bbc / citron

Citron is an experimental quote extraction system created by BBC R&D
Apache License 2.0
25 stars 5 forks source link

False negatives #1

Open sdspieg opened 1 year ago

sdspieg commented 1 year ago

We have applied this library to a corpus of ours and have found the following issues

Any idea whether anything could,be done about this?

jcnewell commented 1 year ago

Unfortunately there is no easy solution. You could annotate the problem cases, add them to the PARC dataset and then retrain the classifiers but this might not eliminate all the errors.