New bounds on D-ary optimal codes |
| |
Authors: | Gonzalo Navarro Nieves Brisaboa |
| |
Affiliation: | a Center for Web Research, Department of Computer Science, Univ. of Chile, Chile b Database Lab., Univ. da Coruña, Facultade de Informática, Spain |
| |
Abstract: | We propose a simple method that, given a symbol distribution, yields upper and lower bounds on the average code length of a D-ary optimal code over that distribution. Thanks to its simplicity, the method permits deriving analytical bounds for families of parametric distributions. We demonstrate this by obtaining new bounds, much better than the existing ones, for Zipf and exponential distributions when D>2. |
| |
Keywords: | Optimal codes Huffman coding Redundancy Frequency distributions D-ary codes Dense codes Information retrieval |
本文献已被 ScienceDirect 等数据库收录! |
|