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

I'm trying to read about Principal Component Analysis and i'm stuck in the identity $Var[\alpha_{1}^{'}X]=\alpha_{1}^{'}\Sigma\alpha_{1}$ where $\Sigma$ is the variance-covariance matrix of $X$, $X=(X_{1},\cdots,X_{p})$ and $\{X_{i}\}$ are random variables, $\alpha_{1}=(a_{1},\cdots,a_{p})$ a coefficients vector, and $\alpha_{1}^{'}$ is the transpose. Developing the LHS of the identity we have $a_{1}^2Var(X_{1})+\cdots+a_{1}^2Var(X_{p})$ wich is not equal to the RHS in general. The identity is satisfied only when the $\{X_{i}\}$ are all independent, so i'm confuse because everywhere I've read and in the book I'am following, Jollife, Principal Component Analysis, there is no such assumption. Thanks in advance for the clarification.

share|cite|improve this question
up vote 1 down vote accepted

The LHS is what you said when the random variables $X_i$ are uncorrelated. In a general case, you have to compute the variance of $\sum_{j=1}^na_jX_j$, which is equal to $$\sum_{i,j=1}^na_ia_jE[X_iX_j]-\sum_{i,j=1}^na_iE[X_i]a_jE[X_j]=\sum_{i,j=1}^na_ia_j\operatorname{cov}(X_i,X_j).$$

share|cite|improve this answer
You're right, I forgot about the assumption of the LHS and thought it was in general. Thanks. – Ivan3.14 Oct 4 '12 at 7:38

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.