Markov Renewal Chain Model for Forecasting Earthquakes in Bangladesh Region

Abstract

In this paper we have proposed a Weibull Markov renewal process to model earthquakes occurred in and around Bangladesh from 1961 to
2013. The process assumes that the sequence of earthquakes is a Markov chain and the sojourn time distribution is a Weibull random variable
that depends only on two successive earthquakes. We estimated the parameters of the models along with transition probabilities using
maximum likelihood method. The transient behavior of earthquake occurrences was investigated in details and probability forecasts were
calculated for different lengths of time interval using the fitted model. We also investigated the stationary behavior of earthquake occurrences
in Bangladesh region.