Listed Companies' Financial Crisis Early-warning Based On Bp Neural Network Research

In 2008,Financial crisis breaks out once again.The influence to the national scope is increasing unceasingly,many countries' financial system and the entity economy have been brought the varying degree impact and the harm.The financial management of every listed company has been confronted with the deeper level's challenge.Taken this as a starting point,the financial crisis early warning of listed company have been studied in this paper.In this paper,the domestic and foreign present research of the financial crisis is analyzed concretely first.The predecessor's research results are classified and integrated.Furthermore,combined with our country concrete national condition,the financial crisis's definition which suits this paper's research is fixed out. Simultaneously,because of deficiency of traditional statistical method in financial crisis early warning application,this article proposes the BP neural network that the advantage and potential had been applied in financial crisis early warning,Combined forecast method is ascertained for another breakthrough point at the same time.Thus the studying mentality has been established based on the BP neural network to research the early warning model of financial crisis for listed company.In this paper, the manufacturing industries of 110 listed companies are selected as a training sample, using principal component analysis and independent sample T test on the original financial index should be screened respectively.Thus the two new indexes are taken for the research index of the signal early warning model,which are the Fisher early warning model and the Logistic early warning model.Finally,the two single prediction result of early-warning model are taken as an input node of BP neural network,then using neural network toolbox of MATLAB7.0,we make a prediction outcome of the financial crisis.The studying result indicate:BP neural networks has overcome the disadvantages of the strictness of the assumption restriction and the demanding samples with the traditional statistical methods in the establishment of early warning models.Because of the strong operational and effective prediction,BP neural networks have a broad prospect in the application of financial crisis...