Early Student Performance Prediction and Progress Tracking Using Machine Learning Techniques

Abstract

In today’s educational environment, a significant amount of student-related data is generated through digital learning platforms and academic systems. Effectively analyzing this data can provide meaningful insights into student learning patterns and outcomes. This study focuses on applying data mining techniques, particularly within the field of Educational Data Mining to predict student performance at an early stage. The proposed system is designed to evaluate students enrolled in an introductory programming course and estimate their likely final performance based on their initial assessments, accounting for approximately 15% of total grading. To achieve accurate predictions, multiple machine learning algorithms from different categories were implemented and tested using the WEKA tool. These algorithms were evaluated based on performance metrics such as accuracy, precision, recall, and F-measure. Among all the models tested, the Decision Tree algorithm demonstrated superior performance in identifying correct outcomes and maintaining balanced evaluation metrics. The developed model not only predicts academic results but also acts as a tracking mechanism, enabling timely feedback. This allows students to recognize their academic standing early and take corrective actions to improve their performance. Ultimately, this approach supports better academic planning and enhances overall student success. Keywords— Data Mining, Education Data Mining, Student Performance Prediction, Machine Learning, WEKA, Decision Tree, J48 Algorithm, F-Measure, Predictive Modelling.