Building a Corporate Tax Risk Model Based on Machine Learning Methods

Abstract

As the economy diversifies and business types increase, the existing tax risk(TR) models in some companies are usually inefficient and difficult to cover all taxpayers in the company, while the company leadership lacks a positive attitude towards risk model building, and more importantly, there is no talent to build risk models professionally, and TRs are usually discovered unintentionally in other business processes rather than through professional risk models, such a working mechanism It is really difficult to truly prevent TRs. Therefore, in order to improve the efficiency of the company's TR identification and management, this paper adopts the decision tree (DT) algorithm of machine learning to build a TR model. Compared with the existing model of the company, the newly built model can effectively identify and classify the company's risk through the decision number method. It is hoped that the model constructed in this paper can also be applied to other companies' TR management.