The PageRank Citation Ranking: Bringing Order to the Web
http://ilpubs.stanford.edu:8090/422/1/1999-66.pdf
This paper is extremely interesting and appears to be one of the earliest papers in which PageRank is discussed. They describe this new “importance” ranking of websites as a way to help search engines and users quickly make sense of the vast heterogeneity of the World Wide Web which is hugely important as the Web continues to grow to obscene numbers of sites and pages. They describe PageRank with high-level mathematics before testing it against NLANR’s proxy data where they compared page rankings to how often certain pages were visited. It’s a shame that their data was hard to compare because of the use of personal email that was a large portion of the proxy data. I think further analysis of that could have been very insightful. They also describe how tracking user hyperlink clicks can lead to “customized” rankings that would be more fitted to the user. This is actually pretty related to machine learning because it is using data to predict an outcome. As we know, Google has been putting a lot of their efforts into expanding the use of machine learning and this includes recommending pages that are user-specific. Another really important thind the paper mentions is that PageRanks are nearly immune to manipulation by commercial interests which is great because the goal of the internet is to free information and not to have companies guide you to what they want you to see.
This paper describes PageRank and its objective manner developed based on an inherently subjective topic. We have discussed PageRank at length in class. They also write about the ever-expanding World Wide Web and how the use of hyperlinks is a great method for determining web page importance. This is all very similar to what we have discussed about hyperlinks and the efficiency of PageRank given the huge number of websites that exist today. Of course, this is dated January 1998, so this paper discusses far fewer websites than we did in class, but the ideas are similar. If you have any thoughts on the matter, I’d love to read your comments!