On 30.9.2002 16:48 Uhr, Greg Fenton <[EMAIL PROTECTED]> wrote:
> I am trying to tweak htsearch parameters to "improve" my search
> results.
>
> I do a search for "foo bar" and a document with the title "foo bar bob"
> is showing up as #4. Marketing wants it as #1. The first few
> documents do contain "foo" and "bar", but not in their titles or in
> other places I would consider "relevant".
Take a look at the title_factor setting:
<http://www.htdig.org/attrs.html#title_factor>
and all the other various _factor settings (some need reindexing to take
effect). With these you can tune your results quite nifty.
I had a similar issue a week a go and it turned out, the PDF contained the
word multiple times whereas the document tith the keyword in the title only
contained it once, therefore the pdf's being listed with a higher priority.
I took me quite some trial and error to get the values for the _factor
settings right. Currently I use:
description_factor: 350
title_factor: 150
heading_factor_1: 60
heading_factor_2: 50
heading_factor_3: 40
heading_factor_4: 30
heading_factor_5: 20
heading_factor_6: 10
multimatch_factor: 4
date_factor: 2
The fastest way however will probably be to add your keyword to the body of
the document with the Title-Match multiple times in an invisible color or as
meta-tags (if you use them while indexing) as already mentioned.
--
<http://www.StefanSeiz.com>
Spamto: <[EMAIL PROTECTED]>
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
htdig-general mailing list <[EMAIL PROTECTED]>
To unsubscribe, send a message to <[EMAIL PROTECTED]> with a
subject of unsubscribe
FAQ: http://htdig.sourceforge.net/FAQ.html
