J4

• 数学 • 上一篇    下一篇

对称熵损失下两个指数总体均值的序约束估计

赵世舜1, 宋洋1, 宋立新2   

  1. 1. 吉林大学 数学研究所, 长春 130012; 2. 大连理工大学 应用数学系, 辽宁省 大连 116024
  • 收稿日期:2006-03-06 修回日期:1900-01-01 出版日期:2007-01-26 发布日期:2007-01-26
  • 通讯作者: 赵世舜

Estimation of Order Means of Two Sample DistributionExponential under Symmetric Entropy Loss

ZHAO Shi shun1, SONG Yang1, SONG Li xin2   

  1. 1. Institute of Mathematics, Jilin University, Changchun 130012, China;2. Department of Applied Mathematics, Dalian University of Technology, Dalian 116024, Liaoning Province, China
  • Received:2006-03-06 Revised:1900-01-01 Online:2007-01-26 Published:2007-01-26
  • Contact: ZHAO Shi shun

摘要: 在对称熵损失下, 讨论了样本容量相等时, 两个指数总体均值λi(i=1,2)的约束极大似然估计i的险, 其中约束为λ1≤λ2. 证明了λ1与λ2具有比经典极大似然估计X1与X2 更小的风险, 并给出了当λ21→∞和n→∞时,λi对Xi(i=1,2)渐近功效e(λi,Xi)的值.

关键词: 序约束, 对称熵损失, 功效, 约束极大似然估计

Abstract: The present paper consider risk of the restricted maximum likelihood estimators (RMLE) of order means of two sample distribution exponential, λ1≤λ2, with the same sample size, under symmetric entropy loss. It proved that RMLE λi have smaller risk than usual sample means Xi(i=1,2). The asymptotic efficiencies e(λi,Xi) of RMLE λi with respect to sample means Xi for λ21→∞ and n→∞ have also been discussed.

Key words: order restriction, symmetric entropy loss, efficiency, restricted maximun likelihood estimator

中图分类号: 

  • O212.1