"Random variables $X$,$Y$ are independent and both have uniform distribution on interval $[0,1]$. The issue is to delineate the probability denisty function of random variables $(X,\frac{Y}{X})$ and $(\frac Y X)$.

I attempted to do it as follows:

Let $(U,V)=G(X,Y)$, where $G(x,y)=(x,\frac y x)$


$\begin{cases} x=u & \text{} \\ y=uv, & \text{} \\ \end{cases}$

and then after some steps

$g_{(u,v)}(u,v)= g_{(x,y)}(G^{-1}(u,v)) \times detDG^{-1}$

where $detDG^{-1}$ =det$ \left[ \begin{array}{cc|c} 1&0\\ v&u \end{array} \right]=u$

therefore I got:

$g_{(u,v)}(u,v)=u\times {indicator}_{[0,1]}(u)\times {indicator}_{[0,\infty]}(v)dudv$

however if we integrate the obtained probability density function we will get $\infty$ instead of $1$, so I must have made some mistake/s. Can anyone help me by improving my reasoning or at least showing when I am making the mistakes?

  • $\begingroup$ Wasn't this asked very recently? Anyway, the problem is that the (u,v)-domain is not [0,1]x[0,oo). If u=x and v=y/x then uv< $__$... $\endgroup$ – Did May 7 '14 at 21:00
  • $\begingroup$ I don't know if i did properly understand your comment, however I tried chaning the domain to $[0,1] \times [0,2]$, so that when I integrate this probability density function i would get 1 as it should be, but I believe that's not correct, since $\frac{Y}{X}$ can equal to any non-negative numer so to numbers greater than 2 as well. $\endgroup$ – user141766 May 7 '14 at 22:35

If $u=x$ and $v=y/x$ and if the $(x,y)$-domain is $0<x<1$ and $0<y<1$, then $x=u$ and $y=uv$ hence the $(u,v)$-domain $D$ is defined by the inequalities $$0<u<1,\qquad0<uv<1,$$ or, equivalently, $$D=\{(u,v)\mid0<u<1,\,0<v<1/u\}. $$ Note that the integral of $u$ on $D$ is $$ \int_0^1u\int_0^{1/u}\mathrm dv\,\mathrm du=\int_0^1u\frac1u\,\mathrm du=1, $$ as it should be for the function $g$ defined by $$ g(u,v)=u\,\mathbf 1_D(u,v), $$ to be a legitimate PDF.


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.