A Berry-Esseen bound for symmetric statistics

Summary The rate of convergence of the distribution function of a symmetric function of N independent and identically distributed random variables to its normal limit is investigated. Under appropriate moment conditions the rate is shown to be (N^–1/2). This theorem generalizes many known results for special cases and two examples are given. Possible further extensions are indicated.Research supported by the U.S. Office of Naval Research, Contract N 00014-80-C-0163     

Geometric convergence to e−z by rational functions with real poles

Summary In this paper, we show that there exists a sequence of rational functions of the formR _n(z)=p_n–1(z)/(1+z/n)ⁿ,n=1, 2, ..., with degp _n–1n–1, which converges geometrically toe ^–z in the uniform norm on [0, +), as well as on some infinite sector symmetric about the positive real axis. We also discuss the usefulness of such rational functions in approximating the solutions of heat-conduction type problems.Research supported in part by the Air Force Office of Scientific Research under Grant AFOSR-74-2688, and by the University of South Florida Research Council.Research supported in part by the Air Force Office of Scientific Research under Grant AFOSR-74-2729, and by the Energy Research and Development Administration (ERDA) under Grant E(11-1)-2075.     

Limit theorems for sums of dependent random variables occurring in statistical mechanics

Summary We study the asymptotic behavior of partial sums S _nfor certain triangular arrays of dependent, identically distributed random variables which arise naturally in statistical mechanics. A typical result is that under appropriate assumptions there exist a real number m, a positive real number , and a positive integer k so that (S _n–nm)/n^1–1/2k converges weakly to a random variable with density proportional to exp(–¦s¦ ^2k/(2k)!). We explain the relation of these results to topics in Gaussian quadrature, to the theory of moment spaces, and to critical phenomena in physical systems.Alfred P. Sloan Research Fellow. Research supported in part by a Broadened Faculty Research Grant at the University of Massachusetts and by National Science Foundation Grant MPS 76-06644Research supported in part by National Science Foundation Grants MPS 74-04870 A01 and MCS 77-20683     

Asymptotic expansions for bivariate von Mises functionals

Summary A Berry-Essen result and asymptotic expansions are derived for the distribution of bivariate von Mises functionals under moment and smoothness conditions.The results apply to the Cramér-von Mises ² — statistic as well as to the Central Limit Theorem in Hilbert space, yielding a convergence rate O(n _–1+) for every >0 on centered ellipsoids.Herrn Professor Dr. Leopold Schmetterer zu Ehren seines sechzigsten Geburtstages gewidmet     

Limit theorems for empirical processes

Summary The empirical measure P _n for iid sampling on a distribution P is formed by placing mass n ^–1 at each of the first n observations. Generalizations of the classical Glivenko-Cantelli theorem for empirical measures have been proved by Vapnik and ervonenkis using combinatorial methods. They found simple conditions on a class C to ensure that sup {|P _n (C) – P(C)|: C C} converges in probability to zero. They used a randomization device that reduced the problem to finding exponential bounds on the tails of a hypergeometric distribution. In this paper an alternative randomization is proposed. The role of the hypergeometric distribution is thereby taken over by the binomial distribution, for which the elementary Bernstein inequalities provide exponential boundson the tails. This leads to easier proofs of both the basic results of Vapnik-ervonenkis and the extensions due to Steele. A similar simplification is made in the proof of Dudley's central limit theorem forn ^1/2(P P _n –P)— a result that generalizes Donsker's functional central limit theorem for empirical distribution functions.This research was supported in part by the Air Force Office of Scientific Research, Contract No. F49620-79-C-0164     

Second order random fields over l p with homogeneous and isotropic increments

Summary A random field over l _p is a stochastic process X(t), where t is an element of l _p .It is said to have homogeneous and isotropic increments if E(X(t) – X(s)) ² is a function of t-s. The subject of this work is the spectral theory of such processes. The main results are: a representation of the field as a series of filtered, orthogonal processes with a real time parameter; a representation as a white noise integral over l _p ;limit theorems for the average of X over a sphere; and, finally, filtering of the orthogonal components.In particular, we mention: (1) The averages over spheres of increasing dimension converge in quadratic mean for p=2 but not for 0<p<2. (2) The limiting distribution of a fixed coordinate of a point uniformly distributed over the l _p -unit sphere in n-space, n , has the density [2(1/p+1)p ^1/p ]^–1exp(-¦x¦ ^p /P).This paper represents results obtained at the Courant Institute of Mathematical Sciences, New York University, under the sponsorship of the National Science Foundation Grants NSF-GP-7378 and NSF-GW-2049.     

Effects of design and error on normal convergence rates in regression problems

Summary We describe the way in which design and experimental error interact to determine convergence rates in central limit theorems for regression estimators. For example, we show that if the convergence rate in a central limit theorem for experimental errors alone isn ^–, wheren is sample size and 0<<1/2, then this rate is maintain in a central limit theorem for intercept and slope parameters if and only if the distribution generating design has finite (2+2)'th moment. We prove that in other circumstances a careful choice of design can substantially improve convergence rates by introducing a degree of symmetry not present in the error distribution. Other results on the relationship between design and error are also derived.     

Wolff type estimates and theH p corona problem in strictly pseudoconvex domains

LetD be a strictly pseudoconvex domain inC ⁿ . We prove that , ϕ a (0,1)-form, admits solutions inL ^p (∂D), 1≤p<∞ and in BMO, under certain Wolff type conditions of ϕ. Some such results (for 1<p<∞) have previously been obtained by Amar in the ball, but under slightly stronger hypotheses. As a corollary we obtain aH ^p -corona result for two generators. Partially supported by the Swedish Natural Sciences Research Council.     

On Marcinkiewicz Integral Operators with Rough Kernels

This paper is devoted to the study on the L^p-mapping properties of Marcinkiewicz integral operators with rough kernels along “polynomial curves” on The boundedness of the Marcinkiewicz integrals for some fixed 1 < p < ∞ are obtained under some size conditions, which essentially improve or extend some well-known results.     

Generalized Bayes minimax estimators of the mean of multivariate normal distribution with unknown variance

We construct a broad class of generalized Bayes minimax estimators of the mean of a multivariate normal distribution with covariance equal to σ²I_p, with σ² unknown, and under the invariant loss δ(X)−θ²/σ². Examples that illustrate the theory are given. Most notably it is shown that a hierarchical version of the multivariate Student-t prior yields a Bayes minimax estimate.     

Positive solutions for logistic type quasilinear elliptic equations on R

In this paper, we consider positive solutions of the logistic type p-Laplacian equation −Δ_pu=a(x)|u|^p−2u−b(x)|u|^q−1u, xR^N (N2). We show that under rather general conditions on a(x) and b(x) for large |x|, the behavior of the positive solutions for large |x| can be determined. This is then used to show that there is a unique positive solution. Our results improve the corresponding ones in J. London Math. Soc. (2) 64 (2001) 107–124 and J. Anal. Math., in press.     

High-dimensional asymptotic expansion of LR statistic for testing intraclass correlation structure and its error bound

This paper deals with the null distribution of a likelihood ratio (LR) statistic for testing the intraclass correlation structure. We derive an asymptotic expansion of the null distribution of the LR statistic when the number of variable p and the sample size N approach infinity together, while the ratio p/N is converging on a finite nonzero limit c(0,1). Numerical simulations reveal that our approximation is more accurate than the classical χ²-type and F-type approximations as p increases in value. Furthermore, we derive a computable error bound for its asymptotic expansion.     

On conditions for optimality in least pth approximation withp → ∞

This paper presents a theoretical discussion of the necessary and sufficient conditions for optimality in generalized nonlinear leastpth approximation problems forp . In the limit, the conditions for a minimax approximation are derived, as is to be expected. Numerical examples involving the modeling of a linear time-invariant fourth-order system by a second-order model and the design of quarter-wave transmission-line transformers illustrate the results.This work was supported by the National Research Council of Canada under Grant No. A7239 and by a Frederick Gardner Cottrell Grant from the Research Corporation. This paper was presented at the 9th Annual Allerton Conference on Circuit and System Theory, Urbana, Illinois, October 6–8, 1971. The authors thank Mrs. J. R. Popovi for helping to correct Example 4.1.     

A Central Limit Theorem for Random Fields of Negatively Associated Processes

A central limit theorem for negatively associated random fields is established under the fairly general conditions. We use the finite second moment condition instead of the finite (2+)th moment condition used by Roussas.⁽¹⁵⁾ A similar result is also given for positively associated sequences.     

Deconvolving a density from contaminated dependent observations

The paper studies the performance of deconvoluting kernel density estimators for estimating the marginal density of a linear process. The data stem from the linear process and are partially, respectively fully contaminated by iid errors with a known distribution. If 1–p denotes the proportion of contaminated observations (and it is, of course, unknown which observations are contaminated and which are not) then for 1–p (0, 1) and under mild conditions almost sure deconvolution rates of orderO(n ^–2/5(logn)^9/10) can be achieved for convergence in . This rate compares well with the existing rates foriid uncontaminated observations. Forp=0 and exponentially decreasing error characteristic function the corresponding rates are of merely logarithmic order. As a by-product the paper also gives a rate of convergence result for the empirical characteristic function in the linear process context and utilizes this to demonstrate that deconvoluting kernel density estimators attain the optimal rate in the dependence case with exponentially decreasing error characteristic function.This work was partially supported by a grant from the Deutsche Forschungsgemeinschaft.     

Linking and Multiplicity Results for the p-Laplacian on Unbounded Cylinders

We consider the p-Laplacian problem[formula]on unbounded cylinders Ω = Ω̃ × R^N − m  R^N, N − m ≥ 2, where Δ_pu = div(|u|^p − 2u), λ is a constant in a certain range, and a  L^N/p(Ω) ∩ L^∞(Ω) is nonnegative, a  0. Using the principle of symmetric criticality, existence and multiplicity are proved under suitable conditions on a and f.     

A Bahri–Lions theorem revisited

In 1988, A. Bahri and P.L. Lions [A. Bahri, P.L. Lions, Morse-index of some min–max critical points. I. Application to multiplicity results, Comm. Pure Appl. Math. 41 (1988) 1027–1037] studied the following elliptic problem:

where Ω is a bounded smooth domain of , 2<p<(2N−2)/(N−2) and f(x,u) is not assumed to be odd in u. They proved the existence of infinitely many solutions under an appropriate growth restriction on f. In the present paper, we improve this result by showing that under the same growth assumption on f the problem admits in fact infinitely many sign-changing solutions. In addition we derive an estimate on the number of their nodal domains. We also deal with the corresponding fourth order equation Δ²u=|u|^p−2u+f(x,u) with both Dirichlet and Navier boundary conditions, as well as with strongly coupled elliptic systems.     

Absolutely continuous Hardy–Sobolev functions in the unit disc

Let f(z) be an analytic function defined in the unit disc whose fractional derivative of order belongs to H^p, 0<p1. We show that as a consequence of a monotonicity condition on the decay of the Taylor coefficients, it is possible to improve the usual radial boundary growth estimate for H^p functions by a logarithmic factor. As a consequence we show that under certain regularity conditions imposed on the decay and oscillations of the absolute values of the function's Taylor coefficients, it is possible to estimate the function's modulus of continuity and modulus of absolute continuity and that a consequence of this is that as p→0, these functions will be generally smoother. Examples are also given of Hardy–Sobolev functions having modulus of absolute continuity different than modulus of continuity.     

On the central limit theorem in R p when p→∞

Summary Let X ₁ , X ₂ , ..., X _n be i.i.d. random vectors in R ^p where p tends to infinity. A theorem is presented showing that the Central Limit Theorem should hold if p ²/n tends to zero. Furthermore, an example is presented with X _i having a mixed multivariate normal distribution (with finite moment generating function) for which a uniform normal approximation to the distribution of the sample mean can not hold if p ²/n does not tend to zero. Research supported in part by National Science Foundation Grants MCS 80-02340, MCS 83-01834, and DMS 85-03785     

On Equivalence of Moduli of Smoothness

It is known that iffW^k_p, thenω_m(f, t)_ptω_m−1(f′, t)_p…. Its inverse with any constants independent offis not true in general. Hu and Yu proved that the inverse holds true for splinesSwith equally spaced knots, thusω_m(S, t)_ptω_m−1(S′, t)_pt²ω_m−2(S″, t)_p…. In this paper, we extend their results to splines with any given knot sequence, and further to principal shift-invariant spaces and wavelets under certain conditions. Applications are given at the end of the paper.