Date of Award

2002

Document Type

Thesis (Undergraduate)

Department

Department of Computer Science

First Advisor

Javed A. Aslam

Abstract

Boosting is a means of using weak learners as subroutines to produce a strong learner with markedly better accuracy. Recent results showing the connection between logistic regression and boosting provide the foundation for an information-theoretic analysis of boosting. We describe the analogy between boosting and gambling, which allows us to derive a new upper bound on training error. This upper bound explicitly describes the effect of noisy data on training error. We also use information-theoretic techniques to derive an alternative upper-bound on testing error which is independent of the size of the weak-learner space.

Original Citation

Listed in the Dartmouth College Computer Science Technical Report Series as TR2002-428.

Share

COinS