The main contributions of this thesis revolve around development of an integrated conversational recommendation system, combining data and information models with community network and interactions to leverage multi-modal information access. We have developed a real time conversational information access community agent that leverages community knowledge by pushing relevant recommendations to users of the community. The recommendations are delivered in the form of web resources, past conversation and people to connect to. The information agent (cobot, for community/ collaborative bot) monitors the community conversations, and is 'aware' of users' preferences by implicitly capturing their short term and long term knowledge models from conversations. The agent leverages from health and medical domain knowledge to extract concepts, associations and relationships between concepts; formulates queries for semantic search and provides socio-semantic recommendations in the conversation after applying various relevance filters to the candidate results. The agent also takes into account users' verbal intentions in conversations while making recommendation decision. One of the goals of this thesis is to develop an innovative approach to delivering relevant information using a combination of social networking, information aggregation, semantic search and recommendation techniques. The idea is to facilitate timely and relevant social information access by mixing past community specific conversational knowledge and web information access to recommend and connect users with relevant information. Language and interaction creates usable memories, useful for making decisions about what actions to take and what information to retain. Cobot leverages these interactions to maintain users' episodic and long term semantic models. The agent analyzes these memory structures to match and recommend users in conversations by matching with the contextual information need. The social feedback on the recommendations is registered in the system for the algorithms to promote community preferred, contextually relevant resources. The nodes of the semantic memory are frequent concepts extracted from user's interactions. The concepts are connected with associations that develop when concepts co-occur frequently. Over a period of time when the user participates in more interactions, new concepts are added to the semantic memory. Different conversational facets are matched with episodic memories and a spreading activation search on the semantic net is performed for generating the top candidate user recommendations for the conversation. The tying themes in this thesis revolve around informational and social aspects of a unified information access architecture that integrates semantic extraction and indexing with user modeling and recommendations.

Community based service-learning is one of the fastest growing reforms in higher education, especially in the field of health care. The increased interest in this phenomenon is based on the demands by government and society that higher education institutions should be more responsive to the needs of the community. Literature, however, reflects that service learning lacks a sound theoretical base to guide teaching and learning due to limited research in this area. This study was, therefore, aimed at exploring the phenomenon knowledge construction in basic nursing programmes in selected South African nursing schools with the intention to generate a middle range theory that may be used to guide the process of knowledge construction in community-based service-learning programmes. This study adopted a qualitative approach and a grounded theory research design by Strauss and Corbin. Two university-based schools of nursing were purposively selected to participate in the study. There were a total number of 16 participants. The collection of data was intensified by the use of multiple sources of data (participant observation, documents analysis and in-depth structured interviews). The data analysis process entailed three phases; open, axial and selective coding. The results of the study revealed that the phenomenon “knowledge construction” is conceptualised as having specific core characteristics, which include the use of authentic health-related problems, academic coaching through scaffolding, academic discourse-dialogue and communities of learners. The findings showed that there are a number of antecedent conditions and contextual circumstances contributing to how knowledge is constructed in a community based service learning programme. The process of knowledge construction emerged as cyclical in nature, with students, facilitators and community members having specific roles to play in the process. A number of intervening variables were identified that had an influence on the expected outcomes on knowledge construction in community based service learning programmes. These findings led to the generation of a conceptual model. Knowledge construction according to this model takes place in an environment which is characterised by interactive learning, collaborative learning, actively learning and inquiry-based learning through continuous reflective learning processes. The main concepts in this conceptual model include concrete learning experiences, continuous reflection, problem posing, problem analysis, knowledge deconstruction and knowledge generation, knowledge verification, knowledge generation, testing of generated knowledge and evaluation of generated knowledge. The sub-concepts include learning through senses, an initial situation, health-related triggers, social interaction, reflection-in action, reflection-on action, hypotheses generation, conceptualisation of learning experiences, information validation and community interventions. Recommendations were categorised into education and training of academic staff, application of the model and further research with regard to quality assurance in CBSL programmes as well as the use of other research designs for similar studies. / Thesis (Ph.D.)-University of KwaZulu-Natal, Durban, 2011.

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal

El test de circuits és una fase del procés de producció que cada vegada pren més importància quan es desenvolupa un nou producte. Les tècniques de test i diagnosi per a circuits digitals han estat desenvolupades i automatitzades amb èxit, mentre que aquest no és encara el cas dels circuits analògics. D'entre tots els mètodes proposats per diagnosticar circuits analògics els més utilitzats són els diccionaris de falles. En aquesta tesi se'n descriuen alguns, tot analitzant-ne els seus avantatges i inconvenients.Durant aquests últims anys, les tècniques d'Intel·ligència Artificial han esdevingut un dels camps de recerca més importants per a la diagnosi de falles. Aquesta tesi desenvolupa dues d'aquestes tècniques per tal de cobrir algunes de les mancances que presenten els diccionaris de falles. La primera proposta es basa en construir un sistema fuzzy com a eina per identificar. Els resultats obtinguts son força bons, ja que s'aconsegueix localitzar la falla en un elevat tant percent dels casos. Per altra banda, el percentatge d'encerts no és prou bo quan a més a més s'intenta esbrinar la desviació.Com que els diccionaris de falles es poden veure com una aproximació simplificada al Raonament Basat en Casos (CBR), la segona proposta fa una extensió dels diccionaris de falles cap a un sistema CBR. El propòsit no és donar una solució general del problema sinó contribuir amb una nova metodologia. Aquesta consisteix en millorar la diagnosis dels diccionaris de falles mitjançant l'addició i l'adaptació dels nous casos per tal d'esdevenir un sistema de Raonament Basat en Casos. Es descriu l'estructura de la base de casos així com les tasques d'extracció, de reutilització, de revisió i de retenció, fent èmfasi al procés d'aprenentatge.En el transcurs del text s'utilitzen diversos circuits per mostrar exemples dels mètodes de test descrits, però en particular el filtre biquadràtic és l'utilitzat per provar les metodologies plantejades, ja que és un dels benchmarks proposats en el context dels circuits analògics. Les falles considerades son paramètriques, permanents, independents i simples, encara que la metodologia pot ser fàcilment extrapolable per a la diagnosi de falles múltiples i catastròfiques. El mètode es centra en el test dels components passius, encara que també es podria extendre per a falles en els actius. / Testing circuits is a stage of the production process that is becoming more and more important when a new product is developed. Test and diagnosis techniques for digital circuits have been successfully developed and automated. But, this is not yet the case for analog circuits. Even though there are plenty of methods proposed for diagnosing analog electronic circuits, the most popular are the fault dictionary techniques. In this thesis some of these methods, showing their advantages and drawbacks, are analyzed.During these last decades automating fault diagnosis using Artificial Intelligence techniques has become an important research field. This thesis develops two of these techniques in order to fill in some gaps in fault dictionaries techniques. The first proposal is to build a fuzzy system as an identification tool. The results obtained are quite good, since the faulty component is located in a high percentage of the given cases. On the other hand, the percentage of successes when determining the component's exact deviation is far from being good.As fault dictionaries can be seen as a simplified approach to Case-Based Reasoning, the second proposal extends the fault dictionary towards a Case Based Reasoning system. The purpose isnot to give a general solution, but to contribute with a new methodology. This second proposal improves a fault dictionary diagnosis by means of adding and adapting new cases to develop aCase Based Reasoning system. The case base memory, retrieval, reuse, revise and retain tasks are described. Special attention to the learning process is taken.Several circuits are used to show examples of the test methods described throughout the text. But, in particular, the biquadratic filter is used to test the proposed methodology because it isdefined as one of the benchmarks in the analog electronic diagnosis domain. The faults considered are parametric, permanent, independent and simple, although the methodology can be extrapolated to catastrophic and multiple fault diagnosis. The method is only focused and tested on passive faulty components, but it can be extended to cover active devices as well.

Les médecins cherchent à résoudre les problèmes de santé posés par des individus. Une solution individualisée tient compte de la singularité du patient concerné. L'individualisation des pratiques est-elle possible et souhaitable? Le cas échéant, selon quelles modalités peut-elle ou doit-elle être réalisée'? La première partie de la thèse vise à montrer: que la question se pose depuis les premières théories de la décision médicale (Hippocrate) ; qu'elle s'est posée de façon aiguë au début du XIX" siècle, avec l'apparition des études statistiques; et que l'observation médicale et son évolution concrétisent la façon dont la documentation des cas et leur individualisation interagissent. La deuxième partie reprend la question dans le contexte contemporain, à travers la naissance de l'"evidence-based medicine", ses critiques et son évolution. La troisième partie montre que l'articulation du raisonnement par règles et du raisonnement par cas modélise de façon opérationnelle une démarche raisonnée d'individualisation des décisions médicales. Ce modèle simple permet de rendre compte du mouvement d'aller-retour entre deux conceptions de l'individualisation et d'en proposer une version équilibrée, mise à l'épreuve dans les domaines de l'évaluation des pratiques et de la littérature médicale. / Physicians try to solve health problems of individual patients. Customized solutions take into account the uniqueness of the patient. Is the individualization of medical decisions possible and desirable'? If so, how can I tor should it be performed? The first part of the thesis shows: that the question arises since the first conceptualizations of medical reasoning (Hippocrates); that is was much debated in the early nineteenth century, when statistical studies were first performed to guide medical decisions; and that the medical observation and its evolution materialize how case documentation and management interact. The second part addresses the issue in the current context, from the birth of evidence-based medicine, its cri tics and its evolution. The third part shows that linking rule-based and case-based reasoning adequately pictures the process of customizing medical decisions. This simple model can account for the movement between two kinds of customization and leads to a balanced approach, tested in the field of practice evaluation and medical literature.

This paper aims to provide an Interactive Learning Environment using the computer to support aid in the diagnosis and treatment of patients with neurological disorders. The study proposes an architecture which facilitates the activities of students in the health area, in decision making, for the advice of physiotherapy for stroke patients. It was used the approach of Case-Based Reasoning (CBR) that has, like general idea, the use of past experiences to the solution of new problems. This work focused on the stages of indexing, representation and retrieval of cases, with the use of metrics, similar characteristics as the Count Features and Tversky s Contrast Model. A prototype was built for the validation of these metrics, proving the efficiency in the recovery of the cases on the basis of cases / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Este trabalho tem como objetivo apresentar um Ambiente Interativo de Aprendizagem utilizando o Computador de apoio ao diagnóstico e auxílio no tratamento de pacientes que apresentam disfunções neurológicas. A pesquisa propõe uma arquitetura que facilite as atividades dos estudantes da área da saúde, na tomada de decisão, para o aconselhamento fisioterápico dos pacientes de acidente vascular cerebral. Utilizou-se a abordagem de Raciocínio Baseado em Casos (RBC) que tem como idéia geral a utilização de experiências passadas para a solução de novos problemas. Este trabalho se concentrou nas fases de indexação, representação e recuperação dos casos, com a utilização de métricas de similaridade como a Contagem de Características e a Regra do Contraste de Tversky. Um protótipo foi construído para a validação dessas métricas, provando a eficiência na recuperação dos casos na base de casos

This work aims to propose a recommendation model of prescription items for Pediatric Oncology based on data extraction from Electronic Patient Record. These data are used as indexed cases to aid providers of medical service based on the similarity of prescriptions, according to the patient's history. From the viewpoint of aid medical education, the modeling objective support the student or health professional in understanding the decision-making process during the prescription items oncological medical treatment, for example drugs, laboratory exams or images exams, diet, gases, care, chemotherapy, radiotherapy. To develop the model, was used the approach of Case Based Reasoning (CBR), through the representation of prescription medical base-case, indexed by their treatment items. During the recovery phase of cases, we used the tool. Data Mining by applying the model of association rule, together with the algorithm "apriori" for obtaining the similarity between cases. To update the case base, a procedure database for performing the process of Extraction, Transformation and Load of the database was developed. The developed model was applied in the database of the Electronic Patient Record of the Santa Casa de Misericordia de Maceio, based on Hospital Management Systems “MV Sistemas”, deployed in the unit since 2005. For the presentation of results, was used the Oracle Data Miner tool, which allowed access to the database and analysis of selected cases by identifying key words contained in the evolution of the clinical condition of the patient. The application of the experiments validate the occurrence of allowed combined application of items of treatment according to the keywords, which can be used as input in the process of making medical decision and tutoring. / Este trabalho tem como objetivo propor um Modelo de Recomendação de itens de prescrição para Oncologia Pediátrica baseado na extração de dados do Prontuário Eletrônico do Paciente. Esses dados são utilizados como casos indexados, para auxiliar os prestadores de serviço médico baseados na similaridade de prescrições, de acordo com o histórico do paciente. Do ponto de vista do apoio a educação médica, a modelagem objetiva apoiar o estudante ou o profissional de saúde no entendimento do processo de tomada de decisão durante a fase de prescrição de itens de tratamento médico oncológico, como, por exemplo: medicamentos, exames de laboratório ou de imagens, dieta, gases, cuidados, quimioterapia, radioterapia. Para o desenvolvimento do modelo, utilizou-se a abordagem de Raciocínio Baseado em Casos (RBC), através da representação de uma base de casos de prescrição médica, indexada por seus itens de tratamento. Durante a fase de recuperação de casos, utilizou-se a ferramenta de Mineração de Dados aplicando-se o modelo de regra de associação, em conjunto com o algoritmo “apriori” visando a obtenção da similaridade entre casos. Para a atualização da base de casos, foi desenvolvido um procedimento de banco de dados para execução do processo de Extração, Transformação e Carga da base de dados. O modelo desenvolvido foi aplicado na base de dados do Prontuário Eletrônico do Paciente da Santa Casa de Misericórdia de Maceió, baseado no sistema de gestão hospitalar MV Sistemas, implantado na unidade desde 2005. Para a apresentação dos resultados, utilizou-se a ferramenta Oracle Data Miner, que possibilitou o acesso ao banco de dados e a análise dos casos selecionados pela identificação de palavras chaves contidas na evolução do estado clínico do paciente. A aplicação dos experimentos permitiu validar a ocorrência de aplicação conjunta de itens de tratamento de acordo com as palavras chaves, o que pode ser utilizado como elemento para o processo de tomada de decisão médica e tutoria.

Dans ce travail, une étude est réalisée sur le développement d'une méthodologie permettant la génération et l'évaluation de nouvelles trajectoires de valorisation pour des déchets. Ainsi, pour répondre à cette problématique, trois sous problèmes ont été identifiés. Le premier concerne un cadre de modélisation permettant la représentation structurée et homogène de chaque trajectoire, ainsi que les indicateurs choisis pour l'évaluation de ces dernières, permettant une sélection ultérieure. Le deuxième se concentre sur le développement d'une méthodologie puis la réalisation d'un outil permettant la génération de nouvelles trajectoires en s'appuyant sur d'autres connues. Enfin, le dernier sous problème concerne le développement d'un second outil développé pour modéliser et estimer les trajectoires générées. La partie de création d'un cadre de modélisation cherche à concevoir des structures globales qui permettent la catégorisation des opérations unitaires sous plusieurs niveaux. Trois niveaux de décomposition ont été identifiés. La Configuration générique de plus haut niveau, qui décrit la trajectoire sous de grandes étapes de modélisation. Le second niveau, Traitement générique propose des ensembles de structures génériques de traitement qui apparaissent régulièrement dans les trajectoires de valorisation. Enfin, le plus bas niveau se focalise sur la modélisation des opérations unitaires. Un second cadre a été créé, plus conceptuel et comportant deux éléments : les blocs et les systèmes. Ces cadres sont ensuite accompagnés par un ensemble d'indicateurs choisis à cet effet. Dans une volonté d'approche de développement durable, un indicateur est sélectionné pour chacune de des composantes : économique, environnemental et social. Dans notre étude, l'impact social se limite à l'estimation du nombre d'emplois créés. Afin de calculer cet indicateur, une nouvelle approche se basant sur les résultats économiques d'une entreprise a été proposée et validée.L'outil de génération de nouvelles trajectoires s'appuie sur l'utilisation de la connaissance en utilisant un système de raisonnement à partir de cas (RàPC). Pour être adapté à notre problématique, la mise en œuvre de ce dernier a impliqué la levée de plusieurs points délicats. Tout d'abord, la structuration des données et plus largement la génération de cas sources sont réalisées par un système basé sur des réseaux sémantiques et l'utilisation de mécanismes d'inférences. Le développement d'une nouvelle méthode de mesure de similarité est réalisé en introduisant la notion de définition commune qui permet de lier les états, qui sont des descriptions de situations, à des états représentant des définitions générales d'un ensemble d'états. Ces définitions communes permettent la création d'ensembles d'états sous différents niveaux d'abstraction et de conceptualisation. Enfin, un processus de décompositions des trajectoires est réalisé afin de résoudre un problème grâce à la résolution de ses sous-problèmes associés. Cette décomposition facilite l'adaptation des trajectoires et l'estimation des résultats des transformations. Basé sur cette méthode, un outil a été développé en programmation logique, sous Prolog. La modélisation et l'évaluation des voies de valorisation se fait grâce à la création d'outil spécifique. Cet outil utilise la méta-programmation permettant la réalisation dynamique de modèle de structure. Le comportement de ces structures est régi par la définition de contraintes sur les différents flux circulants dans l'ensemble de la trajectoire. Lors de la modélisation de la trajectoire, ces contraintes sont converties par un parser permettant la réalisation d'un modèle de programmation par contraintes cohérent. Ce dernier peut ensuite être résolu grâce à des solveurs via une interface développée et intégrée au système. De même, plusieurs greffons ont été réalisés pour analyser et évaluer les trajectoires à l'aide des critères retenus. / In this work, a study is realised about the creation of a new methodology allowing the generation and the assessment of new waste recovery processes. Three elements are proposed for that. The first one is the creation of a modelling framework permitting a structured and homogeneous representation of each recovery process and the criteria used to asses them. The second one is a system and a tool generating new recovery processes from others known. Finally, the last element is another tool to model, to estimate and to asses the generated processes. The creation of a modelling framework tries to create some categories of elements allowing the structuring of unit operations under different levels of description. Three levels have been identified. In the higher level, the Generic operation which describes global structure of operations. The second one is Generic treatment which is an intermediate level between the two others. It proposes here too categories of operations but more detailed than the higher level. The last one is the Unit operation. A second framework has been created. It is more conceptual and it has two components : blocs and systems. These frameworks are used with a set of selected indicators. In a desire of integrating our work in a sustainable development approach, an indicator has been chosen for each of its components: economical, environmental and social. In our study, the social impact is limited to the number of created jobs. To estimate this indicator, we proposed a new method based on economical values of a company. The tool for the generation of new waste recovery processes used the methodology of case-based reasoning CBR which is based on the knowledge management. Some difficult points are treated here to adapt the CBR to our problem. The structuring of knowledge and generally the source case generation is realised by a system based on connections between data and the use of inference mechanisms. The development of a new method for the similarity measure is designed with the introduction of common definition concept which allows linking states, simply put description of objects, to other states under different levels of conceptualizations and abstractions. This point permits creating many levels of description. Finally, recovery process is decomposed from a main problem to some sub-problems. This decomposition is a part of the adaptation mechanism of the selected source case. The realisation of this system is under logic programming with Prolog. This last one permits the use of rules allowing inferences and the backtracking system allowing the exploration to the different possible solution. The modelling and assessment of recovery processes are done by a tool programmed in Python. It uses the meta-programming to dynamically create model of operations or systems. Constraint rules define the behaviour of these models allowing controlling the flux circulating in each one. In the evaluation step, a parser is used to convert theses rules into a homogeneous system of constraint programming. This system can be solved by the use of solvers with an interface developed for that and added to the tool. Therefore, it is possible for the user to add solvers but also to add plug-ins. This plug-ins can make the assessment of the activity allowing to have different kinds of evaluation for the same criteria. Three plug-ins are developed, one for each selected criterion. These two methods are tested to permit the evaluation of the proposed model and to check the behaviour of them and their limits . For these tests, a case-base on waste has been created Finally, for the modelling and assessment tool, a study case about the recovery process of used tyres in new raw material is done.

Submitted in fulfillment of the requirements for the Degree in Masters of Technology in Nursing, Durban University of Technology, Durban, South Africa, 2015. / Background The current health care system in South Africa and its diverse settings of health care delivery system require a nurse who can make decisions, think critically, solve problems and work effectively in a team. Traditional nursing education teaching strategies have over the years relied on didactic and often passive approaches to learning. In pursuit of quality, academics and students must be continually engaged in a process of finding opportunities for improving the teaching and learning process. Purpose of the study The purpose of this study was to evaluate the structure and the process in case based learning at the University of Technology. Methodology This study is qualitative in nature, governed by an interpretive paradigm. This is a case study, which enabled the researcher to merge student interview data with records in order to gain insight into the activities and details of case based learning as practised at the University of Technology under study. Most importantly, the case study method was deemed appropriate for the current study, since case-based learning as a pedagogical approach (and a case) cannot be abstracted from its context for the purposes of study. Case based learning is evaluated in its context namely, the undergraduate nursing programme, using the Donabedian framework of structure, process and product. Results The study recorded that students were positive towards case based learning though some identified dynamics of working in groups as demerits of case based learning. The structures that are in place in the programme and the CBL processes are adequate and support CBL. There are however areas that need attention such as the qualification of the programme coordinator, the size of the class-rooms and the service of the computer laboratory. Conclusion The study found that apart from a few minor discrepancies, case based learning is sufficiently implemented, and experienced as invaluable by students, at the University of Technology under study.

[ES] Recientemente, hay una fuerte y creciente influencia de aplicaciones en línea en nuestro día a día. Más concretamente las redes sociales se cuentan entre las plataformas en línea más usadas, que permiten a usuarios comunicarse e interactuar desde diferentes partes del mundo todos los días. Dado que estas interacciones conllevan diferentes riesgos, y además los adolescentes tienen características que los hacen más vulnerables a ciertos riesgos, es deseable que el sistema pueda guiar a los usuarios cuando se encuentren interactuando en línea, para intentar mitigar la probabilidad de que caigan en uno de estos riesgos. Esto conduce a una experiencia en línea más segura y satisfactoria para usuarios de este tipo de plataformas. El interés en aplicaciones de inteligencia artificial capaces de realizar análisis de sentimientos ha crecido recientemente. Los usos de la detección automática de sentimiento de usuarios en plataformas en línea son variados y útiles. Se pueden usar polaridades de sentimiento para realizar minería de opiniones en personas o productos, y así descubrir las inclinaciones y opiniones de usuarios acerca de ciertos productos (o ciertas características de ellos), para ayudar en campañas de marketing, y también opiniones acerca de personas como políticos, para descubrir la intención de voto en un periodo electoral, por ejemplo. En esta tesis, se presenta un Sistema Multi-Agente (SMA), el cual integra agentes que realizan diferentes análisis de sentimientos y de estrés usando texto y dinámicas de escritura (usando análisis unimodal y multimodal), y utiliza la respuesta de los analizadores para generar retroalimentación para los usuarios y potencialmente evitar que caigan en riesgos y difundan comentarios en plataformas sociales en línea que pudieran difundir polaridades de sentimiento negativas o niveles altos de estrés. El SMA implementa un análisis en paralelo de diferentes tipos de datos y generación de retroalimentación a través del uso de dos mecanismos diferentes. El primer mecanismo se trata de un agente que realiza generación de retroalimentación y guiado de usuarios basándose en un conjunto de reglas y la salida de los analizadores. El segundo mecanismo es un módulo de Razonamiento Basado en Casos (CBR) que usa no solo la salida de los analizadores en los mensajes del usuario interactuando para predecir si su interacción puede generar una futura repercusión negativa, sino también información de contexto de interacciones de usuarios como son los tópicos sobre los que hablan o información sobre predicciones previas en mensajes escritos por la gente que conforma la audiencia del usuario. Se han llevado a cabo experimentos con datos de una red social privada generada en laboratorio con gente real usando el sistema en tiempo real, y también con datos de Twitter.com para descubrir cuál es la eficacia de los diferentes analizadores implementados y del módulo CBR al detectar estados del usuario que se propagan más en la red social. Esto conlleva descubrir cuál de las técnicas puede prevenir mejor riesgos potenciales que los usuarios pueden sufrir cuando interactúan, y en qué casos. Se han encontrado diferencias estadísticamente significativas y la versión final del SMA incorpora los analizadores que mejores resultados obtuvieron, un agente asesor o guía basado en reglas y un módulo CBR. El trabajo de esta tesis pretende ayudar a futuros desarrolladores de sistemas inteligentes a crear sistemas que puedan detectar el estado de los usuarios interactuando en sitios en línea y prevenir riesgos que los usuarios pudiesen enfrentar. Esto propiciaría una experiencia de usuario más segura y satisfactoria. / [CA] Recentment, hi ha una forta i creixent influència d'aplicacions en línia en el nostre dia a dia, i concretament les xarxes socials es compten entre les plataformes en línia més utilitzades, que permeten a usuaris comunicar-se i interactuar des de diferents parts del món cada dia. Donat que aquestes interaccions comporten diferents riscos, i a més els adolescents tenen característiques que els fan més vulnerables a certs riscos, seria desitjable que el sistema poguera guiar als usuaris mentre es troben interactuant en línia, per així poder mitigar la probabilitat de caure en un d'aquests riscos. Açò comporta una experiència en línia més segura i satisfactòria per a usuaris d'aquest tipus de plataformes. L'interés en aplicacions d'intel·ligència artificial capaces de realitzar anàlisi de sentiments ha crescut recentment. Els usos de la detecció automàtica de sentiments en usuaris en plataformes en línia són variats i útils. Es poden utilitzar polaritats de sentiment per a realitzar mineria d'opinions en persones o productes, i així descobrir les inclinacions i opinions d'usuaris sobre certs productes (o certes característiques d'ells), per a ajudar en campanyes de màrqueting, i també opinions sobre persones com polítics, per a descobrir la intenció de vot en un període electoral, per exemple. En aquesta tesi, es presenta un Sistema Multi-Agent (SMA), que integra agents que implementen diferents anàlisis de sentiments i d'estrés utilitzant text i dinàmica d'escriptura (utilitzant anàlisi unimodal i multimodal), i utilitza la resposta dels analitzadors per a generar retroalimentació per als usuaris i potencialment evitar que caiguen en riscos i difonguen comentaris en plataformes socials en línia que pogueren difondre polaritats de sentiment negatives o nivells alts d'estrés. El SMA implementa una anàlisi en paral·lel de diferents tipus de dades i generació de retroalimentació a través de l'ús de dos mecanismes diferents. El primer mecanisme es tracta d'un agent que realitza generació de retroalimentació i guia d'usuaris basant-se en un conjunt de regles i l'eixida dels analitzadors. El segon mecanisme és un mòdul de Raonament Basat en Casos (CBR) que utilitza no solament l'eixida dels analitzadors en els missatges de l'usuari per a predir si la seua interacció pot generar una futura repercussió negativa, sinó també informació de context d'interaccions d'usuaris, com són els tòpics sobre els quals es parla o informació sobre prediccions prèvies en missatges escrits per la gent que forma part de l'audiència de l'usuari. S'han realitzat experiments amb dades d'una xarxa social privada generada al laboratori amb gent real utilitzant el sistema implementat en temps real, i també amb dades de Twitter.com per a descobrir quina és l'eficàcia dels diferents analitzadors implementats i del mòdul CBR en detectar estats de l'usuari que es propaguen més a la xarxa social. Açò comporta descobrir quina de les tècniques millor pot prevenir riscos potencials que els usuaris poden sofrir quan interactuen, i en quins casos. S'han trobat diferències estadísticament significatives i la versió final del SMA incorpora els analitzadors que millors resultats obtingueren, un agent assessor o guia basat en regles i un mòdul CBR. El treball d'aquesta tesi pretén ajudar a futurs dissenyadors de sistemes intel·ligents a crear sistemes que puguen detectar l'estat dels usuaris interactuant en llocs en línia i prevenir riscos que els usuaris poguessen enfrontar. Açò propiciaria una experiència d'usuari més segura i satisfactòria. / [EN] In the present days, there is a strong and growing influence of on-line applications in our daily lives, and concretely Social Network Sites (SNSs) are one of the most used on-line social platforms that allow users to communicate and interact from different parts of the world every day. Since this interaction poses several risks, and also teenagers have characteristics that make them more vulnerable to certain risks, it is desirable that the system could be able to guide users when interacting on-line, to try and mitigate the probability of incurring one of those risks. This would in the end lead to a more satisfactory and safe experience for the users of such on-line platforms. Recently, interest in artificial intelligence applications being able to perform sentiment analysis has risen. The uses of detecting the sentiment of users in on-line platforms or sites are variated and rewarding. Sentiment polarities can be used to perform opinion mining on people or products, and discover the inclinations and opinions of users on certain products (or certain features of them) to help marketing campaigns, and also on people such as politics, to discover the voting intention for example in electoral periods. In this thesis, a Multi-Agent System (MAS) is presented, which integrates agents that perform different sentiment and stress analyses using text and keystroke dynamics data (using both unimodal and multi-modal analysis). The MAS uses the output of the analyzers for generating feedback for users and potentially avoids them from incurring risks and spreading comments in on-line social platforms that could lead to the spread of negative sentiment or high-stress levels. Moreover, the MAS incorporates parallelized analyses of different data types and feedback generation via the use of two different mechanisms. On the one hand, a rule-based advisor agent has been implemented, that generates feedback or guiding for users based on the output of the analyzers and a set of rules. On the other hand, a Case-Based Reasoning (CBR) module that uses not only the output of the different analyzers on the messages of the user interacting, but also context information from user interactions such as the topics being talked about or information about the previous states detected on messages written by people in the audience of the user. Experiments with data from a private SNS generated in a laboratory with real people using the system in real-time, and also with data from Twitter.com have been performed to ascertain the efficacy of the different analyzers implemented and the CBR module on detecting states of the user that propagate more in the network, which leads to discovering which of the techniques is able to better prevent potential risks that users could face when interacting, and in which cases. Significant differences were found and the final version of the MAS incorporates the best-performing analyzer agents, a rule-based advisor agent, and a CBR module. In the end, this thesis aims to help intelligent systems developers to build systems that are able to detect the state of users interacting in on-line sites and prevent risks that they could face, leading to a more satisfactory and safe user experience. / This thesis was funded by the following research projects: Privacy in Social Educational Environments during Child-hood and Adolescence (PESEDIA), Ministerio de Economia y Empresa (TIN2014-55206-R) and Intelligent Agents for Privacy Advice in Social Networks (AI4PRI), Ministerio de Economia y Empresa (TIN2017-89156-R) / Aguado Sarrió, G. (2021). MAS-based affective state analysis for user guiding in on-line social environments [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/164902 / TESIS