Feature Extraction from Raw EEG Signals by Using Second Order Polynomial Fitting Algorithm

The classification of electroencephalogram (EEG) signals is a key issue in the brain computer interface (BCI) technology. Obtaining excellent classification result is directly based on an efficient feature extraction method In the paper, we propose a new method of feature extraction for classification of cursor movement imagery EEG. Second order polynomial fitting algorithm has been applied to imagined EEG signals to extract set of features. Then the extracted features are classified using support vector machine (SVM) and k-nearest neighbor (KNN) algorithms. We obtained significant improvement on classification accuracy for data set 1a, which is a typical representative of one kind of BCI data, as compared to the reported best accuracy in BCI competition 2003.