QE97 : Clustering and analyzing foreshocks for earthquake prediction using data mining methods
Thesis > Central Library of Shahrood University > Geosciences > MSc > 2012
Authors:
Pooyan Ramezani Besheli [Author], Ramazan Ramazani Oomali[Supervisor], mehdi zare[Advisor], [Advisor]
Abstarct: In regarding with the location of country of Iran in the earthquake potential of Alpine-Himalayan the earthquakes are as most destructive natural phenomena which cause humanistic economical and damages in this country. In the past periods, the destructive earthquakes were quaked different areas of our country and were the cause of many of our fellows, which highlights the importance of research in this area. This research investigates the earthquake prediction issues using the foreshocks indexes. In this purpose, first using the available quake data in the earthquake catalogs, we prepared databaxses of foreshocks of the country of Iran. After refining the catalogs and separating the foreshocks from main shocks, we performed the declustering operation algorithm using the Gardner and Noprof (10974) method on the foreshocks of country for foreshocks more than 5. After then we performed the clustering on the foreshocks more than 5 have been performed. For this section, first the self organizator algorithm of Kohnen (SOM) was used to determine the number of clusters and then the K-Means algorithm for 5-15 algorithms have been implemented and regarding the Silhouette index, 6 categories were determined for the foreshocks of country. Then the resulted categories, were grouped using the Cunkan and Tuks tests and the zoning area of Iran were determined and the Zagros - Makran transition Zone were introduces as the main foreshocks of Iran. After determining the main foreshocks, we preceded the research using artificial neural networks for two states of prediction and category. The prediction results were relative weak and with the correlation coefficient were almost 65% but the results are valid with high validity. After the modeling, we introduced the transited zone of Zagros – Makran from tectonical status view and the main Faults were introduced. After all, the earthquake parameters of this zone were investigated. The value a and b and the value of magnitude scale were 5.6, 0.8 and 3.2 respectively. The annual value of „a” was about 4.8. The investigations show the 2 years period variations of earthquake parameters for this zone.
Keywords:
#Earthquake prediction #Foreshocks #Declustering #SOM #Data mining #Artificial neural networks #Transited zone of Zagros – Makran Link
Keeping place: Central Library of Shahrood University
Visitor: