Simultaneous confidence intervals for ratios of means of zero-inflated log-normal populations
Qi Zhang, Jing Xu, Jianxin Zhao, Hua Liang, Xinmin Li
Abstract
This paper studies simultaneous confidence intervals (SCIs) for the ratios of means of zero-heavy log-normal populations. We propose four SCIs by using the Bonferroni adjustment principle. Specifically, we integrate the fiducial generalized pivotal quantity (FGPQ) with the other two statistics to have two FGPQ-based methods and incorporate the variance estimates recovery (MOVER) with FGPQ to have two MOVER-based methods. We develop expedient algorithms to implement the methods, and conduct extensive simulation studies to evaluate and compare the performance of the four methods, which show that our methods perform well. We also apply the methods to analyse a real data set as an illustration.