6
$\begingroup$

I have the continuous symmetric random variable $X$ in $\mathbb{R}$. If I know its distribution function $F(x)$ what are the conditions on $F(x)$ so that $X=Y_1 - Y_2$ where $Y_i$ are also iid continuous symmetric random variables in $\mathbb{R}$.

$\endgroup$
3
  • 1
    $\begingroup$ Can you give some examples and non-examples to help find reasonable conditions? $\endgroup$ Aug 21, 2014 at 13:10
  • 2
    $\begingroup$ One obvious necessary condition is that the characteristic function (aka Fourier transform) $\int e^{itx} dF(x)$ must be nonnegative. $\endgroup$ Aug 21, 2014 at 15:40
  • 4
    $\begingroup$ ... and (some branch of) its square root will then be the characteristic function of $Y_i$. By Bochner's theorem you just have to check that this square root is positive definite. $\endgroup$ Aug 21, 2014 at 16:47

1 Answer 1

1
$\begingroup$

No conditions are necessary. Let $Y_2$ be any continuous symmetric random variable independent of $X$. Then $X + Y_2$ is a continuous symmetric random variable.

Perhaps you wanted $Y_1$ and $Y_2$ to be independent?

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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