Some examples of random walks on free products of discrete groups

Summary We consider the random walk (X_n) associated with a probability p on a free product of discrete groups. Knowledge of the resolvent (or Green's function) of p yields theorems about the asymptotic behaviour of the n-step transition probabilities p^*n(x)=P(X_n= x¦ X₀=e) as n. Woess [15], Cartwright and Soardi [3] and others have shown that under quite general conditions there is behaviour of the type p^*n(x)C_x^– n n^– 3/2. Here we show on the other hand that if G is a free product of m copies ofZ ^r and if (X_n) is the « average » of the classical nearest neighbour random walk on each of the factorsZ ^r, then while it satisfies an « n^–3/2 — law » for r small relative to m, it switches to an n^– r/2 -law for large r. Using the same techniques, we give examples of irreducible probabilities (of infinite support) on the free groupZ ^*m which satisfyn ^– for .     

Limit Theorems for the Maximum Terms of a Sequence of Random Variables with Marginal Geometric Distributions

Let X _n1 ^* , ... X _nn ^* be a sequence of n independent random variables which have a geometric distribution with the parameter p _n = 1/n, and M _n ^* = \max\{X _n1 ^* , ... X _nn ^* }. Let Z ₁, Z₂, Z₃, ... be a sequence of independent random variables with the uniform distribution over the set N _n = {1, 2, ... n}. For each j N _n let us denote X _nj = min{k : Z_k = j}, M _n = max{X_n1, ... X_nn}, and let S _n be the 2nd largest among X _n1, X_n2, ... X_nn. Using the methodology of verifying D(u_n) and D'(u_n) mixing conditions we prove herein that the maximum M _n has the same type I limiting distribution as the maximum M _n ^* and estimate the rate of convergence. The limiting bivariate distribution of (S_n, M_n) is also obtained. Let _n, _n N_n, , and T _n = min{M(A_n), M(B_n)}. We determine herein the limiting distribution of random variable T _n in the case _n , _n/_n > 0, as n .     

On unconditional bases in certain Banach function spaces

Let X be a Banach space, L ([0,1])XL ¹([0,1]), with an unconditional basis. By the well-known stability property in X, there exists a unconditional basis {f _n} _m=1 , where f _n in C([0,1]), nN. In this paper, we introduce the notion that X ^*has the singularity property of X ^*at a point t ₀[0,1]. It is proved that if X ^*has the singularity property at a point t ₀ [0,1], then there exists no orthonormal, fundamental system in C([0,1]) which forms an unconditional basis in X.     

Future independent times and Markov chains

Summary A random timeT is a future independent time for a Markov chain (X _n ) ₀ ifT is independent of (X _T+n ) _n ^/ =0 and if (X _T+n ) _n ^/ =0 is a Markov chain with initial distribution and the same transition probabilities as (X _n ) ₀ . This concept is used (with the conditional stationary measure) to give a new and short proof of the basic limit theorem of Markov chains, improving somewhat the result in the null-recurrent case.This work was supported by the Swedish Natural Science Research Council and done while the author was visiting the Department of Statistics, Stanford University     

Analysis of Henrici's Transformation for Singular Problems

Henrici's transformation is the underlying scheme that generates, by cycling, Steffensen's method for the approximation of the solution of a nonlinear equation in several variables. The aim of this paper is to analyze the asymptotic behavior of the obtained sequence (s _n ^* ) by applying Henrici's transformation when the initial sequence (s _n ) behaves sublinearly. We extend the work done in the regular case by Sadok [17] to vector sequences in the singular case. Under suitable conditions, we show that the slowest convergence rate of (s _n ^* ) is to be expected in a certain subspace N of R ^p . More precisely, if we write s _n ^* =s _n ^* ,N+s _n ^* ,N, the orthogonal decomposition into N and N , then the convergence is linear for (s _n ^* ,N) but ( _n ^* ,N) converges to the same limit faster than the initial one. In certain cases, we can have N=R ^p and the convergence is linear everywhere.     

Rare numbers

Suppose thatX ₁,X ₂,... is a sequence of iid random variables taking values inZ ⁺. Consider the random sequenceA(X)(X ₁,X ₂,...). LetY _n be the number of integers which appear exactly once in the firstn terms ofA(X). We investigate the limit behavior ofn ^–(1–) Y _n for [0, 1].     

On the estimation of ordered means of two exponential populations

Let random samples of equal sizes be drawn from two exponential distributions with ordered means _i . The maximum likelihood estimator _i ^* of _i is shown to have a smaller mean square error than that of the usual estimator X_i, for each i=1,2. The asymptotic efficiency of _i ^* relative to X_i has also been found.     

Free Wreath Product by the Quantum Permutation Group

Let A be a compact quantum group, let nN ^* and let A _aut(X _n ) be the quantum permutation group on n letters. A free wreath product construction A*_w A _aut(X _n ) is introduced. This construction provides new examples of quantum groups, and is useful to describe the quantum automorphism group of the n-times disjoint union of a finite connected graph.     

A Remark on Bernstein,Prokhorov, Bennett,Hoeffding, and Talagrand Inequalities

Let X ₁,..., X_n be independent random variables such that {X_j 1}=1 and E X _j=0 for all j. We prove an upper bound for the tail probabilities of the sum M _n=X₁+...+ X_n. Namely, we prove the inequality {M _nx} 3.7 {S_n x}, where S _n=₁+...+ _n is a sum of centered independent identically distributed Bernoulli random variables such that E S _n ² =ME M _n ² and {_k=1}=E S _n ² /(n+E S _n ² ) for all k (we call a random variable Bernoulli if it assumes at most two values). The inequality holds for x at which the survival function x{S _nx} has a jump down. For remaining x, the inequality still holds provided that we interpolate the function between the adjacent jump points linearly or log-linearly. If necessary, in order to estimate {S _nx} one can use special bounds for binomial probabilities. Up to the factor at most 2.375, the inequality is final. The inequality improves the classical Bernstein, Prokhorov, Bennett, Hoeffding, Talagrand, and other bounds.     

Topological Representation of Sheaf Cohomology of Sites

For a site & (with enough points), we construct a topological space X(&) and a full embedding ^* of the category of sheaves on & into those on X _(&) (i.e., a morphism of toposes :Sh (X_(&)) Sh(&)). The embedding will be shown to induce a full embedding of derived categories, hence isomorphisms H^*(&,A) = H^*(X_(&), ^*A) for any Abelian sheaf A on &. As a particular case, this will give for any scheme Y a topological space X _(Y) and a functorial isomorphism between the étale cohomology H^*(Y _ét,A) and the ordinary sheaf cohomology H^*(X(_(Y),),^*A), for any sheaf A for the étale topology on Y.     

Combined nonparametric inference and state estimation for mixed poisson processes

Summary Given independent, identically distributed copies of a mixed Poisson process N on a LCCB space E, i.e., a Cox process whose directing measure is of the form m ^*, where 0 is a random variable with distribution and m ^* is a measure on E, we construct strongly consistent and asymptotically normal estimators of m ^* and the Laplace transform l . Methods are presented for estimating the directing measure of the (n+1)^st process by combining the data for that process with estimates of appropriate quantities, the latter based on the first n processes. The case where different processes are observed over different sets is addressed.Research supported in part by the Air Force Office of Scientific Research, USAF, grant number AFOSR 82-0029. The United States Government is authorized to reproduce and distribute reprints for Governmental purposes     

Operator stable laws: Series representations and domains of normal attraction

LetX,X ₁,X ₂,... be i.i.d. random vectors in ^d. The limit laws that can arise by suitable affine normalizations of the partial sums,S _n=X ₁+...+X _n, are calledoperator-stable laws. These laws are a natural extension to ^d of the stable laws on. Thegeneralized domain of attraction of [GDOA()] is comprised of all random vectorsX whose partial sums can be affinely normalized to converge to . If the linear part of the affine transformation is restricted to take the formn ^–B for some exponent operatorB naturally associated to thenX is in thegeneralized domain of normal attraction of [GDONA()]. This paper extends the theory of operator-stable laws and their domains of attraction and normal attraction.     

Limiting distributions of homogeneous functions of sample spacings for distributions approaching the uniform distribution

Summary X ₁,...,X _n are independent random variables, identically distributed over the unit interval, with common probability density function 1 + r(x)/n for all sufficiently large n, where is a positive constant, and |r(x)| <D. V ₁, ..., V _n+1 are the sample spacings generated by X ₁,..., X _n . It is shown that in many cases, the asymptotic joint distribution of homogeneous functions of V ₁,..., V _n+1 can be found directly from the asymptotic joint distribution of homogeneous functions of independent exponential random variables.Research supported by NSF Grant GP 3783.     

Estimates of the tail of the stationary density function of certain nonlinear autoregressive processes

We consider the Markov chainX _n+1=T(X _n )+ _n , where { _n ;n1} is a ^d -valued random sequence of independent identically distributed random variables, and the functionT: ^{d d} is measurable and satisfies a suitable growth condition. Under certain conditions involvingT and the probability distribution of _n , we show that this Markov chain is ergodic. Moreover, we obtain sharp upper bounds for the tail of the corresponding stationary probability density function. In our proofs, we make use of the Leray-Schauder fixed-point theorem.     

Asymptotic behavior of the prediction error

Let {X_n} _– be a random process, stationary in the broad sense, with spectral density f() satisfying the singularity condition: · We denote _n ² the mean square prediction error at the prediction of _o by linear forms in X_–1, ... , X_–n. In the paper one investigates the rate of decrease of _n to zero.Translated from Zapiski Nauchnykh Seminarov Leningradskogo Otdeleniya Matematicheskogo Instituta im. V. A. Steklova AN SSSR, Vol. 130, pp. 11–24, 1983.In conclusion, the author wishes to express his gratitude to I. A. Ibragimov for his constant interest and help.     

An example of an irregular random measure

Summary We consider the set of random measures which consists of measurable maps from [0, 1] to the set of measures on . As it is the dual space ofL ₁ ([0, 1];C()), we can equip this space with the weak^* topology. We construct a special random measure , which appears as the weak^* limit of a sequence of Dirac random measures , where (X _n ) _n is a bounded sequence inL _p [0, 1], (1p<2). The special form of this random measure, which oscillates randomly between twoq-stable standard measures on with different normalizations (p<q<2) allows us to prove two properties of (X _n ) _n is equivalent to the unit vector basis ofl _q and has no almost symmetric subsequence.     

On the rate of convergence of the sum of the sample extremes

Summary Let X ₁, X ₂,..., X _n be independent random variables having a common distribution in the domain of normal attraction of a completely asymmetric stable law with characteristic exponent }(0,1) and support bounded below. Let X _n:n X _n:n ^-1...X _n:1 denote the ordered sample. We obtain the rate of convergence of n ^-1/ (X _n:n +...+X _{n:n-k n+1} ) to the stable limit law as both n and k _n ». As a consequence we obtain a representation of the sum X _n:n +...+X _{n:n-k n+1} .     

Craig-Sakamoto's theorem for the Wishart distributions on symmetric cones

A version of Craig-Sakamoto's theorem says essentially that ifX is aN(O,I _n ) Gaussian random variable in ⁿ, and ifA andB are (n, n) symmetric matrices, thenXAX andXBX (or traces ofAXX andBXX) are independent random variables if and only ifAB=0. As observed in 1951, by Ogasawara and Takahashi, this result can be extended to the case whereXX is replaced by a Wishart random variable. Many properties of the ordinary Wishart distributions have recently been extended to the Wishart distributions on the symmetric cone generated by a Euclidean Jordan algebraE. Similarly, we generalize there the version of Craig's theorem given by Ogasawara and Takahashi. We prove that ifa andb are inE and ifW is Wishart distributed, then Tracea.W and Traceb.W are independent if and only ifa.b=0 anda.(b.x)=b.(a.x) for allx inE, where the. indicates Jordan product.Partially supported by NATO grant 92.13.47.     

Strong Laws for Martingale Differences and Independent Random Variables

A strong law of large numbers (SLLN) for martingale differences {X _n,_n,n1} permitting constant, random or hybrid normalizations, is obtained via a related SLLN for their conditional variances E{X _n ² |_n-1}n1. This, in turn, leads to martingale generalizations of known results for sums of independent random variables. Moreover, in the independent case, simple conditions are given for a generalized SLLN which contains the classical result of Kolmogorov when the variables are i.i.d.     

On characterizations of distributions via moments of record values

Summary Let {X _n } _n ⁼¹ be a sequence of i.i.d. random variables having continuous distribution F(x) with E|X| ^l+< for some positive integer l and for some >0. It is shown that for any fixed integer N0 the sequence of moments of record values {E(X _L(n) ) ^l } _n=N characterizes F. Furthermore, this result is applied to the weak convergence of continuous distributions.