The Application Of Logistic Regression And Discrisminant Analysis In The Financial Early Warning Model

If the listed company due to poor management of the financial crisis,it will not only make the operator suffers economic losses,will also bring serious loss of assets to investors.The aim of this paper is to establish a financial crisis early warning model,early to predict whether the company will occur financial crisis.On the one hand,it can help investors to make the right investment decisions,to avoid damage to the assets;on the other hand,also can be to the operator to watch and take timely measures to guard against and defuse the risks.Because the listed company occurs financial crisis is a gradual process,so this paper selects the data of the first three years of financial crisis compared with the financial normal company.We firstly explored financial normal group and ST(financial crisis)whether there were significant differences of the index,then use univariate decision model,the logistic regression model and discriminant analysis model to construct the model.The experimental results showed that: the accuracy of logistic regression model and discriminant analysis model is higher than that of univariate decision model;the general logistic regression model is superior to the component logistic model;fisher discriminant analysis is better than the stepwise discriminant analysis.In T-1,T-2,the discriminant accuracy rate of all models is above 80% while not very good in T-3.