Abstract

Whether listed companies run soundly or not hasdirect impact on development of capital market, therefore,how to forecast financial crisis of listed companiesaccurately has been a widespread topic. Essentially financialcrisis of listed companies is mainly about model patternclassification. Considering that Particle SwarmOptimization (PSO) and Support Vector Machine (SVM)have great performance and strength on classification andregression analysis, this paper puts forward the hybridforecast thought by combination of the two methods aboveas research focus. Firstly via building performanceindicators, the forecast model based on classification isestablished and related parameters are optimized by PSO.Then empirical financial crisis analysis will be conducted onthis method using financial data of listed companies. Thesimulation results indicate that the forecast modelestablished in this paper combines the strength of artificialintelligence and statistics, and can avoid phenomenon ofover fitting and under fitting compared with traditionalmodels. Moreover, with strong generalization ability, themodel is accurate and universal, hence having highapplication value.