An information system is widely recognised as a technological-political system and it can be better understood from a political perspective. This research studies how an information system is socially and politically constructed through the implementation process. It posits that the process of information system implementation can be shaped and impacted by power relations and thus that failing to understand such phenomena will contribute to system failure. The aim of this research is to broaden the understanding of an information system from a power perspective. Particular emphasis is placed on power relations in organisations and how such relations impact the process of system implementation. In order to achieve this aim, the research applies an integrated theoretical framework based on power and discourse perspectives to two qualitative case studies carried out in companies in Thailand. The case studies were undertaken based on an interpretive mode of inquiry, which provides ways of understanding how the process of implementation of information systems was shaped and impacted by human actions and political processes. lnterpretivism helped to explore the complexity of human action and meanings that system participants assigned to those actions leading to the system implementation process. The main findings suggest that information system implementation can be impacted by power relations enacted through discursive actions of textual practices 2 I constituting discourse. Besides, a grand discourse can itself be a source of power that affects the system implementation process. The findings indicate that: i) knowledge integrated in discourse plays a prominent role in power relations, ii) discursive legitimacy could enable and at the same time delimit the exercise of power, and iii) an information system may be used as an apparatus for domination as well as human emancipation. The main findings are taken as significant contributions to both theory and practice in the information systems field. 3

The classification of notated Western art music is a perennial issue. This thesis analyses and models the knowledge organization of notated Western art music in order to elucidate a theoretical understanding of these classification issues and to offer new ways of viewing music classification in the future. This thesis also considers how music classification contributes to developments in general knowledge organization and compares the classification of Western art music across the library and information science (LIS) and music domains. The research is conducted using a number of analytical techniques, including examining music knowledge organization discourse, analysing examples of LIS classification schemes, unpicking discussions of classification in the music domain and analysing composer worklists in the music domain. After ascertaining how music classification fits into theories of faceted classification, three important facets of music are identified: medium, form and genre, and a quasi-facet of function. These three facets are explored in detail over five chapters: the binary vocal/instrumental categorisation; classifying numbers of instruments or voices, accompaniment, arrangements and “extreme” mediums; classifying musical instruments; classifying musical forms and genres; and the quasi-facet of function. Five resulting models of music classification are presented. Model 1 demonstrates the complexities of classifying musical medium, including the interlinked relationships between different parts of musical medium. Model 2 offers a solution to LIS classification’s largely binary view of vocal and instrumental categorisation by suggesting a novel new category: “vocinstrumental”. Model 3 illuminates the entrenched dependencies between facets of music, highlighting one of the structural issues with LIS classifications of music. Model 4 offers an original structure of music classification, proposing a simultaneous faceted and genre-based system. Model 5 compares classification in the music and LIS domains, offering a novel way of considering domain-based classification by codifying various types of relationships between the LIS and domain classifications. This thesis also contributes to the theory and practice of knowledge organization in general through the development of novel frameworks and methodologies to analyse classification schemes: the multiplane approach, reception-infused analysis, webs of Wirkungs (connections) between classification schemes and stress-testing.

L'émergence du Web sémantique a mené à une adoption rapide du format RDF (Resource Description Framework) pour décrire les données et les liens entre elles. Ce modèle de graphe est adapté à la représentation des liens sémantiques entre les objets du Web qui sont identifiés par des IRI. Les applications qui publient et échangent des données RDF potentiellement sensibles augmentent dans de nombreux domaines : bio-informatique, e-gouvernement, mouvements open-data. La problématique du contrôle des accès aux contenus RDF et de l'exposition sélective de l'information en fonction des privilèges des requérants devient de plus en plus importante. Notre principal objectif est d'encourager les entreprises et les organisations à publier leurs données RDF dans l'espace global des données liées. En effet, les données publiées peuvent être sensibles, et par conséquent, les fournisseurs de données peuvent être réticents à publier leurs informations, à moins qu'ils ne soient certains que les droits d'accès à leurs données par les différents requérants sont appliqués correctement. D'où l'importance de la sécurisation des contenus RDF est de l'exposition sélective de l'information pour différentes classes d'utilisateurs. Dans cette thèse, nous nous sommes intéressés à la conception d'un contrôle d'accès pertinents pour les données RDF. De nouvelles problématiques sont posées par l'introduction des mécanismes de déduction pour les données RDF (e.g., RDF/S, OWL), notamment le problème de fuite d'inférence. En effet, quand un propriétaire souhaite interdire l'accès à une information, il faut également qu'il soit sûr que les données diffusées ne pourront pas permettre de déduire des informations secrètes par l'intermédiaire des mécanismes d'inférence sur des données RDF. Dans cette thèse, nous proposons un modèle de contrôle d'accès à grains fins pour les données RDF. Nous illustrons l'expressivité du modèle de contrôle d'accès avec plusieurs stratégies de résolution de conflits, y compris la Most Specific Takes Precedence. Nous proposons un algorithme de vérification statique et nous montrons qu'il est possible de vérifier à l'avance si une politique présente un problème de fuite d'inférence. De plus, nous montrons comment utiliser la réponse de l'algorithme à des fins de diagnostics. Pour traiter les privilèges des sujets, nous définissons la syntaxe et la sémantique d'un langage inspiré de XACML, basé sur les attributs des sujets pour permettre la définition de politiques de contrôle d'accès beaucoup plus fines. Enfin, nous proposons une approche d'annotation de données pour appliquer notre modèle de contrôle d'accès, et nous montrons que notre implémentation entraîne un surcoût raisonnable durant l'exécution / The emergence of the Semantic Web has led to a rapid adoption of the RDF (Resource Description Framework) to describe the data and the links between them. The RDF graph model is tailored for the representation of semantic relations between Web objects that are identified by IRIs (Internationalized Resource Identifier). The applications that publish and exchange potentially sensitive RDF data are increasing in many areas: bioinformatics, e-government, open data movement. The problem of controlling access to RDF content and selective exposure to information based on privileges of the requester becomes increasingly important. Our main objective is to encourage businesses and organizations worldwide to publish their RDF data into the linked data global space. Indeed, the published data may be sensitive, and consequently, data providers may avoid to release their information, unless they are certain that the desired access rights of different accessing entities are enforced properly, to their data. Hence the issue of securing RDF content and ensuring the selective disclosure of information to different classes of users is becoming all the more important. In this thesis, we focused on the design of a relevant access control for RDF data. The problem of providing access controls to RDF data has attracted considerable attention of both the security and the database community in recent years. New issues are raised by the introduction of the deduction mechanisms for RDF data (e.g., RDF/S, OWL), including the inference leakage problem. Indeed, when an owner wishes to prohibit access to information, she/he must also ensure that the information supposed secret, can’t be inferred through inference mechanisms on RDF data. In this PhD thesis we propose a fine-grained access control model for RDF data. We illustrate the expressiveness of the access control model with several conict resolution strategies including most specific takes precedence. To tackle the inference leakage problem, we propose a static verification algorithm and show that it is possible to check in advance whether such a problem will arise. Moreover, we show how to use the answer of the algorithm for diagnosis purposes. To handle the subjects' privileges, we define the syntax and semantics of a XACML inspired language based on the subjects' attributes to allow much finer access control policies. Finally, we propose a data-annotation approach to enforce our access control model, and show that our solution incurs reasonable overhead with respect to the optimal solution which consists in materializing the user's accessible subgraph to enforce our access control model, and show that our solution incurs reasonable overhead with respect to the optimal solution which consists in materializing the user's accessible subgraph

Un des premiers modèles d'indexation de documents qui utilise des termes comme descripteurs était une structure de treillis, cela une vingtaine d'années avant l'arrivée de l'analyse formelle de concepts (FCA pour "Formal Concept Analysis"), qui s'affirme maintenant comme un formalisme théorique important et solide pour l'analyse de données et la découverte de connaissances. Actuellement, la communauté en recherche d'information (RI) s'intéresse particulièrement à des techniques avancées pour la recherche des documents qui relèvent des probabilités et des statistiques. En parallèle, l'intérêt de la communauté FCA au développement de techniques qui font avancer l'état de l'art en RI tout en offrant des fonctionnalités sémantiques lui est toujours bien vivant. Dans cette thèse, nous présentons un ensemble de contributions sur ce que nous avons appelé les systèmes FCA de recherche d'information ("FCA-based IR systems''). Nous avons divisé nos contributions en deux parties, à savoir l'extraction et l'indexation. Pour la récupération, nous proposons une nouvelle technique qui exploite les relations sémantiques entre les descripteurs dans un corpus de documents. Pour l'indexation, nous proposons un nouveau modèle qui permet de mettre en oeuvre un modèle vectoriel d'indexation des documents s'appuyant sur un treillis de concepts (ou treillis de Galois). En outre, nous proposons un modèle perfectionné pour l'indexation hétérogène dans lequel nous combinons le modèle vectoriel et le modèle de recherche booléen. Finalement, nous présentons une technique de fouille de données inspiré de l'indexation des documents, à savoir un modèle d'énumération exhaustive des biclusters en utilisant la FCA. Le biclustering est une nouvelle technique d'analyse de données dans laquelle les objets sont liés via la similitude dans certains attributs de l'espace de description, et non pas par tous les attributs comme dans le "clustering'' standard. En traduisant ce problème en termes d'analyse formelle de concepts, nous pouvons exploiter l'algorithmique associée à la FCA pour développer une technique d'extraction de biclusters de valeurs similaires. Nous montrons le très bon comportement de notre technique, qui fonctionne mieux que les techniques actuelles de biclustering avec énumération exhaustive / One of the first models ever to be considered as an index for documents using terms as descriptors, was a lattice structure, a couple of decades before the arrival of Formal Concept Analysis (FCA) as a solid theory for data mining and knowledge discovery.While the Information Retrieval (IR) community has shifted to more advanced techniques for document retrieval, like probabilistic and statistic paradigms, the interest of the FCA community on developing techniques that would improve the state-of-the-art in IR while providing relevance feedback and semantic based features, never decayed. In this thesis we present a set of contributions on what we call FCA-based IR systems. We have divided our contributions in two sets, namely retrieval and indexing. For retrieval, we propose a novel technique that exploits semantic relations among descriptors in a document corpus and a new concept lattice navigation strategy (called cousin concepts), enabling us to support classification-based reasoning to provide better results compared with state-of-the-art retrieval techniques. The basic notion in our strategy is supporting query modification using "term replacements'' using the lattice structure and semantic similarity. For indexing, we propose a new model that allows supporting the vector space model of retrieval using concept lattices. One of the main limitations of current FCA-based IR systems is related to the binary nature of the input data required for FCA to generate a concept lattice. We propose the use of pattern structures, an extension of FCA to deal with complex object descriptions, in order to support more advanced retrieval paradigms like the vector space model. In addition, we propose an advanced model for heterogeneous indexing through which we can combine the vector space model and the Boolean retrieval model. The main advantage of this approach is the ability of supporting indexing of convex regions in an arbitrary vectorial space built from a document collection. Finally, we move forward to a mining model associated with document indexing, namely exhaustive bicluster enumeration using FCA. Biclustering is an emerging data analysis technique in which objects are related by similarity under certain attributes of the description space, instead of the whole description space like in standard clustering. By translating this problem to the framework of FCA, we are able to exploit the robust machinery associated with the computation of concept lattices to provide an algorithm for mining biclusters based on similar values. We show how our technique performs better than current exhaustive enumeration biclustering techniques.