基于多分类支持向量机的评估模型研究

Scoring Model Based on Support Vector Machine

支持向量机中的参数直接影响其推广能力,针对参数选取的主观性,提出基于改进的遗传算法优化其参数,并将其应用于银行个人信用的五等级分类问题中,针对多分类问题,设计了3个二值分类器,不同分类的参数不同,通过实验证实可以达到更精细的分类效果.

According to the problem that parameters of SVM（support vector machine） are crucial to the model generalization ability,a method to optimize SVM parameters by using improved genetic algorithm is presented in this paper and used to solve five classifications bank personal credit rating problem.For multi-classification problem,three second-class classifiers with different classification parameters are designed in this paper.The results of experiment show that the method is possible and achieving a better classification results.