摘要
针对热定形大样实验数据较少的问题,提出以基于最小二乘支持向量机(1 east squares support vectorm achine,LS-SVM)方法建立棉/氨纶弹力布热定形效率预测模型,该方法基于统计学习理论的原理,能较好地解决小样本、非线性的学习问题。将该方法与传统的多元非线性回归方法进行比较,试验结果表明,该方法具有更高的精度,验证了LS-SVM对热定形效率预测建模是一种可行且有效的方法。
Based on least squares support vector machine （LS - SVM）, a heat - setting efficiency model of cotton knitted fabric containing spandex was proposed in accordance with the features of small samples prediction. The model was established by the principle of statistical learning theory, so that problems contain small samples and non - linear study could be easily solved by this method. The presented method was compared with the traditional method of multivariate nonlinear regression. Test results showed that the former had a higher precision and verified the feasibility and effectiveness of the model for heat - setting efficiency.
出处
《武汉理工大学学报（信息与管理工程版）》
CAS
2010年第1期88-90,95,共4页
Journal of Wuhan University of Technology：Information & Management Engineering
基金
福建省科技计划重点基金资助项目(2009H0031)
关键词
热定形效率
LS—SVM
多元非线性回归
heat - setting efficiency model
LS - SVM
multiple nonlinear regressions
