Evaluate Predictions from a Data Frame

Evaluates model performance from a data frame of predictions, calculating metrics like AUROC, AUPRC, F1 score, etc. This function is designed for use with prediction results, such as the output from apply_dia.

Usage

evaluate_predictions_dia(
  prediction_df,
  threshold_choices = "default",
  pos_class = "Positive",
  neg_class = "Negative"
)

Arguments

prediction_df

A data frame containing predictions. Must contain the columns sample, label (true labels), and score (predicted probabilities).

threshold_choices

A character string specifying the thresholding strategy ("default", "f1", "youden") or a numeric probability threshold value (0-1).

pos_class

A character string for the positive class label used in reporting. Defaults to "Positive".

neg_class

A character string for the negative class label used in reporting. Defaults to "Negative".

Value

A named list containing all calculated performance metrics.

Details

This function strictly requires the label column in prediction_df to adhere to the following format:

• 1: Represents the positive class.

• 0: Represents the negative class.

• NA: Will be ignored during calculation.

The function will stop with an error if any other values are found in the label column.

Examples

# \donttest{
# # Create a sample prediction data frame
# predictions_df <- data.frame(
#   sample = 1:10,
#   label = c(1, 0, 1, 1, 0, 0, 1, 0, 1, 0),
#   score = c(0.9, 0.2, 0.8, 0.6, 0.3, 0.4, 0.95, 0.1, 0.7, 0.5)
# )
#
# # Evaluate the predictions using the 'f1' threshold strategy
# evaluation_results <- evaluate_predictions_dia(
#   prediction_df = predictions_df,
#   threshold_choices = "f1"
# )
#
# print(evaluation_results)
# }