"That's stupidly awesome" or "You're such a jerk :)"
The above is obviously positive, but how would you train a computer to figure that out? So much of our language is contextual and has subtle hints of sentiment that this is a tough problem in natural language processing.
Though there is a great algorithm called Support Vector Machines that can find a close solution! And there's a great Ruby library for you to use as well.
Join us for this talk where we'll go detecting sentiment in tweets using support vector machines. At least join us for the various bouts of swear words and confusing lexicon of the English language.