The Research To Enterprise Financial Early Warning Based On The Performance Omparison Of Multiclassifier

China’s entry into the socialist market economy has brought great opportunities for business development.At the same time, it is also constantly strengthening the degree of competition between enterprises; After entering the capital market, the listing Corporation has a wide development space, but it also faces the crisis of risk. The results of this series in addition to bring the growing difficulty of management of managers, but more is to give corporate investors, stakeholders brings uncertain risks and losses.The market economy is a double-edged sword for business managers, investors and stakeholders. Under the background of today’s economic and financial shocks development, enterprise managers, investors and stakeholders accurate identification market risk, grasp the financial situation of the enterprise investment, make timely correct risk judgment and take corresponding measures, appear particularly important. At the same time, the method of data mining application in the enterprise financial early warning has become the research hot spot.With China A stock market listing Corporation as research sample, based on multiple classifier model in the data mining technology, this paper has developed the multiple classifier financial early warning analysis process which take the China listing Corporation as example. The main job description as follows:(1) Based on the accrual basis and cash basis Perspective, combined with the accrual basis and cash basis research perspectives of past enterprise financial early warning index system of index of complementary and integration,it gets The 6 dimension 36 indicators of enterprise financial early warning system. (2) In the basis of the early warning index system for Chinese listed companies sample data of sample design, using multiple data mining technology of classification algorithm,it sets to establish the multiple classifiers in financial early warning model in training sample, such as neural network, C5.0, CHAID (C), C&R tree (R), QUEST (Q), Logistic, discriminant, SVM, Bayesian network and KNN warning model, etc... (3) Multiple classifiers in financial early warning model prediction accuracy, curve area below, training samples accuracy, model enhance performance index of the degree of the maximum profit rate, maximum profit model were calculated in the test sample sets.(4)After establishing the multiple classifier prediction model, we analyze the early warning model of each classifier performance by using revenue (gains), response (response), enhance the degree (lift), profit (profit), the cumulative rate of return on investment (ROI-return on investment) these five indexes.The main conclusions are as follows:(1)Net assets per share,Operating cash inflow ratio, Investment cash inflow ratio,Operating cash inflow-outflow ratio,Financing cash inflow-outflow ratio,Cash flow per share from operating activities,Net asset earning rate, Accounts payable turnover are the most important indicator of affecting the enterprise in the future if there is a financial crisis in the enterprise financial early warning index system, So the enterprise after three years is the main influencing factors of ST is enterprise profitability and profit quality, enterprise in the future whether it is ST is not only affected by profitability, but also by the impact of corporate earnings quality and cash flow structure. (2) "model, from the data to the data." Data has a decisive effect on the training results of the early warning model. It is more important to pay attention to the standardization of research process than to draw the conclusion of special research, so we need to establish financial early warning research and practice in a general approach:That is, after designing the sample which according to the particularity of the sample data corresponding, it establishes a number of financial warning model, uses multinomial model performance evaluation index to evaluate multiple financial warning model, finally, chooses the better performance of the financial early warning model to predict the results of the analysis.Finally, from the practical point of view, the paper puts forward two suggestions to the enterprise investors, managers and enterprise stakeholders:(1) It needs to emphasis on the financial statements of the financial ratios, but also need to focus on study of the cash flow statement in the financial indicators, combined with accrual and cash based perspective of financial early warning analysis in making the financial crisis early warning of enterprise. (2) In making the financial crisis early warning of enterprise, it should establish multiple classifier prediction model, make comments on the performance of early warning model by using of a number of model performance evaluation index, then select excellent early warning model, make full use of model method to improve the effect of prediction.