	include	VBP
altman	178	4	naive	naive	JJ
altman	178	5	Bayes	Bayes	NNP
altman	178	6	,	,	,
altman	178	7	decision	decision	NN
altman	178	8	trees	tree	NNS
altman	178	9	,	,	,
altman	178	10	and	and	CC
altman	178	11	k	k	NNP
altman	178	12	-	-	HYPH
altman	178	13	nearest	near	JJS
altman	178	14	neighbor	neighbor	NN
altman	178	15	.	.	.
altman	179	1	Regression	regression	NN
altman	179	2	problems	problem	NNS
altman	179	3	try	try	VBP
altman	179	4	to	to	TO
altman	179	5	predict	predict	VB
altman	179	6	an	an	DT
altman	179	7	outcome	outcome	NN
altman	179	8	from	from	IN
altman	179	9	a	a	DT
altman	179	10	continuum	continuum	NN
altman	179	11	of	of	IN
altman	179	12	possibilities	possibility	NNS
altman	179	13	,	,	,
altman	179	14	i.e.	i.e.	FW
altman	179	15	,	,	,
altman	179	16	“	"	``
altman	179	17	based	base	VBN
altman	179	18	on	on	IN
altman	179	19	what	what	WP
altman	179	20	I	-PRON-	PRP
altman	179	21	know	know	VBP
altman	179	22	about	about	IN
altman	179	23	this	this	DT
altman	179	24	book	book	NN
altman	179	25	,	,	,
altman	179	26	what	what	WP
altman	179	27	will	will	MD
altman	179	28	its	-PRON-	PRP$
altman	179	29	retail	retail	JJ
altman	179	30	price	price	NN
altman	179	31	be	be	VB
altman	179	32	?	?	.
altman	179	33	”	"	''
altman	179	34	Regression	regression	NN
altman	179	35	algorithms	algorithm	NNS
altman	179	36	include	include	VBP
altman	179	37	linear	linear	JJ
altman	179	38	regression	regression	NN
altman	179	39	and	and	CC
altman	179	40	regression	regression	NN
altman	179	41	trees	tree	NNS
altman	179	42	.	.	.
altman	180	1	In	in	IN
altman	180	2	unsupervised	unsupervised	JJ
altman	180	3	learning	learning	NN
altman	180	4	,	,	,
altman	180	5	the	the	DT
altman	180	6	ML	ML	NNP
altman	180	7	algorithm	algorithm	NN
altman	180	8	discovers	discover	VBZ
altman	180	9	a	a	DT
altman	180	10	new	new	JJ
altman	180	11	pattern	pattern	NN
altman	180	12	.	.	.
altman	181	1	The	the	DT
altman	181	2	training	training	NN
altman	181	3	data	datum	NNS
altman	181	4	is	be	VBZ
altman	181	5	unlabeled	unlabeled	JJ
altman	181	6	,	,	,
altman	181	7	meaning	mean	VBG
altman	181	8	there	there	EX
altman	181	9	is	be	VBZ
altman	181	10	no	no	DT
altman	181	11	indication	indication	NN
altman	181	12	of	of	IN
altman	181	13	how	how	WRB
altman	181	14	the	the	DT
altman	181	15	data	datum	NNS
altman	181	16	should	should	MD
altman	181	17	be	be	VB
altman	181	18	organized	organize	VBN
altman	181	19	at	at	IN
altman	181	20	the	the	DT
altman	181	21	outset	outset	NN
altman	181	22	.	.	.
altman	182	1	A	a	DT
altman	182	2	common	common	JJ
altman	182	3	example	example	NN
altman	182	4	is	be	VBZ
altman	182	5	clustering	cluster	VBG
altman	182	6	,	,	,
altman	182	7	in	in	IN
altman	182	8	which	which	WDT
altman	182	9	the	the	DT
altman	182	10	algorithm	algorithm	NNP
altman	182	11	groups	group	NNS
altman	182	12	items	item	NNS
altman	182	13	together	together	RB
altman	182	14	based	base	VBN
altman	182	15	on	on	IN
altman	182	16	features	feature	NNS
altman	182	17	it	-PRON-	PRP
altman	182	18	finds	find	VBZ
altman	182	19	mathematically	mathematically	RB
altman	182	20	significant	significant	JJ
altman	182	21	.	.	.
altman	183	1	Perhaps	perhaps	RB
altman	183	2	you	-PRON-	PRP
altman	183	3	have	have	VBP
altman	183	4	a	a	DT
altman	183	5	collection	collection	NN
altman	183	6	of	of	IN
altman	183	7	news	news	NN
altman	183	8	articles	article	NNS
altman	183	9	(	(	-LRB-
altman	183	10	with	with	IN
altman	183	11	no	no	DT
altman	183	12	existing	exist	VBG
altman	183	13	topic	topic	NN
altman	183	14	labels	label	NNS
altman	183	15	)	)	-RRB-
altman	183	16	,	,	,
altman	183	17	and	and	CC
altman	183	18	you	-PRON-	PRP
altman	183	19	want	want	VBP
altman	183	20	to	to	TO
altman	183	21	discover	discover	VB
altman	183	22	common	common	JJ
altman	183	23	themes	theme	NNS
altman	183	24	or	or	CC
altman	183	25	topics	topic	NNS
altman	183	26	that	that	WDT
altman	183	27	appear	appear	VBP
altman	183	28	throughout	throughout	IN
altman	183	29	the	the	DT
altman	183	30	collection	collection	NN
altman	183	31	.	.	.
altman	184	1	The	the	DT
altman	184	2	algorithm	algorithm	NN
altman	184	3	will	will	MD
altman	184	4	not	not	RB
altman	184	5	tell	tell	VB
altman	184	6	you	-PRON-	PRP
altman	184	7	what	what	WP
altman	184	8	the	the	DT
altman	184	9	themes	theme	NNS
altman	184	10	or	or	CC
altman	184	11	topics	topic	NNS
altman	184	12	are	be	VBP
altman	184	13	,	,	,
altman	184	14	but	but	CC
altman	184	15	will	will	MD
altman	184	16	show	show	VB
altman	184	17	which	which	WDT
altman	184	18	articles	article	VBZ
altman	184	19	group	group	NN
altman	184	20	together	together	RB
altman	184	21	.	.	.
altman	185	1	It	-PRON-	PRP
altman	185	2	is	be	VBZ
altman	185	3	then	then	RB
altman	185	4	up	up	IN
altman	185	5	to	to	IN
altman	185	6	the	the	DT
altman	185	7	researcher	researcher	NN
altman	185	8	to	to	TO
altman	185	9	work	work	VB
altman	185	10	out	out	RP
altman	185	11	the	the	DT
altman	185	12	common	common	JJ
altman	185	13	thread	thread	NN
altman	185	14	.	.	.
altman	186	1	In	in	IN
altman	186	2	addition	addition	NN
altman	186	3	to	to	IN
altman	186	4	serving	serve	VBG
altman	186	5	your	-PRON-	PRP$
altman	186	6	research	research	NN
altman	186	7	question	question	NN
altman	186	8	,	,	,
altman	186	9	your	-PRON-	PRP$
altman	186	10	algorithm	algorithm	NN
altman	186	11	should	should	MD
altman	186	12	also	also	RB
altman	186	13	be	be	VB
altman	186	14	a	a	DT
altman	186	15	good	good	JJ
altman	186	16	fit	fit	NN
altman	186	17	for	for	IN
altman	186	18	your	-PRON-	PRP$
altman	186	19	data	datum	NNS
altman	186	20	.	.	.
altman	187	1	Specific	specific	JJ
altman	187	2	considerations	consideration	NNS
altman	187	3	will	will	MD
altman	187	4	vary	vary	VB
altman	187	5	for	for	IN
altman	187	6	each	each	DT
altman	187	7	dataset	dataset	NN
altman	187	8	and	and	CC
altman	187	9	algorithm	algorithm	NN
altman	187	10	,	,	,
altman	187	11	so	so	RB
altman	187	12	make	make	VB
altman	187	13	sure	sure	JJ
altman	187	14	you	-PRON-	PRP
altman	187	15	know	know	VBP
altman	187	16	the	the	DT
altman	187	17	strengths	strength	NNS
altman	187	18	and	and	CC
altman	187	19	weaknesses	weakness	NNS
altman	187	20	of	of	IN
altman	187	21	your	-PRON-	PRP$
altman	187	22	algorithm	algorithm	NN
altman	187	23	and	and	CC
altman	187	24	how	how	WRB
altman	187	25	they	-PRON-	PRP
altman	187	26	relate	relate	VBP
altman	187	27	to	to	IN
altman	187	28	the	the	DT
altman	187	29	unique	unique	JJ
altman	187	30	qualities	quality	NNS
altman	187	31	of	of	IN
altman	187	32	your	-PRON-	PRP$
altman	187	33	dataset	dataset	NN
altman	187	34	.	.	.
altman	188	1	For	for	IN
altman	188	2	example	example	NN
altman	188	3	,	,	,
altman	188	4	algorithms	algorithm	NNS
altman	188	5	differ	differ	VBP
altman	188	6	in	in	IN
altman	188	7	their	-PRON-	PRP$
altman	188	8	abilities	ability	NNS
altman	188	9	to	to	TO
altman	188	10	handle	handle	VB
altman	188	11	datasets	dataset	NNS
altman	188	12	with	with	IN
altman	188	13	a	a	DT
altman	188	14	very	very	RB
altman	188	15	large	large	JJ
altman	188	16	number	number	NN
altman	188	17	of	of	IN
altman	188	18	features	feature	NNS
altman	188	19	,	,	,
altman	188	20	handle	handle	VB
altman	188	21	datasets	dataset	NNS
altman	188	22	with	with	IN
altman	188	23	high	high	JJ
altman	188	24	variance	variance	NN
altman	188	25	,	,	,
altman	188	26	efficiently	efficiently	RB
altman	188	27	process	process	VB
altman	188	28	very	very	RB
altman	188	29	large	large	JJ
altman	188	30	datasets	dataset	NNS
altman	188	31	,	,	,
altman	188	32	and	and	CC
altman	188	33	glean	glean	JJ
altman	188	34	meaningful	meaningful	JJ
altman	188	35	intelligence	intelligence	NN
altman	188	36	from	from	IN
altman	188	37	very	very	RB
altman	188	38	small	small	JJ
altman	188	39	datasets	dataset	NNS
altman	188	40	.	.	.
altman	189	1	Is	be	VBZ
altman	189	2	it	-PRON-	PRP
altman	189	3	important	important	JJ
altman	189	4	that	that	IN
altman	189	5	your	-PRON-	PRP$
altman	189	6	algorithm	algorithm	NN
altman	189	7	be	be	VB
altman	189	8	easy	easy	JJ
altman	189	9	to	to	TO
altman	189	10	explain	explain	VB
altman	189	11	?	?	.
altman	190	1	Some	some	DT
altman	190	2	algorithms	algorithm	NNS
altman	190	3	,	,	,
altman	190	4	such	such	JJ
altman	190	5	as	as	IN
altman	190	6	neural	neural	JJ
altman	190	7	nets	net	NNS
altman	190	8	,	,	,
altman	190	9	function	function	NN
altman	190	10	as	as	IN
altman	190	11	black	black	JJ
altman	190	12	boxes	box	NNS
altman	190	13	,	,	,
altman	190	14	and	and	CC
altman	190	15	it	-PRON-	PRP
altman	190	16	is	be	VBZ
altman	190	17	difficult	difficult	JJ
altman	190	18	to	to	TO
altman	190	19	decipher	decipher	VB
altman	190	20	how	how	WRB
altman	190	21	they	-PRON-	PRP
altman	190	22	arrive	arrive	VBP
altman	190	23	at	at	IN
altman	190	24	their	-PRON-	PRP$
altman	190	25	decisions	decision	NNS
altman	190	26	.	.	.
altman	191	1	Other	other	JJ
altman	191	2	algorithms	algorithm	NNS
altman	191	3	,	,	,
altman	191	4	such	such	JJ
altman	191	5	as	as	IN
altman	191	6	decision	decision	NN
altman	191	7	trees	tree	NNS
altman	191	8	,	,	,
altman	191	9	are	be	VBP
altman	191	10	easy	easy	JJ
altman	191	11	to	to	TO
altman	191	12	understand	understand	VB
altman	191	13	.	.	.
altman	192	1	Can	Can	MD
altman	192	2	you	-PRON-	PRP
altman	192	3	prepare	prepare	VB
altman	192	4	your	-PRON-	PRP$
altman	192	5	data	datum	NNS
altman	192	6	with	with	IN
altman	192	7	a	a	DT
altman	192	8	reasonable	reasonable	JJ
altman	192	9	amount	amount	NN
altman	192	10	of	of	IN
altman	192	11	pre	pre	NN
altman	192	12	-	-	NN
altman	192	13	processing	processing	NN
altman	192	14	?	?	.
altman	193	1	Can	Can	MD
altman	193	2	you	-PRON-	PRP
altman	193	3	find	find	VB
altman	193	4	examples	example	NNS
altman	193	5	of	of	IN
altman	193	6	success	success	NN
altman	193	7	(	(	-LRB-
altman	193	8	or	or	CC
altman	193	9	failure	failure	NN
altman	193	10	)	)	-RRB-
altman	193	11	from	from	IN
altman	193	12	people	people	NNS
altman	193	13	using	use	VBG
altman	193	14	similar	similar	JJ
altman	193	15	datasets	dataset	NNS
altman	193	16	with	with	IN
altman	193	17	the	the	DT
altman	193	18	same	same	JJ
altman	193	19	algorithm	algorithm	NN
altman	193	20	?	?	.
altman	194	1	Asking	ask	VBG
altman	194	2	these	these	DT
altman	194	3	sorts	sort	NNS
altman	194	4	of	of	IN
altman	194	5	questions	question	NNS
altman	194	6	will	will	MD
altman	194	7	help	help	VB
altman	194	8	you	-PRON-	PRP
altman	194	9	to	to	TO
altman	194	10	choose	choose	VB
altman	194	11	an	an	DT
altman	194	12	algorithm	algorithm	NN
altman	194	13	that	that	WDT
altman	194	14	works	work	VBZ
altman	194	15	well	well	RB
altman	194	16	for	for	IN
altman	194	17	your	-PRON-	PRP$
altman	194	18	data	datum	NNS
altman	194	19	,	,	,
altman	194	20	and	and	CC
altman	194	21	will	will	MD
altman	194	22	also	also	RB
altman	194	23	inform	inform	VB
altman	194	24	how	how	WRB
altman	194	25	you	-PRON-	PRP
altman	194	26	prepare	prepare	VBP
altman	194	27	your	-PRON-	PRP$
altman	194	28	data	datum	NNS
altman	194	29	for	for	IN
altman	194	30	optimal	optimal	JJ
altman	194	31	use	use	NN
altman	194	32	.	.	.
altman	195	1	Finally	finally	RB
altman	195	2	,	,	,
altman	195	3	consider	consider	VB
altman	195	4	whether	whether	IN
altman	195	5	or	or	CC
altman	195	6	not	not	RB
altman	195	7	you	-PRON-	PRP
altman	195	8	are	be	VBP
altman	195	9	constrained	constrain	VBN
altman	195	10	by	by	IN
altman	195	11	time	time	NN
altman	195	12	,	,	,
altman	195	13	hardware	hardware	NN
altman	195	14	,	,	,
altman	195	15	or	or	CC
altman	195	16	available	available	JJ
altman	195	17	toolsets	toolset	NNS
altman	195	18	.	.	.
altman	196	1	Different	different	JJ
altman	196	2	algorithms	algorithm	NNS
altman	196	3	require	require	VBP
altman	196	4	different	different	JJ
altman	196	5	amounts	amount	NNS
altman	196	6	of	of	IN
altman	196	7	time	time	NN
altman	196	8	and	and	CC
altman	196	9	memory	memory	NN
altman	196	10	to	to	TO
altman	196	11	train	train	VB
altman	196	12	and/or	and/or	CC
altman	196	13	execute	execute	VB
altman	196	14	.	.	.
altman	197	1	Different	different	JJ
altman	197	2	ML	ML	NNP
altman	197	3	tools	tool	NNS
altman	197	4	offer	offer	VBP
altman	197	5	implementations	implementation	NNS
altman	197	6	of	of	IN
altman	197	7	different	different	JJ
altman	197	8	algorithms	algorithm	NNS
altman	197	9	.	.	.
altman	198	1	Working	work	VBG
altman	198	2	with	with	IN
altman	198	3	machine	machine	NN
altman	198	4	learning	learning	NN
altman	198	5	algorithms	algorithm	NNS
altman	198	6	New	New	NNP
altman	198	7	technologies	technology	NNS
altman	198	8	and	and	CC
altman	198	9	software	software	NN
altman	198	10	advances	advance	NNS
altman	198	11	make	make	VBP
altman	198	12	machine	machine	NN
altman	198	13	learning	learn	VBG
altman	198	14	more	more	RBR
altman	198	15	accessible	accessible	JJ
altman	198	16	to	to	IN
altman	198	17	“	"	``
altman	198	18	lay	lay	VB
altman	198	19	”	"	''
altman	198	20	users	user	NNS
altman	198	21	,	,	,
altman	198	22	by	by	IN
altman	198	23	which	which	WDT
altman	198	24	I	-PRON-	PRP
altman	198	25	mean	mean	VBP
altman	198	26	those	those	DT
altman	198	27	of	of	IN
altman	198	28	us	-PRON-	PRP
altman	198	29	without	without	IN
altman	198	30	advanced	advanced	JJ
altman	198	31	degrees	degree	NNS
altman	198	32	in	in	IN
altman	198	33	mathematics	mathematic	NNS
altman	198	34	or	or	CC
altman	198	35	data	datum	NNS
altman	198	36	science	science	NN
altman	198	37	.	.	.
altman	199	1	Yet	yet	RB
altman	199	2	,	,	,
altman	199	3	the	the	DT
altman	199	4	algorithms	algorithm	NNS
altman	199	5	are	be	VBP
altman	199	6	complex	complex	JJ
altman	199	7	,	,	,
altman	199	8	and	and	CC
altman	199	9	you	-PRON-	PRP
altman	199	10	need	need	VBP
altman	199	11	at	at	IN
altman	199	12	least	least	JJS
altman	199	13	an	an	DT
altman	199	14	intuitive	intuitive	JJ
altman	199	15	understanding	understanding	NN
altman	199	16	of	of	IN
altman	199	17	how	how	WRB
altman	199	18	they	-PRON-	PRP
altman	199	19	work	work	VBP
altman	199	20	if	if	IN
altman	199	21	you	-PRON-	PRP
altman	199	22	hope	hope	VBP
altman	199	23	to	to	TO
altman	199	24	implement	implement	VB
altman	199	25	them	-PRON-	PRP
altman	199	26	correctly	correctly	RB
altman	199	27	.	.	.
altman	200	1	I	-PRON-	PRP
altman	200	2	use	use	VBP
altman	200	3	the	the	DT
altman	200	4	following	follow	VBG
altman	200	5	three	three	CD
altman	200	6	questions	question	NNS
altman	200	7	as	as	IN
altman	200	8	a	a	DT
altman	200	9	guide	guide	NN
altman	200	10	for	for	IN
altman	200	11	understanding	understand	VBG
altman	200	12	an	an	DT
altman	200	13	algorithm	algorithm	NN
altman	200	14	.	.	.
altman	201	1	Keep	keep	VB
altman	201	2	in	in	IN
altman	201	3	mind	mind	NN
altman	201	4	that	that	IN
altman	201	5	any	any	DT
altman	201	6	one	one	CD
altman	201	7	project	project	NN
altman	201	8	will	will	MD
altman	201	9	likely	likely	RB
altman	201	10	make	make	VB
altman	201	11	use	use	NN
altman	201	12	of	of	IN
altman	201	13	several	several	JJ
altman	201	14	complex	complex	JJ
altman	201	15	algorithms	algorithm	NNS
altman	201	16	along	along	IN
altman	201	17	the	the	DT
altman	201	18	way	way	NN
altman	201	19	.	.	.
altman	202	1	These	these	DT
altman	202	2	questions	question	NNS
altman	202	3	help	help	VBP
altman	202	4	ensure	ensure	VB
altman	202	5	that	that	IN
altman	202	6	I	-PRON-	PRP
altman	202	7	have	have	VBP
altman	202	8	the	the	DT
altman	202	9	information	information	NN
altman	202	10	I	-PRON-	PRP
altman	202	11	truly	truly	RB
altman	202	12	need	need	VBP
altman	202	13	,	,	,
altman	202	14	and	and	CC
altman	202	15	avoid	avoid	VB
altman	202	16	getting	get	VBG
altman	202	17	bogged	bogge	VBD
altman	202	18	down	down	RP
altman	202	19	with	with	IN
altman	202	20	details	detail	NNS
altman	202	21	best	well	RBS
altman	202	22	left	leave	VBN
altman	202	23	to	to	IN
altman	202	24	mathematicians	mathematician	NNS
altman	202	25	.	.	.
altman	203	1	·	·	NFP
altman	203	2	What	what	WP
altman	203	3	do	do	VBP
altman	203	4	the	the	DT
altman	203	5	inputs	input	NNS
altman	203	6	and	and	CC
altman	203	7	outputs	output	NNS
altman	203	8	of	of	IN
altman	203	9	the	the	DT
altman	203	10	algorithm	algorithm	NNP
altman	203	11	mean	mean	VB
altman	203	12	?	?	.
altman	204	1	There	there	EX
altman	204	2	are	be	VBP
altman	204	3	two	two	CD
altman	204	4	parts	part	NNS
altman	204	5	to	to	IN
altman	204	6	answering	answer	VBG
altman	204	7	this	this	DT
altman	204	8	question	question	NN
altman	204	9	.	.	.
altman	205	1	First	first	RB
altman	205	2	is	be	VBZ
altman	205	3	the	the	DT
altman	205	4	data	data	NN
altman	205	5	structure	structure	NN
altman	205	6	,	,	,
altman	205	7	e.g.	e.g.	RB
altman	206	1	“	"	``
altman	206	2	this	this	DT
altman	206	3	is	be	VBZ
altman	206	4	a	a	DT
altman	206	5	vector	vector	NN
altman	206	6	with	with	IN
altman	206	7	300	300	CD
altman	206	8	integers	integer	NNS
altman	206	9	.	.	.
altman	206	10	”	"	''
altman	206	11	Second	second	JJ
altman	206	12	is	be	VBZ
altman	206	13	knowing	know	VBG
altman	206	14	what	what	WP
altman	206	15	this	this	DT
altman	206	16	data	data	NN
altman	206	17	describes	describe	VBZ
altman	206	18	,	,	,
altman	206	19	e.g.	e.g.	RB
altman	207	1	“	"	``
altman	207	2	each	each	DT
altman	207	3	vector	vector	NN
altman	207	4	represents	represent	VBZ
altman	207	5	a	a	DT
altman	207	6	document	document	NN
altman	207	7	,	,	,
altman	207	8	and	and	CC
altman	207	9	each	each	DT
altman	207	10	integer	integer	NN
altman	207	11	specifies	specify	VBZ
altman	207	12	the	the	DT
altman	207	13	number	number	NN
altman	207	14	of	of	IN
altman	207	15	times	time	NNS
altman	207	16	a	a	DT
altman	207	17	particular	particular	JJ
altman	207	18	word	word	NN
altman	207	19	appears	appear	VBZ
altman	207	20	in	in	IN
altman	207	21	that	that	DT
altman	207	22	document	document	NN
altman	207	23	.	.	.
altman	207	24	”	"	''
altman	207	25	You	-PRON-	PRP
altman	207	26	also	also	RB
altman	207	27	need	need	VBP
altman	207	28	to	to	TO
altman	207	29	be	be	VB
altman	207	30	aware	aware	JJ
altman	207	31	of	of	IN
altman	207	32	specific	specific	JJ
altman	207	33	implementation	implementation	NN
altman	207	34	details	detail	NNS
altman	207	35	—	—	:
altman	207	36	perhaps	perhaps	RB
altman	207	37	the	the	DT
altman	207	38	input	input	NN
altman	207	39	needs	need	VBZ
altman	207	40	to	to	TO
altman	207	41	be	be	VB
altman	207	42	normalized	normalize	VBN
altman	207	43	in	in	IN
altman	207	44	some	some	DT
altman	207	45	way	way	NN
altman	207	46	,	,	,
altman	207	47	perhaps	perhaps	RB
altman	207	48	the	the	DT
altman	207	49	output	output	NN
altman	207	50	has	have	VBZ
altman	207	51	been	be	VBN
altman	207	52	smoothed	smooth	VBN
altman	207	53	(	(	-LRB-
altman	207	54	a	a	DT
altman	207	55	technique	technique	NN
altman	207	56	that	that	WDT
altman	207	57	compensates	compensate	VBZ
altman	207	58	for	for	IN
altman	207	59	noisy	noisy	JJ
altman	207	60	data	datum	NNS
altman	207	61	or	or	CC
altman	207	62	outliers	outlier	NNS
altman	207	63	)	)	-RRB-
altman	207	64	.	.	.
altman	208	1	This	this	DT
altman	208	2	may	may	MD
altman	208	3	seem	seem	VB
altman	208	4	straightforward	straightforward	JJ
altman	208	5	,	,	,
altman	208	6	but	but	CC
altman	208	7	it	-PRON-	PRP
altman	208	8	can	can	MD
altman	208	9	be	be	VB
altman	208	10	a	a	DT
altman	208	11	lot	lot	NN
altman	208	12	to	to	TO
altman	208	13	keep	keep	VB
altman	208	14	track	track	NN
altman	208	15	of	of	IN
altman	208	16	once	once	IN
altman	208	17	you	-PRON-	PRP
altman	208	18	’ve	have	VB
altman	208	19	gone	go	VBN
altman	208	20	through	through	IN
altman	208	21	several	several	JJ
altman	208	22	layers	layer	NNS
altman	208	23	of	of	IN
altman	208	24	processing	processing	NN
altman	208	25	and	and	CC
altman	208	26	abstraction	abstraction	NN
altman	208	27	.	.	.
altman	209	1	·	·	NFP
altman	209	2	What	what	WDT
altman	209	3	effect	effect	NN
altman	209	4	do	do	VBP
altman	209	5	different	different	JJ
altman	209	6	hyperparameters	hyperparameter	NNS
altman	209	7	have	have	VB
altman	209	8	on	on	IN
altman	209	9	the	the	DT
altman	209	10	algorithm	algorithm	NN
altman	209	11	?	?	.
altman	210	1	Part	part	NN
altman	210	2	of	of	IN
altman	210	3	the	the	DT
altman	210	4	machine	machine	NN
altman	210	5	learning	learning	NN
altman	210	6	process	process	NN
altman	210	7	is	be	VBZ
altman	210	8	tuning	tune	VBG
altman	210	9	hyperparameters	hyperparameter	NNS
altman	210	10	,	,	,
altman	210	11	or	or	CC
altman	210	12	trying	try	VBG
altman	210	13	out	out	RP
altman	210	14	multiple	multiple	JJ
altman	210	15	configurations	configuration	NNS
altman	210	16	until	until	IN
altman	210	17	you	-PRON-	PRP
altman	210	18	get	get	VBP
altman	210	19	satisfying	satisfying	JJ
altman	210	20	results	result	NNS
altman	210	21	.	.	.
altman	211	1	Part	part	NN
altman	211	2	of	of	IN
altman	211	3	the	the	DT
altman	211	4	frustration	frustration	NN
altman	211	5	is	be	VBZ
altman	211	6	that	that	IN
altman	211	7	you	-PRON-	PRP
altman	211	8	ca	can	MD
altman	211	9	n’t	not	RB
altman	211	10	try	try	VB
altman	211	11	every	every	DT
altman	211	12	possible	possible	JJ
altman	211	13	configuration	configuration	NN
altman	211	14	,	,	,
altman	211	15	so	so	CC
altman	211	16	you	-PRON-	PRP
altman	211	17	have	have	VBP
altman	211	18	to	to	TO
altman	211	19	do	do	VB
altman	211	20	some	some	DT
altman	211	21	intelligent	intelligent	JJ
altman	211	22	guesswork	guesswork	NN
altman	211	23	.	.	.
altman	212	1	Twiddling	twiddling	NN
altman	212	2	hyperparameters	hyperparameter	NNS
altman	212	3	can	can	MD
altman	212	4	feel	feel	VB
altman	212	5	enigmatic	enigmatic	JJ
altman	212	6	and	and	CC
altman	212	7	unitutive	unitutive	JJ
altman	212	8	,	,	,
altman	212	9	since	since	IN
altman	212	10	it	-PRON-	PRP
altman	212	11	can	can	MD
altman	212	12	be	be	VB
altman	212	13	difficult	difficult	JJ
altman	212	14	to	to	TO
altman	212	15	predict	predict	VB
altman	212	16	their	-PRON-	PRP$
altman	212	17	impact	impact	NN
altman	212	18	on	on	IN
altman	212	19	the	the	DT
altman	212	20	final	final	JJ
altman	212	21	outcome	outcome	NN
altman	212	22	.	.	.
altman	213	1	The	the	DT
altman	213	2	better	well	RBR
altman	213	3	you	-PRON-	PRP
altman	213	4	understand	understand	VBP
altman	213	5	hyperparameters	hyperparameter	NNS
altman	213	6	and	and	CC
altman	213	7	their	-PRON-	PRP$
altman	213	8	roles	role	NNS
altman	213	9	in	in	IN
altman	213	10	the	the	DT
altman	213	11	ML	ML	NNP
altman	213	12	process	process	NN
altman	213	13	,	,	,
altman	213	14	the	the	DT
altman	213	15	more	more	RBR
altman	213	16	likely	likely	JJ
altman	213	17	you	-PRON-	PRP
altman	213	18	are	be	VBP
altman	213	19	to	to	TO
altman	213	20	make	make	VB
altman	213	21	reasonable	reasonable	JJ
altman	213	22	guesses	guess	NNS
altman	213	23	and	and	CC
altman	213	24	adjustments	adjustment	NNS
altman	213	25	—	—	:
altman	213	26	though	though	IN
altman	213	27	you	-PRON-	PRP
altman	213	28	should	should	MD
altman	213	29	always	always	RB
altman	213	30	be	be	VB
altman	213	31	prepared	prepare	VBN
altman	213	32	for	for	IN
altman	213	33	a	a	DT
altman	213	34	surprise	surprise	NN
altman	213	35	.	.	.
altman	214	1	·	·	NFP
altman	214	2	Can	Can	MD
altman	214	3	you	-PRON-	PRP
altman	214	4	explain	explain	VB
altman	214	5	how	how	WRB
altman	214	6	this	this	DT
altman	214	7	algorithm	algorithm	NN
altman	214	8	works	work	VBZ
altman	214	9	to	to	IN
altman	214	10	a	a	DT
altman	214	11	lay	lie	VBN
altman	214	12	person	person	NN
altman	214	13	and	and	CC
altman	214	14	why	why	WRB
altman	214	15	it	-PRON-	PRP
altman	214	16	’s	’	VBZ
altman	214	17	beneficial	beneficial	JJ
altman	214	18	to	to	IN
altman	214	19	the	the	DT
altman	214	20	project	project	NN
altman	214	21	?	?	.
altman	215	1	There	there	EX
altman	215	2	are	be	VBP
altman	215	3	two	two	CD
altman	215	4	benefits	benefit	NNS
altman	215	5	to	to	TO
altman	215	6	articulating	articulate	VBG
altman	215	7	a	a	DT
altman	215	8	response	response	NN
altman	215	9	to	to	IN
altman	215	10	this	this	DT
altman	215	11	question	question	NN
altman	215	12	.	.	.
altman	216	1	First	first	RB
altman	216	2	,	,	,
altman	216	3	it	-PRON-	PRP
altman	216	4	ensures	ensure	VBZ
altman	216	5	that	that	IN
altman	216	6	you	-PRON-	PRP
altman	216	7	really	really	RB
altman	216	8	understand	understand	VBP
altman	216	9	the	the	DT
altman	216	10	algorithm	algorithm	NNP
altman	216	11	yourself	-PRON-	PRP
altman	216	12	.	.	.
altman	217	1	And	and	CC
altman	217	2	second	second	RB
altman	217	3	,	,	,
altman	217	4	you	-PRON-	PRP
altman	217	5	will	will	MD
altman	217	6	likely	likely	RB
altman	217	7	be	be	VB
altman	217	8	called	call	VBN
altman	217	9	on	on	RP
altman	217	10	to	to	TO
altman	217	11	give	give	VB
altman	217	12	this	this	DT
altman	217	13	explanation	explanation	NN
altman	217	14	to	to	IN
altman	217	15	co	co	NN
altman	217	16	-	-	NNS
altman	217	17	collaborators	collaborator	NNS
altman	217	18	and	and	CC
altman	217	19	other	other	JJ
altman	217	20	stakeholders	stakeholder	NNS
altman	217	21	.	.	.
altman	218	1	A	a	DT
altman	218	2	good	good	JJ
altman	218	3	explanation	explanation	NN
altman	218	4	will	will	MD
altman	218	5	build	build	VB
altman	218	6	excitement	excitement	NN
altman	218	7	around	around	IN
altman	218	8	the	the	DT
altman	218	9	project	project	NN
altman	218	10	,	,	,
altman	218	11	while	while	IN
altman	218	12	a	a	DT
altman	218	13	befuddling	befuddle	VBG
altman	218	14	one	one	PRP
altman	218	15	could	could	MD
altman	218	16	sow	sow	VB
altman	218	17	doubt	doubt	NN
altman	218	18	or	or	CC
altman	218	19	disinterest	disinterest	NN
altman	218	20	.	.	.
altman	219	1	It	-PRON-	PRP
altman	219	2	can	can	MD
altman	219	3	be	be	VB
altman	219	4	difficult	difficult	JJ
altman	219	5	to	to	TO
altman	219	6	strike	strike	VB
altman	219	7	a	a	DT
altman	219	8	balance	balance	NN
altman	219	9	between	between	IN
altman	219	10	general	general	JJ
altman	219	11	summary	summary	NN
altman	219	12	and	and	CC
altman	219	13	technical	technical	JJ
altman	219	14	equations	equation	NNS
altman	219	15	,	,	,
altman	219	16	since	since	IN
altman	219	17	your	-PRON-	PRP$
altman	219	18	stakeholders	stakeholder	NNS
altman	219	19	will	will	MD
altman	219	20	likely	likely	RB
altman	219	21	include	include	VB
altman	219	22	people	people	NNS
altman	219	23	with	with	IN
altman	219	24	diverse	diverse	JJ
altman	219	25	backgrounds	background	NNS
altman	219	26	,	,	,
altman	219	27	so	so	RB
altman	219	28	do	do	VB
altman	219	29	your	-PRON-	PRP$
altman	219	30	best	good	JJS
altman	219	31	and	and	CC
altman	219	32	look	look	VB
altman	219	33	for	for	IN
altman	219	34	opportunities	opportunity	NNS
altman	219	35	for	for	IN
altman	219	36	people	people	NNS
altman	219	37	with	with	IN
altman	219	38	different	different	JJ
altman	219	39	expertises	expertise	NNS
altman	219	40	to	to	TO
altman	219	41	help	help	VB
altman	219	42	refine	refine	VB
altman	219	43	your	-PRON-	PRP$
altman	219	44	team	team	NN
altman	219	45	’s	’s	POS
altman	219	46	understanding	understanding	NN
altman	219	47	of	of	IN
altman	219	48	the	the	DT
altman	219	49	algorithm	algorithm	NN
altman	219	50	.	.	.
altman	220	1	Learning	learn	VBG
altman	220	2	more	more	JJR
altman	220	3	about	about	IN
altman	220	4	the	the	DT
altman	220	5	underlying	underlie	VBG
altman	220	6	math	math	NN
altman	220	7	can	can	MD
altman	220	8	help	help	VB
altman	220	9	you	-PRON-	PRP
altman	220	10	make	make	VB
altman	220	11	better	well	JJR
altman	220	12	,	,	,
altman	220	13	more	more	RBR
altman	220	14	nuanced	nuanced	JJ
altman	220	15	decisions	decision	NNS
altman	220	16	about	about	IN
altman	220	17	how	how	WRB
altman	220	18	to	to	TO
altman	220	19	deploy	deploy	VB
altman	220	20	the	the	DT
altman	220	21	algorithm	algorithm	NN
altman	220	22	,	,	,
altman	220	23	and	and	CC
altman	220	24	is	be	VBZ
altman	220	25	fascinating	fascinating	JJ
altman	220	26	in	in	IN
altman	220	27	its	-PRON-	PRP$
altman	220	28	own	own	JJ
altman	220	29	right	right	NN
altman	220	30	—	—	:
altman	220	31	but	but	CC
altman	220	32	in	in	IN
altman	220	33	most	most	JJS
altman	220	34	cases	case	NNS
altman	220	35	I	-PRON-	PRP
altman	220	36	have	have	VBP
altman	220	37	found	find	VBN
altman	220	38	that	that	IN
altman	220	39	the	the	DT
altman	220	40	above	above	JJ
altman	220	41	three	three	CD
altman	220	42	questions	question	NNS
altman	220	43	provide	provide	VBP
altman	220	44	a	a	DT
altman	220	45	solid	solid	JJ
altman	220	46	foundation	foundation	NN
altman	220	47	for	for	IN
altman	220	48	machine	machine	NN
altman	220	49	learning	learn	VBG
altman	220	50	research	research	NN
altman	220	51	.	.	.
altman	221	1	Tool	tool	NN
altman	221	2	selection	selection	NN
altman	221	3	Tool	tool	NN
altman	221	4	selection	selection	NN
altman	221	5	is	be	VBZ
altman	221	6	an	an	DT
altman	221	7	important	important	JJ
altman	221	8	part	part	NN
altman	221	9	of	of	IN
altman	221	10	your	-PRON-	PRP$
altman	221	11	process	process	NN
altman	221	12	and	and	CC
altman	221	13	should	should	MD
altman	221	14	be	be	VB
altman	221	15	approached	approach	VBN
altman	221	16	thoughtfully	thoughtfully	RB
altman	221	17	.	.	.
altman	222	1	A	a	DT
altman	222	2	good	good	JJ
altman	222	3	approach	approach	NN
altman	222	4	is	be	VBZ
altman	222	5	to	to	TO
altman	222	6	articulate	articulate	VB
altman	222	7	and	and	CC
altman	222	8	prioritize	prioritize	VB
altman	222	9	the	the	DT
altman	222	10	needs	need	NNS
altman	222	11	of	of	IN
altman	222	12	your	-PRON-	PRP$
altman	222	13	team	team	NN
altman	222	14	,	,	,
altman	222	15	and	and	CC
altman	222	16	make	make	VB
altman	222	17	selections	selection	NNS
altman	222	18	that	that	WDT
altman	222	19	meet	meet	VBP
altman	222	20	these	these	DT
altman	222	21	needs	need	NNS
altman	222	22	.	.	.
altman	223	1	I	-PRON-	PRP
altman	223	2	’ve	have	VB
altman	223	3	listed	list	VBN
altman	223	4	some	some	DT
altman	223	5	possible	possible	JJ
altman	223	6	questions	question	NNS
altman	223	7	for	for	IN
altman	223	8	consideration	consideration	NN
altman	223	9	below	below	IN
altman	223	10	,	,	,
altman	223	11	many	many	JJ
altman	223	12	of	of	IN
altman	223	13	which	which	WDT
altman	223	14	you	-PRON-	PRP
altman	223	15	will	will	MD
altman	223	16	recognize	recognize	VB
altman	223	17	as	as	IN
altman	223	18	general	general	JJ
altman	223	19	concerns	concern	NNS
altman	223	20	for	for	IN
altman	223	21	any	any	DT
altman	223	22	tool	tool	NN
altman	223	23	selection	selection	NN
altman	223	24	process	process	NN
altman	223	25	.	.	.
altman	224	1	·	·	NFP
altman	224	2	What	what	WDT
altman	224	3	sorts	sort	NNS
altman	224	4	of	of	IN
altman	224	5	features	feature	NNS
altman	224	6	and	and	CC
altman	224	7	interfaces	interface	NNS
altman	224	8	do	do	VBP
altman	224	9	they	-PRON-	PRP
altman	224	10	offer	offer	VB
altman	224	11	?	?	.
altman	225	1	If	if	IN
altman	225	2	you	-PRON-	PRP
altman	225	3	require	require	VBP
altman	225	4	a	a	DT
altman	225	5	specific	specific	JJ
altman	225	6	algorithm	algorithm	NN
altman	225	7	,	,	,
altman	225	8	the	the	DT
altman	225	9	ability	ability	NN
altman	225	10	to	to	TO
altman	225	11	make	make	VB
altman	225	12	data	datum	NNS
altman	225	13	visualizations	visualization	NNS
altman	225	14	,	,	,
altman	225	15	or	or	CC
altman	225	16	query	query	NN
altman	225	17	interfaces	interface	NNS
altman	225	18	,	,	,
altman	225	19	you	-PRON-	PRP
altman	225	20	can	can	MD
altman	225	21	find	find	VB
altman	225	22	tools	tool	NNS
altman	225	23	to	to	TO
altman	225	24	meet	meet	VB
altman	225	25	these	these	DT
altman	225	26	specific	specific	JJ
altman	225	27	needs	need	NNS
altman	225	28	.	.	.
altman	226	1	·	·	NFP
altman	226	2	How	how	WRB
altman	226	3	well	well	RB
altman	226	4	do	do	VBP
altman	226	5	tools	tool	NNS
altman	226	6	interoperate	interoperate	VB
altman	226	7	with	with	IN
altman	226	8	one	one	CD
altman	226	9	another	another	DT
altman	226	10	,	,	,
altman	226	11	or	or	CC
altman	226	12	with	with	IN
altman	226	13	other	other	JJ
altman	226	14	parts	part	NNS
altman	226	15	of	of	IN
altman	226	16	your	-PRON-	PRP$
altman	226	17	existing	exist	VBG
altman	226	18	systems	system	NNS
altman	226	19	?	?	.
altman	227	1	One	one	CD
altman	227	2	of	of	IN
altman	227	3	the	the	DT
altman	227	4	advantages	advantage	NNS
altman	227	5	of	of	IN
altman	227	6	a	a	DT
altman	227	7	well	well	RB
altman	227	8	-	-	HYPH
altman	227	9	designed	design	VBN
altman	227	10	pipeline	pipeline	NN
altman	227	11	is	be	VBZ
altman	227	12	that	that	IN
altman	227	13	it	-PRON-	PRP
altman	227	14	will	will	MD
altman	227	15	enable	enable	VB
altman	227	16	you	-PRON-	PRP
altman	227	17	to	to	TO
altman	227	18	swap	swap	VB
altman	227	19	out	out	RP
altman	227	20	software	software	NN
altman	227	21	components	component	NNS
altman	227	22	if	if	IN
altman	227	23	the	the	DT
altman	227	24	need	need	NN
altman	227	25	arises	arise	VBZ
altman	227	26	.	.	.
altman	228	1	For	for	IN
altman	228	2	example	example	NN
altman	228	3	,	,	,
altman	228	4	if	if	IN
altman	228	5	your	-PRON-	PRP$
altman	228	6	data	data	NN
altman	228	7	is	be	VBZ
altman	228	8	in	in	IN
altman	228	9	a	a	DT
altman	228	10	format	format	NN
altman	228	11	that	that	WDT
altman	228	12	is	be	VBZ
altman	228	13	interoperable	interoperable	JJ
altman	228	14	with	with	IN
altman	228	15	many	many	JJ
altman	228	16	systems	system	NNS
altman	228	17	,	,	,
altman	228	18	it	-PRON-	PRP
altman	228	19	frees	free	VBZ
altman	228	20	you	-PRON-	PRP
altman	228	21	from	from	IN
altman	228	22	being	be	VBG
altman	228	23	tied	tie	VBN
altman	228	24	down	down	RP
altman	228	25	to	to	IN
altman	228	26	any	any	DT
altman	228	27	specific	specific	JJ
altman	228	28	tool	tool	NN
altman	228	29	.	.	.
altman	229	1	·	·	NFP
altman	229	2	How	how	WRB
altman	229	3	do	do	VBP
altman	229	4	the	the	DT
altman	229	5	tools	tool	NNS
altman	229	6	align	align	VB
altman	229	7	with	with	IN
altman	229	8	the	the	DT
altman	229	9	skill	skill	NN
altman	229	10	sets	set	NNS
altman	229	11	and	and	CC
altman	229	12	comfort	comfort	NN
altman	229	13	levels	level	NNS
altman	229	14	of	of	IN
altman	229	15	your	-PRON-	PRP$
altman	229	16	team	team	NN
altman	229	17	?	?	.
altman	230	1	For	for	IN
altman	230	2	example	example	NN
altman	230	3	,	,	,
altman	230	4	consider	consider	VB
altman	230	5	what	what	WP
altman	230	6	coding	coding	NN
altman	230	7	languages	language	NNS
altman	230	8	your	-PRON-	PRP$
altman	230	9	collaborators	collaborator	NNS
altman	230	10	know	know	VB
altman	230	11	,	,	,
altman	230	12	and	and	CC
altman	230	13	whether	whether	IN
altman	230	14	or	or	CC
altman	230	15	not	not	RB
altman	230	16	they	-PRON-	PRP
altman	230	17	have	have	VBP
altman	230	18	the	the	DT
altman	230	19	capacity	capacity	NN
altman	230	20	to	to	TO
altman	230	21	learn	learn	VB
altman	230	22	a	a	DT
altman	230	23	new	new	JJ
altman	230	24	one	one	NN
altman	230	25	.	.	.
altman	231	1	If	if	IN
altman	231	2	you	-PRON-	PRP
altman	231	3	have	have	VBP
altman	231	4	someone	someone	NN
altman	231	5	who	who	WP
altman	231	6	is	be	VBZ
altman	231	7	already	already	RB
altman	231	8	a	a	DT
altman	231	9	wiz	wiz	NN
altman	231	10	with	with	IN
altman	231	11	a	a	DT
altman	231	12	preferred	preferred	JJ
altman	231	13	spreadsheet	spreadsheet	NN
altman	231	14	program	program	NN
altman	231	15	,	,	,
altman	231	16	see	see	VB
altman	231	17	if	if	IN
altman	231	18	you	-PRON-	PRP
altman	231	19	can	can	MD
altman	231	20	export	export	VB
altman	231	21	data	datum	NNS
altman	231	22	into	into	IN
altman	231	23	a	a	DT
altman	231	24	compatible	compatible	JJ
altman	231	25	file	file	NN
altman	231	26	format	format	NN
altman	231	27	.	.	.
altman	232	1	·	·	NFP
altman	232	2	Are	be	VBP
altman	232	3	the	the	DT
altman	232	4	tools	tool	NNS
altman	232	5	stable	stable	JJ
altman	232	6	,	,	,
altman	232	7	well	well	RB
altman	232	8	-	-	HYPH
altman	232	9	documented	document	VBN
altman	232	10	,	,	,
altman	232	11	and	and	CC
altman	232	12	well	well	RB
altman	232	13	-	-	HYPH
altman	232	14	supported	support	VBN
altman	232	15	?	?	.
altman	233	1	Machine	machine	NN
altman	233	2	learning	learning	NN
altman	233	3	is	be	VBZ
altman	233	4	a	a	DT
altman	233	5	fast	fast	RB
altman	233	6	-	-	HYPH
altman	233	7	changing	change	VBG
altman	233	8	field	field	NN
altman	233	9	,	,	,
altman	233	10	with	with	IN
altman	233	11	new	new	JJ
altman	233	12	algorithms	algorithm	NNS
altman	233	13	,	,	,
altman	233	14	services	service	NNS
altman	233	15	,	,	,
altman	233	16	and	and	CC
altman	233	17	software	software	NN
altman	233	18	features	feature	NNS
altman	233	19	being	be	VBG
altman	233	20	developed	develop	VBN
altman	233	21	all	all	PDT
altman	233	22	the	the	DT
altman	233	23	time	time	NN
altman	233	24	.	.	.
altman	234	1	Something	something	NN
altman	234	2	new	new	JJ
altman	234	3	and	and	CC
altman	234	4	exciting	exciting	JJ
altman	234	5	that	that	WDT
altman	234	6	has	have	VBZ
altman	234	7	n’t	not	RB
altman	234	8	yet	yet	RB
altman	234	9	been	be	VBN
altman	234	10	road	road	NN
altman	234	11	-	-	HYPH
altman	234	12	tested	test	VBN
altman	234	13	may	may	MD
altman	234	14	not	not	RB
altman	234	15	be	be	VB
altman	234	16	worth	worth	JJ
altman	234	17	the	the	DT
altman	234	18	risk	risk	NN
altman	234	19	if	if	IN
altman	234	20	there	there	EX
altman	234	21	is	be	VBZ
altman	234	22	a	a	DT
altman	234	23	more	more	RBR
altman	234	24	dependable	dependable	JJ
altman	234	25	alternative	alternative	NN
altman	234	26	.	.	.
altman	235	1	Furthermore	furthermore	RB
altman	235	2	,	,	,
altman	235	3	there	there	EX
altman	235	4	tends	tend	VBZ
altman	235	5	to	to	TO
altman	235	6	be	be	VB
altman	235	7	more	more	JJR
altman	235	8	scholarship	scholarship	NN
altman	235	9	,	,	,
altman	235	10	documented	document	VBN
altman	235	11	use	use	NN
altman	235	12	cases	case	NNS
altman	235	13	,	,	,
altman	235	14	and	and	CC
altman	235	15	tutorials	tutorial	NNS
altman	235	16	for	for	IN
altman	235	17	older	old	JJR
altman	235	18	,	,	,
altman	235	19	more	more	RBR
altman	235	20	widely	widely	RB
altman	235	21	-	-	HYPH
altman	235	22	adopted	adopt	VBN
altman	235	23	tools	tool	NNS
altman	235	24	.	.	.
altman	236	1	·	·	NFP
altman	236	2	Are	be	VBP
altman	236	3	you	-PRON-	PRP
altman	236	4	concerned	concerned	JJ
altman	236	5	about	about	IN
altman	236	6	speed	speed	NN
altman	236	7	and	and	CC
altman	236	8	scale	scale	NN
altman	236	9	?	?	.
altman	237	1	Do	do	VBP
altman	237	2	n’t	not	RB
altman	237	3	get	get	VB
altman	237	4	bogged	bogge	VBN
altman	237	5	down	down	RP
altman	237	6	with	with	IN
altman	237	7	these	these	DT
altman	237	8	considerations	consideration	NNS
altman	237	9	if	if	IN
altman	237	10	you	-PRON-	PRP
altman	237	11	’re	be	VBZ
altman	237	12	just	just	RB
altman	237	13	trying	try	VBG
altman	237	14	to	to	TO
altman	237	15	get	get	VB
altman	237	16	a	a	DT
altman	237	17	working	working	NN
altman	237	18	pilot	pilot	NN
altman	237	19	off	off	IN
altman	237	20	the	the	DT
altman	237	21	ground	ground	NN
altman	237	22	,	,	,
altman	237	23	but	but	CC
altman	237	24	it	-PRON-	PRP
altman	237	25	can	can	MD
altman	237	26	help	help	VB
altman	237	27	to	to	IN
altman	237	28	at	at	IN
altman	237	29	least	least	JJS
altman	237	30	be	be	VB
altman	237	31	aware	aware	JJ
altman	237	32	of	of	IN
altman	237	33	how	how	WRB
altman	237	34	problems	problem	NNS
altman	237	35	are	be	VBP
altman	237	36	likely	likely	JJ
altman	237	37	to	to	TO
altman	237	38	manifest	manifest	VB
altman	237	39	as	as	IN
altman	237	40	your	-PRON-	PRP$
altman	237	41	volume	volume	NN
altman	237	42	of	of	IN
altman	237	43	data	datum	NNS
altman	237	44	increases	increase	NNS
altman	237	45	,	,	,
altman	237	46	or	or	CC
altman	237	47	as	as	IN
altman	237	48	you	-PRON-	PRP
altman	237	49	integrate	integrate	VBP
altman	237	50	into	into	IN
altman	237	51	time	time	NN
altman	237	52	-	-	HYPH
altman	237	53	sensitive	sensitive	JJ
altman	237	54	workflows	workflow	NNS
altman	237	55	.	.	.
altman	238	1	You	-PRON-	PRP
altman	238	2	and	and	CC
altman	238	3	your	-PRON-	PRP$
altman	238	4	team	team	NN
altman	238	5	can	can	MD
altman	238	6	work	work	VB
altman	238	7	through	through	IN
altman	238	8	these	these	DT
altman	238	9	questions	question	NNS
altman	238	10	and	and	CC
altman	238	11	articulate	articulate	VB
altman	238	12	additional	additional	JJ
altman	238	13	requirements	requirement	NNS
altman	238	14	relevant	relevant	JJ
altman	238	15	to	to	IN
altman	238	16	your	-PRON-	PRP$
altman	238	17	specific	specific	JJ
altman	238	18	context	context	NN
altman	238	19	.	.	.
altman	239	1	Scaling	scale	VBG
altman	239	2	up	up	RP
altman	239	3	Scaling	scale	VBG
altman	239	4	up	up	RP
altman	239	5	in	in	IN
altman	239	6	machine	machine	NN
altman	239	7	learning	learning	NN
altman	239	8	generally	generally	RB
altman	239	9	means	mean	VBZ
altman	239	10	that	that	IN
altman	239	11	you	-PRON-	PRP
altman	239	12	need	need	VBP
altman	239	13	to	to	TO
altman	239	14	work	work	VB
altman	239	15	with	with	IN
altman	239	16	a	a	DT
altman	239	17	larger	large	JJR
altman	239	18	volume	volume	NN
altman	239	19	of	of	IN
altman	239	20	data	datum	NNS
altman	239	21	,	,	,
altman	239	22	or	or	CC
altman	239	23	that	that	IN
altman	239	24	you	-PRON-	PRP
altman	239	25	need	need	VBP
altman	239	26	processes	process	NNS
altman	239	27	to	to	TO
altman	239	28	execute	execute	VB
altman	239	29	faster	fast	RBR
altman	239	30	.	.	.
altman	240	1	Recent	recent	JJ
altman	240	2	advances	advance	NNS
altman	240	3	in	in	IN
altman	240	4	hardware	hardware	NN
altman	240	5	and	and	CC
altman	240	6	software	software	NN
altman	240	7	make	make	VBP
altman	240	8	the	the	DT
altman	240	9	execution	execution	NN
altman	240	10	of	of	IN
altman	240	11	complex	complex	JJ
altman	240	12	computations	computation	NNS
altman	240	13	magnitudes	magnitude	NNS
altman	240	14	faster	fast	RBR
altman	240	15	and	and	CC
altman	240	16	more	more	RBR
altman	240	17	efficient	efficient	JJ
altman	240	18	than	than	IN
altman	240	19	they	-PRON-	PRP
altman	240	20	were	be	VBD
altman	240	21	even	even	RB
altman	240	22	a	a	DT
altman	240	23	decade	decade	NN
altman	240	24	ago	ago	RB
altman	240	25	,	,	,
altman	240	26	and	and	CC
altman	240	27	you	-PRON-	PRP
altman	240	28	can	can	MD
altman	240	29	often	often	RB
altman	240	30	achieve	achieve	VB
altman	240	31	quite	quite	PDT
altman	240	32	a	a	DT
altman	240	33	bit	bit	NN
altman	240	34	by	by	IN
altman	240	35	working	work	VBG
altman	240	36	on	on	IN
altman	240	37	a	a	DT
altman	240	38	personal	personal	JJ
altman	240	39	computer	computer	NN
altman	240	40	.	.	.
altman	241	1	Yet	yet	RB
altman	241	2	,	,	,
altman	241	3	time	time	NN
altman	241	4	is	be	VBZ
altman	241	5	valuable	valuable	JJ
altman	241	6	,	,	,
altman	241	7	and	and	CC
altman	241	8	it	-PRON-	PRP
altman	241	9	can	can	MD
altman	241	10	be	be	VB
altman	241	11	difficult	difficult	JJ
altman	241	12	to	to	TO
altman	241	13	iterate	iterate	VB
altman	241	14	and	and	CC
altman	241	15	experiment	experiment	VB
altman	241	16	effectively	effectively	RB
altman	241	17	when	when	WRB
altman	241	18	individual	individual	JJ
altman	241	19	processes	process	NNS
altman	241	20	take	take	VBP
altman	241	21	too	too	RB
altman	241	22	long	long	JJ
altman	241	23	to	to	TO
altman	241	24	execute	execute	VB
altman	241	25	.	.	.
altman	242	1	There	there	EX
altman	242	2	are	be	VBP
altman	242	3	many	many	JJ
altman	242	4	ML	ML	NNP
altman	242	5	software	software	NN
altman	242	6	packages	package	NNS
altman	242	7	that	that	WDT
altman	242	8	can	can	MD
altman	242	9	help	help	VB
altman	242	10	you	-PRON-	PRP
altman	242	11	make	make	VB
altman	242	12	efficient	efficient	JJ
altman	242	13	use	use	NN
altman	242	14	of	of	IN
altman	242	15	whatever	whatever	WDT
altman	242	16	hardware	hardware	NN
altman	242	17	you	-PRON-	PRP
altman	242	18	have	have	VBP
altman	242	19	,	,	,
altman	242	20	including	include	VBG
altman	242	21	your	-PRON-	PRP$
altman	242	22	personal	personal	JJ
altman	242	23	computer	computer	NN
altman	242	24	.	.	.
altman	243	1	Some	some	DT
altman	243	2	examples	example	NNS
altman	243	3	at	at	IN
altman	243	4	the	the	DT
altman	243	5	time	time	NN
altman	243	6	of	of	IN
altman	243	7	writing	writing	NN
altman	243	8	are	be	VBP
altman	243	9	Apache	Apache	NNP
altman	243	10	Spark	Spark	NNP
altman	243	11	,	,	,
altman	243	12	TensorFlow	TensorFlow	NNP
altman	243	13	,	,	,
altman	243	14	Scikit	Scikit	NNP
altman	243	15	-	-	HYPH
altman	243	16	learn	learn	VB
altman	243	17	,	,	,
altman	243	18	and	and	CC
altman	243	19	Microsoft	Microsoft	NNP
altman	243	20	Cognitive	Cognitive	NNP
altman	243	21	Toolkit	Toolkit	NNP
altman	243	22	,	,	,
altman	243	23	each	each	DT
altman	243	24	with	with	IN
altman	243	25	their	-PRON-	PRP$
altman	243	26	own	own	JJ
altman	243	27	strengths	strength	NNS
altman	243	28	and	and	CC
altman	243	29	applications	application	NNS
altman	243	30	.	.	.
altman	244	1	In	in	IN
altman	244	2	addition	addition	NN
altman	244	3	to	to	IN
altman	244	4	providing	provide	VBG
altman	244	5	libraries	library	NNS
altman	244	6	for	for	IN
altman	244	7	building	building	NN
altman	244	8	and	and	CC
altman	244	9	testing	testing	NN
altman	244	10	models	model	NNS
altman	244	11	,	,	,
altman	244	12	these	these	DT
altman	244	13	software	software	NN
altman	244	14	packages	package	NNS
altman	244	15	optimize	optimize	VBP
altman	244	16	algorithmic	algorithmic	JJ
altman	244	17	performance	performance	NN
altman	244	18	,	,	,
altman	244	19	memory	memory	NN
altman	244	20	resources	resource	NNS
altman	244	21	,	,	,
altman	244	22	data	datum	NNS
altman	244	23	throughputs	throughput	NNS
altman	244	24	,	,	,
altman	244	25	and/or	and/or	CC
altman	244	26	parallel	parallel	JJ
altman	244	27	computations	computation	NNS
altman	244	28	.	.	.
altman	245	1	They	-PRON-	PRP
altman	245	2	can	can	MD
altman	245	3	make	make	VB
altman	245	4	a	a	DT
altman	245	5	remarkable	remarkable	JJ
altman	245	6	difference	difference	NN
altman	245	7	in	in	IN
altman	245	8	both	both	DT
altman	245	9	processing	processing	NN
altman	245	10	speed	speed	NN
altman	245	11	and	and	CC
altman	245	12	the	the	DT
altman	245	13	amount	amount	NN
altman	245	14	of	of	IN
altman	245	15	data	datum	NNS
altman	245	16	you	-PRON-	PRP
altman	245	17	can	can	MD
altman	245	18	comfortably	comfortably	RB
altman	245	19	handle	handle	VB
altman	245	20	.	.	.
altman	246	1	There	there	EX
altman	246	2	are	be	VBP
altman	246	3	also	also	RB
altman	246	4	services	service	NNS
altman	246	5	that	that	WDT
altman	246	6	allow	allow	VBP
altman	246	7	you	-PRON-	PRP
altman	246	8	to	to	TO
altman	246	9	submit	submit	VB
altman	246	10	executable	executable	JJ
altman	246	11	code	code	NN
altman	246	12	and	and	CC
altman	246	13	data	datum	NNS
altman	246	14	to	to	IN
altman	246	15	the	the	DT
altman	246	16	cloud	cloud	NN
altman	246	17	for	for	IN
altman	246	18	processing	processing	NN
altman	246	19	,	,	,
altman	246	20	such	such	JJ
altman	246	21	as	as	IN
altman	246	22	Google	Google	NNP
altman	246	23	AI	AI	NNP
altman	246	24	Platform	Platform	NNP
altman	246	25	.	.	.
altman	247	1	Managing	manage	VBG
altman	247	2	your	-PRON-	PRP$
altman	247	3	own	own	JJ
altman	247	4	hardware	hardware	NN
altman	247	5	upgrades	upgrade	NNS
altman	247	6	is	be	VBZ
altman	247	7	not	not	RB
altman	247	8	without	without	IN
altman	247	9	challenge	challenge	NN
altman	247	10	.	.	.
altman	248	1	You	-PRON-	PRP
altman	248	2	may	may	MD
altman	248	3	be	be	VB
altman	248	4	lucky	lucky	JJ
altman	248	5	enough	enough	RB
altman	248	6	to	to	TO
altman	248	7	have	have	VB
altman	248	8	access	access	NN
altman	248	9	to	to	IN
altman	248	10	a	a	DT
altman	248	11	high	high	RB
altman	248	12	-	-	HYPH
altman	248	13	powered	power	VBN
altman	248	14	computer	computer	NN
altman	248	15	capable	capable	JJ
altman	248	16	of	of	IN
altman	248	17	accelerated	accelerate	VBN
altman	248	18	processing	processing	NN
altman	248	19	.	.	.
altman	249	1	A	a	DT
altman	249	2	common	common	JJ
altman	249	3	example	example	NN
altman	249	4	is	be	VBZ
altman	249	5	a	a	DT
altman	249	6	computer	computer	NN
altman	249	7	with	with	IN
altman	249	8	GPUs	gpu	NNS
altman	249	9	(	(	-LRB-
altman	249	10	graphics	graphic	NNS
altman	249	11	processing	processing	NN
altman	249	12	units	unit	NNS
altman	249	13	)	)	-RRB-
altman	249	14	,	,	,
altman	249	15	which	which	WDT
altman	249	16	break	break	VBP
altman	249	17	complex	complex	JJ
altman	249	18	processes	process	NNS
altman	249	19	into	into	IN
altman	249	20	many	many	JJ
altman	249	21	small	small	JJ
altman	249	22	tasks	task	NNS
altman	249	23	and	and	CC
altman	249	24	run	run	VB
altman	249	25	them	-PRON-	PRP
altman	249	26	in	in	IN
altman	249	27	parallel	parallel	NN
altman	249	28	.	.	.
altman	250	1	However	however	RB
altman	250	2	,	,	,
altman	250	3	these	these	DT
altman	250	4	powerful	powerful	JJ
altman	250	5	machines	machine	NNS
altman	250	6	can	can	MD
altman	250	7	be	be	VB
altman	250	8	prohibitively	prohibitively	RB
altman	250	9	expensive	expensive	JJ
altman	250	10	.	.	.
altman	251	1	Another	another	DT
altman	251	2	scaling	scaling	NN
altman	251	3	technique	technique	NN
altman	251	4	is	be	VBZ
altman	251	5	distributed	distribute	VBN
altman	251	6	or	or	CC
altman	251	7	cluster	clust	JJR
altman	251	8	computing	computing	NN
altman	251	9	,	,	,
altman	251	10	in	in	IN
altman	251	11	which	which	WDT
altman	251	12	complex	complex	JJ
altman	251	13	processes	process	NNS
altman	251	14	are	be	VBP
altman	251	15	distributed	distribute	VBN
altman	251	16	across	across	IN
altman	251	17	multiple	multiple	JJ
altman	251	18	computers	computer	NNS
altman	251	19	,	,	,
altman	251	20	often	often	RB
altman	251	21	in	in	IN
altman	251	22	the	the	DT
altman	251	23	cloud	cloud	NN
altman	251	24	.	.	.
altman	252	1	A	a	DT
altman	252	2	cloud	cloud	NN
altman	252	3	cluster	cluster	NN
altman	252	4	can	can	MD
altman	252	5	bring	bring	VB
altman	252	6	significant	significant	JJ
altman	252	7	cost	cost	NN
altman	252	8	savings	saving	NNS
altman	252	9	,	,	,
altman	252	10	but	but	CC
altman	252	11	managing	manage	VBG
altman	252	12	one	one	CD
altman	252	13	requires	require	VBZ
altman	252	14	specialized	specialized	JJ
altman	252	15	knowledge	knowledge	NN
altman	252	16	and	and	CC
altman	252	17	the	the	DT
altman	252	18	learning	learn	VBG
altman	252	19	curve	curve	NN
altman	252	20	can	can	MD
altman	252	21	be	be	VB
altman	252	22	rather	rather	RB
altman	252	23	steep	steep	JJ
altman	252	24	.	.	.
altman	253	1	It	-PRON-	PRP
altman	253	2	is	be	VBZ
altman	253	3	also	also	RB
altman	253	4	important	important	JJ
altman	253	5	to	to	TO
altman	253	6	note	note	VB
altman	253	7	that	that	IN
altman	253	8	different	different	JJ
altman	253	9	algorithms	algorithm	NNS
altman	253	10	require	require	VBP
altman	253	11	different	different	JJ
altman	253	12	scaling	scaling	NN
altman	253	13	techniques	technique	NNS
altman	253	14	.	.	.
altman	254	1	Some	some	DT
altman	254	2	clustering	clustering	NN
altman	254	3	algorithms	algorithm	NNS
altman	254	4	,	,	,
altman	254	5	for	for	IN
altman	254	6	example	example	NN
altman	254	7	,	,	,
altman	254	8	scale	scale	NN
altman	254	9	well	well	RB
altman	254	10	with	with	IN
altman	254	11	GPUs	gpu	NNS
altman	254	12	but	but	CC
altman	254	13	not	not	RB
altman	254	14	with	with	IN
altman	254	15	distributed	distribute	VBN
altman	254	16	computing	computing	NN
altman	254	17	.	.	.
altman	255	1	Even	even	RB
altman	255	2	with	with	IN
altman	255	3	the	the	DT
altman	255	4	right	right	JJ
altman	255	5	hardware	hardware	NN
altman	255	6	and	and	CC
altman	255	7	software	software	NN
altman	255	8	,	,	,
altman	255	9	scaling	scale	VBG
altman	255	10	up	up	RP
altman	255	11	can	can	MD
altman	255	12	be	be	VB
altman	255	13	a	a	DT
altman	255	14	tricky	tricky	JJ
altman	255	15	business	business	NN
altman	255	16	.	.	.
altman	256	1	ML	ML	NNP
altman	256	2	processes	process	NNS
altman	256	3	tend	tend	VBP
altman	256	4	to	to	TO
altman	256	5	have	have	VB
altman	256	6	dramatic	dramatic	JJ
altman	256	7	spikes	spike	NNS
altman	256	8	in	in	IN
altman	256	9	memory	memory	NN
altman	256	10	or	or	CC
altman	256	11	network	network	NN
altman	256	12	use	use	NN
altman	256	13	,	,	,
altman	256	14	which	which	WDT
altman	256	15	can	can	MD
altman	256	16	tax	tax	VB
altman	256	17	your	-PRON-	PRP$
altman	256	18	systems	system	NNS
altman	256	19	.	.	.
altman	257	1	Not	not	RB
altman	257	2	all	all	DT
altman	257	3	ML	ML	NNP
altman	257	4	algorithms	algorithm	NNS
altman	257	5	scale	scale	VBP
altman	257	6	well	well	RB
altman	257	7	,	,	,
altman	257	8	causing	cause	VBG
altman	257	9	memory	memory	NN
altman	257	10	use	use	NN
altman	257	11	or	or	CC
altman	257	12	execution	execution	NN
altman	257	13	time	time	NN
altman	257	14	to	to	TO
altman	257	15	grow	grow	VB
altman	257	16	exponentially	exponentially	RB
altman	257	17	as	as	IN
altman	257	18	more	more	JJR
altman	257	19	data	datum	NNS
altman	257	20	is	be	VBZ
altman	257	21	added	add	VBN
altman	257	22	.	.	.
altman	258	1	Sometimes	sometimes	RB
altman	258	2	you	-PRON-	PRP
altman	258	3	have	have	VBP
altman	258	4	to	to	TO
altman	258	5	add	add	VB
altman	258	6	additional	additional	JJ
altman	258	7	,	,	,
altman	258	8	complexity	complexity	NN
altman	258	9	-	-	HYPH
altman	258	10	reducing	reduce	VBG
altman	258	11	steps	step	NNS
altman	258	12	to	to	IN
altman	258	13	your	-PRON-	PRP$
altman	258	14	pipeline	pipeline	NN
altman	258	15	to	to	TO
altman	258	16	handle	handle	VB
altman	258	17	data	datum	NNS
altman	258	18	at	at	IN
altman	258	19	scale	scale	NN
altman	258	20	.	.	.
altman	259	1	Some	some	DT
altman	259	2	of	of	IN
altman	259	3	the	the	DT
altman	259	4	more	more	RBR
altman	259	5	common	common	JJ
altman	259	6	machine	machine	NN
altman	259	7	learning	learning	NN
altman	259	8	languages	language	NNS
altman	259	9	,	,	,
altman	259	10	such	such	JJ
altman	259	11	as	as	IN
altman	259	12	Python	Python	NNP
altman	259	13	and	and	CC
altman	259	14	R	R	NNP
altman	259	15	,	,	,
altman	259	16	execute	execute	VB
altman	259	17	relatively	relatively	RB
altman	259	18	slowly	slowly	RB
altman	259	19	,	,	,
altman	259	20	putting	put	VBG
altman	259	21	the	the	DT
altman	259	22	onus	onus	NN
altman	259	23	on	on	IN
altman	259	24	developers	developer	NNS
altman	259	25	to	to	TO
altman	259	26	optimize	optimize	VB
altman	259	27	operations	operation	NNS
altman	259	28	for	for	IN
altman	259	29	efficiency	efficiency	NN
altman	259	30	.	.	.
altman	260	1	In	in	IN
altman	260	2	anticipation	anticipation	NN
altman	260	3	of	of	IN
altman	260	4	these	these	DT
altman	260	5	and	and	CC
altman	260	6	other	other	JJ
altman	260	7	challenges	challenge	NNS
altman	260	8	,	,	,
altman	260	9	it	-PRON-	PRP
altman	260	10	is	be	VBZ
altman	260	11	often	often	RB
altman	260	12	a	a	DT
altman	260	13	good	good	JJ
altman	260	14	idea	idea	NN
altman	260	15	to	to	TO
altman	260	16	start	start	VB
altman	260	17	with	with	IN
altman	260	18	a	a	DT
altman	260	19	scaled	scale	VBN
altman	260	20	-	-	HYPH
altman	260	21	down	down	RP
altman	260	22	pilot	pilot	NN
altman	260	23	or	or	CC
altman	260	24	proof	proof	NN
altman	260	25	of	of	IN
altman	260	26	concept	concept	NN
altman	260	27	,	,	,
altman	260	28	and	and	CC
altman	260	29	not	not	RB
altman	260	30	to	to	TO
altman	260	31	underestimate	underestimate	VB
altman	260	32	the	the	DT
altman	260	33	time	time	NN
altman	260	34	and	and	CC
altman	260	35	resources	resource	NNS
altman	260	36	necessary	necessary	JJ
altman	260	37	to	to	TO
altman	260	38	scale	scale	VB
altman	260	39	up	up	RP
altman	260	40	from	from	IN
altman	260	41	there	there	RB
altman	260	42	.	.	.
altman	261	1	Conclusion	Conclusion	NNP
altman	261	2	New	New	NNP
altman	261	3	technologies	technology	NNS
altman	261	4	make	make	VBP
altman	261	5	it	-PRON-	PRP
altman	261	6	possible	possible	JJ
altman	261	7	for	for	IN
altman	261	8	more	more	JJR
altman	261	9	researchers	researcher	NNS
altman	261	10	and	and	CC
altman	261	11	developers	developer	NNS
altman	261	12	to	to	TO
altman	261	13	leverage	leverage	VB
altman	261	14	the	the	DT
altman	261	15	power	power	NN
altman	261	16	of	of	IN
altman	261	17	machine	machine	NN
altman	261	18	learning	learning	NN
altman	261	19	.	.	.
altman	262	1	Building	build	VBG
altman	262	2	an	an	DT
altman	262	3	effective	effective	JJ
altman	262	4	machine	machine	NN
altman	262	5	learning	learning	NN
altman	262	6	system	system	NN
altman	262	7	means	mean	VBZ
altman	262	8	supporting	support	VBG
altman	262	9	the	the	DT
altman	262	10	entire	entire	JJ
altman	262	11	workflow	workflow	NN
altman	262	12	,	,	,
altman	262	13	from	from	IN
altman	262	14	data	datum	NNS
altman	262	15	acquisition	acquisition	NN
altman	262	16	to	to	IN
altman	262	17	final	final	JJ
altman	262	18	analysis	analysis	NN
altman	262	19	.	.	.
altman	263	1	Practitioners	practitioner	NNS
altman	263	2	must	must	MD
altman	263	3	be	be	VB
altman	263	4	mindful	mindful	JJ
altman	263	5	of	of	IN
altman	263	6	how	how	WRB
altman	263	7	each	each	DT
altman	263	8	implementation	implementation	NN
altman	263	9	decision	decision	NN
altman	263	10	and	and	CC
altman	263	11	subjective	subjective	JJ
altman	263	12	choice	choice	NN
altman	263	13	—	—	:
altman	263	14	from	from	IN
altman	263	15	the	the	DT
altman	263	16	way	way	NN
altman	263	17	you	-PRON-	PRP
altman	263	18	structure	structure	VBP
altman	263	19	and	and	CC
altman	263	20	store	store	VBP
altman	263	21	your	-PRON-	PRP$
altman	263	22	data	datum	NNS
altman	263	23	to	to	IN
altman	263	24	the	the	DT
altman	263	25	algorithms	algorithm	NNS
altman	263	26	you	-PRON-	PRP
altman	263	27	use	use	VBP
altman	263	28	to	to	IN
altman	263	29	the	the	DT
altman	263	30	ways	way	NNS
altman	263	31	you	-PRON-	PRP
altman	263	32	validate	validate	VBP
altman	263	33	your	-PRON-	PRP$
altman	263	34	results	result	NNS
altman	263	35	—	—	:
altman	263	36	will	will	MD
altman	263	37	impact	impact	VB
altman	263	38	the	the	DT
altman	263	39	efficiency	efficiency	NN
altman	263	40	of	of	IN
altman	263	41	operations	operation	NNS
altman	263	42	and	and	CC
altman	263	43	the	the	DT
altman	263	44	quality	quality	NN
altman	263	45	of	of	IN
altman	263	46	learned	learn	VBN
altman	263	47	intelligence	intelligence	NN
altman	263	48	.	.	.
altman	264	1	This	this	DT
altman	264	2	article	article	NN
altman	264	3	has	have	VBZ
altman	264	4	offered	offer	VBN
altman	264	5	some	some	DT
altman	264	6	practical	practical	JJ
altman	264	7	guidelines	guideline	NNS
altman	264	8	for	for	IN
altman	264	9	building	build	VBG
altman	264	10	ML	ML	NNP
altman	264	11	systems	system	NNS
altman	264	12	with	with	IN
altman	264	13	modular	modular	JJ
altman	264	14	,	,	,
altman	264	15	repeatable	repeatable	JJ
altman	264	16	processes	process	NNS
altman	264	17	and	and	CC
altman	264	18	intelligible	intelligible	JJ
altman	264	19	,	,	,
altman	264	20	verifiable	verifiable	JJ
altman	264	21	results	result	NNS
altman	264	22	.	.	.
altman	265	1	There	there	EX
altman	265	2	are	be	VBP
altman	265	3	many	many	JJ
altman	265	4	resources	resource	NNS
altman	265	5	available	available	JJ
altman	265	6	for	for	IN
altman	265	7	further	further	JJ
altman	265	8	research	research	NN
altman	265	9	,	,	,
altman	265	10	both	both	CC
altman	265	11	online	online	RB
altman	265	12	and	and	CC
altman	265	13	in	in	IN
altman	265	14	your	-PRON-	PRP$
altman	265	15	libraries	library	NNS
altman	265	16	,	,	,
altman	265	17	and	and	CC
altman	265	18	I	-PRON-	PRP
altman	265	19	encourage	encourage	VBP
altman	265	20	you	-PRON-	PRP
altman	265	21	to	to	TO
altman	265	22	consult	consult	VB
altman	265	23	with	with	IN
altman	265	24	subject	subject	JJ
altman	265	25	specialists	specialist	NNS
altman	265	26	,	,	,
altman	265	27	data	data	NN
altman	265	28	scientists	scientist	NNS
altman	265	29	,	,	,
altman	265	30	mathematicians	mathematician	NNS
altman	265	31	,	,	,
altman	265	32	programmers	programmer	NNS
altman	265	33	,	,	,
altman	265	34	and	and	CC
altman	265	35	data	datum	NNS
altman	265	36	engineers	engineer	NNS
altman	265	37	.	.	.
altman	266	1	May	May	MD
altman	266	2	your	-PRON-	PRP$
altman	266	3	data	data	NN
altman	266	4	be	be	VB
altman	266	5	clean	clean	JJ
altman	266	6	,	,	,
altman	266	7	your	-PRON-	PRP$
altman	266	8	computations	computation	NNS
altman	266	9	efficient	efficient	JJ
altman	266	10	,	,	,
altman	266	11	and	and	CC
altman	266	12	your	-PRON-	PRP$
altman	266	13	results	result	NNS
altman	266	14	profound	profound	JJ
altman	266	15	.	.	.
altman	267	1	Further	further	JJ
altman	267	2	reading	reading	NN
altman	267	3	I	-PRON-	PRP
altman	267	4	include	include	VBP
altman	267	5	here	here	RB
altman	267	6	a	a	DT
altman	267	7	few	few	JJ
altman	267	8	suggestions	suggestion	NNS
altman	267	9	for	for	IN
altman	267	10	further	further	JJ
altman	267	11	reading	reading	NN
altman	267	12	on	on	IN
altman	267	13	key	key	JJ
altman	267	14	topics	topic	NNS
altman	267	15	.	.	.
altman	268	1	I	-PRON-	PRP
altman	268	2	have	have	VBP
altman	268	3	also	also	RB
altman	268	4	found	find	VBN
altman	268	5	that	that	IN
altman	268	6	in	in	IN
altman	268	7	the	the	DT
altman	268	8	fast	fast	RB
altman	268	9	-	-	HYPH
altman	268	10	changing	change	VBG
altman	268	11	world	world	NN
altman	268	12	of	of	IN
altman	268	13	machine	machine	NN
altman	268	14	learning	learning	NN
altman	268	15	technologies	technology	NNS
altman	268	16	,	,	,
altman	268	17	blogs	blog	NNS
altman	268	18	,	,	,
altman	268	19	internet	internet	NN
altman	268	20	communities	community	NNS
altman	268	21	,	,	,
altman	268	22	and	and	CC
altman	268	23	online	online	JJ
altman	268	24	classes	class	NNS
altman	268	25	can	can	MD
altman	268	26	be	be	VB
altman	268	27	a	a	DT
altman	268	28	great	great	JJ
altman	268	29	source	source	NN
altman	268	30	of	of	IN
altman	268	31	information	information	NN
altman	268	32	that	that	WDT
altman	268	33	is	be	VBZ
altman	268	34	current	current	JJ
altman	268	35	,	,	,
altman	268	36	introductory	introductory	JJ
altman	268	37	,	,	,
altman	268	38	and/or	and/or	CC
altman	268	39	geared	gear	VBD
altman	268	40	toward	toward	IN
altman	268	41	practitioners	practitioner	NNS
altman	268	42	.	.	.
altman	269	1	Tan	Tan	NNP
altman	269	2	,	,	,
altman	269	3	Pang	Pang	NNP
altman	269	4	-	-	HYPH
altman	269	5	Ning	Ning	NNP
altman	269	6	,	,	,
altman	269	7	Michael	Michael	NNP
altman	269	8	Steinbach	Steinbach	NNP
altman	269	9	,	,	,
altman	269	10	and	and	CC
altman	269	11	Vipin	Vipin	NNP
altman	269	12	Kumar	Kumar	NNP
altman	269	13	.	.	.
altman	270	1	2005	2005	CD
altman	270	2	.	.	.
altman	271	1	Introduction	introduction	NN
altman	271	2	to	to	IN
altman	271	3	Data	Data	NNP
altman	271	4	Mining	Mining	NNP
altman	271	5	.	.	.
altman	272	1	Boston	Boston	NNP
altman	272	2	:	:	:
altman	272	3	Pearson	Pearson	NNP
altman	272	4	Addison	Addison	NNP
altman	272	5	Wesley	Wesley	NNP
altman	272	6	.	.	.
altman	273	1	See	see	VB
altman	273	2	chapter	chapter	NN
altman	273	3	2	2	CD
altman	273	4	for	for	IN
altman	273	5	data	data	NN
altman	273	6	preparation	preparation	NN
altman	273	7	strategies	strategy	NNS
altman	273	8	.	.	.
altman	274	1	Later	later	JJ
altman	274	2	chapters	chapter	NNS
altman	274	3	introduce	introduce	VBP
altman	274	4	common	common	JJ
altman	274	5	classification	classification	NN
altman	274	6	and	and	CC
altman	274	7	clustering	cluster	VBG
altman	274	8	algorithms	algorithm	NNS
altman	274	9	.	.	.
altman	275	1	Marz	Marz	NNP
altman	275	2	,	,	,
altman	275	3	Nathan	Nathan	NNP
altman	275	4	and	and	CC
altman	275	5	James	James	NNP
altman	275	6	Warren	Warren	NNP
altman	275	7	.	.	.
altman	276	1	2015	2015	CD
altman	276	2	.	.	.
altman	277	1	Big	Big	NNP
altman	277	2	Data	Data	NNP
altman	277	3	:	:	:
altman	277	4	Principles	principle	NNS
altman	277	5	and	and	CC
altman	277	6	best	good	JJS
altman	277	7	practices	practice	NNS
altman	277	8	of	of	IN
altman	277	9	scalable	scalable	JJ
altman	277	10	real	real	JJ
altman	277	11	-	-	HYPH
altman	277	12	time	time	NN
altman	277	13	data	datum	NNS
altman	277	14	systems	system	NNS
altman	277	15	.	.	.
altman	278	1	Shelter	Shelter	NNP
altman	278	2	Island	Island	NNP
altman	278	3	:	:	:
altman	278	4	Manning	manning	NN
altman	278	5	.	.	.
altman	279	1	“	"	``
altman	279	2	Part	part	NN
altman	279	3	1	1	CD
altman	279	4	:	:	:
altman	279	5	Batch	Batch	NNP
altman	279	6	Layer	Layer	NNP
altman	279	7	”	"	''
altman	279	8	discusses	discuss	VBZ
altman	279	9	immutable	immutable	JJ
altman	279	10	storage	storage	NN
altman	279	11	in	in	IN
altman	279	12	depth	depth	NN
altman	279	13	.	.	.
altman	280	1	Kleppmann	Kleppmann	NNP
altman	280	2	,	,	,
altman	280	3	Martin	Martin	NNP
altman	280	4	.	.	.
altman	281	1	2017	2017	CD
altman	281	2	.	.	.
altman	282	1	Designing	Designing	NNP
altman	282	2	Data	Data	NNP
altman	282	3	-	-	HYPH
altman	282	4	Intensive	Intensive	NNP
altman	282	5	Applications	application	NNS
altman	282	6	:	:	:
altman	282	7	The	the	DT
altman	282	8	Big	Big	NNP
altman	282	9	Ideas	Ideas	NNP
altman	282	10	Behind	behind	IN
altman	282	11	Reliable	reliable	JJ
altman	282	12	,	,	,
altman	282	13	Scalable	Scalable	NNP
altman	282	14	,	,	,
altman	282	15	and	and	CC
altman	282	16	Maintainable	maintainable	JJ
altman	282	17	Systems	Systems	NNPS
altman	282	18	.	.	.
altman	283	1	Boston	Boston	NNP
altman	283	2	:	:	:
altman	283	3	O’Reilly	o’reilly	RB
altman	283	4	.	.	.
altman	284	1	“	"	``
altman	284	2	Chapter	chapter	NN
altman	284	3	10	10	CD
altman	284	4	:	:	:
altman	284	5	Batch	Batch	NNP
altman	284	6	Processing	Processing	NNP
altman	284	7	”	"	''
altman	284	8	is	be	VBZ
altman	284	9	especially	especially	RB
altman	284	10	relevant	relevant	JJ
altman	284	11	if	if	IN
altman	284	12	you	-PRON-	PRP
altman	284	13	are	be	VBP
altman	284	14	interested	interested	JJ
altman	284	15	in	in	IN
altman	284	16	scaling	scale	VBG
altman	284	17	up	up	RP
altman	284	18	.	.	.