On 07/08/2013 12:15 PM, Ira Korshunova wrote:
could you please tell me which corpora you used for training location,
organization and person name finder models for English language in version
1.4.?

The models are trained on MUC 6/7 data plus some fixes and small additions to fix
recognition problems (e.g. a few sentences for Obama where added).

Jörn

Reply via email to