Growth Indicator And Prediction Model Based On Catastrophe Series Method And KMV Model

Not long ago,the private economy symposium was held solemnly in Beijing.Combining with the strategic plan of "Made in China 2025",which has been promoted continuously in recent years,it can be seen that a large number of private enterprises with small and medium size are in the fields of medical treatment,new energy,artificial intelligence and so on.In the future,they will not only continue to be the "sea-stabbing needle" for employment protection,but also shoulder the responsibility of technological innovation throughout China.The responsibility of blooming.A number of enterprises with the above characteristics and through the issuance of stocks into people's vision are listed on the GEM in China.China's GEM was set up on October 30,2009.The original intention is to alleviate and solve the financing difficulties of "two high" and "six new" enterprises.The first 36 enterprises listed on the Shenzhen Stock Exchange.By the end of 2018,the number reached 740.The government's support for high-tech enterprises is increasing,and the enthusiasm of private capital for science and technology companies is also very high.Therefore,it is very necessary to study the growth of these enterprises,to judge which enterprises are diligent in governance,have high growth potential,which enterprises are neglected in operation,and have the suspicion of listing "circle money".The growth of enterprises not only represents the past business results of enterprises,but also means the development potential and brand value of enterprises in the future.Since western countries began to study the quality of enterprises,academia and industry have made great progress in measuring the growth of enterprises.The measurement angle is from single to multiple,and the methods used for evaluation are also tending to be objective and scientific.On the other hand,the development quality of enterprises is a very diverse and comprehensive concept,involving many evaluation fields,which is difficult to take into account scientifically by any measurement methods nowadays.Therefore,there is no unified and recognized index for measuring the growth of enterprises in today's world.Nevertheless,according to objective logic,high-growth enterprises should be better than low-growth enterprises in more indicators.In other words,high-growth enterprises and low-growth enterprises divided by good growth indicators should show enough disparity in each indicator,that is,indicators have good differentiation.Based on this idea,the main focus of this paper is two.First,to construct an effective and differentiated growth index.Secondly,build a model for predicting growth indicators.According to these two research emphases,this paper first calculates the primary growth indicators of enterprises through the method of catastrophe series,mainly based on traditional financial data;then,based on the revised KMV model,it calculates the default distances of enterprises using stock trading data and a small amount of financial data.In order to integrate financial information and market information into the final growth indicators,this paper takes the primary growth indicators and default distance calculated by mutation series as equal weight factors,calculates the final growth indicators;and then proves that the growth indicators after adding KMV default distance have a stronger distinction between high growth and low growth.Finally,nine factors selected according to the characteristics of GEM enterprises are taken as variables to construct models for classification tasks-Gauss Kernel-SVM and tree-based models for continuous value regression.The two kinds of machine learning models have achieved good results in classification and continuous value prediction.