Quick fuzzy backpropagation algorithm

A modification of the fuzzy backpropagation (FBP) algorithm called QuickFBP algorithm is proposed, where the computation of the net function is significantly quicker. It is proved that the FBP algorithm is of exponential time complexity, while the QuickFBP algorithm is of polynomial time complexity. Convergence conditions of the QuickFBP, resp. the FBP algorithm are defined and proved for: (1) single output neural networks in case of training patterns with different targets; and (2) multiple output neural networks in case of training patterns with equivalued target vector. They support the automation of the weights training process (quasi-unsupervised learning) establishing the target value(s) depending on the network's input values. In these cases the simulation results confirm the convergence of both algorithms. An example with a large-sized neural network illustrates the significantly greater training speed of the QuickFBP rather than the FBP algorithm. The adaptation of an interactive web system to users on the basis of the QuickFBP algorithm is presented. Since the QuickFBP algorithm ensures quasi-unsupervised learning, this implies its broad applicability in areas of adaptive and adaptable interactive systems, data mining, etc. applications.

[1]  Witold Pedrycz,et al.  Neurocomputations in Relational Systems , 1991, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  Witold Pedrycz,et al.  Fuzzy-set based models of neurons and knowledge-based networks , 1993, IEEE Trans. Fuzzy Syst..

[3]  Michael J. A. Berry,et al.  Data mining techniques - for marketing, sales, and customer support , 1997, Wiley computer publishing.

[4]  S. Stoeva,et al.  A weight-learning algorithm for fuzzy production systems with weighting coefficients , 1992 .

[5]  Peter Hoschka,et al.  Computers as assistants: a new generation of support systems , 1996 .

[6]  Ronen Feldman,et al.  Data Mining on the Web: A Promising Challenge? , 1998, Künstliche Intell..

[7]  Werner Kinnebrock,et al.  Accelerating the standard backpropagation method using a genetic approach , 1994, Neurocomputing.

[8]  Ivan Bratko,et al.  Machine Learning and Data Mining; Methods and Applications , 1998 .

[9]  Kenneth Steiglitz,et al.  Combinatorial Optimization: Algorithms and Complexity , 1981 .

[10]  Detlef Nauck,et al.  Foundations Of Neuro-Fuzzy Systems , 1997 .

[11]  Madan M. Gupta,et al.  Fuzzy automata and decision processes , 1977 .

[12]  Stefka Stoeva,et al.  A fuzzy backpropagation algorithm , 2000, Fuzzy Sets Syst..

[13]  Christopher R. Westphal,et al.  Data Mining Solutions: Methods and Tools for Solving Real-World Problems , 1998 .

[14]  Alfred Kobsa,et al.  Personalized Hypermedia Information Provision Through Adaptive and Adaptable System Features: User Modelling, Privacy and Security Issues , 1997, IS&N.

[15]  James L. McClelland,et al.  Parallel distributed processing: explorations in the microstructure of cognition, vol. 1: foundations , 1986 .

[16]  Hans-Jörg Bullinger,et al.  Human-computer interaction : ergonomics and user interfaces , 1999 .

[17]  Wolfgang Pohl,et al.  Combining User and Usage Modelling for User-Adaptivity Systems , 1999, HCI.

[18]  D. Signorini,et al.  Neural networks , 1995, The Lancet.

[19]  E. Reingold,et al.  Combinatorial Algorithms: Theory and Practice , 1977 .

[20]  Reinhard Oppermann,et al.  Adaptive user support: ergonomic design of manually and automatically adaptable software , 1994 .

[21]  Julita Vassileva,et al.  Adaptive Hypertext and Hypermedia , 1998, Springer Netherlands.

[22]  M. Sugeno.  FUZZY MEASURES AND FUZZY INTEGRALS—A SURVEY , 1993 .

[23]  Schloss Birlinghoven,et al.  Adaptability and Adaptivity in Learning Systems , 1997 .