报告题目：Imperfect Extreme Ranked Set Sampling: A Framework for Robust and Efficient Inference

报 告 人：陈望学（吉首大学数学与统计学院 教授）

报告时间：2025年12月28日（星期日）9：00-10：00

报告地点：理学院（格物楼）402

报告摘要：Extreme ranked set sampling (ERSS) is a cost-effective data collection technique that improves the efficiency of population mean estimation by quantifying only extreme units. However, its theoretical advantages rely on the often-unrealistic assumption of perfect ranking. This paper introduces a more practical framework, termed imperfect extreme ranked set sampling (IERSS), which formally accounts for judgment errors. We derive the theoretical properties of the IERSS mean estimator, providing explicit expressions for its mean squared error (MSE) and variance. For symmetric populations, the estimator is proven to be unbiased. Our analysis demonstrates that the IERSS estimator consistently outperforms its simple random sampling (SRS) counterpart for symmetric distributions and for small samples from asymmetric distributions, even under significant judgment errors. Furthermore, we explore the efficiency trade-off between IERSS and imperfect ranked set sampling (IRSS), providing practitioners with guidance for choosing the most appropriate sampling design based on the expected ranking quality. Numerical simulations using exponential, Pareto, and normal distributions substantiate our theoretical findings. The results establish IERSS not merely as a theoretical extension but as a pragmatic and superior method for efficient statistical inference in real-world settings where perfect judgment ranking is unattainable.

报告人简介：陈望学，教授，吉首大学数学与统计学院副院长，硕士生导师，主要研究领域为数据采样理论与方法、计算机试验的建模与优化。系芙蓉计划-湖湘青年英才、湖南省青年骨干教师、兼中国现场统计研究会统计调查分会常务理事、副秘书长。主持国家自然科学项目2项。以第一完成人获湖南省自然科学奖三等奖1项。以第一作者和(或)通讯作者在核心权威期刊发表科论文54篇(其中SCI/SSCI论文33篇)。