In order to establish a more applicable individual credit risk assessment model,a PSO-RBF neural network model is constructed,in which the parameters of RBF are trained by PSO algorithm. In this model the global searching capability of PSO and the local optimization efficiency of RBF are combined to overcome the instability of PSO and the shortcoming that trapped into local minimal easily of RBF. The application results indicate that PSO-RBF has an advantage in classification accuracy and assessment accuracy. Therefore, PSO-RBF applies to individual credit risk assessment, showing a good application value.
引用
收藏
页码:493 / 498
页数:6
相关论文
共 2 条
[1]
Henley W.E., 1995, THESIS OPEN U MILTON
[2]
Kennedy J, 1995, 1995 IEEE INTERNATIONAL CONFERENCE ON NEURAL NETWORKS PROCEEDINGS, VOLS 1-6, P1942, DOI 10.1109/icnn.1995.488968