A false negative is a classification error in which a model incorrectly predicts the negative class for an instance that actually belongs to the positive class. It is a fundamental cell of the confusion matrix, often denoted FN, and directly reduces recall (sensitivity). In high-stakes domains such as medical screening or fraud detection, false negatives represent missed true cases and frequently carry asymmetric cost relative to false positives.
Overview
- A false negative arises when a classifier assigns the negative label to an instance whose ground-truth label is positive. Within the confusion matrix it sits opposite the true positive cell, and the count of false negatives appears in the denominator of recall: recall = TP / (TP + FN). A model that minimises false negatives is described as having high sensitivity. The relative cost of a false negative versus a false positive depends entirely on the application: missing a malignant tumour or a fraudulent transaction is typically far more damaging than raising a false alarm, which motivates threshold tuning that trades precision for recall.
Key aspects
- Occupies the FN cell of the confusion matrix, opposite the true positive.
- Lowers recall and sensitivity without affecting precision directly.
- Cost is application-dependent and often asymmetric relative to false positives.
- Reduced by lowering the decision threshold or rebalancing class weights.
- Tracked alongside true positives, false positives and true negatives for full evaluation.
Applications
- Medical screening where a missed diagnosis is a false negative.
- Fraud and intrusion detection where undetected attacks are false negatives.
- Spam and content moderation pipelines balancing miss rate against over-blocking.
- Quality control and defect detection on production lines.