I need to select the most important features from my data frame before starting with nearest neighbours problem.

Which methods are the best to do this? My data frame has around 8 categorical features and 2 continuous features but no target variable. The problem is that I have three categorical features which can only be one hot encoded and once I do that it explodes the data into 47 OHE variables.

So considering these situations, what would be the best method to do the feature selection?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.