1

I'd like to work out $\operatorname{Cov}(\cos(2U), \cos(3U))$ where $U$ is uniformly distributed on $[0, \pi]$.

I believe this involves computing $\mathbb{E}[\cos(2U)\cos(3U)]$. If so, then I first need the joint density, which I'm stuck on.

I can work out the density of $\cos(3U)$; this is similar to this question. $f_{\cos(3U)}(x) = \frac{1}{\pi\sqrt{1 - x^2}}$, $x \in (-1, 1)$. But that and $f_{\cos(2U)}(x)$ are only (directly) useful for individual expected values.

How do I work out the joint density?

(In this answer, the joint density seems to be $1$. Don't know why though.)

User1865345
  • 8,202
johnsmith
  • 335
  • 3
    You don't need the joint density to find this expectation, thanks to https://en.wikipedia.org/wiki/Law_of_the_unconscious_statistician. – StubbornAtom Dec 30 '23 at 07:06
  • Much easier now. – johnsmith Dec 30 '23 at 09:39
  • 1
    If you consider that for a given value of $X=\cos(2U)$, then either (i) $Y=\cos(3U)$ takes one of two possible values, equal in magnitude and opposite in sign (one from $U<\pi/2$, one from $U>\pi/2$), or (ii) $Y$ is $0$; then it's immediately clear that the pair of possible values are equally probable and hence that their corresponding contributions to the covariance cancel. Given $X$ and $Y$ are both bounded, immediately $E(XY)=0$. – Glen_b Dec 31 '23 at 02:50

0 Answers0