I have the following problem: I would like to show that

$A^TA < I$ implies that the eigenvalues of $A$ are less than $1$.

$A$ is a symmetric matrix and $I$ is the identity.

This claim is also made here: Identity minus a matrix times its transpose positive semidefinite But no justification is given.


If $Av=\lambda v$ with $\|v\|=1$, $$ |\lambda|^2=\langle Av,Av\rangle=\langle A^TAv,v\rangle\leq\langle v,v\rangle=1. $$ You don't need $A$ symmetric for this to hold.

The claim in the answer you mention is a bit stronger, although the computation is the same. It says that $$ \|Ax\|^2=\langle Ax,Ax\rangle=\langle A^TAx,x\rangle\leq\langle x,x\rangle=\|x\|^2. $$ So $\|Ax\|\leq\|x\|$ for all $x$.



Let $A=UDU^T$ where $U$ is orthogonal and $D$ is a diagonal matrix, then we can write $UU^T=I$


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.