کد مقاله کد نشریه سال انتشار مقاله انگلیسی نسخه تمام متن
6443812 1640207 2014 11 صفحه PDF دانلود رایگان
عنوان انگلیسی مقاله ISI
Application of support vector machines and relevance vector machines in predicting uniaxial compressive strength of volcanic rocks
ترجمه فارسی عنوان
استفاده از ماشین های بردار پشتیبانی و ماشین های بردار مرتبط در پیش بینی قدرت فشرده سازی یکسانی سنگ های آتشفشانی
موضوعات مرتبط
مهندسی و علوم پایه علوم زمین و سیارات زمین شناسی
چکیده انگلیسی
The uniaxial compressive strength (UCS) of intact rocks is an important and pertinent property for characterizing a rock mass. It is known that standard UCS tests are destructive, expensive and time-consuming task, which is particularly true for thinly bedded, highly fractured, foliated, highly porous and weak rocks. Consequently, prediction models have become an attractive alternative for engineering geologists. In the last several years, a new, alternative kernel-based technique, support vector machines (SVMs), has been popular in modeling studies. Despite superior SVM performance, this technique has certain significant, practical drawbacks. Hence, the relevance vector machines (RVMs) approach has been proposed to recast the main ideas underlying SVMs in a Bayesian context. The primary purpose of this study is to examine the applicability and capability of RVM and SVM models for predicting the UCS of volcanic rocks from NE Turkey and comparing its performance with ANN models. In these models, the porosity and P-durability index representing microstructural variables are the input parameters. The study results indicate that these methods can successfully predict the UCS for the volcanic rocks. The SVM and RVM performed better than the ANN model. When these kernel based models are considered, RVM model found successful in terms of statistical performance criterions (e.g., performance index, PI values for training and testing data are computed as 1.579 and 1.449). These values for SVM are 1.509 and 1.307. Although SVM and RVM models are powerful techniques, the RVM run time was considerably faster, and it yielded the highest accuracy.
ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Journal of African Earth Sciences - Volume 100, December 2014, Pages 634-644
نویسندگان
,