离散GM(1,1)模型在服装流行色预测中的应用  

Fashion color prediction based on discrete GM(1,1) model

作　　者：周捷[1] 李健[1] ZHOU Jie;LI Jian(School of Apparel and Art Design,Xi′an Polytechnic University,Xi′an 710048,China)

机构地区：[1]西安工程大学服装与艺术设计学院,陕西西安710048

出　　处：《西安工程大学学报》2019年第1期23-30,共8页Journal of Xi’an Polytechnic University

基　　金：陕西省科技厅国际科技合作计划项目(2018KW-056)

摘　　要：针对服装流行色数据少、影响因素复杂和预测精度不高等问题,采用离散GM(1,1)模型预测服装流行色。以中国纺织信息中心发布的2013—2019年春夏流行色定案为研究对象,以HSV色彩体系为色彩量化依据,建立服装流行色的预测模型。为验证模型效果,比较GM(1,1)模型(EGM)、原始差分GM(1,1)模型(ODGM)、均值差分GM(1,1)模型(EDGM)和离散GM(1,1)(DGM)的预测结果。结果表明,DGM(1,1)模型提高了预测性能,其平均相对误差为4.493%,残差值趋于稳定,精度等级为优秀,显著优于EGM(1,1)9.622%,略优于ODGM(1,1)4.497%和EDGM(1,1)4.495%。DGM(1,1)模型建模过程简单、预测性能高,为数据序列波动较大的服装流行色预测提供了参考依据和实用指导。In view of the problems of few data,complex influencing factors and low prediction accuracy of fashion color,a discrete GM(1,1) model is proposed to predict fashion color. The study takes the final fashion color case published by China Textile Information Center in spring and summer of 2013-2019 as the research object,and HSV color system as the basis of color quantification,on which a prediction model of fashion color is established. To verify the effect of the model,GM(1,1) model is compared. The results show that the DGM (1,1) model improves the prediction performance with an average relative error of 4.493%,the residual value tends to be stable and the accuracy level is excellent,which is significantly better than 9.622% of EGM (1,1),4.497% of ODGM (1,1) and 4.495% of EDGM (1,1). The process is simple and the prediction is accurate,which provides a reference basis and practical guidance for fashion color prediction with large fluctuation of data series.

关 键 词：服装流行色 GM(1 1)模型 离散GM(1 1)模型 HSV色彩体系 量化预测 

分 类 号：TS941.7[轻工技术与工程—服装设计与工程]