L’immense quantité d'information générée et gérée au quotidien par les systèmes d'information et leurs utilisateurs conduit inéluctablement à la problématique de surcharge d'information. Dans ce contexte, les systèmes de recommandation traditionnels fournissent des informations pertinentes aux utilisateurs. Néanmoins, avec la propagation récente des dispositifs mobiles (smartphones et tablettes), nous constatons une migration progressive des utilisateurs vers la manipulation d'environnements pervasifs. Le problème avec les approches de recommandation traditionnelles est qu'elles n'utilisent pas toute l'information disponible pour produire des recommandations. Davantage d’informations contextuelles pourraient être utilisées dans le processus de recommandation pour aboutir à des recommandations plus précises. Les systèmes de recommandation sensibles au contexte (CARS) combinent les caractéristiques des systèmes sensibles au contexte et des systèmes de recommandation afin de fournir des informations personnalisées aux utilisateurs dans des environnements ubiquitaires. Dans cette perspective où tout ce qui concerne l'utilisateur est dynamique, les contenus qu’il manipule et son environnement, deux questions principales doivent être adressées : i) Comment prendre en compte l'évolution des contenus de l’utilisateur? et ii) Comment éviter d’être intrusif, en particulier dans des situations critiques? En réponse à ces questions, nous avons développé un système de recommandation dynamique et sensible au risque appelé DRARS (Dynamic Risk-Aware Recommender System), qui modélise la recommandation sensible au contexte comme un problème de bandit. Ce système combine une technique de filtrage basée sur le contenu et un algorithme de bandit contextuel. Nous avons montré que DRARS améliore la stratégie de l'algorithme UCB (Upper Confidence Bound), le meilleur algorithme actuellement disponible, en calculant la valeur d'exploration la plus optimale pour maintenir un bon compromis entre exploration et exploitation basé sur le niveau de risque de la situation courante de l'utilisateur. Nous avons mené des expériences dans un contexte industriel avec des données réelles et des utilisateurs réels et nous avons montré que la prise en compte du niveau de risque de la situation de l'utilisateur augmentait significativement la performance du système de recommandation / The vast amount of information generated and maintained everyday by information systems and their users leads to the increasingly important concern of overload information. In this context, traditional recommender systems provide relevant information to the users. Nevertheless, with the recent dissemination of mobile devices (smartphones and tablets), there is a gradual user migration to the use of pervasive computing environments. The problem with the traditional recommendation approaches is that they do not utilize all available information for producing recommendations. More contextual parameters could be used in the recommendation process to result in more accurate recommendations. Context-Aware Recommender Systems (CARS) combine characteristics from context-aware systems and recommender systems in order to provide personalized recommendations to users in ubiquitous environments. In this perspective where everything about the user is dynamic, his/her content and his/her environment, two main issues have to be addressed: i) How to consider content evolution? and ii) How to avoid disturbing the user in risky situations?. In response to these problems, we have developed a dynamic risk sensitive recommendation system called DRARS (Dynamic Risk-Aware Recommender System), which model the context-aware recommendation as a bandit problem. This system combines a content-based technique and a contextual bandit algorithm. We have shown that DRARS improves the Upper Confidence Bound (UCB) policy, the currently available best algorithm, by calculating the most optimal exploration value to maintain a trade-off between exploration and exploitation based on the risk level of the current user's situation. We conducted experiments in an industrial context with real data and real users and we have shown that taking into account the risk level of users' situations significantly increases the performance of the recommender system

A l’avenir, l'agriculture et l'environnement vont pouvoir bénéficier de plus en plus de données hétérogènes collectées par des réseaux de capteurs sans fil (RCSF). Ces données alimentent généralement des outils d’aide à la décision (OAD). Dans cette thèse, nous nous intéressons spécifiquement aux systèmes sensibles et adaptatifs au contexte basés sur un RCSF et un OAD, dédiés au suivi de phénomènes naturels. Nous proposons ainsi une formalisation pour la conception et la mise en œuvre de ces systèmes. Le contexte considéré se compose de données issues du phénomène étudié mais également des capteurs sans fil (leur niveau d’énergie par exemple). Par l’utilisation des ontologies et de techniques de raisonnement, nous visons à maintenir le niveau de qualité de service (QdS) des données collectées (en accord avec le phénomène étudié) tant en préservant le fonctionnement du RCSF. Pour illustrer notre proposition, un cas d'utilisation complexe, l'étude des inondations dans un bassin hydrographique, est considéré. Cette thèse a produit un logiciel de simulation de ces systèmes qui intègre un système de simulation multi-agents (JADE) avec un moteur d’inférence à base de règles (Jess). / In the future, agriculture and environment will rely on more and more heterogeneous data collected by wireless sensor networks (WSN). These data are generally used in decision support systems (DSS). In this dissertation, we focus on adaptive context-aware systems based on WSN and DSS, dedicated to the monitoring of natural phenomena. Thus, a formalization for the design and the deployment of these kinds of systems is proposed. The considered context is established using the data from the studied phenomenon but also from the wireless sensors (e.g., their energy level). By the use of ontologies and reasoning techniques, we aim to maintain the required quality of service (QoS) level of the collected data (according to the studied phenomenon) while preserving the resources of the WSN. To illustrate our proposal, a complex use case, the study of floods in a watershed, is described. During this PhD thesis, a simulator for context-aware systems which integrates a multi-agent system (JADE) and a rule engine (Jess) has been developed.Keywords: ontologies, rule-based inferences, formalization, heterogeneous data, sensors data streams integration, WSN, limited resources, DSS, adaptive context-aware systems, QoS, agriculture, environment.

Mobile learning (m-learning) is still in its infancy, and great efforts should be made so as to investigate the potentials of an educational paradigm shift from the traditional one-size-fits-all teaching approaches to an adaptive learning that can be delivered via mobile devices. Thus, the next challenge has been identified from this implication: How learning design can be implemented so as to benefit from the m-learning characteristics and achieve adaptation and personalization of the individual learning process in different contexts? An important factor for achieving personalized and adaptive m-learning has been the pedagogically meaningful and technically feasible processing of learners’ contextual information. Therefore in this work, design and delivery of personalized educational scenarios are suggested to be re-thought so as to benefit from the affordances of mobile technologies and the learners’ context / El aprendizaje móvil (m-learning) se encuentra todavía en su infancia y grandes esfuerzos se deben hacer para investigar el cambio de paradigma educativo, desde la forma de enseñanza tradicional de 'un modelo único para todos' a un aprendizaje adaptativo que se pueda entregar a través de dispositivos móviles. De esta manera, el siguiente desafío ha sido identificado por esta implicación: ¿Cómo se puede implementar el diseño instruccional con el fin de beneficiarse de las características del m-learning y lograr la adaptación y personalización del proceso de aprendizaje personal en diferentes contextos? Un importante factor para lograr un m-learning personalizado y adaptable ha sido el procesamiento de la información contextual de los estudiantes. Por lo tanto, en este trabajo se sugiere que sean re-pensados el diseño y la entrega de escenarios educativos orientados a la personalización del aprendizaje y que se beneficien de las potencialidades de las tecnologías móviles y el contexto de los estudiantes

Advances in Internet-of-Things integrate sensors and actuators in everyday items or even people transforming our society at an accelerated pace. This occurs in areas such as agriculture, logistics, transport, healthcare, and smart cities and has created new ways to interact with and experience entertainment, (serious) games, education, etc. Common to these domains is the challenge to realize and maintain complex relations with any object or individual globally, with the requirement for immediacy in maintaining relations of varying complexity. Existing architectures for maintaining relations on the Internet, e.g., DNS and search engines are insufficient in meeting these challenges. Their deficiencies mandate the research presented in this dissertation enabling the maintenance of dynamic and multi-criteria relationships among entities in real-time in an Internet-of-Things while minimizing the overall cost for maintaining such context-centric relationships. A second challenge is the need to represent nearness in context-centric relationships, since solutions need to build on what is closely related. The dissertation shows that the proximity on relations can be used to bring about the scalability of maintaining relationships across the IoT. It successfully demonstrates the concept and feasibility of self-organizing context-centric overlay networks for maintaining scalable and real-time relationships between endpoints co-located with associated physical entities. This is complemented by an object model for annotating objects and their relationships as derived and defined over the underpinning context interactions. Complementing measures of nearness are added through a non-metric multi-criteria approach to evaluating the notion of context proximity. A query language and an extension to the publish-subscribe approaches achieves distributed support for discovering such relationships; locating entities relative to a defined hyper-sphere of interest. Furthermore, it introduces adaptive algorithms for maintaining such relationships at minimal overall costs. The results demonstrate the feasibility of moving towards context-centric approaches to immersion and that such approaches are realizable over vast and distributed heterogeneous collections of user and their associated context information.

PhD (Business Management) / Department of Business Management / This thesis demonstrates the researcher’s efforts to put into practice the theoretical foundations of information systems research, in order to come up with a context-aware business intelligence framework (CABIF), for the South African higher education institutions. Using critical realism as the philosophical underpinning and mixed methods research design, a business intelligence (BI) survey was deployed within the South African public higher education institutions to measure the respondents’ satisfaction and importance of business intelligence characteristics. The 258 respondents’ satisfaction and importance of the 34 observed business intelligence variables, were subjected to principal components analysis and design science research to come up with the CABIF. The observable BI variables were drawn from four latent variables namely technology and business alignment; organizational and behavioural strategies; business intelligence domain; and technology strategies. The study yielded good values for all the observed satisfaction and importance business intelligence variables as indicated by the Kaiser- Meyer-Olkin (KMO) Measure of Sampling Adequacy and the Bartlett Test of Sphericity. The data set collected from the survey deployed at the South African public higher education institutions, was reliable and valid based on the Cronbach α values which were all above 0.9. The researcher then used the descriptive and prescriptive knowledge of design science research, and the meta-inferences of the results from the principal components analysis to produce five contexts of CABIF. The BI contexts developed were, the Basic Context; the Business Processes Context which was divided into Macro and Micro business process contexts; the Business Intelligence Context; and the Governance Context. These contexts were extrapolated within the University of Venda’s business processes and this researcher concluded that the CABIF developed, could be inferred within the South African higher education institutions. At the University of Venda, this researcher managed to draw up CABIF based business intelligence tools that spanned from leveraging the existing ICT infrastructure, student cohort analysis, viability of academic entities, strategic enrolment planning and forecasting government block grants. The correlations and regression measures of the technology acceptance variables of the business intelligence tools modelled using CABIF at University of Venda, revealed high acceptance ratio. Overall, this research provides a myriad of conceptual and practical insights into how contextualised aspects of BI directly or indirectly impact on the quality of managerial decision making within various core business contexts of South African higher education institutions.

Les applications réactives et sensibles au contexte sont des applications intelligentes qui observent l’environnement (ou contexte) dans lequel elles s’exécutent et qui adaptent, si nécessaire, leur comportement en cas de changements dans ce contexte, ou afin de satisfaire les besoins ou d'anticiper les intentions des utilisateurs. La recherche dans ce domaine suscite un intérêt considérable tant de la part des académiques que des industriels. Les domaines d'applications sont nombreux: robots industriels qui peuvent détecter les changements dans l'environnement de travail de l'usine pour adapter leurs opérations; systèmes de contrôle automobiles pour observer d'autres véhicules, détecter les obstacles, ou surveiller le niveau d'essence ou de la qualité de l'air afin d'avertir les conducteurs en cas d'urgence; systèmes embarqués monitorant la puissance énergétique disponible et modifiant la consommation en conséquence. Dans la pratique, le succès de la mise en œuvre et du déploiement de systèmes sensibles au contexte dépend principalement du mécanisme de reconnaissance et de réaction aux variations de l'environnement. En d'autres termes, il est nécessaire d'avoir une approche adaptative bien définie et efficace de sorte que le comportement des systèmes peut être modifié dynamiquement à l'exécution. En outre, la concurrence devrait être exploitée pour améliorer les performances et la réactivité des systèmes. Tous ces exigences, ainsi que les besoins en sécurité et fiabilité constituent un grand défi pour les développeurs.C’est pour permettre une écriture plus intuitive et directe d'applications réactives et sensibles au contexte que nous avons développé dans cette thèse un nouveau langage appelé INI. Pour observer les changements dans le contexte et y réagir, INI s’appuie sur deux paradigmes : la programmation événementielle et la programmation à base de règles. Événements et règles peuvent être définis en INI de manière indépendante ou en combinaison. En outre, les événements peuvent être reconfigurésdynamiquement au cours de l’exécution. Un autre avantage d’INI est qu’il supporte laconcurrence afin de gérer plusieurs tâches en parallèle et ainsi améliorer les performances et la réactivité des programmes. Nous avons utilisé INI dans deux études de cas : une passerelle M2M multimédia et un programme de suivi d’objet pour le robot humanoïde Nao. Enfin, afin d’augmenter la fiabilité des programmes écrits en INI, un système de typage fort a été développé, et la sémantique opérationnelle d’INI a été entièrement définie. Nous avons en outre développé un outil appelé INICheck qui permet de convertir automatiquement un sous-ensemble d’INI vers Promela pour permettre un analyse par model checking à l’aide de l’interpréteur SPIN. / Context-aware pervasive computing has attracted a significant research interest from both academy and industry worldwide. It covers a broad range of applications that support many manufacturing and daily life activities. For instance, industrial robots detect the changes of the working environment in the factory to adapt their operations to the requirements. Automotive control systems may observe other vehicles, detect obstacles, and monitor the essence level or the air quality in order to warn the drivers in case of emergency. Another example is power-aware embedded systems that need to work based on current power/energy availability since power consumption is an important issue. Those kinds of systems can also be considered as smart applications. In practice, successful implementation and deployment of context-aware systems depend on the mechanism to recognize and react to variabilities happening in the environment. In other words, we need a well-defined and efficient adaptation approach so that the systems' behavior can be dynamically customized at runtime. Moreover, concurrency should be exploited to improve the performance and responsiveness of the systems. All those requirements, along with the need for safety, dependability, and reliability pose a big challenge for developers.In this thesis, we propose a novel programming language called INI, which supports both event-based and rule-based programming paradigms and is suitable for building concurrent and context-aware reactive applications. In our language, both events and rules can be defined explicitly, in a stand-alone way or in combination. Events in INI run in parallel (synchronously or asynchronously) in order to handle multiple tasks concurrently and may trigger the actions defined in rules. Besides, events can interact with the execution environment to adjust their behavior if necessary and respond to unpredictable changes. We apply INI in both academic and industrial case studies, namely an object tracking program running on the humanoid robot Nao and a M2M gateway. This demonstrates the soundness of our approach as well as INI's capabilities for constructing context-aware systems. Additionally, since context-aware programs are wide applicable and more complex than regular ones, this poses a higher demand for quality assurance with those kinds of applications. Therefore, we formalize several aspects of INI, including its type system and operational semantics. Furthermore, we develop a tool called INICheck, which can convert a significant subset of INI to Promela, the input modeling language of the model checker SPIN. Hence, SPIN can be applied to verify properties or constraints that need to be satisfied by INI programs. Our tool allows the programmers to have insurance on their code and its behavior.

The purpose of this thesis was to design and evaluate The ContexTable, a context-aware system built into a kitchen table. After establishing the current status of the field of context-aware systems and the hurdles and problems being faced, a functioning prototype system was designed and built. The prototype makes it possible to explore established, untested theory and novel solutions to problems faced in the field.

Information Filtering and Recommender Systems have been used and has been implemented in various ways from various entities since the dawn of the Internet, and state-of-the-art approaches rely on Machine Learning and Deep Learning in order to create accurate and personalized recommendations for users in a given context. These models require big amounts of data with a variety of features such as time, location and user data in order to find correlations and patterns that other classical models such as matrix factorization and collaborative filtering cannot. This thesis researches, implements and compares a variety of models with the primary focus of Machine Learning and Deep Learning for the task of music recommendation and do so successfully by representing the task of recommendation as a multi-class extreme classification task with 100 000 distinct labels. By comparing fourteen different experiments, all implemented models successfully learn features such as time, location, user features and previous listening history in order to create context-aware personalized music predictions, and solves the cold start problem by using user demographic information, where the best model being capable of capturing the intended label in its top 100 list of recommended items for more than 1/3 of the unseen data in an offine evaluation, when evaluating on randomly selected examples from the unseen following week. / Informationsfiltrering och rekommendationssystem har använts och implementeratspå flera olika sätt från olika enheter sedan gryningen avInternet, och moderna tillvägagångssätt beror påMaskininlärrning samtDjupinlärningför att kunna skapa precisa och personliga rekommendationerför användare i en given kontext. Dessa modeller kräver data i storamängder med en varians av kännetecken såsom tid, plats och användardataför att kunna hitta korrelationer samt mönster som klassiska modellersåsom matris faktorisering samt samverkande filtrering inte kan. Dettaexamensarbete forskar, implementerar och jämför en mängd av modellermed fokus påMaskininlärning samt Djupinlärning för musikrekommendationoch gör det med succé genom att representera rekommendationsproblemetsom ett extremt multi-klass klassifikationsproblem med 100000 unika klasser att välja utav. Genom att jämföra fjorton olika experiment,så lär alla modeller sig kännetäcken såsomtid, plats, användarkänneteckenoch lyssningshistorik för att kunna skapa kontextberoendepersonaliserade musikprediktioner, och löser kallstartsproblemet genomanvändning av användares demografiska kännetäcken, där den bästa modellenklarar av att fånga målklassen i sin rekommendationslista medlängd 100 för mer än 1/3 av det osedda datat under en offline evaluering,när slumpmässigt valda exempel från den osedda kommande veckanevalueras.

Modern Web or ¿Eternal-Beta¿ applications necessitate a flexible and easy-to-use data management platform that allows the evolutionary development of databases and applications. The classical approach of relational database systems following strictly the ACID properties has to be extended by an extensible and easy-to-use persistency layer with specialized DB features. Using the underlying concept of Software as a Service (SaaS) also enables an economic advantage based on the ¿economy of the scale¿, where application and system environments only need to be provided once but can be used by thousands of users. Within this tutorial, we are looking at the current state-of-the-art from different perspectives. We outline foundations and techniques to build database services based on the SaaS-paradigm. We discuss requirements from a programming perspective, show different dimensions in the context of consistency and reliability, and also describe different non-functional properties under the umbrella of Service-Level agreements (SLA).