Les entreprises et plus généralement les organisations sont confrontées à des enjeux climatiques et économiques avec pour obligation de respecter un cadre légal et des orientations définis à des plus grandes échelles (régionale, nationale et internationale). Une organisation est souvent au fait du but ou de l’objectif à atteindre ; en revanche le moyen d’y parvenir peut nécessiter de l’apprentissage voire de la recherche. Le but de cette thèse est de fournir une méthodologie à l’usage des organisations pour réaliser le management stratégique des projets relatifs à leur transition énergétique. A partir de différents états de l’art sur la planification énergétique et la conception en particulier, nous avons pointé le déficit méthodologique auquel doit faire face une organisation : si les démarches et outils existent lorsqu’un problème est clairement identifié, comment justement identifier un ou des problèmes à partir uniquement d’une formulation de buts ou d’intentions ? La première partie propose une démarche de planification énergétique à l’échelle d’une organisation qui fait émerger, de manière structurée, les problèmes auxquels l’organisation sera potentiellement confrontée. Notre démarche repose sur l’utilisation des BEGES et des méthodes de management de l’énergie/GES d’une part, complétés par des démarches et outils de conception d’autre part. Ces derniers facilitent la consolidation des informations et des données nécessaires pour formuler et structurer les problèmes à résoudre. A l’issue de cette démarche certains problèmes sont formulés sous forme de contradictions et de conflits. La démarche développée est purement qualitative et adaptée au travail de groupe avec des experts. Cependant certaines données numériques traduisent des comportements de systèmes qui sont peu maitrisés par les parties prenantes du projet. La deuxième partie propose une méthode combinant la simulation et l’analyse de données pour identifier les contradictions d’objectifs et de cause qui peuvent ou semblent empêcher l’atteinte des objectifs. Ces contradictions sont formulées de sorte à pouvoir être traitées avec les méthodes de résolution de problèmes inventifs. Le principe d’identification des contradictions d’objectifs repose sur la transformation des réponses expérimentales ou de simulation des systèmes étudiés en données qualitatives binaires et sur l’identification des Paretos optimaux des données ainsi transformées. Les contradictions de causes concernent les facteurs ou paramètres de conception qui induisent les conflits d’objectifs. Nous proposons de les identifier à l’aide d’une méthode d’analyse discriminante binaire à base d’apprentissage supervisé associée à l’ANOVA. Nous montrons sur un cas d’étude, d’une part, comment intégrer cette approche dans la démarche présentée en partie 1 du mémoire, et d’autre part, comment l’utiliser pour obtenir des concepts de solutions dans un contexte multi-objectifs (diminution des consommations d’énergie, des émissions de GES, du coût etc.). / Companies and more generally organizations are confronted with climatic and economic issues, they have to respect a legal framework and orientations defined in larger scales (regional, national and international). An organization usually knows the goal or the objective to be achieved; however the way to do can require learning or even research. The goal of this thesis is to provide a methodology for the use of organizations to realize strategic management of their energy transition projects. From many different states of the art about energy planning and conception in particular, we show the methodological deficit which an organization has to face: if approaches and tools exist when a problem is clearly identified, how actually identify one or several problems from only a goal or intention formulation? The first part proposes an energy planning approach at an organizational scale to bring out in structured way problems which the organization may be confronted. Our approach is based on greenhouse gas emission assessments and energy/GHG management methods which are completed with conception approaches and some tools and methodologies. They facilitate the consolidation of required information and data to formulate and structure problems to solve. As a result of our approach some problems are formulated as contradictions and conflicts. The developed approach is purely qualitative and adapted to workgroup with experts. However some numerical data translate system behaviors which are sparsely mastered by project stakeholders. The second part proposes a combined method of simulation and data analysis to identify objective and cause contradictions which can or seem to prevent achieving the objectives. These contradictions are formulated in such a way to be handled with methods of resolution of inventive problems. The identification of objective contradictions is based on the transformation of experimental or simulation answers of the studied systems in binary qualitative data and on the identification of optimal Pareto of the transformed data. Cause contradictions concern conception factors or parameters which induce objective conflicts. We suggest identifying these contradictions with a binary discriminant analysis method based on supervised learning associated with ANOVA. On one hand, we show on a study case how integrate this initiative into the presented approaches in part 1 and on the other hand, how use it to obtain solution concepts in a multi-objective context (energy consumptions, GHG emissions or cost reduction etc.).

L’objectif de ce travail est de mettre au point une nouvelle approche automatique pour identifier les réseaux de gènes concourant à une même fonction biologique. Ceci permet une meilleure compréhension des phénomènes biologiques et notamment des processus impliqués dans les maladies telles que les cancers. Différentes stratégies ont été développées pour essayer de regrouper les gènes d’un organisme selon leurs relations fonctionnelles : génétique classique et génétique moléculaire. Ici, nous utilisons une propriété connue des réseaux de gènes fonctionnellement liés à savoir que ces gènes sont généralement co-régulés et donc co-exprimés. Cette co-régulation peut être mise en évidence par des méta-analyses de données de puces à ADN (micro-arrays) telles que Gemma ou COXPRESdb. Dans un travail précédent [Al Adhami et al., 2015], la topologie d’un réseau de co-expression de gènes a été caractérisé en utilisant deux paramètres de description des réseaux qui discriminent des groupes de gènes sélectionnés aléatoirement (modules aléatoires, RM) de groupes de gènes avec des liens fonctionnels connus (modules fonctionnels, FM), c’est-à-dire des gènes appartenant au même processus biologique GO. Dans le présent travail, nous avons cherché à généraliser cette approche et à proposer une méthode, appelée TopoFunc, pour améliorer l’annotation existante de la fonction génique. Nous avons d’abord testé différents descripteurs topologiques du réseau de co-expression pour sélectionner ceux qui identifient le mieux des modules fonctionnels. Puis, nous avons constitué une base de données rassemblant des modules fonctionnels et aléatoires, pour lesquels, sur la base des descripteurs sélectionnés, nous avons construit un modèle de discrimination LDA [Friedman et al., 2001] permettant, pour un sous-ensemble de gènes donné, de prédire son type (fonctionnel ou non). Basée sur la méthode de similarité de gènes travaillée par Wang et ses collègues [Wang et al., 2007], nous avons calculé un score de similarité fonctionnelle entre les gènes d’un module. Nous avons combiné ce score avec celui du modèle LDA dans une fonction de fitness implémenté dans un algorithme génétique (GA). À partir du processus biologique d’ontologie de gènes donné (GO-BP), AG visait à éliminer les gènes faiblement co-exprimés avec la plus grande clique de GO-BP et à ajouter des gènes «améliorant» la topologie et la fonctionnalité du module. Nous avons testé TopoFunc sur 193 GO-BP murins comprenant 50-100 gènes et avons montré que TopoFunc avait agrégé un certain nombre de nouveaux gènes avec le GO-BP initial tout en améliorant la topologie des modules et la similarité fonctionnelle. Ces études peuvent être menées sur plusieurs espèces (homme, souris, rat, et possiblement poulet et poisson zèbre) afin d’identifier des modules fonctionnels conservés au cours de l’évolution. / The aim of this work is to develop a new automatic approach to identify networks of genes involved in the same biological function. This allows a better understanding of the biological phenomena and in particular of the processes involved in diseases such as cancers. Various strategies have been developed to try to cluster genes of an organism according to their functional relationships : classical genetics and molecular genetics. Here we use a well-known property of functionally related genes mainly that these genes are generally co-regulated and therefore co-expressed. This co-regulation can be detected by microarray meta-analyzes databases such as Gemma or COXPRESdb. In a previous work [Al Adhami et al., 2015], the topology of a gene coexpression network was characterized using two description parameters of networks that discriminate randomly selected groups of genes (random modules, RM) from groups of genes with known functional relationship (functional modules, FM), e.g. genes that belong to the same GO Biological Process. We first tested different topological descriptors of the co-expression network to select those that best identify functional modules. Then, we built a database of functional and random modules for which, based on the selected descriptors, we constructed a discrimination model (LDA)[Friedman et al., 2001] allowing, for a given subset of genes, predict its type (functional or not). Based on the similarity method of genes worked by Wang and co-workers [Wang et al., 2007], we calculated a functional similarity score between the genes of a module. We combined this score with that of the LDA model in a fitness function implemented in a genetic algorithm (GA). Starting from a given Gene Ontology Biological Process (GO-BP), AG aimed to eliminate genes that were weakly coexpressed with the largest clique of the GO-BP and to add genes that "improved" the topology and functionality of the module. We tested TopoFunc on the 193 murine GO-BPs comprising 50-100 genes and showed that TopoFunc aggregated a number of novel genes to the initial GO-BP while improving module topology and functional similarity. These studies can be conducted on several species (humans, mice, rats, and possibly chicken and zebrafish) to identify functional modules preserved during evolution.

La presente investigación tiene por objetivo, exponer los factores que hacen necesaria la implementación de Business Intelligence con el fin de mejorar la toma de decisiones, que influyen directamente en la eficiencia de los procesos operativos de importación del sector Diagnóstica. Esta investigación se divide en 5 capítulos, en las que se usó el método científico detallado en el libro Metodología de la Investigación, el cual presenta a la investigación como un proceso sistemático, los capítulos están comprendidos por marco teórico, plan de investigación, metodología, desarrollo y aplicación, y análisis de los resultados. Asimismo, se busca identificar que, con esta herramienta informática, se podrá tener la posibilidad de mejorar los procesos operativos de importación. Para finalizar, se comprobará la relación existente entre las dos variables estudiadas Business Intelligence para la mejor toma de decisiones y la eficiencia de los Procesos Operativos, con el fin optimizar la cadena de valor y presentar una herramienta de innovación para el mundo de los negocios internacionales. Para ello, se realizaron análisis cuantitativos y cualitativos, y se empleó la herramienta de SPSS para mostrar el Análisis Factorial y Discriminante. / The objective of this research is to expose the factors that make it necessary to implement Business Intelligence in order to improve decision making, which directly influence the efficiency of the importation processes of the Diagnostic sector. This research is divided into 5 chapters, in which the scientific method detailed in the book Methodology of Research was used, which presents research as a systematic process, the chapters are comprised by theoretical framework, research plan, methodology, development and application, and analysis of results. Likewise, it is sought to identify that, with this computer tool, it will be possible to have the possibility of improving the import operational processes. Finally, the relationship between the two variables studied Business Intelligence for the best decision making and the efficiency of the Operational Processes will be verified, in order to optimize the value chain and present an innovation tool for the world of international business. For this, quantitative and qualitative analyzes were carried out, and the SPSS tool was used to show the Factorial and Discriminant Analysis / Tesis

O uso de dados biométricos da face para verificação automática de identidade é um dos maiores desafios em sistemas de controle de acesso seguro. O processo é extremamente complexo e influenciado por muitos fatores relacionados à forma, posição, iluminação, rotação, translação, disfarce e oclusão de características faciais. Hoje existem muitas técnicas para se reconhecer uma face. Esse trabalho apresenta uma investigação buscando identificar uma face no banco de dados ORL com diferentes grupos de treinamento. É proposto um algoritmo para o reconhecimento de faces baseado na técnica de subespaço LDA (PCA + LDA) utilizando uma rede neural SOM para representar cada classe (face) na etapa de classificação/identificação. Aplicando o método do subespaço LDA busca-se extrair as características mais importantes na identificação das faces previamente conhecidas e presentes no banco de dados, criando um espaço dimensional menor e discriminante com relação ao espaço original. As redes SOM são responsáveis pela memorização das características de cada classe. O algoritmo oferece maior desempenho (taxas de reconhecimento entre 97% e 98%) com relação às adversidades e fontes de erros que prejudicam os métodos de reconhecimento de faces tradicionais. / The use of biometric technique for automatic personal identification is one of the biggest challenges in the security field. The process is complex because it is influenced by many factors related to the form, position, illumination, rotation, translation, disguise and occlusion of face characteristics. Now a days, there are many face recognition techniques. This work presents a methodology for searching a face in the ORL database with some different training sets. The algorithm for face recognition was based on sub-space LDA (PCA + LDA) technique using a SOM neural net to represent each class (face) in the stage of classification/identification. By applying the sub-space LDA method, we extract the most important characteristics in the identification of previously known faces that belong to the database, creating a reduced and more discriminated dimensional space than the original space. The SOM nets are responsible for the memorization of each class characteristic. The algorithm offers great performance (recognition rates between 97% and 98%) considering the adversities and sources of errors inherent to the traditional methods of face recognition.

Por apresentarem mecanismos de transmissão superponíveis, a infecção concomitante pelo vírus da hepatite C (VHC) e pelos vírus linfotrópicos de células T humanas dos tipos 1 (HTLV-1) e 2 (HTLV-2) é esperada. Considerando a relevância dessas infecções em nosso meio e a existência de lacunas no conhecimento da coinfeção VHC/HTLV, conduziu-se este estudo transversal, com o objetivo de comparar uma série de pacientes coinfectados, com indivíduos infectados pelo VHC isoladamente, no tocante a características sócio-demográficas e de exposição aos agentes virais, alterações clínicas e laboratoriais, bem como alterações histológicas do parênquima hepático. Selecionaram-se, com base em algoritmos de diagnóstico sorológico e de biologia molecular, pacientes adultos assistidos em ambulatórios do Hospital das Clínicas da FMUSP entre janeiro de 1993 e agosto de 2005, que apresentaram viremia pelo VHC, associada, ou não, a infecção por HTLV-1 ou HTLV-2, excluindo-se da amostra os coinfectados pelo VHB ou HIV. Coletaram-se dos pacientes selecionados características sócio-demográficas, informações acerca de exposição a vírus de transmissão sexual ou sangüínea, sinais e sintomas clínicos relacionados às infecções causadas pelo VHC ou HTLV, bem como dados laboratoriais hematológicos e de função hepática. Procedeu-se ainda à revisão sistemática dos achados histopatológicos do parênquima hepático, seguindo-se a classificação de Ishak. Compararam-se, então, os grupos VHC, VHC/HTLV-1 e VHC/HTLV-2, empregando-se o teste de X2 para as variáveis categóricas e o teste de Kruskal-Wallis para as variáveis contínuas. Em seguida, pela análise discriminante linear de Fischer, definiram-se funções classificatórias com variáveis que conjuntamente diferenciassem os grupos estudados. Finalmente, a acurácia discriminatória das funções classificatórias foi avaliada por validação cruzada, empregando-se a técnica leave-one-out. Compuseram a população estudada 85 pacientes, sendo 55 no grupo VHC, 24 no grupo VHC/HTLV-1 e 6 no grupo VHC/HTLV-2. À análise bivariada, não se observou diferença significativa entre os grupos no tocante a características sócio-demográficas, hábito de fumar, fatores de exposição às infecções virais, tais como transfusão sangüínea, tatuagem, acupuntura, ou número de parceiros sexuais. Ao contrário, o relato de uso de álcool, drogas endovenosas, ou cocaína inalatória, bem como a parceria sexual com UDEV foi mais freqüente entre os pacientes do grupo VHC/HTLV-2, enquanto o relato de parceiro sexual com hepatite predominou no grupo VHC. Do ponto de vista clínico, apenas a queixa de dor abdominal apresentou-se em freqüência significativamente diferente entre os grupos, sendo mais prevalente no grupo VHC. Em relação aos achados laboratoriais, apesar de contida nos intervalos de normalidade, houve diferença significativa na contagem de plaquetas em sangue periférico, com valores medianos mais elevados nos grupos de coinfectados. As concentrações séricas de aminotransferases e de GGT foram mais altas no grupo VHC. Apesar de freqüentemente encontradas alterações sugestivas de hepatopatia pelo VHC, como fibrose hepática e atividade necroinflamatória, a análise histopatológica não mostrou diferença significativa entre os grupos. À análise discriminante de Fischer, definiram-se funções classificatórias que melhor diferenciam os pacientes estudados, incluindo as variáveis sexo, faixa etária, relato de uso de drogas endovenosas e parceria sexual com indivíduo com hepatite. Por meio de validação cruzada, verificou-se que a acurácia discriminante das funções classificatórias foi alta (87,3%) para a identificação dos infectados pelo VHC isoladamente e intermediária (66,7%) para os coinfectados VHC/HTLV-2. O método não se mostrou, contudo, clinicamente útil na distinção de pacientes com coinfecção VHC/HTLV-1. / Co-infection with hepatitis C virus (HCV) and human T-lymphotropic virus types 1 (HTLV-1) and 2 (HTLV-2) is expected, as these viruses share common infection routes. Due to the relevance of these viral infections in Brazil and the existing gaps in knowledge about HCV/HTLV co-infection, we carried out this cross-sectional survey. A cohort of co-infected patients was compared to HCV-infected subjects, in regard to socio-demographic features, risk factors for viral acquisition, clinical and laboratory data, as well as liver histopathologic findings. Based on established serologic and molecular diagnostic algorithms, we selected HCV-viremic adult patients who attended the Hospital das Clínicas-FMUSP outpatient clinic from January 1993 to August 2005, whether or not they presented co-infection with HTLV-1 or HTLV-2. HBV and HIV-infected individuals were excluded from the sample. We collected patients\' sociodemographic characteristics, risk of exposure to blood-borne or sexually-transmitted viral agents, signs and symptoms related to HCV or HTLV disease, as well as laboratory data that included hematologic counts and liver function tests. Histopathologic findings were systematically reviewed, in accordance to the Ishak\'s scoring system. Patients from the HCV, HCV/HTLV-1 and HCV/HTLV-2 groups, were then compared by means of the X2 or Kruskal-Wallis tests for categorical or continuous variables, respectively. In addition, Fischer\'s linear discriminant analysis was applied to define classification functions that better identified the combined effect of variables important for discrimination of the study groups. Finally, the discriminating accuracy of the model was evaluated by cross-validation, using the leave-one-out technique. The study sample comprised 85 patients, 55 in the HCV group, 24 in the HCV/HTLV-1 group and 6 in the HCV/HTLV-2 group. In bivariable analysis, no significant difference was found among groups in regard to socio-demographic features, smoking, risk factors for viral acquisition, such as blood transfusion, tattooing, acupuncture, or number of sexual partners. In contrast, alcohol consumption, use of intravenous drugs or inhaled cocaine and sexual partnership with an intravenous drug user were more frequent in the HCV/HTLV-2 group, whereas patients in the HCV group more often reported a sexual partner with hepatitis. As far as clinical data are concerned, abdominal pain was the only variable to be reported differently, being more prevalent in the HCV group. Even though within normal ranges, co-infected patients presented higher median platelet counts, whereas aminotransferase and GGT levels were higher among HCV-infected subjects. No significant difference was seen in liver histopathologic findings, though HCV liver disease-associated abnormalities, such as fibrosis and necroinflammatory activity were often found in patients from the three groups. Classification functions, defined by discriminating analysis included as relevant variables sex, age, intravenous drug use and sexual partner with hepatitis. Cross-validation yielded high (87.3%) and intermediate (66,7%) discriminating accuracies for the HCV and HCV/HTLV-2 functions. However, this method was not shown clinically useful to distinguish HCV/HTLV-1 co-infected patients.

Cette thèse s’inscrit dans le cadre du projet "Prendre votre cœur en mains" visant à développer un dispositif médical d’aide à la prescription médicamenteuse pour les insuffisants cardiaques. Dans une première partie, une étude a été menée afin de mettre en évidence la valeur pronostique d’une estimation du volume plasmatique ou de ses variations pour la prédiction des événements cardiovasculaires majeurs à court terme. Deux règles de classification ont été utilisées, la régression logistique et l’analyse discriminante linéaire, chacune précédée d’une phase de sélection pas à pas des variables. Trois indices permettant de mesurer l’amélioration de la capacité de discrimination par ajout du biomarqueur d’intérêt ont été utilisés. Dans une seconde partie, afin d’identifier les patients à risque de décéder ou d’être hospitalisé pour progression de l’insuffisance cardiaque à court terme, un score d’événement a été construit par une méthode d’ensemble, en utilisant deux règles de classification, la régression logistique et l’analyse discriminante linéaire de données mixtes, des échantillons bootstrap et en sélectionnant aléatoirement les prédicteurs. Nous définissons une mesure du risque d’événement par un odds-ratio et une mesure de l’importance des variables et des groupes de variables. Nous montrons une propriété de l’analyse discriminante linéaire de données mixtes. Cette méthode peut être mise en œuvre dans le cadre de l’apprentissage en ligne, en utilisant des algorithmes de gradient stochastique pour mettre à jour en ligne les prédicteurs. Nous traitons le problème de la régression linéaire multidimensionnelle séquentielle, en particulier dans le cas d’un flux de données, en utilisant un processus d’approximation stochastique. Pour éviter le phénomène d’explosion numérique et réduire le temps de calcul pour prendre en compte un maximum de données entrantes, nous proposons d’utiliser un processus avec des données standardisées en ligne au lieu des données brutes et d’utiliser plusieurs observations à chaque étape ou toutes les observations jusqu’à l’étape courante sans avoir à les stocker. Nous définissons trois processus et en étudions la convergence presque sûre, un avec un pas variable, un processus moyennisé avec un pas constant, un processus avec un pas constant ou variable et l’utilisation de toutes les observations jusqu’à l’étape courante. Ces processus sont comparés à des processus classiques sur 11 jeux de données. Le troisième processus à pas constant est celui qui donne généralement les meilleurs résultats / This thesis is part of the "Handle your heart" project aimed at developing a drug prescription assistance device for heart failure patients. In a first part, a study was conducted to highlight the prognostic value of an estimation of plasma volume or its variations for predicting major short-term cardiovascular events. Two classification rules were used, logistic regression and linear discriminant analysis, each preceded by a stepwise variable selection. Three indices to measure the improvement in discrimination ability by adding the biomarker of interest were used. In a second part, in order to identify patients at short-term risk of dying or being hospitalized for progression of heart failure, a short-term event risk score was constructed by an ensemble method, two classification rules, logistic regression and linear discriminant analysis of mixed data, bootstrap samples, and by randomly selecting predictors. We define an event risk measure by an odds-ratio and a measure of the importance of variables and groups of variables using standardized coefficients. We show a property of linear discriminant analysis of mixed data. This methodology for constructing a risk score can be implemented as part of online learning, using stochastic gradient algorithms to update online the predictors. We address the problem of sequential multidimensional linear regression, particularly in the case of a data stream, using a stochastic approximation process. To avoid the phenomenon of numerical explosion which can be encountered and to reduce the computing time in order to take into account a maximum of arriving data, we propose to use a process with online standardized data instead of raw data and to use of several observations per step or all observations until the current step. We define three processes and study their almost sure convergence, one with a variable step-size, an averaged process with a constant step-size, a process with a constant or variable step-size and the use of all observations until the current step without storing them. These processes are compared to classical processes on 11 datasets. The third defined process with constant step-size typically yields the best results

Por apresentarem mecanismos de transmissão superponíveis, a infecção concomitante pelo vírus da hepatite C (VHC) e pelos vírus linfotrópicos de células T humanas dos tipos 1 (HTLV-1) e 2 (HTLV-2) é esperada. Considerando a relevância dessas infecções em nosso meio e a existência de lacunas no conhecimento da coinfeção VHC/HTLV, conduziu-se este estudo transversal, com o objetivo de comparar uma série de pacientes coinfectados, com indivíduos infectados pelo VHC isoladamente, no tocante a características sócio-demográficas e de exposição aos agentes virais, alterações clínicas e laboratoriais, bem como alterações histológicas do parênquima hepático. Selecionaram-se, com base em algoritmos de diagnóstico sorológico e de biologia molecular, pacientes adultos assistidos em ambulatórios do Hospital das Clínicas da FMUSP entre janeiro de 1993 e agosto de 2005, que apresentaram viremia pelo VHC, associada, ou não, a infecção por HTLV-1 ou HTLV-2, excluindo-se da amostra os coinfectados pelo VHB ou HIV. Coletaram-se dos pacientes selecionados características sócio-demográficas, informações acerca de exposição a vírus de transmissão sexual ou sangüínea, sinais e sintomas clínicos relacionados às infecções causadas pelo VHC ou HTLV, bem como dados laboratoriais hematológicos e de função hepática. Procedeu-se ainda à revisão sistemática dos achados histopatológicos do parênquima hepático, seguindo-se a classificação de Ishak. Compararam-se, então, os grupos VHC, VHC/HTLV-1 e VHC/HTLV-2, empregando-se o teste de X2 para as variáveis categóricas e o teste de Kruskal-Wallis para as variáveis contínuas. Em seguida, pela análise discriminante linear de Fischer, definiram-se funções classificatórias com variáveis que conjuntamente diferenciassem os grupos estudados. Finalmente, a acurácia discriminatória das funções classificatórias foi avaliada por validação cruzada, empregando-se a técnica leave-one-out. Compuseram a população estudada 85 pacientes, sendo 55 no grupo VHC, 24 no grupo VHC/HTLV-1 e 6 no grupo VHC/HTLV-2. À análise bivariada, não se observou diferença significativa entre os grupos no tocante a características sócio-demográficas, hábito de fumar, fatores de exposição às infecções virais, tais como transfusão sangüínea, tatuagem, acupuntura, ou número de parceiros sexuais. Ao contrário, o relato de uso de álcool, drogas endovenosas, ou cocaína inalatória, bem como a parceria sexual com UDEV foi mais freqüente entre os pacientes do grupo VHC/HTLV-2, enquanto o relato de parceiro sexual com hepatite predominou no grupo VHC. Do ponto de vista clínico, apenas a queixa de dor abdominal apresentou-se em freqüência significativamente diferente entre os grupos, sendo mais prevalente no grupo VHC. Em relação aos achados laboratoriais, apesar de contida nos intervalos de normalidade, houve diferença significativa na contagem de plaquetas em sangue periférico, com valores medianos mais elevados nos grupos de coinfectados. As concentrações séricas de aminotransferases e de GGT foram mais altas no grupo VHC. Apesar de freqüentemente encontradas alterações sugestivas de hepatopatia pelo VHC, como fibrose hepática e atividade necroinflamatória, a análise histopatológica não mostrou diferença significativa entre os grupos. À análise discriminante de Fischer, definiram-se funções classificatórias que melhor diferenciam os pacientes estudados, incluindo as variáveis sexo, faixa etária, relato de uso de drogas endovenosas e parceria sexual com indivíduo com hepatite. Por meio de validação cruzada, verificou-se que a acurácia discriminante das funções classificatórias foi alta (87,3%) para a identificação dos infectados pelo VHC isoladamente e intermediária (66,7%) para os coinfectados VHC/HTLV-2. O método não se mostrou, contudo, clinicamente útil na distinção de pacientes com coinfecção VHC/HTLV-1. / Co-infection with hepatitis C virus (HCV) and human T-lymphotropic virus types 1 (HTLV-1) and 2 (HTLV-2) is expected, as these viruses share common infection routes. Due to the relevance of these viral infections in Brazil and the existing gaps in knowledge about HCV/HTLV co-infection, we carried out this cross-sectional survey. A cohort of co-infected patients was compared to HCV-infected subjects, in regard to socio-demographic features, risk factors for viral acquisition, clinical and laboratory data, as well as liver histopathologic findings. Based on established serologic and molecular diagnostic algorithms, we selected HCV-viremic adult patients who attended the Hospital das Clínicas-FMUSP outpatient clinic from January 1993 to August 2005, whether or not they presented co-infection with HTLV-1 or HTLV-2. HBV and HIV-infected individuals were excluded from the sample. We collected patients\' sociodemographic characteristics, risk of exposure to blood-borne or sexually-transmitted viral agents, signs and symptoms related to HCV or HTLV disease, as well as laboratory data that included hematologic counts and liver function tests. Histopathologic findings were systematically reviewed, in accordance to the Ishak\'s scoring system. Patients from the HCV, HCV/HTLV-1 and HCV/HTLV-2 groups, were then compared by means of the X2 or Kruskal-Wallis tests for categorical or continuous variables, respectively. In addition, Fischer\'s linear discriminant analysis was applied to define classification functions that better identified the combined effect of variables important for discrimination of the study groups. Finally, the discriminating accuracy of the model was evaluated by cross-validation, using the leave-one-out technique. The study sample comprised 85 patients, 55 in the HCV group, 24 in the HCV/HTLV-1 group and 6 in the HCV/HTLV-2 group. In bivariable analysis, no significant difference was found among groups in regard to socio-demographic features, smoking, risk factors for viral acquisition, such as blood transfusion, tattooing, acupuncture, or number of sexual partners. In contrast, alcohol consumption, use of intravenous drugs or inhaled cocaine and sexual partnership with an intravenous drug user were more frequent in the HCV/HTLV-2 group, whereas patients in the HCV group more often reported a sexual partner with hepatitis. As far as clinical data are concerned, abdominal pain was the only variable to be reported differently, being more prevalent in the HCV group. Even though within normal ranges, co-infected patients presented higher median platelet counts, whereas aminotransferase and GGT levels were higher among HCV-infected subjects. No significant difference was seen in liver histopathologic findings, though HCV liver disease-associated abnormalities, such as fibrosis and necroinflammatory activity were often found in patients from the three groups. Classification functions, defined by discriminating analysis included as relevant variables sex, age, intravenous drug use and sexual partner with hepatitis. Cross-validation yielded high (87.3%) and intermediate (66,7%) discriminating accuracies for the HCV and HCV/HTLV-2 functions. However, this method was not shown clinically useful to distinguish HCV/HTLV-1 co-infected patients.

O uso de dados biométricos da face para verificação automática de identidade é um dos maiores desafios em sistemas de controle de acesso seguro. O processo é extremamente complexo e influenciado por muitos fatores relacionados à forma, posição, iluminação, rotação, translação, disfarce e oclusão de características faciais. Hoje existem muitas técnicas para se reconhecer uma face. Esse trabalho apresenta uma investigação buscando identificar uma face no banco de dados ORL com diferentes grupos de treinamento. É proposto um algoritmo para o reconhecimento de faces baseado na técnica de subespaço LDA (PCA + LDA) utilizando uma rede neural SOM para representar cada classe (face) na etapa de classificação/identificação. Aplicando o método do subespaço LDA busca-se extrair as características mais importantes na identificação das faces previamente conhecidas e presentes no banco de dados, criando um espaço dimensional menor e discriminante com relação ao espaço original. As redes SOM são responsáveis pela memorização das características de cada classe. O algoritmo oferece maior desempenho (taxas de reconhecimento entre 97% e 98%) com relação às adversidades e fontes de erros que prejudicam os métodos de reconhecimento de faces tradicionais. / The use of biometric technique for automatic personal identification is one of the biggest challenges in the security field. The process is complex because it is influenced by many factors related to the form, position, illumination, rotation, translation, disguise and occlusion of face characteristics. Now a days, there are many face recognition techniques. This work presents a methodology for searching a face in the ORL database with some different training sets. The algorithm for face recognition was based on sub-space LDA (PCA + LDA) technique using a SOM neural net to represent each class (face) in the stage of classification/identification. By applying the sub-space LDA method, we extract the most important characteristics in the identification of previously known faces that belong to the database, creating a reduced and more discriminated dimensional space than the original space. The SOM nets are responsible for the memorization of each class characteristic. The algorithm offers great performance (recognition rates between 97% and 98%) considering the adversities and sources of errors inherent to the traditional methods of face recognition.

Cette thèse s’inscrit dans le contexte de l’analyse et de la classification de signaux électroencéphalogrammes (EEG) par des méthodes d’analyse discriminante. Ces signaux multi-capteurs qui sont, par nature, très fortement corrélés spatialement et temporellement sont considérés dans le plan temps-fréquence. En particulier, nous nous intéressons à des signaux de type potentiels évoqués qui sont bien représentés dans l’espace des ondelettes. Par la suite, nous considérons donc les signaux représentés par des coefficients multi-échelles et qui ont une structure matricielle électrodes × coefficients. Les signaux EEG sont considérés comme un mélange entre l’activité d’intérêt que l’on souhaite extraire et l’activité spontanée (ou "bruit de fond"), qui est largement prépondérante. La problématique principale est ici de distinguer des signaux issus de différentes conditions expérimentales (classes). Dans le cas binaire, nous nous focalisons sur l’approche probabiliste de l’analyse discriminante et des modèles de mélange gaussien sont considérés, décrivant dans chaque classe les signaux en termes de composantes fixes (moyenne) et aléatoires. Cette dernière, caractérisée par sa matrice de covariance, permet de modéliser différentes sources de variabilité. Essentielle à la mise en oeuvre de l’analyse discriminante, l’estimation de cette matrice (et de son inverse) peut être dégradée dans le cas de grandes dimensions et/ou de faibles échantillons d’apprentissage, cadre applicatif de cette thèse. Nous nous intéressons aux alternatives qui se basent sur la définition de modèle(s) de covariance(s) particulier(s) et qui permettent de réduire le nombre de paramètres à estimer. / The present thesis finds itself within the framework of analyzing and classifying electroencephalogram signals (EEG) using discriminant analysis. Those multi-sensor signals which are, by nature, highly correlated spatially and temporally are considered, in this work, in the timefrequency domain. In particular, we focus on low-frequency evoked-related potential-type signals (ERPs) that are well described in the wavelet domain. Thereafter, we will consider signals represented by multi-scale coefficients and that have a matrix structure electrodes × coefficients. Moreover, EEG signals are seen as a mixture between the signal of interest that we want to extract and spontaneous activity (also called "background noise") which is overriding. The main problematic is here to distinguish signals from different experimental conditions (class). In the binary case, we focus on the probabilistic approach of the discriminant analysis and Gaussian mixtures are used, describing in each class the signals in terms of fixed (mean) and random components. The latter, characterized by its covariance matrix, allow to model different variability sources. The estimation of this matrix (and of its inverse) is essential for the implementation of the discriminant analysis and can be deteriorated by high-dimensional data and/or by small learning samples, which is the application framework of this thesis. We are interested in alternatives that are based on specific covariance model(s) and that allow to decrease the number of parameters to estimate.

La récente crise financière qui a débuté aux États-Unis en 2007 a révélé les faiblesses du système bancaire international se traduisant par l’effondrement de nombreuses institutions financières aux États-Unis et aussi par l’augmentation de la part des prêts non performants dans les bilans des banques européennes. Dans ce cadre, nous proposons d’abord d’estimer et de tester l’efficacité des modèles de prévisions des défaillances bancaires. L’objectif étant d’établir un système d’alerte précoce (EWS) de difficultés bancaires basées sur des variables financières selon la typologie CAMEL (Capital adequacy, Asset quality, Management quality, Earnings ability, Liquidity). Dans la première étude, nous avons comparé la classification et la prédiction de l’analyse discriminante canonique (CDA) et de la régression logistique (LR) avec et sans coûts de classification en combinant ces deux modèles paramétriques avec le modèle descriptif d’analyse en composantes principales (ACP). Les résultats montrent que les modèles (LR et CDA) peuvent prédire la faillite des banques avec précision. De plus, les résultats de l’ACP montrent l’importance de la qualité des actifs, de l’adéquation des fonds propres et de la liquidité en tant qu’indicateurs des conditions financières de la banque. Nous avons aussi comparé la performance de deux méthodes non paramétriques, les arbres de classification et de régression (CART) et le nouveau modèle régression multivariée par spline adaptative (MARS), dans la prévision de la défaillance. Un modèle hybride associant ’K-means clustering’ et MARS est également testé. Nous cherchons à modéliser la relation entre dix variables financières et le défaut d’une banque américaine. L’approche comparative a mis en évidence la suprématie du modèle hybride en termes de classification. De plus, les résultats ont montré que les variables d’adéquation du capital sont les plus importantes pour la prévision de la faillite d’une banque. Enfin, nous avons étudié les facteurs déterminants des prêts non performants des banques de l’Union Européenne durant la période 2012-2015 en estimant un modèle à effets fixe sur données de panel. Selon la disponibilité des données nous avons choisi un ensemble de variables qui se réfèrent à la situation macroéconomique du pays de la banque et d’autres variables propres à chaque banque. Les résultats ont prouvé que la dette publique, les provisions pour pertes sur prêts, la marge nette d’intérêt et la rentabilité des capitaux propres affectent positivement les prêts non performants, par contre la taille de la banque et l’adéquation du capital (EQTA et CAR) ont un impact négatif sur les créances douteuses. / The recent financial crisis that began in the United States in 2007 revealed the weaknesses of the international banking system resulting in the collapse of many financial institutions in the United States and also the increase in the share of non-performing loans in the balance sheets of European banks. In this framework, we first propose to estimate and test the effectiveness of banking default forecasting models. The objective is to establish an early warning system (EWS) of banking difficulties based on financial variables according to CAMEL’s ratios (Capital adequacy, Asset quality, Management quality, Earnings ability, Liquidity). In the first study, we compared the classification and the prediction of the canonical discriminant analysis (CDA) and the logistic regression (LR) with and without classification costs by combining these two parametric models with the descriptive model of principal components analysis (PCA). The results show that the LR and the CDA can predict bank failure accurately. In addition, the results of the PCA show the importance of asset quality, capital adequacy and liquidity as indicators of the bank’s financial conditions. We also compared the performance of two non-parametric methods, the classification and regression trees (CART) and the newly multivariate adaptive regression splines (MARS) models, in the prediction of failure. A hybrid model combining ’K-means clustering’ and MARS is also tested. We seek to model the relationship between ten financial variables (CAMEL’s ratios) and the default of a US bank. The comparative approach has highlighted the supremacy of the hybrid model in terms of classification. In addition, the results showed that the capital adequacy variables are the most important for predicting the bankruptcy of a bank. Finally, we studied the determinants of non-performing loans from European Union banks during the period 2012-2015 by estimating a fixed effects model on panel data. Depending on the availability of data we have chosen a set of variables that refer to the macroeconomic situation of the country of the bank and other variables specific to each bank. The results showed that public debt, loan loss provisions, net interest margin and return on equity positively affect non performing loans, while the size of the bank and the adequacy of capital (EQTA and CAR) have a negative impact on bad debts.