Deciding the Appropriate Mother Wavelet for Extract Features from Brain Computer Interface Signals

Feature extraction is a very challenging task because the choice of discriminative features directly affects the classification performance of brain computer interface system. The objective of this paper is to investigate the Mother Wavelets' affects on classification results. In order to execute this, we extracted features from three different data sets by using twelve Mother Wavelets. Then we classified the brain computer interface signals with three classification algorithms, including k-nearest neighbor, support vector machine and linear discriminant analysis. The experiments proved that Daubechies and Shannon are the most suitable wavelet families in order to extract more discriminative features from brain computer interface signals.