Detection of alcoholism: An EEG hybrid features and ensemble subspace K-NN based approach

Abstract

The excessive consumption of alcohol affects the brain neuronal system. Electroencephalogram signals convey information regarding alcoholic or normal status of a subject. The paper reports a novel method of detection of alcoholism using EEG hybrid features. Narrow band pass Butterworth filters are designed to separate the EEG rhythms. Linear, nonlinear and statistical feature are extracted to measure the complexity and nonlinearity in EEG signal. Alpha and Gamma rhythm gives very low p-value, indicating that gamma and alpha rhythms are capable to differentiate alcoholic EEG signal from nonalcoholic EEG signal. These rhythm features were applied to ensemble subspace K NN classifier with 10-fold cross validation. The proposed method with ensemble subspace KNN classifier delivers best classification accuracy (98.25%) as compared with other existing techniques.