Skip to main content.
home | support | download

Back to List Archive

Re: Ranking, even with strong bias

From: Peter Karman <peter(at)not-real.peknet.com>
Date: Thu Feb 03 2005 - 21:46:09 GMT
Bill Moseley wrote on 2/3/05 3:16 PM:

> On Thu, Feb 03, 2005 at 12:31:21PM -0800, Tac Tacelosky wrote:
> 
>>We're indexing about 130 XML files, varying from tiny to huge.  They
>>basically consists of products with a title and description, and they're all
>>merged into one giant index (for speed) at the end.  I'd really like the
>>results to be heavily biased if the search term is in the title.  My config
>>file is set to
> 
> 
> One thing you might try is indexing as html, as <title> should rank
> quite a bit higher for html docs.
> 


that's likely the quickest solution. that, plus MetaNamesRank, should help. The 
default RankScheme is probably most helpful if the docs are all about the same 
size. Otherwise, you might try RankScheme 1. If you're using the swish.cgi 
script, I think you want to add a '-R 1' to the swish-e shell call.

indexing as html will artificially inflate the number of occurances whenever a 
word matches in the <title>.

-- 
Peter Karman  .  http://peknet.com/  .  peter(at)not-real.peknet.com
Received on Thu Feb 3 13:46:10 2005