Research On The Financial Crisis Early-warning Of Listed Companies

Along with our country economic system reform and the rapid development of capital markets, it brings companies more development opportunities and more uncertain factors. In this context, enterprises may have a bit inadvertent could fall into financial crisis of the difficulties, and damage the interests of investors and losing the confidence of the market, which is not conducive to the healthy development of capital market. Therefore, constructing enterprise prediction model of financial distress, and to forecast the financial situation of enterprises, and help enterprises avoid the occurrence of financial crisis has become an imminent task.In view of the traditional enterprise prediction model of financial distress is usually based on financial indicator, but not quite comprehensive forecast the financial standing of the enterprise. The thesis tries to select part of non-financial information indexes into the enterprise financial crisis prediction system. Based on the MATLAB neural network theory and SPSS17.0, selecting28enterprises which fall into financial crisis as samples, according to select28normal enterprises as paired samples which belong to the same scale of industry, and then select22financial indexes,7non financial indicators which quantified financial information and non-financial information, using the single factor analysis of variance, correlation analysis to select the primary indicator, eliminating partial index. Finally selecting the independent variable indexes of prediction model of financial distress, and using the empirical analysis method to research prediction model of financial distress of listed enterprises.This thesis based on the theory of financial crisis of listed companies. To give the enterprise financial crisis a reasonable definition and analyze the causes of enterprise financial crisis, which combined with the characteristics of enterprise financial crisis, through the analysis to reasonable construct prediction model of financial distress of the listed enterprises. Secondly, choosing appropriate research samples and preliminary identification of financial crisis early warning index system. This thesis selects the Shenzhen and Shanghai Securities stock28listed enterprises which fall into trouble and28normal enterprises as the research objects, and using the software of SPSS17.0to choose the significant and accurately indicators system from the initial index to find suitable index system of listed enterprises’ financial crisis early warning. Based on MATLAB neural network constructs the prediction model of financial distress of listed companies, and to test the model. Finally, put forward some feasible suggestion to avoid financial crisis of listed companies.Through the research results, not only the financial information can not be ignored in the prediction model of financial distress, but the non financial information introduction also is greatly important to the model of early warning ability, thus the non financial information for enterprise financial crisis warning also has a positive meaning. Meanwhile, building the financial crisis early-warning model tested based on MATLAB neural network-based and found to match the test results and actual rate of86.10%, indicating the MATLAB-based neural network model can be built with the financial crisis, which is a feasibility way.