Expert System for Diagnosing Early Symptoms of COVID-19 Using the Certainty Factor Method

Abstract

Coronavirus Disease 2019 (Covid-19) is a pathogenic virus that is the main cause of respiratory tract diseases that can cause respiratory problems such as lung infections and can cause death. Based on data from the health department, the city of Tangerang said that the level of transmission of Covid-19 in the Tangerang Raya area was included in the red zone category, where the spread of Covid-19 averaged 130 additional cases per day. The main problem found is the difficulty of detecting early symptoms that arise as an indication of being infected with Covid-19 and tools to assist in detecting early symptoms felt by the general public. In this study, a rule-based knowledge representation process for the initial clinical symptoms of Covid-19 infection was carried out using an expert system with the certainty factor method, which aims to measure the level of certainty that has been confirmed by Covid-19 against the initial clinical symptoms felt by the user and makes it easier for users to detect the early symptoms of Covid-19 and helps the government, in this case, the health service center through the availability of the expert system application. Based on the results and discussions that have been carried out in this study, it can be concluded that the results of experiments to detect early clinical symptoms that were implemented using an expert system with the certainty factor method showed empirical calculations in the range of 81-85% confidence level on the results given by the system in initial clinical symptoms entered by the user. The results of the research above were obtained through a process of validating the suitability of initial clinical symptoms to the level of Covid-19 infection by using the forward chaining method and then measuring the level of confidence in the results that have been given based on the established rules. The expert system application with the proposed certainty factor method is the basis for developing mobile-based applications that can be used easily for the wider community