Electricity consumption forecasting using a novel homogeneous and heterogeneous ensemble learning

IF 2.6 4区 工程技术 Q3 ENERGY & FUELS Frontiers in Energy Research Pub Date : 2024-09-04 DOI:10.3389/fenrg.2024.1442502
Hasnain Iftikhar, Justyna Zywiołek, Javier Linkolk López-Gonzales, Olayan Albalawi
{"title":"Electricity consumption forecasting using a novel homogeneous and heterogeneous ensemble learning","authors":"Hasnain Iftikhar, Justyna Zywiołek, Javier Linkolk López-Gonzales, Olayan Albalawi","doi":"10.3389/fenrg.2024.1442502","DOIUrl":null,"url":null,"abstract":"In today’s world, a country’s economy is one of the most crucial foundations. However, industries’ financial operations depend on their ability to meet their electricity demands. Thus, forecasting electricity consumption is vital for properly planning and managing energy resources. In this context, a new approach based on ensemble learning has been developed to predict monthly electricity consumption. The method divides electricity consumption time series into deterministic and stochastic components. The deterministic component, which consists of a secular long-term trend and an annual seasonality, is estimated using a multiple regression model. In contrast, the stochastic part considers the short-run random fluctuations of the consumption time series. It is forecasted by four different time series, four machine learning models, and three novel proposed ensemble models: the time series homogeneous ensemble model, the machine learning ensemble model, and the heterogeneous ensemble model. The study analyzed data on Pakistan’s monthly electricity consumption from 1991-January to 2022-December. The evaluation of the forecasting models is based on three criteria: accuracy metrics (including the mean absolute percent error (MAPE), the mean absolute error (MAE), the root mean squared error (RMSE), and the root relative squared error (RRSE)); an equality forecast statistical test (the Diebold and Mariano’s test); and a graphical assessment. The heterogeneous ensemble model’s forecasting results show lower error values compared to the homogeneous ensemble models and the singles models, with accuracy metrics measured by MAPE, MAE, RMSE, and RRSE at 5.0027, 460.4800, 614.5276, and 0.2933, respectively. Additionally, the heterogeneous ensemble model is statistically significant (p <jats:inline-formula><mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"><mml:mo>&lt;</mml:mo></mml:math></jats:inline-formula> 0.05) and superior to the rest of the models. Also, the heterogeneous ensemble model demonstrates considerable performance with the least mean error, which is comparatively better than the individual and best models reported in the literature and are considered baseline models. Further, the forecast values’ monthly behavior depicts that electricity consumption is higher during the summer season, and this demand will be highest in June and July. The forecast model and graph reveal that electricity consumption rapidly increases with time. This indirectly indicates that the government of Pakistan must take adequate steps to improve electricity production through different energy sources to restore the country’s economic status by meeting the country’s electricity demand. Despite several studies conducted from various perspectives, no analysis has been undertaken using an ensemble learning approach to forecast monthly electricity consumption for Pakistan.","PeriodicalId":12428,"journal":{"name":"Frontiers in Energy Research","volume":"6 3 1","pages":""},"PeriodicalIF":2.6000,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Frontiers in Energy Research","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.3389/fenrg.2024.1442502","RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENERGY & FUELS","Score":null,"Total":0}
引用次数: 0

Abstract

In today’s world, a country’s economy is one of the most crucial foundations. However, industries’ financial operations depend on their ability to meet their electricity demands. Thus, forecasting electricity consumption is vital for properly planning and managing energy resources. In this context, a new approach based on ensemble learning has been developed to predict monthly electricity consumption. The method divides electricity consumption time series into deterministic and stochastic components. The deterministic component, which consists of a secular long-term trend and an annual seasonality, is estimated using a multiple regression model. In contrast, the stochastic part considers the short-run random fluctuations of the consumption time series. It is forecasted by four different time series, four machine learning models, and three novel proposed ensemble models: the time series homogeneous ensemble model, the machine learning ensemble model, and the heterogeneous ensemble model. The study analyzed data on Pakistan’s monthly electricity consumption from 1991-January to 2022-December. The evaluation of the forecasting models is based on three criteria: accuracy metrics (including the mean absolute percent error (MAPE), the mean absolute error (MAE), the root mean squared error (RMSE), and the root relative squared error (RRSE)); an equality forecast statistical test (the Diebold and Mariano’s test); and a graphical assessment. The heterogeneous ensemble model’s forecasting results show lower error values compared to the homogeneous ensemble models and the singles models, with accuracy metrics measured by MAPE, MAE, RMSE, and RRSE at 5.0027, 460.4800, 614.5276, and 0.2933, respectively. Additionally, the heterogeneous ensemble model is statistically significant (p < 0.05) and superior to the rest of the models. Also, the heterogeneous ensemble model demonstrates considerable performance with the least mean error, which is comparatively better than the individual and best models reported in the literature and are considered baseline models. Further, the forecast values’ monthly behavior depicts that electricity consumption is higher during the summer season, and this demand will be highest in June and July. The forecast model and graph reveal that electricity consumption rapidly increases with time. This indirectly indicates that the government of Pakistan must take adequate steps to improve electricity production through different energy sources to restore the country’s economic status by meeting the country’s electricity demand. Despite several studies conducted from various perspectives, no analysis has been undertaken using an ensemble learning approach to forecast monthly electricity consumption for Pakistan.
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
利用新型同质和异质集合学习预测用电量
在当今世界,国家经济是最重要的基础之一。然而,各行各业的财务运营取决于其满足电力需求的能力。因此,预测用电量对于正确规划和管理能源资源至关重要。在此背景下,我们开发了一种基于集合学习的新方法来预测月度用电量。该方法将用电量时间序列分为确定性和随机性两部分。确定性部分由长期趋势和年度季节性组成,采用多元回归模型进行估算。相比之下,随机部分考虑了用电量时间序列的短期随机波动。它由四个不同的时间序列、四个机器学习模型和三个新提出的集合模型进行预测:时间序列同质集合模型、机器学习集合模型和异质集合模型。研究分析了巴基斯坦 1991 年 1 月至 2022 年 12 月的月度用电量数据。预测模型的评估基于三个标准:准确度指标(包括平均绝对百分比误差(MAPE)、平均绝对误差(MAE)、均方根误差(RMSE)和相对平方根误差(RRSE));相等预测统计检验(Diebold 和 Mariano 检验);以及图形评估。与同质集合模型和单一模型相比,异质集合模型的预测结果显示出较低的误差值,用 MAPE、MAE、RMSE 和 RRSE 衡量的准确度指标分别为 5.0027、460.4800、614.5276 和 0.2933。此外,异构集合模型的统计意义显著(p < 0.05),优于其他模型。同时,异质集合模型也表现出相当高的性能,平均误差最小,相对优于文献中报道的单个模型和最佳模型,被认为是基准模型。此外,预测值的月度行为表明,夏季用电量较高,6 月和 7 月的需求量最大。预测模型和图表显示,用电量随着时间的推移迅速增加。这间接表明,巴基斯坦政府必须采取适当措施,通过不同能源提高电力生产,以满足国家的电力需求,从而恢复国家的经济地位。尽管从不同角度进行了多项研究,但还没有人使用集合学习方法对巴基斯坦的月度用电量进行预测分析。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
Frontiers in Energy Research
Frontiers in Energy Research Economics, Econometrics and Finance-Economics and Econometrics
CiteScore
3.90
自引率
11.80%
发文量
1727
审稿时长
12 weeks
期刊介绍: Frontiers in Energy Research makes use of the unique Frontiers platform for open-access publishing and research networking for scientists, which provides an equal opportunity to seek, share and create knowledge. The mission of Frontiers is to place publishing back in the hands of working scientists and to promote an interactive, fair, and efficient review process. Articles are peer-reviewed according to the Frontiers review guidelines, which evaluate manuscripts on objective editorial criteria
期刊最新文献
Grid-integrated solutions for sustainable EV charging: a comparative study of renewable energy and battery storage systems Research on the impact of digitalization on energy companies’ green transition: new insights from China Multi-objective-based economic and emission dispatch with integration of wind energy sources using different optimization algorithms Demand-side management scenario analysis for the energy-efficient future of Pakistan: Bridging the gap between market interests and national priorities Modeling and scheduling of utility-scale energy storage toward high-share renewable coordination
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1