Home › Compare › Precision vs Recall

Precision vs Recall

Precision — at a glance

Category: Evaluation · Difficulty: Beginner

Of the items the model predicted positive, the fraction that are actually positive.

Precision = TP / (TP + FP). High precision means few false alarms — important when acting on a positive prediction is costly (medical procedure, account ban). Always reported alongside recall, since you can trade one for the other.

Read the full Precision definition →

Recall — at a glance

Category: Evaluation · Difficulty: Beginner

Of all the actually positive items, the fraction the model successfully found.

Recall = TP / (TP + FN). High recall means few missed cases — vital when missing a positive is costly (cancer detection, fraud, security). The precision–recall trade-off is at the heart of every classifier tuning decision.

Read the full Recall definition →

Key differences

  • Purpose: Precision is typically used for evaluation problems, while Recall fits evaluation use cases.
  • Complexity: Precision is rated Beginner; Recall is rated Beginner.
  • Definitions: Of the items the model predicted positive, the fraction that are actually positive. vs Of all the actually positive items, the fraction the model successfully found.

Frequently asked questions

What is the difference between Precision and Recall?

Precision: Of the items the model predicted positive, the fraction that are actually positive. Recall: Of all the actually positive items, the fraction the model successfully found.

When should I use Precision instead of Recall?

Use Precision when your problem matches its strengths: Of the items the model predicted positive, the fraction that are actually positive. Use Recall when Of all the actually positive items, the fraction the model successfully found.

Can Precision and Recall be used together?

Yes — many modern AI systems combine Precision and Recall to get the strengths of both approaches.

Is Precision better than Recall?

Neither is universally better. The right choice depends on data, latency, cost, and task. This page breaks down the trade-offs.