How to Transform the Target in Regression

Question 1

WHICH CLASS IN SCIKIT-LEARN IS FOR TRANSFORMING THE TARGET VARIABLE? P341

Answer 1

TransformedTargetRegressor

Question 2

WHAT ARE TWO WAYS WE CAN SCALE THE TARGET VARIABLE? P342

Answer 2

Manually transform the target variable.

Automatically transform the target variable.

Question 3

HOW CAN WE MANUALLY TRANSFORM THE TARGET? P342

Answer 3

Create the transform object, e.g. a MinMaxScaler.
Fit the transform on the training dataset.
Apply the transform to the train and test datasets.
Invert the transform on any predictions made (target_scaler.inverse_transform (y_pred)).

Question 4

WHAT IS THE DOWNSIDE OF MANUALLY TRANSFORMING THE TARGET? P343

Answer 4

You cannot use convenience functions in scikit-learn, such as cross_val_score(), to quickly evaluate a model.

Question 5

HOW DOES TransformedTargetRegressor WORK? P343

Answer 5

TransformedTargetRegressor object wraps a given model (or a pipeline) and a scaling object.
wrapped_model = TransformedTargetRegressor (regressor=model, transformer=MinMaxScaler())

Question 6

HOW CAN WE USE TransformedTargetRegressor? P343

Answer 6

The TransformedTargetRegressor instance can be fit like any other model by calling the fit() function and used to make predictions by calling the predict() function.