Geometry and clustering with metrics derived from separable Bregman divergences

Abstract

Separable Bregman divergences induce Riemannian metric spaces that are isometric to the Euclidean space after monotone embeddings. We investigate fixed rate quantization and its codebook Voronoi diagrams, and report on experimental performances of partition-based, hierarchical, and soft clustering algorithms with respect to these Riemann-Bregman distances.

23 pages

References in corpus (2)

• An elementary introduction to information geometry
• Algorithms and Convergence Results of Projection Methods for Inconsistent Feasibility Problems: A Review