Let $A_1 \in \mathbb{R}^{k\times k}$ be a symmetric and positive definite matrix. Let $x_1,x_2 \in \mathbb{R}^k$. Suppose I have the outer product $x_1 x_1^T$, from this post we know that $x_1$ is an eigenvector of $x_1 x_1^T$ and $||x_1||^2$ is its only non-zero eigenvalue.

We also know that if we apply $A_1 x_1$ then $x_1$ is projected onto the eigenvectors of $A_1$ and scaled by their corresponding eigenvalues given by the symmetric decomposition i.e $A_1 = \sum_i \lambda_i u_i u_i^T $ where the $\lambda_i, u_i$ are one of the $k$ eigenvalues and eigenvectors corresponding to $A_1$.

Is it possible to relate the eigenvalues and eigenvectors of $(A_1x_1 )(A_1x_1)^T + (A_1x_2)( A_1x_2)^T $ to $x_1 x_1^T + x_2 x_2^T $. I am trying to study the effect of $A_1$ on the eigenvalues and eigenvectors of the sum of the outer products.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.