An asymptotic formula for the difference of the $M$-estimates of the regression coefficients of the non-linear model for all $n$ observations and for $n-1$ observations is presented under conditions covering the twice absolutely continuous $\varrho$-functions. Then the implications for the $M$-estimation of the regression model are discussed.
@article{118646,
author = {Asunci\'on Rubio and Francisco Quintana and Jan \'Amos V\'\i \v sek},
title = {Sensitivity analysis of $M$-estimators of non-linear regression models},
journal = {Commentationes Mathematicae Universitatis Carolinae},
volume = {35},
year = {1994},
pages = {111-125},
zbl = {0794.62022},
mrnumber = {1292588},
language = {en},
url = {http://dml.mathdoc.fr/item/118646}
}
Rubio, Asunción; Quintana, Francisco; Víšek, Jan Ámos. Sensitivity analysis of $M$-estimators of non-linear regression models. Commentationes Mathematicae Universitatis Carolinae, Tome 35 (1994) pp. 111-125. http://gdmltest.u-ga.fr/item/118646/
Sensitivity Analysis in Linear Regression, J. Wiley & Sons, New York. | MR 0939610 | Zbl 0648.62066
Residuals and Influence in Regression, Chapman and Hall, New York. | MR 0675263 | Zbl 0564.62054
Robust Statistics - The Approach Based on Influence Functions, J. Wiley & Sons, New York. | MR 0829458 | Zbl 0733.62038
A robust version of the probability ratio test, Ann. Math. Statist. 36, 1753-1758. | MR 0185747 | Zbl 0137.12702
Stability of regression model estimates with respect to subsamples, Computational Statistics 7 183-203. | MR 1178353
Influence function and regression diagnostics, In: Modern Data Analysis, R.L. Launer and A.F. Siegel, eds., Academic Press, New York, 149-169.
Regression analysis (in Czech), Academia, Prague.