Forecasting annual energy consumption using machine learnings: Case of Indonesia

Robi Kurniawan, Shunsuke Managi

Research output: Contribution to journalConference articlepeer-review

1 Citation (Scopus)


To understand the future trajectory of energy consumption, we propose to utilize two different machine learning algorithm, artificial neural networks (ANN) and a model tree. Taking Indonesia as a case, the annual gross energy consumption was estimated by modelling a function of urbanization, real GDP per capita proxy for affluence (economic growth), and real capital use per capita. Utilizing the time period of 1971-2014, we train and test the model. Utilizing the root mean square error and the mean absolute error for model selection, we found the tree-based model has a better performance rather than the ANN. Having more superior performance, the tree-based model was then used to forecast the annual energy consumption for the future years. Using specific scenario, the energy consumption is predicted will increase from 883 kg per capita in 2014 to become 1243 kg per capita in 2040. Providing better accuracy, the approach applied in this study can easily be replicated for other countries. Furthermore, it also can be considered in simulating energy demand and environmental consequence in the future.

Original languageEnglish
Article number012032
JournalIOP Conference Series: Earth and Environmental Science
Issue number1
Publication statusPublished - May 10 2019
Event2019 9th International Conference on Future Environment and Energy, ICFEE 2019 - Osaka, Japan
Duration: Jan 9 2019Jan 11 2019

All Science Journal Classification (ASJC) codes

  • Environmental Science(all)
  • Earth and Planetary Sciences(all)


Dive into the research topics of 'Forecasting annual energy consumption using machine learnings: Case of Indonesia'. Together they form a unique fingerprint.

Cite this