Financial Risk Prediction And Early-warning For Listed Enterprises

Since 1960's, with the increasing enterprise bankruptcy problem, more and more scholars try to predict bankruptcy by quantitative analysis. So far the past 50 years, from linear prediction models such as the multiple linear discriminant analysis to the various non-parameters prediction models of which the neural network is representative, related researches come out endless. However, domestic enterprises prediction of the financial crisis has just begun, and it has great practical and theoretical value to predict a bankruptcy by combining abroad experience and actual situation in China.On this paper, the financial early warning methods, the establishment of model and how to conduct risk prediction have been respectively ananlyissed and introduced. According to acquisition the actual data of listed companies in China ,we used Fisher,Bayes and stepwise discrimination to discriminate analysis and established the multiple linear discriminated models for listed companies before they have financial difficulties. For these forecasts models's establishment is closely linked the samples obtained, so these models all have certain level of presence significamtly different degrees of false-positive rate. The establishment of early warning financial model can only give the general managers and investors a simple feasible prediction method, so the accuracy of the established early financial warning models can meet the requirements. But it can be found through the study that the sooner the companies being listed, the greater the companies might be labeled as ST special treatments.