Hi Everyone,
Just thought I'd accompany the latest post to the backstage blog
(http://backstage.bbc.co.uk/news/archives/2007/11/from_last_years_1.html)
with some examples of muddyboots in action. For those of you who aren't
aware of the project it's probably best to look at
http://muddyboots.rattleresearch.com/cgi-bin/mb.cgi?action=more.
Essentially we're attempting to use Wikipedia and other commons authored
data sources to augment the meta-data around BBC news stories, this
ultimately took the form of automated contextually relevant link
recommendations based off data within Wikipedia and del.icio.us
(although we have some other ideas about how this data could be used ...)
It's still a prototype so it's not production ready by any means, there
are still stories where we are unable to recommend links and there are
others where ambiguity becomes a problem and identifying what context a
story has can be difficult (although we have some ideas around using the
disambiguation data within Wikipedia to improve this).
Here are a few links to stories where I thought muddyboots added some
interest and hopefully a little of that Wikipedia 'browse experience' :
http://muddyboots.rattleresearch.com/cgi-bin/mb.cgi?action=page&id=646
http://muddyboots.rattleresearch.com/cgi-bin/mb.cgi?action=page&id=630
http://muddyboots.rattleresearch.com/cgi-bin/mb.cgi?action=page&id=622
http://muddyboots.rattleresearch.com/cgi-bin/mb.cgi?action=page&id=643
If you'd like to see how those recommendations were arrived at then each
story has a 'View' action which can be used to get a breakdown of each
stage of the muddyboots process, for example :
http://muddyboots.rattleresearch.com/cgi-bin/mb.cgi?action=view&id=622
It's worth noting we only keep the last 50 story submissions in the
system, so these links will eventually 'age' out.
(Disclaimer : I worked on the project)
Thanks,
Rob
-
Sent via the backstage.bbc.co.uk discussion group. To unsubscribe, please
visit http://backstage.bbc.co.uk/archives/2005/01/mailing_list.html.
Unofficial list archive: http://www.mail-archive.com/[email protected]/
