Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks

Tae Young Kim, Sung Bae Cho

Research output: Chapter in Book/Report/Conference proceedingConference contribution

40 Citations (Scopus)


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 languageEnglish
Title of host publicationIntelligent Data Engineering and Automated Learning – IDEAL 2018 - 19th International Conference, Proceedings
EditorsHujun Yin, Paulo Novais, David Camacho, Antonio J. Tallón-Ballesteros
PublisherSpringer Verlag
Number of pages10
ISBN (Print)9783030034924
Publication statusPublished - 2018
Event19th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2018 - Madrid, Spain
Duration: 2018 Nov 212018 Nov 23

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11314 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other19th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2018

Bibliographical note

Publisher Copyright:
© 2018, Springer Nature Switzerland AG.

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • General Computer Science


Dive into the research topics of 'Predicting the Household Power Consumption Using CNN-LSTM Hybrid Networks'. Together they form a unique fingerprint.

Cite this