基于SVM-Mercer的织物悬垂性能评估  

Assessment of fabric drape based on SVM-Mercer

作　　者：房伟[1] 林志贵[1] 黄伟志[1] 

机构地区：[1]天津工业大学信息与通信工程学院

出　　处：《天津工业大学学报》2008年第4期28-30,38,共4页Journal of Tiangong University

基　　金：国家自然科学基金资助项目(60602036);天津市高等学校科技发展基金资助项目(20051210)

摘　　要：结合纺织品悬垂性能参数的特点,在原有径向基核函数基础上,应用SVM-Mercer核进行新的核函数构造,建立了基于SVM-Mercer的织物悬垂性评估模型.文中对该模型的评估结果进行比较,并分析回归精度ε对评估结果的影响.结果表明:基于SVM-Mercer的织物悬垂性能评估模型可行,其评估精度有了一定程度的提高;ε的选取对评估结果有重要影响,在训练过程中宜首先确定.A new assessment model based on SVM-mercer is constructed by integrated the feature of fabrics drape data and the theory of Radial Basis Function （RBF） kernel. The assessment results on the model are compared with the one of reference [1] and the effect of regression accuracy ε on assessment is analyzed. The experimental resuits indicate that the assessment precision is increased and the assessment model is feasible. Meanwhile regression accuracy ε, determined firstly in the training, is important to the result of assessment.

关 键 词：支持向量机 SVM—Mercer 织物悬垂性 

分 类 号：TS101.923.9[轻工技术与工程—纺织工程]