Improving Classification of Tweets Using Linguistic Information from a Large External Corpus

The bag of words representation of documents is often unsatisfactory as it ignores relationships between important terms that do not co-occur literally. Improvements might be achieved by expanding the vocabulary with other relevant word, like synonyms. In this paper we use word-word co-occurence information from a large corpus to expand the vocabulary of another corpus consisting of tweets. Several different methods on how to include the co-occurence information are constructed and tested out on the classification of real twitter data. Our results show that we are able to reduce the number of erroneous classifications by 14% using co-occurence information.