The Research On Finance Early-warning System Of Listed Company In Our Country

The listed company is the foundation of Chinese securities market, and its quality is closely related with the stability and development of the market. But, due to bad management and other reasons, financial distress has already occurred in many listed companies. These companies cannot work normally, some of which almost go into bankruptcy. In order to forecast the probability and seriousness of a company's financial distress and take corresponding measures to avoid losses, it is necessary to establish an efficient financial distress early-warning system(FDES).At the same time, it is also a problem valued by Chinese scholars.On the basis of previous study achievements, this thesis takes Chinese listed companies as the sample .We have chosen 17 financial indexes which reflect listed companies' profit ability , debt paying ability , operation ability , growth ability and companies' scale . and establishes a FDES for listed company of our country by using data excavate modeling method, We hope that this FDES is a useful tool for the related forecasting and decision, and this thesis is benefit to similar studies in our country.The main view and innovation of this thesis (1) It is different from the research of traditional statistical method that this thesis has used the technology of the data mining to carry on research to financial pre-warning. Statistical methods is the main the research approach in our country at present, such as multi-regression, Fishier and Logistic . when using the statistic method to analysis a problem ,It is all to regard a certain assumption at the beginning, then use statistics method to prove or deny this assumption, this kind of analytical method requires the variable to obey normal distribution and the association's variance of samples are equal . This kind of model does not have getting fault-tolerant and learning ability , so it cannot be used widely. But in this thesis we have used data mining technology which is driven by data to analysis financial early-warning problem, this kind of data analysis technology does not need to carry on strict assuming in advance . So it is have a kinder application prospect than statistical methods.(2) Our research adopts the decision tree to establish FDES for the first time in studying at home. Our research selected the decision tree which is one of the main three kinds of data mining technologies to establish a FDES. Decision tree is a top-down classification method, and he constructs out the knowledge representation of the decision type by studying a training sample. The decision tree algorithm has many advantages, such as a lower limited data required, more quickly, higher precision and its tesults is easier to understand and so on. This thesis has proved...