In logistic regression model in Python, how to treat multicollinearity

149 Views Asked by At

In logistic regression model in Python, how to treat multicollinearity in Python in logistic regression model. What do we do when there are high correlated variables from a correlation matrix. Suggest me python steps to reduce multicollinearity

1

There are 1 best solutions below

0
HsergeyF On
  1. You can try regularization. It will add an additional constraint on the weight vector and you can handle multicollinearity. This limit can be set in different ways, but, as a rule, you don't need something more complicated, than L1 or L2 regularization.
  2. Try to reduce dimensions of your df with Factor Analysis. As a result you will get a dataframe of not-correlated factors, wich generally describe the full set of your features.