CODEWITHSUNDEEP
pythonmatplotlibscikit-learnlogistic-regression
Claudio
Claudio

Reputation: 63

Regression summary in Python

I am quite new to Python. I would like to get a summary of a logistic regression like in R. I have created variables x_train and y_train and I am trying to get a logistic regression

import numpy as np
import matplotlib.pyplot as plt
from sklearn import linear_model

clf = linear_model.LogisticRegression(C=1e5)
clf.fit(x_train, y_train)

What I get is:

LogisticRegression(C=100000.0, class_weight=None, dual=False,
    fit_intercept=True, intercept_scaling=1, max_iter=100,
    multi_class='ovr', n_jobs=1, penalty='l2', random_state=None,
    solver='liblinear', tol=0.0001, verbose=0, warm_start=False)

I would like to have a summary with significative levels, R2 ecc.

Upvotes: 3

Views: 30793

Answers (4)

Bharat Vadlamudi
Bharat Vadlamudi

Reputation: 78

import statsmodels.api as sm      
x_train1 = sm.add_constant(x_train1)
lm_1 = sm.OLS(y_train, x_train1).fit()
lm_1.summary()

This is a very use full package for the once who are very much used to R's model summary

For more info refer below articles:

  1. statsmodels.api
  2. stats-models-vs-sklearn

Upvotes: 1

Nick Becker
Nick Becker

Reputation: 4214

I'd recommend taking a look at the statsmodels library. Sk-learn is great (and the other answers provide ways to get at R2 and other metrics), but statsmodels provides a regression summary very similar to the one you're probably used to in R.

As an example:

import statsmodels.api as sm
from sklearn.datasets import make_blobs

x, y = make_blobs(n_samples=50, n_features=2, cluster_std=5.0,
                  centers=[(0,0), (2,2)], shuffle=False, random_state=12)

logit_model = sm.Logit(y, sm.add_constant(x)).fit()
print logit_model.summary()

Optimization terminated successfully.
         Current function value: 0.620237
         Iterations 5
                           Logit Regression Results                           
==============================================================================
Dep. Variable:                      y   No. Observations:                   50
Model:                          Logit   Df Residuals:                       47
Method:                           MLE   Df Model:                            2
Date:                Wed, 28 Dec 2016   Pseudo R-squ.:                  0.1052
Time:                        12:58:10   Log-Likelihood:                -31.012
converged:                       True   LL-Null:                       -34.657
                                        LLR p-value:                   0.02611
==============================================================================
                 coef    std err          z      P>|z|      [95.0% Conf. Int.]
------------------------------------------------------------------------------
const         -0.0813      0.308     -0.264      0.792        -0.684     0.522
x1             0.1230      0.065      1.888      0.059        -0.005     0.251
x2             0.1104      0.060      1.827      0.068        -0.008     0.229
==============================================================================

If you want to add regularization, instead of calling .fit() after the Logit initialization you can call .fit_regularized() and pass in an alpha parameter (regularization strength). If you do this, remember that the C paramater in sk-learn is actually the inverse of regularization strength.

Upvotes: 7

Dmitry
Dmitry

Reputation: 2096

  1. For obtaining of singificance levels you can use sklearn.feature_selection.f_regression.

  2. For obtaining R2 you can use sklearn.metrics.r2_score

Upvotes: 1

Maximilian Peters
Maximilian Peters

Reputation: 31659

You can call clf.score(test_samples, true_values) to get R2.

Significance is not directly provided by sklearn but have at the answer here and this code.

Upvotes: 0

Related Questions