A Statistical Analysis Of Accelerated Life Test Under Parameter Exponential Distribution

Exponential distributions have been particularly useful in the area of reliability and life-testing.An important concept that arises recently in this area is that of censo -red sampling.Under the assumption that the exponential distribution is a reasonable model for a given population,statistical procedures based on sample date from the population will be derived for analyzing a test of sample type censored sampling. Discussion about some problems in one and two parameter exponential distribution hypothesis testing.We design a confidence interval for the exponential distribution parameter based on TypeⅡcensored samples.(1)Suppose that the distribution of the life time of the products follows the one and two parameter exponential law E(λ,μ),in whichλ＞0,is the scale parameter (failure rate),andμ＞0 is the location parameter(life guarantee).The exponential distribution is used as failure distribution in reliability and life test.In chapterⅡ,A statistical method is discussed for the product whose life obeys one or two parameters of exponential distribution to analyze typeⅡcensoring from the accelerated life test,we discuss uniformly minimum variance unbiased estimate and effective estimate of mean life by parameter estimation.(2)We design a confidence lower bound for the exponential distribution parameter based on TypeⅡcensored samples by two-stage procedures.This confidence lower bound has prescribed width and prescribed coverage probability.We design confidence interval for mean life of one parameter exponential distribution in hypothesis testing problem and uniformly minimum variance unbiased estimate of mean life which obeys two parameter exponential distribution in reliability and life test on TypeⅡcensoring.(3)In the fourth chapter,on the basis of the probability statistics discussing,the article gives a system's summary of several problems which is difficult for us to understand in the parameter supposition examination teaching and mainly includes policy-making the basis the rejection region choice as well as the policy-making reliable judgment,then provides the reliable explanation for the policy making"accept zero supposition".we discuss hypothesis testing problom for parameters of one and two parameters exponential distribution.for some test statistics,we also discuss their distributions and have also proved hypothesis testing problem for binomial distribution,poisson distribution and Weibull distribution.(4)Finally,the thesis is summariezed,and we will discuss and research hypothesis testing and interval estimation of the links and differences,in Bayes estimation, we are easy to useθposterior distribution to obtainθconfidence interval and hypothesis testing of Bayes.