Let $A \in \mathrm{R}^{n\times m}$ be a matrix of rank $n$ (with $m>n$). Let $A^+$ be the Moore-Penrose pseudoinverse of $A$. The matrix
$$ A^+ A = A^T(A A^T)^{-1} A $$
has eigenvalues $\lambda_i \in \{0,1 \}$. How can I proof this?
Let $A \in \mathrm{R}^{n\times m}$ be a matrix of rank $n$ (with $m>n$). Let $A^+$ be the Moore-Penrose pseudoinverse of $A$. The matrix
$$ A^+ A = A^T(A A^T)^{-1} A $$
has eigenvalues $\lambda_i \in \{0,1 \}$. How can I proof this?
Copyright © 2021 JogjaFile Inc.