In the fields of epidemiology, actuarial science and risk management of e-commerce, we may often encounter one data scenario that there are a lot of zero value in our research-ing object. And traditional model cannot fit this kind of data very well. In this paper, we proposed zero-inflated exponential distribution and did a research about it through objective Bayesian method. We obtained three different types of noninformative priors (Jeffreys prios, two reference priors and two general second order PMP), and also did posterior analysis based on reference prior. At last, we did simulation study and real data analysis to compare objective Bayesian method with MLE method. |