KohonenAnalyzer Class Reference

#include <KohonenAnalyzer.h>

Inheritance diagram for KohonenAnalyzer:

[legend]Collaboration diagram for KohonenAnalyzer:

[legend]List of all members.


Public Member Functions
	KohonenAnalyzer (uint inputs, const Outs &outputs, PublicValues &ctrl=defaultControl)
virtual void	draw ()
	Shows the current state of the kohonen network (draw point for the most active neuron).
virtual void	calibrate ()=0
	Train the analyzing kohonen while training the mlp.
virtual void	trainAndAnalyze (uint epochs)=0
	do the analysis
uint	getInputCount () const
const KohonenNetwork &	getNet () const
Protected Types
typedef EucledianTopology< AXES >	MyTopology
Protected Member Functions
void	calibrateStep (const Vector &in)
	do one training step. doesn't iterateTime(), caller must take care of it himself
void	analyzeStep (const Vector &in)
Protected Attributes
PublicValues &	control
MyTopology	_topology
StandardKohonenParameters	_parameters
KohonenNetwork	_net

Member Typedef Documentation

typedef EucledianTopology<AXES> MyTopology [protected]

Constructor & Destructor Documentation

KohonenAnalyzer ( uint inputs,

const Outs & outputs,

PublicValues & ctrl = defaultControl

) [inline]

Member Function Documentation

void analyzeStep ( const Vector & in ) [protected]

virtual void calibrate ( ) [pure virtual]

Train the analyzing kohonen while training the mlp.

Implemented in MLPKohonenAnalyzer.

void calibrateStep ( const Vector & in ) [protected]

do one training step. doesn't iterateTime(), caller must take care of it himself

virtual void draw ( ) [virtual]

Shows the current state of the kohonen network (draw point for the most active neuron).

Implements Video::Redrawer.