Automatically deduce scaling from trained model

[koenderks]

I removed the 'scale variables' option from the prediction analysis and now deduce it from the trained model:

• If the features in the trained model were not scaled, the features for the new data will not be scaled either
• If the features in the trained model were scaled, the features for the new data will now be scaled in the same way (with the mean and sd of the original dataset)

I did this because I could not think of a scenario in which the user would scale the features in the prediction analysis if they were unscaled in training, or where the user would leave the features unscaled in the prediction while they were scaled in training. Removing the option entirely and deducing it from the trained model seemed more user-friendly and avoids confusion.

Issues:

• Closes Does it make sense to scale the data in the prediction set with separate mean and sd as the train/test set? #413