指数族广义非线性随机系数模型的变离差检验

数学物理学报

指数族广义非线性随机系数模型的变离差检验

林金官, 韦博成

东南大学数学系南京 210096 江苏教育学院数学系南京 210013

出版日期:2003-10-25 发布日期:2003-10-25
基金资助:
国家社会科学基金(02BTJ001)和东南大学博士后基金资助项目

Testing for Varying Dispersion in Exponential Family Nonlinear Models with Random Coefficients

LIN Jin-Guan, WUIBo-Cheng

东南大学数学系南京 210096 江苏教育学院数学系南京 210013

Online:2003-10-25 Published:2003-10-25
Supported by:
国家社会科学基金(02BTJ001)和东南大学博士后基金资助项目

摘要/Abstract

摘要：

指数族广义非线性随机系数模型是Smith & Heitjan［１０］和 Wei et al［１１］所研究模型的推广。该文分别在模型离差 (dispersion) 的权不变和变异时，讨论了指数族广义非线性随机系数模型的变离差的检验问题，得到了score检验统计量。并利用欧洲野兔数据，分别对正态分布模型、Γ 分布模型和
逆高斯分布模型说明检验方法的有效性。

关键词: 指数族分布, 非线性模型;随机系数;变离差;Score 函数;Score检验

Abstract:

Exponential family nonlinear models with random coeff icients are the extension of models studied by Smith & Heitjan (1993) and Wei et al (1998). This paper dea ls with this type of models and discusses the tests of varying dispersion for constant and nonconstan t weights of dispersion respectively. Several score test statistics are obtain ed . European rabbit data (Ratkowsky (1983) are presented for normal, Γand inverse aussian nonlinear models respecti vely to illustrate the methodology.

Key words: Exponential family distribution, Nonlinear models, Random coefficients, Varying dispersion;Score function, Score , test , statistics.

中图分类号:

62J02

林金官, 韦博成. 指数族广义非线性随机系数模型的变离差检验[J]. 数学物理学报, 2003, 23(5): 535-544.

LIN Jin-Guan, WUIBo-Cheng. Testing for Varying Dispersion in Exponential Family Nonlinear Models with Random Coefficients[J]. Acta mathematica scientia,Series A, 2003, 23(5): 535-544.

参考文献

[1]Cook R D， Weisberg S.Diagnostic for heteroscedasticity in regression. Biometrika,1983,70: 1-10

[2]Simonoff J S, Tsai C L.Improved tests for nonconstant variance in regression based on the modified profile likelihood.Applied Statitics,1994,43: 357-370

[3]韦博成. 加权非线性回归的Score检验及其局部影响分析.应用概率统计,1995,11(2):147-156

[4]韦博成,胡跃清. 非线性回归模型相关性和异方差性的检验. 工程数学学报,1994,11(4): 1-12

[5]Smyth G K. Generalized linear models with varying dispersion.J Roy Statist Soc Ser B,1989, 51: 47-60

[6]Cox D R.Some remarks on overdispersion. Biometrika, 1983, 70:269-274

[7]Breslow N E.Tests of hypotheses in overdispersed Poisson regression and other quasi likelihood models.J Amer Statist Assoc, 1990, 85: 565-571

[8]Dean C B. Testing for overdispersion in poisson and binomi al regression models. J Amer Statist Assoc,1992, 87:451-459

[9]Ganio L M, Schafer D W. Diagnostics for overdispersion. J Amer Statist Assoc, 1992,87 :795-804

[10]Smith P J, Heitjan D F. Testing and adjusting for departures from nominal dispersion in generalized linear models. Appl Statist, 1993, 42(1):31-41

[11]Wei B C, Shi J Q, Fung W K,etal.Testing for varying dispersion in exponentialfamily nonlinear models. Ann Inst Statist Math, 1998,50:277-294

[12]Wei B C. Exponential family nonlinear models.Singapore: SpringerVerlag,1998

[13]BarndorffNielsen O E, Cox D R. Asymptotic technique for use in statistics.London: Chapmam and Hall,1989

[14]JacqminGadda H, Commenges D. Tests of homogeneity for generalized line ar models.J Amer Statist Assoc,1995,90:1237-1246

[15]Dean C B, Lawless J F. Testes for detecting overdispersion in poisso n regression models.J Amer Statist Assoc, 1988, 84: 467-472

[16]Ratkowsky D A. Nonlinear regression in modeling. New York: Marcal Dekk er, 1983