It is often interesting to refer to crisp subsets roughly corresponding
to a given fuzzy subset (to use criteria of decision-making, for example).
The simplest way to perform this approximation is to fix a lower limit, noted , with degrees of membership taken into account. Thus, the crisp subset of associated with A for threshold is built by selecting all the elements of which belong to A with a degree at least equal to .
More precisely, for any value of [0, 1], the -cut (or subset of level ) of a fuzzy subset A of is defined as subset .
The associated characteristic function
is defined by:
|
(30) |