Even if I don't define the scoring parameter, it scores and makes a decision for best estimator, but documentation says the default value for scoring is "None", so what is it using to score when I don't define a metric or list of metrics?
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
2
From the User Guide:
By default, parameter search uses the
scorefunction of the estimator to evaluate a parameter setting. These are thesklearn.metrics.accuracy_scorefor classification andsklearn.metrics.r2_scorefor regression...
answered May 10, 2021 at 15:16
-
$\begingroup$ Thank you, I didn't know they had defaults in function of classificator or regressor, just seeing "score" was driving me mad. Deep dive into he documentation is needed. $\endgroup$ Commented May 11, 2021 at 23:29
-
$\begingroup$ Each estimator's
scoredocstring says what it measures. But classifiers and regressors consistently use accuracy and r2, respectively. $\endgroup$– Ben Reiniger ♦Commented May 12, 2021 at 1:13