Early Detection of Mental Health Using Social Media Posts and Machine Learning Techniques

Early Detection of Mental Health Using Social Media Posts and Machine Learning Techniques

Zahid Nawaz Khan

Department of BECHLOR OF VOCATIONAL IN ARTIFICIAL INTELLIGENCE AND DATA SCIENCE ,

Anjuman-I-Islam's AbduL Razzaq Kalsekar Polytechnic, New Panvel, Maharashtra, India

Abstract - Mental health disorders such as depression and anxiety are growing concerns worldwide. With the surge in the usage of social media platforms like Reddit, many individuals tend to express their feelings and thoughts online. This paper presents a machine learning-based system that leverages Reddit posts to identify early signs of mental health conditions. The proposed solution uses natural language processing (NLP) techniques, including text cleaning, tokenization, and classification through traditional ML algorithms. The model is deployed with an interactive Streamlit-based web app. The dataset used in this research was sourced from Kaggle, specifically the "Reddit Mental Health Data" repository. The implementation and results showcase promising accuracy in predicting depression and anxiety levels.

Key Words: Mental Health Detection, Reddit Data, Machine Learning, NLP, Streamlit, Depression, Anxiety, Classification