Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks

Tae Young Kim; Sung Bae Cho

doi:10.1007/978-3-030-03493-1_50

Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks

Tae Young Kim, Sung Bae Cho

College of Computing

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

31 Citations (Scopus)

Abstract

Prediction of power consumption is an integral part of the operation and planning of the electricity supply company. In terms of power supply and demand, For the stable supply of electricity, the reserve power must be prepared. However, it is necessary to predict electricity demand because electricity is difficult to store. In this paper, we propose a CNN-LSTM hybrid network that can extract spatio-temporal information to effectively predict the house power consumption. Experiments have shown that CNN-LSTM hybrid networks, which linearly combine convolutional neural network (CNN), long short-term memory (LSTM) and deep neural network (DNN), can extract irregular features of electric power consumption. The CNN layer is used to reduce the spectrum of spatial information, the LSTM layer is suitable for modeling temporal information, the DNN layer generates a predicted time series. The CNN-LSTM hybrid approach almost completely predicts power consumption. Finally, the CNN-LSTM hybrid method achieves higher root mean square error (RMSE) than traditional predictive methods for the individual household power consumption data sets provided by the UCI repository.

Original language	English
Title of host publication	Intelligent Data Engineering and Automated Learning – IDEAL 2018 - 19th International Conference, Proceedings
Editors	Hujun Yin, Paulo Novais, David Camacho, Antonio J. Tallón-Ballesteros
Publisher	Springer Verlag
Pages	481-490
Number of pages	10
ISBN (Print)	9783030034924
DOIs	https://doi.org/10.1007/978-3-030-03493-1_50
Publication status	Published - 2018
Event	19th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2018 - Madrid, Spain Duration: 2018 Nov 21 → 2018 Nov 23

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11314 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	19th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2018
Country/Territory	Spain
City	Madrid
Period	18/11/21 → 18/11/23

Bibliographical note

Funding Information:
This research was supported by Korea Electric Power Corporation. (Grant number: R18XA05).

Funding Information:
This research was supported by Korea Electric Power Corporation. (Grant

Publisher Copyright:
© 2018, Springer Nature Switzerland AG.

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-030-03493-1_50

Cite this

Kim, T. Y., & Cho, S. B. (2018). Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks. In H. Yin, P. Novais, D. Camacho, & A. J. Tallón-Ballesteros (Eds.), Intelligent Data Engineering and Automated Learning – IDEAL 2018 - 19th International Conference, Proceedings (pp. 481-490). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11314 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-03493-1_50

Kim, Tae Young ; Cho, Sung Bae. / Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks. Intelligent Data Engineering and Automated Learning – IDEAL 2018 - 19th International Conference, Proceedings. editor / Hujun Yin ; Paulo Novais ; David Camacho ; Antonio J. Tallón-Ballesteros. Springer Verlag, 2018. pp. 481-490 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{2c34be11da4e408dba8f145b4b110f35,

title = "Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks",

abstract = "Prediction of power consumption is an integral part of the operation and planning of the electricity supply company. In terms of power supply and demand, For the stable supply of electricity, the reserve power must be prepared. However, it is necessary to predict electricity demand because electricity is difficult to store. In this paper, we propose a CNN-LSTM hybrid network that can extract spatio-temporal information to effectively predict the house power consumption. Experiments have shown that CNN-LSTM hybrid networks, which linearly combine convolutional neural network (CNN), long short-term memory (LSTM) and deep neural network (DNN), can extract irregular features of electric power consumption. The CNN layer is used to reduce the spectrum of spatial information, the LSTM layer is suitable for modeling temporal information, the DNN layer generates a predicted time series. The CNN-LSTM hybrid approach almost completely predicts power consumption. Finally, the CNN-LSTM hybrid method achieves higher root mean square error (RMSE) than traditional predictive methods for the individual household power consumption data sets provided by the UCI repository.",

author = "Kim, {Tae Young} and Cho, {Sung Bae}",

note = "Funding Information: This research was supported by Korea Electric Power Corporation. (Grant number: R18XA05). Funding Information: This research was supported by Korea Electric Power Corporation. (Grant Publisher Copyright: {\textcopyright} 2018, Springer Nature Switzerland AG.; 19th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2018 ; Conference date: 21-11-2018 Through 23-11-2018",

year = "2018",

doi = "10.1007/978-3-030-03493-1_50",

language = "English",

isbn = "9783030034924",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "481--490",

editor = "Hujun Yin and Paulo Novais and David Camacho and Tall{\'o}n-Ballesteros, {Antonio J.}",

booktitle = "Intelligent Data Engineering and Automated Learning – IDEAL 2018 - 19th International Conference, Proceedings",

address = "Germany",

}

Kim, TY & Cho, SB 2018, Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks. in H Yin, P Novais, D Camacho & AJ Tallón-Ballesteros (eds), Intelligent Data Engineering and Automated Learning – IDEAL 2018 - 19th International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11314 LNCS, Springer Verlag, pp. 481-490, 19th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2018, Madrid, Spain, 18/11/21. https://doi.org/10.1007/978-3-030-03493-1_50

Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks. / Kim, Tae Young; Cho, Sung Bae.

Intelligent Data Engineering and Automated Learning – IDEAL 2018 - 19th International Conference, Proceedings. ed. / Hujun Yin; Paulo Novais; David Camacho; Antonio J. Tallón-Ballesteros. Springer Verlag, 2018. p. 481-490 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11314 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks

AU - Kim, Tae Young

AU - Cho, Sung Bae

N1 - Funding Information: This research was supported by Korea Electric Power Corporation. (Grant number: R18XA05). Funding Information: This research was supported by Korea Electric Power Corporation. (Grant Publisher Copyright: © 2018, Springer Nature Switzerland AG.

PY - 2018

Y1 - 2018

N2 - Prediction of power consumption is an integral part of the operation and planning of the electricity supply company. In terms of power supply and demand, For the stable supply of electricity, the reserve power must be prepared. However, it is necessary to predict electricity demand because electricity is difficult to store. In this paper, we propose a CNN-LSTM hybrid network that can extract spatio-temporal information to effectively predict the house power consumption. Experiments have shown that CNN-LSTM hybrid networks, which linearly combine convolutional neural network (CNN), long short-term memory (LSTM) and deep neural network (DNN), can extract irregular features of electric power consumption. The CNN layer is used to reduce the spectrum of spatial information, the LSTM layer is suitable for modeling temporal information, the DNN layer generates a predicted time series. The CNN-LSTM hybrid approach almost completely predicts power consumption. Finally, the CNN-LSTM hybrid method achieves higher root mean square error (RMSE) than traditional predictive methods for the individual household power consumption data sets provided by the UCI repository.

AB - Prediction of power consumption is an integral part of the operation and planning of the electricity supply company. In terms of power supply and demand, For the stable supply of electricity, the reserve power must be prepared. However, it is necessary to predict electricity demand because electricity is difficult to store. In this paper, we propose a CNN-LSTM hybrid network that can extract spatio-temporal information to effectively predict the house power consumption. Experiments have shown that CNN-LSTM hybrid networks, which linearly combine convolutional neural network (CNN), long short-term memory (LSTM) and deep neural network (DNN), can extract irregular features of electric power consumption. The CNN layer is used to reduce the spectrum of spatial information, the LSTM layer is suitable for modeling temporal information, the DNN layer generates a predicted time series. The CNN-LSTM hybrid approach almost completely predicts power consumption. Finally, the CNN-LSTM hybrid method achieves higher root mean square error (RMSE) than traditional predictive methods for the individual household power consumption data sets provided by the UCI repository.

UR - http://www.scopus.com/inward/record.url?scp=85057083690&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85057083690&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-03493-1_50

DO - 10.1007/978-3-030-03493-1_50

M3 - Conference contribution

AN - SCOPUS:85057083690

SN - 9783030034924

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 481

EP - 490

BT - Intelligent Data Engineering and Automated Learning – IDEAL 2018 - 19th International Conference, Proceedings

A2 - Yin, Hujun

A2 - Novais, Paulo

A2 - Camacho, David

A2 - Tallón-Ballesteros, Antonio J.

PB - Springer Verlag

T2 - 19th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2018

Y2 - 21 November 2018 through 23 November 2018

ER -

Kim TY, Cho SB. Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks. In Yin H, Novais P, Camacho D, Tallón-Ballesteros AJ, editors, Intelligent Data Engineering and Automated Learning – IDEAL 2018 - 19th International Conference, Proceedings. Springer Verlag. 2018. p. 481-490. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-03493-1_50

Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks

Abstract

Publication series

Other

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this