Deriving the joint probability distribution of a transformed rv

179 Views Asked by At

Suppose that A and B are 2 r.v with joint probability distribution given by $f_{AB}(a,b) = -\frac{1}{2}(ln(a)+ln(b))$, if $0 \lt a \lt 1$ and $0 \lt b \lt 1$, and 0 otherwise.

Define $X = A+B$ and $Y = \frac{A}{B}$.

How do I derive the joint distribution of $f_{XY}(x,y)$?

Progress: I know that X will take values $0 \lt x \lt 2$, and the Jacobian = $2y$ . Any help will be much appreciated.

1

There are 1 best solutions below

4
On BEST ANSWER

The jacobian results to me

$$|J|=\frac{x}{(y+1)^2}$$

and after standard calculations the joint density results to me

$$f_{XY}(x,y)=\frac{-x}{2(y+1)^2}\log\frac{x^2y}{(y+1)^2}\cdot\left[\mathbb{1}_{(0;1]}(x)\mathbb{1}_{(0;+\infty)}(y)+\mathbb{1}_{(1;2)}(x)\mathbb{1}_{\left(x-1;\frac{1}{x-1}\right)}(y)\right]$$

This because when $x\in(0;1]$ there are no problems but when $1<x<2$, in order to have

$$0<\frac{xy}{y+1}<1$$

And

$$0<\frac{x}{y+1}<1$$

You need to have

$$x-1<y<\frac{1}{x-1}$$


Alternative notation


Joint support can alternatively be expressed by

$$0<x<\min\left(\frac{y+1}{y};y+1\right)$$

$$0<y<+\infty$$

To understand this alternative notation it is enough to do a drawing of support region

enter image description here