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 Da...@zycus.com on 2011/09/21 17:49:24 UTC
Custom Score-Similarity
Hi,
I need some insight into the how the document matching actually happens
and how score is generated.
The basic need driving this desire is to search on multiple fields using
boolean queries, and computing similarity score, for each field for each
hit document so as to device a method to smartly select a document from a
hitlist.
The similarity between fields may want to leverage idf's(Lucene
statistics). :(
Your Help in sharing online resources or even better your inputs will be
of great help.
Thanks
Darshan.
Re: Custom Score-Similarity
Posted by Da...@zycus.com.
Thanks Ian, I found it a couple of hours back. Now Trying to figure out
the code flow. Debugging has never been this insightful :)
Regards,
Darshan Pandit
From: Ian Lea <ia...@gmail.com>
To: java-user@lucene.apache.org
Date: 22-09-2011 14:31
Subject: Re: Custom Score-Similarity
http://lucene.apache.org/java/3_4_0/scoring.html would be a good place to
start.
--
Ian.
On Wed, Sep 21, 2011 at 4:49 PM, <Da...@zycus.com> wrote:
> Hi,
> I need some insight into the how the document matching actually happens
> and how score is generated.
> The basic need driving this desire is to search on multiple fields using
> boolean queries, and computing similarity score, for each field for each
> hit document so as to device a method to smartly select a document from
a
> hitlist.
> The similarity between fields may want to leverage idf's(Lucene
> statistics). :(
>
> Your Help in sharing online resources or even better your inputs will be
> of great help.
>
> Thanks
> Darshan.
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org
Re: Custom Score-Similarity
Posted by Ian Lea <ia...@gmail.com>.
http://lucene.apache.org/java/3_4_0/scoring.html would be a good place to start.
--
Ian.
On Wed, Sep 21, 2011 at 4:49 PM, <Da...@zycus.com> wrote:
> Hi,
> I need some insight into the how the document matching actually happens
> and how score is generated.
> The basic need driving this desire is to search on multiple fields using
> boolean queries, and computing similarity score, for each field for each
> hit document so as to device a method to smartly select a document from a
> hitlist.
> The similarity between fields may want to leverage idf's(Lucene
> statistics). :(
>
> Your Help in sharing online resources or even better your inputs will be
> of great help.
>
> Thanks
> Darshan.
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org