Contains features for bins of a probability (or a score). More...

#include <merkmal.hh>

Inheritance diagram for BinnedMMGroup:

Public Member Functions
	BinnedMMGroup (string name, int monotonic=0)

void	reset ()

void	removeOrigs ()

void	addProb (Double p)

void	trainBins (int maxNumBins)

void	monotonify ()

void	printBoundaries ()

int	getIndex (Double p)

Double	factor (Double p)

void	write (ostream &out)

void	read (ifstream &in)

Double	getMinProb (float qthresh)

Double *	getFactor (int index)

string	verbalDescription (int index)

void	smooth ()

int	getNumPars ()

Public Member Functions inherited from MMGroup
	MMGroup (string name)

void	registerPars (Parameters *parameters, double alphavalue=1.0)

void	addCount (int index, int count=1)

int	getOffset ()

void	setOffset (int offset)

void	setName (string name)

Public Attributes
int	monotonic

int	nbins

vector< Double >	origprobs

vector< Double >	bb

vector< Double >	avprobs

Additional Inherited Members
Protected Attributes inherited from MMGroup
string	parname

int	numPars

Detailed Description

Contains features for bins of a probability (or a score).

E.g. when a probability is mapped to a piecewise constant function. Used to make features for signal models where a pattern probability previously existed. Avoids overfitting.

Author: Mario Stanke

Member Function Documentation

◆ getFactor()

Double * BinnedMMGroup::getFactor ( int index )

virtual

Implements MMGroup.

◆ getNumPars()

int BinnedMMGroup::getNumPars ( )

inlinevirtual

Reimplemented from MMGroup.

◆ smooth()

void BinnedMMGroup::smooth ( )

inlinevirtual

Reimplemented from MMGroup.

◆ verbalDescription()

string BinnedMMGroup::verbalDescription ( int index )

virtual

Implements MMGroup.

The documentation for this class was generated from the following files:

include/merkmal.hh
src/merkmal.cc

Public Member Functions