Quadratic prediction and quadratic sufficiency in finite populations |
| |
Authors: | Xu-Qing Liu Dong-Dong Wang Jian-Ying Rong |
| |
Institution: | aDepartment of Computing Science, Huaiyin Institute of Technology, Huai’an 223003, PR China;bDepartment of Foundation Courses, Huai’an College of Information Technology, Huai’an 223003, PR China |
| |
Abstract: | The problem of quadratic prediction for population quadratic quantities in finite populations has been considered in the literature. In this paper, we mainly aim at extending the ordinary quadratic prediction problems to a general case, and derive the representations of the two essentially unique optimal predictors: one is an optimal invariant quadratic unbiased predictor, and the other is an optimal invariant quadratic (potentially) biased predictor. Further, we show that the two predictors are nonnegative and reasonable by considering an extreme situation, and apply resulting conclusions to a special model with a compound symmetric variance matrix. In addition, we propose a notion of quadratic sufficiency with regard to the optimal prediction problems by employing materials derived in the first part, and investigate corresponding characterizations in detail. |
| |
Keywords: | Finite population Linear model Nonnegativity OIQBP OIQUP Optimal invariant quadratic (potentially) biased predictor Optimal invariant quadratic unbiased predictor Quadratic prediction Quadratic sufficiency |
本文献已被 ScienceDirect 等数据库收录! |
|