View Item 

Probabilistic approach to the dynamic ensemble selection using measures of competence and diversity of base classifiers

Access Status
Fulltext not available
Authors
Date
2011
Type
Conference Paper

Metadata
Show full item record
Citation
Lysiak, R. and Kurzynski, M. and Woloszynski, T. 2011. Probabilistic approach to the dynamic ensemble selection using measures of competence and diversity of base classifiers, in Corchado, E. and Kurzynski, M. and Wozniak, M. (ed), Proceedings of the 6th International Conference on Hybrid Artificial Intelligence Systems (HAIS 2011), May 23-25 2011, pp. 229-236. Wroclaw: Springer.
Source Title
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
DOI
ISBN
School
Department of Mechanical Engineering
URI
Collection
Abstract

In the paper measures of classifier competence and diversity using a probabilistic model are proposed. The multiple classifier system (MCS) based on dynamic ensemble selection scheme was constructed using both measures developed. The performance of proposed MCS was compared against three multiple classifier systems using six databases taken from the UCI Machine Learning Repository and the StatLib statistical dataset. The experimental results clearly show the effectiveness of the proposed dynamic selection methods regardless of the ensemble type used (homogeneous or heterogeneous). © 2011 Springer-Verlag.