Weibull分布等分位数数据填充算法及其应用  

QUANTILE-FILLING ALGORITHM FOR WEIBULL DISTRIBUTION AND ITS APPLICATION

作　　者：姜宁宁[1] 

机构地区：[1]中国科学院数学与系统科学研究院,北京100080

出　　处：《系统科学与数学》2008年第6期662-668,共7页Journal of Systems Science and Mathematical Sciences

基　　金：中国科学院国防科技创新基金资助

摘　　要：基于作者曾提出的处理非随机删失数据的迭代算法,针对Weibull分布相同定时截尾型试验数据,通过改变数据的填补方式,在保证由算法所得的参数估计的相合性和不变性的前提下,使得改进后的算法所得的虚拟完全数据更接近于真实的完全数据.模拟研究表明,对于删失数据情形下可靠度的区间估计问题,改进后的算法更加精确.Based on the iteration algorithm which is proposed to deal with non-random censoring data in past paper, the data-filling method is modified so that the virtual complete data obtained from the algorithm is much closer to the real complete data in the case of Weibull distribution under type Ⅰ censoring data. The simulation result shows that this algorithm is superior to the algorithm in past paper for interval estimator of reliability under censoring data.

关 键 词：矩不变准则 非随机删失 不变性 相合性 区间估计 

分 类 号：O212.1[理学—概率论与数理统计]