Machine Learning based Mental Health Analysis using Social Media Data

Social media is a pivotal aspect of modern communication, influencing how individuals connect, share, and engage. This study investigates its effects on mental well-being through the application of sophisticated Machine Learning techniques, including Random Forest (97% accuracy), Decision Tree (91.75% accuracy), XGBoost (95% accuracy), and Naïve Bayes (92% accuracy). By employing sentiment analysis and natural language processing on user-generated content, the research identifies emotional trends, behavior patterns, and early indicators of psychological distress. The findings highlight age-specific tendencies, particularly among teenagers and young adults, where excessive use of social media platforms is linked to heightened risks of anxiety, depression, and isolation, despite the platforms' potential to offer support networks and mental health resources. The investigation proposes AI-driven approaches to encourage balanced social media usage, facilitating early intervention and addressing mental health issues. Thorough validation and independent data analysis confirm the models' reliability, providing valuable insights for mental health experts, policymakers, and platform developers to create user-focused solutions that promote healthier online interactions.