Skip to content

Adds documentation about racing metrics #128

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
+56 −118
Open

Adds documentation about racing metrics #128

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0a5df45
Fixes #113
frans-ixpantia Sep 19, 2025
e9474b8
updated tests (due to RNG differences)
topepo Sep 19, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion R/tune_race_anova.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,8 @@
#' data frame should have columns for each parameter being tuned and rows for
#' tuning parameter candidates. An integer denotes the number of candidate
#' parameter sets to be created automatically.
#' @param metrics A [yardstick::metric_set()] or `NULL`.
#' @param metrics A [yardstick::metric_set()] or `NULL`. The first metric in
#' metrics is the one that will be optimized.
#' @param eval_time A numeric vector of time points where dynamic event time
#' metrics should be computed (e.g. the time-dependent ROC curve, etc). The
#' values must be non-negative and should probably be no greater than the
Expand Down
3 changes: 2 additions & 1 deletion R/tune_race_win_loss.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,8 +51,9 @@
#' statistical analysis is updated. More candidate parameters may be excluded
#' with each new resample that is processed.
#'
#' The [control_race()] function contains are parameter for the significance cutoff
#' The [control_race()] function contains parameters for the significance cutoff
#' applied to the Bradley-Terry model results as well as other relevant arguments.
#' The first metric in the `metrics` argument is the one that will be optimized.
#'
#' ## Censored regression models
#'
Expand Down
32 changes: 20 additions & 12 deletions tests/testthat/_snaps/sa-misc.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,9 +6,9 @@
Message
Optimizing roc_auc
Initial best: 0.85731
1 ( ) accept suboptimal roc_auc=0.85682 (+/-0.01022)
2 ( ) accept suboptimal roc_auc=0.85238 (+/-0.01078)
3 ( ) accept suboptimal roc_auc=0.85138 (+/-0.0109)
1 ( ) accept suboptimal roc_auc=0.85432 (+/-0.01039)
2 <3 new best roc_auc=0.86674 (+/-0.009097)
3 <3 new best roc_auc=0.87245 (+/-0.008363)

---

Expand All @@ -18,9 +18,9 @@
Message
Optimizing roc_auc
Initial best: 0.85325
1 ( ) accept suboptimal roc_auc=0.85313 (+/-0.0106)
2 ( ) accept suboptimal roc_auc=0.85181 (+/-0.01065)
3 ( ) accept suboptimal roc_auc=0.85165 (+/-0.01055)
1 ( ) accept suboptimal roc_auc=0.85268 (+/-0.01056)
2 <3 new best roc_auc=0.85618 (+/-0.01015)
3 <3 new best roc_auc=0.85799 (+/-0.009991)

---

Expand All @@ -30,9 +30,9 @@
Message
Optimizing roc_auc
Initial best: 0.86616
1 ( ) accept suboptimal roc_auc=0.86399 (+/-0.01081)
2 <3 new best roc_auc=0.86768 (+/-0.009563)
3 <3 new best roc_auc=0.87329 (+/-0.008273)
1 ( ) accept suboptimal roc_auc=0.86568 (+/-0.01027)
2 <3 new best roc_auc=0.86654 (+/-0.009714)
3 <3 new best roc_auc=0.8688 (+/-0.009487)

---

Expand All @@ -42,7 +42,15 @@
Message
Optimizing roc_auc
Initial best: 0.86616
1 ( ) accept suboptimal roc_auc=0.86399 (+/-0.01081)
2 <3 new best roc_auc=0.86768 (+/-0.009563)
3 <3 new best roc_auc=0.87329 (+/-0.008273)
1 ( ) accept suboptimal roc_auc=0.86568 (+/-0.01027)
2 <3 new best roc_auc=0.86654 (+/-0.009714)
3 <3 new best roc_auc=0.8688 (+/-0.009487)

# tune_sim_anneal with wrong type

Code
tune_sim_anneal(1)
Condition
Error in `tune_sim_anneal()`:
! The first argument to `tune_sim_anneal()` should be either a model or workflow.

134 changes: 31 additions & 103 deletions tests/testthat/_snaps/sa-overall.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,79 +12,43 @@
Optimizing roc_auc
Initial best: 0.73008
i Fold1, Repeat1: preprocessor 1/1
v Fold1, Repeat1: preprocessor 1/1
i Fold1, Repeat1: preprocessor 1/1, model 1/1
v Fold1, Repeat1: preprocessor 1/1, model 1/1
i Fold1, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold1, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold2, Repeat1: preprocessor 1/1
v Fold2, Repeat1: preprocessor 1/1
i Fold2, Repeat1: preprocessor 1/1, model 1/1
v Fold2, Repeat1: preprocessor 1/1, model 1/1
i Fold2, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold2, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold3, Repeat1: preprocessor 1/1
v Fold3, Repeat1: preprocessor 1/1
i Fold3, Repeat1: preprocessor 1/1, model 1/1
v Fold3, Repeat1: preprocessor 1/1, model 1/1
i Fold3, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold3, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold1, Repeat2: preprocessor 1/1
v Fold1, Repeat2: preprocessor 1/1
i Fold1, Repeat2: preprocessor 1/1, model 1/1
v Fold1, Repeat2: preprocessor 1/1, model 1/1
i Fold1, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold1, Repeat2: preprocessor 1/1, model 1/1 (predictions)
i Fold2, Repeat2: preprocessor 1/1
v Fold2, Repeat2: preprocessor 1/1
i Fold2, Repeat2: preprocessor 1/1, model 1/1
v Fold2, Repeat2: preprocessor 1/1, model 1/1
i Fold2, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold2, Repeat2: preprocessor 1/1, model 1/1 (predictions)
i Fold3, Repeat2: preprocessor 1/1
v Fold3, Repeat2: preprocessor 1/1
i Fold3, Repeat2: preprocessor 1/1, model 1/1
v Fold3, Repeat2: preprocessor 1/1, model 1/1
i Fold3, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold3, Repeat2: preprocessor 1/1, model 1/1 (predictions)
1 ( ) accept suboptimal roc_auc=0.72145 (+/-0.003605)
1 <3 new best roc_auc=0.74036 (+/-0.002433)
i Fold1, Repeat1: preprocessor 1/1
v Fold1, Repeat1: preprocessor 1/1
i Fold1, Repeat1: preprocessor 1/1, model 1/1
v Fold1, Repeat1: preprocessor 1/1, model 1/1
i Fold1, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold1, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold2, Repeat1: preprocessor 1/1
v Fold2, Repeat1: preprocessor 1/1
i Fold2, Repeat1: preprocessor 1/1, model 1/1
v Fold2, Repeat1: preprocessor 1/1, model 1/1
i Fold2, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold2, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold3, Repeat1: preprocessor 1/1
v Fold3, Repeat1: preprocessor 1/1
i Fold3, Repeat1: preprocessor 1/1, model 1/1
v Fold3, Repeat1: preprocessor 1/1, model 1/1
i Fold3, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold3, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold1, Repeat2: preprocessor 1/1
v Fold1, Repeat2: preprocessor 1/1
i Fold1, Repeat2: preprocessor 1/1, model 1/1
v Fold1, Repeat2: preprocessor 1/1, model 1/1
i Fold1, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold1, Repeat2: preprocessor 1/1, model 1/1 (predictions)
i Fold2, Repeat2: preprocessor 1/1
v Fold2, Repeat2: preprocessor 1/1
i Fold2, Repeat2: preprocessor 1/1, model 1/1
v Fold2, Repeat2: preprocessor 1/1, model 1/1
i Fold2, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold2, Repeat2: preprocessor 1/1, model 1/1 (predictions)
i Fold3, Repeat2: preprocessor 1/1
v Fold3, Repeat2: preprocessor 1/1
i Fold3, Repeat2: preprocessor 1/1, model 1/1
v Fold3, Repeat2: preprocessor 1/1, model 1/1
i Fold3, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold3, Repeat2: preprocessor 1/1, model 1/1 (predictions)
2 <3 new best roc_auc=0.73173 (+/-0.003018)
2 ( ) accept suboptimal roc_auc=0.73543 (+/-0.009624)

# variable interface

Expand All @@ -100,79 +64,43 @@
Optimizing roc_auc
Initial best: 0.73008
i Fold1, Repeat1: preprocessor 1/1
v Fold1, Repeat1: preprocessor 1/1
i Fold1, Repeat1: preprocessor 1/1, model 1/1
v Fold1, Repeat1: preprocessor 1/1, model 1/1
i Fold1, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold1, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold2, Repeat1: preprocessor 1/1
v Fold2, Repeat1: preprocessor 1/1
i Fold2, Repeat1: preprocessor 1/1, model 1/1
v Fold2, Repeat1: preprocessor 1/1, model 1/1
i Fold2, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold2, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold3, Repeat1: preprocessor 1/1
v Fold3, Repeat1: preprocessor 1/1
i Fold3, Repeat1: preprocessor 1/1, model 1/1
v Fold3, Repeat1: preprocessor 1/1, model 1/1
i Fold3, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold3, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold1, Repeat2: preprocessor 1/1
v Fold1, Repeat2: preprocessor 1/1
i Fold1, Repeat2: preprocessor 1/1, model 1/1
v Fold1, Repeat2: preprocessor 1/1, model 1/1
i Fold1, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold1, Repeat2: preprocessor 1/1, model 1/1 (predictions)
i Fold2, Repeat2: preprocessor 1/1
v Fold2, Repeat2: preprocessor 1/1
i Fold2, Repeat2: preprocessor 1/1, model 1/1
v Fold2, Repeat2: preprocessor 1/1, model 1/1
i Fold2, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold2, Repeat2: preprocessor 1/1, model 1/1 (predictions)
i Fold3, Repeat2: preprocessor 1/1
v Fold3, Repeat2: preprocessor 1/1
i Fold3, Repeat2: preprocessor 1/1, model 1/1
v Fold3, Repeat2: preprocessor 1/1, model 1/1
i Fold3, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold3, Repeat2: preprocessor 1/1, model 1/1 (predictions)
1 ( ) accept suboptimal roc_auc=0.72145 (+/-0.003605)
1 <3 new best roc_auc=0.74036 (+/-0.002433)
i Fold1, Repeat1: preprocessor 1/1
v Fold1, Repeat1: preprocessor 1/1
i Fold1, Repeat1: preprocessor 1/1, model 1/1
v Fold1, Repeat1: preprocessor 1/1, model 1/1
i Fold1, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold1, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold2, Repeat1: preprocessor 1/1
v Fold2, Repeat1: preprocessor 1/1
i Fold2, Repeat1: preprocessor 1/1, model 1/1
v Fold2, Repeat1: preprocessor 1/1, model 1/1
i Fold2, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold2, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold3, Repeat1: preprocessor 1/1
v Fold3, Repeat1: preprocessor 1/1
i Fold3, Repeat1: preprocessor 1/1, model 1/1
v Fold3, Repeat1: preprocessor 1/1, model 1/1
i Fold3, Repeat1: preprocessor 1/1, model 1/1 (extracts)
i Fold3, Repeat1: preprocessor 1/1, model 1/1 (predictions)
i Fold1, Repeat2: preprocessor 1/1
v Fold1, Repeat2: preprocessor 1/1
i Fold1, Repeat2: preprocessor 1/1, model 1/1
v Fold1, Repeat2: preprocessor 1/1, model 1/1
i Fold1, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold1, Repeat2: preprocessor 1/1, model 1/1 (predictions)
i Fold2, Repeat2: preprocessor 1/1
v Fold2, Repeat2: preprocessor 1/1
i Fold2, Repeat2: preprocessor 1/1, model 1/1
v Fold2, Repeat2: preprocessor 1/1, model 1/1
i Fold2, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold2, Repeat2: preprocessor 1/1, model 1/1 (predictions)
i Fold3, Repeat2: preprocessor 1/1
v Fold3, Repeat2: preprocessor 1/1
i Fold3, Repeat2: preprocessor 1/1, model 1/1
v Fold3, Repeat2: preprocessor 1/1, model 1/1
i Fold3, Repeat2: preprocessor 1/1, model 1/1 (extracts)
i Fold3, Repeat2: preprocessor 1/1, model 1/1 (predictions)
2 <3 new best roc_auc=0.73173 (+/-0.003018)
2 ( ) accept suboptimal roc_auc=0.73543 (+/-0.009624)

---

Expand All @@ -183,9 +111,9 @@
Message
There were 2 previous iterations
Optimizing roc_auc
2 v initial roc_auc=0.73173 (+/-0.003018)
3 <3 new best roc_auc=0.74172 (+/-0.008775)
4 <3 new best roc_auc=0.7909 (+/-0.009887)
2 v initial roc_auc=0.74036 (+/-0.002433)
3 <3 new best roc_auc=0.75103 (+/-0.009282)
4 <3 new best roc_auc=0.76617 (+/-0.01037)

---

Expand All @@ -205,39 +133,39 @@
set.seed(40)
rf_res_finetune <- tune_sim_anneal(wf_rf, resamples = bt, initial = rf_res)
Message
i Creating pre-processing data to finalize unknown parameter: mtry
i Creating pre-processing data to finalize 1 unknown parameter: "mtry"
Optimizing roc_auc
Initial best: 0.84994
1 ( ) accept suboptimal roc_auc=0.84375 (+/-0.007727)
2 + better suboptimal roc_auc=0.84943 (+/-0.007036)
3 ( ) accept suboptimal roc_auc=0.84371 (+/-0.007903)
4 + better suboptimal roc_auc=0.84825 (+/-0.008036)
5 ( ) accept suboptimal roc_auc=0.84479 (+/-0.00814)
6 + better suboptimal roc_auc=0.84816 (+/-0.007283)
7 ( ) accept suboptimal roc_auc=0.84381 (+/-0.007999)
8 <3 new best roc_auc=0.85014 (+/-0.007172)
9 ( ) accept suboptimal roc_auc=0.84344 (+/-0.007818)
10 + better suboptimal roc_auc=0.84802 (+/-0.007281)
Initial best: 0.84917
1 ( ) accept suboptimal roc_auc=0.84467 (+/-0.007843)
2 <3 new best roc_auc=0.84956 (+/-0.007013)
3 ( ) accept suboptimal roc_auc=0.84384 (+/-0.008042)
4 + better suboptimal roc_auc=0.84952 (+/-0.00757)
5 ( ) accept suboptimal roc_auc=0.84357 (+/-0.008148)
6 + better suboptimal roc_auc=0.84848 (+/-0.007499)
7 ( ) accept suboptimal roc_auc=0.84391 (+/-0.008097)
8 + better suboptimal roc_auc=0.84853 (+/-0.007036)
9 ( ) accept suboptimal roc_auc=0.84471 (+/-0.007812)
10 x restart from best roc_auc=0.84935 (+/-0.007313)

---

Code
set.seed(40)
rf_res_finetune <- tune_sim_anneal(wf_rf, resamples = bt)
Message
i Creating pre-processing data to finalize unknown parameter: mtry
i Creating pre-processing data to finalize 1 unknown parameter: "mtry"
Optimizing roc_auc
Initial best: 0.84418
1 <3 new best roc_auc=0.84839 (+/-0.007753)
2 ( ) accept suboptimal roc_auc=0.84384 (+/-0.008085)
3 <3 new best roc_auc=0.84857 (+/-0.007615)
4 ( ) accept suboptimal roc_auc=0.8435 (+/-0.007746)
5 + better suboptimal roc_auc=0.84804 (+/-0.00774)
6 ( ) accept suboptimal roc_auc=0.84338 (+/-0.007515)
7 <3 new best roc_auc=0.84923 (+/-0.007371)
8 ( ) accept suboptimal roc_auc=0.84389 (+/-0.007938)
9 <3 new best roc_auc=0.84926 (+/-0.007163)
10 ( ) accept suboptimal roc_auc=0.84397 (+/-0.00741)
Initial best: 0.84369
1 <3 new best roc_auc=0.84929 (+/-0.007889)
2 ( ) accept suboptimal roc_auc=0.84461 (+/-0.008084)
3 + better suboptimal roc_auc=0.84785 (+/-0.007989)
4 ( ) accept suboptimal roc_auc=0.84325 (+/-0.007817)
5 + better suboptimal roc_auc=0.84917 (+/-0.007376)
6 ( ) accept suboptimal roc_auc=0.84458 (+/-0.007566)
7 + better suboptimal roc_auc=0.84871 (+/-0.007483)
8 ( ) accept suboptimal roc_auc=0.84431 (+/-0.007694)
9 x restart from best roc_auc=0.84861 (+/-0.007929)
10 ( ) accept suboptimal roc_auc=0.84522 (+/-0.008113)

# incompatible parameter objects

Expand Down
2 changes: 1 addition & 1 deletion tests/testthat/test-win-loss-overall.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@ test_that("one player is really bad", {
add_formula(y ~ .) |>
add_model(rf_spec)

grid <- tibble(min_n = c(1, 40))
grid <- tibble(min_n = c(1, 100))
ctrl <- control_race(burn_in = 2, alpha = .05, randomize = TRUE)
set.seed(3355)
tuning_results <- tune_race_win_loss(
Expand Down
Loading