Statistical Inference Of Multiple Population Comparisons In Certain Situations

The comparison of multiple population has high research value in real life.In this paper two-parameter exponential distribution and Pareto distribution are selected to research.We consider hypothesis test to compare the parameters come from different population,parameters Bootstrap,Fiducial inference and MOVER method respectively.We firstly compare the pth-quantiles of two-parameter exponential distributions,consider the parameters Bootstrap,Fiducial inference and MOVER method.Then we consider the mutiple testing problem of several populations.We use same methods to compare the mean values of different Pareto distributions.To estimate performance of three methods,we simulate all the methods and make a comparison regarding how these methods effect in different circumstances.Simulation show that three methods result good in most cases,the performance of parameters Bootstrap is more consistent when the sample size is small.Finally we confirm the result of simulation through the case study.