Optimizing stacking ensemble by an ant colony optimization approach
Document Type
Conference paper
Source Publication
Genetic and Evolutionary Computation Conference, GECCO'11 - Companion Publication
Publication Date
1-1-2011
First Page
7
Last Page
8
Keywords
aco, ensemble, metaheuristics, stacking
Abstract
An ensemble is a collective decision making system which applies some strategy to combine the predictions of classifiers to generate its prediction on new instances. Stacking is a well-known approach among the ensembles. It is not easy to find a suitable ensemble configuration for a specific dataset. Ant Colony Optimization (ACO) is a popular metaheuristic approach which could be a solution to find configurations. In this work, we propose a new Stacking construction method which applies ACO in the Stacking construction process to generate domain-specific configurations. The experiment results show that the new approach can achieve promising results on 18 datasets compared with some well-known ensemble approaches.
DOI
10.1145/2001858.2001863
ISBN
9781450306904
Publisher Statement
Copyright © 2011 Authors. Access to external full text or publisher's version may require subscription.
Full-text Version
Publisher’s Version
Language
English
Recommended Citation
Chen, Y., & Wong, M. L. (2011). Optimizing stacking ensemble by an ant colony optimization approach. In GECCO '11 Proceedings of the 13th annual conference companion on Genetic and evolutionary computation (pp.7-8). doi: 10.1145/2001858.2001863