|
Activity Number:
|
82
- Computer Experiments, Statistical Engineering, and Applications in Physical Sciences
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 29, 2018 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
|
Abstract #330563
|
Presentation
|
|
Title:
|
Variance Components Estimators OPE, NOPE and AOPE in Linear Mixed Effects Models
|
|
Author(s):
|
Subir Ghosh*
|
|
Companies:
|
Univ. of California, Riverside
|
|
Keywords:
|
Mixed Effects;
Design of Experiments;
Variance Components;
Optimum Estimators;
Robustness;
Replications
|
|
Abstract:
|
The optimum variance component estimation method yielding the uniformly minimum variance quadratic unbiased estimator (UMIVQUE) is often difficult or impossible to carry out for the full data. When it is impossible to find the UMIVQUE, a near optimum estimator (NOPE) is proposed under some distributional assumptions for the data but without assuming an exact functional form. The proposed general method of finding NOPE provides an exact closed form expression. When it is easy to find the optimum estimators (OPEs) for subsets of the data generated from a replicated experiment, a general method of finding an average optimum estimator (AOPE) is also proposed. The performance comparison of NOPE and AOPE are made individually with the other optimum estimators : MLE and REMLE, by simulated data for a designed experiment. The performance comparison of AOPE is made under four constraints on variance components. A real data is analyzed to obtain AOPE in addition to MLE and REMLE. The robustness properties of AOPE and NOPE in comparison with the other other estimators including the MOM are also investigated.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2018 program
|
