The goal of ensemble methods is to combine the predictions of several base estimators built with a given learning algorithm in order to improve generalizability

Every estimator has its advantages and drawbacks. Its generalization error can be decomposed in terms of bias, variance and noise. The