基于M^3-DGMF的债券评级算法研究

Study on the Bond Rating Algorithm Based on M^3-DGMF

KNN、SVM等算法已经较为普遍地使用在信用评级中,但债券评级数据有着数量庞大、样本向量不一致、多分量等特征,经典算法在解决这种问题时往往局限性较大。采取双高斯合成函数的最小最大模块化神经网络的监督聚类算法对债券评级数据进行学习、泛化,实验结果表明该算法具有较大优势。

KNN, SVM and other algorithms have been widely used in credit rating, but the bond rating data has the characteris?tics of large number, inconsistent sample vectors, multi-component, and so on.The classical algorithms are often limited in solving this problem.In this paper, the supervised clustering algorithm based on the Min-Max Modular Neural Network of the double Gauss kernel mixed- function is used to learn and generalize the bond rating data.The experimental results show that the algorithm has great advantages.