Search results for "Computational Mathematic"
showing 10 items of 987 documents
Adaptive Distance-Based Pooling in Convolutional Neural Networks for Audio Event Classification
2020
In the last years, deep convolutional neural networks have become a standard for the development of state-of-the-art audio classification systems, taking the lead over traditional approaches based on feature engineering. While they are capable of achieving human performance under certain scenarios, it has been shown that their accuracy is severely degraded when the systems are tested over noisy or weakly segmented events. Although better generalization could be obtained by increasing the size of the training dataset, e.g. by applying data augmentation techniques, this also leads to longer and more complex training procedures. In this article, we propose a new type of pooling layer aimed at …
Avoiding strange attractors in efficient parametric families of iterative methods for solving nonlinear problems
2019
[EN] Searching zeros of nonlinear functions often employs iterative procedures. In this paper, we construct several families of iterative methods with memory from one without memory, that is, we have increased the order of convergence without adding new functional evaluations. The main aim of this manuscript yields in the advantage that the use of real multidimensional dynamics gives us to decide among the different classes designed and, afterwards, to select its most stable members. Moreover, we have found some elements of the family whose behavior includes strange attractors of different kinds that must be avoided in practice. In this sense, Feigenbaum diagrams have resulted an extremely …
More restrictive Gray codes for some classes of pattern avoiding permutations
2009
In a recent article [W.M.B. Dukes, M.F. Flanagan, T. Mansour, V. Vajnovszki, Combinatorial Gray codes for classes of pattern avoiding permutations, Theoret. Comput. Sci. 396 (2008) 35-49], Dukes, Flanagan, Mansour and Vajnovszki present Gray codes for several families of pattern avoiding permutations. In their Gray codes two consecutive objects differ in at most four or five positions, which is not optimal. In this paper, we present a unified construction in order to refine their results (or to find other Gray codes). In particular, we obtain more restrictive Gray codes for the two Wilf classes of Catalan permutations of length n; two consecutive objects differ in at most two or three posit…
Application of Operator Splitting Methods in Finance
2016
Financial derivatives pricing aims to find the fair value of a financial contract on an underlying asset. Here we consider option pricing in the partial differential equations framework. The contemporary models lead to one-dimensional or multidimensional parabolic problems of the convection-diffusion type and generalizations thereof. An overview of various operator splitting methods is presented for the efficient numerical solution of these problems.
BEM application on an external problem comparison with both theoretical and finite elements results and observations on divergence strip
1992
Abstract By means of a computer program the Boundary Element Method is applied to a central hole in an undefined plate with uniform load along the boundary. Results are compared with those obtained by Kirsch's theoretical solution and a previous analysis by the Finite Element Method. The calculus of percentage error shows the advantage of the Boundary Element Method on the external problem with regard to the Finite Element Method. The error causes near the boundary internal points are analysed with the existence of a strip, where the result is not reliable in evidence.
On the convergence of a finite volume method for the Navier–Stokes–Fourier system
2020
Abstract The goal of the paper is to study the convergence of finite volume approximations of the Navier–Stokes–Fourier system describing the motion of compressible, viscous and heat-conducting fluids. The numerical flux uses upwinding with an additional numerical diffusion of order $\mathcal O(h^{ \varepsilon +1})$, $0<\varepsilon <1$. The approximate solutions are piecewise constant functions with respect to the underlying polygonal mesh. We show that the numerical solutions converge strongly to the classical solution as long as the latter exists. On the other hand, any uniformly bounded sequence of numerical solutions converges unconditionally to the classical solution of t…
An adaptive rectangular mesh administration and refinement technique with application in cancer invasion models
2022
We present an administration technique for the bookkeeping of adaptive mesh refinement on (hyper-)rectangular meshes. Our technique is a unified approach for h-refinement on 1-, 2- and 3D domains, which is easy to use and avoids traversing the connectivity graph of the ancestry of mesh cells. Due to the employed rectangular mesh structure, the identification of the siblings and the neighbouring cells is greatly simplified. The administration technique is particularly designed for smooth meshes, where the smoothness is dynamically used in the matrix operations. It has a small memory footprint that makes it affordable for a wide range of mesh resolutions over a large class of problems. We pre…
Objective function design for robust optimality of linear control under state-constraints and uncertainty
2009
We consider a model for the control of a linear network flow system with unknown but bounded demand and polytopic bounds on controlled flows. We are interested in the problem of finding a suitable objective function that makes robust optimal the policy represented by the so-called linear saturated feedback control. We regard the problem as a suitable differential game with switching cost and study it in the framework of the viscosity solutions theory for Bellman and Isaacs equations. © 2009 EDP Sciences, SMAI.
Comparison of Internal Clustering Validation Indices for Prototype-Based Clustering
2017
Clustering is an unsupervised machine learning and pattern recognition method. In general, in addition to revealing hidden groups of similar observations and clusters, their number needs to be determined. Internal clustering validation indices estimate this number without any external information. The purpose of this article is to evaluate, empirically, characteristics of a representative set of internal clustering validation indices with many datasets. The prototype-based clustering framework includes multiple, classical and robust, statistical estimates of cluster location so that the overall setting of the paper is novel. General observations on the quality of validation indices and on t…
Quantum chemical modelling of polarons and perovskite solid solutions
2001
Abstract Following our previous study [J. Phys.: Condens. Matter 10 (1998) 6271] of a single Nb impurity and Nb clusters in KTaO 3 , we present results of the calculations for a series of perovskite KNb x Ta 1− x O 3 (KTN) solid solutions ( x =0, 0.125, 0.25, 0.75, 1). The quantum chemical method of the intermediate neglect of the differential overlap (INDO) combined with the large unit cell (LUC) periodic model is used. According to the INDO calculations, Nb impurity becomes off-center in KTaO 3 already at the lowest studied concentrations ( x =0.125), in a good agreement with XAFS measurements. We compare our results with previous ab initio FP-LMTO calculations. Quantum chemical calculati…