Local limit theorems for transition densities of Markov chains converging to diffusions

We consider triangular arrays of Markov chains that converge weakly to a diffusion process. Local limit theorems for transition densities are proved. Received: 28 August 1998 / Revised version: 6 September 1999 / Published online: 14 June 2000     

Deterministic equivalents of additive functionals of recurrent diffusions and drift estimation

Let X _t be a one-dimensional Harris recurrent diffusion, with a drift depending on an unknown parameter θ belonging to some metric compact Θ. We firstly show that all integrable additive functionals of X _t are asymptotically equivalent in probability to some deterministic process v _t . Then we use this result to study the behavior of the maximum likelihood estimator for the parameter θ. Under mild regularity assumptions, we find an upper rate of its convergence as a function of v _t , extending some recent results for ergodic diffusions.      

Functional law of iterated logarithm for additive functionals of reversible Markov processes

1.IntroductionandMainResultsAssumethat(Xt),.T(T~NorAl)isaPolishspaceE-valuedMarkovprocess,definedon(fi,F,(R),(ot),(P-c)..E),withitssemigroupoftransitionkernels(Pt).Here(ot)isthesemigroupofshiftsonfisuchthatX.(otw)~X. t(w),Vs,tET;(R)isthenaturalfiltration.Throughoutthispaperweassumethat(Pt)issymmetricandergodicwithrespectto(w.r.t.forshort)aprobabilitymeasurepon(E,e)(eistheBorela--fieldofE),i.e.,.Symmetry:(Ptf,g)~(f,Pig):~isfptgdp,acET,if,gCL'(P);.ErgodicitytFOranyfEL'(P),ifPtf~f…     

The Law of the Iterated Logarithm for Functionals of Harris Recurrent Markov Chains: Self Normalization

Let {X _n } _n0 be a Harris recurrent Markov chain with state space E, transition probability P(x, A) and invariant measure , and let f be a real measurable function on E. We prove that with probability one,

A note on some laws of the iterated logarithm

Some function space laws of the iterated logarithm for Brownian motion with values in finite and infinite dimensional vector spaces are shown to follow from Hincin's classical law of the iterated logarithm and some martingale techniques. A law of the iterated logarithm for Brownian motion in a differentible manifold is also stated.     

A law of the iterated logarithm for Markov chains on ℕ0 associated with orthogonal polynomials

We derive laws of the iterated logarithm for Markov chains on the nonnegative integers whose transition probabilities are associated with a sequence of orthogonal polynomials. These laws can be applied to a large class of birth and death random walks and random walks on polynomial hypergroups. In particular, the results of our paper lead immediately to a law of the iterated logarithm for the growth of the distance of isotropic random walks on infinite distance-transitive graphs as well as on certain finitely generated semigroups from their starting points.     

Waiting time distributions of runs in higher order Markov chains

We consider a {0,1}-valuedm-th order stationary Markov chain. We study the occurrences of runs where two 1’s are separated byat most/exactly/at least k 0’s under the overlapping enumeration scheme wherek≥0 and occurrences of scans (at leastk ₁ successes in a window of length at mostk, 1≤k ₁≤k) under both non-overlapping and overlapping enumeration schemes. We derive the generating function of first two types of runs. Under the conditions, (1) strong tendency towards success and (2) strong tendency towards reversing the state, we establish the convergence of waiting times of ther-th occurrence of runs and scans to Poisson type distributions. We establish the central limit theorem and law of the iterated logarithm for the number of runs and scans up to timen.     

A law of the iterated logarithm for the product limit estimator with doubly censored data

We establish the law of the iterated logarithm for the product limit estimator, when the data are subject to double censoring. This investigation extends the results available for the model for singly censored data.     

The Compact Law of the Iterated Logarithm for Multivariate Stochastic Approximation Algorithms

Abstract

We consider a sequence (Z _n ) _n≥1 defined by a general multivariate stochastic approximation algorithm and assume that (Z _n ) converges to a solution z* almost surely. We establish the compact law of the iterated logarithm for Z _n by proving that, with probability one, the limit set of the sequence (Z _n  ? z*) suitably normalized is an ellipsoid. We also give the law of the iterated logarithm for the l ^p norms, p ∈ [1, ∞], of (Z _n  ? z*).     

随机向量自正则和的重对数律

本文给出了独立随机向量序列自正则和的重对数律成立的一个充分条件.     

A tail empirical process approach to some nonstandard laws of the iterated logarithm

We show that a number of nonstandard laws of the iterated logarithm have limiting constants which may be expressed as the extreme values of functionals off, wheref varies over suitable compact sets of functions. By solving the corresponding extremal problems, we show how these constants are generated. In addition, several new laws are presented.     

On the strong law of large numbers for functionals of countable nonhomogeneous Markov chains

We prove a strong law of large numbers for functionals of nonhomogeneous Markov chains. The approach is analytic and different from the usual one.     

A metric discrepancy result for the sequence of powers of minus two

The law of the iterated logarithm for discrepancies of _{(−2)kt}k${\left\{{(-2)}^{k}t\right\}}_k$ is proved. This result completes the concrete determination of the law of the iterated logarithm for discrepancies of the geometric progression with integer ratio, and reveals the fact that 2 is the only positive integer θ>1$\theta >1$ such that fractional parts of _{(−θ)kt}k${\left\{{(-\theta )}^{k}t\right\}}_k$ converge to uniform distribution faster than those of _{θkt}k${\left\{{\theta }^{k}t\right\}}_k$ a.e. t$t$.     

On convergence of the Gauss-Newton method for convex composite optimization

The local quadratic convergence of the Gauss-Newton method for convex composite optimization f=h∘F is established for any convex function h with the minima set C, extending Burke and Ferris’ results in the case when C is a set of weak sharp minima for h. Received: July 24, 1998 / Accepted: November 29, 2000?Published online September 3, 2001     

On the convergence of the DFP algorithm for unconstrained optimization when there are only two variables

Let the DFP algorithm for unconstrained optimization be applied to an objective function that has continuous second derivatives and bounded level sets, where each line search finds the first local minimum. It is proved that the calculated gradients are not bounded away from zero if there are only two variables. The new feature of this work is that there is no need for the objective function to be convex. Received: June 16, 1999 / Accepted: December 24, 1999?Published online March 15, 2000     

On the superlinear convergence of the variable metric proximal point algorithm using Broyden and BFGS matrix secant updating

In previous work, the authors provided a foundation for the theory of variable metric proximal point algorithms in Hilbert space. In that work conditions are developed for global, linear, and super–linear convergence. This paper focuses attention on two matrix secant updating strategies for the finite dimensional case. These are the Broyden and BFGS updates. The BFGS update is considered for application in the symmetric case, e.g., convex programming applications, while the Broyden update can be applied to general monotone operators. Subject to the linear convergence of the iterates and a quadratic growth condition on the inverse of the operator at the solution, super–linear convergence of the iterates is established for both updates. These results are applied to show that the Chen–Fukushima variable metric proximal point algorithm is super–linearly convergent when implemented with the BFGS update. Received: September 12, 1996 / Accepted: January 7, 2000?Published online March 15, 2000     

The low of the iterated logarithm for empirical processes under absolute regularity

The compact law of the iterated logarithm for empirical processes whose underlying sequence satisfies a -mixing condition is considered. In particular, we show a compact law of the iterated logarithm for VC subgraph classes of functions, for classes of functions which satisfy the bracketing condition in Doukhanet al. ⁽⁶⁾ and for some classes of smooth functions.Research partially supported by NSF Grant DMS-93-02583.     

Laws of the iterated logarithm for partial sum processes indexed by functions

We establish a bounded and a compact law of the iterated logarithm for partial sum processes indexed by classes of functions. We assume a growth condition on the metric entropy under bracketing. Examples show that our results are sharp. As a corollary we obtain new results for weighted sums of independent identically distributed random variables.     

多指标独立随机变量的重对数律及其应用

对于多指标独立同分布的随机变量序列,在某些更广泛的正则化序列下,本文给出了重对数律成立的充分必要条件.作为应用,本文讨论了正则和极大值函数的矩存在的充分必要条件.