Usefulness of Unsupervised Ensemble Learning Methods for Time Series Forecasting of Aggregated or Clustered Load

  • Peter Laurinec
  • Mária Lucká
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10785)


This paper presents a comparison of the impact of various unsupervised ensemble learning methods on electricity load forecasting. The electricity load from consumers is simply aggregated or optimally clustered to more predictable groups by cluster analysis. The clustering approach consists of efficient preprocessing of data gained from smart meters by a model-based representation and the K-means method. We have implemented two types of ensemble learning methods to investigate the performance of forecasting on clustered or simply aggregated load: bootstrap aggregating based and the newly proposed clustering based. Two new bootstrap aggregating methods for time series analysis methods were newly proposed in order to handle the noisy behaviour of time series. The smart meter datasets used in our experiments come from Ireland and Slovakia, where data from more than 3600 consumers were available in both cases. The achieved results suggest that for extremely fluctuate and noisy time series unsupervised ensemble learning is not useful. We have proved that in most of the cases when the time series are regular, unsupervised ensemble learning for forecasting aggregated and clustered electricity load significantly improves accuracy.


Load forecasting Clustering Bagging Ensemble learning 



This work was partially supported by the Scientific Grant Agency of The Slovak Republic, Grant No. VG 1/0752/14 and STU Grant scheme for Support of Young Researchers.


  1. 1.
    Laurinec, P., Lucká, M.: Comparison of representations of time series for clustering smart meter data. In: Proceedings of WCECS, pp. 458–463 (2016)Google Scholar
  2. 2.
    Laurinec, P., Lóderer, M., Vrablecová, P., Lucká, M., Rozinajová, V., Ezzeddine, A.B.: Adaptive time series forecasting of energy consumption using optimized cluster analysis. In: Proceedings of IEEE ICDMW, pp. 398–405 (2016)Google Scholar
  3. 3.
    Adhikari, R., Verma, G., Khandelwal, I.: A model ranking based selective ensemble approach for time series forecasting. Procedia Comput. Sci. 48, 14–21 (2015)CrossRefGoogle Scholar
  4. 4.
    Shen, W., Babushkin, V., Aung, Z., Woon, W.L.: An ensemble model for day-ahead electricity demand time series forecasting. In: e-Energy 2013, pp. 51–62. ACM (2013)Google Scholar
  5. 5.
    Grmanová, G., Laurinec, P., Rozinajová, V., Ezzedine, A.B., Lucká, M., Lacko, P., Vrablecová, P., Návrat, P.: Incremental ensemble learning for electricity load forecasting. Acta Polytech. Hung. 13(2), 97–117 (2016)Google Scholar
  6. 6.
    Shahzadeh, A., Khosravi, A., Nahavandi, S.: Improving load forecast accuracy by clustering consumers using smart meter data. In: Proceedings of IJCNN (2015)Google Scholar
  7. 7.
    Wijaya, T.K., Vasirani, M., Humeau, S., Aberer, K.: Cluster-based aggregate forecasting for residential electricity demand using smart meter data. In: Proceedings of IEEE International Conference on Big Data, pp. 879–887 (2015)Google Scholar
  8. 8.
    Arthur, D., Vassilvitskii, S.: K-means++: the advantages of careful seeding. In: SODA 2007, pp. 1027–1035 (2007)Google Scholar
  9. 9.
    Davies, D.L., Bouldin, D.W.: A cluster separation measure. IEEE Trans. Pattern Anal. Mach. Intell. 2, 224–227 (1979)CrossRefGoogle Scholar
  10. 10.
    Cleveland, R.B., et al.: Seasonal-trend decomposition procedure based on LOESS. J. Off. Stat. 6, 3–73 (1990)Google Scholar
  11. 11.
    Box, G.E.P., Jenkins, G.M.: Time Series Analysis: Forecasting and Control. Holden-Day, San Francisco (1970)zbMATHGoogle Scholar
  12. 12.
    Holt, C.C.: Forecasting seasonals and trends by exponentially weighted moving averages, vol. 52. ONR Research Memorandum, Carnegie Inst. of Tech. (1957)Google Scholar
  13. 13.
    Hyndman, R.J., Koehler, A.B., Snyder, R.D., Grose, S.: A state space framework for automatic forecasting using exponential smoothing methods. Int. J. Forecast. 18(3), 439–454 (2002)CrossRefGoogle Scholar
  14. 14.
    Breiman, L., Friedman, J.H., Olshen, R.A., Stone, C.J.: Classification and Regression Trees. Chapman and Hall/CRC, Wadsworth (1984)zbMATHGoogle Scholar
  15. 15.
    Hyndman, R.J., Khandakar, Y.: Automatic time series forecasting: the forecast package for R. J. Stat. Softw. 27(3), 1–22 (2008)CrossRefGoogle Scholar
  16. 16.
    Strasser, H., Weber, C.: On the asymptotic theory of permutation statistics. Math. Methods Stat. 8, 220–250 (1999)MathSciNetzbMATHGoogle Scholar
  17. 17.
    Breiman, L.: Bagging predictors. Mach. Learn. 24(2), 123–140 (1996)MathSciNetzbMATHGoogle Scholar
  18. 18.
    Bergmeir, C., Hyndman, R.J., Benítez, J.M.: Bagging exponential smoothing methods using STL decomposition and Box-Cox transformation. Int. J. Forecast. 32(2), 303–312 (2016)CrossRefGoogle Scholar
  19. 19.
    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 KDD, pp. 226–231 (1996)Google Scholar
  20. 20.
    Ankerst, M., Breunig, M.M., Kriegel, H.P., Sander, J.: OPTICS: ordering points to identify the clustering structure. In: Proceedings of ACM SIGMOD, pp. 49–60 (1999)Google Scholar

Copyright information

© Springer International Publishing AG, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Faculty of Informatics and Information TechnologiesSlovak University of Technology in BratislavaBratislavaSlovak Republic

Personalised recommendations