En Sat, 24 Jan 2009 15:08:08 -0200, S.Selvam Siva <s.selvams...@gmail.com>
escribió:
I am developing spell checker for my local language(tamil) using python.
I need to generate alternative word list for a miss-spelled word from the
dictionary of words.The alternatives must be as much as closer to the
miss-spelled word.As we know, ordinary string comparison wont work here .
Any suggestion for this problem is welcome.
I think it would better to add Tamil support to some existing library like
GNU aspell: http://aspell.net/
You are looking for "fuzzy matching":
http://en.wikipedia.org/wiki/Fuzzy_string_searching
In particular, the Levenshtein distance is widely used; I think there is a
Python extension providing those calculations.
--
Gabriel Genellina
--
http://mail.python.org/mailman/listinfo/python-list
