摘要
In order to improve the efficiency of the support vector machine (SVM) for classification to deal with a large amount of samples, the least squares support vector machine (LSSVM) for classification methods is introduced into the reliability analysis. To reduce the coraputational cost, the solution of the SVM is transformed from a quadratic programming to a group of linear equations. The numerical results indicate that the reliability method based on the LSSVM for classification has higher accuracy and requires less computational cost than the SVM method.
In order to improve the efficiency of the support vector machine (SVM) for classification to deal with a large amount of samples, the least squares support vector machine (LSSVM) for classification methods is introduced into the reliability analysis. To reduce the coraputational cost, the solution of the SVM is transformed from a quadratic programming to a group of linear equations. The numerical results indicate that the reliability method based on the LSSVM for classification has higher accuracy and requires less computational cost than the SVM method.
基金
supported by the National High-Tech Research and Development Program of China (863 Program) (No.2006AA04Z405)