## ----------------------------------------------------------------------------- knitr::opts_chunk$set( collapse = TRUE, comment = "#>" ) ## ----------------------------------------------------------------------------- library(yardstick) library(dplyr) data("hpc_cv") hpc_cv |> group_by(Resample) |> slice(1:3) ## ----------------------------------------------------------------------------- hpc_cv |> filter(Resample == "Fold01") |> accuracy(obs, pred) ## ----------------------------------------------------------------------------- hpc_cv |> group_by(Resample) |> accuracy(obs, pred) ## ----------------------------------------------------------------------------- class_metrics <- metric_set(accuracy, kap) hpc_cv |> group_by(Resample) |> class_metrics(obs, estimate = pred) ## ----------------------------------------------------------------------------- library(knitr) library(dplyr) yardns <- asNamespace("yardstick") fns <- lapply(names(yardns), get, envir = yardns) names(fns) <- names(yardns) get_metrics <- function(fns, type) { where <- vapply(fns, inherits, what = type, FUN.VALUE = logical(1)) paste0("`", sort(names(fns[where])), "()`") } all_metrics <- bind_rows( tibble(type = "class", metric = get_metrics(fns, "class_metric")), tibble(type = "class prob", metric = get_metrics(fns, "prob_metric")), tibble( type = "ordered prob", metric = get_metrics(fns, "ordered_prob_metric") ), tibble(type = "numeric", metric = get_metrics(fns, "numeric_metric")), tibble( type = "dynamic survival", metric = get_metrics(fns, "dynamic_survival_metric") ), tibble( type = "integrated survival", metric = get_metrics(fns, "integrated_survival_metric") ), tibble( type = "static survival", metric = get_metrics(fns, "static_survival_metric") ), tibble( type = "linear predictor survival", metric = get_metrics(fns, "linear_pred_survival_metric") ), tibble( type = "quantile", metric = get_metrics(fns, "quantile_metric") ) ) kable(all_metrics, format = "html")