How to prove the following relation:

$$ \begin{array}{ccc} Au_i=\lambda_i u_i & \Rightarrow & A=\Sigma\lambda_iu_iu_i^T \end{array} $$

My try:

$$ \left\{ \begin{array}{lc} Au_i=\lambda_i u_i & (1)\\ AU=U \Lambda & (2)\\ U^TAU=U^TU\Lambda=\Lambda & (3)\\ UU^TAUU^T=U\Lambda U^T & (4) \\ A=U\Lambda U^T & (5)\\ A=\Sigma \lambda_i u_iu_i^T & (6) \end{array} \right. $$

Is the mentioned proof correct?

Thanks for any suggestions.

  • $\begingroup$ What is $U$? Is it an orthogonal matrix? Give more details - I don't have the book that you are referencing and I bet others don't either. $\endgroup$ – Sean Roberson Jul 26 '17 at 20:39

I think you have to state that $\{ u_i \in \mathbb{R}^n | i = 1, \ldots , n \}$ forms an orthonormal basis to justify that $U^TU=I$ and $UU^T=I$.

You might like to state explicitly the definition of $U$ and $\Lambda$ as well.

Name: Orthogonal Diagonalization.

Remark: Since you are able to write $A=U\Lambda U^T$, A is symmetric.


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.