Estimation and Inference for Varying-coefficient Models by Using Penalized Splines

摘要

This article provides a unified method for the estimation and inference on varying- coefficient models (VCM) using penalized spline functions. Penalized splines are used to estimate the smooth but arbitrary nonparametric coefficient functions. Some testing methods such as generalized maximum likelihood ratio test and generalized cross validation test are developed from the connection between penalized spline estimation of VCM and linear mixed model. Some simulation studies are done to evaluate the performance of the proposed method.