EARTHQUAKES DATA SIMULATION BY USING COPULA METHODS AND SELF-ORGANIZING NEURAL NETWORKS (SONN) TO DEVELOP TECHNIQUES FOR FORECASTING

Abstract eng:
Recent advances made in forecasting Earthquakes using clustering analysis techniques are being run by numerical simulations. In this paper, the Gaussian Copula clustering technique is used to obtain Earthquake patterns. Copulas methods can involve recognizing precursory seismic patterns before a large earthquake within a specific region occurs. The observed data represent seismic activities situated around IRAN in the 1970-2014 time intervals. This technique is based on applying cluster analysis of earthquake patterns to observe and synthetic seismic catalogue. Earthquakes are first classified into different clusters, and then, patterns are discovered before large earthquakes via Copulas and SONN simulation. The results of the experiments show that recognition rates achieved within this system are much higher than those achieved only during the feature map is used on the seismic silence and the Doughnut pattern before large earthquakes. INTRODUCTION Pattern recognition technique has been shown to elegantly and powerfully realize solutions to problems in Seismology and earthquake forecasting. A few applications of advanced statistical methods to seismology have been carried out. For example Alexander et al. (1992), Leach et al. (1993), Tsvang et al. (1993), Taylor et. al. (1989), Allamehzadeh and Nassery (1999) have applied artificial neural networks to the explosion seismology for discriminating natural earthquakes versus explosions. The other goal of pattern recognition in seismology is to identify earthquake prone area using Gaussian Copula. Allamehzadeh (2104) make previous applications of pattern recognition to earthquake locations in 2015, for strictly predictive purposes in central Asia and Anatolia. A pattern is a suite of traits that characterizes a group of objects, such as earthquake epicenters, and distinguishes this group of objects from another group, such as places that will not be epicenters. The other methodology is developed by Allamehzadeh and Mokhtari that has been applied to many seismic regions of the world for the identification of seismogenic nodes (Allamehzadeh et al. 2003; 2009; Madahizadeh and Allamehzadeh (2011). Recent earthquakes in each of the regions studied have proved the reliability of the results obtained.. Murat and Rudman (1992), Wang (1992), Wang and Mendel (1992), McCormack et al. (1993), and Roth and Tarantola (1994) have applied artificial neural networks to reflection and refraction seismic studies. Katz and Aki (1992) have developed an approach to earthquake prediction using neural network techniques. Serono and Patnaik (1993) have applied neural networks to phase identification at threecomponent stations. Tung et al. (1994) have applied neural networks to predict the spatial distribution of the Modified Mercalli intensity for the California area. Dai and Macbeth (1995) have performed a study to test the ability of an ANN to detect and pick local seismic arrivals. In many respects the above lists summarizes the features of conventional earth science data, and are the main reasons for the increasing popularity of advanced statistical techniques in geosciences. In this paper, Copulas methods and SONN are used for pattern recognition of earthquake distribution. More recently copula function has been used in other fields such as climate, oceanography, hydrology, geodesy, reliability, evolutionary computation and engineering. By using copula theory, a joint distribution can be built with a copula function and, possibly, several different marginal distributions. Copula theory has been used also for modeling multivariate distributions. (Smith 2003; Nelsen 2006)

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket