Predicting global horizontal irradiance using regression-based machine learning models in high-altitude environments
Global solar radiation is a fundamental component of the Earth's energy balance and plays a critical role in integrating renewable energy into electrical systems. However, its pronounced variability, particularly in high-altitude regions, constrains energy stability and complicates planning processe...
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Artículo |
| Lenguaje: | inglés |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | Enlace permanente (DOI o Handle) Ver en el sistema de origen |
| Etiquetas: |
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1881700520916156416 |
|---|---|
| author | Milton Edward Humpiri Flores David Mamani-Pari Danny Lévano Rodríguez Roel Dante Gómez Apaza Esteban Tocto-Cano Ángel Rosendo Condori-Coaquira Javier Linkolk López‐Gonzales |
| author_browse | Danny Lévano Rodríguez David Mamani-Pari Esteban Tocto-Cano Javier Linkolk López‐Gonzales Milton Edward Humpiri Flores Roel Dante Gómez Apaza Ángel Rosendo Condori-Coaquira |
| author_facet | Milton Edward Humpiri Flores David Mamani-Pari Danny Lévano Rodríguez Roel Dante Gómez Apaza Esteban Tocto-Cano Ángel Rosendo Condori-Coaquira Javier Linkolk López‐Gonzales |
| author_sort | Milton Edward Humpiri Flores |
| collection | Producción científica |
| description | Global solar radiation is a fundamental component of the Earth's energy balance and plays a critical role in integrating renewable energy into electrical systems. However, its pronounced variability, particularly in high-altitude regions, constrains energy stability and complicates planning processes. This study evaluated regression-based and machine-learning models for 1-day-ahead prediction of global horizontal irradiance (GHI) in Juliaca, Peru (>3,800 m.a.s.l.), using 11 years (2010–2020) of daily NASA POWER data ( n = 4,000 observations after cleaning; 80/20 chronological train/test split: training from 2010 to 2018 and testing from 2018 to 2020). During data auditing, we identified that the same-day clearness index ( K T ) is algebraically related to the same-day GHI through the extraterrestrial radiation term, which would otherwise inflate the apparent predictive accuracy. Therefore, we restricted same-day predictors to cyclical solar geometry terms and used lagged specific humidity, clearness index, and irradiance values from the previous 1–3 days as legitimate autoregressive information. Under this leakage-free design, Linear Regression achieved the best performance ( R 2 = 0.390, MAE = 0.677 kWh m −2 d −1 , and RMSE = 0.896 kWh m −2 d −1 ), narrowly outperforming gradient boosting ( R 2 = 0.384) and polynomial regression ( R 2 = 0.381). Random Forest and shallow neural networks followed closely, with all six models exhibiting remarkably comparable performance. Feature-importance and SHAP analyses for the Random Forest model showed that previous-day irradiance and previous-day (lag-1) humidity were the dominant predictors. This pattern may be consistent with short-persistence cloud processes; however, cloud observations are required to verify this interpretation directly. The modest but realistic explanatory power obtained here, together with the diagnostic procedure used to achieve it, offers a template for detecting and avoiding under-reported leakage risk in solar radiation forecasting studies that use clearness-index-derived predictors. |
| format | Article |
| id | openalex:W7212791171 |
| institution | UPeU |
| language | English |
| publishDate | 2026 |
| publishDateRange | 2026 |
| publishDateSort | 2026 |
| record_format | dspace |
| spelling | openalex:W7212791171Predicting global horizontal irradiance using regression-based machine learning models in high-altitude environmentsMilton Edward Humpiri FloresDavid Mamani-PariDanny Lévano RodríguezRoel Dante Gómez ApazaEsteban Tocto-CanoÁngel Rosendo Condori-CoaquiraJavier Linkolk López‐Gonzalesglobal radiationsolar irradiance forecastinghigh-altitude environmentclearness indexdata leakagespecific humidityNASA POWERrandom forestgradient boostinglinear regressionpolynomial regressionshallow neural networksfeature importanceSHAPcloud processesrenewable energy integrationSolar Radiation and PhotovoltaicsSolar Thermal and Photovoltaic SystemsMeteorological Phenomena and SimulationsGlobal solar radiation is a fundamental component of the Earth's energy balance and plays a critical role in integrating renewable energy into electrical systems. However, its pronounced variability, particularly in high-altitude regions, constrains energy stability and complicates planning processes. This study evaluated regression-based and machine-learning models for 1-day-ahead prediction of global horizontal irradiance (GHI) in Juliaca, Peru (>3,800 m.a.s.l.), using 11 years (2010–2020) of daily NASA POWER data ( n = 4,000 observations after cleaning; 80/20 chronological train/test split: training from 2010 to 2018 and testing from 2018 to 2020). During data auditing, we identified that the same-day clearness index ( K T ) is algebraically related to the same-day GHI through the extraterrestrial radiation term, which would otherwise inflate the apparent predictive accuracy. Therefore, we restricted same-day predictors to cyclical solar geometry terms and used lagged specific humidity, clearness index, and irradiance values from the previous 1–3 days as legitimate autoregressive information. Under this leakage-free design, Linear Regression achieved the best performance ( R 2 = 0.390, MAE = 0.677 kWh m −2 d −1 , and RMSE = 0.896 kWh m −2 d −1 ), narrowly outperforming gradient boosting ( R 2 = 0.384) and polynomial regression ( R 2 = 0.381). Random Forest and shallow neural networks followed closely, with all six models exhibiting remarkably comparable performance. Feature-importance and SHAP analyses for the Random Forest model showed that previous-day irradiance and previous-day (lag-1) humidity were the dominant predictors. This pattern may be consistent with short-persistence cloud processes; however, cloud observations are required to verify this interpretation directly. The modest but realistic explanatory power obtained here, together with the diagnostic procedure used to achieve it, offers a template for detecting and avoiding under-reported leakage risk in solar radiation forecasting studies that use clearness-index-derived predictors.Articlehttps://doi.org/10.3389/frai.2026.1870055https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2026.1870055/pdfW7212791171enFrontiers in Artificial Intelligence2624-82122026-09-14 |
| spellingShingle | global radiation solar irradiance forecasting high-altitude environment clearness index data leakage specific humidity NASA POWER random forest gradient boosting linear regression polynomial regression shallow neural networks feature importance SHAP cloud processes renewable energy integration Solar Radiation and Photovoltaics Solar Thermal and Photovoltaic Systems Meteorological Phenomena and Simulations Milton Edward Humpiri Flores David Mamani-Pari Danny Lévano Rodríguez Roel Dante Gómez Apaza Esteban Tocto-Cano Ángel Rosendo Condori-Coaquira Javier Linkolk López‐Gonzales Predicting global horizontal irradiance using regression-based machine learning models in high-altitude environments |
| title | Predicting global horizontal irradiance using regression-based machine learning models in high-altitude environments |
| title_full | Predicting global horizontal irradiance using regression-based machine learning models in high-altitude environments |
| title_fullStr | Predicting global horizontal irradiance using regression-based machine learning models in high-altitude environments |
| title_full_unstemmed | Predicting global horizontal irradiance using regression-based machine learning models in high-altitude environments |
| title_short | Predicting global horizontal irradiance using regression-based machine learning models in high-altitude environments |
| title_sort | predicting global horizontal irradiance using regression based machine learning models in high altitude environments |
| topic | global radiation solar irradiance forecasting high-altitude environment clearness index data leakage specific humidity NASA POWER random forest gradient boosting linear regression polynomial regression shallow neural networks feature importance SHAP cloud processes renewable energy integration Solar Radiation and Photovoltaics Solar Thermal and Photovoltaic Systems Meteorological Phenomena and Simulations |
| topic_facet | global radiation solar irradiance forecasting high-altitude environment clearness index data leakage specific humidity NASA POWER random forest gradient boosting linear regression polynomial regression shallow neural networks feature importance SHAP cloud processes renewable energy integration Solar Radiation and Photovoltaics Solar Thermal and Photovoltaic Systems Meteorological Phenomena and Simulations |
| url | https://doi.org/10.3389/frai.2026.1870055 https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2026.1870055/pdf |
| work_keys_str_mv | AT miltonedwardhumpiriflores predictingglobalhorizontalirradianceusingregressionbasedmachinelearningmodelsinhighaltitudeenvironments AT davidmamanipari predictingglobalhorizontalirradianceusingregressionbasedmachinelearningmodelsinhighaltitudeenvironments AT dannylevanorodriguez predictingglobalhorizontalirradianceusingregressionbasedmachinelearningmodelsinhighaltitudeenvironments AT roeldantegomezapaza predictingglobalhorizontalirradianceusingregressionbasedmachinelearningmodelsinhighaltitudeenvironments AT estebantoctocano predictingglobalhorizontalirradianceusingregressionbasedmachinelearningmodelsinhighaltitudeenvironments AT angelrosendocondoricoaquira predictingglobalhorizontalirradianceusingregressionbasedmachinelearningmodelsinhighaltitudeenvironments AT javierlinkolklopezgonzales predictingglobalhorizontalirradianceusingregressionbasedmachinelearningmodelsinhighaltitudeenvironments |