Predicting probabilities in CatBoost regressor

Question

Does CatBoost regressor have a method to predict the probabilities of each prediction? I see one for CatBoost classifier (https://catboost.ai/en/docs/concepts/python-reference_catboostclassifier_predict_proba) but not for regressor.

Answer 1

There is no predict_proba method in the Catboost regressor, but you can specify the output type when you call predict on the trained model.

model = CatBoostRegressor(loss_function='RMSE', silent=True)
model.fit(x_train, y_train)

y_pred = model.predict(x_test, prediction_type='Probability')