liveBook · Manning

False positive

As explained in Grokking Machine Learning, Second Edition

A false positive is a classification outcome in which a model predicts the positive class for an example whose true class is negative.

A false positive occurs when a classification model predicts the positive class for an example whose true class is negative. It is an incorrect positive prediction. section 7.5 section 7.18

Everyday classification examples

In a medical model, a false positive is a healthy person incorrectly diagnosed as sick. In a spam filter, it is a legitimate ham email incorrectly classified as spam. The positive label depends on the task: sickness in the first example and spam in the second. section 7.5

Use the confusion matrix to distinguish this error from related outcomes:

  • False positive: the actual class is negative, but the prediction is positive. section 7.5
  • False negative: the actual class is positive, but the prediction is negative. section 7.5
  • True positive and true negative are correct positive and negative predictions. section 7.5

False positives matter because different applications assign different costs to errors. Precision falls when a model produces many false positives, while recall focuses on finding positive cases and tolerates more mistaken positives. section 7.18 section 7.7

Figure 7.1 Examples of two models that are widely used in real life and that we’ll use throughout this chapter. On the left, a coronavirus model where the people are diagnosed as healthy or sick. On the right, a spam-detection model where the emails are classified as spam or ham. For each model, we have highlighted some of their errors and separated them as false positives and false negatives.

Compare false positives and false negatives visually in medical diagnosis and spam detection. section 7.5

How Grokking Machine Learning, Second Edition covers it

How deep it goes
Defines the error, contrasts it with three confusion-matrix outcomes, and relates it to precision and recall.
Examples
Uses coronavirus diagnosis, spam filtering, recommendation systems, pedestrian detection, and voice assistants as error-cost examples.
What it assumes
Assumes binary positive and negative labels while defining these labels and prediction outcomes from zero.
What you can do after
Evaluate classification results by locating false positives and deciding when precision should receive greater emphasis.
How it connects
Connects false positives with false negatives, true positives, true negatives, confusion matrices, precision, recall, and F-scores.

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, Second Edition. The section links point to the parts of the book it draws on.