Clearspace allows for sorting of search results using 4 sort criteria, including "relevance".
Relevance ranking for search results can cover a variety of metrics, including number of matched words, frequency of occurrence of matched words, word position (e.g in event title versus event description), tag matching and others.
Is there documentation available describing the logic used by CS when applying relevance to the search results?
Thanks,
Patrick
hi Patrick,
We use an open source project called Lucene for searching, they have a great document that talks about how hits are scored here:
http://lucene.apache.org/java/docs/scoring.html
We obviously do some of the work of getting items into the search index and determining what's indexed and how it's indexed, going into the details here would be a really long post. You can get a good idea of how items are indexed by checking out the source code for com.jivesoftware.community.impl.search.LocalSearchManager. Start with the toSearchDocument(DocumentBean bean) method, take a look at the
private Field field(int objectType, IndexField name, Object value) implementation and the class com.jivesoftware.community.search.BoostParser class.
I'll also ask about getting some of this documented.
Cheers,
AJ
Gary, the documentation looks like it moved to: http://lucene.apache.org/java/2_3_2/scoring.html
Nova, here is a link to the search tips in the documentation which might help, and in general the best way to influence search results as an end-user is to choose a descriptive subject and add descriptive tags to the content. These two elements help with search relevance since the subject and tags of content are weighted more heavily than the body for relevance.
I hope that helps, and if you'd like more information, please ask.
Jive combines collaboration software, community software & social networking software into the leading SBS solution.
© Copyright 2000–2009 Jive Software. All rights reserved.
915 SW Stark St., Suite 400, Portland, OR 97205