Stacking Stacked Generalization

Data Leakage

Critical problem in stacking if predictions on the training set are generated by models already trained on these same data, leading to overestimation of performance.

← Kembali