Emotion classification using EEG signals and machine learning methods

Abstract

In this paper, a method of emotion classification from EEG signals is presented. This method comprises four steps of pre-processing by low-pass filtering, feature extraction using the discrete wavelet transform and the Wigner–Ville distribution, dimensionality reduction using linear dis-criminant analysis, and classification. The k nearest neighbors, random forests, and support vector machines were used as classification models. The results of this study showed that the features extracted by the wavelet transform and Wigner–Ville distribution led to the improvement of classification accuracy compared to other studies. In addition, the highest accuracy of 94.1% for the classification of 4 emotions was obtained using the features of the TP9 electrode and the support vector machine classifier.