AN INTELLIGENT STUDENT LIFESTYLE ANALYTICS FRAMEWORK FOR SUPPORTING ACADEMIC SUCCESS

Abstract

Student academic performance is influenced by various academic and lifestyle factors that interact in complex ways. This study developed an Intelligent Student Lifestyle Analytics for Academic Performance Optimization by integrating Educational Data Mining and machine learning techniques to model students' academic and behavioral characteristics and provide data-driven recommendations for improving learning outcomes. The study utilized the publicly available Student Lifestyle Dataset consisting of 2,000 student records with variables related to study habits, sleep duration, physical activity, extracurricular participation, social activities, and stress level. Several machine learning algorithms, including Logistic Regression, Support Vector Machine, Decision Tree, Random Forest, and Gradient Boosting, were developed and evaluated using a 70:30 train-test split and five-fold cross-validation. Results revealed that study hours exhibited the strongest positive correlation with Grade Point Average (GPA) (r = 0.734), while physical activity showed a moderate negative relationship (r = -0.341). Among the classification models, Gradient Boosting achieved the highest performance with an accuracy of 87.67% and an F1-score of 85.52%. For GPA prediction, Ridge Regression produced the best results with an R² of 0.529 and an RMSE of 0.207. Feature importance analysis identified study hours as the most influential predictor of academic performance. The findings demonstrate the applicability of Student Digital Twin technology and machine learning in monitoring student behaviors, predicting academic outcomes, and supporting personalized interventions for academic performance optimization.