A new statistical approach to model the counts of novel coronavirus cases.

This study proposes new statistical tools to analyze the counts of the daily coronavirus cases and deaths. Since the daily new deaths exhibit highly over-dispersion, we introduce a new two-parameter discrete distribution, called discrete generalized Lindley, which enables us to model all kinds of dispersion such as under-, equi-, and over-dispersion. Additionally, we introduce a new count regression model based on the proposed distribution to investigate the effects of the important risk factors on the counts of deaths for OECD countries. Three data sets are analyzed with proposed models and competitive models. Empirical findings show that air pollution, the proportion of obesity, and smokers in a population do not affect the counts of deaths for OECD countries. The interesting empirical result is that the countries with having higher alcohol consumption have lower counts of deaths.