Estimation For Kumaraswamy Distribution Based On Complete And Censored Data

It was in 1980 that the Kumaraswamy distribution was fisrt put forward by Kumaraswamy. Kumaraswamy distribution is a random variable values between the range of [0,1] on the double parameter continuity distribution. In this thesis,we discuss the parameter estimation of Kumaraswamy distribution with the con-sideration of complete data and censored data and study the problem of following aspects:Fisrtly, this paper discuss the bayesian and maximum Likelihood estimation for Kumaraswamy distribution using both simple sampling and ranked set. sam-pling techniques based on the complete data. Comparing the bias and the mean square error and efficiency using the method of numerical simulation, the results show that the parameter estimation under ranked set sampling is superior to the parameter estimation under simple sampling.Secondly, this paper discuss using EM algorithm to the maximum likeli-hood estimation and three kinds of loss function to bayesian estimation of Ku-maraswamy distribution based on the progressive Type-II hybrid censored data.Because we can not directly solve the complicated integral form. So we use lindley approximation method to obtain estimates of expression. Through the numerical simulation method for comparing the mean value and mean square error and con-fidence interval estimates using the method of numerical simulation. The results show that the bayesian estimation is better than maximum likelihood estimation.Finally, this paper discuss the maximum likelihood estimation and bayesian estimation of parameters from Kumaraswamy distribution and the confidence interval of parameters of different types under adaptive progressive Type-? cen-sored data. By numerical simulation, the results show that maximum likelihood estimation is quite close to bayesian estimation under the noninformative gamma distribution. Because of the bayesian estimation computation is time-consuming,So that the maximum likelihood estimation is better.