ReducedVarianceInputCorrelatorComputeRstar Method

Computes the scores matrix transformed to match the desired correlation matrix. uses a varaince reduction technique in which the transformation matrix is adjusted so that the final sample correlation matrix will be much closer to the desired correlation matrix, especially for correlation values of 0.

Namespace: CenterSpace.NMath.Core
Assembly: NMath (in NMath.dll) Version: 7.4

Syntax

C++

Copy

protected override DoubleMatrix ComputeRstar(
	int numSamples,
	int numVaraibles,
	DoubleSymmetricMatrix desiredCorrelationMatrix
)

Protected Overrides Function ComputeRstar ( 
	numSamples As Integer,
	numVaraibles As Integer,
	desiredCorrelationMatrix As DoubleSymmetricMatrix
) As DoubleMatrix

protected:
virtual DoubleMatrix^ ComputeRstar(
	int numSamples, 
	int numVaraibles, 
	DoubleSymmetricMatrix^ desiredCorrelationMatrix
) override

abstract ComputeRstar : 
        numSamples : int * 
        numVaraibles : int * 
        desiredCorrelationMatrix : DoubleSymmetricMatrix -> DoubleMatrix 
override ComputeRstar : 
        numSamples : int * 
        numVaraibles : int * 
        desiredCorrelationMatrix : DoubleSymmetricMatrix -> DoubleMatrix

Parameters

numSamples Int32: Sample size.
numVaraibles Int32: Number of variables in each sample.
desiredCorrelationMatrix DoubleSymmetricMatrix: Desired correlation matrix.

Return Value

DoubleMatrix

Reference

ReducedVarianceInputCorrelator Class

CenterSpace.NMath.Core Namespace