summary.conf_mat: Summary Statistics for Confusion Matrices

Description

Various statistical summaries of confusion matrices are produced and returned in a tibble. These include those shown in the help pages for sens(), recall(), and accuracy(), among others.

Usage

# S3 method for conf_mat
summary(
  object,
  prevalence = NULL,
  beta = 1,
  estimator = NULL,
  event_level = yardstick_event_level(),
  ...
)

Value

A tibble containing various classification metrics.

Arguments

object: An object of class conf_mat().
prevalence: A number in (0, 1) for the prevalence (i.e. prior) of the event. If left to the default, the data are used to derive this value.
beta: A numeric value used to weight precision and recall for f_meas().
estimator: One of: "binary", "macro", "macro_weighted", or "micro" to specify the type of averaging to be done. "binary" is only relevant for the two class case. The other three are general methods for calculating multiclass metrics. The default will automatically choose "binary" or "macro" based on estimate.
event_level: A single string. Either "first" or "second" to specify which level of truth to consider as the "event". This argument is only applicable when estimator = "binary". The default uses an internal helper that generally defaults to "first", however, if the deprecated global option yardstick.event_first is set, that will be used instead with a warning.
...: Not currently used.

Relevant Level

There is no common convention on which factor level should automatically be considered the "event" or "positive" result when computing binary classification metrics. In yardstick, the default is to use the first level. To alter this, change the argument event_level to "second" to consider the last level of the factor the level of interest. For multiclass extensions involving one-vs-all comparisons (such as macro averaging), this option is ignored and the "one" level is always the relevant result.

Examples

Run this code

data("two_class_example")

cmat <- conf_mat(two_class_example, truth = "truth", estimate = "predicted")
summary(cmat)
summary(cmat, prevalence = 0.70)

library(dplyr)
library(tidyr)
data("hpc_cv")

# Compute statistics per resample then summarize
all_metrics <- hpc_cv %>%
  group_by(Resample) %>%
  conf_mat(obs, pred) %>%
  mutate(summary_tbl = lapply(conf_mat, summary)) %>%
  unnest(summary_tbl)

all_metrics %>%
  group_by(.metric) %>%
  summarise(
    mean = mean(.estimate, na.rm = TRUE),
    sd = sd(.estimate, na.rm = TRUE)
  )