Dans une période marquée par un bouleversement des systèmes d’information et de santé et de la place de la maladie dans la société, la question du rapport au savoir en santé devient essentielle. La relation soignant-soigné, anciennement vécue sur un mode passif, est aujourd’hui un échange actif de savoirs entre deux individus et deux mondes sociaux, partenaires. Ces constats remettent en cause les rapports soignant-soigné, entre savoir savant et savoir profane, et les modèles de pratiques existants. En alliant les atouts des sciences de l’éducation à celles des modélisations mathématiques, nous avons caractérisé comment les patients atteints d’asthme ou d’hypertension artérielle pulmonaire échangeaient de l’information et des savoirs avec les pharmaciens de ville et hospitaliers. La méthodologie générale faisait appel à la stratégie de triangulation et se divisait en 4 parties : une analyse de la littérature, un remue-méninges, une analyse de contenu de 39 entretiens semi-dirigés, une étude statistique utilisant l’analyse des correspondances simples basée sur un questionnaire diffusé à 124 patients. La nature de l’échange de savoirs (ES) était composée de 3 dimensions interdépendantes que nous avons modélisée : « Modèle 3 C : cure, care et coordination ». L’intensité et la nature de l’ES variaient selon le type de pharmacien impliqué, la maladie, sa durée, son grade de sévérité, l’âge, le niveau d’apprenance, les représentations des patients vis-à-vis des pharmaciens et des médicaments. Le partenariat avec les professionnels de santé, le patient et les aidants, constituait une composante indispensable et facilitatrice de l’ES. L’ES contribuait à l’autoformation des patients pour acquérir des compétences d’autosoins et mieux gérer leur maladie chronique et ses impacts. Nous avons mis en évidence que le pharmacien s’apparentait à un facilitateur de l’autoformation des patients, de l’éducation diffuse et du bricolage des savoirs ; la pharmacie, officinale ou hospitalière constituait l’embryon d’un tiers-lieu. / In this period of change characterized by a disruption of information and health systems, the relationship issue of knowledge becomes more essential. The healthcare professional-patient relationship, formerly based on a passive mode, has become an active exchange of knowledge between two individuals and two social worlds, seen as partners. These observations challenge the healthcare professional -patient relationship, between scholar and lay knowledge, and existing practice models. Combining the strengths of Education Sciences to those of mathematical modeling, this work allows us accurately characterizing how patients with asthma or pulmonary arterial hypertension shared information and knowledge with both community and hospital pharmacists. This study uses a triangulation strategy and combines 4 parts: a literature analysis, a brainstorming, a content analysis of 39 semi-directed interviews and a correspondance analysis based on a questionnaire submitted to 124 patients. The nature of knowledge exchange consisted in 3 interrelated dimensions that we modeled : “3C Model: Cure, Care and Coordination”. The exchange intensity and nature varied with the type of pharmacist involved, the pathology, the severity and disease duration, the patient age, the knowledge level. The patient representations towards pharmacists and medicine also influenced the nature. We identified that the partnership between healthcare professionals, patients and caregivers was a fundamental component and a facilitator of knowledge exchange. We found that the knowledge exchange contributed to the self-training of patient to acquire self-care skills and better manage their chronic disease and its impacts. Finally, this study allowed highlighting (1) the pharmacist was a facilitator of patients selftraining, diffuse education and self-made knowledge; (2) the community and hospital pharmacies were the location where all these take place, as a third place.

Made available in DSpace on 2015-05-07T15:08:48Z (GMT). No. of bitstreams: 1 ArquivoTotal.pdf: 1257067 bytes, checksum: f4b64bb389b89dba21b808b61beed8a9 (MD5) Previous issue date: 2012-08-07 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / This work grew out of my need to understand the processes that give meaning to my academic trajectory. Since graduating, I was provoked by one of my teachers to write about my memories of learning. At the time, it seemed a meaningless activity. But, as time went by, the idea started to get my attention and I sought to understand, if, indeed, there is a distance between theory and practice. During the dissertation orientation sessions, we considered the possibility of contextualizing my experiences of teaching and learning in light of inclusion policies as something challenging and at the same time thought-provoking, because of its innovative nature. Having accepted the challenge, I chose core research questions that I attempt to respond to in this dissertation, such as: how did I become an educator? What are the challenges in the Brazilian educational scenario that characterize the process of inclusion/exclusion? In face of the questions raised, the main goal of this research is to develop an (auto) biographical account of my learning experiences as a visually impaired person who experienced the everyday challenges of inclusion/exclusion at school. The guiding axis of this dissertation departs from my experiences within the family and at school and it is based on a process of listening to the people who have contributed to this process of development and self-development, like my family, my teachers and colleagues who shared learning experiences with me. In short, I want to analyze the teaching and learning processes experienced by me, assessing the contributions of the school curriculum to the training and (self-) training processes through which I have become a pedagogue. Methodologically, I will use life stories derived from oral and written sources (documents). As an operational strategy, I take the (auto) biographical perspective, since my life experiences are considered the trails of my trajectory, and by walking through those trails, I will recollect the contributions of school curriculum to my training and self-training process. / O presente trabalho surgiu da minha necessidade de compreender as marcas que dão sentido a minha trajetória acadêmica. Desde a graduação, fui provocada por uma das minhas professoras para escrever sobre minhas memórias de aprendizagem. Na ocasião, pareceu uma atividade sem sentido. Mas, com o tempo fui amadurecendo a ideia e busquei compreender, se, de fato, há distanciamento entre teoria e prática. Nas sessões de orientação da dissertação cogitamos a possibilidade de contextualizar meu percurso de ensino e aprendizagem no âmbito das políticas de inclusão como algo desafiador e, ao mesmo tempo, instigante devido seu caráter inovador. Desafio aceito elegi questões centrais de pesquisa que tento responder nesta dissertação, entre elas: Como me tornei pedagoga? Quais os desafios no cenário educacional brasileiro que caracterizam o processo de inclusão/exclusão? Diante das questões postas, o objetivo principal dessa pesquisa é (auto) biografar as minhas experiências de aprendizagem como deficiência visual, que viveu no cotidiano escolar os desafios da inclusão/exclusão. O eixo norteador desta dissertação parte de minhas vivências no seio familiar e escolar a partir de um processo de escuta das pessoas que contribuíram para este processo de formação e (auto)formação como a minha família, meus professores e professoras e colegas que compartilhei as experiências de aprendizagem. Em síntese, pretendo analisar os processos de ensino e aprendizagem vivenciados por mim, avaliando as contribuições do currículo escolar no meu processo de formação e (auto) formação que me tornaram uma pedagoga. Metodologicamente, utilizarei as histórias de vida centradas em fontes orais e escritas(documentos). Como estratégia operacional parto do método (auto) biográfico, uma vez que as minhas experiências de vida se constituem as trilhas do meu caminhar e, neste caminhar, vou resgatando as contribuições do currículo escolar no meu processo de formação e (auto)formação.

The creation of a high-quality optical character recognition system (OCR) requires a large amount of labeled data. Obtaining, or in other words creating, such a quantity of labeled data is a costly process. This thesis focuses on several methods which efficiently use unlabeled data for the training of an OCR neural network. The proposed methods fall into the category of self-training algorithms. The general approach of all proposed methods can be summarized as follows. Firstly, the seed model is trained on a limited amount of labeled data. Then, the seed model in combination with the language model is used for producing pseudo-labels for unlabeled data. Machine-labeled data are then combined with the training data used for the creation of the seed model and they are used again for the creation of the target model. The successfulness of individual methods is measured on the handwritten ICFHR 2014 Bentham dataset. Experiments were conducted on two datasets which represented different degrees of labeled data availability. The best model trained on the smaller dataset achieved 3.70 CER [%], which is a relative improvement of 42 % in comparison with the seed model, and the best model trained on the bigger dataset achieved 1.90 CER [%], which is a relative improvement of 26 % in comparison with the seed model. This thesis shows that the proposed methods can be efficiently used to improve the OCR error rate by means of unlabeled data.

Next-generation sequencing has generated enormous amount of DNA and RNA sequences that potentially carry volumes of genetic information, e.g. protein-coding genes. The thesis is divided into three main parts describing i) GeneMarkS-2, ii) GeneMarkS-T, and iii) MetaGeneTack. In prokaryotic genomes, ab initio gene finders can predict genes with high accuracy. However, the error rate is not negligible and largely species-specific. Most errors in gene prediction are made in genes located in genomic regions with atypical GC composition, e.g. genes in pathogenicity islands. We describe a new algorithm GeneMarkS-2 that uses local GC-specific heuristic models for scoring individual ORFs in the first step of analysis. Predicted atypical genes are retained and serve as ‘external’ evidence in subsequent runs of self-training. GeneMarkS-2 also controls the quality of training process by effectively selecting optimal orders of the Markov chain models as well as duration parameters in the hidden semi-Markov model. GeneMarkS-2 has shown significantly improved accuracy compared with other state-of-the-art gene prediction tools. Massive parallel sequencing of RNA transcripts by the next generation technology (RNA-Seq) provides large amount of RNA reads that can be assembled to full transcriptome. We have developed a new tool, GeneMarkS-T, for ab initio identification of protein-coding regions in RNA transcripts. Unsupervised estimation of parameters of the algorithm makes unnecessary several steps in the conventional gene prediction protocols, most importantly the manually curated preparation of training sets. We have demonstrated that the GeneMarkS-T self-training is robust with respect to the presence of errors in assembled transcripts and the accuracy of GeneMarkS-T in identifying protein-coding regions and, particularly, in predicting gene starts compares favorably to other existing methods. Frameshift prediction (FS) is important for analysis and biological interpretation of metagenomic sequences. Reads in metagenomic samples are prone to sequencing errors. Insertion and deletion errors that change the coding frame impair the accurate identification of protein coding genes. Accurate frameshift prediction requires sufficient amount of data to estimate parameters of species-specific statistical models of protein-coding and non-coding regions. However, this data is not available; all we have is metagenomic sequences of unknown origin. The challenge of ab initio FS detection is, therefore, twofold: (i) to find a way to infer necessary model parameters and (ii) to identify positions of frameshifts (if any). We describe a new tool, MetaGeneTack, which uses a heuristic method to estimate parameters of sequence models used in the FS detection algorithm. It was shown on several test sets that the performance of MetaGeneTack FS detection is comparable or better than the one of earlier developed program FragGeneScan.

La compréhension automatique de la parole est au confluent des deux grands domaines que sont la reconnaissance automatique de la parole et l'apprentissage automatique. Un des problèmes majeurs dans ce domaine est l'obtention d'un corpus de données conséquent afin d'obtenir des modèles statistiques performants. Les corpus de parole pour entraîner des modèles de compréhension nécessitent une intervention humaine importante, notamment dans les tâches de transcription et d'annotation sémantique. Leur coût de production est élevé et c'est la raison pour laquelle ils sont disponibles en quantité limitée.Cette thèse vise principalement à réduire ce besoin d'intervention humaine de deux façons : d'une part en réduisant la quantité de corpus annoté nécessaire à l'obtention d'un modèle grâce à des techniques d'apprentissage semi-supervisé (Self-Training, Co-Training et Active-Learning) ; et d'autre part en tirant parti des réponses de l'utilisateur du système pour améliorer le modèle de compréhension.Ce dernier point touche à un second problème rencontré par les systèmes de compréhension automatique de la parole et adressé par cette thèse : le besoin d'adapter régulièrement leurs modèles aux variations de comportement des utilisateurs ou aux modifications de l'offre de services du système

Modelos para aprendizado não supervisionado podem fornecer restrições complementares úteis para melhorar a capacidade de generalização de classificadores. Baseando-se nessa premissa, um algoritmo existente, denominado de C3E (Consensus between Classification and Clustering Ensembles), recebe como entradas estimativas de distribuições de probabilidades de classes para objetos de um conjunto alvo, bem como uma matriz de similaridades entre esses objetos. Tal matriz é tipicamente construída por agregadores de agrupadores de dados, enquanto que as distribuições de probabilidades de classes são obtidas por um agregador de classificadores induzidos por um conjunto de treinamento. Como resultado, o C3E fornece estimativas refinadas das distribuições de probabilidades de classes como uma forma de consenso entre classificadores e agrupadores. A ideia subjacente é de que objetos similares são mais propensos a compartilharem o mesmo rótulo de classe. Nesta tese, uma versão mais simples do algoritmo C3E, baseada em uma função de perda quadrática (C3E-SL), foi investigada em uma abordagem que permitiu a estimação automática (a partir dos dados) de seus parâmetros críticos. Tal abordagem faz uso de um nova estratégia evolutiva concebida especialmente para tornar o C3E-SL mais prático e flexível, abrindo caminho para que variantes do algoritmo pudessem ser desenvolvidas. Em particular, para lidar com a escassez de dados rotulados, um novo algoritmo que realiza aprendizado semissupervisionado foi proposto. Seu mecanismo explora estruturas intrínsecas dos dados a partir do C3E-SL em um procedimento de autotreinamento (self-training). Esta noção também inspirou a concepção de um outro algoritmo baseado em aprendizado ativo (active learning), o qual é capaz de se autoadaptar para aprender novas classes que possam surgir durante a predição de novos dados. Uma extensa análise experimental, focada em problemas do mundo real, mostrou que os algoritmos propostos são bastante úteis e promissores. A combinação de classificadores e agrupadores resultou em modelos de classificação com grande potencial prático e que são menos dependentes do usuário ou do especialista de domínio. Os resultados alcançados foram tipicamente melhores em comparação com os obtidos por classificadores tradicionalmente usados. / Unsupervised learning models can provide a variety of supplementary constraints to improve the generalization capability of classifiers. Based on this assumption, an existing algorithm, named C3E (from Consensus between Classification and Clustering Ensembles), receives as inputs class probability distribution estimates for objects in a target set as well as a similarity matrix. Such a similarity matrix is typically built from clusterers induced on the target set, whereas the class probability distributions are obtained by an ensemble of classifiers induced from a training set. As a result, C3E provides refined estimates of the class probability distributions, from the consensus between classifiers and clusterers. The underlying idea is that similar new objects in the target set are more likely to share the same class label. In this thesis, a simpler version of the C3E algorithm, based on a Squared Loss function (C3E-SL), was investigated from an approach that enables the automatic estimation (from data) of its critical parameters. This approach uses a new evolutionary strategy designed to make C3E-SL more practical and flexible, making room for the development of variants of the algorithm. To address the scarcity of labeled data, a new algorithm that performs semi-supervised learning was proposed. Its mechanism exploits the intrinsic structure of the data by using the C3E-SL algorithm in a self-training procedure. Such a notion inspired the development of another algorithm based on active learning, which is able to self-adapt to learn new classes that may emerge when classifying new data. An extensive experimental analysis, focused on real-world problems, showed that the proposed algorithms are quite useful and promising. The combination of supervised and unsupervised learning yielded classifiers of great practical value and that are less dependent on user-defined parameters. The achieved results were typically better than those obtained by traditional classifiers.

This thesis describes new gene finding methods for eukaryotic gene prediction. The current methods for deriving model parameters for gene prediction algorithms are based on curated or experimentally validated set of genes or gene elements. These training sets often require time and additional expert efforts especially for the species that are in the initial stages of genome sequencing. Unsupervised training allows determination of model parameters from anonymous genomic sequence with. The importance and the practical applicability of the unsupervised training is critical for ever growing rate of eukaryotic genome sequencing. Three distinct training procedures are developed for diverse group of eukaryotic species. GeneMark-ES is developed for species with strong donor and acceptor site signals such as Arabidopsis thaliana, Caenorhabditis elegans and Drosophila melanogaster. The second version of the algorithm, GeneMark-ES-2, introduces enhanced intron model to better describe the gene structure of fungal species with posses with relatively weak donor and acceptor splice sites and well conserved branch point signal. GeneMark-LE, semi-supervised training approach is designed for eukaryotic species with small number of introns. The results indicate that the developed unsupervised training methods perform well as compared to other training methods and as estimated from the set of genes supported by EST-to-genome alignments. Analysis of novel genomes reveals interesting biological findings and show that several candidates of under-annotated and over-annotated fungal species are present in the current set of annotated of fungal genomes.

Orientadora: Profa. Dra. Maria Inês Ribas Rodrigues / Dissertação (mestrado) - Universidade Federal do ABC, Programa De Pós-Graduação em Ensino, História, Filosofia das Ciências e Matemática, 2017. / O crescente uso das Tecnologias da Informação e Comunicação (TIC) no âmbito escolar, exigem mudanças na prática do professor. Para tanto, deve ocorrer uma reflexão sobre a sua prática, para que as potencialidades desses aparatos tecnológicos em relação ao processo de ensino e aprendizagem sejam alcançadas. No contexto da Formação Continuada, têm-se os conhecimentos que os professores devem desenvolver ao longo das suas práxis, dentre eles o conhecimento pedagógico das tecnologias. A respeito disso, tem-se o TPACK (Conhecimento Tecnológico e Pedagógico do Conteúdo), voltado para o conhecimento tecnológico do professor atrelado ao conhecimento pedagógico. Assim, esse trabalho segue a linha de pesquisa sobre a Formação de Professores de Ciências e Matemática, e pretende contribuir com discussões acerca das concepções e práticas pedagógicas desses profissionais no tocante ao uso das TIC no contexto escolar. A investigação foi realizada no segundo semestre de 2016, em duas escolas públicas, localizadas na cidade de São Paulo, envolvendo cinquenta e cinco sujeitos. Com uma metodologia de natureza qualitativa, que teve como método o estudo de caso, a presente pesquisa possui como objetivo analisar a formação continuada com abordagem na inserção das TIC por professores de Ciências dessas escolas, destacando suas opiniões sobre o uso desses recursos na sua prática, e as dificuldades de inseri-las como ferramenta pedagógica no contexto das instituições em que atuam. Além de verificar como os alunos se comportam diante do uso das TIC em sala de aula, e quais são os tipos de atividades, dessa natureza, que eles julgam acrescentar de forma positiva na sua aprendizagem. Para tanto, os procedimentos utilizados com o intuito de alcançar esses objetivos foram: entrevistas individuais com os professores e os coordenadores; intervenções por meio de softwares educacionais, onde tanto a prática do professor quanto o comportamento dos alunos foram observadas; e por meio de questionários, direcionados aos alunos. A análise dos dados foi feita por meio da técnica de Análise Textual Discursiva, que aponta para uma necessidade de os professores usarem as TIC. Entretanto, existe uma deficiência na formação que os auxiliem no uso desses recursos tecnológicos, e que muitos dos cursos de formação continuada não suprem as necessidades oriundas da sala de aula, pois são baseados muito mais na teoria do que na prática. Em relação às dificuldades de utilizar as TIC na concepção dos sujeitos envolvidos, dentre outras, destacam-se: a infraestrutura das escolas atrelada à carga horária dos professores; a receptividade dos alunos; a poucas habilidades por parte dos professores, oriundas da falta de capacitação. Na perspectiva dos alunos, as aulas realizadas com aporte de algum recurso tecnológico ou mesmo alguma prática, possuem uma aceitação maior, por contribuir na aprendizagem de conteúdos de Ciências. Nesse sentido, acredita-se que, os métodos tradicionais de ensino, embora não sejam na sua totalidade insatisfatórios, precisam ser reavaliados, para que a prática sobreponha as aulas extremamente teóricas. / The increasing use of information and communication technologies (ICT) does not require schooling. To do so, develop a reflection on their practice, so that the potentialities say technological apparatuses in relation to the process of teaching and learning and entrepreneurship. In the context of Continuing Education, we have the knowledge that teachers must develop throughout their praxis, among them the pedagogical knowledge of technologies. In this respect, we have the TPACK (Technological and Pedagogical Knowledge of Content), aimed at the technological knowledge of the teacher linked to pedagogical knowledge. Thus, this work follows a line of research on Teacher Training in Science and Mathematics, and send projects of discussions about the conceptions and pedagogical practices of people who are not concerned with the use of ICT in the school context. A survey was conducted without a semester of 2016, in two public schools, located in the city of São Paulo, involving fifty-five subjects. With a methodology of a qualitative nature, which has as its study method, a constant research has as objective to analyze a continuous training with approach in the insertion of ICT by science teachers for schools, highlighting their opinions on the use of resources in their practice, and as insertion difficulties as a pedagogical tool in the context of the institutions in which they operate. In addition to checking how students are, if they are common in the classroom, they are all kinds of activities of that nature, and they are judged positively in their learning. To do this, the procedures used to achieve the objectives: individual interviews with teachers and coordinators; interventions through educational software, where both a teacher's practice and student behavior were observed; and through questionnaires, directed to the students. An analysis of the data was done using the Discursive Textual Analysis technique, which points to a need for teachers to use ICT. There are a number of nonsupreme continuing training courses as needs arising in the classroom because they are based much more on theory than on practice. In relation to the difficulties of use as ICT in the conception of the subjects involved, among others, the following stand out: a school infrastructure linked to the teachers' workload; a receptivity of the students; to a few skills on the part of the teachers, due to lack of training. From the perspective of the students, the classes carried out with the contribution of some technological resource or even some practice, have a greater acceptance, for contributing to the learning of Science contents. In this sense, it is believed that traditional teaching methods, while not altogether unsatisfactory, need to be reevaluated, so that the practice overlaps extremely theoretical classes.

Made available in DSpace on 2014-12-17T14:36:34Z (GMT). No. of bitstreams: 1 IsabelCASRN_TESE_capa_ ate_pag93.pdf: 3436787 bytes, checksum: 6a7f97533ac4a7e79efea2cd5f9d7109 (MD5) Previous issue date: 2013-10-14 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / autoforma??o maternal: cen?rios de uma educa??o vivencial humanescente apresenta a pesquisa realizada com gestantes e seus filhos com idade de at? um ano de vida, da comunidade de Barreta, situada no munic?pio de N?sia Floresta/RN, acompanhados pela equipe de estrat?gia sa?de da fam?lia. A problem?tica se prende ?s fragilidades na concep??o de uma pr?tica educativa humanescente que vise o cuidado integral do bin?mio m?e e filho. Constituindo assim a garantia ao direito da m?e ter uma gesta??o saud?vel e da crian?a de viver uma inf?ncia feliz. Objetivando descrever e interpretar como o desenvolvimento das viv?ncias integrativas de educa??o em sa?de, contribui para o processo de autoforma??o maternal. Partindo deste pensamento de uma pr?tica educativa humanescente, optou-se pelos seguintes pressupostos: a pedagogia vivencial humanescente (CAVALCANTI, 2006) e seus princ?pios. Estes estar?o, durante a pesquisa, relacionados ? teoria da complexidade (MORIN, 2005), ? teoria autopoi?tica (MATURANA; VARELA, 2001), ?s abordagens transdisciplinares (MORAES, LA TORRE, 2008), bem como aos princ?pios do SUS. A pesquisa, de abordagem qualitativa, adota princ?pios da pesquisa-a??o, utilizando as seguintes ferramentas: a observa??o participante existencial; a escuta sens?vel; a fotorreportagem; pr?ticas corporais transdisciplinares e as viv?ncias l?dicas integrativas. O laborat?rio se deu na Unidade B?sica de Sa?de de Barreta. Os encontros foram desenvolvidos atrav?s de experi?ncias vivenciadas, nas quais foram abordadas a autoimagem e autoestima, a modelagem, a natureza e o corpo, a Shantalla e as pr?ticas integrativas complementares em sa?de. As viv?ncias experenciadas permitiram um olhar sens?vel sobre as emo??es e sentimentos, resultando na corporaliza??o da sensibilidade, ludicidade, criatividade e reflexividade, contribuindo no processo de autoforma??o maternal humanescente

Modelos para aprendizado não supervisionado podem fornecer restrições complementares úteis para melhorar a capacidade de generalização de classificadores. Baseando-se nessa premissa, um algoritmo existente, denominado de C3E (Consensus between Classification and Clustering Ensembles), recebe como entradas estimativas de distribuições de probabilidades de classes para objetos de um conjunto alvo, bem como uma matriz de similaridades entre esses objetos. Tal matriz é tipicamente construída por agregadores de agrupadores de dados, enquanto que as distribuições de probabilidades de classes são obtidas por um agregador de classificadores induzidos por um conjunto de treinamento. Como resultado, o C3E fornece estimativas refinadas das distribuições de probabilidades de classes como uma forma de consenso entre classificadores e agrupadores. A ideia subjacente é de que objetos similares são mais propensos a compartilharem o mesmo rótulo de classe. Nesta tese, uma versão mais simples do algoritmo C3E, baseada em uma função de perda quadrática (C3E-SL), foi investigada em uma abordagem que permitiu a estimação automática (a partir dos dados) de seus parâmetros críticos. Tal abordagem faz uso de um nova estratégia evolutiva concebida especialmente para tornar o C3E-SL mais prático e flexível, abrindo caminho para que variantes do algoritmo pudessem ser desenvolvidas. Em particular, para lidar com a escassez de dados rotulados, um novo algoritmo que realiza aprendizado semissupervisionado foi proposto. Seu mecanismo explora estruturas intrínsecas dos dados a partir do C3E-SL em um procedimento de autotreinamento (self-training). Esta noção também inspirou a concepção de um outro algoritmo baseado em aprendizado ativo (active learning), o qual é capaz de se autoadaptar para aprender novas classes que possam surgir durante a predição de novos dados. Uma extensa análise experimental, focada em problemas do mundo real, mostrou que os algoritmos propostos são bastante úteis e promissores. A combinação de classificadores e agrupadores resultou em modelos de classificação com grande potencial prático e que são menos dependentes do usuário ou do especialista de domínio. Os resultados alcançados foram tipicamente melhores em comparação com os obtidos por classificadores tradicionalmente usados. / Unsupervised learning models can provide a variety of supplementary constraints to improve the generalization capability of classifiers. Based on this assumption, an existing algorithm, named C3E (from Consensus between Classification and Clustering Ensembles), receives as inputs class probability distribution estimates for objects in a target set as well as a similarity matrix. Such a similarity matrix is typically built from clusterers induced on the target set, whereas the class probability distributions are obtained by an ensemble of classifiers induced from a training set. As a result, C3E provides refined estimates of the class probability distributions, from the consensus between classifiers and clusterers. The underlying idea is that similar new objects in the target set are more likely to share the same class label. In this thesis, a simpler version of the C3E algorithm, based on a Squared Loss function (C3E-SL), was investigated from an approach that enables the automatic estimation (from data) of its critical parameters. This approach uses a new evolutionary strategy designed to make C3E-SL more practical and flexible, making room for the development of variants of the algorithm. To address the scarcity of labeled data, a new algorithm that performs semi-supervised learning was proposed. Its mechanism exploits the intrinsic structure of the data by using the C3E-SL algorithm in a self-training procedure. Such a notion inspired the development of another algorithm based on active learning, which is able to self-adapt to learn new classes that may emerge when classifying new data. An extensive experimental analysis, focused on real-world problems, showed that the proposed algorithms are quite useful and promising. The combination of supervised and unsupervised learning yielded classifiers of great practical value and that are less dependent on user-defined parameters. The achieved results were typically better than those obtained by traditional classifiers.