Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1149580 | Journal of Statistical Planning and Inference | 2009 | 12 Pages |
Abstract
This paper studies the optimality of circular neighbor-balanced designs (CNBDs) for total effects when the one-sided or two-sided neighbor effects are present in the model and the observation errors are correlated according to a first-order circular autoregressive (AR(1,CC)) process. Some optimality results under some specified conditions are provided and the efficiency of a CNBD relative to the optimal block design is investigated. In order to discuss the efficiency of a CNBD among all possible block designs with the same size, the optimal equivalence classes of sequences under the one-sided neighbor effects model are characterized and the efficiencies of CNBDs with blocks of small size are illustrated.
Related Topics
Physical Sciences and Engineering
Mathematics
Applied Mathematics
Authors
Mingyao Ai, Yunlong Yu, Shuyuan He,