Feature Importance

import pandas as pd
import data_describe as dd

from sklearn.naive_bayes import GaussianNB

import warnings
warnings.simplefilter("ignore")

from sklearn.datasets import load_breast_cancer
data = load_breast_cancer()
df = pd.DataFrame(data.data, columns=list(data.feature_names))
df['target'] = data.target
df.head(1)

	mean radius	mean texture	mean perimeter	mean area	mean smoothness	mean compactness	mean concavity	mean concave points	mean symmetry	mean fractal dimension	...	worst texture	worst perimeter	worst area	worst smoothness	worst compactness	worst concavity	worst concave points	worst symmetry	worst fractal dimension	target
0	17.99	10.38	122.8	1001.0	0.1184	0.2776	0.3001	0.1471	0.2419	0.07871	...	17.33	184.6	2019.0	0.1622	0.6656	0.7119	0.2654	0.4601	0.1189	0

1 rows × 31 columns

Default

Pass in the data frame and the column name of the response variable

dd.importance(df, 'target')

Text(0.5, 1.0, 'Feature Importance')

Show negative importance

Use truncate=False to disable removal of negative importance values.

dd.importance(df, 'target', truncate=False)

Text(0.5, 1.0, 'Feature Importance')

Getting the data values

You can get the importance values by setting return_values to True.

dd.importance(df, 'target', return_values=True)

array([0.00035398, 0.00561404, 0.00106195, 0.        , 0.00069865,
       0.00035709, 0.00033225, 0.00386586, 0.        , 0.        ,
       0.00105884, 0.00176681, 0.00105263, 0.00247477, 0.        ,
       0.        , 0.        , 0.        , 0.        , 0.        ,
       0.00738705, 0.00597423, 0.        , 0.00386896, 0.00069554,
       0.00034467, 0.        , 0.00525074, 0.        , 0.        ])

Alternate Model

You can also specify a different model type using the estimator argument.

dd.importance(df, 'target', estimator=GaussianNB())

Text(0.5, 1.0, 'Feature Importance')