数学物理学报

• 论文 • 上一篇    下一篇

非线性再生散度随机效应模型参数置信域的曲率表示

张文专;唐年胜;王学仁   

  1. 东南大学数学系 南京 210018; 贵州财经学院数学与统计系 贵阳 550004
  • 收稿日期:2004-01-08 修回日期:2005-08-11 出版日期:2006-08-25 发布日期:2006-08-25
  • 通讯作者: 张文专
  • 基金资助:
    国家自然科学数学天元基金(10226005)、云南省自然科学基(2004A0002)、
    贵州省省长基金和东南大学博士后基金资助

Confidence Regions in Nonlinear Reproductive Dispersion Mixed Models in Terms of Curvatures

Zhang Wenzhuan;Tang Niansheng ; Wang Xueren   

  1. Department of Mathematics, Southeast University, Nanjing 210018;
    Institute of Mathematics and Statistics, Guizhou Finance and Economics College, Guiyang 550004
  • Received:2004-01-08 Revised:2005-08-11 Online:2006-08-25 Published:2006-08-25
  • Contact: Zhang Wenzhuan

摘要: 该文基于Laplace逼近建立了非线性再生散度随机效应模型在Euclid空间中的几何结构, 并在此基础上研究了此模型参数和子集参数的置信域, 进一步推广和发展了 Hamilton, Watts 和 Bates[1]关于正态非线性回归模型, Wei[2,3]关于嵌入模型和指数族非线性模型, Zhu, Tang 和 Wei[4]关于半参数非线性模型,唐年胜、韦博成和王学仁[5]关于非线性再生散度模型, Tang 和 Wang[6]关于拟似然非线性模型等的结果.

关键词: 非线性再生散度随机效应模型, Laplace逼近, 曲率, Score 统计量, 置信域

Abstract: A modified geometric framework is proposed for nonlinear reproductive dispersion mixed models (NRDMMs) based on the Laplace approximate marginal likelihood. Three improved approximate confidence regions of parameters and subset paramters are investigated for NRDMMs based on the proposed curvatures. The results for nonlinear regression models and exponential family nonlinear models are extended to the complicated nonlinear reproductive dispersion mixed models including exponential family nonlinear models and nonlinear reproductive dispersion models as their special cases.

Key words: Nonlinear reproductive dispersion mixed models, Laplace approximation, Curvatures, Score statistic, Confidence region

中图分类号: 

  • 62F25