2
$\begingroup$

Suppose X, Y, W are independent random variables such that X ∼ GAM(2,3), Y ∼ N(1,4) and W ∼ BIN(10,1/4). Let U = 2X − 3Y and V = Y − W . Find cov(U, V ).

I know that cov(U, V) = E(U, V) - E(U)E(V). I've found E(U) and E(V) but I don't know how to find E(U, V).

$\endgroup$
1
  • $\begingroup$ You know, or can look up, the mean and variance of the various distributions. But in the second simpler way in the answer, all we need is the cov of $Y$ and $Y$, which is the variance of $Y$. We were told that $Y$ has variance $4$, so there is no work to do. $\endgroup$ – André Nicolas Apr 19 '16 at 6:11
2
$\begingroup$

You mean that you don't know how to find $E(UV)$. This is $E((2X-3Y)(Y-W))$. Expand the product, and use the linearity of expectation. We find that $$E(UV)=2E(XY)-2E(XW)-3E(Y^2)+3E(YW).$$ Now you can use independence and your knowledge of the various distributions to finish the calculation of $E(UV)$.

However, this is not the best way to find the covariance of $U$ and $V$. Instead, use the bilinearity of covariance. We have $$\text{Cov}(UV)=2\text{Cov}(X,Y)-2\text{Cov}(X,W)-3\text{Cov}(Y,Y)+3\text{Cov}(Y,W).$$ Almost all these covariances are $0$!

$\endgroup$
2
  • $\begingroup$ Almost all of them are 0? Aren't they all 0 since X, Y, W are all independent? $\endgroup$ – clover20 Apr 19 '16 at 6:11
  • 1
    $\begingroup$ The Cov of $Y$ and $Y$ is not $0$, it is the variance of $Y$. The other three are $0$. $\endgroup$ – André Nicolas Apr 19 '16 at 6:13

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.