جستجو در تالارهای گفتگو

LEAST SQUARES SUPPORT VECTOR MACHINE FOR CONSTITUTIVE MODELING OF CLAY X. Zhou and J. Shen Abstract Constitutive modeling of clay is an important research in geotechnical engineering. It is difficult to use precise mathematical expressions to approximate stress-strain relationship of clay. Artificial neural network (ANN) and support vector machine (SVM) have been successfully used in constitutive modeling of clay. However, generalization ability of ANN has some limitations, and application of SVM in large scale function approximation problems is limited during optimization. In this paper, least squares support vector machine (LSSVM) is proposed to simulate stress-strain relationship of clay. LSSVM is a robust type of SVM, maintains the good features of SVM and also has its own unique advantages. LSSVM offers an effective alternative for mimicking constitutive modeling of clay. The good performance of the LSSVM models is demonstrated by learning and prediction of constitutive relationship of Fujinomori clay under undrained and drained conditions. In the present study, three versions of LSSVM models are built by considering more history points. The results prove that the LSSVM based models are superior to Modified Cam-clay model. Keywords Constitutive modeling, Artificial neural network (ANN), Support vector machine (SVM), Least squares support vector machine (LSSVM), Fujinomori clay چکیده مدل سازی ساختاری خاک رس از جمله تحقیقات مهم در مهندسی ژئوتکنیک است. استفاده از عبارات دقیق ریاضی برای تخمین رابطه تنش-کرنش خاک رس دشوار است. شبکه عصبی مصنوعی (ANN) و ماشین بردار پشتیبان (SVM) با موفقیت در مدل سازی ساختاری خاک رس استفاده شده اند. با این حال، توانایی تعمیم شبکه عصبی با برخی از محدودیتها روبروست و استفاده از SVM در مسائل تقریب تابع در مقیاس بزرگ در طول بهینه سازی محدود است. در این مقاله، حداقل مربعات ماشین بردار پشتیبانی (LSSVM) برای شبیه سازی رابطه تنش-کرنش خاک رس پیشنهاد شده است. LSSVM یک نوع قوی از SVM است که ویژگی های خوب SVM را حفظ کرده و همچنین دارای مزایای منحصر به فرد خود است. LSSVM جایگزین موثری برای تقلید مدل سازی ساختاری خاک رس ارائه می دهد. عملکرد خوب مدل LSSVM با یادگیری و پیش بینی رابطه ساختاری خاک رس Fujinomori تحت شرایط زهکشی نشده و خشک نشان داده می‌شود. در مطالعه حاضر، سه نسخه از مدل LSSVM با در نظر گرفتن نقاط تاریخی بیشتر ساخته شده است. نتایج ثابت می‌کند که مدل های بر اساس LSSVM نسبت به مدل بادامک خاک رس اصلاح شده برتری دارند. منبع دانلود 28-11-4.pdf

Application of Support Vector Machine for Crash Injury Severity Prediction: A Model Comparison Approach. Aghayan I., Hadji Hosseinlou M., Metin Kunt M. J. Civil Eng. Urban., 5(5): 193-199, 2015; pii:S225204301500031-5 Abstract The study presented in this paper investigated the application of using support vector machine with different kernel functions for crash injury severity prediction. A support vector machine model was developed for predicting the injury severity related to individual crashes based on crash data. The models were developed using the input parameters of driver’s age and gender, the use of a seat belt, the type and safety of a vehicle, weather conditions, road surface, speed ratio, crash time, crash type, collision type and traffic flow. Also, three injury levels were considered as output parameters for this study (i.e. no injury, evident injury and fatality). The overall prediction accuracy of the support vector machine model was compared to the multi-layer perceptron, genetic algorithm, combined genetic algorithm and pattern search. The results demonstrated that the constructed multi-layer perceptron’s performance was slightly better than the support vector machine for injury severity prediction. Whereas, support vector machine involves much lower computational cost than multi-layer perceptron because of using a straight forward algorithm in learning phase. The percent of prediction accuracy for the multi-layer perceptron model was 86.2%, which was higher than the support vector machine model with polynomial kernel (81.4%) followed by the combination of the genetic algorithm and pattern search (78.6%) and genetic algorithm (78.1%). The classification results of the two-level (no-injury and evidence injury/fatality) support vector machine found to be 85.3% was higher than multi-class classification (81.4%). Keywords: Crash Injury Severity Prediction, Genetic Algorithm, Multi-Layer Perceptron, Pattern Search, Support Vector Machine [Full text-PDF] منبع: [Hidden Content] J. Civil Eng. Urban., 5 (5) 193-199, 2015.pdf