MultiClass_ECOC Class Reference

Multiclass classification using error-correcting output code. More...

#include <multiclass_ecoc.h>

Inheritance diagram for MultiClass_ECOC:

[legend]Collaboration diagram for MultiClass_ECOC:


Public Member Functions
	MultiClass_ECOC ()
	MultiClass_ECOC (std::istream &is)
virtual const id_t &	id () const
virtual MultiClass_ECOC *	create () const
	Create a new object using the default constructor.
virtual MultiClass_ECOC *	clone () const
	Create a new object by replicating itself.
REAL	model_weight (UINT n) const
const ECOC_TABLE &	ECOC_table () const
void	set_ECOC_table (const ECOC_TABLE &)
void	set_ECOC_table (ECOC_TYPE)
void	set_ECOC_table (UINT, const ECOC_VECTOR &)
UINT	n_class () const
virtual bool	support_weighted_data () const
	Whether the learning model/algorithm supports unequally weighted data.
virtual REAL	c_error (const Output &out, const Output &y) const
	Error measure for classification problems.
virtual void	set_train_data (const pDataSet &, const pDataWgt &=0)
	Set the data set and sample weight to be used in training.
virtual void	train ()
	Train with preset data set and sample weight.
virtual void	reset ()
virtual Output	operator() (const Input &) const
virtual Output	get_output (UINT idx) const
	Get the output of the hypothesis on the idx-th input.
virtual REAL	margin (UINT) const
	Report the (unnormalized) margin of the example i.
virtual REAL	margin_of (const Input &, const Output &) const
	Report the (unnormalized) margin of an example (x, y).
REAL	cost () const
	The in-sample exponential cost defined in my paper.
Protected Member Functions
virtual REAL	ECOC_distance (const Output &, const ECOC_VECTOR &) const
virtual REAL	ECOC_distance (REAL, int, REAL, REAL=0) const
const std::vector< REAL > &	distances (const Input &) const
const std::vector< REAL > &	distances (UINT) const
bool	is_full_partition (const ECOC_VECTOR &) const
	Does the partition only consist of -1 and +1?
virtual void	setup_aux ()
	Prepare auxiliary variables for current n_in_agg.
virtual bool	ECOC_partition (UINT, ECOC_VECTOR &) const
virtual pLearnModel	train_with_partition (ECOC_VECTOR &) const
virtual REAL	assign_weight (const ECOC_VECTOR &, const LearnModel &) const
virtual void	update_aux (const ECOC_VECTOR &)
	Update those auxiliary variables after this round of learning.
virtual bool	serialize (std::ostream &, ver_list &) const
virtual bool	unserialize (std::istream &, ver_list &, const id_t &=NIL_ID)
Protected Attributes
std::vector< REAL >	lm_wgt
	hypothesis weight
ECOC_TABLE	ecoc
ECOC_TYPE	ecoc_type
UINT	nclass
	number of classes
std::vector< REAL >	labels
	class labels
std::vector< UINT >	ex_class
	class number of examples
std::vector< REAL >	local_d

Detailed Description

Multiclass classification using error-correcting output code.

Definition at line 30 of file multiclass_ecoc.h.

Constructor & Destructor Documentation

MultiClass_ECOC ( ) [inline]

Definition at line 44 of file multiclass_ecoc.h.
References LearnModel::set_dimensions().
Referenced by MultiClass_ECOC::clone(), and MultiClass_ECOC::create().

MultiClass_ECOC ( std::istream & is ) [inline, explicit]

Definition at line 46 of file multiclass_ecoc.h.

Member Function Documentation

virtual REAL assign_weight ( const ECOC_VECTOR & ,

const LearnModel &

) const [inline, protected, virtual]

Reimplemented in AdaBoost_ECOC.
Definition at line 121 of file multiclass_ecoc.h.
Referenced by MultiClass_ECOC::train().

REAL c_error ( const Output & out,

const Output & y

) const [virtual]

Error measure for classification problems.

Parameters:

out is the output from the learned hypothesis.

y is the real output.

Returns:
Classification error between out and y. The error measure is not necessary symmetric. A commonly used measure is out != y.

Reimplemented from LearnModel.
Definition at line 124 of file multiclass_ecoc.cpp.
References LABEL_EQUAL, and LearnModel::n_output().

virtual MultiClass_ECOC* clone ( ) const [inline, virtual]

Create a new object by replicating itself.

Returns:
A pointer to the new copy.
The code for a derived class Derived is always
return new Derived(*this);
Though seemingly redundant, it helps to copy an object without knowing the real type of the object.
See also:
C++ FAQ Lite 20.6

Implements Aggregating.
Reimplemented in AdaBoost_ECOC, and AdaBoost_ERP.
Definition at line 51 of file multiclass_ecoc.h.
References MultiClass_ECOC::MultiClass_ECOC().

REAL cost ( ) const

The in-sample exponential cost defined in my paper.

Definition at line 399 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::distances(), MultiClass_ECOC::ex_class, LearnModel::ptd, and LearnModel::ptw.

virtual MultiClass_ECOC* create ( ) const [inline, virtual]

Create a new object using the default constructor.
The code for a derived class Derived is always
return new Derived();

Implements Aggregating.
Reimplemented in AdaBoost_ECOC, and AdaBoost_ERP.
Definition at line 50 of file multiclass_ecoc.h.
References MultiClass_ECOC::MultiClass_ECOC().

const std::vector< REAL > & distances ( UINT ) const [protected]

Definition at line 263 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::ecoc, MultiClass_ECOC::ECOC_distance(), MultiClass_ECOC::get_output(), Aggregating::lm, MultiClass_ECOC::lm_wgt, and Aggregating::n_in_agg.

const std::vector< REAL > & distances ( const Input & ) const [protected]

Note:
It might be unsafe to use a reference to a local variable as the return value

Definition at line 243 of file multiclass_ecoc.cpp.
References Aggregating::lm, MultiClass_ECOC::lm_wgt, Aggregating::n_in_agg, and LearnModel::n_output().
Referenced by MultiClass_ECOC::cost(), MultiClass_ECOC::get_output(), MultiClass_ECOC::margin(), MultiClass_ECOC::margin_of(), MultiClass_ECOC::operator()(), and AdaBoost_ECOC::setup_aux().

REAL ECOC_distance ( REAL ,

int ,

REAL ,

REAL = 0

) const [protected, virtual]

Definition at line 348 of file multiclass_ecoc.cpp.

REAL ECOC_distance ( const Output & ,

const ECOC_VECTOR &

) const [protected, virtual]

Definition at line 337 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::lm_wgt.
Referenced by MultiClass_ECOC::distances().

bool ECOC_partition ( UINT ,

ECOC_VECTOR &

) const [protected, virtual]

Reimplemented in AdaBoost_ECOC, and AdaBoost_ERP.
Definition at line 360 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::ecoc, and Aggregating::size().
Referenced by AdaBoost_ERP::ECOC_partition(), AdaBoost_ECOC::ECOC_partition(), and MultiClass_ECOC::train().

const ECOC_TABLE& ECOC_table ( ) const [inline]

Definition at line 55 of file multiclass_ecoc.h.
References MultiClass_ECOC::ecoc.

Output get_output ( UINT idx ) const [virtual]

Get the output of the hypothesis on the idx-th input.

Note:
It is possible to cache results to save computational effort.

Reimplemented from LearnModel.
Definition at line 307 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::distances(), MultiClass_ECOC::ecoc, GET_BEST_CLASS, MultiClass_ECOC::labels, and LearnModel::ptw.
Referenced by MultiClass_ECOC::distances().

virtual const id_t& id ( ) const [virtual]

Returns:
Class ID string (class name)

Implements Object.
Reimplemented in AdaBoost_ECOC, and AdaBoost_ERP.
Referenced by MultiClass_ECOC::train().

bool is_full_partition ( const ECOC_VECTOR & ) const [protected]

Does the partition only consist of -1 and +1?

Definition at line 353 of file multiclass_ecoc.cpp.
Referenced by AdaBoost_ECOC::smpwgt_with_partition(), AdaBoost_ERP::train_with_partition(), AdaBoost_ECOC::train_with_partition(), and AdaBoost_ECOC::update_aux().

REAL margin ( UINT idx ) const [virtual]

Report the (unnormalized) margin of the example i.

Note:
It is possible to cache results to save computational effort.

Reimplemented from LearnModel.
Definition at line 328 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::distances(), MultiClass_ECOC::ex_class, GET_MARGIN, and LearnModel::ptw.

REAL margin_of ( const Input & ,

const Output &

) const [virtual]

Report the (unnormalized) margin of an example (x, y).

Reimplemented from LearnModel.
Definition at line 320 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::distances(), GET_MARGIN, LABEL2INDEX, LABEL_EQUAL, and MultiClass_ECOC::labels.

REAL model_weight ( UINT n ) const [inline]

Definition at line 54 of file multiclass_ecoc.h.
References MultiClass_ECOC::lm_wgt.

UINT n_class ( ) const [inline]

Definition at line 59 of file multiclass_ecoc.h.
References MultiClass_ECOC::nclass.
Referenced by AdaBoost_ECOC::ECOC_partition(), AdaBoost_ECOC::setup_aux(), and AdaBoost_ECOC::smpwgt_with_partition().

Output operator() ( const Input & ) const [virtual]

Implements LearnModel.
Definition at line 301 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::distances(), GET_BEST_CLASS, and MultiClass_ECOC::labels.

void reset ( ) [virtual]

Todo:
Keep the initial ECOC settings

Reimplemented from Aggregating.
Definition at line 129 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::ecoc, MultiClass_ECOC::ecoc_type, MultiClass_ECOC::lm_wgt, lemga::NO_TYPE, and Aggregating::reset().

bool serialize ( std::ostream & ,

ver_list &

) const [protected, virtual]

Reimplemented from Aggregating.
Definition at line 20 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::ecoc, Aggregating::lm, MultiClass_ECOC::lm_wgt, SERIALIZE_PARENT, and Aggregating::size().

void set_ECOC_table ( UINT ,

const ECOC_VECTOR &

)

Definition at line 79 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::ecoc, and Aggregating::size().

void set_ECOC_table ( ECOC_TYPE )

Definition at line 89 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::ecoc_type, lemga::NO_TYPE, lemga::ONE_VS_ALL, and lemga::ONE_VS_ONE.

void set_ECOC_table ( const ECOC_TABLE & )

Definition at line 72 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::ecoc, Aggregating::max_n_model, Aggregating::set_max_models(), and Aggregating::size().
Referenced by MultiClass_ECOC::train().

void set_train_data ( const pDataSet & pd,

const pDataWgt & pw = 0

) [virtual]

Set the data set and sample weight to be used in training.
If the learning model/algorithm can only do training using uniform sample weight, i.e., support_weighted_data() returns false, a ``boostrapped'' copy of the original data set will be generated and used in the following training. The boostrapping is done by randomly pick samples (with replacement) w.r.t. the given weight pw.
In order to make the life easier, when support_weighted_data() returns true, a null pw will be replaced by a uniformly distributed probability vector. So we have the following invariant
Invariant:
support_weighted_data() == (ptw != 0)

Parameters:

pd gives the data set.

pw gives the sample weight, whose default value is 0.

See also:
support_weighted_data(), train()

Reimplemented from Aggregating.
Definition at line 139 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::ecoc_type, LABEL_EQUAL, MultiClass_ECOC::labels, LearnModel::n_samples, lemga::NO_TYPE, LearnModel::ptd, and Aggregating::set_train_data().

virtual void setup_aux ( ) [inline, protected, virtual]

Prepare auxiliary variables for current n_in_agg.

Reimplemented in AdaBoost_ECOC.
Definition at line 118 of file multiclass_ecoc.h.
Referenced by MultiClass_ECOC::train().

virtual bool support_weighted_data ( ) const [inline, virtual]

Whether the learning model/algorithm supports unequally weighted data.

Returns:
true if supporting; false otherwise. The default is false, just for safety.

See also:
set_train_data()

Reimplemented from LearnModel.
Definition at line 61 of file multiclass_ecoc.h.

void train ( ) [virtual]

Train with preset data set and sample weight.

Implements LearnModel.
Definition at line 194 of file multiclass_ecoc.cpp.
References LearnModel::_n_out, MultiClass_ECOC::assign_weight(), MultiClass_ECOC::ecoc, MultiClass_ECOC::ECOC_partition(), MultiClass_ECOC::id(), MultiClass_ECOC::labels, Aggregating::lm, Aggregating::lm_base, MultiClass_ECOC::lm_wgt, Aggregating::max_n_model, Aggregating::n_in_agg, LearnModel::ptd, LearnModel::ptw, LearnModel::set_dimensions(), MultiClass_ECOC::set_ECOC_table(), MultiClass_ECOC::setup_aux(), MultiClass_ECOC::train_with_partition(), and MultiClass_ECOC::update_aux().

pLearnModel train_with_partition ( ECOC_VECTOR & ) const [protected, virtual]

Reimplemented in AdaBoost_ECOC, and AdaBoost_ERP.
Definition at line 370 of file multiclass_ecoc.cpp.
References dataset::append(), MultiClass_ECOC::ex_class, Aggregating::lm_base, LearnModel::ptd, and LearnModel::ptw.
Referenced by MultiClass_ECOC::train().

bool unserialize ( std::istream & ,

ver_list & ,

const id_t & = NIL_ID

) [protected, virtual]

Reimplemented from Aggregating.
Definition at line 46 of file multiclass_ecoc.cpp.
References MultiClass_ECOC::ecoc, Aggregating::lm, MultiClass_ECOC::lm_wgt, Object::NIL_ID, and UNSERIALIZE_PARENT.

virtual void update_aux ( const ECOC_VECTOR & ) [inline, protected, virtual]

Update those auxiliary variables after this round of learning.

Reimplemented in AdaBoost_ECOC.
Definition at line 124 of file multiclass_ecoc.h.
Referenced by MultiClass_ECOC::train().