evaluate_performance: Evaluate Model Performance (ROC and Precision-Recall)

In quickSentiment: A Fast and Flexible Pipeline for Text Classification

Evaluate Model Performance (ROC and Precision-Recall)

Description

Evaluate Model Performance (ROC and Precision-Recall)

Usage

evaluate_performance(predicted_probs, actual_classes, positive_label)

Arguments

predicted_probs	Numeric vector of predicted probabilities for the positive class.
actual_classes	Factor or character vector of the actual true labels.
positive_label	Character string. The target class you want to evaluate.

Value

An object of class 'quickSentiment_eval', which is a list containing the following metrics:

target_class	Character. The specific positive label used for the evaluation.
auc_roc	Numeric. The Area Under the Receiver Operating Characteristic curve.
best_threshold_roc	Numeric. The optimal probability threshold that maximizes Youden's J statistic.
auc_pr	Numeric. The Area Under the Precision-Recall curve.
best_threshold_pr	Numeric. The probability threshold that maximizes the F1-Score.
accuracy_at_best	Numeric. The overall accuracy of the model if 'best_threshold_pr' is applied.
roc	S3 object containing the ROC curve data and metrics.
prc	S3 object containing the Precision-Recall curve data and metrics.
threshold_summary	A data frame summarizing Accuracy, Precision, Recall, and F1 at 0.1 threshold increments.

quickSentiment documentation built on Aug. 29, 2026, 1:07 a.m.