如何在 logit statsmodel python 中包含交互变量？

Question

我正在研究 Logistic 回归模型，我正在使用 statsmodels api 的 logit。我无法弄清楚如何向模型提供交互项。

Answer 1

可以使用公式界面，在公式里面使用冒号:，例如：

import statsmodels.api as sm
import statsmodels.formula.api as smf
import numpy as np
import pandas

np.random.seed(111)
df = pd.DataFrame(np.random.binomial(1,0.5,(50,3)),columns=['x1','x2','y'])

res1 = smf.logit(formula='y ~ x1 + x2 + x1:x2', data=df).fit()

res1.summary()
                           Logit Regression Results                           
==============================================================================
Dep. Variable:                      y   No. Observations:                   50
Model:                          Logit   Df Residuals:                       46
Method:                           MLE   Df Model:                            3
Date:                Thu, 04 Feb 2021   Pseudo R-squ.:                 0.02229
Time:                        10:03:59   Log-Likelihood:                -32.463
converged:                       True   LL-Null:                       -33.203
Covariance Type:            nonrobust   LLR p-value:                    0.6869
==============================================================================
                 coef    std err          z      P>|z|      [0.025      0.975]
------------------------------------------------------------------------------
Intercept     -0.9808      0.677     -1.449      0.147      -2.308       0.346
x1             0.4700      0.851      0.552      0.581      -1.199       2.139
x2             0.9808      0.863      1.137      0.256      -0.710       2.671
x1:x2         -1.1632      1.229     -0.946      0.344      -3.572       1.246
==============================================================================

如何在 logit statsmodel python 中包含交互变量？

How to include interaction variables in logit statsmodel python?

python

statsmodels

logistic-regression