R yardstick gain_capture 增益捕獲

gain_capture(data, ...)

# S3 method for data.frame
gain_capture(
  data,
  truth,
  ...,
  estimator = NULL,
  na_rm = TRUE,
  event_level = yardstick_event_level(),
  case_weights = NULL
)

gain_capture_vec(
  truth,
  estimate,
  estimator = NULL,
  na_rm = TRUE,
  event_level = yardstick_event_level(),
  case_weights = NULL,
  ...
)

gain_curve() 計算完整增益曲線。

其他類概率指標：average_precision() , brier_class() , classification_cost() , mn_log_loss() , pr_auc() , roc_auc() , roc_aunp() , roc_aunu()

# ---------------------------------------------------------------------------
# Two class example

# `truth` is a 2 level factor. The first level is `"Class1"`, which is the
# "event of interest" by default in yardstick. See the Relevant Level
# section above.
data(two_class_example)

# Binary metrics using class probabilities take a factor `truth` column,
# and a single class probability column containing the probabilities of
# the event of interest. Here, since `"Class1"` is the first level of
# `"truth"`, it is the event of interest and we pass in probabilities for it.
gain_capture(two_class_example, truth, Class1)
#> # A tibble: 1 × 3
#>   .metric      .estimator .estimate
#>   <chr>        <chr>          <dbl>
#> 1 gain_capture binary         0.879

# ---------------------------------------------------------------------------
# Multiclass example

# `obs` is a 4 level factor. The first level is `"VF"`, which is the
# "event of interest" by default in yardstick. See the Relevant Level
# section above.
data(hpc_cv)

# You can use the col1:colN tidyselect syntax
library(dplyr)
hpc_cv %>%
  filter(Resample == "Fold01") %>%
  gain_capture(obs, VF:L)
#> # A tibble: 1 × 3
#>   .metric      .estimator .estimate
#>   <chr>        <chr>          <dbl>
#> 1 gain_capture macro          0.743

# Change the first level of `obs` from `"VF"` to `"M"` to alter the
# event of interest. The class probability columns should be supplied
# in the same order as the levels.
hpc_cv %>%
  filter(Resample == "Fold01") %>%
  mutate(obs = relevel(obs, "M")) %>%
  gain_capture(obs, M, VF:L)
#> # A tibble: 1 × 3
#>   .metric      .estimator .estimate
#>   <chr>        <chr>          <dbl>
#> 1 gain_capture macro          0.743

# Groups are respected
hpc_cv %>%
  group_by(Resample) %>%
  gain_capture(obs, VF:L)
#> # A tibble: 10 × 4
#>    Resample .metric      .estimator .estimate
#>    <chr>    <chr>        <chr>          <dbl>
#>  1 Fold01   gain_capture macro          0.743
#>  2 Fold02   gain_capture macro          0.727
#>  3 Fold03   gain_capture macro          0.796
#>  4 Fold04   gain_capture macro          0.748
#>  5 Fold05   gain_capture macro          0.730
#>  6 Fold06   gain_capture macro          0.754
#>  7 Fold07   gain_capture macro          0.730
#>  8 Fold08   gain_capture macro          0.747
#>  9 Fold09   gain_capture macro          0.710
#> 10 Fold10   gain_capture macro          0.731

# Weighted macro averaging
hpc_cv %>%
  group_by(Resample) %>%
  gain_capture(obs, VF:L, estimator = "macro_weighted")
#> # A tibble: 10 × 4
#>    Resample .metric      .estimator     .estimate
#>    <chr>    <chr>        <chr>              <dbl>
#>  1 Fold01   gain_capture macro_weighted     0.759
#>  2 Fold02   gain_capture macro_weighted     0.745
#>  3 Fold03   gain_capture macro_weighted     0.811
#>  4 Fold04   gain_capture macro_weighted     0.734
#>  5 Fold05   gain_capture macro_weighted     0.733
#>  6 Fold06   gain_capture macro_weighted     0.730
#>  7 Fold07   gain_capture macro_weighted     0.737
#>  8 Fold08   gain_capture macro_weighted     0.730
#>  9 Fold09   gain_capture macro_weighted     0.681
#> 10 Fold10   gain_capture macro_weighted     0.737

# Vector version
# Supply a matrix of class probabilities
fold1 <- hpc_cv %>%
  filter(Resample == "Fold01")

gain_capture_vec(
   truth = fold1$obs,
   matrix(
     c(fold1$VF, fold1$F, fold1$M, fold1$L),
     ncol = 4
   )
)
#> [1] 0.7428922

# ---------------------------------------------------------------------------
# Visualize gain_capture()

# Visually, this represents the area under the black curve, but above the
# 45 degree line, divided by the area of the shaded triangle.
library(ggplot2)
autoplot(gain_curve(two_class_example, truth, Class1))