Evaluating the impact of a small number of areas on spatial estimation

Aswi, Aswi and Cramb, Susanna and Duncan, Earl and Mengersen, Kerrie L. (2020) Evaluating the impact of a small number of areas on spatial estimation. International Journal of Health Geographics. pp. 1-14.

	Text (Jurnal Internasional) 04_Aswi_2020_journal_Evaluating the impact of a small number of areas on spatial estimation.pdf - Published Version Download (709kB)
	Text (Hasil Penilaian Peer Review Jurnal Internasional) 04 Hasil Penilaian Peer Review.pdf - Supplemental Material Download (1MB)
	Text (Hasil Turnitin Jurnal Internasional) 04Turnitin Evaluating the impact of a small number of areas on spatial estimation.pdf - Supplemental Material Download (4MB)
	Text (Bukti Korespondensi IJHG) Bukti Korespondensi IJHG_Evaluating the impact of a small number of areas on spatial estimation.pdf Download (383kB)

Abstract

Background: There is an expanding literature on diferent representations of spatial random efects for diferent types of spatial correlation structure within the conditional autoregressive class of priors for Bayesian spatial models. However, little is known about the impact of these diferent priors when the number of areas is small. This paper aimed to investigate this problem both in the context of a case study of spatial analysis of dengue fever and more generally through a simulation study. Methods: Both the simulation study and the case study considered count data aggregated to a small area level in a region. Five diferent conditional autoregressive priors for a simple Bayesian Poisson model were considered: inde�pendent, Besag-York-Mollié, Leroux, and two variants of a localised clustering model. Data were simulated with eight diferent sizes of areal grids, ranging from 4 to 2500 areas, and two diferent levels of both spatial autocorrelation and disease counts. Model goodness-of-ft measures and model estimates were compared. A case study involving dengue fever cases in 14 local areas in Makassar, Indonesia, was also considered. Results: The simulation study showed that model performance varied under diferent scenarios. When areas had low autocorrelation and high counts, and the number of areas was at most 25, the BYM, Leroux and localised G = 2models performed similarly and better than the independent and localised G = 3 models. However, when the num�ber of areas were at least 100, all models performed diferently, and the Leroux model performed the best. Overall, the Leroux model performed the best for every scenario especially when there were at least 16 areas. Based on the case study, the comparative performance of spatial models may also vary for a small number of areas, especially when the data have a relatively large mean and variance over areas. In this case, the localised model with G=3 was a better choice. Conclusion: Detecting spatial patterns can be difcult when there are very few areas. Understanding the character�istics of the data and the relative infuence of alternative conditional autoregressive priors is essential in selecting an appropriate Bayesian spatial model.

Item Type:	Article
Uncontrolled Keywords:	Bayesian spatial estimation, Conditional autoregressive (CAR), Few areas
Subjects:	FMIPA > STATISTIKA - (S1) FMIPA KARYA ILMIAH DOSEN Universitas Negeri Makassar > KARYA ILMIAH DOSEN
Divisions:	KOLEKSI KARYA ILMIAH UPT PERPUSTAKAAN UNM MENURUT FAKULTAS > KARYA ILMIAH DOSEN KARYA ILMIAH DOSEN
Depositing User:	Dr. Aswi Aswi
Date Deposited:	01 Oct 2021 09:56
Last Modified:	19 Dec 2022 13:45
URI:	http://eprints.unm.ac.id/id/eprint/21174

Actions (login required)

View Item