Networks and machine learning to detect Value Added Tax fraud

Abstract

Value Added Tax (VAT) fraud erodes public revenue and puts legitimate businesses at a disadvantaged position thereby impacting inequality. Identifying and combating VAT fraud before it occurs is therefore important for welfare. This paper proposes flexible machine learning algorithms which detect fraudulent transactions by utilising the information provided by the VAT network structure. It does so by recognising that VAT fraud is a group or community activity which has to be detected by relying on heterogeneous data of huge dimension. VAT fraud detection is implemented through a combination of a suitably constructed Laplacian matrix with classification algorithms that rely on scalable machine learning techniques.