Artificial Intelligence-Based Chatbot to Support Public Health Services in Indonesia

Abstract

The aim of this study is to build an artificial intelligence chatbot application to support public health services. The chatbot acts as an information service that can replace the role of humans. The analysis of functional needs was obtained from information submitted by one of the heads of public health centers in Indonesia. This study uses the Scrum method with pregame stages to produce a plan consisting of functional and non-functional requirements analysis and conceptual design of the chatbot, which will be developed using Unified Modeling Language (UML) diagrams. The process of finding answers uses the matching graph master technique, which is a backtrack matching that utilizes a depth-first search strategy. There are 6 topics of chatbot services, including service schedules, health information, registration, diseases, drugs, and early care services for chatbot users. Tests conducted on these 6 topics showed an average correct answer ratio of 93.1% out of a total of 251 questions. The result of the usability measurement on the chatbot application that has been built obtained a system usability scale value of 80.1, indicating that the developed chatbots are acceptable for use.