Research And Analysis Of Data Mining In The Field Of Securities CRM System Customer Segmentation

With the deepening of the process of world economic integration and China’s accession to WTO, China’s securities industry has gained rapid development. The attendant competition has become fiercer than ever, and the traditional "product-centric" model of competition gradually becomes unable to meet customer needs, so the "customer-centric" model approaches gradually. Customer relationship management system is tailored to meet this need. It integrates advanced modern management concepts, data mining, and other related information technology. The enterprises can use it to analyze the demand patterns of existing and potential customers to win maximum benefits.First, this paper discusses the necessity and prospects of using CRM in the securities industry, and then analyzes the theory bases of customer relationship management and data mining technology, focusing on the connotation of customer relationship management and the specific content and implementation methods of data mining technology.Secondly, basing on the characteristics of the practical application on the securities industry and theoretical basis of data mining, we design a securities industry CRM system integrating data mining functionality. We emphasize on introducing the specific architecture of the system, the business processes, the data collection and cleaning, the database design, the establishment and organization of data warehouse, the data analysis and reporting implementation, CRM system specific functions and module design, etc. The practice shows that the system can meet the actual needs of the securities industry and has good scalability and robustness.Finally, an improved k-means clustering algorithm is proposed basing on in-depth analysis of the characteristics of the actual needs of the Securities CRM. The new algorithm adopts improved center initialization method to reduce the influence of isolated points and noisy points, uses the triangle inequality to reduce the number of iterations and the computational consumption. We use the improve algorithm in the securities industry to study the customer subdivision, and get an effective customer subdivision. With this subdivision, the enterprises can provide different customers with services of different levels, which can improve customer satisfaction, enhance the competitiveness of enterprises.