Clustering of conceptual graphs with sparse data

Jean-Gabriel Ganascia

pp. 156-169

in: Karl E. Wolff, Heather D. Pfeiffer, Harry Delugach (eds), Conceptual structures at work, Berlin, Springer, 2004

Abstract

This paper gives a theoretical framework for clustering a set of conceptual graphs characterized by sparse descriptions. The formed clusters are named in an intelligible manner through the concept of stereotype, based on the notion of default generalization. The cognitive model we propose relies on sets of stereotypes and makes it possible to save data in a structured memory.