Detailed Description
class SVMOcas
Definition at line 30 of file SVMOcas.h.
List of all members.
Public Member Functions |
| CSVMOcas (void) |
| CSVMOcas (E_SVM_TYPE type) |
| CSVMOcas (float64_t C, CDotFeatures *traindat, CLabels *trainlab) |
virtual | ~CSVMOcas () |
virtual EClassifierType | get_classifier_type () |
virtual bool | train (CFeatures *data=NULL) |
void | set_C (float64_t c_neg, float64_t c_pos) |
float64_t | get_C1 () |
float64_t | get_C2 () |
void | set_epsilon (float64_t eps) |
float64_t | get_epsilon () |
void | set_bias_enabled (bool enable_bias) |
bool | get_bias_enabled () |
void | set_bufsize (int32_t sz) |
int32_t | get_bufsize () |
Protected Member Functions |
virtual const char * | get_name () const |
Static Protected Member Functions |
static void | compute_W (float64_t *sq_norm_W, float64_t *dp_WoldW, float64_t *alpha, uint32_t nSel, void *ptr) |
static float64_t | update_W (float64_t t, void *ptr) |
static int | add_new_cut (float64_t *new_col_H, uint32_t *new_cut, uint32_t cut_length, uint32_t nSel, void *ptr) |
static int | compute_output (float64_t *output, void *ptr) |
static int | sort (float64_t *vals, float64_t *data, uint32_t size) |
static void | print (ocas_return_value_T value) |
Protected Attributes |
bool | use_bias |
int32_t | bufsize |
float64_t | C1 |
float64_t | C2 |
float64_t | epsilon |
E_SVM_TYPE | method |
float64_t * | old_w |
float64_t | old_bias |
float64_t * | tmp_a_buf |
float64_t * | lab |
float64_t ** | cp_value |
uint32_t ** | cp_index |
uint32_t * | cp_nz_dims |
float64_t * | cp_bias |
Constructor & Destructor Documentation
constructor
- Parameters:
-
Definition at line 30 of file SVMOcas.cpp.
constructor
- Parameters:
-
| C | constant C |
| traindat | training features |
| trainlab | labels for training features |
Definition at line 37 of file SVMOcas.cpp.
Member Function Documentation
int add_new_cut |
( |
float64_t * |
new_col_H, |
|
|
uint32_t * |
new_cut, |
|
|
uint32_t |
cut_length, |
|
|
uint32_t |
nSel, |
|
|
void * |
ptr | |
|
) |
| | [static, protected] |
add new cut
- Parameters:
-
| new_col_H | new col H |
| new_cut | new cut |
| cut_length | length of cut |
| nSel | nSel |
| ptr | ptr |
Definition at line 187 of file SVMOcas.cpp.
int compute_output |
( |
float64_t * |
output, |
|
|
void * |
ptr | |
|
) |
| | [static, protected] |
compute output
- Parameters:
-
Definition at line 274 of file SVMOcas.cpp.
compute W
- Parameters:
-
| sq_norm_W | square normed W |
| dp_WoldW | dp W old W |
| alpha | alpha |
| nSel | nSel |
| ptr | ptr |
Definition at line 300 of file SVMOcas.cpp.
bool get_bias_enabled |
( |
|
) |
|
check if bias is enabled
- Returns:
- if bias is enabled
Definition at line 111 of file SVMOcas.h.
get buffer size
- Returns:
- buffer size
Definition at line 123 of file SVMOcas.h.
get C1
- Returns:
- C1
Definition at line 81 of file SVMOcas.h.
get C2
- Returns:
- C2
Definition at line 87 of file SVMOcas.h.
get classifier type
- Returns:
- classifier type SVMOCAS
Reimplemented from CClassifier.
Definition at line 57 of file SVMOcas.h.
get epsilon
- Returns:
- epsilon
Definition at line 99 of file SVMOcas.h.
virtual const char* get_name |
( |
void |
|
) |
const [protected, virtual] |
static void print |
( |
ocas_return_value_T |
value |
) |
[static, protected] |
void set_bias_enabled |
( |
bool |
enable_bias |
) |
|
set if bias shall be enabled
- Parameters:
-
| enable_bias | if bias shall be enabled |
Definition at line 105 of file SVMOcas.h.
void set_bufsize |
( |
int32_t |
sz |
) |
|
set buffer size
- Parameters:
-
Definition at line 117 of file SVMOcas.h.
set C
- Parameters:
-
| c_neg | new C constant for negatively labeled examples |
| c_pos | new C constant for positively labeled examples |
Definition at line 75 of file SVMOcas.h.
set epsilon
- Parameters:
-
Definition at line 93 of file SVMOcas.h.
sort
- Parameters:
-
| vals | vals |
| data | data |
| size | size |
Definition at line 263 of file SVMOcas.cpp.
bool train |
( |
CFeatures * |
data = NULL |
) |
[virtual] |
train SVM classifier
- Parameters:
-
| data | training data (parameter can be avoided if distance or kernel-based classifiers are used and distance/kernels are initialized with train data) |
- Returns:
- whether training was successful
Reimplemented from CClassifier.
Definition at line 54 of file SVMOcas.cpp.
update W
- Parameters:
-
- Returns:
- something floaty
Definition at line 160 of file SVMOcas.cpp.
Member Data Documentation
cutting plane index
Definition at line 211 of file SVMOcas.h.
cutting plane dimensions
Definition at line 213 of file SVMOcas.h.
sparse representation of cutting planes
Definition at line 209 of file SVMOcas.h.
The documentation for this class was generated from the following files: