Financial Knowledge Service Research And Platform Realization Based On Analysis Of User Investment Behavior

With the continuous development of science and technology,Internet information technology has been widely applied to people’s daily lives,"Internet finance" is the typical case of information technology to promote the development of the financial sector.At present,many Internet financial service platforms mainly provide market information services,financial transaction services and other functional services,lacking personalized and innovative services for different users and user groups.In this paper,we use data mining methods to analyze the investment behavior of users.As the user investment behavior reflects the preference of the user’s investment,the analysis result of the investment behavior can be used to achieve personalized services for different users and user groups,at the same time,by effectively summarizing,analyzing and evaluating user investment behavior,it also helps investors to find their own investment characteristics and shortcomings,and achieve some innovative financial knowledge services.The main content of this paper includes the following aspects:Acquisition and processing of financial heterogeneous Information.Any knowledge service system is inseparable from the support of information,information acquisition and processing is to the most basic aspects of knowledge services.The heterogeneous information used in this paper includes stock quotation data,financial text data and user investment data,and we adopt different methods to acquire and deal with different kinds of data.Analysis and prediction of individual investment behavior.This paper acquires characteristics and shortcomings of user investment behavior through analyzing user investment behaviors which are recorded in the real transaction data and simulated transaction data.The problem of user investment behavior prediction is defined and described.The user investment behavior is predicted by the linear model,the random forest,the gradient boosting decision tree and ensemble models,respectively.The precision is 0.51,the recall rate is 0.69,the F1 value is 0.58.All indexes are better than random prediction.User investment group discovery.User investment behavior has not only individuality,independence,but also has a group effect.In this paper,we use the stock plate as a user label,the user investment group partition is realized by means of k-means and spectral clustering algorithm,in addition,we use dimensionality reduction method and threshold method to achieve feature selection and redundancy elimination.During this process,the research of user investment group visualization based on dimension reduction method is carried out,and the distribution of user investment group is displayed in the form of scatter graph.Construction of financial knowledge service platform.With the financial heterogeneous information acquisition and pretreatment,and the intelligent computing technology and distributed information processing technology combined with stock and bond trading rules,a financial knowledge service platform is built and provides users with online simulation trading and user investment behavior analysis knowledge services,among them,simulation trading system is characterized by support for bond trading,and user investment behavior analysis system supports users to upload investment data,which make it more open.