English

Greedy Algorithms for Optimal Distribution Approximation

Information Theory 2016-07-28 v1 math.IT

Abstract

The approximation of a discrete probability distribution t\mathbf{t} by an MM-type distribution p\mathbf{p} is considered. The approximation error is measured by the informational divergence D(tp)\mathbb{D}(\mathbf{t}\Vert\mathbf{p}), which is an appropriate measure, e.g., in the context of data compression. Properties of the optimal approximation are derived and bounds on the approximation error are presented, which are asymptotically tight. It is shown that MM-type approximations that minimize either D(tp)\mathbb{D}(\mathbf{t}\Vert\mathbf{p}), or D(pt)\mathbb{D}(\mathbf{p}\Vert\mathbf{t}), or the variational distance pt1\Vert\mathbf{p}-\mathbf{t}\Vert_1 can all be found by using specific instances of the same general greedy algorithm.

Keywords

Cite

@article{arxiv.1601.06039,
  title  = {Greedy Algorithms for Optimal Distribution Approximation},
  author = {Bernhard C. Geiger and Georg Böcherer},
  journal= {arXiv preprint arXiv:1601.06039},
  year   = {2016}
}

Comments

5 pages

R2 v1 2026-06-22T12:34:56.880Z