Random forest regressor has a negative score

Asked Nov 22 '18 at 17:12

Active Nov 22 '18 at 20:06

Viewed 480 times

I am using a RandomForestRegressor. When I check the score for the model with the training data, it's

Rregressor.score(X_train,y_train) 0.8357837327169805

but when I check the score using the test set, it's negative

Rregressor.score(X_validate,y_validate) -0.5022815681820925

Is this overfitting? Is this the right way to check how good a model is? What are the ways to improve the performance?

edited Nov 22 '18 at 20:06

asked Nov 22 '18 at 17:12

AnoojNair

It can be overfitting or a mistake in data preparation logic. – Michael M Nov 22 '18 at 17:15

0 Answers0