In Section 3, Unit 5 the input features are scaled but not the output feature. Shouldn't the output feature also be scaled and the predicted valued can be found by doing a prediction and unscaling the values?
Hi Jan,
You are correct scaling the target is another approach. You can scale the target variable first and then train using it. Then predict the target variable and finally multiply with the scaling factor. For a regression problem, this above method can be used. Do compare the test results of the scaled and unscaled models to compare whichever has the better R2 score and use that.