Dictionary-based compression
Arithmetic algorithms as well as Huffman algorithms are all based on a statistical model, namely an alphabet and the probability distribution of a source. The compression efficiency for a given source depends on the alphabet size and how close its probability distribution of the statistics is to those of the source. The coding method also affects the compression efficiency. For a variable length code, the lengths of the codewords have to satisfy the Kraft inequality in order to be uniquely decodable. This, in one way, provides theoretically guidance on how far a compression algorithm can go; in another, it restricts the performance of these compression algorithms.
In this chapter, we look at a set of algorithms ...
Get Fundamental Data Compression now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
