Feature Selection ToolboxFST3 Library / Documentation

FST::Indexed_Matrix< DATATYPE, DIMTYPE, SUBSET > Class Template Reference

Matrix representation and operations, allows operation in selected subspace. More...

#include <indexed_matrix.hpp>

List of all members.

Public Types

typedef boost::shared_ptr< SUBSET > PSubset

Public Member Functions

 Indexed_Matrix (const DIMTYPE n, bool LUstorage=false)
 Indexed_Matrix (const Indexed_Matrix &im)
Indexed_Matrixoperator= (Indexed_Matrix const &im)
void reset (const DIMTYPE n_max, bool LUstorage=false)
 reallocates
void redim (const DIMTYPE n=0)
 keeps main allocations, just resets index arrays and adjusts _n; n==0 here equals _n_max
DATATYPE & at_raw (const DIMTYPE row, const DIMTYPE col)
DATATYPE & at (const DIMTYPE row, const DIMTYPE col)
DATATYPE * get_row_raw (const DIMTYPE row)
DATATYPE * get_row (const DIMTYPE row)
void narrow_to (const PSubset sub)
 fill _index[] -> access to data is then mapped to sub-matrix
void denarrow ()
 reset _index[] -> access is reset to the full matrix
DIMTYPE get_n_max () const
DIMTYPE get_n () const
DIMTYPE get_d () const
void set_all_raw_to (const DATATYPE val)
void add_raw (const Indexed_Matrix &mr)
void add_constmul_raw (const Indexed_Matrix &mr, const DATATYPE mul)
void copy_constmul_raw (const Indexed_Matrix &mr, const DATATYPE mul)
void copy_raw (const Indexed_Matrix &mr)
void LUdecompose (Indexed_Matrix &result)
 returns matrix+vector+sign
void solve_equations (Indexed_Vector< DATATYPE, DIMTYPE, SUBSET > &result_vec, const Indexed_Matrix &LU_matrix, Indexed_Vector< DATATYPE, DIMTYPE, SUBSET > &coef_vec)
 returns vector
void invert (Indexed_Matrix &result, const Indexed_Matrix &LU_matrix)
 returns matrix
DATATYPE determinant (const Indexed_Matrix &LU_matrix)
 returns scalar

Protected Attributes

DIMTYPE _n_max
 allocated space can hold _n_max*_n_max matrix
DIMTYPE _n
 full feature set dimensionality
boost::scoped_array< DATATYPE > _data
 raw matrix data
boost::scoped_array< DIMTYPE > _permut
 LU permutation index.
int _sign
 permutation sign (renamed obsolete d)
DIMTYPE _d
 virtual sub-matrix (as represented by _index) dimensionality
boost::scoped_array< DIMTYPE > _index
 maps virtual sub-matrix _d dimensions to original _n dimensions
boost::scoped_array< DATATYPE > _tmp_rowscale
 (lazy allocation)
boost::scoped_array< DIMTYPE > _tmp_whereis
 (lazy allocation)
boost::scoped_ptr
< Indexed_Vector< DATATYPE,
DIMTYPE, SUBSET > > 
_tmp_eq_coefs
 invert coefs (lazy allocation)
boost::scoped_ptr
< Indexed_Vector< DATATYPE,
DIMTYPE, SUBSET > > 
_tmp_eq_roots
 to store/pass intermediate results in invert() etc. (lazy allocation)

Friends

std::ostream & operator<< (std::ostream &os, const Indexed_Matrix &mr)

Detailed Description

template<typename DATATYPE, typename DIMTYPE, class SUBSET>
class FST::Indexed_Matrix< DATATYPE, DIMTYPE, SUBSET >

Matrix representation and operations, allows operation in selected subspace.


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

Generated on Thu Mar 31 11:38:36 2011 for FST3Library by  doxygen 1.6.1