A Comparative Study of Administrative and Managerial Response Models for Effectively Controlling the Spread and the Aftermaths of COVID-19 Pandemic

Abstract

COVID-19 is a global pandemic that has emerged in Wuhan, Hubei Province, China in early December 2019 as “pneumonia of unknown etiology”. The spread of the virus has shown very rapid growth within a couple of months and soon became a public health emergency of international concern that threaten the health and safety of global population. World Health Organization (WHO) officially declared COVID-19 as global pandemic on March 11, 2020 soon after there were over 118,000 coronavirus infections in over 110 countries and territories around the world. COVID-19’s rate of fatalities is almost 20% and the rate of transmission is exponential in the number of infected subjects. All countries in the world have been affected by the COVID-19 outbreak. They all have responded with limited information and facing a lot of uncertainties. They have to respond at all levels of the institutions and society. There were various calls of response from different directions, domains and experts. The approach for the response to the pandemic also greatly varies from country to country. Countries have been challenged with a lot of issues such as underestimation, lack of establishment of central task force, conflicting orders from different authorities, late response or void, self-medication problem, information sharing and social media, shortage of protection gear, shortage of ventilators, identification of isolation centers, social challenges, migration and workers problem and supply chain disruption challenges response with different magnitude. The aim of the research was to analyze the COVID-19 spread in Ethiopia by developing machine learning models to predict COVID-19 cases, recoveries and deaths due to aforesaid pandemic for the short term and long-term period. The output is believed to be useful for timely administrative and managerial response for COVID-19 and resembling epidemics in the future. In order to conduct the experiments, a pre-contextual analysis have been carried out by gathering the relevant data through various platforms. The process of forecast modeling is done using time series analysis techniques and forecasting models for the short- and long-term predictions respectively between three days to ten days are developed using the world data as well as Ethiopia data. Long-short-term memory has superior performance than exponential smoothening, with a root mean squared error of 13.19 and a mean absolute percentage error of 13.12 for the daily cases forecast. Reports in the form of statistics of number of cases, recoveries and deaths along with progression were continuously published on the online dashboard of tableau.