Use a single, arbitrary learning algorithm but manipulate training data to make it learn multiple models. Data1 Data2 … Data m; Learner1 = Learner2 = …
