Performance of Support Vector Machine and Multi-layered Neural Network for EEG-based BCI Mobile Robot Control

Abstract

This paper presents comparison results between support vector machine (SVM) and multi-layered neural network (NN) for EEG-based BCI mobile robot control. SVM and NN are functional and well-studied tools based on machine learning for classification. The authors had already implemented some NNs for the BCI in our previous studies, however, the performance was not enough for practical use. In this study, the authors employed SVM for the same purpose and compared with the previous results by the NNs. From these results, it was shown that the machine learning algorithms might be complementary to each other depending on the EEG data as input signals.