Entropy Analysis Based Differential Evolution Approach for Emotion Classification for EEG

Abstract

Electroencephalography (EEG) signal processing is having its significance in various applications related to the emotion recognition and classification. The behavior monitoring, behavior class identification, emotion class identification are the major aspects for classification of EEG signal. In this paper, a feature adaptive differential evolution (DE) approach is defined to perform emotion classification. In this work, we used discrete wavelet transform (DWT), for extracting the statistical features from the EEG signal for classifying the emotions. The signal features are obtained in the form of entropy value, mean value and standard deviation analysis. The experimentation is applied on different trained and testing sets. DE approach has been used to give the accuracy up to 85%. The results showed that the effective analytical results are derived from the model