Bayesian linear prediction in finite populations |
| |
| Authors: | Heleno Bolfarine |
| |
| Institution: | (1) Instituto de Matematica e Estatistica, Universidade de São Paulo, Agencia Iguatemi, Caixa Postal 20570, 01498 São Paulo, Brasil |
| |
| Abstract: | In this paper, Bayesian linear prediction of the total of a finite population is considered in situations where the observation error variance is parameter dependent. Connections with least squares prediction (Royall (1976, J. Amer. Statist. Assoc., 71, 657–664)) in mixed linear models (Theil (1971, Principles of Econometrics, Wiley, New York)), are established. Extensions to the case of dynamic (state dependent) superpopulation models are also proposed. |
| |
| Keywords: | Bayes linear prediction parameter dependent error variance mixed linear model dynamic superpopulation model |
| 本文献已被 SpringerLink 等数据库收录！ |
|
