Spaces:
Running
Running
NER Tendency to label consistently rather than fully accurately.
#1
by
awacke1
- opened
I notice within NER systems that I see that if a NER entity is found then NER resolution around it seems to include word influence around it - for example Script bumping twice in a row when storyboards would classify as storyboard if not preceded by a correct Script prediction. Is there a way to correct that in model?
Hi, thanks for the feedback, this model was trained for Italian language, so the performance with English language is not very good, however I will try to solve the problem and improve the performance as soon as possible.
DeepMount00
changed discussion status to
closed