Skip to content

Suitable for both Eastern and Western Armenian? #1

@AngledLuffa

Description

@AngledLuffa

I'm working on preparing models for the next release of Stanford's Stanza python software, https://stanfordnlp.github.io/stanza

The next release of Universal Dependencies distinguishes between Eastern and Western Armenian. Is it suitable to use the word vectors you host for both dialects?

https://github.com/UniversalDependencies/UD_Armenian-ArmTDP/
https://github.com/UniversalDependencies/UD_Western_Armenian-ArmTDP/

If I count the words from the UD datasets that are present in the Glove 200 file, for example, 97% of the words in the Eastern Armenian dataset appear here, and only 88% of the words in the Western Armenian dataset appear in the word vectors. This makes me think there's a bit of an issue with the coverage of these word vectors. If these vectors are not ideal, do you have any recommendations for others that would be, or do you have any intention of adding more Western Armenian words to the dataset?

Thanks!

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions