Forecasting natural gas consumption in Turkey using fractional non-linear grey Bernoulli model optimized by grey wolf optimization (GWO) algorithm
Özcan, Tuncay and Konyalioglu, Aziz Kemal and Apaydın, Tuğçe Beldek (2024) Forecasting natural gas consumption in Turkey using fractional non-linear grey Bernoulli model optimized by grey wolf optimization (GWO) algorithm. Euro-Mediterranean Journal for Environmental Integration. ISSN 2365-7448 (https://doi.org/10.1007/s41207-024-00618-9)
Preview |
Text.
Filename: Ozcan-etal-2024-EMJEI-Forecasting-natural-gas-consumption-in-Turkey-using-fractional-non-linear-grey-Bernoulli-model_.pdf
Accepted Author Manuscript License: Download (1MB)| Preview |
Abstract
Natural gas stands as an indispensable energy source, integrated to the daily operations of countries worldwide, serving as a primary energy input for various industries, homes, and sectors. The predominant driver behind the escalating trend in natural gas consumption is rooted in its distinctive environmental profile, characterized by a relatively lower carbon emissions footprint. Recognized as the most environmentally friendly among fossil fuels, natural gas has become the preferred choice, reflecting a conscious effort to mitigate environmental impact and promote sustainability in energy consumption patterns in the world. Especially, in developing countries like Turkey, effective management of energy resources and the formulation of policies centered on the production and consumption of natural gas necessitate accurate forecasting. This study, thus, focuses on forecasting natural gas consumption in Turkey, employing the Fractional Nonlinear Grey Bernoulli Model (FANGBM(1,1)) optimized by Grey Wolf Optimizer (GWO). First, the parameters are optimized using GWO for an accurate forecasting to be used through the metaheuristic model FANGBM(1,1). After using GWO-FANGBM(1,1) model to forecast natural gas consumption in Turkey, a comparative study has been performed including GM(1,1) and GWO-GM(1,1). The predictive performance of these models is compared with ARIMA and linear regression. Notably, numerical results reveal that the proposed hybrid model GWO-FANGBM(1,1) model surpasses other grey models, such as GM(1,1) and GWO-GM(1,1), as well as statistical methods like ARIMA and linear regression. Numerical results show that the proposed hybrid model, GWO-FANGBM(1,1), achieves superior prediction accuracy with a MAPE of 5.82%, an RMSE of 3857.12, and an MAE of 3062.00, outperforming GM(1,1), GWO-GM(1,1), ARIMA, and LR. The originality of the study is supported by the fact that a hybrid approach named as GWO-FANGBM(1,1) has not been used in the literature to forecast natural gas consumption in Turkey with an accurate parameter optimization.
ORCID iDs
Özcan, Tuncay, Konyalioglu, Aziz Kemal ORCID: https://orcid.org/0000-0002-2443-5063 and Apaydın, Tuğçe Beldek;Persistent Identifier
https://doi.org/10.17868/strath.00090448-
-
Item type: Article ID code: 90448 Dates: DateEvent27 August 2024Published27 August 2024Published Online30 July 2024Accepted10 January 2024SubmittedSubjects: Social Sciences > Economic Theory Department: Strathclyde Business School > Hunter Centre for Entrepreneurship, Strategy and Innovation Depositing user: Pure Administrator Date deposited: 03 Sep 2024 08:53 Last modified: 25 Sep 2024 10:56 Related URLs: URI: https://strathprints.strath.ac.uk/id/eprint/90448