Credit evaluation based on support vector machine

The paper uses the learning algorithm of support vector machine to separate both 106 listed companies of China in 2000 and 80 borrowers of a national commercial bank of China in 2001 into two patterns respectively by using two different kernel functions: polynomial function and radial basis function. The experimental results show that, under the circumstance of LIBSVM, the learning algorithms of support vector machine adopted two different kernel functions have very high classification accuracy rate by selecting appropriate parameters. To the two different samples of the paper, the classification accuracy rates are all 100%.