Research Of Financial Distress Prediction Model Based On Rough Sets And Support Vector Machines

Economic globalization is the direction of the current world economic development. Since China’s accession to the WTO, domestic enterprises will face more and more uncertain macroeconomic situation and the increasingly fierce market competition, the survival and development face unprecedented challenges. Enterprises will inevitably encounter the crisis in the survival and development process, the best way is to follow the principle of ahead of the crisis early warning and precautions.It has come to the point how can regulators, investors and other stakeholders take measures to prevent in time before the financial crisis. With the constant improvement of China’s securities market regulation, meanwhile, it helps to protecting the listed company financial crisis early warning.Financial early warning, also called the financial failure warning, which is on the basis of the financial accounting information, with the help of enterprise provide financial statements, business plans, and other related information, through ratio analysis, factor analysis, analysis of a variety of ways or by discriminate model, the risk index system, the financial situation of the building of enterprise business activities, to urge the enterprise operator to take timely measures to avoid potential risk evolution into a loss. As the pace of capital market continues to quicken, more attention is payed on investment. Consequently, how to make full use of the accounting information of listed companies to identify and control the investment risk effectively has been a great question.Support vector machine (SVM) was put forward by Vapnik in the Middle1990s, which is a new method of machine learning based on statistical learning theory. In his research, he achieved empirical risk minimization by seeking for structure risk minimization. Also, it was successfully applied in pattern recognition, regression analysis, time sequence prediction, etc. At present the selection of SVM parameters mainly depends on experience, however, there is no way to guide the SVM parameter selection.Rough set theory is put forward by a Polish mathematician Pawlak in1982, a kind of incompleteness and uncertainty of mathematical tool. It does not require any additional information about the data given in advance. What’s more, it can simplify the data on the premise of retaining key information and the minimum expression of knowledge, and it also plays well on how to identify and evaluate the data dependencies between, obtain the minimum rule from empirical data, in machine learning, decision analysis, process control and other fields that has been widely used.Combining RS and SVM theory, this paper based on RS and SVM is proposed financial crisis early warning model of listed companies. At the same time, according to parameter selection problem, we use genetic algorithm to optimize the parameters of the SVM algorithm to realize the automation of the SVM parameters selection.Through empirical study, this paper constructed it can reduce the training time, and has higher prediction accuracy based on rough sets and support vector machine (SVM) of the financial crisis early warning model is simply using support vector machine (SVM).In addition, this article at the end of the early warning model is proved to exist some limitations, and puts forward some Suggestions for subsequent research.