Internship← Artificial Intelligence & Machine Learning learning pathArtificial Intelligence & Machine Learning · Module 6 of 10
Model evaluation
Measure model quality correctly.
📖 Key ideas
- Confusion matrix: TP, FP, TN, FN.
- Accuracy, precision, recall, F1 score.
- ROC curve and AUC for ranking quality.
- Cross-validation gives reliable estimates.
- Overfitting: great on training, poor on test.
💡 Real-life example: Evaluating a model is grading an exam: not just total marks, but which questions were wrong.
PrecisionCorrect positives among predicted positives
RecallFound positives among actual positives
F1 scoreBalance of precision and recall
Cross-validationRotating train/test folds