基于粒子算法和支持向量机的财务危机预警模型

Hybrid Particle Swarm Optimization and Support Vector Machine for Bankruptcy Prediction

基于特征集的选择、核函数参数的优化对支持向量机(SVM)模型的预测性能有着重要的影响,提出了一个粒子算法-支持向量机(PSO-SVM)模型.该模型采用PSO对特征集和核函数参数同时进行优化,从而提高SVM模型的预测结果.将所提出的PSO-SVM模型应用到财务危机预警中,取得了较佳的预测结果.

As feature subset selection and parameters are important for the performance of SVM-based model, a PSO-SVM model was provided which uses particle swarm optimization （PSO） to optimize both a feature subset and parameters of SVM simultaneously so as to improve the prediction result. Finally, the PSO-SVM model was applied to bankruptcy prediction, which shows a better performance than pure SVM- based model.