search

mkearney / textfeatures

👷‍♂️ A simple package for extracting useful features from character objects 👷‍♀️
https://textfeatures.mikewk.com
Other
166 stars 17 forks source link

word2vec vs LDA topic modeling? #11

Open juliasilge opened 4 years ago

juliasilge commented 4 years ago

Thank you so much for your excellent work on this package! 🙌

I have some questions about the way the w results are presented/documented. The function documentation uses some language about word2vec, but in the functions that create these results, only functions from text2vec for topic modeling are called, not any vector embedding functions.

The text2vec package has functions for LDA topic modeling and also for word embeddings, specifically an implementation of GloVe. If I am understanding correctly, it looks like only the first are used in this package, but the language describing the results refers to the second.

What are your thoughts on this? Do you think the language and naming should be updated?

juliasilge commented 4 years ago

cc @EmilHvitfeldt because this will impact textrecipes