This project utilizes nonparametric density and regression to construct a technique that can accurately and consistently model different countries cumulative growth curve through phase divisions, while also employing clustering tools to determine which variable derived from the phase divisions is the most important and successful predictor of the different phases each country undergoes. 11 different countries, United States, Italy, South Korea, Taiwan, India, Brazil, Russia, Sweden, Peru, Spain, and South Africa are analyzed. These countries are chosen using stratified sampling based on how many cases and deaths they have, how they are handling their infection curve, and the overall situation of the pandemic in their country. To begin, the first and second nonparametric densities are utilized to divide up the infection curve into separate phases, which are then modeled using regression. The model type (linear, quadratic, or logarithmic), duration, slope, and r2 for the first three phases for every country are further analyzed using multivariate correlation, clustering variables, and hierarchical clustering. The authors draw links from the analysis to real life events in order to help predict the COVID-19 model of similar countries based on the most important parameter and decisions.
0 Comments
|