In this paper we discuss the problem of finding optimal prefix-free codes for unequal letter costs, a variation of the classical Huffman coding problem. Our problem consists of finding a minimal cost prefix-free code in which the encoding alphabet consists of unequal cost (length) letters, with leng
Optimal binary prefix-condition codes with ordinal probabilities
โ Scribed by Greenberg, Irwin
- Publisher
- John Wiley and Sons
- Year
- 1990
- Tongue
- English
- Weight
- 432 KB
- Volume
- 41
- Category
- Article
- ISSN
- 0002-8231
No coin nor oath required. For personal study only.
โฆ Synopsis
Probability estimates for letters in a finite alphabet are derived in order to apply the Huffman algorithm to determine an optimal binary prefix-condition code. The estimates maximize the entropy subject to constraints specifying second-level orderings of the letters' iikeiihoods. These are compared to estimates obtained from another probability estimation technique described in the decision theory literature using the English aiphabet as an example.
๐ SIMILAR VOLUMES
Decoding convolutional codes with binary digital modulation in intersymbol inter-Jerence (ISI) channels is studied. The receiver structure is a whitened matched filter (WMF) whose transi[br function is determined by the ISI channel. Decoding t~f the output sequence can be per/brined in two steps or