Spatial filter selection with LASSO for EEG classification

摘要: Spatial filtering is an important step of preprocessing for electroencephalogram (EEG) signals. Extreme energy ratio (EER) a recently proposed method to learn spatial filters EEG classification. It selects several eigenvectors from top and end the eigenvalue spectrum resulting spectral decomposition construct group as filter bank. However, that strategy has some limitations in are often selected improperly. Therefore features filtered by bank do not contain enough discriminative information or severely overfit on small training samples. This paper utilize one penalized feature selection strategies called LASSO aid us termed can better filters. Then two different classification methods presented evaluate our Their excellent performances demonstrate stronger generalization ability bank, shown experimental results.