基于ARIMA构建SWECPX模型解决电商需求预测问题

• 作者： 向俊坤 ¹ 郁佳幸 ¹ 高贺 ¹ 孙浩翔 ²
• 作者单位：
  1.  中国人民解放军空军预警学院

  2.  上海交通大学
• 发布时间：2024-10-17

摘要: 本文针对电商需求预测问题，基于促销节日因素S(Sale)和仓库因素C(Category)，借助Matlab、Excel软件进行数据预处理，以ARIMA时间序列模型为核心，建立SWECPX(Sale Ware Effect Category Product X)模型，使用Matlab软件中的X-12-ARIMA选项等方法进行求解，实现了对商品需求量的准确预测，取得较好的1-wrmape指标测试效果。本文最大的创新点是提出了SWECPX模型，对影响商品需求量的要素S和C进行区分和求解，使对商品需求量预测更加精确，1-wrmape值较高。当每日的商品需求量处于较低水平时，预测效果的提升尤为显著，其预测值几乎与实际值相同。因此，我们期望SWECPX模型可以为电商仓储平台的决策提供切实的参考和借鉴。

Abstract: This paper addresses the problem of e-commerce demand prediction, based on the promotion festival factor S (Sales) and warehouse factor C (Category), using Matlab and Excel software for data preprocessing. With the ARIMA time series model as the core, an SWECPX (Sales Ware Eff ect Category Product X) model is established. The X-12-ARIMA option in Matlab software is used to solve the problem, resulting in accurate forecasting of commodity demand and good 1-wrmap indicator testing results. The biggest innovation of this paper is the proposal of the SWECPX model, which distinguishes and solves the factors S and C that affect the commodity demand, making the forecasting of commodity demand more accurate, with a higher 1-wrmap value. When the daily demand for goods is at a low level, the improvement in forecasting accuracy is particularly significant, with the forecasted values almost identical to the actual values. Therefore, we expect the SWECPX model to provide practical references and insights for decision-making in e-commerce warehousing platforms.

• 分类： 应用经济学 >> 贸易经济
• 稿件状态： 已在期刊出版
期刊名称： 《中国商论》 2024年 第8期
• 引用： PSSXiv:202410.01496 (或此版本 PSSXiv:202410.01496V1)
  DOI:10.19699/j.cnki.issn2096-0298.2024.08.029
  CSTR:32012.36.PSSXiv.202410.01496
• 推荐引用方式： 向俊坤,郁佳幸,高贺,孙浩翔.基于ARIMA构建SWECPX模型解决电商需求预测问题.中国商论,2024(8).[DOI:10.19699/j.cnki.issn2096-0298.2024.08.029] (点此复制)