You are viewing a plain text version of this content. The canonical link for it is here.
Posted to java-user@lucene.apache.org by "Madhu Sasidhar, MD" <ma...@gmail.com> on 2005/04/26 15:22:23 UTC

multi word synonym

I have found the previous discussions on multi word synonyms as as well as 
the section on synonym injection in Hatcher's book, but have not been able 
to come up with a satisfactory solution. I am indexing text that has several 
multi word synonyms. Some of the synonyms may have single words as one of 
the synonym. These words should only match for the appropriate multi word 
phrase searches - in other words, "chest pain", should only match a query 
for "chest pain" and not for "chest" or "pain". In addition, it will match 
angina (synonym). So, in this sentence:
Lab results for alpha 1 antitrypsin level....
I would like to index 'alpha-1-antitrypsin', 'antitrypsin', 'antitrypsin, 
alpha 1', 'A1AT' as synonyms for the phrase alpha-1-antitrypsin in the 
sentence. Thanks in advance...
madhu

Re: multi word synonym

Posted by Paul Libbrecht <pa...@activemath.org>.
If I understand well... it would be easy to do so if you do not wish to 
use phrase matches... you could just add a field (with the same name) 
for each token...

I think that, if you wish phrase-matches (or the span-ones) then Lucene 
can't help you... but I'm quite a newbie on this topic.

Is there a hope this becomes different in Lucene 1.9 or 2.0 ??

My dream would be to have the position increments living in a tree... 
you know.... and... XML tree...

thanks

paul

Le 26 avr. 05, à 15:22, Madhu Sasidhar, MD a écrit :
> I have found the previous discussions on multi word synonyms as as 
> well as
> the section on synonym injection in Hatcher's book, but have not been 
> able
> to come up with a satisfactory solution. I am indexing text that has 
> several
> multi word synonyms. Some of the synonyms may have single words as one 
> of
> the synonym. These words should only match for the appropriate multi 
> word
> phrase searches - in other words, "chest pain", should only match a 
> query
> for "chest pain" and not for "chest" or "pain". In addition, it will 
> match
> angina (synonym). So, in this sentence:
> Lab results for alpha 1 antitrypsin level....
> I would like to index 'alpha-1-antitrypsin', 'antitrypsin', 
> 'antitrypsin,
> alpha 1', 'A1AT' as synonyms for the phrase alpha-1-antitrypsin in the
> sentence. Thanks in advance...
> madhu


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org