Nonparametric Generation of Synthetic Data for Small Geographic Areas

摘要: Computing and releasing statistics for small geographic areas is a common task many statistical agencies, but public-use microdata these much less due to data confidentiality concerns. Accessing the restricted usually only possible within research center (RDC). This arrangement inconvenient researchers who must travel large distances and, in some cases, pay sizeable usage fee access nearest RDC. An alternative dissemination method that has been explored release synthetic data. In general, consists of imputed values drawn from predictive model based on observed Data preserved because no actual are released. The typically standard, parametric distribution, often key variables interest do not follow strict forms. this paper, we apply nonparametric generating continuous collected areas. evaluated using 2005-2007 American Community Survey. analytic validity assessed by comparing (baseline) inferences obtained with those