Bagging, or Bootstrap Aggregating, is an Ensemble Learning technique in Machine Learning. It builds multiple models from resampled data subsets, then combines their predictions to reduce variance and improve overall accuracy, effectively combatting Overfitting.