Hello,
I don’t understand this warning:
“Caution!
Be aware that we use train_test_split here for didactic purposes, to show the scikit-learn API.”
In “real life” what would you have used? because in the processing the result of this didactic purpose is really used!
In really like, one should use cross-validation as previously demonstrated. Without cross-validation, you will not be aware of the uncertainty of the statistical performance of a model.
the uncertainty of your estimation of the statistical performance of a model.
Thank you, the link with the cross-validation had escaped me.
I am tagging this for v3 since this we should probably mention that cross-validation may be preferred