Local Government Debt Risk Early Warning By Machine Learning Method

Since 2008,the debt of local governments in China has been growing rapidly,and the annual growth rate has been much higher than the GDP growth rate,which the debt burden of local governments is constantly increasing,and this situation has also received widespread attention from all walks of life.Therefore,in 2011,the debt problems of local governments in China were audited.According to various statistics,by the end of 2018,Chinese local government debt totaled 1.89 trillion yuan,of which the debt ratio as an indicator to judge the debt situation of local governments reached 76.6%by the end of 2018.By the end of 2018,nearly 90%of the local governments in China have borrowed money to varying degrees.For this reason,the Party Central Committee and the State Council attach great importance to the risk prevention and control of local government debt.At the same time,in 2018,the Ministry of Finance of China put forward that the issue scale of local government debt should be rationalized and standardized,and that the risk prevention and control should be placed in a prominent position so as to enhance the awareness and ability of risk prevention and control,The early warning and prevention mechanism of debt risk,which belongs to the nature of local government,is also clearly required to be established,so that the local management responsibilities can be implemented.And if the government wants to effectively control the risk,it must strictly let the local government debt risk be assessed.Firstly,the relevant literature at home and abroad is sorted out.At the same time,the concept and theoretical basis of local government debt are also described.Based on machine learning algorithm,this paper discusses whether it is suitable for early warning of local government debt risk.Moreover,the local government debt be predicted by a scientific and feasible machine learning algorithm.The local government debt risk is selected as the object of prediction.Local government debt from 2012 to 2017 was taken as a research sample.Then,the method of analysis and comparison is used to make the scope,structure,content and investment of local government debt be studied in detail.The characteristics and causes of local government debt risk in a region is taken as the main research object.And early warning index system was build.Then,the European distance and Pearson correlation coefficient clustering method of repeated stratification are mainly used in the evaluation of early warning indicators.Therefore,the risk factors are began to study,sorted out the score of risk factors and demarcated the corresponding critical value.The corresponding critical values are calibrated.The object of training and learning is carried out simulation training,machine learning method is used to estimate the debt risk of local government in A region in 2018.The results show it can be seen that the early warning results of local government debt risk calculated by machine learning algorithm are relatively reliable and can be adopted.However,in 2018,the risk level of the structural risk factor of debt structure in A region is rated as medium risk,the risk level of narrow debt risk factor is high risk,the risk factor of income structure is medium risk,and the total risk factor of F is medium risk,so the debt situation of local government in A region is relatively safe.Finally,Application Suggestions and policy Suggestions of local government debt risk based on machine learning algorithm are proposed.