Real-Time Emotion Classification Using EEG Data Stream in E-Learning Contexts.

Area of Science:

• Affective Computing
• Educational Technology
• Machine Learning

Background:

• Emotions significantly impact learning outcomes in both traditional and online environments.
• Existing machine learning models for emotion classification are often offline, limiting real-time application in continuous data streams.
• There is a need for systems that can classify emotions in real-time to adapt e-learning experiences.

Purpose of the Study:

• To propose a novel real-time emotion classification system (RECS) for e-learning applications.
• To develop an online learning approach for emotion classification using electroencephalogram (EEG) signals.
• To evaluate the system's effectiveness and compare it with existing offline and online methods.

Main Methods:

• A real-time emotion classification system (RECS) was developed using Logistic Regression (LR).
• The RECS model was trained online using the Stochastic Gradient Descent (SGD) algorithm on continuous EEG data streams.
• The DEAP dataset, a standard benchmark for emotion classification, was utilized for validation.

Main Results:

• The proposed RECS effectively classifies emotions in real-time from streaming EEG data.
• The system achieved superior accuracy and F1-score compared to other offline and online emotion classification approaches.
• The developed system demonstrates practical applicability within an e-learning context.

Conclusions:

• Online training with SGD enables real-time emotion classification from EEG signals, crucial for adaptive e-learning.
• The RECS system offers a promising solution for enhancing e-learning by responding to learners' emotional states.
• This approach advances the field of affective computing in educational technology.