Sciences sociales

Géographie et médias

L’agenda géomédiatique international : analyse multidimensionnelle des flux d’actualité

Claude Grasland, Robin Lamarche-Perrin, Benjamin Loveluck, and Hugues Pecout. In L’Espace Géographique, vol. 45, issue 2016/1, p. 25-43. Éditions Belin, Paris, 2016.

Notre image du monde dépend dans une large mesure des flux d’information que nous recevons de l’étranger via les médias de masse. Nous proposons dans cet article un cadre d’analyse quantitative des flux médiatiques – marqueurs possibles des dynamiques contemporaines de mondialisation et de régionalisation – reposant sur le concept d’« agenda géomédiatique ».

L’agenda géomédiatique désigne ici le processus de sélection des unités territoriales qui sont portées à l’attention du public par les médias. Nous présentons pour cela trois modèles permettant d’identifier les ressemblances et les spécificités géographiques et temporelles de différents médias, et ainsi d’analyser la formation de l’actualité internationale selon trois perspectives distinctes.
Lire la suite

Télécharger l’article (FR)

Télécharger l’article (EN)

Télécharger les diapositives


Building Optimal Macroscopic Representations of Complex Multi-agent Systems. Application to the Spatial and Temporal Analysis of International Relations through News Aggregation

Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. In N.T. Nguyen, Ryszard Kowalczyk, Juan M. Corchado, and Javier Bajo (eds.), Transactions on Computational Collective Intelligence, vol. XV, LNCS 8670, p. 1-27. Springer-Verlag Berlin, Heidelberg, 2014.

Identification of International Media Events by Spatial and Temporal Aggregation of RSS Flows of Newspapers. Application to the Case of the Syrian Civil War between May 2011 and December 2012

Timothée Giraud, Claude Grasland, Robin Lamarche-Perrin, Yves Demazeau and Jean-Marc Vincent. In Proceedings of the 18th European Colloquium on Theoretical and Quantitative Geography (ECTQG’13), p. 112-114. 2013.

The design and the debugging of large-scale MAS require abstraction tools in order to work at a macroscopic level of description. Agent aggregation provides such abstractions by reducing the complexity of the system’s microscopic representation. Since it leads to an information loss, such a key process may be extremely harmful for the analysis if poorly executed.

This paper presents measures inherited from information theory to evaluate abstractions and to provide the experts with feedback regarding the quality of generated representations. Several evaluation techniques are applied to the spatial and temporal aggregation of an agent-based model of international relations. The information from on-line newspapers constitutes a complex microscopic representation of the agent states. Our approach is able to evaluate geographical abstractions used by the domain experts in order to provide efficient and meaningful macroscopic representations of the world global state.
Lire la suite

Télécharger l’article

Télécharger les diapositives

The research project GEOMEDIA (ANR Corpus, 2013-2015) elaborates an international observatory of mediatized events, based on the collection of RSS flows feeded by 100 newspapers in French and English languages. The aim of this presentation is (1) to describe the complexity of the information contained in RSS flows according to space, time and media dimensions.

(2) To derive basic solutions for the identification of international events on the basis of time aggregation procedures. (3) To analyze the spatial interactions between countries through an analysis of co-quotations in RSS flows; (4) to check the existence of interactions between time and space dimensions.
Lire la suite

Télécharger l’article

Télécharger les diapositives


Sociologie et réseaux

Multidimensional Outlier Detection in Temporal Interaction Networks: An Application to Political Communication on Twitter

Audrey Wilmet and Robin Lamarche-Perrin. ArXiv, 2019.

In social network Twitter, users can interact with each other and spread information via retweets. These millions of interactions may result in media events whose influence goes beyond Twitter framework. In this paper, we thoroughly explore interactions to provide a better understanding of the emergence of certain trends. First, we consider an interaction on Twitter to be a triplet (s,a,t) meaning that user s, called the spreader, has retweeted a tweet of user a, called the author, at time t. We model this set of interactions as a data cube with three dimensions: spreaders, authors and time.

Then, we provide a method which builds different contexts, where a context is a set of features characterizing the circumstances of an event. Finally, these contexts allow us to find relevant unexpected behaviors, according to several dimensions and various perspectives: a user during a given hour which is abnormal compared to its usual behavior, a relationship between two users which is abnormal compared to all other relationships, etc. We apply our method to a set of retweets related to the 2017 French presidential election and show that one can build interesting insights regarding political organization on Twitter.
Lire la suite

Télécharger l’article

Télécharger les diapositives


Mining Political Opinion on Twitter: Challenges and Opportunities of Multiscale Approaches

Marta Severo and Robin Lamarche-Perrin. In Gilles Bastin and Paola Tubaro (eds.), Revue française de sociologie: “Big Data, Sociétés et Sciences Sociales”. Presses de Sciences Po, Paris, 2018.

Link weights recovery in heterogeneous information networks

Hông-Lan Botterman and Robin Lamarche-Perrin. ArXiv, 2019.

Social research on public opinion has been affected by the recent deluge of new digital data on the Web, from blogs and forums to Facebook pages and Twitter accounts. This fresh type of information useful for mining opinions is emerging as an alternative to traditional techniques, such as opinion polls. Firstly, by building the state of the art of studies of political opinion based on Twitter data, this paper aims at identifying the relationship between the chosen data analysis method and the definition of political opinion implied in these studies.

Secondly, it aims at investigating the feasibility of performing multiscale analysis in digital social research on political opinion by addressing the merits of several methodological techniques, from content-based to interaction-based methods, from statistical to semantic analysis, from supervised to unsupervised approaches. The end result of such an approach is to identify future trends in social science research on political opinion.
Lire la suite

Télécharger l’article (EN)

Télécharger l’article (FR)

Socio-technical systems usually consists of many intertwined networks, each connecting different types of objects (or actors) through a variety of means. As these networks are co-dependent, one can take advantage of this entangled structure to study interaction patterns in a particular network from the information provided by other related networks. A method is hence proposed and tested to recover the weights of missing or unobserved links in heterogeneous information networks (HIN) – abstract representations of systems composed of multiple types of entities and their relations.

Given a pair of nodes in a HIN, this work aims at recovering the exact weight of the incident link to these two nodes, knowing some other links present in the HIN. To do so, probability distributions resulting from path-constrained random walks i.e., random walks where the walker is forced to follow only a specific sequence of node types and edge types, capable to capture specific semantics and commonly called a meta-path, are combined in a linearly fashion in order to approximate the desired result. This method is general enough to compute the link weight between any types of nodes. Experiments on Twitter and bibliographic data show the applicability of the method.
Lire la suite

Télécharger l’article

Télécharger les diapositives