Text this: A conditional random field based approach for high-accuracy part-of-speech tagging using language-independent features