Predictive Models for Forecasting Learner Achievement: A Data Mining and Machine Learning Approach

Abstract

The study aims to identify key factors influencing learner achievement and develop an early detection model for at-risk students. After the process of exploratory data analysis and feature engineering, predictive models using three machine learning algorithms: logistic regression (LR), decision tree (DT), and support vector machine (SVM) were developed and evaluated. Results show that SVM outperformed the others across all performance metrics. SEMESTER_GPA, PROGNAME, and INTAKESEM emerged as the most significant predictors.