An important contribution to anti-fraud and supply-chain security comes from the development of Risk Analysis tools targeted to the discovery of suspicious containerized transportations. In this work we present the Anomalous Container Itinerary Detection (ACID) framework that analyses Container Status Messages to discover irregular container shipments. The system has been developed at JRC as part of an in-house global route-based risk analysis facility for containers monitoring. It adopts a flexible and modular design and its preliminary prototype applies an SVM one-class classifier to detect anomalies. The experimental evaluation %analyzed more than 300 thousand containers, demonstrates that the analysis module may be set to detect efficiently the expected number of suspicious itineraries, which can be further investigated by Customs authorities thanks to the web-based visualization tool provided with the system.
展开▼
