Modelling of The Number of Malaria Sufeffers in Indonesia Using Bayesian Generalized Linear Models
Vera Maya Santi(a*), Anang Kurnia(b), Kusman Sadik(b)
a) Department of Statistics, Faculty of Mathematics and Natural Science, Universitas Negeri Jakarta
b) Department of Statistics, Faculty of Mathematics and Natural Science, Institut Pertanian Bogor
Abstract
Generalized Linear Models (GLM) has been used for modeling various types of data where the distribution of response variables is an exponential family. Common examples include those for binomial and Poisson response data. The GLM regression model determines the structure of the explanatory variable or covariate information, where the link function specifically determines the relationship between the regression model and the expected value of the observation. Bayesian techniques can now be applied to complex modeling problems where they could not have been applied previously. This method is a simpler model than traditional frequentist techniques. Estimating the regression model parameters is done by using Bayesian GLM. In this paper, we study conducted modeling for the number of malaria sufferers in Indonesia using the Bayesian GLM approach with several prior distributions. There are 6 independent variables that have a significant effect on the regression model, that is population density, Gini ratio, proper sanitation access, healthy zoning, integrated control and total sanitation. Based on Akaike Information Criterion (AIC) and standard error, the Bayesian GLM estimation results for Cauchy and Normal prior distribution will converge to the same value as that obtained by GLM.
Keywords: Malaria, Generalized Linear Models, Regression Parameter, AIC, Standard Error, Bayesian GLM.
Topic: Mathematics