Open AccessForecasting dengue incidence in Bali by means latent Bayesian count data model
Author(s) -
Anna Chadidjah,
I Gede Nyoman Mindra Jaya
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1776/1/012042
Subject(s) - dengue fever , incidence (geometry) , poisson regression , poisson distribution , count data , bayesian probability , statistics , statistical model , econometrics , bayesian inference , virology , mathematics , environmental health , medicine , population , geometry
Dengue disease is a viral infectious disease caused by DENV-1, DENV-2, DENV-3, and DENV-4. It does not only cause a public health problem. However, it may also cause social and economic conditions. Forecasting dengue incidence is a crucial part of an early warning system (EWS), which is needed in controlling and preventing dengue disease. Forecasting is a statistical tool used to obtain future information based on historical data. However, the models had been developed commonly applied for continuous data. It was very rare in modeling count data. We develop a model based on a Poisson log-linear model to accommodate count data. We use the Latent Bayesian approach to estimate the parameters model. We apply the model for forecasting dengue incidence in Bali. We used data from 2011 to 2016 to forecast the dengue incidences in periods 2017-2020. We found a Poisson model with Random Walk order one prior and Half Cauchy hyperprior distribution is the bet model for forecasting dengue incidence in Bali. We found the dengue incidence decrease from 2017 to 2020 where the highest incidence rates always occur from January to May. This condition is thought to be related to the rainfall period.