Methods for Predicting Learners’ Cognitive Load in E-Learning Environments Using Eye-Tracking Data

The paper evaluates machine learning methods for the task of predicting learners’ cognitive load in e-learning environments using gaze tracking data. The main objective of the study is real-time adaptive intervention to prevent cognitive overload and increase learner engagement in the learning process. The study considers supervised learning techniques such as Support Vector Machines (SVM), Random Forest and Logistic Regression using simulated learner gaze tracking data. The research problem and objectives of the study are clearly defined. The study contains a comprehensive literature review that examines cognitive load theory, gaze tracking and machine learning techniques in educational contexts. The methodology focuses on developing and training models using k-fold cross validation to ensure robustness. Measures such as accuracy, precision, recall and F-score are used to evaluate the methods. The results of the study show that Random Forest is the most effective method, demonstrating its ability to capture complex patterns in gaze tracking data. The key contribution of this study is the novel application of intelligent methods to predict cognitive load from gaze tracking data, which enhances the predictive power of machine learning methods. The article highlights the importance of implementing these methods in real time and validation on real learner data, as well as addressing the ethical issues surrounding the use of gaze tracking data in educational settings.