<![CDATA[An increased use of social media platforms by university students has raised some concerns about their possible addiction, which might result in negative effects on their mental health, sleep and academic achievement. Addiction assessment to social media can be done through questionnaire-based tools that cannot effectively capture the interrelations between variables. In this study, the development and evaluation of machine learning regression models for predicting social media addiction of university students will be discussed. The dataset comprises the survey data having the dependent variable, Addicted Score, and three independent variables: Average Daily Usage Hours, Mental Health Score, and Sleep Hours per Night. Three regression algorithms, including Linear Regression, Decision Tree Regression and Support Vector Machine (SVM) have been compared using Orange Data Mining software. Models have been evaluated based on the fivefold cross-validation with five performance metrics: MSE, RMSE, MAE, MAPE and R². Based on the experiment, it was found that Linear Regression had the highest accuracy with R² of 0.915, MSE of 0.215, and RMSE of 0.464. The R² score of Decision Tree Regression was found to be 0.907, while for SVM it was 0.879. It can be seen from these results that there is linearity amongst the relationships between the different variables in the dataset which made Linear Regression give superior generalization performance over other regressors. This paper has shown that the use of regression using machine learning gives an efficient way of predicting the level of social media addiction amongst university students.]]>
Copyrights © 2026
