Дисертації з теми "Recommandation de données d'apprentissage"
Оформте джерело за APA, MLA, Chicago, Harvard та іншими стилями
Ознайомтеся з топ-50 дисертацій для дослідження на тему "Recommandation de données d'apprentissage".
Біля кожної праці в переліку літератури доступна кнопка «Додати до бібліографії». Скористайтеся нею – і ми автоматично оформимо бібліографічне посилання на обрану працю в потрібному вам стилі цитування: APA, MLA, «Гарвард», «Чикаго», «Ванкувер» тощо.
Також ви можете завантажити повний текст наукової публікації у форматі «.pdf» та прочитати онлайн анотацію до роботи, якщо відповідні параметри наявні в метаданих.
Переглядайте дисертації для різних дисциплін та оформлюйте правильно вашу бібліографію.
Joshi, Bikash. "Algorithmes d'apprentissage pour les grandes masses de données : Application à la classification multi-classes et à l'optimisation distribuée asynchrone." Thesis, Université Grenoble Alpes (ComUE), 2017. http://www.theses.fr/2017GREAM046/document.
This thesis focuses on developing scalable algorithms for large scale machine learning. In this work, we present two perspectives to handle large data. First, we consider the problem of large-scale multiclass classification. We introduce the task of multiclass classification and the challenge of classifying with a large number of classes. To alleviate these challenges, we propose an algorithm which reduces the original multiclass problem to an equivalent binary one. Based on this reduction technique, we introduce a scalable method to tackle the multiclass classification problem for very large number of classes and perform detailed theoretical and empirical analyses.In the second part, we discuss the problem of distributed machine learning. In this domain, we introduce an asynchronous framework for performing distributed optimization. We present application of the proposed asynchronous framework on two popular domains: matrix factorization for large-scale recommender systems and large-scale binary classification. In the case of matrix factorization, we perform Stochastic Gradient Descent (SGD) in an asynchronous distributed manner. Whereas, in the case of large-scale binary classification we use a variant of SGD which uses variance reduction technique, SVRG as our optimization algorithm
Labiadh, Mouna. "Méthodologie de construction de modèles adaptatifs pour la simulation énergétique des bâtiments." Thesis, Lyon, 2021. http://www.theses.fr/2021LYSE1158.
Predictive modeling of energy consumption in buildings is essential for intelligent control and efficient planning of energy networks. One way to perform predictive modeling is through machine learning approaches. Alongside their good performance, these approaches are time efficient and facilitates the integration of buildings into smart environments. However, accurate machine learning models rely heavily on collecting relevant building operational data in a sufficient amount, notably when deep learning is used. In the field of buildings energy, historical data are not available for training, such is the case in newly built or newly renovated buildings. Moreover, it is common to verify the energy efficiency of buildings before construction or renovation. For such cases, only a contextual description about the future building and its design is available. The goal of this dissertation is to address the predictive modeling tasks of building energy consumption when no historical data are available for the given target building. To that end, existing data collected from multiple different source buildings are leveraged. This is increasingly relevant with the growth of open data initiatives in various sectors, namely building energy. The main idea is to transfer knowledge across building models. There is little research at the intersection of building energy modeling and knowledge transfer. An important challenge arises when dealing with multi-source data, since large domain shift may exist between different sources and also between each source and the target. As a contribution, a two-fold query-adaptive methodology is developed for cross-building predictive modeling. The first process recommends relevant training data to a target building solely by using a minimal contextual description on it (metadata). Contextual descriptions are provided as user queries. To enable a task-specific recommendation, a deep similarity learning framework is used. The second process trains multiple predictive models based on recommended training data. These models are combined together using an ensemble learning framework to ensure a robust performance. The implementation of the proposed methodology is based on microservices. Logically independent workflows are modeled as microservices with single purposes and separate data sources. Building metadata and time series data collected from multiple sources are integrated into an unified ontology-based view. Experimental evaluation of the predictive model factory validates the effectiveness and the applicability for the use case of building energy modeling. Moreover, because of its generic design, the methodology for query-adaptive cross-domain predictive modeling can be re-used for a diverse range of use cases in different fields
Servajean, Maximilien. "Recommandation diversifiée et distribuée pour les données scientifiques." Thesis, Montpellier 2, 2014. http://www.theses.fr/2014MON20216/document.
In many fields, novel technologies employed in information acquisition and measurement (e.g. phenotyping automated greenhouses) are at the basis of a phenomenal creation of data. In particular, we focus on two real use cases: plants observations in botany and phenotyping data in biology. Our contributions can be, however, generalized to Web data. In addition to their huge volume, data are also distributed. Indeed, each user stores their data in many heterogeneous sites (e.g. personal computers, servers, cloud); yet he wants to be able to share them. In both use cases, collaborative solutions, including distributed search and recommendation techniques, could benefit to the user.Thus, the global objective of this work is to define a set of techniques enabling sharing and discovery of data in heterogeneous distributed environment, through the use of search and recommendation approaches.For this purpose, search and recommendation allow users to be presented sets of results, or recommendations, that are both relevant to the queries submitted by the users and with respect to their profiles. Diversification techniques allow users to receive results with better novelty while avoiding redundant and repetitive content. By introducing a distance between each result presented to the user, diversity enables to return a broader set of relevant items.However, few works exploit profile diversity, which takes into account the users that share each item. In this work, we show that in some scenarios, considering profile diversity enables a consequent increase in results quality: surveys show that in more than 75% of the cases, users would prefer profile diversity to content diversity.Additionally, in order to address the problems related to data distribution among heterogeneous sites, two approaches are possible. First, P2P networks aim at establishing links between peers (nodes of the network): creating in this way an overlay network, where peers directly connected to a given peer p are known as his neighbors. This overlay is used to process queries submitted by each peer. However, in state of the art solutions, the redundancy of the peers in the various neighborhoods limits the capacity of the system to retrieve relevant items on the network, given the queries submitted by the users. In this work, we show that introducing diversity in the computation of the neighborhood, by increasing the coverage, enables a huge gain in terms of quality. By taking into account diversity, each peer in a given neighborhood has indeed, a higher probability to return different results given a keywords query compared to the other peers in the neighborhood. Whenever a query is submitted by a peer, our approach can retrieve up to three times more relevant items than state of the art solutions.The second category of approaches is called multi-site. Generally, in state of the art multi-sites solutions, the sites are homogeneous and consist in big data centers. In our context, we propose an approach enabling sharing among heterogeneous sites, such as small research teams servers, personal computers or big sites in the cloud. A prototype regrouping all contributions have been developed, with two versions addressing each of the use cases considered in this thesis
Collobert, Ronan. "Algorithmes d'Apprentissage pour grandes bases de données." Paris 6, 2004. http://www.theses.fr/2004PA066063.
Pradel, Bruno. "Evaluation des systèmes de recommandation à partir d'historiques de données." Paris 6, 2013. http://www.theses.fr/2013PA066263.
This thesis presents various experimental protocols leading to abetter offline estimation of errors in recommender systems. As a first contribution, results form a case study of a recommendersystem based on purchased data will be presented. Recommending itemsis a complex task that has been mainly studied considering solelyratings data. In this study, we put the stress on predicting thepurchase a customer will make rather than the rating he will assign toan item. While ratings data are not available for many industries andpurchases data widely used, very few studies considered purchasesdata. In that setting, we compare the performances of variouscollaborative filtering models from the litterature. We notably showthat some changes the training and testing phases, and theintroduction of contextual information lead to major changes of therelative perfomances of algorithms. The following contributions will focus on the study of ratings data. Asecond contribution will present our participation to the Challenge onContext-Aware Movie Recommendation. This challenge provides two majorchanges in the standard ratings prediction protocol: models areevaluated conisdering ratings metrics and tested on two specificsperiod of the year: Christmas and Oscars. We provides personnalizedrecommendation modeling the short-term evolution of the popularitiesof movies. Finally, we study the impact of the observation process of ratings onranking evaluation metrics. Users choose the items they want to rateand, as a result, ratings on items are not observed at random. First,some items receive a lot more ratings than others and secondly, highratings are more likely to be oberved than poor ones because usersmainly rate the items they likes. We propose a formal analysis ofthese effects on evaluation metrics and experiments on the Yahoo!Musicdataset, gathering standard and randomly collected ratings. We showthat considering missing ratings as negative during training phaseleads to good performances on the TopK task, but these performancescan be misleading favoring methods modeling the popularities of itemsmore than the real tastes of users
Ben, Ellefi Mohamed. "La recommandation des jeux de données basée sur le profilage pour le liage des données RDF." Thesis, Montpellier, 2016. http://www.theses.fr/2016MONTT276/document.
With the emergence of the Web of Data, most notably Linked Open Data (LOD), an abundance of data has become available on the web. However, LOD datasets and their inherent subgraphs vary heavily with respect to their size, topic and domain coverage, the schemas and their data dynamicity (respectively schemas and metadata) over the time. To this extent, identifying suitable datasets, which meet specific criteria, has become an increasingly important, yet challenging task to supportissues such as entity retrieval or semantic search and data linking. Particularlywith respect to the interlinking issue, the current topology of the LOD cloud underlines the need for practical and efficient means to recommend suitable datasets: currently, only well-known reference graphs such as DBpedia (the most obvious target), YAGO or Freebase show a high amount of in-links, while there exists a long tail of potentially suitable yet under-recognized datasets. This problem is due to the semantic web tradition in dealing with "finding candidate datasets to link to", where data publishers are used to identify target datasets for interlinking.While an understanding of the nature of the content of specific datasets is a crucial prerequisite for the mentioned issues, we adopt in this dissertation the notion of "dataset profile" - a set of features that describe a dataset and allow the comparison of different datasets with regard to their represented characteristics. Our first research direction was to implement a collaborative filtering-like dataset recommendation approach, which exploits both existing dataset topic proles, as well as traditional dataset connectivity measures, in order to link LOD datasets into a global dataset-topic-graph. This approach relies on the LOD graph in order to learn the connectivity behaviour between LOD datasets. However, experiments have shown that the current topology of the LOD cloud group is far from being complete to be considered as a ground truth and consequently as learning data.Facing the limits the current topology of LOD (as learning data), our research has led to break away from the topic proles representation of "learn to rank" approach and to adopt a new approach for candidate datasets identication where the recommendation is based on the intensional profiles overlap between differentdatasets. By intensional profile, we understand the formal representation of a set of schema concept labels that best describe a dataset and can be potentially enriched by retrieving the corresponding textual descriptions. This representation provides richer contextual and semantic information and allows to compute efficiently and inexpensively similarities between proles. We identify schema overlap by the help of a semantico-frequential concept similarity measure and a ranking criterion based on the tf*idf cosine similarity. The experiments, conducted over all available linked datasets on the LOD cloud, show that our method achieves an average precision of up to 53% for a recall of 100%. Furthermore, our method returns the mappings between the schema concepts across datasets, a particularly useful input for the data linking step.In order to ensure a high quality representative datasets schema profiles, we introduce Datavore| a tool oriented towards metadata designers that provides rankedlists of vocabulary terms to reuse in data modeling process, together with additional metadata and cross-terms relations. The tool relies on the Linked Open Vocabulary (LOV) ecosystem for acquiring vocabularies and metadata and is made available for the community
Shu, Wu. "Contributions à la détection des anomalies et au développement des systèmes de recommandation." Thèse, Université de Sherbrooke, 2012. http://hdl.handle.net/11143/6563.
Elati, Mohamed. "Apprentissage de réseaux de régulation génétique à partir de données d'expression." Paris 13, 2007. http://www.theses.fr/2007PA132031.
Bonis, Thomas. "Algorithmes d'apprentissage statistique pour l'analyse géométrique et topologique de données." Thesis, Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLS459/document.
In this thesis, we study data analysis algorithms using random walks on neighborhood graphs, or random geometric graphs. It is known random walks on such graphs approximate continuous objects called diffusion processes. In the first part of this thesis, we use this approximation result to propose a new soft clustering algorithm based on the mode seeking framework. For our algorithm, we want to define clusters using the properties of a diffusion process. Since we do not have access to this continuous process, our algorithm uses a random walk on a random geometric graph instead. After proving the consistency of our algorithm, we evaluate its efficiency on both real and synthetic data. We then deal tackle the issue of the convergence of invariant measures of random walks on random geometric graphs. As these random walks converge to a diffusion process, we can expect their invariant measures to converge to the invariant measure of this diffusion process. Using an approach based on Stein's method, we manage to obtain quantitfy this convergence. Moreover, the method we use is more general and can be used to obtain other results such as convergence rates for the Central Limit Theorem. In the last part of this thesis, we use the concept of persistent homology, a concept of algebraic topology, to improve the pooling step of the bag-of-words approach for 3D shapes
Aleksandrova, Marharyta. "Factorisation de matrices et analyse de contraste pour la recommandation." Thesis, Université de Lorraine, 2017. http://www.theses.fr/2017LORR0080/document.
In many application areas, data elements can be high-dimensional. This raises the problem of dimensionality reduction. The dimensionality reduction techniques can be classified based on their aim: dimensionality reduction for optimal data representation and dimensionality reduction for classification, as well as based on the adopted strategy: feature selection and feature extraction. The set of features resulting from feature extraction methods is usually uninterpretable. Thereby, the first scientific problematic of the thesis is how to extract interpretable latent features? The dimensionality reduction for classification aims to enhance the classification power of the selected subset of features. We see the development of the task of classification as the task of trigger factors identification that is identification of those factors that can influence the transfer of data elements from one class to another. The second scientific problematic of this thesis is how to automatically identify these trigger factors? We aim at solving both scientific problematics within the recommender systems application domain. We propose to interpret latent features for the matrix factorization-based recommender systems as real users. We design an algorithm for automatic identification of trigger factors based on the concepts of contrast analysis. Through experimental results, we show that the defined patterns indeed can be considered as trigger factors
Chulyadyo, Rajani. "Un nouvel horizon pour la recommandation : intégration de la dimension spatiale dans l'aide à la décision." Thesis, Nantes, 2016. http://www.theses.fr/2016NANT4012/document.
Nowadays it is very common to represent a system in terms of relationships between objects. One of the common applications of such relational data is Recommender System (RS), which usually deals with the relationships between users and items. Probabilistic Relational Models (PRMs) can be a good choice for modeling probabilistic dependencies between such objects. A growing trend in recommender systems is to add spatial dimensions to these objects, and make recommendations considering the location of users and/or items. This thesis deals with the (not much explored) intersection of three related fields – Probabilistic Relational Models (a method to learn probabilistic models from relational data), spatial data (often used in relational settings), and recommender systems (which deal with relational data). The first contribution of this thesis deals with the overlapping of PRM and recommender systems. We have proposed a PRM-based personalized recommender system that is capable of making recommendations from user queries in cold-start systems without user profiles. Our second contribution addresses the problem of integrating spatial information into a PRM
Ben, Ticha Sonia. "Recommandation personnalisée hybride." Thesis, Université de Lorraine, 2015. http://www.theses.fr/2015LORR0168/document.
Face to the ongoing rapid expansion of the Internet, user requires help to access to items that may interest her or him. A personalized recommender system filters relevant items from huge catalogue to particular user by observing his or her behavior. The approach based on observing user behavior from his interactions with the website is called usage analysis. Collaborative Filtering and Content-Based filtering are the most widely used techniques in personalized recommender system. Collaborative filtering uses only data from usage analysis to build user profile, while content-based filtering relies in addition on semantic information of items. Hybrid approach is another important technique, which combines collaborative and content-based methods to provide recommendations. The aim of this thesis is to present a new hybridization approach that takes into account the semantic information of items to enhance collaborative recommendations. Several approaches have been proposed for learning a new user profile inferring preferences for semantic information describing items. For each proposed approach, we address the sparsity and the scalability problems. We prove also, empirically, an improvement in recommendations accuracy against collaborative filtering and content-based filtering
Falip, Joris. "Structuration de données multidimensionnelles : une approche basée instance pour l'exploration de données médicales." Thesis, Reims, 2019. http://www.theses.fr/2019REIMS014/document.
A posteriori use of medical data accumulated by practitioners represents a major challenge for clinical research as well as for personalized patient follow-up. However, health professionals lack the appropriate tools to easily explore, understand and manipulate their data. To solve this, we propose an algorithm to structure elements by similarity and representativeness. This method allows individuals in a dataset to be grouped around representative and generic members who are able to subsume the elements and summarize the data. This approach processes each dimension individually before aggregating the results and is adapted to high-dimensional data and also offers transparent, interpretable and explainable results. The results we obtain are suitable for exploratory analysis and reasoning by analogy: the structure is similar to the organization of knowledge and decision-making process used by experts. We then propose an anomaly detection algorithm that allows complex and high-dimensional anomalies to be detected by analyzing two-dimensional projections. This approach also provides interpretable results. We evaluate these two algorithms on real and simulated high-dimensional data with up to thousands of dimensions. We analyze the properties of graphs resulting from the structuring of elements. We then describe a medical data pre-processing tool and a web application for physicians. Through this intuitive tool, we propose a visual structure of the elements to ease the exploration. This decision support prototype assists medical diagnosis by allowing the physician to navigate through the data and explore similar patients. It can also be used to test clinical hypotheses on a cohort of patients
Poirier, Damien. "Des textes communautaires à la recommandation." Phd thesis, Université d'Orléans, 2011. http://tel.archives-ouvertes.fr/tel-00597422.
Delporte, Julien. "Factorisation matricielle, application à la recommandation personnalisée de préférences." Phd thesis, INSA de Rouen, 2014. http://tel.archives-ouvertes.fr/tel-01005223.
Moin, Afshin. "Les Techniques De Recommandation Et De Visualisation Pour Les Données A Une Grande Echelle." Phd thesis, Université Rennes 1, 2012. http://tel.archives-ouvertes.fr/tel-00724121.
Moin, Afshin. "Les techniques de recommandation et de visualisation pour les données à une grande échelle." Rennes 1, 2012. https://tel.archives-ouvertes.fr/tel-00724121.
Nous avons assisté au développement rapide de la technologie de l'information au cours de la dernière décennie. D'une part, la capacité du traitement et du stockage des appareils numériques est en constante augmentation grâce aux progrès des méthodes de construction. D'autre part, l'interaction entre ces dispositifs puissants a été rendue possible grâce à la technologie de réseautage. Une conséquence naturelle de ces progrès, est que le volume des données générées dans différentes applications a grandi à un rythme sans précédent. Désormais, nous sommes confrontés à de nouveaux défis pour traiter et représenter efficacement la masse énorme de données à notre disposition. Cette thèse est centrée autour des deux axes de recommandation du contenu pertinent et de sa visualisation correcte. Le rôle des systèmes de recommandation est d'aider les utilisateurs dans le processus de prise de décision pour trouver des articles avec un contenu pertinent et une qualité satisfaisante au sein du vaste ensemble des possibilités existant dans le Web. D'autre part, la représentation correcte des données traitées est un élément central à la fois pour accroître l’utilité des données pour l'utilisateur final et pour la conception des outils d'analyse efficaces. Dans cet exposé, les principales approches des systèmes de recommandation ainsi que les techniques les plus importantes de la visualisation des données sous forme de graphes sont discutées. En outre, il est montré comment quelques-unes des mêmes techniques appliquées aux systèmes de recommandation peuvent être modifiées pour tenir compte des exigences de visualisation
Gras, Benjamin. "Les oubliés de la recommandation sociale." Thesis, Université de Lorraine, 2018. http://www.theses.fr/2018LORR0017/document.
A recommender system aims at providing relevant resources to a user, named the active user. To allow this recommendation, the system exploits the information it has collected about the active user or about resources. The collaborative filtering (CF) is a widely used recommandation approach. The data exploited by CF are the preferences expressed by users on resources. CF is based on the assumption that preferences are consistent between users, allowing a user's preferences to be inferred from the preferences of other users. In a CF-based recommender system, at least one user community has to share the preferences of the active user to provide him with high quality recommendations. Let us define a specific preference as a preference that is not shared by any group of user. A user with several specific preferences will likely be poorly served by a classic CF approach. This is the problem of Grey Sheep Users (GSU). In this thesis, I focus on three separate questions. 1) What is a specific preference? I give an answer by proposing associated hypotheses that I validate experimentally. 2) How to identify GSU in preference data? This identification is important to anticipate the low quality recommendations that will be provided to these users. I propose numerical indicators to identify GSU in a social recommendation dataset. These indicators outperform those of the state of the art and allow to isolate users whose quality of recommendations is very low. 3) How can I model GSU to improve the quality of the recommendations they receive? I propose new recommendation approaches to allow GSU to benefit from the opinions of other users
Khemiri, Rym. "Vers l'OLAP collaboratif pour la recommandation des analyses en ligne personnalisées." Thesis, Lyon 2, 2015. http://www.theses.fr/2015LYO22015/document.
The objective of this thesis is to provide a collaborative approach to the OLAP involving several users, led by an integrated personalization process in decision-making systems in order to help the end user in their analysis process. Whether personalizing the warehouse model, recommending decision queries or recommending navigation paths within the data cubes, the user need an efficient decision-making system that assist him. We were interested in three issues falling within data warehouse and OLAP personalization offering three major contributions. Our contributions are based on a combination of datamining techniques with data warehouses and OLAP technology. Our first contribution is an approach about personalizing dimension hierarchies to obtain new analytical axes semantically richer for the user that can help him to realize new analyzes not provided by the original data warehouse model. Indeed, we relax the constraint of the fixed model of the data warehouse which allows the user to create new relevant analysis axes taking into account both his/her constraints and his/her requirements. Our approach is based on an unsupervised learning method, the constrained k-means. Our goal is then to recommend these new hierarchy levels to other users of the same user community, in the spirit of a collaborative system in which each individual brings his contribution. The second contribution is an interactive approach to help the user to formulate new decision queries to build relevant OLAP cubes based on its past decision queries, allowing it to anticipate its future analysis needs. This approach is based on the extraction of frequent itemsets from a query load associated with one or a set of users belonging to the same actors in a community organization. Our intuition is that the relevance of a decision query is strongly correlated to the usage frequency of the corresponding attributes within a given workload of a user (or group of users). Indeed, our approach of decision queries formulation is a collaborative approach because it allows the user to formulate relevant queries, step by step, from the most commonly used attributes by all actors of the user community. Our third contribution is a navigation paths recommendation approach within OLAP cubes. Users are often left to themselves and are not guided in their navigation process. To overcome this problem, we develop a user-centered approach that suggests the user navigation guidance. Indeed, we guide the user to go to the most interesting facts in OLAP cubes telling him the most relevant navigation paths for him. This approach is based on Markov chains that predict the next analysis query from the only current query. This work is part of a collaborative approach because transition probabilities from one query to another in the cuboids lattice (OLAP cube) is calculated by taking into account all analysis queries of all users belonging to the same community. To validate our proposals, we present a support system user-centered decision which comes in two subsystems: (1) content personalization and (2) recommendation of decision queries and navigation paths. We also conducted experiments that showed the effectiveness of our analysis online user centered approaches using quality measures such as recall and precision
Firoozeh, Nazanin. "Semantic-oriented Recommandation for Content Enrichment." Thesis, Sorbonne Paris Cité, 2018. http://www.theses.fr/2018USPCD033.
In this thesis, we aim at enriching the content of an unstructured document with respect to a domain of interest. The goal is to minimize the vocabulary and informational gap between the document and the domain. Such an enrichment which is based on Natural Language Processing and Information Retrieval technologies has several applications. As an example, flling in the gap between a scientifc paper and a collection of highly cited papers in a domain helps the paper to be better acknowledged by the community that refers to that collection. Another example is to fll in the gap between a web page and the usual keywords of visitors that are interested in a given domain so as it is better indexed and referred to in that domain, i.e. more accessible for those visitors. We propose a method to fll that gap. We first generate an enrichment collection, which consists of the important documents related to the domain of interest. The main information of the enrichment collection is then extracted, disambiguated and proposed to a user,who performs the enrichment. This is achieved by decomposing the problem into two main components of keyword extraction and topic detection. We present a comprehensive study over different approaches of each component. Using our findings, we propose approaches for extracting keywords from web pages, detecting their under lying topics, disambiguating them and returning the ones related to the domain of interest. The enrichment is performed by recommending discriminative sets of semantically relevant keywords, i.e. topics, to a user. The topics are labeled with representative keywords and have a level of granularity that is easily interpretable. Topic keywords are ranked by importance. This helps to control the length of the document, which needs to be enriched, by targeting the most important keywords of each topic. Our approach is robust to the noise in web pages. It is also knowledge-poor and domain-independent. It, however, exploits search engines for generating the required data but is optimized in the number of requests sent to them. In addition, the approach is easily tunable to different languages. We have implemented the keyword extraction approach in 12 languages and four of them have been tested over various domains. The topic detection approach has been implemented and tested on English and French. However, it is on French language that the approaches have been tested on a large scale : the keyword extraction on roughly 400 domains and the topic detection on 80 domains.To evaluate the performance of our enrichment approach, we focused on French and we performed different experiments on the proposed keyword extraction and topic detection methods. To evaluate their robustness, we studied them on 10 topically diverse domains.Results were evaluated through both user-based evaluations on a real application context and by comparing with baseline approaches. Our results on the keyword extraction approach showed that the statistical features are not adequate for capturing words importance within a web page. In addition, we found our proposed approach of keyword extraction to be effective when applied on real applications. The evaluations on the topic detection approach also showed that it can electively filter out the keywords which are not related to a target domain and that it labels the topics with representative and discriminative keywords. In addition, the approach achieved a high precision in preserving the semantic consistency of the keywords within each topic. We showed that our approach out performs a baseline approach, since the widely-used co-occurrence feature between keywords is notivenough for capturing their semantic similarity and consequently for detecting semantically consistent topics
Duthil, Benjamin. "De l'extraction des connaissances à la recommandation." Phd thesis, Montpellier 2, 2012. http://tel.archives-ouvertes.fr/tel-00771504.
Berti-Équille, Laure. "La qualité des données et leur recommandation : modèle conceptuel, formalisation et application a la veille technologique." Toulon, 1999. http://www.theses.fr/1999TOUL0008.
Technological Watch activities are focused on information qualification and validation by human expertise. As a matter of facf, none of these systems can provide (nor assist) a critical and qualitative analysis of data they store and manage- Most of information systems store data (1) whose source is usually unique, not known or not identified/authenticated (2) whose quality is unequal and/or ignored. In practice, several data may describe the same entity in the real world with contradictory values and their relative quality may be comparatively evaluated. Many techniques for data cleansing and editing exist for detecting some errors in database but it is determinant to know which data have bad quality and to use the benefit of a qualitative expert judgment on data, which is complementary to quantitative and statistical data analysis. My contribution is to provide a multi-source perspective to data quality, to introduce and to define the concepts of multi-source database (MSDB) and multi-source data quality (MSDQ). My approach was to analyze the wide panorama of research in the literature whose problematic have some analogies with technological watch problematic. The main objective of my work was to design and to provide a storage environment for managing textual information sources, (more or less contradictory) data that are extracted from the textual content and their quality mcta-data. My work was centered on proposing : the methodology to guide step-by-step a project for data quality in a multi-source information context, the conceptual modeling of a multi-source database (MSDB) for managing data sources, multi-source data and their quality meta-data and proposing mechanisms for multi-criteria data recommendation ; the formalization of the QMSD data model (Quality of Multi-Source Data) which describes multi-source data, their quality meta-data and the set of operations for manipulating them ; the development of the sQuaL prototype for implementing and validating my propositions. In the long term, the perspectives are to develop a specific dccisional information system extending classical functionalities for (1) managing multi-source data (2) taking into account their quality meta-data and (3) proposing data-quality-based recommendation as query results. The ambition is to develop the concept of "introspective information system" ; that is to say, an information system thai is active and reactive concerning the quality of its own data
Séguéla, Julie. "Fouille de données textuelles et systèmes de recommandation appliqués aux offres d'emploi diffusées sur le web." Thesis, Paris, CNAM, 2012. http://www.theses.fr/2012CNAM0801/document.
Last years, e-recruitment expansion has led to the multiplication of web channels dedicated to job postings. In an economic context where cost control is fundamental, assessment and comparison of recruitment channel performances have become necessary. The purpose of this work is to develop a decision-making tool intended to guide recruiters while they are posting a job on the Internet. This tool provides to recruiters the expected performance on job boards for a given job offer. First, we identify the potential predictors of a recruiting campaign performance. Then, we apply text mining techniques to the job offer texts in order to structure postings and to extract information relevant to improve their description in a predictive model. The job offer performance predictive algorithm is based on a hybrid recommender system, suitable to the cold-start problem. The hybrid system, based on a supervised similarity measure, outperforms standard multivariate models. Our experiments are led on a real dataset, coming from a job posting database
Ducoffe, Mélanie. "Active learning et visualisation des données d'apprentissage pour les réseaux de neurones profonds." Thesis, Université Côte d'Azur (ComUE), 2018. http://www.theses.fr/2018AZUR4115/document.
Our work is presented in three separate parts which can be read independently. Firstly we propose three active learning heuristics that scale to deep neural networks: We scale query by committee, an ensemble active learning methods. We speed up the computation time by sampling a committee of deep networks by applying dropout on the trained model. Another direction was margin-based active learning. We propose to use an adversarial perturbation to measure the distance to the margin. We also establish theoretical bounds on the convergence of our Adversarial Active Learning strategy for linear classifiers. Some inherent properties of adversarial examples opens up promising opportunity to transfer active learning data from one network to another. We also derive an active learning heuristic that scales to both CNN and RNN by selecting the unlabeled data that minimize the variational free energy. Secondly, we focus our work on how to fasten the computation of Wasserstein distances. We propose to approximate Wasserstein distances using a Siamese architecture. From another point of view, we demonstrate the submodular properties of Wasserstein medoids and how to apply it in active learning. Eventually, we provide new visualization tools for explaining the predictions of CNN on a text. First, we hijack an active learning strategy to confront the relevance of the sentences selected with active learning to state-of-the-art phraseology techniques. These works help to understand the hierarchy of the linguistic knowledge acquired during the training of CNNs on NLP tasks. Secondly, we take advantage of deconvolution networks for image analysis to present a new perspective on text analysis to the linguistic community that we call Text Deconvolution Saliency
Albaric, Franck. "Problèmes posés par le recueil des données concernant la prise en charge extra-hospitalière de l'arrêt cardio-respiratoire : application du style d'Utstein." Lyon 1, 1994. http://www.theses.fr/1994LYO1261B.
Tran, Nguyen Minh-Thu. "Abstraction et règles d'association pour l'amélioration des systèmes de recommandation à partir de données de préférences binaires." Paris 13, 2011. http://www.theses.fr/2011PA132016.
In recent years, recommendation systems have been extensively explored in order to help the user facing the increasing information on Internet. Those systems are used in e-commerce (Amazon, eBay, Netflix. . . ), entertainment, online news, etc. In the domain of e-commerce, the available data is often difficult to exploit to build robust recommendations : binary data, long tail of the distribution of preferences and everlasting adding or removing of items. In fact, most recommender systems focus on the most popular items because the new items or those of the "long tail" are associated with little or no preference. To improve the performance of these systems, we propose to search for association rules between abstracted items. First, the abstraction of the items can lead to a considerable reduction of the long tail effect. Second, the extraction of abstract association rules can be used to identify items to be recommended. . Two algorithms are introduced : AbsTopk, based on the rules in the space of abstract and ACReco combining items in the space of abstract and concrete items by pair. These algorithms were evaluated quantitatively (relevance) and qualitatively (novelty and diversity) on a real database of an online e-commerce site. The empirical results presented show the interest of the proposed approach
Karoui, Hajer. "Système coopératif de type égal-à-égal pour la recommandation : Application à la gestion et la recommandation de références bibliographiques." Phd thesis, Université Paris-Nord - Paris XIII, 2007. http://tel.archives-ouvertes.fr/tel-00299935.
Deux problématiques se présentent : comment obtenir les références pertinentes et comment choisir des agents avec qui collaborer ? Pour résoudre ces problèmes, nous nous sommes basés sur l'exploitation des historiques des interactions entre les agents.
Le RàPC est utilisée pour deux finalités :
a)déterminer pour une requête, des agents intéressants à interroger ;
b)chercher pour une requête, des références pertinentes.
Laloë, Thomas. "Sur quelques problèmes d'apprentissage supervisé et non supervisé." Phd thesis, Université Montpellier II - Sciences et Techniques du Languedoc, 2009. http://tel.archives-ouvertes.fr/tel-00455528.
Goulon-Sigwalt-Abram, Aurélie. "Une nouvelle méthode d'apprentissage de données structurées : applications à l'aide à la découverte de médicaments." Phd thesis, Université Pierre et Marie Curie - Paris VI, 2008. http://pastel.archives-ouvertes.fr/pastel-00004103.
Draidi, Fady. "Recommandation Pair-à-Pair pour Communautés en Ligne à Grande Echelle." Phd thesis, Université Montpellier II - Sciences et Techniques du Languedoc, 2012. http://tel.archives-ouvertes.fr/tel-00766963.
Ghenname, Mérième. "Le web social et le web sémantique pour la recommandation de ressources pédagogiques." Thesis, Saint-Etienne, 2015. http://www.theses.fr/2015STET4015/document.
This work has been jointly supervised by U. Jean Monnet Saint Etienne, in the Hubert Curien Lab (Frederique Laforest, Christophe Gravier, Julien Subercaze) and U. Mohamed V Rabat, LeRMA ENSIAS (Rachida Ahjoun, Mounia Abik). Knowledge, education and learning are major concerns in today’s society. The technologies for human learning aim to promote, stimulate, support and validate the learning process. Our approach explores the opportunities raised by mixing the Social Web and the Semantic Web technologies for e-learning. More precisely, we work on discovering learners profiles from their activities on the social web. The Social Web can be a source of information, as it involves users in the information world and gives them the ability to participate in the construction and dissemination of knowledge. We focused our attention on tracking the different types of contributions, activities and conversations in learners spontaneous collaborative activities on social networks. The learner profile is not only based on the knowledge extracted from his/her activities on the e-learning system, but also from his/her many activities on social networks. We propose a methodology for exploiting hashtags contained in users’ writings for the automatic generation of learner’s semantic profiles. Hashtags require some processing before being source of knowledge on the user interests. We have defined a method to identify semantics of hashtags and semantic relationships between the meanings of different hashtags. By the way, we have defined the concept of Folksionary, as a hashtags dictionary that for each hashtag clusters its definitions into meanings. Semantized hashtags are thus used to feed the learner’s profile so as to personalize recommendations on learning material. The goal is to build a semantic representation of the activities and interests of learners on social networks in order to enrich their profiles. We also discuss our recommendation approach based on three types of filtering (personalized, social, and statistical interactions with the system). We focus on personalized recommendation of pedagogical resources to the learner according to his/her expectations and profile
Piworwarski, Benjamin. "Techniques d'apprentissage pour le traitement d'informations structurées : application à la recherche d'information." Paris 6, 2003. http://www.theses.fr/2003PA066567.
Omidvar, Tehrani Behrooz. "Optimization-based User Group Management : Discovery, Analysis, Recommendation." Thesis, Université Grenoble Alpes (ComUE), 2015. http://www.theses.fr/2015GREAM038/document.
User data is becoming increasingly available in multiple domains ranging from phone usage traces to data on the social Web. User data is a special type of data that is described by user demographics (e.g., age, gender, occupation, etc.) and user activities (e.g., rating, voting, watching a movie, etc.) The analysis of user data is appealing to scientists who work on population studies, online marketing, recommendations, and large-scale data analytics. However, analysis tools for user data is still lacking.In this thesis, we believe there exists a unique opportunity to analyze user data in the form of user groups. This is in contrast with individual user analysis and also statistical analysis on the whole population. A group is defined as set of users whose members have either common demographics or common activities. Group-level analysis reduces the amount of sparsity and noise in data and leads to new insights. In this thesis, we propose a user group management framework consisting of following components: user group discovery, analysis and recommendation.The very first step in our framework is group discovery, i.e., given raw user data, obtain user groups by optimizing one or more quality dimensions. The second component (i.e., analysis) is necessary to tackle the problem of information overload: the output of a user group discovery step often contains millions of user groups. It is a tedious task for an analyst to skim over all produced groups. Thus we need analysis tools to provide valuable insights in this huge space of user groups. The final question in the framework is how to use the found groups. In this thesis, we investigate one of these applications, i.e., user group recommendation, by considering affinities between group members.All our contributions of the proposed framework are evaluated using an extensive set of experiments both for quality and performance
Rochd, El Mehdi. "Modèles probabilistes de consommateurs en ligne : personnalisation et recommandation." Thesis, Aix-Marseille, 2015. http://www.theses.fr/2015AIXM4086.
Research systems have facilitated access to information available on the web using mechanisms for collecting, indexing and storage of heterogeneous content. They generate data resulting from the activity of users on Internet (queries, logfile). The next step is to analyze the data using data mining tools in order to improve the response’s quality of these systems, or to customize the response based on users’ profiles. Some actors, such as the company Marketshot, are positioned as intermediaries between consumers and professionals. Indeed, they link potential buyers with the leading brands and distribution networks through their websites. For such purposes, these intermediaries have developed effective portals, and have stored large volumes of data related to the activity of users on their websites. These data repositories are exploited to respond positively to the needs of users as well as those of professionals who seek to understand the behavior of their customers and anticipate their purchasing actions. My thesis comes within the framework of searching through the data collected from the web. The idea is to build models that explain the correlation between the activities of users on websites of aid for the purchase, and sales trends of products in « real life ». In fact, my research concerns probabilistic learning, in particular Topic Models. It involves modeling the users’ behavior from uses of trader websites
Nana, jipmo Coriane. "Intégration du web social dans les systèmes de recommandation." Thesis, Université Paris-Saclay (ComUE), 2017. http://www.theses.fr/2017SACLC082/document.
The social Web grows more and more and gives through the web, access to a wide variety of resources, like sharing sites such as del.icio.us, exchange messages as Twitter, or social networks with the professional purpose such as LinkedIn, or more generally for social purposes, such as Facebook and LiveJournal. The same individual can be registered and active on different social networks (potentially having different purposes), in which it publishes various information, which are constantly growing, such as its name, locality, communities, various activities. The information (textual), given the international dimension of the Web, is inherently multilingual and intrinsically ambiguous, since it is published in natural language in a free vocabulary by individuals from different origin. They are also important, specially for applications seeking to know their users in order to better understand their needs, activities and interests. The objective of our research is to exploit using essentially the Wikpédia encyclopedia, the textual resources extracted from the different social networks of the same individual in order to construct his characterizing profile, which can be exploited in particular by applications seeking to understand their users, such as recommendation systems. In particular, we conducted a study to characterize the personality traits of users. Many experiments, analyzes and evaluations were carried out on real data collected from different social networks
Jerbi, Houssem. "Personnalisation d'analyses décisionnelles sur des données multidimensionnelles." Phd thesis, Toulouse 1, 2012. http://tel.archives-ouvertes.fr/tel-00695371.
Feuilloy, Mathieu. "Étude d'algorithmes d'apprentissage artificiel pour la prédiction de la syncope chez l'homme." Phd thesis, Université d'Angers, 2009. http://tel.archives-ouvertes.fr/tel-00465008.
Tadlaoui, Mohammed. "Système de recommandation de ressources pédagogiques fondé sur les liens sociaux : Formalisation et évaluation." Thesis, Lyon, 2018. http://www.theses.fr/2018LYSEI053/document.
With the increasing amount of educational content produced daily by users, it becomes very difficult for learners to find the resources that are best suited to their needs. Recommendation systems are used in educational platforms to solve the problem of information overload. They are designed to provide relevant resources to a learner using some information about users and resources. The present work fits in the context of recommender systems for educational resources, especially systems that use social information. We have defined an educational resource recommendation approach based on research findings in the area of recommender systems, social networks, and Technology-Enhanced Learning. We rely on social relations between learners to improve the accuracy of recommendations. Our proposal is based on formal models that calculate the similarity between users of a learning environment to generate three types of recommendation, namely the recommendation of 1) popular resources; 2) useful resources; and 3) resources recently consulted. We have developed a learning platform, called Icraa, which integrates our recommendation models. The Icraa platform is a social learning environment that allows learners to download, view and evaluate educational resources. In this thesis, we present the results of an experiment conducted for almost two years on a group of 372 learners of Icraa in a real educational context. The objective of this experiment is to measure the relevance, quality and usefulness of the recommended resources. This study allowed us to analyze the user’s feedback on the three types of recommendations. This analysis is based on the users’ traces which was saved with Icraa and on a questionnaire. We have also performed an offline analysis using a dataset to compare our approach with four base line algorithms
Nurbakova, Diana. "Recommendation of activity sequences during distributed events." Thesis, Lyon, 2018. http://www.theses.fr/2018LYSEI115/document.
Multi-day events such as conventions, festivals, cruise trips, to which we refer to as distributed events, have become very popular in recent years, attracting hundreds or thousands of participants. Their programs are usually very dense, making it challenging for the attendees to make a decision which events to join. Recommender systems appear as a common solution in such an environment. While many existing solutions deal with personalised recommendation of single items, recent research focuses on the recommendation of consecutive items that exploits user's behavioural patterns and relations between entities, and handles geographical and temporal constraints. In this thesis, we first formulate the problem of recommendation of activity sequences, classify and discuss the types of influence that have an impact on the estimation of the user's interest in items. Second, we propose an approach (ANASTASIA) to solve this problem, which aims at providing an integrated support for users to create a personalised itinerary of activities. ANASTASIA brings together three components, namely: (1) estimation of the user’s interest in single items, (2) use of sequential influence on activity performance, and (3) building of an itinerary that takes into account spatio-temporal constraints. Thus, the proposed solution makes use of the methods based on sequence learning and discrete optimisation. Moreover, stating the lack of publicly available datasets that could be used for the evaluation of event and itinerary recommendation algorithms, we have created two datasets, namely: (1) event attendance on board of a cruise (Fantasy_db) based on a conducted user study, and (2) event attendance at a major comic book convention (DEvIR). This allows to perform evaluation of recommendation methods, and contributes to the reproducibility of results
Guàrdia, Sebaoun Elie. "Accès personnalisé à l'information : prise en compte de la dynamique utilisateur." Thesis, Paris 6, 2017. http://www.theses.fr/2017PA066519/document.
The main goal of this thesis resides in using rich and efficient profiling to improve the adequation between the retrieved information and the user's expectations. We focus on exploiting as much feedback as we can (being clicks, ratings or written reviews) as well as context. In the meantime, the tremendous growth of ubiquitous computing forces us to rethink the role of information access platforms. Therefore, we took interest not solely in performances but also in accompanying users through their access to the information. Through this thesis, we focus on users dynamics modeling. Not only it improves the system performances but it also brings some kind of explicativity to the recommendation. Thus, we propose to accompany the user through his experience accessing information instead of constraining him to a given set of items the systems finds fitting
Bouzayane, Sarra. "Méthode de classification multicritère, incrémentale et périodique appliquée à la recommandation pour l'aide au transfert des savoirs dans les MOOCs." Thesis, Amiens, 2017. http://www.theses.fr/2017AMIE0029/document.
The thesis deals with the problem of knowledge transfer in mediated environments in the era of massive data. We propose a Multicriteria Approach for the Incremental Periodic Prediction (MAI2P) of the decision class to which an action is likely to belong. The MAI2P method is based on three phases. The first consists of three steps : the construction of a family of criteria for the characterization of actions ; the construction of a representative set of “Reference actions” for each of the decision classes ; and the construction of a decision table. The second phase is based on the DRSA-Incremental algorithm that we propose for the inference and the updating of the set of decision rules following the sequential increment of the “Reference actions” set. The third phase is meant to classify the “Potential Actions” in one of the predefined decision classes using the set of inferred decision rules. The MAI2P method is validated especially in the context of the Massive Open Online Courses (MOOCs), which are e-courses characterized by a huge amount of data exchanged between a massive number of learners. It allows the weekly prediction of the three decision classes : Cl1 of the “At risk learners”, those who intend to give up the MOOC; Cl2 of the “Struggling learners”, those who have pedagogical difficulties but have no plan to abandon it ; and Cl3 of the “Leader learners”, those who can support the other two classes of learners by providing them with all the information they need. The prediction is based on data from all the previous weeks of the MOOC in order to predict the learner profile for the following week. A recommender system KTI-MOOC (Recommender system for Knowledge Transfer Improvement within a MOOC) is developed to recommend to each “At risk learner” or “Struggling learner” a personalized list of “Leader learners”. This system is based on the demographic filtering technique and aims to promote the individual appropriation, of the exchanged information, for each learner
Rajaonarivo, Hiary Landy. "Approche co-évolutive humain-système pour l'exploration de bases de données." Thesis, Brest, 2018. http://www.theses.fr/2018BRES0114/document.
This thesis focus on a proposition that helps humans during the exploration of database. The particularity of this proposition relies on a co-evolution principle between the user and an intelligent interface. It provides a support to the understanding of the domain represented by the data. A metaphor of living virtual museum is adopted. This museum evolves incrementally according to the user's interactions. It incarnates both the data and the semantic information which are expressed by a knowledge model specific to the domain of the data. Through the topological organization and the incremental evolution, the museum personalizes online the user's exploration. The approach is insured by three main mechanisms: the evaluation of the user profile modelled by a dynamical weighting of the semantic information, the use of this dynamic profile to establish a recommendation as well as the incarnation of the data in the living museum. The approach has been applied to the heritage domain as part of the ANTIMOINE project, funded by the National Research Agency (ANR). The genericity of the latter has been demonstrated through its application to a database of publications but also using various types of interfaces (website, virtual reality).Experiments have validated the hypothesis that our system adapts itself to the user behavior and that it is able, in turn, to influence him.They also showed the comparison between a 2D interface and a 3D interface in terms of quality of perception, guidance, preference and efficiency
Dupas, Rémy. "Apport des méthodes d'apprentissage symbolique automatique pour l'aide à la maintenance industrielle." Valenciennes, 1990. https://ged.uphf.fr/nuxeo/site/esupversions/7ab53b01-cdfb-4932-ba60-cb5332e3925a.
Peoples, Bruce E. "Méthodologie d'analyse du centre de gravité de normes internationales publiées : une démarche innovante de recommandation." Thesis, Paris 8, 2016. http://www.theses.fr/2016PA080023.
“Standards make a positive contribution to the world we live in. They facilitate trade, spreadknowledge, disseminate innovative advances in technology, and share good management andconformity assessment practices”7. There are a multitude of standard and standard consortiaorganizations producing market relevant standards, specifications, and technical reports in thedomain of Information Communication Technology (ICT). With the number of ICT relatedstandards and specifications numbering in the thousands, it is not readily apparent to users howthese standards inter-relate to form the basis of technical interoperability. There is a need todevelop and document a process to identify how standards inter-relate to form a basis ofinteroperability in multiple contexts; at a general horizontal technology level that covers alldomains, and within specific vertical technology domains and sub-domains. By analyzing whichstandards inter-relate through normative referencing, key standards can be identified as technicalcenters of gravity, allowing identification of specific standards that are required for thesuccessful implementation of standards that normatively reference them, and form a basis forinteroperability across horizontal and vertical technology domains. This Thesis focuses on defining a methodology to analyze ICT standards to identifynormatively referenced standards that form technical centers of gravity utilizing Data Mining(DM) and Social Network Analysis (SNA) graph technologies as a basis of analysis. As a proofof concept, the methodology focuses on the published International Standards (IS) published bythe International Organization of Standards/International Electrotechnical Committee; JointTechnical Committee 1, Sub-committee 36 Learning Education, and Training (ISO/IEC JTC1 SC36). The process is designed to be scalable for larger document sets within ISO/IEC JTC1 that covers all JTC1 Sub-Committees, and possibly other Standard Development Organizations(SDOs).Chapter 1 provides a review of literature of previous standard analysis projects and analysisof components used in this Thesis, such as data mining and graph theory. Identification of adataset for testing the developed methodology containing published International Standardsneeded for analysis and form specific technology domains and sub-domains is the focus ofChapter 2. Chapter 3 describes the specific methodology developed to analyze publishedInternational Standards documents, and to create and analyze the graphs to identify technicalcenters of gravity. Chapter 4 presents analysis of data which identifies technical center of gravitystandards for ICT learning, education, and training standards produced in ISO/IEC JTC1 SC 36.Conclusions of the analysis are contained in Chapter 5. Recommendations for further researchusing the output of the developed methodology are contained in Chapter 6
Baro, Johanna. "Modélisation multi-échelles de la morphologie urbaine à partir de données carroyées de population et de bâti." Thesis, Paris Est, 2015. http://www.theses.fr/2015PEST1004/document.
Since a couple of decades the relationships between urban form and travel patterns are central to reflection on sustainable urban planning and transport policy. The increasing distribution of regular grid data is in this context a new perspective for modeling urban structures from measurements of density freed from the constraints of administrative division. Population density data are now available on 200 meters grids covering France. We complete these data with built area densities in order to propose two types of classified images adapted to the study of travel patterns and urban development: classifications of urban fabrics and classifications of morphotypes of urban development. The construction of such classified images is based on theoretical and experimental which raise methodological issues regarding the classification of a statistically various urban spaces. To proceed exhaustively those spaces, we proposed a per-pixel classification method of urban fabrics by supervised transfer learning. Hidden Markov random fields are used to take into account the dependencies in the spatial data. The classifications of morphotypes are then obtained by broadening the knowledge of urban fabrics. These classifications are formalized from chorematique theoretical models and implemented by qualitative spatial reasoning. The analysis of these classifications by methods of quantitative spatial reasoning and factor analysis allowed us to reveal the morphological diversity of 50 metropolitan areas. It highlights the relevance of these classifications to characterize urban areas in accordance with various development issues related to the density or multipolar development
Moreau, Aurélien. "How fuzzy set theory can help make database systems more cooperative." Thesis, Rennes 1, 2018. http://www.theses.fr/2018REN1S043/document.
In this thesis, we are interested in how we can leverage fuzzy logic to improve the interactions between relational database systems and humans. Cooperative answering techniques aim to help users harness the potential of DBMSs. These techniques are expected to be robust and always provide answer to users. Empty set (0,00 sec) is a typical example of answer that one may wish to never obtain. The informative nature of explanations is higher than that of actual answers in several cases, e.g. empty answer sets and plethoric answer sets, hence the interest of robust cooperative answering techniques capable of both explaining and improving an answer set. Using terms from natural language to describe data --- with labels from fuzzy vocabularies --- contributes to the interpretability of explanations. Offering to define and refine vocabulary terms increases the personalization experience and improves the interpretability by using the user's own words. We propose to investigate the use of explanations in a cooperative answering setting using three research axes: 1) in the presence of a plethoric set of answers; 2) in the context of recommendations; 3) in the context of a query/answering problem. These axes define cooperative techniques where the interest of explanations is to enable users to understand how results are computed in an effort of transparency. The informativeness of the explanations brings an added value to the direct results, and that in itself represents a cooperative answer
Ngo, Duy Hoa. "Amélioration de l'alignement d'ontologies par les techniques d'apprentissage automatique, d'appariement de graphes et de recherche d'information." Phd thesis, Université Montpellier II - Sciences et Techniques du Languedoc, 2012. http://tel.archives-ouvertes.fr/tel-00767318.
Sakout, Andaloussi Kenza, and Andaloussi Kenza Sakout. "Amélioration de l'expérience d'apprentissage dans un système hypermédia adaptatif éducatif grâce aux données extraites et inférées à partir des réseaux sociaux." Doctoral thesis, Université Laval, 2019. http://hdl.handle.net/20.500.11794/37918.
Avec l'émergence des formations en ligne accessibles pour tous, la personnalisation de l'apprentissage devient de plus en plus cruciale et présente de nouveaux défis aux chercheurs du domaine. Il est actuellement nécessaire de tenir compte de l'hétérogénéité du public cible et lui présenter des contenus éducatifs adaptés à ses besoins et sa façon d'apprendre afin de lui permettre de profiter au maximum de ces formations et éviter le décrochage. Ce travail de recherche s'inscrit dans le cadre des travaux sur la personnalisation de l'apprentissage à travers les systèmes hypermédias adaptatifs utilisés en éducation (SHAE). Ces systèmes ont la vocation de personnaliser le processus d'apprentissage selon des critères bien spécifiques, tels que les pré-requis ou plus souvent les styles d'apprentissage, en générant un chemin d'apprentissage adéquat. Les SHAE se basent généralement sur trois modèles principaux à savoir le modèle apprenant, le modèle du domaine et le modèle d'adaptation. Bien que la personnalisation du processus d'apprentissage offerte par les SHAE actuels soit avantageuse pour les apprenants, elle présente encore certaines limites. D'un côté, juste le fait de personnaliser l'apprentissage augmente les chances que le contenu présenté à l'apprenant lui soit utile et sera ainsi mieux compris. Mais d'un autre côté, la personnalisation dans les SHAE existants se contente des critères niveau de connaissances et style d'apprentissage, et elle s'applique seulement à certains aspects qui n'ont pas évolué depuis leur création, à savoir le contenu, la présentation et la navigation. Ceci remet en question la pertinence des objets d'apprentissage attribués aux apprenants et la motivation de ces derniers à faire usage des SHAE sachant que ceux-ci se basent essentiellement sur les questionnaires pour la constitution de leur modèle apprenant. Suite à une étude empirique d'une cinquantaine de SHAE existants, révélant leurs atouts et limites, certains objectifs de recherche ont été identifiés afin d'améliorer l'expérience d'apprentissage à travers ces systèmes. Ces objectifs visent à établir un modèle de SHAE capable de (i) déterminer les données du modèle apprenant de façon implicite à partir des réseaux sociaux tout en répondant aux standards associés à ce modèle afin de construire le modèle apprenant; (ii) favoriser la collaboration entre les différents apprenants qui seraient mieux motivés à apprendre en collaborant; (iii) personnaliser, de façon automatique, de nouveaux aspects à savoir l'approche pédagogique, la collaboration et le feedback selon les traits de personnalité de l'apprenant en plus des trois volets existants. Un modèle de SHAE a été proposé pour répondre à ces objectifs. Ce modèle permet d’extraire les données personnelles de l'utilisateur à partir de ses réseaux sociaux et de prédire ses traits de personnalité selon son interaction avec ces réseaux. Par la suite, il est possible d'adapter les objets d'apprentissage, sur la base d'un système de recommandation, à ces traits de personnalité en plus du style d'apprentissage et du niveau de connaissances des apprenants. L'adaptation aux traits de personnalité de l'apprenant selon le modèle Big Five a permis de personnaliser de nouveaux aspects tels l'approche pédagogique, le type de collaboration et le feedback. Un prototype, "ColadaptLearn", conçu à partir de ce modèle et expérimenté avec un ensemble d'étudiants a permis de valider les choix du prototype pour les objets d'apprentissage, selon les règles préétablies, en les confrontant aux choix faits par les étudiants. Ces données ont été utilisées pour développer un réseau bayésien permettant de prédire les objets d'apprentissage adéquats aux futurs apprenants. Les résultats de l’expérimentation ont montré qu'il y a une bonne concordance entre les choix du prototype et ceux des apprenants, en plus d'une satisfaction de ces derniers par rapport aux feedbacks reçus, ce qui appuie le rajout des nouveaux aspects proposés. Comme suite à cette thèse, il est envisageable d'appliquer le modèle proposé dans des environnements d'apprentissage plus larges de types cours en ligne ouverts et massifs, jeu sérieux ou même des formations mobiles, ce qui contribuerait à mieux valider les propos amenés. Il est aussi possible d’utiliser des techniques d'apprentissage automatique autres que les réseaux bayésiens pour la prédiction des objets d'apprentissage adaptés. Finalement, il serait intéressant d'explorer d'autres sources de données qui pourraient fournir plus d'informations sur l'apprenant de façon implicite tels ses centres d'intérêt ou ses émotions auxquels un SHAE pourrait s'adapter.
With the growth of online learning accessible to all, learning personalization is becoming increasingly crucial and presents new challenges for researchers. It is currently essential to take into account the heterogeneity of the target audience and adapt educational content to their needs and learning style in such a way that they are able to fully benefit from these learning forms and prevent them from dropping out. This research work addresses learning personalization through adaptive educational hypermedia systems (AEHS). These systems are designed to customize the learning process according to specific criteria, such as prerequisites or, more often, learning styles, by generating a suitable learning path. AEHS are generally based on three main models: the learning model, the domain model and the adaptation model. Although the learning process customization offered by current AEHS is beneficial to learners, it still has some limitations. On one hand, just the fact of personalizing learning increases the likelihood that the content presented to the learner will be useful and thus better understood. But on the other hand, customization in existing AEHS is limited to the criteria knowledge level and learning style and applies only to certain aspects which have not evolved since their creation, namely content, presentation and navigation. This questions the relevance of the learning objects assigned to learners and their motivation to use such AEHS, knowing that they rely essentially on questionnaires to build their learner model. After conducting an empirical study of 50 existing AEHS, revealing their strengths and limitations, some research objectives were identified to improve the learning experience through such systems. These objectives aim to establish an AEHS model which is able to (i) implicitly identify the learning model data on the basis of social networks while meeting the associated standards; (ii) promote collaboration between different learners who would be better motivated to learn while collaborating; (iii) automatically customize new aspects such as the teaching approach, collaboration and feedback according to learners' personality traits in addition to the three existing ones. An AEHS model has been proposed to meet these objectives. This model makes it possible to extract the user's personal data from his social networks and to predict his personality traits depending on his interaction with these networks. Thereafter, it is possible to adapt the learning objects, on the basis of a recommendation system, to these personality traits in addition to the criteria learning style and knowledge level. Adapting to the learner's personality traits according to the Big Five model enabled the customization of new aspects such as the pedagogical approach, the collaboration type and the feedback. A prototype, "ColadaptLearn", based on this model and experimented with a group of students, validated the prototype's choices for learning objects while confronting them to the students' choices. These data were then used to build a Bayesian network to predict the appropriate learning objects for future learners. The experimental results showed that there is a good match between the prototype choices and those of learners, in addition to learners' satisfaction regarding the feedback received, which supports the addition of the proposed new aspects. As a follow-up to this thesis, it is possible to apply the proposed model in a larger learning environment such as massive open online courses (MOOC), serious games or mobile learning, which would help to validate the proposals made. It is also possible to use other automatic learning techniques than Bayesian networks to predict suitable learning objects. Finally, it would be interesting to explore other data sources that could implicitly provide more information about the learner, such as his or her interests or emotions that an SHAE could adapt to.
With the growth of online learning accessible to all, learning personalization is becoming increasingly crucial and presents new challenges for researchers. It is currently essential to take into account the heterogeneity of the target audience and adapt educational content to their needs and learning style in such a way that they are able to fully benefit from these learning forms and prevent them from dropping out. This research work addresses learning personalization through adaptive educational hypermedia systems (AEHS). These systems are designed to customize the learning process according to specific criteria, such as prerequisites or, more often, learning styles, by generating a suitable learning path. AEHS are generally based on three main models: the learning model, the domain model and the adaptation model. Although the learning process customization offered by current AEHS is beneficial to learners, it still has some limitations. On one hand, just the fact of personalizing learning increases the likelihood that the content presented to the learner will be useful and thus better understood. But on the other hand, customization in existing AEHS is limited to the criteria knowledge level and learning style and applies only to certain aspects which have not evolved since their creation, namely content, presentation and navigation. This questions the relevance of the learning objects assigned to learners and their motivation to use such AEHS, knowing that they rely essentially on questionnaires to build their learner model. After conducting an empirical study of 50 existing AEHS, revealing their strengths and limitations, some research objectives were identified to improve the learning experience through such systems. These objectives aim to establish an AEHS model which is able to (i) implicitly identify the learning model data on the basis of social networks while meeting the associated standards; (ii) promote collaboration between different learners who would be better motivated to learn while collaborating; (iii) automatically customize new aspects such as the teaching approach, collaboration and feedback according to learners' personality traits in addition to the three existing ones. An AEHS model has been proposed to meet these objectives. This model makes it possible to extract the user's personal data from his social networks and to predict his personality traits depending on his interaction with these networks. Thereafter, it is possible to adapt the learning objects, on the basis of a recommendation system, to these personality traits in addition to the criteria learning style and knowledge level. Adapting to the learner's personality traits according to the Big Five model enabled the customization of new aspects such as the pedagogical approach, the collaboration type and the feedback. A prototype, "ColadaptLearn", based on this model and experimented with a group of students, validated the prototype's choices for learning objects while confronting them to the students' choices. These data were then used to build a Bayesian network to predict the appropriate learning objects for future learners. The experimental results showed that there is a good match between the prototype choices and those of learners, in addition to learners' satisfaction regarding the feedback received, which supports the addition of the proposed new aspects. As a follow-up to this thesis, it is possible to apply the proposed model in a larger learning environment such as massive open online courses (MOOC), serious games or mobile learning, which would help to validate the proposals made. It is also possible to use other automatic learning techniques than Bayesian networks to predict suitable learning objects. Finally, it would be interesting to explore other data sources that could implicitly provide more information about the learner, such as his or her interests or emotions that an SHAE could adapt to.
Bordes, Antoine. "Nouveaux Algorithmes pour l'Apprentissage de Machines à Vecteurs Supports sur de Grandes Masses de Données." Phd thesis, Université Pierre et Marie Curie - Paris VI, 2010. http://tel.archives-ouvertes.fr/tel-00464007.
Hussain, Syed Fawad. "Une Nouvelle Mesure de Co-Similarité : Applications aux Données Textuelles et Génomique." Phd thesis, Grenoble, 2010. http://tel.archives-ouvertes.fr/tel-00525366.