Why a Single Train/Test Split Can Mislead Your ML Model

A single 80/20 train/test split can produce a 91% accuracy that is due to luck, not skill. K-fold cross-validation splits data into k folds, trains on k-1 folds and validates on the held-out fold k times, yielding a stable mean ± std score. This technique ensures every data point is used for both training and validation, reducing variance and providing an honest performance estimate.