The empirical study of individual housing loan credit risk based on proximal support vector machine

Abstract

The purpose of this paper is to reduce the default rate of personal housing loan and accurately predict whether or not the borrower defaults. Based on the data of individual housing loan, this paper employs a proximal support vector machine (PSVM) to explore the credit risk factors. Then the paper constructed the credit risk assessment system of individual housing loan. The data of individual housing loan was from China Construction Bank of Shaanxi branch in Xi'an market. The empirical results not only show that PSVM can accurately predict credit risk assessment of personal housing loan, but also can quickly and accurately judge whether or not the borrower break a contract. © 2011 Springer-Verlag Berlin Heidelberg.