pr_auc()
is a metric that computes the area under the precision
recall curve. See pr_curve()
for the full curve.
pr_auc(data, ...) # S3 method for data.frame pr_auc(data, truth, ..., estimator = NULL, na_rm = TRUE) pr_auc_vec(truth, estimate, estimator = NULL, na_rm = TRUE, ...)
data  A 

...  A set of unquoted column names or one or more

truth  The column identifier for the true class results
(that is a 
estimator  One of 
na_rm  A 
estimate  If 
A tibble
with columns .metric
, .estimator
,
and .estimate
and 1 row of values.
For grouped data frames, the number of rows returned will be the same as the number of groups.
For pr_auc_vec()
, a single numeric
value (or NA
).
Macro and macroweighted averaging is available for this metric.
The default is to select macro averaging if a truth
factor with more
than 2 levels is provided. Otherwise, a standard binary calculation is done.
See vignette("multiclass", "yardstick")
for more information.
There is no common convention on which factor level should
automatically be considered the "event" or "positive" result.
In yardstick
, the default is to use the first level. To
change this, a global option called yardstick.event_first
is
set to TRUE
when the package is loaded. This can be changed
to FALSE
if the last level of the factor is considered the
level of interest by running: options(yardstick.event_first = FALSE)
.
For multiclass extensions involving onevsall
comparisons (such as macro averaging), this option is ignored and
the "one" level is always the relevant result.
pr_curve()
for computing the full precision recall curve.
Other class probability metrics:
average_precision()
,
gain_capture()
,
mn_log_loss()
,
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. pr_auc(two_class_example, truth, Class1)#> # A tibble: 1 x 3 #> .metric .estimator .estimate #> <chr> <chr> <dbl> #> 1 pr_auc binary 0.946#  # 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") %>% pr_auc(obs, VF:L)#> # A tibble: 1 x 3 #> .metric .estimator .estimate #> <chr> <chr> <dbl> #> 1 pr_auc macro 0.611# 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")) %>% pr_auc(obs, M, VF:L)#> # A tibble: 1 x 3 #> .metric .estimator .estimate #> <chr> <chr> <dbl> #> 1 pr_auc macro 0.611#> # A tibble: 10 x 4 #> Resample .metric .estimator .estimate #> <chr> <chr> <chr> <dbl> #> 1 Fold01 pr_auc macro 0.611 #> 2 Fold02 pr_auc macro 0.620 #> 3 Fold03 pr_auc macro 0.689 #> 4 Fold04 pr_auc macro 0.680 #> 5 Fold05 pr_auc macro 0.620 #> 6 Fold06 pr_auc macro 0.650 #> 7 Fold07 pr_auc macro 0.607 #> 8 Fold08 pr_auc macro 0.650 #> 9 Fold09 pr_auc macro 0.628 #> 10 Fold10 pr_auc macro 0.603# Weighted macro averaging hpc_cv %>% group_by(Resample) %>% pr_auc(obs, VF:L, estimator = "macro_weighted")#> # A tibble: 10 x 4 #> Resample .metric .estimator .estimate #> <chr> <chr> <chr> <dbl> #> 1 Fold01 pr_auc macro_weighted 0.746 #> 2 Fold02 pr_auc macro_weighted 0.743 #> 3 Fold03 pr_auc macro_weighted 0.789 #> 4 Fold04 pr_auc macro_weighted 0.754 #> 5 Fold05 pr_auc macro_weighted 0.737 #> 6 Fold06 pr_auc macro_weighted 0.743 #> 7 Fold07 pr_auc macro_weighted 0.748 #> 8 Fold08 pr_auc macro_weighted 0.756 #> 9 Fold09 pr_auc macro_weighted 0.711 #> 10 Fold10 pr_auc macro_weighted 0.737# Vector version # Supply a matrix of class probabilities fold1 < hpc_cv %>% filter(Resample == "Fold01") pr_auc_vec( truth = fold1$obs, matrix( c(fold1$VF, fold1$F, fold1$M, fold1$L), ncol = 4 ) )#> [1] 0.6109931