An Empirical Investigation of Brute Force to choose Features, Smoothers and Function Approximators

Abstract

The generalization error of a function approximator, feature set or smoother can be estimated directly by the leave-one-out cross-validation error. For memory-based methods, this is computationally feasible. We describe an initial version of a general memory-based learning system (GMBL): a large collection of learners brought into a widely applicable machine-learning family. We present ongoing investigations into search algorithms which, given a dataset, nd the family members and features that generalize best. We also describe GMBL's application to two noisy, di cult problems|predicting car engine emissions from pressure waves, and controlling a robot billiards player with redundant state variables.