Jan 7, 2008 - In particular a kind of Neural Network named Radial Basis Function Network has .... In Carozza [18] a new algorithm for function approximation.
Park and Sandberg [36], [37] proved that RBF neural networks with one layer of radial basis functions are capable of universal approximation. Under certain mild ...
Oct 12, 2014 - diagnosis, the complete training data describing the input- output relationship .... y, we say that decision vector x weakly dominates (or simply dominates) the .... proaches for the problem of inductive supervised learning within the
The radial-basis-function network is trained by simulated frequency ..... Neural networks are known as offering the ability of skillfully approximating highly ...
POLO II â Universidade de Coimbra, Pinhal de Marrocos, 3030 Coimbra. Phone: +351 39 790000, Fax: +351 39 701266 email: {cpereira,dourado}@dei.uc.pt. 2.
However, current detection techniques, such as Pap smear and colposcopy, fail ... and diagnostic (colposcopy) programs are currently in place, approximately ...
The fundamental operation in most of the neural network models existing in the ... anthropoid robot with large initial conditions and unknown plant dynamics.
J. C. QUADRADO ... E33-335 kW, in figure 2 which presents the output .... 10. 15. 20. 25. Wind speed [m/s]. Outpu t pow er. [kW. ] Proceedings of the 5th WSEAS ...
high performance flight vehicle such as F-15 military aircraft. The baseline dynamic inversion controller is augmented with a Self-Organizing Radial Basis ...
May 16, 2018 - Hannah Jessie Rani R.â¯, Aruldoss Albert Victoire T.â¯*. Department of ... Traditional training algorithms in general suffer and trap in local.
Dec 22, 2016 - networks and significant amino acid pairs ..... 6 Sequence logo for 22 GTP binding proteins in transport proteins (generated from WebLogo).
May 7, 2016 - of the radial basis function collocation method and the proposed method ..... the delay differential equations with nonlinear delay function, d3 y.
Robert J. Howlett and Lakhmi C. Jain, Radial Basis Function Networks 2: New Advan- ces in Design, Physica-Verlag: Heidelberg, 2001. 6. Ahalt, S. C. and ...
Jul 19, 2013 - Sensor Routing Topology Control in Underground Mine Rescue. Operation Based on ... relationship; such that when real accident happens, the.
... and Biophysics, Johnson Research Foundation, University of Pennsylvania, School .... excitation wavelengths: 337, 380 and 460 nm (Ramanujam et al., 1996). ...... Cothren, R. M., Richards-Kortum, R. R., Rava, R. P., Boyce, G. A., Doxtader, ...
Oct 8, 2013 - Furthermore, HFC-PSSA is exploited here to optimize the proposed ...... [30] J. J. Hu, E. D. Goodman, K. S. Seo, and M. Pei, âAdaptive hierar-.
of researchers 6], 11] and they have been proved to be capable of universal function approximation 14]. RBF networks have been applied to sev- eral real-world ...
Keywords:ANN, biodiesel, radial basis function, coefficient of determination, MAPE. ... many areas like function-approximation and predictions. [2]. In this work, a ...
be useful in related contexts, namely radial basis function network models, and ..... a new algorithm for function approximation from noisy data was presented.
pixel-based color features are used to develop object/non-object RBF classifiers. ... of its approximation ability as well as the construction of its archi- tecture.
Abstract. The quality of Radial Basis Functions (RBF) and other nonlinear ... UBF and apply them to a number of classification and function approximation tasks.
The network is trained by means of complexity regularization involving empirical risk minimization. Bounds on the expected risk in terms of the sample size are.
19 Jul 2013 - [6] J. H. Saleh and A. M. Cummings, âSafety in the mining industry and the .... [31] S. Zarifzadeh, A. Nayyeri, and N. Yazdani, âEfficient construc-.
(RBF centers, scaling parameters, and output layer weights) simultaneously. This, we call three-phase learning in RBF networks. A practical advantage of two- ...
This article describes a new structure to create a RBF neural network; this new structure has 4 main characteristics: ... function approximation. Other important ...
A New Radial Basis Function Networks Structure: Application to time series prediction. I.Rojas, H.Pomares, J.Gonzalez, E.Ros, M.Salmeron, J.Ortega,A.Prieto Department of Architecture and Computer Technology. University of Granada. Spain.
Abstract This article describes a new structure to create a RBF neural network; this new structure has 4 main characteristics: firstly, the special RBF network architecture uses regression weights to replace the constant weights normally used. These regression weights are assumed to be functions of input variables. The second characteristic is the normalization of the activation of the hidden neurons (weighted average) before aggregating the activations, which, as observed by various authors, produces better results than the classical weighted sum architecture. The third aspect is that a new type of nonlinear function is proposed: the pseudo-gaussian function (PGBF). With this, the neural system gains flexibility, as the neurons possess an activation field that does not necessarily have to be symmetric with respect to the centre or to the location of the neuron in the input space. In addition to this new structure, we propose, as the fourth and final feature, a sequential learning algorithm, which is able to adapt the structure of the network; with this, it is possible to create new hidden units and also to detect and remove inactive units.
I. INTRODUCTION The output of the classical RBF neural networks is defined as the linear combination of the radial basis function layer, as follows:
where the radial basis functions ai are the nonlinear functions, usually gaussian function [4]. The structure of the neural system proposed is modified using a pseudo-gaussian function (PG) in which two scaling parameters U are introduced, which eliminates the symmetry restriction and provides the neurons in the hidden layer with greater flexibility with respect to function approximation. Other important characteristics of the proposed neural system are that the activation of the hidden neurons is normalized and that instead of using a single parameter for the output weights, these are functions of the input variables which leads to a significant reduction in the number of hidden units compared with the classical RBF network. The -* output FRBFof the neural network is:
j?* ( x , ) =
‘=Iv
;withw; = y . b
Being the output of a hidden neuron is computed as:
