Abstract
In this paper, it is shown that four-layer regular fuzzy neural networks can serve as universal approximators for the sendograph-metric-continuous fuzzy-valued functions. The proof is constructive. We propose a principled method to design four-layer regular fuzzy neural neural network to approximate the target functions. In the previous work, a step function is used as the activation function. To improve the approximation accuracy, in the present work, we also consider using a semi-linear sigmoidal function as the activation function. Then it shows how to design the regular fuzzy neural networks (RFNNs) when the activation functions are the semi-linear sigmoidal function and the step function, respectively. After analyze the approximation accuracy of these two classes of RFNNs, it is found that the former has a much better performance than the latter in approximation accuracy. This conclusion also holds when the target functions satisfy other types of continuity. So the results in this paper can also be used to improve the related work. At last, we give a simulation example to validate the theoretical results.
Similar content being viewed by others
References
Aliev RA, Pedrycz W, Guirimov B, Aliev RR, Ilhan U, Babagil M, Mammadli S (2011) Fuzzy neural networks with fuzzy clustering and differential evolution optimization. Inf Sci 181(9):1591–1608
Bafghi A, Safabakhsh R, Sadeghiyan B (2008) Finding the differential characteristics of block ciphers with neural networks. Inf Sci 178:3118–3132
Barron A (1993) Universal approximation bounds for superpositions of a sigmoidal function. IEEE Trans Inf Theory 39(3):930–945
Boulkroune A, M’Saad M, Farza M (2012a) Adaptive fuzzy tracking control for a class of mimo nonaffine uncertain systems. Neurocomputing 93:48–55
Boulkroune A, M’Saad M, Farza M (2012b) Fuzzy approximation-based indirect adaptive controller for multi-input multi-output non-affine systems with unknown control direction. IET Control Theory Appl 6:2619–2629
Buckley JJ, Hayashi Y (1994) Can fuzzy neural nets approximate continuous fuzzy functions. Fuzzy Sets Syst 61:43–51
Buckley JJ, Feuring T, Hayashi Y (2002) Solving fuzzy equations using evolutionary algorithms and neural nets. Soft Comput 6(2):116–123
Cao FL, Lin SB, Xu ZB (2010) Constructive approximate interpolation by neural networks in the metric space. Math Comput Model 52:1674–1681
Chen TP, Chen H, Liu RW (1995) Approximation capability in \({C}(\overline{{R}}^n)\) by multilayer feedforward netwrks and related problems. IEEE Trans Neural Netw 6:25–29
Diamond P, Kloeden P (1994) Metric space of fuzzy sets—theory and application. World Scientific, Singapore
Fukuoka Y, Matsuki H, Minamitani H (1998) A modified backpropagation method to avoid false local minima. Neural Netw 11:1059–1072
Greco G, Moschen M, Rezende E (1998) On the variational convergence of fuzzy sets in metric spaces. Annali dell’Universita di Ferrara 44(1):27–39
Hamzaçebi C (2008) Improving artificial neural networks’ performance in seasonal time series forecasting. Inf Sci 178:4550–4559
Hornik K (1991) Approximation capabilities of multilayer feedforward networks. Neural Netw 4:251–257
Huang GB, Zhu QY, Siew CK (2006) Universal approximation using incremental constructive feedforward networks with random hidden nodes. IEEE Trans Neural Netw 17(4):879–892
Huang H (2010) Some notes on zadeh’s extensions. Inf Sci 180:3806–3813
Huang H, Wu CX (2009a) Approximation capabilities of multilayer fuzzy neural networks on the set of fuzzy-valued functions. Inf Sci 179:2762–2773
Huang H, Wu CX (2009b) Approximation of level continuous fuzzy-valued functions by multilayer regular fuzzy neural networks. Math Comput Model 49:1311–1318
Huang H, Wu CX (2011) Approximation of fuzzy functions by regular fuzzy neural networks. Fuzzy Sets Syst 177:60–79
Kadirkamanathan V, Niranjan M (1993) A function estimation approach to sequential learning with neural networks. Neural Comput 5:954–975
Leshno M, Lin V, Pinkus A, Schocken S (1993) Multilayer feedforward networks with a nonpolynomial activation function can approximate any function. Neural Netw 6:861–867
Li TS, Tong SC, Feng G (2010) A novel robust adaptive-fuzzy-tracking control for a class of nonlinear multi-input/multi-output systems. IEEE Trans Fuzzy Syst 18(1):150–160
Liu PY (2000) Analyses of regular fuzzy neural networks for approximation capabilities. Fuzzy Sets Syst 114:329–338
Liu PY (2001) Universal approximations of continuous fuzzy-valued function by multi-layer regular fuzzy neural networks. Fuzzy Sets Syst 119:313–320
Liu PY, Li HX (2005) Approximation analysis of feedforward regular fuzzy neural network with two hidden layers. Fuzzy Sets Syst 150:373–396
Liu YJ, Tong SC, Li TS (2011) Observer-based adaptive fuzzy tracking control for a class of uncertain nonlinear mimo systems. Fuzzy Sets Syst 164:25–44
Rumelhart DE, Hinton GE, Williams RJ (1986) Learning representations by backpropagation errors. Nature 323:533–536
Scarselli F, Tsoi AC (1998) Universal approximation using feedforward neural networks: a survery of some existing methods, and some new results. Neural Netw 11(1):15–37
Shin D, Na SY, Kim JY, Baek SJ (2008) Fuzzy neural networks for obstacle pattern recognition and collision avoidance of fish robots. Soft Comput 12(7):715–720
Tong SC, Li YM, Zhang HG (2011) Adaptive neural network decentralized backstepping output-feedback control for nonlinear large-scale systems with time delays. IEEE Trans Neural Netw 22(7):1073–1086
Tong SC, Li YM, Shi P (2012) Observer-based adaptive fuzzy backstepping output feedback control of uncertain mimo pure-feedback nonlinear systems. IEEE Trans Fuzzy Syst 20(4):771–785
Wang XZ, Chen AX, Feng HM (2011) Upper integral network with extreme learning mechanism. Neurocomputing 74(16):2520–2525
Wu CX, Ma M (1991) The basic of fuzzy analysis (in Chinese). National Defence Industry press, Beijing
Zhao JW, Park DS, Lee J, Cao FL (2012) Generalized extreme learning machine acting on a metric space. Soft Comput 16:1503–1514
Acknowledgments
This work was supported by the National Natural Science Foundation of China (Grant No. 61103052). The author would like to thanks the anonymous referees for their valuable comments that improve the presentation of this paper.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by V. Piuri.
Rights and permissions
About this article
Cite this article
Huang, H., Wu, C. Approximation of fuzzy-valued functions by regular fuzzy neural networks and the accuracy analysis. Soft Comput 18, 2525–2540 (2014). https://doi.org/10.1007/s00500-014-1232-x
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00500-014-1232-x