Rates of convergence in the operator-stable limit theorem (Q1908204)

When one deals with random vectors, then the normalizations by linear operators (matrices in \(\mathbb{R}^d\)) seem to be the most natural [see the reviewer and \textit{J. D. Mason}, ``Operator-limit distributions in probability theory'' (J. Wiley, New York, 1993)]. For i.i.d. summands it leads to the operator-stable distributions that are characterized by their operator exponents. The authors study speed of convergence to operator-stable distributions under different kinds of probability metrics. In particular, they introduce a new metric depending on the operator exponent.