Key points are not available for this paper at this time.
Background: Advancements in ICT have driven the widespread adoption of e-learning platforms like Moodle, enhancing education delivery and experience. While research has primarily focused on predicting learner performance using Moodle data, the role of cognitive load has been underexplored. Purpose: This study aimed to compare machine learning classifiers in predicting student cognitive load and performance within Moodle. Methodology: Conducted at the Technical University of Mombasa in Kenya between November and December 2023, the experiment involved 415 undergraduate students in a four-week online course powered by Moodle LMS. Participants were randomly assigned to a treatment group containing cognitive load management interventions or a control group, without interventions. Using Moodle logs, the predictive performance of Naïve Bayes (NB), Random Forests (RF), Support Vector Machines (SVM), and K-Nearest Neighbours (KNN) was analyzed with Python's Jupyter package. Findings: Results showed that NB had high precision (96.96%) and accuracy (93.04%) with minimal training time, while SVM also performed well with higher training time. RF excelled in accuracy but required more computational resources. Conclusion: The study suggests NB and SVM effectively predict cognitive load and performance. This knowledge can be utilized by LMS and instructional designers to advance data-driven student interventions in supporting student success in e-learning environments
Ogwoka et al. (2024) studied this question.