The class of Empirical Best Predictors (EBP) is widely used in survey sampling and small area estimation due to possibility of prediction of any population or subpopulation parameters. Their good stochastic properties requires that a set of relatively strong assumptions should be met. We present results of simulation studies where properties of the predictor are widely studied including different cases of model misspecification. A fast algorithm of computing EBP is also applied.
1 September 2006 in Australian & New Zealand Journal of Statistics
In the survey sampling estimation or prediction of both population’s and subopulation’s (domain’s) characteristics is one of the key issues. In the case of the estimation or prediction of domain’s characteristics one of the problems is looking for additional sources of information that can be used to increase the accuracy of estimators or predictors. One of these sources may be spatial and temporal autocorrelation. Due to the mean squared error (MSE) estimation, the standard assumption is that random variables are independent for population elements from different domains. If the assumption is taken into account, spatial correlation may be assumed only inside domains. In the paper, we assume some special case of the linear mixed model with two random components that obey assumptions of the first-order spatial autoregressive model SAR(1) (but inside groups of domains instead of domains) and first-order temporal autoregressive model AR(1). Based on the model, the empirical best linear unbiased predictor will be proposed together with an estimator of its MSE taking the spatial correlation between domains into account.
In the paper the empirical best linear unbiased predictor of the subpopulation total is proposed under some longitudinal model where both temporal and spatial moving average models of profile specific random components are taken into account. Two estimators of the mean square error of the predictor are proposed as well. Considerations are supported by two Monte Carlo simulation studies and the case study.
We consider longitudinal data and the problem of prediction of subpopulation (domain) characteristics that can be written as a linear combination of the variable of interest, including cases of small or zero sample sizes in the domain and time period of interest. We consider the empirical version of the predictor proposed by Royall (1976) showing that it is a generalization of the empirical version of the predictor presented by Henderson (1950). We propose a parametric bootstrap MSE estimator of the predictor. We prove its asymptotic unbiasedness and derive the order of its bias. Considerations are supported by Monte Carlo simulation analyses to compare its accuracy (not only the bias) with other MSE estimators, including jackknife and weighted jackknife MSE estimators that we adapt for the considered predictor.