FORECASTING OF LARGE PEAK GROUND ACCELERATION USING COACTIVE ANFIS and SVM NEURAL NETWORK
عنوان دوره: هجدهمین کنفرانس ژئوفیزیک ایران
نویسندگان
1دانشگاه سمنان
2دانشجو پژوهشگاه زلزله شناسی
3Seismology (IIEES) 19395/3913,Tehran, Iran
4هیات علمی
چکیده
Strong ground motions have important affects for the site, As a practical engineering design. specially acceleration Strong ground motions is one of the key factors in potential analysis
the destruction is caused by earthquakes.
In this paper, to estimate maximum peak ground acceleration in an area, two artificial neural network was used, by name Coactive Anfis and Support vector machine neural network , In C-Anfis network , Because the rules of fuzzy logic are combined with neural algorithms, One Strong network with high flexibility can be resulted .
After different tests, Coactive ANFIS (C-ANFIS) network has maximum output correlation coefficient (0.82), Also has the least mean square error (LSSE=0.075). and SVM Network has maximum correlation coefficient (R=0.9955) and (LSSE=0.0014) . Therefore , these two neural network are good neural network which can estimate possible peak acceleration more than (1g) in an area.
کلیدواژه ها
 
Title
FORECASTING OF LARGE PEAK GROUND ACCELERATION USING COACTIVE ANFIS and SVM NEURAL NETWORK
Authors
mohammadbagher nasrollahnejad, ALI Nasrollahnejad, Mostafa Allameh Zadeh, Golam Javan Doloei
Abstract
Strong ground motions have important affects for the site, As a practical engineering design. specially acceleration Strong ground motions is one of the key factors in potential analysis
the destruction is caused by earthquakes.
In this paper, to estimate maximum peak ground acceleration in an area, two artificial neural network was used, by name Coactive Anfis and Support vector machine neural network , In C-Anfis network , Because the rules of fuzzy logic are combined with neural algorithms, One Strong network with high flexibility can be resulted .
After different tests, Coactive ANFIS (C-ANFIS) network has maximum output correlation coefficient (0.82), Also has the least mean square error (LSSE=0.075). and SVM Network has maximum correlation coefficient (R=0.9955) and (LSSE=0.0014) . Therefore , these two neural network are good neural network which can estimate possible peak acceleration more than (1g) in an area.
Keywords
(Strong ground motions, engineering design, Coactive Anfis neural network, Least mean sum square error, Support vector machine Neural Network )