How do you normalise the train+validation sets together?

Question

This question is somewhat related to: Is it correct to join training and validation set before inferring on test-set?

As far as I understand, normalisation in general is done in the following way:

1. Split the data
2. Normalise train data
3. Use mean+std from 2 for normalising validation and test data
4. Train model and tune hyperparameters

Now we have a model we are happy with and hyperparams. Above question suggests that it's good then train a model using the train+validation data together. But I am confused, do I leave them normalised as they are and combine them together? Or do I calculate a fresh new normalisation on the combined sets and then recalculate test data with the new normalisation values?

Thank you!

Answer 1

If you are going for the approach with train, validation and test sets and want to train the model using both train and validation sets than yes, it would be appropriate to perform new normalization on the data in its raw form. Than just like before, you use the mean and standard deviation to normalize test set data.