Research On The Model Of Tax Audit Selection In H City Based On Logistic Regression To Identify Corporate Tax Evasion

Tax revenue concerns the rise and fall of a nation and the safety of its political regime.However,with the constant development of China's society and economy,illegal means of taxation have also been constantly changed and updated.An increasing number of criminals engaged in tax evasion have cropped up and tax evasion cases also occur frequently.Therefore,to maximally prevent and limit the tax evasion behaviors of taxpayers has become the top priority in tax inspection in recent years.If tax authorities want to succeed in this fight against tax evasion,they must sort out,analyze and study various patterns and means of tax evasion that are quite common in the present economic market,conduct thorough research,figure out their development laws,and produce a general working framework that can be replicated and promoted,and can effectively prevent and limit criminal behaviors of tax evasion to more vigorously crack down on illegal tax behaviors such as tax evasion,standardize market economic order,and enhance the tax compliance of taxpayers.This thesis adopts the empirical research method,offers analysis by integrating common tax evasion models with corporate financial data,establishes a case selection indicator system,adopts logistic regression method,and constructs a case selection model that can identify companies engaged in tax evasion.First of all,this thesis illustrates the background and significance for tax inspection case selection and sorts out and comments on relevant literature on the research of tax inspection case selection.Secondly,this thesis briefly introduces the definition,functions and basic work procedures of tax inspection and offers a business analysis on tax inspection case selection.Then,the thesis introduces the current status of tax inspection case selection in H city,points out major problems in real work,and puts forward the necessity of establishing tax inspection case selection models.Through analyzing the reasons,types and means of tax evasion,the thesis summarizes five types of common means of tax evasion adopted by taxpayers and points out ways to quickly and accurately identify tax evasion means.Thirdly,based on consulting relevant literature and research results and summarizing the practical experience of inspectors for tax departments,the thesis extracts financial indicators from corporate financial statements and tax returns as independent variables and builds a case selection indicator system.Selecting thetax-related data of sample companies in Huhhot city,the thesis adopts the logistic regression analysis method to construct the tax inspection case selection model for identifying tax evasion by companies in Huhhot city.Through the real inspection results of selected samples,the thesis tests the feasibility of the case selection model.Finally,the thesis puts forward the conditions and line of thinking for the practical application of the case selection models.This thesis selects 200 samples of companies in Huhhot city and adopts the logistic regression analysis method to construct the tax inspection case selection model.The result of research shows that the logistic regression model has high accuracy,the accuracy rate of honest enterprise reaches 79.9%,the accuracy rate of tax evasion enterprise reaches 76.2%,and the comprehensive accuracy rate reaches 79.5%.These data reflect that we can use the information technology in the process of case selection.Besides,the the case selection models can effectively improve the accuracy of identifying tax evasion and improve work efficiency.