The Predictive Modelling of Student Academic Performance Using Machine Learning Approaches

In the field of education, machine learning techniques have been applied in numerous studies covering a wide range of topics, including student enrollment, graduation forecasts, failure rates, retention, and academic performance. The use of predictive analytics in machine learning offers valuable insights to educators, potentially aiding in the improvement of students’ outcomes through the analysis of historical data. However, based on a review of existing literature, research focusing on the use of machine learning and predictive analytics to enhance student performance in Malaysian higher education remains limited, specifically among Islamic Universities. The primary objective of this study is to create the most effective predictive model for forecasting students’ final grades by employing machine learning methods such as Multinomial Logistic Regression, Decision Trees, Random Forest, K-Nearest Neighbors, Naïve Bayes, and Support Vector Machine. This research utilizes a dataset comprising student records from the Business Statistics course at Universiti Islam Pahang Sultan Ahmad Shah, spanning from 2013 to 2022. The findings reveal that the Decision Tree model is the most accurate, with a 0.60 accuracy rate in predicting students’ performance levels. This optimal model is instrumental in enabling lecturers to identify students at risk of failing at an early stage.