M. Tech. (Department of Information Technology, Faculty of Applied and Computer Sciences), Vaal University of Technology. / The increasing demand for knowledge representation and exchange on the semantic web has resulted in an increase in both the number and size of ontologies. This increased features in ontologies has made them more complex and in turn difficult to select, reuse and maintain them. Several ontology evaluations and ranking tools have been proposed recently. Such evaluation tools provide a metrics suite that evaluates the content of an ontology by analysing their schemas and instances. The presence of ontology metric suites may enable classification techniques in placing the ontologies in various categories or classes. Machine Learning algorithms mostly based on statistical methods used in classification of data makes them the perfect tools to be used in performing classification of ontologies. In this study, popular Machine Learning algorithms including K-Nearest Neighbors, Support Vector Machines, Decision Trees, Random Forest, Naïve Bayes, Linear Regression and Logistic Regression were used in the classification of ontologies based on their complexity metrics. A total of 200 biomedical ontologies were downloaded from the Bio Portal repository. Ontology metrics were then generated using the OntoMetrics tool, an online ontology evaluation platform. These metrics constituted the dataset used in the implementation of the machine learning algorithms. The results obtained were evaluated with performance evaluation techniques, namely, precision, recall, F-Measure Score and Receiver Operating Characteristic (ROC) curves. The Overall accuracy scores for K-Nearest Neighbors, Support Vector Machines, Decision Trees, Random Forest, Naïve Bayes, Logistic Regression and Linear Regression algorithms were 66.67%, 65%, 98%, 99.29%, 74%, 64.67%, and 57%, respectively. From these scores, Decision Trees and Random Forests algorithms were the best performing and can be attributed to the ability to handle multiclass classifications.

Master of Science / Department of Computing and Information Sciences / Doina Caragea / In recent years, we have witnessed a significant increase in the number, size and diversity of the available data sources in many application domains. Data sources in a particular domain are autonomously created and maintained, and therefore distributed and semantically heterogeneous. In this thesis, we focused on the problem of querying such semantically heterogeneous data sources from a user's perspective. We approach this problem by using the concepts of ontologies and mappings between ontologies. A system for answering queries in a transparent way to the user has been designed and implemented. The main components of this system are an ontology mapping algorithm that maps user ontologies to data source ontologies, and a query processing engine that maps user queries to queries that can be answered by the data sources in the system. We have shown that machine learning algorithms can also be incorporated in the system, thus making it possible to learn machine learning classifiers (in particular, generative models such as Naïve Bayes) from distributed, semantically heterogeneous data sources. Because many data sources today are relational in nature, in this work we have dealt specifically with relational data sources, as opposed to flat files, XML or object oriented data sources. However, our system can be easily extended to other types of data sources.

The Semantic Web, as the foreseen successor of the current Web, is envisioned to be a semantically enriched information space usable by machines or agents that perform sophisticated tasks on behalf of their users. The realisation of the Semantic Web prescribe the development of a comprehensive and functional layered architecture for the increasingly semantically expressive languages that it comprises of. A functional architecture is a model specified at an appropriate level of abstraction identifying system components based on required system functionality, whilst a comprehensive architecture is an architecture founded on established design principles within Software Engineering. Within this study, an argument is formulated for the development of a comprehensive and functional layered architecture through the development of a Semantic Web status model, the extraction of the function of established Semantic Web technologies, as well as the development of an evaluation mechanism for layered architectures compiled from design principles as well as fundamental features of layered architectures. In addition, an initial version of such a comprehensive and functional layered architecture for the Semantic Web is constructed based on the building blocks described above, and this architecture is applied to several scenarios to establish the usefulness thereof. In conclusion, based on the evidence collected as result of the research in this study, it is possible to justify the development of an architectural model, or more specifically, a comprehensive and functional layered architecture for the languages of the Semantic Web. / Computing / PHD (Computer Science)

Σε αυτή τη διδακτορική διατριβή μελετήσαμε τις δυνατότητες αξιοποίησης των οντολογιών στην αναπαράσταση γνώσης σε συστήματα απανταχού και κινητού υπολογίζειν. / In this thesis we studied the use of ontologies for knowledge representation in ubiquitous and mobile computing.

Τα τελευταία χρόνια γίνεται ολοένα όλο και πιο έντονη η επιθυμία, τόσο από εκπαιδευτικούς ή μαθητές, αλλά και από άτομα μεγαλύτερης ηλικίας που θέλουν να αναπτύξουν την γνώση τους σε κάποιο αντικείμενο, για την δημιουργία εκπαιδευτικών μηχανών (λογισμικών) που θα μπορούν να αντικαταστήσουν σε μεγάλο βαθμό τον ρόλο του εκπαιδευτικού. Η προστιθέμενη αξία ενός εκπαιδευτικού λογισμικού θα μπορούσε να είναι η εύκολη πρόσβαση σε μεγάλο όγκο πληροφοριών, η πιο συστηματική εκμάθηση, καθώς και η εξοικονόμηση χρόνου και εκπαιδευτικών πηγών (εννοώντας τους εκπαιδευτικούς ως φυσικά πρόσωπα). Το ζητούμενο δεν είναι απλά η δημιουργία ενός εκπαιδευτικού λογισμικού αλλά ενός "σωστά" δομημένου εκπαιδευτικού συστήματος. Αυτό σημαίνει ότι ο εκπαιδευόμενος θα μπορεί να αντλεί σωστά και μεθοδικά πληροφορία από αυτό, όπως ακριβώς θα έκανε αν είχε στην διάθεσή του έναν καταρτισμένο εκπαιδευτικό. Στην παρούσα λοιπόν εργασία, μέσα από ένας συνδυασμό νέων τεχνολογιών όπως είναι αυτή των οντολογιών και του σημασιολογικού ιστού καθώς επίσης και θεωριών συσχετιζόμενων με την εκπαίδευση, παρουσιάζονται τα βήματα για δημιουργία ενός διαδραστικού crowdsoursing εκπαιδευτικού συστήματος. Παρουσιάζεται ένα σύστημα που με απλά λόγια θα είναι σε θέση να εξυπηρετεί μαθητές και εκπαιδευτικούς αλλά και οποιονδήποτε άλλο ενδιαφερόμενο, να προσφέρει μεθοδική εκμάθηση, να συλλέγει πληροφορία από τους χρήστες του την οποία να επεξεργάζεται και να την διαθέτει σε αυτούς σε ξανά βελτιωμένη και εμπλουτισμένη. / -

Présentation de travaux sur la création d'environnements collaboratifs synchrones réutilisables pour télé-TPs.

This thesis presents a framework for semantic Web Service discovery using descriptive (non-functional) service characteristics in a large-scale, multi-domain setting. The framework uses Web Ontology Language for Services (OWL-S) to design a template for describing non-functional service parameters in a way that facilitates service discovery, and presents a layered scheme for organizing ontologies used in service description. This service description scheme serves as a core for desigining the four main functions of a service directory: a template-based user interface, semantic query expansion algorithms, a two-level indexing scheme that combines Bloom filters with a Distributed Hash Table, and a distributed approach for storing service description. The service directory is, in turn, implemented as an extension of the Open Service Discovery Architecture. <br /><br /> The search algorithms presented in this thesis are designed to maximize precision and completeness of service discovery, while the distributed design of the directory allows individual administrative domains to retain a high degree of independence and maintain access control to information about their services.

19th IEEE International Conference on Computational Science and Engineering (CSE 2016), is the event, in a series of highly successful International Conferences on Computational Science and Engineering, held mainly as the International Workshop on High Performance Scientific and Engineering Computing for 11 editions. August 24-26, 2016 - Paris, France / Currently multimedia contents dominate the information exchanged in Internet, particularly through social networks. Each actor on the Internet becomes producer and consumer of contents. Nevertheless, social network and other traditional collaborative environments present limitations regarding content selection, categorization, aggregation, linking and interoperability, and usage control and privacy. In [1], we proposed the architecture (based on a peer-to-peer infrastructure and Semantic Web) of a MultiMedia Digital EcoSystem (MMDES), as a new environment for collaboration and sharing of multimedia resources, multimedia processings, as well as for computing and storage capabilities. In this paper, we describe MMDES framework and functionalities related to managing the collective knowledge and equilibrium in MMDES. We also describe the implementation of MMDES using a mobile platform in order to provide resources’ sharing for the Archivo Nacional de Arte Rupestre (ANAR) in Venezuela

Cette thèse porte sur l'exploitation des connaissances de domaine dans un processus de découvertes de sources de données biologiques sur le Web. Tout d'abord, des ensembles de métadonnées sont utilisés pour décrire le contenu et la qualité des sources de données. Ensuite, en s'appuyant sur ces métadonnées, les sources sont organisées dans un treillis de concepts en fonction de leurs caractéristiques communes. Le treillis de concepts constitue le support de la découverte de sources de données qui s'effectue de deux manières différentes et complémentaires : par navigation et par interrogation. Dans les deux cas la découverte de sources de données peut être guidée par des connaissances du domaine. Lors d'une découverte de sources de données par navigation, les connaissances sont utilisées soit pour réduire l'espace de recherche soit pour orienter la navigation vers des concepts sectionnés. Lors d'une découverte de sources de données par interrogation, les connaissances du domaine sont soit exprimées sous la forme de préférences entre métadonnées dans la requête soit utilisées pour l'enrichissement (ou reformulation) de la requête. Pour assurer une prise en compte des connaissances du domaine plus fidèle, nous avons introduit les treillis de concepts multivalués. L'organisation des sources de données sous la forme d'un treillis de concepts multivalués permet de contrôler la taille de l'espace de recherche et d'augmenter la flexibilité et les performances du processus de découverte dans ses deux modes. La navigation peut être effectuée dans des treillis de différents niveaux de spécialisation avec la possibilité d'effectuer des zooms dynamiques permettant le passage d'un treillis à l'autre. L'interrogation bénéficie d'une augmentation de l'expressivité dans les requêtes. / This thesis deals with knowledge-based biological data sources discovery. First, domain ontologies are used for encoding metadata describing the content of biological data sources. Then the data sources are organized into a concept lattice according to their common metadata. The data source discovery process can be performed either by navigation into the obtained concept lattice or by defining queries to be inserted into the concept lattice. In both cases, domain knowledge can be used to guide the discovery. In the case of navigation, domain knowledge is used to reduce the search space and/or to guide the navigation to some concepts rather than others. In the case of querying, domain knowledge is used to express preferences between the query keywords or to refine the query. In order to take more advantage of domain knowledge, we introduce many-valued concept lattices. Several many-valued concept lattices with different levels of precision can be built from the data sources metadata set based on domain knowledge. The use of such many-valued concept lattices allows to improve the discovery process in its both forms. In the case of navigation, it is possible to consider more than one lattice and to dynamically switch from one lattice to another in a zooming operation. In the case of querying, more complex expressive queries can be defined and inserted into the many-valued concept lattice.

Les systèmes d'information biomédicaux actuels reposent sur l'exploitation de données provenant de sources multiples. Les Systèmes d'Organisation de la Connaissance (SOC) permettent d'expliciter la sémantique de ces données, ce qui facilite leur gestion et leur exploitation. Bénéficiant de l'évolution des technologies du Web sémantique, un nombre toujours croissant de SOCs a été élaboré et publié dans des domaines spécifiques tels que la génomique, la biologie, l'anatomie, les pathologies, etc. Leur utilisation combinée, nécessaire pour couvrir tout le domaine biomédical, repose sur la définition de mises en correspondance entre leurs éléments ou mappings. Les mappings connectent les entités des SOCs liées au même domaine via des relations sémantiques. Ils jouent un rôle majeur pour l'interopérabilité entre systèmes, en permettant aux applications d'interpréter les données annotées avec différents SOCs. Cependant, les SOCs évoluent et de nouvelles versions sont régulièrement publiées de façon à correspondre à des vues du domaine les plus à jour possible. La validité des mappings ayant été préalablement établis peut alors être remis en cause. Des méthodes sont nécessaires pour assurer leur cohérence sémantique au fil du temps. La maintenance manuelle des mappings est une possibilité lorsque le nombre de mappings est restreint. En présence de SOCs volumineux et évoluant très rapidement, des méthodes les plus automatiques possibles sont indispensables. Cette thèse de doctorat propose une approche originale pour adapter les mappings basés sur les changements détectés dans l'évolution de SOCs du domaine biomédical. Notre proposition consiste à comprendre précisément les mappings entre SOCs, à exploiter les types de changements intervenant lorsque les SOCs évoluent, puis à proposer des actions de modification des mappings appropriées. Nos contributions sont multiples : (i) nous avons réalisé un travail expérimental approfondi pour comprendre l'évolution des mappings entre SOCs; nous proposons des méthodes automatiques (ii) pour analyser les mappings affectés par l'évolution de SOCs, et (iii) pour reconnaître l'évolution des concepts impliqués dans les mappings via des patrons de changement; enfin (iv) nous proposons des techniques d'adaptation des mappings à base d'heuristiques. Nous proposons un cadre complet pour l'adaptation des mappings, appelé DyKOSMap, et un prototype logiciel. Nous avons évalué les méthodes proposées et le cadre formel avec des jeux de données réelles contenant plusieurs versions de mappings entre SOCs du domaine biomédical. Les résultats des expérimentations ont démontré l'efficacité des principes sous-jacents à l'approche proposée. La maintenance des mappings, en grande partie automatique, est de bonne qualité. / Modern biomedical information systems require exchanging and retrieving data between them, due to the overwhelming available data generated in this domain. Knowledge Organization Systems (KOSs) offer means to make the semantics of data explicit which, in turn, facilitates their exploitation and management. The evolution of semantic technologies has led to the development and publication of an ever increasing number of large KOSs for specific sub-domains like genomics, biology, anatomy, diseases, etc. The size of the biomedical field demands the combined use of several KOSs, but it is only possible through the definition of mappings. Mappings interconnect entities of domain-related KOSs via semantic relations. They play a key role as references to enable advanced interoperability tasks between systems, allowing software applications to interpret data annotated with different KOSs. However, to remain useful and reflect the most up-to-date knowledge of the domain, the KOSs evolve and new versions are periodically released. This potentially impacts established mappings demanding methods to ensure, as automatic as possible, their semantic consistency over time. Manual maintenance of mappings stands for an alternative only if a restricted number of mappings are available. Otherwise supporting methods are required for very large and highly dynamic KOSs. To address such problem, this PhD thesis proposes an original approach to adapt mappings based on KOS changes detected in KOS evolution. The proposal consists in interpreting the established correspondences to identify the relevant KOS entities, on which the definition relies on, and based on the evolution of these entities to propose actions suited to modify mappings. Through this investigation, (i) we conduct in-depth experiments to understand the evolution of KOS mappings; we propose automatic methods (ii) to analyze mappings affected by KOS evolution, and (iii) to recognize the evolution of involved concepts in mappings via change patterns; finally (iv) we design techniques relying on heuristics explored by novel algorithms to adapt mappings. This research achieved a complete framework for mapping adaptation, named DyKOSMap, and an implementation of a software prototype. We thoroughly evaluated the proposed methods and the framework with real-world datasets containing several releases of mappings between biomedical KOSs. The obtained results from experimental validations demonstrated the overall effectiveness of the underlying principles in the proposed approach to adapt mappings. The scientific contributions of this thesis enable to largely automatically maintain mappings with a reasonable quality, which improves the support for mapping maintenance and consequently ensures a better interoperability over time.