Research On Multi Granularity Recommendation Model Of Internet Financial Products

With the continuous development of Internet and information technology,the national advocate and actively promote the development and application of “Internet plus”.Consequently,Internet Finance emerges as the times require.The emergence of Internet financial products has greatly changed the trading mode of traditional financial products.Traditional financial enterprises are difficult to provide financial products that matching the users demand,thus it can’t satisfy the rear user of the long tail curve which demand for convenience and availability of financial services,thereby reducing its earnings.In addition,the lack of information interaction between users and financial institutions and the time-consuming and labor-consuming purchase of financial products,which leads to an increase in costs.In the state of information asymmetry,how to balance the benefits between users and financial institutions,and how financial institutions efficiently provide users with financial products and grasp users needs accurately,which has become the two major problems to be solved at present.Therefore,this paper in the incomplete information condition,the core of the research is the construction of the multi-granularity recommendation model of Internet financial products,it aims to recommend appropriate Internet financial products for users,and analyze the appropriate user groups of Internet financial products,so as to enhance both users and financial institutions’ income and satisfaction.The specific work mainly includes the following aspects:(1)Analysis on the development and research status of Internet financial products.At present,there are many researches on the development status of Internet financial products at home and abroad.However,the relevant research on the recommendation of Internet financial products was started late and the research results are relatively few.It is of great significance to research on the model of financial products recommendation because of the information of asymmetry in thefinancial market itself.(2)The dynamic game analysis of incomplete information on Internet financial products.In the state of information asymmetry,this paper studies the income of users and financial institutions.This paper constructs a two-party game model by depicting game tree diagrams of incomplete dynamic information,analyzes the one-time game process between financial institutions and users,and draws the user and financial institution’s returns under different conditions.The final analysis of high-quality financial products balanced the income of both sides,which effectively alleviates the problem of information asymmetry.(3)Analysis and extraction of characteristics of Internet financial products.The advantages and disadvantages of financial products greatly influenced on the income of users and financial institutions.This paper analyzes the properties of financial products and user preferences to control the quality of financial products.It uses Principal component analysis(PCA)and attribute granulation to extract the main attributes of users and Internet financial products,which lays the foundation for the construction of a recommended model.(4)Multi-granularity recommendation model construction.Based on the collaborative filtering recommendation algorithm,this paper finally establishes a multi-granular recommendation model based on financial products through the analysis of the recommendation algorithm.Three sets of comparative experiments were designed in this thesis to study the effects of changing the recommended number,selecting different attribute grains,and changing the granularity of attribute granularity to the recommendation result.Ultimately the experimental results shows that the recommendation model based on internet financial products has a good applicability in the Internet financial market.