Glossary term
Glossary term
Evaluation and Benchmarks
The process of determining whether a new (novel) example comes from the same distribution as the training set. In other words, after training on the training set, novelty detection determines whether a new example (during inference or during additional training) is an outlier.
Contrast with outlier detection.
Created for this library
A manufacturing team uses novelty detection on sensor streams to flag equipment behaviors that were not seen in historical training data.
A cybersecurity team uses novelty detection on network traffic to surface unusual patterns that might indicate a new attack pattern.
A retail analytics team uses novelty detection on customer behaviors to surface new shopping patterns that the segmentation model has not seen before.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License