Precision
Precision (Positive Predictive Value) · Also known as: Positive Predictive Value, PPV
Precision measures the proportion of positive predictions that were actually correct. It answers the question: 'Of all the cases we predicted as positive, how many were truly positive?' Precision is critical in scenarios where false positives are costly.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use precision when false positives are costly or inconvenient, such as spam filtering (false positives annoy users by filtering legitimate email), medical testing for expensive treatments (false positives lead to unnecessary treatment), or content moderation (false positives unfairly flag legitimate content). Precision is less important when missing positive cases (high false negatives) is the primary concern.
Strengths & limitations
- Directly measures confidence in positive predictions; high precision means positive predictions are reliable.
- Interpretable: precision = 1.0 means all positive predictions were correct.
- Essential for applications where false positives have high cost or social impact.
- Complements recall to give a complete picture of classification performance.
- Ignores false negatives; a model could have high precision while missing many positive cases.
- Can be artificially high if the model is overly conservative (makes very few positive predictions).
- Does not account for class imbalance or the relative costs of different error types.
- Alone, precision is incomplete without considering recall.
Frequently asked
When should I prioritize precision over recall?
Prioritize precision when the cost of a false positive is high. In spam filtering, a false positive wastes user time and trust. In medical screening for expensive treatments, a false positive leads to unnecessary medical procedures. In contrast, for cancer screening tests, recall is often more important because missing a cancer case is very costly.
Can a model have high precision but low recall?
Yes. A model could predict positive only in cases where it is very confident, correctly predicting few cases (high precision) but missing many actual positive cases (low recall). This is a conservative strategy suitable only if false positives are extremely costly and missing cases is acceptable.
How do precision and recall relate to false positives and false negatives?
Precision is directly related to false positives: it decreases as false positives increase. Recall is directly related to false negatives: it decreases as false negatives increase. To improve precision, reduce false positives. To improve recall, reduce false negatives. The trade-off between them is fundamental in classification.
Sources
- Fawcett, T. (2006). An introduction to ROC analysis. Pattern Recognition Letters, 27(8), 861-874. DOI: 10.1016/j.patrec.2005.10.010 ↗
- Powers, D. M. (2011). Evaluation: From Precision, Recall and F-Measure to ROC, Informedness, Markedness and Correlation. Journal of Machine Learning Technologies, 2(1), 37-63. link ↗
How to cite this page
ScholarGate. (2026, June 3). Precision (Positive Predictive Value). ScholarGate. https://scholargate.app/en/model-evaluation/precision
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- AccuracyModel Evaluation↔ compare
- F1-ScoreModel Evaluation↔ compare
- Matthews Correlation CoefficientModel Evaluation↔ compare
- Recall (Sensitivity)Model Evaluation↔ compare
- SpecificityModel Evaluation↔ compare