Predicting Academic Achievement Using Machine Learning

Abstract

Predicting Academic Achievement Using Machine Learning Plays A Significant Role In Shaping A Student’s Future Educational And Career Opportunities. In This Study, We Propose A Machine Learning-based System To Predict Student Academic Achievement Using Performance Scores And Socioeconomic Data. The Project Utilizes Three Datasets: The Students Performance Dataset (subject-wise Scores: Math, Reading, Writing), A Scholarship Dataset (reflecting Financial And Social Backgrounds), And A Tamil Nadu Colleges Dataset (for Regional Mapping And Future Integration). After Preprocessing And Balancing The Data Using SMOTE, We Developed And Evaluated Several Machine Learning Models—Support Vector Machine (SVM), Random Forest, Decision Tree, And Boost. Among Them, The SVM Model Delivered The Highest Accuracy And Was Deployed Through A Flask-based Web Application. Explainable AI Techniques, Including SHAP And LIME, Were Utilized To Interpret Feature Importance And Ensure Model Transparency. This System Provides A Data-driven Approach To Identify Struggling Students Early And Offer Personalized Academic Support.

Keywords