1. 7
  1.  

  2. 2

    Huh. I hadn’t really put it together that Twitter would be more of an opportunity for large service providers if they could mechanically classify tweets. It makes sense that sarcasm is the biggest issue for sentiment analyzers, and I hope someone solves it soon! It would make all sorts of NLP interactions with computers a lot more pleasant if they could pick up on it.