Implementing statistical learning methods through Bayesian networks (Part 2): bayesian evaluations for results of black toner analyses in forensic document examination

This paper presents and discusses the use of Bayesian procedures - introduced through the use of Bayesian networks in Part I of this series of papers - for 'learning' probabilities from data. The discussion will relate to a set of real data on characteristics of black toners commonly used in printing and copying devices. Particular attention is drawn to the incorporation of the proposed procedures as an integral part in probabilistic inference schemes (notably in the form of Bayesian networks) that are intended to address uncertainties related to particular propositions of interest (e.g., whether or not a sample originates from a particular source). The conceptual tenets of the proposed methodologies are presented along with aspects of their practical implementation using currently available Bayesian network software.

Interpretation: Printed Document Examination and Evidence

The literature dealing with the interpretation of results of examinations performed on "printed" documents is very limited. The absence of published literature reflects the absence of formal guidelines to help scientists assess the relationship between a questioned document and a particular printing technology. Generally, every printout, independent of the printing technology, may bear traces induced by characteristics of manufacture and/or acquired features of the printing device. A logical approach to help the scientist in the formal interpretation of such findings involves the consideration of a likelihood ratio. Three examples aim to show the application of this approach.

A simple logical approach to questioned envelopes examination.

This case report outlines research undertaken as the result of a document examination case in which two envelopes were involved. The combination of the circumstances of the case and the results of the examination allows a simple application of a logical approach to pre-assess the probability that an envelope (or a package) potentially recovered at a suspect's home comes from the same batch (same source) as questioned envelopes. This highlights that it is useful to examine envelopes.

Documents : Dating a document

The determination of the age of a document is a very frequent query, however it is also one of the most challenging and controversial areas of questioned document examination. Several approaches were defined to address this problem. The first is based on the introduction date of raw constituents (like paper or ink) on the market. The second approach is based on the aging of documents, which is unfortunately not only influenced by passing time, but also by storage conditions and document composition. The third approach considers the relative age of documents and aims at reconstructing their chronology. The three approaches are equally complex to develop and encounter quantity of problems which are not negligible. This article aims to expose the potential applications and limitations of current ink dating methods. Method development and validation and the interpretation of evidence prove to be essential criteria for the dating of documents.

Effects of Different Postural Conditions on Handwriting Variability

During their career, forensic document examiners will inevitably be confronted with handwriting carried out under unusual conditions (UnC). A questioned document signed on top of a car or on a vertical surface like a wall are two examples. These atypical circumstances may give rise to more variability of the signatures or written words, in particular if the body was in non-traditional writing position without the traditional support. Few studies were devoted to handwriting variability under unusual writing conditions. The current study investigates whether individual variability changes with special writing conditions. In a previous study (Sciacca & al, 2009), we found that eight repetitions were sufficient to obtain a correct estimation of the variance. In the present study, twelve subjects were asked to write two word sets eight times in upper and eight times in lower case, under different conditions : sitting and writing on a horizontal (usual condition UC) or vertical support; and standing, kneeling or laying while writing on a horizontal or vertical support (unusual conditions UnC). Words were written on a pen tablet, normalized in space and time and then averaged. The variance of the eight words was measured under all conditions. Results showed only an increase in variability under the laying and kneeling / vertical UnC. Within the five other postural conditions tested, handwriting was shown to be very stable.

Quantitative Assessment of Handwriting Evidence: The Value of the Shape of the Letter "A"

To provide a quantitative support to the handwriting evidence evaluation, a new method was developed through the computation of a likelihood ratio based on a Bayesian approach. In the present paper, the methodology is briefly described and applied to data collected within a simulated case of a threatening letter. Fourier descriptors are used to characterise the shape of loops of handwritten characters "a" of the true writer of the threatening letter, and: 1) with reference characters "a" of the true writer of the threatening letter, and then 2) with characters "a" of a writer who did not write the threatening letter. The findings support that the probabilistic methodology correctly supports either the hypothesis of authorship or the alternative hypothesis. Further developments will enable the handwriting examiner to use this methodology as a helpful assistance to assess the strength of evidence in handwriting casework.

Implementing statistical learning methods through Bayesian networks. Part 1: a guide to Bayesian parameter estimation using forensic science data

As a thorough aggregation of probability and graph theory, Bayesian networks currently enjoy widespread interest as a means for studying factors that affect the coherent evaluation of scientific evidence in forensic science. Paper I of this series of papers intends to contribute to the discussion of Bayesian networks as a framework that is helpful for both illustrating and implementing statistical procedures that are commonly employed for the study of uncertainties (e.g. the estimation of unknown quantities). While the respective statistical procedures are widely described in literature, the primary aim of this paper is to offer an essentially non-technical introduction on how interested readers may use these analytical approaches - with the help of Bayesian networks - for processing their own forensic science data. Attention is mainly drawn to the structure and underlying rationale of a series of basic and context-independent network fragments that users may incorporate as building blocs while constructing larger inference models. As an example of how this may be done, the proposed concepts will be used in a second paper (Part II) for specifying graphical probability networks whose purpose is to assist forensic scientists in the evaluation of scientific evidence encountered in the context of forensic document examination (i.e. results of the analysis of black toners present on printed or copied documents).

Atitude dos docentes do segundo ciclo face à inclusão escolar de alunos com necessidades educativas especiais

O presente trabalho tem por base o estudo do tipo de atitude revelada pelos docentes de 2ºciclo face à inclusão escolar de alunos com necessidades educativas especiais. A metodologia seguida passou pela entrevista aplicada a cinco docentes de vários grupos disciplinares, observação direta a dois docentes em contexto de sala de aula e análise documental. A análise dos resultados obtidos permite afirmar que os docentes estão sensibilizados para as vantagens da escola inclusiva, contudo, consideram que, para a escola inclusiva ter sucesso, é imprescindível investir na formação de professores, sobretudo, na área das necessidades educativas especiais. Para além de referido anteriormente salienta-se ainda, a necessidade de maior número de horas de apoio por parte dos docentes de Educação Especial, mais recursos materiais, turmas mais reduzidas e apoio de outros técnicos especializados. A inclusão dos alunos considerados com necessidades educativas especiais no ensino regular implica mudanças ao nível das atitudes e das práticas pedagógicas de todos os intervenientes no processo ensino e aprendizagem, da organização e da gestão na sala de aula e na própria Escola enquanto instituição.

O gerenciamento de riscos na prevenção de perdas trabalhistas : um estudo de caso no Grupo Eberle Mundial

Este trabalho tem como objetivo geral verificar como uma empresa gerencia seus riscos trabalhistas com vistas a preveni-los. É um estudo de caso único, de natureza descritiva e qualitativa. Foi desenvolvido junto ao Grupo Eberle Mundial, através de entrevista do Gerente de Recursos Humanos, questionário auto-aplicado por sete membros do corpo executivo da organização, exame de documentos e análise de artigos de revistas e de jornais. Paralelamente, buscou experiências e opiniões de quatro especialistas da área trabalhista, também mediante questionário auto-aplicado. Em face dos resultados obtidos, o trabalho descreve a forma de gerenciamento de riscos trabalhistas adotada pelo Grupo Eberle Mundial e atende os objetivos específicos nos termos adiante sintetizados. Demonstra que o Grupo mantém rotinas para identificação de riscos trabalhistas, porém, não pratica a definição prévia do que sejam riscos relevantes e não relevantes. Identifica uma diversidade de riscos trabalhistas percebidos pelo Grupo e verifica que esses não abrangem riscos operacionais. Mostra que o Grupo criou métodos e instrumentos de trabalho próprios para o gerenciamento de seus riscos trabalhistas, atua de forma permanente na prevenção desses e faz a revisão anual de seus processos, buscando chegar à incorporação dessa filosofia à cultura da empresa, contudo, não mantém um sistema que integre seus procedimentos, na forma proposta pela literatura. Confirma que, através do projeto implantado pela nova administração, o Grupo que, em 1998, estava com sua sobrevivência comprometida, até 2003 já reduziu em mais de 50% as perdas financeiras que vinha sofrendo e restabeleceu sua credibilidade e imagem internas e externas. Constata que existe comprometimento da alta direção do Grupo com o gerenciamento de riscos trabalhistas, todavia, esse comprometimento não é do conhecimento de todos membros do corpo executivo da organização. Aponta que o Grupo não mantém área específica para gerenciamento de seus riscos trabalhistas, nem tem claro quem são os profissionais responsáveis por esses procedimentos.

Desenvolvimento das competências técnicas dos peritos documentoscópio da Polícia Federal

A Documentoscopia é a maior área de perícia da Criminalística da PF, respondendo por 24,49% de toda a produção de laudos do Sistema Nacional de Criminalística. Apesar disso, não possui área de concurso ou graduação específicas, e o desenvolvimento das competências da área depende quase que exclusivamente da capacitação oferecida e executada internamente, dentro da instituição e do ambiente de trabalho. Considerando os planejamentos estratégicos da Direção Geral e da Diretoria Técnico-Científica da PF, que manifestaram a importância da valorização de seus servidores por meio da capacitação contínua e da gestão de competências como estratégia para se alcançar suas missões, vê-se a relevância no adequado estudo e desenvolvimento das competências na área da perícia documentoscópica. O presente trabalho tem por objetivo analisar se as competências técnicas dos peritos documentoscópico da Polícia Federal elencadas na matriz da função técnica da PF estão em consonância com as elencadas pela ONU para os examinadores forenses de documentos, e se essas competências estão sendo desenvolvidas nas ações de capacitação oferecidas pela ANP voltadas para a área. Foram identificadas algumas lacunas, ou seja, recomendações da ONU que encontram correspondência nas elencadas na matriz, mas não são desenvolvidas pelas ações de capacitação, além da discrepância quanto à carga horária dos cursos. Algumas sugestões para a minimização ou eliminação dessas lacunas foram colocadas, e outras considerações foram feitas, principalmente voltadas à maior oferta de capacitação, à especialização profissional, à instituição de testes de proficiência e da mentoria.

Evaluation of the evidential value of the elemental composition of glass, ink and paper by laser-based microspectrochemical methods

Elemental analysis can become an important piece of evidence to assist the solution of a case. The work presented in this dissertation aims to evaluate the evidential value of the elemental composition of three particular matrices: ink, paper and glass. In the first part of this study, the analytical performance of LIBS and LA-ICP-MS methods was evaluated for paper, writing inks and printing inks. A total of 350 ink specimens were examined including black and blue gel inks, ballpoint inks, inkjets and toners originating from several manufacturing sources and/or batches. The paper collection set consisted of over 200 paper specimens originating from 20 different paper sources produced by 10 different plants. Micro-homogeneity studies show smaller variation of elemental compositions within a single source (i.e., sheet, pen or cartridge) than the observed variation between different sources (i.e., brands, types, batches). Significant and detectable differences in the elemental profile of the inks and paper were observed between samples originating from different sources (discrimination of 87–100% of samples, depending on the sample set under investigation and the method applied). These results support the use of elemental analysis, using LA-ICP-MS and LIBS, for the examination of documents and provide additional discrimination to the currently used techniques in document examination. In the second part of this study, a direct comparison between four analytical methods (µ-XRF, solution-ICP-MS, LA-ICP-MS and LIBS) was conducted for glass analyses using interlaboratory studies. The data provided by 21 participants were used to assess the performance of the analytical methods in associating glass samples from the same source and differentiating different sources, as well as the use of different match criteria (confidence interval (±6s, ±5s, ±4s, ±3s, ±2s), modified confidence interval, t-test (sequential univariate, p=0.05 and p=0.01), t-test with Bonferroni correction (for multivariate comparisons), range overlap, and Hotelling's T2 tests. Error rates (Type 1 and Type 2) are reported for the use of each of these match criteria and depend on the heterogeneity of the glass sources, the repeatability between analytical measurements, and the number of elements that were measured. The study provided recommendations for analytical performance-based parameters for µ-XRF and LA-ICP-MS as well as the best performing match criteria for both analytical techniques, which can be applied now by forensic glass examiners.

Evaluation of the Evidential Value of the Elemental Composition of Glass, Ink and Paper by Laser-Based Micro-Spectrochemical Methods

Elemental analysis can become an important piece of evidence to assist the solution of a case. The work presented in this dissertation aims to evaluate the evidential value of the elemental composition of three particular matrices: ink, paper and glass. In the first part of this study, the analytical performance of LIBS and LA-ICP-MS methods was evaluated for paper, writing inks and printing inks. A total of 350 ink specimens were examined including black and blue gel inks, ballpoint inks, inkjets and toners originating from several manufacturing sources and/or batches. The paper collection set consisted of over 200 paper specimens originating from 20 different paper sources produced by 10 different plants. Micro-homogeneity studies show smaller variation of elemental compositions within a single source (i.e., sheet, pen or cartridge) than the observed variation between different sources (i.e., brands, types, batches). Significant and detectable differences in the elemental profile of the inks and paper were observed between samples originating from different sources (discrimination of 87 – 100% of samples, depending on the sample set under investigation and the method applied). These results support the use of elemental analysis, using LA-ICP-MS and LIBS, for the examination of documents and provide additional discrimination to the currently used techniques in document examination. In the second part of this study, a direct comparison between four analytical methods (µ-XRF, solution-ICP-MS, LA-ICP-MS and LIBS) was conducted for glass analyses using interlaboratory studies. The data provided by 21 participants were used to assess the performance of the analytical methods in associating glass samples from the same source and differentiating different sources, as well as the use of different match criteria (confidence interval (±6s, ±5s, ±4s, ±3s, ±2s), modified confidence interval, t-test (sequential univariate, p=0.05 and p=0.01), t-test with Bonferroni correction (for multivariate comparisons), range overlap, and Hotelling’s T2 tests. Error rates (Type 1 and Type 2) are reported for the use of each of these match criteria and depend on the heterogeneity of the glass sources, the repeatability between analytical measurements, and the number of elements that were measured. The study provided recommendations for analytical performance-based parameters for µ-XRF and LA-ICP-MS as well as the best performing match criteria for both analytical techniques, which can be applied now by forensic glass examiners.