You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
With test case "chukumwong" and "ckwong", python-Levenshtein currently returns 0.5111111111111111 (halflen=3, matches=2) while lingpipe returns 0.8666666666666667 (halflen=4, matches=6).
The text was updated successfully, but these errors were encountered:
According to http://www.census.gov/srd/papers/pdf/rr91-9.pdf and http://en.wikipedia.org/wiki/Jaro–Winkler_distance, the halflen should be the length of the longer string / 2 - 1. So, this line in Levenshtein.c:
halflen = (len1 + 1)/2;
should be:
halflen = (len2 / 2) - 1;
With test case "chukumwong" and "ckwong", python-Levenshtein currently returns 0.5111111111111111 (halflen=3, matches=2) while lingpipe returns 0.8666666666666667 (halflen=4, matches=6).
The text was updated successfully, but these errors were encountered: