Financial Crisis Warning Research For Listed Enterprise Based On BP Neural Network

The health of the financial situation of the enterprise will have direct affect to the survival and development of it.Especially for listed enterprises,it also determines the investor's confidence in the enterprise.The financial crisis will affect the development of enterprises,reduce the investor's confidence for the enterprise but increase the difficulty for enterprise in bank financing.Therefore,financial crisis is a worldwide topic for research,which every country or enterprise will come across.Financial crisis is a gradual process.It is predicable due to its character of timing,though it will take different time period,either long or short,for enterprise to go from having healthy financial situation to having financial crisis.It can effectively avoid the deterioration of the financial crisis by using scientific method to predict financial crisis,and give warning for enterprise whenever it is appropriate.This has practical significance for enterprise itself,investors and regulatory authorities.Through effective financial crisis warning,enterprise can have themselves self-examined timely,reduce the possibility for enterprise of financial crisis,and meanwhile,the resolution of financial crisis will increase the confidence of investors for the enterprise and it will be easier for enterprise to obtain a bank loan.Furthermore,the regulatory authorities can make use of financial crisis warning to inform the enterprise of the possible crisis to ensure the market order.Currently,we can make use of multivariate variable financial crisis warning analysis,regression financial crisis warning analysis,etc.,to give warning for financial crisis for enterprise.And most of them are using statistic methods.It tends to have discrepancies because there will have problems as following in statistic method,including some of the financial indicators cannot meet the statistic assumptions;it is too complex to process data;there are too much computation,and so on.This article selected the domestic A-share listed manufacturing enterprises as the research object,and make use of the financial indicator data for(a-3)years from the enterprise data to predict the financial crisis situation of the enterprise for year a.at the same time,the BP neural network will be used to construct the financial crisis warning model.The biggest advantage of BP neural network is that it will find out the law between seemingly irregular data,as well as expanding the scope of application by gradually verification.This article has selected financial indicator data from 101 domestic A-share listed manufacturing enterprises,and 90 of them are training data with 15 enterprises which had financial crisis and 75 enterprises which have healthy financial situation,sampled according to ratio of 1:5.After finished training of the model,data from 15 enterprises which had financial crisis and 75 enterprises which have healthy financial situation will be used for verification of this model.The result tells that the model has an accuracy of 90%,and the new financial crisis warning model created and trained using BP neutral network is effective for listed manufacturing enterprises compared with logistic regression model and has practical value and prospect among them.Lastly,the article will take TBEA Stock Co.Ltd.as a case analysis,to analyze the financial indicator data from 2010 to 2012,based on which the article will have prediction and verification.