Neste trabalho foram revisados os principais modelos, para a avaliação do risco de crédito e para o provisionamento de perdas com clientes, concluindo-se com uma proposta de adoção de um modelo estatístico, com o objetivo de medir o risco associado ao financiamento e empréstimo a clientes, com o conseqüente impacto na mensuração dos ativos. Sem o objetivo de exaurir o assunto, foram adotados os passos relacionados a seguir para o desenvolvimento do tema até a proposição final. Na introdução, são feitas as justificativas sobre o tema, qual a questão problema associada ao tema e os desafios da contabilidade quanto à mensuração dos ativos. Em relação à gestão de riscos, são relacionados os tipos de riscos em geral, detalhado o risco de crédito em particular e avaliados os modelos de concessão de crédito. Sobre a constituição da Provisão para Devedores Duvidosos, foram pesquisados os principais autores de contabilidade e de finanças, onde se constatou proposições semelhantes, que podem ser resumidas em 4 modelos de provisionamento para Perdas com Devedores Duvidosos: 1. Baixa – "Write-off"; 2. Percentual sobre as vendas; 3. Percentual sobre o montante de contas a receber; 4. Idade da carteira – "aging". Em seguida são analisadas as correlações entre os modelos de previsão de insolvência e as perdas com crédito, onde é possível identificar que os modelos de insolvência são úteis para a concessão do crédito, mas pouco utilizados para a estimativa da perda provável com devedores duvidosos. Em 21 de dezembro de 1999, o Banco Central do Brasil, emitiu a Resolução 2.682, na qual recomenda às instituições financeiras que alterem suas metodologias de provisionamento para perdas com devedores duvidosos. O Banco Central, entretanto, não indica qual modelo utilizar, deixando a cargo de cada instituição o desenvolvimento dos modelos. Utilizando a norma do Banco Central como referência, e procurando um embasamento científico para a constituição da PDD, é proposto um modelo para a sua constituição, modelo esse testado e avaliado, tanto em conformidade às normas do Banco Central, como com orientação gerencial. Para tanto, foi desenvolvido um modelo estatístico, aplicando-se a técnica da regressão logística, a 202 clientes de uma instituição financeira, onde foi possível concluir-se que o uso do modelo, na constituição da Provisão para Devedores Duvidosos, poderá trazer benefícios na mensuração do real valor dos investimentos em contas a receber. / This project aims at evaluating the used models and proposing the adoption of new models, for the Allowance for Doubtful Accounts, with the objective of measuring the risk related to customers financing and loan activities, and the resulting impact in assets measurements. In order to achieve this goal and try not to over exploit the subject, the following steps related to the development of the theme were adopted. In the introduction the theme is explained; the main issue associated to the theme and the challenges the accounting has to face concerning the assets measurements. As for the risk management the general kinds of risks are described, particularly the credit risk and the credit concession models are evaluated. Referring to the Allowance for Doubtful Accounts constitution, most meaningful authors in the field of Accounting and Finance were researched and similar propositions underlined their writings; that can be summarized in 4 allowance models for Doubtful Accounts. 1. Write off; 2. Percentage over the sales; 3. Percentage over receivables; 4. Aging. Then the correlations between the models of insolvency prediction and the credit losses are analyzed where it is possible to verify that the insolvency models are useful for credit concession, though not very much used for estimating probable loss with Doubtful Accounts. The Central Bank of Brazil (Banco Central do Brasil) issued the act number 2682 on the 21st of December, 1999, that urges all financial institutions to change their methodologies of Allowance for Doubtful Accounts. The Central Bank, however, does not indicate the model to be used, leaving the task of developing the models to each institution. Based on the policy of Central Bank and keeping a scientific approach to the constitution of Allowance for Doubtful Accounts, a model for their constitution in the portfolio is proposed. Such model is tested and evaluated not only, according to the rules of Central Bank of Brazil, but also in terms of management orientation. Having this purpose in mind a statistic model was developed, using LOGIT Regression applied to 202 customers of a financial institution where it was possible to come to the conclusion that the use of the model in the constitution of the Allowance for Doubtful Accounts can bring benefits in measuring the real value of investments in Receivables.

Le développement des véhicules autonomes a reçu une attention croissant ces dernières années, notamment les secteurs de la défense et de l'industrie automobile. L'intérêt pour l'industrie automobile est motivé par la conception de véhicules sûrs et confortables. Une raison commune derrière la plupart des accidents de la circulation est due au manque de vigilance du conducteur sur la route. Cette thèse se trouve dans le problématique de l'estimation des risques de collision pour un véhicule dans les secondes qui suivent en condition de circulation urbaines. Les systèmes actuellement disponibles dans le commerce sont pour la plupart conçus pour prévenir les collisions avant, arrières, ou latérales. Ces systèmes sont généralement équipés d'un capteur de type radar, à l'arrière, à l'avant ou sur les côtés pour mesurer la vitesse et la distance aux obstacles. Les algorithmes pour déterminer le risque de collision sont fondés sur des variantes du TTC (time-to-collision en anglais). Cependant, un véhicule peut se trouver dans des situations où les routes ne sont pas droites et l'hypothèse que le mouvement est linéaire ne tient pas pour le calcul du TTC. Dans ces situations, le risque est souvent sous-estimé. De plus, les instances où les routes ne sont pas tout droit se trouve assez souvent dans les environnement urbain ; par exemple, les rond point ou les intersections. Un argument de cette thèse est que, savoir simplement qu'il y ait un objet à une certaine position et à une instance spécifique dans le temps ne suffit pas à évaluer sa sécurité dans le futur. Un système capable de comprendre les comportements de déplacement du véhicule est indispensable. En plus, les contraintes environnementales doivent être prises en considération. Le cas le plus simple du mouvement « libre » est d'abord traité. Dans cette situation il n'ya pas de contraintes environnementales ou de comportement explicite. Ensuite, les contraintes environnementales des routes sur trafic urbain et le comportement des conducteurs des véhicules sont introduits et pris en compte explicitement. Cette thèse propose un modèle probabiliste pour les trajectoires des véhicules fondé sur le processus gaussien (GP). Son avantage est le pouvoir d'exprimer le mouvement dans le futur indépendamment de la discrétisation d'espace et d'état. Les comportements des conducteurs sont modélisés avec une variante du modèle de Markov caché. La combinaison de ces deux modèles donne un modèle probabiliste de l'évolution complète du véhicule dans le temps. En plus, une méthode générale pour l'évaluation probabiliste des risques de collision est présentée, où différentes valeurs de risque, chacun avec sa propre sémantique.

Benzene is an important toxic chemical in urban air and known human carcinogen released substantially by mobile sources. It's important to understand the spatial variation of benzene concentrations in order to understand exposures of susceptible sub-populations such as children and minority groups. Current monitoring networks use large and expensive air samplers that require electricity and restrict the location and number of samplers, not allowing for fine spatial resolution data. The goals of this study are to develop and evaluate protocols for passive sampling and analysis of ambient benzene concentrations, and conduct a pilot study investigating small-scale variations over an area where children are likely to be exposed. Protocols were developed for the use and analysis of the Radiello RAD130 passive sampler for field sampling over the spatial scale of a city park adjacent to an elementary school. A pilot study was conducted from 4/27/11-5/4/11, where 11 samplers were exposed for a seven day sampling period at the park. After sampler exposure, benzene concentrations were determined through solvent desorption followed by analysis using a Varian gas chromatograph with mass spectrometer. Co-location with the existing regulatory active sampler in the county and of two samplers at the same site was done to evaluate the accuracy and precision of the methods, respectively. Health risk estimates were calculated using risk assessment guidance from the U.S. and California Environmental Protection Agencies. Concentrations over the park were found to range from 0.23 0.34 µg m^-3 with a coefficient of variation of 11%. A relative percent difference of 3% was found between the co-located sampler and the active sampler, and a 14% relative percent difference was found between the two duplicate samplers. The variation in health risk from concentration variation due to sampler placement contributed less to the overall uncertainty in the estimates than the uncertainty built in to the calculation parameters of inhalation unit risk and cancer potency factor, as estimated by the U.S. EPA and California EPA, respectively. These results suggest that the exposure of an individual at the park would be characterized sufficiently for standard health risk analysis through the use of one sampler. Further research is necessary into using passive samplers over both the same spatial scale in other areas, as well as on a larger scale to determine intra-urban benzene concentration distributions. The protocols developed here will be used in a future planned study of benzene concentration measurements to characterize neighborhood-scale exposures in Hillsborough County.

Older people admitted to hospital are at high risk of rehospitalization and medication errors. We have demonstrated, in a randomized controlled trial, that a clinical pharmacist intervention reduces the incidence of revisits to hospital for patients aged 80 years or older admitted to an acute internal medicine ward. The aims of this thesis were to further study the effects of the intervention and to investigate possibilities of targeting the intervention by identifying predictors of treatment response or adverse health outcomes. The effect of the pharmacist intervention on the appropriateness of prescribing was assessed, by using three validated tools. This study showed that the quality of prescribing was improved for the patients in the intervention group but not for those in the control group. However, no association between the appropriateness of prescribing at discharge and revisits to hospital was observed. Subgroup analyses explored whether the clinical pharmacist intervention was equally effective in preventing emergency department visits in patients with few or many prescribed drugs and in those with different levels of inappropriate prescribing on admission. The intervention appeared to be most effective in patients taking fewer drugs, but the treatment effect was not altered by appropriateness of prescribing. The most relevant risk factors for rehospitalization and mortality were identified for the same study population, and a score for risk-estimation was constructed and internally validated (the 80+ score). Seven variables were selected. Impaired renal function, pulmonary disease, malignant disease, living in a nursing home, being prescribed an opioid and being prescribed a drug for peptic ulcer or gastroesophageal reflux disease were associated with an increased risk, while being prescribed an antidepressant drug (tricyclic antidepressants not included) was linked with a lower risk. These variables made up the components of the 80+ score. Pending external validation, this score has potential to aid identification of high-risk patients. The last study investigated the occurrence of prescription errors when patients with multi-dose dispensed (MDD) drugs were discharged from hospital. Twenty-five percent of the MDD orders contained at least one medication prescription error. Almost half of the errors were of moderate or major severity, with potential to cause increased health-care utilization.

Dans un marché de l'assurance vie en Afrique subsaharienne francophone à la traîne, mais promis à un bel avenir en cas d'émergence de solutions techniques et commerciales endogènes, la thèse propose des outils théoriques et opérationnels adaptés à son développement. Cette démarche s'inscrit en parallèle des actions entreprises par l'autorité de contrôle régionale (la CIMA) pour fournir aux assureurs de la région des outils adaptés. En effet, la CIMA a initié des travaux pour la construction de nouvelles tables réglementaires d'expérience, ce qui a permis de fournir des références fiables et pertinentes pour la mortalité de la population assurée dans la région. Toutefois, certaines problématiques techniques utiles n'ont pas été développées dans ces travaux de construction. La thèse leur accorde alors une attention particulière. Ainsi, d'une part, la thèse permet de fournir des outils pour tenir compte des différences de mortalité entre pays de la région, tout en limitant les risques systématiques liés aux fluctuations d'échantillonnage (dues à la petite taille des échantillons de données par pays). Il apparaît notamment que si la modélisation indépendante de chaque pays n'est pas appropriée, les modèles d'hétérogénéité à facteurs observables, tels que le modèle de Cox ou de Lin et Ying, permettent d'atteindre cet objectif. On précise toutefois ici que ces modèles d'hétérogénéité ne permettent pas de supprimer le risque systématique lié aux fluctuations d'échantillonnage lors de l'estimation du modèle, ils engendrent seulement une réduction de ce risque en contrepartie d'une augmentation du risque systématique lié au choix du modèle. D'autre part, la thèse permet également de fournir des outils pour modéliser la mortalité d'expérience future dans la région. En absence de données sur les tendances passées de la mortalité d'expérience, ni le modèle classique de Lee-Carter ni ses extensions ne sont applicables. Une solution basée sur un ajustement paramétrique, une hypothèse sur la forme de l'évolution du niveau de mortalité (évolution linaire ou exponentielle) et un avis d'expert sur l'espérance de vie générationnelle à un âge donné est alors proposée (ces travaux s'appuient sur le modèle de Bongaarts). Ensuite, dans un second temps, en supposant disposer de données sur les tendances passées (ce qui pour mémoire n'est pas le cas à ce stade dans la région, mais devrait l'être dans les prochaines années), la thèse propose une modélisation de la mortalité future à partir d'une référence de mortalité externe et une analyse des risques systématiques associés (risques liés aux fluctuations d'échantillonnage et au choix de la référence de mortalité) / In a market of life insurance in sub-Saharan Francophone behind, but with a bright future in the event of emergence of endogenous technical and commercial solutions, the thesis provides theoretical and operational tools adapted to its development. This approach is in parallel with actions taken by the supervisory authority regional (CIMA) to provide insurance tools in the region. Indeed, CIMA has initiated work to construct new tables regulatory experience, which has provided reliable and relevant references for mortality of the insured population in the region. However, some useful technical issues were not developed in such construction. The thesis then give them special attention. Thus, on the one hand, the theory can provide tools to account for differences in mortality between countries in the region, while limiting the risks associated with systematic sampling fluctuations (due to small sample sizes data countries). Particular, it appears that if the model independent of each country is not appropriate, models of heterogeneity with observable factors, such as the Cox or Lin and Ying model, can achieve this goal. However, it says here that these models of heterogeneity does not eliminate the systematic risk due to sampling fluctuations when estimating the model, they generate only a reduction of this risk in exchange for an increase in systematic risk associated the choice of model. On the other hand, the thesis can also provide tools to model future mortality experience in the region. In the absence of data on past trends in mortality experience, nor the classical Lee-Carter or its extensions are applicable. A solution based on a parametric adjustment, an assumption about the form of changes in the level of mortality (exponential or linear trend) and an expert opinion on the generational life expectancy at a given age is then proposed (this work based on the model of Bongaarts). Then in a second time, assuming availability of data on past trends (which for the record is not the case at this stage in the region, but should be in the coming years), the thesis proposes a model of future mortality from an external reference mortality and analyzes associated systematic risk (risk of sampling fluctuations and on the choice of the reference of mortality)

京都大学 / 新制・論文博士 / 博士(医学) / 乙第13398号 / 論医博第2222号 / 新制||医||1051(附属図書館) / (主査)教授 妹尾 浩, 教授 中山 健夫, 教授 西浦 博 / 学位規則第4条第2項該当 / Doctor of Medical Science / Kyoto University / DFAM

When analyzing the flood risk using multi-criteria analysis it is necessary to estimate

Autonomous vehicles promise to play an important role aiming at increased efficiency and safety in road transportation. Although we have seen several examples of autonomous vehicles out on the road over the past years, how to ensure the safety of autonomous vehicle in the uncertain and dynamic environment is still a challenging problem. This thesis studies this problem by developing a risk-aware decision making framework. The system that integrates the dynamics of an autonomous vehicle and the uncertain environment is modelled as a Partially Observable Markov Decision Process (POMDP). A risk measure is proposed based on the Responsibility-Sensitive Safety (RSS) distance, which quantifying the minimum distance to other vehicles for ensuring safety. This risk measure is incorporated into the reward function of POMDP for achieving a risk-aware decision making. The proposed risk-aware POMDP framework is evaluated in two case studies. In a single-lane car following scenario, it is shown that the ego vehicle is able to successfully avoid a collision in an emergency event where a vehicle in front of it makes a full stop. In the merge scenario, the ego vehicle successfully enters the main road from a ramp with a satisfactory distance to other vehicles. As a conclusion, the risk-aware POMDP framework is able to realize a trade-off between safety and usability by keeping a reasonable distance and adapting to other vehicles behaviours. / Autonoma fordon förutspås spela en stor roll i framtiden med målen att förbättra effektivitet och säkerhet för vägtransporter. Men även om vi sett flera exempel av autonoma fordon ute på vägarna de senaste åren är frågan om hur säkerhet ska kunna garanteras ett utmanande problem. Det här examensarbetet har studerat denna fråga genom att utveckla ett ramverk för riskmedvetet beslutsfattande. Det autonoma fordonets dynamik och den oförutsägbara omgivningen modelleras med en partiellt observerbar Markov-beslutsprocess (POMDP från engelskans “Partially Observable Markov Decision Process”). Ett riskmått föreslås baserat på ett säkerhetsavstånd förkortat RSS (från engelskans “Responsibility-Sensitive Safety”) som kvantifierar det minsta avståndet till andra fordon för garanterad säkerhet. Riskmåttet integreras i POMDP-modellens belöningsfunktion för att åstadkomma riskmedvetna beteenden. Den föreslagna riskmedvetna POMDP-modellen utvärderas i två fallstudier. I ett scenario där det egna fordonet följer ett annat fordon på en enfilig väg visar vi att det egna fordonet kan undvika en kollision då det framförvarande fordonet bromsar till stillastående. I ett scenario där det egna fordonet ansluter till en huvudled från en ramp visar vi att detta görs med ett tillfredställande avstånd till andra fordon. Slutsatsen är att den riskmedvetna POMDP-modellen lyckas realisera en avvägning mellan säkerhet och användbarhet genom att hålla ett rimligt säkerhetsavstånd och anpassa sig till andra fordons beteenden.

Radiation from Computed Tomography (CT) is now the major contributor to population radiation dose. Despite controversy around the dose-effect relationship of radiation from CT, the linear non-threshold (LNT) theory is endorsed by many authorities, and constitutes the basis of cancer risk estimates. The purpose of this study was (1) a literature review of radiobiological theories, and methods of dose saving stategies in CT; (2) to highlight the importance of dose saving in CT, and to demonstrate how dose can be saved in a radiology department: Following a 40% reduction in reference X-ray tube current for a CT of the urinary tract, the effecitve dose and estimated lifetime attributable risk of incident cancer due to this CT in a group (n=103) were reduced by 37% and 38% in an age and sex-matched group respectively. The literature review showed that the public health implications of CT radiation exposure remain uncertain. / Health Studies / M.A. (Public Health)

Thesis (PhD (Statistics and Actuarial Sciences))--Stellenbosch University, 2008. / We consider the problem of model assessment by risk estimation. Various approaches to risk estimation are considered in a uni ed framework. This a discussion of various complexity dimensions and approaches to obtaining bounds on covering numbers is also presented. The second type of training sample interval estimator discussed in the thesis is Rademacher bounds. These bounds use advanced concentration inequalities, so a chapter discussing such inequalities is provided. Our discussion of Rademacher bounds leads to the presentation of an alternative, slightly stronger, form of the core result used for deriving local Rademacher bounds, by avoiding a few unnecessary relaxations. Next, we turn to a discussion of PAC-Bayesian bounds. Using an approach developed by Olivier Catoni, we develop new PAC-Bayesian bounds based on results underlying Hoe ding's inequality. By utilizing Catoni's concept of \exchangeable priors", these results allowed the extension of a covering number-based result to averaging classi ers, as well as its corresponding algorithm- and data-dependent result. The last contribution of the thesis is the development of a more exible shell decomposition bound: by using Hoe ding's tail inequality rather than Hoe ding's relative entropy inequality, we extended the bound to general loss functions, allowed the use of an arbitrary number of bins, and introduced between-bin and within-bin \priors". Finally, to illustrate the calculation of these bounds, we applied some of them to the UCI spam classi cation problem, using decision trees and boosted stumps. framework is an extension of a decision-theoretic framework proposed by David Haussler. Point and interval estimation based on test samples and training samples is discussed, with interval estimators being classi ed based on the measure of deviation they attempt to bound. The main contribution of this thesis is in the realm of training sample interval estimators, particularly covering number-based and PAC-Bayesian interval estimators. The thesis discusses a number of approaches to obtaining such estimators. The rst type of training sample interval estimator to receive attention is estimators based on classical covering number arguments. A number of these estimators were generalized in various directions. Typical generalizations included: extension of results from misclassi cation loss to other loss functions; extending results to allow arbitrary ghost sample size; extending results to allow arbitrary scale in the relevant covering numbers; and extending results to allow arbitrary choice of in the use of symmetrization lemmas. These extensions were applied to covering number-based estimators for various measures of deviation, as well as for the special cases of misclassi - cation loss estimators, realizable case estimators, and margin bounds. Extended results were also provided for strati cation by (algorithm- and datadependent) complexity of the decision class. In order to facilitate application of these covering number-based bounds,