Skip to main content

Decision support, Bayesian Networks and Food Security

Professor Jim Smith has had an enduring interest in Food Security issues for a number of years. In 1999 he was one of a 4 person team commissioned to review the requirements for statistical, economic and decision support systems for the Food Standards agency in the UK. He also is part of a joint team across York and Warwick Universities to support the EFSA. Within this contract. the Warwick Team takes the lead in its Crisis Decision Support and Epideiological Modelling.

His expertise centres on the formal integration of expert judgements concerning components of a high dimensional stochastic in order to provide coherent Bayesian decision support and has worked in a number of application areas of this technology. In particular over about 15 years he was part of a team integrating countermeasures for decision support after a nuclear accident. This included the modeling and then the evaluation of the efficacy of various coutermeasures to address the effects of such an accident in food contamination.

Jim Smith is an expert in the selection and estimation of graphical models and Bayes Nets - especially dynamic ones - as they apply to the types of dynamically evolving scenarios illustrated above. His research in these areas has been supported through a number of EPSRC, EU grants, and Home Office Contracts over the years.

He is currently part of a three person team resourced by Warwick University to investigate the feasibility of using such graphical methods and domain expert judgements suported by any avaliable observational and experimental evidence to provide integrated probabilistic modelling for decision support under scenarios posing a threat to food security.

JIm has recently won a three year £400k EPSRC award to investigate the development of systems to explore the effectiveness of countermeasures for addressing food insecurities and is also collaborating in a new COST Network 'Expert Judgment Network: Bridging The Gap Between Scientific Uncertainty And Evidence-Based Decision Making'.

A few recent relevant publications:

Xiang, Y., Smith, J.Q. and Kroes, K. (2011) "Multiagent Bayesian Forecasting of Structural Time invariant Dynamic Systems with Graphical Models" International J. of Approximate Reasoning 52, no. 7, 960--977; MR2835054

Rigat, F. and Smith, J.Q. (2011) "Study of Key Interventions into Terrorism using Bayesian Networks" (SRG/09/44) Home Office Final Report (44 pages)

Smith, J.Q. and Freeman, G. (2011) "Distributional Kalman Filters for Bayesian Forecasting and Closed Form Recurrences" J. of Forecasting, Vol 30, No. 1 210 -224

Smith, J.Q. (2010) " Bayesian Decision Analysis: Principles and Practice" Cambridge University Press

Thwaites, P. Smith, J.Q. and Riccomagno, E. (2010) "Causal Analysis with Chain Event Graphs" Artificial Intelligence, 174, 889–909

Smith, J.Q. and Daneshkhah, A. (2010) "On the Robustness of Bayesian Networks to Learning from Non-conjugate Sampling" International J. of Approximate Reasoning, 51, 558 -572


Professor Jim Smith