In [1]:

    

%matplotlib inline
%config InlineBackend.figure_format='retina'

from __future__ import absolute_import, division, print_function
import matplotlib as mpl
from matplotlib import pyplot as plt
from matplotlib.pyplot import GridSpec
import seaborn as sns
import numpy as np
import pandas as pd
import warnings

from sklearn.datasets import make_classification
from sklearn.ensemble import RandomForestClassifier


    

In [2]:

    

n_features = 20

nonsense_column_names = [''.join(np.random.choice(['ab', 'cd', 'ef', '  ', 'gh'],
                                                  size=np.random.randint(1, 12)))
                                  for _ in range(n_features)]
nonsense_column_names


    

    
Out[2]:





['cd',
 'efefef  gh  ',
 '  efef  cdabcd  gh',
 'ef  efghghcd    ',
 'efababefghef  ',
 'cdcdgh',
 'cd',
 'ghgh',
 'cd',
 '    cdefcd  ghab',
 'ghghgh',
 'cdabefabcdabef',
 'cd',
 'cdcdcd  ghab  abcd',
 'cdabefefefefghgh',
 '    cdgh    cdabef  ',
 'ghefcdcdcd',
 'ef',
 'efghcdef  ef  ghabcd  ',
 'cd']

In [3]:

    

X, y = make_classification(n_features=n_features)
X = pd.DataFrame(data=X, columns=nonsense_column_names)


    

In [4]:

    

clf = RandomForestClassifier(min_samples_leaf=5)
clf.fit(X, y)


    

    
Out[4]:





RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
            max_depth=None, max_features='auto', max_leaf_nodes=None,
            min_impurity_split=1e-07, min_samples_leaf=5,
            min_samples_split=2, min_weight_fraction_leaf=0.0,
            n_estimators=10, n_jobs=1, oob_score=False, random_state=None,
            verbose=0, warm_start=False)

In [5]:

    

clf.feature_importances_


    

    
Out[5]:





array([ 0.00106261,  0.01539515,  0.07398669,  0.00974474,  0.01253549,
        0.02184585,  0.00558645,  0.15462275,  0.00463382,  0.0021489 ,
        0.05179587,  0.        ,  0.        ,  0.01428622,  0.52419094,
        0.01502966,  0.        ,  0.02913962,  0.04103748,  0.02295775])

In [6]:

    

rf_importances = pd.DataFrame({'name':X.columns,
                               'importance':clf.feature_importances_
                              }).sort_values(by='importance', 
                                              ascending=False).reset_index(drop=True)


    

In [7]:

    

rf_importances


    

    
Out[7]:







  

    

      

      
importance
      
name
    
  
  

    

      
0
      
0.524191
      
cdabefefefefghgh
    
    

      
1
      
0.154623
      
ghgh
    
    

      
2
      
0.073987
      
efef  cdabcd  gh
    
    

      
3
      
0.051796
      
ghghgh
    
    

      
4
      
0.041037
      
efghcdef  ef  ghabcd
    
    

      
5
      
0.029140
      
ef
    
    

      
6
      
0.022958
      
cd
    
    

      
7
      
0.021846
      
cdcdgh
    
    

      
8
      
0.015395
      
efefef  gh
    
    

      
9
      
0.015030
      
cdgh    cdabef
    
    

      
10
      
0.014286
      
cdcdcd  ghab  abcd
    
    

      
11
      
0.012535
      
efababefghef
    
    

      
12
      
0.009745
      
ef  efghghcd
    
    

      
13
      
0.005586
      
cd
    
    

      
14
      
0.004634
      
cd
    
    

      
15
      
0.002149
      
cdefcd  ghab
    
    

      
16
      
0.001063
      
cd
    
    

      
17
      
0.000000
      
cdabefabcdabef
    
    

      
18
      
0.000000
      
cd
    
    

      
19
      
0.000000
      
ghefcdcdcd
    
  

In [8]:

    

sns.set_context('poster', font_scale=1.3)


    

In [9]:

    

fig, ax = plt.subplots(figsize=(12, 8))
rf_importances.set_index('name').sort_values(by='importance').plot.barh(ax=ax)
ax.set_xlabel('Feature Importance')
ax.set_ylabel('Feature')
ax.legend(loc='best')
fig.tight_layout();


    

In [11]:

    

fig, ax = plt.subplots(figsize=(12, 8))
rf_importances.set_index('name').plot.bar(ax=ax)
ax.set_xlabel('Feature')
ax.set_ylabel('Feature Importance')
fig.autofmt_xdate()
fig.tight_layout()


    

In [ ]: