Public Member Functions | Static Public Member Functions | Protected Member Functions | Protected Attributes

CDotFeatures Class Reference


Detailed Description

Features that support dot products among other operations.

DotFeatures support the following operations:

\[r = {\bf x} \cdot {\bf x'}\]

\[r = {\bf x} \cdot {\bf z}\]

\[ {\bf z'} = \alpha {\bf x} + {\bf z} \]

Definition at line 43 of file DotFeatures.h.

Inheritance diagram for CDotFeatures:
Inheritance graph
[legend]

List of all members.

Public Member Functions

 CDotFeatures (int32_t size=0)
 CDotFeatures (const CDotFeatures &orig)
 CDotFeatures (CFile *loader)
virtual ~CDotFeatures ()
virtual int32_t get_dim_feature_space ()=0
virtual float64_t dot (int32_t vec_idx1, int32_t vec_idx2)=0
virtual float64_t dense_dot (int32_t vec_idx1, const float64_t *vec2, int32_t vec2_len)=0
virtual void add_to_dense_vec (float64_t alpha, int32_t vec_idx1, float64_t *vec2, int32_t vec2_len, bool abs_val=false)=0
virtual void dense_dot_range (float64_t *output, int32_t start, int32_t stop, float64_t *alphas, float64_t *vec, int32_t dim, float64_t b)
virtual void dense_dot_range_subset (int32_t *sub_index, int32_t num, float64_t *output, float64_t *alphas, float64_t *vec, int32_t dim, float64_t b)
virtual int32_t get_nnz_features_for_vector (int32_t num)=0
float64_t get_combined_feature_weight ()
void set_combined_feature_weight (float64_t nw)
virtual void get_feature_matrix (float64_t **dst, int32_t *num_feat, int32_t *num_vec)
virtual void get_feature_vector (float64_t **dst, int32_t *len, int32_t num)
void benchmark_add_to_dense_vector (int32_t repeats=5)
void benchmark_dense_dot_range (int32_t repeats=5)
virtual void * get_feature_iterator (int32_t vector_index)=0
virtual bool get_next_feature (int32_t &index, float64_t &value, void *iterator)=0
virtual void free_feature_iterator (void *iterator)=0

Static Public Member Functions

static void * dense_dot_range_helper (void *p)

Protected Member Functions

void display_progress (int32_t start, int32_t stop, int32_t v)

Protected Attributes

float64_t combined_weight
 feature weighting in combined dot features

Constructor & Destructor Documentation

CDotFeatures ( int32_t  size = 0 )

constructor

Parameters:
sizecache size

Definition at line 51 of file DotFeatures.h.

CDotFeatures ( const CDotFeatures orig )

copy constructor

Definition at line 57 of file DotFeatures.h.

CDotFeatures ( CFile loader )

constructor

Parameters:
loaderFile object via which to load data

Definition at line 64 of file DotFeatures.h.

virtual ~CDotFeatures (  ) [virtual]

Definition at line 66 of file DotFeatures.h.


Member Function Documentation

virtual void add_to_dense_vec ( float64_t  alpha,
int32_t  vec_idx1,
float64_t vec2,
int32_t  vec2_len,
bool  abs_val = false 
) [pure virtual]

add vector 1 multiplied with alpha to dense vector2

Parameters:
alphascalar alpha
vec_idx1index of first vector
vec2pointer to real valued vector
vec2_lenlength of real valued vector
abs_valif true add the absolute value

Implemented in CCombinedDotFeatures, CExplicitSpecFeatures, CHashedWDFeatures, CHashedWDFeaturesTransposed, CImplicitWeightedSpecFeatures, CPolyFeatures, CSimpleFeatures< ST >, CSNPFeatures, CSparseFeatures< ST >, CSparsePolyFeatures, CWDFeatures, CSimpleFeatures< float64_t >, CSimpleFeatures< uint16_t >, and CSparseFeatures< float64_t >.

void benchmark_add_to_dense_vector ( int32_t  repeats = 5 )

run benchmark for add_to_dense_vec

Definition at line 290 of file DotFeatures.cpp.

void benchmark_dense_dot_range ( int32_t  repeats = 5 )

run benchmark for dense_dot_range

Definition at line 313 of file DotFeatures.cpp.

virtual float64_t dense_dot ( int32_t  vec_idx1,
const float64_t vec2,
int32_t  vec2_len 
) [pure virtual]

compute dot product between vector1 and a dense vector

Parameters:
vec_idx1index of first vector
vec2pointer to real valued vector
vec2_lenlength of real valued vector

Implemented in CCombinedDotFeatures, CExplicitSpecFeatures, CHashedWDFeatures, CHashedWDFeaturesTransposed, CImplicitWeightedSpecFeatures, CPolyFeatures, CSimpleFeatures< ST >, CSNPFeatures, CSparseFeatures< ST >, CSparsePolyFeatures, CWDFeatures, CSimpleFeatures< float64_t >, CSimpleFeatures< uint16_t >, and CSparseFeatures< float64_t >.

void dense_dot_range ( float64_t output,
int32_t  start,
int32_t  stop,
float64_t alphas,
float64_t vec,
int32_t  dim,
float64_t  b 
) [virtual]

Compute the dot product for a range of vectors. This function makes use of dense_dot alphas[i] * sparse[i]^T * w + b

Parameters:
outputresult for the given vector range
startstart vector range from this idx
stopstop vector range at this idx
alphasscalars to multiply with, may be NULL
vecdense vector to compute dot product with
dimlength of the dense vector
bbias

note that the result will be written to output[0...(stop-start-1)]

Reimplemented in CCombinedDotFeatures, and CHashedWDFeaturesTransposed.

Definition at line 38 of file DotFeatures.cpp.

void * dense_dot_range_helper ( void *  p ) [static]

Compute the dot product for a range of vectors. This function is called by the threads created in dense_dot_range

Reimplemented in CHashedWDFeaturesTransposed.

Definition at line 202 of file DotFeatures.cpp.

void dense_dot_range_subset ( int32_t *  sub_index,
int32_t  num,
float64_t output,
float64_t alphas,
float64_t vec,
int32_t  dim,
float64_t  b 
) [virtual]

Compute the dot product for a subset of vectors. This function makes use of dense_dot alphas[i] * sparse[i]^T * w + b

Parameters:
sub_indexindex for which to compute outputs
numlength of index
outputresult for the given vector range
alphasscalars to multiply with, may be NULL
vecdense vector to compute dot product with
dimlength of the dense vector
bbias

Reimplemented in CCombinedDotFeatures, and CHashedWDFeaturesTransposed.

Definition at line 124 of file DotFeatures.cpp.

void display_progress ( int32_t  start,
int32_t  stop,
int32_t  v 
) [protected]

display progress output

Parameters:
startminimum value
stopmaximum value
vcurrent value

Definition at line 217 of file DotFeatures.h.

virtual float64_t dot ( int32_t  vec_idx1,
int32_t  vec_idx2 
) [pure virtual]

compute dot product between vector1 and vector2, appointed by their indices

Parameters:
vec_idx1index of first vector
vec_idx2index of second vector

Implemented in CCombinedDotFeatures, CExplicitSpecFeatures, CHashedWDFeatures, CHashedWDFeaturesTransposed, CImplicitWeightedSpecFeatures, CPolyFeatures, CSimpleFeatures< ST >, CSNPFeatures, CSparseFeatures< ST >, CSparsePolyFeatures, CWDFeatures, CSimpleFeatures< float64_t >, CSimpleFeatures< uint16_t >, and CSparseFeatures< float64_t >.

virtual void free_feature_iterator ( void *  iterator ) [pure virtual]
float64_t get_combined_feature_weight (  )

get combined feature weight

Returns:
combined feature weight

Definition at line 148 of file DotFeatures.h.

virtual int32_t get_dim_feature_space (  ) [pure virtual]

obtain the dimensionality of the feature space

(not mix this up with the dimensionality of the input space, usually obtained via get_num_features())

Returns:
dimensionality

Implemented in CCombinedDotFeatures, CExplicitSpecFeatures, CHashedWDFeatures, CHashedWDFeaturesTransposed, CImplicitWeightedSpecFeatures, CPolyFeatures, CSimpleFeatures< ST >, CSNPFeatures, CSparseFeatures< ST >, CSparsePolyFeatures, CWDFeatures, CSimpleFeatures< float64_t >, CSimpleFeatures< uint16_t >, and CSparseFeatures< float64_t >.

virtual void* get_feature_iterator ( int32_t  vector_index ) [pure virtual]

iterate over the non-zero features

call get_feature_iterator first, followed by get_next_feature and free_feature_iterator to cleanup

Parameters:
vector_indexthe index of the vector over whose components to iterate over
Returns:
feature iterator (to be passed to get_next_feature)

Implemented in CCombinedDotFeatures, CExplicitSpecFeatures, CHashedWDFeatures, CHashedWDFeaturesTransposed, CImplicitWeightedSpecFeatures, CPolyFeatures, CSimpleFeatures< ST >, CSNPFeatures, CSparseFeatures< ST >, CSparsePolyFeatures, CWDFeatures, CSimpleFeatures< float64_t >, CSimpleFeatures< uint16_t >, and CSparseFeatures< float64_t >.

void get_feature_matrix ( float64_t **  dst,
int32_t *  num_feat,
int32_t *  num_vec 
) [virtual]

get a copy of the feature matrix (in feature space) num_feat,num_vectors are returned by reference

Parameters:
dstdestination to store matrix in
num_featnumber of features (rows of matrix)
num_vecnumber of vectors (columns of matrix)

Reimplemented in CSimpleFeatures< float64_t >.

Definition at line 255 of file DotFeatures.cpp.

void get_feature_vector ( float64_t **  dst,
int32_t *  len,
int32_t  num 
) [virtual]

get a copy of the feature matrix (in feature space) num_feat,num_vectors are returned by reference

Parameters:
dstdestination to store matrix in
lennumber of features
numindex of vector

Reimplemented in CSimpleFeatures< float64_t >.

Definition at line 277 of file DotFeatures.cpp.

virtual bool get_next_feature ( int32_t &  index,
float64_t value,
void *  iterator 
) [pure virtual]

iterate over the non-zero features

call this function with the iterator returned by get_first_feature and call free_feature_iterator to cleanup

Parameters:
indexis returned by reference (-1 when not available)
valueis returned by reference
iteratoras returned by get_first_feature
Returns:
true if a new non-zero feature got returned

Implemented in CCombinedDotFeatures, CExplicitSpecFeatures, CHashedWDFeatures, CHashedWDFeaturesTransposed, CImplicitWeightedSpecFeatures, CPolyFeatures, CSimpleFeatures< ST >, CSNPFeatures, CSparseFeatures< ST >, CSparsePolyFeatures, CWDFeatures, CSimpleFeatures< float64_t >, CSimpleFeatures< uint16_t >, and CSparseFeatures< float64_t >.

virtual int32_t get_nnz_features_for_vector ( int32_t  num ) [pure virtual]

get number of non-zero features in vector

(in case accurate estimates are too expensive overestimating is OK)

Parameters:
numwhich vector
Returns:
number of sparse features in vector

Implemented in CCombinedDotFeatures, CExplicitSpecFeatures, CHashedWDFeatures, CHashedWDFeaturesTransposed, CImplicitWeightedSpecFeatures, CPolyFeatures, CSimpleFeatures< ST >, CSNPFeatures, CSparseFeatures< ST >, CSparsePolyFeatures, CWDFeatures, CSimpleFeatures< float64_t >, CSimpleFeatures< uint16_t >, and CSparseFeatures< float64_t >.

void set_combined_feature_weight ( float64_t  nw )

set combined kernel weight

Parameters:
nwnew combined feature weight

Definition at line 154 of file DotFeatures.h.


Member Data Documentation

feature weighting in combined dot features

Definition at line 248 of file DotFeatures.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