An Introduction to Heavy-Tailed and Subexponential by Sergey Foss, Dmitry Korshunov, Stan Zachary

By Sergey Foss, Dmitry Korshunov, Stan Zachary

Heavy-tailed chance distributions are a massive part within the modeling of many stochastic platforms. they're often used to competently version inputs and outputs of laptop and information networks and repair amenities equivalent to name facilities. they're a necessary for describing danger procedures in finance and likewise for coverage premia pricing, and such distributions happen certainly in versions of epidemiological unfold. the category comprises distributions with strength legislation tails comparable to the Pareto, in addition to the lognormal and likely Weibull distributions.

One of the highlights of this new version is that it contains difficulties on the finish of every bankruptcy. bankruptcy five is usually up to date to incorporate fascinating purposes to queueing concept, threat, and branching procedures. New effects are provided in an easy, coherent and systematic way.

Graduate scholars in addition to modelers within the fields of finance, coverage, community technology and environmental reviews will locate this e-book to be a necessary reference.

Show description

Read Online or Download An Introduction to Heavy-Tailed and Subexponential Distributions PDF

Similar mathematical physics books

An Introduction to Chaos in Nonequilibrium Statistical Mechanics

This publication is an advent to the purposes in nonequilibrium statistical mechanics of chaotic dynamics, and in addition to using thoughts in statistical mechanics very important for an knowing of the chaotic behaviour of fluid structures. the elemental thoughts of dynamical platforms thought are reviewed and easy examples are given.

Labyrinth of Thought: A History of Set Theory and Its Role in Modern Mathematics

"José Ferreirós has written a magisterial account of the background of set concept that's panoramic, balanced, and interesting. not just does this e-book synthesize a lot past paintings and supply clean insights and issues of view, however it additionally includes a significant innovation, a full-fledged remedy of the emergence of the set-theoretic process in arithmetic from the early 19th century.

Computational Physics: Problem Solving with Python

Using computation and simulation has turn into a vital a part of the medical method. with the ability to rework a idea into an set of rules calls for major theoretical perception, special actual and mathematical figuring out, and a operating point of competency in programming. This upper-division textual content offers an surprisingly large survey of the themes of contemporary computational physics from a multidisciplinary, computational technological know-how perspective.

Extra resources for An Introduction to Heavy-Tailed and Subexponential Distributions

Sample text

Fn (x) ≥ (1 − ε )n−1 . F 1 (x) + . . 28) now follows by letting ε → 0. 29 yields the following corollary. 30. Let the distribution F on R be long-tailed (F ∈ L). Then, for any n ≥ 2, lim inf x→∞ F ∗n (x) ≥ n. 29. 31. Let the distributions F and G on R be such that F is long-tailed (F ∈ L). Then, lim inf x→∞ F ∗ G(x) ≥ 1. 30) Proof. Let ξ and η be independent random variables with respective distributions F and G. For any fixed a, F ∗ G(x) ≥ P{ξ > x − a, η > a} = F(x − a)G(a). For every ε > 0 there exists a such that G(a) ≥ 1 − ε .

Let ξ1 and ξ2 be two independent random variables with distribution F. 37 that, for some c < ∞, P{η1 + η2 > x, η1 > h(x), η2 > h(x)} ≤ cP{ξ1 + ξ2 > x, ξ1 > h(x), ξ2 > h(x)}. 6, together with the immediately preceding remark, imply that P{ξ1 + ξ2 > x, ξ1 > h(x), ξ2 > h(x)} = o(F(x)). Hence the result follows. 3 Subexponentiality and Weak Tail-Equivalence We start with the definition of weak tail-equivalence and then use this property to establish a number of powerful results. 9. Two distributions F and G with right-unbounded supports are called weakly tail-equivalent if there exist c1 > 0 and c2 < ∞ such that, for any x > 0, c1 ≤ F(x) ≤ c2 .

Suppose that F is whole-line subexponential and that the function h is such that h(x) → ∞ as x → ∞. Let the distributions G1 , G2 be such that, for i = 1, 2, we have Gi (x) = O(F(x)) as x → ∞. If η1 and η2 are independent random variables with distributions G1 and G2 , then P{η1 + η2 > x, η1 > h(x), η2 > h(x)} = o(F(x)) as x → ∞. Proof. Let ξ1 and ξ2 be two independent random variables with distribution F. 37 that, for some c < ∞, P{η1 + η2 > x, η1 > h(x), η2 > h(x)} ≤ cP{ξ1 + ξ2 > x, ξ1 > h(x), ξ2 > h(x)}.

Download PDF sample

Rated 4.54 of 5 – based on 15 votes