Search

The occurrence of earthquakes is increasing almost every year in Indonesia. From January 2014 to December 2017, there was around 16,645 earthquakes with magnitude ≥4 Richter Scale occurred. This study is the first part of earthquake risk modeling that we conducted. This study aims to analyze the relationship of frequency and magnitude of the earthquakes by using Poisson Regression and Generalized Poisson Regression. The data used in this study is frequency and magnitude data of earthquakes occurred in Indonesia. The data were selected by selecting earthquakes with magnitude ≥4 Richter Scale in the period January 2014 to December 2017 (4 years). The dependent variable is frequency, meanwhile the magnitude is independent. The frequency of earthquakes is the rounded value of natural log (Ln) transformation of cumulative frequency of earthquakes occurred in time period, and tested that it follows poisson distribution. The Poisson Regression analysis was done for the first, then the analysis continued by using Generalized Poisson Regression to observe whether there is equidispersion effect. The result of two models was compared then continued by selecting the best model based on the smallest of Akaike Information Criterion (AIC). According to the result of Poisson Regression as well as Generalized Poisson Regression, the magnitude is significantly affect the frequency. Based on AIC, the best model of frequency-magnitude relationship is presented by Poisson Regression model, μ = exp (4.048 – 0.3935x). [ABSTRACT FROM AUTHOR]

Corona Virus Disease 2019 (COVID-19) is a new virus that can be contagious and its worst effects can lead to death. COVID-19 first appeared in Wuhan, China until it finally spread throughout the country, one of which is Indonesia. The spread of COVID-19 cases in Indonesia itself is quite rapid until finally the World Health Organization (WHO) designates COVID-19 cases as pandemics. Based on current conditions, this paper discuss about predict positive case data of COVID-19 at five locations in East Java (Malang City, Batu City, Pasuruan Regency, Malang Regency, Pasuruan City) using a space-time model namely Generalized Space-Time Autoregressive (GSTAR). Considering that COVID-19 is very easy to spread not only depending on the time but also the proximity between locations, the GSTAR method is good enough to be used to predict the assumption of parameters between heterogeneous locations. The estimation used is OLS with the location weight of cross-correlation normalization. The results of this study obtained the GSTAR(21)-OLS model is the best model to predict the number of positive cases of COVID-19 in.five locations in East Java by weighting the normalization of cross-correlation based on the smallest RMSE value in data out sample. Forecast results for the next 10 days of positive cases of COVID-19 in.all five locations show not very significant changes. [ABSTRACT FROM AUTHOR]