采用基因库构建的季节性服装需求预测  

Demand forecasting method for seasonal clothing based on constructed gene database

作　　者：张细香[1,2] 马卫民[2] 刘建勋[3] 

机构地区：[1]嘉兴学院商学院,浙江嘉兴314001 [2]同济大学经济与管理学院,上海200092 [3]嘉兴学院图书馆,浙江嘉兴314001

出　　处：《纺织学报》2014年第5期142-148,共7页Journal of Textile Research

基　　金：国家自然科学基金资助项目(71071113);教育部人文社会科学研究青年基金项目(12YJCZH284)

摘　　要：季节性新款服装的历史销售数据少,生命周期短,故利用数据序列的趋势信息进行预测的方法不合适。鉴于服装企业的销售点(POS)数据中包含许多特征信息,利用层次聚类与动态时间弯曲距离从POS数据中进行特征信息抽取,以构建服装销售基因库。并提出基于基因库进行季节性服装需求预测的方法,根据新款服装信息、销售信息计算其与基因库中基因的相似度,再以相似度查找相似基因,用相似基因的信息预测新款服装的需求。这种方法可为服装企业利用POS数据进行定量预测,构建快速响应系统提供解决方案。最后,对某知名女装企业2011年POS数据抽取了基因库,并对其2012年3月1日到2012年7月31日的新款库存单位进行预测,结果表明该预测方法是有效的。New style seasonal clothing has these features： few historical sale data and short life cycle. Therefore, it＇s not appropriate to forecast the demand of new style seasonal clothing using conventional forecast method based on data sequence. However, there exist lots of characteristic information in apparel enterprises＇ POS data. The method of hierarchical clustering and dynamic time warping （DTW） was employed to extract characteristic information from POS data in order to build a sale gene database. And a forecasting method was put forward to figure out the demand of new style seasonal clothing based on the constructed gene database, which computed the similarity between a new clothing and a gene of the characteristic database, searched the nearest gene, and computed the demand of a new style seasonal clothing based on the information from the nearest gene. The proposed method provides a solution for apparel companies to forecast its demand quantitatively based on their POS data, and can be used to build a quick response system to provide solutions. An experiment was carried out, which extracted the gene database based on 2011 POS data of a famous women apparel enterprise, and the new stock keeping units sold between 1st, March,2012 and 31th, July,2012 were used to forecast its demand based on the constructed gene database. And the result proved that the proposed foreeasting method is workable.

关 键 词：需求预测 特征库抽取 相似度 层次聚类 动态时间弯曲距离 

分 类 号：F272.1[经济管理—企业管理]