Research On The Dynamic Financial Distress Prewaring Of Hi-tech Manufactuirng Companies

High-tech manufacturing with high technology and high additional value is the mainforce in manufacturing. Research on the financial early warning of high-tech manufacturingenterprises can help companies avoid the financial crisis, guarantee the benign operation ofhigh-tech enterprises and then promote the development of the national economy. This thesisreviews and summarizes the domestic and foreign research on the financial early-warningresearch literatures. On the basis of advantages and disadvantages of different kinds ofanalysis research, this paper consider that a combined model based on artificial intelligence isan effective method of modern financial early warning research. On the basis of the actualsituation of listed companies in China, this thesis define the special treated company as theconcept of financial crisis and take the high-tech manufacturing listed companies as theresearch object. According to the construction principle of the early warning index system, thethesis build the whole financial early warning indicator system which is suitable for high-techmanufacturing listed companies.Because the emergence of enterprise financial is a continuous dynamic developmentprocess, this thesis studies the early warning of the high-tech manufacturing listed companies’financial crisis from short-and long-term perspective. As the short-term dynamic warning isbased on quarter, put the quarterly financial panel data into the integrated model composed ofGM (1,1) and BP neural network, and then take "Si Da Gao Ke" listed company as anexample to validate the model. The result shows that the GM-BP neural network caneffectively reflect the trend of the company’s financial situation and it has strong timeliness.As the long-term dynamic warning is based on year, put the dates of T-2and T-3intoLogistic-BP neural network model, then compare the results with logistic regression analysis’result and BP neural network model’s result. The result proves that Logistic-BP neuralnetwork model can reflect the occurrence mechanism of financial crisis. And then take "XinHua Zhi Yao" listed company as an application example which also proves the model’svalidity and high accuracy.This thesis’ main research conclusions are as follows:Firstly, the screening and streamlining in financial early-warning indicators is apre-requisite for establishing an early-warning model effectively.Secondly, the short-term financial early-warning model and long-term financial earlywarning model for high-tech manufacturing listed companies all have high pre-warningaccuracy. These models can help entrepreneurs notice the changes in financial condition andmake a reasonable judgment to avoid the financial crisis through rational decisions.Thirdly, the financial indicators’ sequential trend is an important index of financial evaluation and the longitudinal information has a significant effect on the financial crisis earlywarning. The combination of short-term and long-term financial early warning model to buildstatic and dynamic binding can fully exploit the corporate financial information, timely andeffective development trends also reflects the trend in financial position.Fourthly, the mixed model combined with different forecasting methods can takeadvantage of each method and improve the generalization ability which is also the trend offuture research.