Pendugaan Parameter Fungsi Intensitas Proses Poisson Non- Homogen pada Pencemaran Gas Ozon di Kota Jakarta

Abstract

Non-homogeneous Poisson process can be applied in various fields. One of them is modeling ozone gas pollution in the city of Jakarta. Amount of ozone gas is assumed to follow the model of the non-homogeneous Poisson process. Intensity functions for Poisson process used are Weibull distribution, Musa Okumoto distribution, Goel Okumoto distribution and Goel Okumoto General distribution. Estimation of model parameter uses Bayesian method via Markov chain Monte Carlo method. The best model is chosen by using the deviance information criterion (DIC) and graphical approach. Based on the distribution used, Goel Okumoto General distribution is the best model for Jakarta city, because it has the smallest DIC value and it has estimation of expected value which is closed to the plot of the actual accumulation of ozone gas.