SENSITIVITY ANALYSIS IN LINEAR REGRESSION

Based on a multivariate linear regression model, we propose several generalizations to the multivariate classical and modified Cook’s distances in order to detect one or more of influential observations including the case of linear transformations of the estimated regression parameter. For those distances, we derived the exact distributions and point out a method to extend the calculation of exact distributions for several other metrics available in the literature, for the univariate and multivariate cases. The results are extended to elliptical families not under the assumption of normality. An application is described in order to exemplify the methodology.

[1]  Chester L. Olson,et al.  Comparative Robustness of Six Tests in Multivariate Analysis of Variance , 1974 .

[2]  David A. Belsley,et al.  Regression Analysis and its Application: A Data-Oriented Approach.@@@Applied Linear Regression.@@@Regression Diagnostics: Identifying Influential Data and Sources of Collinearity , 1981 .

[3]  S. Weisberg,et al.  Residuals and Influence in Regression , 1982 .

[4]  Muni S. Srivastava,et al.  An introduction to applied multivariate statistics , 1984 .

[5]  T. J. Page.  Multivariate Statistics: A Vector Space Approach , 1984 .

[6]  C. Caroni,et al.  Residuals and influence in the multivariate linear model , 1987 .

[7]  D. Fraser,et al.  On generalization of the analysis of variance , 1988 .

[8]  K. Fang,et al.  Generalized Multivariate Analysis , 1990 .

[9]  T. W. Anderson,et al.  Statistical Inference in Elliptically Contoured and Related Distributions , 1990 .

[10]  A. Fielding.  Sensitivity Analysis in Linear Regression , 1990 .

[11]  A. Hossain,et al.  A comparative study on detection of influential observations in linear regression , 1991 .

[12]  Arjun K. Gupta,et al.  Elliptically contoured models in statistics , 1993 .

[13]  J. A. Díaz-García,et al.  Proof of the conjectures of H. uhlig on the singular multivariate beta and the Jacobian of a certain matrix transformation , 1997 .

[14]  Heleno Bolfarine,et al.  Local influence in elliptical linear regression models , 1997 .

[15]  Donald E. Ramirez,et al.  14 Some exact properties of cook's DI , 1998, Order statistics.

[16]  A. Rukhin.  Matrix Variate Distributions , 1999, The Multivariate Normal Distribution.

[17]  J. A. Díaz-García,et al.  SINGULAR ELLIPTICAL DISTRIBUTION: DENSITY AND APPLICATIONS , 2002 .

[18]  José A. Díaz-García,et al.  Influence Diagnostics for Elliptical Multivariate Linear Regression Models , 2003 .

[19]  Connie M. Borror,et al.  Methods of Multivariate Analysis, 2nd Ed. , 2004 .

[20]  José A. Díaz-García,et al.  A note on the Cook's distance , 2004 .

[21]  John Fox,et al.  Applied Regression Analysis and Generalized Linear Models , 2008 .