0
$\begingroup$

Given two random uniform variables, $U$ and $V$, that are uniformly distributed over [0,1], how do you calculate the joint pdf of $X$, $Y$ where $X = F(U,V)$ and $Y = G(U,V)$ and where is the joint pdf zero.

Specifically, where $$X = + \sqrt(U) \\ Y = UV$$

I understand this is similar to Joint PDF of two random variables and their sum but I am not interested in the marginal distribution and looking for more of an explanation.

$\endgroup$
  • $\begingroup$ Maybe you want $Y = G(U, V)$ where $G$ is another function that can be different from $F$. $\endgroup$ – Tunococ Sep 4 '13 at 23:15
  • $\begingroup$ @Tunococ, yes and I've updated the question. I tried to pose the question as general as possible and then added the specific HW question once the general form is explained. $\endgroup$ – Elpezmuerto Sep 4 '13 at 23:21
1
$\begingroup$

For any $x, y$ in $[0, 1]$, \begin{align} \Pr(X \le x, Y \le y) & = \Pr(\sqrt U \le x, UV \le y) \\ & = \Pr\left(\sqrt U \le x, U \le \frac yV\right) \\ & = \int_0^1 \Pr\left(U \le x^2, U \le \frac yV \mid V = v\right) dv \\ & = \int_0^1 \min\left\{x^2, \frac yv\right\} dv \\ & = \begin{cases} x^2 & ; x^2 \le y \\ \int_0^{y/x^2} x^2 dv + \int_{y/x^2}^1 \frac yv dv & ; x^2 \ge y \\ \end{cases}\\ & = \begin{cases} x^2 & ; x^2 \le y \\ y - y \log y + 2y\log x & ; x^2 \ge y \\ \end{cases} \end{align} To obtain the probability density function, you take the derivative with respect to $x$ and $y$: $$ f_{X,Y}(x,y) = \begin{cases} 0 & ; x^2 \le y \\ 2/x & ; x^2 \ge y \\ \end{cases} $$

$\endgroup$
  • $\begingroup$ Thanks, this really helps. Just a few questions to clarify. First, what does PR stand for? Second, why have the conditional probabilities with respect to $y$? Third, the derivative of $x^2$ is $2x$, not $0$? $\endgroup$ – Elpezmuerto Sep 5 '13 at 12:36
  • $\begingroup$ First, $\Pr$ stands for "probability". Second, you could condition on $X$ or $Y$. I chose $Y$ because I thought the calculation would be easier with this choice. Third, you are right, but then you have to take the derivative of $2x$ with respect to $y$. $\endgroup$ – Tunococ Sep 6 '13 at 9:39

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.