Take the 2-minute tour ×
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.

I am having trouble making sense of this. I know what independence of random variables means. Suppose $X$ and $Y$ are independent, standard normal (Gaussian) random variables. Then, it is supposed to be that $X^2 + Y^2$ and $\frac X Y$ are also independent random variables.

I just cannot intuitively make sense of this. Where does that fact that they are standard normal Gaussian come into play? Also, if we know that $\frac X Y \leq k$ for some fixed $k$, then it seems obvious that it should affect the probability distribution of $X^2 +Y^2$.

share|improve this question
add comment

2 Answers 2

Consider the random point $(X,Y)$ in $\mathbb{R}^2$. The ratio $X/Y$ tells us what angle the segment from $(0,0)$ to $(X,Y)$ makes with the $x$-axis, while $X^2+Y^2$ tells us how far $(X,Y)$ is from $(0,0)$.

The distribution of $(X,Y)$ is symmetric under rotations, so the distribution of the angle $\Theta$ is uniform and independent of the radius $R=\sqrt{X^2+Y^2}$.

This intuitive explanation can be made more rigorous by converting to polar coordinates.


The joint density of two independent standard normals $(X,Y)$ is $$f(x,y)={1\over 2\pi} \exp(-(x^2+y^2)/2).$$ Converting to polar coordinates we get the joint density of $(R,\Theta)$ as $$g(r,\theta)={r\over 2\pi}\exp(-r^2/2)={1\over 2\pi}\cdot r\exp(-r^2/2).$$ This product form of $g(r,\theta)$ shows that $\Theta$ and $R$ are independent.

share|improve this answer
    
where does gaussian standard normal come into play?? –  Rama Apr 5 '12 at 14:33
1  
For independent standard normals, the joint distribution is invariant under rotations. This is the crucial fact that makes $X/Y$ and $X^2+Y^2$ independent. I've added a better explanation to my answer. –  Byron Schmuland Apr 5 '12 at 14:44
add comment

Additional remark to Byron's answer: It can actually be shown that an (isotropic) Normal distribution is the only rotationally symmetric distribution for which $X$ and $Y$ as well as $X/Y$ and $X^2 + Y^2$ are independent.

See e.g.

Ali, M. M. (1980). Characterization of the Normal Distribution Among the Continuous Symmetric Spherical Class. Journal of the Royal Statistical Society B, 42(2), 162-164.

share|improve this answer
add comment

Your Answer

 
discard

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.