Precision

Precision

Precision is a metric used in classification tasks to measure how many of the predicted positive results are actually correct. It is also known as the positive predictive value.

Definition

$\text{Precision}=\frac{TP}{TP+FP}$

Where:

• TP = True Positives – correct positive predictions
• FP = False Positives – incorrect positive predictions

Precision helps to answer the question: "Of all the items labeled as positive, how many are truly positive?"

Simple Example

Imagine a spam filter that marked 100 emails as spam. Out of these, 80 were actually spam, and 20 were not.

• TP = 80
• FP = 20

Then,

$\text{Precision}=\frac{80}{80+20}=\frac{80}{100}=0.8=80\mathrm{\%}$

This means that 80% of emails flagged as spam were truly spam.

When to Use Precision

Precision is especially important when the cost of false positives is high.

Real-World Scenarios

• Medical testing: Avoiding telling a healthy person they are sick.
• Email spam detection: Ensuring important emails aren't marked as spam.
• Search engines: Returning highly relevant search results.

High vs Low Precision

• High Precision: Most positive predictions are correct.
• Low Precision: Many false alarms (false positives).

Related Metrics

• Recall
• F1 Score
• Accuracy
• Confusion Matrix

SEO Keywords

precision in machine learning, positive predictive value, classification metric, ML model accuracy, spam detection precision, precision formula