Forum Moderators: open
I have looked closer into a new European search engine project/study that tries to make a new kind of "smart" search engines that learns and get smarter as they go along.
Well, reading about this project I was thinking of the current "theme" engines and how they work - Iīm a novice in this - sorry have to ask...
How does the "theme" engine deside whatīs a theme? Is it based on the keywords only and do they learn something?
Lets say you have a mobile entertainment site :-) I guess ringtone and ringtones belongs to the same theme, but what about ring tones, tones, tunes, tone etc. My question is where does the "theme" end? To us people ringtones, logos, mobile games are the same "theme" and belong to mobile entertainment - how about the theme engines. Same thing with mobile phone, cell phone and wireless phone?
This is a basic question, but I didnīt find any clear answer to it. The idea of the "european" SE -project I refered to was to create a SE that learns and sees the theems beyong keywords (and much more...). Could be this isnīt anything new?
The project page: [cosco.hiit.fi ]
If not already known to you, you might want to have a look at some of these:
[labs.google.com...] (similar terms-sets)
[labs.google.com...] (clustering)
www.vivisimo.com (clustering on the left)
www.teoma.com (refine option on the right)
Topic sensitive Pagerank [webmasterworld.com]
Outride and personalised search [webmasterworld.com]
I think search engines would love to implement better theme based results yesterday (preventing Pagerank buying, guestbook signing etc).
