patternmodel.h File Reference

Contains classes for Pattern Models. More...

#include "patternstore.h"
#include "classencoder.h"
#include "algorithms.h"
#include <limits>
#include <cmath>
#include <cstdint>
#include <map>
#include <set>
#include <sstream>
#include <array>
#include <exception>
#include "bz2stream.h"

Go to the source code of this file.

Classes
class	NoSuchPattern
class	PatternModelOptions
	Options for Pattern Model loading and training. More...
class	PatternModelInterface
	Basic read-only interface for pattern models, abstract base class. More...
class	PatternSetModel
	A pattern model based on an unordered set, does not hold data, only patterns. Very suitable for loading constraint models. More...
class	PatternModel< ValueType, ValueHandler, MapType, PatternType >
	A model mapping patterns to values, gigh-level interface. More...
class	IndexedPatternModel< MapType, PatternType >
	An indexed model mapping patterns to values, high-level interface. This is a specialised subclass of PatternMap. More...
class	PatternPointerModel< ValueType, ValueHandler, MapType >
class	IndexedPatternPointerModel< MapType >

Typedefs
typedef PatternMap< uint32_t, BaseValueHandler< uint32_t >, uint64_t >	t_relationmap
typedef PatternMap< double, BaseValueHandler< double >, uint64_t >	t_relationmap_double
typedef PatternMap< uint32_t, BaseValueHandler< uint32_t >, uint64_t >::iterator	t_relationmap_iterator
typedef PatternMap< double, BaseValueHandler< double >, uint64_t >::iterator	t_relationmap_double_iterator

Enumerations
enum	ModelType {   UNINDEXEDPATTERNMODEL = 10, UNINDEXEDPATTERNPOINTERMODEL = 11, INDEXEDPATTERNMODEL = 20, INDEXEDPATTERNPOINTERMODEL = 21,   PATTERNSETMODEL = 30, PATTERNALIGNMENTMODEL = 40 }
enum	ReverseIndexType { NONE = 0, QUICK = 1, COMPACT = 2 }

Functions
int	getmodeltype (const std::string &filename)
double	comparemodels_loglikelihood (const Pattern pattern, std::vector< PatternModel< uint32_t > * > &models)
void	comparemodels_loglikelihood (std::vector< PatternModel< uint32_t > * > &models, PatternMap< double > *resultmap, bool conjunctiononly=false, std::ostream *output=NULL, ClassDecoder *classdecoder=NULL)

Detailed Description

Contains classes for Pattern Models.

Author

Maarten van Gompel (proycon) proyc.nosp@m.on@a.nosp@m.napro.nosp@m.y.nl

LICENSE

Licensed under GPLv3

DESCRIPTION

Contains classes for Pattern Models

Typedef Documentation

typedef PatternMap<uint32_t,BaseValueHandler<uint32_t>,uint64_t> t_relationmap

A relationmap is just a pattern map, the map is specific for a pattern and holds patterns that are in a relation with the first pattern. The value of the map is an integer that expresses how often the relationship occurs.

typedef PatternMap<double,BaseValueHandler<double>,uint64_t> t_relationmap_double

A relationmap_double is just a pattern map, the map is specific for a pattern and holds patterns that are in a relation with the first pattern. The value of the map is an double that expresses the weight of the relation.

typedef PatternMap<double,BaseValueHandler<double>,uint64_t>::iterator t_relationmap_double_iterator

typedef PatternMap<uint32_t,BaseValueHandler<uint32_t>,uint64_t>::iterator t_relationmap_iterator

Enumeration Type Documentation

enum ModelType

Defines the various types of pattern models

Enumerator
UNINDEXEDPATTERNMODEL
UNINDEXEDPATTERNPOINTERMODEL
INDEXEDPATTERNMODEL
INDEXEDPATTERNPOINTERMODEL
PATTERNSETMODEL
PATTERNALIGNMENTMODEL

enum ReverseIndexType

Defines Reverse Index Types

Enumerator
NONE
QUICK
COMPACT

Function Documentation

double comparemodels_loglikelihood	(	const Pattern	pattern,
		std::vector< PatternModel< uint32_t > * > &	models
	)

Computation of log likelihood for a single pattern, accross corpora.

Computation is as in Rayson and Garside (2000), Comparing corpora using frequency profiling. In proceedings of the workshop on Comparing Corpora, held in conjunction with the 38th annual meeting of the Association for Computational Linguistics (ACL 2000). 1-8 October 2000, Hong Kong, pp. 1 - 6: http://www.comp.lancs.ac.uk/~paul/publications/rg_acl2000.pdf

void comparemodels_loglikelihood	(	std::vector< PatternModel< uint32_t > * > &	models,
		PatternMap< double > *	resultmap,
		bool	conjunctiononly,
		std::ostream *	output,
		ClassDecoder *	classdecoder
	)

Computation of log likelihood between patterns in corpora

Computation is as in Rayson and Garside (2000), Comparing corpora using frequency profiling. In proceedings of the workshop on Comparing Corpora, held in conjunction with the 38th annual meeting of the Association for Computational Linguistics (ACL 2000). 1-8 October 2000, Hong Kong, pp. 1 - 6: http://www.comp.lancs.ac.uk/~paul/publications/rg_acl2000.pdf

int getmodeltype

(

const std::string &

filename

)

Extracts the model type (one of ModelType) from a file