EEG Signal Classification using Linear Predictive Cepstral Coefficient Features

Abstract

An electroencephalogram (EEG) is a procedure that records brain wave patterns, which are used to identify abnormalities related to the electrical activities of the brain. In this study an effective algorithm is proposed to automatically classify EEG clips into two different classes: normal and abnormal. For categorizing the EEG data, feature extraction techniques such as linear predictive coefficients (LPC) and linear predictive cepstral coefficients (LPCC) are used. Support vector machines (SVM) is used to classify the EEG clip into their respective classes by learning from training data.