An invariance principle for sums and record times of regularly varying\n stationary sequences

Bojan Basrak, Hrvoje Planinić, Philippe Soulier

arXiv (Cornell University) · 2016 · 37 citations · 18 references

DOIFull text

Open access

Concepts

Abstract

We prove a sequence of limiting results about weakly dependent stationary and\nregularly varying stochastic processes in discrete time. After deducing the\nlimiting distribution for individual clusters of extremes, we present a new\ntype of point process convergence theorem. It is designed to preserve the\nentire information about the temporal ordering of observations which is\ntypically lost in the limit after time scaling. By going beyond the existing\nasymptotic theory, we are able to prove a new functional limit theorem. Its\nassumptions are satisfied by a wide class of applied time series models, for\nwhich standard limiting theory in the space $D$ of \\cadlag\\ functions does not\napply.\n To describe the limit of partial sums in this more general setting, we use\nthe space~$E$ of so--called decorated \\cadlag\\ functions. We also study the\nrunning maximum of partial sums for which a corresponding functional theorem\ncan be still expressed in the familiar setting of space $D$.\n We further apply our method to analyze record times in a sequence of\ndependent stationary observations, even when their marginal distribution is not\nnecessarily regularly varying. Under certain restrictions on dependence among\nthe observations, we show that the record times after scaling converge to a\nrelatively simple compound scale invariant Poisson process.\n

References

18