Revivify: A Depression Detection and Control System using Tweets and Automated Chatbot

Abstract

Mental Health is a stigma in India and on a global scale. Ignorance of mental health on our part has created a world where those suffering from it cannot talk about it openly and often feel uncomfortable disclosing it to others or even professional therapists. To address this problem, we propose a digital system that detects signs of anxiety and also suggests methods for depression control. Revivify performs a comprehensive analysis of a user’s mental state using different techniques. We have used tweets, patient health questionnaires, depression anxiety stress scale (DASS), and personalized responses as our dataset. Our system uses Feed Forward Neural Networks, Latent Dirichlet Allocation, and Random Forest Classifier algorithm to classify the user responses and tweets into one of the nine levels of anxiety and depression. Random Forest Classifier gives the highest accuracy. Further, the chatbot also suggests various blogs and provides helpline numbers for damage control. This system is a cost-effective solution to detect depression.