http://www.ejournals.eu/Schedae-Informaticae/2016/Volume-25/art/9008/

Abstract

Abstract. There is a strong research eort towards developing models that can achieve state-of-the-art results without sacricing interpretability and simplicity. One of such is recently proposed Recursive Random Support Vector Machine (R 2 SVM) model, which is composed of stacked linear models. R 2 SVM was reported to learn deep representations outperforming many strong classiers like Deep Convolutional Neural Network. In this paper we try to analyze it both from theoretical and empirical perspective and show its important limitations. Analysis of similar model Deep Representation Extreme Learning Machine (Dr ELM) is also included. It is concluded that models in its current form achieves lower accuracy scores than Support Vector Machine with Radial Basis Function kernel.