Managing risk in the modern world - Applications of Bayesian networks
A Knowledge Transfer report from the London Mathematical Society and the KTN for Industrial Mathematics, authored by Norman Fenton and Martin Neil.

Executive summary

Businesses and governments must often assess and manage risk in areas where there is little or no direct historical data to draw upon, or where relevant data is difficult to identify. For example, the Barings Bank collapse in 1995 was not due to credit or market risk, where banks have sufficient data for prediction and mitigation of risk, but rather it was due to what is now called operational risk – the results of failures in everyday operational processes. The challenges are similarly acute when the source of the risk is novel: terrorist attacks, ecological disasters, major project failures, and more general failures of novel systems, market-places and business models.

Even though we may have little or no historical data, there is often an abundance of expert (but subjective) judgement, as well as diverse information and data on indirectly related risks. These are the types of situation that can be successfully addressed using Bayesian Networks (BNs), even when classical, data-driven approaches to risk assessment are not possible. BNs describe "webs" of causes and effects, using a graphical framework that provides for the rigorous quantification of risks and the clear communication of results. They can combine historical data with expert judgement.

During the last decade, researchers have incorporated BN techniques into easy-to-use toolsets, which in turn have enabled the development of decision support systems in a diverse set of application domains, including medical diagnosis, safety assessment, forensics, procurement, equipment fault diagnosis and software quality. Further technology and tool advancements since 2000 mean that end-users, rather than just researchers, are now able to develop and deploy their own BN-based solutions. As a result, BN methods are beginning to penetrate mainstream business practice. Recent commercial case studies provide evidence of impressive returns on investment from these techniques.

Both the practice and research of BNs are mushrooming. This report provides a snapshot of this dynamic and exciting area, including an introduction to the underpinning ideas, recent case studies, emerging areas of application, current research challenges, and a summary of the key players. The full report is available by following the link below, or in hard copy by writing to the Industrial Mathematics KTN.

Download 'BayesianNetworks.pdf' (841 Kb).