... the Association for Computational Linguistics:shortpapers, pages 42–47,Portland, Oregon, June 19-24, 2011.c2011 Association for Computational LinguisticsPart-of -Speech Tagging for Twitter: ... 2010).than for Standard English text. For example, apos-trophes are often omitted, and there are frequentlywords like ima (short for I’m gonna) that cut acrosstraditional POS categories. Therefore, ... performance (Sch¨utze andPedersen, 1993). We used 1.9 million tokens from 134,000 unlabeled tweets to construct distributionalfeatures from the successor and predecessor proba-bilities for...