Binary Classification by SVM based neural Trees and Nonlinear SVMs

When performing classification of large set of samples, neural trees (NTs) are preferably used. To circumvent the problem of poor generalization of neural trees, hybrid neural trees have been proposed. Recently hybrid SVM based neural tree has been shown to be an effective binary classifier. In this paper, we examine the performance of SVM based neural trees relative to the nonlinear SVMs. We observe that nonlinear SVMs are more effective, though at higher computational cost. Our conclusions will provide important guidelines in data mining applications on real world datasets