0
$\begingroup$

Let $A$ be an $n\times n$ symmetric positive definite matrix with eigenvalues and eigenvectors $\lambda_1\ge\lambda_2\ge\cdots\ge\lambda_n>0$ and $v_1,v_2,\cdots,v_n$ respectively.

We know that the largest eigenvalue of $A$ can be obtained by "trace maximization" or "Rayleigh quotient maximization". I also noted that the largest eigenvalue of $A$ can be obtained via

$$\underset{x \in \mathbb{R}^n}{\text{minimize}} \quad \|A-xx'\|_F^2 \tag{OP}$$

Specifically, $x = \pm \sqrt \lambda_1 v_1$ are minimizers of $\text{(OP)}$.

Is there a specific name for this optimization problem? Could you please direct me to some literature. Thanks!

$\endgroup$
0

1 Answer 1

3
$\begingroup$

It's the symmetric version of the low-rank approximation problem.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.