In order to improve the prediction accuracy of compressive strength of concrete,103 groups of concrete data were collected as the samples.We selected seven kinds of ingredients from the concrete samples, using Grid-SV...In order to improve the prediction accuracy of compressive strength of concrete,103 groups of concrete data were collected as the samples.We selected seven kinds of ingredients from the concrete samples, using Grid-SVM, PSO-SVM, and GA-SVM models to establish the prediction model of cubic meter compressive strength of concrete.The experimental results show that SVM model based on Grid optimization algorithm,SVM model based on Particle swarm optimization algorithm,SVM model based on Genetic optimization algorithm mean square error respectively are 0.001, 0.489 8, and 0.304 2, correlation coefficients are 0.994 8, 0.994 6, and 0.993 0. It is shown that cubic meter compressive strength prediction method based on Grid-SVM model is the best optimization algorithm.展开更多
基金Funded by Natioanl Natural Science Foundation of Chin a(Nos.2012BAJ11B00,41301588,41471339,41571514)the Center for Materials Research and Analysis,Wuhan University of Technology
文摘In order to improve the prediction accuracy of compressive strength of concrete,103 groups of concrete data were collected as the samples.We selected seven kinds of ingredients from the concrete samples, using Grid-SVM, PSO-SVM, and GA-SVM models to establish the prediction model of cubic meter compressive strength of concrete.The experimental results show that SVM model based on Grid optimization algorithm,SVM model based on Particle swarm optimization algorithm,SVM model based on Genetic optimization algorithm mean square error respectively are 0.001, 0.489 8, and 0.304 2, correlation coefficients are 0.994 8, 0.994 6, and 0.993 0. It is shown that cubic meter compressive strength prediction method based on Grid-SVM model is the best optimization algorithm.
