M.: Knowledge Extraction from an XML Data Flow: Building a Taxonomy based on Clustering Technique

Abstract Similarity between XML documents can be studied in different ways, in particular depending of the fact that we take into consideration only the structure of the document, its content, or both of them. In this paper we choose a structure-based approach, based on the fuzzyfication of the XML documents representing instances of a given XML data model, belonging to a common domain of interest. After a process of data flow analysis we obtain similarity values between documents, used to create similarity classes representing the domain. Using a clustering technique and a graph based approach it is possible build a taxonomy that can be represented through an ontology language