Deep learning versus traditional machine learning methods for aggregated energy demand prediction

N.G. Paterakis, E. Mocanu, M. Gibescu, B. Stappers, W. van Alst

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

10 Citations (Scopus)
8 Downloads (Pure)

Abstract

In this paper the more advanced, in comparison with traditional machine learning approaches, deep learning methods are explored with the purpose of accurately predicting the aggregated energy consumption. Despite the fact that a wide range of machine learning methods have been applied to probabilistic energy prediction, the deep learning ones certainly represent the state-of-the-art artificial intelligence methods with remarkable success in a spectrum of practical applications. In particular, the use of Multi Layer Perceptrons, recently enhanced with deep learning capabilities, is proposed. Furthermore, its performance is compared with the most commonly used machine learning methods, such as Support Vector Machines, Gaussian Processes, Regression Trees, Ensemble Boosting and Linear Regression. The analysis of the day-ahead energy prediction demonstrates that different prediction methods present significantly different levels of accuracy in the case of a challenging dataset that comprises an interesting mix of consumers, wind and solar generation. The results show that Multi Layer Perceptrons outperform all the eight methods used as a benchmark in this study.

Original languageEnglish
Title of host publication2017 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2017 - Proceedings
Place of PublicationPiscataway
PublisherInstitute of Electrical and Electronics Engineers
Pages1-6
Number of pages6
ISBN (Electronic)9781538619537
DOIs
Publication statusPublished - 16 Jan 2018
Event2017 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe 2017) - Politecnico di Torino, Torino, Italy
Duration: 26 Sep 201729 Sep 2017
Conference number: 7
http://sites.ieee.org/isgt-europe-2017/

Conference

Conference2017 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe 2017)
Abbreviated titleISGT Europe 2017
CountryItaly
CityTorino
Period26/09/1729/09/17
Internet address

Fingerprint

Learning systems
Multilayer neural networks
Linear regression
Artificial intelligence
Support vector machines
Energy utilization
Deep learning

Keywords

  • deep learning
  • energy consumption
  • energy prediction
  • forecasting
  • machine learning

Cite this

Paterakis, N. G., Mocanu, E., Gibescu, M., Stappers, B., & van Alst, W. (2018). Deep learning versus traditional machine learning methods for aggregated energy demand prediction. In 2017 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2017 - Proceedings (pp. 1-6). Piscataway: Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ISGTEurope.2017.8260289
Paterakis, N.G. ; Mocanu, E. ; Gibescu, M. ; Stappers, B. ; van Alst, W. / Deep learning versus traditional machine learning methods for aggregated energy demand prediction. 2017 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2017 - Proceedings. Piscataway : Institute of Electrical and Electronics Engineers, 2018. pp. 1-6
@inproceedings{c81c67e417334b0e872e75a35e82e95f,
title = "Deep learning versus traditional machine learning methods for aggregated energy demand prediction",
abstract = "In this paper the more advanced, in comparison with traditional machine learning approaches, deep learning methods are explored with the purpose of accurately predicting the aggregated energy consumption. Despite the fact that a wide range of machine learning methods have been applied to probabilistic energy prediction, the deep learning ones certainly represent the state-of-the-art artificial intelligence methods with remarkable success in a spectrum of practical applications. In particular, the use of Multi Layer Perceptrons, recently enhanced with deep learning capabilities, is proposed. Furthermore, its performance is compared with the most commonly used machine learning methods, such as Support Vector Machines, Gaussian Processes, Regression Trees, Ensemble Boosting and Linear Regression. The analysis of the day-ahead energy prediction demonstrates that different prediction methods present significantly different levels of accuracy in the case of a challenging dataset that comprises an interesting mix of consumers, wind and solar generation. The results show that Multi Layer Perceptrons outperform all the eight methods used as a benchmark in this study.",
keywords = "deep learning, energy consumption, energy prediction, forecasting, machine learning",
author = "N.G. Paterakis and E. Mocanu and M. Gibescu and B. Stappers and {van Alst}, W.",
year = "2018",
month = "1",
day = "16",
doi = "10.1109/ISGTEurope.2017.8260289",
language = "English",
pages = "1--6",
booktitle = "2017 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2017 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers",
address = "United States",

}

Paterakis, NG, Mocanu, E, Gibescu, M, Stappers, B & van Alst, W 2018, Deep learning versus traditional machine learning methods for aggregated energy demand prediction. in 2017 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2017 - Proceedings. Institute of Electrical and Electronics Engineers, Piscataway, pp. 1-6, 2017 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe 2017), Torino, Italy, 26/09/17. https://doi.org/10.1109/ISGTEurope.2017.8260289

Deep learning versus traditional machine learning methods for aggregated energy demand prediction. / Paterakis, N.G.; Mocanu, E.; Gibescu, M.; Stappers, B.; van Alst, W.

2017 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2017 - Proceedings. Piscataway : Institute of Electrical and Electronics Engineers, 2018. p. 1-6.

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

TY - GEN

T1 - Deep learning versus traditional machine learning methods for aggregated energy demand prediction

AU - Paterakis, N.G.

AU - Mocanu, E.

AU - Gibescu, M.

AU - Stappers, B.

AU - van Alst, W.

PY - 2018/1/16

Y1 - 2018/1/16

N2 - In this paper the more advanced, in comparison with traditional machine learning approaches, deep learning methods are explored with the purpose of accurately predicting the aggregated energy consumption. Despite the fact that a wide range of machine learning methods have been applied to probabilistic energy prediction, the deep learning ones certainly represent the state-of-the-art artificial intelligence methods with remarkable success in a spectrum of practical applications. In particular, the use of Multi Layer Perceptrons, recently enhanced with deep learning capabilities, is proposed. Furthermore, its performance is compared with the most commonly used machine learning methods, such as Support Vector Machines, Gaussian Processes, Regression Trees, Ensemble Boosting and Linear Regression. The analysis of the day-ahead energy prediction demonstrates that different prediction methods present significantly different levels of accuracy in the case of a challenging dataset that comprises an interesting mix of consumers, wind and solar generation. The results show that Multi Layer Perceptrons outperform all the eight methods used as a benchmark in this study.

AB - In this paper the more advanced, in comparison with traditional machine learning approaches, deep learning methods are explored with the purpose of accurately predicting the aggregated energy consumption. Despite the fact that a wide range of machine learning methods have been applied to probabilistic energy prediction, the deep learning ones certainly represent the state-of-the-art artificial intelligence methods with remarkable success in a spectrum of practical applications. In particular, the use of Multi Layer Perceptrons, recently enhanced with deep learning capabilities, is proposed. Furthermore, its performance is compared with the most commonly used machine learning methods, such as Support Vector Machines, Gaussian Processes, Regression Trees, Ensemble Boosting and Linear Regression. The analysis of the day-ahead energy prediction demonstrates that different prediction methods present significantly different levels of accuracy in the case of a challenging dataset that comprises an interesting mix of consumers, wind and solar generation. The results show that Multi Layer Perceptrons outperform all the eight methods used as a benchmark in this study.

KW - deep learning

KW - energy consumption

KW - energy prediction

KW - forecasting

KW - machine learning

UR - http://www.scopus.com/inward/record.url?scp=85046290718&partnerID=8YFLogxK

U2 - 10.1109/ISGTEurope.2017.8260289

DO - 10.1109/ISGTEurope.2017.8260289

M3 - Conference contribution

AN - SCOPUS:85046290718

SP - 1

EP - 6

BT - 2017 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2017 - Proceedings

PB - Institute of Electrical and Electronics Engineers

CY - Piscataway

ER -

Paterakis NG, Mocanu E, Gibescu M, Stappers B, van Alst W. Deep learning versus traditional machine learning methods for aggregated energy demand prediction. In 2017 IEEE PES Innovative Smart Grid Technologies Conference Europe, ISGT-Europe 2017 - Proceedings. Piscataway: Institute of Electrical and Electronics Engineers. 2018. p. 1-6 https://doi.org/10.1109/ISGTEurope.2017.8260289