Bayesian Networks and Probabilistic Inference in Forensic Science

✍ Scribed by Prof Franco Taroni, Colin Aitken, Prof Paolo Garbolino, Dr Alex Biedermann

• Publisher: Wiley
• Year: 2006
• Tongue: English
• Leaves: 364
• Series: Statistics in practice
• Category: Library

✦ Synopsis

The amount of information forensic scientists are able to offer is ever increasing, owing to vast developments in science and technology. Consequently, the complexity of evidence does not allow scientists to cope adequately with the problems it causes, or to make the required inferences. Probability theory, implemented through graphical methods, specifically Bayesian networks, offers a powerful tool to deal with this complexity, and discover valid patterns in data. Bayesian Networks and Probabilistic Inference in Forensic Science provides a unique and comprehensive introduction to the use of Bayesian networks for the evaluation of scientific evidence in forensic science.

• Includes self-contained introductions to both Bayesian networks and probability.
• Features implementation of the methodology using HUGIN, the leading Bayesian networks software.
• Presents basic standard networks that can be implemented in commercially and academically available software packages, and that form the core models necessary for the reader’s own analysis of real cases.
• Provides a technique for structuring problems and organizing uncertain data based on methods and principles of scientific reasoning.
• Contains a method for constructing coherent and defensible arguments for the analysis and evaluation of forensic evidence.
• Written in a lucid style, suitable for forensic scientists with minimal mathematical background.
• Includes a foreword by David Schum.

The clear and accessible style makes this book ideal for all forensic scientists and applied statisticians working in evidence evaluation, as well as graduate students in these areas. It will also appeal to scientists, lawyers and other professionals interested in the evaluation of forensic evidence and/or Bayesian networks.