Improving classification of tweets using word-word co-occurrence information from a large external corpus

Abstract

Classifying tweets is an intrinsically hard task as tweets are\nshort messages which makes traditional bags of words based\napproach ine cient. In fact, bags of words approaches ig-\nnores relationships between important terms that do not\nco-occur literally.\nIn this paper we resort to word-word co-occurence informa-\ntion from a large corpus to expand the vocabulary of another\ncorpus consisting of tweets. Our results show that we are\nable to reduce the number of erroneous classi cations by\n14% using co-occurence information.