Generalization of the rule interpolation method resulting always in acceptable conclusion

Rok, strany: 2001, 73 - 91

This paper deals with the extension of the MACI (Modified Alpha Cut based Interpolation) method. The original MACI method (see, e.g., [D. Tikk, P. Baranyi: Comprehensive analysis of a new fuzzy rule interpolation method, IEEE Trans. on Fuzzy Systems, 8 (2000), 281–296]), was derived from the original alpha cut based interpolation method, usually termed KH interpolation (see, e.g., [L.T. Kóczy, K. Hirota: Approximate reasoning by linear rule interpolation and general approximation, Internat. J. Approx. Reason. 9 (1993), 197–225], [L. T. Kóczy, K. Hirota: Size reduction by interpolation in fuzzy rule bases, IEEE Trans. on SMC, 27 (1997), 14–25]), to eliminate the abnormality of the KH interpolation methods, however, maintaining its advantageous computational behaviour. The MACI (and the KH) methods work over convex and normal fuzzy (CNF) sets. In certain systems, e.g., in hierarchical reasoning schemes, these properties cannot always be ensured. Hence, in this paper we propose algorithm for the modification of MACI method to extend its applicability for such systems where nonconvex and subnormal fuzzy sets can occur. The skeleton of the algorithms can be used also for other fuzzy reasoning techniques.

ISO 690:

Tikk, D., Baranyi, P., Gedeon, T., Muresan, L. 2001. Generalization of the rule interpolation method resulting always in acceptable conclusion. In Tatra Mountains Mathematical Publications, vol. 21, no.1, pp. 73-91. 1210-3195.

APA:

Tikk, D., Baranyi, P., Gedeon, T., Muresan, L. (2001). Generalization of the rule interpolation method resulting always in acceptable conclusion. Tatra Mountains Mathematical Publications, 21(1), 73-91. 1210-3195.