Résumé
The use of XML in data exchanges has extensively grown for the last few years. Furthermore, huge volumes of data are sent through the Internet. Today data mining tools are required in order to provide users with automatic tools to deal with the heterogeneity of the data in order to query it. Some approaches have been proposed in order to mine sub-structures (subtrees) from such XML databases. However, they are usually based on crisp methods. Here we propose to soften these approaches to extract more understandable and relevant knowledge.