摘要
针对当前纺织印染品市场对产品的销售预测需求不断变化,基于某公司海量纺织印染品销售资料数据,提出一种基于多元线性回归分析的方案,采用数据挖掘的相关算法,对纺织印染品出库数据进行回归预测分析。首先提出销售热度的概念,用以衡量某件纺织印染品在一定时间范围内的销售综合情况;其次找出影响销售热度这一因变量的影响因素,对部分含有离散型数据的如样品类型数据进行离散化处理,通过逐步回归算法,剔除掉一些影响度小的自变量比如颜色、尺寸等,最后筛选得出面料规格、特殊样式、宽度、样式这些变量,并通过数据分析,得出纺织印染品的面料规格、特殊样式、样式与纺织印染品销售热度有显著的关系,建立了多元线性回归预测模型,完成了销售预测,公司的运营数据证明了方案的有效性。
Concerning the increasing demand for sales forecasting in the field of current textile printing and dyeing products,a multi-linear regression-based prediction scheme based on data mining related algorithms was proposed to realize the regression prediction analysis of the textile printing and dyeing data of a company.Firstly,a concept of sales enthusiasm was proposed to measure the sales situation of a certain textile printing and dyeing product in a certain time range.Secondly,the factors that affect the sales enthusiation were determined,and some data such as sample type data containing discrete data were discretized.What’s more,some small independent variables such as color,size were eliminated through stepwise regression algorithm.Finally the fabric specifications,special style,width,style variables were selected;and through the data analysis,for the textile printing and dyeing products,the fabric specifications,special styles and styles have significant relationships with the sales of textile printing and dyeing products,so a multivariate linear regression forecasting model was established and the sales forecast was completed.The company s operational data proved the effectiveness of the proposed analysis system.
作者
李锋
袁琐云
LI Feng;YUAN Suoyun(College of Computer Science and Technology,Donghua University,Shanghai 201620,China)
出处
《计算机应用》
CSCD
北大核心
2019年第S02期271-274,共4页
journal of Computer Applications
基金
国家973计划项目(2017YFB0309800)
上海市自然科学基金资助项目(16ZR1401100)
关键词
纺织印染品
销售预测分析
多元线性回归预测
逐步回归算法
textile printing and dyeing product
sales forecast analysis
multiple linear regression prediction
stepwise regression algorithm
作者简介
李锋(1969—),男,江苏徐州人,教授,博士,主要研究方向:推荐算法、深度学习;通信作者:袁琐云(1995—),女,浙江杭州人,硕士研究生,主要研究方向:数据挖掘,电子邮箱sally_082213@163.com。
