Difference between Novelty Detection and Outlier Detection?
Updated May 5, 2026
Short answer
Outlier detection includes anomalies in training; novelty does not[cite: 1].
Deep explanation
Outlier detection assumes training data is 'dirty'. Novelty detection assumes training data is 'pure' (normal only)[cite: 1].
Real-world example
Detecting a new animal species in a forest (novelty)[cite: 1].
Common mistakes
- Training a novelty detector on data containing anomalies[cite: 1].
Follow-up questions
- Which is harder?