A Comparative Study of Time Series Anomaly Detection Models for Industrial Control Systems

Kim, Bedeuro; Alawami, Mohsen Ali; Kim, Eun-Soo; Oh, Sanghak; Park, Jeongyong; Kim, Hyoungshick

doi:10.3390/s23031310

Cited by 48 publications

(22 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [23], the authored proposed a complex method for detecting anomaly from real-time data using recurrence and fractal analysis. In [24], the authors conducted a comparative analysis of five time-series anomaly detection models. In [25],…”

Section: Of 20mentioning

confidence: 99%

See 1 more Smart Citation

An Intuitionistic Fuzzy Rough Set Based Classification for Anomaly Detection

Mazarbhuiya¹,

Sahmoudi²

2023

Preprint

View full text Add to dashboard Cite

The challenging issues of Computer Network and Databases are not only the intrusion detection but also the reduction of false positive and increase of detection rate. In any intrusion detection system, anomaly detection mainly focuses on modeling the normal behavior of the users and detecting the deviations from normal behavior which are assumed to be potential intrusions or treat. Several techniques have already been successfully tried for this purpose. However, the normal and suspicious behavior are hard to predict as there is no precise boundary differentiat-ing one from another. Here rough set theory and fuzzy set theory come into the picture. In this article, a hybrid approach based on rough set theory and intuitionistic fuzzy set theory is pro-posed for the detection of anomaly. The proposed approach is a classification approach which takes the advantages of softness properties both rough and fuzzy set theory to deal with uncer-tainty in the dataset. The algorithm classifies the data instances in such a way that they can be expressed using natural language. The experimental results with a real world dataset and a syn-thetic dataset show that the proposed algorithm has normal true positive rates of 91.989% and 96.99% and attack true positive rates of 91.289% and 96.29% respectively

show abstract

Section: Of 20mentioning

confidence: 99%

“…Kim et al [24], conducted a comparative analysis of five time-series anomaly detection models. In [25], the authors applied an ensemble learning model to analyze and forecast anomaly of the enormous system logs.…”

Section: Related Workmentioning

confidence: 99%

An Intuitionistic Fuzzy Rough Set Based Classification for Anomaly Detection

Mazarbhuiya¹,

Sahmoudi²

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Faced with vast amounts of monitoring data and numerous metrics, traditional static anomaly thresholding and manual filtering methods have become impractical(W. Zhang, Jia, Zhu, & Yan, 2017). Modern industrial anomaly detection has shifted towards artificial intelligence, where machines autonomously perform equipment state monitoring and anomaly detection, eliminating manual intervention (Kim et al, 2023;Shaoke, Xiaohu, Yanjing, & Jun, 2021).…”

Section: Introductionmentioning

confidence: 99%

Industrial multivariate time-series data anomaly detection incorporating attention mechanisms and adversarial training

Yang,

Chu,

et al. 2024

Preprint

View full text Add to dashboard Cite

To address the challenges faced in industrial anomaly detection, including data sample imbalance, lack of anomaly labels, and complex spatiotemporal relationships in high-dimensional data, this paper proposes a novel multi-modal time-series anomaly detection model that combines attention mechanisms and adversarial training. In this model, the first step involves utilizing graph attention mechanisms to extract sequence correlation features from multi-modal time-series data, which are then summed with the original data to form a dual-feature-based data representation. Subsequently, a self-supervised learning approach is employed to input this data representation into a variational autoencoder's encoding-decoding network for reconstruction. Anomaly detection is performed by analyzing the error between the input and reconstructed data. The model also employs spatiotemporal attention mechanisms and adversarial training during reconstruction to enhance feature extraction and model generalization. By comparing our proposed model to five commonly used baseline models, we demonstrate its effectiveness in detecting anomalies in scenarios involving high-dimensional data and imbalanced abnormal samples, demonstrating superior anomaly detection performance, as well as excellent performance on real industrial production and processing datasets.

show abstract

“…They are leveraging the advantages of unsupervised learning and featureless data to uncover unique temporal patterns. The problem of finding abnormalities in featureless data has drawn several researchers' attention [ 26 , 27 ] which, although resource-effective, might not be able to detect intricate patterns in time-series data. These novelty detection techniques frequently offer insightful information on comprehending temporal behaviours that are just beginning to occur.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised novelty detection for time series using a deep learning approach

Hossen,

Hoque,

Aziz

et al. 2024

Heliyon

View full text Add to dashboard Cite

A Comparative Study of Time Series Anomaly Detection Models for Industrial Control Systems

Cited by 48 publications

References 53 publications

An Intuitionistic Fuzzy Rough Set Based Classification for Anomaly Detection

An Intuitionistic Fuzzy Rough Set Based Classification for Anomaly Detection

Industrial multivariate time-series data anomaly detection incorporating attention mechanisms and adversarial training

Unsupervised novelty detection for time series using a deep learning approach

Contact Info

Product

Resources

About