Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
These last years, many analysis methods have been proposed to extract knowledge from social networks. As for the traditional data mining domain, these network-based approaches can be classified according to two main families. The approaches based on predictive modelling, which encompass the techniques that analyse current and historical facts to make predictive assumptions about future or unknown events. The approaches based on descriptive modelling, which cover the set of techniques that aim to summarize the data by identifying some relevant features in order to describe how things organize and actually work. In this paper, we review the main descriptive modelling methods of social networks and show for each of them the resulting useful knowledge on a running example. We particularly emphasize on the most recent methods that combine information available on both the network structure and the node attributes in order to provide original description models taking into account the context.
EISSN: 1877-0509

hal-01312383
https://hal.univ-antilles.fr/hal-01312383
DOI : 10.1016/j.procs.2015.05.505

Éditeur(s) : HAL CCSD
Résumé : International audience
More and more animal species are endangered every day on earth. In order to study their adaptation to world and climate change and their chances of survival, numerous initiatives have been taken that mostly need human intrusion into animal communities. Today mobile devices enable researchers to go beyond this limit. In this paper, we propose an original solution that consists on a new framework for detecting individual songs in a bird population and identifying remotely by this way their collective behavior in movements without human interaction. Movement patterns are elicited by analyzing data collected via wireless sensors fitted with microphone. Whereas similar methods use mobile devices fitted on some specimens, we rather propose fixed sensors. We demonstrate that this solution provides a good answer to technical constraints assessed by the context and we discuss results of experimental simulations that allow to define optimized parameters for the architecture to be set up on the ground. Experimental results are provided and show the relative impact of different parameters such as the number of sensors or the population size on the detection rate.
Proceedings of the 8th International Conference on Advances in Mobile Computing and Multimedia (MOMM 2010)

hal-00601641
https://hal.archives-ouvertes.fr/hal-00601641

Résumé : Les réseaux de capteurs constituent un axe de recherche très fertile ces dernières années. Cette technique se développe dans différents domaines comme l'environnement, l'industrie, le commerce, la médecine, l'armée, etc. Les réseaux de capteurs sont difficiles à concevoir parce qu'ils sont fortement contraints en énergie et que tous les éléments ont potentiellement une influence sur la durée de vie du système. Nous proposons un outil permettant l'aide au bon paramétrage et aux choix de paramètres optimaux pour la stabilité des applications.Dans cette thèse, nous nous sommes intéressés à deux problématiques : une classification des paramètres pour un outil d'aide à la décision pour la configuration d'un réseau de capteurs et la seconde, un outil de test de conformité du système dans un environnement réel. Le document est divisé en deux parties où la première partie est un état de l'art de différents protocoles existants et la deuxième partie décrit notre contribution dans ces deux problématiques.Dans la première contribution, nous avons analysé l'impact de la couverture radio puis de la topologie sur les performances d'un réseau de capteurs. Nous étudions le taux de perte et le niveau d'énergie pour en déduire la stabilité d'une application. Puis nous avons proposé une étude menant à une classification pour notre outil d'aide à la décision. Notre classification est basée sur une étude de divers paramètres de la couche MAC, physique, protocole de routage, nombre de nœuds et type d'application.Dans la deuxième contribution, nous nous sommes focalisés sur une approche pragmatique permettant de tester la conformité d'un réseau de capteurs dans un environnement réel. Pour tester la conformité dans un environnement réel, nous proposons une architecture d'exécution de test sur un réseau de capteurs réel. Ceci dans un but d'assurer un niveau correct de conformité et la stabilité de celui-ci durant son fonctionnement.
Wireless sensor networks is one of the hotest research topic in the last few years. This technology can be applied for different fields such as environment, industry, trading, medicine, military etc. Wireless sensor networks are hard to conceive because they require a lot of energy and because each of its component may have an influence on the lifetime of the whole system. What we suggest is a tool allowing to choose the correct and optimal parameters for the reliability of the applications.In this thesis, we focused on two major problems : firstly, a classification of the parameters for a tool allowing to make decisions about the configuration of a wireless sensors network, and secondly, a tool testing the compliance of the system with a real environment. The document is divided into two parts : the first part states the different protocols that exist, and the second part describes our contributions to those topics.In the first contribution, we analyzed how influential the radio cover and the network topology are on the network performances. Then, we deduced from the study of the loss rate and of the level of energy, the reliability of the application. Next, we suggested a study leading to a classification for our decision making tool. For this classification, we studied various parameters related to the MAC layer, the Physical layer, the network layer, the application layer the number of nodes involved in the network.In the second contribution, we adopted a pragmatic approach so we could test the conformity of a wireless sensors network in a real environment. In order to test its conformity in a real environment, we suggested a structured test execution on a real wireless sensor network. This task has been suggested in order to check the conformance level of the network while it was working.

http://www.theses.fr/2012AGUY0586/document

Éditeur(s) : HAL CCSD RNTI
Résumé : International audience
Nous avons tous déjà eu l’occasion d’effectuer des recherches d’ordremédical sur Internet. Si certains sites spécialisés se refusent à tout diagnosticen ligne, préférant le renvoi vers des professionnels de santé, d’autres en revancheconduisent souvent à des déclarations alarmistes faisant état de situationshumaines difficiles. Dans ce travail, nous étudions l’ampleur de ce phénomèneet montrons que quel que soit le syndrome recherché, les résultats obtenusconduisent toujours à l’énoncé des mots "cancer" ou "tumeur".
RNTI. Extraction et Gestion des Connaissances (EGC), Jan 2014, Rennes, France
Rennes, France

hal-01514251
https://hal.archives-ouvertes.fr/hal-01514251

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
In this paper, we present an original and formal framework, the D2SNet model designed to combine both the social network evolution and the diffusion dynamics among individuals. We have conducted experiments on three social networks that show identical characteristics as real social networks. A formal definition of the model is provided and we describe its implementation in a simulation tool. We represent human behaviors and information dissemination strategies by standard and synthetic scheme. In a first step, we study the impact of network growing strategies only and we highlight important parameters such as the evolution speed and mainly the kind of strategies that favour or not the spread. Then we study a more complete evolution strategy that involves link creation and deletion. We provide a deep analysis on the impact of each parameter such as evolution speed, creation and deletion probabilities and dynamic human behaviors on the diffusion amplitude and coverage. Our study gives a novel and useful insight in the diffusion process in a dynamic context.
ISSN: 0747-5632

hal-00786164
https://hal.archives-ouvertes.fr/hal-00786164
DOI : 10.1016/j.chb.2012.06.004

Éditeur(s) : HAL CCSD springer
Résumé : International audience
In this paper, we address the problem of flood prediction in complex situations. We present an original solution in order to achieve the main goals of accuracy, flexibility and readability. We propose the SM2D modular data driven approach that provides predictive models for each sub-process of a global hydrological process. We show that this solution improves the predictive accuracy regarding a global approach. The originality of our proposition is threefold: (1) the predictive model is defined as a set of aggregate variables that act as classifiers, (2) an evolutionary technique is implemented to find best juries of such classifiers and (3) the flood process complexity problem is addressed by searching for sub-models on sub-processes identified partly by spatial criteria. The solution has proved to perform well on flash flood phenomena of tropical areas known to be hardly predictable. It was indeed successfully applied on a real caribbean river dataset after both preprocessing and preliminary analysis steps presented in the paper.
Lecture notes in computer science
Singapour, Singapore

hal-00878780
https://hal.archives-ouvertes.fr/hal-00878780
DOI : 10.1007/978-3-642-40897-7_13

Éditeur(s) : HAL CCSD
Résumé : International audience
In this work, we focus on the issue of diffusion in dynamic networks. We begin by proposing a generic framework that merges two essential aspects of the problem: diffusion and network dynamics. Then, our framework is used for comparing effects of two aspects of the human behavior, sociability and network dynamics. On the one hand the dynamics of the network is related to the movements of individuals in a social space that lead ephemeral social proximity contacts. On the other hand the diffusion phenomenon is studied through the epidemic paradigm. The results obtained by simulations provide original insight about how and why the "social dynamics" of individuals, related to the proximity in a social space and its evolutions, influences diffusion phenomena through a population.
International Conference on Advances in Social Networks Analysis and Mining
Istanbul, Turkey

hal-00786162
https://hal.archives-ouvertes.fr/hal-00786162
DOI : 10.1109/ASONAM.2012.104

Éditeur(s) : HAL CCSD
Résumé : International audience
The extraction of knowledge from social networks is an area that has experienced significant growth in recent years. Indeed, thanks to the improvement of storage and calculation capacities, and the heterogeneity of data that can currently be extracted, much effort has been made to go beyond traditional knowledge, by proposing new kinds of patterns that take into account the context. However, while many works were interested in designing new patterns of knowledge or in optimizing existing approaches, few studies have been focused in merging patterns and on the useful knowledge emerging from such fusions. In this work, we focus on two network clustering approaches, able to extract two distinct kinds of patterns, and we seek to understand both the intersections that can exist between them and the knowledge that emerges from their fusion. The first is the classical nodes clustering approach that consists in searching for communities into a network. The second is the search for frequent conceptual links, a new link clustering approach that aims identifying frequent links between groups of nodes sharing common attributes. We propose a set of original measures that aim to evaluate the amount of shared information between these patterns when they are extracted from a same network. These measures are applied to three datasets and demonstrate the interest in simultaneously considering several sources of knowledge.
Advances in Knowledge Discovery and Management (AKDM)

hal-01514227
https://hal.archives-ouvertes.fr/hal-01514227
DOI : 10.1007/978-3-319-45763-5_13

Éditeur(s) : HAL CCSD
Résumé : International audience
A key issue for any project that aims to study endangered animals is the counting process. In this work, we propose an algorithm for automatic counting of singing birds in their habitat by using wireless sensors fitted with microphone. Sensors are used to record audio samples of bird songs and samples are analyzed to extract fingerprints that permit to recognize bird species. This paper details the counting algorithm that uses graph theory and audio inputs comparison. Finally, we demonstrate our scheme efficiency through experimentations.
International Symposium on a World of Wireless, Mobile and Multimedia Networks
Lucca, Italy

hal-00618563
https://hal.archives-ouvertes.fr/hal-00618563
DOI : 10.1109/WoWMoM.2011.5986215

Éditeur(s) : HAL CCSD Elseiver
Résumé : International audience
Social networks are used on a daily basis by millions of individuals who post millions of messages on several topics. The data recorded by these networks provide useful information in order to predict or detect events in the real world. Some diffused messages are misinformation or false rumours, and so, can be the cause of panic or stress situations. In order to avoid and anticipate these critical situations and understand the diffusion phenomenon in general, it is necessary to study and model the propagation of the information. In this sense, several varieties of models have been proposed and some researchers have attempted to identify parameters involved in the information diffusion. In this paper, we introduce well-known diffusion models that generally simplify drastically the process and we present also a survey of more advanced works whether recent or not studying factors that influence the information diffusion. Finally, we give essential perspectives of research toward a more realistic coverage of information diffusion phenomena.
International Conference on Ambient Systems, Networks and Technologies( ANT)
Madeira, Portugal

hal-01586512
https://hal.archives-ouvertes.fr/hal-01586512
DOI : 10.1016/j.procs.2017.05.404

Éditeur(s) : HAL CCSD
Résumé : International audience
In this work, we propose a novel approach for the discovery of frequent patterns in a social network on the basis of both vertex attributes and link frequency. With an analogy to the traditional task of mining frequent item sets, we show that the issue addressed can be formulated in terms of a conceptual analysis that elicits conceptual links. A social-based conceptual link is a synthetic representation of a set of links between groups of vertexes that share similar internal properties. We propose a first algorithm that optimizes the search into the concept lattice of conceptual links and extracts maximal frequent conceptual links. We study the performances of our solution and give experimental results obtained on a sample example. Finally we show that the set of conceptual links extracted provides a conceptual view of the social network.
IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM)
Istanbul, Turkey

hal-00786161
https://hal.archives-ouvertes.fr/hal-00786161
DOI : 10.1109/ASONAM.2012.15

Éditeur(s) : HAL CCSD
Résumé : International audience
The paper proposes a new knowledge discovery method called MAX-FLMin for extracting frequent patterns in social networks. Unlike traditional approaches that mainly focus on the network topological structure, the originality of our solution is its ability to exploit information both on the network structure and the attributes of nodes in order to elicit specific regularities that we call "Frequent Links". This kind of patterns provides relevant knowledge about the groups of nodes most connected within the network. First, we detail the method proposed to extract maximal frequent links from social networks. Second, we show how the extracted patterns are used to generate aggregated networks that represent the initial social network with more semantics. Qualitative and quantitative studies are conducted to evaluate the performances of our algorithm in various configurations.
Database (Weston, Conn.) and Expert Systems Applications
Vienna, Austria

hal-00767052
https://hal.archives-ouvertes.fr/hal-00767052
DOI : 10.1007/978-3-642-32600-4_35

Éditeur(s) : HAL CCSD
Résumé : National audience
Traditional network generation models attempt to replicate global structural properties (degree distribution, average distance, clustering coefficient, communities, etc.) through synthetic link formation mechanisms such as triadic closure or preferential attachment. In this work, we study the evolution of a very big communication network coming from mobile telephony and we analyse the link formation process. A first study conducted on the standard mechanisms allows observing that several mechanisms are responsible for the properties observed in this network. In a second study, we characterize more precisely the link formation process by searching for correlations between the probability of creating a new link and some individual properties such as the degree, the clustering coefficient and the age of the nodes.
Conférence sur les modèles et l'analyse des réseaux : Approches mathématiques et informatiques
Saint-Etienne, France

hal-00873990
https://hal.archives-ouvertes.fr/hal-00873990

Éditeur(s) : HAL CCSD
Résumé : Works presented in this report were related to the extraction of interesting models in data mining and in the context of information system modeling.
Les travaux présentés dans ce mémoire, ont été développés sur le thème de la découverte de motifs intéressants à travers la fouille de données et mis en oeuvre dans le cadre de la conception de systèmes d'information. Ils sont essentiellement consacrés aux problèmes soulevés par l'étape de fouille pour la découverte de modèles et de motifs fréquents. Ils sont à la fois d'ordre méthodologique et applicatif.
https://tel.archives-ouvertes.fr/tel-01059407

tel-01059407
https://tel.archives-ouvertes.fr/tel-01059407
https://tel.archives-ouvertes.fr/tel-01059407/document
https://tel.archives-ouvertes.fr/tel-01059407/file/HDR.pdf

Éditeur(s) : HAL CCSD
Résumé : International audience
Geographic Information Systems (GIS) help addressing geographical and environmental issues by providing information about a region or a city as a set of maps (layers), each one displaying information about a given theme like roads, vegetation, tourist spots or museums for instance. By combining different layers on a region, one can associate a given area to characteristics from their related themes. Indeed, the information from two or more layers might be merged and then transformed into a new layer as defined in map "algebra". When a theme vocabulary is organized as a taxonomy with concepts linked by is-a relationships, there are different ways to annotate an area with a concept depending on the level selected into the layer taxonomy. In this paper, we present an heuristic- based approach for an optimal merging of such layers in a GIS. Our goal is to generate new layers which sum up information from several themes in a most useful way. Two optimization criteria are considered, the average size of resulting areas and the average informative value of their resulting annotation. We demonstrate the validity of the proposed solution, firstly, on a formal example, and then, on a real world application.
Proceedings of the 4th International Conference on Advanced Geographic Information Systems, Applications, and Services
Valencia, Spain

hal-00679555
https://hal.archives-ouvertes.fr/hal-00679555

Éditeur(s) : HAL CCSD
Résumé : International audience
In the network modeling area, the most widely used definition of a "pattern" is that of a subgraph, a notion that considers only the network topological structure. While this definition has been very useful for extracting subgraphs frequently found in a network or a set of networks, it does not take into account the node attributes, an intrinsic component of social networks that often provides relevant information on the role or the position of a node in a network. In this paper, we propose a novel approach for extracting frequent patterns in social networks, called frequent link mining, based on the search for particular patterns that combine information on both network structure and node attributes. This kind of patterns, that we call frequent links, provides knowledge on the groups of nodes connected in the social network. In this article, we detail the method proposed for extracting frequent links and discuss its flexibility and its complexity. We demonstrate the efficiency of our solution by carrying out qualitative and quantitative studies.
Advances in Database (Weston, Conn.)s and Information Systems
Poznań, Poland

hal-00767050
https://hal.archives-ouvertes.fr/hal-00767050
DOI : 10.1007/978-3-642-33074-2_28

Éditeur(s) : HAL CCSD
Résumé : International audience
Modelling techniques for river hydrologic forecasting systems have taken advantage of machine learning methods especially for flood prediction. But current solutions, mostly based on artificial neural networks do not always meet end users requirements on the readability and the understandability of predictive models. In this paper, we present a new version of our original solution based on the concept of aggregate variables in order to predict flash flood events from observed water level and/or rain measurements, particularly in the context of caribbean watersheds in which flash flood are much uncertain. We combine aggregate variables in juries. Juries of aggregate variables are trained and tested using a typical 10-fold cross validation scheme. Best juries are searched through an evolutionary approach that is optimized. Different parameters are set up like aggregation periods and jury sizes to prove the efficiency of the proposed approach compared to classical solutions.
Congress on Evolutionary Computation (CEC) 2013
Cancun, Mexico

hal-00847582
https://hal.archives-ouvertes.fr/hal-00847582
DOI : 10.1109/CEC.2013.6557656

Éditeur(s) : HAL CCSD
Résumé : International audience
While the impact of network properties on information spread- ing is now widely studied, in uence of network dynamics is very little known. In this paper, we study how evolution mechanisms tradition- ally observed within social networks can a ect information di usion. We present an approach that merges two models: model of informa- tion di usion through social networks and model of network evolution. Since epidemics provide a reference in application domains of information spreading, we measure the impact of basic network structure changes on epidemic peak value and timing. Then we investigate observed trends in terms of changes appearing in the network structure. Our results provide promising results on how and why network dynamics is a strong param- eter to integrate in requirements for information spreading modelling.
World Summit on the Knowledge Society
Mykonos, Greece

hal-00768405
https://hal.archives-ouvertes.fr/hal-00768405

Éditeur(s) : HAL CCSD Lavoisier
Résumé : International audience
In this paper, we present an heuristic based approach for feature selection in the context of flood prediction. Features are complex variables that represent aggregate values. We apply a preprocessing method on data in order to elicit relevant information that could not be easily accessible initially because it is split through several lines of a dataset. A genetic algorithm is used in order to search for the features that may prove the best performances for flood prediction.
ISSN: 1633-1311

hal-00840738
https://hal.archives-ouvertes.fr/hal-00840738
DOI : 10.3166/isi.16.3.113-139

Éditeur(s) : HAL CCSD Lavoisier
Résumé : National audience
Dans cette étude, nous présentons, une méthode permettant de découvrir, par le biais de variables agrégées et d’un algorithme génétique, des informations non connues a priori et utiles pour la prédiction de crues. Cette méthode se prête particulièrement bien aux données dans lesquelles des informations à propos d’un même élément sont dispersées à travers plusieurs lignes du jeu de données initial.
In this paper, we present an heuristic based approach for feature selection in the context of flood prediction. Features are complex variables that represent aggregate values. We apply a preprocessing method on data in order to elicit relevant information that could not be easily accessible initially because it is split through several lines of a dataset. A genetic algorithm is used in order to search for the features that may prove the best performances for flood prediction.
ISSN: 1633-1311

hal-01184510
https://hal.archives-ouvertes.fr/hal-01184510