The Research On The Application Of Data Mining In Analytical Customer Relation Management

The application of data mining in CRM (Customer Relationship Management) as a heated topic, in recent years,has already attracted extensive concern from many scholars and practitioners. CRM is a repeated cyclical process which transforms customer information into positive customer relationship. Business enterprises, by establishing convenient communication with customers, carrying out the customer care, and creating higher value for the customers, can raise the customers' satisfaction and loyalty, and then carry out higher profits and long-term development of the business. Data mining is a powerful tool exploring useful knowledge from a large amount of data, and it is one of the key techniques carrying out the CRM. Business enterprises, on the basis of the data about customer and detailed transaction, can make use of data mining to discover valuable knowledge about customers such as customer characteristics and their buying patterns, and guide practice in customer relationship management.Operational CRM pays attention to combining resources inside enterprises, and analytical CRM aims at increasing system's commercial analysis and aid decision ability. So this paper researches on the application of data mining in analytical customer relation management, and realizes the analysis decision ability in enterprises' CRM.This paper clarifies the data mining technology, CRM idea, and the application of data mining in CRM. It also introduces the RFM and customer value matrix technology. It designs and constructs the analytical MB(Marine Bunker)CRM based on the operational MBCRM. Combining RFM and customer value matrix, it defines three subject-oriented mining models such as the classification of customers' buying behavior, the analysis of customers' categorization characteristic and the analysis of customers' selling, aiming at task features and business needs. This paper proposes mining modeling algorithm, designs mining process, and then, achieves them through procedural coding, and draws a lot of heuristic rules. Finally, this paper presents a number of experience and future research prospects.