The Study On Financial Crisis Of Listed Companies Based On Pre-warning Combination Model Of Probit And SVM

Under the premise of market economy,the company faced the"survival of the fittest" at the early stage of establishment.In the face of an increasingly large trading market,the number and scale of listed companies constantly challenge the quality of supervision,so the control of financial risk is particularly impor-tant.Financial distress usually goes through a long period from the incubation period to the outbreak period.So the financial crisis of company is predictive and predictable.If we establish a financial crisis pre-warning model,we can efficiently predict,identify and control its development trend.And we can make useful mea-sures improve the status quoreduce loss and make the stakeholders obtain the biggest benefit when risks burgeon.In addition,the current statistical software has powerful functions,the theory is mature,and the China securities regulatory commission has formulated norms for the financial information disclosure of listed companies,which makes the data sources required for modeling more accurate and provides conditions for the establishment,of a reasonable pre-warning model.Therefore,if we can correctly predict the financial crisis of the company,it is not only related to the interests of investors and creditors,but also can help the se-curities regulatory authority to formulate more effective methods to stabilize the market.Therefore,the study on pre-warning financial-crisis of listed companies has very important theoretical significance and practical significance.Firstly,this paper introduces the definition and the cause of the financial crisis,then it summarizes and reviews existing research status,and also deeply analyses the influencing factors of the financial crisis on listed companies in China.We select five aspects to build the financial early warning index system,like the cash flow,solvency,growth ability,profitability and assets operation ability,and then combine theory with demonstration on this basis.The ST companies are chosen as the symbol of the financial crisis.In two stock markets of Shanghai and Shen-zhen,we choose A-share listed companies which were capped from 2014 and 2018 as this study object,and also select 660(including 102 in 2014,120 in 2015,153 in 2016,159 in 2017 and 126 in 2018)listed companies of different industries as this study samples.Then we choose the first 3 years of financial datas of listed com-pany which were capped and successively built Probit regression model,support vector machine(SVM)model and Probit-SVM pre-warning combination model to make empirical analysis.The empirical results showed that the combined pre-warning model had higher stability and classification accuracy compared with the two single models.The prediction of the model is also stronger and has a greater advantage for the financial crisis pre-warning of listed companies.