ANALYSIS OF EMOTIONS USING EEG DATA AND MACHINE LEARNING

• Published in: Annals of DAAAM & proceedings p. 158
• Main Authors: Potekhin, Vyacheslav V, Unal, Ogul
• Format: Journal Article
• Language: English
• Published: DAAAM International Vienna 01.01.2021
• Subjects: Analysis; Data mining; Electroencephalography; Machine learning
• ISSN: 1726-9679; 1726-9679
• DOI: 10.2507/32nd.daaam.proceedings.025

Emotions might be one of the most significant differences between a machine and a human being. Currently, Machine Learning can learn and make predictions using Data in many different fields such as Medicine. In this paper, the goal is to evaluate how human emotions are being executed then test the performance of ML in terms of Emotion analysis. DREAMER Data was used to perform all the tasks. The classification of rankings for 23 subjects' was done. The reactions of subjects and EEG data were recorded while 23 participants were watching 18 different short movies. Each candidate has a ranking, between one to five, based on their affective state after each stimuli in terms of valence, arousal, and dominance. Afterward, Using Generative Adversarial Network (GAN), synthetic data will be created and analysed using supervised and unsupervised learning algorithms. Finally, all results will be compared. The research can help to evaluate the basic human emotions for robots or devices in Medicine. Robotics or cyber-physical machines in healthcare are already growing every day therefore, the quality of surgeries, treatments, or medical assistance can be improved. Keywords: Machine Learning; Emotion Analysis; Electroencephalography; Brain Computer Interface; Generative Adversarial Networks