TORONTO -                                                       

1. Abstract

A chatbot is an artificial intelligence (AI) software that simulates a conversation - or a chat - with a user in natural language through messaging applications, websites, mobile apps or over a telephone.

There are several types of chatbots, such as button-based, rule-based, recognition-based, self-learning, and voice bots. In this paper, various methods, and libraries, including Sci-kit Learn, Matplotlib, Word Cloud, Pandas, natural language processing toolkit, and Flask have been adopted to explore, analyze, and prepare data, and build, test and deploy a model, the best of which, with 88 percent accuracy, predicts users’ symptoms and answers their queries. Two engines are built, defined as Engine One for symptom prediction, and Engine Two for COVID FAQ.

2. Introduction

This is a rule-based chatbot for the novel coronavirus (COVID-19) based on frequently asked questions (FAQ) from the World Health Organization (WHO). The WHO provides answers to FAQ about the coronavirus. Since the virus that causes COVID is still changing due to its frequent mutations, and because the scientists studying it are still learning about it, it can be challenging to find consolidated and reliable information. The WHO COVID dataset available in the GitHub repository is thus the one this project selected.

2.1 Workflow COVID-19 chatbot - symptom prediction

This section is dedicated to the application of the chatbot and how it works behind the scenes. Everything begins when