支持向量机对羊毛/铜氨纤维混纺形状记忆织物的记忆性能预测  

Predicting shape memory property of wool/cuprammonium rayon blended shape memory fabric using support vector machine

作　　者：金关秀[1] 胡金莲[2] 吕晶[2] 王志均 JIN Guanxiu;HU Jinlian;LV Jing;WANG Zhijun(Zhejiang Industry Polytechnic College,Shaoxing,Zhejiang 312000,China;The HongKong Polytechnic University,Hong Kong 999077,China;Zhejiang Shaoxing Yinqiao Textile Co.,Ltd.,Shaoxing,Zhejiang 312016,China)

机构地区：[1]浙江工业职业技术学院,浙江绍兴312000 [2]香港理工大学,香港999077 [3]浙江绍兴银桥纺织有限公司,浙江绍兴312016

出　　处：《毛纺科技》2019年第7期6-10,共5页Wool Textile Journal

摘　　要：为有效预测羊毛/铜氨纤维混纺形状记忆织物的记忆性能,以形状记忆纤维含量、织物总紧度和组织系数的不同组合制备了27种织物试样,并进行形状记忆性能的测试。输入形状记忆纤维含量、织物总紧度和组织系数,建立支持向量机模型,对织物形状记忆性能,即织物经、纬向的形变固定率和形状回复率进行预测。以交叉验证法对模型结构参数进行优化。结果显示:该模型对各项形状记忆性能指标值的预测精度均高于98%,而变异系数则低于2%,后续的验证实验印证了支持向量机模型具有很强的预测性能。To predict the shape memory properties of wool/cuprammonium rayon blended shape memory fabric effectively,twenty seven fabric samples were produced by varying shape memory fiber content,fabric tightness and weave texture coefficient.The shape memory properties of these samples were measured.Shape memory fiber content,fabric tightness and weave texture coefficient being chosen as the inputs,a model based on support vector machine was established to predict the shape memory properties,that is,the deformation fixity ratio and the shape recovery ratio for both warp and weft direction of the fabrics.Cross validation was applied to optimize the structural parameters of the support vector machine model.The results showed that the values of prediction precision for all shape memory properties were higher than 98%,and all their variation coefficients were lower than 2%.The result of the verification experiment confirmed the strong predictive power of the support vector machine model.

关 键 词：支持向量机 羊毛/铜氨纤维混纺织物 形状记忆纤维 形变固定率 形状回复率 预测模型 

分 类 号：TS131.8[轻工技术与工程—纺织材料与纺织品设计]