Personalized Recommendation Based On User Behavior Analiysis Of Securities And Financial Products

Under the background of big data, traditional industries such as retail, finance and health begin to transform into a “data driven enterprise”, they realize that recording user’s online behavior and excavating user’s intentions and preferences can provide strong support for enterprise operation and marketing. The securities produce a large number of data that has high quality and great value. With the increased awareness of financial management, the research of user behavior in financial management also receive more attention.With the rapid development of science and technology, the information that securities users can obtain growth explosively, but our ability of selecting information is limited. Securities and users need to solve a problem, that using personalized recommendation technology to solve information overload phenomenon. This article is against this background, through the analysis of user behavior patterns, research suitable recommendation algorithm of financial products, and then form a complete personalized recommendation system of securities finance.I complete three main works in this thesis:(1) Analyze the user’s behavior of buying financial products, count the number of purchase, financial product sales and study the activity of users. Analyze the user’s behavior through human dynamics, and then find out that the behavior pattern mainly shows the characteristics of “strong paroxysmal and weak memory”, just as other human behaviors.(2) Study collaborative filtering algorithm and hybrid algorithm, and apply them into the personalized recommendation system of securities financial products. And then put forward two extension recommendation strategy based on the actual usage scenario, which are hot recommendation based on user clustering and personalized recommendation based on user’s real-time behavior.(3) Participate in the design and implement of personalized recommendation system, detail the framework design of personalized-financial-product-marketing module, and the process of offline analysis and online recommendation, response to users within seconds through incremental recommendation. Achieve the promotion in the recommendation results both from algorithm and marketing aspects.