摘要
为了推动数据分析技术在纺织服装企业中的实践应用,从数据分析思维出发,梳理常用的定性和定量数据分析方法以及应用场景,并以时间序列预测分析为例,对某纺织服装上市公司的营业收入进行预测,根据过去11年各季度营业收入数据,预测出该公司2021年4个季度的营业收入分别为6.68、6.42、6.98、11.61亿元,模型均方根误差为0.615,与真实值6.28、6.81、6.81、11.90亿元的相对误差分别为6.42%、5.77%、2.46%、2.45%,均小于10%,说明时间序列分解预测适用于纺织服装等具有强季节同周期性的行业。
In order to promote the practical application of data analysis technology in textile and garment enterprises,started from the thinking of data analysis,commonly used qualitative and quantitative data analysis methods and application scenarios were sorted out,and time series forecasting analysis was taken as an example to predict operating income of a listed company for the four quarters of 2021,based on the data of the past 11 years,it predicted that the operating income in the four quarters of 2021 will be 668,642,698,and 1161 million yuan,the root mean square error of the model is 0.615,and the relative errors of 6.42%,5.77%,2.46% and 2.45% with the true value 6.28,6.81,6.81 and 11.90 are all less than 10%,and the relative error between the forecast data and the known real date of first quarter is 6.37%,indicating that the time series decomposition forecast is applicable to the sales forecast of enterprises with strong seasonal cycles such as textile and clothing.
作者
刘正源
王保鲁
刘凤坤
LIU Zhengyuan;WANG Baolu;LIU Fengkun(Brand Office,China National Textile and Apparel Council,Beijing 100020,China;Beijing Institute of Fashion Technology,Beijing 100029,China;China Textile Information Center,Beijing 100025,China)
出处
《毛纺科技》
CAS
北大核心
2022年第7期121-125,共5页
Wool Textile Journal
基金
北京市教委科研社科一般项目(SM202110012004)。
关键词
数据分析
时间序列
预测
纺织服装
data analysis
time series
forecast
textile and garment