DataTerminology.comNEW
Also known as: Sensitivity, True Positive Rate
The fraction of actual positives correctly identified by a classifier: TP divided by TP plus FN.
DESCRIPTION
Recall measures completeness: how many of the true positive instances a model recovers. It is crucial in tasks where missing positives is costly, such as disease detection or fraud flagging. Recall trades off with precision based on decision thresholds. Evaluation often considers precision–recall curves, F1 scores, and domain-specific costs. Calibrating thresholds and class weights helps optimize recall when warranted.
ALTERNATIVE DEFINITIONS
Wikipedia
Recall is the fraction of relevant instances that were retrieved.
Concise definition stressing the notion of relevant instances retrieved.
scikit-learn
Recall is the ratio tp / (tp + fn) where tp is true positives and fn is false negatives.
Provides explicit formula linking recall to confusion matrix components.
Google Developers
Recall is the percentage of actual positives that the model correctly identified as positive.
Defines recall in plain terms with percentage interpretation.
RELATED TERMS
F1 Score Specificity ROC Curve Confusion Matrix False Positive Rate Precision-Recall Curve Balanced Accuracy Type II Error Threshold
COMPETING TERMS
RELATIONS
- Narrower: Precision and Recall
KNOWLEDGE AREAS
AI Data Science Testing & Validation
TARGET AUDIENCE
Data Scientist AI Engineer MLOps Engineer Data Analyst
EXTERNAL SOURCES
- Wikipedia - Precision and recall
- scikit-learn - Model evaluation metrics including recall
- Google Developers - Crash course on precision and recall
TOP QUESTIONS
When is Recall more important than Precision?
When missing a positive is costly, such as medical screening or fraud detection, recall should be prioritized.
Category: Advice
How do I increase Recall for a classifier?
Lower the decision threshold, adjust class weights, oversample positives, or optimize for recall-related metrics.
Category: Practical
What is the trade-off between Recall and Precision?
Improving recall often reduces precision due to more positive predictions, affecting false positives.
Category: Concept
How does Recall relate to the ROC curve?
Recall equals true positive rate, which is plotted on the ROC y-axis.
Category: Concept
What is a common mistake with optimizing Recall?
Maximizing recall without monitoring precision can flood systems with false alarms.
Category: Mistakes
How does class imbalance affect Recall?
Severe imbalance can reduce recall for minority classes; rebalancing techniques can help.
Category: Risk
Which metric balances Recall and Precision?
The F1 score is the harmonic mean of precision and recall.
Category: Concept
How do thresholds influence Recall?
Lower thresholds increase recall by classifying more instances as positive.
Category: Practical
Can Recall be perfect?
Yes, but often at the cost of very low precision, depending on the problem.
Category: Limits