No abstract

Abstract-Identifying mammalian vigilance states has recently become an important topic in biological science research. The biological researchers concern not only to improve the accuracy rate for classifying the vigilance states, but also to extract the meaningful frequency bands. In this study, we propose a novel feature selection to extract the critical frequency bands of rat's EEG signals. The proposed algorithm adopts the concept of neighborhood relation during adding and eliminating a candidate feature. In the experiments, the proposed method shows better accuracy rate, and find out the feature subset which locate on the critical frequency bands for recognizing rat's vigilance states.