precision-recall curve

A precision-recall curve is a diagnostic visualization that plots precision against recall for a binary classifier across all possible decision thresholds. Each point on the curve corresponds to a specific threshold: recall (true positive rate) on the x-axis measures the proportion of actual positives correctly identified, while precision (positive predictive value) on the y-axis measures the proportion of predicted positives that are actually positive. The curve reveals the tradeoff between finding more positives and maintaining confidence in those finds — a classifier that perfectly separates classes produces a curve hugging the top-left corner, while a random classifier traces a diagonal. This curve is used to compare classifiers and select operating thresholds, especially when class distributions are imbalanced and accuracy is misleading. [formal: curve precision-recall | substrate: mind | horizon: hours | explicit: yes | epoch: 0.01]