基于缴费行为的用电客户细分研究

Electricity Customer Segmentation Study Based on Payment Behavior

摘要: 对用电客户缴费行为进行有针对性的细分研究，可有效提升电力公司电费收缴工作效率。研究用电客户缴费行为特点，在传统RFM模型的基础上，提出适用于电力客户缴费行为的指标体系和客户细分方法。首先，建立了基于用电客户缴费行为细分的WFM指标体系；然后，通过聚类优度分析获取最优聚类簇个数，采用SOM神经网络对用电客户进行聚类，对各类群体的属性与总体属性进行比较分析，得到客户细分结果；最后，将该方法应用于内蒙古赤峰市宁城县用电客户缴费明细数据中，提取出四类不同特征的客户群体，并分别提出相应的缴费渠道营销策略，验证了该方法的可行性和有效性。

Abstract: It is effective to improve electricity companies’ efficiency of collecting electricity charge by subdividing the payment behavior of electricity customers and adopting the improvement measures targeted. Based on the traditional RFM model, this paper proposes an index system and customer segmentation method which is suitable for the power customer system. Firstly, the WFM index system based on the electricity customer’s payment behavior is established. Then, K-means clustering method is used to obtain the optimal number of cluster with the best clustering goodness. The SOM neural network is used to cluster the customers, and the attributes of each group are compared with the overall attributes. So, the customer segmentation results are obtained. Finally, this method is applied to the detail data of electricity customer in Ningcheng County, Chifeng City, Inner Mongolia. Four different types of customer groups are extracted, and the corresponding marketing strategies of payment channels are put forward respectively. So, the feasibility and validity of this method are verified.