سال انتشار: ۱۳۸۹
محل انتشار: اولین کنفرانس ملی محاسبات نرم و فن آوری اطلاعات
تعداد صفحات: ۳
Elaheh Haji Sotoudeh – Sama Organization affiliated with Islamic Azad University_ Torbat jam Branch, Department of Computer Engineering ,Torbat jam, Iran
An answer for the question: whether big earthquakes are predictable at all? is researched. Seismic events belong to the phenomena that every time has addressed with high complex problems. At the other hand, fractal properties of earthquake statistics established beyond doubt and are in agreement with modern models of seismicity. In this paper, an analysis of spatial and temporal changes of two parameters _d_value and b_value_ of seismic regime carried out. d_value is the fractal correlation dimension of the earthquake distribution, and b_value is the slop of the seismic recurrence curve. Our data is the earthquakes occurring at the window that surround turkey from 20 years ago up to now. Unlike the previous works, we try to find a relation between d_value and b_value variations in time and earthquake size variation at the same time, via Neural Networks. However researchers attempt to discover this relationship in some mathematical manners, but because of complexity of the phenomena, using a model free tool such as neural networks, seem to be beneficial. The possibilities to use the variations of the parameters of seismic regime to forecast earthquakes are discussed.