Universal Penalized Least Squares Estimation for Semiparametric Regression Models
HU Hong-chang (College of Mathematics and Statistics, Wuhan University, Wuhan 430072; Department of Mathematics, Hubei Normal University, Hangshi 435002 )
First, the universal penalized least squares method is given in this paper. Using the method, the semiparametric regression model is considered, and the estimators of parameter and non-parameter are obtained. Then, the universal penalized least squares estimators are compared with penalized least squares estimators. At last, the validity of the method is illuminated by simulating examples.