基于NP算法的CRM中客户识别特征的选择  

Feature selection for customer recognition in CRM based on nested pratitions algorithm

作　　者：路晓伟[1] 蒋馥[1] 侯立文[1] 

机构地区：[1]上海交通大学管理学院

出　　处：《系统工程学报》2005年第6期600-605,共6页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(编号:70271038)

摘　　要：客户识别对于CRM的实施具有重要意义,客户特征选择是客户识别中的重要问题.嵌套分割算法(NP算法)是一种新型的系统优化方法,通过对其四个算子进行确定,将其应用于具有组合优化特征的CRM中客户识别中的客户特征选择问题.并通过将NP算法应用于某人寿保险公司的客户特征选择问题,说明了该方法的有效性.该方法不但能够保证以概率1收敛于最优解,而且能够提高客户特征选择的效率.Customer recognition is of great importance to the implementation of CRM（Customer Relationship Management）, and the selection of customer features is an important problem in customer recognition. Nested Partitions（NP） Algorithm is a new kind of algorithm of system optimization. By identifying its four operators, the algorithm is used to solve the selection problem of customer features in the customer recognition in CRM that has the properties of combinatorial optimization. The validity of the method is demonstrated by its application to the customer features selection problem of a life insurance company.The NP algorithm can not only converge to the global maximum of the problem with probability one, but also improve the efficiency of the customer features selection.

关 键 词：客户关系管理 客户识别 嵌套分割算法 特征选择 模式识别 

分 类 号：F273[经济管理—企业管理]