基于序列算子和GM(1,1)模型的服装流行色预测  

A predictive study of fashion color based on buffer operator and GM(1,1)models

作　　者：张晓丹 周捷[1] 王振洁 ZHANG Xiaodan;ZHOU Jie;WANG Zhenjie(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：《纺织高校基础科学学报》2020年第3期38-44,共7页Basic Sciences Journal of Textile Universities

基　　金：陕西省科技厅国际科技合作计划项目(2018KW-056)。

摘　　要：针对服装流行色预测精度不高且原始数据受到冲击扰动等问题,基于中国纺织中心发布的2000—2019年春夏流行色定案中冷、暖色系所占比例,分别在长序列、宽间距序列、短序列下采用5种序列算子对原始序列进行优化,再分别与GM(1,1)模型结合,选择出预测精度最优的模型组合。结果表明,在长序列中紧邻均值生成算子和GM(1,1)模型结合预测效果较好。2019年冷、暖色系平均相对误差为0.73%,精度等级为优秀。In view of the low accuracy of fashion color prediction and the impact disturbance of raw data,based on the proportion of cold and warm color system in the spring and summer 2019 popular color fixation case released by the China Textile Center,5 kinds of sequence operators,in the long sequence,wide interval sequence and short sequence,were used to optimize the original sequence,respectively,with GM(1,1).The model was combined to select the optimal model combination of prediction accuracy.The results show that the near-mean generation operator and GM(1,1)model in the long sequence are well combined with prediction.The average relative error of the cold and warm color system in 2019 was 0.73%,and the accuracy rating was excellent.

关 键 词：服装流行色 序列算子 长序列 GM(1 1)模型 宽间距序列 短序列 

分 类 号：TS941.13[轻工技术与工程—服装设计与工程]