PLearn::ClassDistanceProportionCostFunction Class Reference

#include <ClassDistanceProportionCostFunction.h>

Inheritance diagram for PLearn::ClassDistanceProportionCostFunction:

[legend]Collaboration diagram for PLearn::ClassDistanceProportionCostFunction:

[legend]List of all members.

Public Member Functions
	ClassDistanceProportionCostFunction ()
	PLEARN_DECLARE_OBJECT (ClassDistanceProportionCostFunction)
virtual real	evaluate (const Vec &output, const Vec &target) const
	returns K(x1,x2)
Private Types
typedef Kernel	inherited

---

Detailed Description

This is useful for multiclass problems when the scores represent negative distances to each class, It returns distance_to_right_class/(distance_to_right_class+distance_to_closest_other_classes)

Definition at line 64 of file ClassDistanceProportionCostFunction.h.

---

Member Typedef Documentation

typedef Kernel PLearn::ClassDistanceProportionCostFunction::inherited [private]

Reimplemented from PLearn::Kernel. Definition at line 66 of file ClassDistanceProportionCostFunction.h.

---

Constructor & Destructor Documentation

PLearn::ClassDistanceProportionCostFunction::ClassDistanceProportionCostFunction (   )  [inline]

Definition at line 69 of file ClassDistanceProportionCostFunction.h.

---

Member Function Documentation

real PLearn::ClassDistanceProportionCostFunction::evaluate (  const Vec &  output, const Vec &  target )  const [virtual]

returns K(x1,x2) Implements PLearn::Kernel. Definition at line 51 of file ClassDistanceProportionCostFunction.cc. References PLearn::argmax(), PLearn::TVec< T >::length(), PLearn::max(), PLERROR, and PLearn::Vec.

PLearn::ClassDistanceProportionCostFunction::PLEARN_DECLARE_OBJECT (  ClassDistanceProportionCostFunction   )

---

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

• ClassDistanceProportionCostFunction.h
• ClassDistanceProportionCostFunction.cc

---