Bibliographic data provide well established means of the information resource description. Nowadays, there is obvious need to make this data more web-friendly and web-compatible, while web is mostly seen as the only place where people are seeking information. The proposed solution is to change the traditional bibliographic data representation format. Semantic Web initiative presents enhancement of the current web with machine processable data semantics (meaning). The main goal is to manage more effective information processing by web applications that brings up many benefits for web users. Linked data are seen as practical means to reach the semantic web vision. This initiative promotes best practices of publishing, connecting and sharing structural data on the web. This thesis reports the transformation of bibliographic records' format in case of Database of publishing activities at University of Economics, Prague (PCVSE). Benefits of linked data publishing model are demonstrated on the realized projects. The main part describes practical experiences gained during the implementation of RDF framework in phases of data modelling, interlinking and exposing the dataset. The linked data publishing model is seen there as an optimal web-compatible bibliographic data representation format.

O crescimento exponencial do uso das mídias sociais na Web está transformando a maneira de como as pessoas tratam as informações, interagem com elas e compartilham conhecimento. Da mesma forma, as organizações estão mudando a maneira de interagir com seus funcionários, parceiros e consumidores. Novas aplicações na Internet têm surgido para proporcionar confiança aos usuários e incentivá-los a interagir e conectá-los uns aos outros e a conteúdos disponibilizados. Essas aplicações podem identificar comportamentos, extrair opiniões e retornar informações de interesse dos usuários e das organizações de maneira a auxiliar a tomada de decisão. Essas aplicações proporcionam grande volume de dados e demandam complexos processos de análise. Essas análises abrem oportunidades para o desenvolvimento de novas soluções que agregam mais valor aos usuários de produtos e serviços disponibilizadas na Web. Empresas e instituições de pesquisa têm desenvolvidos meios para tratar o grande volume de dados e identificar oportunidades de negócio. O uso de modelos que permitem entender esse fenômeno coletivo tem aumentado nos últimos anos por, basicamente, duas razões: a necessidade de descobrir, organizar e representar o conhecimento empírico relacionado a um determinado domínio de interesse e a necessidade de disseminar mecanismos para auxiliar os tomadores de decisão. Nesse contexto, ontologias de domínio têm sido bastante utilizadas como forma de organização e representação do conhecimento. No entanto, são poucos os modelos ou aplicações que extraem, organizam e representam o conhecimento (implícito e explícito) contextualizado de grupos de pessoas que atuam coletivamente para resolver problemas comuns ou produzir algo novo. Este trabalho de pesquisa propõe um modelo de referência para extração da inteligência coletiva (IC) para suporte à tomada de decisão. O modelo foi inicialmente desenvolvido para caso do planejamento estratégico de TI para ser utilizada por órgãos de governo. Como parte do modelo, foi desenvolvida uma inovadora ontologia de domínio denominada ITMPvoc. De seu processo de construção e validação, o modelo extrai a IC que é contextualiza segundo o referencial 5W1H (What, Who, Why, Where, When, How) e aplicada para suporte à decisão em situações específicas. Outras instâncias do modelo para dois casos de uso são também apresentadas. São elas: extração da IC e suporte à decisão para alertas de doenças na agricultura e para alertas sobre adoção de software livre por municípios. Os resultados demonstram que os modelos de extração da IC de comunidades ou organizações humanas podem melhorar os complexos processos de tomada de decisão em colaboração. Verificou-se também que a melhoria do processo de tomada de decisão se dá de duas maneiras. A primeira pela compreensão mais ampla pela comunidade dos conceitos e seus relacionamentos de causa e efeito mapeados pelo referencial 5W1H. A segunda pela composição mais adequada dos componentes What, Who, Why, Where, When e How em função do contexto. Ambas maneiras contribuem para o enriquecimento do conhecimento sobre os domínios considerados. / The exponential growth in the use of social media on the Web is transforming the which people treat information, interact with it, and share knowledge. Similarly, organizations are changing the way they interact with their employees, partners, and consumers. New Internet applications have emerged to provide users with confidence and encourage them to interact and connect to each other and to access contents made available. These applications can identify behaviors, extract opinions, and return information of interest to users and organizations in order to support decision making. These applications provide large amounts of data and require complex analysis processes. These analyzes open opportunities for the development of new solutions that add value to users of products and services available on the Web. Companies and research institutions have developed means to handle the large volume of data and identify business opportunities. The use of models that allow to understand this collective phenomenon has increased in recent years for basically two reasons: the need to discover, organize and represent empirical knowledge related to a particular domain of interest and the need to disseminate mechanisms to support the decision-makers. In this context, domain ontologies have been widely used as a form of organization and representation of the knowledge. However, there are few models or applications that extract, organize, and represent the contextualized (implicit and explicit) knowledge of groups of people who act collectively to solve common problems or produce something new. This research proposes a reference model for extracting the collective intelligence (CI) for decision making support. The model was initially developed for strategic planning of IT to be used by government organizations. As part of the model, an innovative domain ontology called ITMPvoc was developed. From its construction and validation process, the model extracts the CI that is contextualized according to the 5W1H (What, Who, Why, Where, How) framework and it is applied for decision making support in specific situations. Other instances of the model are also presented for two use cases. They are: extraction and decision making support based on CI for (I) early warning disease in agriculture and (ii) early warning in adoption of free software by municipalities. The results demonstrate that the CI extraction model from human communities or organizations can improve complex collaborative decision-making processes. It was also found that the improvement of the decision-making process occurs in two ways. The first is by the community\'s broader understanding of concepts and their cause-and-effect relationships mapped by the 5W1H framework. The second is the most appropriate composition of the What, Who, Why, Where, When, and How components, according to the context. Both ways contribute to the enrichment of the knowledge about the considered domains.

Die Studentenkonferenz Informatik Leipzig 2011 bietet die Möglichkeit, die Identifikation für das Studienfach Informatik und die Begeisterung für IT-Themen allgemein bei Studierenden zu wecken. Bei der Studentenkonferenz reichten Studierende kurze Artikel über Studien-, Abschlussarbeiten oder in der Freizeit absolvierte informatikrelevante Projekte ein. Andere Studierende, Doktoranden und wissenschaftliche Mitarbeiter der Leipziger Hochschulen bewerteten und diskutierten die eingereichten Arbeiten. Interessante und gut ausgearbeitete Einreichungen wurden zur Präsentation auf der Konferenz angenommen. Dieses Buch beinhaltet die überarbeiteten Beiträge der studentischen Autoren. Eine Studentenkonferenz unterscheidet sich kaum von einer anderen wissenschaftlichen Konferenz. Die Themenvielfalt kann allerdings durch die Breite der vertretenen Themen größer sein und die wissenschaftliche Innovation ist bei der Bewertung der Arbeiten nicht immer das primäre Kriterium. Eine Studentenkonferenz hilft, das kreative Potential von Studierenden besser sichtbar zu machen und Studierende für die Informatik und die Forschung zu begeistern. Außerdem stärkt sie den Austausch zwischen verschiedenen Disziplinen innerhalb der Informatik und fördert insbesondere das gegenseitige Verständnis von Lehrkräften und Studierenden. In diesem Jahr wurde am Institut für Angewandte Informatik (InfAI) e.V. zum zweiten Mal die Studentenkonferenz Informatik Leipzig (SKIL 2011) organisiert. Initiiert und maßgeblich organisiert wurde die SKIL 2011 von den Forschungsgruppen Agile Knowledge Engineering and Semantic Web (AKSW) und Service Science and Technology (SeSaT) der Universität Leipzig. Die Konferenz fand am 02. Dezember 2011 in Leipzig statt.

O crescimento exponencial do uso das mídias sociais na Web está transformando a maneira de como as pessoas tratam as informações, interagem com elas e compartilham conhecimento. Da mesma forma, as organizações estão mudando a maneira de interagir com seus funcionários, parceiros e consumidores. Novas aplicações na Internet têm surgido para proporcionar confiança aos usuários e incentivá-los a interagir e conectá-los uns aos outros e a conteúdos disponibilizados. Essas aplicações podem identificar comportamentos, extrair opiniões e retornar informações de interesse dos usuários e das organizações de maneira a auxiliar a tomada de decisão. Essas aplicações proporcionam grande volume de dados e demandam complexos processos de análise. Essas análises abrem oportunidades para o desenvolvimento de novas soluções que agregam mais valor aos usuários de produtos e serviços disponibilizadas na Web. Empresas e instituições de pesquisa têm desenvolvidos meios para tratar o grande volume de dados e identificar oportunidades de negócio. O uso de modelos que permitem entender esse fenômeno coletivo tem aumentado nos últimos anos por, basicamente, duas razões: a necessidade de descobrir, organizar e representar o conhecimento empírico relacionado a um determinado domínio de interesse e a necessidade de disseminar mecanismos para auxiliar os tomadores de decisão. Nesse contexto, ontologias de domínio têm sido bastante utilizadas como forma de organização e representação do conhecimento. No entanto, são poucos os modelos ou aplicações que extraem, organizam e representam o conhecimento (implícito e explícito) contextualizado de grupos de pessoas que atuam coletivamente para resolver problemas comuns ou produzir algo novo. Este trabalho de pesquisa propõe um modelo de referência para extração da inteligência coletiva (IC) para suporte à tomada de decisão. O modelo foi inicialmente desenvolvido para caso do planejamento estratégico de TI para ser utilizada por órgãos de governo. Como parte do modelo, foi desenvolvida uma inovadora ontologia de domínio denominada ITMPvoc. De seu processo de construção e validação, o modelo extrai a IC que é contextualiza segundo o referencial 5W1H (What, Who, Why, Where, When, How) e aplicada para suporte à decisão em situações específicas. Outras instâncias do modelo para dois casos de uso são também apresentadas. São elas: extração da IC e suporte à decisão para alertas de doenças na agricultura e para alertas sobre adoção de software livre por municípios. Os resultados demonstram que os modelos de extração da IC de comunidades ou organizações humanas podem melhorar os complexos processos de tomada de decisão em colaboração. Verificou-se também que a melhoria do processo de tomada de decisão se dá de duas maneiras. A primeira pela compreensão mais ampla pela comunidade dos conceitos e seus relacionamentos de causa e efeito mapeados pelo referencial 5W1H. A segunda pela composição mais adequada dos componentes What, Who, Why, Where, When e How em função do contexto. Ambas maneiras contribuem para o enriquecimento do conhecimento sobre os domínios considerados. / The exponential growth in the use of social media on the Web is transforming the which people treat information, interact with it, and share knowledge. Similarly, organizations are changing the way they interact with their employees, partners, and consumers. New Internet applications have emerged to provide users with confidence and encourage them to interact and connect to each other and to access contents made available. These applications can identify behaviors, extract opinions, and return information of interest to users and organizations in order to support decision making. These applications provide large amounts of data and require complex analysis processes. These analyzes open opportunities for the development of new solutions that add value to users of products and services available on the Web. Companies and research institutions have developed means to handle the large volume of data and identify business opportunities. The use of models that allow to understand this collective phenomenon has increased in recent years for basically two reasons: the need to discover, organize and represent empirical knowledge related to a particular domain of interest and the need to disseminate mechanisms to support the decision-makers. In this context, domain ontologies have been widely used as a form of organization and representation of the knowledge. However, there are few models or applications that extract, organize, and represent the contextualized (implicit and explicit) knowledge of groups of people who act collectively to solve common problems or produce something new. This research proposes a reference model for extracting the collective intelligence (CI) for decision making support. The model was initially developed for strategic planning of IT to be used by government organizations. As part of the model, an innovative domain ontology called ITMPvoc was developed. From its construction and validation process, the model extracts the CI that is contextualized according to the 5W1H (What, Who, Why, Where, How) framework and it is applied for decision making support in specific situations. Other instances of the model are also presented for two use cases. They are: extraction and decision making support based on CI for (I) early warning disease in agriculture and (ii) early warning in adoption of free software by municipalities. The results demonstrate that the CI extraction model from human communities or organizations can improve complex collaborative decision-making processes. It was also found that the improvement of the decision-making process occurs in two ways. The first is by the community\'s broader understanding of concepts and their cause-and-effect relationships mapped by the 5W1H framework. The second is the most appropriate composition of the What, Who, Why, Where, When, and How components, according to the context. Both ways contribute to the enrichment of the knowledge about the considered domains.

La nature temporelle de l'audiovisuel impose de passer par le biais de la description pour enrichir les documents et donc les exploiter. Nous soutenons qu'une représentation de la structure et du contenu des documents est nécessaire. Par structure, nous entendons la structure documentaire c'est-à-dire l'organisation méréologique des éléments qui composent le document, tandis que le contenu est une structure conceptuelle, c'est-à-dire une catégorisation de ces éléments. Après une revue des propositions actuelles de modélisation des documents audiovisuels, issues de l'ingénierie documentaire et de l'ingénierie des connaissances, nous montrons qu'aucun des langages étudiés ne permet de traiter ces deux aspects de manière satisfaisante. Nous proposons alors une architecture générale permettant la représentation formelle de la structure et du contenu des documents audiovisuels, qui engendrera une base de connaissances sur laquelle il est possible d'effectuer des raisonnements. Cette architecture se compose d'une ontologie de l'audiovisuel, dont on traduit une partie dans un langage documentaire pour contrôler la structure logique des documents, et d'une ontologie de domaine pour décrire formellement leur contenu. Nous avons développé l'outil DOE (Differential Ontology Editor), qui implémente la méthodologie de construction d'ontologies utilisée. Nous montrons finalement la pertinence de l'approche à l'aide de deux expérimentations utilisant un corpus de vidéos annoté, illustrant ainsi les types d'inférences possibles.

L'essor d'Internet et du Web a favorisé la mise en ligne de nombreuses informations disponibles, potentiellement utiles pour la veille technologique et scientifique d'une entreprise. Différentes techniques de recherche d¤information sur le Web ont été proposées afin de construire des outils permettant d'affiner la recherche pour obtenir des résultats pertinents. Cependant, dans le contexte du Web actuel, malgré de grandes avancées dans le champ de la recherche d'information, ces outils ont montré leurs limites en termes de précision et de rappel. L'application des technologies du Web Sémantique, en particulier des ontologies, semble donc intéressante pour améliorer les performances de la tâche de veille technologique et scientifique sur le Web. Les travaux de cette thèse se sont déroulés dans le cadre d'une coopération entre le Centre Scientifique et Technique du Bâtiment (CSTB) et l'équipe ACACIA de l'INRIA Sophia Antipolis. L'objectif principal est d'exploiter les technologies du Web Sémantique pour développer un système de veille (OntoWatch), guidé par des ontologies, pour collecter, capturer, filtrer, classer et structurer le contenu du Web en provenance de plusieurs sources d¤information dans un scénario d'aide à la veille technologique et scientifique. Dans une première partie, nous modélisons le processus de veille technologique et scientifique du CSTB reposant sur le modèle général de veille proposé par Lesca. Puis nous identifions les apports potentiels de l¤ontologie dans les différentes étapes et nous construisons une ontologie dédiée au système de veille. Cette ontologie intègre une partie d'une ontologie existante et des vocabulaires offerts dans des thésaurus du domaine du CSTB. Ensuite, nous proposons des algorithmes utilisant une ontologie pour améliorer la recherche des documents sur le Web, puis générer automatiquement les annotations sémantiques (représentées dans le langage RDF) sur ces documents. Ces annotations alimentent dans le système les bases d'annotations, sur lesquelles repose la recherche sémantique d'informations. <br />Enfin, nous proposons une architecture multi-agents pour l'implémentation du système OntoWatch. Nous nous focalisons en particulier sur la conception des sous-sociétés d'agents dédiées à la recherche et à l'annotation automatique des documents sur le Web.

RDF est un langage de représentation des connaissances dédié à l'annotation des ressources dans le Web Sémantique. Bien que RDF peut être lui-même utilisé comme un langage de requêtes pour interroger une base de connaissances RDF (utilisant la conséquence RDF), la nécessité d'ajouter plus d'expressivité dans les requêtes a conduit à définir le langage de requêtes SPARQL. Les requêtes SPARQL sont définies à partir des patrons de graphes qui sont fondamentalement des graphes RDF avec des variables. Les requêtes SPARQL restent limitées car elles ne permettent pas d'exprimer des requêtes avec une séquence non-bornée de relations (par exemple, Existe-t-il un itinéraire d'une ville A à une ville B qui n'utilise que les trains ou les bus?"). Nous montrons qu'il est possible d'étendre la syntaxe et la sémantique de RDF, définissant le langage PRDF (pour Path RDF) afin que SPARQL puisse surmonter cette limitation en remplaçant simplement les patrons de graphes basiques par des graphes PRDF. Nous étendons aussi PRDF à CPRDF (pour Constrained Path RDF) permettant d'exprimer des contraintes sur les sommets des chemins traversés (par exemple, "En outre, l'une des correspondances doit fournir une connexion sans fil."). Nous avons fourni des algorithmes corrects et complets pour répondre aux requêtes (la requête est un graphe PRDF ou CPRDF, la base de connaissances est un graphe RDF) basés sur un homomorphisme particulier, ainsi qu'une analyse détaillée de la complexité. Enfin, nous utilisons les graphes PRDF ou CPRDF pour généraliser les requêtes SPARQL, définissant les extensions PSPARQL et CPSPARQL, et fournissons des tests expérimentaux en utilisant une implémentation complète de ces deux langages.

Le Web Sémantique d'Entreprise (WSE) est une approche particulière de la Gestion des Connaissances d'une entreprise pour la prochaine génération du Web Sémantique. Dans la réalité, les organisations vivent dans un environnement hétérogène, dynamique et en cours d'évolution qui mène souvent à des changements externes et internes requérant l'évolution de leur système de gestion des connaissances. Peu de recherches actuelles font face aux changements du système de gestion des connaissances. L'objectif de cette thèse est de contribuer à lever cette limitation. Dans ce manuscrit, nous présentons une nouvelle approche de la gestion de l'évolution du WSE. Nous nous focalisons en particulier sur l'évolution de l'ontologie et de l'annotation sémantique qui sont deux composants importants du WSE. Nous nous intéressons à deux scénarios d'évolution de l'ontologie : (i) avec trace et (ii) sans trace de changements ontologiques effectués. Ces deux scénarios sont fréquents dans les situations réelles et ils peuvent entraîner des inconsistances au niveau des annotations sémantiques reposant sur cette ontologie modifiée. Pour chacun des contextes d'évolution, nous développons des approches équivalentes : une approche procédurale et une approche basée sur des règles en vue de gérer l'évolution des annotations sémantiques et de détecter et de corriger les annotations sémantiques inconsistantes. Ces propositions ont été implémentées et validées dans le système CoSWEM qui facilite la gestion de l'évolution du WSE. Ce système a été expérimenté dans le cadre des projets PALETTE et E-WOK_HUB sur un ensemble de données réelles et évolutives provenant de ces projets.

<p>It is not easy to apply XML in e-commerce development for achieving interoperability in heterogeneous environments. One of the reasons is a multitude of XML-based Frameworks for Internet Commerce (XFIC), or industrial standards. This thesis surveys 15 frameworks, i.e., ebXML, eCo Framework, UDDI, SOAP, BizTalk, cXML, ICE, Open Applications Group, RosettaNet, Wf-XML, OFX, VoiceXML, RDF, WSDL and xCBL.</p><p>This thesis provides three models to systematically understand how the 15 frameworks meet the requirements of e-commerce. A hierarchical model is presented to show the purpose and focus of various XFIC initiatives. A relationship model is given to show the cooperative and competitive relationships between XFIC. A chronological model is provided to look at the development of XFIC. In addition, the thesis offers guidelines for how to apply XFIC in an e-commerce development.</p><p>We have also implemented a B2B e-procurement system. That not only demonstrates the feasibility of opensource or freeware, but also validates the complementary roles of XML and Java: XML is for describing contents and Java is for automating XML documents (session handling). Auction-based dynamic pricing is also realized as a feature of interest. Moreover, the implementation shows the suitability of e-procurement for educational purposes in e-commerce development.</p> / Report code: LiU-Tek-Lic-2001:19.

This thesis concerns catalytic combustion for gas turbineapplication using a low heating-value (LHV) gas, derived fromgasified waste. The main research in catalytic combustionfocuses on methane as fuel, but an increasing interest isdirected towards catalytic combustion of LHV fuels. This thesisshows that it is possible to catalytically combust a LHV gasand to oxidize fuel-bound nitrogen (NH3) directly into N2without forming NOX. The first part of the thesis gives abackground to the system. It defines waste, shortly describesgasification and more thoroughly catalytic combustion. The second part of the present thesis, paper I, concerns thedevelopment and testing of potential catalysts for catalyticcombustion of LHV gases. The objective of this work was toinvestigate the possibility to use a stable metal oxide insteadof noble metals as ignition catalyst and at the same timereduce the formation of NOX. In paper II pilot-scale tests werecarried out to prove the potential of catalytic combustionusing real gasified waste and to compare with the resultsobtained in laboratory scale using a synthetic gas simulatinggasified waste. In paper III, selective catalytic oxidation fordecreasing the NOX formation from fuel-bound nitrogen wasexamined using two different approaches: fuel-lean andfuel-rich conditions. Finally, the last part of the thesis deals with deactivationof catalysts. The various deactivation processes which mayaffect high-temperature catalytic combustion are reviewed inpaper IV. In paper V the poisoning effect of low amounts ofsulfur was studied; various metal oxides as well as supportedpalladium and platinum catalysts were used as catalysts forcombustion of a synthetic gas. In conclusion, with the results obtained in this thesis itwould be possible to compose a working catalytic system for gasturbine application using a LHV gas. <b>Keywords:</b>Catalytic combustion; Gasified waste; LHVfuel; RDF; Biomass; Selective catalytic oxidation; NH3; NOX;Palladium; Platinum; Hexaaluminate; Garnet; Spinel;Deactivation; Sulfur; Poisoning