Public Member Functions | Protected Member Functions | Protected Attributes

CGMNPLib Class Reference


Detailed Description

class GMNPLib Library of solvers for Generalized Minimal Norm Problem (GMNP).

Generalized Minimal Norm Problem to solve is

min 0.5*alpha'*H*alpha + c'*alpha

subject to sum(alpha) = 1, alpha(i) >= 0

H [dim x dim] is symmetric positive definite matrix. c [dim x 1] is an arbitrary vector.

The precision of the found solution is given by the parameters tmax, tolabs and tolrel which define the stopping conditions:

UB-LB <= tolabs -> exit_flag = 1 Abs. tolerance. UB-LB <= UB*tolrel -> exit_flag = 2 Relative tolerance. LB > th -> exit_flag = 3 Threshold on lower bound. t >= tmax -> exit_flag = 0 Number of iterations.

UB ... Upper bound on the optimal solution. LB ... Lower bound on the optimal solution. t ... Number of iterations. History ... Value of LB and UB wrt. number of iterations.

The following algorithms are implemented: ..............................................

For more info refer to V.Franc: Optimization Algorithms for Kernel Methods. Research report. CTU-CMP-2005-22. CTU FEL Prague. 2005. ftp://cmp.felk.cvut.cz/pub/cmp/articles/franc/Franc-PhD.pdf .

Definition at line 66 of file GMNPLib.h.

Inheritance diagram for CGMNPLib:
Inheritance graph
[legend]

List of all members.

Public Member Functions

 CGMNPLib (void)
 CGMNPLib (float64_t *vector_y, CKernel *kernel, int32_t num_data, int32_t num_virtual_data, int32_t num_classes, float64_t reg_const)
virtual ~CGMNPLib ()
int8_t gmnp_imdm (float64_t *vector_c, int32_t dim, int32_t tmax, float64_t tolabs, float64_t tolrel, float64_t th, float64_t *alpha, int32_t *ptr_t, float64_t **ptr_History, int32_t verb)
void get_indices2 (int32_t *index, int32_t *c, int32_t i)

Protected Member Functions

float64_tget_kernel_col (int32_t a)
float64_tget_col (int32_t a, int32_t b)
float64_t kernel_fce (int32_t a, int32_t b)
virtual const char * get_name () const

Protected Attributes

float64_tdiag_H
float64_t ** kernel_columns
float64_tcache_index
int32_t first_kernel_inx
int64_t Cache_Size
int32_t m_num_data
float64_t m_reg_const
float64_tm_vector_y
CKernelm_kernel
int32_t first_virt_inx
float64_tvirt_columns [3]
int32_t m_num_virt_data
int32_t m_num_classes

Constructor & Destructor Documentation

CGMNPLib ( void   ) 

default constructor

Definition at line 80 of file GMNPLib.cpp.

CGMNPLib ( float64_t vector_y,
CKernel kernel,
int32_t  num_data,
int32_t  num_virtual_data,
int32_t  num_classes,
float64_t  reg_const 
)

constructor

Parameters:
vector_y vector y
kernel kernel
num_data number of data
num_virtual_data number of virtual data
num_classes number of classes
reg_const reg const

Definition at line 100 of file GMNPLib.cpp.

~CGMNPLib (  )  [virtual]

Definition at line 143 of file GMNPLib.cpp.


Member Function Documentation

float64_t * get_col ( int32_t  a,
int32_t  b 
) [protected]

get col

Parameters:
a a
b b
Returns:
col at a, b

Definition at line 212 of file GMNPLib.cpp.

void get_indices2 ( int32_t *  index,
int32_t *  c,
int32_t  i 
)

get indices2

Parameters:
index index
c c
i i

Definition at line 194 of file GMNPLib.cpp.

float64_t * get_kernel_col ( int32_t  a  )  [protected]

get kernel col

Parameters:
a a
Returns:
col at a

Definition at line 161 of file GMNPLib.cpp.

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

Implements CSGObject.

Definition at line 140 of file GMNPLib.h.

int8_t gmnp_imdm ( float64_t vector_c,
int32_t  dim,
int32_t  tmax,
float64_t  tolabs,
float64_t  tolrel,
float64_t  th,
float64_t alpha,
int32_t *  ptr_t,
float64_t **  ptr_History,
int32_t  verb 
)

-------------------------------------------------------------- GMNP solver based on improved MDM algorithm 1.

Search strategy: u determined by common rule and v is optimized.

Usage: exitflag = gmnp_imdm( &get_col, diag_H, vector_c, dim, tmax, tolabs, tolrel, th, &alpha, &t, &History ); --------------------------------------------------------------

Definition at line 260 of file GMNPLib.cpp.

float64_t kernel_fce ( int32_t  a,
int32_t  b 
) [protected]

kernel fce

Parameters:
a a
b b
Returns:
something floaty

Definition at line 477 of file GMNPLib.cpp.


Member Data Documentation

float64_t* cache_index [protected]

cache index

Definition at line 148 of file GMNPLib.h.

int64_t Cache_Size [protected]

cache size

Definition at line 152 of file GMNPLib.h.

float64_t* diag_H [protected]

diag H

Definition at line 144 of file GMNPLib.h.

int32_t first_kernel_inx [protected]

first kernel inx

Definition at line 150 of file GMNPLib.h.

int32_t first_virt_inx [protected]

index of first used column

Definition at line 163 of file GMNPLib.h.

float64_t** kernel_columns [protected]

kernel columns

Definition at line 146 of file GMNPLib.h.

CKernel* m_kernel [protected]

kernel

Definition at line 160 of file GMNPLib.h.

int32_t m_num_classes [protected]

number of classes

Definition at line 169 of file GMNPLib.h.

int32_t m_num_data [protected]

num data

Definition at line 154 of file GMNPLib.h.

int32_t m_num_virt_data [protected]

number of virt data

Definition at line 167 of file GMNPLib.h.

float64_t m_reg_const [protected]

reg const

Definition at line 156 of file GMNPLib.h.

float64_t* m_vector_y [protected]

vectory

Definition at line 158 of file GMNPLib.h.

float64_t* virt_columns[3] [protected]

cache for three columns

Definition at line 165 of file GMNPLib.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