电子商务环境下基于点击流的需求预测模型和企业库存优化

Clickstream-based demand forecasting model andcorporation inventory optimization under E-commerce environment

基于电子商务平台中较有代表性的两大数据:点击量和购买量数据之间的关系,通过点击量和购买量的历史信息进行需求预测,并对需求预测模型的有效性进行理论证明。以电商企业为研究对象,以库存成本最小化为目标,运用单阶段报童模型对电子商务企业的库存模型进行优化,证明模型的有效性、适用条件及其库存成本降低的概率下限,分析适用于这一优化模型的企业特征,并为企业提出相应的库存策略。基于天猫平台实际数据,对加入需求预测模型的报童模型的性质进行了实证研究。结果表明,并非所有电商企业都能通过应用点击-购买需求模型对库存管理进行优化,电商企业应根据企业自身特征采用相应的库存策略以降低其库存成本。

Based on the relationship of two represent- ative data sets in E-commerce： data of clicking times and buying times, the demand was predicted dependent on historical clicking and buying information, and the forecasting model was theoretically validated. Considering E-commerce corporation with the minimum inventory cost as the optimal objective, this paper optimized the single period newsvendor model of E-commerce corporation, validated the model, its applicable criteria and the minimum probability of inventory cost reduction, analyzed the characteristics of corporation applying to this model, and also provided the corresponding inventory strategy for E-commerce corporations. By using data from Tmall, this paper launched an empirical research to validate the new model and its propositions. The results show that not every E-commerce corporation can improve the inventory management by applying the click-buy demand forecasting model, and E-commerce corporations should reduce inventory cost considering their own characteristics.