Sensitivity Analysis of Bayesian Networks Used in Forensic Investigations

Kwan, Mei Po; Overill, Richard E.; Chow, Kam-Pui; Tse, Hayson; Law, Frank Y. W.; Lai, Pierre K. Y.

doi:10.1007/978-3-642-24212-0_18

Cited by 11 publications

(8 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Forward and backward inference is recognized as a crucial characteristic in BNs. These attributes, along with a rigorous probabilistic background, have led BNs to successful applications [1], [2], [5], [6]. In the technical literature, BNs are also known as Bayesian expert systems, probabilistic graphic networks, belief networks, or Bayes nets.…”

Section: Bayesian Networkmentioning

confidence: 99%

“…Their ability to model competing hypotheses about the causes of failure makes them a suitable tool for forensic assessments. BNs have been mainly used in medicine, legal, and forensic science, mostly to support decisions in circumstances where uncertainty or imprecision prevails [1]- [5]. Forensic engineering has partially used BNs for drawing reasonable conclusions about the causes of failures.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Testing the Causes of a Levee Failure Using Bayesian Networks

Garcia-Feria

Colmenares

Pérez

2022

Ing.

View full text Add to dashboard Cite

Context: Forensic geotechnical engineering aims to determine the most likely causes leading to geotechnical failures. Standard practice tests a set of credible hypotheses against the collected evidence using backward analysis and complex but deterministic geotechnical models. Geotechnical models involving uncertainty are not usually employed to analyze the causes of failure, even though soil parameters are uncertain, and evidence is often incomplete. Method: This paper introduces a probabilistic model approach based on Bayesian Networks to test hypotheses in light of collected evidence. Bayesian networks simulate patterns of human reasoning under uncertainty through a bidirectional inference process known as “explaining away.” In this study, Bayesian Networks are used to test several credible hypotheses about the causes of levee failures. Probability queries and the K-Most Probable Explanation algorithm (K-MPE) are used to assess the hypotheses. Results: This approach was applied to the analysis of a well-known levee failure in Breitenhagen, Germany, where previous forensic studies found a multiplicity of competing explanations for the causes of failure. The approach allows concluding that the failure was most likely caused by a combination of high phreatic levels, a conductive layer, and weak soils, thus allowing to discard a significant number of competing explanations. Conclusions: The proposed approach is expected to improve the accuracy and transparency of conclusions about the causes of failure in levee structures.

show abstract

Section: Bayesian Networkmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Testing the Causes of a Levee Failure Using Bayesian Networks

Garcia-Feria

Colmenares

Pérez

2022

Ing.

View full text Add to dashboard Cite

show abstract

“…With every anticipated item of digital evidence successfully recovered the posterior probability in favour of the prosecution hypothesis was ca.97.2%. While a credible alternative hypothesis was not available for this case against which to compare the result, it corresponds to an LR of ca.34.7 in favour of the hypothesis; however, both single-parameter and multi-parameter sensitivity analyses resulted in minimal perturbations to that value [ 26 , 27 ].…”

Section: Proposition Plausibility Metricsmentioning

confidence: 99%

Quantitative evaluation of the results of digital forensic investigations: a review of progress

Overill

Collie

2021

Forensic Sciences Research

View full text Add to dashboard Cite

Unlike conventional forensics, digital forensics does not at present generally quantify the results of its investigations. It is suggested that digital forensics should aim to catch-up with other forensic disciplines by using Bayesian and other numerical methodologies to quantify its investigations' results. Assessing the plausibility of alternative hypotheses (or propositions, or claims) which explain how recovered digital evidence came to exist on a device could assist both the prosecution and the defence sides in criminal proceedings: helping the prosecution to decide whether to proceed to trial and helping defence lawyers to advise a defendant how to plead. This paper reviews some numerical approaches to the goal of quantifying the relative weights of individual items of digital evidence and the plausibility of hypotheses based on that evidence. The potential advantages enabling the construction of cost-effective digital forensic triage schemas are also outlined. KEY POINTS The absence of quantified results from digital forensic investigations, unlike those of conventional forensics, is highlighted. A number of approaches towards quantitative evaluation of the results of digital forensic investigations are reviewed. The significant potential benefits accruing from such approaches are discussed.

show abstract

“…Determining the weights and beliefs for the edges and items of evidence poses a problem for our framework as well as for other symbolic approaches. For some items of evidence the weights as well as the probabilities can be elicited based on statistical analysis and forensic evidence (Kwan et al 2011;Fenton and Neil 2012;Zhang and Thai 2016). To test the robustness of Bayesian networks with respect to minor changes in subjective beliefs, (Fenton, Neil, and Lagnado 2013) propose to apply sensitivity analysis on the nodes in question.…”

Section: Related Workmentioning

confidence: 99%

Explainable Automated Reasoning in Law using Probabilistic Epistemic Argumentation

Ibs¹,

Potyka²

2020

Preprint

View full text Add to dashboard Cite

Applying automated reasoning tools for decision support and analysis in law has the potential to make court decisions more transparent and objective. Since there is often uncertainty about the accuracy and relevance of evidence, non-classical reasoning approaches are required. Here, we investigate probabilistic epistemic argumentation as a tool for automated reasoning about legal cases. We introduce a general scheme to model legal cases as probabilistic epistemic argumentation problems, explain how evidence can be modeled and sketch how explanations for legal decisions can be generated automatically. Our framework is easily interpretable, can deal with cyclic structures and imprecise probabilities and guarantees polynomial-time probabilistic reasoning in the worstcase.

show abstract

Sensitivity Analysis of Bayesian Networks Used in Forensic Investigations

Cited by 11 publications

References 10 publications

Testing the Causes of a Levee Failure Using Bayesian Networks

Testing the Causes of a Levee Failure Using Bayesian Networks

Quantitative evaluation of the results of digital forensic investigations: a review of progress

Explainable Automated Reasoning in Law using Probabilistic Epistemic Argumentation

Contact Info

Product

Resources

About