An Improved Estimation for Mixed Effect Model with Application to Health Care Expenditure

• Activity Number: 353
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Lingjun He*+ and Jianwei Chen
• Companies: San Diego State University and San Diego State University
• Keywords: Asymptotic conditional bias and variance ; local linear smoother ; two-step estimation ; nonparametric regression ; random effects ; health care expenditure in China

Abstract

Mixed-effect models arise frequently when data are collected from different clusters or from longitudinal studies. They provide investigators a powerful way to adjust for the measured and unmeasured covariates that might affect the regression function. This paper investigates an estimation for a nonparametric mixed effects model where we model the unknown functional using the local linear smoother. Although the proposed estimation methods are applicable for general mixed effects models, we develop the asymptotic properties of the proposed estimators and show that it is an efficient method for estimating unknown function and predicting random effects. Our simulation results demonstrate that the proposed estimator is outperformed without considering random effects when there are random effects in the model. A real application from the health care expenditure in China is presented to illustrate the proposed method.