Chatbot for Mental health

Framework

The dataset is confidential; however, I thought to make the open-source to make a compilation of making different chatbots from scracth in Python. Motivation behind this project In 2017, the National Mental Health Survey reported that one in seven people in India suffered from mental disorders, including depression and anxiety. The increasing awareness of mental health has made it a primary concern of development. Nearly 150 million people in India needed interventions, where the low and middle class faced more burden than the well-off people. This project is an attempt to make mental health more accessible. This conversational agent can be complemented with clinicians to make it more effective and fruitful. Overview of the bots trained The dataset was picked up from kaggle - Mental Health FAQ. This dataset consists of 98 FAQs about Mental Health. It consists of 3 columns - QuestionID, Questions, Answers.

What it does

How we built it

Future goals I want to research the possibilities of the generative-based chatbot further. The current encoder-decoder model cannot capture all the dependencies in the decoder layer due to the compact nature of LSTM. Attention layers can be added after LSTM layers to decode each output dynamically.