A framework for tunable anomaly detection

Md Rakibul Alam, Ilias Gerostathopoulos, Christian Prehofer, Alessandro Attanasi, Tomas Bures

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

Abstract

As software architecture practice relies more and more on runtime data to inform decisions in continuous experimentation and self-adaptation, it is increasingly important to consider the quality of the data used as input to the different decision-making and prediction algorithms. One issue in data-driven decisions is that real-life data coming from running systems can contain invalid or wrong values which can bias the result of data analysis. Data-driven decision-making should therefore comprise detection and handling of data anomalies as an integral part of the process. However, currently, anomaly detection is either absent in runtime decision-making approaches for continuous experimentation and self-adaptation or difficult to tailor to domain-specific needs. In this paper, we contribute by proposing a framework that simplifies the detection of data anomalies in timeseries-outputs of running systems. The framework is generic, since it can be employed in different domains, and tunable, since it uses expert user input in tailoring anomaly detection to the needs and assumptions of each domain. We evaluate the feasibility of the framework by successfully applying it to detecting anomalies in a real-life timeseries dataset from the traffic domain.

Original languageEnglish
Title of host publicationProceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages201-210
Number of pages10
ISBN (Electronic)9781728105284
DOIs
Publication statusPublished - 30 Apr 2019
Externally publishedYes
Event2019 IEEE International Conference on Software Architecture, ICSA 2019 - Hamburg, Germany
Duration: 25 Mar 201929 Mar 2019

Publication series

NameProceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019

Conference

Conference2019 IEEE International Conference on Software Architecture, ICSA 2019
CountryGermany
CityHamburg
Period25/03/1929/03/19

Fingerprint

Decision making
Software architecture

Keywords

  • Anomaly detection
  • Data anomalies
  • Data-driven decisions
  • Experimentation
  • Self-adaptive systems

Cite this

Alam, M. R., Gerostathopoulos, I., Prehofer, C., Attanasi, A., & Bures, T. (2019). A framework for tunable anomaly detection. In Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019 (pp. 201-210). [8703916] (Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICSA.2019.00029
Alam, Md Rakibul ; Gerostathopoulos, Ilias ; Prehofer, Christian ; Attanasi, Alessandro ; Bures, Tomas. / A framework for tunable anomaly detection. Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 201-210 (Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019).
@inproceedings{4bb15c193985471383da7845de8869d6,
title = "A framework for tunable anomaly detection",
abstract = "As software architecture practice relies more and more on runtime data to inform decisions in continuous experimentation and self-adaptation, it is increasingly important to consider the quality of the data used as input to the different decision-making and prediction algorithms. One issue in data-driven decisions is that real-life data coming from running systems can contain invalid or wrong values which can bias the result of data analysis. Data-driven decision-making should therefore comprise detection and handling of data anomalies as an integral part of the process. However, currently, anomaly detection is either absent in runtime decision-making approaches for continuous experimentation and self-adaptation or difficult to tailor to domain-specific needs. In this paper, we contribute by proposing a framework that simplifies the detection of data anomalies in timeseries-outputs of running systems. The framework is generic, since it can be employed in different domains, and tunable, since it uses expert user input in tailoring anomaly detection to the needs and assumptions of each domain. We evaluate the feasibility of the framework by successfully applying it to detecting anomalies in a real-life timeseries dataset from the traffic domain.",
keywords = "Anomaly detection, Data anomalies, Data-driven decisions, Experimentation, Self-adaptive systems",
author = "Alam, {Md Rakibul} and Ilias Gerostathopoulos and Christian Prehofer and Alessandro Attanasi and Tomas Bures",
year = "2019",
month = "4",
day = "30",
doi = "10.1109/ICSA.2019.00029",
language = "English",
series = "Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "201--210",
booktitle = "Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019",
address = "United States",

}

Alam, MR, Gerostathopoulos, I, Prehofer, C, Attanasi, A & Bures, T 2019, A framework for tunable anomaly detection. in Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019., 8703916, Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019, Institute of Electrical and Electronics Engineers Inc., pp. 201-210, 2019 IEEE International Conference on Software Architecture, ICSA 2019, Hamburg, Germany, 25/03/19. https://doi.org/10.1109/ICSA.2019.00029

A framework for tunable anomaly detection. / Alam, Md Rakibul; Gerostathopoulos, Ilias; Prehofer, Christian; Attanasi, Alessandro; Bures, Tomas.

Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019. Institute of Electrical and Electronics Engineers Inc., 2019. p. 201-210 8703916 (Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019).

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - A framework for tunable anomaly detection

AU - Alam, Md Rakibul

AU - Gerostathopoulos, Ilias

AU - Prehofer, Christian

AU - Attanasi, Alessandro

AU - Bures, Tomas

PY - 2019/4/30

Y1 - 2019/4/30

N2 - As software architecture practice relies more and more on runtime data to inform decisions in continuous experimentation and self-adaptation, it is increasingly important to consider the quality of the data used as input to the different decision-making and prediction algorithms. One issue in data-driven decisions is that real-life data coming from running systems can contain invalid or wrong values which can bias the result of data analysis. Data-driven decision-making should therefore comprise detection and handling of data anomalies as an integral part of the process. However, currently, anomaly detection is either absent in runtime decision-making approaches for continuous experimentation and self-adaptation or difficult to tailor to domain-specific needs. In this paper, we contribute by proposing a framework that simplifies the detection of data anomalies in timeseries-outputs of running systems. The framework is generic, since it can be employed in different domains, and tunable, since it uses expert user input in tailoring anomaly detection to the needs and assumptions of each domain. We evaluate the feasibility of the framework by successfully applying it to detecting anomalies in a real-life timeseries dataset from the traffic domain.

AB - As software architecture practice relies more and more on runtime data to inform decisions in continuous experimentation and self-adaptation, it is increasingly important to consider the quality of the data used as input to the different decision-making and prediction algorithms. One issue in data-driven decisions is that real-life data coming from running systems can contain invalid or wrong values which can bias the result of data analysis. Data-driven decision-making should therefore comprise detection and handling of data anomalies as an integral part of the process. However, currently, anomaly detection is either absent in runtime decision-making approaches for continuous experimentation and self-adaptation or difficult to tailor to domain-specific needs. In this paper, we contribute by proposing a framework that simplifies the detection of data anomalies in timeseries-outputs of running systems. The framework is generic, since it can be employed in different domains, and tunable, since it uses expert user input in tailoring anomaly detection to the needs and assumptions of each domain. We evaluate the feasibility of the framework by successfully applying it to detecting anomalies in a real-life timeseries dataset from the traffic domain.

KW - Anomaly detection

KW - Data anomalies

KW - Data-driven decisions

KW - Experimentation

KW - Self-adaptive systems

UR - http://www.scopus.com/inward/record.url?scp=85065791633&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85065791633&partnerID=8YFLogxK

U2 - 10.1109/ICSA.2019.00029

DO - 10.1109/ICSA.2019.00029

M3 - Conference contribution

T3 - Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019

SP - 201

EP - 210

BT - Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Alam MR, Gerostathopoulos I, Prehofer C, Attanasi A, Bures T. A framework for tunable anomaly detection. In Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019. Institute of Electrical and Electronics Engineers Inc. 2019. p. 201-210. 8703916. (Proceedings - 2019 IEEE International Conference on Software Architecture, ICSA 2019). https://doi.org/10.1109/ICSA.2019.00029