大数据背景下的商品需求预测与分仓规划

商品需求预测对于电商企业意义重大,对阿里电商平台的交易数据进行挖掘以获取有效特征,利用特征建立模型对未来两周这些商品的需求进行动态预测,并基于预测结果和成本最小的原则提出分仓规划建议.预测模型选择随机森林做回归,然后在残差分析的基础上建立报童模型求解分仓的库存规划.对特征数量众多的电商交易数据挖掘所建立的模型有助于电商企业进行有效的商品需求预测并据此制定成本更低的分仓规划.

Forecasting of Commodities Demand and Warehouse Planning Under the Background of Big Data

Forecasting of commodities demand is significant for electronic commerce enterprises.In this paper,we extract useful features by mining trading data of Ali electronic commerce platform and build a RF(random forest) model based on them to forecast demand of all the goods in next two weeks.Following minimum cost principle we build newsboy model by difference analysis to optimize predicted results and provide an inventory planning suggestion on basis of them.The models built by mining massive data with many features in this paper are helpful for those companies to get effective demand forecasting results and make lower cost warehouse planning according to them.