Empirical Evaluation of Autoencoder Models for Anomaly Detection in Packet-based NIDS

• DOI: 10.36227/techrxiv.24043608
• Publication Date: 2023-08-31T04:25:13Z
• Authors: Soumyadeep Hore; Quoc Nguyen; Yulun Xu; Ankit Shah; Nathaniel Bastian; Trung Le
• Keywords: Autoencoder; Anomaly (physics); Performance metric

<p>Packet-based network intrusion detection systems (NIDS) allow for real-time detection, making this research area crucial. This study compares autoencoder models anomaly in packet-based NIDS. It presents a framework implementing an autoencoder-based NIDS using packet data. A novel metric reconstruction error autoencoders is introduced. evaluated at different thresholds to compare how accurately it detects traffic anomalies. The efficacy of showcased across various attacks and adversarial samples from public data sets. analysis emphasizes the strengths limitations detecting anomalies traffic. insights provide valuable guidance researchers practitioners developing mechanisms.</p>