Closed alessandro-russo closed 2 years ago
Hi all,
in the Jaccard measure for strings, the similarity computed from overlap, lengthA and lengthB (see here) is currently:
overlap
lengthA
lengthB
overlap / (lengthA + lengthA - overlap)
I guess lengthA is erroneously used twice; the similarity should be:
overlap / (lengthA + lengthB - overlap)
Thanks!
Thanks for reporting this
Thanks for reporting this. I have already fixed. See https://github.com/dice-group/LIMES/tree/hotfix.
https://github.com/dice-group/LIMES/tree/hotfix
Hi all,
in the Jaccard measure for strings, the similarity computed from
overlap
,lengthA
andlengthB
(see here) is currently:I guess
lengthA
is erroneously used twice; the similarity should be:Thanks!