Well Bot—A Reinforcement Learning and Sentiment Analysis-Based Chatbot Implementation for Mental Well-Being

As the world becomes more and more competitive, the number of people experiencing stress, anxiety, and other mental health problems is rapidly rising. In today’s society, stress and pressure are affecting everyone, including school-going children and elderly people. We believe that an individual’s mental health is equally important as their physical health. But unfortunately, there are not enough resources for mental health care or adequate facilities in Indian society; thus, individuals struggle on their own, without assistance or support. Therefore, we suggest a solution for designing a chatbot that will serve as a communication channel for users. The users can let out their feelings and thus get relieved of the stress and/or anxiety that is clogged up. The chatbot application will send pop-up notifications to the user based on inputs of the user’s mood from Google Health installed on the mobile. The sentiment analysis module will generate appropriate notifications and converse with the user once they begin chatting with the chatbot. The chatbot uses Google’s Dialogflow services, including natural language processing and reinforcement learning, to generate responses. This will benefit people who are depressed and are afraid to express their feelings out loud or fear being judged.