Zugl.: Diss. University Duisburg-Essen, 2007.

Thesis (Ph. D.)--Ohio State University, 2006. / Full text release at OhioLINK's ETD Center delayed at author's request

Made available in DSpace on 2015-03-26T13:32:22Z (GMT). No. of bitstreams: 1 texto completo.pdf: 445502 bytes, checksum: 7d042d373f0007ddc1967700e0e477c8 (MD5) Previous issue date: 2014-10-22 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Brucellosis is an infectious and contagious disease caused by bacteria of the genus Bru- ceIIa. It produces a characteristic infection in animals and may aIso contaminate humans. The tests used for diagnosis of bruceIIosis in Brazil are conducted in samples obtained from animals suspected of having the disease, sIaughtered or that died at the farm. In Brazil there are few vaIidation studies of diagnostic tests for bruceIIosis presenting statistical methodologies for the estimation of sensitivity and Specificity satisfactorily. The present work used the methodology proposed by Joseph, Gyorkos e Coupal (1995) to obtain es- timates of sensitivity and Specificity of screening test Buffered Acidified Antigen (AAT) and confirmatory tests Mercaptoethanol (2-ME) and Bacteriological Diagnosis (DBAC), these tests are in accordance with the National Program for Control and Eradication of AnimaI Brucellosis and Tuberculosis (PNCEBT), which began in Brazil in 2001. The study was complemented with comparisons of the estimates in three different scenario: (i) when the resuIt of onIy one test is avaiIabIe (ii) when results of two test are avaiIabIe, and (iii) results of three tests. The data contained samples from 175 animaIs, obtained by convenience from material sent from aII regions of BraziI, to the Laboratory of Diag- nosis of Bacterial Diseases of LANAGRO-MG, between the years 2008-2011. Statistical computations and Gibbs Sampler algorithm were impIemented in OpenBUGS. Results showed that the estimated prevalence of bovine bruceIIosis in suspected animals is 79%, which means that for every 100 animals suspected of having the disease, 79 are diagnosed with it. Regarding performance measures, AAT was more sensitive for diagnosing Bovine BruceIIosis, 2-ME more Specific for not diagnosing Bovine Brucellosis and DBac showed 100% specific for not diagnosing the disease and Iess sensitive for diagnosing the disease. / A brucelose é uma doença infectocontagiosa provocada por bactérias do gênero Brucella que produz infecção característica nos animais, podendo contaminar o homem. Os testes para diagnóstico da brucelose utiIizados no Brasil são reaIizados a partir de amostras obti- das em animais com suspeita da enfermidade abatidos ou mortos na propriedade. No país são poucos os estudos de vaIidação de testes diagnósticos para brucelose que apresentam metodologias estatísticas para a estimação da sensibilidade e da especificidade satisfatori- amente. Neste trabalho, empregou a metodologia proposta por Joseph, Gyorkos e Coupal (1995) para obter estimativas da sensibilidade e da especificidade do teste de triagem Antígeno Acidificado Tamponado (AAT) e dos testes confirmatórios Mercaptoetanol (2- ME) e Diagnóstico Bacteriológico (DBac), testes estes, em conformidade ao Programa Nacional de Controle e Erradicação da Brucelose e da Tuberculose Animal (PNCEBT) inserido no Brasil em 2001. De forma complementar, comparou-se as estimativas em três cenário distintos: quando dispõe do resultado de somente um teste; (ii) quando dispõe dos resultados de dois testes; e (iii) quando dispõe dos resultados de três testes. A amostra conteve 175 animais, obtida por conveniência a partir de material encaminhado, de todas as regiões do BrasiI, ao Laboratório de Diagnóstico de Doenças Bacterianas do LANAGRO-MG, entre os anos de 2008 a 2011. Os códigos para obter as estimativas foram impIementados no OpenBUGS, por meio do algoritmo Gibbs Sampler. Os resuIta- dos apontaram que a prevalência estimada para brucelose bovina em animais suspeitos é de 79%, ou seja, de cada 100 animais com suspeita da doença, 79 são diagnósticos como doentes. Em relação às medidas de desempenho, AAT se mostrou mais sensível para diagnosticar a Brucelose Bovina, o 2-ME mais Especifico para não diagnosticar a Brucelose Bovina e o DBac mostrou-se 100% específico para não diagnosticar a doença e menos sensível para diagnosticar a doença.

Ce travail contribue à la théorie de la décision possibiliste et plus précisément à la prise de décision séquentielle dans le cadre de la théorie des possibilités, à la fois au niveau théorique et pratique. Bien qu'attrayante pour sa capacité à résoudre les problèmes de décision qualitatifs, la théorie de la décision possibiliste souffre d'un inconvénient important : les critères d'utilité qualitatives possibilistes comparent les actions avec les opérateurs min et max, ce qui entraîne un effet de noyade. Pour surmonter ce manque de pouvoir décisionnel, plusieurs raffinements ont été proposés dans la littérature. Les raffinements lexicographiques sont particulièrement intéressants puisqu'ils permettent de bénéficier de l'arrière-plan de l'utilité espérée, tout en restant "qualitatifs". Cependant, ces raffinements ne sont définis que pour les problèmes de décision non séquentiels. Dans cette thèse, nous présentons des résultats sur l'extension des raffinements lexicographiques aux problèmes de décision séquentiels, en particulier aux Arbres de Décision et aux Processus Décisionnels de Markov possibilistes. Cela aboutit à des nouveaux algorithmes de planification plus "décisifs" que leurs contreparties possibilistes. Dans un premier temps, nous présentons des relations de préférence lexicographiques optimistes et pessimistes entre les politiques avec et sans utilités intermédiaires, qui raffinent respectivement les utilités possibilistes optimistes et pessimistes. Nous prouvons que les critères proposés satisfont le principe de l'efficacité de Pareto ainsi que la propriété de monotonie stricte. Cette dernière garantit la possibilité d'application d'un algorithme de programmation dynamique pour calculer des politiques optimales. Nous étudions tout d'abord l'optimisation lexicographique des politiques dans les Arbres de Décision possibilistes et les Processus Décisionnels de Markov à horizon fini. Nous fournissons des adaptations de l'algorithme de programmation dynamique qui calculent une politique optimale en temps polynomial. Ces algorithmes sont basés sur la comparaison lexicographique des matrices de trajectoires associées aux sous-politiques. Ce travail algorithmique est complété par une étude expérimentale qui montre la faisabilité et l'intérêt de l'approche proposée. Ensuite, nous prouvons que les critères lexicographiques bénéficient toujours d'une fondation en termes d'utilité espérée, et qu'ils peuvent être capturés par des utilités espérées infinitésimales. La dernière partie de notre travail est consacrée à l'optimisation des politiques dans les Processus Décisionnels de Markov (éventuellement infinis) stationnaires. Nous proposons un algorithme d'itération de la valeur pour le calcul des politiques optimales lexicographiques. De plus, nous étendons ces résultats au cas de l'horizon infini. La taille des matrices augmentant exponentiellement (ce qui est particulièrement problématique dans le cas de l'horizon infini), nous proposons un algorithme d'approximation qui se limite à la partie la plus intéressante de chaque matrice de trajectoires, à savoir les premières lignes et colonnes. Enfin, nous rapportons des résultats expérimentaux qui prouvent l'efficacité des algorithmes basés sur la troncation des matrices. / This work contributes to possibilistic decision theory and more specifically to sequential decision-making under possibilistic uncertainty, at both the theoretical and practical levels. Even though appealing for its ability to handle qualitative decision problems, possibilisitic decision theory suffers from an important drawback: qualitative possibilistic utility criteria compare acts through min and max operators, which leads to a drowning effect. To overcome this lack of decision power, several refinements have been proposed in the literature. Lexicographic refinements are particularly appealing since they allow to benefit from the expected utility background, while remaining "qualitative". However, these refinements are defined for the non-sequential decision problems only. In this thesis, we present results on the extension of the lexicographic preference relations to sequential decision problems, in particular, to possibilistic Decision trees and Markov Decision Processes. This leads to new planning algorithms that are more "decisive" than their original possibilistic counterparts. We first present optimistic and pessimistic lexicographic preference relations between policies with and without intermediate utilities that refine the optimistic and pessimistic qualitative utilities respectively. We prove that these new proposed criteria satisfy the principle of Pareto efficiency as well as the property of strict monotonicity. This latter guarantees that dynamic programming algorithm can be used for calculating lexicographic optimal policies. Considering the problem of policy optimization in possibilistic decision trees and finite-horizon Markov decision processes, we provide adaptations of dynamic programming algorithm that calculate lexicographic optimal policy in polynomial time. These algorithms are based on the lexicographic comparison of the matrices of trajectories associated to the sub-policies. This algorithmic work is completed with an experimental study that shows the feasibility and the interest of the proposed approach. Then we prove that the lexicographic criteria still benefit from an Expected Utility grounding, and can be represented by infinitesimal expected utilities. The last part of our work is devoted to policy optimization in (possibly infinite) stationary Markov Decision Processes. We propose a value iteration algorithm for the computation of lexicographic optimal policies. We extend these results to the infinite-horizon case. Since the size of the matrices increases exponentially (which is especially problematic in the infinite-horizon case), we thus propose an approximation algorithm which keeps the most interesting part of each matrix of trajectories, namely the first lines and columns. Finally, we reports experimental results that show the effectiveness of the algorithms based on the cutting of the matrices.

Severe declines in the number of some songbirds over the last 40 years have caused heated debate amongst interested parties. Many factors have been suggested as possible causes for these declines, including an increase in the abundance and distribution of an avian predator, the Eurasian sparrowhawk Accipiter nisus. To test for evidence for a predator effect on the abundance of its prey, we analyse data on 10 species visiting garden bird feeding stations monitored by the British Trust for Ornithology in relation to the abundance of sparrowhawks. We apply Bayesian hierarchical models to data relating to averaged maximum weekly counts from a garden bird monitoring survey. These data are essentially continuous, bounded below by zero, but for many species show a marked spike at zero that many standard distributions would not be able to account for. We use the Tweedie distributions, which for certain areas of parameter space relate to continuous nonnegative distributions with a discrete probability mass at zero, and are hence able to deal with the shape of the empirical distributions of the data. The methods developed in this thesis begin by modelling single prey species independently with an avian predator as a covariate, using MCMC methods to explore parameter and model spaces. This model is then extended to a multiple-prey species model, testing for interactions between species as well as synchrony in their response to environmental factors and unobserved variation. Finally we use a relatively new methodological framework, namely the SPDE approach in the INLA framework, to fit a multi-species spatio-temporal model to the ecological data. The results from the analyses are consistent with the hypothesis that sparrowhawks are suppressing the numbers of some species of birds visiting garden feeding stations. Only the species most susceptible to sparrowhawk predation seem to be affected.

In this thesis we provide a unifying framework for two decades of work in an area of Machine Learning known as cost-sensitive Boosting algorithms. This area is concerned with the fact that most real-world prediction problems are asymmetric, in the sense that different types of errors incur different costs. Adaptive Boosting (AdaBoost) is one of the most well-studied and utilised algorithms in the field of Machine Learning, with a rich theoretical depth as well as practical uptake across numerous industries. However, its inability to handle asymmetric tasks has been the subject of much criticism. As a result, numerous cost-sensitive modifications of the original algorithm have been proposed. Each of these has its own motivations, and its own claims to superiority. With a thorough analysis of the literature 1997-2016, we find 15 distinct cost-sensitive Boosting variants - discounting minor variations. We critique the literature using {\em four} powerful theoretical frameworks: Bayesian decision theory, the functional gradient descent view, margin theory, and probabilistic modelling. From each framework, we derive a set of properties which must be obeyed by boosting algorithms. We find that only 3 of the published Adaboost variants are consistent with the rules of all the frameworks - and even they require their outputs to be calibrated to achieve this. Experiments on 18 datasets, across 21 degrees of cost asymmetry, all support the hypothesis - showing that once calibrated, the three variants perform equivalently, outperforming all others. Our final recommendation - based on theoretical soundness, simplicity, flexibility and performance - is to use the original Adaboost algorithm albeit with a shifted decision threshold and calibrated probability estimates. The conclusion is that novel cost-sensitive boosting algorithms are unnecessary if proper calibration is applied to the original.

Les attributs environnementaux, de plus en plus présents sur les produits de consommation, sont des attributs de croyance, renvoyant à un bien non marchand (i.e. la protection de l'environnement) et à des valeurs morales, ce qui les rend difficiles à évaluer par les consommateurs. Ce travail doctoral s'intéresse précisément à la manière dont les consommateurs utilisent un attribut environnemental (certifié ou non certifié) dans leur évaluation de produit. Cette problématique de recherche est abordée selon trois questions de recherche. La première s'intéresse à l'effet du nombre d'attributs (1 vs. 2) sur le poids associé à l'attribut environnemental. Cette question renvoie à l'étude d'un biais appelé effet d'inclusion qui est caractérisé par une insensibilité au nombre d'attributs. La deuxième question étudie le biais lié au mode d'évaluation sur le poids associé à un attribut environnemental. Les effets de deux critères du mode d'évaluation sont étudiés : le mode de réponse (monétaire vs. non monétaire) et le mode de présentation (jointe vs. séparée). La troisième prolonge l'étude du mode d'évaluation et de son effet sur le poids de l'attribut environnemental en étudiant la validité prédictive du mode d'évaluation pour estimer les préférences des consommateurs. Ces biais liés au contexte d'évaluation sont étudiés, montrés et discutés au moyen de cinq expérimentations et pour différents attributs environnementaux et stimuli. La validité prédictive du mode d'évaluation met en évidence l'intérêt d'utiliser le mode « consentement à payer ». D'un point de vue théorique, cette recherche montre l'intérêt de prendre en compte l'effet d'inclusion et le biais lié au mode d'évaluation pour des personnes intéressées par l'évaluation d'attributs environnementaux et montre également la limite potentielle à l'ajout d'un attribut environnemental sur un produit de consommation. D'un point de vue méthodologique, ce travail doctoral propose une approche permettant de comparer les poids obtenus entre les modes d'évaluation. Cette recherche suggère aux chefs de produits de prendre en compte ces caractéristiques contextuelles pour améliorer la précision de l'estimation des préférences des consommateurs pour un produit comportant un attribut environnemental. Ainsi, la valeur associée à un attribut environnemental varie lorsque cet attribut est seul sur le produit ou en présence d'un autre attribut. De même, le mode d'évaluation « consentement à payer » semble mieux prédire les préférences réelles pour un attribut environnemental que le mode d'évaluation « choix entre deux options ». / Environmental attributes are increasingly being included in consumption products. Because environmental attributes are credence attributes, they refer to a public good (i.e. environmental protection), and they reference moral values, they are difficult for consumers to assess. This doctoral research specifically focuses on how consumers use an environmental attribute (certified or uncertified) in product evaluation. This core question is addressed through three research questions. The first deals with the effect of attribute number (1 vs. 2) on the weight given to environmental attributes. This study focuses on a counterintuitive effect named the embedding effect, characterized by insensitivity to the number of environmental attributes present. The second question concerns the effect of the method of value measurement on the weight given to an environmental attribute. The effects of two criteria are studied: response mode (pricing vs. non pricing) and evaluation mode (joint vs. separate). The third and last question extends the study of evaluation measure and its effect on the weight of an environmental attribute by testing the predictive validity of the evaluation measure on consumer preferences. These research questions are investigated with five experiments that employ various environmental attributes and stimuli. From a theoretical perspective, this research shows the importance of considering the embedding effect and evaluation measurement bias for people interested in environmental attributes evaluation. It also underlines the potential limit of adding an environmental attribute to a consumption product. In terms of utility measurement, the thesis shows, counterintuitively, that willingness to pay is a better measure of environmental values than is choice. From a methodological standpoint, this doctoral thesis proposes an approach to enable attribute weight comparisons across measures. This research suggests to product managers that they need to be aware of these contextual factors when assessing and predicting consumer preferences for a product with an environmental attribute. If the environmental attribute in the finished product will be alone versus in conjunction with another environmental attribute, for example, affects attribute utility. Likewise, willingness to pay appears to be a better predictor of actual preference for environmental attribute than is choice.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The recent insertion of the electronic process befalls wrapped in a series of questions, since in relation to the more practical aspects of procedural routine to the most sensitive issues, particularly in relation to the exponential use of the technique, which requires a deep reflection. Alongside this context, it is notorious double face of crisis, which is revealed in a model of the problem of law and the philosophical and interpretive parameter, and, to some extent, prevents the "happening" of decision theory of Constitutionalism Contemporary. This dissertation, based on this situation, seek to answer the following question: to what extent the civil process of electronic nature, to be conceived in the wake of new information and communication technologies (TIC`s), puts at risk what has been She won in terms of the theory of decidability, based on the Constitutionalism Contemporary assumptions? Therefore, the "methodology" approach used is the "method" phenomenological-hermeneutic and, as regards methods of procedure, monographic and historical methods will be adopted. It is concluded that there are risks inherent in the extensive use of technology, which, when used to promote the process of virtualization, deepens the risk of a poor justice meaning, dematerialized, destemporalizada and dehumanized, to strengthen and enhance the hallucinatory search referring quantitative results in practice due to acceleration of pressing technique. The big concern is imminent deepening the gap between factual world (case) and legal world, exacerbating the difficulty of inserting the judge-interpreter in the hermeneutic situation, substantially changing the hermeneutic circle and the ontological difference. It remains, therefore, a long road still to be traveled in order to speak of a real "happening" of Constitutionalism Contemporary decision theory and, therefore, it is essential to discuss / prevent that put into question what has been achieved. / A recente inserção do processo eletrônico sobrevém envolta em uma série de questionamentos, desde em relação a aspectos mais práticos do cotidiano processual a questões mais delicadas, em especial no que tange ao exponencial uso da técnica, que requerem uma profunda reflexão. A par desse contexto, é notória uma crise de dupla face, que se revela em um problema do modelo do direito e do parâmetro filosófico-interpretativo, e, em determinada medida, impede o acontecer da teoria da decisão do Constitucionalismo Contemporâneo. A presente dissertação, com base nessa conjuntura, buscará responder ao seguinte questionamento: em que medida o processo civil de cariz eletrônico, ao ser concebido no bojo das novas tecnologias da informação e comunicação (TIC`s), coloca em risco o que já se conquistou em termos de teoria da decidibilidade, pautada nos pressupostos do Constitucionalismo Contemporâneo? Para tanto, a metodologia de abordagem utilizada será o método fenomenológicohermenêutico e, no que concerne aos métodos de procedimento, serão adotados os métodos monográfico e histórico. Conclui-se que existem riscos inerentes ao uso extensivo da técnica, a qual, ao ser utilizada para a promoção da virtualização do processo, aprofunda os riscos de uma justiça carente de significação, desmaterializada, destemporalizada e desumanizada, por fortalecer e incrementar a alucinante busca de resultados quantitativos na prática jurisdicional, devido a premente aceleração da técnica. A grande preocupação é com um iminente aprofundamento do abismo entre mundo fático (caso concreto) e mundo jurídico, agravando a dificuldade da inserção do juiz-intérprete na situação hermenêutica, alterando substancialmente o círculo hermenêutico e a diferença ontológica. Resta, portanto, um longo caminho ainda a ser percorrido para poder-se falar em um verdadeiro acontecer da teoria da decisão do Constitucionalismo Contemporâneo e, para tanto, é imprescindível discutir/prevenir que se ponha em xeque o que já se conquistou.

Submitted by ERIKA REGINA DA SILVA GALLO null (erika.regina@hotmail.com) on 2016-09-15T19:21:48Z No. of bitstreams: 1 Versão Final.pdf: 1593845 bytes, checksum: e99e5b67cc74cf377bfd115091bec7e9 (MD5) / Approved for entry into archive by Juliano Benedito Ferreira (julianoferreira@reitoria.unesp.br) on 2016-09-21T20:08:06Z (GMT) No. of bitstreams: 1 gallo_ers_me_arafcl.pdf: 1593845 bytes, checksum: e99e5b67cc74cf377bfd115091bec7e9 (MD5) / Made available in DSpace on 2016-09-21T20:08:06Z (GMT). No. of bitstreams: 1 gallo_ers_me_arafcl.pdf: 1593845 bytes, checksum: e99e5b67cc74cf377bfd115091bec7e9 (MD5) Previous issue date: 2016-09-09 / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / Neste trabalho temos a intenção de contribuir com o debate na área de economia e finanças comportamentais ao realizar um estudo sobre o comportamento do mercado acionário ao incluirmos o viés de comportamento aversão à perda aos agentes que ali operam. Para tanto, em um primeiro momento, fez-se uma concisa revisão da teoria da decisão na escola econômica e seus desdobramentos em finanças, no que tange à Hipótese dos Mercados Eficientes. Em um segundo momento, apresentamos um modelo de agente aplicado ao mercado acionário que foi programado em software livre NetLogo, cujo método é, em parte, baseado em modelos de agentes já programados para mercados financeiros artificiais e, ao mesmo tempo, parcialmente novo ao propor a realização de testes com parâmetros diferentes dos utilizados por outros autores – a saber: aversão à perda. Os resultados encontrados sugerem que a subjetividade humana presente na tomada de decisão, isto é, quando os agentes possuem aversão à perda, faz com que o movimento do mercado acionário artificial apresente alguns ruídos. Destarte, concluímos que os experimentos realizados nos oferecem indícios de que há certa fragilidade em alguns pressupostos da Hipótese dos Mercados Eficientes. / In this work, we intend to contribute to the debate in behavioral economics and finance to conduct a study on how Stock Markets behaves by including the loss averse agent's bias on this environment. At first, we built a concise review of the decision theory on economic school and its developments in finance, regarding to Efficient Market Hypothesis. Secondly, we presented an agent-based model applied to the stock market which has been programmed in the free software NetLogo, whose has, in part, agent-based models already programmed for artificial financial markets and at the same time, is partially new to propose conduction tests, which differ from other parameters used by several authors such as: loss aversion. The results suggest that human's subjectivity, inherently presented in decision-making, that is, when agents have loss aversion, it may cause the artificial stock market movement to present some noise. Thus, we conclude that the experiments give us evidence that there is some fragility on the Efficient Market Hypothesis. / CNPq: 130136/2015-8

No âmbito da Assistência Social, existe a necessidade de se identificar as famílias em situação de vulnerabilidade e risco social, processo esse chamado de “Busca Ativa”, para que as famílias nesta situação possam ser assistidas adequadamente. O Ministério do Desenvolvimento Social e Combate à Fome do Brasil orienta que seja realizado o cruzamento de bases de dados como forma de realizar a Busca Ativa, mas não disponibiliza nenhuma ferramenta para realização desse processo. Este trabalho busca identificar e aplicar técnicas de mineração de dados para apoiar a identificação das famílias em situação de vulnerabilidade e risco social. Os resultados obtidos em experimentos preliminares demonstraram que na maioria dos casos os modelos gerados preveem sempre a classe majoritária. Após realizar um balanceamento manual das classes removendo algumas amostras os experimentos foram repetidos e indicaram que os resultados estavam sendo diretamente afetados devido ao desbalanceamento das classes. Por esse motivo foram utilizados diversos métodos específicos para realizar o balanceamento das amostras a fim de que todas as classes possuíssem a mesma quantidade de amostras. Após realizar o balanceamento das amostras novos experimentos foram realizados. Durante a análise dos resultados foi observado que com as medidas padrões de avaliação de aprendizado de máquina não estava sendo possível identificar qual método havia obtido o melhor resultado. Em função disso um método de qualidade de ranking foi utilizado juntamente com a medida Recall para avaliar os resultados. / In the current Brazilian Government there is a Social Assistance policy that is highly concerned about helping families who might be at social risk and vulnerability. The process of identification of these families is known as “active search”. The task of active search is defined in a document by the Brazilian Ministry of Social Development and Fight Against Hunger. This document provides the main guidelines about how to perform the active search. However, despite the task’s importance, there are still no tool to help the social assistants with this task. This work aim to investigate the use of data mining techniques to identify the families in vulnerability and social risk situations. The results obtained in preliminary experiments showed that the classification models created always predict the majority class. After balancing manually the datasets by removing some examples the experiments were repeated and showed that the results were being directly influenced by the imbalanced data. Because of it was used a bunch of sampling methods to produce the same amount of examples in each class. After proceed with the sampling of the examples new experiments were proceeded. During the result’s evaluation it was realized that the standard metrics used in machine learn were not being able to identify wich method obtained the best result. Due to this situation a ranking quality method was used combined with the Recall metric to evaluate the results.