Research On Intelligent Early Warning Of Expenditure Control In Administrative Institutions Based On Machine Learning

The arrival of the era of big data and artificial intelligence has accelerated the process of informatization of internal control in administrative institutions,and promoted the transformation of internal control in administrative institutions from traditional models to intelligent models.More and more administrative institutions have realized the cloud operation of the business by establishing a financial cloud platform.Under the financial cloud platform,the internal control of administrative institutions has a qualitative improvement in cost control,efficiency,and quality compared with traditional control models.Expenditure control,as an important part of internal control of administrative institutions,has always been the focus of attention of state organs,relevant departments,and the public.However,the expenditure control of administrative institutions under the financial cloud platform still lacks timely and effective supervision.Therefore,in the context of intelligence,the use of a large amount of financial and business data accumulated by the financial cloud platform to achieve intelligent early warning of expenditure control can help managers and financial personnel to find the risks in unit expenditure control in a timely manner,and further strengthen the expenditure of administrative institutions It is of great significance to control and promote the informatization of internal control.In view of this,this paper takes administrative expenditure control under the financial cloud platform as the research object,on the basis of comprehensive analysis of the expenditure control objectives,key control points,risk points and other contents of administrative institutions,determine the target and content of early warning of expenditure control,and build a machine-based Intelligent early warning framework for learning expenditure control of administrative institutions.The thesis conducts early warning research on the control of the three aspects of expenditure approval process integrity,standard compliance and rationality.First,for the early warning of the integrity control of the expenditure approval process of administrative institutions,based on the analysis of the expenditure approval system and control status of the administrative institutions,the integrity control of the expenditure approval process was designed from the aspects of coverage,efficiency and quality of the approval process Early warning indicators,using the K-means algorithm to implement unsupervised clustering of early warning data,and dividing the early warning levels of various types of data on the basis of analyzing the characteristics of the clustering indicators to achieve intelligent early warning of the integrity control of the expenditure approval process under the financial cloud platform;Secondly,for the compliance control and early warning of the expenditure standards of administrative institutions,based on the analysis of the existing expenditure standard regulations,control status and relevant expenditure data characteristics,the standard compliance control and early warning indicators were rationally designed,using the C4.5 decision tree The algorithm realizes intelligent early warning of compliance control of expenditure standards;third,for the rational control and early warning of expenditure of administrative institutions,combined with budget preparation and execution related theories,scientifically design early warning indicators of expenditure rational control from the two levels of expenditure structure and progress,The Kohonen-SVM combination algorithm is used to realize intelligent early warning of reasonable expenditure control.