The Effect of the Breakdown Point Value on the Asymptotic Efficiency of the S-Estimator for Penalized Regression Splines Model: A Simulation Study

Ibrahim, Sabrin Mohamed Disoky

doi:10.21608/acj.2018.36219

The Effect of the Breakdown Point Value on the Asymptotic Efficiency of the S-Estimator for Penalized Regression Splines Model: A Simulation Study

Document Type : Original Article

Author

Sabrin Mohamed Disoky Ibrahim

Statistics,Mathematics and Insurance Department Faculty of Commerce Damnhour University Damnhour Egypt

10.21608/acj.2018.36219

Abstract

The penalized regression spline model is one of the most popular models that can be used for smoothing data in which it is difficult to determine the appropriate functional form that expresses the relationship between the dependent variable and the independent variable in the simple linear regression models. In practice, however, data containing outliers can be encountered, so there is a need for using robust estimators for that model such as Sestimator. A general feature of the S estimator in the linear regression models is that this estimator can have a breakpoint point 50% but this is accompanied by a low asymptotic efficiency, and if a breakdown point of the estimator is less than 50% will be accompanied by a relative increase in the relative efficiency of the estimator. Therefore, the current research is concerned with the study of the effect of the difference of the breakdown point on the performance of the S estimator for penalized regression spline model by conducting a simulation study

Keywords