同时调用SKLearn的cross_val_score和cross_val_predict？

Question

I am running a nested-cross validation with cross_val_score by passing a GridSearchCV object to it. Then I follow up with cross_val_predict to get model predictions for graphing. Like this:

gs = GridSearchCV(mymodel, myparams)
score = cross_val_score(gridsearch, X_train, y_train)
prediction = cross_val_predict(gs, X_train, y_train)

This seems computationally redundant; is there a way to get the cross-validated predictions from cross_val_score , or do I need to manually iterate through the folds of a CV object to do this in one step?

Answer 1

From what I am seeing in the documentation it appears that cross_val_score and cross_val_predict obtain their values by different processes and it is not recommended for the two to be combined.

https://scikit-learn.org/stable/modules/cross_validation.html#obtaining-predictions-by-cross-validation