PageRank explained

The first company to receive the patent for outbound links accounting system was Google Inc. This algorithm was named PageRank (PR). Here goes some information about this algorithm and its possible affection the search results ranking process.

PageRank (PR) index calculation is made separately for each web-page. This index is defined according to the citation of other Internet pages linking to this one. So there is a certain loop.

The main task lies in the necessity to find the criteria defining the Internet page importance. To calculate the PageRank (PR) index a theoretical page attendance was chosen as the main criteria.

Let’s look over the model reflecting the user site surfing by clicking links. Let’s think that a user starts surfing from a certain page. Then one goes to other sites by the links on the page. There is some probability that the user will leave this site and start surfing from another random page. The PageRank (PR) algorithm estimates this probability as 0.15 for each hop. So we have that the probability that the user goes on his journey by some link among the presented on the page is 0.85. If we do it infinite times than we’ll see that the user will often visit the popular sites and the sites of low popularity will be visited rarely.

So PageRank (PR) for any Internet page shows the probability of that the user arrives at this very page. It’s important that the sum of the probabilities of the user’s presence at all of all Internet pages at the moment is 1 because the user has to be at one of the pages anyway.

As it’s difficult to work with probabilities a system of comparatively simple conversion was suggested, using which PageRank (PR) index obtains a certain value. Some kind of it can be seen by users in Google Toolbar where every page has a PageRank value from 0 to 10.

The model described above shows that:

  • Any Internet page has a non-zero PageRank (PR) value even if none of links lead to it. But this value is minor.
  • Any page having links to other Internet pages gives them some part of its weight. Besides the more links lead to other sites the lesser part is given by any of the links to other Internet sites.
  • Moreover not the whole PageRank (PR) value is given away. We must remember that every hop has a 15% probability that a user may start surfing from another random Internet page.

And now let’s try to understand how the PageRank index can affect the search results ranking. The word ‘can’ is here on purpose because PageRank (PR) index is not used in Google algorithms separately today. The PageRank (PR) influence is easy: when the search engine has chosen some documents according to the user’s request they will be sorted by the PageRank values. Then based on simple logical assumption we get the documents with many high-quality outbound links will have the most valuable information for a user.

So thanks to the PageRank algorithm usage the pages that were popular and needed without any search are put to the top of the search result list.

Оставьте комментарий

Ваш адрес email не будет опубликован. Обязательные поля помечены *