Credit risk assessment using regression model on P2P lending

Abstract

The financial market crashed after Lehman Shock in 2008 which creates the risk averting environment for banks. Under this kind of environment, it has become difficult for new and small businesses to have access to loans. Since then, P2P (Peer to Peer) lending becomes more and more popular around the world. However, most of the researches who had studied about credit risk on P2P lending consider only the event that the borrower will default instead of the amount of loss. In this work, we consider Net Return Rate (NRR) as the criterion to label the data for prediction training. We train the regression model to assess credit risk. The proposed model predicts the amount of profit from a borrower. In our results, by using our proposed credit risk assessment model, an investor of P2P lending can measure the risk with better accuracy and the proposed model can also predict the amount of profit from a loan.