Wind speed and power forecasting using Bayesian optimized machine learning models in Gabal Al-Zayt, Egypt
Abstract Accurate wind speed and power forecasts are essential for applications involving renewable wind energy. Ten machine learning techniques, including single and ensemble models, are compared, and evaluated in this study over a range of time scales. The outcomes of the wind speed prediction (WS...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-08-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-13140-x |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849333090492612608 |
|---|---|
| author | Nehal Elshaboury Haytham Elmousalami |
| author_facet | Nehal Elshaboury Haytham Elmousalami |
| author_sort | Nehal Elshaboury |
| collection | DOAJ |
| description | Abstract Accurate wind speed and power forecasts are essential for applications involving renewable wind energy. Ten machine learning techniques, including single and ensemble models, are compared, and evaluated in this study over a range of time scales. The outcomes of the wind speed prediction (WSP) model are used as inputs for the wind power prediction (WPP) model in a wind speed and power integration prediction system. The accuracy of various machine learning models is compared using several evaluation metrics, such as Pearson’s correlation coefficient (R), explained variance (EV), mean absolute percentage error (MAPE), mean square error (MSE), and concordance correlation coefficient (CCC). For WSP, the light gradient boosting machine (LGBM), extreme gradient boosting, and bagged decision tree (BDT) algorithms accurately predict wind speed across different time scales, with MAPE, MSE, EV, R, and CCC values ranging from 2.641 to 12.274%, 0.044 to 0.953, 0.888 to 0.994, 0.943 to 0.997, and 0.939 to 0.997, respectively. For WPP, the LGBM and BDT algorithms demonstrate strong predictive performance across different time scales, with MAPE, MSE, EV, R, and CCC values ranging from 0.277 to 186.710%, 0.927 to 9444.576, 0.970 to 1.000, 0.985 to 1.000, and 0.985 to 1.000, respectively. |
| format | Article |
| id | doaj-art-dec4665e40ea4485a4dcb6ddcfc8fbd4 |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-08-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-dec4665e40ea4485a4dcb6ddcfc8fbd42025-08-20T03:45:59ZengNature PortfolioScientific Reports2045-23222025-08-0115112510.1038/s41598-025-13140-xWind speed and power forecasting using Bayesian optimized machine learning models in Gabal Al-Zayt, EgyptNehal Elshaboury0Haytham Elmousalami1 Construction and Project Management Research Institute, Housing and Building National Research CenterDepartment of Infrastructure Engineering, Faculty of Engineering and Information Technology, The University of MelbourneAbstract Accurate wind speed and power forecasts are essential for applications involving renewable wind energy. Ten machine learning techniques, including single and ensemble models, are compared, and evaluated in this study over a range of time scales. The outcomes of the wind speed prediction (WSP) model are used as inputs for the wind power prediction (WPP) model in a wind speed and power integration prediction system. The accuracy of various machine learning models is compared using several evaluation metrics, such as Pearson’s correlation coefficient (R), explained variance (EV), mean absolute percentage error (MAPE), mean square error (MSE), and concordance correlation coefficient (CCC). For WSP, the light gradient boosting machine (LGBM), extreme gradient boosting, and bagged decision tree (BDT) algorithms accurately predict wind speed across different time scales, with MAPE, MSE, EV, R, and CCC values ranging from 2.641 to 12.274%, 0.044 to 0.953, 0.888 to 0.994, 0.943 to 0.997, and 0.939 to 0.997, respectively. For WPP, the LGBM and BDT algorithms demonstrate strong predictive performance across different time scales, with MAPE, MSE, EV, R, and CCC values ranging from 0.277 to 186.710%, 0.927 to 9444.576, 0.970 to 1.000, 0.985 to 1.000, and 0.985 to 1.000, respectively.https://doi.org/10.1038/s41598-025-13140-xEnsemble machine learningBayesian optimizationWind speed predictionWind power predictionDeveloping countries |
| spellingShingle | Nehal Elshaboury Haytham Elmousalami Wind speed and power forecasting using Bayesian optimized machine learning models in Gabal Al-Zayt, Egypt Scientific Reports Ensemble machine learning Bayesian optimization Wind speed prediction Wind power prediction Developing countries |
| title | Wind speed and power forecasting using Bayesian optimized machine learning models in Gabal Al-Zayt, Egypt |
| title_full | Wind speed and power forecasting using Bayesian optimized machine learning models in Gabal Al-Zayt, Egypt |
| title_fullStr | Wind speed and power forecasting using Bayesian optimized machine learning models in Gabal Al-Zayt, Egypt |
| title_full_unstemmed | Wind speed and power forecasting using Bayesian optimized machine learning models in Gabal Al-Zayt, Egypt |
| title_short | Wind speed and power forecasting using Bayesian optimized machine learning models in Gabal Al-Zayt, Egypt |
| title_sort | wind speed and power forecasting using bayesian optimized machine learning models in gabal al zayt egypt |
| topic | Ensemble machine learning Bayesian optimization Wind speed prediction Wind power prediction Developing countries |
| url | https://doi.org/10.1038/s41598-025-13140-x |
| work_keys_str_mv | AT nehalelshaboury windspeedandpowerforecastingusingbayesianoptimizedmachinelearningmodelsingabalalzaytegypt AT haythamelmousalami windspeedandpowerforecastingusingbayesianoptimizedmachinelearningmodelsingabalalzaytegypt |