|
Activity Number:
|
359
- SPEED: Computing, Graphics, and Programming Statistics
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 1, 2017 : 11:35 AM to 12:20 PM
|
|
Sponsor:
|
Section for Statistical Programmers and Analysts
|
|
Abstract #325142
|
|
|
Title:
|
A Mixture-Of-Regressions Model with Measurement Errors in the Response
|
|
Author(s):
|
Xiaoqiong Fang* and Derek Young
|
|
Companies:
|
University of Kentucky and University of Kentucky
|
|
Keywords:
|
Astronomy data ;
Bootstrap ;
EM algorithm ;
Measurement error ;
Mixture model ;
Regression
|
|
Abstract:
|
Research on mixture-of-regressions models is primarily limited to directly observed variables. However, the presence of measurement error imposes additional challenges for estimation. We consider a mixture-of- linear-regressions model where the variance of the measurement error is roughly known, which occurs with some astronomy data. Estimation is accomplished using an EM algorithm framework with a weighted least squares (WLS) estimator that has already been developed for the non-mixture set- ting. The sampling behavior of our proposed model's estimates is examined by a simulation study. This includes a bootstrap method for estimating the standard errors of the estimates for the proposed model. We also demonstrate the efficacy of this approach on a real astronomy dataset involving the flux measurements of gamma-ray bursts, where the variance of the measurement error for the flux measurements (the response) are known. Our results for this data problem are compared with estimates obtained for other traditional models, including the linear regression model and the mixture- of-linear-regressions model.
|
Authors who are presenting talks have a * after their name.
