Mental Health Chatbot Behaviour Classification and Heuristic Analysis

To combat any kind of mental health issues, seeking help from a professional is not everyone’s cup of tea especially in a developing country like India. Chatbots are introduced as an alternative yet efficient solution for the same. But not every chatbot can be transparent with data privacy and coherent with helping the one in need. This is where our proposed model comes into picture. This model is a text-to-text-based chatbot which works on dual model surveys that include rule-based and retrieval-based surveys. Rule-based part includes age and live emotion detection using deep neural networks and openCV. Retrieval-based model includes natural language processing techniques to process the data that has been given by the user. The model uses R-CNN algorithm and includes TF-IDF and cosine similarity index to give the best response based on the perception of the user. The proposed model yields better results than any other existing text-to-text chatbot.