A policy distribution service for proactive fraud management over financial data streams

With the ever-increasing number of unmanned customer interaction channels (e.g. internet and mobile banking), financial platforms are becoming subject to an expanding collection of system attacks due to their unique potential for large scale monetary theft. Significant challenge therefore remains in keeping fraud policy repositories current to prevent financial loss and customer inconvenience, due to the rapid rate at which fraudsters re-engineer their methods and deploy new system attacks. Accordingly, fraud policy development based upon mining of an institution's transactional data provides only an isolated view of industry behaviour with no support for the exchange of fraud policy data with other institutions that are likely to experience similar behavioural attacks as fraudsters horizontally permeate their efforts through analogous sector organisations. This paper proposes a Policy Distribution Service for the automated collection, analysis and distribution of emerging fraud policy definitions through a collaborative policy sharing model. A key aspect of the service is the creation of an industry wide policy knowledge base within a centralised mediator component and distribution of policies using a common modelling formalism which maybe mapped into multiple disparate target fraud management implementations. © 2008 IEEE.