Research On Association Rules Mining And Outlier Analysis And Their Application In Analytical Customer Relationship Management

The major reason that data mining has attracted a great deal of attention in the information industry in recent years is duo to the wide availability of huge amounts of data and imminent need for turning such data into useful information and knowledge.Data mining is one of the important approach to achieve the transition. The thesis pays attention to the association rules and outlier problem in data mining, studies probability & statistics and Liapunov theorem as well as their applications in cross sales and risk evalutions of Analytical Customer Relationship Management system.The thesis has six chapters. In chapter 1, an introduction on the definition,reason for in vogue and types of the customer relationship management and the background, definition and function of data mining as well as their applications in CRM are given.Chapter 2 summarizes all kinds of mining methods of association rules.In chapter 3,the application of probability & statistics in mining negative association rules is studied.Traditional mining methods of association rules do not consider the correlation negative association among attributes,which induces that the mining result is misleading and not full.To solve this problem, a new mining method based on probability & statistics is suggested and corresponding algorithms are given in chapter 3.Chapter 4 summarizes all kinds of mining methods of outliers.The application of Liapunov throrem in solving outliers problem is studied.The traditional methods of mining outliers is complicated in calculating or difficult to understand or not easy to confirm parameters.To solve the problems, a new mining method based on Liapunov theorem is suggested and corresponding algorithms are given in chapter 5.Chapter 6 discribes the architecture and realization or the Analytical Customer Relationship Management, adds algorithms and models discussed in former chapters to ACRM system, and accomplishes an elementary ACRM system.