Exercise. KFold and scoring

Module 7. Evaluating model performance Regression metrics
#1

Hi,

In the exercise it is said “Use a KFold cross-validation with 10 folds” so in the beginning I did not use cv=10 but really tried to use KFold(n_splits=10) but I had an error.
But I think my error is rather linked with my second question : for scoring, I first used 'accuracy' and in the previous exercise I had no problem doing that but here I have the following error :
…in _check_targets

raise ValueError("{0} is not supported".format(y_type))
ValueError: continuous is not supported

And I don’t really understand why I have bug here whereas previously If I remember well I had no problem passing ‘accuracy’ as an argument.

Thank you very much,

Geoffrey

#2

It looks like if you are trying to compute the accuracy (a classification metric) on a continuous target (regression problem).

The message indicates you that you should use a regression metric with this type of target.