Measure Performance on Test Set

In this section, you will transform the test set using your full pipeline and measure the performance of your best model on the test set. Follow the steps below, using the specified variable names for automatic grading through CodeGrade.

Based on all previous cross$-$validation results, pick your best model.

Use the previously created column transformers to transform the test set, both with and without the G$1/$G$2$ columns.

Using your best model, measure its performance on the test set to estimate the generalization error.

Instructions for Submission

Fit Best Model: If you haven't already, fit your best model to both sets of your transformed training data.

Transform the Test Set: Use your column transformers to transform the test set $($X$\_$test$),$ both with and without the G$1/$G$2$ columns. Name these transformed datasets X$\_$test$\_$transformed$\_$with$\_$grades and X$\_$test$\_$transformed$\_$without$\_$grades.

Evaluate Performance: Measure the performance of your best$-$fitted models on the transformed test sets using Root Mean Squared Error $($RMSE$)$ and R$-$squared $($R$^2)$ metrics. Save these variables as:

rmse$\_$with$\_$grades

r$2\_$with$\_$grades

rmse$\_$without$\_$grades

r$2\_$without$\_$grades