Detecting Temporal and Spatial Anomalies in Users’ Activities for Security Provisioning in Computer Networks
Aleks Huč · Informatica · 2023
The paper summarizes a Doctoral Thesis that focuses on two new approaches for detecting anomalies in computer networks based on network flows. The approaches use incremental hierarchical clustering algorithms and monitor changes in the data structures to detect anomalies. Both approaches achieved prediction performance comparable to the state-of-the-art supervised approaches (F1 score over 0.90), even when taking into account that our approaches see every data point only once and then discard it and they operate without the prerequisite learning phase with labeled data.