VerticaPy

Python API for Vertica Data Science at Scale

Example: Methods in a Binary Classification Model¶

In this example, we use the 'Titanic' dataset to demonstrate the methods available to binary classification models.

In [1]:

from verticapy.datasets import load_titanic
titanic = load_titanic()
display(titanic)

	123 pclass Int	123 survived Int	Abc Varchar(96)	Abc gender Char(1)	123 age Numeric(6,3)	123 sibsp Int	123 parch Int	Abc ticket Varchar(24)	123 fare Numeric(7,4)	Abc cabin Char(10)	Abc embarked Char(1)	Abc boat Char(4)	123 body Int	Abc homedest Varchar(64)	123 split Float
1	1	0		f	2.000	1	2	113781	151.5500	C22 C26	S	[null]	[null]	Montreal, PQ / Chesterville, ON	0.663092836970463
2	1	0		m	30.000	1	2	113781	151.5500	C22 C26	S	[null]	135	Montreal, PQ / Chesterville, ON	0.161721162730828
3	1	0		f	25.000	1	2	113781	151.5500	C22 C26	S	[null]	[null]	Montreal, PQ / Chesterville, ON	0.527904006186873
4	1	0		m	39.000	0	0	112050	0.0000	A36	S	[null]	[null]	Belfast, NI	0.936014187755063
5	1	0		m	71.000	0	0	PC 17609	49.5042	[null]	C	[null]	22	Montevideo, Uruguay	0.40874306159094
6	1	0		m	47.000	1	0	PC 17757	227.5250	C62 C64	C	[null]	124	New York, NY	0.0771260927431285
7	1	0		m	[null]	0	0	PC 17318	25.9250	[null]	S	[null]	[null]	New York, NY	0.360885151894763
8	1	0		m	24.000	0	1	PC 17558	247.5208	B58 B60	C	[null]	[null]	Montreal, PQ	0.78418595273979
9	1	0		m	36.000	0	0	13050	75.2417	C6	C	A	[null]	Winnipeg, MN	0.392075354699045
10	1	0		m	25.000	0	0	13905	26.0000	[null]	C	[null]	148	San Francisco, CA	0.760954634286463
11	1	0		m	45.000	0	0	113784	35.5000	T	S	[null]	[null]	Trenton, NJ	0.083728444064036
12	1	0		m	42.000	0	0	110489	26.5500	D22	S	[null]	[null]	London / Winnipeg, MB	0.187408909434453
13	1	0		m	41.000	0	0	113054	30.5000	A21	S	[null]	[null]	Pomeroy, WA	0.464197367895395
14	1	0		m	48.000	0	0	PC 17591	50.4958	B10	C	[null]	208	Omaha, NE	0.989756429567933
15	1	0		m	[null]	0	0	112379	39.6000	[null]	C	[null]	[null]	Philadelphia, PA	0.197682256577536
16	1	0		m	45.000	0	0	113050	26.5500	B38	S	[null]	[null]	Washington, DC	0.567665293579921
17	1	0		m	[null]	0	0	113798	31.0000	[null]	S	[null]	[null]	[null]	0.221641422947869
18	1	0		m	33.000	0	0	695	5.0000	B51 B53 B5	S	[null]	[null]	New York, NY	0.738475721329451
19	1	0		m	28.000	0	0	113059	47.1000	[null]	S	[null]	[null]	Montevideo, Uruguay	0.366747971624136
20	1	0		m	17.000	0	0	113059	47.1000	[null]	S	[null]	[null]	Montevideo, Uruguay	0.268853009678423
21	1	0		m	49.000	0	0	19924	26.0000	[null]	S	[null]	[null]	Ascot, Berkshire / Rochester, NY	0.513710137922317
22	1	0		m	36.000	1	0	19877	78.8500	C46	S	[null]	172	Little Onn Hall, Staffs	0.453328180825338
23	1	0		m	46.000	1	0	W.E.P. 5734	61.1750	E31	S	[null]	[null]	Amenia, ND	0.169697929639369
24	1	0		m	[null]	0	0	112051	0.0000	[null]	S	[null]	[null]	Liverpool, England / Belfast	0.0851756541524082
25	1	0		m	27.000	1	0	13508	136.7792	C89	C	[null]	[null]	Los Angeles, CA	0.0152691958937794
26	1	0		m	[null]	0	0	110465	52.0000	A14	S	[null]	[null]	Stoughton, MA	0.916533817071468
27	1	0		m	47.000	0	0	5727	25.5875	E58	S	[null]	[null]	Victoria, BC	0.568090511020273
28	1	0		m	37.000	1	1	PC 17756	83.1583	E52	C	[null]	[null]	Lakewood, NJ	0.627309911651537
29	1	0		m	[null]	0	0	113791	26.5500	[null]	S	[null]	[null]	Roachdale, IN	0.747528738109395
30	1	0		m	70.000	1	1	WE/P 5735	71.0000	B22	S	[null]	269	Milwaukee, WI	0.335725951706991
31	1	0		m	39.000	1	0	PC 17599	71.2833	C85	C	[null]	[null]	New York, NY	0.481107967440039
32	1	0		m	31.000	1	0	F.C. 12750	52.0000	B71	S	[null]	[null]	Montreal, PQ	0.0527018732391298
33	1	0		m	50.000	1	0	PC 17761	106.4250	C86	C	[null]	62	Deephaven, MN / Cedar Rapids, IA	0.97885028552264
34	1	0		m	39.000	0	0	PC 17580	29.7000	A18	C	[null]	133	Philadelphia, PA	0.157812902005389
35	1	0		f	36.000	0	0	PC 17531	31.6792	A29	C	[null]	[null]	New York, NY	0.644055349286646
36	1	0		m	[null]	0	0	PC 17483	221.7792	C95	S	[null]	[null]	[null]	0.279342690482736
37	1	0		m	30.000	0	0	113051	27.7500	C111	C	[null]	[null]	New York, NY	0.767035169992596
38	1	0		m	19.000	3	2	19950	263.0000	C23 C25 C2	S	[null]	[null]	Winnipeg, MB	0.490231013158336
39	1	0		m	64.000	1	4	19950	263.0000	C23 C25 C2	S	[null]	[null]	Winnipeg, MB	0.235716909635812
40	1	0		m	[null]	0	0	113778	26.5500	D34	S	[null]	[null]	Westcliff-on-Sea, Essex	0.390792083926499
41	1	0		m	[null]	0	0	112058	0.0000	B102	S	[null]	[null]	[null]	0.71977791050449
42	1	0		m	37.000	1	0	113803	53.1000	C123	S	[null]	[null]	Scituate, MA	0.114295000443235
43	1	0		m	47.000	0	0	111320	38.5000	E63	S	[null]	275	St Anne's-on-Sea, Lancashire	0.831041638739407
44	1	0		m	24.000	0	0	PC 17593	79.2000	B86	C	[null]	[null]	[null]	0.967385483207181
45	1	0		m	71.000	0	0	PC 17754	34.6542	A5	C	[null]	[null]	New York, NY	0.591355679091066
46	1	0		m	38.000	0	1	PC 17582	153.4625	C91	S	[null]	147	Winnipeg, MB	0.870173128321767
47	1	0		m	46.000	0	0	PC 17593	79.2000	B82 B84	C	[null]	[null]	New York, NY	0.99003749457188
48	1	0		m	[null]	0	0	113796	42.4000	[null]	S	[null]	[null]	[null]	0.403090257197618
49	1	0		m	45.000	1	0	36973	83.4750	C83	S	[null]	[null]	New York, NY	0.323395622661337
50	1	0		m	40.000	0	0	112059	0.0000	B94	S	[null]	110	[null]	0.170912061352283
51	1	0		m	55.000	1	1	12749	93.5000	B69	S	[null]	307	Montreal, PQ	0.737730374792591
52	1	0		m	42.000	0	0	113038	42.5000	B11	S	[null]	[null]	London / Middlesex	0.356435600901023
53	1	0		m	[null]	0	0	17463	51.8625	E46	S	[null]	[null]	Brighton, MA	0.61207019072026
54	1	0		m	55.000	0	0	680	50.0000	C39	S	[null]	[null]	London / Birmingham	0.96926672779955
55	1	0		m	42.000	1	0	113789	52.0000	[null]	S	[null]	38	New York, NY	0.536786375567317
56	1	0		m	[null]	0	0	PC 17600	30.6958	[null]	C	14	[null]	New York, NY	0.56807480356656
57	1	0		f	50.000	0	0	PC 17595	28.7125	C49	C	[null]	[null]	Paris, France New York, NY	0.809493057662621
58	1	0		m	46.000	0	0	694	26.0000	[null]	S	[null]	80	Bennington, VT	0.982430178439245
59	1	0		m	50.000	0	0	113044	26.0000	E60	S	[null]	[null]	London	0.696268314495683
60	1	0		m	32.500	0	0	113503	211.5000	C132	C	[null]	45	[null]	0.267956059193239
61	1	0		m	58.000	0	0	11771	29.7000	B37	C	[null]	258	Buffalo, NY	0.533579262206331
62	1	0		m	41.000	1	0	17464	51.8625	D21	S	[null]	[null]	Southington / Noank, CT	0.596509403316304
63	1	0		m	[null]	0	0	113028	26.5500	C124	S	[null]	[null]	Portland, OR	0.430671361740679
64	1	0		m	[null]	0	0	PC 17612	27.7208	[null]	C	[null]	[null]	Chicago, IL	0.355260608484969
65	1	0		m	29.000	0	0	113501	30.0000	D6	S	[null]	126	Springfield, MA	0.0263086690101773
66	1	0		m	30.000	0	0	113801	45.5000	[null]	S	[null]	[null]	London / New York, NY	0.309636837337166
67	1	0		m	30.000	0	0	110469	26.0000	C106	S	[null]	[null]	Brockton, MA	0.0152337534818798
68	1	0		m	19.000	1	0	113773	53.1000	D30	S	[null]	[null]	New York, NY	0.364044237649068
69	1	0		m	46.000	0	0	13050	75.2417	C6	C	[null]	292	Vancouver, BC	0.475262716412544
70	1	0		m	54.000	0	0	17463	51.8625	E46	S	[null]	175	Dorchester, MA	0.30979929654859
71	1	0		m	28.000	1	0	PC 17604	82.1708	[null]	C	[null]	[null]	New York, NY	0.408943912014365
72	1	0		m	65.000	0	0	13509	26.5500	E38	S	[null]	249	East Bridgewater, MA	0.478623170405626
73	1	0		m	44.000	2	0	19928	90.0000	C78	Q	[null]	230	Fond du Lac, WI	0.148798908572644
74	1	0		m	55.000	0	0	113787	30.5000	C30	S	[null]	[null]	Montreal, PQ	0.126385363517329
75	1	0		m	47.000	0	0	113796	42.4000	[null]	S	[null]	[null]	Washington, DC	0.625826764851809
76	1	0		m	37.000	0	1	PC 17596	29.7000	C118	C	[null]	[null]	Brooklyn, NY	0.980584268225357
77	1	0		m	58.000	0	2	35273	113.2750	D48	C	[null]	122	Lexington, MA	0.496911391615868
78	1	0		m	64.000	0	0	693	26.0000	[null]	S	[null]	263	Isle of Wight, England	0.189228599891067
79	1	0		m	65.000	0	1	113509	61.9792	B30	C	[null]	234	Providence, RI	0.280916610732675
80	1	0		m	28.500	0	0	PC 17562	27.7208	D43	C	[null]	189	?Havana, Cuba	0.485971889225766
81	1	0		m	[null]	0	0	112052	0.0000	[null]	S	[null]	[null]	Belfast	0.392282282235101
82	1	0		m	45.500	0	0	113043	28.5000	C124	S	[null]	166	Surbiton Hill, Surrey	0.0708799788262695
83	1	0		m	23.000	0	0	12749	93.5000	B24	S	[null]	[null]	Montreal, PQ	0.089022418949753
84	1	0		m	29.000	1	0	113776	66.6000	C2	S	[null]	[null]	Isleworth, England	0.545445928815752
85	1	0		m	18.000	1	0	PC 17758	108.9000	C65	C	[null]	[null]	Madrid, Spain	0.513165891636163
86	1	0		m	47.000	0	0	110465	52.0000	C110	S	[null]	207	Worcester, MA	0.164120414061472
87	1	0		m	38.000	0	0	19972	0.0000	[null]	S	[null]	[null]	Rotterdam, Netherlands	0.431144091999158
88	1	0		m	22.000	0	0	PC 17760	135.6333	[null]	C	[null]	232	[null]	0.174215304432437
89	1	0		m	[null]	0	0	PC 17757	227.5250	[null]	C	[null]	[null]	[null]	0.641582700423896
90	1	0		m	31.000	0	0	PC 17590	50.4958	A24	S	[null]	[null]	Trenton, NJ	0.814205173403025
91	1	0		m	[null]	0	0	113767	50.0000	A32	S	[null]	[null]	Seattle, WA	0.195884848246351
92	1	0		m	36.000	0	0	13049	40.1250	A10	C	[null]	[null]	Winnipeg, MB	0.653516355436295
93	1	0		m	55.000	1	0	PC 17603	59.4000	[null]	C	[null]	[null]	New York, NY	0.10073209204711
94	1	0		m	33.000	0	0	113790	26.5500	[null]	S	[null]	109	London	0.278607521904632
95	1	0		m	61.000	1	3	PC 17608	262.3750	B57 B59 B6	C	[null]	[null]	Haverford, PA / Cooperstown, NY	0.0629091840237379
96	1	0		m	50.000	1	0	13507	55.9000	E44	S	[null]	[null]	Duluth, MN	0.963238372234628
97	1	0		m	56.000	0	0	113792	26.5500	[null]	S	[null]	[null]	New York, NY	0.462777572218329
98	1	0		m	56.000	0	0	17764	30.6958	A7	C	[null]	[null]	St James, Long Island, NY	0.0639691490214318
99	1	0		m	24.000	1	0	13695	60.0000	C31	S	[null]	[null]	Huntington, WV	0.0569538441486657
100	1	0		m	[null]	0	0	113056	26.0000	A19	S	[null]	[null]	Streatham, Surrey	0.725290278671309

Rows: 1-100 of 1234 | Columns: 15

Let's create a logistic regression to predict the survival of the passengers. We'll use the age and the fare as predictors.

In [2]:

from verticapy.learn.linear_model import LogisticRegression
model = LogisticRegression("public.LR_titanic")
model.fit("public.titanic", ["age", "fare"], "survived")

Out[2]:


=======
details
=======
predictor|coefficient|std_err |z_value |p_value 
---------+-----------+--------+--------+--------
Intercept| -0.09135  | 0.15559|-0.58709| 0.55714
   age   | -0.01439  | 0.00475|-3.02599| 0.00248
  fare   |  0.01546  | 0.00212| 7.29445| 0.00000


==============
regularization
==============
type| lambda 
----+--------
 l2 | 1.00000


===========
call_string
===========
logistic_reg('public.LR_titanic', 'public.titanic', '"survived"', '"age", "fare"'
USING PARAMETERS optimizer='cgd', epsilon=0.0001, max_iterations=100, regularization='l2', lambda=1, alpha=0)

===============
Additional Info
===============
       Name       |Value
------------------+-----
 iteration_count  |  5  
rejected_row_count| 238 
accepted_row_count| 996

Fitting the model creates new model attributes, making methods easier to use.

In [3]:

model.X

Out[3]:

['"age"', '"fare"']

In [4]:

model.y

Out[4]:

'"survived"'

In [5]:

model.input_relation

Out[5]:

'public.titanic'

In [6]:

model.test_relation

Out[6]:

'public.titanic'

Since we didn't write a test relation when fitting the model, the model will use the training relation as the test relation.

Let's compute the accuracy of the model.

In [7]:

model.score(method = "accuracy")

Out[7]:

0.6969205834683955

The 'score' method uses the 'y' attribute and the model prediction in the 'test_relation' to compute the accuracy of the model. You can change these attributes at any time to deploy the models on different columns.

Models have many useful attributes. For example, the 'coef_' attribute gives us the p-value of the model.

In [8]:

model.coef_

	Abc predictor Varchar(65000)	123 coefficient Float	123 std_err Float	123 z_value Float	123 p_value Float
1	Intercept	-0.0913487583375232	0.155594583418985	-0.587094719689177	0.557140093691283
2	age	-0.0143850235204285	0.00475381848744904	-3.02599343210255	0.00247817685818187
3	fare	0.0154603623341147	0.00211946971061135	7.29444835031647	2.99885239324488e-13

Out[8]:

You can view other attributes using the 'get_attr' method.

In [9]:

model.get_attr()

	Abc attr_name Varchar(128)	Abc Long varchar(32000000)	123 #_of_rows Integer
1	details		3
2	regularization		1
3	iteration_count		1
4	rejected_row_count		1
5	accepted_row_count		1
6	call_string		1

Out[9]:

PRC, ROC or lift charts can help you visualize your model.

In [10]:

model.roc_curve()
model.prc_curve()
model.lift_chart()

Let's look at the SQL query for our model.

In [11]:

display(model.deploySQL())

PREDICT_LOGISTIC_REG("age", "fare" USING PARAMETERS model_name = 'public.LR_titanic', type = 'probability', match_by_pos = 'true')

You can evaluate the quality of your model with the 'classification_report' method.

In [12]:

model.classification_report()

	value
auc	0.6974762740166146
prc_auc	0.6003540469187277
accuracy	0.6969205834683955
log_loss	0.281741003041208
precision	0.6194968553459119
recall	0.43777777777777777
f1_score	0.5769062584198693
mcc	0.31193616529653234
informedness	0.2834410430839003
markedness	0.34329598198346645
csi	0.3450087565674256
cutoff	0.999

Out[12]:

You can also add the prediction to your vDataFrame.

In [13]:

model.predict(titanic, name = "pred_survived")

	123 pclass Int	123 survived Int	Abc Varchar(96)	Abc gender Char(1)	123 age Numeric(6,3)	123 sibsp Int	123 parch Int	Abc ticket Varchar(24)	123 fare Numeric(7,4)	Abc cabin Char(10)	Abc embarked Char(1)	Abc boat Char(4)	123 body Int	Abc homedest Varchar(64)	123 split Float	123 pred_survived Float
1	1	0		f	2.000	1	2	113781	151.5500	C22 C26	S	[null]	[null]	Montreal, PQ / Chesterville, ON	0.663092836970463	0.902287093955018
2	1	0		m	30.000	1	2	113781	151.5500	C22 C26	S	[null]	135	Montreal, PQ / Chesterville, ON	0.161721162730828	0.860580339225997
3	1	0		f	25.000	1	2	113781	151.5500	C22 C26	S	[null]	[null]	Montreal, PQ / Chesterville, ON	0.527904006186873	0.868988361039809
4	1	0		m	39.000	0	0	112050	0.0000	A36	S	[null]	[null]	Belfast, NI	0.936014187755063	0.342456861096143
5	1	0		m	71.000	0	0	PC 17609	49.5042	[null]	C	[null]	22	Montevideo, Uruguay	0.40874306159094	0.414029417624538
6	1	0		m	47.000	1	0	PC 17757	227.5250	C62 C64	C	[null]	124	New York, NY	0.0771260927431285	0.939923180247281
7	1	0		m	[null]	0	0	PC 17318	25.9250	[null]	S	[null]	[null]	New York, NY	0.360885151894763	[null]
8	1	0		m	24.000	0	1	PC 17558	247.5208	B58 B60	C	[null]	[null]	Montreal, PQ	0.78418595273979	0.967395967987415
9	1	0		m	36.000	0	0	13050	75.2417	C6	C	A	[null]	Winnipeg, MN	0.392075354699045	0.635075716220126
10	1	0		m	25.000	0	0	13905	26.0000	[null]	C	[null]	148	San Francisco, CA	0.760954634286463	0.487751219756109
11	1	0		m	45.000	0	0	113784	35.5000	T	S	[null]	[null]	Trenton, NJ	0.083728444064036	0.45268401682983
12	1	0		m	42.000	0	0	110489	26.5500	D22	S	[null]	[null]	London / Winnipeg, MB	0.187408909434453	0.429216842625225
13	1	0		m	41.000	0	0	113054	30.5000	A21	S	[null]	[null]	Pomeroy, WA	0.464197367895395	0.447792562926084
14	1	0		m	48.000	0	0	PC 17591	50.4958	B10	C	[null]	208	Omaha, NE	0.989756429567933	0.499713369289623
15	1	0		m	[null]	0	0	112379	39.6000	[null]	C	[null]	[null]	Philadelphia, PA	0.197682256577536	[null]
16	1	0		m	45.000	0	0	113050	26.5500	B38	S	[null]	[null]	Washington, DC	0.567665293579921	0.418678120562726
17	1	0		m	[null]	0	0	113798	31.0000	[null]	S	[null]	[null]	[null]	0.221641422947869	[null]
18	1	0		m	33.000	0	0	695	5.0000	B51 B53 B5	S	[null]	[null]	New York, NY	0.738475721329451	0.380187438266754
19	1	0		m	28.000	0	0	113059	47.1000	[null]	S	[null]	[null]	Montevideo, Uruguay	0.366747971624136	0.558247748355842
20	1	0		m	17.000	0	0	113059	47.1000	[null]	S	[null]	[null]	Montevideo, Uruguay	0.268853009678423	0.59683358559402
21	1	0		m	49.000	0	0	19924	26.0000	[null]	S	[null]	[null]	Ascot, Berkshire / Rochester, NY	0.513710137922317	0.40269570450922
22	1	0		m	36.000	1	0	19877	78.8500	C46	S	[null]	172	Little Onn Hall, Staffs	0.453328180825338	0.647904295062985
23	1	0		m	46.000	1	0	W.E.P. 5734	61.1750	E31	S	[null]	[null]	Amenia, ND	0.169697929639369	0.548033368992097
24	1	0		m	[null]	0	0	112051	0.0000	[null]	S	[null]	[null]	Liverpool, England / Belfast	0.0851756541524082	[null]
25	1	0		m	27.000	1	0	13508	136.7792	C89	C	[null]	[null]	Los Angeles, CA	0.0152691958937794	0.836841368048104
26	1	0		m	[null]	0	0	110465	52.0000	A14	S	[null]	[null]	Stoughton, MA	0.916533817071468	[null]
27	1	0		m	47.000	0	0	5727	25.5875	E58	S	[null]	[null]	Victoria, BC	0.568090511020273	0.408093385463076
28	1	0		m	37.000	1	1	PC 17756	83.1583	E52	C	[null]	[null]	Lakewood, NJ	0.627309911651537	0.659723620113947
29	1	0		m	[null]	0	0	113791	26.5500	[null]	S	[null]	[null]	Roachdale, IN	0.747528738109395	[null]
30	1	0		m	70.000	1	1	WE/P 5735	71.0000	B22	S	[null]	269	Milwaukee, WI	0.335725951706991	0.499846330243495
31	1	0		m	39.000	1	0	PC 17599	71.2833	C85	C	[null]	[null]	New York, NY	0.481107967440039	0.610568134710964
32	1	0		m	31.000	1	0	F.C. 12750	52.0000	B71	S	[null]	[null]	Montreal, PQ	0.0527018732391298	0.566271370083388
33	1	0		m	50.000	1	0	PC 17761	106.4250	C86	C	[null]	62	Deephaven, MN / Cedar Rapids, IA	0.97885028552264	0.69736239132171
34	1	0		m	39.000	0	0	PC 17580	29.7000	A18	C	[null]	133	Philadelphia, PA	0.157812902005389	0.451851682120462
35	1	0		f	36.000	0	0	PC 17531