Both of these practices help to remove unnecessary variables in the training set to help the model perform better. Most importantly, it helps to pick a machine learning model that will work better ...