go-v.orgdoi

Intrusion Detection in Industrial Control Systems Using ML-Based Log Analysis

Doaa Mohammad Majed, Aya Falah, Vishal Jain · Digitalen Hochschulbibliothek Sachsen-Anhalt (Universitäts- und Landesbibliothek Sachsen-Anhalt) · 2026

  • DOI 10.25673/123736
  • Publisher Universitäts Frauenklinik
  • Type other
  • Issued 2026

Not cached yet. This record was fetched live for this request and is stored after the lookup. Reopening the page should serve the cached copy.

Digitalen Hochschulbibliothek Sachsen-Anhalt (Universitäts- und Landesbibliothek Sachsen-Anhalt) · Universitäts Frauenklinik · other · 2026

Abstract

The foundation of operational critical infrastructure is known as Industrial Control Systems (ICS) and due to their growing interconnection they have become targets of advanced cyber-attacks. The conventional intrusion detection systems (IDS) are limited because of being rule-based and are unable to identify zero-day exploits and subtle anomalies. This paper offers an intrusion detection system, which is a machine learning (ML) system that will examine ICS log data to detect the anomalies efficiently and accurately. The process includes systematized data gathering, preprocessing, feature discovery and mixed ML modeling by autoencoders and classifiers.

The results of the experiment prove that the proposed system is more precise, has higher recall, and AUC values compared to traditional approaches. The distribution of anomaly scores indicates the existence of a distinct boundary between normal behavior and attack behavior, whereas the analysis of the feature importance offers operational information on the important log parameters. The framework has an AUC of 0.984 indicative of its strength and capability to be used in real-time. Also, the architecture of the model is modular which facilitates future scalability and explainability.

The superiority of the system to other existing log based and network based IDS models is proved by comparative benchmarking. The research points out the opportunities of ML to improve the ICS cybersecurity using data-driven, adaptive, and explainable ways to do it.

Open the resolved locator · 10.25673/123736

Metadata licence: OpenAlex metadata (CC0)

Registry
OpenAlex DOI index + API
Fetched
2026-11-20 02:01 UTC
Cache state
not cached yet (bulk-index+openalex)
Locator
https://opendata.uni-halle.de//handle/1981185920/125670
Metadata licence
OpenAlex metadata (CC0)
Scope
Metadata as returned by the resolver; abstracts are never generated here.