Research On Theapplication Of Data Mining Technology In Banking Crm

With China's economic developing and growing rapidly, the banking environment has undergone tremendous changes. The competition among banks is becoming increasingly intense. Commercial banks gradually take customer as the center, using customer relationship management, customer knowledge management and other related theories, methods and tools to manage and service customers. Taking into account the hugeness of bank customers and the difference between clients, it is necessary to take a customer classification study, so that banks can accurately grasp the status of existing customers and take different marketing policies for different types of customers.In view of customer classification's important role in the bank, this article uses a research method combining theory and application. After pre-processing the customer credit data, we use the decision tree classification algorithm, which is the most popular data mining method, to build a customer classification mining model based on bank's CRM.The paper first use C4.5 algorithm (release 8) to build a customer classification model, we accessed the result of model forecast and found the predicted result is unsatisfactory. Then, after analyzing the characteristic of bank customer data, we found the shortcomings of C4.5 algorithm and proposed an improved algorithm named IC4.5. IC4.5 algorithm mainly improved in the following aspects: We proposed an improved program on the discreization of continuous attributes combining Fayyad's theorem. We put forward an improved discrete attribute splitting technology, which generates a binary decision tree and reduces the complexity of the decision tree to some extent. We studied and implemented two kinds of post-pruning techniques REP and EBP. Finally, IC4.5 algorithm was applied to customer classification. Through comparison and evaluation of the model effection, the customer classification model built by IC4.5 algorithm was excellent and can be used for decision support.Combining the characteristics of bank's operation and data, data mining technique was applied to the customer classification and got a good result. The result can help banks to better understand customer value, implement differential management based on customer classification and optimize the allocation of bank resources.