基于粒子群协同优化算法的供应链金融信用风险评价模型

摘要

针对供应链金融模式下信用风险评价精度受信用特征子集与模型参数影响的问题,提出一种粒子群协同优化信用风险评价模型.该模型在充分论证供应链金融风险特征指标体系的基础上,利用二进制粒子群算法优选特征子集,并对支持向量机(SVM)参数协同优化.对供应链金融信用风险评估进行实验,并与传统径向基支持向量机和主成分分析特征抽取方法对比,结果表明,该模型优选的特征子集和SVM参数能显著提高信用风险评价精度.%Aiming at the problem that the accuracy of credit risk evaluation of supply chain finance mode was affected by credit feature subset and model parameters ,we proposed a credit risk evaluation model with particle swarm cooperative optimization . On the basis of fully demonstrating the characteristic index system of supply chain financial risk ,we used the binary particle swarm algorithm to optimize the feature subset and optimize parameters of support vector machines .We carried out an experiment on the risk evaluation of supply chain financial credit ,and compared it with traditional radial basis support vector machines and feature extraction method of principal component analysis . The results show that the selected feature subset and SVM parameters of the proposed model cansignificantly improve the accuracy of credit risk evaluation .