NLADesign Class Reference

Simple description of NLA design, viewed from user's perspective. More...

Public Member Functions
	NLADesign ()
	NLADesign (const NLADesign &sd)
	~NLADesign ()
void	clear ()
bool	isSupervised () const
NLADesign &	operator= (const NLADesign &sd)

Static Public Member Functions
static int	finalNrHiddenNodes (int usrsz, int nrinp)

Public Attributes
bool	classification_
TypeSet< int >	hiddensz_
BufferStringSet	inputs_
BufferStringSet	outputs_

Detailed Description

Simple description of NLA design, viewed from user's perspective.

Note: Currently NN only. If hiddensz == 0, it will be set to nrinputs / 2, with a minimum of 1. If nr of outputs == 0, unsupervised network will be assumed. That means the actual nr of output nodes is 2 (segment and match). If classification is true, two extra output nodes will be added ('Classification' and 'Confidence').

<>

Constructor & Destructor Documentation

NLADesign() [1/2]

NLADesign::NLADesign ( )

inline

~NLADesign()

NLADesign::~NLADesign ( )

inline

NLADesign() [2/2]

NLADesign::NLADesign ( const NLADesign &  sd )

inline

Member Function Documentation

clear()

void NLADesign::clear ( )

inline

finalNrHiddenNodes()

static int NLADesign::finalNrHiddenNodes ( int  usrsz, int  nrinp  )

inlinestatic

isSupervised()

bool NLADesign::isSupervised ( ) const

inline

operator=()

NLADesign& NLADesign::operator= ( const NLADesign &  sd )

inline

Member Data Documentation

classification_

bool NLADesign::classification_

hiddensz_

TypeSet<int> NLADesign::hiddensz_

inputs_

BufferStringSet NLADesign::inputs_

outputs_

BufferStringSet NLADesign::outputs_