人工神经网络在生丝抱合与其结构关系研究中的应用  

Application of Artifical Neural Network to Study of Relationship between Raw Silk Cohesion and Structure

作　　者：蔡煜东[1] 姚林声 

机构地区：[1]中国科学院上海冶金研究所

出　　处：《纺织学报》1993年第10期16-18,共3页Journal of Textile Research

摘　　要：本文运用人工神经网络的“反向传播”模型,研究了生丝抱合与其结构之间的关系,并选取一组标样进行了具体研究,其最大拟合相对误差不超过4.2%,最大预测相对误差不超过1.8%,结果表明:该方法性能良好,司望成为生丝抱合定量研究的一种有效手段。The relationship between raw silk cohesion and structure has been studied with the back-pro-pagation model,which is used as one of the typical artifical neural network models in this paper.And a group of samples has been selected as an item of the study,its maximum fitting rclativecrror doesn't exceed 4.2%,and a maxmium predicating relative error is within 1.8%.The resultsindicate that the perfcrence of the neural network approach is good,and therefore it might bereferred as an effective technique for quantitative study of raw silk cohesion.

关 键 词：人工神经网络 生丝 抱合 结构 

分 类 号：TS143[轻工技术与工程—纺织材料与纺织品设计]