PlanetMath (more info)
 Math for the people, by the people.
Encyclopedia | Requests | Forums | Docs | Wiki | Random | RSS  
Login
create new user
name:
pass:
forget your password?
Main Menu
Owner confidence rating: High Entry average rating: No information on entry rating
Cauchy random variable (Definition)

$ X$ is a Cauchy random variable with parameters $ \theta\in\mathbb{R}$ and $ \beta>0\in\mathbb{R}$, commonly denoted $ X\sim Cauchy(\theta,\beta)$ if

$\displaystyle f_X(x)=\frac{1}{\pi\beta[1+(\frac{x-\theta}{\beta})^2]}.$    

Cauchy random variables are used primarily for theoretical purposes, the key point being that the values $ E[X]$ and $ Var[X]$ are undefined for Cauchy random variables.



"Cauchy random variable" is owned by mathcam. [ full author list (2) | owner history (1) ]
(view preamble)

View style:

Other names:  Cauchy distribution
Log in to rate this entry.
(view current ratings)

Cross-references: point, parameters
There are 4 references to this entry.

This is version 6 of Cauchy random variable, born on 2001-10-26, modified 2007-06-24.
Object id is 531, canonical name is CauchyRandomVariable.
Accessed 8706 times total.

Classification:
AMS MSC60A10 (Probability theory and stochastic processes :: Foundations of probability theory :: Probabilistic measure theory)

Pending Errata and Addenda
None.
[ View all 4 ]
Discussion
Style: Expand: Order:
forum policy
Why is the expected value not zero if theta = 0? by sprocketboy on 2005-08-15 16:56:24
I encountered a Cauchy distribution when solving a problem in "Communications Engineering" (3rd Edition) by Proakis and Salehi.

Let X and Y be independent Gaussian random variables, each with distribution N(0, o^2). ('o' is how I am writing "sigma" here).

Find the pdf of the random variable R = X / Y. What is its mean and variance?

After a little calculation, I came up with:

f(r) = 1 / (pi * (1 + r^2)), which looks like a Cauchy pdf with beta = 1 and theta = 0. The transformation r = tan(u) allows you to calculate that the integral of f(r)dr from -inf to inf is unity, which is what you would expect.

However, when I try to calculate E(r), I get zero. In fact, the nth moment seems to evaluate to zero for all odd n. If n is even, then the nth moment evaluates to infinity.

I tried calculating E(r) by evaluating the integral r*f(r)dr from -T to T, where T>0, then taking the limit as T->inf. I still get zero.

The denominator is never zero for nonzero beta, and a plot of f(r) shows it to have even symmetry about theta, which in my case is zero.

What am I missing, here?
[ reply | up ]

Interact
post | correct | update request | add derivation | add example | add (any)