Hi Andrew, The example you provide can only partially be performed using a rule based stemmer, such as those uesd by Snowball. Most stemmers are capable of stemming eating, eats, and eaten to eat. However they will not stem ate to eat.
While in theory you could consturuct some form of dictionary to help with these verbal irregularities, it would be an very complex task. Damien. > Hi Andrew, > > ahg <[EMAIL PROTECTED]> wrote on 24/04/2007 12:18:22: > >> Hi, all, >> >> I'm looking for a simple, straightforward example of how to use the >> Snowball stemmer to make Lucene search results return all variants of >> the terms searched for. >> >> For example, if I search for "eat", I'd like Lucene to find "eating", >> "eaten", "ate", etc. >> > > Take a look at the snowball test: > org.apache.lucene.analysis.snowball.TestSnowball > (under src/test) - it does something very similar. > >> In particular, I'm not clear on whether everything required to do this >> is the contrib/snowball/ directory, or do I have to find an external >> dictionary somewhere? > > No, snowball stemmers are algorithmic/rule based, no dictionaries. > >> >> In summary: I just need a quick explanation as to what resources I need >> and how to put them together. >> >> (By the way, I'm actually looking to do this with Spanish rather than >> English, but examples for any language would be great.) >> >> Many thanks in advance, >> Andrew Green > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
