Faculty Advisor

Nandram, Balgobin

Abstract

A proposed new ensemble method for classification is introduced. The goal of the new method is to reduce CPU time while producing more accurate classification. This is achieved by using a smaller re-sampling rate compared to bagging, which is an existing ensemble method. Two variations of the new method are explored, and are compared to bagging. The accuracy of the two methods are comparable to bagging while the CPU time of the new ensemble procedure is reduced significantly.

Publisher

Worcester Polytechnic Institute

Date Accepted

January 2003

Major

Actuarial Mathematics

Project Type

Major Qualifying Project

Accessibility

Restricted-WPI community only

Advisor Department

Mathematical Sciences

Share

COinS