时 间：2025年8月18日(周一) 14：00 – 17：30
地 点：普陀校区理科大楼A1514室
报告人：加州大学戴维斯分校 蒋继明教授
主持人：华东师范大学 於州教授
摘 要：
The mean squared prediction error (MSPE) is widely used in small area estimation (SAE), but it treats over-prediction and under-prediction equally, despite their different consequences. We propose an asymmetric MSPE (AMSPE) measure that assigns different weights to these errors, leading to a new best predictor (BP) and its empirical version (EBP). Under the normality assumption, the AMSPE-based BP is shown to have a simple expression involving the MSPE-based BP and a weight-dependent constant. We also develop a second-order unbiased estimator for the area-specific AMSPE of the EBP. Theoretical and empirical properties of the proposed methods are evaluated, and an iterative procedure for determining the AMSPE weight with global linear convergence is introduced. A real-data example demonstrates the effectiveness of the approach. This work is in collaboration with Haiqiang Ma (Jiangxi University of Finance and Economics, China) and Thuan Nguyen (Oregon Health and Science University, USA).
报告人简介：
蒋继明，加州大学戴维斯分校教授、统计学系主任，研究方向包括混合效应模型、模型选择、小域估计、纵向数据分析、生物遗传学、统计渐近理论等，已在AoS、JASA、JRSSB、Biometrika等统计学国际顶级期刊发表论文100多篇，出版专著6本。蒋继明教授曾获聘国家高层次人才，并担任AoS、JASA等多个统计学国际顶级期刊的编委，是美国科学促进会(AAAS)、美国统计协会(ASA)、国际数理统计学会(IMS)的Fellow, 也是国际统计学会(ISI)的Elected Member，曾获得 Outstanding Statistical Application Award (ASA, 1998), Distinguished Alumni Award (National Institute of Statistical Science, 2015), the 31st Morris Hansen Lecture (Washington Statistical Society, 2023)，Award for Outstanding Contribution to Small Area Estimation (2024)。