Research On Enterprise Tax Screening Method Based On Data Mining

Data mining technology has been widely used in various fields because of its excellent data analysis capabilities.However,there are few applications for tax auditing and selection.There are still problems in tax auditing selection,including traditional methods,relying on experience,and low efficiency.Since the three-period tax information system of Golden Taxes was put into operation,more and more corporate declaration data were available to the tax authorities.The introduction of data mining technology in the selection of tax audits has increased the efficiency of data utilization and improved the efficiency of tax auditing.Very important significance.This paper analyzes the process of tax auditing selection,data sources and current problems,studies the data mining theory,and converts taxation auditing cases into data classification and clustering problems.In the classification analysis,the industrial manufacturing and commodity circulation industries were selected,and the existing tax inspection case data was used to apply the KNN(nearest neighbor algorithm)to establish a classification model for small-scale taxpayers and verify the model,which proved that the classification analysis can be introduced into The selection of tax auditors can greatly improve the efficiency of selection.In the cluster analysis,two industries,manufacturing and commodity circulation,were selected for the two types of taxes: value-added tax and corporate income tax.The nine audit selection indicators for two consecutive tax years were used to build the general taxpayer based on the DBSCAN clustering algorithm.Selection model.The historical data was used to verify the model,and the suspect enterprises selected by the model were included in the tax inspection and selection plan of the current year.By comparing with the traditional methods of the year,the operability and practicality of the selected model were verified.Cluster analysis can improve the quality of tax audit selection.