Research And Application Of Customer Value For Citic Securities Based On Data Mining Model

Securities Co. financial customer value includes the current value, potential value and customer growth value. The current value is mainly refers to customers in the Securities Co. for a variety of business for the Securities Co. profit income; potential value is the present value of the customer is possible for the Securities Co.s to bring monetary value in the next period of time, the potential value is an objective existence, but not immediately available, need to dig; growth value with value changes itself. Time and customer so that the customer value to the continued growth of Securities Co. brings the value, growth value although there are many variable factors, but also cannot be ignored. Securities Co. to evaluation and calculation of customer value, must analyze business transactions and customer, the customer and the Securities Co.'s consumer records and other information, and these information scattered in different Securities Co.ing system, Securities Co. deposit management system, CRM system, credit management system, credit card business system, online Securities Co.ing system. Because the system of different construction time, data sources, formats, tre ent is not the same, customer information scattered in various business systems, data warehouse Securities Co. must use data mining technology to integrate the scattered information, establish a unified, and on the basis of analysis and mining, to find customers value.Customer value evaluation model of CITIC Securities Co. Zhejiang branch including customer current value evaluation and customer potential value evaluation of two parts. Three aspects of the value evaluation from income contribution, customer value and cost savings value evaluation; potential value, including customer future profits, customer loyalty and customer credit degree three aspects were evaluated respectively. The clustering analysis of Securities Co. customers through the K-means clustering algorithm, data mining to disperse in Securities Co. business system of customer information, dig out the group behavior of customers, find customers in common, access to customer classification rules. The results of decision tree classification model established by clustering, classification of Securities Co. customers, according to the level of customer value customers will be divided into 16 categories, summarizes the characteristics of each type of customer, to grasp their investment philosophy, the formulation of marketing strategy, the greater value of customer resources. Securities Co.s can also through the analysis of classification or clustering to group of customers, and then by the mode analysis to predict which people may become its customers, to help managers to find potential customers.Finally, The thesis take jin-billion industry company as an example, makes an empirical analysis on the customer value prediction model and classification model, through comparing the calculated value with the model's predictions, predictive ability and practical calculation model is close to the results, the prediction results to the CITIC Securities customer segmentation has a certain reference value. The results show that, based on the model of customer classification data mining technology for Securities Co. in the customer relationship management to provide a scientific and effective analysis method, provides the reliable basis for the Securities Co. decision.