<![CDATA[The advancement of digital technologies has strengthened the use of data driven approaches in understanding the factors that shape students’ academic achievement. This study aims to examine how daily habits, lifestyle patterns, and environmental conditions contribute to exam performance using the Student Habits vs Academic Performance dataset from Kaggle, which contains 1,000 student records covering behavioral, health related, and socioenvironmental attributes. Guided by the CRISPDM framework, the research includes data preparation, exploratory analysis, and predictive modeling using two regression techniques: Linear Regression and Random Forest Regressor. The predictive models were developed to estimate exam scores based on several key variables, including study duration, attendance rate, sleep quality, leisure activities, and parental education level. The results show that Linear Regression achieved the highest accuracy, with an MAE of 4.19, an RMSE of 5.15, and an R² of 0.897, indicating that approximately 89.65% of score variability can be explained by the selected features. Meanwhile, the Random Forest model recorded a slightly lower R² of 0.850, suggesting that the dominant relationships in the dataset follow a largely linear pattern. These findings highlight that consistent study routines, regular attendance, adequate sleep, and supportive home environments are strongly associated with improved academic outcomes. The study emphasizes the importance of interpretable machine learning models in educational analytics and offers insights that may support data informed interventions aimed at enhancing student performance.]]>
Copyrights © 2026
