Day 3 - Regression Evaluation Metrics: RMSE, MAE, MAPE, R² and Residual Analysis

Where classification asks "right or wrong?", regression asks "how far off?" Residual (prediction error) is raw material for all metrics. Today covers RMSE, MAE, MAPE, R² computation, character, meaning, plus residual plots revealing model flaws hiding in single scores.

Residual: Starting Point

Per sample i, residual is the gap:

residual_i = y_i (truth) - ŷ_i (prediction)

All regression metrics compress these gaps into one number. Difference: how errors are handled — squared?, absolute?, ratio?

MAE: Mean Absolute Error

MAE = (1/n) Σ |y_i - ŷ_i|

Absolute error average. Traits: