Modeling of Epidemic Transmission and Predicting the Spread of Infectious Disease

Abstract

Ismail Husein, Dwi Noerjoedianto, Muhammad Sakti, Abeer Hamoodi Jabbar

China’s Wuhan was the epicenter of coronavirus, reported in December 2019. A month later, there was an intensive outbreak inception. Indeed, epidemiologists and virologists forecasted that the peak of the crisis would be in three months and disappear by the end of the fourth month. However, the virus contagion’s healing, prediction, and diagnosis continue to pose critical challenges. In this study, the central objective was to engage in COVID-19 analysis to discern its general diagnosis index via artificial intelligence (AI) application. The motivation was to steer improvements in how accurately the disease could be diagnosed and allow for timely clinical interventions. The sample entailed 85 undiagnosed persons and 32 diagnosed individuals. The research context was in Zhejiang province. To screen the important indexes, four AI technology types were employed. To deal with the problem, some of the feature selection methods that were employed included recursive feature elimination, Gradient boosted feature selection, and the multi-objective decomposition ensemble optimizer (ARMED).
In the results, it was noted that out of 18 indexes that were achieved in relation to coronavirus diagnosis, those that matched with the 2019 China virus diagnosis clinical guide included Amyloid-A in the laboratory, 2019 novel coronavirus RNA, Eosinophil rate, Eosinophil count, and white blood cells (WBC). Overall, the method developed was accurate relative to COVID-19 prediction and diagnosis, upon which the rate of confirmed diagnosis could be improved and pave the way for timely clinical interventions.