Abstract

Creating a reliable synthetic micropopulation dataset at small area level is challenging in the microsimulation modelling approach of small area estimation. Although a range of methods are in use for generating spatial microdata, none of these methods can consider the entire scenario of the overall population at small area level. As a result, a newly generated micropopulation dataset from those approaches often leads to inconsistent data for many small areas and to validate the outputs of model built on such a synthetic microdata is also difficult. In this paper, we propose a novel approach of Bayesian reweighting tools for generating synthetic spatial micropopulation data at statistical local area levels in Australia.

The new system uses MCMC simulation with a joint posterior density based iterative algorithm. Results demonstrate that the new method takes consideration of the complete scenario of micropopulation data units at small area level, and it can produce statistically reliable small area estimates and their variance estimations as well as credible intervals.
Original languageEnglish
Pages1
Number of pages3
Publication statusPublished - 2012
EventThe 21st Australian Statistical Conference (ASC 2012) and the 8th Australian Conference on Teaching Statistics (OZCOTS 2012): ASC2012 - Adelaide Convention Centre, Adelaide, Australia
Duration: 09 Jul 201213 Jul 2012
https://www.statsoc.org.au/Past-Conferences

Conference

ConferenceThe 21st Australian Statistical Conference (ASC 2012) and the 8th Australian Conference on Teaching Statistics (OZCOTS 2012)
Country/TerritoryAustralia
CityAdelaide
Period09/07/1213/07/12
Internet address

Fingerprint

Dive into the research topics of 'A Bayesian reweighting technique for simulating small area micropopulation datasets in Australia'. Together they form a unique fingerprint.

Cite this