| With the development of office informatization in my country,the business models of major enterprises have gradually changed with the development of society and technology.In 2012,the VAT reform was gradually carried out nationwide,and the business tax of enterprises was changed to value-added tax.This policy reduced double taxation for small and medium-sized enterprises and lowered the tax burden of enterprises.Therefore,the standardization of value-added tax management is very important for an enterprise.In the context of the prevalence of big data,it is an inevitable trend to integrate big data technology into the tax management system.In addition to the management of input and output items and value-added tax,the tax declaration system under the background of big data also needs to realize the use of relevant tax data to predict the taxation risk of the company.The result of the risk prediction can be known.How big are the tax-related risks of each enterprise,and assist each enterprise in self-inspection of tax work,and standardize the tax management of the enterprise.In order to realize the office automation and intelligence of the tax staff of Company A,the company decided to develop a tax declaration and tax risk prediction system.The core idea of the main writing of this thesis is based on demand investigation and demand analysis,outline design,detailed design and testing of tax declaration and tax risk prediction system,and then integrated into the relevant research of tax risk prediction algorithm.The implementation technology framework of the system is Spring,Spring MVC and Spring JDBC.When realizing the systematic taxation risk prediction,the research of the prediction algorithm model is first carried out.This part of the content will be written as one of the key chapters of this thesis.The taxation risk prediction model algorithm research mainly selects three algorithms: random forest,multi-layer perceptron,and support vector machine.The three algorithms are respectively constructed for the prediction model,and the company’s historical tax data is processed by the three algorithms.Comparing the experimental results and finally applying the best taxation risk prediction model derived from the experimental results to the system.Up to now,the system has been put into use,and the initial operation results have become a benign trend.The research and development of this system,for financial personnel,helps tax personnel to free themselves from heavy and complicated taxation work,and realizes the automation and intelligence of tax management work;for company managers,it helps them to predict the results based on tax risks.Make more scientific decisions on the company’s tax management work;for other departments of the company,based on the tax risk prediction results of each branch,they can evaluate the work of their company and department to understand the company’s current operation. |
