# When using GridSearchCV with regression tree how to interpret mean_test_score?

I am using GridSearchCV to tune hyperparameters of regression decision tree. When I do, I get mean_test_score but I thought it would return mean MSE since it is a regressor. how to interpret mean_test_score? Is there a way to tweak GridSearchCV so it returns mean MSE?

here is my code

tree_reg = GridSearchCV(DecisionTreeRegressor(criterion="mse"), {
"min_samples_split":[2,3,4],
"min_samples_leaf":[1,2,3]
}, cv=5, return_train_score=False)

tree_reg.fit(X, y)

pd.DataFrame(tree_reg.cv_results_)
>>> params                   split0_test_score  .... mean_test_score
{"min_samples_leaf":2,        0.998782             0.9989933
"min_samples_split":3}

{"min_samples_leaf":2,        0.998823             0.998930
"min_samples_split":4}
...


what does mean_test_score mean?

By default, GridSearchCV uses the score method of its estimator; see the last paragraph of the scoring parameter on the docs:
And DecisionTreeRegressor.score (indeed, all/most regressors) uses R^2.
In response to your edit: you can specify scoring='neg_mean_squared_error'. But note too that there's a linear relationship between MSE and R^2, so optimizing either of these are equivalent.