Functional data analysis: exploratory tools on Covid-19 pandemic

Abstract

Coronavirus disease 2019, also known as Covid-19, is caused by a novel coronavirus called the severe acute respiratory syndrome coronavirus. The Covid-19 disease has become a severe threat to all countries globally, and the World Health Organization has labeled it a global pandemic. This study employs a modern statistical method known as functional data analysis, to examine the pattern of Covid-19 incidents in Malaysia. Smoothing functional data, which includes the summary of functional data, functional principal component, and functional outliers, are used to investigate the disease's characteristics. The functional principal component will capture the variation of the smoothing curves while their scores are utilised to cluster the Covid-19 cases. In addition, the functional graphical methods consist of rainbow plot, functional bagplot, and functional high density region are then used to identify outliers that were not visible in the original data plot. Functional data analysis provides more comprehensive methods than conventional statistical methods in identifying the pattern of Covid-19 incidence particularly in considering their temporal dynamics.