detectiecurve

Detectiecurve is a graphical representation used to evaluate the performance of a detector or classifier. In many applications it describes how the probability of detection (true positives) varies with the chosen decision threshold or with the false positive rate, providing a compact summary of trade-offs between sensitivity and specificity. The term is used in Dutch literature as a generic label; in English contexts it is often referred to as a detection curve or, more specifically, as a receiver operating characteristic (ROC) curve or a sensitivity–specificity curve, depending on the axis choices.

Construction of a detectiecurve typically involves a dataset with known positives and negatives. The detector is run across a range of thresholds, and for each threshold the numbers of true positives and false positives are counted. From these counts, rates such as the true positive rate and false positive rate are computed. Plotting these rates for the various thresholds yields the curve; some variations plot detection probability against the threshold level or against signal strength.

Interpretation of the curve focuses on the trade-off between misses and false alarms. A curve that rises quickly and remains high indicates strong performance. The area under the curve (AUC) is a common single-number summary of overall ability. Detectiecurves are used to compare detectors, select appropriate thresholds, or assess robustness across different data conditions.

Applications span signal processing, radar and sonar, medical imaging, quality control, fraud and spam detection, and environmental monitoring, among others. Related concepts include ROC curves, precision–recall curves, sensitivity, specificity, and threshold tuning.