How It Works
Positive Likelihood Ratio (PLR)
A single number that says how much a positive result from a test or classifier should raise your confidence that something is really there. It is the true-positive rate divided by the false-positive rate (sensitivity ÷ (1 − specificity)). A PLR of 1 means a positive result tells you nothing; around 10 or more means a positive is strong evidence. It is used to judge medical tests and, increasingly, AI detectors and classifiers.
Origin · no single documented coiner
A long-standing tool of statistics and clinical epidemiology rather than one person’s coinage; it was popularized for everyday diagnostic use by evidence-based-medicine texts in the 1980s and 1990s, including work by David Sackett and colleagues.