Handling Class Imbalance in Multiclass Datasets by using a Neighborhood based Adaptive Heterogeneous Oversampling Ensemble Classifier

Abstract

Classification of multiclass datasets with the complexity of skewed data distribution is a widely discussed research area. In this paper, a novel Neighborhood based Adaptive Heterogeneous Oversampling Ensemble classifier is proposed to address the class imbalance in multidass datasets. The proposed algorithm is examined on five datasets. The performance results are compared with the benchmarking algorithms. The results revealed that the proposed method performs better than the benchmarking algorithms.