查看sklearn的线性回归系数的p value

可以用statsmodel.api模块。比如说逻辑回归

>>> import statsmodels.api as sm
>>> log_reg = sm.Logit(df['y'], df[['a', 'b', 'c']])
>>> clf = log_reg.fit()
Optimization terminated successfully.
         Current function value: 0.692762
         Iterations 3

如果要查看系数的p value，调用clf.summary()就可以了，这个方法对OLS回归也是一样的。

>>> clf.summary()
Logit Regression Results
Dep. Variable:  y   No. Observations:   1000
Model:  Logit   Df Residuals:   997
Method: MLE Df Model:   2
Date:   Tue, 17 Oct 2017    Pseudo R-squ.:  -0.1341
Time:   10:19:37    Log-Likelihood: -692.76
converged:  True    LL-Null:    -610.86
                LLR p-value:    1.000
     coef    std err z   P>|z|   [95.0% Conf. Int.]
a   0.1240  0.219   0.565   0.572   -0.306 0.554
b   -1.9466 6.245   -0.312  0.755   -14.186 10.293
c   3.7477  6.230   0.602   0.547   -8.463 15.958

sklearn不支持查看系数的p-value，可以参考statsmodels模块。