Self-Assessment Chatbot for COVID-19 prognosis using Deep Learning-based Natural Language Processing (NLP)

Abstract

Chatbots are computer programs that communicate with humans using natural language processing. Conversational programming is the most crucial part of the development of a chatbot. Existing chatbots are based on simple rule-based techniques. They are therefore limited in terms of functionality, context, and coherence, with narrow domains. Some of these systems also lack flexibility and do not emulate actual human conversation. The available chatbots use machine learning algorithms to answer questions and have structured conversations. However, they lack flexibility and do not emulate actual human conversation. The study proposes a chatbot that will help users self-diagnose symptoms related to COVID-19. Even today, the creation of a good chatbot model remains a significant challenge in Natural Language Processing (NLP) using Machine Learning (ML). This study suggests a newly developed method of creating a chatbot by using deep learning: The Recurrent Neural Network (RNN) will be used to model sequence data since it is modeled after the functionality of the human brain. Based on the classification of the available response classes, this study outputs both the chatbot response and the probability value. The system achieves a response time of 0.3 seconds and an accuracy of 95.35%. The study suggests a newly developed method of creating a chatbot by using deep learning with multiple processing layers, the deep learning approach provides better learning for chatbots to make intelligent decisions. Deep learning-based NLP chatbots result in better text classification accuracy with the ability to understand sarcasm, humor, and conversations better.