AAIA: Advanced in AI Audit

Time series cross-validation preserves the chronological order of monthly observations by evaluating the model on later periods than those used for training. This detects changes in the data distribution over time and prevents future data from being mixed into training folds, which could conceal data drift.