Abstract

Spatial convolution (Poisson-Lognormal) model with Bayesian approach is developed into spatiotemporal form by adding temporal trend to analyze DHF relative risk. The DHF data had been considered as a 2-level hierarchy phenomenon. The developed model is divided into two models, e.i. Bayesian Poisson-Lognormal 2-level (BP2L) spatio-temporal due to the spatial random effects and extended of BP2L (EoBP2L) spatio-temporal due to the spatio-temporal random effects. The works of the models were demonstrated by using MCMC Gibbs sampler to analyze DHF data on 31 districts in Surabaya city during 120 months (2001-2010) using covariate such as temperature, humidity, rainfall, and population density. Based on virtue of relative risk visualization, MC error, and deviance, the EoBP2L spatio-temporal not only has better performance than the BP2L, but it also break up Surabaya city into two zones of DHF hot spot; Sawahan and Tambaksari district. January is the best time for DHF intervention every year in both hot spots.

Original languageEnglish
Pages (from-to)39-47
Number of pages9
JournalInternational Journal of Applied Mathematics and Statistics
Volume47
Issue number17
Publication statusPublished - 2013

Keywords

  • Bayesian spatio-temporal
  • Convolution
  • DHF
  • Deviance
  • MCMC Gibbs sampler
  • Poisson-lognormal
  • Random effects

Fingerprint

Dive into the research topics of 'New look for DHF relative risk analysis using bayesian poisson-lognormal 2-level spatio-temporal'. Together they form a unique fingerprint.

Cite this