Usage
binary_prediction_scores(x, observed = NULL, threshold = 0.5, epsilon = 1e-12)
Arguments
- x
A binary expected-response gp3bayes_prediction or numeric event
probabilities.
- observed
Optional binary outcomes when x is numeric.
- threshold
Classification threshold used only for threshold summaries.
- epsilon
Probability truncation used for finite log loss.
Value
A one-row data frame of descriptive predictive scores.
Examples
binary_prediction_scores(c(0.1, 0.8, 0.7, 0.2), c(0, 1, 1, 0))
#> n brier log_loss auc threshold accuracy sensitivity specificity
#> 1 4 0.045 0.2270806 1 0.5 1 1 1
#> balanced_accuracy automatic_decision
#> 1 1 FALSE
