Find in Library

Anomaly detection is an important aspect of data mining, where the main objective is to identify anomalousor unusual data from a given dataset. However, there is no formal categorization of application-specificanomaly detection techniques for big data and this ignites a confusion for the data miners. In this paper, wecategorise anomaly detection techniques based on nearest neighbours, clustering and statistical approachesand investigate the performance analysis of these techniques in critical infrastructure applications such asSCADA systems. Extensive experimental analysis is conducted to compare representative algorithms fromeach of the categories using seven benchmark datasets (both real and simulated) in SCADA systems. Theeffectiveness of the representative algorithms is measured through a number of metrics. We highlighted theset of algorithms that are the best performing for SCADA systems.