Unveiling Student Satisfaction in Online Learning: Leveraging Artificial Neural Networks for Predictive Insights

Abstract

Amidst the surge in online learning, understanding factors contributing to student satisfaction becomes paramount. This study utilizes an Artificial Neural Network to predict satisfaction levels in online courses, considering variables like Online self-efficacy, Instructional design, Perceived Usefulness, Perceived System Quality, Assessment, and Learner Content Interaction, which impact attitude and behavioral intention. Primary data collected via a survey of 382 postgraduate management students from Tamil Nadu, India, was used to train the ANN model, enhancing its accuracy through techniques like data preprocessing, feature selection, and model optimization. Leveraging ANN capabilities, the study aims to identify key factors influencing student satisfaction in online learning, offering insights into their intricate relationship. The ANN demonstrated strong predictive performance. Behavioral Intention achieved 97.00% during training and 92.17% during testing, while Attitude reached 99.25% in training and 96.52% in testing. The Satisfaction model showed 98.13% in training and 97.39% in testing. Confusion matrices and key metrics like True Positive Rates (TPR) and True Negative Rates (TNR) highlighted the model's reliability, with TPR above 88% and TNR over 96% for all outcomes. These findings underscore the potential of ANNbased models for assessing student satisfaction in online learning environments and highlight the significant role of Attitude and Behavioral Intention in satisfaction forecasting. Moreover, the predictive prowess of the ANN model provides educators and institutions with a practical tool to assess and improve student satisfaction in online learning environments.