Ray Baxter
5/24/2008 4:35:00 PM
On May 24, 2008, at 9:14 AM, Axel Etzold wrote:
> well, this is what tree-tagger (see tags output below, for the tagset
> see my previous post) says:
> England are bound to lose the match. (proper noun singular) (nobody
> is perfect).
The collective noun in American English is singular, while in British
English the collective noun is plural. In American English, we would
say "England is bound to lose the match," so your results are correct,
if the language under consideration is American English. (Although I'm
not sure what to make of the plural verb.)
> Parts-of-speech tagging uses a Bayesian decision model, requiring
> training on a set of human-tagged text.
Did you train tree-tagger on a data set of American English?
Ray