Add implementations for all basic supervised learning algorithms in scikit-learn. I'd like to not have to implement different functions for everything that operates almost identically, but I'm not sure how well that works with the generic function approach.