Singular / Plural differences in words

[aschecht]

Many of my users are getting confounded by searches failing to find singular/plural differences.

For example, my content contains the word "gallstone" but people who search for the word "gallstones" aren't finding it. I know that technically this is how a search engine is supposed to work (since the word gallstones isn't present) but is there a simple way to make it so that whether they searched for gallstone or gallstones it would still find the included content (gallstone)? THis would be especially awesome in the search suggestions section. Right now it gives me totally unrelated things - “Did you mean: closed or Clostridium or Cluster”" instead of "Did you mean: gallstone" which would be perfect.

Any suggestions on how to improve this for my users?

Andrew

[Ray]

This has been discussed in detail in this previous thread:
http://www.wrensoft.com/forum/showthread.php?t=537

Unfortunately, there is no simple solution that will work universally. But if there is a small number of words which are especially important to you with singular/plural forms, you can create a list of synonyms to achieve this, or include all the word variants as meta keywords. More details in the abovementioned thread.

[aschecht]

Thanks for the reply. Interesting discussion in the old thread.
I'll look into adding metadata. Seems like the best (although not an easy) solution.

Andrew

[aschecht]

Synonym Limit

I think I'm going to use the synonyms feature rather than adding metadata to
solve the singular/plural issue and also to solve the problem with "apostrophe s" (ie. people looking for Raynaud's disease might search for "Raynaud" which wouldn't match).

I may end up with one to two thousand synonyms on my list which will be easy to manage with the import option.

The question: Will 2,000 synonyms on the list cause any problems?

Andrew

[Ray]

There shouldn't be. But note that synonyms count towards you "Total unique words indexed" count and thus, your limits specified. Also remember to enable "apostrophe" for the word join rules.

[aschecht]

OK. Thanks.

Related question. If I designate a character as one that should separate rather than join words, would the search engine know when searching for a term with that character to look for the individual components?

For example, I define the dash (-) as a character that should separate words. I have in my content "abcd-efgh". Since I defined the dash to separate words, I'd expect this to be indexed as two separate words: abcd and efgh. If someone searched for the string "abcd-efgh" (no quotes) would the search engine know that anything with a dash would have been indexed as separate words and then adapt the search to find abcd and efgh rather than coming up empty since there's no single word indexed as "abcd-efgh"?

Andrew

[David]

The indexer and the search scripts both follow the same rule. So if you break words on a dash, both the indexer and the script will break words in the same way. In effect the dash is treated as white space is you are breaking words with a dash in them.

[aschecht]

Cool. Thanks.

Andrew