Kernel-based methods for source identification using very small particles from carpet fibers

Armstrong, Douglas; Neumann, Cédric; Saunders, Christopher P.; Gantz, Donald T.; Miller, J. J.; Stoney, David A.

doi:10.1016/j.chemolab.2016.10.004

Cited by 4 publications

(20 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We warn the reader that other models involving scores have been proposed (Armstrong et al, 2017;Swofford et al, 2018;Ausdemore et al, 2019;Hendricks et al, 2019) but are not considered to be score-based likelihood ratios. These models do not rely on the ratio of the likelihoods of the score in two sampling distributions.…”

Section: Score-based Likelihood Ratiosmentioning

confidence: 98%

Defence against the modern arts: the curse of statistics—Part II: ‘Score-based likelihood ratios’

Neumann

Ausdemore

2020

Law, Probability and Risk

View full text Add to dashboard Cite

For several decades, legal and scientific scholars have argued that conclusions from forensic examinations should be supported by statistical data and reported within a probabilistic framework. Multiple models have been proposed to quantify the probative value of forensic evidence. Unfortunately, several of these models rely on adhoc strategies that are not scientifically sound. The opacity of the technical jargon used to present these models and their results, and the complexity of the techniques involved make it very difficult for the untrained user to separate the wheat from the chaff. This series of papers is intended to help forensic scientists and lawyers recognise limitations and issues in tools proposed to interpret the results of forensic examinations. This paper focuses on tools that have been proposed to leverage the use of similarity scores to assess the probative value of forensic findings. We call this family of tools "score-based likelihood ratios". In this paper, we present the fundamental concepts on which these tools are built, we describe some specific members of this family of tools, and we explore their convergence to the Bayes factor through an intuitive geometrical approach and through simulations. Finally, we discuss their validation and their potential usefulness as a decision-making tool in forensic science.

show abstract

Section: Score-based Likelihood Ratiosmentioning

confidence: 98%

Defence against the modern arts: the curse of statistics—Part II: ‘Score-based likelihood ratios’

Neumann

Ausdemore

2020

Law, Probability and Risk

View full text Add to dashboard Cite

show abstract

“…The first stage of the approach is essentially a standard statistical hypothesis test, where we wish to test whether the characteristics of the trace objects in E u are indistinguishable from those of the objects in E s obtained from

scriptS

using a formal α ‐level test. This section extends the results presented by Armstrong et al 8 to develop a statistical test that relies on a kernel function to measure the level of similarity between pairs of high‐dimensional, heterogenous, and complex random vectors.…”

Section: First Stage: Testing Indistinguishabilitymentioning

confidence: 60%

“…It needs to satisfy only two requirements: it must be a symmetric function, that is, κ ( x i , x j )= κ ( x j , x i ); and it must ensure that the marginal distribution of s ij is normal to satisfy the assumption made on the score model in equation . The assumption of normality is the main assumption made by Armstrong et al 8 when developing their model; it is reasonable for high‐dimensional objects and can be satisfied through careful design of the kernel function 20…”

Section: First Stage: Testing Indistinguishabilitymentioning

confidence: 99%

“…In most situations, Ψ is not known and must be learned from s n . Armstrong et al 8 show that an analytical solution to estimate Ψ from s n exists. Instead of replacing Ψ by a point estimate,

bold \hat{Ψ}

, in equation , we integrate out the uncertainty associated with the model parameters by considering the posterior distributions of θ ,

σ_{a}^{2}

, and

σ_{e}^{2}

, given s n .…”

Section: First Stage: Testing Indistinguishabilitymentioning

confidence: 99%

“…Below, we propose a formal statistical method to test the hypothesis that two groups of chemical spectra are indistinguishable (Parker's similarity stage). We extend the work published by Armstrong et al 8 to develop a generic α ‐level test for comparing sets of high‐dimensional, heterogenous, and complex random vectors, in which we account for the uncertainty on the model's parameters. We propose various mathematical developments that result in a computationally efficient Algorithm and make it useable in practice.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Two‐stage approach for the inference of the source of high‐dimensional and complex chemical data in forensic science

Ausdemore

Neumann

Saunders

et al. 2020

Journal of Chemometrics

View full text Add to dashboard Cite

Forensic chemists are often criticised for the lack of quantitative support for the conclusions of their examinations. While scholars advocate for the use of a Bayes factor to quantify the weight of forensic evidence, it is often impossible to assign the necessary probability measures to perform likelihood‐based inference on chemical data. To address this issue, we leverage the properties of kernel functions to offer a method that allows for statistically supporting the inference of the identity of source of sets of trace and control objects by way of a single test. Our method is generic in that it can be easily tailored to any type of data encountered in forensic chemistry, and our method does not depend on the dimension or the type of the considered data. The application of our method to paint evidence analysed by FTIR shows that this type of evidence carries substantial probative value. Finally, our approach can easily be extended to other types of chemical evidence such as glass, fibres, and dust.

show abstract

Skin surface debris as an archive of environmental traces: an investigation through the naked eye, episcopic microscope, ED-XRF, and SEM–EDX

et al. 2023

View full text Add to dashboard Cite

Traces from bodies can be of various nature, for example of biological or inorganic origin. Some of these historically have received more consideration than others in forensic practice. Samplings of gunshot residues or biological fluid traces are commonly standardized, whereas macroscopically invisible environmental traces are usually ignored. This paper simulated the interaction between a cadaver and a crime scene by placing skin samples on the ground of five different workplaces and inside the trunk of a car. Traces on samples were then investigated through different approaches: the naked eye, episcopic microscope, Scanning Electron Microscopy (SEM) with Energy-dispersive X-ray spectroscopy (EDX) and Energy Dispersive X-Ray Fluorescence (ED-XRF). The purpose is to provide the forensic scientist with the awareness of the value of debris on skin and then to highlight implications for forensic investigations. Results demonstrated that even naked eye observation can reveal useful trace materials, for defining the possible surrounding environment. As a next step, the episcopic microscope can increase the number of visible particulates and their analysis. In parallel, the ED-XRF spectroscopy can be useful to add a first chemical composition to the morphological data. Finally, the SEM–EDX analysis on small samples can provide the greatest morphological detail and the most complete chemical analysis, although limited, like the previous technique, to inorganic matrices. The analysis of debris on the skin, even with the difficulties due to the presence of contaminants, can provide information on the environments involved in criminal events that can add to the investigation framework.

show abstract

Kernel-based methods for source identification using very small particles from carpet fibers

Cited by 4 publications

References 4 publications

Defence against the modern arts: the curse of statistics—Part II: ‘Score-based likelihood ratios’

Defence against the modern arts: the curse of statistics—Part II: ‘Score-based likelihood ratios’

Two‐stage approach for the inference of the source of high‐dimensional and complex chemical data in forensic science

Skin surface debris as an archive of environmental traces: an investigation through the naked eye, episcopic microscope, ED-XRF, and SEM–EDX

Contact Info

Product

Resources

About