Public Member Functions | Protected Member Functions | Protected Attributes

CANOVAKernel Class Reference


Detailed Description

ANOVA (ANalysis Of VAriances) kernel.

Formally described as

\[ K_d(x,z) = \sum_{1\le i_1<i_2<\dots<i_d\le n} \prod_{j=1}^d x_{i_j} z_{i_j} \]

with d(cardinality)=1 by default this function is computed recusively

Definition at line 37 of file ANOVAKernel.h.

Inheritance diagram for CANOVAKernel:
Inheritance graph
[legend]

List of all members.

Public Member Functions

 CANOVAKernel ()
 CANOVAKernel (int32_t cache, int32_t d)
 CANOVAKernel (CSimpleFeatures< float64_t > *l, CSimpleFeatures< float64_t > *r, int32_t d, int32_t cache)
virtual ~CANOVAKernel ()
virtual bool init (CFeatures *l, CFeatures *r)
virtual EKernelType get_kernel_type ()
virtual EFeatureType get_feature_type ()
virtual EFeatureClass get_feature_class ()
virtual const char * get_name () const
int32_t get_cardinality ()
void set_cardinality (int32_t value)
float64_t compute_rec1 (int32_t idx_a, int32_t idx_b)
float64_t compute_rec2 (int32_t idx_a, int32_t idx_b)

Protected Member Functions

virtual float64_t compute (int32_t idx_a, int32_t idx_b)
void init ()
void allocate_arrays ()
virtual void cleanup ()
void register_params ()
virtual void load_serializable_post () throw (ShogunException)

Protected Attributes

int32_t cardinality
 degree parameter of kernel
float64_tDP
 array for compute_recursive1
int32_t DP_len
 length of DP
float64_tKD
 arrays for compute_recursive2
float64_tKS
 KS.
float64_tvec_pow
 vec pow

Constructor & Destructor Documentation

CANOVAKernel (  ) 

default constructor

Definition at line 17 of file ANOVAKernel.cpp.

CANOVAKernel ( int32_t  cache,
int32_t  d 
)

constructor

Parameters:
cache size of cache
d kernel parameter cardinality

Definition at line 23 of file ANOVAKernel.cpp.

CANOVAKernel ( CSimpleFeatures< float64_t > *  l,
CSimpleFeatures< float64_t > *  r,
int32_t  d,
int32_t  cache 
)

constructor

Parameters:
l features left-side
r features right-side
d kernel parameter cardinality
cache cache size

Definition at line 30 of file ANOVAKernel.cpp.

~CANOVAKernel (  )  [virtual]

Definition at line 39 of file ANOVAKernel.cpp.


Member Function Documentation

void allocate_arrays (  )  [protected]

allocate arrays

Definition at line 124 of file ANOVAKernel.cpp.

void cleanup (  )  [protected, virtual]

clean up kernel

Reimplemented from CKernel.

Definition at line 142 of file ANOVAKernel.cpp.

float64_t compute ( int32_t  idx_a,
int32_t  idx_b 
) [protected, virtual]

compute kernel for specific feature vectors corresponding to [idx_a] of left-side and [idx_b] of right-side

Parameters:
idx_a left-side index
idx_b right-side index
Returns:
kernel value

Reimplemented from CDotKernel.

Definition at line 55 of file ANOVAKernel.cpp.

float64_t compute_rec1 ( int32_t  idx_a,
int32_t  idx_b 
)

compute rec 1

Parameters:
idx_a 
idx_b 
Returns:
rec1

Definition at line 75 of file ANOVAKernel.cpp.

float64_t compute_rec2 ( int32_t  idx_a,
int32_t  idx_b 
)

computer rec 2

Parameters:
idx_a 
idx_b 
Returns:
rec2

Definition at line 94 of file ANOVAKernel.cpp.

int32_t get_cardinality (  ) 

getter for degree parameter

Returns:
kernel parameter cardinality

Definition at line 90 of file ANOVAKernel.h.

virtual EFeatureClass get_feature_class (  )  [virtual]
Returns:
class of features

Reimplemented from CDotKernel.

Definition at line 80 of file ANOVAKernel.h.

virtual EFeatureType get_feature_type (  )  [virtual]
Returns:
type of features

Reimplemented from CDotKernel.

Definition at line 75 of file ANOVAKernel.h.

virtual EKernelType get_kernel_type (  )  [virtual]
Returns:
kernel type

Implements CDotKernel.

Definition at line 70 of file ANOVAKernel.h.

virtual const char* get_name ( void   )  const [virtual]
Returns:
name of kernel

Reimplemented from CDotKernel.

Definition at line 85 of file ANOVAKernel.h.

void init ( void   )  [protected]

init

array for compute_recursive1

arrays for compute_recursive2

Reimplemented from CKernel.

Definition at line 113 of file ANOVAKernel.cpp.

bool init ( CFeatures l,
CFeatures r 
) [virtual]

initialize kernel with features

Parameters:
l features left-side
r features right-side
Returns:
true if successful

Reimplemented from CDotKernel.

Definition at line 44 of file ANOVAKernel.cpp.

void load_serializable_post ( void   )  throw (ShogunException) [protected, virtual]

Can (optionally) be overridden to post-initialize some member variables which are not PARAMETER::ADD'ed. Make sure that at first the overridden method BASE_CLASS::LOAD_SERIALIZABLE_POST is called.

Exceptions:
ShogunException Will be thrown if an error occurres.

Reimplemented from CKernel.

Definition at line 158 of file ANOVAKernel.cpp.

void register_params (  )  [protected, virtual]

register params

Reimplemented from CKernel.

Definition at line 164 of file ANOVAKernel.cpp.

void set_cardinality ( int32_t  value  ) 

setter for degree parameter

Parameters:
value kernel parameter cardinality

Definition at line 95 of file ANOVAKernel.h.


Member Data Documentation

int32_t cardinality [protected]

degree parameter of kernel

Definition at line 148 of file ANOVAKernel.h.

float64_t* DP [protected]

array for compute_recursive1

Definition at line 151 of file ANOVAKernel.h.

int32_t DP_len [protected]

length of DP

Definition at line 153 of file ANOVAKernel.h.

float64_t* KD [protected]

arrays for compute_recursive2

Definition at line 156 of file ANOVAKernel.h.

float64_t* KS [protected]

KS.

Definition at line 158 of file ANOVAKernel.h.

float64_t* vec_pow [protected]

vec pow

Definition at line 160 of file ANOVAKernel.h.


The documentation for this class was generated from the following files:
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Defines

SHOGUN Machine Learning Toolbox - Documentation