normalizeParameterMatrix

Method of mdss.

Row-normalize the model’s CP parameter matrix.

Source: src/model/@mdss/normalizeParameterMatrix.m

Syntax

sys = normalizeParameterMatrix(sys, factor)

Description

normalizeParameterMatrix rescales the parameter matrix of the model’s CPNTensor so that each equation row is normalized to its largest-magnitude coefficient, then multiplied by factor.

The signed parameter matrix is first reassembled from its three stored parts,

$\Phi = \Phi_\text{cont} + \Phi_\text{one} - \Phi_\text{minusOne},$

each row is scaled by factor / max(abs(row)), and the result is split back into the three stored matrices: entries equal to 1 go to parameterMatrixOne, entries equal to -1 to parameterMatrixMinusOne, and the remainder to parameterMatrixContinuous.

Normalizing keeps the coefficient magnitudes comparable across successive model manipulations; it is used, for example, at the end of algebraicElimination.

Input arguments

Argument	Description
`sys`	The `mdss` model whose parameter matrix is normalized.
`factor`	Scalar applied to every row after normalization.

Output arguments

Output	Description
`sys`	The model with its parameter matrix row-normalized.

Example