lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jack Krupansky" <j...@basetechnology.com>
Subject Re: Boost on specific fields
Date Tue, 30 Jul 2013 13:16:18 GMT
Can you point to a specific portion of your query that leads you to "expect" 
some particular document to score higher? Please keep in mind that Solr does 
not yet have the ability to read your mind - it simply does its best to 
follow your specific instructions.

Doing a keyword match on a field (has_image) will work fine to "filter" 
documents, but won't affect their score unless it is used in an "OR" manner 
as opposed to a simple "AND" filtering manner. IOW, if all matched documents 
must have this field set to 'Y', then it won't give any document a higher 
score relative to the other matched documents.

You mention wanting a higher count in a field to rank higher - the only way 
I know to do that is using a function query with the "boost" or "bf" 
parameter - what does your function query look like and how did you decide 
whether the boost should be additive or multiplicative?

-- Jack Krupansky

-----Original Message----- 
From: sivaprasad
Sent: Tuesday, July 30, 2013 3:43 AM
To: solr-user@lucene.apache.org
Subject: Boost on specific fields

Hi,

I have indexed product information , the product name, no of reviews and
has_image fields. For example, any two products has "laptop" in the product
name and the user issued "laptop" as the search query, both the products are
coming as part of search results. Here, I am applying boosting on the no of
reviews and the has_image(This will be "Y" Or "N") and I am expecting the
product which has no of reviews count is more and the has_image="Y" should
come first. But, in some of the cases ,  I am not getting what I am
expecting from the Solr.

Is it right way to boost no of reviews and has_image fields?

-Siva



--
View this message in context: 
http://lucene.472066.n3.nabble.com/Boost-on-specific-fields-tp4081224.html
Sent from the Solr - User mailing list archive at Nabble.com. 


Mime
View raw message