区间统计量及其分布  

INTERVAL STATISTIC AND ITS DISTRIBUTION

作　　者：傅惠民[1] 

机构地区：[1]北京航空航天大学小样本技术研究中心,北京100083

出　　处：《机械强度》2005年第6期752-757,共6页Journal of Mechanical Strength

基　　金：国防科技预研项目(413200204)资助

摘　　要：提出区间统计量和区间秩统计量的概念,建立其概率密度函数和联合分布函数,给出区间统计量的均值、方差、协方差和区间秩统计量的平均秩、中位秩、百分位秩计算公式。区间统计量是对顺序统计量的补充,当失效数据较少难以统计分析时,可以通过区间统计量增加其数据,提高统计分析的精度。在此基础上,提出一种区间数据最佳线性无偏估计方法,并对正态分布、极值分布、Weibull分布等常用的位置—尺度分布进行讨论,该方法能够对产品贮存寿命进行小样本预测和可靠性评定。Interval statistics and interval rank statistics are presented respectively, and their distributions and united distributions are derived. Not only the means, the variances and the covarianees of interval statistics, but also the mean ranks, the median ranks and the percent ranks of interval rank statistics are given. Interval statistics and order statistics can complement each other, and the information among order statistics can fully be exploited and applied by interval statistics. Then the best linear unbiased estimate method for in- terval data of normal distribution, Weibull distribution, extreme value distribution or other location-scale parameter models is presented. The presented method has higher precision, and solves the problem of reliability assessment with small sample of storage life.

关 键 词：区间统计量 区间秩统计量 顺序统计量 秩统计量 区间数据 最佳线性无偏估计 贮存寿命 可靠性小样本 

分 类 号：O241.7[理学—计算数学] TB115[理学—数学]