Pagerank Algorithm Revealed?
Posted by Jon Lee in Web Development, tags: google, pagerank, statistics
I attended a talk at the local University discussing job opportunities at Google in the field of statistics. There were surprisingly many positions available that required at least a masters in statistics, some of which I might try to apply for!
Anyway, the speaker gave an overview of how Google’s advertising model worked and how advanced statistical analysis plays a key role in determining how to order both search and ad results.
Also, he spent a good 10 minutes discussing the PageRank algorithm and how it worked. On one of his slides, there was a short little formula defining the PageRank of a particular page.
I thought I had struck a gold mine and quickly scribbled it down, thinking I had discovered the holy grail of the Internet — Google’s PageRank algorithm. Of course, I was being naive and I quickly realized this obviously isn’t the actual algorithm. In fact, this was nothing more than what can be found on the Wikipedia entry.
What is interesting to note however, is that the PageRank is nothing more than the eigenvalues of an adjacency matrix. Essentially, what this means is that it is something that can be calculated very quickly without heavy computational resources. I suspect the algorithm can be made more complicated but the increase in computational power would be too great for even Google to handle. Considering the billions and billions of pages on the Internet, even an algorithm that is marginally slower can make a big difference.
That’s just my two cents on Google’s PageRank. Sorry if I misled you into thinking you would find the actual PageRank algorithm here! I doubt the algorithm in its entirety would fit comfortably onto a single post anyway. 
Popularity: 4% [?]
Entries (RSS)