Volume 18, Issue 48 (2-2007)
Abstract
Efron's bootstrap method can only be used to estimate the precision measures of estimators when observations are independent. For spatial data that are spatially correlated, the moving block bootstrap method is usually used. But, in this method, the boundary observations have less chance of presence in blocks resampling than the other observations. In this paper, the new separate block bootstrap method is introduced and an algorithm is given for estimating the precision measures of estimators. A simulation study is carried out to compare the efficiency of the separate block bootstrap method with moving block bootstrap. It is shown that, with their method we can estimate the bias of sample mean with no error, and the estimator for variance of sample mean is consistent.
