1
Vote

Ratcliff/Obershelp and Intersect

description

Hi!

not sure if this project is being maintaned, but thought I'll report a bug and just say it's amazing library!

So,
I noticed that Ratcliff/Obershelp similarity function doesn't return 100% match when comparing "customs" vs "customs", so I looked inside the function and found that it is using Intersect which doesn't count duplicates letters, so the match percentage is slightly lower for words with duplicate letters.

It is supposed to use LCS algorithm for that, and Intersect simply doesn't do the job :(

I even used example in this paper - http://www.morfoedro.it/doc.php?n=223&lang=en

Hope we can fix this together!

Thanks,
Pavel

comments