#include <ruleset.hh>

Collaboration diagram for GarpRuleSet:

Public Member Functions
	GarpRuleSet (int size)
	~GarpRuleSet ()
int	size ()
int	numRules ()
void	clear ()
void	trim (int rules)
void	filter (PerfIndex perfIndex, double threshold)
GarpRule *	get (int index)
int	insert (PerfIndex perfIndex, GarpRule *rule)
int	replace (int index, GarpRule *rule)
int	remove (int index)
int	add (GarpRule *rule)
int	findSimilar (GarpRule *rule)
Scalar	getValue (const Sample &x) const
void	performanceSummary (PerfIndex perfIndex, double best, double worst, double *average)
void	gatherRuleSetStats (int gen)
void	log () const
Private Attributes
GarpRule **	_rules
int	_size
int	_numRules

Detailed Description

Stores a set of Garp rules. Consists of a valid Garp model. Provide basic methods for manipulating the set of rules as a whole.

Definition at line 53 of file ruleset.hh.

Constructor & Destructor Documentation

GarpRuleSet::GarpRuleSet ( int size )

Constructor.

Parameters:

size	Defines the maximum number of rules that can be stored in the rule set

Definition at line 54 of file ruleset.cpp.

References _numRules, _rules, _size, and size().

Here is the call graph for this function:

GarpRuleSet::~GarpRuleSet ( )

Definition at line 66 of file ruleset.cpp.

References _rules, and clear().

Here is the call graph for this function:

Member Function Documentation

int GarpRuleSet::add ( GarpRule * rule )

Adds new rule at the end of rule set.

Parameters:

rule	Pointer to rule to be added to rule set. Once rule is added, rule set gains control over memory allocated to rule after insertion, i.e., rule set will deallocate memory for that rule when rule is eliminated.

Returns:: Number of rules after addition on success, 0 on failure.

Definition at line 244 of file ruleset.cpp.

References _numRules, _rules, and _size.

Referenced by Garp::_setConfiguration(), Garp::colonize(), Garp::select(), TEST(), and testHelperAddRules().

void GarpRuleSet::clear ( )

Remove all rules from rule set.

Definition at line 91 of file ruleset.cpp.

References trim().

Referenced by Garp::select(), TEST(), and ~GarpRuleSet().

Here is the call graph for this function:

void GarpRuleSet::filter	(	PerfIndex	perfIndex,
		double	threshold
	)

Eliminates rules with performance below a certain threshold from the rule set.

Parameters:

perfIndex	Index on performance array that will be considered for elimination of rules.
threshold	Minimum performance value a rule must present to remain in the rule set. Rules with performance value below threshold are removed from rule set.

Definition at line 122 of file ruleset.cpp.

References _numRules, and _rules.

Referenced by Garp::iterate(), and TEST().

int GarpRuleSet::findSimilar ( GarpRule * rule )

Checks whether there is rule in the rule set that is similar to a given rule.

Parameters:

rule	Pointer to rule

Returns:: Index of first similar rule found in rule set or -1 if there are no similar rules in rule set.

Note:: : Similarity between rules is defined on the Rule class hierarchy by abstract method Rule->similar() and its concrete implementations in subclasses.

Definition at line 267 of file ruleset.cpp.

References _numRules, and _rules.

Referenced by Garp::keepFittest().

void GarpRuleSet::gatherRuleSetStats ( int gen )

Definition at line 344 of file ruleset.cpp.

References _numRules, _rules, GarpRule::getPerformance(), GarpRule::getPrediction(), PerfUtil, and GarpRule::type().

Here is the call graph for this function:

GarpRule * GarpRuleSet::get ( int index )

Returns pointer to rule at a given position in rule set

Parameters:

index Index of rule to be returned

Returns:: Pointer to rule at given position. This pointer should not be deallocated by caller.

Definition at line 201 of file ruleset.cpp.

References _numRules, and _rules.

Referenced by Garp::_getConfiguration(), Garp::crossover(), Garp::evaluate(), Garp::keepFittest(), Garp::mutate(), printPerfs(), Garp::select(), and TEST().

Scalar GarpRuleSet::getValue ( const Sample & x ) const

Returns rule prediction for a given data vector.

Returns:: Prediction value for the data vector based on GARP model

Definition at line 285 of file ruleset.cpp.

References _numRules, _rules, and GarpRule::getPrediction().

Referenced by Garp::getValue().

Here is the call graph for this function:

int GarpRuleSet::insert	(	PerfIndex	perfIndex,
		GarpRule *	rule
	)

Inserts a rule in the rule set, preserving sort order

Parameters:

perfIndex	Index on performance array that will be considered during insertion to find correct position for new rule in the rule set.
rule	pointer to the new rule being inserted. Rule set gains control over memory allocated to rule after insertion, i.e., rule set will deallocate memory for that rule when rule is eliminated.

Returns:: Index where rule was inserted in the rule set.

Definition at line 165 of file ruleset.cpp.

References _numRules, _rules, and GarpRule::getPerformance().

Referenced by Garp::keepFittest().

Here is the call graph for this function:

void GarpRuleSet::log ( ) const

Definition at line 332 of file ruleset.cpp.

References _numRules, _rules, Log::info(), Log::instance(), and GarpRule::log().

Here is the call graph for this function:

int GarpRuleSet::numRules ( )

Returns the number of rules currently stored in the rule set.

Definition at line 83 of file ruleset.cpp.

References _numRules.

Referenced by Garp::_getConfiguration(), Garp::colonize(), Garp::crossover(), Garp::evaluate(), Garp::keepFittest(), Garp::mutate(), printPerfs(), Garp::select(), and TEST().

void GarpRuleSet::performanceSummary	(	PerfIndex	perfIndex,
		double *	best,
		double *	worst,
		double *	average
	)

Returns performance statistics of the rule set.

Parameters:

perfIndex	Base calculation on this performance array index.
best	Returns best performance value found in rule set.
worst	Returns worst performance value found in rule set.
average	Returns average performance of rules in rule set.

Definition at line 304 of file ruleset.cpp.

References _numRules, _rules, and GarpRule::getPerformance().

Referenced by Garp::iterate(), Garp::select(), and TEST().

Here is the call graph for this function:

int GarpRuleSet::remove ( int index )

Remove rule at a given position in rule set

Parameters:

index Index of rule to be removed.

Returns:: 1 on success, 0 on failure.

Definition at line 225 of file ruleset.cpp.

References _numRules, and _rules.

Referenced by Garp::keepFittest().

int GarpRuleSet::replace	(	int	index,
		GarpRule *	rule
	)

Replaces rule at a given position in rule set by another rule provided

Parameters:

index	Index of rule to be replaced.
rule	Pointer to rule that will replace old one. Once rule is added, rule set gains control over memory allocated to rule after insertion, i.e., rule set will deallocate memory for that rule when rule is eliminated.