Dataset Analysis for Anomaly Detection on Critical Infrastructures

Anomaly Detection techniques allow to create robust security measures that provides early detection and are able to identify novel attacks that could not be prevented otherwise. Datasets represent a critical component in the process of designing and evaluating any kind of anomaly detection method. For this reason, in this paper we present the evaluation of two datasets showing the dependencies that arise between the techniques employed and the dataset itself. We also describe the characteristics that have to be taken into account while selecting a dataset to evaluate a detection algorithm in a critical infrastructure context.