CSparseFeatures< ST > Class Template Reference


Detailed Description

template<class ST>
class shogun::CSparseFeatures< ST >

Template class SparseFeatures implements sparse matrices.

Features are an array of TSparse, sorted w.r.t. vec_index (increasing) and withing same vec_index w.r.t. feat_index (increasing);

Sparse feature vectors can be accessed via get_sparse_feature_vector() and should be freed (this operation is a NOP in most cases) via free_sparse_feature_vector().

As this is a template class it can directly be used for different data types like sparse matrices of real valued, integer, byte etc type.

Definition at line 74 of file SparseFeatures.h.

Inheritance diagram for CSparseFeatures< ST >:
Inheritance graph
[legend]

List of all members.

Public Member Functions

 CSparseFeatures (int32_t size=0)
 CSparseFeatures (TSparse< ST > *src, int32_t num_feat, int32_t num_vec, bool copy=false)
 CSparseFeatures (ST *src, int32_t num_feat, int32_t num_vec)
 CSparseFeatures (const CSparseFeatures &orig)
 CSparseFeatures (char *fname)
virtual ~CSparseFeatures ()
void free_sparse_feature_matrix ()
void free_sparse_features ()
virtual CFeaturesduplicate () const
ST get_feature (int32_t num, int32_t index)
ST * get_full_feature_vector (int32_t num, int32_t &len)
void get_full_feature_vector (ST **dst, int32_t *len, int32_t num)
virtual int32_t get_nnz_features_for_vector (int32_t num)
TSparseEntry< ST > * get_sparse_feature_vector (int32_t num, int32_t &len, bool &vfree)
ST sparse_dot (ST alpha, TSparseEntry< ST > *avec, int32_t alen, TSparseEntry< ST > *bvec, int32_t blen)
ST dense_dot (ST alpha, int32_t num, ST *vec, int32_t dim, ST b)
void add_to_dense_vec (float64_t alpha, int32_t num, float64_t *vec, int32_t dim, bool abs_val=false)
void free_sparse_feature_vector (TSparseEntry< ST > *feat_vec, int32_t num, bool free)
TSparse< ST > * get_sparse_feature_matrix (int32_t &num_feat, int32_t &num_vec)
void get_sparse_feature_matrix (TSparse< ST > **dst, int32_t *num_feat, int32_t *num_vec, int64_t *nnz)
void clean_tsparse (TSparse< ST > *sfm, int32_t num_vec)
TSparse< ST > * get_transposed (int32_t &num_feat, int32_t &num_vec)
virtual void set_sparse_feature_matrix (TSparse< ST > *src, int32_t num_feat, int32_t num_vec)
ST * get_full_feature_matrix (int32_t &num_feat, int32_t &num_vec)
void get_full_feature_matrix (ST **dst, int32_t *num_feat, int32_t *num_vec)
virtual bool set_full_feature_matrix (ST *src, int32_t num_feat, int32_t num_vec)
virtual bool apply_preproc (bool force_preprocessing=false)
virtual int32_t get_size ()
bool obtain_from_simple (CSimpleFeatures< ST > *sf)
virtual int32_t get_num_vectors ()
int32_t get_num_features ()
int32_t set_num_features (int32_t num)
virtual EFeatureClass get_feature_class ()
virtual EFeatureType get_feature_type ()
void free_feature_vector (TSparseEntry< ST > *feat_vec, int32_t num, bool free)
int64_t get_num_nonzero_entries ()
float64_tcompute_squared (float64_t *sq)
float64_t compute_squared_norm (CSparseFeatures< float64_t > *lhs, float64_t *sq_lhs, int32_t idx_a, CSparseFeatures< float64_t > *rhs, float64_t *sq_rhs, int32_t idx_b)
CLabelsload_svmlight_file (char *fname, bool do_sort_features=true)
void sort_features ()
bool write_svmlight_file (char *fname, CLabels *label)
virtual int32_t get_dim_feature_space ()
virtual float64_t dot (int32_t vec_idx1, int32_t vec_idx2)
virtual float64_t dense_dot (int32_t vec_idx1, const float64_t *vec2, int32_t vec2_len)
virtual const char * get_name () const
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()
template<>
EFeatureType get_feature_type ()

Protected Member Functions

virtual TSparseEntry< ST > * compute_sparse_feature_vector (int32_t num, int32_t &len, TSparseEntry< ST > *target=NULL)

Protected Attributes

int32_t num_vectors
 total number of vectors
int32_t num_features
 total number of features
TSparse< ST > * sparse_feature_matrix
 array of sparse vectors of size num_vectors
CCache< TSparseEntry< ST > > * feature_cache

Constructor & Destructor Documentation

CSparseFeatures ( int32_t  size = 0  ) 

constructor

Parameters:
size cache size

Definition at line 81 of file SparseFeatures.h.

CSparseFeatures ( TSparse< ST > *  src,
int32_t  num_feat,
int32_t  num_vec,
bool  copy = false 
)

convenience constructor that creates sparse features from the ones passed as argument

Parameters:
src dense feature matrix
num_feat number of features
num_vec number of vectors
copy true to copy feature matrix

Definition at line 94 of file SparseFeatures.h.

CSparseFeatures ( ST *  src,
int32_t  num_feat,
int32_t  num_vec 
)

convenience constructor that creates sparse features from dense features

Parameters:
src dense feature matrix
num_feat number of features
num_vec number of vectors

Definition at line 120 of file SparseFeatures.h.

CSparseFeatures ( const CSparseFeatures< ST > &  orig  ) 

copy constructor

Definition at line 128 of file SparseFeatures.h.

CSparseFeatures ( char *  fname  ) 

constructor

Parameters:
fname filename to load features from

Definition at line 152 of file SparseFeatures.h.

virtual ~CSparseFeatures (  )  [virtual]

Definition at line 157 of file SparseFeatures.h.


Member Function Documentation

void add_to_dense_vec ( float64_t  alpha,
int32_t  num,
float64_t vec,
int32_t  dim,
bool  abs_val = false 
) [virtual]

add a sparse feature vector onto a dense one dense+=alpha*sparse

Parameters:
alpha scalar to multiply with
num index of feature vector
vec dense vector
dim length of the dense vector
abs_val if true, do dense+=alpha*abs(sparse)

Implements CDotFeatures.

Definition at line 469 of file SparseFeatures.h.

virtual bool apply_preproc ( bool  force_preprocessing = false  )  [virtual]

apply preprocessor

Parameters:
force_preprocessing if preprocssing shall be forced
Returns:
if applying was successful

Definition at line 805 of file SparseFeatures.h.

void clean_tsparse ( TSparse< ST > *  sfm,
int32_t  num_vec 
)

clean TSparse

Parameters:
sfm sparse feature matrix
num_vec number of vectors in matrix

Definition at line 547 of file SparseFeatures.h.

virtual TSparseEntry<ST>* compute_sparse_feature_vector ( int32_t  num,
int32_t &  len,
TSparseEntry< ST > *  target = NULL 
) [protected, virtual]

compute feature vector for sample num if target is set the vector is written to target len is returned by reference

NOT IMPLEMENTED!

Parameters:
num num
len len
target target

Definition at line 1346 of file SparseFeatures.h.

float64_t* compute_squared ( float64_t sq  ) 

compute a^2 on all feature vectors

Parameters:
sq the square for each vector is stored in here
Returns:
the square for each vector

Definition at line 928 of file SparseFeatures.h.

float64_t compute_squared_norm ( CSparseFeatures< float64_t > *  lhs,
float64_t sq_lhs,
int32_t  idx_a,
CSparseFeatures< float64_t > *  rhs,
float64_t sq_rhs,
int32_t  idx_b 
)

compute (a-b)^2 (== a^2+b^2+2ab) usually called by kernels'/distances' compute functions works on two feature vectors, although it is a member of a single feature: can either be called by lhs or rhs.

Parameters:
lhs left-hand side features
sq_lhs squared values of left-hand side
idx_a index of left-hand side's vector to compute
rhs right-hand side features
sq_rhs squared values of right-hand side
idx_b index of right-hand side's vector to compute

Definition at line 961 of file SparseFeatures.h.

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

compute dot product between vector1 and a dense vector

Parameters:
vec_idx1 index of first vector
vec2 pointer to real valued vector
vec2_len length of real valued vector

Implements CDotFeatures.

Definition at line 1311 of file SparseFeatures.h.

ST dense_dot ( ST  alpha,
int32_t  num,
ST *  vec,
int32_t  dim,
ST  b 
)

compute the dot product between dense weights and a sparse feature vector alpha * sparse^T * w + b

Parameters:
alpha scalar to multiply with
num index of feature vector
vec dense vector to compute dot product with
dim length of the dense vector
b bias
Returns:
dot product between dense weights and a sparse feature vector

Definition at line 440 of file SparseFeatures.h.

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

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

Parameters:
vec_idx1 index of first vector
vec_idx2 index of second vector

Implements CDotFeatures.

Definition at line 1290 of file SparseFeatures.h.

virtual CFeatures* duplicate (  )  const [virtual]

duplicate feature object

Returns:
feature object

Implements CFeatures.

Definition at line 187 of file SparseFeatures.h.

void free_feature_vector ( TSparseEntry< ST > *  feat_vec,
int32_t  num,
bool  free 
)

free feature vector

Parameters:
feat_vec feature vector to free
num index of vector in cache
free if vector really should be deleted

Definition at line 901 of file SparseFeatures.h.

void free_sparse_feature_matrix (  ) 

free sparse feature matrix

Definition at line 165 of file SparseFeatures.h.

void free_sparse_feature_vector ( TSparseEntry< ST > *  feat_vec,
int32_t  num,
bool  free 
)

free sparse feature vector

Parameters:
feat_vec feature vector to free
num index of this vector in the cache
free if vector should be really deleted

Definition at line 501 of file SparseFeatures.h.

void free_sparse_features (  ) 

free sparse feature matrix and cache

Definition at line 176 of file SparseFeatures.h.

virtual int32_t get_dim_feature_space (  )  [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

Implements CDotFeatures.

Definition at line 1279 of file SparseFeatures.h.

ST get_feature ( int32_t  num,
int32_t  index 
)

get a single feature

Parameters:
num number of feature vector to retrieve
index index of feature in this vector
Returns:
sum of features that match dimension index and 0 if none is found

Definition at line 199 of file SparseFeatures.h.

virtual EFeatureClass get_feature_class (  )  [virtual]

get feature class

Returns:
feature class SPARSE

Implements CFeatures.

Definition at line 887 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the LONGREAL feature can deal with

Returns:
feature type LONGREAL

Implements CFeatures.

Definition at line 1470 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the DREAL feature can deal with

Returns:
feature type DREAL

Implements CFeatures.

Definition at line 1461 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the SHORTREAL feature can deal with

Returns:
feature type SHORTREAL

Implements CFeatures.

Definition at line 1452 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the ULONG feature can deal with

Returns:
feature type ULONG

Implements CFeatures.

Definition at line 1443 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the LONG feature can deal with

Returns:
feature type LONG

Implements CFeatures.

Definition at line 1434 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the UINT feature can deal with

Returns:
feature type UINT

Implements CFeatures.

Definition at line 1425 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the INT feature can deal with

Returns:
feature type INT

Implements CFeatures.

Definition at line 1416 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the WORD feature can deal with

Returns:
feature type WORD

Implements CFeatures.

Definition at line 1407 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the SHORT feature can deal with

Returns:
feature type SHORT

Implements CFeatures.

Definition at line 1398 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the BYTE feature can deal with

Returns:
feature type BYTE

Implements CFeatures.

Definition at line 1389 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the CHAR feature can deal with

Returns:
feature type CHAR

Implements CFeatures.

Definition at line 1380 of file SparseFeatures.h.

EFeatureType get_feature_type (  )  [virtual]

get feature type the BOOL feature can deal with

Returns:
feature type BOOL

Implements CFeatures.

Definition at line 1371 of file SparseFeatures.h.

virtual EFeatureType get_feature_type (  )  [virtual]

get feature type

Returns:
templated feature type

Implements CFeatures.

void get_full_feature_matrix ( ST **  dst,
int32_t *  num_feat,
int32_t *  num_vec 
)

gets a copy of a full feature matrix (swig compatible) num_feat,num_vectors are returned by reference

Parameters:
dst full feature matrix
num_feat number of features in matrix
num_vec number of vectors in matrix

Definition at line 681 of file SparseFeatures.h.

ST* get_full_feature_matrix ( int32_t &  num_feat,
int32_t &  num_vec 
)

gets a copy of a full feature matrix num_feat,num_vectors are returned by reference

Parameters:
num_feat number of features in matrix
num_vec number of vectors in matrix
Returns:
full feature matrix

Definition at line 646 of file SparseFeatures.h.

void get_full_feature_vector ( ST **  dst,
int32_t *  len,
int32_t  num 
)

get the fully expanded dense feature vector num

Parameters:
dst feature vector
len length is returned by reference
num index of feature vector

Definition at line 263 of file SparseFeatures.h.

ST* get_full_feature_vector ( int32_t  num,
int32_t &  len 
)

converts a sparse feature vector into a dense one preprocessed compute_feature_vector caller cleans up

Parameters:
num index of feature vector
len length is returned by reference
Returns:
dense feature vector

Definition at line 231 of file SparseFeatures.h.

virtual const char* get_name (  )  const [virtual]
Returns:
object name

Implements CSGObject.

Definition at line 1333 of file SparseFeatures.h.

virtual int32_t get_nnz_features_for_vector ( int32_t  num  )  [virtual]

get number of non-zero features in vector

Parameters:
num which vector
Returns:
number of non-zero features in vector

Implements CDotFeatures.

Definition at line 295 of file SparseFeatures.h.

int32_t get_num_features (  ) 

get number of features

Returns:
number of features

Definition at line 862 of file SparseFeatures.h.

int64_t get_num_nonzero_entries (  ) 

get number of non-zero entries in sparse feature matrix

Returns:
number of non-zero entries in sparse feature matrix

Definition at line 914 of file SparseFeatures.h.

virtual int32_t get_num_vectors (  )  [virtual]

get number of feature vectors

Returns:
number of feature vectors

Implements CFeatures.

Definition at line 856 of file SparseFeatures.h.

virtual int32_t get_size (  )  [virtual]

get memory footprint of one feature

Returns:
memory footprint of one feature

Implements CFeatures.

Definition at line 835 of file SparseFeatures.h.

void get_sparse_feature_matrix ( TSparse< ST > **  dst,
int32_t *  num_feat,
int32_t *  num_vec,
int64_t *  nnz 
)

get the pointer to the sparse feature matrix (swig compatible) num_feat,num_vectors are returned by reference

Parameters:
dst feature matrix
num_feat number of features in matrix
num_vec number of vectors in matrix
nnz number of nonzero elements

Definition at line 533 of file SparseFeatures.h.

TSparse<ST>* get_sparse_feature_matrix ( int32_t &  num_feat,
int32_t &  num_vec 
)

get the pointer to the sparse feature matrix num_feat,num_vectors are returned by reference

Parameters:
num_feat number of features in matrix
num_vec number of vectors in matrix
Returns:
feature matrix

Definition at line 517 of file SparseFeatures.h.

TSparseEntry<ST>* get_sparse_feature_vector ( int32_t  num,
int32_t &  len,
bool &  vfree 
)

get sparse feature vector for sample num from the matrix as it is if matrix is initialized, else return preprocessed compute_feature_vector

Parameters:
num index of feature vector
len number of sparse entries is returned by reference
vfree whether returned vector must be freed by caller via free_sparse_feature_vector
Returns:
sparse feature vector

Definition at line 314 of file SparseFeatures.h.

TSparse<ST>* get_transposed ( int32_t &  num_feat,
int32_t &  num_vec 
)

compute and return the transpose of the sparse feature matrix which will be prepocessed. num_feat, num_vectors are returned by reference caller has to clean up

Parameters:
num_feat number of features in matrix
num_vec number of vectors in matrix
Returns:
transposed sparse feature matrix

Definition at line 567 of file SparseFeatures.h.

CLabels* load_svmlight_file ( char *  fname,
bool  do_sort_features = true 
)

load features from file

Parameters:
fname filename to load from
do_sort_features if true features will be sorted to ensure they are in ascending order
Returns:
label object with corresponding labels

Definition at line 1024 of file SparseFeatures.h.

bool obtain_from_simple ( CSimpleFeatures< ST > *  sf  ) 

obtain sparse features from simple features

Parameters:
sf simple features
Returns:
if obtaining was successful

Definition at line 842 of file SparseFeatures.h.

virtual bool set_full_feature_matrix ( ST *  src,
int32_t  num_feat,
int32_t  num_vec 
) [virtual]

creates a sparse feature matrix from a full dense feature matrix necessary to set feature_matrix, num_features and num_vectors where num_features is the column offset, and columns are linear in memory see above for definition of sparse_feature_matrix

Parameters:
src full feature matrix
num_feat number of features in matrix
num_vec number of vectors in matrix

Definition at line 716 of file SparseFeatures.h.

int32_t set_num_features ( int32_t  num  ) 

set number of features

Sometimes when loading sparse features not all possible dimensions are used. This may pose a problem to classifiers when being applied to higher dimensional test-data. This function allows to artificially explode the feature space

Parameters:
num the number of features, must be larger than the current number of features
Returns:
previous number of features

Definition at line 875 of file SparseFeatures.h.

virtual void set_sparse_feature_matrix ( TSparse< ST > *  src,
int32_t  num_feat,
int32_t  num_vec 
) [virtual]

set feature matrix necessary to set feature_matrix, num_features, num_vectors, where num_features is the column offset, and columns are linear in memory see below for definition of feature_matrix

Parameters:
src new sparse feature matrix
num_feat number of features in matrix
num_vec number of vectors in matrix

Definition at line 630 of file SparseFeatures.h.

void sort_features (  ) 

ensure that features occur in ascending order, only call when no preprocessors are attached

Definition at line 1191 of file SparseFeatures.h.

ST sparse_dot ( ST  alpha,
TSparseEntry< ST > *  avec,
int32_t  alen,
TSparseEntry< ST > *  bvec,
int32_t  blen 
)

compute the dot product between two sparse feature vectors alpha * vec^T * vec

Parameters:
alpha scalar to multiply with
avec first sparse feature vector
alen avec's length
bvec second sparse feature vector
blen bvec's length
Returns:
dot product between the two sparse feature vectors

Definition at line 382 of file SparseFeatures.h.

bool write_svmlight_file ( char *  fname,
CLabels label 
)

write features to file using svm light format

Parameters:
fname filename to write to
label Label object (number of labels must correspond to number of features)
Returns:
true if successful

Definition at line 1239 of file SparseFeatures.h.


Member Data Documentation

CCache< TSparseEntry<ST> >* feature_cache [protected]

feature cache

Definition at line 1364 of file SparseFeatures.h.

int32_t num_features [protected]

total number of features

Definition at line 1358 of file SparseFeatures.h.

int32_t num_vectors [protected]

total number of vectors

Definition at line 1355 of file SparseFeatures.h.

TSparse<ST>* sparse_feature_matrix [protected]

array of sparse vectors of size num_vectors

Definition at line 1361 of file SparseFeatures.h.


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

SHOGUN Machine Learning Toolbox - Documentation