This article is an extended version of a paper presented in the WSOM'2012 conference [1]. We display a combination of factorial projections, SOM algorithm and graph techniques applied to a text mining problem. The corpus contains 8 medieval manuscripts which were used to teach arithmetic techniques to merchants.Among the techniques for Data Analysis, those used for Lexicometry (such as Factorial Analysis) highlight the discrepancies between manuscripts. The reason for this is that they focus on the deviation from the independence between words and manuscripts. Still, we also want to discover and characterize the common vocabulary among the whole corpus.Using the properties of stochastic Kohonen maps, which define neighborhood between inputs in a non-deterministic way, we highlight the words which seem to play a special role in the vocabulary. We call them fickle and use them to improve both Kohonen map robustness and significance of FCA visualization. Finally we use graph algorithmic to exploit this fickleness for classification of words.

L’impact de l’environnement digital sur les pratiques historiennes est généralement réduit à une transformation des conditions de diffusion des produits de l’activité historienne. Nous montrons que le développement de nouvelles techniques de traitement des données a un impact sur la recherche historique qui a une certaine spécificité. Les données historiques sont rarement originellement numériques. La production de données adaptées à l’activité historienne suppose la mise en place de plates-formes complexes dont l’élaboration suppose une collaboration avec des physiciens et des informaticiens. Les données produites sont souvent incomplètes et inégalement documentées, ce qui suppose un paramétrage fin des outils statistiques utilisés, ce qui implique des échanges avec des mathématiciens. Nous en concluons que cette configuration contribue à redessiner la carte des relations professionnelles des historiens

No abstract

In the last two decades many random graph models have been proposed to extract knowledge from networks. Most of them look for communities or, more generally, clusters of vertices with homogeneous connection profiles. While the first models focused on networks with binary edges only, extensions now allow to deal with valued networks. Recently, new models were also introduced in order to characterize connection patterns in networks through mixed memberships. This work was motivated by the need of analyzing a historical network where a partition of the vertices is given and where edges are typed. A known partition is seen as a decomposition of a network into subgraphs that we propose to model using a stochastic model with unknown latent clusters. Each subgraph has its own mixing vector and sees its vertices associated to the clusters. The vertices then connect with a probability depending on the subgraphs only, while the types of edges are assumed to be sampled from the latent clusters. A variational Bayes expectation-maximization algorithm is proposed for inference as well as a model selection criterion for the estimation of the cluster number. Experiments are carried out on simulated data to assess the approach. The proposed methodology is then applied to an ecclesiastical network in Merovingian Gaul. An R code, called Rambo, implementing the inference algorithm is available from the authors upon request.

International audienceIn this paper we present a combination of factorial projections and of SOM algorithm applied to a text mining problem. The corpus consists of 8 medieval texts which were used to teach arithmetic techniques to merchants. Classical Factorial Component Analysis (FCA) gives nice representations of the selected words in association with the texts, but the quality of the representation is poor in the center of the graphs and it is not easy to look for the successive projections to conclude. So using the nice properties of Kohonen maps, we can highlight the words which seems to play a special role in the vocabulary since they are associated with very different words from a map to another. Finally we show that combination of both representations is a powerful help to text analysis