The False Positive Rate (FPR) is a binary-classification metric: the proportion of actual negatives that the classifier wrongly labels positive. From the confusion matrix it is \( \mathrm{FPR} = \mathrm{FP} / (\mathrm{FP} + \mathrm{TN}) \), defined only when the negative class is non-empty, and it equals one minus specificity. Because the denominator counts only actual negatives, FPR is independent of class balance (prevalence) but it does move with the decision threshold: lowering the threshold raises both the true positive rate and the FPR, which is exactly the trade-off a Receiver Operating Characteristic (ROC) curve plots, with FPR on the horizontal axis. Compute it by counting false positives and true negatives on a labelled evaluation set at a fixed threshold, never from accuracy or from the positive class alone.
The false positive rate of a diagnostic test is the proportion of people without the condition who are incorrectly classified as testing positive, calculated as the number of false positives divided …