Kismiantini, I.G.N.M. Jaya
Diarrhea becomes a severe problem for children under five years of age. A preventive action is needed to minimize the negative effect of diarrhea. Gender risk assessment may be necessary to control diarrhea transmission as different sexes have distinct healthy behaviour. We develop a collection of candidate models of Bayesian shared component random intercept models to identify a gender group's spatial risk of diarrhea. The candidate models include Poisson-Inverse Gamma and Negative Binomial-Inverse Gamma models with different values hyperprior parameters. The results showed that the Negative-Inverse Gamma model performed better than the Poisson-Inverse Gamma with respect to the existence of overdispersion in the data. The spatial patterns of diarrhea for men and women were similar. In some sub-districts, however, it indicated that women at two different locations had a higher risk of diarrhea compared to men. © Published under licence by IOP Publishing Ltd.
Department of Statistics, Universitas Negeri Yogyakarta, Indonesia; Department of Statistics, Universitas Padjadjaran, Indonesia
