Part of Speech Tagging

• Part of Speech (POS) tagging is assigning one or more tags to each word in a corpus, indicating that word's part of speech.
• In almost all cases, there is a single correct tag that should be assigned, so a POS-tagger's capabilities are judged by its ability to eliminate the other possible, but wrong, tags.
• A classic example:
  

  Time N V

  flies N V

  like N V P

  an D

  arrow. N

• To do POS-tagging, we need a tagset (the set of parts of speech) and a POS-tagging algorithm.
• POS-tagging is not parsing. (Why not?)