Données semi-structurées. Découverte, maintenance et analyse de tendances

Mining knowledge from structured data has been extensively addressed in the few past years. However, with the growing popularity of the Web, the number of semi structured documents available is rapidly increasing and it is judicious to assume that a query on document structure is almost as important as a query on data. In this paper, we propose an approach to extract such structures. Moreover, manipulated data is constantly being updated; we extend our approach to take into account source evolutions in a knowledge extraction process. Finally, we show that it is possible to analyze trends during the different data sources evolutions. MOTS-CLÉS : données semi structurées, extraction de connaissances, évolutions des sources de données, tendances.