Misclassification analysis for the class imbalance problem

In classification, the class imbalance issue normally causes the learning algorithm to be dominated by the majority classes and the features of the minority classes are sometimes ignored. This will indirectly affect how human visualise the data. Therefore, special care is needed to take care of the learning algorithm in order to enhance the accuracy for the minority classes. In this study, the use of misclassification analysis is investigated for data re-distribution. Several under-sampling techniques and hybrid techniques using misclassification analysis are proposed in the paper. The benchmark data sets obtained from the University of California Irvine (UCI) machine learning repository are used to investigate the performance of the proposed techniques. The results show that the proposed hybrid technique presents the best performance in the experiment.