列联表数据的局部影响分析

数学物理学报 ›› 2011, Vol. 31 ›› Issue (2): 518-527.

列联表数据的局部影响分析

何利平¹|石磊²

1.昆明医学院公共卫生学院卫生统计教研室昆明 650031|2. 云南财经大学统计与数学学院昆明 650221

收稿日期:2008-12-05 修回日期:2010-03-29 出版日期:2011-04-25 发布日期:2011-04-25
基金资助:
国家自然科学基金(10761010)、国家社科基金(XTJ001)和云南省自然科学基金资助

Local Influential Analysis in Contingency Tables

HE Li-Ping², SHI Lei²

1.Faculty of Public Health, Kunming Medical University, Kunming 650031|2.Statistics and Mathematics College, Yunnan University of Finance and Economics, Kunming 650221

Received:2008-12-05 Revised:2010-03-29 Online:2011-04-25 Published:2011-04-25
Supported by:
国家自然科学基金(10761010)、国家社科基金(XTJ001)和云南省自然科学基金资助

摘要/Abstract

摘要：

该文研究列联表离散数据的局部影响分析.基于幂散布度量(Power-divergence Measure)定义了离散数据模型的I距离函数,影响图和一致法曲率,并针对离散数据构建了一种局部影响分析方法.利用该方法给出了二维列联表独立模型的局部影响分析结果,并通过实例分析验证了该方法的有效性.

关键词: 离散数据, 局部影响分析, 幂散布度量, 二维列联表, I距离函数, 一致法曲率

Abstract:

This paper studies the local influence in contingency tables. Based on power-divergence statistics, the I distance function, associated influential graph and the conformal normal curvature are constructed, and a new local influence method is proposed to detect the influential cells in r×c contingency table. The results are applied to detect the influential observations in the independent model of the contingency tables. Finally an example is used for illustration.

Key words: Discrete data, Local influence, Power-divergence statistics, Contingence table, distance function, Conformal normal curvature

中图分类号:

62F03

何利平, 石磊. 列联表数据的局部影响分析[J]. 数学物理学报, 2011, 31(2): 518-527.

HE Li-Ping, SHI Lei. Local Influential Analysis in Contingency Tables[J]. Acta mathematica scientia,Series A, 2011, 31(2): 518-527.

参考文献

[1] Andersen E B. Diagnostics in categorical data analysis. J Roy Statist Soc, 1992, 54B: 781--791

[2]  Andy H Lee, John S Yick. A perturbation approach to outlier detection in two-way contingency tables. Austral \& New Zealand J Statist, 1999, 41(3): 305--314

[3]  Cook R D. Detection of influential observations in linear regression. Technometrics, 1977, 19: 15--18

[4]  Cook R D. Assessment of local influence. J Roy Statist Soc, 1986, 48B: 133--169

[5]  Cressie N, Read T R C. Multinomial goodness of fit tests. J Roy Statist Soc, 1984, 46B: 440--464

[6]  Cressie N, Read T R C. Pearson χ² and the loglikelihood ratio statistic G²: a comparative review. Int Statist Rev, 1989, 57: 19--43

[7]  Garcia-Heras J, Munoz-Garcia J, Munoz-Pichardo J M, Pardo L. Influence measures based on Cressie-Read divergence measures in multivariate linear model. Comm Statist Theory and Method, 2006, 35: 2055--2073

[8] Poon W Y, Poon Y S. Conformal normal curvature and assessment of local influence.J R Statist, 1999, 61B: 51--61

[9]  Read T R C, Cressie N. Goodness-of-fit Statistics for Discrete Multivariate Data. New York: Springer-Verlag, 1988

[10]  Shi L, Wang X R. Assessment of local influence in multivariate analysis. Acta Mathematica Scientia, 1996, 16(3): 257--270

[11]  Shi L. Local influence in principal components. Biometrika, 1997, 84(1): 175--186

[12]  Shi L, Wang X R. Local influence in ridge regression model. Computational Statistics and Data Analysis, 1999, 31(1): 341--353

[13] Shi L, Ojeda M. Local influence in multilevel regression for growth curves. Journal of Multivariate Analysis, 2004, 91: 282--304

[14]  石磊, 陈飞. 具有相关结构线性模型的局部影响分析. 数学物理学报, 2007, 27(1): 118--130

[15]  Simonoff J S. Detecting outlying cells on two-way contingency tables via backwards stepping.Technometrics, 1998, 30: 339--345

[16]  Wu X, Luo Z. Second-order approach to local influence. J Roy Statist Soc, 1993, 55B: 929--936