Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

If $X$ is an $n \times p$ matrix of rank $r$ and $C = AX$ for some $q \times n$ matrix $A$ with rank$(A) = q$, how do I show that rank $(X(I-C^{-}C))=$ rank$(X)-$ rank$(C)$? I can show that rank $(X(I-C^{-}C))\geq $ rank$(X)-$ rank$(C)$, but how do I get the reverse inequality?

Note : $C^{-}$ is a generalized inverse of $C$.

Any help would be appreciated.

share|cite|improve this question

$C^-$ is denoted $C^+$. The fact that $A$ has maximal rank is useless.

1) We show that, for every $X,C$, one has $rank(X(I_p-C^+C))=dim(\ker(C))-dim(\ker(X)\cap \ker(C))$.

Proof: $C^+C$ is the orthogonal projection onto $im(C^*)$, that is, onto the orthogonal of $\ker(C)$. Then $I_p-C^+C$ is the orthogonal projection onto $\ker(C)$. Thus $im(I-C^+C)=\ker(C)$ and $im(X(I-C^+C))=X(\ker(C))$. Let $\ker(C)=(\ker(C)\cap\ker(X))\oplus Z$ ; then $X(\ker(C))=X(Z)$ and $X$ is an isomorphism from $Z$ onto $X(\ker(C))$ and we are done.

2) Here $\ker(X)\subset \ker(C)$ ; we deduce that $rank(X(I_p-C^+C))=dim(\ker(C))-dim(\ker(X))=rank(X)-rank(C)$.

EDIT: @Dunn,@Gigili, I see that you are unable to send a simple "thank you". Incredible rudeness.

share|cite|improve this answer
I'm sorry, but what should I thank you for? – Gigili Jan 15 '14 at 13:46
@Gigili, you offered a bounty for a possible answer. I am not interested by your bounty, but, at least, I think that you can indicate that you have read the answer – loup blanc Jan 16 '14 at 9:45

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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