| Popis: |
The main aim of the article is to discuss the discriminative features of EEG for alcoholism analysis. EEG signals are non-stationary and complex for analysis. The paper reports a brief discussion on different feature extraction and selection methods that have been studied for alcoholism analysis. Features can be extracted from direct EEG signals or decomposing signals into EEG bands. A wide range of features can be obtained from EEG. The paper reports a review on various feature selection techniques like statistical t-test, Mann-Whitney U test, Kruskal-Wallis plot, and principal component analysis. Reported results in the literature indicate the necessity of proper feature extraction and feature selection. |
