Why Cross-Validation
The train-test split lottery, understanding why multiple splits estimate generalization reliably, and the sealed test set protocol.
Cross-Validation Methods
Comparing Leave-One-Out (LOO-CV), Leave-P-Out (LPO-CV), and K-Fold Cross-Validation, their statistical mechanics, and computational tradeoffs.
K-Fold and Stratified K-Fold
Guidelines for choosing K, resolving class imbalance with Stratified K-Fold, and analyzing the complete cross-validation pipeline.
Data Leakage and Best Practices
Preventing data leakage with scikit-learn Pipelines, unbiased hyperparameter tuning with Nested Cross-Validation, and specialized splitting strategies.