Google Books: how does it work?

Alexis Madrigal, writing in the The Atlantic (1 November 2010), outlines how the searching algorithm used for Google Books has recently been improved.  The article indicates that the algorithm is somewhat different from that used for Google’s web search and is being “tweaked” as the design team gains more experience in the kind of searching that is used to locate titles of books.  Google’s web search makes extensive use of the Page Rank process, whereby full account is taken of the number of pages linking to a site when assessing its likely relevance.  This, however, does not happen with the digitized copies of books that Google has stored, so some experimentation was necessary in order to develop a set of some one hundred predictive indicators of “title” relevance, such as how often the title is searched for on the web, recent book sales, the number of libraries that hold the title, and how often an older book has been reprinted.  The result was released as “Rich Results” and commenced on 1 November 2010.  Why not try it now?

Tags: , , ,

No comments yet.

Leave a Reply