issues with negation of verbs
paride92 opened this issue · 3 comments
After running th function get_narratives it recognize 'B-ARGM-NEG_highdim', 'B-ARGM-NEG_lowdim' as true even if there is not a negation in the original sentence. As a consequence when inspect the narratives this is the result.
Original raw tweet:
police attacking kids demonstrating ddlzan
High-dimensional narrative:
police not_attack kid
Low-dimensional narrative:
country not_attack child
Clearly the the narrative should be something like "police attack kid". Do you have any suggestion on how to solve the issue? it happen quite often, sometimes I obtain the "not_" appendix even though there is no verb
Original raw tweet:
After not approving the DDLZan, the Italian political class confirms that Italy has a long way to go, that we are too far behind culturally and socially
High-dimensional narrative:
italy not_ long way far culturally socially
Low-dimensional narrative:
italy not_ text
There was an error in the tutorial regarding negations. I fixed it today. Does this solve your problem?
Now it seems to work fine! Thanks. Btw, I noticed that now I can run the script only with python 3.7, while with python 3.8, it crashes just after loading the word embeddings. It is not a problem for me, but I just want to signal the issue since in the repository it is written that both 3.7 and 3.8 can be used.
Thanks for the info. Not sure why, tox is meant to test the code on Python 3.7, 3.8, and 3.9 before every push.