2
$\begingroup$

I know that one can generalise the classical CLT in terms of heavy tail distributions, namely, for any i.i.d. random variables $X_i$, $$\frac{X_1+\cdots+X_n}{n^{1/\alpha}}\rightarrow S(\alpha,\beta,\gamma,\delta)$$ in distribution sense, whenever $X_i$ belongs to the domain of attraction of its corresponding limit. When $\alpha$ takes $2$, this law becomes the classical CLT.

To push this to multivariate versions, I also see https://en.wikipedia.org/wiki/Multivariate_stable_distribution, in which is multivariate stable distribution is introduced.

Where can I find a reference for such multivariate generalised central limit theorem? A textbook reference would be great.


P.S. This is also an existing question in MSE: https://math.stackexchange.com/questions/3786947/reference-for-multivariate-generalised-clt?noredirect=1#comment7800922_3786947

$\endgroup$
2
$\begingroup$

$\newcommand\al\alpha\newcommand\R{\mathbb R}$The domains of attraction to multidimensional stable distributions are characterized by Rvačeva's Theorems 4.1 (p. 194, for $\al=2$) and 4.2 (p. 196, for $\al<2$).

The domains of strict attraction to multidimensional strictly stable distributions are characterized by Shimura's Theorems 3.1 (p. 354, for $\al=1$), 3.3 (p. 356, for $\al\in(0,1)$), 3.4 (p. 356, for $\al\in(1,2)$), and 3.5 (p. 356, for $\al=2$).

$\endgroup$

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.