Next up: @markjrieke is Introducing work boots: generate prediction intervals from tidymodels workflows #RStudioConf2022 #RStudioConf #RStats 
Xgboost can't generate prediction intervals. workboots lets you add them (I missed the photo of his slide with intervals, he tweeted a link to his slides, I'll add that in a sec) #RStudioConf2022 #RStudioConf #RStats
workboots lets you use powerful models but also get prediction intervals via bootstrapping #RStudioConf2022 #RStudioConf #RStats
To use workboots make sure you have:
Normally distributed residuals
Time to run it (it takes longer, slow models might be too slow)
An actual need for the fancy package. #RStudioConf2022 #RStudioConf #RStats
Normally distributed residuals
Time to run it (it takes longer, slow models might be too slow)
An actual need for the fancy package. #RStudioConf2022 #RStudioConf #RStats
I recommend clicking through to the slides for the code examples. You can add ranges to predictions and also generate (ranges of) feature importance. #RStudioConf2022 #RStudioConf #RStats
• • •
Missing some Tweet in this thread? You can try to
force a refresh
