workflows

workflows copied to clipboard

first pass at postprocessing proof-of-concept

4

Mostly just pattern-matches `recipe` and `model_fit` implementations and hooks into the existing post-processing infrastructure. See tests for up-to-date examples. Previous PR description, outdated A fast and loose proof-of-concept for integrating...

simonpcouch

[This issue](https://forum.posit.co/t/predictors-dropped-with-a-recipe-step/184469) describes how, once the data are given to a workflow, prediction always requires it as-is. This will be increasingly important for supervised feature selection. Look at my initial...

topepo

For external packages and upcoming feature selection APIs, we'll need a convenient way to determine the predictor column names in the data given to `fit.workflow()`. This is a standalone file...

topepo

I was expecting to be able to fit a workflow with loading only {workflows} but while the `fit.workflow()` method is exported, the generic is not. ``` r library(workflows) library(magrittr) workflow()...

hfrick

I believe those in https://github.com/tidymodels/workflows/blob/1413997d256cb0e4206211f5875615434317be85/vignettes/extras/getting-started.Rmd and the hex logo are the only images needing it. See https://posit.co/blog/knitr-fig-alt/ for examples.

simonpcouch

`use_standalone("r-lib/rlang", "types-check")` instead of home grown argument checkers

simonpcouch

axe_data does not remove pre$mold$extras

3

The axe_data.workflow() function does not remove the mold$extras like it does for outcomes and predictors. If a column of data is set to be another role, the model will retain...

era127

Workflows which use case weights calculated in a recipe from non-predictor columns error when attempting to `predict()`.

1

## The problem Workflows which use dynamic case weights (calculated from non-predictor columns) error when attempting to `predict()`. This is following a string of issues, most recently https://github.com/tidymodels/hardhat/issues/240 , and...

mikemahoney218

PR 2/3 to address tidymodels/tune#652. Follows up on tidymodels/parsnip#955.

simonpcouch

Custom probability thresholds for binary classification ?

2

Dear Authors: I am aware that there is a [plan](https://community.rstudio.com/t/predict-a-class-using-a-threshold-different-than-the-0-5-default-with-tidymodels/56273) to add the feature of custom probability threshold for classification in workflow package. But it seems not published yet. I...

yangxg