文献类型：期刊文献

中文题名：基于ARIMA时间序列模型的销售量预测分析

英文题名：Forecast and Analysis of Sales Volume Based on ARIMA Time Series Model

作者：葛娜[1];孙连英[2];赵平[1];万莹[1]

第一作者：葛娜

机构：[1]北京联合大学智慧城市学院;[2]北京联合大学城市轨道交通与物流学院

第一机构：北京联合大学智慧城市学院

年份：2018

卷号：32

期号：4

起止页码：27-33

中文期刊名：北京联合大学学报

外文期刊名：Journal of Beijing Union University

语种：中文

中文关键词：ARIMA模型;销售预测;参数估计;误差分析

外文关键词：ARIMA model;Sales forecast;Parameter estimation;Error analysis

摘要：有效分析历史运营数据与未来销售量之间的关系是企业制定重大销售战略的重要支撑。针对某企业多年的产品销售量原始数据,基于时间序列模型对企业产品的销量进行预测,以ARIMA模型(Autoregressive Integrated Moving Average Model)建立预测模型得到销售量预测值与销售记录值的对应关系。从数据检验及预处理、模型识别与定阶、参数估计、模型适应性检验、模型预测和误差分析6个方面对模型的预测效果进行检验。结果表明,ARIMA(1,2,4)模型能够较好地描述销售量变化趋势,可为企业制定行之有效的销售战略和产业布局提供依据。
Effective analysis of the relationship between historical operational data and future sales is an important support for companies to develop significant sales strategies. Based on the original product sales data of a company for many years, the sales volume of the enterprise products is predicted based on the time series model, and the predictive model is established by the ARIMA model （Autoregressive Integrated Moving Average Model） to obtain the corresponding relationship between the sales volume predicted value and the sales record value. From the aspects of data inspection and preprocessing, model identification and ordering, parameter estimation, model adaptability test, model prediction and error analysis, the prediction effect of the model is tested. The results show that the ARIMA （1, 2, 4） model can better describe the trend of sales volume, which can provide a basis for enterprises to develop effective sales strategies and industrial layout.