search

TommyJones / tidylda

Implements an algorithim for Latent Dirichlet Allocation using style conventions from the [tidyverse](https://style.tidyverse.org/) and [tidymodels](https://tidymodels.github.io/model-implementation-principles/index.html).
Other
41 stars 3 forks source link

Make gallery of example useage #28

Open TommyJones opened 4 years ago

TommyJones commented 4 years ago

Absolute necessity before any CRAN release.

Show some basic analyses:

  1. Vanilla LDA analysis + prediction
  2. Skipgram embeddings + ML analyses classification summarization embedding out-of-vocabulary words under the model
  3. In-development extensions model updates transfer learning seeding topics
TommyJones commented 3 years ago

For vanilla analyses, demonstrate tidy ways to get functionality you'd normally get from textmineR

TommyJones commented 2 years ago

Not making this a dependency to get on CRAN. It'll be a lot of work (barrier to getting it up there) and I'm ambivalent about incurring the computational costs of running vignettes. This makes running checks on textmineR time consuming. The CRAN linux distro has a "no vignettes" flag for this reason.

Vignettes are needed, but it might be best to keep them simple and then create a separate static site that's a gallery of using tidylda. That introduces risk of stale code, but also decreases the burden of updates.

TommyJones commented 2 years ago

For any future users of tidylda: if you have projects or work you want to showcase, comment below. I am considering setting up a gallery rather than vignettes (or in addition to) per the comment above. Any help with examples would be greatly appreciated.