liveBook · Manning

False positive

As explained in Grokking Machine Learning

A false positive is a classification error in which a model predicts the positive class for a negative example.

Imagine a medical model that labels sick patients as positive. If it labels a healthy patient as sick, that prediction is a false positive. More generally, a false positive is a negatively labeled data point that a model incorrectly predicts as positive. section 7.2

Use the confusion matrix to distinguish this error from the other prediction outcomes:

  • True positive: a positive example correctly predicted as positive. section 7.2
  • False negative: a positive example incorrectly predicted as negative. section 7.2
  • True negative: a negative example correctly predicted as negative. section 7.2

Spam detection

When spam is the positive class, a ham email incorrectly marked as spam is a false positive. A spam model with 5 false positives, 30 true positives, 10 false negatives, and 55 true negatives makes this error five times. section 7.2

Why the error matters

False positives and false negatives do not always have equal importance. Precision is useful when false positives should be minimized, such as in spam filtering; recall is useful when false negatives matter more, such as in medical diagnosis. section 7.4

How Grokking Machine Learning covers it

How deep it goes
Defines the error, contrasts it with three other outcomes, locates it in a confusion matrix, and connects it to precision.
Examples
Uses coronavirus diagnosis, spam filtering, confusion-matrix tables, and recommendation, pedestrian-detection, and voice-assistant exercises.
What you can do after
Evaluate models by counting false positives and choosing precision when avoiding them is especially important.
How it connects
Connects false positives to false negatives, confusion matrices, accuracy, precision, recall, F-score, sensitivity, and specificity.

Also explained in

Compare how 26 books explain false positive →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Grokking Machine Learning. The section links point to the parts of the book it draws on.