Showing papers in "Statistics & Probability Letters in 2004"

TL;DR: Sub-fractional Brownian motion (sub-fBm) as mentioned in this paper is a Gaussian process that is intermediate between Bm and fractional Bm, but the increments on non-overlapping intervals are more weakly correlated and their covariance decays polynomially at a higher rate.

TL;DR: A wide class of bivariate copulas depending on two univariate functions which generalizes many known families of copulas is studied, which exhibit several properties concerning symmetry, dependence concepts, and concordance ordering.

TL;DR: It is shown that a large family of margin-based loss functions are Fisher consistent for classification, that is, the population minimizer of the loss function leads to the Bayes optimal rule of classification.

TL;DR: In this paper, a new test to the multivariate Behrens-Fisher problem is obtained by modifying Nel and Van der Merwe's (Commist. Theory Methods 15 (1986) 3719) test.

TL;DR: In this article, the authors suggest a new way to make proposals more acceptable using a secondary Markov chain to modify proposed moves at little extra programming cost, at the expense of increased complexity.

TL;DR: In this article, a measure of discrepancy between past-life distributions is proposed, based on Kullback-Leibler discrimination information and of discrimination information introduced by Ebrahimi and Kirmani.

TL;DR: In this paper, the authors consider a sequential test procedure which detects possible changes in the mean of observations satisfying a weak invariance principle and show that the limiting distribution of the delay time for m→∞ is normal under a suitable standardization provided the change appeared sufficiently soon after m.

TL;DR: In this paper, a test for a change in the parameters of a GARCH(p,q) model is proposed based on approximate likelihood scores and does not require the observations to have finite variance.

TL;DR: In this paper, a kernel method is introduced to estimate a spatial conditional regression under mixing spatial processes, which requires no assumption on the regressor and allows the mixing coefficients decreasing to zero slowly.

TL;DR: In this paper, alternative state space representations for ARMA models have been explored, with minimal state order and appealing Kalman filter steady state properties, and expressions for smoother output and connections to classical infinite sample representations.

TL;DR: In this paper, the authors consider three major types of nonparametric regression tests that are based on kernel and local polynomial smoothing techniques and establish asymptotic power comparisons systematically under the fixed and contiguous alternatives, and are also illustrated through nonasymptotic investigations and finite-sample simulation studies.

TL;DR: In this paper, a new generalization of the Borel-Cantelli Lemma is obtained. But this generalization is special case of the Lemma for special cases.

TL;DR: In this paper, a penalized likelihood estimation procedure is developed for heteroscedastic regression, which is able to estimate both the mean and variance functions simultaneously without parametric assumption for either.

TL;DR: It is demonstrated that approximate posterior expectations computed from discretized LGCPs converge to the exact posterior expectations when the cell sizes of the discretization tends to zero.

TL;DR: In this paper, a wavelet procedure is used to establish the asymptotic normalization of the parametric components and the error variance of the model under weaker conditions than those assumed in the previous literature.

TL;DR: The Gini test as mentioned in this paper is a new test for goodness-of-t based on the Gini index which is the sum over all pairs, of the absolute di6erences of the observed spacings.

TL;DR: In this paper, distribution-free confidence intervals are derived to estimate the fixed quantiles of an arbitrary unknown distribution, based on current records of an iid sequence from that distribution.

TL;DR: A unifying framework for Bayesian analysis in discrete nonparametric settings and the general structure of the posterior and predictive distributions and an explicit updating mechanism for the posterior are developed.

TL;DR: In this article, the authors introduce threshold GARCH(1,1) processes to which Box-Cox transformations are applied, which accommodates asymmetries in conditional variances through a threshold.

TL;DR: For the prior distribution uniform over C, the mean squared error risk of the generalized Bayes estimator is less than or equal to that of X for θ ∈ C if and only if C is a cone, and θ is an apex of the cone.

TL;DR: In this article, the authors study maxisets for linear procedures in the framework of the heteroscedastic white noise model and compare the performance of linear and nonlinear estimates by comparing their respective maxi sets.

TL;DR: In this paper, a lower bound of χ 2 (D ) was obtained along with the sufficient and necessary condition for achieving it and connections between χ2(D ) and other supersaturated design criteria are discussed.

TL;DR: In this article, the authors study the almost sure convergence for ρ -mixing random variable sequences and obtain some new results which extend and improve the corresponding results of Jamison et al.

TL;DR: In this paper, the authors introduced minimum secondary aberration (MSA) and maximum secondary estimation capacity (MSEC) criteria for discriminating among rival FFSP designs, which is an improvement and generalization of the related results in (Statist. Sinica 12 (2002) 885).

TL;DR: In this paper, characterizations of the GIG and gamma distributions through the constancy of regressions of Vr on U are considered, which preserves a bivariate probability measure which is a product of the generalized inverse Gaussian (GIG) and gamma distribution.

TL;DR: In this article, the limit distribution for properly normalized products is proved when the underlying distribution is in the domain of attraction of a stable law with exponent 1 and has a finite first moment.

TL;DR: In this paper, three quantilequantile (Q-Q) plots are derived from a characterization for the multivariate normal distribution, which can be easily used for detecting a possible departure from multinormality in high-dimensional data analysis.

TL;DR: The authors proposed a method that combines smoothing spline estimates of different smoothness to form a final improved estimate, which is straightforward to implement, computationally inexpensive, and gives reliable performances in simulations.

TL;DR: In this article, the estimation of time-varying ARMA models subject to Markovian changes in regime is considered and conditions ensuring consistency and asymptotic normality are given.

TL;DR: A random number generator that produces uniform [0,1) variates directly, as 64-bit floating point numbers, without the customary floating of integers is described.