A credit assessment method for tax-paying enterprises based on generalized maximum flow

Abstract

The invention discloses a tax credit evaluation method based on generalized maximum flow. Firstly, the Taxpayer Global Network (TGN) is constructed by discovering explicit social relations and mining implicit social relations among tax-paying enterprises, and characterizing the weights of relationship edges between tax-paying enterprises; secondly, the influence of tax-paying credit The transfer process is modeled analogously to the maximum flow problem in the network flow, using the generalized maximum flow modeling to influence the path selection, and using the attenuation function to model the weakening of the influence with the growth of the path, and obtain the enterprise's correlation evaluation; finally, comprehensively consider the taxpaying enterprise itself And the interactive relationship between taxpayers and affiliated enterprises, using the normalized fusion method, the individual taxpayer evaluation and the associated evaluation are fused to obtain a comprehensive tax credit evaluation result.

Description

technical field [0001] The invention relates to a method for identifying groups suspected of tax evasion by using computer technology, in particular to a credit evaluation method for taxpaying enterprises based on generalized maximum flow. Background technique [0002] With the continuous advancement of my country's credit system construction, the social value and social influence of tax credit have been increasing day by day. It has been widely used in bidding, financing and other fields, and has become an important asset for enterprises to participate in market competition. How to use corporate tax data to construct a model for tax credit assessment has become an urgent problem to be solved. [0003] The existing tax credit assessment methods are mainly based on the taxpayer's fulfillment of tax obligations, and the taxpayer's tax credit rating within a certain period of time. A combination of annual evaluation index scores and direct judgment grades is adopted: usually b...

AI Technical Summary

Problems solved by technology

[0010] The methods described in the above documents mainly have the following problems: Documents 1 and 3 respectively use the neural network model and the decision tree model to classify the tax-related data to obtain the tax credit level of the enterprise. However, both of them only use the historical tax payment data of the enterprise itself. , without considering the interactive relationship with affiliated enterprises; although Document 2 considers the interactive relationship between affiliated enterprises, the interactive relationship involved is only transaction relationship, and does not consider the path selection and Influence attenuation problem

Images