A Novel Evolving Ensemble Approach Based on Bootstrapping and Dynamic Selection of Diverse Learners

摘要

In this paper, a novel evolving ensemble approach based on bootstrapping and dynamic selection of diverse learners is presented. This approach combines the use of evolving Fuzzy-rule-based (FRB) classifiers with the dynamic generation of ensembles based on its diversity. In this sense, the method proposed keeps the properties of the evolving classifiers and the advantages of the use of ensembles.