Fault diagnosis based on pruned ensemble

Abstract

A new fault diagnosis method based on ensemble pruning is proposed. Ensemble pruning means to search for a good subset of ensemble members that performs as well as, or better than, the original ensemble. Margin distribution on training sets is thought as an important factor to improve the generalization performance of classifiers. In this paper, based on the margin loss minimization, a new ensemble pruning algorithm is proposed and utilized in fault diagnosis. Experiment results show the effectiveness of the proposed technique.