Anomaly Detection: Different Machine Learning Techniques, A Review

Abstract: Anomaly is something that deviates from normal, standard or unexpected. Anomaly detection in different applications has its own importance. The most common reason is that the unexpected behaviour always results in some kind of loss - it can either be theft of important data or damage to the system itself. Many anomaly detection techniques have been developed specific to application or to data. In this paper we have compiled a few machine learning algorithms that can be used for anomaly detection which can help researchers to select a particular algorithm for anomaly detection Keywords: anomaly, anomaly detection, intrusion detection, outlier, supervised, unsupervised