Research On Enterprise Income Tax Assessment Based On BP Neural Network

In modern tax collection and administration,tax assessment plays an increasingly important role.It is of great significance in promoting the scientific and refined tax administration.How to scientifically and effectively assess taxation of enterprises,especially real estate companies,has always been a major difficulty in the work of tax authorities at all levels.Since the introduction of the tax assessment mechanism in China,tax authorities have used the indicator peak method for tax assessment,and have not established a scientific tax assessment model.The professional quality of tax assessment staff is not high enough to understand the purpose of tax assessment and confuse it with tax audits.Taxpayer's rights are damaged.In recent years,the rise of artificial intelligence,especially the improvement of neural network technology that is good at solving non-linear problems,has set a new direction for the construction of a tax assessment model for the real estate industry.Applying neural network technology to taxation assessment and realizing the identification of tax risk enterprises based on tax-related information of enterprises is a new idea for the current taxation authority to carry out tax assessment work.This paper analyzes the characteristics of China's real estate industry and its corporate income tax assessment.It puts forward that its salient problems are lacking of legal basis for assessment,analysis of assessment indicators being incomplete,index peak method being not suitable for the economic status,lack of professional assessment team and the professional quality of assessment staff being not high.In order to solve these problems,this paper establishes an index system from four aspects: solvency,profitability,operational capacity,and corporate income tax status.Based on this,it proposes a design concept based on BP neural network,and introduces its principles and implementation process in detail.A real estate enterprise income tax assessment model is established.Taking the real estate company in Qingyuan City of Guangdong Province as the research object,the feasibility of BP neural network in tax assessment is verified.It is hoped that this technology will be used to improve the accuracy of tax assessment options,to better detect corporate tax evasion and to increase tax compliance.