I've got a system that searches the Gorniad's new indexes against terms which come from the user and are cross-referenced against a db of relations. It's commercial, so I can't give it to you, but it took about a day to write it.
Not very AI, IMO, but then what is...? Lee At 17:01 24/02/2002 -0700, Sean M. Burke wrote: >This is an AI question more than anything else, so I'll sling it here, >unless anyone knows of anyplace else more relevent for it: > >The Guardian, www.guardian.co.uk, seems to encourage people to pass around >their content, as long as it's credited (See: >http://www.guardian.co.uk/Distribution/CDistr_Help/ ). And I enjoy >reading The Guardian, so for a few weeks now, I've been basically using >their "content distribution" facilities to just mirror each day's new >stories to my hard drive, so I can read the paper while I'm offline. (I'm >not actually redistributing the content to others.) It's about 120 >stories of basically plaintext every day, about 500K a day. > >So after a while of having these stories download every day, I started >thinking that it wouldn't be hard to make up a system where people >selected the /kinds/ of news they're interested in; and, then based on >that, a program could point them at the most relevent stories in a given >day's news. But with finer grain than just "I like stories in the tech >Technology section, and nothing from the Sports section!" More like an >Amazon recommendation system, but for news stories. > >I'm sure systems like this must exist, but I've never seen >one. Ideas? Examples? > >-- >Sean M. Burke [EMAIL PROTECTED] http://www.spinn.net/~sburke/
