Coreference, Gazetteers and More
- There are a lot of other things that can be done in GATE and
other TE systems.
- Coreference involves multiple phrases refering to the same
object. E.g. I went to the store and it was closed.
- It's more than pronominal reference, but you need it to
bind documents together. The best Coref results are in the 90s.
- Gazetteers (or Gazateers for supporters of Paul Gascoine) are
also really important. They're lists of proper nouns. As
proper nouns, like Santa Claus, are usually relatively solid,
these lists can be used to give some solidity to the TE.
- The lists are of course extracted into the tag set.
- We can't parse well, so in most TE systems, full discourse analysis is
out. However, things can be combined across sentences.