[FoRK] A better PageRank --- prediction!

Lucas Gonze lgonze at panix.com
Sun May 29 23:15:22 PDT 2005


N.B.: PageRank hasn't been a central part of search engine algos for quite 
a while.  This is because it always converges on the big link factories 
like Yahoo, etc.

(It's still in use, just limited to being one weighted factor among many. 
This is why Webjay usually ranks higher for any artist that the artist's 
own page.).


On Mon, 30 May 2005, Jeff Bone wrote:

>
> PageRank is what I call a "psuedo-intelligent" algorithm.  It' very 
> effective, but it only simulates intelligence by exploiting the topological 
> "connectedness" of the Web.
>
> The next, and better, version of PageRank or similar will take into account 
> not only topological connectedness but also will discount, as per my recent 
> suggestion about "weighting" opinion in mailing list conversation, 
> self-similarity among topologically-related pieces of content.  It will 
> finally and utterly eliminate, for example, propagandistic influence 
> (deliberate and otherwise) of particularly infectious memes, and will distill 
> out the real "reasoning" (modulo legitimate difference of opinion) on the 
> Web.  Still psuedo-intelligent, but nonetheless a quantum leap over 
> previously less-intelligent algorithms.
>
> Zip saves humanity, film at 11.
>
> "And a child shall lead the way."  I'm reminded of the post in which 
> (no-longer-child) Aaron Swartz reciting a common saw, but in a particularly 
> (to me, at least) clear and concise way:  the relationship between 
> compressibility, content, and similarity:
>
>   http://www.aaronsw.com/2002/codemag/
>
> $0.02,
>
> jb
>
> _______________________________________________
> FoRK mailing list
> http://xent.com/mailman/listinfo/fork


More information about the FoRK mailing list