Great post from CitizenNet on a new patent we were just granted:
CitizenNet has always invested heavily in the core technologies that comprise our platform; we always pursue what is best for our customers based on a solid foundation of technology, not merely sales.
One area of focus for us is natural language processing specific to social; simply applying old NLP techniques to social doesn’t work — there’s too much slang, too little text, and a Tweet rarely resembles a long form document. To that end, short social text can be very ambiguous. Take two messages:
“I am going to no doubt — love them!” vs “I am going to no doubt”
If you were searching for messages about the band No Doubt, which one would you include?
Read more. Link.