Final Technical Report for Hybrid Algorithms and Oblivious Decision Graphs Using MLC

Abstract : MLC++ is a Machine Learning Library of C++ classes. Over 350 different sites have copied the MLC++ kit, and machine learning research in the robotics lab at Stanford is enhanced through the use of the library. The main topics are: (1) Hybrid decision tree and nearest-neighbor, (2) Stacking and Bagging, and (3) Oblivious decision graphs.