Shape of the fuzzy conclusion generated by linear interpolation of trapezoidal IF...THEN rules

Rok, strany: 1995, 83 - 93

The method of linear rule interpolation is one of the possibilities of approximate reasoning in sparse fuzzy rule bases. (This method uses the Extension Principle and an extended definition of the Minkowski distance for comparable fuzzy terms.) In most practical applications, the shape of the membership functions of the terms in the rules is triangular or trapezoidal, so it is interesting to examine the shape of the generated conclusion in these cases. We find that in general the piecewise linear shape is lost in the conclusion, and this means more computational time for the reasoning. It is also examined if the conclusion maintains normality, as multistep interpolative reasoning requires the preservation of this property. The obtained conditions are interpreted from a semantical point of view.

ISO 690:

Kóczy, L., Kovács, S. 1995. Shape of the fuzzy conclusion generated by linear interpolation of trapezoidal IF...THEN rules. In Tatra Mountains Mathematical Publications, vol. 6, no.2, pp. 83-93. 1210-3195.

APA:

Kóczy, L., Kovács, S. (1995). Shape of the fuzzy conclusion generated by linear interpolation of trapezoidal IF...THEN rules. Tatra Mountains Mathematical Publications, 6(2), 83-93. 1210-3195.