What about the original paper they wrote? It's actually easier to understand than that. |
Ah, it's not that bad. Just do what I do; think of it as a Singular Value Decomposition on a really really massive matrix. I'd start with Golub and van Lo*n if you'd like to work up to the math..
Philipp, your software wouldn't let me write L*an. It's the author's name. :) |
"Ah, it's not that bad. Just do what I do; think of it as a Singular Value Decomposition on a really really massive matrix. I'd start with Golub and van Löan if you'd like to work up to the math.."
did ANYONE understand that? ;o)
PS the löan filter is a great idea sir spam fighter, the beginnings of your own Bayesian filter! |
Hello! do you know if there is something inside about pageRank's philosophy ? thanks |
Palpitt, the book has 3 introductory chapters before the math starts, so there is also discussion of their philosophy. Turns out in 1998 (according to the book) two groups independently came up with "make the link democracy rank pages" approaches. The one was Brin/ Page with PageRank/ Google/ Backrub, the other was Jon Kleinberg with HITS (Hypertext Induced Topic Search). The book says 3 factors made PageRank win over HITS as dominant model: - query independence - virtual immunity to spamming - Google's huge business success
I guess the time was ripe, as ever so often. If you read The Google Story you'll also see that while time was ripe, the business world wasn't – the Google founders weren't able to sell their algo to an existing search company (who around that time were suffering portalitis). |
theres an mp3 of matt cutts interview on search here
http://www.clickz.com/experts/search/results/article.php/3602456 |
thanks a lot, Philipp, for these explanations ;) |
Here is the HTML version of the PDF paper they released in Oct 2004 –
http://72.14.203.104/search?q=cache:iZ3uom5emKYJ:meyer.math.ncsu.edu/Meyer/PS_Files/DeeperInsidePR.pdf+&hl=en&gl=us&ct=clnk&cd=1 |
Travis: I understood that – but it helps to have a math degree from MIT :-)
|