The MS social network and brand monitoring application LookingGlass uses algorithms to determine the correlation between keywords and keyword clusters to consumer sentiment.  Some of the linguists who worked on the original grammar checker for MS Word (back in the day) are now building and maintaining the machine learning that goes into LookingGlass.

I Love My Panasonic RP-HC500 Noise Cancelling Headphones to Death

Interesting correlation:  our expert linguists in MS Research found a very high correlation between the occurence of the word “death” and the positive sentiment of a review in certain verticals. Sure seemed like either a bug or a life-imitating-art reflection from Mad Men Season 1.

What our linguists discovered is that the phrase (or variations thereof) “I love it to death” occurred at a high frequency in the most positive reviews for products in certain verticals.  

“Discovering unexpected correlations are one of the great benefits of using machine learning,” said Michael Gamon, one of the key developers of the LookingGlass algorithm.

Popularity: 5% [?]