Constraint that relaxes the equivalence threshold for strings (primitive type name
). It uses a similarity measure specified by similarity-measure
, which returns a value between 0.0
and 1.0
, where 0.0
means no similarity, and 1.0 means absolute similarity. You cannot use non-string concepts for this key.
Note that this function is not used by the NLU system and has nothing to do with user inputs. It should not be used, for example, as a way to catch typos.
You can alter the threshold at which the equivalence function returns true
, false
or uncertain
.
fuzzy-string-equality {
true-tolerance (0.8)
uncertain-tolerance (0.55)
similarity-measure (EditToken)
stop-words ("the with and - &")
}
fuzzy-string-equality
for structures takes a property as a value,
unlike the primitive version of fuzzy-string-equality
:
fuzzy-string-equality (viv.address.streetAddress) {
true-tolerance (0.8)
uncertain-tolerance (0.7)
similarity-measure (EditToken)
}
similarity-measure optional | Specifies the algorithm used to calculate the similarity between the two strings |
stop-words optional | Allows you to specify any stop words that should be removed before performing the similarity-measure calculation |
true-tolerance optional | Specifies the threshold below which the equivalence function will return true |
uncertain-tolerance optional | Specifies the threshold of which below uncertain-tolerance but above true-tolerance will return uncertain |
Copyright 2025 Samsung All rights reserved