Opening your learning space
Starting the app …
Opening your learning space
Starting the app …
Precision is the fraction of positive predictions that are correct among all cases classified as positive.
Free to browse · flashcards or a multiple-choice quiz · import from Essential
Formula
dimensionless, between 0 and 1; convention-dependent when TP+FP=0
| Symbol | Meaning | Unit |
|---|---|---|
| True positive predictions | count | |
| False positive predictions | count |
\mathrm{Precision}=\frac{TP}{TP+FP}The positive predictive value asks: when a model predicts positive, how often is that prediction correct? It depends on the decision threshold and class distribution in the evaluated data. When TP+FP=0 the metric is undefined, so software must choose a convention. Precision ignores false negatives and true negatives; it does not measure completeness or overall accuracy. Comparisons should use the same class, threshold and test population.
The positive predictive value asks: when a model predicts positive, how often is that prediction correct? It depends on the decision threshold and class distribution in the evaluated data. When TP+FP=0 the metric is undefined, so software must choose a convention. Precision ignores false negatives and true negatives; it does not measure completeness or overall accuracy. Comparisons should use the same class, threshold and test population.
Precision is the fraction of positive predictions that are correct among all cases classified as positive.
True positives from precision
Keep the classification threshold fixed.
Use the example for Precision (positive predictive value) to explain which inputs determine the requested output.
The example connects concrete inputs to its result.
Precision is the fraction of positive predictions that are correct among all cases classified as positive.
The relation applies only under its stated assumptions.
Applying the relation to a different case without checking its stated assumptions.
Check the validity range before substituting values: The positive predictive value asks: when a model predicts positive, how often is that prediction correct? It depends on the decision threshold and class distribution in the evaluated data. When.
Formula cluster
Precision is the fraction of positive predictions that are correct among all cases classified as positive.
A filter marks TP=36 messages correctly as relevant and FP=9 incorrectly. Precision=36/(36+9)=0.80: four out of five positive flags are correct.
Assess positive alerts when false alarms are costly, such as review queues or medical follow-up.
Here is how to work through a typical Precision (positive predictive value) (Precision = TP/(TP+FP)) task step by step:
Task
Use the example for Precision (positive predictive value) to explain which inputs determine the requested output.
Solution path
The example connects concrete inputs to its result.
Precision is the fraction of positive predictions that are correct among all cases classified as positive.
The relation applies only under its stated assumptions.
Precision considers every case the model marked positive and asks what fraction was actually positive. A value of 0.80 means four out of five positive flags were correct. It does not count missed positives or correct negatives. The metric is useful when a false alarm triggers costly review. The score depends on the threshold and prevalence of positive cases. Record the test population, positive class and decision threshold so results can be compared meaningfully. A changed base rate can shift precision even when other model properties remain stable.
Divide true positives TP by all positive predictions TP+FP. With TP=36 and FP=9, precision is 36/45=0.80. State which class is positive; reversing labels changes TP and FP. The denominator must be nonzero. If there are no positive predictions, the metric is undefined and a software convention must be reported. The score depends on the threshold and prevalence of positive cases. Record the test population, positive class and decision threshold so results can be compared meaningfully. A changed base rate can shift precision even when other model properties remain stable.
A classifier often produces a score that is converted into a positive or negative decision using a threshold. A higher threshold often yields fewer positive decisions and may reduce false alarms, but can miss true positives. TP and FP therefore change, as does precision. The threshold is part of the reported metric; scores from different thresholds need context. The score depends on the threshold and prevalence of positive cases. Record the test population, positive class and decision threshold so results can be compared meaningfully. A changed base rate can shift precision even when other model properties remain stable.
No. Precision considers only cases predicted positive. Accuracy counts all correct predictions, TP and TN, over all cases. With rare positives, a model can have high accuracy while its positive alerts are unreliable. Report recall and the confusion matrix as well to show false alarms and missed cases. The score depends on the threshold and prevalence of positive cases. Record the test population, positive class and decision threshold so results can be compared meaningfully. A changed base rate can shift precision even when other model properties remain stable.
When TP+FP=0, the classifier predicted no positives. The empirical positive predictive value cannot be determined because there is no positive prediction to assess. Libraries may warn, assign zero by convention or omit the case. Report that handling, especially for small test sets; it is not a mathematically measured precision of zero. The score depends on the threshold and prevalence of positive cases. Record the test population, positive class and decision threshold so results can be compared meaningfully. A changed base rate can shift precision even when other model properties remain stable.
Quanta exam set
Curated exam set for "Precision (positive predictive value)":
Question (front)
Which formula describes Precision (positive predictive value)?
Answer in your set
Question (front)
How do you rearrange Precision (positive predictive value) for True positives from precision?
Answer in your set
Question (front)
Which common mistake happens with Precision (positive predictive value)?
Answer in your set
+ 7 more cards: units, variables, derivation, example, exam task
The formula set contains 10 cards. After importing, review them and start practicing; your ratings determine later review intervals.
This list names subject references and further reading. A page or section number appears when it is part of the reference; these are not verbatim quotations.
Create a curated FSRS exam set for Precision = TP/(TP+FP): formula recall, variables, derivation, rearrangement, worked example, common mistakes and exam context.
Curated formula set · LaTeX · FSRS spaced repetition · import from Essential