Scoring used for the test score

M6. Ensemble of models M6. Introductory example to ensemble models
#1

Hi,

What is the default metric used in the scoring of cross_validate for regression problems ?
I understand the scoring with a negative mean absolute error and a negative mean squared error.
But in the example of california housing, the score seems to be between 0 and 1 but I don’t understand how it is computed and what it represents.
In a classification problem, the default scoring can be understood as a probability of success but what for regression problems ?

Thank you !

#2

cross_validate uses the default scoring strategy for a given estimator. For classifcation it is accuracy and for regression is the coefficient of determination (R^2).

You can find the info of the default scoring strategy in the documentation for the score method, e.g.

1 Like
#3

Thank you, it is clear for me.