How to compensate for missing features in some samples in Logistic regression?

21 Views Asked by At

I am trying to fit a logistic regression function for data with some rows that have missing features values. (E.g Row one contain information for 7 features, and in row two feature number 6 doesn't contain any value) I am not allowed to delete any information. and not allowed to fill the missing values with any technique. How can I train the model to deal with such a case? So the equation will know to compensate for such cases. Also, different scenarios, if I already have the logistic equation and I want to use it on some test set with missing features values, is that possible?