You can start with the Relevance Scores chapter in the Search Developer's Guide:
http://docs.marklogic.com/4.2doc/docapp.xqy#display.xqy?fname=http://pubs/4.2doc/xml/search-dev-guide/relevance.xml -Danny From: [email protected] [mailto:[email protected]] On Behalf Of Rajesh TR Sent: Thursday, July 14, 2011 2:02 PM To: General MarkLogic Developer Discussion Subject: Re: [MarkLogic Dev General] Relevancy score influence across 2 different data Hi Kelly, Thanks for your response. I am using ML4. Can you point to the doc on how to boost relevance scores? I have read the documentation but couldn't find details on how to do this. * Changing document quality is not preferred. How to determine how much of document quality should be inputed to achieve the desired order results between the two content set * Length of document based tuning might be more appropriate - If you can point to some docs it will help. Thanks and Regards, Rajesh TR. On Fri, Jul 8, 2011 at 3:19 PM, Kelly Stirman <[email protected]<mailto:[email protected]>> wrote: Hi Rajesh, I think the best way to do this is with document quality. This is a setting applied to each document that boosts relevance scores. You can optionally increase, decrease, or ignore this boost at query time. Changing document quality requires updating the document. I'm not sure what version of MarkLogic you are using. In recent versions there are options for normalization of relevance score based on length of documents that may be worth investigating. This is a database setting. Kelly Message: 1 Date: Thu, 7 Jul 2011 16:03:17 -0400 From: Rajesh TR <[email protected]<mailto:[email protected]>> Subject: [MarkLogic Dev General] Relevancy score influence across 2 different data To: General MarkLogic Developer Discussion <[email protected]<mailto:[email protected]>> Message-ID: <CAB4uqzoED4HYLRmD89dn=zxuivc86gbfrf8zkkyavfn+mxk...@mail.gmail.com<mailto:zxuivc86gbfrf8zkkyavfn%[email protected]>> Content-Type: text/plain; charset="iso-8859-1" Hi, I have come across atypical situation and wondering if Marklogic can handle this scenario. *Content Type 1: Book (1 MB content)* *Content Type 2: Course (1kb is the maximum content describing agenda of course video).* Search "Management" across both Books and Courses. The number of "hits in Books" Vs number of" hits in Courses" is skewing the relevancy badly so that Courses will never be listed first even though the there is a perfectly matching course. Has anybody across such a situation before? Is there a solution to it? NOTE: we cannot search only on Description because we will miss out so many books. One solution that I can think of is to give higher weights for Description field. But I want to hear from the experts if there is something like "Function Queries" or "Boost Functions" available? If so how, any guidance on how to use it. _______________________________________________ General mailing list [email protected]<mailto:[email protected]> http://developer.marklogic.com/mailman/listinfo/general -- Thanks and Regards, Rajesh TR.
_______________________________________________ General mailing list [email protected] http://developer.marklogic.com/mailman/listinfo/general
