Kenneth Benoit

icon indicating a website link http://kenbenoit.net icon indicating an email [email protected]

icon company London School of Economics and Political Science icon location London icon location Professor of Computational Social Science; Director, LSE Data Science Institute: Director of @quanteda Initiative

Results 308 comments of Kenneth Benoit

Issue 2329 2

And: Should we increment this now to v4.1.0?

Issue 2329 2

@koheiw can you resolve the conflict in `src/tokens_xptr.cpp` and update with the latest `master`?

Issue 2329 2

@koheiw could you pls resolve the conflict in `tokens_compound.R`? I think I know the right code section but wanted to be sure.

Always remove paddings in `dfm()`

So we would not be able to count paddings when creating a dfm? That seems like a loss. What if we want to know the total removals by document?

Add featfreq.tokens() and docfreq.tokens()

I think we discussed this before, and didn't implement them simply due to naming: tokens objects don't have "features". But we could define tokens methods for these, I'm fine with...

Adding document-vector matrix?

That could work well, especially with corpus_reshape, since the vector representation could then apply to any reshaped subset of a document that people wanted to define. Would we need to...

Headings are not separated from sentences

I remember this being one of the earliest problems I encountered as a **quanteda** use case, over 10 years ago. That's why the pattern exists in `corpus_segment()`, because I was...

Consider moving append_key to tokens_annotate()

Makes sense, but we should deprecate it rather than delete it, for a version or three.