The Covid-19 Chatbot Application Using A Natural Language Processing Approach

Abstract

Cases exposed to the Covid-19 virus in Indonesia until June 2021 continue to experience a spike in increases, to handle it, various government policies continue to be rolled out and the public needs to be given correct, precise and fast information so that mutual awareness can be built to suppress cases exposed to COVID-19. With this background, this study aims to design and build a COVID-19 chatbot system based on artificial intelligence based on the Natural Language Processing algorithm. This chatbot is expected to be a place to ask questions about all things related to covid-19 so that it can become a personal assistant with two-way communication that can be accessed quickly for 24 hours. This chatbot system was built using the Python programming language, Node.js server and MariaDB as the database. As a client, this chatbot is integrated with the popular instant messaging application in Indonesia, namely WhatsApp. The data set used to train the chatbot was 369 question data and spread into 46 question tags. Testing the chatbot system using blackbox testing, and to test the expected output, the chatbot was tested using 350 testing data and the accuracy rate of the chatbot in answering reached 54%.