FYI. While you are creating this package. Might be also interesting to consider an interface to the Starspace wrapper (https://CRAN.R-project.org/package=ruimtehol). It's particularly nice in multi-label classification which other models lack a bit.
The vignette documentation at https://cran.r-project.org/web/packages/ruimtehol/vignettes/ground-control-to-ruimtehol.pdf shows an example of classifying questions in Belgian Parliament on more than 1400 labels.
Only downside is that the there is a package failure due to multithreading on Mac OS on cran.
FYI. While you are creating this package. Might be also interesting to consider an interface to the Starspace wrapper (https://CRAN.R-project.org/package=ruimtehol). It's particularly nice in multi-label classification which other models lack a bit. The vignette documentation at https://cran.r-project.org/web/packages/ruimtehol/vignettes/ground-control-to-ruimtehol.pdf shows an example of classifying questions in Belgian Parliament on more than 1400 labels. Only downside is that the there is a package failure due to multithreading on Mac OS on cran.
We've had a thread here on evaluating different models for text classification purposes: https://github.com/bnosac/ruimtehol/issues/11