eigenfactor.org and journal-ranking.com evaluate the influence of scientific journals, newspapers, universities, and so on by applying Google's PageRank algorithm to the citations in scientific publications: Just as webpages link to other webpages with related information scientific papers link to related work by citations. More links or citations signify higher importance.
This has been known for ages and there is even a company that counts citations and condenses them into an impact factor measuring the relative importance of scientific journals.
However, so far this has been a simple one-pass process: All citations were weighted equally. There was no means to capture credibility of a source. This lead to strange results: E.g., in 2003 the highest-ranked journal was the "Annual Review of Immunlogy" with an ISI citation index of about 50. In comparison, Nature ranked 6th with an index of about 30. There was clearly something wrong.
According to eigenfactor.org the PageRank algorithm ranks Nature first, Science second, and the Journal of Biological Chemisty third. I cannot comment on the last one, but Nature and Science sound about right.
References
J. Bollen, M. A. Rodriguez, H. Van de Sompel:
Journal Status
arXiv:cs.DL/0601030, January, 2006, PDF
S. Brin, L. Page:
The anatomy of a large-scale hypertextual web search engine
Proc. of the 17th Intl. Conf. on WWW, April, 1998, PDF
www.journal-ranking.com:
Assessing the influence of Operations Research and Management Science papers - Can analysis match perceptions
White paper, February, 2007, PDF
