1
$\begingroup$

I'm having trouble finding the Jacobian when trying to compute a distribution. If $(X,Y)$ is a point on a unit disk with radius $1$, I'd like to find the density of the distance between the point and the centre of the disk.

So the joint density function of $X$ and $Y$ is

$$ f_{XY}(x,y) = \begin{cases} 1/\pi, & x^2 + y^2 \leq 1,\\ 0, & \text{otherwise} \end{cases}. $$

I put $U = \sqrt{X^2+Y^2}$ and use an auxiliary random variable $V = \mathrm{arctan}(Y/X).$ So there is a 2-to-1 mapping.

Now I wonder what the Jacobian might be. Or, well, the book tells me it's $u$, but I don't know how to get there myself. Using wolfram alpha (admittedly I'm probably just using it wrong) gets me the answer $0$, not $u$.

Normally, I would work with simpler auxiliary variables, which have helped me compute the Jacobian for other problems, but in this case I'm adviced to use $\mathrm{arctan}(Y/X)$ and that could be what makes this so hard for me.

$\endgroup$

1 Answer 1

0
$\begingroup$

Full solution, using Dirac delta: $$ p_U(u)=\int_D dx dy \frac{1}{\pi}\delta\left(u-\sqrt{x^2+y^2}\right)\ . $$ Making a change to polar coordinates $x=r\cos\theta$ and $y=r\sin\theta$, whose Jacobian is $=r$, $$ p_U(u)=\frac{1}{\pi}\int_0^{2\pi}d\theta\int_0^1 dr\ r\ \delta(r-u)=2u\qquad 0\leq u\leq 1\ , $$ which is correctly normalized, $\int_0^1 du\ p_U(u)=1$.


About the Jacobian, I am not sure what the problem is [and also not sure about the statement '2-to-1' you made...you had two variables $X,Y$ and introduce two other variables $U,V$]. The Jacobian between the old variables $(X,Y)$ and the new ones $(U,V)$ is $$ \det\begin{pmatrix} \frac{\partial u}{\partial x} & \frac{\partial u}{\partial y}\\ \frac{\partial v}{\partial x} & \frac{\partial v}{\partial y} \end{pmatrix} = \det\begin{pmatrix} \frac{x}{\sqrt{x^2+y^2}} & \frac{y}{\sqrt{x^2+y^2}} \\ -\frac{y}{x^2+y^2} & \frac{x}{x^2+y^2} \end{pmatrix}=\frac{1}{\sqrt{x^2+y^2}}=\frac{1}{u} $$ (obviously, then, the Jacobian between the new variables and the old ones is $u$).

$\endgroup$
1
  • $\begingroup$ Ah, normally I would have computed the elements as $\frac{\partial x}{\partial u}$, I didn't know I could do it the way you did it, so that was very helpful. :) What I mean by saying it is 2-to-1 is that $U$ is the same for $(X,Y)$ as it is for $(-X,-Y)$. I guess it wasn't really relevant for my question, so I apologize if it was confusing. $\endgroup$ Feb 23, 2016 at 6:11

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .