The Research On Medium And Small Businesses Financial Early Warning System

The small and medium-sized enterprise in our country is a group which has the most vigor and vitality ,and already became the important strength which supports the national economic growth.However, along with more and more intense of the market competition, the management and operations of the SMEs are facing more and more stern challenge, the financial risk becomes to the important attribute.Therefore, the studies on existence of the financial risk and establishing the early warning system of the SMEs, become the difficult problem which the SMEs superintendent's have to face.This article is precisely take the SMEs as the object of study, from financial risk elementary theory obtaining, studies the type as well as the reasons of the SMEs finance risk, and the early warning question.This paper applied the support vector machine (SVM), which is currently relatively popular machine learning algorithms to the company's financial crisis in the field of early warning. SVM offers a new modeling thought and method for studying the companies' financial crisis prediction,and widens the research range of this field. In the course of the study, this paper, divides the listed company which according to financial situation good or bad into two types, one is non-ST and the other ST established model, and with BP neural network were compared and then try to be listed in this paper the company is divided into three categories, and the establishment of LS-SVM model, and achieved good results. in accordance with the current practices of the securities market, The results showed that two models of SVM authoritative and operational characteristics, in the nonlinear classification has an advantage to overcome the traditional linear model assumptions shortage and improve the accuracy of the warning.There are five parts in this paper. In the first part,the background , meaning, content and method is introduced at the aim of recapitulation .And the second one is literature colligation. The third part on the financial risks for SMEs , the concept of the financial early warning system, nature. In the fourth part, firstly is introduction of support vector machine on the theory of knowledge, and then statistical theory and financial management theory are used to choose financial index as Warning variables. Finally the listed companies' financial crisis model base on SVM is setup and implemented by mat1ab program. In the last part this thesis try to establish the business financial early warning system: early warning organization, early warning analysis methods, application Process, and the Countermeasures for avoiding and eliminating the crisis.