model_metrics_curves: Plot model metrics across classification threshold

In mai00fti/prettyPROC: Print pretty PR and ROC curves

Plot model metrics across classification threshold

Description

This function uses the tibble that can be generated with the package function [get_threshold_data()] and plots the values of selected metrics in dependency of the threshold. This can be inspiring to evaluate the selection of a threshold for different learning aims. A selected threshold value can be provided to set a vertical mark in the plot

Usage

model_metrics_curves(df, metrics, plot_title, threshold = NA)

Arguments

df	- The tibble that can be generated with the package function [get_threshold_data()]
metrics	- A character vectore of metrics that will be plotted. Note, that some of the returned metrics mean principally the same: Recall = Sensitivity, Precision = PPV, BalancedAccuracy = roc_auc. The available metrics are: "Sensitivity", "Specificity", "Pos Pred Value", "Neg Pred Value", "Precision", "Recall", "F1", "Prevalence", "Detection Rate", "Detection Prevalence", "Balanced Accuracy", "TP", "FN", "FP", "TN", "P", "N", "N_samples", "pr_baseline", "P_pred", "N_pred", "fpr", "tpr", "tnr", "fnr", "roc_auc"
plot_title	- A title for your plot
threshold	- The threshold that should be marked with a vertical dashed line. Optional.

Value

A ggplot2 object is returned which can be extended by further ggplot functions, be showed and saved. NULL is returned, if none of the requested metrics is available

Examples

y_true <- sample(c(0,1), replace = TRUE, size = 1000)
y_predicted <- runif(1000)
data <- get_threshold_data(truth = y_true, prediction = y_predicted)
plot <- model_metrics_curves(df = data,
                             metrics = c("Sensitivity", "F1", "Balanced Accuracy"),
                             plot_title = "SENS, F1 and balanced Accuracy across classification threshold",
                             threshold = 0.5)
show(plot)

mai00fti/prettyPROC documentation built on Aug. 16, 2024, 4:48 p.m.