University of Twente Student Theses

Login
This website will be unavailable due to maintenance December 1st between 8:00 and 12:00 CET.

Unsupervised Anomaly-Based Network Intrusion Detection Using Auto Encoders for Practical Use

Keijer, BSc J.S. (2021) Unsupervised Anomaly-Based Network Intrusion Detection Using Auto Encoders for Practical Use.

[img] PDF
5MB
Abstract:In academic literature, many intrusion detection systems are proposed, both signature and anomaly based. In commercial solutions, signature-based detection is the norm. This type of detection does not provide defense against types of attacks not seen before or unseen variations of known attacks. With the rise of zero-day attacks and evolving types of attacks, signature-based detection does not give complete protection. For anomaly-based detection academic works propose both supervised and unsupervised methods. Since it is often unfeasible to create labeled datasets in a commercial setting, unsupervised methods are the most likely solution. In literature on unsupervised methods, Auto Encoders show results which could make them viable for commercial production. This work aims to address the lack of implementation of anomaly-based methods in commercial use. For this, a method is proposed for processing network data to increase the performance and efficiency of Auto Encoders, for unsupervised anomaly-based intrusion detection.
Item Type:Essay (Master)
Clients:
Northwave Nederland B.V., Utrecht, Netherlands
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science MSc (60300)
Link to this item:https://purl.utwente.nl/essays/86521
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page