The common practice for Hyper-parameter tuning so to split the data into 3. The first set everyone agrees on is called the Training Set. The remaining two can have their names switched round but typically referred to as Test Set and Validation Set

In any case, one of the sets (in my example the Test Set) is used to evaluate the model when tuning Hyper-parameters. Once satisfied, the final set of data (Validation in my example) can then be used to independently evaluate the optimised model.

Love podcasts or audiobooks? Learn on the go with our new app.