广义Shiryaev-Roberts变点检测程序的三阶渐近最优性
统计理论
2015-03-17 v3 统计理论
摘要
考虑了简单变点问题背景下Shiryaev-Roberts检测程序的几种变体：从R_0=0开始(原始Shiryaev-Roberts程序)、从固定r>0的R_0=r开始，以及从具有准平稳分布的R_0开始。对操作特性进行了比较。随着虚警平均运行长度趋于无穷大，差异逐渐消失。结果表明，从专门设计的点r开始或从随机“准平稳”点开始的Shiryaev-Roberts程序是三阶渐近最优的。
引用
@article{arxiv.1005.1129,
title = {Third-order Asymptotic Optimality of the Generalized Shiryaev-Roberts Changepoint Detection Procedures},
author = {Alexander G. Tartakovsky and Moshe Pollak and Aleksey S. Polunchenko},
journal= {arXiv preprint arXiv:1005.1129},
year = {2015}
}
备注
28 pages, 4 figures, Theory of Probability and Its Applications (to appear)