Network Flow Analysis for Unsupervised Anomaly Detection using Hierarchical Temporal Memory

Summary—The detection of anomalous users in network traffic flows constitutes a fundamental element of modern cybersecurity. This article analyzes the performance of the Hierarchical Temporal Memory (HTM) algorithm for identifying anomalies in computer networks using the UGR'16 data set. Using a detailed study of network flow metrics, the effectiveness of HTM is compared with conventional approaches, demonstrating its superior performance and flexibility in dynamic traffic scenarios.