Research On Financial Crisis Prediction System Based On Rough-Sets And Neural Network

The primary objective of the research is to demonstrate the feasibility and validity of the financial crisis prediction with combined Rough Set and Artificial Neural Network (ANN) prediction system. It first imposes the information reduction technology of the Rough Set to reduce the corporation's financial data, then uses the improved BP algorithm with momentum accession method and self-adapt parameter adjust measure to train the network combined with checking out the effect of the prediction model and it instances the Rare Earth Hi-Tech to analyze its financial crisis. At the end of the text it summarizes some issues that should be cared in the prediction system research.Firstly, it makes the enterprise face more fierce competition and austere challenge with unsure environment and risk factors nowadays. Therefore financial crisis prediction becomes the research hotspot. Totally speaking, there are four kinds of financial crisis prediction method which are single variable analysis, MDA, Logistic regression analysis and ANN technique.Secondly, along with the appearance and development of the Rough set and ANN technique, dynamic prediction research on financial crisis has become possible and the developing trend. Based on this actuality, this text established a Rough set-ANN financial crisis warning system fit for the situation of our nation and that its test result shows the model has some feasibility and validity. This system is based on our country's negotiable securities market and present situation of the corporation that come into market, uses the study production of country's inside and outside, cites the A-stock of the Shenzhen and Shanghai market as the stylebook.Again, the main way of constructing Rough sets-ANN financial crisis prediction is that making use of the Rough sets'attribute reduction technique to reduce the financial index firstly and then imposing the ANN learning function to train network so as to drop out enterprise's crisis.Lastly, in this paper, it has used some related technology of the Rough set and ANN, while adopted SPSS, Rosetta and Matlab engineering software to analyze and dispose the data.