基于PCA与SOM神经网络算法的客户价值细分模型研究  

A Research of Customer Value Segmentation Model Based on PCA and SOM Neural Network

作　　者：胥学峰 邢德恩 宗炫君 

机构地区：[1]都城伟业集团有限公司,北京100020 [2]国网江苏省电力公司经济技术研究院,江苏南京210008

出　　处：《电气自动化》2017年第3期49-52,56,共5页Electrical Automation

基　　金：国家电网公司科技项目<基于云计算的多产业辅助决策支撑体系研究与应用>

摘　　要：针对房地产客户价值管理问题,以某大型房地产企业的普通住宅业务为研究对象,构建了基于PCA与SOM神经网络算法的房地产客户价值细分模型。首先采用PCA主成份分析法将输入变量重组为线性不相关的综合指标,然后采用SOM神经网络算法对客户价值进行聚类分析,最后针对聚类结果,分析不同目标客户群的购买模式和价值特征,旨在为房地产企业的营销过程提供决策支持。Facing real estate customer value management and taking common residence business of a large real estate enterprise as research object, this paper establishes a real estate customer value segmentation model based on PCA and SOM neural network. Firstly, input variables are reorganized into linear and irrelevant composite indicators in the PCA method. Then, a cluster analysis is made of customer value in the SOM neural network algorithm. Finally, with respect to cluster results, this paper analyzes purchasing patterns and value characteristics of different target customer groups so as to provide decision support for the marketing process of the real estate enterprise.

关 键 词：客户价值细分 SOM 主成份分析法 自组织性 评价指标 

分 类 号：TM744[电气工程—电力系统及自动化]