Also, using normalized data allows you to interpret the model weights more easily (larger magnitudes mean more effect on the predicted y value). Linear regression with interactions is ... However, for ...