A hybrid architecture that includes Radial Basis Functions (RBF) and projection based hidden units is introduced together with a simple gradient based training algorithm. Classification and regression results are demonstrated on various data sets and compared with several variants of RBF networks. In particular, best classification results are achieved on the vowel classification data [1].

Radial basis function (RBF) methods have shown the potential to be a universal grid free method for the numerical solution of partial differential equations. Both global and compactly supported basis functions may be used in the methods to achieve a higher order of accuracy. In this paper, we take advantage of the grid free property of the methods and use an adaptive algorithm to choose the loc...

In this paper a neural network for approximating function is described. The activation functions of the hidden nodes are the Radial Basis Functions (RBF) whose parameters are learnt by a two-stage gradient descent strategy. A new growing radial basis functions-node insertion strategy with different radial basis functions is used in order to improve the net performances. The learning strategy is...

In this paper a neural network for solving partial differential equations is described. The activation functions of the hidden nodes are the radial basis functions (RBF) whose parameters are learnt by a two-stage gradient descent strategy. A new growing RBF-node insertion strategy with different RBF is used in order to improve the net performances. The learning strategy is able to save computat...

Recently, error estimates have been made available for divergencefree radial basis function (RBF) interpolants. However, these results are only valid for functions within the associated reproducing kernel Hilbert space (RKHS) of the matrix-valued RBF. Functions within the associated RKHS, also known as the “native space” of the RBF, can be characterized as vector fields having a specific smooth...

We propose a method for optimizing the complexity of Radial basis function (RBF) networks. The method involves two procedures: adaptation (training) and selection. The first procedure adaptively changes the locations and the width of the basis functions and trains the linear weights. The selection procedure performs the elimination of the redundant basis functions using an objective function ba...

we compare performance of adaptive schemes which are based on radial-basis functions and kalman filters for fast extraction of auditory evoked potentials. moreover, we propose a new method based on evoked potential modeling in the kalman filter framework, which can improve the accuracy compared to the existing methods. simulation results show that adaptive schemes and the kalman method are not ...

The local radial basis function (RBF) interpolation method enables very large-scale data sets to be handled efficiently, overcoming the drawbacks of global interpolation which produces highly ill-conditioned linear systems. Whereas there have been intensive studies on the accuracy of global RBF interpolation, the error analysis of local RBF interpolation is much less investigated. In this regar...

Function approximation has been found in many applications. The radial basis function (RBF) network is one approach which has shown a great promise in this sort of problems because of its faster learning capacity. A traditional RBF network takes Gaussian functions as its basis functions and adopts the least-squares criterion as the objective function, However, it still suffers from two major pr...

one of the most important problems in geodesy is the unification of height datum. generally in geodesy; there are two types of height systems, the geometrical height based on ellipsoid and the physical height based on gravity-defined surface (zhang et al, 2009).local height datum is determined according to mean sea level (msl). in regarding to mismatch of mean sea level and geoid, on the one ha...