TRandSVDLRApx< T_value > Class Template Reference

Uses randomized SVD to compute low rank approximation (WARNING: O(n²) complexity) More...

#include <TLowRankApx.hh>

Inheritance diagram for TRandSVDLRApx< T_value >:

Public Member Functions
virtual std::unique_ptr< TMatrix< value_t > >	build (const TBlockCluster *bcl, const TTruncAcc &acc) const
virtual std::unique_ptr< TMatrix< value_t > >	build (const TBlockIndexSet &block_is, const TTruncAcc &acc) const
Public Member Functions inherited from TLowRankApx< T_value >
virtual bool	has_statistics () const
	indicate if algorithm provides statistics

Detailed Description

template<typename T_value>
class Hpro::TRandSVDLRApx< T_value >

     TRandSVDLRApx uses randomized singular value decomposition to approximate 
     a given matrix block. For the approximation, the complete matrix block has
     to be evaluated, hence complexity is O(n²).

Member Function Documentation

build() [1/2]

template<typename T_value >

virtual std::unique_ptr< TMatrix< value_t > > build ( const TBlockCluster *  bcl, const TTruncAcc &  acc  ) const

virtual

build low rank matrix for block cluster bcl with rank defined by accuracy acc

Reimplemented from TLowRankApx< T_value >.

build() [2/2]

template<typename T_value >

virtual std::unique_ptr< TMatrix< value_t > > build ( const TBlockIndexSet &  block_is, const TTruncAcc &  acc  ) const

virtual

build low rank matrix for block index set block_is with rank defined by accuracy acc

Implements TLowRankApx< T_value >.