A machine learning enabled affective E-learning system model.

Area of Science:

• Educational Technology
• Affective Computing
• Machine Learning in Education

Background:

• Personalized learning is crucial for effective education.
• Traditional e-learning systems often lack adaptivity to individual learner states.
• Monitoring and responding to student emotions can enhance the learning experience.

Purpose of the Study:

• To propose an e-learning system model for content personalization based on real-time student emotions.
• To utilize brainwave data and machine learning for emotion recognition and adaptive content recommendation.
• To evaluate the system's impact on student engagement, satisfaction, and learning outcomes.

Main Methods:

• Brainwave data collected using portable Electroencephalogram (EEG).
• Supervised machine learning (K-nearest neighbours - KNN) for real-time emotion recognition.
• Reinforcement learning for adaptive content recommendation based on emotional states.
• Experimental design with control and experimental groups (n=30).

Main Results:

• K-nearest neighbours (KNN) achieved 74.3% accuracy, 70.8% precision, and 69.3% recall in recognizing emotional states.
• The proposed e-learning system showed no significant impact on learning and engagement.
• A significant enhancement in student satisfaction was observed compared to traditional e-learning systems (p < 0.05).

Conclusions:

• The developed e-learning system effectively recognizes student emotions using EEG and KNN.
• While not improving learning or engagement, the system demonstrably increases student satisfaction.
• This research highlights the potential of emotion-aware adaptive learning systems in education.