MathOverflow is a question and answer site for professional mathematicians. 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

I'm reading a paper that outlines the calculation of a covariance matrix like the following:


What is the order of this matrix? My interpretation of the formula is something akin to dot products, but it's producing a vector not the expected matrix. Can someone explain where my interpretation is incorrect?

share|cite|improve this question
A link to the paper would help, but to me it seems most plausible that the $x_i$'s are column vectors (say of length $n$) and hence $x_ix_i^T$ has $(j,k)$ entry $x_i^jx_i^k$, and all matrices are $n$-by-$n$. If you're working over the real numbers, this looks like a decomposition of a symmetric positive semidefinite matrix as a sum of positive rank one matrices. – Jonas Meyer Apr 22 '10 at 20:13
Yes, now that I have looked at the paper, it seems to be implicitly assumed that vectors are columns by default. Notice for example that the coefficients of $x_i$ with respect to an orthonormal basis $(e_j)_j$are given on page 3 by $a_j=x_i^Te_j$. – Jonas Meyer Apr 22 '10 at 20:24
Very helpful, thank you! – fbrereto Apr 22 '10 at 20:29

Your Answer


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

Browse other questions tagged or ask your own question.