Exploring Student Performance Patterns Using Tree-Based Techniques

Due to its direct impact on the development and progress of nations, predicting student performance has acquired considerable importance in modern society. The evaluation of student performance measures the effectiveness of educational institutions and their capacity to influence the next generation. As a result, enhancing the educational process has become a necessity, compelling governments and institutions to devote significant resources to its ongoing development. Based on the Student Grade Data obtained from the Binus Online Learning platform at Bina Nusantara University, this study analyzes and predicts student performance using tree-based methods, specifically Decision Tree and Random Forest. The dataset includes pupil information and variables pertaining to performance. By contrasting the performance of these tree-based models, it is possible to gain insight into their accuracy and efficacy in predicting student outcomes. The experimental results demonstrate that both the Decision Tree and Random Forest models predict student performance with high accuracy. These results demonstrate the potential of tree-based methods in educational data analysis and prediction, providing educators, administrators, and policymakers with valuable insights for identifying at-risk students and implementing timely interventions to enhance educational outcomes.