Minimax estimation of the mean of spherically symmetric distributions under general quadratic loss

For X one observation on a p-dimensional (p ≥ 4) spherically symmetric (s.s.) distribution about θ, minimax estimators whose risks dominate the risk of X (the best invariant procedure) are found with respect to general quadratic loss, L(δ, θ) = (δ − θ)′ D(δ − θ) where D is a known p × p positive definite matrix. For C a p × p known positive definite matrix, conditions are given under which estimators of the form δ_a,r,C,D(X) = (I − (ar(|X|²)) D^−1/2CD^1/2 |X|⁻²)X are minimax with smaller risk than X. For the problem of estimating the mean when n observations X₁, X₂, …, X_n are taken on a p-dimensional s.s. distribution about θ, any spherically symmetric translation invariant estimator, δ(X₁, X₂, …, X_n), with have a s.s. distribution about θ. Among the estimators which have these properties are best invariant estimators, sample means and maximum likelihood estimators. Moreover, under certain conditions, improved robust estimators can be found.     

Estimation of a Mean Vector in a Two-Sample Problem

We consider the problem of estimating a p-dimensional vector μ₁ based on independent variables X₁, X₂, and U, where X₁ is N_p(μ₁, σ²Σ₁), X₂ is N_p(μ₂, σ²Σ₂), and U is σ²χ²_n (Σ₁ and Σ₂ are known). A family of minimax estimators is proposed. Some of these estimators can be obtained via Bayesian arguments as well. Comparisons between our results and the one of Ghosh and Sinha (1988, J. Multivariate Anal.27 206-207) are presented.     

Maximum Likelihood Estimation of Means and Variances from Normal Populations Under Simultaneous Order Restrictions

For k normal populations with unknown means μ_i and unknown variances σ²_i, i = 1, ..., k, assume that there are some order restrictions among the means and variances, respectively, for example, simple order restrictions: μ₁ ≤ μ₂ ≤ ... ≤ μ_k and σ²₁ ≥ σ²₂ ≥ ... ≥ σ²_k > 0. Some properties of maximum likelihood estimation of μ_is and σ²_i are discussed and an algorithm of obtaining the maximum likelihood estimators under the order restrictions is proposed.     

An identity for the Wishart distribution with applications

Let S_p×p have a Wishart distribution with unknown matrix Σ and k degrees of freedom. For a matrix T(S) and a scalar h(S), an identity is obtained for E_∑tr[h(S)T∑⁻¹]. Two applications are given. The first provides product moments and related formulae for the Wishart distribution. Higher moments involving S can be generated recursively. The second application concerns good estimators of ∑ and ∑⁻¹. In particular, identities for several risk functions are obtained, and estimators of ∑ (∑⁻¹) are described which dominate aS(bS⁻¹), a ≤ 1/k (b ≤ k − p − 1). [3] Ann. Statist. 7 No. 5; (1980) Ann. Statist. 8 used special cases of the identity to find unbiased risk estimators. These are unobtainable in closed form for certain natural loss functions. In this paper, we treat these case as well. The dominance results provide a unified theory for the estimation of ∑ and ∑⁻¹.     

Borel ideals vs. Borel sets of countable relations and trees

For every μ < ω₁, let I_μ be the ideal of all sets S ω^μ whose order type is <ω^μ. If μ = 1, then I₁ is simply the ideal of all finite subsets of ω, which is known to be Σ⁰₂-complete. We show that for every μ < ω₁, I_μ is Σ⁰_2μ-complete. As corollaries to this theorem, we prove that the set WO_ωμ of well orderings Rω × ω of order type <ω^μ is Σ⁰_2μ-complete, the set LP_μ of linear orderings R ω × ω that have a μ-limit point is Σ⁰_2μ+1-complete. Similarly, we determine the exact complexity of the set LT_μ of trees T ^<ωω of Luzin height <μ, the set WR_μ of well-founded partial orderings of height <μ, the set LR_μ of partial orderings of Luzin height <μ, the set WF_μ of well-founded trees T ^<ωω of height <μ(the latter is an old theorem of Luzin). The proofs use the notions of Wadge reducibility and Wadge games. We also present a short proof to a theorem of Luzin and Garland about the relation between the height of ‘the shortest tree’ representing a Borel set and the complexity of the set.     

Rook-by-rook rook theory: Bijective proofs of rook and hit equivalences

Suppose μ and ν are integer partitions of n, and N>n. It is well known that the Ferrers boards associated to μ and ν are rook-equivalent iff the multisets [μ_i+i:1iN] and [ν_i+i:1iN] are equal. We use the Garsia–Milne involution principle to produce a bijective proof of this theorem in which non-attacking rook placements for μ are explicitly matched with corresponding placements for ν. One byproduct is a direct combinatorial proof that the matrix of Stirling numbers of the first kind is the inverse of the matrix of Stirling numbers of the second kind. We also prove q-analogues and p,q-analogues of these results. We also use the Garsia–Milne involution principle to show that for any two rook boards B and B^′, if B and B^′ are bijectively rook-equivalent, then B and B^′ are bijectively hit-equivalent.     

Summability of orthogonal expansions of several variables

Summability of spherical h-harmonic expansions with respect to the weight function ∏_j=1^d |x_j|^2κ_j (κ_j0) on the unit sphere S^d−1 is studied. The main result characterizes the critical index of summability of the Cesàro (C,δ) means of the h-harmonic expansion; it is proved that the (C,δ) means of any continuous function converge uniformly in the norm of C(S^d−1) if and only if δ>(d−2)/2+∑_j=1^d κ_j−min_1jd κ_j. Moreover, it is shown that for each point not on the great circles defined by the intersection of the coordinate planes and S^d−1, the (C,δ) means of the h-harmonic expansion of a continuous function f converges pointwisely to f if δ>(d−2)/2. Similar results are established for the orthogonal expansions with respect to the weight functions ∏_j=1^d |x_j|^2κ_j(1−|x|²)^μ−1/2 on the unit ball B^d and ∏_j=1^d x_j^κ_j−1/2(1−|x|₁)^μ−1/2 on the simplex T^d. As a related result, the Cesàro summability of the generalized Gegenbauer expansions associated to the weight function |t|^2μ(1−t²)^λ−1/2 on [−1,1] is studied, which is of interest in itself.     

A one-parameter family of unimodal, concave, polynomial maps of the interval exhibiting multiplicities

This paper shows there exists a polynomial map, p, of the interval [0, 1] onto itself that is concave, symmetric about the point and such that, when parameterized {μp}, 0 ≤ μ ≤ 1, there exist three distinct values of the parameter μ₀ < μ₁ < μ₂ such that RLR³C = K(μ₀p) ≠ K(μ₁p) ≠ K(μ₂p) = RLR³C. There is also given an explicit construction of a C¹ family with the same properties.     

Continuity in G

For a discrete group G, we consider βG, the Stone– ech compactification of G, as a right topological semigroup, and G^*=βGG as a subsemigroup of βG. We study the mappings λ_p^* :G^*→G^*and μ^* :G^*→G^*, the restrictions to G^* of the mappings λ_p :βG→βG and μ :βG→βG, defined by the rules λ_p(q)=pq, μ(q)=qq. Under some assumptions, we prove that the continuity of λ_p^* or μ^* at some point of G^* implies the existence of a P-point in ω^*.     

Best approximation in L(μ, X), II

The object of this paper is to prove the following theorem: If Y is a closed subspace of the Banach space X, then L¹(μ, Y) is proximinal in L¹(μ, X) if and only if L^p(μ, Y) is proximinal in L^p(μ, Y) for every p, 1 < p < ∞. As an application of this result we prove that if Y is either reflexive or Y is a separable proximinal dual space, then L¹(μ, Y) is proximinal in L¹(μ, X).     

The Pointwise Densities of the Cantor Measure

Let be the classical middle-third Cantor set and let μ be the Cantor measure. Set s = log 2/log 3. We will determine by an explicit formula for every point x the upper and lower s-densities Θ*^s(μ , x), Θ*^s(μ , x) of the Cantor measure at the point x, in terms of the 3-adic expansion of x. We show that there exists a countable set F such that 9(Θ*^s(μ , x))^{− 1/s} + (Θ*^s(μ , x))^{− 1/s} = 16 holds for x \F. Furthermore, for μ_C almost all x, Θ*^s(μ , X) − 2 · 4^{− s} and Θ*^s(μ , x) = 4^{− s}. As an application, we will show that the s-dimensional packing measure of the middle-third Cantor set is 4^s.     

Existence of maximal solutions

Let Ω be a plane bounded region. Let U = {U_μ(P):μ(P)εL∞(Ω), u_μ ε H²_{2, 0}(Ω) and a(P, μ(P))u_μ,xx + 2b(P, μ(P))u_μ,xy + c(P, μ(P))u_μ,vv = ƒ(P) for P ε Ω; here we are given a(P, X), b(P, X), c(P, X) ε L_∞(Ω × E¹), ƒ(P) ε L_p(Ω) with p > 2, and our partial differential equation is uniformly elliptic. The functions μ(P) are called profiles. We establish sufficient conditions—which when they apply are constructive—that there exist a μ₀ ε L_∞(Ω) such that u_μ0 (P) u_μ(P) for all P ε Ω and for each μ ε L_∞(Ω). Similar results are obtained for a difference equation and convergence is proved.     

Inequalities of Rafalson type for algebraic polynomials

For a positive Borel measure dμ, we prove that the constantcan be represented by the zeros of orthogonal polynomials corresponding to dμ in case (i) dν(x)=(A+Bx)dμ(x), where A+Bx is nonnegative on the support of dμ and (ii) dν(x)=(A+Bx²)dμ(x), where dμ is symmetric and A+Bx² is nonnegative on the support of dμ. The extremal polynomials attaining the constant are obtained and some concrete examples are given including Markov-type inequality when dμ is a measure for Jacobi polynomials.     

Commutators for Fourier multipliers on Besov Spaces

If T is any bounded linear operator on Besov spaces B_p^σ_j,q_j(Rⁿ)(j=0,1, and 0<σ₁<σ<σ₀), it is proved that the commutator [T,T_μ]=TT_μ−T_μT is bounded on B_p^σ,q(Rⁿ), if T_μ is a Fourier multiplier such that μ is any (possibly unbounded) symbol with uniformly bounded variation on dyadic coronas.     

On the degree of weak convergence of a sequence of finite measures to the unit measure under convexity

This is a study of the degree of weak convergence under convexity of a sequence of finite measures μ_j on ^k, k 1, to the unit measure δ_x0. LetQ denote a convex and compact subset of ^k, let ƒ ε C^m(Q), m 0, satisfy a convexity condition and let μ be a finite measure on Q. Using standard moment methods, upper bounds and best upper bounds are obtained for ¦∝_Qƒdμ − ƒ(x₀)¦. They sometimes lead to sharp inequalities which are attained for particular μ and ƒ. These estimates are better than the corresponding ones found in the literature.     

Compact causal data interpolation

Consider a Hilbert space equipped with a time-structure, i.e., a resolution E of the identity on defined on subsets of some linearly ordered set Λ. For which x and y in is it possible to find a causal (time respecting) compact operator T, so that Tx = y? When T is required to be a Hilbert-Schmidt operator and (Λ, E) is sufficiently regular, this question is answered in terms of the “time-densities” of x and y. The condition is that the integral ∝_gLμ_x({s t})⁻¹ dμ_y(t) should be finite, where μ_x and μ_y are the measures on Λ given by μ_x(Ω) = ¦|E(Ω)x¦|² and μ_y(Ω) = ¦|E(Ω)y¦|². Further a solution is given for the related problem of minimizing the sum of ¦|Tx − y¦|² and the squared Hilbert-Schmidt norm ¦|R¦|₂² of T.     

Completions of -algebras

A μ-algebra is a model of a first-order theory that is an extension of the theory of bounded lattices, that comes with pairs of terms (f,μ_x.f) where μ_x.f is axiomatized as the least prefixed point of f, whose axioms are equations or equational implications.Standard μ-algebras are complete meaning that their lattice reduct is a complete lattice. We prove that any nontrivial quasivariety of μ-algebras contains a μ-algebra that has no embedding into a complete μ-algebra.We then focus on modal μ-algebras, i.e. algebraic models of the propositional modal μ-calculus. We prove that free modal μ-algebras satisfy a condition–reminiscent of Whitman’s condition for free lattices–which allows us to prove that (i) modal operators are adjoints on free modal μ-algebras, (ii) least prefixed points of Σ₁-operations satisfy the constructive relation μ_x.f=_n≥0fⁿ(). These properties imply the following statement: the MacNeille–Dedekind completion of a free modal μ-algebra is a complete modal μ-algebra and moreover the canonical embedding preserves all the operations in the class of the fixed point alternation hierarchy.     

Existence of three solutions for a class of elliptic eigenvalue problems

In this paper, we consider a problem of the type −Δu = λ(f(u) + μg(u)) in Ω, u¦∂Ω = 0, where Ω Rⁿ is an open-bounded set, f, g are continuous real functions on R, and λ, μ ε R. As an application of a new approach to nonlinear eigenvalues problems, we prove that, under suitable hypotheses, if ¦μ¦ is small enough, then there is some λ > 0 such that the above problem has at least three distinct weak solutions in W₀^1,2(Ω).     

Representation of Reproducing Kernels and the Lebesgue Constants on the Ball

For the weight function (1−x²)^μ−1/2 on the unit ball, a closed formula of the reproducing kernel is modified to include the case −1/2<μ<0. The new formula is used to study the orthogonal projection of the weighted L² space onto the space of polynomials of degree at most n, and it is proved that the uniform norm of the projection operator has the growth rate of n^(d−1)/2 for μ<0, which is the smallest possible growth rate among all projections, while the rate for μ0 is n^μ+(d−1)/2.     

Nonnegative Minimum Biased Quadratic Estimation in Mixed Linear Models

The problem of nonnegative quadratic estimation of a parametric function γ(β, σ)=β′Fβ+∑^r_i=1 f_iσ²_i in a general mixed linear model {y, Xβ, V(σ)=∑^r_i=1 σ²_iV_i} is discussed. Necessary and sufficient conditions are given for y′A₀y to be a minimum biased estimator for γ. It is shown how to formulate the problem of finding a nonnegative minimium biased estimator of γ as a conic optimization problem, which can be efficiently solved using convex optimization techniques. Models with two variance components are considered in detail. Some applications to one-way classification mixed models are given. For these models minimum biased estimators with minimum norms for square of expectation β² and for σ²₁ are presented in explicit forms.