ComplexMeasure Class Reference

#include <complexmeasure.h>

Inherits AbstractMeasure.

Inherited by TrackableMeasure.

Inheritance diagram for ComplexMeasure:

Inheritance graph
[legend]
Collaboration diagram for ComplexMeasure:

Collaboration graph
[legend]
List of all members.

Public Member Functions

 ComplexMeasure (const char *measureName, ComplexMeasureMode mode, int numberBins)
 creates a new complex measure.
virtual void addObservable (double &observedValue, double minValue, double maxValue)
 adds a observed variable to the measure.
virtual ~ComplexMeasure ()
virtual void step ()
 defined by AbstractMeasure.

Protected Member Functions

void calculatePInf ()
 calculates the Predictive Information
void updateEntropy (int binNumber)
 updates the entropy.
void computeEntropy ()
 computes the entropy.
void initF ()
 inits F, neccessary after each call of addObservable()

Protected Attributes

std::list< double * > observedValueList
std::list< Discretisizer * > discretisizerList
ComplexMeasureMode mode
int numberBins
long fSize
int historySize
int * binNumberHistory
int historyIndex
int * historyIndexList
int historyIndexNumber
int historyInterval
matrix::SparseArray< long,
int > 
F

Constructor & Destructor Documentation

ComplexMeasure ( const char *  measureName,
ComplexMeasureMode  mode,
int  numberBins 
)

creates a new complex measure.

the calculated things are such like mutual information, entropy, joint entropy and so on. it`s possible to add new ones, see above for the ComplexMeasureModes. Don"t forget! to add observed values! with the method

See also:
addObservable
Parameters:
measureName the name of the measure, needed for PlotOptions and HUDSM
mode the measure you like to have
numberBins in earlier versions named as intervalCount. For complex measures the observedValue has to be discretisized, this does the ComplexMeasure with the class Discretisizer for you.

~ComplexMeasure (  )  [virtual]


Member Function Documentation

void addObservable ( double &  observedValue,
double  minValue,
double  maxValue 
) [virtual]

adds a observed variable to the measure.

Parameters:
observedValue address of the observed value
minValue minimum value the observed value can become
maxValue maximum value the observed value can become

void calculatePInf (  )  [protected]

calculates the Predictive Information

void computeEntropy (  )  [protected]

computes the entropy.

uses the normal rule with O(m*n*o) costs

void initF (  )  [protected]

inits F, neccessary after each call of addObservable()

void step (  )  [virtual]

defined by AbstractMeasure.

This method is called from StatisticTools for updating the measure in every simStep (ODE).

Implements IMeasure.

Reimplemented in TrackableMeasure.

void updateEntropy ( int  binNumber  )  [protected]

updates the entropy.

uses update rule with O(1) costs

Parameters:
binNumber the bin number


Member Data Documentation

int* binNumberHistory [protected]

std::list<Discretisizer*> discretisizerList [protected]

matrix::SparseArray<long, int> F [protected]

long fSize [protected]

int historyIndex [protected]

int* historyIndexList [protected]

int historyIndexNumber [protected]

int historyInterval [protected]

int historySize [protected]

ComplexMeasureMode mode [protected]

int numberBins [protected]

std::list<double*> observedValueList [protected]


The documentation for this class was generated from the following files:
Generated on Fri Oct 30 16:29:02 2009 for Robot Simulator of the Robotics Group for Self-Organization of Control by  doxygen 1.4.7