An entropy-based diversity measure for classifier combining and its application to face classifier ensemble thinning

Authors:
Wenyao Liu;Zhaohui Wu;Gang Pan
Affiliations:
Department of Computer Science and Engineering, Zhejiang University, Hangzhou, P.R China;Department of Computer Science and Engineering, Zhejiang University, Hangzhou, P.R China;Department of Computer Science and Engineering, Zhejiang University, Hangzhou, P.R China
Venue:
SINOBIOMETRICS'04 Proceedings of the 5th Chinese conference on Advances in Biometric Person Authentication
Year:
2004

Citing 5
Cited 2

On Combining Classifiers

IEEE Transactions on Pattern Analysis and Machine Intelligence
An Experimental Comparison of Three Methods for Constructing Ensembles of Decision Trees: Bagging, Boosting, and Randomization

Machine Learning
Person Identification Using Multiple Cues

IEEE Transactions on Pattern Analysis and Machine Intelligence
Measures of Diversity in Classifier Ensembles and Their Relationship with the Ensemble Accuracy

Machine Learning
Classifier Combinations: Implementations and Theoretical Issues

MCS '00 Proceedings of the First International Workshop on Multiple Classifier Systems

True Path Rule Hierarchical Ensembles

MCS '09 Proceedings of the 8th International Workshop on Multiple Classifier Systems
Improving bagging performance through multi-algorithm ensembles

PAKDD'11 Proceedings of the 15th international conference on New Frontiers in Applied Data Mining

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we introduce a new diversity measure for classifier combining, called Entropy-based Pair-wise Diversity Measure (EBPDM) Its application to help removing redundant classifiers from a face classifier ensemble is conducted The preliminary experiments on UC Irvine repository and AT&T face database demonstrate that, compared with other diversity measures, the proposed measure is comparable at predicting the performance of multiple classifier systems, and is able to make classifier ensembles smaller without loss in performance.