Scholarlink

<p>The COVID-19 outbreak necessitated a rapid transition to eLearning in higher education institutions worldwide, including Uganda, where infrastructural and digital literacy challenges compounded this shift. Predicting student satisfaction with eLearning systems helps institutions evaluate how well these platforms are working, assess their future potential, and make informed decisions. This supports better use of resources, prevents investment in ineffective systems, and enables timely interventions to improve teaching and learning quality. This study developed and evaluated machine learning models to predict student satisfaction based on perceptual data. Various machine learning predictive algorithms were trained on the processed and augmented dataset and tested on the original processed data, including ensemble methods (XGBoost, Random Forest, AdaBoost, Gradient Boosting), traditional classifiers (Logistic Regression, Decision Tree, Support Vector Machine, K-Nearest Neighbors), and neural networks (Multi-Layer Perceptron). Model performance was evaluated using metrics such as accuracy, precision, recall, and F1-score to identify the most effective approach for predicting student satisfaction levels. Among the evaluated models, K-Nearest Neighbors (KNN) achieved the highest mean accuracy of 88.4%, followed closely by XGBoost at 86.6%. The Friedman test indicated statistically significant differences in performance across models (χ²(8) = 39.79, <i>p</i> &lt; 0.001). Post-hoc Nemenyi tests identified KNN and XGBoost as significantly outperforming several other classifiers, underscoring their effectiveness for predicting student satisfaction. These findings demonstrate the potential of machine learning models in accurately predicting student satisfaction within eLearning environments. Identifying patterns in usability, content quality, and support services can enable institutions to leverage these predictive insights to optimize resource allocation, improve instructional design, and implement timely interventions that enhance the overall quality and effectiveness of online education.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Using machine learning and perceptual data to predict student satisfaction of eLearning systems in Ugandan institutions of higher education

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月6日

查看全文
  • Simon Peter Khabusi,
  • Patience Atukunda,
  • John Othieno

摘要

The COVID-19 outbreak necessitated a rapid transition to eLearning in higher education institutions worldwide, including Uganda, where infrastructural and digital literacy challenges compounded this shift. Predicting student satisfaction with eLearning systems helps institutions evaluate how well these platforms are working, assess their future potential, and make informed decisions. This supports better use of resources, prevents investment in ineffective systems, and enables timely interventions to improve teaching and learning quality. This study developed and evaluated machine learning models to predict student satisfaction based on perceptual data. Various machine learning predictive algorithms were trained on the processed and augmented dataset and tested on the original processed data, including ensemble methods (XGBoost, Random Forest, AdaBoost, Gradient Boosting), traditional classifiers (Logistic Regression, Decision Tree, Support Vector Machine, K-Nearest Neighbors), and neural networks (Multi-Layer Perceptron). Model performance was evaluated using metrics such as accuracy, precision, recall, and F1-score to identify the most effective approach for predicting student satisfaction levels. Among the evaluated models, K-Nearest Neighbors (KNN) achieved the highest mean accuracy of 88.4%, followed closely by XGBoost at 86.6%. The Friedman test indicated statistically significant differences in performance across models (χ²(8) = 39.79, p < 0.001). Post-hoc Nemenyi tests identified KNN and XGBoost as significantly outperforming several other classifiers, underscoring their effectiveness for predicting student satisfaction. These findings demonstrate the potential of machine learning models in accurately predicting student satisfaction within eLearning environments. Identifying patterns in usability, content quality, and support services can enable institutions to leverage these predictive insights to optimize resource allocation, improve instructional design, and implement timely interventions that enhance the overall quality and effectiveness of online education.