DEVELOPMENT OF A FUZZY INFERNENCE SYSTEM BASED WEB APPLICATION FOR EARLY PREDICTION OF CORONAVIRUS DISEASE

Abstract

COVID-19 or Coronavirus is the most crucial health crisis of present times globally. Starting from China every part of the world has fallen a victim to this pattern changing, life taking virus. Due to its high transmission power everyday thousands are attacked and dying, eventually leading to severe disruption in social, economic, academic arenas across the globe. All countries are in an unseen race to decline the rate of positive cases, spread and deaths caused by disease through isolation and treatment. The most prominent way of achieving this is testing. But people are mostly unaware or in dilemma about testing as maximum symptoms of COVID-19 coincides with regular flue. Delay in identifying COVID-19 may result in postponed medical intervention, enabling the virus to advance unchecked and potentially induce severe respiratory complications. Prompt diagnosis plays a pivotal role in commencing timely treatment, averting complications, and mitigating the risk of virus-related mortality. Consulting doctors or going for tests puts them in vulnerable situation and exposes them to the disease, this being a highly infectious one. While staying untested is risky too. To help in this scenario this study proposes a fuzzy inference system-based model to predict the possibility of being COVID-19 positive based on symptoms provided as input by users to the model. Fuzzy being a robust model, with capability to understand and process vague and imprecise data, is better suited over other models to interpret human inputs of COVID-19 symptoms. Different membership functions have been used to map the symptoms to fuzzy values. Then various defuzzification methods are used to observe the result and to perform a comparative study. From the comparative analysis, COA method has proven the best fitted result. Thus, COA has been used for the web application tool. The system outputs the possibility in the considered scale and suggests the need of test. The result of this system is interpreted by identifying cases in three separate categories: Not Needed, Stay Isolated and Test Immediately. An output score between 0 to 3 indicates that test is not needed. Output score between 2 to 5 provides indication to stay isolated and beyond 5 is a state where test is recommended to be done immediately.