Skip to content

Pull a Classification Performance Metric/Statistic

pull_stat.Rd

Indexes into the S3 summary method for confusion_matrix class object.

Usage

pull_stat(
  x,
  which = c("Sensitivity", "Recall", "Specificity", "Precision", "PPV", "NPV",
    "Accuracy", "MCC", "AUC", "Brier", "Error", "F_measure", "G_mean", "Wt_Acc")
)

Arguments

x

A summary_confusion_matrix class object, from a call to summary(calc_confusion(...)).

which

character(1). Matched string for the test desired statistic.

Value

double(1). The classification performance statistic.

Details

Optional statistics are restricted to the output of the summary method. One of:

  • "Sensitivity"

  • "Recall"

  • "Specificity"

  • "Precision"

  • "PPV"

  • "NPV"

  • "Accuracy"

  • "Error"

  • "F_measure"

  • "G_mean"

  • "Wt_Acc"

See also

calc_confusion()

Author

Stu Field

Examples

n <- 20
withr::with_seed(22, {
  true  <- sample(c("control", "disease"), n, replace = TRUE)
  pred  <- runif(n)
})

# Summary of confusion matrix
c_mat <- calc_confusion(true, pred, pos_class = "disease") |> summary()
class(c_mat)
#> [1] "summary_confusion_matrix" "list"                    

# pull out elements
pull_stat(c_mat, "Spec")
#> [1] 0.5714286
pull_stat(c_mat, "Sens")
#> [1] 0.4615385
pull_stat(c_mat, "Recall")    # same
#> [1] 0.4615385
pull_stat(c_mat, "Accuracy")
#> [1] 0.5
pull_stat(c_mat, "Error")     # 1 - accuracy
#> [1] 0.5
pull_stat(c_mat, "PPV")
#> [1] 0.6666667
pull_stat(c_mat, "NPV")
#> [1] 0.3636364
pull_stat(c_mat, "F_measure")
#> [1] 0.5454545
pull_stat(c_mat, "Wt_Acc")
#> [1] 0.489011