Predicting Student Adaptability in Online Education: A Comparative Study of Machine Learning Models and Copula-Based Analysis

Abstract

The rapid shift to online education has underscored the need to understand and predict students’ adaptability levels to ensure effective learning outcomes. This study aims to classify students’ adaptability in online education using a range of machine learning models, including Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, K-Nearest Neighbors, Naive Bayes, Neural Network (MLP), and Gradient Boosting. The analysis is based on a dataset from Kaggle that includes features such as demographic information, educational background, and technological access. In addition to traditional machine learning approaches, the Copula method was applied to explore dependencies among features, enhancing the interpretability of the models’ predictions. The models were evaluated using several performance metrics, including Accuracy, Sensitivity, Specificity, Precision, Negative Predictive Value, and F-Score. Logistic Regression emerged as the most effective model, achieving an accuracy score of 99%, demonstrating superior performance across multiple metrics. These findings offer valuable insights for educators and policymakers, highlighting the potential of machine learning models, complemented by Copula-based analysis, to enhance our understanding of student adaptability and guide the development of targeted interventions in online education.