Search Engine Journal » PageRank http://www.searchenginejournal.org Wed, 28 Dec 2011 01:41:11 +0000 en hourly 1 http://wordpress.org/?v=3.3.1 Increase web site visitors http://www.searchenginejournal.org/increase-web-site-visitors.html http://www.searchenginejournal.org/increase-web-site-visitors.html#comments Sun, 21 Jun 2009 04:58:37 +0000 admin http://searchenginejournal.org/?p=233 On realizing the triumphal procession of PageRank, one can’t help thinking about its increasing for oneâs page. It is intuitively clear, that the more the recourse, on which there is a link, is authoritative, the more it raises PageRank of the page, to which it is linked. And vice versa, the more links are on the page, the lower will be its investment to increase your PageRank – one more proof of vainness of participating in FFA (Free For All – sites, which content a set of links with free adding).
An optimal topology of reciprocal linked pages is less evident. For example, pages which are organized into “ring” (when every page links to its neighbour on the left and on the right, the last links to the first, the first links to the last) will have one and the same PageRank independently of the amount of pages in the ring (if not to make ranging according to the sum, everybody will have PageRank equal to 1).

The same is valid for starts or the case, when everybody link to everybody and probably this statement is valid for all symmetric topologies. Much more perspective from the point of view of increasing PageRank are dissymmetric topologies. The statement about vainness of creating free hosters’ empty (but linked) sites isn’t so evident.

For example, you can organize links exchange on 5 sites so, that one of them will have PageRank 15 times higher than the minimal non-zero PageRank. You can make sure of it on writing a small program.

]]> http://www.searchenginejournal.org/increase-web-site-visitors.html/feed 0 PageRank Explained http://www.searchenginejournal.org/pagerank-explained.html http://www.searchenginejournal.org/pagerank-explained.html#comments Thu, 18 Jun 2009 12:03:54 +0000 admin http://searchenginejournal.org/?p=47 Appraisal of link importance.
Term “Link Popularity” is a bit incorrect. It would be much more close to what it means if it would be called “Link Topology”, because this method considers relationship of links along with quantity. However, as a result of analysis we receive “importance” of a page. This is not what “relevancy” is. Relevancy shows how contents of your page correspond to a particular search query. “Importance” shows value of page, regardless of it’s contents. Any inbound link states that this page has some value and it increases it’s “importance” this way. The more rating it has, the more important” it is.
Not all links are making equal contribution in page’s rating. Some of linking pages can be more important than others and so on, thus outbound link from such page is more important.
So, “Important page is a page that has links from important pages. Exclusive circle? Yes, it’s rather easy to understand subconsciously. For instance a link from NASA will be more important than a link from your cousin’s Kate homepage – not because NASA loves you more, but because there are thousands of sites linking to NASA and just a couple of them linking to Kate’s.
How the “importance” is measured.
Though it is easy to understand on instance of relationship between two pages, measuring of importance of milliards of related pages seems hopelessly complicated. Indeed this is really complicated, but not hopelessly – everything’s almost easy. Such measures demand lots calculations, but fortunately we shan’t invent anything new. We can just take ready formulae from scientific sources.
Larry Page and Sergey Brin, the founders of Google and first developers of it’s algorithm have published “The Anatomy of Large Hypertext Search Engine”. You can download it from http://www.searchenginejournal.org in PDF format. The document describes the Page Rank technology – method of appraisal of page important, measured proceeding from pages linking to the appreciated page.

So, the Page Rank formula. It looks complicated, but it just looks so. In practice, you will need just a little knowledge of algebra (I don’t know whether algebra is studied in such volumes in schools of Her Majesty’s Land and US, but in my country math is studied since 7 y.o; algerbra and senior math since 10 y.o.)

For instance, there are page A, which has inbound links from other pages. Let’s call them T1, T2, T3, and so on up to Tn.
No math yet, we’ll just give names to things that we are going to speak about. Imagine that A is your homepage and T1-Tn – other pages, which contain hyperlinks pointing your page. For instance, T2 can be a homepage of your cousin Kate (if this helps in understanding)

PageRank of page A is calculated using the following formula:

PR(A) = (1-d)+d [PR(T1)/C(T1)+PR(T2)/C(T2)+PR(T3)/C(T3)+…+PR(Tn)/C(Tn)]
In case if it looks complicated for you, let’s divide it in three groups:
PR (A) means PageRank of a page A – value we are trying to calculate. This expression just defines the problem – all calculations will be on the other side of “=”.
( 1-d) + d – fade ratio. Don’t pay attention to it. Page and Brin recommend to measure it equally to 0,85. so we will set it 0,85 and forget about it. Though it is important if you create a search engine, our calculations allow taking ready value. We are just going to calculate expression in brackets, multiply it by 0,85 and add 0,15 to the result, as it is mentioned in formula.

Now let’s get back to the expression in brackets and write it as follows:
[ PR(T1)/C(T1)+
PR(T2)/C(T2)+
PR(T3)/C(T3)+…+
PR(Tn)/C(Tn)]
It’s easy to see that T1, T2 and T3 are that pages, which link to A. I hope it’s easy to make calculations with these simple formulas you have received after dividing. Obvious difficulty is just in quantity of calculations.
PR means Page Rank of T1, T2… Tn pages. The only novelty that appears in this formula is C – quantity of hyperlinks on the given page. C(T2) is common quantity of outbound links on T2 page, e.g. links of such kind:

http://www.searchenginejournal.org

This link is an inbound link for page, where it points.
Having united these three components, which we have previously divided, we can define sequence of actions applying this formula to any particular page.
Create a list of all pages, which link to this page.
Define the following values for each page:
PageRank, outbound links,
Divide PageRank by outbound links (e.g. if PageRank is 6 and there are three outbound links the result will be 6:3=2. )
Make sum of such results for each inbound link.
Apply fade ratio to the result.

]]> http://www.searchenginejournal.org/pagerank-explained.html/feed 2