You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
English wordnet is a fork of princeton wordnet that is developped in opensource, as such it is much more actively maintained, hence it is both more correct and more complete.
NLTK integrating english wordnet corpus as an option, would make it more discoverable and accessible.
The complexity seems low, as if wordnet english can be exported to the same format as princeton wordnet then nltk would support wordnet english out of the box?
If so what format does NLTK expect?
The text was updated successfully, but these errors were encountered:
Alongside that open-source english Wordnet, @goodmami is working on wn, which may also be worth checking out as an alternative to NLTK's Wordnet functionality. As far as I can tell, there's some good work being done there.
However, to put this all in a broader context, the NLTK team has been looking to deprecate the existing NLTK Wordnet corpus reader and opt for a separate one for quite some while now (e.g. #2423). That said, I don't believe there is consensus on how to move forward with this deprecation fully. Wordnet is one of the big use cases of NLTK, so we need to be careful here.
In short, this is a difficult discussion, but I imagine we would eventually move to deprecate the NLTK Wordnet in favor of wn, if the latter can be maintained and improved more effectively. Or perhaps we would just leave the choice with our users. There is no consensus on this yet, I believe.
Thank you for this great answer :). While NLTK and wn supports english wordnet, english wordnet has 2 upstream issues (including one regarding sense ordering, as pointed by ekaf) that are not yet fixed so in the meantime I will stick with NLTK classic wordnet 3.0/3.1.
English wordnet is a fork of princeton wordnet that is developped in opensource, as such it is much more actively maintained, hence it is both more correct and more complete.
NLTK integrating english wordnet corpus as an option, would make it more discoverable and accessible.
@tomaarsen friendly ping
The complexity seems low, as if wordnet english can be exported to the same format as princeton wordnet then nltk would support wordnet english out of the box?
If so what format does NLTK expect?
The text was updated successfully, but these errors were encountered: