CPN2mat

Method of CPNTensor.

Convert explicit CPN structure and parameter matrices of mss models into the full multilinear parameter matrix.

Source: src/tensor/@CPNTensor/CPN2mat.m

Syntax

F = obj.CPN2mat(structureMatrix, parameterMatrix)
F = obj.CPN2mat(structureMatrixTrue,...
    structureMatrixFalse,...
    structureMatrixContinuous,...
    parameterMatrixOne, parameterMatrixMinusOne,...
    parameterMatrixContinuous)

Description

The function supports both standard CPN representations and hybrid CPN representations containing true, false and continuous variables.

The resulting parameter matrix is expressed in the canonical monomial basis and can be directly converted into a TTTensor using TTTensor.fromFullParameterMatrix.

This utility is primarily intended for model conversion and validation of Tensor Train representations.

Input arguments

Argument Description
obj CPNTensor object.
structureMatrix Structure matrix.
parameterMatrix Parameter matrix.
structureMatrixTrue Sparse logical mask of monomial factors vlv_l of signals 𝐯=[v1,…,vnv]⊤\mathbf{v} = [v_1,\dots,v_{n_v}]^\top.
structureMatrixFalse Sparse logical mask of complemented factors 1−vl1-v_l.
structureMatrixContinuous Sparse weights ss for general factors 1−|s|+svl1-\lvert s\rvert + s\,v_l.
parameterMatrixOne Sparse logical mask of coefficient +1+1.
parameterMatrixMinusOne Sparse logical mask of coefficient −1-1.
parameterMatrixContinuous Sparse remaining coefficient values.

Output arguments

Output Description
F Sparse multilinear parameter matrix in monomial basis.

See also

CPNTensor · mss

MyToolbox Documentation | Generated automatically by CI/CD pipeline