Enterprise Tax Irsk Measure And Control

In recent years，Many large enterpri，，ses fell into the “tax gatescandal,because the lackof tax risk prevention and control. Such as Enron, WorldCom, Parmalat, those foreigncompanies and GuoMei,SuNing,ChuangWei,WaHaH’a,China PingAn and many other famousenterprises have trapped themselves in it.If the tax risk handle wrong, the enterprises maysuffer legal sanctions, financial loss, damage to reputation and undertake unnecessary taxburden. When seriously, they can cause the bankruptcy of the enterprise. Therefore,understanding the cause of the enterpriser tax risk and analysis its main influencing factors,measuring and controlling it is very necessary.This paper combined with the theory of risk and risk management, it defines the tax riskconcept, Analysis the causes and characteristics of enterprise tax risk. On the basis of this,combined with " Big enterprise tax risk management guidelines (Trial) whichw as Pro mulg at e d by the State Administration of Taxation in May,2009，Mainly from theenterprise management philosophy, the financial situation of enterprises, business status, thetax department statu so of the corporate, tax law and implementation, and other external factors.This six aspects analysis the tax risk of enterprise. And designed the questionnaire, afteranalysis questionnaire, use AHP to get the weight of influence factors, The business concept ofenterprise tax risk impact most heavily weighted. Secondly, is the enterprise’s tax departmentstatus,and then other external factors. Then followed by tax law and implementation, thefinancial situation of enterprises and the management condition of the enterprise.After got therelative weights of the factors, Then using the rough set to the tax risk influence factors ofknowledge reduction and attribute reduction, delete the redundant data,making the preparationfor the next t BP neural network construction. Rough set theory and neural network fusion has’many characteristics and advantages.lt s suitable for processing and solving unstructured,nonlinear complex problems. In the BP neural network model, first of all, accessing to relevantfactors of enterprise ‘s tax risk, than,the expert to affects the enterprise tax risk factorinformation for preliminary evaluation, According to some standards, quantified theinformation of the factors. Secondly, using rough set module to the quantitative information ofattribute reduction and knowledge reduction, extracting minimization rule set,and than,simplify the rules and knowledge training neural network, determining the neural networkweights and continuous learning speed, After the learning of neural networks, we will get thefinal output of classification and prediction. Finally, the prediction result is transformed into natural language, and the enterprise’ s tax risk is measured.