CustomScore Class Reference

Interface, defines the behavior of the scoring function for the client application. More...

#include <CustomScore.h>

Inheritance diagram for CustomScore:

[legend]List of all members.

Public Types
typedef vector< EvScores >	LabelScore
Public Member Functions
vector< FeatVec * >	calcVec (const RegionPrediction *, dsu::Strand_t) const
virtual double	getProb (const AnnotationItem &, EvScores::Bnd_t, dsu::Strand_t) const=0
const vector< DataDescrip * > &	getFeatMaps () const
vector< DataDescrip * > &	getFeatMaps ()
virtual LabelScore	score (const RegionPrediction &, const string &) const=0
virtual void	score (vector< RegionPrediction * > &, const string &) const=0
virtual bool	parseOnly () const
	an unfortunate hack at the moment, allows the client application to turn off the dynamic programming scoring, allows the application to quickly create the evidence sets with out the scoring matrix (might want to move to Options)
Protected Member Functions
	CustomScore (unsigned size=(unsigned) EvScores::eNumBnd)
Protected Attributes
vector< DataDescrip * >	_ddVec
Friends
class	InputData

---

Detailed Description

Interface, defines the behavior of the scoring function for the client application.

Each version of the combiner must derive an instance of this class This is the labeling function that takes an evidence set E (RegionPrediction) returns a 2-tuple with a weight for coding, and a weight for noncoding

The user is expected to access the feature maps which map evidence into their feature vector representation, so that the approriate weights can be applied

---

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

• CustomScore.h
• CustomScore.cpp