Evaluation of Residential Loan by Combining RVM and Logistic Regression
Qinrong Meng · 2011
A combining forecast model is proposed to evaluate the residential loan, which improves the accuracy of a single evaluation model. Firstly, the Relevance Vector Machine (RVM) model and logistic regression model are trained by the financial data respectively. Then the weighted average rule is used to fuse these two models based on a weight training procedure. Finally, the combining model is employed to evaluate the real house loan data. The experiments show that the combining evaluation modal is super to a single model and behaves robust.