L'ultramétrique inférieure maximum d'une dissimilarité à valeurs dans un inf-demi treillis

Mathématiques et sciences humaines. Mathematics and social sciences

Benkaraache, Taoufik¹ 

关键词: terms clustering;    knowledge acquisition;    ontology;    concept learning;    corpus analysis;    statistical data analysis;    text-mining;   

DOI  :  10.4000/msh.2785

学科分类：数学（综合）

来源: College de France * Ecole des Hautes Etudes en Sciences Sociales (E H E S S)

【 摘 要 】

Dissimilarity functions are usually defined to take values in the set R+ of nonnegative reals. This set has various properties as, for instance, those related to the intrinsec linear order of R. But many authors have shown that a number of basic results relative to dissimilarity functions are preserved when R+ is replaced by any partially ordered set. In this paper we propose two methods for the fitting of a given dissimilarity approximation by the subdominant ultrametric when this dissimilarity takes values in a meet semi-lattice.

【 授权许可】

Unknown