Knowledge and knowledge work are key factors of today’s successful companies. This study devises an approach for increasing the performance of knowledge work by shifting it towards a process orientation. Business process management and workflow management are methods for structured and predefined work but are not flexible enough to support knowledge work in a comprehensive way. Case-based reasoning (CBR) uses the knowledge of previously experienced cases in order to propose a solution to a problem. CBR can be used to retrieve, reuse, revise, retain and store functional and process knowledge. The aim of the research was to develop an approach that combines CBR and process execution to improve knowledge work. The research goals are: a casedescription for knowledge work that can be integrated into a process execution system and that contains both functional and process knowledge; a similarity algorithm for the retrieval of functional and procedural knowledge; and an adaptation mechanism that deals with the different granularities of solution parts. This thesis contains a profound literature framework and follows a design science research (DSR) strategy. During the awareness phase of the design science research process, an application scenario was acquired using the case study research method, which is the admission process for a study programme at a university. This application scenario is used to introduce and showcase the combined CBR and process execution approach called ICEBERG-PE, which consists of a case model and CBR services. The approach is implemented as a prototype and can be instantiated using the ICEBERG-PE procedure model, a specific procedure model for ontology-based, CBR projects. The ICEBERG-PE prototype has been evaluated using triangulated evaluation data and different evaluation settings to confirm that the approach is transferable to other contexts. Finally, this thesis concludes with potential recommendations for future research. / Computing / D. Phil. (Information Systems)

Made available in DSpace on 2016-06-02T20:00:37Z (GMT). No. of bitstreams: 1 3691.pdf: 6551508 bytes, checksum: ba7667ef8adc647d17a16ee824f5c9ca (MD5) Previous issue date: 2011-06-30 / Financiadora de Estudos e Projetos / Considering the significant and growing environmental degradation and the quality of life in cities, as well as the demand for systematization of information applicable to their control and recovery, this study aimed to develop a decision support system for management of public green spaces in areas of permanent preservation of urban water bodies. For its development were adopted as reference the guidelines of territorial scenario planning method through which main aspects were analyzed related to the problem and systematized actions concerning to the study and solution proposals. The main materials used correspond to the works of academic and technical literature and applied standardization, as well as the related court cases, raised through search terms in library collections and databases of legislation and jurisprudence. From the analysis of these materials, were studied the dynamics of the urban environment and degradation causes of its water bodies; discussed the effective functions of their permanent preservation areas and the intervening factors to their performance; as well as analyzed legal and related court aspects. As a result, were obtained a proposal of a structured key decision model for the evaluation and selection of locational alternatives and their functional-space aptitude for implementation of public green spaces, as a solution established by CONAMA Resolution n. 369 of 2006, which regulates the intervention exceptional cases in permanent preservation areas. Finally, the conclusion was that the achieved results can contribute as an important reference for this issue approach, however, further studies on the refinement of the proposed system are recommended, aiming at optimizations that improve its application and performance. / Considerando a expressiva e crescente degradação ambiental e da qualidade de vida da população nas cidades, assim como, a demanda pela sistematização de informações aplicáveis ao seu controle e recuperação, esta dissertação teve como objetivo o desenvolvimento de um sistema de suporte à decisão para gestão de áreas verdes de domínio público em áreas de preservação permanente de corpos hídricos urbanos. Para o seu desenvolvimento foram adotadas como referências as diretrizes do método de planejamento por cenários territoriais, mediante o qual foram analisados os principais aspectos correlatos ao problema e sistematizadas as ações concernentes ao estudo e proposição de soluções. Os principais materiais utilizados corresponderam às obras da literatura técnico-acadêmica e da normatização aplicada, bem como aos casos judiciais correlatos, levantados através de termos de busca junto aos acervos bibliográficos e bancos de dados legislativos e jurisprudências. A partir da análise destes materiais, foram estudadas a dinâmica do meio urbano e as causas de degradação dos seus corpos hídricos; discutidas as efetivas funções das suas áreas de preservação permanente e os fatores intervenientes ao seu desempenho; assim como, analisados os aspectos legais e judiciais relacionados. Como resultado, obteve-se a proposta de um modelo estruturado em chaves de decisão para avaliação e seleção de alternativas locacionais e de sua aptidão espaço-funcional para implantação de áreas verdes públicas, como solução instituída pela Resolução Conama n. 369 de 2006, que regulamenta os casos excepcionais de intervenção em áreas de preservação permanente. Por fim, conclui-se que os resultados alcançados podem contribuir como uma importante referência para o tratamento da questão, contudo, recomenda-se que estudos complementares avancem no refinamento do sistema proposto, visando otimizações que aprimorem sua aplicação e desempenho.

Artificiell Intelligens (AI) har flera användningsområden som bland annat robotik, ansiktsigenkänning och stöd vid beslutsfattande. Organisationer kommer använda AI mer för att möta utmaningar inom Human Resources (HR) de närmaste fem åren vilket pekar på att AI sannolikt kommer bli en vanligare förekomst inom rekryteringsprocessen. En av de viktigaste tillgångarna i ett företag är dess anställda och felaktiga rekryteringar kan komma att medföra stora kostnader. Med maskininlärning och AI-system som beslutsfattare kan det vara av vikt att fundera på vad det är för data som dessa system förses med då en av riskerna med maskininlärning inom AI är att man inte vet vad maskinerna lär sig när den lär sig själv. En större datamängd behöver i sig inte medföra mer subjektiva resultat men risken att direkt koda in diskriminering finns fortfarande eftersom den data AI-system förses med i sig kan innehålla bias. Det har även visat sig att kandidater inte vill bli bedömda på politiska åsikter, relationer eller annat som kan tas fram via big data och data mining. Syftet med studien är att skapa en djupare förståelse för vad som behövs för att automatisera rekryteringsprocessen med hjälp av AI och maskininlärning samt till att utforma en lista på hur AI kan vara ett stöd för företag att ha i åtanke vid en möjlig implementering. Till studien har tre metoder till empiriinsamling valts ut varav samtliga med en kvalitativ ansats. Intervjuer och en enkät har samlat in de data som analyserats i Excel 2016 samt Google Docs. Intervjuerna utfördes i flera skeden och riktade sig mot två anställda på varsitt rekryteringsföretag. Enkäten riktade sig främst till individer som kommer att ta/har tagit examen inom det närmaste året. Urvalet har skett enligt studiens syfte och vid enstaka tillfällen har ett bekvämlighetsurval gjorts. Resultatet visar att rekryterarna lägger mycket tid på att screena kandidater och gör det manuellt. Enkäten visar att kommande kandidater främst är neutrala i sin tillit till att screening utförs av ett AI-system. Respondenten i uppföljningsintervjun säger att en automatisering med AI hade underlättat arbetet och håller med det enkätrespondenterna anser om fördelar och nackdelar med AI men skulle samtidigt inte lita på resultatet. Vidare tror respondenten att det är den automatiserade vägen rekryteringsprocessen kommer att gå. Resultatet av studien kan komma att nyttjas av rekryteringsföretag som funderar på att införa AI i sina rekryteringsprocesser. / Artificial Intelligence (AI) has several areas of use such as robotics, facial recognition and decision-making support. Organizations will use AI more to meet challenges within Human Resources (HR) over the next five years, indicating that AI is likely to become a more common occurrence in the recruitment process. One of the most important assets of a company is its employees and incorrect recruitments can lead to high costs. With machine learning and AI systems as decision makers it may be important to think about what data is provided to these systems, since one of the risks of machine learning within AI is that you do not know what the machines learn as they learn themselves. A larger amount of data does not necessarily lead to more subjective results, but the risk of directly encode discrimination still exists because of the data the AI system is provided with can contain bias. It has also been found that candidates do not want to be judged on political views, relationships or anything that can be gained through big data and data mining. The purpose of the study is to provide a deeper understanding of what is needed to automate the recruitment process using AI and machine learning and to design a list of how AI can be a support for companies to keep in mind during a possible implementation. The study has chosen three methods for empirical gathering, all of which are qualitative. Interviews and a survey has collected the data which is analyzed in Excel 2016 as well as Google Docs. The interviews were conducted in several stages and aimed towards two employees working at two different recruiting companies. The survey was aimed primarily towards individuals who will have graduated within this year. The selection of participants has been made for the purpose of the study and on some occasions a comfort selection has been made. The result shows that the recruiters spend a lot of time screening candidates and do this manually. The survey shows that future candidates have a neutral stance when it comes to trusting in an AI system performing the screening process. The respondent in the follow-up interview says that automation using AI would facilitate the work and agrees with the survey respondents considering the pros and cons of AI, but at the same time would not rely on the results. Further, the respondent believes that it is in the automated way the recruitment process will continue. The result of the study may be used by recruitment companies that are considering introducing AI into their recruitment processes.

Made available in DSpace on 2014-12-17T14:54:59Z (GMT). No. of bitstreams: 1 CeciliaRML_TESE.pdf: 2405004 bytes, checksum: b79db46caa488d92a5933e45e80ca647 (MD5) Previous issue date: 2011-06-10 / The area of the hospital automation has been the subject a lot of research, addressing relevant issues which can be automated, such as: management and control (electronic medical records, scheduling appointments, hospitalization, among others); communication (tracking patients, staff and materials), development of medical, hospital and laboratory equipment; monitoring (patients, staff and materials); and aid to medical diagnosis (according to each speciality). This thesis presents an architecture for a patient monitoring and alert systems. This architecture is based on intelligent systems techniques and is applied in hospital automation, specifically in the Intensive Care Unit (ICU) for the patient monitoring in hospital environment. The main goal of this architecture is to transform the multiparameter monitor data into useful information, through the knowledge of specialists and normal parameters of vital signs based on fuzzy logic that allows to extract information about the clinical condition of ICU patients and give a pre-diagnosis. Finally, alerts are dispatched to medical professionals in case any abnormality is found during monitoring. After the validation of the architecture, the fuzzy logic inferences were applied to the trainning and validation of an Artificial Neural Network for classification of the cases that were validated a priori with the fuzzy system / A ?rea da automa??o hospitalar tem sido alvo de muitas pesquisas, abordando problemas pertinentes que podem ser automatizados, como: gerenciamento e controle (prontu?rio eletr?nico, marca??o de consulta, internamento, entre outros); comunica??o (rastreamento de pacientes, materiais e funcion?rios); desenvolvimento de equipamentos m?dicos, hospitalares e laboratoriais; monitoramento (pacientes, materiais e funcion?rios); e aux?lio ao diagn?stico m?dico (de acordo com cada especialidade). Esta tese de doutorado apresenta uma Arquitetura de um Sistema Inteligente de Monitoramento e Envio de Alertas de Pacientes (SIMAp). A arquitetura est? baseada em t?cnicas de sistemas inteligentes e aplicada na automa??o hospitalar, mais especificamente em Unidade de Terapia Intensiva (UTI) para monitoramento de pacientes. O objetivo do SIMAp ? a transforma??o dos dados do monitor multiparam?trico em informa??es, por meio do conhecimento dos especialistas e dos par?metros de normalidade dos sinais vitais de pacientes, utilizando l?gica fuzzy na extra??o das informa??es a respeito do quadro cl?nico de pacientes internados em UTI. Por fim, alertas s?o gerados e podem ser enviados para a equipe m?dica, caso seja encontrada alguma anormalidade no monitoramento. Ap?s a valida??o da arquitetura, as infer?ncias oriundas do modelo fuzzy foram aplicadas no treinamento e valida??o de uma RNA para a classifica??o das situa??es previstas no modelo, resultando no pr?-diagn?sticos

Made available in DSpace on 2014-12-17T14:55:04Z (GMT). No. of bitstreams: 1 GlauciaRMAS_TESE.pdf: 2163942 bytes, checksum: 5778dd8818ffc286b87137c2a56b9fc0 (MD5) Previous issue date: 2012-05-14 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / Breast cancer, despite being one of the leading causes of death among women worldwide is a disease that can be cured if diagnosed early. One of the main techniques used in the detection of breast cancer is the Fine Needle Aspirate FNA (aspiration puncture by thin needle) which, depending on the clinical case, requires the analysis of several medical specialists for the diagnosis development. However, such diagnosis and second opinions have been hampered by geographical dispersion of physicians and/or the difficulty in reconciling time to undertake work together. Within this reality, this PhD thesis uses computational intelligence in medical decision-making support for remote diagnosis. For that purpose, it presents a fuzzy method to assist the diagnosis of breast cancer, able to process and sort data extracted from breast tissue obtained by FNA. This method is integrated into a virtual environment for collaborative remote diagnosis, whose model was developed providing for the incorporation of prerequisite Modules for Pre Diagnosis to support medical decision. On the fuzzy Method Development, the process of knowledge acquisition was carried out by extraction and analysis of numerical data in gold standard data base and by interviews and discussions with medical experts. The method has been tested and validated with real cases and, according to the sensitivity and specificity achieved (correct diagnosis of tumors, malignant and benign respectively), the results obtained were satisfactory, considering the opinions of doctors and the quality standards for diagnosis of breast cancer and comparing them with other studies involving breast cancer diagnosis by FNA. / O c?ncer de mama, apesar de ser uma das principais causas de morte entre as mulheres em todo o mundo, ? uma doen?a que pode ser curada se for diagnosticada precocemente. Uma das principais t?cnicas utilizadas na detec??o de c?ncer de mama ? a Fine Needle Aspirate FNA (ou Pun??o Aspirativa por Agulha Fina) que, dependendo do caso cl?nico, necessita da an?lise de v?rios m?dicos especialistas para a efetiva??o do diagn?stico. Entretanto, a realiza??o de tais diagn?sticos e a emiss?o de segundos pareceres t?m sido prejudicadas pela dispers?o geogr?fica dos m?dicos e/ou a dificuldade na concilia??o de tempo para realizar trabalhos em conjunto. Inserindo-se nessa realidade, esta tese de doutorado utiliza intelig?ncia computacional no apoio ? tomada de decis?o m?dica para a realiza??o de telediagn?sticos. Para tanto apresenta um m?todo fuzzy destinado a auxiliar o diagn?stico de c?ncer de mama, capaz de processar e classificar dados extra?dos de esfrega?os de tecidos mam?rios obtidos por FNA. Este m?todo est? integrado a um ambiente virtual para realiza??o de telediagn?stico colaborativo, cujo modelo foi desenvolvido prevendo a incorpora??o de M?dulos de Pr?-Diagn?stico para apoio ? tomada de decis?o m?dica. No desenvolvimento do m?todo fuzzy, o processo de aquisi??o do conhecimento foi realizado pela extra??o e an?lise dos dados num?ricos em base de dados padr?o ouro e por entrevistas e discuss?es com m?dicos especialistas. O m?todo foi testado e validado com casos reais e, em fun??o da sensibilidade e da especificidade alcan?adas (diagn?stico correto de tumores, respectivamente, malignos e benignos), os resultados obtidos foram satisfat?rios, considerando tanto os pareceres de m?dicos e os padr?es de qualidade para diagn?stico de c?ncer de mama quanto a compara??o com outros estudos realizados envolvendo diagn?stico de c?ncer de mama por FNA.

Le développement et l’usage croissants de systèmes pervasifs, dotés de fonctionnalités et de moyens de communication de plus en plus sophistiqués, offrent de fantastiques potentialités de services, en particulier pour l’e-Santé et la télémédecine, au bénéfice de tout citoyen, patient ou professionnel de santé. L’un des challenges sociétaux actuels est de permettre une meilleure exploitation des services disponibles pour l’ensemble des acteurs impliqués dans un domaine donné. Mais la multiplicité des services offerts, la diversité fonctionnelle des systèmes, et l’hétérogénéité des besoins nécessitent l’élaboration de modèles de connaissances de ces services, des fonctions de ces systèmes et des besoins. En outre, l’hétérogénéité des environnements informatiques distribués, la disponibilité et les capacités potentielles des diverses ressources humaines et matérielles (instrumentation, services, sources de données, etc.) requises par les différentes tâches et processus, la variété des services qui fournissent des données aux utilisateurs, et les conflits d’interopérabilité entre schémas et sources de données sont autant de problématiques que nous avons à considérer au cours de nos travaux de recherche. Notre contribution vise à optimiser la qualité de services en environnement ambiant et à réaliser une exploitation intelligente de ressources ubiquitaires. Pour cela, nous proposons un méta-modèle de connaissances des principaux concepts à prendre en compte en environnement pervasif. Ce méta-modèle est basé sur des ontologies décrivant les différentes entités précitées dans un domaine donné ainsi que leurs relations. Puis, nous l’avons formalisé en utilisant un langage standard de description des connaissances. A partir de ce modèle, nous proposons alors une nouvelle méthodologie de construction d’un framework architectural, que nous avons appelé ONOF-PAS. ONOF-PAS est basé sur des modèles ontologiques, une base de règles, un moteur d’inférence, et des composants orientés objet permettant la gestion des différentes tâches et le traitement des ressources. Il s’agit d’une architecture générique, applicable à différents domaines. ONOF-PAS a la capacité d’effectuer un raisonnement à base de règles pour gérer les différents contextes d’utilisation et aider à la prise de décision dans des environnements hétérogènes dynamiques, tout en tenant compte de la disponibilité et de la capacité des ressources humaines et matérielles requises par les diverses tâches et processus exécutés par des systèmes d’information pervasifs. Enfin, nous avons instancié ONOF-PAS dans le domaine de la télémédecine pour traiter le scénario de l’orientation des patients ou de personnes victimes de problèmes de santé en environnement hostile telles que la haute montagne ou des zones géographiquement isolées. Un prototype d’implémentation de ces scénarios, appelé T-TROIE a été développé afin de valider le framework ONOF-PAS. / The growing development and use of pervasive systems, equipped with increasingly sophisticated functionalities and communication means, offer fantastic potentialities of services, particularly in the eHealth and Telemedicine domains, for the benifit of each citizen, patient or healthcare professional. One of the current societal challenges is to enable a better exploitation of the available services for all actors involved in a given domain. Nevertheless, the multiplicity of the offered services, the systems functional variety, and the heterogeneity of the needs require the development of knowledge models of these services, systems functions, and needs. In addition, the distributed computing environments heterogeneity, the availability and potential capabilities of various human and material resources (devices, services, data sources, etc.) required by the different tasks and processes, the variety of services providing users with data, the interoperability conflicts between schemas and data sources are all issues that we have to consider in our research works. Our contribution aims to empower the intelligent exploitation of ubiquitous resources and to optimize the quality of service in ambient environment. For this, we propose a knowledge meta-model of the main concepts of a pervasive environment, such as Actor, Task, Resource, Object, Service, Location, Organization, etc. This knowledge meta-model is based on ontologies describing the different aforementioned entities from a given domain and their interrelationships. We have then formalized it by using a standard language for knowledge description. After that, we have designed an architectural framework called ONOF-PAS (ONtology Oriented Framework for Pervasive Applications and Services) mainly based on ontological models, a set of rules, an inference engine, and object oriented components for tasks management and resources processing. Being generic, extensible, and applicable in different domains, ONOF-PAS has the ability to perform rule-based reasoning to handle various contexts of use and enable decision making in dynamic and heterogeneous environments while taking into account the availability and capabilities of the human and material resources required by the multiples tasks and processes executed by pervasive systems. Finally, we have instantiated ONOF-PAS in the telemedicine domain to handle the scenario of the transfer of persons victim of health problems during their presence in hostile environments such as high mountains resorts or geographically isolated areas. A prototype implementing this scenario, called T-TROIE (Telemedicine Tasks and Resources Ontologies for Inimical Environments), has been developed to validate our approach and the proposed ONOF-PAS framework.

Os Sistemas Integrados de Gestão, também conhecidos como ERP (Enterprise Re-source Planning), vem tendo ampla utilização nas organizações a partir dos anos 90. Por exigir um investimento de elevado valor financeiro para a sua implantação, os responsáveis pela sua aquisição devem tomar cuidados especiais, uma vez que os seus resultados positivos ou negativos somente surgem após longo período de im-plantação, às vezes após muitos anos. Sendo um problema complexo, repleto de incertezas e riscos, os decisores tomam muito de seu tempo para analisar os diver-sos critérios e funcionalidades das ofertas de sistemas recebidas. A tese objetiva apresentar um procedimento que possibilite às empresas, em particular as de pe-queno e médio porte, um procedimento que as permita analisar quando do interesse da aquisição de um ERP, qual das ofertas disponíveis estará mais adequada às su-as necessidades de negócio, baseado em um método multicritérios de apoio à deci-são. A revisão da literatura analisa os Sistemas de Informação (SI) informatizados e os principais papéis desempenhados por eles: apoio às operações, apoio à vanta-gem competitiva e apoio à decisão. A seleção das ofertas propostas foi efetuada uti-lizando o método de Estudo de Casos múltiplos em empresas que adquiriram esses sistemas,ghy utilizando o ANP (Analytic Network Process) como instrumento de pesquisa. Para se estabelecer uma classificação dos critérios utilizados na análise foi utilizada a Técnica Delphi, a qual foi realizada junto a especialistas em Tecnologia de Informação. O resultado obtido pelo Estudo de Casos mostrou que o procedimen-to proposto é válido e pode ser utilizado por empresas de todos os portes. / The use of Integrated Management Systems, also known as ERP (Enterprise Re-source Planning), are widely accepted by organizations since beginning of the ni-neties. As its implementation means a high financial value investment, the respon-sible team for its acquisition has to take special cares, once their positive or nega-tive results will appear only after long implementation period, often after many years. As it is a complex decision problem, evolving uncertainties and risks, the decision agents spend a lot of time to analyze the several criteria and functional-ities from received offers. This thesis presents a proposal which makes possible the companies, particularly those of small or medium sizes, which allows to analyze during the ERP acquisition phase, the available offers more adapted to their business requirements, based on a multi-criteria support decision method. The literature revision analyzes the computerized Information Systems (IS) and the main roles carried out by them: operations support, competitive advantage support and decision support. In order to define the criteria set used in the multi-criteria analysis, the Delphi Method was used and it was answered by Information Technology experts. These criteria was used to classify the ERP\'s offers using the multiple cases study using ANP (Analytic Network Process) as research tool.. The results obtained by case study in four companies were used to validate several propositions.

A necessidade de investimento em saneamento no Brasil é essencial, pois está vinculada à melhoria da qualidade de vida da sociedade. Os projetos de saneamento exigem altos investimentos e, para garantir a prestação dos serviços, é necessário um sistema complexo de infraestrutura. Os elevados custos envolvidos e a limitação de recursos financeiros fazem com que seja preciso estabelecer prioridades para execução de projetos de saneamento. Com efeito, o objetivo desta pesquisa é propor um modelo multicritério de decisão para apoiar decisões de hierarquia de projetos de abastecimento de água e esgotamento sanitário. A pesquisa abordada tem enfoque qualitativo, sendo também vista como metodológica, pois sua finalidade é envolver métodos e procedimentos adotados como científicos. Assim, traz como escopo apoiar e estruturar o processo de decisão em que são definidos: os elementos (intervenientes, alternativas potenciais, critérios, problemática); tipos de decisão em grupo; escolha dos métodos multicritérios (PROMETHEE II & GAIA e ELECTRE IV); modelagem de preferência; sistemas de apoio à decisão (D-SIGHT e ELECTRE III-IV); avaliação de resultados e análise de sensibilidade. Procura-se garantir que a pesquisa tenha caráter prático, razão por que foi realizada a aplicação numérica do modelo no contexto da bacia dos rios Piracicaba, Capivari e Jundiaí, região sudeste do Brasil. / The need of investment in water supply and sewage projects in Brazil is substantial to improve the quality of life. These projects require high investments and, mostly, to ensure the provision of these services it is necessary a complex infrastructure. Due to the high costs associated with the lack of resources, it is relevant to prioritize projects. Thus, the purpose of this research is to propose a multicriteria decision model to support decisions hierarchy of water supply and sewage projects. This work has a qualitative and methodological approach; the goal is to inquire a scientific procedure. The object is to structure the decision-making process defined by its main concepts (actors, potential alternatives, criteria, problems), group decision making; selection of multiple criteria methods (PROMETHEE II & GAIA, ELECTRE IV); preference modeling, decision support systems (D-SIGHT and ELECTRE III-IV), evaluation and sensitivity analysis. This study seeks a practical purpose, so the proposed model, is applied in the basin of Piracicaba, Capivari e Jundiaí. The contribution here to aid similar situations where is necessary to establish priorities of sanitation projects.

Submitted by Irene Nascimento (irene.kessia@ufpe.br) on 2016-09-27T18:39:30Z No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) RicardoDeAndradeAraujo.pdf: 2136922 bytes, checksum: 3bf9d638152b4cc1870ed7c533772fae (MD5) / Made available in DSpace on 2016-09-27T18:39:30Z (GMT). No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) RicardoDeAndradeAraujo.pdf: 2136922 bytes, checksum: 3bf9d638152b4cc1870ed7c533772fae (MD5) Previous issue date: 2016-01-01 / CNPQ / Este trabalho apresenta um estudo sobre séries temporais financeiras, em alta-frequência, na tentativa de identificar as características do seu fenômeno gerador e, baseado neste estudo, propor um modelo, composto por uma combinação balanceada entre operadores lineares e operadores não-lineares crescentes e decrescentes, capaz de prever este tipo particular de série temporal. Para o processo de aprendizagem, é proposto um método baseado em gradiente descendente, utilizando ideias do algoritmo de retropropagação do erro (back propagation, BP) e uma abordagem alternativa para superar o problema da não-diferenciabilidade dos operadores não-lineares. Uma análise experimental é conduzida com o modelo proposto, utilizando um conjunto de séries temporais financeiras, em alta-frequência, do mercado de ações Brasileiro: Banco do Brasil SA, Banco Bradesco SA, Brasil Foods SA, BR Malls Participações SA e Companhia Energética Minas Gerais. Nestes experimentos, um conjunto relevante de medidas é utilizado para avaliar o desempenho preditivo do modelo proposto, e os resultados alcançados superam aqueles obtidos utilizando técnicas estatísticas, neurais e híbridas apresentadas na literatura. Também, são realizadas simulações com um sistema de apoio à decisão, baseado em previsão, para compra e venda de ações, tendo em vista demonstrar o desempenho econômico expressivo do modelo proposto no mercado de ações, em alta-frequência. / This work presents a study about high-frequency financial time series to identify the characteristics of their generator phenomenon and, based on such study, to propose a model, composed of a balanced combination of linear operators and increasing and decreasing nonlinear operators, able to predict this kind of time series. For the learning process, it is proposed a descent gradient-based method, using ideas from the back propagation (BP) algorithm and a systematic approach to overcome the problem of nondifferentiability of nonlinear operators. An experimental analysis is conducted with the proposed model, using a set of highfrequency financial time series of the Brazilian stock market: Banco do Brasil SA, Banco Bradesco SA, Brasil Foods SA, BR Malls Participações SA and Companhia Energética Minas Gerais. In these experiments, a relevant set of measures are used to assess the prediction performance of the proposed model, and the achieved results overcome those obtained by statistical, neural and hybrid techniques presented in the literature. Also, it is performed simulations with a prediction-based decision support system, for buy and sale of stocks, to demonstrate the significant economic performance of the proposed model in real high-frequency stock market

Orientador: Fernando José Von Zuben / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-15T23:45:59Z (GMT). No. of bitstreams: 1 Ferreira_HamiltonMelo_M.pdf: 539116 bytes, checksum: 3ef220fd5b659d2f8f794d853924ae71 (MD5) Previous issue date: 2008 / Resumo: As concessionárias de energia elétrica deixam de faturar anualmente expressivos valores devido a perdas comerciais, as quais são originadas principalmente por fraudes cometidas por parte dos consumidores e por medidores defeituosos. A detecção automática dos pontos específicos onde ocorrem tais perdas é uma tarefa complexa, dada a grande quantidade de consumidores, a grande variedade de perfis de consumo de energia elétrica e o alto custo de cada inspeção. Este trabalho propõe o uso de técnicas de aprendizado de máquina para a incorporação de processamento inteligente na identificação das fontes de perdas comerciais, usando os dados reais fornecidos pela concessionária de energia elétrica AES Eletropaulo. Além da manipulação dos dados e análise de propostas alternativas presentes na literatura, quatro estratégias de classificação foram implementadas e comparadas, sendo que o algoritmo de indução C4.5 produziu os resultados mais consistentes em termos de especificidade e confiabilidade, tomadas como critérios de desempenho / Abstract: The electric power concessionaires miss along the year significant amount of revenue due to commercial losses, which are mainly caused by frauds produced by consumers and defective sensors. The automatic detection of the specific sites where the losses are located is a complex task, given the high number of consumers, the great variety of electric power consumption profiles, and the high cost of each inspection. This work proposes the use of machine learning techniques capable of incorporating intelligent processing in the identification of the sources of commercial losses, using real data provided by the electric power concessionaire AES Eletropaulo. Besides data manipulation and analysis of alternative proposals presented in the literature, four classification strategies have been implemented and compared. The C4.5 algorithm has produced the most consistent results in terms of specificity and confiability, taken as performance criteria / Mestrado / Engenharia de Computação / Mestre em Engenharia Elétrica