Octane v1.01.20 - The Open Compression Toolkit for C++

Modelers

Detailed Description

The Modeler class is for statistical compression methods.

It's job is to continually maintain a probability distribution over the next symbol to be parsed. This probability information can be used by a coder (like an arithmetic or huffman coder) to optimally code the parsed symbol.

Compounds

class OctaneModeler

The Modeler class is for statistical compression methods. More...

class SampleModeler

This is a sample modeler which simply ignores all 'training' information and assigns equal probability to each symbol in the parser (sometimes called a -1 order model). More...

class OctaneModeler_WeightVectored

This is a derived modeler class which has some default functions for modelers which work with a vector of weights as probabilities (very common). More...

class ZeroOrderModeler

A simple zero-order modeler, that builds its probability model by counting frequencies in a training file. More...

Generated on 20 May 2004 by doxygen 1.3.3


Compounds
class	OctaneModeler
	The Modeler class is for statistical compression methods. More...
class	SampleModeler
	This is a sample modeler which simply ignores all 'training' information and assigns equal probability to each symbol in the parser (sometimes called a -1 order model). More...
class	OctaneModeler_WeightVectored
	This is a derived modeler class which has some default functions for modelers which work with a vector of weights as probabilities (very common). More...
class	ZeroOrderModeler
	A simple zero-order modeler, that builds its probability model by counting frequencies in a training file. More...