Design And Realization Of Precision Marketing System Based On Data Mining Technology

As the trend of modern marketing, Precision Marketing is more and more popular to the business by precision customer-orientation, high-technology procedure and widespread application, along with the development of database, network and other computer technology. However, the technological cost is often prohibitive for SME, therefore, it's meanful to design and implement a platform for SME to achieve low-cost precision marketing. At the same time, as a bridge of users and companies, Best Tone system of China Telecom has accumulated vast record datas and the platform is based on knowledge discovery of the record data.In this thesis, the Precision Marketing System (PMS) was designed to providing the SME a platform for precision marketing. The system handles the call record which was obtained from the China Telecom's database by data analysis and mining technology. Also, based on Best Tone, some application services for SME were designed and clients can customize services to achieve precision customers and marketing with low cost through SMS or voice.Transaction database contains tranactions consisted by industry categories of interest, we chosed single-layer and boolean association rules and mined associative categories by classical association rules mining algorithm, Apriori. The mining results were applied in the services which spread an industry's marketing business to the users of the associative industries. In addition, the increasing record datas induced updating of transaction database, then led to the problem of incremental updating for Association Rules. After studying the incremental data mining algorithms, Fast Updating algorithm (FUP) and Pruning and Fast Updating algorithm(PFUP), the thesis presented an improved parallel, pruning and fast updating (PPFUP) algorithm to solve these problems of PFUP overloading the processor and decreases executing efficiency. PPFUP algorithm bases on shared-nothing workstations, partitions the locally-large itemsets and assigns them to different workstations, and synchronizes the tasks of computing on two databases to abate the stress of memory.Considering the gradual expansion of services, the system was designed to be B / S five-layer system framework based on the J2EE. System used JAVA language, JDBC specification, and the XML (extensible Markup Language), so that system can be upgraded smoothly and be interated with a variety of other systems.