I'm getting a "numpy.linalg.linalg.LinAlgError: Singular matrix" error when fitting a logistic regression model.
I know its coming from feature X which is 0 for all samples and when I'm looking at X_train.corr() its NAN.
I can't exclude this feature and I'm using statsmodels because I need cov_params() which sklearn does not have (which is why I cant uses method = bfgs also)
Correlation matrix:
