Python 中系数的置信区间？

Question

In R, you can get confidence intervals for each coefficient in a Logistic Regression as shown here ( https://www.r-bloggers.com/example-9-14-confidence-intervals-for-logistic-regression-models/ ).

Can you do this in sci-kit learn in Python? I was exploring, but I couldn't find a way.

Answer 1

I don't think you can get that from sci-kit learn, one option is to use statsmodels in python, which is very similar to R:

import statsmodels.api as sm
import pandas as pd

df = pd.read_csv("http://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data",
                 header=None,names=["s_wid","s_len","p_wid","p_len","species"])

y = np.array(df['species'] == "Iris-virginica").astype(int)
X = sm.add_constant(df.iloc[:,:4])
model = sm.Logit(y, X)
result = model.fit()

result.summary()

Logit Regression Results
Dep. Variable:  y   No. Observations:   150
Model:  Logit   Df Residuals:   145
Method: MLE Df Model:   4
Date:   Wed, 17 Jun 2020    Pseudo R-squ.:  0.9377
Time:   00:25:21    Log-Likelihood: -5.9493
converged:  True    LL-Null:    -95.477
Covariance Type:    nonrobust   LLR p-value:    1.189e-37
coef    std err z   P>|z|   [0.025  0.975]
const   -42.6378    25.708  -1.659  0.097   -93.024 7.748
s_wid   -2.4652 2.394   -1.030  0.303   -7.158  2.228
s_len   -6.6809 4.480   -1.491  0.136   -15.461 2.099
p_wid   9.4294  4.737   1.990   0.047   0.145   18.714
p_len   18.2861 9.743   1.877   0.061   -0.809  37.381