Advertisement

A Framework to Monitor Machine Learning Systems Using Concept Drift Detection

  • Xianzhe Zhou
  • Wally Lo FaroEmail author
  • Xiaoying Zhang
  • Ravi Santosh Arvapally
Conference paper
Part of the Lecture Notes in Business Information Processing book series (LNBIP, volume 353)

Abstract

As more and more machine learning based systems are being deployed in industry, monitoring of these systems is needed to ensure they perform in the expected way. In this article we present a framework for such a monitoring system. The proposed system is designed and deployed at Mastercard. This system monitors other machine learning systems that are deployed for use in production. The monitoring system performs concept drift detection by tracking the machine learning system’s inputs and outputs independently. Anomaly detection techniques are employed in the system to provide automatic alerts. We also present results that demonstrate the value of the framework. The monitoring system framework and the results are the main contributions in this article.

Keywords

Monitoring system Concept drift Machine learning Anomaly detection Framework 

References

  1. 1.
    Arvapally, R.S., Hicsasmaz, H., Lo Faro, W.: Artificial intelligence applied to challenges in the fields of operations and customer support. In: 2017 IEEE International Conference on Big Data (IEEE Big Data), Boston, pp. 3562–3569 (2017)Google Scholar
  2. 2.
    Gama, J., Zliobaite, L., Bifet, A., Pechenizkiy, M., Bouchachia, A.: A survey on concept drift adaptation. J. ACM Comput. Surv. (CSUR) 46, 44:1–44:37 (2014)zbMATHGoogle Scholar
  3. 3.
    Pastorello, G., et al.: Hunting data rouges at scale: data quality control for observational data in research infrastructures. In: 2017 IEEE 13th International Conference on e-Science (e-Science), Auckland, pp. 446–447 (2017)Google Scholar
  4. 4.
    Gamage, S., Premaratne, U.: Detecting and adapting to concept drift in continually evolving stochastic processes. In: ACM Proceedings of the International Conference on Big Data and Internet of Thing, London, pp. 109–114 (2017)Google Scholar
  5. 5.
    Webb, G., Lee, L.K., Goethals, B., Petitjean, F.: Analyzing concept drift and shift from sample data. J. Data Min. Knowl. Discov. 32(5), 1–21 (2018)MathSciNetGoogle Scholar
  6. 6.
    Gholipur, A., Hosseini, M.J., Beigy, H.: An adaptive regression tree for non-stationary data streams. In: Proceedings of the 28th Annual ACM Symposium on Applied Computing (ACM), Coimbra, pp. 815–817 (2013)Google Scholar
  7. 7.
    Jadhav, A., Deshpande, L.: An efficient approach to detect concept drifts in data streams. In: IEEE 7th International Advance Computing Conference (IEEE IACC), Hyderabad, pp. 28–32 (2017)Google Scholar
  8. 8.
    Chandola, V., Banerjee, A., Kumar, V.: Anomaly detection: a survey. ACM Comput. Surv. (CSUR) 41(3), 15 (2009)CrossRefGoogle Scholar
  9. 9.
    Ding, M., Tian, H.: PCA-based network traffic anomaly detection. J. Tsinghua Sci. Technol. 21(5), 500–509 (2016)CrossRefGoogle Scholar
  10. 10.
    Zhang, L., Veitch, D., Kotagiri, R.: The role of KL divergence in anomaly detection. In: Proceedings of the ACM SIGMETRICS Joint International Conference on Measurement and Modeling of Computer Systems, San Jose, pp. 123–124 (2011)Google Scholar
  11. 11.
    Laptev, N., Amizadeh, S., Flint, I.: Generic and scalable framework for automated time-series anomaly detection. In: Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (ACM), Sydney, pp. 1939–1947 (2015)Google Scholar
  12. 12.
    Pedregosa, F., et al.: Scikit-learn: machine learning in Python. J. Mach. Learn. Res. 12, 2825–2830 (2011)MathSciNetzbMATHGoogle Scholar
  13. 13.
    Ester, M., Kriegel, H.P., Sander, J., Xu, X.: A density based algorithm for discovering clusters in large spatial databases with noise. In: Proceedings of the Second International Conference on Knowledge Discovery and Data Mining, Portland, pp. 226–231 (1996)Google Scholar
  14. 14.
    Liu, F.T., Ting, K.M., Zhou, Z.H.: Isolation based anomaly detection. ACM Trans. Knowl. Discov. Data (TKDD) 6(1), 3 (2012)Google Scholar
  15. 15.
    Liu, A., Zhang, G., Lu, J.: Fuzzy time windowing for gradual concept drift adaptation. In: IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), Naples, pp. 1–6 (2017)Google Scholar
  16. 16.
    Geng, Y., Zhang, J.: An ensemble classifier algorithm for mining data streams based on concept drift. In: 10th International Symposium on Computational Intelligence and Design (ISCID), Hangzhou, pp. 227–230 (2017)Google Scholar
  17. 17.
    Hu, H., Kantardzic, M.M., Lyu, L.: Detecting different types of concept drifts with ensemble framework. In: 17th IEEE International Conference on Machine Learning and Applications (ICMLA), Orlando, pp. 344–350 (2018)Google Scholar
  18. 18.
    Senaratne, H., Broring, A., Schreck, T., Lehle, D.: Moving on Twitter: using episodic hotspot and drift analysis to detect and characterise spatial trajectories: In: 7th ACM SIGSPATIAL International Workshop on Location - Based Social Networks (LBSN), Dallas (2014)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Xianzhe Zhou
    • 1
  • Wally Lo Faro
    • 1
    Email author
  • Xiaoying Zhang
    • 1
  • Ravi Santosh Arvapally
    • 1
  1. 1.MastercardO’FallonUSA

Personalised recommendations