NYU: Description of the MENE Named Entity System as Used in MUC-7

Borthwick, Sterling, Agichtein, Grishman 1998

Borthwick et al describe MENE, a maxent named entity classification system. It tags one word at a time, in sequential order, with one of the following tags:

• _start
• _continue
• _end
• _unique
• other

It uses the following binary features:

• word features for words n-2...n (e.g., word[n-1]='Mr.')
• section features (e.g., section='preamble')
• wordlist features (e.g., word[n] in first_names)
• other systems (e.g., proteus.tag(n)=person_start)

Any feature that occurs atleast 3 times is used. The results are quite good (f-measure around 84%), and get even better when the other-system features are used (f-measure around 92%).

@InProceedings{borthwick1998,
  author =       {Andrew Borthwick and John Sterling and
                  Eugene Agichtein and Ralph Grishman},
  title =        {NYU: Description of the MENE Named Entity
                  System as Used in MUC-7},
  booktitle =    {Proceedings of the Seventh Message
                  Understanding Conference (MUC-7)},
  year =         1998
}