X-Boost  2.3.8
Public Member Functions | Public Attributes | List of all members
WeightedPattern< Descriptor > Struct Template Reference

#include <WeightedPattern.h>

Inherits Pattern< Descriptor >.

Public Member Functions

 WeightedPattern (const Pattern< Descriptor > &src)
 
 WeightedPattern (const Descriptor &des, int cat)
 
float GetWeightedCategory () const
 
bool correct () const
 
- Public Member Functions inherited from Pattern< Descriptor >
 Pattern (const Descriptor &d, int cat)
 

Public Attributes

double d
 Current weight associated to this pattern.
 
int test
 Result on the last evaluate (-1,0,1) of Weak Classifier (binary problem)
 
- Public Attributes inherited from Pattern< Descriptor >
int category
 Pattern category (0..k-1 for multiclass problem, or -1|+1 for binary problem)
 
float weight
 weight associated to this pattern
 

Additional Inherited Members

- Public Types inherited from Pattern< Descriptor >
typedef Descriptor DataType
 Data Type.
 

Detailed Description

template<class Descriptor>
struct WeightedPattern< Descriptor >

Extend Pattern with a weight. All information related to input pattern used for AdaBoost training. contains both static information and dynamic information. Extend a pattern with data required by AdaBoost or similiar algorithm

See also
Pattern

Member Function Documentation

template<class Descriptor >
bool WeightedPattern< Descriptor >::correct ( ) const
inline

test if last weak classifier, classifies correctly this pattern

Note
abstension not works
template<class Descriptor >
float WeightedPattern< Descriptor >::GetWeightedCategory ( ) const
inline

return weighted category (overriding unweighted one) [Binary Problem] [AdaBoost]


The documentation for this struct was generated from the following file: