Online Program Home
My Program

Abstract Details

Activity Number: 125
Type: Contributed
Date/Time: Monday, August 1, 2016 : 8:30 AM to 10:20 AM
Sponsor: Biometrics Section
Abstract #320757 View Presentation
Title: Classification of Multivariate EEG Records via $\Epsilon$-Complexity of Continuous Vector-Functions
Author(s): Alexandra Piryatinska* and Boris Darkhovsky and Nathanael Aff
Companies: San Francisco State University and Institute for Systems Analysis and San Francisco State University
Keywords: EEG records ; classification ; epsilon-complexity ; multivariate time series

We propose a methodology for classification of relatively-short multivariate EEG records. This methodology is based on the theory of the ?-complexity of continuous functions which is extended here to the case of vector-functions. This extension permits us to handle multichannel EEG recordings. The essence of the methodology is to use the ?-complexity coefficients as features to classify different types of vector-functions representing EEG-records corresponding to different mental states. We apply our methodology to the problem of classification of two sets of multichannel EEG-records. The first one is a classification of EEG records into alcoholic and control groups. The second one is related to a group of healthy adolescents and a group of adolescents with schizophrenia. We have established that in both cases we obtained an accurate classification in the 4-dimensional spaces of ?-complexity coefficients. For classification in these cases the best results were achieved by the Random Forest classifier. The obtained results indicate the effectiveness of the proposed methodology.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

Copyright © American Statistical Association