Almost sure uniform convergence of a random Gaussian field conditioned on a large linear form to a non random profile
Publication date: Available online 21 January 2019Source: Statistics & Probability LettersAuthor(s): Philippe MounaixAbstractWe investigate the realizations of a random Gaussian field on a finite domain of Rd in the limit where a given linear functional of the field is large. We prove that if its variance is bounded, the field converges uniformly and almost surely to a non random profile depending only on the covariance and the considered linear functional of the field. This is a significant improvement of the weaker L2-convergence in probability previously obtained in the case of conditioning on a large quadratic func...
Source: Statistics and Probability Letters - January 21, 2019 Category: Statistics Source Type: research

Locally optimal designs for mixed binary and continuous responses
Publication date: Available online 18 January 2019Source: Statistics & Probability LettersAuthor(s): Soohyun Kim, Ming-Hung KaoAbstractWe consider locally optimal designs for experiments involving bivariate responses of mixed data types. To efficiently obtain optimal designs, we adapt a recently proposed complete class approach and derive sharp bounds for the number of design points of candidate designs. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 20, 2019 Category: Statistics Source Type: research

Pointwise eigenfunction estimates and mean Lp-norm blowup of a system of semilinear SPDEs with symmetric Lévy generators
Publication date: Available online 19 January 2019Source: Statistics & Probability LettersAuthor(s): Eugenio Guerrero, José Alfredo López-MimbelaAbstractWe investigate semilinear systems of the form dui(t,x)=[Aiui(t,x)+Gi(u3−i(t,x))]dt+κiui(t,x)dWt,x∈D,i=1,2,with Dirichlet boundary conditions, where D⊂Rd is a bounded smooth domain, Ai is the generator of a pure-jump symmetric Lévy procees in Rd, Gi is a convex locally Lipschitz function such that Gi(z)≥z1+βi for z≥0, βi>0 and κi are constants, i=1,2, and {Wt,t≥0} is a one-dimensional standard Brownia...
Source: Statistics and Probability Letters - January 20, 2019 Category: Statistics Source Type: research

A concentration inequality for inhomogeneous Neyman–Scott point processes
Publication date: May 2019Source: Statistics & Probability Letters, Volume 148Author(s): Jean-François Coeurjolly, Patricia Reynaud-BouretAbstractIn this note, we prove some non-asymptotic concentration inequalities for functionals, called innovations, of inhomogeneous Neyman–Scott point processes, a particular class of spatial point process models. Innovation is a functional built from the counting measure minus its integral compensator. The result is then applied to obtain almost sure rate of convergence for such functionals. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 18, 2019 Category: Statistics Source Type: research

Centered Sobolev inequality and exponential convergence in Φ-entropy
Publication date: Available online 15 January 2019Source: Statistics & Probability LettersAuthor(s): Lingyan Cheng, Liming WuAbstractIn this short paper we find that the Sobolev inequality 1p−2∫fpdμ2p−∫f2dμ≤C∫|∇f|2dμ(p≥0) is equivalent to the exponential convergence of the Markov diffusion semigroup (Pt) to the invariant measure μ, in some Φ-entropy. We provide the estimate of the exponential convergence in total variation and a bounded perturbation result under the Sobolev inequality. Finally in the one-dimensional case we get some two-sided estimates of the Sobolev ...
Source: Statistics and Probability Letters - January 16, 2019 Category: Statistics Source Type: research

Feynman–Kac penalizations of rotationally symmetric α-stable processes
Publication date: Available online 15 January 2019Source: Statistics & Probability LettersAuthor(s): Yunke Li, Masayoshi TakedaAbstractIn Yano et al. (2009), limit theorems (so called, Feynman–Kac penalizations) for the one-dimensional recurrent symmetric α-stable process weighted and normalized by negative (killing) Feynman–Kac functionals with small potential are studied. In this paper, we deal with the same problem for negative potentials diverging at infinity. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 16, 2019 Category: Statistics Source Type: research

A bound of the β-mixing coefficient for point processes in terms of their intensity functions
Publication date: Available online 16 January 2019Source: Statistics & Probability LettersAuthor(s): Arnaud PoinasAbstractWe prove a general inequality on β-mixing coefficients of point processes depending uniquely on their nth order intensity functions. We apply this inequality in the case of determinantal point processes and show that the rate of decay of the β-mixing coefficients of a wide class of DPPs is optimal. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 16, 2019 Category: Statistics Source Type: research

Feynman–Kac penalisations of rotationally symmetric α-stable processes
Publication date: Available online 15 January 2019Source: Statistics & Probability LettersAuthor(s): Yunke Li, Masayoshi TakedaAbstractIn Yano et al. (2009), limit theorems (so called, Feynman–Kac penalisations) for the one-dimensional recurrent symmetric α-stable process weighted and normalized by negative (killing) Feynman–Kac functionals with small potential are studied. In this paper, we deal with the same problem for negative potentials diverging at infinity. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 15, 2019 Category: Statistics Source Type: research

Editorial Board
Publication date: April 2019Source: Statistics & Probability Letters, Volume 147Author(s): (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 12, 2019 Category: Statistics Source Type: research

On the properties of a Takács distribution
Publication date: Available online 10 January 2019Source: Statistics & Probability LettersAuthor(s): Lucio Barabesi, Luca PratelliAbstractThe Takács law – which includes the generalized Dickman law as a special case – arises as a limiting distribution of normalized sums. We provide some properties of this law and a feasible method for computing the probability density and distribution functions of the Takács random variable. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 11, 2019 Category: Statistics Source Type: research

Unbiasedness of LRT on the homogeneity of means in MANOVA under general order restrictions
Publication date: Available online 10 January 2019Source: Statistics & Probability LettersAuthor(s): Mustafa Hamdan, Xiaomi HuAbstractIn this paper a likelihood ratio test is developed through matrix projections for the hypothesis on the homogeneity of mean vectors in MANOVA. It is assumed that the mean vectors are under a general multivariate order restriction. The multivariate order relation in the restriction is a generalization of orders in real numbers. Based on the study of the structure of the acceptance region with respect to a mirror image transformation, and the properties of transformed matrix distributions,...
Source: Statistics and Probability Letters - January 11, 2019 Category: Statistics Source Type: research

Hölder’s identity
Publication date: Available online 11 January 2019Source: Statistics & Probability LettersAuthor(s): W.D. Brinda, Jason M. Klusowski, Dana YangAbstractWe clarify that Hölder’s inequality can be stated more generally than is often realized. This is an immediate consequence of an analogous information-theoretic identity which we call Hölder’s identity. We also explain Andrew R. Barron’s original use of the identity. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 11, 2019 Category: Statistics Source Type: research

Density deconvolution from grouped data with additive errors
Publication date: Available online 11 January 2019Source: Statistics & Probability LettersAuthor(s): Cao Xuan Phuong, Le Thi Hong ThuyAbstractWe study the problem of estimating a density based on grouped data with additive errors. We propose a consistent estimator for the density and compute upper bounds on convergence rate of L2-risk in some cases of error density. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 11, 2019 Category: Statistics Source Type: research

On computing maximum likelihood estimates for the negative binomial distribution
Publication date: Available online 8 January 2019Source: Statistics & Probability LettersAuthor(s): Udika Bandara, Ryan Gill, Riten MitraAbstractThis paper proves that starting value for the Newton–Raphson method can be chosen so that the method is guaranteed to find the maximum likelihood estimate of the parameters of a negative binomial distribution when it exists. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 9, 2019 Category: Statistics Source Type: research

The smallest eigenvalues of random kernel matrices: Asymptotic results on the min kernel
Publication date: Available online 4 January 2019Source: Statistics & Probability LettersAuthor(s): Lu-Jing Huang, Yin-Ting Liao, Lo-Bin Chang, Chii-Ruey HwangAbstractThis paper investigates asymptotic properties of the smallest eigenvalue of the random kernel matrix Mn=1nk(Xi,Xj)i,j=1n, where k(x,y)=min{x,y} is the min kernel function and X1,X2,…,Xn are i.i.d. random variables in [0,1]. We prove that under certain conditions, the smallest eigenvalue converges in L1 to zero with the rate of convergence O(n−3). In addition, if the underlying distribution of Xi’s has a bounded density, the distribution...
Source: Statistics and Probability Letters - January 4, 2019 Category: Statistics Source Type: research

On mutual information estimation for mixed-pair random variables
Publication date: Available online 2 January 2019Source: Statistics & Probability LettersAuthor(s): Aleksandr Beknazaryan, Xin Dang, Hailin SangAbstractWe study the mutual information estimation for mixed-pair random variables. One random variable is discrete and the other one is continuous. We develop a kernel method to estimate the mutual information between the two random variables. The estimates enjoy a central limit theorem under some regular conditions on the distributions. The theoretical results are demonstrated by simulation study. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 3, 2019 Category: Statistics Source Type: research

Pricing contingent convertible bonds: An analytical approach based on two-dimensional stochastic processes
Publication date: Available online 2 January 2019Source: Statistics & Probability LettersAuthor(s): Geon Ho Choe, Hyun Jin Jang, Young Hoon NaAbstractIn this paper, we study fast pricing methods for contingent convertible bonds (CoCos). Based on two-dimensional stochastic processes, we propose two pricing models for CoCos; a dynamic capital-ratio model and a dynamic debt-equity model. Under these frameworks, we derive analytic-form formulae for CoCos with fixed and floored conversion prices. Many practical implications for analyzing CoCos are observed through numerical tests by choosing the plausible model parameters o...
Source: Statistics and Probability Letters - January 3, 2019 Category: Statistics Source Type: research

Independence test for large sparse contingency tables based on distance correlation
Publication date: Available online 3 January 2019Source: Statistics & Probability LettersAuthor(s): Qingyang ZhangAbstractWe propose a new approach to testing independence in a sparse contingency table based on distance correlation measure. We derive the explicit formula of the distance correlation between two categorical variables and suggest a simple permutation test for practical implementation. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 3, 2019 Category: Statistics Source Type: research

Editorial Board
Publication date: March 2019Source: Statistics & Probability Letters, Volume 146Author(s): (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - January 1, 2019 Category: Statistics Source Type: research

Extremal properties of the multivariate extended skew-normal distribution, Part B
Publication date: April 2019Source: Statistics & Probability Letters, Volume 147Author(s): B. Beranger, S.A. Padoan, Y. Xu, S.A. SissonAbstractThe skew-normal and related families are flexible and asymmetric parametric models suitable for modelling a diverse range of systems. We show that the multivariate maximum of a high-dimensional extended skew-normal random sample has asymptotically independent components and derive the speed of convergence of the joint tail. To describe the possible dependence among the components of the multivariate maximum, we show that under appropriate conditions an approximate multivariate e...
Source: Statistics and Probability Letters - January 1, 2019 Category: Statistics Source Type: research

Stochastic monotonicity of MLEs of the mean for exponentially distributed lifetimes under hybrid censoring
Publication date: Available online 24 December 2018Source: Statistics & Probability LettersAuthor(s): Thomas van Bentum, Erhard CramerAbstractA generalized Three Monotonicities Lemma is established. Assuming exponentially distributed lifetimes, it is used to prove the stochastic monotonicity of the MLE of the mean under various hybrid censoring schemes where it had been conjectured but remained unproven so far. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - December 25, 2018 Category: Statistics Source Type: research

A concentration inequality for inhomogeneous Neymann–Scott point processes
Publication date: Available online 21 December 2018Source: Statistics & Probability LettersAuthor(s): Jean-François Coeurjolly, Patricia Reynaud-BouretAbstractIn this note, we prove some non-asymptotic concentration inequalities for functionals, called innovations, of inhomogeneous Neymann–Scott point processes, a particular class of spatial point process models. Innovation is a functional built from the counting measure minus its integral compensator. The result is then applied to obtain almost sure rate of convergence for such functionals. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - December 24, 2018 Category: Statistics Source Type: research

Power and level robustness of a test for composite hypotheses under independent non-homogeneous data
Publication date: Available online 23 December 2018Source: Statistics & Probability LettersAuthor(s): Abhik Ghosh, Ayanendranath BasuAbstractWe consider a class of density power divergence based tests for composite hypotheses under non-homogeneous data. This note provides a rigorous derivation of the power and level influence functions to theoretically justify their robustness with applications to fixed-carrier linear regressions. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - December 24, 2018 Category: Statistics Source Type: research

A concentration inequality for inhomogeneous Neymann-Scott point processes
Publication date: Available online 21 December 2018Source: Statistics & Probability LettersAuthor(s): Jean-François Coeurjolly, Patricia Reynaud-BouretAbstractIn this note, we prove some non-asymptotic concentration inequalities for functionals, called innovations, of inhomogeneous Neymann-Scott point processes, a particular class of spatial point process models. Innovation is a functional built from the counting measure minus its integral compensator. The result is then applied to obtain almost sure rate of convergence for such functionals. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - December 22, 2018 Category: Statistics Source Type: research

Nonzero-sum risk-sensitive finite-horizon continuous-time stochastic games
Publication date: Available online 16 December 2018Source: Statistics & Probability LettersAuthor(s): Qingda WeiAbstractThis paper concerns the nonzero-sum games for continuous-time jump processes with unbounded transition rates under the risk-sensitive finite-horizon cost criterion. The state space is a countable set and the costs are allowed to be unbounded in the game model. Under the suitable optimality conditions, by introducing an appropriate topology for the set of all randomized Markov multi-strategies and employing the risk-sensitive finite-horizon optimality equations of the players, we prove the existence of...
Source: Statistics and Probability Letters - December 16, 2018 Category: Statistics Source Type: research

Removal of the points that do not support an E-optimal experimental design
Publication date: Available online 14 December 2018Source: Statistics & Probability LettersAuthor(s): Radoslav Harman, Samuel RosaAbstractWe propose a method for removing design points that cannot support any E-optimal experimental design of a linear regression model with uncorrelated observations. The proposed method can be used to reduce the size of some large E-optimal design problems such that they can be efficiently solved by semidefinite programming. This paper complements the results of Pronzato [Pronzato, L., 2013. A delimitation of the support of optimal designs for Kiefer’s ϕp-class of criteria. Statis...
Source: Statistics and Probability Letters - December 15, 2018 Category: Statistics Source Type: research

Cointegrated linear processes in Bayes Hilbert space
Publication date: Available online 12 December 2018Source: Statistics & Probability LettersAuthor(s): Won-Ki Seo, Brendan K. BeareAbstractA cointegrated linear process in Bayes Hilbert space is isomorphic to a cointegrated linear process in a Hilbert space of centered square-integrable real functions. We illustrate the use of this isomorphism for modeling nonstationary time series of probability densities. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - December 13, 2018 Category: Statistics Source Type: research

Extremal properties of the multivariate extended skew-normal distribution
Publication date: Available online 7 December 2018Source: Statistics & Probability LettersAuthor(s): B. Beranger, S.A. Padoan, Y. Xu, S.A. SissonAbstractThe skew-normal and related families are flexible and asymmetric parametric models suitable for modelling a diverse range of systems. We show that the multivariate maximum of a high-dimensional extended skew-normal random sample has asymptotically independent components and derive the speed of convergence of the joint tail. To describe the possible dependence among the components of the multivariate maximum, we show that under appropriate conditions an approximate mult...
Source: Statistics and Probability Letters - December 8, 2018 Category: Statistics Source Type: research

Extremal properties of the univariate extended skew-normal distribution, Part A
Publication date: Available online 5 December 2018Source: Statistics & Probability LettersAuthor(s): B. Beranger, S.A. Padoan, Y. Xu, S.A. SissonAbstractWe consider the extremal properties of the highly flexible univariate extended skew-normal distribution. We derive the well-known Mills’ inequalities and Mills’ ratio for the extended skew-normal distribution and establish the asymptotic extreme-value distribution for the maximum of samples drawn from this distribution. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - December 6, 2018 Category: Statistics Source Type: research

A note on large deviation probabilities for empirical distribution of branching random walks
Publication date: Available online 1 December 2018Source: Statistics & Probability LettersAuthor(s): Wanlin ShiAbstractWe consider a branching random walk on R started from the origin. Let Zn(⋅) be the counting measure which counts the number of individuals at the nth generation located in a given set. For any interval A⊂R, it is well known that Zn(nA)Zn(R) converges a.s. to ν(A) under some mild conditions, where ν is the standard Gaussian measure. In this note, we study the convergence rate of PZ̄nnσ2A−ν(A)≥Δ,for a small constant Δ∈(0,1−ν(A)). Our work comple...
Source: Statistics and Probability Letters - December 2, 2018 Category: Statistics Source Type: research

Testing multivariate scatter parameter in elliptical model based on forward search method
Publication date: Available online 1 December 2018Source: Statistics & Probability LettersAuthor(s): Chitradipa ChakrabortyAbstractIn this article, we establish a test for multivariate scatter parameter in elliptical model, where the location parameter is known, and the scatter parameter is estimated by the multivariate forward search method. The consistency property of the test, along with its performances for various simulated data in comparison with a classical one, is also studied here. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - December 2, 2018 Category: Statistics Source Type: research

On probability of high extremes of Gaussian fields with a smooth random trend
Publication date: Available online 1 December 2018Source: Statistics & Probability LettersAuthor(s): Goran Popivoda, Siniša StamatovićAbstractLet ξ1≔ξ1(t),t∈Rn be a centered locally ((E,α),Dt)-stationary Gaussian field, ξ2≔ξ2(t),t∈Rn be a centered differentiable in mean-square sense Gaussian field and η≔η(t),t∈Rn particular smooth random field, being independent of ξi, i=1,2. In this paper we discuss the asymptotics of Psupt∈Mk(ξi(t)+η(t))>u,asu→∞, where Mk⊂Rn is a smooth k-dimensional compact manifold, 0
Source: Statistics and Probability Letters - December 2, 2018 Category: Statistics Source Type: research

Model selection using mass-nonlocal prior
Publication date: Available online 1 December 2018Source: Statistics & Probability LettersAuthor(s): Guiling Shi, Chae Young Lim, Tapabrata MaitiAbstractBayesian variable selection is an active research area. The paper proposes a variable selection technique by taking advantages of two distinctive approaches of Bayesian variable selection, namely spike and slab prior and nonlocal prior. Contrary to the local priors, nonlocal priors put zero mass at null values of the parameters. The proposed method uses posterior median as the parameter estimates and asymptotic consistency is established. The Bayesian implementation is...
Source: Statistics and Probability Letters - December 2, 2018 Category: Statistics Source Type: research

On Brownian exit times from perturbed multi-strips
Publication date: Available online 30 November 2018Source: Statistics & Probability LettersAuthor(s): M. Lifshits, A. NazarovAbstractWe consider the tail probabilities for Brownian exit time from a class of perturbed multi-strips in Euclidean plane. Under some assumptions we prove that the long stays in a perturbed multi-strip are more likely than those in a strip of the same width. This effect is related to the existence of the trapped modes in waveguides. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - December 1, 2018 Category: Statistics Source Type: research

Ergodicity conditions for a double mixed Poisson autoregression
Publication date: Available online 29 November 2018Source: Statistics & Probability LettersAuthor(s): Abdelhakim Aknouche, Nacer DemoucheAbstractWe propose a double mixed Poisson autoregression in which the intensity, scaled by a unit mean independent and identically distributed (i.i.d.) mixing process, has different regime specifications according to the state of a finite unobserved i.i.d. chain. Under some contraction in mean conditions, we show that the proposed model is strictly stationary and ergodic with a finite mean. Applications to various count time series models are given. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 29, 2018 Category: Statistics Source Type: research

Confidence regions in step–stress experiments with multiple samples under repeated type-II censoring
Publication date: Available online 22 November 2018Source: Statistics & Probability LettersAuthor(s): S. Bedbur, U. KampsAbstractIn a multi-sample general step–stress model with pre-specified numbers of observations under all stress levels, confidence regions for associated parameters are provided with minimum volume, minimum coverage probabilities of false parameters, or based on divergence measures. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 27, 2018 Category: Statistics Source Type: research

Averaging principle for the heat equation driven by a general stochastic measure
Publication date: Available online 24 November 2018Source: Statistics & Probability LettersAuthor(s): Vadym RadchenkoAbstractWe study the one-dimensional stochastic heat equation in the mild form driven by a general stochastic measure μ, for μ we assume only σ-additivity in probability. The time averaging of the equation is considered, uniform a. s. convergence to the solution of the averaged equation is obtained. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 26, 2018 Category: Statistics Source Type: research

On the action of the Hilbert transform on ℓ1N-valued functions
Publication date: Available online 24 November 2018Source: Statistics & Probability LettersAuthor(s): Adam OsękowskiAbstractLet H be a separable Hilbert space. The periodic Hilbert transform H is bounded as an operator on Lp(T;ℓ1N(H)), 1
Source: Statistics and Probability Letters - November 24, 2018 Category: Statistics Source Type: research

Confidence regions in step-stress experiments with multiple samples under repeated type-II censoring
Publication date: Available online 22 November 2018Source: Statistics & Probability LettersAuthor(s): S. Bedbur, U. KampsAbstractIn a multi-sample general step-stress model with pre-specified numbers of observations under all stress levels, confidence regions for associated parameters are provided with minimum volume, minimum coverage probabilities of false parameters, or based on divergence measures. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 23, 2018 Category: Statistics Source Type: research

A short note on the dependence structure of random vectors
Publication date: Available online 22 November 2018Source: Statistics & Probability LettersAuthor(s): José M. González-Barrios, Eduardo Gutiérrez-Peña, Raúl RuedaAbstractIn this note we define the concept of exhaustive dependence for a random vector X. In the case where the random vector X is not exhaustively dependent (e.d.), we give an easy decomposition of X in e.d. subvectors which are mutually independent. We also give sufficient conditions on bivariate subvectors of X in order for X to be e.d. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 23, 2018 Category: Statistics Source Type: research

Equivalent distortion risk measures on moment spaces
Publication date: Available online 22 November 2018Source: Statistics & Probability LettersAuthor(s): Dries Cornilly, Steven VanduffelAbstractWe show that maximizing distortion risk measures over the set of distributions with given mean is equivalent to maximizing their concave counterpart. In the case of Value-at-Risk and Tail Value-at-Risk the equivalence also holds when adding information on higher moments. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 22, 2018 Category: Statistics Source Type: research

A simple model-free survival conditional feature screening
Publication date: Available online 19 November 2018Source: Statistics & Probability LettersAuthor(s): Xiaolin Chen, Yahui Zhang, Xiaojing Chen, Yi LiuAbstractWe propose a new, simple, model-free conditional screening approach for survival data. Sure screening and consistency in ranking properties are rigorously established, and simulation studies are conducted to examine the finite-sample performance of the proposed method. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 21, 2018 Category: Statistics Source Type: research

A central limit theorem for marginally coupled designs
Publication date: Available online 20 November 2018Source: Statistics & Probability LettersAuthor(s): Sumin Wang, Dongying Wang, Fasheng SunAbstractIn this paper, we derive a central limit theorem for marginally coupled designs that are intended for computer experiments with both qualitative and quantitative factors. This result is useful for establishing confidence intervals for estimators in various statistical applications. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 21, 2018 Category: Statistics Source Type: research

Double robustness without weighting
Publication date: Available online 20 November 2018Source: Statistics & Probability LettersAuthor(s): Myoung-jae Lee, Sanghyeok LeeAbstractIn contrast to the well-known propensity score (PS), the lesser known ‘ prognostic score (PGS)’ balances the potential untreated response. This paper shows that ‘double robustness’ can be achieved by controlling both PS and PGS in various ways in a “method-blind” manner. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 21, 2018 Category: Statistics Source Type: research

Couplings for processes with independent increments
Publication date: Available online 17 November 2018Source: Statistics & Probability LettersAuthor(s): David CriensAbstractStrassen’s theorem states that the usual stochastic order of two processes is equivalent to the existence of a pathwise ordered coupling. In the case of processes with independent increments, we construct explicit pathwise ordered couplings under suitable conditions on the characteristics of the processes. To formalize the couplings we use independent decompositions, a redistribution mechanism for jumps and Itô’s map, which relates general Lévy measures to a reference L&eacu...
Source: Statistics and Probability Letters - November 18, 2018 Category: Statistics Source Type: research

A generalization of Expected Shortfall based capital allocation
Publication date: Available online 16 November 2018Source: Statistics & Probability LettersAuthor(s): Li Xun, Yangzhi Zhou, Yong ZhouAbstractIn this paper, a generalization of Expected Shortfall based capital allocation is explored, which is a class of allocation rules based on Orlicz risk measure with different Young function. We propose the estimating equation estimator of the Orlicz risk measure based capital allocation. The properties of consistency and asymptotic normality of the estimator are derived. Simulations illustrate that the proposed estimator performs well with moderate sample sizes. (Source: Statistics ...
Source: Statistics and Probability Letters - November 17, 2018 Category: Statistics Source Type: research

Sliced Latin hypercube designs with both branching and nested factors
Publication date: Available online 16 November 2018Source: Statistics & Probability LettersAuthor(s): Hao Chen, Jinyu Yang, Dennis K.J. Lin, Min-Qian LiuAbstractOne special kind of sliced Latin hypercube designs (SLHDs) for computer experiments with branching and nested factors is proposed here, where not only the whole design is an SLHD, but all its slices are also SLHDs. In addition, the SLHD in the first layer has a flexible number of slices, and the slice numbers of the SLHDs in the second layer can be flexible (either the same or different). The construction method is easy to implement, and the resulting designs a...
Source: Statistics and Probability Letters - November 17, 2018 Category: Statistics Source Type: research

On moments of integral exponential functionals of additive processes
Publication date: Available online 16 November 2018Source: Statistics & Probability LettersAuthor(s): Paavo Salminen, Lioudmila VostrikovaAbstractFor real-valued additive process (Xt)t≥0 a recursive equation is derived for the entire positive moments of functionals Is,t=∫stexp(−Xu)du,0≤s
Source: Statistics and Probability Letters - November 17, 2018 Category: Statistics Source Type: research

An elementary derivation of the Chinese restaurant process from Sethuraman’s stick-breaking process
Publication date: Available online 16 November 2018Source: Statistics & Probability LettersAuthor(s): Jeffrey W. MillerAbstractThe usual proof that Sethuraman’s stick-breaking process gives rise to the Chinese restaurant process is indirect, relying on abstract properties of the Dirichlet process. This note provides a direct elementary proof, without using any measure theory. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 17, 2018 Category: Statistics Source Type: research

Dimension walks on Sd×R
Publication date: Available online 16 November 2018Source: Statistics & Probability LettersAuthor(s): N.H. Bingham, Tasmin L. SymonsAbstractWe verify that the established one- and two-step recurrences for positive definite functions on spheres Sd extend to the spatio-temporal case of Sd×R. (Source: Statistics and Probability Letters)
Source: Statistics and Probability Letters - November 17, 2018 Category: Statistics Source Type: research