| Customer Relationship Management (CRM) is an important marketing strategy formobile communication company, and the accurate custom classification is thefoundation of high efficiency of CRM in the mobile communication company.Customer classification is performed to divide the customers into different setsaccording to their attributes, so as to analyze their purchasing patterns and therefore topredict their future activities.The accurate customer classification in mobile communication company deals witha large amount of fuzzy factors, and the criterions for classification also vary fordifferent goals. In order to solve it, in this thesis an intelligent data mining is proposedto help to mine the inner rule, and provide some suggestions for the manager of mobilecommunication company. The cost of mobile communication company can beremarkably reduced by the customer classification technology. In order to solve thecustomer classification problem in JinHua mobile communication company, in thisthesis, we proposed a customer classification model based on fuzzy ISODATAclustering in the CRM system, which extract the features of customers and introduce theconcept of soft clustering membership degree. Compared with the hard clusteringmembership degree, this new defined membership degree can describe thecharacteristics of customers. Moreover, this thesis analyzed and studied the startingvalue selection method in the Fuzzy ISODATA algorithm, and used the method ofmaximal matrix element to ascertain the number of classification, through thetheoretical analysis and empirical results.Finally, the improved fuzzy ISODATA algorithm is introduced. The algorithmreduced sensitivity to the starting value. The algorithm can highly improve clusteringanalyze and obtains a stable result, so it presents an efficient way of improving thecontribution value of custom service. Subsequently it will help to establish one-to-oneservice system and further difference-oriented customer management. |
PDF Full Text Request