Google PageRank Patent analysis

[jpservicez1]

Does anyone know where on this site or any other site where i can find a thoroughly analysed doc/posting of google pagerank patent filing/submission.?

The reason i'm asking this is because with all the issues discussed on this site in regards with the changes recently on google indexing. All lot more is involve and this document might just have the answer for me.

[tedster]

The first thing to say is that PageRank, the pure PR thing, is an essentially simple formula that's been the same since Larry and Sergey were still at Stanford. So pure Page Rank analysis is all over the web by now -- but the Page Rank patent is only the most basic of what has now become about 250 patents filed by Google relating to search technology and how to rank web pages for queries.

Actually, the USPTO is a bit hard to use for this kind of discovery, because in the US many of the patents have been filed under individual names and "Google" is not found anywhere in the filing. In the EuropePatent Office [european-patent-office.org], more of the patents are filed with the Google name as the owner, and that's where a friend of mine located the 250 or so. (Warning: search on that site can be unbearably slow!)

My own opinion is that one key to this moment in Google is last year's historical information patent [appft1.uspto.gov] and its 137 various points. Last year, many people thought that patent filing was just FUD. Now, I see evidence that Big Daddy gave Google the tool and the "elbow room" it needed to start bringing these factors to life.

To answer your question most directly, the original PageRank paper is here [www-db.stanford.edu]. It is an essentially simple formula for calculating a ranking factor called PR for any page, based on the number of inbound links from other pages and how many links those pages have. While the math needed to caluculate PR has a complexity (it's an "iterative" calculation that goes through pages over and over, until PR values settle down and no longer show large variations on every iteration) the basic equation is not complex in its concept:

We assume page A has pages T1...Tn which point to it (i.e., are citations). The parameter d is a damping factor which can be set between 0 and 1. We usually set d to 0.85. There are more details about d in the next section. Also C(A) is defined as the number of links going out of page A. The PageRank of a page A is given as follows:

PR(A) = (1-d) + d (PR(T1)/C(T1) + ... + PR(Tn)/C(Tn))

Note that the PageRanks form a probability distribution over web pages, so the sum of all web pages' PageRanks will be one.

So the PR of PageA is calculated by looking at all the pages that link to PageA, dividing that PR by the number of links on the page. Mix in a "damping factor" to keep PR from flying off to infinity and there you have it.

Todays Google, as I said earlier, only uses PR as a small part of the overall algorithm, which one newspaper recently reported now contains more than 200 factors.

[tedster]

An even simpler way to think about PageRank is this. Imagine a "random walk" around the web. Just start anywhere, on some random page. Now click on one of that page's links, also chosen at random. Then do the same on the next page you find, and keep on going.

The PR of YourPage indicates how likely it is that you will end up finding YourPage by this method.

[jpservicez1]

tedster..

thanks for your reply..

Actually you have hit the nail on the head...when you mentioned "historical information patent"..i was looking for google patent filing doc that is relevant to the changes we are seeing now.

And i feel on this forum the likes of you and other SEO experts should help us understand what's in these documents so that one can use the information/analysis provided to understand and adhere to the changes that google has now introduce.

My point is we need a complete list of the likely changes and direction google is taking now so that all webmasters can practice those changes on their sites..