The Early-warning Of Listed Companies Financial Crisis Of Statistical Analysis

With the global economy integration trend intensified and information waves increased, our corporations not only achieve new market chances but also face huge survival crisis. Financial crisis is the concentration during the process of production management and particularly featured. Therefore, the early-warning of corporation financial crisis is an important question in the academic and practical circle. In order to make Chinese capital market step into a healthy road. All participators in the capital market, including investors, intermediaries, listed companies, supervision organizations and legislative institutions, face a huge task of improving theory level and strengthening practical experiences. Especially, at present, our country is encouraging to develop organization investors to cultivate the stable strength in the capital market. Therefore, it is significant to find early-warning problems of listed company in our capital market with quantitative research method. The article combines domestic relevant situation and progress with overseas aimed at many financial early-warning problems in our capital market. After researching the background, meaning and purpose, the research structure is determined.The article firstly introduces the concept of financial warning, forewarning model and the influences of financial warning. In addition, the article describes the theory and statistical thoughts of the principle component analysis and techniques of discriminant analysis. Then the article defines the listed company involving the financial crisis, choosing 34 companies which listed firstly by ST in 2014 and 34 companies whose finance is normal among Shenzhen and Shanghai A-shares for research training sample. The article uses the principle component analysis to choose 14 representative financial indexes among 23 financial indexes established at first. After significance testing and normality testing, 6 financial indexes chosen at last will be created. The article chooses the optimal Fisher linear discrimination model to establish our financial warning model and gives the early-warning section. The accuracy rate of the warning model is 83.82%. This early-warning is proved to have economic meaning and practical value.