Predicting Academic Performance of University Students Using Machine Learning: A Case Study in the UK

Students’ performance is one of the key success factors in educational institutions. Understanding the performance of students can help identify issues, enabling real-time action where and when necessary. Further, early identification and improvement of students’ academic performance at all levels has been a major challenge in these institutions. Students may experience some difficulties which can impair their study and can negatively impact their academic performance. These issues can be efficiently addressed if students’ data is pre-analysed, and students’ performance predicted early to allow immediate decisions on support. Early prediction by educators and policy makers can assist in improving student and class performance. This work applied machine learning algorithms to analyse significant factors that influence students’ academic performance, which could be used to inform decisions on support, or to identify and notify the students who require assistance; thus, taking effective steps to improving their performance. We used the academic records of computer-science students at the University of Huddersfield from 2017-2022, which provided several features that useful in predicting the students’ performance. Evaluation of the results showed that decision tree and Ada Boost regression have higher accuracy score.