Discrimination for Non-performing Loans Recovery: A Method of Support Vector Machines Based on Wavelet Transform

According to the complexity of financial system, the model of discrimination for non-performing loans recovery based on support vector machines(SVM) and wavelet transform integrated is proposed, which in order to improve the accuracy and reliability of risk assessment. First we performed model input optimization using the wavelet transform, and then selected the radial basis function(RBF) as the kernel function of wavelet-SVM, At last, we compared with the original SVM, experimental results show that the method is a feasible and effective recognition method with higher generalization performance.