Machine Learning-Based E-Learners’ Engagement Level Prediction Using Benchmark Datasets

The wide adoption of e-learning especially during and after the pandemic has given rise to the concern of learners' motivation and involvement.E-leaner engagement level recognition over time has become critical since there is little to no physical interaction.In this paper, a benchmark dataset was utilized in predicting learners' engagement levels in a blended e-learning system.Information Gain feature ranker was leveraged to ascertain the significance of the features.This study performed a comparative study on some machine learning algorithms including; Decision Tree, Naïve Bayes, Random Forest, Logistics Regression, Stochastic Gradient Descent, LogitBoost, Sequential Minimal Optimization, Voted Perceptron, and AdaptiveBoost.Each model was accessed using the 10-fold cross-validation.We measure the performance of the models before and after feature selection.The predictive results show that Sequential Minimal Optimization outperformed other models by attaining an accuracy of 90% with precision, recall, and f-measure values of 0.895, 0.897, and 0.895 respectively.