Machine Learning-Based Detection of Stress Levels in Students with Social Media Addiction

The epidemic has shown that social networks (SNs) are prevalent and necessary for human connections. However, SNs are easy to access, which may make them addictive, especially for kids. The study literature suggests several psychological and physiological risk factors for social media addiction. The suggested method includes preprocessing, feature selection, and model training. The content was preprocessed to remove all URLs to guarantee that web links would not interfere with the analysis. The mRMR method and Chi-Square test were the most practical and effective feature selection methods. Training the model using the Attention-Based GCN. The recommended model outperformed the popular Random Forest (RF) and established GCN models. Attention-Based GCN boosted accuracy to 92.84%. It systematically identifies social media addiction risk factors using cutting-edge feature selection, model training, and preprocessing. Results suggest Attention-Based GCN can overcome social media addiction identification challenges.