Stanford's Global InfoBase

[msgraph]

Global InfoBase Project [www-db.stanford.edu]

The Global InfoBase project has been around for some time but rarely is information posted on it.

This part I find very interesting:

Personalization: We are currently exploring how to adapt search and ranking mechanisms to take into account user preferences. For example, the "page rank" algorithm used by the Google search engine (www.google.com) can be modified to give bookmark pages additional weight in the recursive page rank evaluation.

Some relation to this?

Personalized Results: Exploring The Future Of Google [webmasterworld.com]

Sounds similar to me.

[brotherhood of LAN]

Interesting stuff. But at the end of it, all they are saying is they will have a database with an algo.

The added concept of personalised results is something I've not read much about (I guess this is time to), but it seems to make sense.

It seems more realistic that two different people typing in the same word will not be receiving the same "relevance" due to the differences in what they want returned. Very interesting idea, sounds very computational and resource hungry from my own short-viewed standpoint!

Looks like I am going to have to read up your earlier post! But the whole concept sounds interesting.

In the long run, seems SEO will be less important and more emphasis on the actual market you want to attract as opposed to SEO'ing a page to be noticed........or something along those lines IMO ;)

[rjohara]

Ha, that's what the voting buttons are going to be for on the Google toolbar! They won't have any influence on PR in the main database nor could they - everyone was wondering how those buttons could be used since sending lots of negative votes for someone else's site obviously would not be a fair practice. But the voting buttons will allow users to create their own "correction factor" for Google's default rankings based on personal interests. This is the very thing Amazon does by allowing you to vote a recommendation on an item - it doesn't alter the ranking of the item per se, it alters what recommendations are presented to you. Bingo.

There are some similarities to neural network theory here, I think, although that isn't a subject I really know a lot about. The system can become self-reinforcing for individual users over time, and more and more refined results will be presented. The drawback to this behavior in neural networks is that they can get in a rut and make it difficult for new connections to override old ones when conditions change. I suppose one can program around that in software, though, more easily than in wetware (brains). Very interesting. The future is now clear.

[brotherhood of LAN]

maybe OT but,

those who thought "smily-ing their site to heaven" helped them in rankings may be the only people in the world who are actually seing those number 1 rankings due to "personal results" ;)

[wasmith]

How could such a algo be implemented, short of a personal database that takes your own shortcuts into account as the begining weight for whats important i _believe_ (i do not use such strong terms on a regular bases) it can not be implemented.

_but_ a [groupŚcultureŚcommunityŚwebsite] version may be practical. In fact i would be fascinated to search based on placing a country code as the starting point for determining the PR of sites. It has been alleged that google began using yahoo as its starting point (but what if another starting point? what would be different?)