Model-Free Classification of Multi-Channel EEG via the Epsilon-Complexity Theory

• Activity Number: 229 - Statistical Process Monitoring of High-Volume Data Streams
• Type: Topic Contributed
• Date/Time: Monday, July 30, 2018 : 2:00 PM to 3:50 PM
• Sponsor: Quality and Productivity Section
• Author(s): Alexandra Piryatinska* and Boris Darkhovsky
• Companies: San Francisco State University and Institute for Systems Analysis FRC CRC RAS,
• Keywords: binary classification; EEG records; epsilon complexity

Abstract

We will present a new methodology for binary classification of EEG records, which correspond to two different mental states. We assume that these records are restrictions of continuous vector-functions on a uniform grid. Our methodology is based on the theory of the ?-complexity of continuous vector-functions. We utilize the ?-complexity coefficients as features for classification of EEG-records. Firstly we estimate the ?-complexity coefficients of the original signal and its finite differences. Secondly, we utilize Random Forest (RF) or Support Vector Machine (SVM) classifiers. We applied our methodology to the problem of classification of multichannel-EEG records related to a group of healthy adolescents (39 subjects) and a group of adolescents with schizophrenia (45 subjects). We were able to classify subjects in four-dimensional feature space. We found that the random forest (RF) classifier provides a superior result. In particular, out-of-bag accuracy in the case of RF was 85.3\%. Using 10-fold cross-validation (CV), RF gave an average accuracy of 84.5\% on a test set, whereas SVM gave an accuracy of 81.07\%. We note that the highest accuracy on CV was 89.3\%.