Relevance Vector Machine Applied to EEG Signals Classification

Abstract

The electroencephalogram (EEG) is a complex and aperiodic time series, which is a sum over a very large number of neuronal membrane potentials. Despite the rapid advances of neuroimaging techniques, EEG recording con- tinues playing an important role in both the diagnosis of neurological diseases and understanding of the psychological process. In order to extract relevant information of brain electrical activity, a variety of computerized-analysis me- thods have been used. In this paper, we propose the use of a recently developed machine-leaning technique - relevance vector machine (RVM) - for EEG signals classification. RVM is based on Bayesian estimation theory, which has as distinctive feature the fact that it can yield a sparse decision function defined only by a very small number of so-called re- levance vectors. From the experimental results, we can see that estimation and classification based on RVM perform well in EEG signals classification problem compared with traditional approach support vector machine (SVM), which indicates that this classification method is valid and has promising application.