A Zero-and-One Inflated Cosine Geometric Distribution and Its Application

Count data containing both excess zeros and ones occur in many fields, the zero-and-one inflated distribution is suitable for analyzing them. Herein, we construct confidence intervals (CIs) parameters of cosine geometric (ZOICG) constructed by using five methods: a Wald CI based on maximum likelihood estimate, equal-tailed Bayesian CIs uniform or Jeffreys prior, highest posterior density prior. Their efficiencies were compared terms their coverage probabilities average lengths via simulation study. The results show that prior performed best most cases. number new daily COVID-19-related deaths Luxembourg 2020 involving with high proportion analyzed. It was found ZOICG model appropriate this scenario.