Research On Data Mining Classification Technologies And Its Application In CRM

In recent years, the core ideas in modern marketing concepts and business operations have been transferring to Customer Relationship Management (CRM). CRM is a process of transforming the information of customers to positive customer relationship. With the growing complexity of customer information and the significant accumulation of customer data, they gradually become key factors in business success that to analyze complicated customer data, to find out the trends of customer behavior and to create real value of customers for the firms. The decision-makers strongly desire to transform massive data to valuable information and knowledge. Data mining technologies supply powerful supports to such needs. The effective application of data mining to CRM can get valuable knowledge and rules, from massive customer data, for decision making in firms. Classification methods are the most popular data mining technologies applied to business and industry right now. Classification and its application in CRM have become the common areas of concern from both scholars and entrepreneurs.This dissertation firstly introduces the fundamental theories of CRM, data mining and the application of data mining technologies in CRM, and recommends the detailed process for the application of data mining methodologies to CRM specified problems.Then, this thesis introduces the general process of classification and frequently used classification algorithm, as well as systematic summary and conclusion of currently existed evaluation indexes of classification algorithm. To recover the shortfalls of those indexes in dealing with specific problems of CRM, this thesis creates a new systematic index of evaluation, which supplies some new ideas to the evaluation of data mining algorithms. This thesis also proposes a new classification algorithm, based on thorough study of current algorithms and the practical needs of CRM.Finally, this thesis applies the new algorithm and evaluation index to the specific problem of customer finding in automobiles sales industry, and constructs a detailed solution for the application of classification techniques in this area. The results of real data modeling proves the scientific and practicality of that new algorithm and evaluation index. Based on these work, this thesis proposes a system framework for application of classification techniques to the CRM area. This system framework has some representativeness in solving the applied problems of CRM using data mining techniques, and supplies some new concepts and ideas.With the development of data mining classification technologies, the study of application for these techniques in CRM has increasingly important research value, which offers the CRM more market value and better application prospective in the e-commerce ear. Based on the above arguments, research work in this thesis has important academic value and practical meaning, for data mining technologies and application of them in CRM.