Next: Further Reading/Information
Up: Lossless Compression Algorithms (Entropy
Previous: Lempel-Ziv-Welch (LZW) Algorithm
- Huffman maps
fixed length symbols to variable length codes. Optimal only when symbol
probabilities are powers of 2.
- Arithmetic maps entire message to real number range based on
statistics. Theoretically optimal for long messages, but optimality
depends on data model. Also can be CPU/memory intensive.
- Lempel-Ziv-Welch is a dictionary-based compression method.
It maps a variable number of symbols to a fixed length code.
- Adaptive algorithms do not need a priori estimation of probabilities,
they are more useful in real applications.
Dave Marshall
10/4/2001