统计过程控制在网络销售中的应用

The Application of Statistical Process Control in Network Sales

现有网络销售的研究中,人们非常关心订货量,订货周期,销售价格和利润等变量。从统计过程控制的角度,给出了监控网络销售变量参数变化的一般框架。鉴于网络销售变量分布的多样性,建立了多元的Max-EWMA控制图用来监控网络销售变量的均值向量的漂移。通过蒙特卡洛模拟比较了Max-EWMA控制图与MEWMA控制图和T2控制图的性能。以马氏距离度量漂移的大小,以AR(1)模型刻画变量间的相关性。模拟结果表明,Max-EWMA 控制图对监控中小漂移比较有效,具有较理想的性能。

In the existing research of online sales, people are very concerned about the variables such as order quantity, order cycle, sales price and profit. From the perspective of statistical process control, a general framework for monitoring the change of network sales variable parameters is given. In view of the diversity of the distribution of online sales variables, a multivariate Max-EWMA control chart is established to monitor the drift of the mean vector of online sales variables. The performances of Max-EWMA control chart, MEWMA control chart and T2 control chart are compared by Monte Carlo simulation. Mahalanobis distance is used to measure the drift, and AR(1) model is used to describe the correlation between variables. The simulation results show that the Max-EWMA control chart is effective in monitoring small and medium drift, and has better performance.