Title :
Review of novelty detection methods
Author :
Dubravko Miljković
Author_Institution :
Hrvatska elektroprivreda, Vukovarska 37, 10000 Zagreb, Croatia
Abstract :
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.
Keywords :
"Neurons","Taxonomy","Training data","Signal processing","Learning systems","Algorithm design and analysis","Phase detection","Data preprocessing","Feature extraction","Bellows"
Conference_Titel :
MIPRO, 2010 Proceedings of the 33rd International Convention
Print_ISBN :
978-1-4244-7763-0
