Une approche MDS hybride pour l'exploration visuelle interactive

The paper presents a novel technique for the exploration of an information space where elements are associated with a taxonomy and where each element has an associated attribute vector. A hybrid MDS method is at the core of the method. The underlying algorithm embeds elements in the 2D plane, trying to place “similar” elements close to one another. The dynamic character of the algorithm makes it a well suited tool for interactive exploration. Our first results presented here confirm our approach as a tool relevant for finding merging opportunities between companies based on various economic characteristics. CATEGORIES AND SUBJECT DESCRIPTORS: H.4 [Information Systems Applications]: I.3.6 [Computer Graphics]: Methodology and Techniques − Interaction Techniques. GENERAL TERMS: Algorithms, Management

• Type: Conference
• Year: 2005
• Where: IHM
• Authors: Fabien Jourdan, Guy Melançon, Christophe Douy, Alexandre Gasne
• Keywords: Algorithm Embeds Elements, CATEGORIES AND SUBJECT, Hybrid Mds Method, IHM 2005, Signal Processing