A Comparative Study Of Financial Distress Warning Of Listed Companies Bases On BP Neural Network And Logistic Model

With economic development, the environment of the enterprises are facing moreand more volatile in the diversification of investment and operation, increasedcompetition and technological development brought about structural friction, makingthe modern enterprise than ever facing more difficult to control the risk. Warning isparticularly important before the crisis hit. Enterprise risk is not limited to thefinancial aspects of the business failure often comes from the weak financial riskcontrol. Looking In recent years, several monster bankruptcies, given its fatal blow isoften the financial risks of runaway. Each enterprise in the course of business, shouldalways consider the financial early warning management to stifle before the crisis inthe controllable range.This paper studies the Chinese manufacturing listed companies, theoreticallymore suitable for the prediction of the BP neural network technology to establishfinancial early warning model to predict the crisis may be coming in the future,whether it is more accurate and effective.This paper first describes the theoretical basis of the financial early warningresearch defined financial risk related concepts. Clarify the basic theory, combinedwith the results of previous studies, the establishment of a financial early warningindicator system model. Follow by listed companies in China’s manufacturingindustry as a study sample financial early warning model and research. Select69non-ST and69ST companies as the study sample. Select a representative indicator offinancial and non-financial indicators. Screened by principal component analysismethod, the representative of the financial indicators, the index system wasestablished in the inclusion of non-financial indicators. Principal component analysisis used to extract representative indicators. Finally, factor analysis index system, theestablishment of the BP neural network model is used to calculate the predictionaccuracy rate. In order to compare the BP neural network in the validity of the studyof financial early warning at the same time using a logistic regression model forcomparison.The total correctly judge the rate BP neural network financial early-warningmodel is97%and the total correct judgment rate of logistic regression model is94%.In general, the BP neural network model is better. Therefore it can be concluded:Chinese manufacturing listed companies to establish financial early warning model based on BP neural network technology to predict the possible arrival of the crisis onthe future, more accurate and effective.