What's the difference between recall and precision?

Recall: what fraction of actual positives did you find? Precision: what fraction of positive predictions were correct? Recall focuses on not missing positives; precision focuses on not having false alarms.

How do I improve recall?

Lower the classification threshold (more liberal "yes" decisions), add features that help identify hard-to-find positives, address class imbalance, or use recall-focused training objectives.

When is recall more important than precision?

When false negatives are costly/dangerous: medical diagnosis, security threats, safety-critical defects. When you can afford to investigate false positives but can't afford to miss real cases.

What is the F1 score?

F1 is the harmonic mean of precision and recall, balancing both. F1 = 2 × (Precision × Recall) / (Precision + Recall). Useful when you need good precision AND recall.

Book Free Assessment Calculator

Recall

Of all actual positives, what proportion did the model identify. High recall means few false negatives - the model finds most of the positive cases.

In-Depth Explanation

Recall (sensitivity, true positive rate) measures how well the model finds all positive cases. It answers: "Of everything that was actually positive, how much did we catch?"

Recall formula: Recall = True Positives / (True Positives + False Negatives)

Recall focus:

Minimises false negatives
Important when missing positives is costly
Trade-off with precision typically

High recall needed when:

Missing positive cases is dangerous/costly
Comprehensive coverage is required
False negatives have severe consequences

Examples:

Disease screening (missing cancer is catastrophic)
Security threats (missing attacks is dangerous)
Defect detection (missing defects causes recalls)

Business Context

Optimise for recall when missing positives is dangerous. Better to investigate false alarms than miss real threats.

How Clever Ops Uses This

We configure AI systems for appropriate recall levels for Australian businesses, especially in safety-critical or high-stakes applications.

Example Use Case

"A manufacturing defect detector with 95% recall catches 95 out of 100 defective products - only 5 slip through."

Frequently Asked Questions

Related Terms

Precision F1 Score Confusion Matrix

Learn More

Model Selection and Evaluation: Choosing the Right AI Model for Your Use Case

Learn how to select the optimal AI model for your needs by comparing capabilities, costs, and performance. Includes evaluation frameworks, benchmarking strategies, and migration guidance.

Read article

Real-Time Processing Recall