The PageRank algorithm used by Google
expresses the ranking of a web page in terms of two components:
- a base value, (1 - d), usually 0.15
- d * ∑i ∈ links PRi / ni where PRi is
the page rank of a page that links to this page, and ni
is the number of links from that page.
The PageRank values can be approximated by relaxation by using
this formula repeatedly within MapReduce.
Each page is initially given a PageRank of 1.0; the sum of all
values will always equal the number of pages.
- Map: Share the love: each page distributes its PageRank
equally across the pages it links to.
- Reduce: Each page sums the incoming values, multiplies
by 0.85, and adds 0.15 .