Geometry and clustering with metrics derived from separable Bregman divergences
This work addresses clustering and quantization problems in machine learning, but appears incremental as it applies known divergences to existing algorithms.
The paper tackled the problem of quantization and clustering using metrics derived from separable Bregman divergences, which induce Riemannian metric spaces isometric to Euclidean space, and reported experimental performances of various clustering algorithms with respect to these distances.
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.