Construction and application of GA-SVM model for personal credit scoring

SVM is new computational technique suitable for small samples' classification. The parameters of SVM are crucial for the model's classification performance. Aiming at the randomness of the parameters determined in SVM, this paper constructed a GA-SVM model by using GA to search the optimal parameters of SVM. Through GA's fitness function to control the type II error rate, the model was used for personal credit scoring. The application results indicate that GA-SVM model gets high classification accuracy with 0.00% of type II error rates on training samples and testing samples and the model shows strong robustness.