Research On Key Technologies Of Bank Customer Relationship Management Based On Data Mining

It is a pressing mission for financial industry to make research on how to implement data warehouse and data mining technology in CRM (Customer Relationship Management) aspect. The related research aspect includes data warehouse and data mining technologies, construction of CRM systems and design of more effective data mining algorithms. This paper discusses some of the key technologies in the implementation process of CRM systems in financial industry ,meanwhile studies the bank CRM technology which basing on data-mining in detail.Typical applications of traditional data mining technology in financial industry include classification of bank customers, precaution of bank customer loss, financial fraud analysis and data-mining based bank credit card analysis. This paper presents a summary of these applications and analyzes those using concrete cases. This paper presents a brief summary and analysis toward these traditional applications, and show a new conclusion and solution basing on the practical specific case.The chapter 3 of the paper focuses on the clustering analysis of bank customers. By comparing the decision tree approach and rough set approach, this chapter presents a new clustering algorithm for bank customers based on multi-variable decision, which is a combination of rough set approach and decision tree approach. The test results prove this approach enhances the efficiency of clustering analysis in some degree.Temporal data mining and analysis in financial industry is one of the hot research areas at present. This paper attempts to convert temporal data into strings and uses related string pattern match algorithms in the analysis of the financial temporal data. In the research of string algorithms, attentions will usually be given to the regularities in the strings, e.g. repetition. SEED is a general repetition in that if a super-string of a given string can be constructed by the combination of the substring of this given string, the substring is a Seed of the given string. This paper discusses the creation of Seeds of a given string in temporal sequence prediction and proposes an effective algorithm to compute all the Seeds of a given string of length n, with a computation complexity of O(nlogn).Agent and MAS (Multi-Agent Systems) are catching the research attention as important technologies in the application of AI and intelligent software in distributed computing. Supplying software entities that incorporate domain knowledge, act according to the psychological states and provide semantic interactions, cooperation and coordination, will not only provide strong supports for effective coordination but also pave the ground for a new distributed computing oriented environment with the character of open, reconstruct ability and scalability. This paper studies MAS-related theories and proposes a customer sales coordination framework by applyingmulti-Agent coordination framework in cross-sales analysis of bank customers.In the chapter 6 of this paper, analyses of the data mining related technologies and their practical application domains are provided from a technological and application perspective to problems arising from the project implementation. By combining the application background of Shanghai Pudong Development Bank, this paper presents a design for a bank customer analysis system.This paper provides a helpful guidance and reference to different application domains including applying structural and non-structural data mining in domestic financial industry, deploying enterprise intelligence, CRM, market analysis, competition analysis and dynamic market demand analysis etc.