Adaptive Information System for Data Approximation Problems

An adaptive information system is constructed for the approximation of a multidimensional data base. Such a construction problem can be regarded as an approximation of a multivariate real-valued function that is known in a discrete number of points. That set of points forming the multidimensional data base is called a training set TRE for the adaptive information system. The constructed system contains a module of one-conditional fuzzy rules consequent parts of which are artificial feed-forward neural networks. The numerical data contained in TRE are used to construct premise parts of each rule and membership functions of fuzzy sets involved in them. They are specified with the help of a clustering analysis of TRE. For each fuzzy rule the neural network appearing as its consequent part is trained on the corresponding cluster as on its subdomain In that way the cluster analysis can supply a knowledge in designing the approximator.

[1]  C. Micchelli,et al.  Approximation by superposition of sigmoidal and radial basis functions , 1992 .

[2]  James J. Buckley,et al.  Numerical relationships between neural networks, continuous functions, and fuzzy systems , 1993 .

[3]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1992, Math. Control. Signals Syst..

[4]  Zbigniew Michalewicz,et al.  Evolutionary Domain Covering of an Inference System for Function Approximation , 1998, Evolutionary Programming.

[5]  Chuen-Tsai Sun,et al.  Functional equivalence between radial basis function networks and fuzzy inference systems , 1993, IEEE Trans. Neural Networks.

[6]  Ken-ichi Funahashi,et al.  On the approximate realization of continuous mappings by neural networks , 1989, Neural Networks.

[7]  A. E. Eiben,et al.  Evolutionary Programming VII , 1998, Lecture Notes in Computer Science.

[8]  Michio Sugeno,et al.  Fuzzy identification of systems and its applications to modeling and control , 1985, IEEE Transactions on Systems, Man, and Cybernetics.

[9]  Charles A. Micchelli,et al.  Dimension-independent bounds on the degree of approximation by neural networks , 1994, IBM J. Res. Dev..

[10]  David J. Hand,et al.  Discrimination and Classification , 1982 .

[11]  Ah Chung Tsoi,et al.  Universal Approximation Using Feedforward Neural Networks: A Survey of Some Existing Methods, and Some New Results , 1998, Neural Networks.

[12]  Kurt Hornik,et al.  Approximation capabilities of multilayer feedforward networks , 1991, Neural Networks.

[13]  Witold Kosiński,et al.  General mapping approximation problems solving by neural and fuzzy inference networks , 1997 .

[14]  Michael R. Anderberg,et al.  Cluster Analysis for Applications , 1973 .