Sparse Sequential Dirichlet Coding
Information Theory
2012-06-19 v1 math.IT
Abstract
This short paper describes a simple coding technique, Sparse Sequential Dirichlet Coding, for multi-alphabet memoryless sources. It is appropriate in situations where only a small, unknown subset of the possible alphabet symbols can be expected to occur in any particular data sequence. We provide a competitive analysis which shows that the performance of Sparse Sequential Dirichlet Coding will be close to that of a Sequential Dirichlet Coder that knows in advance the exact subset of occurring alphabet symbols. Empirically we show that our technique can perform similarly to the more computationally demanding Sequential Sub-Alphabet Estimator, while using less computational resources.
Cite
@article{arxiv.1206.3618,
title = {Sparse Sequential Dirichlet Coding},
author = {Joel Veness and Marcus Hutter},
journal= {arXiv preprint arXiv:1206.3618},
year = {2012}
}
Comments
7 pages