This thesis explores how organizations in the shipping and banking industry can apply anomaly detection to identify unwanted or risky events that threaten their operations. It consists of two parts. In the first part, we study how freight forwarders and customs authorities can apply anomaly detection to combat miscoding and smuggling in international shipping. Miscoding and smuggling are fraud schemes in which fraudsters provide falsified information about the goods in transit to evade shipping restrictions or customs duties. We develop a fraud detection system based on a Bayesian network to detect such fraud automatically in shipment data and compare the effectiveness of the system with traditional audit methods. In the second part, we study how central banks can apply anomaly detection to identify liquidity risk at banks from the transaction logs generated by financial market infrastructures. Liquidity risk arises when a bank manages its liquidity inadequately and is no longer able to meet its payment obligations. Identifying early warning signs of such risk is of importance to central banks to supervise the financial activities of banks. We develop various models to identify liquidity risk and evaluate their usefulness for bank supervision.
|Qualification||Doctor of Philosophy|
|Award date||13 Nov 2019|
|Place of Publication||Tilburg|
|Print ISBNs||978 90 5668 609 3|
|Publication status||Published - 2019|