基于粗糙集和SOM网络的客户市场细分方法  

A Method for Customer Market Segmentation Based on Rough Sets and SOM Network

作　　者：秦海鸥[1] 陶俊才[2] 

机构地区：[1]盐城师范学院信息科学与技术学院,江苏盐城224002 [2]南昌大学信息工程学院,江西南昌330031

出　　处：《南昌大学学报（理科版）》2009年第4期392-394,398,共4页Journal of Nanchang University(Natural Science)

摘　　要：市场细分对提高企业核心竞争力具有重要作用。针对目前一些市场细分方法的不足,提出了基于粗糙集和SOM网络的市场细分方法。该方法首先对客户消费数据库进行离散化处理和属性约简,然后采用SOM网络进行自动聚类,形成稳定的客户群,客户群的数量可通过设定SOM网络竞争层神经元个数而改变。实验表明该方法自动化程度较高、适合海量数据的处理。Customer market segmentation plays an important role in promoting the core competition of enterprises. This paper analyzes the defeats of current market segmentation methods and puts forward a method for customer market segmentation based on rough sets and SOM network. In order to form stable customer groups, the method diseretizes customers＇ consumption database and reduce attributes, then clusters by SOM network. The number of customer groups can be changed by setting the number of competitive - layer neurons in SOM network. The results of experiments show the method has the advantages of high degree of automation and it is suited to process a flood of data.

关 键 词：市场细分 粗糙集 SOM网络 聚类 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]