Multiway dissimilarities and generalizations of hypergraphs without triangles. (Dissimilarités multivoies et généralisations d’hypergraphes sans triangles.) (French) Zbl 0910.62062

Summary: Multiway dissimilarities naturally generalize the usual two-way dissimilarities. In this paper, particular classes of multiway dissimilarities are studied as well as some models for deriving from a dissimilarity of a given number of ways one of a greater or a lower number of ways. An application to clustering systems specification has led to a one-to-one correspondence between a class of multiway dissimilarities and a family of stratified clustering systems.


62H30 Classification and discrimination; cluster analysis (statistical aspects)
05C65 Hypergraphs
91C20 Clustering in the social and behavioral sciences
Full Text: Numdam EuDML