7.3 User Navigation: The PageRank Proxy
We have developed a web proxy application that annotates each link that a user sees with its PageRank. This is quite useful, because users receive some information about the link before they click on it. In Figure 7 is a screen shot from the proxy. The length of the red bars is the log of the URL's PageRank. We can see that major organizations, like Stanford University, receive a very high ranking followed by research groups, and then people, with professors at the high end of the people scale. Also notice ACM has a very high PageRank, but not as high as Stanford University. Interestingly, this PageRank annotated view of the page makes an incorrect URL for one of the professors glaringly obvious since the professor has an embarrassingly low PageRank. Consequently this tool seems useful for authoring pages as well as navigation. This proxy is very helpful for looking at the results from other search engines, and pages with large numbers of links such as Yahoo's listings. The proxy can help users decide which links in a long listing are more likely to be interesting. Or, if the user has some idea where the link they are looking for should fall in the "importance" spectrum, they should be able to scan for it much more quickly using the proxy.
7.4 Other Uses of PageRank
The original goal of PageRank was a way to sort backlinks so if there were a large number of backlinks for a document, the "best" backlinks could be displayed first. We have implemented such a system. It turns out this view of the backlinks ordered by PageRank can be very interesting when trying to understand your competition. For example, the people who run a news site always want to keep track of any significant backlinks the competition has managed to get. Also, PageRank can help the user decide if a site is trustworthy or not. For example, a user might be inclined to trust information that is directly cited from the Stanford homepage.
8 Conclusion
In this paper, we have taken on the audacious task of condensing every page on the World Wide Web into a single number, its PageRank. PageRank is a global ranking of all web pages, regardless of their content, based solely on their location in the Web's graph structure.
Using PageRank, we are able to order search results so that more important and central Web pages are given preference. In experiments, this turns out to provide higher quality search results to users. Th intuition behind PageRank is that it uses information which is external to the Web pages themselves - their backlinks, which provide a kind of peer review. Furthermore, backlinks from "important" pages are more significant than backlinks from average pages. This is encompassed in the recursive definition of PageRank (Section 2.4).
PageRank could be used to separate out a small set of commonly used documents which can answer most queries. The full database only needs to be consulted when the small database is not adequate to answer a query. Finally, PageRank may be a good way to help find representative pages to display for a cluster center.
We have found a number of applications for PageRank in addition to search which include traffic estimation, and user navigation. Also, we can generate personalized PageRanks which can create a view of Web from a particular perspective.
Overall, our experiments with PageRank suggest that the structure of the Web graph is very useful for a variety of information retrieval tasks.