The Research Of Chinese Listed Corporations' Financial Crisis Forewarning Based On Data Mining

Listed company Exposure to the stock market,its production and operation are particularly related to the external economic environment,In the face of more complex economic environment,whether the companies can continue to operate? How the parties obtained information about the financial situation when companies felt into crisis? This is a worthy of attention and study question.In this paper,while selecting 32 listed companies under Special Treatment by the Stock Exchange in 2007 and 32 non-ST listed companies as a paired sample,and selecting their Financial indicators from 2003-2005 as sample data,we constructed Logistic,Neural Network(NN) and Decision Tree financial distress forewarning models using data mining techniques,The result of study indicates that the forewarning models has good prediction especially neural network model,the forecast accuracy of the NN model is high to the training sample.The forecast accuracy is 88%in use of NN model to predict forecast samples included 26 listed companies.As a result,this article of building a financial crisis early wanting model has a scientific and practical.The research results can be referenced by the listed-companies,investors and creditors to take effective risk preventing measures.