Title: Predicting world electricity generation by sources using different machine learning algorithms

Authors: Mehmet Hakan Özdemir; Batin Latif Aylak; Murat İnce; Okan Oral

Addresses: Department of Business Administration, Turkish-German University, Istanbul, Turkey ' Department of Industrial Engineering, Turkish-German University, Istanbul, Turkey ' Department of Computer Technologies, Isparta University of Applied Sciences, Isparta, Turkey ' Department of Mechatronics Engineering, Akdeniz University, Antalya, Turkey

Abstract: Electrical energy plays a crucial role in both social and economic growth. It is thought to be an essential part of industrial manufacturing. In addition to its contribution to industry, electrical energy is essential for addressing the needs of people on a daily basis. Therefore, electricity generation prediction is crucial for accurate electricity planning and energy usage, with machine learning (ML) algorithms becoming popular for their ability to extract complex relationships and make precise predictions. With the data from the period 2000-2022, this study predicts world electricity generation for 2023 by different energy sources employing seven different ML algorithms, namely long short-term memory (LSTM), artificial neural network (ANN), linear regression (LR), support vector regression (SVR), decision tree regression (DTR), random forest regression (RFR) and eXtreme gradient boosting (XGBoost). The algorithms were also contrasted in the study, and it was discovered that LSTM produced the most accurate predictions. [Received: June 16, 2023; Accepted: August 19, 2023]

Keywords: energy; electricity generation; machine learning; prediction; long short-term memory; LSTM; artificial neural network; ANN; support vector regression; SVR; decision tree regression; DTR; random forest regression; RFR.

DOI: 10.1504/IJOGCT.2024.136028

International Journal of Oil, Gas and Coal Technology, 2024 Vol.35 No.1, pp.98 - 115

Received: 14 Jun 2023
Accepted: 19 Aug 2023
Published online: 12 Jan 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article