PCA-Based Low-Complexity Anomaly Detection 


Vol. 46,  No. 6, pp. 941-955, Jun.  2021
10.7840/kics.2021.46.6.941


PDF
  Abstract

As the Internet of Things (IoT) technology has been rapidly developed, even low-end devices, such as smart light bulbs and sensors, are now able to be remotely operated via the Internet. The growth of the IoT industry leads to emerging open-source platforms, but they have a critical weakness of security vulnerabilities. Such security issue becomes much more important in low-end IoT devices. The conventional anomaly detection approaches could not be the solution in low-end devices because they have limited hardware capabilities, like low specifications of CPU and memory. Therefore, it is important to develop an anomaly detection algorithm with low computational complexity. In this paper, we propose a linear transform-based low-complexity anomaly detection solution that can be operated in low-end devices. Using principal component analysis (PCA), a principal subspace is built from normal datasets. Then, the principal components are obtained by projecting the collected IoT data onto the principal subspace. We further propose a modified Mahalanobis distance which detects anomalies from principal components. In simulation results, it is shown that the proposed solution outperforms existing methods even though it requires lower computational complexity than the others.

  Statistics
Cumulative Counts from November, 2022
Multiple requests among the same browser session are counted as one view. If you mouse over a chart, the values of data points will be shown.


  Cite this article

[IEEE Style]

H. Kye and M. Kwon, "PCA-Based Low-Complexity Anomaly Detection," The Journal of Korean Institute of Communications and Information Sciences, vol. 46, no. 6, pp. 941-955, 2021. DOI: 10.7840/kics.2021.46.6.941.

[ACM Style]

Hyoseon Kye and Minhae Kwon. 2021. PCA-Based Low-Complexity Anomaly Detection. The Journal of Korean Institute of Communications and Information Sciences, 46, 6, (2021), 941-955. DOI: 10.7840/kics.2021.46.6.941.

[KICS Style]

Hyoseon Kye and Minhae Kwon, "PCA-Based Low-Complexity Anomaly Detection," The Journal of Korean Institute of Communications and Information Sciences, vol. 46, no. 6, pp. 941-955, 6. 2021. (https://doi.org/10.7840/kics.2021.46.6.941)

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)






pISSN : 1226 - 4717
eISSN : 2287 - 3880