STUDENTS PERFORMANCE PREDICTION

Abstract

Student academic performance prediction is essential for identifying learners who may require timely academic support and intervention. This study presents a machine learning-based Student Performance Prediction System that analyzes academic and behavioral attributes, including attendance, internal assessment marks, previous semester performance, assignment submission, study hours, and classroom participation. The proposed framework performs data preprocessing through data cleaning, missing-value handling, feature selection, normalization, and categorical encoding before applying Decision Tree, Random Forest, Support Vector Machine, and Logistic Regression algorithms. Students are classified into three performance categories: Excellent, Average, and Poor. Model effectiveness is evaluated using accuracy, precision, recall, F1-score, and confusion matrix analysis. Experimental comparison indicates that Random Forest provides the highest prediction performance. The developed web-based system supports early identification, visualization, reporting, and timely academic intervention for improved educational decision-making. Keywords: Student Performance Prediction, Machine Learning, Random Forest, Academic Performance, Educational Data Mining, Predictive Analytics.