-
Notifications
You must be signed in to change notification settings - Fork 1
Open
Description
When comparing the analyzed syntax with a state of the art model such as google nlp api, the analysis is not quite correct. Take for instance a sentence such as the house had thick and wide stone walls
On the demo page (and using the nodejs module), the syntax outputs the wide and stone as a MWE, part of the walls. Which is incorrect.
Using google nlp, the syntax output for the same sentence is wide is conj and stone is nn, which is correct.
the above mentioned result can be verified here: https://cloud.google.com/natural-language/
Metadata
Metadata
Assignees
Labels
No labels