I highlight an experiment by the Annals of Improbable Research involving mailing various unusual items. While two-thirds reached their destination, I doubt the Indian postal service would be as successful with such a test.
I'm highlighting the upcoming Ig Nobel Prizes from the Annals of Improbable Research. These awards honor science that makes you laugh and then think. Don't miss the bizarre guide to Britney Spears and semiconductor physics.
I analyzed Calvin and Hobbes to find the top 20 statistically improbable phrases like "chocolate frosted sugar bombs." I highlight the technical challenges of tokenization, such as handling contractions, and the difficulty of weighting longer, naturally rarer phrases.
I updated my Statistically Improbable Phrases application on Google AppEngine to include interactive stopword filtering, dynamic frequency adjustments, and visual controls for font size and contrast to help users better identify unique subjects in text.
I transcribed Calvin and Hobbes and compared its word frequency against a large English corpus to automatically identify characteristic themes. This 'statistically improbable phrases' method highlights unique recurring concepts like Spaceman Spiff, Transmogrifier, and tuna.
I ported my Perl script to Python to launch a Statistically Improbable Phrases tool on Google AppEngine. The app analyzes URL text against a custom corpus, generating tag clouds where font size and color represent word frequency and improbability.