APPROXIMATION OF FUNCTIONAL-RELATIONSHIPS TO FUZZY OBSERVATIONS

For approximating functional relationships to fuzzy observations there exist several approaches joining some communities: The functional relationship is still kept a fuzzy function even if the observations are crisp. The case of crisp observations often requires specific considerations. In this paper an approximation method using Hausdorff-like distances is proposed naturally containing the crisp observation case and leading to a crisp function for crisp observations. Additional, the proposed method leads to well-known approximation principles from classical theory and can be realized by standard computer software. A simple example completes the considerations.