Classifier subset selection based on classifier representation and clustering ensemble

Ensemble pruning can improve the performance and reduce the storage requirements of an integration system. Most ensemble pruning approaches remove low-quality or redundant classifiers by evaluating the classifiers' competence and relationships via their predictions. However, finding the best way to represent classifiers and create ensemble diversity is still a worthy research problem in the ensemble pruning field. To confront this issue, we discuss whether properties other than predictions can represent classifiers and propose a new classifier selection method, classifier-representation- and clustering-ensemble-based ensemble pruning (CRCEEP). In the proposed method, two new classifier-representation-learning methods, local-space- and relative-transformation-based representation, are proposed to obtain more information about classifiers. CRCEEP incorporates the clustering ensemble method to group classifiers and prune redundant learners. Finally, accurate and diverse classifiers are integrated to improve classification performance. Extensive experiments were carried out on UCI datasets, and the experimental results verify CRCEEP's effectiveness and the necessity of classifier representation.

Automated summary

Ensemble pruning improves system performance and reduces storage requirements in integration systems. Most approaches evaluate the competence and relationships of classifiers by analyzing their predictions to remove low-quality or redundant classifiers. However, finding the best way to represent classifiers and create ensemble diversity remains a research problem. To address this, we propose a new classifier selection method called CRCEEP, which incorporates two new classifier representation learning methods and a clustering ensemble method. Extensive experiments on UCI datasets demonstrate the effectiveness of CRCEEP and the importance of classifier representation.