Predictive Identification Of At-Risk Students: Using Student Information System Data

Abstract

As electronic systems are increasingly utilized in education, vast amounts of data have been accumulated in educational databases. Educational Data Mining (EDM) is an emerging research field that aims to explore data from educational institutions. Extracting knowledge from educational data can facilitate a better understanding and improvement of educational processes. Predicting student performance has become a hot topic, as it can help decision-makers identify the factors contributing to student success or failure. In this study, we utilized EDM techniques to create a framework for predicting student academic performance, using pre-admission information and first-year subject marks. In the first step, the most important attributes of the data were identified and then used an EDM algorithm to extract knowledge from the selected attributes. Our experimental results revealed that the random forest algorithm produced the most accurate predictions, achieving an accuracy rate of more than 78%.