New Clustering methods for interval data

Marie Chavent (1) , Francisco de A.T. de Carvahlo (2) , Yves Lechevallier (3) , Rosanna Verde (4)

1 IMB - Institut de Mathématiques de Bordeaux
2 CIn - Centro de Informatica UFPE [Recife]
3 AxIS - Usage-centered design, analysis and improvement of information systems
4 Dipartimento di Strategie Aziendali e Metodologie Quantitative

In this paper we propose two clustering methods for interval data based on the dynamic cluster algorithm. These methods use different homogeneity criteria as well as different kinds of cluster representations (prototypes). Some tools to interpret the final partitions are also introduced. An application of one of the methods concludes the paper.

Mots clés: Dynamic clustering; interval data; distances; prototypes

Domaines: Autres [stat.ML]; Apprentissage [cs.LG]

Marie Chavent, Francisco de A.T. de Carvahlo, Yves Lechevallier, Rosanna Verde. New Clustering methods for interval data. Computational Statistics, 2006, 21, pp.211-229. ⟨10.1007/s00180-006-0260-0⟩. ⟨hal-00260959⟩