A Novel Nonparallel Plane Proximal SVM for Imbalance Data Classification

The research of imbalance data classification is the hot point in the field of data mining. Conventional classifiers are not suitable to the imbalanced learning tasks since they tend to classify the instances to the majority class which is the less important class. This paper pays close attention to the uniqueness of uneven data distribution in imbalance classification problems. Without change the original imbalance training data, this paper indicated the advantages of proximal classifier for imbalance data classification. In order to improve the accuracy of classification, this paper proposed a new model named LSNPPC, based the classical proximal SVM models which find two nonparallel planes for data classification. The LS-NPPC model is applied to six UCI datasets and one real application. The results indicate the effectiveness of the proposed model for imbalanced data classification problems.