This paper describes an approach for predicting the diameter of a jet-grout column using the support vector regression(SVR)technique,which is regarded as a novel learning machine based upon recent advances in statisti...This paper describes an approach for predicting the diameter of a jet-grout column using the support vector regression(SVR)technique,which is regarded as a novel learning machine based upon recent advances in statistical theory,in which the combined effects of the construction(construction methods and jetting parameters)and soil properties(soil type and shearing resistance)are considered.Four different kernel functions,namely,a linear kernel function,polynomial kernel function,radial basis kernel function,and sigmoid kernel function,are integrated into the SVR technique.A large amount of field measured data on the diameter of jet-grout column are retrieved from the published literature for training and testing purposes.The results indicate that the SVR technique with a radial basis kernel function provides predictions closest to the measured results,whereas the prepared design charts enable the ability to significantly widen the application of the proposed approach to the areas of ground improvement and environmental protection.展开更多
基金The Project Supported by Natural Science Basic Research Plan in Shaanxi Province of China(Grant no.2019JQ-114)the National Nature Science Foundation of China(NSFC)(Grant nos.41702287 and 41807245)the Fundamental Research Funds for the Central Universities(Grant no.300102218517).
文摘This paper describes an approach for predicting the diameter of a jet-grout column using the support vector regression(SVR)technique,which is regarded as a novel learning machine based upon recent advances in statistical theory,in which the combined effects of the construction(construction methods and jetting parameters)and soil properties(soil type and shearing resistance)are considered.Four different kernel functions,namely,a linear kernel function,polynomial kernel function,radial basis kernel function,and sigmoid kernel function,are integrated into the SVR technique.A large amount of field measured data on the diameter of jet-grout column are retrieved from the published literature for training and testing purposes.The results indicate that the SVR technique with a radial basis kernel function provides predictions closest to the measured results,whereas the prepared design charts enable the ability to significantly widen the application of the proposed approach to the areas of ground improvement and environmental protection.
