In this paper, we deal with the asymptotic distribution of the maximum increment of a random walk with a regularly varying jump size distribution. This problem is motivated by a longstanding problem on change point detection for epidemic alternatives. It turns out that the limit distribution of the maximum increment of the random walk is one of the classical extreme value distributions, the Fréchet distribution. We prove the results in the general framework of point processes and for jump sizes taking values in a separable Banach space.Keywords: Banach space valued random element; epidemic change point; extreme value theory; Fréchet distribution; maximum increment of a random walk; point process convergence; regular variation This is an electronic reprint of the original article published by the ISI/BS in Bernoulli, 2010, Vol. 16, No. 4, 1016-1038. This reprint differs from the original in pagination and typographic detail.
1350-7265 c 2010 ISI/BSThe limit distribution of the maximum increment of a random walk 1017 mention that nothing seems to be known about the distributional properties of T n . There exist several approaches to replace the original problem by a more tractable one. One way is to restrict the range over which the maximum is taken to ℓ n ≤ ℓ ≤ n − ℓ n for some ℓ n → ∞ satisfying ℓ n = o(n); see, for example, Yao [24]. Alternatively, one can change the normalizing constants ℓ(1 − ℓ/n) in a suitable way; see, for example, Račkauskas and Suquet [20].Statistics of type T n appear in the context of tests for change points in the mean under epidemic alternatives. This problem can be formulated as follows: given that X 1 , . . . , X n are independent random variables, test the null hypothesis of constant mean • H 0 : EX 1 = EX 2 = · · · = EX n = µ against the epidemic alternative • H A : There exist integers 1 ≤ k * < m * < n such that EX 1 = · · · = EX k * = EX m * +1 = · · · = EX n = µ, EX k * +1 = · · · = EX m * = ν and µ = ν.