Good reference of putting all tidymodels packages together in a rather complicated ML project


I'm looking for another reference guide other than this article A gentle intro to tidymodels which would explain in more detail the process of using rsample, recipes, dials and parsnip for a fairly complicated ML project where I'd like to, e.g.: apply a number of preprocessing steps with recipes on a nested, re-sampled dataset with rsample in order to hypertune and find the best model parameters with dials and eventually fit the best model with parsnip.

Is there any guideline that would discuss this A to Z? There's a lot of good articles on individual packages websites but I've already read them all and they never really show the full picture. I'd like already to start using all of these packages together instead of reverting back to caret to do the final modelling piece.

For discussions related to modeling, machine learning and deep learning. Related packages include caret, modelr, yardstick, rsample, parsnip, tensorflow, keras, cloudml, and tfestimators.


Not sure how useful these are but i have seen this which looks pretty good....

1 Like

Something like this is on our to-do list but we probably won't get to it until the end of the year. We have a few more core packages to release before then.

If you want to give more details (here or over email), I'd be happy to make suggestions.

1 Like

Hey! Thanks for coming back to me. I took a look at that great blogpost that @john.smith provided and started applying the overall framework on our real case example. However, I started coming across this issue when using the glmnet backend.

I was able to reproduce the problem on the iris dataset and posted and issue on the parsnip github page:

It would be great if you could take a look at it and let me know whether I'm missing something or if I found some type of bug.


This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.