来自于Multiple-Outlier模型的最小次序统计量序性质（英文）  

Ordering Properties of Minima from Multiple-Outlier Models

作　　者：程美芳[1] 方龙祥[1] 杨芳[1] 

机构地区：[1]安徽师范大学数学计算机科学学院,芜湖241002

出　　处：《应用概率统计》2017年第3期317-330,共14页Chinese Journal of Applied Probability and Statistics

基　　金：supported by the National Natural Science Foundation of China(Grant No.11201003);the Provincial Natural Science Research Project of Anhui Colleges(Grant No.KJ2016A263);the Natural Science Foundation of Anhui Province(Grant No.1408085MA07);the PhD Research Startup Foundation of Anhui Normal University(Grant No.2014bsqdjj34)

摘　　要：本文中,我们研究来自于两个multiple-outlier模型的最小次序统计量的随机比较,其中两个模型中独立同分布的随机变量个数不同.令X_(1:n)(p,q)和X_(1:n~*)(p~*,q~*)分别表示来自于X_1,…,X_p,X_(p+1),…,X_n和X_1,…,X_(p),X_(p~*+1),…,X_(n)的最小次序统计量,这里q=n-p,q~*=n~*-p~*.在参数(p,q)和(p~*,q~*)满足某些优化序条件下,我们根据普通随机序,失效率序和似然比序给出了X_(1:n)(p,q)和X_(1:n~*)(p~*,q~*)的序比较.In this paper, we compare the smallest order statistics arising from multiple-outliermodels when the numbers of independent and identically distributed random variables are dif-ferent. Let Xi ： n（ p ,q ） and Xi：n ＊ （p＊ ,q＊） denote the smallest order statistics among Xi,...,X p, Xp ＋i,...,X n and Xi,...,X p ＊ ,Xp ＊＋i,...,X n ＊, respectively, where q = n - p and q＊ = n＊ - p＊. We then prove that Xi：n （p,q） and Xi：n ＊（p＊,q＊） are ordered in terms of the usual stochastic order, hazard rate order and likelihood ratio order under the majorization relationship between （p,q） and（p＊,q＊）.

关 键 词：multiple-outlier模型 普通随机序 失效率序 似然比序 最小次序统计量 比例失效率模型 

分 类 号：O212.4[理学—概率论与数理统计] O213.2[理学—数学]