Research On The Rating Model Of Credit Users' Credit In The Industry Of Network Loan

The borrower's credit assessment has become a hot issue of widespread concern in the International lending market.At present,the evaluation methods of credit loans both at home and abroad are vulnerable to fraud during the data acquisition process,affecting the correctness of the the credit rating of borrowers and lacking a credit rating method that has stability over a specific period of time.In this paper,the credit evaluation method of borrowers in mutual-money industries is studied with the goal of establishing a low-cost,difficult to be tampered with and imperceptible credit evaluation method.Firstly,we introduce five theories of user credit calculation model:homogeneity principle,small-world theory,semantic community depicting users,the mapping of virtual reality space and the principle of information complementarity.Based on these five basic principles,a virtual space model for portraying user portraits,a realistic space model for constructing assets capabilities of users in real life,and a virtual-reality combination model for providing information complementary mechanisms and information interaction are proposed.Then this paper models the content,behavior and emotion of users in the virtual space by extracting the available information from mass virtual space data and correctly analyzing user level.The user content information mainly analyzes the content of the browsing history of the user and the published text of the social platform,and the content text is clustered to obtain the theme of the user under different credit ratings.The user behavior information mainly explores the behavior patterns of the users in the virtual space and constructs the behavior network for the users' comments and forwarding behaviors on the social platforms.The user's emotional information mainly analyzes the six-dimensional emotions of love,joy,anger,scaredness,fear and sadness on the social platforms.Next,in view of the problem that the parameters used by user credit in real space are easily deceived,four kinds of centrality of users in social network behavior are proposed as the evaluation index to measure the credit behavior network of users.Four kinds of centrality are used to describe the user's topological structure,and the classic classification algorithm is used for the classification of user credit rating.The questionnaire is used to get the emotion of the label,and the method of synonym expansion is applied to obtain the vocabulary with certain emotion.Event side and emotion calculation model is constructed and the influence of user's inner emotional world on user's credit rating is studied.Finally,this paper empirically analyzes the differences in the topics of interest of different credit rating users,distinguishes users with different credit ratings by extracting four central features of users,and studies the characteristics of emotional trends of users in different credit ratings in virtual space,and help the company to complete credit rating and loan quota evaluation without user's perception.