Chapter : Lesson 3

Episode 9 - Evaluation Metrics for Regression

face Josiah Wang

Summary:

  • Common metrics for regression:
    • Mean Squared Error (MSE) - penalises large errors more than smaller errors
    • Mean Absolute Error (MAE) - less sensitive to outliers
    • Root Mean Squared Error (RMSE) - more interpretable, less sensitive to outliers than MSE. Most common and preferred.