Optimal Feature-Centric Approach for EEG-Based Human Emotion Identification

Emotions are intrinsic to human nature, playing a vital role in human cognition. Emotions are closely intertwined with rational decision-making, perception, human interaction, and human intelligence. EEG has emerged as a promising tool for investigating the neural aspects of emotions. The primary objective of this study is to identify distinguishing features in EEG data and determine suitable classification techniques for categorizing of human emotion EEG-based approach for detecting mental states i.e., relaxation, neutrality, and concentration. To validate this research, a publicly available dataset EEG Brainwave mental state is utilized. Twenty-five temporal domain features were extracted from this dataset. Correlation-based feature selection technique is applied to select the significant features for the classification of three mental states. The selected EEG features are classified using four distinct classifiers. It is evident from the results, that the Random Forest achieves the highest classification accuracy of 98.61%. The proposed framework outperforms prior emotion classification methods in terms of electrode count and accuracy.