Laboratoire des systèmes solaires (L2S). 50 av. du Lac Léman,
73375 LE BOURGET DU LAC - CEDEX
Verleysen, M., Giron, A. and Fertil, B. (2007)
tool for visualization and exploration of
IEEE transactions on Neural Networks, 18(5), pp 1265-1279.
of high-dimensional data is generally achieved by a projection
into a low (usually 2- or 3-) dimensional space. Visualization is
facilitate the understanding of datasets by preserving some
"essential" information. This paper presents DD-HDS (Data Driven
High-Dimensional Scaling), a non-linear Multi-Dimensional Scaling (MDS)
relying on the Force Directed Placement (FDP) paradigm to help
discover features of interest in data sets. Through a specific
distances taking into account the concentration of measures phenomenon,
symmetric handling of short distances in the original and output
method is particularly adapted to the projection of high-dimensional
single user-defined parameter in the optimization procedure implements
compromise between local neighborhood preservation and global mapping.
projection of low- and high-dimensional examples illustrates the
advantages of the proposed algorithm.