摘要
提出了一种将经过改进的离散粒子群(PSO)算法和支持向量机(SVM)相结合的算法,以选择最优的指标集并用于财务危机预警。将此算法应用于上市公司的数据,检验模型提前3年的预警效果,最后与常见的主成分分析方法与SVM相结合的模型进行对比,结果证明了PSO-SVM模型的合理性和优越性。
Financial crisis warning model is proposed by combination of modified particle swarm optimization with support vector machine. Taking the choused listed companies＇ data as the model sample, a test is given to compare the proposed method with the present PCA-SVM model. The test results illustrate the reasonability and superiority of the proposed method.
出处
《管理学报》
2007年第5期588-592,共5页
Chinese Journal of Management
关键词
财务危机预警
上市公司
粒子群算法
支持向量机
financial crisis warning
listed company
particle swarm optimization
support vector machine
