Abstract
The use of Radial Basis Function Neural Networks (RBFNNs) to solve functional approximation problems has been addressed many times in the literature. When designing an RBFNN to approximate a function, the first step consists of the initialization of the centers of the RBFs. This initialization task is very important because the rest of the steps are based on the positions of the centers. Many clustering techniques have been applied for this purpose achieving good results although they were constrained to the clustering problem. The next step of the design of an RBFNN, which is also very important, is the initialization of the radii for each RBF. There are few heuristics that are used for this problem and none of them use the information provided by the output of the function, but only the centers or the input vectors positions are considered. In this paper, a new algorithm to initialize the centers and the radii of an RBFNN is proposed. This algorithm uses the perspective of activation grades for each neuron, placing the centers according to the output of the target function. The radii are initialized using the center’s positions and their activation grades so the calculation of the radii also uses the information provided by the output of the target function. As the experiments show, the performance of the new algorithm outperforms other algorithms previously used for this problem.
Similar content being viewed by others
References
Baraldi A and Blonda P (1999a). A survey of fuzzy clustering algorithms for pattern recognition – Part I. IEEE Trans Syst Man Cybern 29(6): 786–801
Baraldi A and Blonda P (1999b). A survey of fuzzy clustering algorithms for pattern recognition – Part II. IEEE Trans Syst Man Cybern 29(6): 786–801
Benoudjit N and Verleysen M (2003). On the kernel widths in radial basis function networks. Neural Process Lett 18(2): 139–154
Bezdek JC (1981). Pattern recognition with fuzzy objective function algorithms. Plenum, New York
Bors AG (2001). Introduction of the Radial Basis Function (RBF) networks. OnLine Symp Electron Eng 1: 1–7
Box GEP and Jenkins GM (1976). Time series analysis, forecasting and control. Holden Day, San Francisco CA
Broomhead DS and Lowe D (1988). Multivariate functional interpolation and adaptive networks. Complex Syst 2: 321–355
Chen T and Chen H (1995). Approximation capability to functions of several variables, nonlinear functionals and operators by radial basis function networks. IEEE Trans Neural Networks 6(4): 904–910
Cherkassky V and Lari-Najafi H (1991). Constrained topological mapping for nonparametric regression analysis. Neural Networks 4(1): 27–40
Cherkassky V, Gehring D and Mulier F (1996). Comparison of adaptative methods for function estimation from samples. IEEE Trans Neural Networks 7: 969–984
Duda RO and Hart PE (1973). Pattern classification and scene analysis. Wiley, New York
Elanayar S and Shin YC (1994). Radial basis function neural networks for approximation and estimation of nonlinear stochastic dynamic systems. IEEE Trans Neural Networks 5: 594–603
Friedman J (1981). Projection pursuit regression. J Am Statist Assoc 76: 817–823
Friedman JH (1991). Multivariate adaptive regression splines (with discussion). Ann Stat 19: 1–141
Gersho A (1979). Asymptotically optimal block quantization. IEEE Trans Inform Theory 25(4): 373–380
González J, Rojas I, Pomares H, Ortega J and Prieto A (2002). A new clustering technique for function aproximation. IEEE Trans Neural Networks 13(1): 132–142
González J, Rojas I, Ortega J, Pomares H, Fernández F and Díaz A (2003). Multiobjective evolutionary optimization of the size, shape, and position parameters of radial basis function networks for function approximation. IEEE Trans Neural Networks 14(6): 1478–1495
Guillén A, Rojas I, González J, Pomares H, Herrera L, Valenzuela O and Prieto A (2005a). Improving clustering technique for functional approximation problem using fuzzy logic: ICFA algorithm. Lecture Notes Comp Sci 3512: 272–280
Guillén A, Rojas I, González J, Pomares H, Herrera L J (2005b) RBF centers initialization using fuzzy clustering technique for function approximation problems. Fuzzy Econ Rev 10(2): 27–44
Karayiannis NB (1999). Reformulated radial basis neural networks trained by gradient descent. IEEE Trans Neural Networks 10(3): 657–671
Karayiannis NB and Mi GW (1997). Growing radial basis neural networks: merging supervised and unsupervised learning with network growth techniques. IEEE Trans Neural Networks 8: 1492–1506
Karayiannis NB, Balasubramanian M, Malki HA (2003) Evaluation of cosine radial basis function neural networks on electric power load forecasting. In: Proceedings of the International Joint Conference on Neural Networks, vol 3, pp 2100–2105
Leski JM (2003). Generalized weighted conditional fuzzy clustering. IEEE Trans Fuzzy Syst 11(6): 709–715
Moody J and Darken C (1989). Fast learning in networks of locally-tunned processing units. Neural Comput 1(2): 281–294
Musavi MT, Ahmed W, Chan K, Faris K and Hummels D (1992). On the training of radial basis functions classifiers. Neural Networks 5(4): 595–603
Park J and Sandberg JW (1991). Universal approximation using radial basis functions network. Neural Comput 3: 246–257
Park J and Sandberg I (1993). Approximation and radial basis function networks. Neural Comput 5: 305–316
Patanè G and Russo M (2001). The enhanced-LBG algorithm. Neural Networks 14(9): 1219–1237
Pedrycz W (1998). Conditional fuzzy clustering in the design of radial basis function neural networks. IEEE Trans Neural Networks 9(4): 601–612
Poggio T, Girosi F (1990) Networks for approximation and learning. In: Proceedings of the IEEE, vol 78, pp 1481–1497
Pomares H (2000) Nueva metodología para el dise no automático de sistemas difusos. PhD dissertation, University of Granada, Spain
Randolph-Gips M M, Karayiannis N B (2003) Cosine radial basis function neural networks. In: Proceedings of the International Joint Conference on Neural Networks, vol 3, pp 96–101
Rojas I, Anguita M, Prieto A and Valenzuela O (1998). Analysis of the operators involved in the definition of the implication functions and in the fuzzy inference proccess. Int J Approx Reason 19: 367–389
Runkler TA and Bezdek JC (1999). Alternating cluster estimation: a new tool for clustering and function approximation. IEEE Trans Fuzzy Syst 7(4): 377–393
Schilling RJ, Carroll JJ and Al-Ajlouni AF (2001). Approximation of nonlinear systems with radial basis function neural networks. IEEE Trans Neural Networks 12(1): 1–15
Sigitani T, Iiguni Y and Maeda H (1999). Image interpolation for progressive transmission by using radial basis function networks. IEEE Trans Neural Networks 10(2): 381–389
Uykan Z, Gzelis C, Celebei ME and Koivo HN (2000). Analysis of input–output clustering for determining centers of RBFN. IEEE Trans Neural Networks 11(4): 851–858
Zhang J and Leung Y (2004). Improved possibilistic C–means clustering algorithms. IEEE Trans Fuzzy Syst 12: 209–217
Zhu Q, Cai Y and Liu L (1999). A global learning algorithm for a RBF network. Neural Networks 12: 527–540
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Guillén, A., Rojas, I., González, J. et al. Output value-based initialization for radial basis function neural networks. Neural Process Lett 25, 209–225 (2007). https://doi.org/10.1007/s11063-007-9039-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11063-007-9039-8