On the Earthquake Modeling by Using Bayesian Mixture Poisson Process

  • Aisah, Nur Iriawan, Achmad Choiruddin

Abstract

An earthquake is a natural disaster causing adverse impacts and damages. Indonesia
is one of the most earthquake occurrence countries due to its location on the active
Pacific Rings of Fire, especially in the Sulawesi-Maluku region. In many types of previous
research, earthquake is commonly analyzed based on its marks such as magnitude, depth,
or time. In this paper, we investigate the relationship between earthquakes and their
triggers as covariates namely fault, subduction, and volcano. The covariates are defined
as the distance of the earthquake’s epicenter to the nearest fault, subduction, and
volcano. The mixture Poisson point process model (MPPM) is proposed to model
earthquake datasets improving the common single unimodal Poisson process model for
clustered or nonhomogeneous earthquake patterns involving the covariates. Parameter
estimation is done using the Bayesian Markov Chain Monte Carlo (MCMC) approach
which employs the Gibbs sampling method. The results showed that the Poisson mixture
process which carries significant earthquake intensity characteristics is better than the
Poisson unimodal because it has a smaller Watanabe Akaike Information Criterion
(WAIC) value.

Published
2020-05-01
How to Cite
Aisah, Nur Iriawan, Achmad Choiruddin. (2020). On the Earthquake Modeling by Using Bayesian Mixture Poisson Process. International Journal of Advanced Science and Technology, 29(7s), 3350-3358. Retrieved from https://sersc.org/journals/index.php/IJAST/article/view/17622