Review of novelty detection methods

Novelty detection is the identification of new or unknown data or signals that a machine learning system is not aware of during training. Novelty detection methods try to identify outliers that differ from the distribution of ordinary data. This paper is a short review of novelty detection and its methods.