Application And Research Of Self-Organizing Feature Map On Customer Classification

Customer relationship management for enterprises master the needs of customers is very important, it can help enterprises master the content of demand and customer demand trend, and improve the relationships between customers and enterprises, and effective management and mining customers resources, so as to gain more profits for enterprises. It is the one of core ideas for CRM that the customers is the most important resources, and it is the essential for enterprises that the proper master of consumer behavior. If you want to better understand customer needs, it is necessary that customer classification and provide a variety of services. There are many customer classification method. For example, RFM analysis, customer value matrix analysis, ABC analysis, and so on.In RFM analysis, when the increase in F, M also will be increased, between the F and M have are collinearity, which would reduce the effectiveness of RFM, the classification of results will have an impact on. In this paper, using (M|—) instead of M, by M /F be so good on the settlement between the F and M of the collinearity, and then use SOM network and the method of RF(M|—) combining the customer classification, according to different customers attribute to customers of the different collections, and through the different collection, analysis and projections, the consumption patterns and thus for different customers to take various measures, greater customer satisfaction, and let customers get more confidence. In RF(M|—) method, the first enterprise customers and records of transactions into RFM indicators, and then through the M get the (M|—) . After using SOM operate the conversion of data clustering, clusters are different for each customer cluster analysis, judging the tendency to deviate from the enterprise or the tendency of loyalty to help enterprises accurate positioning customer buying behavior.Currently CRM business is the largest in North America and Western Europe, domestic CRM market is not large, is still in the conceptual stage, nor did a more mature customer classification system. In view of this paper has been developed SOM customer classification system, the system is based on the data source, records of transactions from the impact of industry, have a certain commonality.