Nested cross-validation and groups

M3. Hyperparameter tuning M3. Evaluation and hyperparameter tuning
#1

Hi, thank you very much for this lesson on nested cross-validation. It is still a new concept for me so I wanted to ask the question on how it would work when the data is split into groups.

For example, in the case where the data consists of different individuals with multiple recordings, and it is important to not have recordings of each individual in both the train and validation set, how would nested cross-validation be implemented?

I apologize in advance if my question is not very clear. I will try to rephrase it if necessary.

Thank you once again for this wonderful mooc!

#2

This is relatively covered in the Sample grouping notebook in Module 7. The only difference is is that there we are not tuning any parameter of the LogisticRegression, which could be done using sklearn.model_selection.GroupKFold for both the inner and outer loops of the nested cross-validation.