Next Article in Journal
Emerging Trends and Challenges in Fog and Edge Computing for the Internet of Things
Previous Article in Journal
Acknowledgment to Reviewers of IoT in 2021
Article

Assessing Feature Representations for Instance-Based Cross-Domain Anomaly Detection in Cloud Services Univariate Time Series Data

1
ADAPT Research Centre, Trinity College Dublin, D02 W272 Dublin, Ireland
2
ADAPT Research Centre, Dublin City University, D09 RFK0 Dublin, Ireland
3
Huawei Research, D02 R156 Dublin, Ireland
4
ADAPT Research Centre, Technological University Dublin, D07 EWV4 Dublin, Ireland
*
Author to whom correspondence should be addressed.
These authors contributed equally to this work.
These authors contributed equally to this work.
Academic Editor: Javier Berrocal
IoT 2022, 3(1), 123-144; https://doi.org/10.3390/iot3010008
Received: 7 December 2021 / Revised: 30 December 2021 / Accepted: 10 January 2022 / Published: 29 January 2022
In this paper, we compare and assess the efficacy of a number of time-series instance feature representations for anomaly detection. To assess whether there are statistically significant differences between different feature representations for anomaly detection in a time series, we calculate and compare confidence intervals on the average performance of different feature sets across a number of different model types and cross-domain time-series datasets. Our results indicate that the catch22 time-series feature set augmented with features based on rolling mean and variance performs best on average, and that the difference in performance between this feature set and the next best feature set is statistically significant. Furthermore, our analysis of the features used by the most successful model indicates that features related to mean and variance are the most informative for anomaly detection. We also find that features based on model forecast errors are useful for anomaly detection for some but not all datasets. View Full-Text
Keywords: time series analysis; anomaly detection; data representation; cloud monitoring; AIOPS time series analysis; anomaly detection; data representation; cloud monitoring; AIOPS
Show Figures

Figure 1

MDPI and ACS Style

Agrahari, R.; Nicholson, M.; Conran, C.; Assem, H.; Kelleher, J.D. Assessing Feature Representations for Instance-Based Cross-Domain Anomaly Detection in Cloud Services Univariate Time Series Data. IoT 2022, 3, 123-144. https://doi.org/10.3390/iot3010008

AMA Style

Agrahari R, Nicholson M, Conran C, Assem H, Kelleher JD. Assessing Feature Representations for Instance-Based Cross-Domain Anomaly Detection in Cloud Services Univariate Time Series Data. IoT. 2022; 3(1):123-144. https://doi.org/10.3390/iot3010008

Chicago/Turabian Style

Agrahari, Rahul, Matthew Nicholson, Clare Conran, Haytham Assem, and John D. Kelleher. 2022. "Assessing Feature Representations for Instance-Based Cross-Domain Anomaly Detection in Cloud Services Univariate Time Series Data" IoT 3, no. 1: 123-144. https://doi.org/10.3390/iot3010008

Find Other Styles

Article Access Map by Country/Region

1
Back to TopTop