Wednesday, May 19, 2010

Next Up: An Irony Meter...

One tricky part about internet (or any written) communication is the lack of visual and audio clues that help impart intent or meaning. Sarcasm, in particular, is difficult to convey. Israeli researchers have tried to tackle that problem and have unveiled SASI: a Semi-supervised Algorithm for Sarcasm Identification.
SASI, a Semi-supervised Algorithm for Sarcasm Identification, can recognize sarcastic sentences in product reviews online with pretty astounding 77 percent precision. To create such an algorithm, the team scanned 66,000 product reviews, with three different human annotators tagging sentences for sarcasm. The team then identified certain sarcastic patterns that emerged in the reviews and created a classification algorithm that puts each statement into a sarcastic class.
77% is better than some people I know!

[via Omni Brain]


rob said...

Amazon reviews are the perfect test usage for this algorithm. However, I hope that this algorithm will not be used to remove sarcastic reviews from amazon, some of them are very entertaining.