|
|
Posted by W Luke on 02/10/05 15:10
On Wed, 09 Feb 2005 10:56:51 -0800 (PST), Matthew Weier O'Phinney
<matthew@garden.org> wrote:
> * W Luke <wtluke@gmail.com>:
> > I've been fascinated by Flickr's, del.icio.us and other sites' usage
> > of these Weighted Lists. It's simple but effective and I really want
> > to use it for a project I'm doing.
> >
> > So I had a look at Nick Olejniczak's plugin for Wordpress (available
> > here: www.nicholasjon.com) but am struggling to understand the logic
> > behind it.
> >
> > What I need is to dump all words (taken from the DB) from just one
> > column into an array. Filter out common words
> > (the,a,it,at,you,me,he,she etc), then calculate most frequent words to
> > provide the weighted list. Has anyone attempted this?
>
> Funny you should mention this -- I'm working on something like this
> right now for work.
>
> Basically, you need to:
>
> * define a list of common words to skip
> * define weighting (I weight items in a title and in text differently,
> for instance -- usually you weight by which field you're using); store
> weighting in an associative array
> * define a weights array (associative array of word => score)
> * separate all text from the column into words (build a words array)
> * loop over the words array
> * skip if the word is a common word
> * increment word element in weights array by the weight
>
> The sticky issues are: what is a word (you'll need to build a regexp for
> that), and how will you weight words (usually by field). Once you have
> all this, you populate a database table for use as a reverse lookup.
Thanks Matthew - a similar logic to what I'm using, although mine (so
far) is a little clunky. I might ping you later on today if that's
ok, to talk a bit more and share notes.
Regards,
--
Will The Corridor of Uncertainty http://www.cricket.mailliw.com/
- Sanity is a madness put to good use -
Navigation:
[Reply to this message]
|