Stemmer and Synonym analyzer

I am planning on building an analyzer that has stemming,
stopwords and
synonyms.  I am planning on using the Snowball Porter
stemmer and the
WordNet synonym engine.

Does it make sense to stem the synonym index?  

I do not want to stem the term “history” and  then try
to find the synonym. 
The stem of “history” is “histori” which will not
have a synonym in the
index unless I originally stemmed all the terms in the
synonym index.
(synonyms(stem(tokenstream), stopwordlist))

Alternatively, I can find the synonyms of the token stream
and then stem all
of them.  This solution should not require stemming the
synonym index.
(stem(synonyms(tokenstream)), stopwordlist)

Does anyone have any experience combining a stemmer and a
synonym analyzer
together?



-- 
View this message in context: http://www.nabble.com/Stemmer-and-Syn
onym-analyzer-tf4685221.html#a13388811
Sent from the Lucene - Java Users mailing list archive at
Nabble.com.


------------------------------------------------------------
---------
To unsubscribe, e-mail: java-user-unsubscribelucene.apache.org
For additional commands, e-mail: java-user-helplucene.apache.org