pr_curve()
constructs the full precision recall curve and returns a
tibble. See pr_auc()
for the area under the precision recall curve.
pr_curve(data, ...) # S3 method for data.frame pr_curve(data, truth, ..., na_rm = TRUE) autoplot.pr_df(object, ...)
data  A 

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

truth  The column identifier for the true class results
(that is a 
na_rm  A 
object  The 
A tibble with class pr_df
or pr_grouped_df
having
columns .threshold
, recall
, and precision
.
pr_curve()
computes the precision at every unique value of the
probability column (in addition to infinity).
There is a ggplot2::autoplot()
method for quickly visualizing the curve. This works for
binary and multiclass output, and also works with grouped data (i.e. from
resamples). See the examples.
If a multiclass truth
column is provided, a onevsall
approach will be taken to calculate multiple curves, one per level.
In this case, there will be an additional column, .level
,
identifying the "one" column in the onevsall calculation.
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.
Compute the area under the precision recall curve with pr_auc()
.
Other curve metrics:
gain_curve()
,
lift_curve()
,
roc_curve()
#  # 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_curve(two_class_example, truth, Class1)#> # A tibble: 501 x 3 #> .threshold recall precision #> <dbl> <dbl> <dbl> #> 1 Inf 0 1 #> 2 1.00 0.00388 1 #> 3 1.00 0.00775 1 #> 4 1.00 0.0116 1 #> 5 1.00 0.0155 1 #> 6 1.00 0.0194 1 #> 7 1.00 0.0233 1 #> 8 1.00 0.0271 1 #> 9 1.00 0.0310 1 #> 10 1.00 0.0349 1 #> # … with 491 more rows#  # `autoplot()` # Visualize the curve using ggplot2 manually library(ggplot2) library(dplyr) pr_curve(two_class_example, truth, Class1) %>% ggplot(aes(x = recall, y = precision)) + geom_path() + coord_equal() + theme_bw()# Multiclass onevsall approach # One curve per level hpc_cv %>% filter(Resample == "Fold01") %>% pr_curve(obs, VF:L) %>% autoplot()# Same as above, but will all of the resamples hpc_cv %>% group_by(Resample) %>% pr_curve(obs, VF:L) %>% autoplot()