In this paper we present a Multi-Agent System (MAS) for extracting and analysing users interaction in a Collaborative Knowledge Management system called KnowCat. The proposed MAS employs Web Use Mining and Web Structure Mining techniques in order to detect the most relevant interactions of the Know-Cat users and therefore should have more weight in the Knowledge Crystallization mechanism of KnowCat. More concretely, the MAS extracts the users interaction information and analyses whether they

more »

... working in the system in an organised or disorganised way. The obtained results in this study give us evidence that organised users contribute positively in a good performance of the Knowledge Crystallization mechanism of KnowCat.