For classification problems, conf_mat_resampled() computes a separate confusion matrix for each resample then averages the cell counts.

conf_mat_resampled(x, parameters = NULL, tidy = TRUE)

Arguments

x

An object with class tune_results that was used with a classification model that was run with control_*(save_pred = TRUE).

parameters

A tibble with a single tuning parameter combination. Only one tuning parameter combination (if any were used) is allowed here.

tidy

Should the results come back in a tibble (TRUE) or a matrix.

Value

A tibble or matrix with the average cell count across resamples.

Examples

library(parsnip) library(rsample) library(dplyr) data(two_class_dat, package = "modeldata") set.seed(2393) res <- logistic_reg() %>% set_engine("glm") %>% fit_resamples(Class ~ ., resamples = vfold_cv(two_class_dat, v = 3), control = control_resamples(save_pred = TRUE)) conf_mat_resampled(res)
#> # A tibble: 4 x 3 #> Prediction Truth Freq #> <fct> <fct> <dbl> #> 1 Class1 Class1 123 #> 2 Class1 Class2 25.7 #> 3 Class2 Class1 22.7 #> 4 Class2 Class2 92.3
conf_mat_resampled(res, tidy = FALSE)
#> Class1 Class2 #> Class1 123.00000 22.66667 #> Class2 25.66667 92.33333