0
$\begingroup$

Assume X$_1$, ... , X$_n$ are i.i.d distributed random variables. And each them of has the same probability density function as f(x) = $\frac{1}{\pi}$$\frac{1}{1+ x^2}$. How can we compute the probability density function for X = $\frac{1}{n}$ $\sum_{i=1}^{n}$X$_i$?

Some idea is to write the cumulative distribution function:

F$_{X_1+X_2+...+X_n}$(a) = P(X$_1$ + X$_2$ + ... + X$_n$ $\leq$ a) = $\idotsint$ f$_{X_1}$(x$_1$)f$_{X_2}$(x$_2$) $\cdots$ $\cdots$ f$_{X_n}$(x$_n$)d$_{x_1}$d$_{x_2}$$\cdots$d$_{x_n}$

But feel hard to continue. Any suggestions?

$\endgroup$
  • 3
    $\begingroup$ Try using the characteristic function (if you know about this). Note that $X_i$ come from a standard Cauchy distribution (see en.wikipedia.org/wiki/Cauchy_distribution; $x_0=0$ and $\gamma=1$ there for the standard Cauchy distribution). See stats.stackexchange.com/questions/238246/… for a proof with the characteristic function. $\endgroup$ – Minus One-Twelfth Mar 28 at 8:31
  • $\begingroup$ thanks, is there any way to calculate it? btw, why can we say they have same pdfs if their characteristic functions are the same? $\endgroup$ – Jonny Mar 28 at 21:44
2
$\begingroup$

It also has density $\frac 1 {\pi} \frac 1 {1+x^{2}}$. The best way to prove this is to use characteristic functions: since $Ee^{itX_1}=e^{-|t|}$ and $(e^{-|\frac t n|})^{n}=e^{-|t|}$ we get the conclusion easily.

$\endgroup$
  • $\begingroup$ Thanks, btw, how can we prove the characteristic function for standard Cauchy distribution? $\endgroup$ – Jonny Mar 28 at 19:57
  • $\begingroup$ ignore previous comment. any suggestion on traditional way to calculate it rather than using characteristic functions? $\endgroup$ – Jonny Mar 28 at 22:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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