A COMPOUND POISSON MODEL FOR LEARNING DISCRETE BAYESIAN NETWORKS

We introduce here the concept of Bayesian networks, in compound Poisson model, which provides a graphical modeling framework that encodes the joint probability distribution for a set of random variables within a directed acyclic graph. We suggest an approach proposal which offers a new mixed implicit estimator. We show that the implicit approach applied in compound Poisson model is very attractive for its ability to understand data and does not require any prior information. A comparative study between learned estimates given by implicit and by standard Bayesian approaches is established. Under some conditions and based on minimal squared error calculations, we show that the mixed implicit estimator is better than the standard Bayesian and the maximum likelihood estimators. We illustrate our approach by considering a simulation study in the context of mobile communication networks.     

Numerical Simulation of Two-Phase Flow through Heterogeneous Porous Media

A mixed finite element method is combined to finite volume schemes on structured and unstructured grids for the approximation of the solution of incompressible flow in heterogeneous porous media. A series of numerical examples demonstrates the effectiveness of the methodology for a coupled system which includes an elliptic equation and a nonlinear degenerate diffusion–convection equation arising in modeling of flow and transport in porous media.     

Screening model of magnetotransport hysteresis observed in bilayer quantum Hall systems

We report on theoretical and experimental investigations of a novel hysteresis effect that has been observed on the magnetoresistance of quantum Hall bilayer systems. Extending to these system a recent approach, based on the Thomas–Fermi–Poisson nonlinear screening theory and a local conductivity model, we are able to explain the hysteresis as being due to screening effects such as the formation of “incompressible strips”, which hinder the electron density in a layer within the quantum Hall regime to reach its equilibrium distribution.     

Fixed Point Theory for 1-Set Weakly Contractive Operators in Banach Spaces

In this work, using an analogue of Sadovskii＇s fixed point result and several important inequalities we investigate and give new existence theorems for the nonlinear operator equation F（x） =μx, （μ≥1） for some weakly sequentially continuous, weakly condensing and weakly 1-set weakly contractive operators with different boundary conditions. Correspondingly, we can obtain some applicable fixed point theorems of Leray-Schauder, Altman and Furi-Pera types in the weak topology setting which generalize and improve the corresponding results of [3,15,16].     

Unwinding of poly(ethylene oxide) chains grafted on silica under a shear flow of cyclohexane, observed by spin-labelling

Poly(ethylene oxide) chains of molecular weight 2000 have been grafted on pyrogenic silica, spin-labelled at their free end and put into contact with cyclohexane. The electron paramagnetic resonance of the nitroxide free radical gives evidence of two specific environments: namely, trains adsorbed on the solid with a restricted motion and loops and tails protruding into solution with a high mobility. The dynamic parameters such as the rotational correlation time and the static ones such as the fraction of each population have been evaluated. Under a flow of solvent the unwinding of the chains has been observed with increasing flow rate in our experimental range. Some mechanisms are proposed to explain this behaviour. Received: 27 April 1999 Accepted in revised form: 3 June 1999     

Phase separation in semicrystalline blends of poly(phenylene sulfide) and poly(ethylene terephthalate). II. Effect of poly(phenylene sulfide) homopolymer solubilization of PPS‐graft‐PET copolymer on morphology and crystallization behavior

The morphology and crystallization behavior of poly(phenylene sulfide) (PPS) and poly(ethylene terephthalate) (PET) blends compatibilized with graft copolymers were investigated. PPS‐blend‐PET compositions were prepared in which the viscosity of the PPS phase was varied to assess the morphological implications. The dispersed‐phase particle size was influenced by the combined effects of the ratio of dispersed‐phase viscosity to continuous‐phase viscosity and reduced interfacial tension due to the addition of PPS‐graft‐PET copolymers to the blends. In the absence of graft copolymer, the finest dispersion of PET in a continuous phase of PPS was achieved when the viscosity ratio between blend components was nearly equal. As expected, PET particle sizes increased as the viscosity ratio diverged from unity. When graft copolymers were added to the blends, fine dispersions of PET were achieved despite large differences in the viscosities of PPS and PET homopolymers. The interfacial activity of the PPS‐graft‐PET copolymer appeared to be related to the molecular weight ratio of the PPS homopolymer to the PPS segment of the graft copolymer (M_H/M_A). With increasing solubilization of the PPS graft copolymer segment by the PPS homopolymer, the particle size of the PET dispersed phase decreased. In crystallization studies, the presence of the PPS phase increased the crystallization temperature of PET. The magnitude of the increase in the PET crystallization temperature coincided with the viscosity ratio and extent of the PPS homopolymer solubilization in the graft copolymer. © 2000 John Wiley & Sons, Inc. J Polym Sci B: Polym Phys 38: 599–610, 2000     

Semi-Parametric Estimation Using Bernstein Polynomial and a Finite Gaussian Mixture Model

The central focus of this paper is upon the alleviation of the boundary problem when the probability density function has a bounded support. Mixtures of beta densities have led to different methods of density estimation for data assumed to have compact support. Among these methods, we mention Bernstein polynomials which leads to an improvement of edge properties for the density function estimator. In this paper, we set forward a shrinkage method using the Bernstein polynomial and a finite Gaussian mixture model to construct a semi-parametric density estimator, which improves the approximation at the edges. Some asymptotic properties of the proposed approach are investigated, such as its probability convergence and its asymptotic normality. In order to evaluate the performance of the proposed estimator, a simulation study and some real data sets were carried out.