Emotion Classification Using EEG Signals

This paper proposes a 3D emotional model for classifying emotions of a user while watching a musical video. A standard dataset DEAP (Database for Emotion Analysis using Physiological Signals) Dataset is used for studying and analyzing the human emotions using EEG signals. Participants are shown various videos across multiple trials and their corresponding EEG signals are recorded. The emotional states are classified on the basis of various parameters such as arousal, valence, dominance, and liking for a particular set of video. After relevant pre-processing and noise removal a 3D Emotional model is constructed. The 3D Emotional Model comprising of 8 octants within a Valence-Arousal-Dominance space gives rises to 8 different emotional states namely relaxed, peaceful, bored, disgust, nervous, sad, surprised and excited. The resultant emotional state obtained provides useful insight into the thinking and behaviour of participants in certain scenarios. The EEG based emotional classification can aid the developers to provide relevant recommendations to the user on the basis of his emotional state. Machine learning algorithms like Naive Bayes, Support Vector Machine (SVM) when applied to the proposed 3D Emotional model classifies the emotion aptly with an accuracy of 78.06% and 58.90% respectively.