Which statement is true:

Group of answer choices

Feature normalization is needed when different features have different scales but we are using a model that features should be on the same scale

Feature normalization would increase variability in the data.

We can use different scalers to transform training and test data set if we have already split the data.

When normalizing features using a scaler, we fit the scaler on the training data and then use the fit scaler to transform the test data but not the training data.