ML Model Performance Visualizer

Data is randomly generated to demonstrate various regression and classification metrics. Try dragging the points!

Regression Metrics

• Mean Absolute Error (MAE): average lengths of the residual segments

• Mean Squared Error (MSE): average area of the squares formed by the residuals

• Concordance Index (C-index): the agreement in the ordering of predicted and actual values

Classification Metrics (i.e., for binary outcomes 0/1)

• Brier Score: average squared difference between predicted probabilities and actual outcomes

• Expected Calibration Error (ECE): the difference between predicted probabilities and observed frequencies

• Area Under the ROC Curve (AUC): the agreement in the ordering of predicted probabilities for positive and negative classes