Deep Learning Based Chatbot Framework for Mental Health Therapy

Anushka Bhagchandani,Aryan Nayak

DOI: https://doi.org/10.1007/978-981-16-5689-7_24

2022-01-01

Abstract:With the rising popularity of chatbots, the research on their underlying technology has expanded to provide increased support to the users. One such sphere has been mental health support. As we train chatbots to better understand human emotions, we can also employ them to assist users in dealing with their emotions and improving their mental well-being. This paper presents a novel approach toward building a chatbot framework that can converse with the users and also provide therapeutic advice based on assessment of the user’s mood. The framework employs sentiment analysis for analyzing the user behavior which classifies the use of our chatbot architecture. Depending on the classification, the framework present two trained chatbot model based on self-attention mechanism to engage user in generic or therapy based conversations. Hence, the framework is designed with an emphasis on using natural language processing and machine learning techniques to ameliorate the onset of mental health disorders.