PREDICTIVE ANALYSIS ON STUDENT PERFORMANCE USING MACHINE LEARNING ALGORITHM

ABSTRACT This study focuses on the development and deployment of a supervised machine learning system designed to predict student academic performance, with the goal of supporting academic planning and early intervention strategies. The research employed a Linear Regression model, selected for its simplicity and effectiveness in modeling continuous variables, to forecast students’ future Grade Point Averages (GPAs) based on historical academic records. The system was built using Python and the Flask web framework, ensuring seamless integration between the front-end user interface and back-end processing logic. Data was collected through a custom-designed HTML form where students’ details such as name, matriculation number, academic level, and previous GPAs were submitted and stored in a structured format, either as CSV files or within a relational database. Prior to model training, the data underwent preprocessing, including cleaning, normalization, and transformation to ensure accuracy and consistency. The model was trained using input features derived from previous academic performance, and the resulting regression equation was used to generate predictions for subsequent academic levels (200, 300, and 400 levels). Evaluation of the model’s performance was conducted using the Mean Absolute Error (MAE) metric, which was calculated to be 0.7409. Given the CGPA value range of 3.03 to 4.70, this error represents approximately 44% of the range, indicating moderate predictive accuracy. Predicted results were dynamically rendered on a results page in a tabulated format using HTML and JavaScript, and made available for download. The system’s modular and object-oriented architecture facilitates ease of maintenance, scalability, and potential integration with institutional academic systems. By providing a practical and functional application of machine learning in the educational domain, this study contributes significantly to the field of learning analytics. It enhances understanding of how predictive models can inform decision-making, support personalized academic guidance, and improve overall educational outcomes through data-driven approaches.