Definition Of all the items a model flagged as positive, the share that were actually positive.
Example 80 real fraud cases out of 100 flagged gives a precision of 80%.
For your role Your role Choose your role Accounts payable Credit controllers FP&A analysts Finance business partners Finance directors and CFOs Financial controllers Management accountants Payroll specialists Account executives Sales managers Sales operations analysts Sales representatives Bid and tender managers Commercial directors New to business Pricing analysts Revenue managers Senior leaders Campaign managers Communications managers Customer insight analysts Digital marketing managers Marketing analysts Marketing directors HR business partners HR directors Learning and development managers People analysts People managers Recruiters Operations analysts Operations directors Operations managers AI product managers Heads of product Product analysts Product managers Product owners Machine learning engineers Analytics engineers Data analysts Data architects Data engineers Data governance leads Data privacy officers Data scientists Information architects Platform engineers Report builders Business analysts Corporate development managers Heads of strategy Strategy analysts
Choose your role to see what this term means in your work.
Why it matters It shows how trustworthy a model’s positive predictions are.
Where it is used AI model reports and fraud detection reviews.
How to calculate it True positives ÷ (true positives + false positives) × 100
Worked example 80 real fraud cases out of 100 flagged gives a precision of 80%.
