基于用户行为研究的精确营销系统设计与实现-design and implementation of precise marketing system based on user behavior research.docxVIP

II II Abstract With the rapid development of Mobile Internet, the new Internet services continute toimpact the traditional ISP services, in this situation, China Telecom proposed shift from “call operation” to “traffic operation”and “be the leader of smart pipes”, but how to “be the leader of smart pipes” and do the “traffic operation” good, it needs to know the habit and preferences of the mobile Internet users, to treat every people individually due to their habit. Thispaperis mainly about how to obtain user preferences and accurate selling. This paper starts from the following aspects including business requirements, system design, data collection, data analysisdata mining and selling tests.It introduced the method of use DPI system to capture user’s Internet using records, and obtain user’s preferences from WEB view records with spider technology. And introduced the dimensions of customer portraits and the details of data mining based on customer portraits. This paper also introduced the real selling tests on the results of data mining, and made comparative assessment of selling results. This paper’s innovations are: it proposed an exactly way to improve the selling success rate with user behavior analysis, whichfilled the disconnect between user behavior analysis and selling. It also proposed the detailed steps of create user portraits and the specific method of data mining based on the massive Internet data, this is basically referential. In addition, the system of this paper is self-corrected, it can correct the analysis modules with user feedbacks. The subject of this paper comes from the project “Research on optimization strategies for mobile Internet traffic operation support system” and “Field experiments ofthe key technology in integrated platform for user behavior analysis”. The proposed system has been deployed in Telecom JiangSu, and used in daily operations. Key words: Data Analysis, Data Mining, User Behavior Analysis, User Portrait, Accurate Selli