Detection And Prediction Of Future Mental Disorder From Social Media Data Using ML, Ensemble Learnind And LLM

Abstract

The increasing prevalence of mental health disorders such as depression, anxiety, and stress has highlighted the need for early detection and proactive intervention strategies. With the widespread use of social media, users generate vast amounts of textual data that reflect their emotions, behaviors, and psychological states. This study proposes an intelligent framework for the detection and prediction of future mental disorders using social media data by integrating machine learning (ML), ensemble learning techniques, and large language models (LLMs). The system utilizes natural language processing (NLP) to preprocess and extract features such as sentiment, linguistic patterns, and behavioral trends, while LLMs provide deep contextual understanding of text. Multiple ML models including Support Vector Machines, Random Forest, and Logistic Regression are combined using ensemble methods like bagging, boosting, and stacking to improve accuracy and robustness. The framework is designed to both detect existing mental health conditions and predict future risks based on historical data patterns. Experimental results indicate improved performance in terms of accuracy, precision, recall, and F1-score compared to traditional approaches. The proposed system offers a scalable and efficient solution for continuous mental health monitoring, enabling early intervention while addressing ethical concerns such as data privacy, bias, and responsible AI usage.