PUBLIC

NumPerfHoldOutC::NumPerfHoldOutC(RealT) NumPerformanceC::Body(void) NumPerformanceC::Body(void) const NumPerformanceC::Copy(void) const NumPerformanceC::operator=(const NumPerformanceC &) NumPerformanceC::PMC(const NumVLDataSetC &,NumClassifySupervisedC &,const NumErrorC &) const NumPerformanceC::GetInfo(void) const NumPerformanceC::GetName(void) const NumPerformanceC::Save(ostream &) const RCHandleAC::Abstract(void) RCHandleAC::IsHandleType(const DT &) const RCHandleAC::CheckHandleType(const DT &) const RCHandleAC::Save(ostream &) const RCHandleAC::Copy(void) const RCHandleC::IsValid(void) const RCHandleC::IsValidObject(void) const RCHandleC::operator=(const RCHandleC &) RCHandleC::operator==(const RCHandleC &) const RCHandleC::operator!=(const RCHandleC &) const RCHandleC::Hash(void) const RCHandleC::IsConst(void) const RCHandleC::IsNotConst(void) const RCHandleC::SetConst(void) const RCHandleC::Invalidate(void) RCHandleC::Body(void) RCHandleC::Body(void) const

NumPerfHoldOutC
Hold out method for classification performance evaluation
include	"amma/Num/NumPerfHoldOut.hh"
User Level:	Default
Library:	NumPerformance
Example:	testNumPerformance.cc
Section:	Pattern Recognition.Performance Evaluation.Implementation
In Scope:	std

Comments:
Class for performing hold out method for classification performance. The proportion of sample data to be used for testing must be specified. The classifier is then trained using the remaing data and tested using the extracted test set. The PMC is the average error count for all samples in the test set.

public NumPerformanceC

NumPerfHoldOutC(RealT proportion)

Constructor

proportion	amount of sample data to use for test set

NumPerformanceBC & Body()

Access body.

const NumPerformanceBC & Body() const

Access body.

NumPerformanceC Copy() const

Makes a deep copy

NumPerformanceC & operator=(const NumPerformanceC & oth)

Assignment operator

RealT PMC(const NumVLDataSetC & data,NumClassifySupervisedC & classifier,const NumErrorC & error = NumErrorCountC()) const

Determines probability of misclassification

data	data set containing feature vectors and correct labels
classifier	classifier to test
error	error counting method

A derived class must provide a method for splitting the data into test and training data and then determining the average probability of misclassification of the given classifier using the supplied error counting method.

const StringC GetInfo() const

Gets string describing object

const StringC GetName() const

Gets type name of the object

BooleanT Save(ostream & out) const

Writes object to stream, cna be loaded using constructor

RCAbstractC Abstract()

Create an abstract handle to this object.

BooleanT IsHandleType(const DT &) const

Is handle of given type ?

void CheckHandleType(const DT & dummy) const

Check handle type. Throw an expception if not.

BooleanT Save(ostream & out) const

Save to ostream.

RCHandleAC<NumPerformanceBC> Copy() const

Creat a deep copy of this object.
protected:

BooleanT IsValid() const

Check its a valid handle.

BooleanT IsValidObject() const

Check its a valid handle, and that the object is pointed to is also valid.

RCHandleC<NumPerformanceBC> & operator=(const RCHandleC<NumPerformanceBC> & oth)

Assignment.

BooleanT operator==(const RCHandleC<NumPerformanceBC> & oth) const

Comparison operator.

BooleanT operator!=(const RCHandleC<NumPerformanceBC> & oth) const

Comparison operator.

UIntT Hash() const

Hash function.x

BooleanT IsConst() const

Is object constant ?

BooleanT IsNotConst() const

Is object not constant ?

void SetConst(void) const

Lock the object.
This is const as a convience though it actual modified the object, often object you wish to ensure are constant already have const set. protected:

void Invalidate()

Turn this into an invalid handle.

NumPerformanceBC & Body()

Direct access to body.

const NumPerformanceBC & Body() const

Constant access to body.