Methods for Detecting Anomalies in Network Traffic based on One-Class SVM Technology

This article presents research and application of the One-Class Support Vector Machines (One-Class SVM) method for detecting anomalies in network traffic. The paper provides a comprehensive overview of network anomaly detection challenges, introduces a methodological framework for applying One-Class SVM, presents experimental results using the CICIDS2017 dataset, and discusses the performance metrics and practical implications of the proposed approach. The research demonstrates that One-Class SVM achieves high accuracy in identifying both known and previously unseen network anomalies without requiring examples of malicious activity at the training stage.