市场化取向改革中客户分档标准制定方法

Method for setting customer rating criterion in market-oriented reform

为贯彻落实2015年全国烟草工作会议精神,国家烟草专卖局扩大了卷烟营销市场化取向改革试点工作。其中,客户分档标准制定是卷烟营销工作的基础,其好坏直接影响到货源投放和新品培育工作。客户分档标准基于客户经营能力,可以根据进货量、进货金额等因素评分。根据客户评分,本文使用数据挖掘中的k-means聚类算法自动对其分类,并使用肘方法确定最佳分类数目。客户类别确定后,本文使用近邻分类kNN算法,加以适当调整,使之可根据客户商圈、业态等客户属性确定客户类别用于指导新入网客户的客户分类工作。本文提出的方法避免了依赖经验完成客户分档过程的主观性,提高了准确性,使得客户分档更加公平公正,同时也减少了客户分档的工作量。

Customer rating is buik based on their management capacity and will be graded upon their purchase quantity and their capital input in the purchase. According to customer scores, this paper classified customers automatically by using k-means clustering algorithm and determined optimum customer clustering numbers by elbow method. After customer clustering was determined, this paper utilized modified kNN algorithm to determine customer clustering based on customer attributes. This method boasted efficiency and accuracy and was designed to provide reference for new customer clustering.