线性模型基于M估计的统计诊断与影响分析
M Estimators and Influence Diagnostics in Linear Models
摘要:

为了克服实际观测数据与既定模型之间可能存在的较大偏离，目前有两种常用的处理方法：稳健统计与统计诊断。M方法是最重要的稳健统计之一，也是线性回归分析中是最受重视和研究成果最多的方法之一。所以本文结合M估计方法分析数据的统计诊断，从而得到异常点或强影响点。本文给出了参数估计偏离的表达式及几个诊断统计量，最后通过两个实例验证了本文所提方法的可行性。

Abstract:  In order to overcome the large deviation between the actual observed data and established models. There are two common methods: the robust estimation and statistical diagnostics. M estimator is a robust es- timation, it is the method which got the most attention and research results in linear regression. Therefore, in- fluence diagnostics and M estimator are used to judge the impact of outliers or strong influence points in this paper. Moreover, the expression of parameter estimation deviate and diagnostic statistics are given. Finally, the proposed methods are applied to two data sets.

 

文章引用：姜荣, 钱伟民. 线性模型基于M估计的统计诊断与影响分析[J]. 统计学与应用, 2012, 1(2): 31-36. http://dx.doi.org/10.12677/SA.2012.12007

参考文献

[1] R. D. Cook, S. Weisberg. Residual and influence in regression. New York: Chapman and Hall, 1982.
[2] G. Seber, C. J. Wild. Nonlinear Regression. New York: Wiley, 1989.
[3] 姜荣, 邵明江, 钱伟民. 半参数非线性模型中的t-型估计和影响分析[J]. 华东师范大学学报(自然科学版), 2011, 3: 1-11.
[4] B. C. Wei. Exponential family nonlinear models. Singapore: Springer-Verlag, 1994.
[5] R. J. Beckman, C. J. Nachtsheim and R. D. Cook. Diagnostics for mixed-model analysis of variance. Technometrics, 1987, 29(4): 413-426.
[6] 吴小燕, 赵林城, 杨亚宁. 线性模型中M估计分布的随机加权方法逼近[J]. 系统科学与数学, 2008, 28(9): 1092-1100.
[7] R. Jiang, X. Yang and W. Qian. Random weighting M-estimation for linear errors-in-variables models. Journal of the Korean Sta- tistical Society, 2012, 41(4): 505-514.
[8] 孙慧慧, 林金官. 基于M估计的线性混合模型的局部影响分析[J]. 应用概率统计, 2012, 28(2): 217-223.
[9] 韦博成, 林金官, 解锋昌. 统计诊断[M]. 北京: 高等教育出版社, 2009.
[10] R. D. Cook. Assessment of local influence (with dis-cussion). Jour- nal of the Royal Statistical Society Series B, 1986, 48(2): 133- 169.
[11] L. A. Escobar, W. Q. Meeker. Assessing influence in regression analysis with censored data. Biometrics, 1992, 48(2): 507-508.
[12] B. C. We. Exponential family nonlinear models. Singapore: Springer-Verlag, 1994.
[13] R. D. Cook, N. Holschuh, and S. Weisberg. A notes on alterna- tive outlier model. Journal of the Royal Statistical Society Series B, 1982, 44(3): 370-376.
[14] S. Weisberg, Applied linear regression. New York: Wiley, 1985.
[15] 陈希孺, 赵林城. 线性模型中的M方法[M]. 上海: 上海科学技术出版社, 1996.

为你推荐