[Xapian-devel] GSoC - Improving Spelling Correction

晓兵 hitandyliu at gmail.com
Mon Apr 11 08:45:25 BST 2011

Hi Everyone,

I’m a student in Peking University, China, my name is Xiaobing Liu, I’m
interest in Spelling correction project.

I think Xapian’s Spelling correction is based on edit distance with trusted
dictionary. I have some questions about the spelling correction algorithm.

1.      The Spelling correction algorithm depends on which dictionary, the
dictionary in inverted index or other trusted dictionary?

2.      A simple approach to get the correction words is traversing the
dictionary and ranking for each word, when the dictionary is huge, this
approach should be quite time consuming, how Xapian solve this problem
effectively ?

3.       Since search engine always suggest only the top one correction for
spelling, improving the aspect  “only a single suggested correction is
available via the API” may be not so useful.

Any discussion is appreciated.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.xapian.org/pipermail/xapian-devel/attachments/20110411/a847b8bc/attachment.htm>

More information about the Xapian-devel mailing list