Sign up ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Given a (continuous) bivariate random vector $(X,Y)$ with a probability density:

$$f_{XY}(x,y)=\left\{ \begin{array}{l}\tfrac{1}{y}e^{-(y+\tfrac{x}{y})} & \text{if }x>0 \text{ and } y>0 \\ 0 & \text{otherwise}\end{array} \right.$$

Find the marginal density $f_X(x)$ and $f_Y(y)$ and show whether $X$ and $Y$ are stochastically independent.

In my attempt, I tried to calculate the marginal densities as follows:

$$\begin{align*}f_Y(y) &= \int_{-\infty}^{\infty}f_{XY}(x,y)\mathrm{d}x \\ &= \int_{0}^{\infty}\tfrac{1}{y}e^{-(y+\tfrac{x}{y})}\mathrm{d}x\\ &= -\lim_{N \to \infty} \left[ e^{-(y+\tfrac{x}{y})} \right]_0^N \\ &= -\lim_{N \to \infty} (e^{-(y+\tfrac{N}{y})}-e^{-(y+\tfrac{0}{y})})\\ &= e^{-y}-0 = \frac{1}{e^y} \end{align*}$$ $\\$ $$\begin{align*}f_X(x) &= \int_{-\infty}^{\infty}f_{XY}(x,y)\mathrm{d}y \\ &= \int_{0}^{\infty}\tfrac{1}{y}e^{-(y+\tfrac{x}{y})}\mathrm{d}y\\ \end{align*}$$

But I get stuck at solving this last integrals. How should this be solved?

Thanks in advance.

share|cite|improve this question
This looks a lot like homework, and if so, please add the homework tag. Also, could you please check to make sure that you have copied the problem correctly? If $X$ and $Y$ are stochastically independent, then you should be able to write $f_{X,Y}(x,y)$ as $g(x)h(y)$ and this does not seem obviously so. – Dilip Sarwate Jan 11 '13 at 2:01
Actually, it's an exercise from last years exam for which I'm preparing. And I saw my mistake, it should have read "show whether X and Y are stochastically independent". – Jeroen Jan 11 '13 at 2:12

1 Answer 1

The expression for $f_X(x)$ is not elementary. In fact, the integral is a defining integral for the Macdonald function: $$ f_X(x) = \int_0^\infty \exp\left(-y - \frac{x}{y}\right)\frac{\mathrm{d}y}{y} = \left.\left( \underbrace{\left(\frac{z}{2}\right)^\nu \int_0^\infty \exp\left(-t-\frac{z^2}{4t} \right) \frac{\mathrm{d}t}{t^{\nu+1}}}_{2 K_\nu\left(z\right)}\right)\right|_{\nu=0,z=2 \sqrt{x}} = 2 K_0\left(2 \sqrt{x}\right) $$ where $x$ was assumed positive. Anyhow, you readily conclude that $f_{X,Y}(x,y) \not= f_X(x) f_Y(y)$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.