Statistical Inference Of Burr Distribution Under Type-Ⅱ Doubly Censored Scheme

Because of the industry and technology innovation,people pay more attention to thereliability of the product.In order to make clear the reliability of the products tested andput forward suggestions to improve product reliability, So we often need to make life ex-periment.the cutting tail is most commonly used.In dealing with these practical problems,incomplete data often is got and called the missing data, due to a variety of subjective orobjective reasons result in data loss. type-II doubly censored scheme is a common data inthe life experiment. Previous scholars mainly use fgure forecast method and regressionmethod to analyze the type-II doubly censored scheme. Because of good properties ofright thick tail of Burr distribution, in this paper, the system discussed the Bayes estima-tion for Burr distribution parameter under type-II doubly censored scheme under will bediscussed systematically, and the bilateral destiny truncated introduce missing data intoconstant and life experiment, get the test results.The frst chapter, the background and research situation at home and abroad oftype-II doubly censored scheme are introducesd.Defnition of type-II doubly censoredscheme,constant stress accelerated life test and two assumptions are gived.Finally theBurr distribution are given.The second chapter, Considering the type-II doubly censored scheme samples, param-eters of maximum likelihood estimation is given,the available Newton-Raphson iterationmethod to fnd the approximate solution;Bayes estimation and hierarchical Bayes estima-tion for Burr distribution parameter were given under q-Symmetric Entropy Loss Functionand composite LINEX loss function.The third chapter,in the life test type-II doubly censored scheme will be truncatedmissing data is applied to the more constant stress test and accelerated life test usingconstraints in the Bayes method, gives the Bayes estimation of failure rate under diferentstress level and accelerate the Bayes estimation of factor. The fourth chapter,mixed two parameters of the Burr distribution is an importantstatistical model, application of the conventional method is hard to get the parameter esti-mates.This chapter applies the ECM, Newton Raphson algorithm iterative expressions ofparameters estimation are given,and validate the estimated by data simulation is feasible.