AI-Driven Chatbot for Mental Health Analysis Using Transformer Models

Abstract

This Mental Wellness Chatbot is a software-based solution created for the purpose of solving the incomplete alleviation of depression symptoms in a person. This mental wellness Companion is a chatbot that act as a normal chatbot while mainly focus on the mental health of the user. To accomplish this, we are going to be use various techniques and algorithms to create this chatbot. We are going to use API token of a pre-trained hugging face model in order to create this chatbot rather than using the Large Language Model (LLM) locally since executing the chatbot locally requires a lot of GPU power in the device in order to create and embed the data that we are providing as a dataset. Mainly we are going to be using a pre-trained model from the open-source hugging face model and we are not going to be using OpenAI for completing it. For the database we are going to be using the vector database Facebook AI Similarity Search (FAISS) instead of Mangodb. The mental health Chatbot app is like a reliable friend in the pocket, always ready to lend an ear and offer support whenever you need it. Through gentle questions and personalized suggestions, it helps to navigate the emotions and find ways to feel better. By tracking persons progress over time, it creates a valuable record of emotional journey, empowering to understand oneself better. With 24/7 availability and a commitment to privacy and confidentiality, it provides a safe space for a person to express openly and honestly. Whether the person looking for coping strategies, educational insights, or connections to additional resources, the Chatbot app is there to guide a person on their path to emotional well-being.