Global approximation to arbitrary cost functions: A Bayesian approach with application to US banking

This paper proposes and estimates a globally flexible functional form for the cost function, which we call Neural Cost Function (NCF). The proposed specification imposes a priori and satisfies globally all the properties that economic theory dictates. The functional form can be estimated easily using Markov Chain Monte Carlo (MCMC) techniques or standard iterative SURE. We use a large panel of U.S. banks to illustrate our approach. The results are consistent with previous knowledge about the sector and in accordance with mathematical production theory.     

多时滞随机神经网络的稳定性

通过构建李雅普偌夫函数的方法和利用半鞅收敛定理对一类随机时滞神经网络的全局指数稳定进行了分析,提出了易于判定随机时滞神经网络几乎必然指数稳定性新的代数判据,推广了[1]中的主要结论.     

Automatic Smoothing Spline Projection Pursuit

Abstract

A highly flexible nonparametric regression model for predicting a response y given covariates {x_k}^d _k=1 is the projection pursuit regression (PPR) model ? = h(x) = β₀ + Σ_jβ_jf_j(α^T _jx) where the f_j , are general smooth functions with mean 0 and norm 1, and Σ^d _k=1α² _kj=1. The standard PPR algorithm of Friedman and Stuetzle (1981) estimates the smooth functions f_j using the supersmoother nonparametric scatterplot smoother. Friedman's algorithm constructs a model with M _max linear combinations, then prunes back to a simpler model of size M ≤ M _max, where M and M _max are specified by the user. This article discusses an alternative algorithm in which the smooth functions are estimated using smoothing splines. The direction coefficients α_j, the amount of smoothing in each direction, and the number of terms M and M _max are determined to optimize a single generalized cross-validation measure.     

Multivariate Jackson-type inequality for a new type neural network approximation

In this paper, we introduce a new type neural networks by superpositions of a sigmoidal function and study its approximation capability. We investigate the multivariate quantitative constructive approximation of real continuous multivariate functions on a cube by such type neural networks. This approximation is derived by establishing multivariate Jackson-type inequalities involving the multivariate modulus of smoothness of the target function. Our networks require no training in the traditional sense.     

An <Emphasis Type="Italic">L</Emphasis><Subscript>\infty</Subscript>/<Emphasis Type="Italic">L</Emphasis><Subscript>1</Subscript>-Constrained Quadratic Optimization Problem with Applications to Neural Networks

Pattern formation in associative neural networks is related to a quadratic optimization problem. Biological considerations imply that the functional is constrained in the L _\infty norm and in the L ₁ norm. We consider such optimization problems. We derive the Euler–Lagrange equations, and construct basic properties of the maximizers. We study in some detail the case where the kernel of the quadratic functional is finite-dimensional. In this case the optimization problem can be fully characterized by the geometry of a certain convex and compact finite-dimensional set.     

Modelling the Temperature Time‐dependent Speed of Mean Reversion in the Context of Weather Derivatives Pricing

In this paper, in the context of an Ornstein–Uhlenbeck temperature process, we use neural networks to examine the time dependence of the speed of the mean reversion parameter α of the process. We estimate non‐parametrically with a neural network a model of the temperature process and then compute the derivative of the network output w.r.t. the network input, in order to obtain a series of daily values for α. To our knowledge, this is the first time that this has been done, and it gives us a much better insight into the temperature dynamics and temperature derivative pricing. Our results indicate strong time dependence in the daily values of α, and no seasonal patterns. This is important, since in all relevant studies performed thus far, α was assumed to be constant. Furthermore, the residuals of the neural network provide a better fit to the normal distribution when compared with the residuals of the classic linear models used in the context of temperature modelling (where α is constant). It follows that by setting the mean reversion parameter to be a function of time we improve the accuracy of the pricing of the temperature derivatives. Finally, we provide the pricing equations for temperature futures, when α is time dependent.     

A mended hybrid learning algorithm for radial basis function neural networks to improve generalization capability

A two-step learning scheme for radial basis function neural networks, which combines the genetic algorithm (GA) with the hybrid learning algorithm (HLA), is proposed in this paper. It is compared with the methods of the GA, the recursive orthogonal least square algorithm (ROLSA) and another two-step learning scheme for RBF neural networks, which combines the K-means clustering with the HLA (K-means + HLA). Our proposed method has the best generalization performance.     

Neural networks and finite-order approximations

Email: Curry{at}Cardiff.ac.uk This paper investigates the approximation properties of standardfeedforward neural networks (NNs) through the application ofmultivanate Thylor-series expansions. The capacity to approximatearbitrary functional forms is central to the NN philosophy,but is usually proved by allowing the number of hidden nodesto increase to infinity. The Thylor-series approach does notdepend on such limiting cases, lie paper shows how the seriesapproximation depends on individual network weights. The roleof the bias term is taken as an example. We are also able tocompare the sigmoid and hyperbolic-tangent activation functions,with particular emphasis on their impact on the bias term. Thepaper concludes by discussing the potential importance of ourresults for NN modelling: of particular importance is the trainingprocess.     

Rotorcraft parameter estimation using radial basis function neural network

Increased emphasis on rotorcraft performance and operational capabilities has resulted in accurate computation of aerodynamic stability and control parameters. System identification is one such tool in which the model structure and parameters such as aerodynamic stability and control derivatives are derived. In the present work, the rotorcraft aerodynamic parameters are computed using radial basis function neural networks (RBFN) in the presence of both state and measurement noise. The effect of presence of outliers in the data is also considered. RBFN is found to give superior results compared to finite difference derivatives for noisy data.     

B_a空间中神经网络和平移网络的逼近

讨论了具一个隐层单元的神经网络在B_a空间中逼近的特征性定理并给出了逼近估计.对于平移网络,建立了Favard型估计.Orlicz空间中的相应结果均作为应用而给出.     

DENSENESS OF RADIAL-BASIS FUNCTIONS IN L^2（R^n） AND ITS APPLICATIONS IN NEURAL NETWORKS

The authors discuss problems of approximation to functions in L2(Rn) and operators fromL2(Rn1) to L2(Rn2) by Radial-Basis Functions. The results obtained solve the problem ofcapability of RBF neural networks, a basic problem in neural networks.     

B_a空间中神经网络和平移网络的逼近

讨论了具一个隐层单元的神经网络在Ba空间中逼近的特征性定理并给出了逼近估计.对于平移网络,建立了Favard型估计.Orlicz空间中的相应结果均作为应用而给出.     

Feedback control of bursting and multistability in chaotic systems

With a modulated CO₂ laser as a standard model of periodically driven multistable systems, we experimentally demonstrate that a small-amplitude optoelectronic feedback perturbation can efficiently transform a bursting chaotic system to a nonchaotic one. Numerical simulations are in excellent agreement with the experimental results. The control has also been equally effective in the case of a driven FitzHugh-Nagumo model of Neuroscience.     

An application of Liapunov''s method for the analysis of neural networks

The stability is studied of a class of nonlinear dynamical systems which possess many nonlinearities and many equilibrium states. As a special case, the analyzed class of systems includes analog neural networks. Sufficient conditions for the nonoscillatory behaviour of these systems, in the form of frequency domain criteria, are presented. The main result is proved relying on a suitable Liapunov function which is subsequently used for the simultaneous computation of regions of attraction for each stable equilibrium.     

广义小波变换及其在人工神经网络中的应用

相应于非线性系统用人工神经网络的逼近问题，本文引入了一种新的小波变换并研究了其性质.作为推论，本文给出了在Lp范数下单个隐层前馈神经网络逼近定理的构造性证明.     

Robust global synchronization of complex networks with neutral-type delayed nodes

In this paper, the problems of robust global exponential synchronization for a class of complex networks with time-varying delayed couplings are considered. Each node in the network is composed of a class of delayed neural networks described by a nonlinear delay differential equation of neutral-type. Since model errors commonly exist in practical applications, the parameter uncertainties are involved in the considered model. Sufficient conditions that ensure the complex networks to be robustly globally exponentially synchronized are obtained by using the Lyapunov functional method and some properties of Kronecker product. An illustrative example is presented to show the effectiveness of the proposed approach.     

Support vector regression for warranty claim forecasting

Forecasting the number of warranty claims is vitally important for manufacturers/warranty providers in preparing fiscal plans. In existing literature, a number of techniques such as log-linear Poisson models, Kalman filter, time series models, and artificial neural network models have been developed. Nevertheless, one might find two weaknesses existing in these approaches: (1) they do not consider the fact that warranty claims reported in the recent months might be more important in forecasting future warranty claims than those reported in the earlier months, and (2) they are developed based on repair rates (i.e., the total number of claims divided by the total number of products in service), which can cause information loss through such an arithmetic-mean operation.To overcome the above two weaknesses, this paper introduces two different approaches to forecasting warranty claims: the first is a weighted support vector regression (SVR) model and the second is a weighted SVR-based time series model. These two approaches can be applied to two scenarios: when only claim rate data are available and when original claim data are available. Two case studies are conducted to validate the two modelling approaches. On the basis of model evaluation over six months ahead forecasting, the results show that the proposed models exhibit superior performance compared to that of multilayer perceptrons, radial basis function networks and ordinary support vector regression models.