| |
Abstract:
Perceptron Decision Trees (also known as Linear Machine DTs,
etc.) are analysed in order that data dependent Structural Risk
Minimization cann be applied. Data dependent analysis is performed
which indicates that choosing the maximal margin hyperplanes at the
decision nodes will improve the generalization. The analysis is
performed by defining a real valued function class computed by
decision trees and bounding its generalization error using a novel
technique. Experiments are performed to test the approach.
|
