Early Warning And Investigation Countermeasures Of Crime Of Issuing False Invoice

In recent years,with the implementation of the policy of "Replacing Business Tax with Value-Added Tax" and "Streamlining Administration,Delegating Powers,and Improving Regulation and Services",criminals have been wildly issuing false invoice,and such cases have shown a trend of high frequency in economic crimes.Tax department and public security department are facing increasingly serious crime situation that will bring them a new challenge.By studying the current trend of falsely making invoice crime,the difficulties of investigation in such cases are analyzed,with the features of tax information from enterprises conducted false invoice as the breakthrough point,machine learning method is introduced to build a risk pre-warning assessment model constructed on the basis of Support Vector Machine to detect enterprises issuing false invoice.Therefore an effect of early finding,early warning and early crime fight can be achieved.Based on the theory of investigation,this paper adopts machine learning method to study the crime of making false invoice in the following aspects: The first part give a comprehensive description of the relevant research background and a systematical refection of the literature reviews,besides,the research method of quantitative analysis is also proposed in this section.The second part give a definition of the crime concept of issuing false invoice and of the falsely issuing behavior.It also give a description of three main crime types,"including falsely out with no false in","false in with no false out" and "false in and false out".The third part give an analysis of the crime trend of issuing false invoice and of the difficulties of investigation.At present,such cases are facing with the challenges of low roll back loss rate,poor cooperation in investigation and difficult in obtaining evidence.To better solve this kind of case investigation and improve investigation department initiative to crack down on such crimes,three steps are designed in part four—firstly,establishing a risk pre-warning assessment model to detect enterprises issuing false invoice,secondly,enterprises are classified,according to the risk levels,into 5 groups: A,B,C,D,E,lastly,collected data is used to make empirical analysis,the result of which shows that the accuracy rate of the model is 97%,proving a better adaptability than other models.In the fifth part,corresponding investigation countermeasures are formulated for enterprises with different risk levels,first,for low-risk enterprises,scientific prevention work mode is established as a main method,second,in the case of medium risk enterprises,third-rate factorsare mainly checked in data collision,and peripheral investigation was conducted on key enterprises suspected of "false";finally,for high-risk enterprises,the tax research and judgment platform is used to timely link up cases and lock criminal gangs,check and control abnormal illegal funds,strengthen the coordination mechanism among various departments,comprehensively fix the evidence of crime,and launch a cluster campaign to arrest criminal suspects in time when necessary.The risk pre-warning model constructed on the basis of machine learning methods can provide a real-time alarm,which make it possible for the investigation department to achieve a shift of passively accepting a challenge to taking the initiative.