PLearn::AddCostToLearner Class Reference

#include <AddCostToLearner.h>

Inheritance diagram for PLearn::AddCostToLearner:

[legend]Collaboration diagram for PLearn::AddCostToLearner:

[legend]List of all members.


Public Member Functions
	AddCostToLearner ()
virtual void	build ()
	Simply calls inherited::build() then build_().
virtual void	makeDeepCopyFromShallowCopy (map< const void , void > &copies)
	Transforms a shallow copy into a deep copy.
	PLEARN_DECLARE_OBJECT (AddCostToLearner)
virtual int	outputsize () const
	Returns the size of this learner's output, (which typically may depend on its inputsize(), targetsize() and set options).
virtual void	forget ()
	(Re-)initializes the PLearner in its fresh state (that state may depend on the 'seed' option) And sets 'stage' back to 0 (this is the stage of a fresh learner!).
virtual void	train ()
	The role of the train method is to bring the learner up to stage==nstages, updating the train_stats collector with training costs measured on-line in the process.
virtual void	computeOutput (const Vec &input, Vec &output) const
	Computes the output from the input.
virtual void	computeCostsFromOutputs (const Vec &input, const Vec &output, const Vec &target, Vec &costs) const
	Computes the costs from already computed output.
virtual TVec< string >	getTestCostNames () const
	Returns the names of the costs computed by computeCostsFromOutpus (and thus the test method).
virtual TVec< string >	getTrainCostNames () const
	Returns the names of the objective costs that the train method computes and for which it updates the VecStatsCollector train_stats.
virtual void	setExperimentDirectory (const string &the_expdir)
	Overrridden to forward to the sub_learner.
virtual void	setTrainingSet (VMat training_set, bool call_forget=true)
	Declares the train_set Then calls build() and forget() if necessary Note: You shouldn't have to overload this in subclasses, except in maybe to forward the call to an underlying learner.
virtual void	setTrainStatsCollector (PP< VecStatsCollector > statscol)
	Sets the statistics collector whose update() method will be called during training.
Public Attributes
bool	check_output_consistency
int	combine_bag_outputs_method
bool	compute_costs_on_bags
TVec< string >	costs
bool	force_output_to_target_interval
real	from_max
real	from_min
bool	rescale_output
bool	rescale_target
PP< PLearner >	sub_learner
real	to_max
real	to_min
Static Protected Member Functions
void	declareOptions (OptionList &ol)
	Declares this class' options.
Protected Attributes
Mat	bag_outputs
	Used to store the outputs of the sub_learner for each sample in a bag.
int	bag_size
	Used to count the number of instances in a bag.
VarArray	cross_entropy_prop
	Propagation path for the cross entropy cost.
Var	cross_entropy_var
	Variable giving the cross entropy cost.
Vec	desired_target
	Used to store the target when computing a cost.
real	fac
	A precomputed factor for faster mapping.
real	output_max
	Constraints on the output given the costs being computed.
real	output_min
Var	output_var
	Its value is sub_learner_output[0].
Vec	sub_learner_output
	Used to store the sub_learner_output.
Vec	sub_input
	Used to store the input given to the sub_learner, when it needs to be copied in a separate place.
Var	target_var
	Its value is desired_target[0].
Private Types
typedef PLearner	inherited
Private Member Functions
void	build_ ()
	This does the actual building.
Private Attributes
Vec	combined_output
	Global storage to save memory allocations.

Member Typedef Documentation

typedef PLearner PLearn::AddCostToLearner::inherited [private]

Reimplemented from PLearn::PLearner.
Definition at line 57 of file AddCostToLearner.h.

Constructor & Destructor Documentation

PLearn::AddCostToLearner::AddCostToLearner ( )

Definition at line 80 of file AddCostToLearner.cc.

Member Function Documentation

void PLearn::AddCostToLearner::build ( ) [virtual]

Simply calls inherited::build() then build_().

Reimplemented from PLearn::PLearner.
Definition at line 155 of file AddCostToLearner.cc.
References build_().

void PLearn::AddCostToLearner::build_ ( ) [private]

This does the actual building.

Reimplemented from PLearn::PLearner.
Definition at line 164 of file AddCostToLearner.cc.
References PLearn::accessElement(), bag_outputs, costs, PLearn::cross_entropy(), cross_entropy_prop, cross_entropy_var, desired_target, PLearn::endl(), fac, from_max, from_min, PLearn::TVec< string >::length(), PLearn::max(), PLearn::min(), output_max, output_min, output_var, PLERROR, PLearn::propagationPath(), REAL_MAX, rescale_output, rescale_target, PLearn::TVec< T >::resize(), PLearn::TMat< T >::resize(), sub_learner, sub_learner_output, target_var, to_max, to_min, and PLearn::var().
Referenced by build().

void PLearn::AddCostToLearner::computeCostsFromOutputs ( const Vec & input,

const Vec & output,

const Vec & target,

Vec & costs

) const [virtual]

Computes the costs from already computed output.

Implements PLearn::PLearner.
Definition at line 216 of file AddCostToLearner.cc.
References PLearn::abs(), bag_outputs, bag_size, check_output_consistency, PLearn::TMat< T >::column(), combine_bag_outputs_method, combined_output, compute_costs_on_bags, cross_entropy_prop, cross_entropy_var, desired_target, fac, PLearn::TVec< T >::fill(), force_output_to_target_interval, PLearn::VarArray::fprop(), from_min, PLearn::TMat< T >::length(), PLearn::TVec< T >::length(), PLearn::max(), MISSING_VALUE, output_max, output_min, PLERROR, PLWARNING, PLearn::powdistance(), PLearn::pownorm(), rescale_output, rescale_target, PLearn::TVec< T >::resize(), PLearn::TMat< T >::resize(), sub_learner, sub_learner_output, PLearn::TVec< T >::subVec(), to_max, to_min, and PLearn::TMat< T >::width().

void PLearn::AddCostToLearner::computeOutput ( const Vec & input,

Vec & output

) const [virtual]

Computes the output from the input.

Implements PLearn::PLearner.
Definition at line 380 of file AddCostToLearner.cc.
References sub_learner.

void PLearn::AddCostToLearner::declareOptions ( OptionList & ol ) [static, protected]

Declares this class' options.

Reimplemented from PLearn::PLearner.
Definition at line 96 of file AddCostToLearner.cc.
References PLearn::declareOption(), and PLearn::OptionList.

void PLearn::AddCostToLearner::forget ( ) [virtual]

(Re-)initializes the PLearner in its fresh state (that state may depend on the 'seed' option) And sets 'stage' back to 0 (this is the stage of a fresh learner!).

Implements PLearn::PLearner.
Definition at line 388 of file AddCostToLearner.cc.
References bag_size, and sub_learner.

TVec< string > PLearn::AddCostToLearner::getTestCostNames ( ) const [virtual]

Returns the names of the costs computed by computeCostsFromOutpus (and thus the test method).

Implements PLearn::PLearner.
Definition at line 397 of file AddCostToLearner.cc.
References PLearn::TVec< T >::append(), costs, PLearn::TVec< string >::length(), and sub_learner.

TVec< string > PLearn::AddCostToLearner::getTrainCostNames ( ) const [virtual]

Returns the names of the objective costs that the train method computes and for which it updates the VecStatsCollector train_stats.

Implements PLearn::PLearner.
Definition at line 412 of file AddCostToLearner.cc.
References sub_learner.

void PLearn::AddCostToLearner::makeDeepCopyFromShallowCopy ( map< const void *, void * > & copies ) [virtual]

Transforms a shallow copy into a deep copy.

Definition at line 424 of file AddCostToLearner.cc.
References bag_outputs, combined_output, costs, cross_entropy_prop, cross_entropy_var, PLearn::deepCopyField(), desired_target, output_var, sub_input, sub_learner, sub_learner_output, target_var, and PLearn::varDeepCopyField().

int PLearn::AddCostToLearner::outputsize ( ) const [virtual]

Returns the size of this learner's output, (which typically may depend on its inputsize(), targetsize() and set options).

Implements PLearn::PLearner.
Definition at line 443 of file AddCostToLearner.cc.
References sub_learner.

PLearn::AddCostToLearner::PLEARN_DECLARE_OBJECT ( AddCostToLearner )

virtual void PLearn::AddCostToLearner::setExperimentDirectory ( const string & the_expdir ) [inline, virtual]

Overrridden to forward to the sub_learner.

Reimplemented from PLearn::PLearner.
Definition at line 191 of file AddCostToLearner.h.
References sub_learner.

void PLearn::AddCostToLearner::setTrainingSet ( VMat training_set,

bool call_forget = true

) [virtual]

Declares the train_set Then calls build() and forget() if necessary Note: You shouldn't have to overload this in subclasses, except in maybe to forward the call to an underlying learner.

Reimplemented from PLearn::PLearner.
Definition at line 451 of file AddCostToLearner.cc.
References compute_costs_on_bags, PLearn::VMat::length(), PLERROR, sub_learner, and PLearn::VMat::width().

virtual void PLearn::AddCostToLearner::setTrainStatsCollector ( PP< VecStatsCollector > statscol ) [inline, virtual]

Sets the statistics collector whose update() method will be called during training.
Note: You shouldn't have to overload this in subclasses, except maybe to forward the call to an underlying learner.
Reimplemented from PLearn::PLearner.
Definition at line 195 of file AddCostToLearner.h.
References setTrainStatsCollector(), and sub_learner.
Referenced by setTrainStatsCollector().

void PLearn::AddCostToLearner::train ( ) [virtual]

The role of the train method is to bring the learner up to stage==nstages, updating the train_stats collector with training costs measured on-line in the process.

Implements PLearn::PLearner.
Definition at line 479 of file AddCostToLearner.cc.
References sub_learner.

Member Data Documentation

Mat PLearn::AddCostToLearner::bag_outputs [mutable, protected]

Used to store the outputs of the sub_learner for each sample in a bag.

Definition at line 71 of file AddCostToLearner.h.
Referenced by build_(), computeCostsFromOutputs(), and makeDeepCopyFromShallowCopy().

int PLearn::AddCostToLearner::bag_size [mutable, protected]

Used to count the number of instances in a bag.

Definition at line 74 of file AddCostToLearner.h.
Referenced by computeCostsFromOutputs(), and forget().

bool PLearn::AddCostToLearner::check_output_consistency

Definition at line 111 of file AddCostToLearner.h.
Referenced by computeCostsFromOutputs().

int PLearn::AddCostToLearner::combine_bag_outputs_method

Definition at line 112 of file AddCostToLearner.h.
Referenced by computeCostsFromOutputs().

Vec PLearn::AddCostToLearner::combined_output [mutable, private]

Global storage to save memory allocations.

Definition at line 60 of file AddCostToLearner.h.
Referenced by computeCostsFromOutputs(), and makeDeepCopyFromShallowCopy().

bool PLearn::AddCostToLearner::compute_costs_on_bags

Definition at line 113 of file AddCostToLearner.h.
Referenced by computeCostsFromOutputs(), and setTrainingSet().

TVec<string> PLearn::AddCostToLearner::costs

Definition at line 114 of file AddCostToLearner.h.
Referenced by build_(), getTestCostNames(), and makeDeepCopyFromShallowCopy().

VarArray PLearn::AddCostToLearner::cross_entropy_prop [mutable, protected]

Propagation path for the cross entropy cost.

Definition at line 77 of file AddCostToLearner.h.
Referenced by build_(), computeCostsFromOutputs(), and makeDeepCopyFromShallowCopy().

Var PLearn::AddCostToLearner::cross_entropy_var [protected]

Variable giving the cross entropy cost.

Definition at line 80 of file AddCostToLearner.h.
Referenced by build_(), computeCostsFromOutputs(), and makeDeepCopyFromShallowCopy().

Vec PLearn::AddCostToLearner::desired_target [protected]

Used to store the target when computing a cost.

Definition at line 83 of file AddCostToLearner.h.
Referenced by build_(), computeCostsFromOutputs(), and makeDeepCopyFromShallowCopy().

real PLearn::AddCostToLearner::fac [protected]

A precomputed factor for faster mapping.

Definition at line 86 of file AddCostToLearner.h.
Referenced by build_(), and computeCostsFromOutputs().

bool PLearn::AddCostToLearner::force_output_to_target_interval

Definition at line 115 of file AddCostToLearner.h.
Referenced by computeCostsFromOutputs().

real PLearn::AddCostToLearner::from_max

Definition at line 116 of file AddCostToLearner.h.
Referenced by build_().

real PLearn::AddCostToLearner::from_min

Definition at line 117 of file AddCostToLearner.h.
Referenced by build_(), and computeCostsFromOutputs().

real PLearn::AddCostToLearner::output_max [protected]

Constraints on the output given the costs being computed.

Definition at line 89 of file AddCostToLearner.h.
Referenced by build_(), and computeCostsFromOutputs().

real PLearn::AddCostToLearner::output_min [protected]

Definition at line 90 of file AddCostToLearner.h.
Referenced by build_(), and computeCostsFromOutputs().

Var PLearn::AddCostToLearner::output_var [protected]

Its value is sub_learner_output[0].

Definition at line 93 of file AddCostToLearner.h.
Referenced by build_(), and makeDeepCopyFromShallowCopy().

bool PLearn::AddCostToLearner::rescale_output

Definition at line 118 of file AddCostToLearner.h.
Referenced by build_(), and computeCostsFromOutputs().

bool PLearn::AddCostToLearner::rescale_target

Definition at line 119 of file AddCostToLearner.h.
Referenced by build_(), and computeCostsFromOutputs().

Vec PLearn::AddCostToLearner::sub_input [mutable, protected]

Used to store the input given to the sub_learner, when it needs to be copied in a separate place.

Definition at line 100 of file AddCostToLearner.h.
Referenced by makeDeepCopyFromShallowCopy().