On 8/11/07, Karen Coyle <kcoyle_at_kcoyle.net> wrote:
> Take a look at: http://library.lanl.gov/cgi-bin/getfile?LA-UR-05-8420.pdf
>
> They used usage data to create clusters of activity, and then user data
> to link those clusters. So this creates clusters of "hot" and also
> creates paths of searching that others could be allowed to follow (or
> could be shown).
Yeah, that's cool. The problem, though, is the lack of a model that
deals with the contextual decline of "hot" over time (or even
contextualize it *in* time) and through other subject proxies /
contexts. This is where AI over statistics will do a better job. One
day. :)
Alex
--
---------------------------------------------------------------------------
Project Wrangler, SOA, Information Alchemist, UX, RESTafarian, Topic Maps
------------------------------------------ http://shelter.nu/blog/ --------
Received on Fri Aug 10 2007 - 10:36:33 EDT