Semiparametric Estimation for Measurement Error Models with Validation Data

• Activity Number: 135
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Yuhang Xu* and Yehua Li and Jae-kwang Kim
• Companies: and Iowa State University and Iowa State University
• Keywords: Semiparametric ; measurement error ; kernel smoothing

Abstract

We propose kernel-based semi-parametric estimators for linear/nonlinear models with measurement error when a validation data set is available in addition to a primary data set. In particular, the measurement error model is completely unspecified and modeled by data-driven smoothing methods. The proposed estimators are shown to be consistent and asymptotically normal no matters whether the response variable is observed or not in the validation data set. Simulation studies and a real data analysis are performed to illustrate the performance of the proposed methods.