How to solve matrix equations with inconsistent dimensions?

29 Views Asked by At

I'm confused by something in one of the papers I'm reading, but I'm not sure if there's any math in it that I'm not aware of.

$\mathbf{E}=\mathbf{R}\mathbf{e}$, where $\mathbf{E}\in\mathbb{C}^{n_1\times n_3}, \mathbf{R}\in\mathbb{C}^{n_1\times n_2}, \mathbf{e}\in\mathbb{C}^{n_2\times n_3}$

Currently $\mathbf{E}$ and $\mathbf{R}$ are known and it is necessary to find $\mathbf{e}$. Then the article uses the following formula to find $\mathbf{e}$

$\mathbf{R}^T\mathbf{E}=(\mathbf{R}^T\mathbf{R}+\gamma\mathbf{I})\mathbf{e}$, where $\gamma$ is the regularization factor.

I would like to know why such a transformation needs to be made and what theory it is based on. If there's any material I can reference, feel free to provide it to me. I've attached the article I'm reading, but it's in the field of computational electromagnetics.

Eqn. (2)~(3) in https://ieeexplore.ieee.org/document/10068421

Any help is appreciated.

1

There are 1 best solutions below

0
On

The problem is solved.

The least-squares solution of Equation $\mathbf{E}=\mathbf{R}\mathbf{e}$ is $\mathbf{R}^T\mathbf{E}=\mathbf{R}^T\mathbf{R}\mathbf{e}$.

When the number of matrix conditions is too large, the ridge regression technique will be introduced. Therefore the $\gamma\mathbf{I}$ is added.