ensemble methods (e.g., bagging, boosting, stacking)

Use multiple learning algorithms to obtain better predictive performance than could be obtained from any of the constituent learning algorithms.