Mental Health Evaluation During Internet Blackouts: A Machine Learning Approach

Authors: M. T. Islam, T. Islam, M. A. I. Rafi, and T. R. Sakib 

Abstract— Internet has become essential for modern communication and access to services. However, blackouts during politically sensitive periods can disrupt daily life, causing anxiety and distress. This study examines the psychological effects of internet blackouts during the Bangladesh Quota Movement in July 2024, when internet access was shutdown to control information flow. The disruption significantly affected communication, financial transactions, and access to essential services, exacerbating anxiety, stress, tension, and feelings of isolation among citizens. A survey of 980 participants with 20 questions assessed behavioral, emotional, and psychological impacts. Results showed stress levels ranging from minimal to extreme, reflecting widespread distress. The study also used machine learning models to predict stress levels. The Decision Tree model showed limited predictive power, with around 55% accuracy, while the Random Forest model improved to 67%. XGBoost performed better than both, achieving over 94% in all metrics, demonstrating better accuracy. These findings highlight the potential of advanced algorithms to model mental health impacts, enabling policymakers to develop targeted interventions and allocate resources efficiently, ultimately minimizing the psychological toll of future internet disruptions and improving overall preparedness.

Keywords— Internet Blackout, Quota Movement, Mental Health, Emotional, Behavioural, Environment

Full Paper