| Recently in these years,interntet loan has rapidly developed in world wide range alonging with the development of Internet finace.For example,peer to peer(p2p)has gained a great development under the background of micro-enterprise,some representations like YIRENDAI and RENRENDAI in China.interntet loan is different from the traditional bank loan in some ways the traditional loan operates mainly through bank and online,in this way the bank can have useful personal information to build a rik controling model.While interntet loan proceeds offline and there are varied information which combines useful and useless information together,it is a tough work to distinguish those we need and the useless.So it is obvious that the traditional ways of risk control do not match the requirement of today's Internet loan.The house-loan system provides online loan service,which mainly includes loan application,qualification certificating and approval process.The current loan system processes are complecate and it costs many resource.This paper uses history data of the loan system to build a prediction model to predict the possibilityof loan cheat.With the help of the prediction model,company's lost would be well controlled in a acceptablerange.After testing,the accuracy of the prediction model is more than 95%.For default users,the prediction accuracy is more than 90%.The total loan time decreases from 8 days to 3 days,and the human resource decreases to the half of the original.In a word,the prediction model makes the loan process much simpler and much more efficient.The first chapter of this paper summarize the background and recent reseach,the second chapter analysis the requirement and design the main frame of the decision engine,the Third chapter analysis the relative techniques,the fourth chapter proceeds the data processing,the fifth chapter present the detailed design and implementation of the decision engine and the final chapter is about the test. |
PDF Full Text Request