DOMINO - Advanced data analytics for compliance in e-commerce

The flow of border crossing e-commerce has been a considerable problem for Dutch Customs and for society. Much of this flow consists of cheap throwaway products that are characterized by a very high degree of non-compliance to quality standards and customs procedures. In an earlier Topsector project (Dafoe) we have shown the reasons for intransparency in e-commerce chains and the shortcomings of the current shipment driven customs enforcement approach.

In this project, we will develop new ways to supervise e-commerce flows, by leveraging data as well as automated learning techniques for risk assessment. We aim to develop a risk assessment approach based on one-armed bandit algorithms, that will help both declarants and customs leverage the large amount of data in a new and effective way.