AlgorithmImpl Class Reference

#include <Algorithm.hh>

Inheritance diagram for AlgorithmImpl:

Collaboration diagram for AlgorithmImpl:

Public Types
typedef std::map< icstring, std::string >	ParamSetType

Public Member Functions
	AlgorithmImpl (AlgMetadata const *metadata)
virtual	~AlgorithmImpl ()
void	setParameters (int nparam, AlgParameter const *param)
void	setParameters (const ParamSetType &)
std::string const	getID () const
AlgMetadata const *	getMetadata () const
AlgorithmPtr	getFreshCopy ()
virtual int	supportsModelProjection () const
Model	createModel (const SamplerPtr &samp, CallbackWrapper *func=0)
void	setSampler (const SamplerPtr &samp)
virtual int	initialize ()=0
virtual int	iterate ()
virtual int	finalize ()
virtual int	done () const
virtual int	getConvergence (Scalar *const val) const
virtual float	getProgress () const
virtual int	needNormalization ()
Normalizer *	getNormalizer () const
void	setNormalization (const SamplerPtr &samp) const
void	setNormalization (const EnvironmentPtr &env) const
virtual Scalar	getValue (const Sample &x) const =0
virtual Model	getModel () const
ConfigurationPtr	getConfiguration () const
void	setConfiguration (const ConstConfigurationPtr &)
Public Member Functions inherited from Configurable
virtual	~Configurable ()

Protected Member Functions
virtual void	_getConfiguration (ConfigurationPtr &) const
virtual void	_setConfiguration (const ConstConfigurationPtr &)
int	dimDomain ()
int	getParameter (std::string const &name, std::string *value)
int	getParameter (std::string const &name, double *value)
int	getParameter (std::string const &name, float *value)
int	getParameter (std::string const &name, int *value)

Protected Attributes
SamplerPtr	_samp
Normalizer *	_normalizerPtr
ParamSetType	_param

Private Types
typedef ParamSetType::value_type	ParamSetValueType

Private Attributes
AlgMetadata const *	_metadata

Friends
class	ReferenceCountedPointer< AlgorithmImpl >
class	ReferenceCountedPointer< const AlgorithmImpl >

Additional Inherited Members
Private Member Functions inherited from ReferenceCountedObject
	ReferenceCountedObject ()
virtual	~ReferenceCountedObject ()=0

Detailed Description

Base class for all distribution modeling algorithms. Provide methods to all derived classes so that they can do data sampling and access the environmental layers to build the model.

Definition at line 76 of file Algorithm.hh.

Member Typedef Documentation

typedef std::map< icstring, std::string > AlgorithmImpl::ParamSetType

Definition at line 84 of file Algorithm.hh.

typedef ParamSetType::value_type AlgorithmImpl::ParamSetValueType

private

Definition at line 257 of file Algorithm.hh.

Constructor & Destructor Documentation

AlgorithmImpl::AlgorithmImpl ( AlgMetadata const *  metadata )

explicit

Definition at line 54 of file Algorithm.cpp.

References Log::debug(), and Log::instance().

Here is the call graph for this function:

AlgorithmImpl::~AlgorithmImpl ( )

virtual

Definition at line 70 of file Algorithm.cpp.

References _normalizerPtr, Log::debug(), and Log::instance().

Here is the call graph for this function:

Member Function Documentation

virtual void AlgorithmImpl::_getConfiguration ( ConfigurationPtr &  ) const

inlineprotectedvirtual

Reimplemented in Csm, AquaMaps, Garp, AbstractBestSubsets, BestSubsets, EnvironmentalDistance, NNAlgorithm, GarpAlgorithm, RfAlgorithm, EnvelopeScore, ConsensusAlgorithm, SvmAlgorithm, NicheMosaic, Bioclim, DistanceToAverage, MinimumDistance, and VirtualNicheGenerator.

Definition at line 199 of file Algorithm.hh.

Referenced by getConfiguration().

virtual void AlgorithmImpl::_setConfiguration ( const ConstConfigurationPtr &  )

inlineprotectedvirtual

Reimplemented in Csm, AquaMaps, Garp, AbstractBestSubsets, BestSubsets, EnvironmentalDistance, NNAlgorithm, GarpAlgorithm, RfAlgorithm, ConsensusAlgorithm, EnvelopeScore, SvmAlgorithm, NicheMosaic, Bioclim, DistanceToAverage, MinimumDistance, and VirtualNicheGenerator.

Definition at line 200 of file Algorithm.hh.

Referenced by setConfiguration().

Model AlgorithmImpl::createModel	(	const SamplerPtr &	samp,
		CallbackWrapper *	func = 0
	)

Definition at line 362 of file Algorithm.cpp.

References _normalizerPtr, _samp, CallbackWrapper::abortionRequested(), Normalizer::computeNormalization(), done(), error, Log::error(), finalize(), getModel(), getProgress(), Log::info(), initialize(), Log::instance(), iterate(), needNormalization(), CallbackWrapper::notifyModelCreationProgress(), setNormalization(), and setSampler().

Here is the call graph for this function:

int AlgorithmImpl::dimDomain ( )

inlineprotected

All protected methods should not be called before initialized(). This is because the sampler object _samp is not initialized in the constructor and the algorithm's parameters were not set yet.Dimension of problem domain: number of independent variables (climatic + soil) added to the number of dependent variables (occurrence prediction).

Definition at line 212 of file Algorithm.hh.

virtual int AlgorithmImpl::done ( ) const

inlinevirtual

Return != 0 if algorithm finished.

Reimplemented in Csm, AquaMaps, Garp, AbstractBestSubsets, BestSubsets, NNAlgorithm, NicheMosaic, RfAlgorithm, EnvelopeScore, ConsensusAlgorithm, Bioclim, DistanceToAverage, MinimumDistance, SvmAlgorithm, GarpAlgorithm, VirtualNicheGenerator, and EnvironmentalDistance.

Definition at line 140 of file Algorithm.hh.

Referenced by createModel().

virtual int AlgorithmImpl::finalize ( )

inlinevirtual

Called after the training phase has finished, ie after done() returned not zero.

Definition at line 137 of file Algorithm.hh.

Referenced by createModel().

ConfigurationPtr AlgorithmImpl::getConfiguration ( ) const

virtual

Implements Configurable.

Definition at line 86 of file Algorithm.cpp.

References _getConfiguration(), _metadata, _normalizerPtr, _param, Configurable::getConfiguration(), AlgMetadata::id, and AlgMetadata::version.

Here is the call graph for this function:

virtual int AlgorithmImpl::getConvergence ( Scalar *const  val ) const

inlinevirtual

Returns the algorithm's convergence value at the moment

Reimplemented in Csm, AquaMaps, Garp, AbstractBestSubsets, BestSubsets, RfAlgorithm, EnvelopeScore, ConsensusAlgorithm, Bioclim, SvmAlgorithm, GarpAlgorithm, and VirtualNicheGenerator.

Definition at line 143 of file Algorithm.hh.

AlgorithmPtr AlgorithmImpl::getFreshCopy

(

)

Returns a copy of this algorithm in the same state as it was before being initialized (ie., no model, same parameters).

Definition at line 254 of file Algorithm.cpp.

References _metadata, _param, Log::error(), AlgMetadata::id, Log::instance(), and AlgorithmFactory::newAlgorithm().

Here is the call graph for this function:

std::string const AlgorithmImpl::getID ( ) const

inline

Definition at line 100 of file Algorithm.hh.

AlgMetadata const* AlgorithmImpl::getMetadata ( ) const

inline

Definition at line 102 of file Algorithm.hh.

Model AlgorithmImpl::getModel ( ) const

virtual

Definition at line 536 of file Algorithm.cpp.

Referenced by createModel().

Normalizer * AlgorithmImpl::getNormalizer

(

)

const

Returns the normalizer used by the algorithm. This method was created for super algorithms like "best subsets", so that the super algorithm can retrieve and use the same normalizer needed by the sub algorithm.

Definition at line 338 of file Algorithm.cpp.

References _normalizerPtr, and Normalizer::getCopy().

Referenced by ConsensusAlgorithm::initialize().

Here is the call graph for this function:

int AlgorithmImpl::getParameter ( std::string const &  name, std::string *  value  )

protected

Returns an algorithm parameter of string type.

Parameters

name	Parameters name.
value	Filled with parameter's value.

Returns

Zero if the parameter does not exists or the parameters were not set yet.

Referenced by AquaMaps::_getAndCheckParameter(), ConsensusAlgorithm::_setConfiguration(), EnvironmentalDistance::_setConfiguration(), Garp::_setConfiguration(), EnvironmentalDistance::initialize(), VirtualNicheGenerator::initialize(), GarpAlgorithm::initialize(), MinimumDistance::initialize(), DistanceToAverage::initialize(), SvmAlgorithm::initialize(), Bioclim::initialize(), ConsensusAlgorithm::initialize(), RfAlgorithm::initialize(), CsmBS::initialize(), NNAlgorithm::initialize(), BestSubsets::initialize(), NicheMosaic::initialize(), Garp::initialize(), AbstractBestSubsets::initialize(), AquaMaps::initialize(), SvmAlgorithm::needNormalization(), AbstractBestSubsets::needNormalization(), DgGarpBestSubsets::transferParametersToAlgorithm(), and GarpBestSubsets::transferParametersToAlgorithm().

int AlgorithmImpl::getParameter ( std::string const &  name, double *  value  )

protected

Returns an algorithm parameter of floating point type.

Parameters

name	Parameters name.
value	Filled with parameter's value.

Returns

Zero if the parameter does not exists or the parameters were not set yet.

int AlgorithmImpl::getParameter ( std::string const &  name, float *  value  )

protected

int AlgorithmImpl::getParameter ( std::string const &  name, int *  value  )

protected

Returns an algorithm parameter of int point type.

Parameters

name	Parameters name.
value	Filled with parameter's value.

Returns

Zero if the parameter does not exists or the parameters were not set yet.

virtual float AlgorithmImpl::getProgress ( ) const

inlinevirtual

Returns progress so far (between 0.0 and 1.0)

Reimplemented in AquaMaps, Garp, AbstractBestSubsets, BestSubsets, NicheMosaic, NNAlgorithm, RfAlgorithm, ConsensusAlgorithm, and GarpAlgorithm.

Definition at line 146 of file Algorithm.hh.

Referenced by createModel().

virtual Scalar AlgorithmImpl::getValue ( const Sample &  x ) const

pure virtual

The algorithm must return the occurrence probability at the given environment conditions.

Parameters

x	Environmental conditions.

Returns

The occurrence probability in the range [0,1].

Implemented in Csm, AquaMaps, Garp, AbstractBestSubsets, BestSubsets, NNAlgorithm, NicheMosaic, RfAlgorithm, EnvelopeScore, ConsensusAlgorithm, Bioclim, DistanceToAverage, MinimumDistance, SvmAlgorithm, GarpAlgorithm, VirtualNicheGenerator, and EnvironmentalDistance.

virtual int AlgorithmImpl::initialize ( )

pure virtual

Initiate a new training.

Implemented in Csm, AquaMaps, AbstractBestSubsets, Garp, NicheMosaic, BestSubsets, NNAlgorithm, CsmBS, EnvelopeScore, RfAlgorithm, ConsensusAlgorithm, Bioclim, DistanceToAverage, MinimumDistance, SvmAlgorithm, GarpAlgorithm, VirtualNicheGenerator, and EnvironmentalDistance.

Referenced by createModel().

virtual int AlgorithmImpl::iterate ( )

inlinevirtual

One step further on the training. Return 0 if something wrong happened.

Reimplemented in Csm, AquaMaps, Garp, AbstractBestSubsets, BestSubsets, NicheMosaic, NNAlgorithm, EnvelopeScore, RfAlgorithm, ConsensusAlgorithm, Bioclim, DistanceToAverage, MinimumDistance, SvmAlgorithm, GarpAlgorithm, and VirtualNicheGenerator.

Definition at line 132 of file Algorithm.hh.

Referenced by createModel().

virtual int AlgorithmImpl::needNormalization ( )

inlinevirtual

The algorithm should return != 0 if it needs normalization of environmental variables (non categorical ones).

Reimplemented in AbstractBestSubsets, NNAlgorithm, RfAlgorithm, ConsensusAlgorithm, and SvmAlgorithm.

Definition at line 155 of file Algorithm.hh.

Referenced by createModel().

void AlgorithmImpl::setConfiguration ( const ConstConfigurationPtr &  config )

virtual

Implements Configurable.

Definition at line 121 of file Algorithm.cpp.

References _normalizerPtr, _param, _setConfiguration(), Log::error(), Log::instance(), Configurable::setConfiguration(), and UNUSED.

Here is the call graph for this function:

void AlgorithmImpl::setNormalization

(

const SamplerPtr &

samp

)

const

Normalize the given environment.

Parameters

samp	Sampler to normalize.

Definition at line 350 of file Algorithm.cpp.

References _normalizerPtr.

Referenced by createModel(), SvmAlgorithm::initialize(), NNAlgorithm::initialize(), and AbstractBestSubsets::needNormalization().

void AlgorithmImpl::setNormalization

(

const EnvironmentPtr &

env

)

const

Normalize the given environment.

Parameters

env	Environment to normalize.

Definition at line 356 of file Algorithm.cpp.

References _normalizerPtr.

void AlgorithmImpl::setParameters	(	int	nparam,
		AlgParameter const *	param
	)

Configure the algorithm's parameters.

Parameters

nparam	Number of parameters.
param	Vector with all parameters. The address 'param' points to must exists while this object is used.

Definition at line 229 of file Algorithm.cpp.

References _param, AlgParameter::id(), and AlgParameter::value().

Here is the call graph for this function:

void AlgorithmImpl::setParameters

(

const ParamSetType &

params

)

Definition at line 245 of file Algorithm.cpp.

References _param.

void AlgorithmImpl::setSampler

(

const SamplerPtr &

samp

)

Set the sampler object. Need to be called before start the algorithm initialization or iteration.

Definition at line 221 of file Algorithm.cpp.

References _samp.

Referenced by createModel().

virtual int AlgorithmImpl::supportsModelProjection ( ) const

inlinevirtual

If algorithm returns 0 then it does not support projection of models to other environmental data layers. By default all algorithms support that feature.

Definition at line 113 of file Algorithm.hh.

Friends And Related Function Documentation

friend class ReferenceCountedPointer< AlgorithmImpl >

friend

Definition at line 79 of file Algorithm.hh.

friend class ReferenceCountedPointer< const AlgorithmImpl >

friend

Definition at line 80 of file Algorithm.hh.

Member Data Documentation

AlgMetadata const* AlgorithmImpl::_metadata

private

Definition at line 253 of file Algorithm.hh.

Referenced by getConfiguration(), and getFreshCopy().

Normalizer* AlgorithmImpl::_normalizerPtr

protected

Definition at line 247 of file Algorithm.hh.

Referenced by createModel(), DistanceToAverage::DistanceToAverage(), EnvironmentalDistance::EnvironmentalDistance(), Garp::Garp(), GarpAlgorithm::GarpAlgorithm(), GarpBestSubsets::GarpBestSubsets(), getConfiguration(), getNormalizer(), SvmAlgorithm::initialize(), NNAlgorithm::initialize(), MinimumDistance::MinimumDistance(), SvmAlgorithm::needNormalization(), AbstractBestSubsets::needNormalization(), NNAlgorithm::NNAlgorithm(), setConfiguration(), setNormalization(), SvmAlgorithm::SvmAlgorithm(), and ~AlgorithmImpl().

ParamSetType AlgorithmImpl::_param

protected

Definition at line 249 of file Algorithm.hh.

Referenced by getConfiguration(), getFreshCopy(), AbstractBestSubsets::iterate(), setConfiguration(), and setParameters().

SamplerPtr AlgorithmImpl::_samp

protected

Definition at line 245 of file Algorithm.hh.

Referenced by NicheMosaic::cleanOccurrences(), Garp::colonize(), createModel(), Garp::crossover(), EnvironmentalDistance::initialize(), VirtualNicheGenerator::initialize(), GarpAlgorithm::initialize(), Bioclim::initialize(), MinimumDistance::initialize(), SvmAlgorithm::initialize(), DistanceToAverage::initialize(), ConsensusAlgorithm::initialize(), RfAlgorithm::initialize(), EnvelopeScore::initialize(), NNAlgorithm::initialize(), Garp::initialize(), NicheMosaic::initialize(), AquaMaps::initialize(), Csm::initialize(), VirtualNicheGenerator::iterate(), ConsensusAlgorithm::iterate(), BestSubsets::iterate(), AbstractBestSubsets::iterate(), AquaMaps::iterate(), SvmAlgorithm::needNormalization(), NNAlgorithm::needNormalization(), AbstractBestSubsets::needNormalization(), Csm::SamplerToMatrix(), and setSampler().

---

The documentation for this class was generated from the following files:

• Algorithm.hh
• Algorithm.cpp