Affect modeling in POOLE III using EEG signals and facial features

• Main Author: ILagan, Gerald Arthur S.
• Format: text
• Language: English
• Published: Animo Repository 2011
• Subjects: Computer Sciences
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/2637
• Institution: De La Salle University

In the field emotion detection, cameras and psychological sensors are some mediums used to gather data. ITS developers have created ITS with detailed model which makes an affective gap between machine and the learner. The Programmers Object-Oriented Learning Environment III (POOLE III) is a system that teaches object-oriented programming to students. However, POOLE III lacks an interactive avatar that will help further the student in learning or studying. This thesis presents an affective emotion detection module incorporated into POOLE III to give the student a more interactive learning environment and provide feedbacks to the student. Classification algorithm such as Naïve Bayes, Bayes Network, KNN and C4.5 were used to train the facial features (x and y coordinates of facial points), EEG features and both. Both facial and EEG features achieved the highest accuracies in all classification algorithms. Naïve Bayes achieved 48.82%, Bayes Network achieved 89.31%, kNN with k=3 achived 95.23%, and C4.5 achieved 95.73% using facial and EEG features. The research was able to detect the emotions in a learning context.