Reputation: 2766
I'm learning PageRank algorithm and from Wikipedia, it gives the following formula:
From the formula, the page rank is calculated from the ranks of pages linking to it. Also, they give a simple example of four nodes A,B,C,D. Initially, each node has a page rank of 0.25. Therefore, if nodes B,C,D link to node A and there are no other links, PR(A) = 0.15 + 0.85*(0.25 + 0.25 + 0.25) = 0.7875
and PR(B) = PR(C) = PR(D) = 0.15
. But the sum of page ranks is not equal to 1 which is the sum of ranks at the inital step. Am I wrong for this calculation?
I've read another tutorial, and in their calculation, the sum of ranks is always the same. Can anyone explain me where I'm wrong? Many thanks.
Upvotes: 1
Views: 1730
Reputation: 18576
The formula that keeps the sum equal to 1 is (1 - d) / N + d * (...)
, not the one in your post (it's okay that the sum is not one with your formula). The Wikipedia article is clear about it:
The damping factor adjusts the derived value downward. The original paper, however, gave the following formula, which has led to some confusion:
... (your formula goes here)
The difference between them is that the PageRank values in the first formula sum to one, while in the second formula each PageRank is multiplied by N and the sum becomes N.
Upvotes: 1