Statistical diagnosis and gross error test for semiparametric linear model

This paper systematically studies the statistical diagnosis and hypothesis testing for the semiparametric linear regression model according to the theories and methods of the statistical diagnosis and hypothesis testing for parametric regression model. Several diagnostic measures and the methods for gross error testing are derived. Especially, the global and local influence analysis of the gross error on the parameter X and the nonparameter s are discussed in detail; at the same time, the paper proves that the data point deletion model is equivalent to the mean shift model for the semiparametric regression model. Finally, with one simulative computing example, some helpful conclusions are drawn.