Metric learning for event-related potential component classification in EEG signals

In this paper, we introduce a metric learning approach for the classification process in the recognition procedure for P300 waves in electroencephalographic (EEG) signals. We show that the accuracy of support machine vector (SVM) classification is significantly improved by learning a similarity metric from the training data instead of using the default Euclidean metric. The effectiveness of the algorithm is validated through experiments on the dataset II of the brain-computer interface (BCI) Competition III(P300 speller).