Financial Early Warning System Study Based On Combination BP Network

When fierce competition accompany with the opportunity, the market nowadays is more risk than any other days. So, as for the public companies, they are now facing a critical problem of being financial disabled, in other means is that they will have financial crisis or even worse being forced to quit list. If a company runs into such situation, it will surely bring threat to the existence and development of that company, and what's more brings enormous loss to investors, creditors as well. So to establish a effective and practical early warning model on financial crisis, in order to meet the increasingly urgent needs of relevant interests party is not only an academic problem, but also becomes the important factor influencing sound development of capital market of our country, and has very important realistic meanings.In order to solve the existing problems that lies in the current financial early warning indexes and that the invalidation caused by financial information distortion, the thesis made a series of cash flow prediction indexes regarding manufacturing industry company as the research object. The thesis also made out a demonstration that supplied references to the study of financial early warning in our country. Firstly, by way of summarizing the current study situation about early warning of the financial crisis, the thesis analyzed the theory and the method of financial crisis early warning.Secondly, after giving the meaning and standard of our financial early warning system, the thesis creates ten indexes both based on cash flow, then we have them both under the T examination to prove their capability of doing financial early warning, meanwhile we give the identify of the early warning's standard of this thesis.Thirdly, after introduce some common methods for analysis financial crisis; the thesis formed a new combination forecast model using combination forecast theory to combine the BP(Back Propagation)neural network with the factor analysis. The thesis use 152 public companies as the study objects, after doing the experiment the result shows that this combination forecast model has the advantages of both tradition statistics model and artificial intelligence model, and the accurate radio is 83.92%, that means it will have great value in practical use.Finally, the thesis give some advises from three different aspects such as cash flow management, internal control and business strategy, to reduce the risk of being financial crisis, after having financial early warninganalysis. Thus this thesis provided a new way to the financial early warning for the manufacturing industry company.