JSM Preliminary Online Program
This is the preliminary program for the 2009 Joint Statistical Meetings in Washington, DC.

The views expressed here are those of the individual authors
and not necessarily those of the ASA or its board, officers, or staff.

Back to main JSM 2009 Program page

Activity Number: 521
Type: Contributed
Date/Time: Wednesday, August 5, 2009 : 2:00 PM to 3:50 PM
Sponsor: Section on Statistical Learning and Data Mining
Abstract - #305386
Title: GA-Boost: A Genetic Algorithm for Robust Boosting
Author(s): Brian Gray and Dong-Yop Oh*+
Companies: The University of Alabama and The University of Alabama
Address: Dept of Info Systems, Statistics, and Mgt Science, Tuscaloosa, AL, 35487-0226,
Keywords: AdaBoost ; classification ; decision tree ; predictive modeling ; weak classifier
Abstract:

Boosting is an important tool for predictive modeling. Like other ensemble techniques, boosting has better performance than single decision trees, but it is a black-box model that does not lend itself to easy interpretation. In this paper, we provide examples showing that traditional boosting, as defined by AdaBoost, is non-resistant to outliers, which can lead to overfitting. We propose a new method, GA-Boost, based on a genetic algorithm construction of the boosting ensemble that is more resistant to outliers and results in simpler predictive models than boosting.

  • The address information is for the authors that have a + after their name.
  • Authors who are presenting talks have a * after their name.

Back to the full JSM 2009 program


JSM 2009 For information, contact jsm@amstat.org or phone (888) 231-3473. If you have questions about the Continuing Education program, please contact the Education Department.
Revised September, 2008