I saw this earlier thread: Predict a class using a threshold different than the 0.5 default with tidymodels
I ran into the same question but just wanted to ask if passing in our own probability thresholds for classification was still on the todolist for the next release of workflows?
I found the probably package which could be a workaround for now: https://probably.tidymodels.org/articles/where-to-use.html
For discussions related to modeling, machine learning and deep learning. Related packages include