Research On Financial Crisis Prediction Based On Dea And SVM

With the deepening development of market economy, Chinese enterprises have gained more development opportunities, so that the pace of development of China’s enterprises is relatively rapidly. But at the same time, it also makes the Chinese market has become more complicated and risky, enterprises are facing with the chanllenge of the competitive environment becomes more complex and intense. Financial crisis will trigger a series of disastrous consequences such as investment losses, employee unemployment, credit can not be recovered, even more frightening is that the financial crisis will affect the country’s economic development, financial security and social stability. Therefore, the research on financial crisis early warning becomes very necessary and urgent.Currently, the financial crisis early warning methods experienced trend analysis, discriminant analysis, Logistic regression analysis, BP neural network and improve the traditional methods and so on phase. However, relying solely on the traditional knowledge of financial theory and subjective experience to assess the financial condition, obviously can not meet the requirements of a modern market economy. So we need to find new theories, new methods, or use new tools, new ways to analyze the financial situation of enterprises.Support vector machines (SVM) from the statistical learning theory and based on VC dimension theory and structural risk minimization principle. Due to the excellent learning and generalization ability, support vector machine has been widespread concern in the field of financial early warning in recent years. Data envelopment analysis (DEA) was introduced to determine the relative efficiency of a set of similar Decision Making Units (DMU), where each DMU uses multiple inputs to produce a number of outputs. DEA is able to provide measures for the efficiency of a corporation, thus DEA is employed as a tool to predict corporate failure in many financial failure prediction literatures. Financial crisis prediction is researched based on DEA-SVM model in the study. The main topics are followed.Firstly, establish a new indicator system based on SEDEA-GRA. This chapter integrated SE-DEA and GRA for indicators selection. Grey relational degrees between financial situations and financial indicators are the feature weights on account of that GRDs can provide effective predicting information for the financial crisis of the listed companies. Then, using financial data listed on China’s securities market for empirical analysis.Secondly, establish a hybrid model based on DEA and SVM. In the proposed method (DEA-SVM), DEA is employed as a tool to evaluate the input/output efficiency. Then, this paper established a new indicator system which includes the efficiency values, further more, the hybrid model of DEA and SVM was constructed. Finanly, using the listed companies of China’s securities market for empirical analysis.This article uses financial data of Chinese listed companies do empirical analysis, the empirical results validate the proposed warning ideas, warning indicators and warning method are not only feasible and effective.