Evaluation metrics

A trained model needs a report card: MAE, RMSE and R² each tell a slightly different story.

• MAE — average error size, in the target’s units. Easy to explain.
• RMSE — like MAE but squares first, so large errors dominate.
• R² — fraction of variance explained, from 0 toward 1 (perfect).

Always report metrics on the TEST set — grading on training data is grading answers the model already saw.