<![CDATA[Online education has gained unprecedented prominence since the COVID-19 pandemic, highlighting the crucial role of student adaptability in ensuring academic success. In this context, accurately predicting adaptability levels categorized (Low, Moderate, and High) has become a vital research focus. This study investigates such classification using a Decision Tree algorithm and addresses a key limitation found in previous research, which suggested the need to implement oversampling techniques to reduce class confusion, particularly between adjacent categories. A public dataset comprising 1,205 student records was used, originally published by Suzan et al. (2021). The dataset was significantly imbalanced (Moderate: 625, Low: 480, High: 100), implementing the application of two resampling strategies: random undersampling and SMOTE (Synthetic Minority Over-sampling Technique). Results show a clear performance gap between the two approaches. The undersampled model achieved an accuracy of 76.67%, while the SMOTE based model reached 95% accuracy. Improvements were particularly notable in minority class performance.]]>
Copyrights © 2026
