数学物理学报 ›› 1997, Vol. 17 ›› Issue (S1): 184-194.

• 论文 • 上一篇    下一篇

ASSESSMENT OF LOCAL INFLUENCE IN MULTIVARIATE REGRESSION MODEL

石磊1, 任仕泉2   

  1. 1. Department. of Statistics, Yunnan University, Kunming 650091, China;
    2. Dep. of Appl. Math., ChenDu Sci. and Tech. University, ChenDu 610065, China
  • 收稿日期:1996-01-30 出版日期:1997-12-26 发布日期:1997-12-26
  • 基金资助:
    Reseaxch is supported by National Natural Science Foundation of China and Science Foundation of Yunnan Province.

ASSESSMENT OF LOCAL INFLUENCE IN MULTIVARIATE REGRESSION MODEL

Shi Lei1, Ren Shiquan2   

  1. 1. Department. of Statistics, Yunnan University, Kunming 650091, China;
    2. Dep. of Appl. Math., ChenDu Sci. and Tech. University, ChenDu 610065, China
  • Received:1996-01-30 Online:1997-12-26 Published:1997-12-26
  • Supported by:
    Reseaxch is supported by National Natural Science Foundation of China and Science Foundation of Yunnan Province.

PDF (PC)

6

摘要： In this article, authors introduce a method to assess local influence of observations on the parameter estimates and prediction in multivariate regression model. The diagnostics under the perturbations of error variance, response variables and explanatory variables are derived, and the results are compared with those of casedeletion. Two examples are analyzed for illustration.

关键词: Influence graph, local influence, multivariate regression model, perturbation scheme

Abstract: In this article, authors introduce a method to assess local influence of observations on the parameter estimates and prediction in multivariate regression model. The diagnostics under the perturbations of error variance, response variables and explanatory variables are derived, and the results are compared with those of casedeletion. Two examples are analyzed for illustration.

Key words: Influence graph, local influence, multivariate regression model, perturbation scheme

版权所有 © 《数学物理学报》杂志社
地址： 武汉市71010号信箱(430071)　电话: 027-87199206(中文版)　027-87199087(英文版)
E-mail: actams@apm.ac.cn
本系统由北京玛格泰克科技发展有限公司设计开发