The multivariate regression model is considered with p regressors. A latent vector with p binary entries serves to identify one of two types of regression coef®cients: those close to 0 and those not. Specializing our general distributional setting to the linear model with Gaussian errors and using natural conjugate prior distributions, we derive the marginal posterior distribution of the binary...

Nonparametric regression techniques such as spline smoothing and local tting depend implicitly on a parametric model. For instance, the cubic smoothing spline estimate of a regression function based on observations ti; Yi is the minimizer of P(Yi (ti))2 + R ( 00)2. Since R ( 00)2 is zero when is a line, the cubic smoothing spline estimate favors the parametric model (t) = 0+ 1t: Here we conside...

In the common polynomial regression of degree m we determine the design which max imizes the minimum of the D e ciency in the model of degree m and the D e ciencies in the models of degree m j m k j k given The resulting designs allow an e cient estimation of the parameters in the chosen regression and have reasonable e ciencies for checking the goodness of t of the assumed model of degree m by...

Abstract. This work presents a patch-based multi-atlas segmentation approach based on non-negative least squares regression. Our approach finds a weighted linear combination of local image patches that best models the target patch, jointly for all considered atlases. The local coe cients are optimised with the constraint of being positive or zero and serve as weights, of the underlying segmenta...

In this paper, a property of Schur polynomial of real coe cients and real Toeplitz matrix is given. Suppose that the vector of real coe cients of a Schur polynomial annihilates a real Toeplitz matrix, then the Toeplitz matrix is in fact a zero matrix.

Convergence of stochastic processes with jumps to di!usion processes is investigated in the case when the limit process has discontinuous coe$cients. An example is given in which the diffusion approximation of a queueing model yields a di!usion process with discontinuous di!usion and drift coe$cients. c © 2002 Elsevier Science B.V. All rights reserved. MSC: 60B10; 60K25

Average interpolating wavelets are used to construct multilevel representations of large data sets. The data is partitioned into disjoint blocks, and each block is transformed separately. Properties of wavelet coe cients guarantees that this partitioning does not introduce arti cial boundary e ects. Coe cients are organized in a data structure that is well suited for real-time visualization wit...

We show how lattice paths and the re ection principle can be used to give easy proofs of unimodality results. In particular, we give a \one-line" combinatorial proof of the unimodality of the binomial coe cients. Other examples include products of binomial coe cients, polynomials related to the Legendre polynomials, and a result connected to a conjecture of Simion.

We quantify the identifying power of special regressors in heteroskedastic binary regressions with median-independent or conditionally symmetric errors. We measure the identifying power using two criteria: the set of regressor values that help point identify coe¢ cients in latent payo¤s as in (Manski 1988); and the Fisher information of coe¢-cients as in (Chamberlain 1986). We …nd for median-in...

We report the design approach of a home-built rotary substrate holder on simple rotation stages for low-pressure vacuum applications. The result designing and fabricating that it can be successfully used to rotate in application without pressure fl uctuations leakage. Then we investigated suitable thin fi lm deposition by optimal multilayer coatings under three variable operating conditions: Sp...