Empirical Study On Financial Crisis Prediction Of Listed Manufacturing Companies

Security Market plays a significant role on guiding investment and optimizes the allocation of financial resources. The Security Market is growing rapidly since established in 1995.Until the end of 2013, the number of the listed companies has been to 2489. Beneath the surface of prosperity, there is the hidden problem. The number of the listed companies with financial crisis is rising. The key financial indicators such as debt ratio failed to reach the standard of Multinational Corporation. In the crude reality of the highly competitive environmental scenario, the domestic companies have shaky foundation and weak abilities to resist the financial crisis. Thus, establishing an effective financial forecasting model is to be at hand for many people including stockholders,creditor,government and stakeholders.The paper firstly defined the meaning of financial crisis and introduced relevant literature in and abroad. After that, summarize the merits and shortcomings of different kind of financial forecasting model.Based on that, the paper chose the Support Vector Machine to establish the forecasting model. Support Vector Machine Model has no strict supposes and good training approximation and generalization. To improve precision further, use PSO method to optimize the parameters.The paper selected the new ST Company as sample among the listed companies of manufacturing industry. Using GA method to optimize the parameter and getting the best parameters. The result of Empirical Study is that the precision of the forecasting ModelThe article also used Logistic model, BP neural network and PSO-SVM model to compare and the results was compared to each other. The result of cooperation is that the GA-SVM model has better precision. It is expected that the study may provide good reference for financial forecasting study.