Inference Methods for Partially Redundant Rule Bases
Ralf Mikut, Jens Jäkel, Lutz Gröll · 2000
In this paper, a new inference strategy applicable to redundant or contradictory fuzzy rules is introduced. Both characteristics result mainly from a data-based generation of fuzzy systems where linguistic hedges are used to get an abstract description and where different rules’ premises are overlapping. It is shown, that common fuzzy operators fail in these cases and that the newly introduced switching fuzzy operators solve these problems. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.