TSAGen : Synthetic Time Series Generation for KPI Anomaly Detection

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review

16 Scopus Citations
View graph of relations

Author(s)

  • Chengyu Wang
  • Kui Wu
  • Tongqing Zhou
  • Guang Yu
  • Zhiping Cai

Detail(s)

Original languageEnglish
Pages (from-to)130-145
Journal / PublicationIEEE Transactions on Network and Service Management
Volume19
Issue number1
Online published21 Jul 2021
Publication statusPublished - Mar 2022
Externally publishedYes

Abstract

A key performance indicator (KPI) consists of critical time series data that reflect the runtime states of network systems (e.g., response time and available bandwidth). Despite the importance of KPI, datasets for KPI anomaly detection available to the public are very limited, due to privacy concerns and the high overhead in manually labelling the data. The insufficiency of public KPI data poses a great barrier for network researchers and practitioners to evaluate and test what-if scenarios in the development of artificial intelligence for IT operations (AIOps) and anomaly detection algorithms. To tackle the difficulty, we develop a univariate time series generation tool called TSAGen, which can generate KPI data with anomalies and controllable characteristics for KPI anomaly detection. Experiment results show that the data generated by TSAGen can be used for comprehensive evaluation of anomaly detection algorithms with diverse user-defined what-if scenarios.

© 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.

Research Area(s)

  • AIOps, fault injection, time series anomaly detection, Time series generation

Citation Format(s)

TSAGen: Synthetic Time Series Generation for KPI Anomaly Detection. / Wang, Chengyu; Wu, Kui; Zhou, Tongqing et al.
In: IEEE Transactions on Network and Service Management, Vol. 19, No. 1, 03.2022, p. 130-145.

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review