摘要：Cyber threats are nowadays a major danger to critical infrastructures and to homeland security. For several years now, the focus have been targeted at the physical protection of critical infrastructures. Currently, experts realize that the critical infrastructure can be also attacked via the application layer of computer networks. In order to efficiently protect such critical systems, the huge amount of data has to be efficiently analyzed and correlated. Therefore, this paper focuses on the overview of the advanced data stream processing methods to be applied in the domain of cybersecurity and critical infrastructure protection. The major contribution of this work is the analysis of such innovative aspects as concept drift analysis deployed as the pre-processing step dedicated for anomaly detection systems to counter cyber attacks. Moreover, we discuss the different challenges in data streams analysis including data imbalance and provide solid reasoning why applying a concept drift detector is crucial when designing a modern cybersecurity systems.
关键词：anomaly detection; concept drift; cybersecurity; data imbalance; data science; data stream; machine learning
Loading...
