Fuzzy logic inference for unsupervised anomaly detection
Tetiana Gladkykh, Taras Hnot, Volodymyr Solskyy · 2016
We are proposing the solution for unsupervised anomaly detection, which allows to detect unexpected activity of user or network equipment, based on the analysis of mutual dependencies of the separate slices of network activity. Proposed solution based on automation building of Fuzzy Logic Inference System, that describe general patterns of analyzed activity and is the development of Association Rules based approach. As the part of complex solution, this model can be used for discovering both typical network activity anomalies and new elements of anomaly.