SUSTAINABLE AI-NLP FRAMEWORK FOR EARLY DETECTION OF MENTAL HEALTH RISKS IN SOCIAL MEDIA POST

Authors:

Potharaju Aravind, Jahanara Begum, Mannem Hema, Rathla Rahul, Vemula Dhanunjay

Abstract-- Early identification of mental health disorders is vital but is frequently hindered by conventional methods of assessment. In this work, we apply natural language processing (NLP) and machine learning (ML) techniques to predict mental health disorders from social media posts. Raw textual data is preprocessed and then TF-IDF, embeddings, and sentiment scores are used to represent the text as features. SVM, Random Forest, LSTM and BERT. The results demonstrate that transformer based models are better than classical based models, suggesting that social media analytics can provide a scalable and non-intrusive tool for early screening of individuals for mental health disorders with due consideration to privacy and ethical issues.

Keywords— Mental health prediction, Social media posts, Natural Language Processing (NLP) , Machine Learning, BERT, LSTM , TF-IDF , Sentiment analysis , Early detection , Privacy and ethics