Topic: Prediction Metrics and Calibration

Mean Absolute Error in Outcome Units

Mean Prediction Error and Systematic Bias

Metric Choice Under Asymmetric Prediction Costs

Interpreting Negative Out-of-Sample R-Squared

How Outcome Transformations Change Error Metrics

Choosing Between MAE and RMSE

Choosing Bins for Calibration Plots

Calibration as Agreement Between Predictions and Outcomes

Calibration Intercept and Slope

Baseline Predictions and Skill Scores