This link from Wired News: Judging a Book by Its Contents takes a brief look at work being done at Amazon.com to pull statistically improbable phrases from books. These phrases, in Amazon's words: "are the most distinctive phrases in the text of books in the Search Inside! program." What's really cool is that Amazon pages provide links to other books that have similarly improbable phrases.
The Wired article mentions that this is an example of an automatically created tag - something I suspect we'll see used much more frequently. For the fun of it I've been playing with using Mac OS X.4's Spotlight feature to access infrequently used phrases to probe the content of my computer.
Comments