Machine learning-based framework for predicting user satisfaction in e-Learning systems

Abstract

The usability of eLearning systems is of paramount importance in determining the effectiveness and user satisfaction. This study introduces a machine learning-based framework to predict users satisfaction on eLearning System aiming to create user-centered platforms that cater to diverse learners satisfaction. The study employed machine learning models such as Support Vector Machines, Decision Trees and Neural Networks to predict user satisfaction towards usability of eLearning System. OC2 (Optimal Course Content & Online Collaboration Lab) dataset was subject into the three said models to predict user’s satisfaction in eLearning System. The results obtained from these models shows a promising performance and a perfect classification of both satisfied and unsatisfied users in eLearning System. All the three models achieved and accuracy, precision, Recall and FI score of 100% which shows there is no misclassification in the three models. This proves the modes underscore its reliability in predicting users’ satisfaction level. The outstanding accuracy of machine learning models in predicting satisfaction levels demonstrates their effectiveness as dependable tools for assessing usability. This study can be extended by employing diverse dataset with different factors in identifying various usability issues and improving the design and functionality of e-Learning Systems. Also other models apart from Support Vector Machine, Decision Tree, and Neural Network can also be applied to this study to know the performance of the models in predicting the usability scores based on the identified factors on the dataset. Future research can extend this study by utilizing a more diverse dataset with additional factors to further refine the identification of usability issues and improve system design. Additionally, alternative machine learning models beyond Support Vector Machines (SVM), Decision Trees (DT), and Neural Networks (NN) can be explored to assess their effectiveness in predicting usability scores based on the identified factors. Also leveraging Deep learning model will enhance the study to know the stage of user satisfaction on the e-learning system.