Evaluation Metrics

Classification

Metric	What it measures	When to use
Accuracy	% correct	Balanced classes only
Precision	Of predicted positives, how many are correct	When false positives are costly (spam filter)
Recall	Of actual positives, how many were found	When false negatives are costly (disease screening)
F1	Harmonic mean of precision & recall	Imbalanced classes, need balance
AUC-ROC	Area under ROC curve	Comparing models, threshold-independent

Confusion matrix

                Predicted
              Pos    Neg
Actual  Pos   TP     FN
        Neg   FP     TN

Precision = TP / (TP + FP)
Recall    = TP / (TP + FN)
F1        = 2 × Precision × Recall / (Precision + Recall)

from sklearn.metrics import classification_report, confusion_matrix
 
print(classification_report(y_true, y_pred))
print(confusion_matrix(y_true, y_pred))

Regression

Metric	What it measures	Notes
MSE	Average squared error	Penalizes large errors
RMSE	√MSE	Same units as target
MAE	Average absolute error	Robust to outliers
R²	Proportion of variance explained	1.0 = perfect, 0 = predicts mean

Links

• Loss Functions
• Cross-Validation
• Bias-Variance Tradeoff