Study of the Prediction of Micro-Loan Default Based on Logit Model

Abstract

With the development of network, there are more and more online loan platforms, such as LendingClub, which is a popular short-term micro-loans network platform. Since the risks of loan default have a great influence on the capital security and financial order, it is necessary to build a model which can help estimate the risk of loan before making decisions, and this paper achieved this by using Logit model. Using the Logit model for the regression analysis of data provided by LendingClub platform, this study determined the main factors affecting default risks, and got a model which can be used to check default risks of borrowers in advance. The results include the accuracy index of the model and the visualization of the relationship between the main factors, which can be used to deeper seek for the reason of loan default and put forward some improvements to the model.