The Random Walker Modeling
There is another way of interpreting PageRank (which may have besides been the base of the formulation of the algorithm!), based on the modeling of the random surfing of a Net surfer on the Web.
Let us imagine that a Net surfer walks on the Web by reading pages, jumping from one to another by clicking randomly on a link. In certain cases he may be blocked in a group of pages linked together but apart from the remainder of the Web. In this case it jumps randomly on another page of the Web. This abrupt change can also come directly if the walker who wearied that pages he was visiting, or if he did not find what it was looking for...
The PageRank of a page can then be seen as the probability that at a given time this surfer is precisely on this page.
The more many other pages link back to it, the more this probability, especially if these pages are also significant: it is the second term of the formula .
The first term models the probability that it remains on the same page, without following outgoing links. The factor D in the formula can thus be seen as the probability that the surfer jumps on another page.
It is time to pass to the conclusion on PageRank!
